AbstractsJ Clin Invest. 1972;51(6):1a-56a. https://doi.org/10.1172/JCI106960.

Find the latest version:

https://jci.me/106960/pdf

ABSTRACTSExplanation of symbols: No symbol = Member; * = Nonmember; ** = Emeritus or senior member

1. Characterization of the Hepatitis-Associated AntigenSubtypes by Radioimmunoassay. R. D. AACH,* E.HACKER,* AND C. PARKER,* St. Louis, Mo. (introduced byD. Alpers).

The recent recognition of different antigenic subtypes ofhepatitis-associated antigen (HAA) raises questions as tothe immunologic contribution of each determinant (a, d, andy) and the efficacy of a given antiserum in detection of HAA.Antibodies were produced in animals by immunization withHAA of different antigenic subtypes. HAA of dissimilarsubtypes were radiolabeled with 'I and radioimmunoassayswere performed as previously described (1971. Proc. Nat.Acad. Sci. U.S.A. 68: 1056). It was found that these anti-sera varied widely in their ability to recognize a, y, and ddeterminants. Circulating HAA in patients ill with hepatitiscould be separated into different subgroups and the relativecontribution of a, d, or y determinants characterized. 20patients were screened, 16 patients with HAA-positive acuteviral hepatitis and 4 with chronic persistent hepatitis. HAAof 9 of the 16 patients with acute disease were subgroup Y(a'y') and 7 subgroup D (a+d+). Three of the four patientswith chronic persistent hepatitis had HAA of subgroup D,and one had HAAof subgroup Y. These studies emphasize:(a) the ability to readily distinguish antigenic subtypes byradioimmunoassay. This provides a sensitive system which canbe used as an epidemiologic tool, and which can be used torelate subtypes with patterns of hepatic disease. (b) Anti-sera raised in animals show wide variability in the recogni-tion of individual determinants. This difference in specificitymust be taken into account in the selection of antiserum ofantiserum for routine screening of blood donors.

2. Chromosomal Aberrations in Newborns Exposed toHeroin In Utero. CYRIL A. L. ABRAMS* AND PEI-YULIAO,* New York (introduced by Gerald B. Phillips"*).

Equivocal evidence of chromosomal abnormalities has beenfound upon the administration of various psychotropic drugs.Opiates have received scant attention in this respect, anddefinitive information concerning their cytogenetic effects islacking. We examined peripheral blood lymphocytes of 16newborns aged i-31 days whose mothers had used heroinduring pregnancy, and 14 newborn controls aged 1-13 dayswhose mothers were not drug users. Both groups of new-borns received vitamin K before blood samples were taken.In the heroin group, five newborns had received medicationfor withdrawal symptoms, and one had had a viral illnessbefore sampling. 10 mothers in the heroin group had receivedmethadone during pregnancy. Cultures were set up in parallelfor heroin subjects and controls. Whole blood inoculum(heelstick) was incubated for 72 hr, and colcemide added2 hr before harvesting. 100 metaphases per subject wereanalyzed. The heroin group showed 81 chromatid breaks,29 dicentrics, 28 fragments, 28 gaps, 9 isochromatid breaks,5 deletions, and 3 bizarre forms in the 1600 mitoses analyzed.

The control group showed 20 chromatid breaks and 4 gapsin the 1400 mitoses analyzed. The mean values for damagedchromosomes and for damaged cells in the heroin group werefound to be 6 times higher than the corresponding values inthe control group (P <0.0001). Although other drugs andvarious environmental factors must be considered as possiblecauses of the chromosomal aberrations observed, it is per-missible to infer that heroin itself played the major role.(Research supported by NIH grant AM-05531.)

3. Myocardial Damage after Aorto-Coronary Vein By-pass Surgery. STEPHEN ACHUFF,* LAWRENCEGRIFFITH,*J. O'NEAL HUMPHRIES,* C. RICHARDCONTI,* ROBERTBRAW-LEY,* VINCENT Gorr,* AND RICHARD Ross, Baltimore, Md.

60 of the first 153 patients receiving aorto-coron~ry veinbypass surgery have been reevaluated with coronary arteri-ography, single-plane left ventriculography, and stress test-ing (exercise and atrial pacing) an average of 5.3 monthsafter surgery. In this group, electrocardiography (ECG) ab-normalities compatible with recent myocardial damage wereseen in 35 patients (58%) within 10 days postoperatively.The 60 patients herein reported are comparable with thesurgical population as a whole with respect to mean dura-tion of angina (40 months), incidence of previous myocardialinfarction (63%o ), New York Heart Association functionalclass (95% class III or IV because of angina), and 86%had significant stenosis of two or all three major coronaryarteries. Good to excellent symptomatic relief was present in82% and improved stress tolerance in 85%. Bypass patencyfor the entire group was 64%. Hemodynamic parameters ofleft ventricular end diastolic pressure, cardiac index, andejection fraction showed no significant differences betweenpre- and postoperative studies. The ECG abnormalities inthe above 35 patients were supported by the angiographicfindings as follows: coronary arteriography revealed totalocclusion of a previously patent major coronary artery in60%, significant worsening of previously recognized lesionsin another 11%, and with ventriculography segmental con-tractility was judged to be reduced in another 23%o. Sympto-matic improvement in this group with myocardial damagewas present in 86%, stress tolerance improved in 87%, andbypass patency in 53%. These data may be interpreted assuggesting that clinical improvement was related to bypasspatency, infarction of ischemic myocardium, or a combina-tion of the two. (Research supported by grants HE 05584,FR 35 from NIH.)

4. Changes in Liver and Skeletal Muscle Transport ofAmino Acids Induced by Diets. SIAMAK ADIBI,* SWAMIK-KAN NALLATHAMBI,* AND THOMASMODESTO,* Pittsburgh,Pa. (introduced by I. Arthur Mirsky**).

Present studies were performed to characterize the changesin amino acid transport processes of above tissues in rela-tion to substrate specificity and dietary composition. The in

I a

vivo transport of cycloleucine-"C (CYC) and a-aminoiso-butyric acid-14C (AIB) was investigated by determining thein vivo distribution ratio (DR = intracellular concentra-tion/extracellular concentration) of these substrates after i.v.injection in rats. Starvation even for 1 day significantlyincreased the liver DRof CYCby 32%o (P < 0.01). Starva-tion produced an even more dramatic increase in the liverDR of AIB (284%). While the muscle DR of CYC wasnot affected during the starvation, the DR of AIB wasmarkedly reduced (67%, P<0.01). To distinguish whetherthe caloric or amino acid deprivation was responsible forthe above changes, similar studies were repeated in ratswhich were force-fed a protein-free diet. Protein deprivationresulted in significant increases in the liver DR of bothamino acids, but the muscle DR of neither amino acid wasaffected. To determine whether the alteration in hormonalmilieu accounted for the transport changes, the followingstudies were performed. (a) Liver slices of starved rats,incubated in an ionic media containing only CYC, showedmarked increases in the uptake of this amino acid. (b) Star-vation or protein deprivation for 1 day did not changeplasma glucagon levels, but insulin levels were promptlyreduced after starvation. It thus appears that caloric restric-tion is the critical dietary factor for selective amino acidtransport changes in the muscle, while liver transportchanges for both substrates result from either caloric orprotein deprivation.

5. Interrelationship between Cell pH and Metabolism inPotassium Depletion. SHELDON ADLER,* Pittsburgh, Pa.(introduced by Philip Troen**).

In potassium depletion we demonstrated that cell pH(pHt) in rat diaphragm muscle is unchanged at extracellu-lar pH (pHe) 7.40, but it is decreased in both extracellularacidosis and alkalosis. Therefore, changes in metabolism inthis state could be due either to changes in cell [K] or pH,.Since citrate decarboxylation in diaphragm is pH dependent,this reaction was employed to determine the relative effectof these variables on metabolism. Intact rat diaphragmswere incubated simultaneously for 4 hr in a Krebs-Ringerbicarbonate solution containing 5 or 0 mEq potassium.Hemidiaphragms were then removed and incubated at thesame pH, in a 1 mEq potassium bath. 14CO2 productionfrom 1,5-citrate-"C was measured. At pH. 7.42 differencesbetween normal and potassium-depleted tissue were not sig-nificant. At pH. 6.90, however, citrate decarboxylation was432 dpm/mg in depleted tissue compared with 372 in nor-mals (P < 0.001). pH, was 6.73 and 6.89, respectively.Similarly, at pHe 7.65 decarboxylation was 188 and 165(P < 0.025) in potassium depletion and normals, respectively,whereas pH, was 7.16 and 7.23. When pH. was 6.94 inpotassium depletion and 6.65 in normals thus making pH,almost identical in both groups, there were no significantdifferences in decarboxylation. These results indicate that itis pH,, not pH. or tissue potassium, which is important inaltering muscle metabolism in potassium depletion. Our pre-vious studies indicate that when pH. is increased in thisstate, pH, is closer to that found at a physiologic pH. of7.40. It is proposed, therefore, that the lack of clinicalsymptoms often found in potassium depletion may be due

to normal pHi despite both an extracellular alkalosis anda reduction in muscle potassium.

6. Intracellular Composition and Transepithelial Trans-port: Inulin as a Potential Source of Experimental Er-ror. Q. ALAWQATI,* A. LEAF,** A. D. C. MACKNIGHT,*AND M. M. CBIAN,* Boston, Mass.

It has become increasingly apparent that the volume andelectrolyte content of epithelial cells is an important deter-minant of transport activity. Inulin has been widely used inthe measurement of the composition of epithelial cells underthe assumption that its volume of distribution accuratelymeasures the extracellular space. Recent studies from ourlaboratory have raised questions concerning the validity ofthis assumption. Inulin and sucrose spaces were simulta-neously measured in cells scraped from the paired halves ofurinary bladders from the toad; one hemibladder was incu-bated for 1 hr and the other for 6, 12, or 20 hr. After 1 hrthe nonsucrose space was 3.2+0.1 kg/kg dry weight (dw)and did not change significantly after 6, 12, or 20 hr. Thenoninulin space after 1 hr was 4.0+0.1 kg/kg dw, and de-creased significantly after 6 hr in one or three sets of ex-periments. After 12 hr, the noninulin space fell by 1.8+0.3(P < 0.005) and after 20 hr, decreased by 0.6+0.2 kg/kgdw (P <0.05). Using sucrose as an extracellular marker,vasopressin increased cell water by 0.38+0.14 kg/kg dw (P< 0.05), sodium by 54±+12 mEq/kg dw (P < 0.01), andchloride by 38± 10 (P < 0.02). Cell potassium was un-changed. Since the sucrose space is consistently larger thanthe inulin space and is unchanged with prolonged incuba-tion, sucrose is the more suitable marker. The results indi-cate that vasopressin increases the permeability for sodiumentry into the epithelial cells, consistent with previous datafrom this laboratory using inulin as an extracellular marker.(Supported by the American Heart Association grant 71-847,NIH grant HE-06664, and John A. Hartford Foundation,Inc.)

7. The Response of the Blood Coagulation System afterAcute Thrombic Stroke. NORMAALKJAERSIG,* BENEDICTELAURSEN,* AND ANTHONYFLETCHER,** St. Louis, Mo.

20 patients with acute cerebral thrombosis were followedserially for 7 days after the ictus. Plasma fibrinogen chro-matography (Fletcher, 1970. Trans. Ass. Amer. Physicians.83: 159), a method for quantifying fibrinogen complexes,uncomplexed fibrinogen, and other fibrinogen derivatives,detected the presence of blood hypercoagulable/thromboticstates. Such states were diagnosed when fibrinogen complexconcentration exceeded 20% of total fibrinogen. Serial assaysfor plasminogen, fibrinogen, ai-antitrypsin, a2-macroglobulin,and antithrombin III were obtained on all samples. Meanplasma fibrinogen rose steadily from 300 mg/100 ml to 350mg/100 ml over the observation period and mean a,-macro-globulin approximately 35%, but changes in other com-ponent mean values were insignificant. Mean fibrinogen firstderivative concentration was elevated throughout the obser-vation period, a finding indicative of enhanced plasmafibrinolysis. 49%o of the total plasma samples examinedshowed blood hypercoagulable/thrombotic patterns. Com-parison of plasma samples containing > 20%o fibrinogen com-

2 a

plexes with those containing < 20% fibrinogen complexesshowed antithrombin III, plasminogen, ai-antitrypsin, anda,-macroglobulin levels to be significantly depressed (P <0.01 for days 1-3) in plasma exhibiting hypercoagulable/thrombotic chromatographic findings. Also, the concentrationof fibrinogen first derivative was significantly decreased(P < 0.05) in these latter samples. Depression of anti-thrombin III concentration in samples showing blood hyper-coagulable/thrombotic chromatographic patterns indicatedthat thrombin had recently been formed in these plasmas.The concomitant decrease in plasminogen and of the plasmininhibitor ar-macroglobulin, together with the finding of in-creased fibrinogen first derivative, is confirmatory of thepresence of a secondary plasma fibrinolytic response. Thefindings demonstrate that a blood hypercoagulable statefollows the acute ictus in a high proportion of patients withcerebral thrombosis. Since concomitant activation of theplasminogen system also occurs, it may be inferred thatintravascular coagulation complicates the blood hyperco-agulable state. (Supported by HE-03745 and NB 06833.)

8. The Isolation of Human Vitamin B,-Binding Pro-teins Utilizing Affinity Chromatography. ROBERTH. AL-LEN* AND PHILIP W. MAJERUS, St. Louis, Mo.

We have developed a method of affinity chromatographywhich is a potent tool for isolation of the trace B,2-bindingproteins. The affinity ligand was prepared by partial acidhydrolysis of the amide groups of the propionamide side-chains of the corrin ring of cyanocobalamin (Ba2). The re-sultant mixtures of mono-, di-, and tricarboxylic Bu deriva-tives were separated by chromatography on QAE-Sephadex.Monocarboxyl-B, was coupled covalently to the free aminogroup of 3,3'-diaminodipropylamine-substituted Sepharose us-ing 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide thus re-generating native Bn1 stably coupled to Sepharose (yield, 0.7/Amoles B,2/ml Sepharose). Bu-Sepharose bound (> 95%)the Bua-binding proteins of human leukocytes, plasma, andgastric juice. After extensive washing the B12-binding pro-tein is eluted in high yield (60-95%o) with 7.5 M guanidine-HCl. We have purified the Ba-binding protein of chronicmyelogenous leukemia granulocytes 10,000-fold using thistechnique alone. The purified leukocyte binder saturated withB,2 gives a single band on polyacrylamide disc gel electro-phoresis which is red in unstained gels. The molecular weightof the leukocyte binder is 120,000 as determined both bySDS-polyacrylamide gel electrophoresis and Sephadex G-150chromatography. The protein binds 2.6 moles of Bn/120,000g protein (Lowry) indicating that the leukocyte binder hasmultiple Bvu-binding sites. Human serum transcobalamin II(TCII) has been purified 760,000-fold starting with Cohnfraction III from 1000 liters of plasma. This preparation ofTCII binds 7.9 ,moles B2/g protein, gives a single red bandon disc gel electrophoresis, but contains 50-75% contaminat-ing protein based on the disc gel. Further purification is inprogress. (Research supported by grants from NIH andACS.)

9. Increased Susceptibility to Infection in Type II Es-sential Hypercatabolism of C3. CHESTERA. ALPER, KURTJ. BLOCH, AND FRED S. ROSEN, Boston, Mass.

Wehave previously described a young man with a lifelonghistory of infections, hypercatabolism of C3, and multipledefects in complement-mediated functions and deficiencies inplasma proteins affecting C3. The present patient is a 34 yrold woman with a history of meningococcal, 8-hemolyticstreptococcal, and pneumococcal infections. Humoral anti-body production, cellular immunity, serum immunoglobulins,and white cell phagocytic function were normal. All comple-ment components were normal in concentration except forC3 which was 7 mg/100 ml or 5% of normal. Approximately50% of the immunochemically detected C3 was in the inac-tivated form, C3i. Metabolic studies with l"I-labeled C3revealed a 2-3 times elevated fractional catabolic rate of6% per hr. Complement-mediated functions such as theenhancement of phagocytosis by serum were diminished butwere normalized by the addition of C3 in vitro. In the pre-vious patient, GBG (properdin Factor B), C3 proactivatoractivity, GBGase inhibitor, and C3 inactivator were unde-tectable, but, in sharp contrast, these proteins were presentin normal concentrations in the present patient's serum. C3added to her serum and incubated at 370C was convertedand inactivated more rapidly than in normal serum. Theaddition of normal serum increased the rate of C3 break-down in vitro. The factor in the patient's serum causing C3breakdown was a 6S heat-labile 8-pseudoglobulin requiringMg"+ for its action. The cofactor in normal serum hadsimilar physicochemical characteristics but was slightly lessheat labile. The abnormality in this patient is thus quitedifferent from that described in the previous patient, but isalso associated with undue susceptibility to infection andincreased breakdown of C3 in vivo.

10. ABOBlood Group Activity Associated with HumanIntestinal Disaccharidases. DAVID H. ALPERS* AND JOHNJ. KELLY,* St. Louis, Mo. (introduced by Arthur Z. Eisen.)

Intestinal disaccharidases are glycoproteins resistant totrypsin treatment. We have purified human maltase andsucrase about 300-fold to homegeneity on gel electrophoresis.Carbohydrate content was measured by paper chromatogra-phy and specific enzymatic analysis, and by colorimetrictechniques. Carbohydrate measured accounted for almost 40%of the isolated protein. Sucrase from a blood type A patientcontained per milligram of protein 132 ,ug hexosamine, 20,ug glucose, 209 ,ug galactose, 36 ,ug mannose, 83 ,tg fucose,and 154 ,ug xylose, for a total of 634 ,ug carbohydrate. Mal-tase from a blood type B patient contained per milligram ofprotein 56 ,ug hexosamine, 45 ,ug glucose, 329 lAg galactose,14 lAg mannose, 42 ,ug fucose, and 125 ,ug xylose, for a totalof 611 ,ug carbohydrate. Sialic acid could not be detected.Amino acid analysis revealed 17%o of residues as threonineand serine. These data suggested the carbohydrate contentof blood group substance, also high in galactose, fucose, andhexosamine. Pure disaccharidase inhibited type-specific ABred cell agglutination, at least to a concentration of 1 ,ug/ml.Group D agglutination was not affected. Inhibition was notdue to a small contamination with pure blood group sub-stance for the following reasons: (a) disaccharidase proteinand activity did not enter a 4%o acrylamide gel after incuba-tion with specific antibody; (b) preincubation with specificantibody shifted the disaccharidase in sucrose gradients from11.2 to 14S; and (c) periodate treatment rendered the en-

3 a

zyme more trypsin-sensitive by 20% while decreasing agglu-tination inhibition by 50%. We conclude that intestinaldisaccharidases have a carbohydrate composition and struc-ture similar to blood group substance, and that this carbo-hydrate prevents intraluminal hydrolysis of the protein bypancreatic enzymes. (Supported from grant AM-14038 fromNIH.)

11. Inhibition of Hemoglobin Synthesis by Cyanate.BLANCHE P. ALTER, YUET WAI KAN, AND DAVID G.NATHAN, Boston, Mass.

Cyanate inhibits sickling and prolongs red cell survivalin sickle cell anemia. Since cyanate carbamylates many pro-teins, several abnormalities in red cell function might beanticipated. We have found that cyanate markedly inhibitshemoglobin synthesis in human and rabbit reticulocytes. In-corporation of radioactive amino acid into hemoglobin byhuman sickle reticulocytes, bone marrow erythroblasts, orrabbit reticulocytes was inhibited by as little as 5 mmcya-nate and abolished by 50 mm. Both a- and 8-chains wereequally affected. Thrice washing of cells exposed to 50 mmcyanate for 30 min restored 30% of the synthetic activity.Transport of radioactive amino acid (lys or leu) into thecell was evaluated by Sephadex G-25 chromatography ofhemolysates prepared from washed cells previously incubatedwith 50 mmcyanate and amino acid for 2 hr. Free intra-cellular radioactive amino acid was not decreased and wasnot carbamylated as shown by iontopheresis at pH 1.9. Theinfluence of cyanate on amino acylation of tRNA was evalu-ated with leucine-"C, rabbit liver tRNA, and ribosome-freerabbit reticulocyte lysate. Cyanate did not inhibit acylationand the acylated amino acid, released from tRNA by 0.1 NNaOH, was not carbamylated. After incubation of rabbitreticulocytes with methionine-'S and cyanate, polysomeswere examined by sucrose density gradients. Cyanate in-duced degradation of polysomes to monosomes, suggestinginhibition of initiation rather than blockade of terminationas the basis of the protein synthetic defect. We concludethat cyanate profoundly inhibits initiation of hemoglobinsynthesis. If this effect occurs in other tissues, the chemicalmay not be suitable for therapy, at least in growing children.

12. Influence of Increased Glucose Availability and Man-nitol on Performance of Hypoxic Myocardium. EzRA A.AMSTERDAM,*DUANEFOLEY,* RASHID A. MASsuMI,* ROB-ERT ZELIS,* AND DEAN T. MASON, Davis, Calif.

Support of the ischemic, failing heart by positive inotropicdrugs is limited by the unfavorable alteration in myocardialoxygen supply-demand these agents produce. Other pharma-cologic approaches which enhance cardiac performance with-out accompanying deleterious effects would be advantageous.Therefore, the effects of glucose, insulin, and mannitol oncontractile performance of isolated, supported right ven-tricular cat papillary muscles were studied during normoxia(95% 02, 5% GO2) and hypoxia (95% N2, 5% CO). Inseven muscles with 100 mg/100 ml glucose, control peakisometric tension fell to 18% (4.8 to 0.8 g/mm'). Duringsupport with 100 mg/100 ml glucose, hypoxia-induced re-duction of tension was attenuated (P<0.05) to 27% (4.4to 1.4). In six additional muscles, augmentation of glucose

from 100 mg/100 ml to 1000 mg/100 ml during hypoxia in-creased tension (0.7 to 1.1, P < 0.01). Addition of insulin(200 mU/ml) to six muscles during hypoxia-100 mg/100 mlglucose increased tension from 1.5 to 1.8 (P < 0.05). Duringhypoxia-1000 mg/100 ml glucose, insulin increased tensionfrom 1.4 to 2.2 (P <0.01). Mannitol, equiosmolar to 1000mg/100 ml glucose, augmented tension during normoxia-100mg/100 ml glucose (5.4 to 6.4, P < 0.05) and improved(P < 0.05) hypoxia-induced reduction of tension to 31% ofcontrol (6.4 to 2.0) compared to decrease of tension of 18%of control with hypoxia-100 mg/100 ml glucose alone. There-fore, augmented glucose availability and mannitol have pro-tective effects on function of the hypoxic myocardium andmay provide additional clinical approaches for support ofthe ischemic, failing heart. (Supported by NIH ResearchGrant MH-19489.)

13. Glycolytic Rates and Glycogen Content of DiabeticRat Jejunal Mucosa. JAMES W. ANDERSON*ANDALBERT L.JONES, San Francisco, Calif. (introduced by Marvin H.Sleisenger**).

Diabetes is associated with significant increases in intes-tinal transport of sugars, amino acids, and sodium. How-ever, the metabolic alterations in diabetic intestine are lesswell characterized. Previously we have demonstrated thatthe activities of certain key glycolytic and pentose phosphatepathway enzymes are significantly increased in jejunum ofdiabetic rats. In the present investigation, glucose utiliza-tion and lactate production as well as glycogen concentra-tions and localization in jejunal mucosa of nonfasted controland alloxan-diabetic rats and 72-hr fasted rats were studied.Glucose utilization was 4.0+0.2 gmoles/min per g (mean+SEM) and lactate production was 4.8±0.2 ,umoles/min perg in control jejunal homogenates. These values are several-fold higher than rates reported for jejunal supernates and20-fold higher than values reported for liver. When com-pared with control values, both glucose utilization and lactateproduction were significantly reduced in jejunum of fastedrats while both were increased in diabetic rats. Glycogencontent and intracellular location have not been well definedin rat jejunal mucosa. Control rats had 444+41 jg glyco-gen/g mucosa whereas fasted rats had 214+21 ,ug/g. Dia-betic rats, however, had 818+149 ,ug glycogen/g mucosa.Electron microscopy demonstrated an accumulation of par-ticles of the size and appearance of f8-glycogen within jejunalabsorptive cells of diabetics. These particles were particu-larly prominent within the region of the terminal web butwere also found scattered at random throughout the apicalcytoplasm. The diabetic state, therefore, produces significantalterations in glucose metabolism of intestinal mucosa whichinclude increased glucose utilization, lactate production, andglycogen accumulation. These biochemical alterations un-doubtedly are related to reported alterations in intestinaltransport in the diabetic animal.

14. Active Role of Blood Cells in Amino Acid Transportin Man. THOMAS T. AOKI,* MURRAY F. BRENNAN,*WALTERA. MULLER,* AND GEORGEF. CAHILL, JR., Boston,Mass.

4 a

Studies of amino acid metabolism in human subjects havebeen predicated on plasma levels, while the potential contri-bution of the cellular constituents of blood remained unas-sessed. To ascertain both the role of blood cells in aminoacid transport, in particular glutamate, and the ability ofinsulin to increase glutamate uptake by muscle, five normalhuman volunteers underwent eight forearm studies in eachof which insulin was infused for 90 min into the brachialartery (venous effluent insulin concentration at 90 min = 188+58 JAU/ml). Enzymatically-determined plasma glutamatearterio-deep venous (A-DV) differences were not signifi-cantly increased during the entire procedure, while, in sharpcontrast, whole blood glutamate A-DV differences were sig-nificantly increased (O min = 8±3, 60 min = 26+9, 90 min =25±7, 120 min = 24±9, 150 min = 18±6, 180 min = 13±7,240 min= 4±8 gmoles/liter of whole blood, mean +SEM) at60 and 90 min (P < 0.05, paired t test). Using the hemato-crit, the glutamate content of the blood cells was calculated.At rest, a net movement of glutamate from arterial plasmainto both blood cells and muscle was found. During theinfusion, however, glutamate moved out of both plasma andcell compartments into muscle. Insulin thus appears to becapable of increasing glutamate uptake by muscle. Of fargreater importance, this change could only be documentedwhen whole blood rather than plasma glutamate determina-tions were performed. Therefore, insulin affects both thedirection and magnitude of the dynamic interplay that exists,with respect to glutamate and perhaps other amino acids aswell, between plasma, blood cells, and muscle with the bloodcells playing a pivotal role. (Research supported by grantsfrom NIH.)

15. Brain Edema after Rapid Lowering of Plasma Glu-cose in Normal and Diabetic Rabbits. ALLEN I. ARIEFF*AND CHARLESR. KLEEMAN,** Los Angeles, Calif.

Cerebral edema frequently complicates diabetic coma; themechanism is unknown. Hyperglycemia (plasma glucose =1100 mg/100 ml) in normal and diabetic rabbits was inducedby glucose (14C-labeled) infusion; measurements were madeof Na, K, glucose, Cl, lactate, amino acids, sorbitol, inositol,H20, 'C activity, and osmoles (by cryoscopy) in brain,plasma, and CSF. Idiogenic osmoles (Id Osm) were de-fined as brain osmoles-sum of all measured solutes. In bothgroups, during 4 hr of hyperglycemia (brain and plasmaOsm= 342 mOsm/kg) there was no loss of brain H20. Indiabetics, brain Na + K increased by 10%, but in normalsthere was no increase in any of the aforementioned solutes;therefore, the brain generated Id Osm. Brain "C/plasma XCactivity suggested these Id Osm were not glucose metabo-lites. When plasma glucose was then rapidly lowered withinsulin, cerebral edema developed in both normals and dia-betics, with t in brain content (mEq/kg dry wt) of Na(234- >266), K (406-4485, H20 (376-4454 g/100 g drywt), Id Osm (32 mM/kg H20) and "C activity (+400%)and an osmotic gradient between brain (328 mOsm/kg) andplasma (290 mOsm/kg), but no change in brain lactate,sorbitol, amino acids, or inositol. When plasma glucose waslowered with peritoneal dialysis (PD), brain edema did notoccur. These data show: (a) during hyperglycemia the braingenerates Id Osm in both normals and diabetics but in the

latter, 40% of the T in brain Osm is Na + K; (b) loweringplasma glucose rapidly with insulin causes brain edema whilelowering plasma glucose with PD does not; (c) there ap-pears to be a specific action of insulin on brain, affecting Kand Na transport and causing Id Osm generation. (Sup-ported by USPHSGrant NS05905.)

16. Antibody Specificity, Immunochemical Heterogeneity,and the Interpretation of Measurements of Plasma Para-thyroid Hormone by Radioimmunoassay. CLAUDE AR-NAUD, RALPH GOLDSMITH,* JULIANNA BISCHOFF,* GLENSIZEMORE,* SUSANOLDHAM,* AND JUDITH LARSEN,* Roch-ester, Minn.

This investigation was done to explain the discrepanciesin the reported results of studies of endogenous plasmaPTH measured by radioimmunoassay. 10 anti-PTH antiserafrom different laboratories were systematically evaluated.We observed large differences in the values for immuno-reactive PTH (iPTH) in the same hyperparathyroid(HPT) plasma. Each antiserum was characterized bydetermining its individual affinity for two preparations ofPTH purified from parathyroid tumor culture medium whichare immunologically similar if not identical to the 9000(PTH-9) and the 6-8000 (PTH-7) M.W. species of PTHwhich are the major components of iPTH in HPT plasma.Two of the antisera had unique properties: in mixtures ofvarying ratios of PTH-9 and PTH-7, GP 1 M reacted al-most exclusively with PTH-7 and CH 14 M with PTH-9.Using these two antisera, we found that PTH-9 had a muchshorter half-life and its measurement with CH 14 M morereliably localized parathyroid pathology (during differentialvenous catheterization) than did measurement of PTH-7with GP 1 M. On the other hand, greater discriminationbetween normal and primary HPT was found with GP 1 M(PTH-7) than with CH 14 M (PTH-9). This was evenmore striking in end-stage renal failure, in which PTH-7was increased 3 to 20-fold in all patients, whereas PTH-9was increased in only 70% of such patients. We concludethat (a) the differences in the results of iPTH measure-ments in different laboratories are most likely explained bydifferences in antiserum specificities for the two major circu-lating species of PTH, (b) assays of PTH-7 more likelyreflect the steady state status of circulating PTH, and (c)assays of PTH-9 reflect the acute secretory status of theparathyroid gland. (NIH AM12302.)

17. Reduced Catechol-O-Methyltransferase (COMT) Ac-tivity in the Liver and Increased Pressor Response toNorepinephrine (NE). NUZHET0. ATUK* ANDTHOMASC.WESTFALL,* Charlottesville, Va. (introduced by Julian I.Kitay).

The liver is the largest source of COMT, a major enzymefor metabolism of circulating catecholamines. The finding oftransient hypertension and increased urinary norepinephrineduring episodes of jaundice in a woman with recurrent intra-hepatic cholestasis prompted studies to define the mechanismof altered catecholamine metabolism. During icteric (I) andnonicteric (NI) episodes, the following were measured: (a)NE metabolism by NE-'C infusion, (b) COMTand mono-

S a

amine oxidase (MAO) activity in liver biopsies, and (c)pressor response to graded i.v. infusion of NE. After injec-tion of NE-`4C urine was collected at intervals and assayedfor radioactive NE, O-methylated and O-methylated-deami-nated metabolites. During 0-2 hr total recovery of injectedradioactivity in both I and NI phases was less (24 and25%) than in normals (37±4.9%o) and there was decreasedformation of O-methylated metabolites; during 2-48 hr therecovery of radioactivity was increased (69% I and 79%o NI)as compared to normals (61+9.7%). Liver MAOactivitywas normal, but COMTactivity was reduced in both I andNI phases (0.25 and 0.34 nmoles/g per hr). Five controlsamples contained 1.43-2.65 nmoles/g per hr. Also, the pa-tient showed increased responsiveness to injected NE in bothphases. The data suggest that decreased hepatic COMTinthis patient was associated with delayed NE metabolism andincreased pressor response to NE. The initial decreased ex-cretion of radioactivity may represent a compensatory in-crease in NE uptake by tissue in the presence of delayedmetabolism. Since metabolism of NE was similar in the twophases, and excretion of endogenous NE was increased inthe I phase, it is suggested that during the I phase thereis increased NE synthesis. Increased production of NEcoupled with decreased metabolism may explain the hyper-tension in the I phase.

18. Uptake of Radioiron-Labeled Hemoglobin and Trans-ferrin by Isolated Hepatic Kupffer and ParenchymalCells. MICHIYAsu AWAI,* BARBARA CHIPMAN,* AND EL-MER B. BROWN, St. Louis, Mo.

Although the hepatic uptake of radioiron after injectionof hemoglobin or transferrin has been studied, little is knownabout the initial cellular distribution of these forms of ironin the liver. Recent techniques permit the separation ofparenchymal cells from Kupffer cells so that the distribu-tion of iron bound to transferrin or hemoglobin could belocalized in specific cells of the liver at various times afterinjection. Hepatic parenchymal cells were isolated from ratsby two methods: by incubation of perfused liver with tetra-phenylboron or by mechanical tissue homogenization. Kupffercells were isolated using a pronase method. Purity of cellseparation was estimated by phase microscopy of phago-cytosed carbon particles and trypan blue, or by radioisotopiclabeled colloidal gold and technetium-sulfur-colloid in Kupffercells. 7 mg of hemoglobin-'Fe was injected intravenouslyand was removed 10 to 20 times more effectively by paren-chymal than by Kupffer cells of rats killed after 3-24 hr.Blocking Kupffer cell uptake by prior injection of heat-denatured albumin produced no change in total liver orparenchymal cell uptake of hemoglobin-'Fe. Distribution of'Fe in hepatic parenchymal and Kupffer cells was measured3-48 hr after injection of transferrin-6Fe into rats withmarrow erythroid hypoplasia produced by hypoxic poly-cythemia. At 3 hr 'Fe was almost equally distributed inparenchymal and Kupffer cells, at 6 hr maximal 'Fe uptakeoccurred in parenchymal cells, but at 48 hr uptake of 'Feby Kupffer cells predominated. Extension of these techniquesshould allow the exploration of various factors influencingiron uptake by the two major hepatic cells types. (Researchsupported by grants from NIH.)

19. In Vitro Inhibition of Phagocytosis and Disruptionof Microfilaments by Cytochalasin B. STANTONG. Ax-LINE* AND EvE REAVEN,* Palo Alto, Calif. (introduced byJack Remington).

The role of cytoplasmic microfilaments as a possible medi-ator of phagocytosis has been investigated using cytochalasinB (CB), an agent that interferes with cell locomotion. Pre-treatment of cultivated mouse peritoneal macrophages withCB (5-10 ,ug/ml) for 2-48 hr reduced the phagocytic up-take of polyvinyl toluene particles by 80%. The inhibitoryeffect of CB on both the attachment and interiorizationstages of phagocytosis was rapidly reversed by removal ofCB from the media. Observations by time-lapse photomi-croscopy showed that undulating plasma membrane activitywas completely inhibited within 10 min of CB application,resulting in irregular cytoplasmic retraction, rounding of thecentrosphere region, and shortening of pseudopods at sitesof attachment to glass. Electron microscopic examination ofmacrophages treated for periods from 10 min to 24 hr withCB showed disaggregation of oriented bundles of 40-50 Acytoplasmic microfilaments. In addition, cytoplasm specifi-cally associated with shortened pseudopods and invaginatedplasma membranes showed an increased density of the non-oriented subplasmalemmal microfilamentous network. Largerfilaments (70-100 A) and other cell organelles were un-affected by CB. The observed drug effects were calcium-independent, and protein synthesis was insensitive to CB.The morphological studies suggest that CB-induced micro-filament disorganization may be responsible for membraneimmobilization causing inhibition of phagocytosis. (Sup-ported by USPHS Grant AI-10055 and the VeteransAdministration.)

20. Effect of Group A Streptococcal Antibody on HeartValve Tissue Biosynthesis. ELIA M. AYOUB, MICHAELDENNIS,* AND GLENND. SWANK,* Gainesville, Fla.

The immunological relationship between antibody to theGroup A streptococcus and heart valve tissue was studied.A soluble product which reacted on agar diffusion with anti-body to the Group A streptococcus was obtained from human(rheumatic) and porcine mitral valves. Using the radio-immune precipitin technique, the soluble valve antigen in-hibited the reaction of purified streptococcal Group Acarbohydrate with Group A antiserum. The effect of humoralantibody on the biosynthetic activity of valve tissue was alsoinvestigated. Freshly obtained valve tissue was minced, incu-bated in tissue-culture medium containing radiolabeled pre-cursors (glucosamine-'C and proline-3H). The uptake ofthese precursors by the tissue, in the presence and absenceof antibody to valve tissue and Group A streptococcus, wasmeasured. The data obtained revealed that antibody to valvetissue produced almost complete inhibition of uptake of bothprecursors. However, while antibody to the Group A strep-tococcus produced inhibition of the uptake of glucosamine-"C,an enhanced uptake of proline-3H was observed in the pres-ence of this antibody. Similar results were obtained when thetissue was preincubated with the antiserum, then washedbefore incubation with the labeled precursor. Parallel studieswith human mitral valves were performed using proline-'Honly and showed enhanced uptake of this precursor in the

6 a

presence of streptococcal Group A antibody. The results ob-tained in this study confirm the presence of a commonantigenic component between the Group A streptococcalcarbohydrate and heart valve tissue. In addition, studieswith labeled precursors suggest that Group A antiseruminhibits synthesis of valvar glycopeptide-mucopolysaccharide,as reflected by inhibition of glucoasmine uptake, and stimu-lates synthesis of collagen-protein, as reflected by enhancedproline uptake. (Supported by NIH grant AI09645-03 andFlorida Heart Association grant 72 AG 16.)

21. Purine Reutilization in Human Lymphocytes. JOHNAYVAzIAN* AND SOLOMONSKUPP,* New York (introducedby Aaron J. Marcus**).

Human lymphocytes in culture, stimulated to transforma-tion by phytohemagglutinin, were used to study the relativeimportance and degree of reutilization of purines. In theinitial study lymphocytes undergoing transformation fromthe resting stage through initial mitosis were observed. Thecells were pulse-labeled early in transformation with hy-poxanthine-8-"C. Initially 70%o of the incorporated labelappeared in the mononucleotides and the remainder in RNA.Portions were taken daily for 3 days thereafter. At thecompletion of transformation 94% of the labeled purineswere found in nucleic acids. Cell survival was 95% asindicated by the increase in DNA. Despite the turnover andinterconversion of purine compounds during transformation,the total recovered "C activity for each portion remainedconstant with the last having 92% of the activity of thefirst. In the second study lymphocytes were labeled duringthe first synthetic phase and cultures followed for 9 daysthrough three to four cellular replications. The cells weredoubly-labeled with hypoxanthine-8-"C and thymidine-'H.Excess label was removed and chase media substituted whichinhibited reutilization of label from degenerating cells. Cor-rection was made for cell loss by using the decay in thespecific activity of DNA-3H to determine the mean numberof synthetic phases undergone by the cultures and thendetermining the theoretical yield of DNA assuming 100%cell survival. The theoretical DNA was compared to theobserved and cell loss calculated. During growth and repli-cation the labeled purines shifted from the nucleotides andRNAinto DNAbut despite this metabolic activity the cor-rected "C recoveries demonstrated an efficiency of reutiliza-tion of purines that exceeded 95%. (Supported by the Vet-erans Administration and NIH Grant AM-14059.)

22. Genetic Polymorphism of Parotid Basic (Pb) Pro-teins. EDWIN A. AZEN* AND WILMA B. BIAS,* Madison,Wis. and Baltimore, Md. (introduced by Fritz Bach).

A frequent polymorphism with three phenotypes showingmulti-banded patterns was observed among 90 randomly col-lected samples from Blacks and was studied by acid-ureastarch-gel. Autosomal docominant inheritance was postulatedand supported by excellent fit (P=0.89) of observed (64,23, 3) and expected (63.4, 24.3, 2.3) values for the pheno-types (1-1, 1-2, 2-2, respectively), as well as results offamily studies including 32 members. In two families withparents of alternative homozygous types (1-1 X 2-2), allnine children were heterozygotes (1-2). One heterozygous

variant, identical electrophoretically to those in Blacks, wasfound among 101 randomly collected samples from Whites,the rest being type 1-1. There were no variants in 19 sam-ples from Orientals, all being type 1-1. Gene frequencieswere: for Blacks, Pb' = 0.84; Pb' = 0.16; for Whites, Pb'

0.995; Pb2 - 0.005. A proteolytic degradative process in-volving these proteins and blocked by protease inhibitorswas found in parotid saliva on incubation studies, and thisexplained, in part, interesting variations in band mobilitiesand intensities within phenotypic classes. The Pb proteinsare distinct from parotid fluid amylase, IgA, lysozyme, andalbumin. There was nonidentity between Pb proteins andproducts of other gene loci tested, including ABO, Rh,MNSs, Kell, Kidd, Lutheran, Duffy, Lewis, Secretor, hapto-globin, transferrin, group-specific component, and hemoglobin(,-chain determinant). This polymorphism should be usefulfor genetic research since the frequency is high, and itappears virtually restricted to Blacks (like Fv [a-b-] andJSa). (Research supported by grant AMO-0955-07 fromNIH.)

23. Depression of Peripheral Thyroxine Turnover inHeroin Addicts. F. AzIzi,* A. VAGENAKIS,* A. CORMAN,*V. PATCH,* A. SAMARAWERA,*L. BRAVERMAN, AND S.INGBAR,** Boston, Mass.

Since the activity of hepatic drug-metabolizing enzymesand the rate of peripheral thyroxine turnover seem closelyrelated, and since morphine depresses drug metabolism insome species, we studied thyroxine metabolism in 6 addictstaking "street heroin" (H), 10 undergoing early withdrawal(W), and 10 on prolonged methadone maintenance (M).Compared to serum thyroxine concentration in 21 matchedcontrols (7.6±1.9 ug/ml), values were significantly increasedin H (11.3+2.7) and W (9.6±2.1). Fractional thyroxineturnover rates (per cent per day) and clearances (liters/day) were significantly decreased in H (8.0±0.9; 0.9±0.2,respectively) and W(7.6+1.1; 0.9±0.2). Despite individualabnormalities, mean values for serum thyroxine (8.8+2.5),fractional thyroxine turnover (10.7±1.7), and clearance(1.2+0.3) were not significantly altered in M. Despite thesedifferences between groups H and Wand group M, thy-roxine-binding capacities of TBG in the three groups(28±7; 27+5; 26±5 ,ug/100 ml, respectively) were equallyand significantly increased (normal, 18±4). Values for theper cent of free thyroxine were correspondingly decreased.Lack of the usual inverse relation between increased TBGand retarded thyroxine turnover was evident not only forthe three groups as a whole, but also when values in indi-vidual patients were examined. Mild abnormalities in liverfunction were present in many, but not all, patients in thethree groups. However, as judged from lack of correlationin individual patients and from previous studies of cirrhosisand hepatitis, disturbances in thyroxine binding and turn-over did not seem to result from hepatocellular disease.Conclusions: Abnormalities in thyroxine transport and turn-over frequently occur in patients taking heroin or methadone,particularly the former. These probably result from a directeffect of these agents on the cellular, presumably hepato-cellular, disposition of the hormone. (Supported by grantsAM07917 and AM07953 from NIH.)

7 a

24. Stimulation of Adenosine Diphosphate Uptake inRat Liver Mitochondria by Triiodo-L-Thyronine. BER-NARD M. BABIOR,* SIDNEY H. INGBAR,** SUSAN M. CREA-GAN,* AND RUBY S. KIPNES,* Boston, Mass.

Entry of ADP into the mitochondrion largely dependsupon a carrier-mediated process in which ADP is ex-changed for ATP, a process that is specifically inhibited bythe complex organic acid, atractyloside. Since the rate ofmitochondrial substrate oxidation is limited by the avail-ability of ADP (Lehninger, The Mitochondrion, Chapter 7),we have investigated whether thyroid hormone-induced in-crease in mitochondrial oxygen consumption could resultfrom stimulation of mitochondrial ADP uptake. Atractylo-side-sensitive uptake of ADP-14C and ADP-dependent suc-cinate oxidation by rat liver mitochondria were studied inspecimens obtained 20 hr and 3 days after animals had beengiven a single large dose (20 ,ug/100 g) of triiodo-L-thyronine(T8) i.p. No significant effect of Ts on ADP uptake wasobserved at 20 hr. By 3 days, however, a time when wholebody oxygen consumption response to T3 reaches a peak, asignificant increase in atractyloside-sensitive ADP uptakewas seen in Ta-treated animals as compared to saline-treatedcontrols (2.35+0.58 vs. 1.81±0.63 nmoles/min per mg pro-tein; P < 0.20). This effect could not be ascribed to in-creased oxygen consumption, since ADP uptake was mea-sured in the absence of oxidizable substrate. In both controland T3-treated mitochondria, a significant correlation be-tween ADP-dependent oxygen consumption and atractylo-side-sensitive ADPuptake was evident. Moreover regressioncurves for these functions in the two groups were virtuallyidentical, indicating that T3, although stimulatory, did notalter the stoichiometric relationship between oxygen con-sumption and ADPuptake. No increase in ADPuptake wasobserved when T, (10' to 101 M) was added to suspensionsof mitochondria from normal rats, either 1 hr or imme-diately before measurements were made. Conclusions: Theobservations are consistent with the interpretation that theTi-induced increase in mitochondrial oxygen consumption issecondary to an increase in carrier-mediated ADP uptake.(Research supported by grant AM-09753 from NIH.)

25. De Novo Lipid Synthesis as a Repair Mechanismfor Membrane Damage in Mammalian Erythroid Cells.SAMIR K. BALLAS* AND EDWARDR. BURKA, Philadelphia,Pa.

Because the mammalian reticulocyte is a nondividing cell,lipid turnover in the cell membrane has been presumed torepresent exchange rather than de novo synthesis. The im-portance of the cell membrane in activities other than celldivision has prompted investigation of the relationship be-tween erythroid cell lipid metabolism and a functionallyintact cell membrane. Membranes of rabbit and humanerythroid cells, incubated under various conditions in thepresence of labeled lipid precursors, were isolated and theradioactivity in neutral (NL) and phospholipids (PL) wasdetermined. Reticulocytes incorporated acetate, glycerol, fattyacids (FA), and inorganic phosphorus into total cell lipids(TCL). Partial hydrolysis of PL showed glycerol-1'C in-corporation via a-glycerophosphate into the three carbonbackbone of PL, confirming de novo synthesis. Glycerol is

also incorporated into NL via the Emden-Meyerhof pathway.Acetate-"'C appeared in NL and the acyl and 3-C backboneof PL, indicating metabolism to a-glycerophosphate via theEmden-Meyerhof pathway. Since acetate and glycerol arenot incorporated into erythrocyte TCL, enzymes necessaryfor de novo lipid synthesis disappear during cell maturation.Puromycin, known only as an inhibitor of protein synthesis,stimulated glycerol incorporation into PL, but not NL, inrabbit and human cells in a concentration-dependent manner(maximum 174±44% at 10' M) at a step subsequent toformation of a-glycerophosphate. Agents causing transientdamage to the cell membrane, such as 10' M butanol or 1.5X 10' M vincristine, also specifically stimulated the de novosynthesis of PL. The data indicate that the mammalianreticulocyte can synthesize total lipids from simple pre-cursors. The pathway for de novo PL synthesis is stimu-lated, at a step subsequent to the formation of a-glycero-phosphate, by puromycin by an unknown mechanism and byagents causing membrane damage. The capacity for increasedPL synthesis may represent a repair mechanism for the cellmembrane under conditions of stress. (Supported by NIH.)

26. Pathogenesis of Alcoholic Hyperlipemia. ENRIQUEBARAONA,* ROMANOC. PIROLA,* AND CHARLES S. LIEBER,New York.

Alcohol produces hyperlipemia, an effect which is muchgreater in fed than in fasted rats and which we found tobe associated with increased lipoprotein production. Recently,because of enhanced lymph lipid output and intestinal fattyacid esterification in vitro, the intestine was incriminated inthe development of the hyperlipemia observed in fasted rats16 hr after ethanol. To assess this mechanism in the post-prandial state, rats were given ethanol acutely (3 g/kgintragastrically) or isocaloric carbohydrate in liquid dietscontaining tripalmitin-14C. In the 1st hr after ethanol, intes-tinal lymph flow increased by 63% (P <0.01), incorpora-tion of dietary fatty acids into lymph lipids by 165%(P < 0.01), and lymph lipid output by 49% (P < 0.05),compared to controls. However, no postprandial hyperlipemiaensued and fatty acid esterification by intestinal slices wasdecreased. By contrast, hyperlipemia developed after admin-istration of a dietary load of fat (with or even withoutethanol) in rats given 36% of calories as alcohol for 3-4 wkbut not in littermates pair-fed control diets. Lymph flowand lipid output were not increased whereas fatty acidesterification by intestinal slices was enhanced. Thus, esteri-fication in vitro did not correlate with lymph lipid output.Moreover, after lymph diversion, chronic ethanol feedingstill produced hyperlipemia (compared to controls), pro-vided adequate and equal intravenous lipid loads (lymph orchylomicrons) were administered; incorporation of i.v.lysine-'H into serum lipoproteins also remained significantlyincreased. Thus, acute ethanol administration increased lymphlipid output but did not produce postprandial hyperlipemia,whereas in animals fed ethanol chronically, hyperlipemiadeveloped without increased lymph lipid output or despitereplacement of intestinal lymph by a constant i.v. fat load;this suggests a nonintestinal, most likely hepatic, mechanism.(Supported by USPHSgrants AM 12511, MH15558, andthe Veterans Administration.)

8 a

27. Chemical Bonding in Sputum Gels. A. D. BARTON,*J. V. POWERS,* AND R. V. LoURENgO,* Chicago, Ill. (intro-duced by Morton D. Bogdonoff**).

We investigated the types of chemical bonding responsiblefor the gel structure of thick sputum because abnormalbronchial secretions may impair mucociliary clearance, ob-struct the airways, and facilitate the persistence of chronicinfection. We studied the dispersal of sputum samples rang-ing from mucoid to grossly purulent, by several agents thatcleave secondary bonds; the material not dispersed was sepa-rated by sedimentation. Extraction with 30% propylene gly-col (PG) dispersed mucoid sputum almost completely.Mucopurulent sputum contains deoxyribonucleoprotein(DPN) fibrils that constitute part of the gel structure;this material resisted dispersal by PG, but was dispersedby extraction with N-acetylcysteine (NAC). Part of themucopurulent sputum gel resisted dispersal by all of thesecondary bond agents tested, including urea, sodium do-decylsulfate, concentrated salt solutions, etc. The residuewas rich in a y-glutamyl transpeptidase (GGT) whose levelof activity may indicate the existing inflammatory reaction,since it parallels the degree of purulence. The gel structureof mucoid sputum appears to be due to secondary bondingsince PG dispersed it. In mucopurulent sputum the DPNfibril material may be cross-linked with S-S-bonds, since itis dispersed by NAC but not by PG. Part of the muco-purulent sputum gel appears to be cross-linked by othermore stable covalent bonds since it resists dispersal by allof the secondary bond agents tested. The presence of aGGTsuggests that transpeptidation may play a role in suchcross-linking. (Research supported by NIH grant HE13824and Veterans Administration funds.)

28. Molecular Basis of Glucocorticoid Response: Spe-cific DNABinding of a Steroid-Receptor Complex. JOHND. BAXTER,* Guy G. ROUSSEAU,* STEPHEN J. HIGGINS,*AND GORDONM. TOMKINS,** San Francisco, Calif.

The mechanism of glucocorticoid hormone action has beenstudied in two systems: induction of specific protein syn-thesis in cultured rat hepatoma (HTC) cells, and lysis ofcultured mouse lymphoma cells. In HTC cells, inducersteroids (S) bind reversibly to specific cytosol receptorprotein (B). This interaction leads to an initial confor-mational change in the receptor (SB*) followed by a second,temperature-dependent alteration in its properties (SB**).The resulting complex (SB**): (a) binds in the nucleuswith high affinity to a fixed number (... 20,000 per nucleus,1 pmole/mg DNA) of DNA-containing sites; (b) bindswith a similar affinity to a fixed number of sites (7 pmoles/mg DNA) on purified HTC cell (but not Micrococcuslysodeikticus) DNA. Thus binding to DNA simulatesbinding to nuclei, although in the latter chromosomal pro-teins may restrict the number of available sites. Proges-terone, an anti-inducer in HTC cells, reversibly binds to thecytosol receptors but does not promote the conformationalchanges required for nuclear binding in intact cells, or forhigh-affinity binding to purified DNA. Development of steroidresistance in lymphoma cells is associated with a markedreduction (by 90%) in cytosol receptors and further altera-tions in their properties. Nuclei from resistant cells retain

the capacity for binding steroid-receptor complexes fromsensitive cells. These studies support the following hypothe-ses: (a) glucocorticoids are allosteric effectors which induceconformational changes in specific cytoplasmic receptors;(b) the steroid-receptor complex then binds to specific siteson nuclear DNA to affect gene transcription and therebyprotein synthesis; and (c) steroid unresponsiveness mayresult from alterations in the cytoplasmic receptor proteins.(Research supported by NIH and ACS.)

29. Specificity of the Change in the Amino Acid Com-position of the Diabetic Glomerular Basement Mem-brane. PAUL J. BEISSWENGER,* Philadelphia, Pa. (intro-duced by Albert I. Winegrad).

The human glomerular basement membrane (GBM) is aglycoprotein of the collagen family containing 7% carbo-hydrate. In previous studies GBMfrom patients with long-standing diabetes was found to differ from that of age-matched controls in that they contained increasedhydroxylysine and decreased lysine although the totallysine plus hydroxylysine was unaltered. This suggestedthat a greater fraction of the lysine residues was hy-droxylated after peptide synthesis in the diabetic. The dia-betic GBMalso contained an increase in glucosyl-galactosecovalently linked to hydroxylysine residues (P. J. Beiss-wenger and R. G. Spiro. 1970. Science (Washington). 168:596). To determine the specificity of these changes theamino acid composition of GBMwas determined in non-diabetics with well characterized renal diseases associatedwith proteinuria. The amino acid composition of GBMfrompatients with accelerated essential hypertension, systemiclupus erythematosus, severe nephrosclerosis, rapidly pro-gressive glomerulonephritis, and postpartum necrotizingrenal arteriolitis was determined and compared with GBMfrom age-matched nondiabetic controls without renal disease.In contrast to diabetic GBM, the amino acid compositionof the GBMof nondiabetics with renal disease and protein-uria did not differ from that of age-matched normal controls.In all patients the ratio of lysine/hydroxylysine in the GBMwas greater than 1.0; in the GBMof diabetics the ratioaveraged 0.65. In the nondiabetic patients with renal diseasemore than 20% of the residues are glycine; hydroxyprolineand proline were present in nearly equal quantities; andthere was a significant amount of half-cystine as in the nor-mal human GBM. These observations suggest that thechanges observed in the composition of the diabetic GBMmay be unique to diabetes mellitus. (Research supported bythe American Heart Association.)

30. Serum Inhibitors of Chemotactic Factors. JEFFREY L.BERENBERG*AND PETER A. WARD,* Washington, D. C. andFarmington, Conn. (introduced by Irwin H. Lepow**).

The present studies demonstrate the existence in humanserum of an inhibitor of chemotactic factors. Using in vitromicropore filter techniques and human or rabbit neutrophilsas the indicator cells, chemotactic responses of leukocytes tothe complement-derived chemotactic factors (C3a, C5a,C567) and a chemotactic factor derived from Escherichiacoli have been studied. The chemotactic inhibitor isolatedfrom human serum blocks the activity of all chemotactic

9 a

factors tested, is soluble in ammonium sulfate (at 45%saturation), and in sucrose density gradient ultracentrifuga-tion as well as gel filtration appears in a biphasic distri-bution with activity near the IgG and albumin markers.The inhibitor acts directly on the chemotactic factor ratherthan on the leukocyte. Evidence for binding of inhibitor tochemotactic factor (C3a) has been negative. Demonstrationof the inhibitor in normal human serum requires fractiona-tion and/or concentration of serum, whereas serum frommost patients with agammaglobulinemia (Bruton type) or

Hodgkins disease is rich in inhibitor activity. In these patho-logical sera inhibitor activity is at least fourfold increasedover the level in normal serum. To date all studies suggestthe inhibitor present in pathological sera is qualitativelysimilar to, or identical with, the inhibitor present in normalhuman serum. The amount of chemotactic activity generatedin serum by addition of zymosan is significantly less ininhibitor-rich serum compared to normal (inhibitor-poor)serum. The existence in serum of inhibitors of chemotacticfactors may represent a control mechanism of these inhibi-tors. While the chemotactic factor inhibitor(s) differs fromthe anaphylatoxin inactivator in many respects, evidence doesnot yet permit a conclusive statement about the relationshipof these two inhibitors (inactivators). (Supported in partby NIH Grant AI-09651.)

31. Defective Bromsulphalein Metabolism in Patientswith Gilbert's Syndrome (GS). PAUL D BERK* AND TER-RENCE F. BLASCHKE,* Bethesda, Md. (introduced by Na-thaniel I. Berlin**).

Although the defect in bilirubin metabolism in GS hasbeen extensively studied, hepatic transport of other organicanions in this condition has received little attention. Wehaveexamined the metabolism of bromsulphalein (BSP) (5mg/kg) and indocyanine green (ICG, 0.5 mg/kg) in 26patients in whom GS was diagnosed on clinical grounds,confirmed by studies of bilirubin kinetics (24), and supportedby liver biopsy findings (18). Plasma BSP concentration at45 min was abnormal (0.52 ;1.52 mg/100 ml; nl < 0.50) in 9of the 26 patients. Complete BSP disappearance curves were

obtained in 18 of the patients and 12 normal volunteers(NV's) and analyzed by digital computer. In all instances,the curve was resolvable into a sum of two exponentials.In eight patients (group I), curves were indistinguishablefrom normal, and BSP clearance (CBSP) was 4.3±0.6 ml/min per kg (nl 4.4±0.7). 10 of the patients, including all9 with increased 45 min retention, had abnormal curves.In four of these (group II), initial disappearance rate fromthe plasma was normal, but increased 45 min retention re-

sulted from an increased intercept of the second exponential.In six (group III), initial disappearance rate was reduced.CBSP in group II (3.1±0.6 ml/min per kg) and group III(2.6+0.4 ml/min per kg) were both significantly reducedcompared to NV's (P < 0.01, P < 0.001, respectively). Com-partmental analysis indicates that the curves in group IIIresult from defective hepatic BSP uptake, while those ofgroup II are due to an abnormality at a later stage of theBSP transport process. There were no significant correla-tions between abnormalities of specific parameters of BSPmetabolism and corresponding parameters of bilirubin me-

10

tabolism in these studies. In contrast to the results withBSP, CICG in 16 patients with GS (8.6±2.4 ml/min per kg)was not significantly different (P > 0.6) from the value in17 NV's (8.9+1.6). The data indicate that the hepatic de-fect in GS may involve organic anions other than bilirubin,and illustrate the heterogeneity of the patients classifiedunder that diagnosis.

32. The Regulation of Intracellular Glutathione Levels.E. BEUTLER, N. V. PANIKER,* AND K. G. BLUME,* Duarte,Calif.

Red cell reduced glutathione (GSH) levels are increasedin some patients with myeloproliferative disorders. The ad-ministration of drugs such as diaminodiphenylsulfone(DDS) or methylene blue (MB) also increases GSHlevels. At one time it was believed that the increase ofGSHfound under these conditions might be due to reduc-tion of intracellular GSSG to GSH. However, with thedevelopment of an improved method for GSSGdetermina-tion we were able to demonstrate that about 99.75%o of theglutathione in red cells was already in the reduced state.The synthesis of red cell glutathione is accomplished in twosteps. First a peptide bond is formed between glutamic acidand cysteine by y-glutamylcysteine synthetase (GC-S) andthen the product, y-glutamylcysteine, forms a bond withglycine through the action of GSH synthetase (GSH-S).The activity of these two enzymes was measured in a groupof patients with myeloproliferative disorders, includingchronic granulocytic leukemia, myelfibrosis, and polycy-themia vera. Patients with elevated red cell GSH levelswere found to have increased levels of GSH-S, but normalGC-S activity. No change in the kinetic properties of theenzyme was observed. In studies of rabbits we found MBand DDS administration increased the GSHcontent of redcells both in vivo and in vitro. MB and DDS failed toaffect GC-S, but produced a kinetic change in the secondenzyme, GSH-S, lowering its Km for -y-glutamylcysteine.Since the levels of this peptide in red cells is much lowerthan the Km this increases the rate of the GSH-S reactionwithout affecting the amount of enzyme present. Thesestudies indicate that the level of GSHin red cells is regu-lated by the activity of GSH-S. Pathophysiological altera-tions of the activity of this enzyme alter the level of GSHin erythrocytes.

33. Glucose Stimulation of Active Sodium Transport inthe Human Jejunum. HENRY J. BIrNDER,* New Haven,Conn. (introduced by H. M. Spiro**).

Alternate theories have been advanced to explain themechanism of the stimulation of sodium absorption in thehuman jejunum by glucose. One proposes active Na trans-port based on the Na gradient hypothesis associated withcarrier-mediated Na entry into the epithelial cell, and theother suggests that glucose-stimulated Na movement in theabsence of HCO3 is not an active process but secondary tobulk water flow. To test these hypotheses 10 pieces of nor-mal jejunum at laparotomy from four patients were placedin an Ussing chamber. Identical HCO3-free, a-free 140mmNa isethionate solutions bathed both the mucosa andserosa. Spontaneous transmural potential difference (PD)

and short-circuit current (I.,) were monitored before andafter the addition of 10 mmglucose. Without glucose mean(+SE) PD and 1L. were - 1.4+0.2 mv (mucosa negative)and 14+2 Asa/cm', respectively. Glucose caused an immediateand sustained increase of both PD and Is to -7.1+0.5 mvand 81+7 ,ua/cm'. Simultaneous unidirectional Na fluxes(Jet) with '2Na and "Na under short-circuit conditionswere determined on seven pieces of tissue from two patients.Before glucose addition Jn'et was 0.6+0.2 ,uEq/hr - cm' whenthe I,, was 0.5±0.1 /AEq/hr * cm'. Glucose increased bothJint (2.5±0.3 /AEq/hr - cm') and Le (3.0±0.2 AEq/hr - cm').These studies provide convincing evidence that active glu-cose-stimulated electrogenic Na transport is present in thehuman jejunum although they do not exclude the existenceof other possible mechanisms. (Supported by grants fromNIH and John A. Hartford Foundation.)

34. Control of Embryonic Cartilage Growth by Endo-genous Cyclic Adenosine-3'5' Monophosphate. B. M.BIRCH,* H. K. DELCHER,* J. L. RENDALL,* AND H. E. LEBO-VITZ,* Durham, N. C. (introduced by S. H. Appel).

Wehave shown that serum sulfation factor (SSF)-stimu-lated 'SO4 incorporation into embryonic chick cartilage is par-tially inhibited by exogenous cyclic AMP(CAMP) in a dose-related manner. SSF stimulation of cartilage had no effecton tissue CAMPlevels. The present study was designed todetermine whether endogenous CAMP levels alter the re-sponse of chick embryonic cartilage to SSF. 3SO4 incorpora-tion into pelvic cartilage from 12- to 14-day chick embryoswas determined using a modification of Hall's technique.Cartilage CAMPwas assayed by the competitive protein-binding technique of Gilman. Glucagon and insulin did noteffect 'SO4 incorporation by cartilage, nor tissue CAMPlevels in the presence or absence of SSF. Theophylline at2.5 and 5 mmcaused a dose-related inhibition of SSF-stimu-lated 'SOs incorporation (15 and 46%, respectively). Cor-responding cartilage CAMPlevels increased 150 and 292%above control values. A positive correlation existed betweentissue CAMPlevels and inhibition of 'SOs uptake. ATP,5'AMP, and adenosine at 1 mmall significantly inhibitedSSF-stimulated 'SOs incorporation into embryonic chickcartilage (29, 44, and 50%). Tissue CAMPlevels rose incartilages incubated with 1 mmATP, 5'AMP, and adenosine(41, 52, and 62%). These adenosine compounds at 1 mmdonot cross-react with our CAMP-specific protein kinase.Measurements of chromatographically isolated CAMPfromcartilages incubated with the various agents confirmed theelevations noted in tissue extracts. This data supports thethesis that CAMPacts as a modulator of SSF action oncartilage. It also demonstrates that incubation with ATP,5'AMP, and adenosine leads to elevated cartilage CAMPlevels. (Research supported by grants from the NIH andVA.)

35. Pyrogen Production In Vitro by Lymphoid Tissue inMalignant Lymphoma. PHYLLIS BODEL, New Haven, Conn.

The mechanism of fever in malignant lymphoma is un-known. Endogenous, pyrogen (EP) production by blood ortissue leukocytes has been clearly implicated as the cause ofexperimental fever in several animal models. Therefore blood,

spleen, and lymph node cells from patients with lymphomaswere examined for their capacity to release EP. Cell suspen-sions prepared from operative specimens were incubated forsuccessive 24-hr periods in sterile, pyrogen-free tissue culturemedium; blood leukocytes were incubated overnight. Super-nates were cultured and injected into rabbits for pyrogenassay. Blood leukocytes from febrile or afebrile patients, likethose from normal individuals, rarely produced EP spon-taneously, although pyrogen release always occurred after aphagocytic stimulus: However, 5 of 10 preparations of spleencells and 4 of 5 lymph node cell suspensions from patientswith Hodgkin's disease released EP spontaneously. By con-trast, no pyrogen was detected after comparable incubationsof eight preparations of spleen cells from patients withsplenomegaly resulting from various nonmalignant disorders.EP was 'produced in all instances after addition of heat-killed staphylococci. Pyrogen from tissues was clearly dif-ferent from endotoxin, and appeared identical with leuko-cyte pyrogen. However, unlike blood leukocytes, spontane-ously active spleens and lymph nodes from patients withlymphoma usually continued to produce EP for 48-72 hr.Production was suppressed by inhibitors of protein synthesis.Spontaneous pyrogen release did not correlate with (a)presence of fever in the 'patient, (b) pathological diagnosisof tissue infiltration, or (c) numbers of granulocytes in thetissues. Mononuclear leukocytes were clearly responsible forEP production by lymph node cell suspensions. These studiesdemonstrate for the first time spontaneous and prolongedproduction of EP by lymphoid tissue from patients withmalignant disease.

36. Fractionation of Human Lymphocytes: Isolation ofTwo Populations Differing in Their Response to Con-canavalin-A. DAVID BOLDT,* SR. ANN MARIE SKINNER,*AND STUARTKORNFELD, St. Louis, Mo.

Lymphocytes represent a functionally heterogeneous popu-lation of cells despite their similar morphology. Techniquesfor separation of lymphocyte subclasses have been identifiedfor several laboratory species, yet in man, fractionationof lymphocytes into functionally distinct subpopulations hasproved to be more difficult. We have identified two humanlymphocyte subpopulations differing in responsiveness to theplant mitogen concanavalin-A (Con-A). When peripheralblood lymphocytes (responsive to Con-A) are passed througha nylon column, a subpopulation of lymphocytes responsiveto Con-A is removed and another subpopulation of cells rela-tively unresponsive to Con-A emerges from the column.Con-A-stimulated DNA synthesis is 6.5-fold greater and thepercentage blast formation is 4- to 5-fold greater in the re-sponsive vs. the unresponsive population. Mixing responsivewith unresponsive cells fails to induce responsiveness in thelatter indicating that a "helper" cell is not involved. Thefailure to respond to Con-A is specific for that mitogen sinceboth populations respond equally to phytohemagglutinin andthe pokeweed mitogen. Fluorescein-conjugated Con-A bindsto all cells in both populations, the predominant pattern beinghomogeneous staining. Cap formation occurred in 10% of theresponsive population vs. 2% in the unresponsive population.Binding studies with 'I-labeled Con-A demonstrate that theresponsive population possesses four times as many Con-A

11 a

receptors per cell while both cell populations bind the mito-gen with the same affinity (Ka of 10-20 ,ug/ml). From thesedata we conclude that (a) nylon columns fractionate lym-phocytes into distinct subpopulations, characterized by dif-ferent abilities to respond to various plant mitogens; and (b)binding of a plant mitogen to the lymphocyte cell surface isnecessary but not sufficient to induce blast transformation.(Research supported by grants from NIH.)

37. Starvation: Dissociation of Renin and Aldosteroneand the Insensitive Nephron. PHILIP BOULTER,* RICHARDSPARK,* AND RONALDARKY, Boston, Mass.

Early starvation is characterized by natriuresis, and sub-sequent carbohydrate refeeding by antinatriuresis. To investi-gate these phenomena, sequential analyses of plasma reninactivity (PRA) and aldosterone secretory rate (ASR) wereperformed in eight healthy obese subjects, after equilibra-tion, while undergoing 10- and 15-day periods of fasting withcarbohydrate refeeding on a fixed 50 mEq Na -87 mEq Kintake. During early fasting, urinary sodium excretion (UNa)increased from a mean control of 43.2 to a peak of 83.7 mEq/24° while mean ASR increased from 393 to 636 ,ug/24°.PRA was dissociated from the rising ASR, decreasing froma mean control of 2.6-1.4 ng/ml per hr. Carbohydrate refeed-ing induced an abrupt antinatriuresis, mean UN. of 1.4 mEq/240, and a fall in mean ASR from 686 to 323 itg/24'. Withrefeeding, PRA was again dissociated from the falling ASRby increasing from 2.71 to 7.68 ng/ml per hr. Persistentnatriuresis occurring despite rising ASR suggested tubularinsensitivity to mineralo-corticoid in early fasting. Adminis-tration of fludrocortisone (0.6 mg/day) for 2 days duringearly fasting failed to prevent either the natriuresis or therise in ASR, whereas a similar dose of fludrocortisone ad-ministered on days 11 and 12 of fasting resulted in promptantinatriuresis and suppression of endogenous aldosteronesecretion. The natriuresis of early starvation and the anti-natriuresis of carbohydrate refeeding are characterized bydissociations of the renin-aldosterone system. ASR riseswhile PRA falls during natriuresis; ARS falls and PRArises with antinatriuresis. Early starvation is further charac-terized by renal tubular insensitivity to endogenous and exo-genous mineralo-corticoids. Restoration of tubular sensitivityto mineralo-corticoid is demonstrable as fasting progresses.These data suggest that renal tubular utilization of availablemetabolic substrates is a major determinant of mineralo-corticoid action. (Research supported by NIH.)

38. On the Nature of the Circulating Natriuretic Factorin Chronic Uremia. JACQUES BOURGOIGNIE,* R. WILLIAMScHMIDr,* Kuo HWANG,* TEWFIK NAWAR,* SAULO KLAHR,LEWIS CHASE,* AND NEAL S. BRICKER, St. Louis, Mo.

The striking and progressive increase in natriuresis pernephron that occurs in patients with advancing renal diseaseon a normal salt intake suggests that natriuretic forces aremobilized increasingly as GFR falls. Thus, if a natriuretichormone is among these forces, the opportunity for its detec-tion might be enhanced greatly in patients with advanceduremia. Wehave described a gel filtration fraction of uremicserum that inhibits sodum transport in vitro and is natriure-

tic in the rat. It is unclear, however, whether this factor is aphysiologic mediator of sodium excretion or a uremic toxinretained in blood because of failure of excretion. The presentstudies were performed to further characterize this material.The natriuretic effects of the fraction could not be repro-duced by the uremic "toxins" guanidinosuccinic acid ormethylguanidine. On the other hand, incubation of the frac-tion with leucine aminopeptidase abolished its natriuretic ac-tivity suggesting that the active factor might be a peptide.In addition, the active fraction increased 3'5' cyclic AMPex-cretion. Phosphate excretion also increased; however, theactive material does not appear to be related to parathyroidhormone. Nor does it appear to be a prostaglandin, angio-tensin, thyrocalcitonin, or vasopressin. The relationship ofthe fraction to sodium balance is suggested by the observa-tion that no natriuretic activity could be detected in the serumfraction from severely uremic patients with marked edemaformation. Finally, in uremic dogs in which natriuresis pernephron was prevented by decreasing sodium intake in pro-portion to GFR, no natriuretic activity was found. The ob-servations thus are consistent with the view that the natri-uretic factor of uremia may be a peptide hormone with aphysiologic role.

39. Cyclic Adenosine Monophosphate (cAMP) and Leu-kocyte Function: Effects of Choleragen. H. R BOURNE,*L. M. LICHTENSTEIN, AND C. S. HENNEY,* San Francisco,Calif., and Baltimore, Md.

Choleragen, the exotoxin of Vibrio cholerae, increasescAMP content of gut and other tissues. Wetested the effectof choleragen on two leukocyte functions thought to be regu-lated by cAMP in vitro: (a) IgE-mediated release of hista-mine from leukocytes of allergic donors (H-release) ; (b)Cytolysis of allogeneic mast cells by splenic lymphocytes ob-tained from specifically immunized mice (lymphocyte cyto-lytic activity, LCA). Choleragen, 10 ng/ml, inhibited bothH-release and LCA by 75-90% and caused a parallel four- tosevenfold rise in cAMP content of human leukocytes andmouse splenic lymphocytes. These effects of choleragen werecompetitively blocked by a specific cholera antitoxin andcholeragenoid. Choleragen's effects were undetectable before30 min, but increased thereafter to a plateau at 60-90 min.The rise in cAMPwas associated with increased basal adenylcyclase activity, and was not prevented by cycloheximide,1 X 10' M. Antitoxin and choleragenoid completely preventedthe rise in cAMP, measured at 60-90 min, only if added tocells simultaneously with choleragen; the block was unde-tectable if the antagonists were added at 10 min or later.Also, repeated washing of leukocytes after 10 min of ex-posure to choleragen (at either 00 or 37°C) failed to pre-vent the subsequent cAMP accumulation, the increase inadenyl cyclase activity, or the inhibition of histamine release.The choleragen effect on leukocyte cAMP contrasted withthose of prostaglandins, catecholamines, and histamine, whichwere rapid (maximal at 10 min) and easily reversed bywashing. These results suggest that choleragen stimulatescAMP production by a unique mechanism involving rapidand irreversible binding to the leukocyte, followed by a de-layed stimulation of adenyl cyclase. Further, they confirmprevious reports implicating leukocyte cAMP as an inhibitor

12 a

of both immediate (H-release) and delayed (LCA) hyper-sensitivity in vitro. (Research supported by NIH grants.)

40. Clinical Studies on the Luteinizing Hormone andFollicular-Stimulating Hormone Releasing Activityof pGLU-HIS-TRP-SER-TYR-GLY-ARG-LEU-PRO-GLY-NH2. C. Y. BOWERS,* J. MALACARA,* F. GOMEZ,*S. HERNANDEZ,* J. CHANG,* AND K. FOLKERS,* New Or-leans, La., Leon, Mexico, and Austin, Tex. (introduced byG. E. Burch).

To study the problem of whether DP is the physiologicalregulator of both luteinizing hormone (LH) and follicular-stimulating hormone (FSH) or only LH, serum LH andFSH levels were measured before and 15, 30, 45, and 60min after a single intravenous injection of 50 or 200 ,ug ofsynthetic DP in 30 normal men (NM), 30 normal women(NW), 8 postmenopausal women (PW) before and duringoral medication with 5 mg diethylstilbesterol (DES) daily,4 NMbefore and during medication with 100 mg testos-terone (T) given intramuscularly five times daily, and 5women with amenorrhea (A). The greatest mean A mIU/ml rises in NWstudied three times at monthly intervalswere: LH 90, 62, 67, and FSH 18, 3.5, 5.8. Largest A FSHrises (mean 37) were obtained in 13 of 89 tests performedin these NWand the associated mean A LH rise was > 100.Nine of these 13 NWwere tested near the midcycle. Theother A mean rise results were: NMLH 55 and FSH 3.5;PWLH 126 and FSH 30 and after 5 wk of DES Rx LHwas 36 and FSH 4.4. The three daily mean values of theNMbefore T were LH 41, 50, 50, and FSH 3.6, 4.9, 2.9,and during T Rx were LH 24, 34, 34, 54, 95, and FSH0.6, 1.2, 0.9, 0, 0. The mean A LH rise in the women withA was 109 and A FSH rise 8.4. In conclusion, DP releasesboth LH and FSH in NWand NM. The LH and FSHsurge at ovulation could be explained solely by the LH-FSHactivity of the DP; thus, it appears to be the hypothalamic-ovulation-hormone. However, because of the consistentlyhigh LH and low FSH activity of the DP, it is unlikelythat DP regulates both LH and FSH except at ovulation.Thus the DP seems to be physiologically only LH-RH.

41. Identification of Puberty by Synchronization of Lu-teinizing Hormone Release with Sleep. ROBERTM. BOYAR,*JORDAN FINKELSTEIN,* HOWARDROFFWARG,* AND LEONHELLMAN,** New York.

Luteinizing hormone (LH) and follicle-stimulating hor-mone (FSH) were measured in plasma by radioimmuno-assay every 20 min for 24 hr in 12 normal children (ages8-15) in various stages of sexual maturation and 5 normaladult men (ages 21-40). Polygraphic monitoring was carriedout simultaneously to precisely identify sleep onset, wakeful-ness, and sleep stages. Four prepubertal children showedLH concentrations in the prepubertal range (2.0-4.0 mIU/ml) throughout the 24 hr period. Four boys in early pubertyshowed LH secretory patterns during the day that wereindistinguishable from prepubertal children (2.5-4.5 mIU/ml); however, with onset of night sleep, there was animmediate increase of LH secretory activity that resultedin mean LH concentrations two- to fourfold higher (7.1-10.1

mIU/ml) than during wakefulness (P <0.001). By experi-mentally delaying sleep onset, synchronization of this LHsecretory "program" with actual sleep was clearly demon-strated. The number of LH secretory episodes during eithernight or day sleep corresponded to the number of sleepcycles. Sleep LH secretory activity was independent of FSHwhich did not show any sleep synchronization. Two childrenwith complete sexual maturation showed LH secretory epi-sodes indistinguishable from the 24-hr LH patterns of fivenormal adult men with no significant difference betweensleep and awake periods. These data show that (a) in earlypuberty, LH secretory activity becomes synchronized withsleep; (b) this CNS "program" for LH release is lost withadvancing sexual maturation; (c) there is absence of con-currence between LH and FSH release throughout pubertysuggesting different control mechanisms. This finding mayconstitute a new biological index for early detection ofpuberty.

42. Intracellular Potassium: the Regulator of Aldos-terone Production. JOHN E. BOYD* AND PATRICK J. MUL-ROW, New Haven, Conn.

Small acute increases in plasma [K+] stimulate aldos-terone secretion, while chronic K loading of rats increasessecretion, although plasma [K+], ACTH, and renin activitymay be unchanged (Boyd et al. 1971. Endocrinology. 88:556). Using the in vitro rat adrenal gland technique, wehave investigated this problem further and have shown thatan increase in [K+] of 0.35 mEq/liter (3.85-4.20) in theincubation media stimulated aldosterone production 43%(P < 0.01) without affecting gluco-corticoid production (B).It seems unlikely that such small changes in E.C.F. [K+]altered membrane potential and hence steroid secretion,rather, the data suggest intracellular K+ is the regulatorof aldosterone production. The effect of K+ requires Na-K-ATPase activity. Procedures which inhibit Na-K-ATPasereduce aldosterone production. The absence of Na+ ormarked increase in [K+] to 24 mEq/liter in the incubationmedia reduce aldosterone production. The addition of ouabain(5 X 10i M), which is known to reduce rat adrenal cell K@,blocks the stimulation of aldosterone production by in-creased [K+], 1 vs. 8 mEq [K+], %A aldo 90, with ouabain13, P < 0.01. Increasing the [K+] further partially over-comes the ouabain inhibition. Thallium, which has a higheraffinity than K+ for Na-K-ATPase, also inhibits aldosteroneproduction. None of these manipulations have a consistenteffect on B. Furthermore, the activity of the Na-K-ATPaseis greater in the outer portion (containing the zona glomeru-losa cells) than the inner portion of the adrenal gland(1.0±0.14 SE VS. 0.44+0.16 aM Pi/mg protein per hr P <0.01). Addition of ACTH in vitro stimulates aldosteroneand B production; ouabain inhibits the aldosterone but notthe B stimulation (% A aldo and B, 24 and 716, withouabain -32, 573). Similar results were found when cyclic3'S' AMPwas added to the media (% A aldo and B, 21, 432,with ouabain - 36, 403). Conclusion: aldosterone produc-tion in the zona glomerulosa cell is sensitive to changes inintracellular K+ while steroidogenesis by the fasciculata-reticularis cells is not. The site of action of intracellularK+ appears to be beyond 3'S' cyclic AMPformation.

13 a

43. The Influence of Inorganic Phosphate (Pi) on RedCell (RBC) Metabolism: In Vitro and In Vivo Studies.MICHAEL C. BRAIN, ROBERT T. CARD,* AND NORMANL.JONES,* Hamilton, Ontario, Canada.

Pathological changes in plasma Pi are known to affectRBC adenosine triphosphate (ATP) and 2,3-diphospho-glycerate (DPG) levels. The effects of Pi concentrationnear to the physiological range have been studied. NormalRBC were incubated at 20% hematocrit in isotonic bicar-bonate-glucose buffer containing 0.5-5.0 mMPi in air and5% CO, at controlled pH 7.40+0.01 for 5 hr. Levels ofRBCglucose-6-phosphate (G-6-P), fructose-i, 6-diphosphate(FDP), triose, DPG, 3-phosphoglycerate (3-PG), glucose,lactate, and ATP were measured every 60 min. After 3 hrincubation at concentrations of Pi between 1.5 and 5.0 mm,total triose (FDP and triose) rose 10-fold, (correlation toPi concentration; r = 0.9), glucose consumption twofold,and DPG less than 5%o. Pyruvate (0.2 mM) addition ledto a rapid sustained fall in total triose and a rise in DPG(mean 0.7 mM/liter RBC), and transient rises in 3-PG andG-6-P. The intravenous infusion of 20-75 mmoles of iso-tonic phosphate (pH 7.4) into male volunteers resulted ina two- to threefold rise in plasma Pi (control 1.0±0.3 mM),

a 10-20% rise in DPG, but no rise in total triose. Sus-tained muscular exercise at 80% of V02 max on a bicycleergometer resulted in an 80% rise in plasma Pi and a 10%rise in DPG. A rise in Pi concentration above 1.5 mmin-fluences RBCmetabolism and DPGlevels, probably throughphosphofructokinase, and may enhance tissue oxygen de-livery and, by shifting hemoglobin-oxygen dissociation, mayaccount for the "anemia of childhood" when lowered levelsof hemoglobin coincide with the hyperphosphatemia ofgrowing children. (Research supported by a grant fromM.R.C. Canada.)

44. Weight Gain and Work Efficiency in Normal Vol-unteers. GEORGEA. BRAY, BRIAN J. WHIPP,* AND SANKARKOYAL,* Torrance, Calif.

This study was designed to investigate overeating and theefficiency of coupling food energy to muscular work. Fournormal males were studied while eating 1800 cal/m' andagain after a 28 day supplement of 4000 cal/day. Subjects per-formed cycle ergometer exercise to steady state at "0,"25, 50, 75, and 100 w and also a step test. Signals from apneumotachograph and rapidly responding O0 and C02 an-alyzers were processed by a digital minicomputer. Minuteventilation (VE), COa production (VC02), 0,2 uptake(V02), respiratory exchange ratio (R), and alveolar 02 andCO, tensions were calculated and displayed breath-by-breath. Weight gain ranged from 3 to 10.8 kg. Assumingthat half the added weight was fat, only 12-44o of totalcaloric intake was stored. Distance walked daily did notincrease during the weight gain. Fecal mass rose 50-100%o(106-173 g/day and 117-275 g/day). V02 at rest increasedslightly. S7O2 was higher at each work rate on the cycleergometer following weight gain in proportion to the re-duced R. Resulting work efficiency (Whipp and Wasser-man. 1969. J. Appl. Physiol. 26: 644.) approximated 30%before and after weight gain. This efficiency was unalteredby ingestion of a 1000 cal breakfast 30 min before cycling.

Vi increased in proportion to VCO2. Changes in thesefunctions after weight gain were attributable to changes ofR. Efficiency during the step test did not deteriorate withweight gain. We conclude that the ingested calories, whichcannot be accounted for by accumulated weight, were notdissipated by decreased efficiency of coupling oxidation tomuscular work.

45. Direct Assessment of Pressures, Flows, and Resist-ances Across Single Glomerular Units of the Rat Kid-ney. B. M. BRENNER,* J. TROY,* R. MACINNES,* AND T.DAUGHARTY,* San Francisco, Calif. (introduced by T. B.Bradley, Jr.).

Pressures, flows, and resistances were determined for ac-cessible glomerular capillaries and single afferent (AA) andefferent (EA) arterioles on the renal surface of adultMunich-Wistar rats. Using a servo-null device, mean pres-sures were recorded in glomerular (PGC) and peritubular(Pc) capillaries. From measurements of nephron GFR andsurface filtration fraction (SFF), glomerular (GPF = GFR/SFF) and EA plasma flows (EAPF = GPF- GFR) werecalculated. From knowledge of the pressure drop (AP)across AA (mean arterial presure, [AP], - Poc) and EA(Paa-Pc) and blood flow through these single vessels(GPF or EAPF/l-Hct), resistances to flow through AA(RA) and EA (RE) were calculated. In 11 hydropenic rats(AP > 150 cmH2O), PGC and Pac/AP averaged 62 cmH2O+2 SE and 0.38±0.01, respectively, indicating a large APacross AA. GPF, and therefore AAPF, averaged 71 nl/min+6. RA averaged 4.5±0.4 X 10's dynes - sec * cm . P0 andEAPF averaged 9.4 cmHs0+0.6 and 47 nl/min+6. At theseflows since AP across EA was less than across AA, REwas likewise less than RA, averaging 2.9±0.2 X 1010 dynes -

sec * cm-, or 40%o of the total arteriolar resistance (RT=RA+ RE). Graded reductions in renal perfusion pressure(to AP 125, 100, <85 cmH20) resulted in progressive de-clines in Pac (59, 52, 46 cmHs0) which were dispropor-tionately less than in AP; hence Pac/AP increased (0.44,0.49, 0.59). This tendency to preserve Pac (and thereforeGFR) resulted from progressive reductions in RA (3.2, 2.4,2.1 X 1010 dynes - sec - cm') while RE remained nearly con-stant (2.6, 2.4, 2.9). Thus, the AAserves as the principalvariable resistor in maintaining Pac in the face of largereductions in perfusion pressure. (Research supported byNIH Grant AM13888 and VA Grant 01/1073.1.)

46. Effects of Elevation of Plasma Sodium Concentra-tion on Tubular Reabsorption of Sodium during Meral-luride or Ethacrynic Acid Diuresis. EMANUELH. BREs-LER* AND KRISTIN T. NIELSEN,* New Orleans, La. (intro-duced by George E. Burch**).

Clearance studies at varying levels of PN. in the hyper-natremic range have shown that as PN. is elevated TNiI isproportionately elevated. Recently, however, dogs given etha-crynic acid and chlorothiazide have been reported not todisplay this relationship, suggesting that this type of glo-merulotubular balance does not hold for the proximal tubule.In this study, 19 hydropenic anesthetized dogs were givenmeralluride or ethacrynic acid after varying degrees of

14 a

stable hypernatremia had been produced by saline infusions.Urine collections were obtained via surgically-introducedureteral catheters. Glomerular filtration (GFR) was deter-mined by either exogenous creatinine or inulin clearance.Clearance periods were selected during the height of diure-sis. The known covariance of TNa and GFR was factoredout by plotting TNa/GFR against PNa. Although the varia-bility in diuretic response tended to obscure the previouslyreported relationship between TN. and PNa found in non-diuretic animals, a significant linear relationship betweenTNa/GFR and PN. remains (P <0.001). In addition, whenTN. was plotted against GFR using different symbols forpoints obtained at PN. < 162 mEq/liter and those obtainedat PN. > 171 mEq/liter two regressions were obtained whichwere found to be significantly different (P <0.001). Whena similar plot was made for T1120 against GFR these tworegressions were not found to be significantly different.These results show that TN. remains proportional to PNa evenduring strong meralluride or ethacrynic acid diuresis.

47. Biologic Action of 1,25-Dihydroxycholecalciferol inUremic Man. ARNOLDS. BRICKMAN,* JACK W. COBURN,*AND ANTHONYW. NORMAN,* Los Angeles and Riverside,Calif. (introduced by Seymour Dayton).

Considerable evidence indicates that 1,25-dihydroxychole-calciferol (1,25-diOHCC) is the active form of cholecalci-ferol (vitamin D3). The kidney is the only known organcapable of producing this hormone from its precurser, 25-hydroxycholecalciferol. Thus, destruction of renal paren-chyma in chronic renal disease might impair the productionof 1,25-diOHCC and be responsible for "vitamin D resist-ance" in uremia. The possibility that exogenously adminis-tered 1,25-diOHCC may bypass this defect and reverse cer-tain abnormalities of calcium metabolism in uremia prompteda study of its effects in uremic man. 1,25-diOHCC was givento three uremic patients in doses of 50-100 U/day (1 U=0.025 ;Lg) for 3-7 days. During metabolic balance study,one anephric patient, dialyzed thrice weekly and with overtsecondary hyperparathyroidism, received 100 U/day for 6days. Serum calcium increased by 5 mg/100 ml and phos-phorus by 1.5 mg/100 ml during treatment. Absorption of "7Caincreased from 24 to 47% (normal range for his dietary cal-cium intake, 17-37%), and mean fecal calcium fell from 740to 550 mg/day. Another patient, with stable chronic uremia,received 100 U/day for 7 days; serum calcium increased from6.4 to 8.5 mg/100 ml and phosphorus from 6.6 to 7.5 mg/100ml; 47Ca absorption was 35%, compared with 18+6% (mean+ SD) in all uremics. A third patient, undergoing regularhemodialysis, showed no response to 50 U/day for 3 days.Thus, extremely small amounts of 1,25-diOHCC may aug-ment calcium absorption and raise serum calcium level inuremic man. The latter may occur, in part, due to mobiliza-tion of bone mineral. These results are consistent with thetheory that impaired production of 1,25-diOHCC in uremiamay account for certain abnormalities of calcium homeo-stasis. This agent may hold future promise in the manage-ment of disordered calcium metabolism in uremia. (Researchsupported by USPHSgrant AM14750 and Contract PH 43-68-1040).

48. Factors Influencing Gelation of Hemoglobin S. ROBINW. BRIEHL, Bronx, N. Y.

Factors which inhibit or enhance gelation of hemoglobinS may reflect precipitating causes of sickle crises and mayalso offer information on gel structure and specific inter-molecular interactions. Hemolysates of blood from homo-zygous S patients were stripped of 2,3-diphosphoglycerate(DPG) on Sephadex G-25, dialyzed into bis Tris [bis (2-hydroxyethyl) iminotris (hydroxymethyl)methane] or potas-sium phosphate buffer, deoxygenated in tonometers with ab-sorption cells and viscometers attached, and studied for vis-cosity as a function of increasing temperature up to thegelation temperature. Deoxygenation was carried out withnitrogen followed by anaerobic addition of small amounts ofsodium dithionite to ensure total absence of oxyhemoglobin.Hemoglobin concentrations were 10-18 mm (heme). Endpoints were sharp, the sol-gel transition occurring whollywithin less than 10. The solutions were then subject to changein DPGconcentration, pH, ionic strength, or buffer concen-tration anaerobically and end points determined again. In-crease in DPGfrom 0 to 6 mmlowered end points by 4° to90 in 0.05 M bis Tris with 0.01 or 0.1 M NaCl. Decrease inpH lowered and increase raised the temperature end points.In potassium phosphate increase in buffer from 0.1 to 1 Mlowered end points. However, increase in NaCl from 0.1 to1 M raised end points. Therefore, DPG, inorganic phosphate,and increasing acidity favor gelation whereas increasing ionicstrength alone inhibits it. Thus the region on the moleculardyad where DPGbinds may be concerned in intermolecularinteractions; and the presence of charge interactions is sup-ported (in addition to hydrophobic interactions). If sym-metry is retained in the helical or rod-like structures of thegel, the molecular dyads lie perpendicular to the cylinderaxis; such symmetry restricts the possible locations of DPGin the gel helix or rod. (Research supported by a grant fromthe American Heart Association.)

49. Neural Control of the Metabolic and Hormonal Re-sponses to Glucopenic Stress in Man. ROBERTG. BRODOWS,*F. XAVIER PI-SUNYER,* DONALDS. SCHALCH,* AND ROBERTG. CAMPBELL,* Rochester, N. Y. and New York. (introducedby Christine Waterhouse**).

The effect of autonomic denervation on the metabolic andhormonal responses during cellular glucopenic-induced stresshas been examined. 2-Deoxy-glucose (2 DG), a competitiveinhibitor of glucose metabolism, was administered intraven-ously to normal volunteers (N = 9) and to two patients, onewith complete cervical cord transection (C-6) and anotherwith idiopathic orthostatic hypotension, anhidrosis, and mildcarbohydrate intolerance. Before, during, and following a 20min infusion of 2 DG (50 mg/kg) plasma concentrations offree fatty acids (FFA), glycerol, lactate, total catechola-mines, glucose, immunoreactive insulin (IRI), human growthhormone (HGH), and cortisol were determined for periodsup to 150 min. In control subjects, the initial elevation ofFFA, glycerol, glucose, HGH, and cortisol correspondedwith the rise in total catecholamines, with maximal levelsattained at 40 min postinfusion; lactate rose at a slower rate,reaching peak levels at 160 min. No change in IRI was noted.In contrast, 2 DG-induced glucopenic stress in the autonomic

15 a

denervated subjects was characterized by no detectable cate-cholamine release or rise in FFA, glycerol, glucose, lactate,or IRI. However, a twofold increase of HGHrelease abovethe mean maximal response of controls was noted in thecord-sectioned patient, while no significant rise of HGHwasseen in the patient with idiopathic orthostatic hypotension.Cortisol responsiveness in both patients was no differentthan controls. Conclusion: In man, the integrity of the auto-nomic nervous system is essential for both the lipolytic andglycogenolytic response to glucopenic-induced stress. It isunlikely that HGHand cortisol together or independentlyplay a significant immediate counter-regulatory role duringperiods of sudden, carbohydrate deprivation. (Research sup-ported by grants from NIH and Nutrition Foundation.)

50. Augmentation of Leukemic Lymphocyte Reactivityto Phytohemagglutinin by Salivary Receptor. JEROME I.BRODY, Philadelphia, Pa.

The purpose of this investigation was to determine whetherthe response of the leukemic lymphocyte to phytohemagglu-tinin (PHA), restricted because of membrane receptor de-ficit, could be augmented by providing additional binding sitesfor the mitogen from an external source. To delineate thefeasibility of such a maneuver, normal peripheral blood lym-phocytes were incubated with serial concentrations of Pro-nase, a sialoglycoprotein-releasing enzyme, until, under con-trolled culture conditions, their Tdr-14C incorporation afterPHA stimulation dropped to 50% of that obtained beforeenzyme treatment, but with cell viability maintained. Lyophi-lized saliva (LS), an acknowledged source of PHA binder,with potency standardized by hemagglutination inhibitionassay, was added to a portion of enzyme-altered lympho-cytes for a set period, the cells washed, and these and theportion not incubated with LS placed in culture with PHA.Duplicate cultures of leukemic lymphocytes from patientswith chronic lymphocytic leukemia were also cultured withPHA, one vial holding plain cells and the other lymphocytespreviously incubated with lyophilate as described. Leukemiclymphocytes never were exposed to Pronase. The diminutionof Tdr-'4C incorporation by normal, enzyme-treated cells wasreversed by incubation of these lymphocytes with lyophilizedsaliva. More significantly, salivary receptor doubled PHAreactivity of leukemic lymphocytes in terms of their abilityto synthesize DNA. These observations indicate that, for thefirst time, a deficient number of PHA receptors on the neo-plastic lymphocyte may be augmented by an external sub-stance with sufficient complementary surface binding to per-mit increased synthetic activity of the leukemic cell. Thismodification of the anomalous leukemic cell membrane mayprovide a logical base for further attempts at making thisinherently immunoincompetent cell more amenable to im-munogenic stimuli and a better participant in host defense.(Research supported by NIH grants.)

51. Hepatic and Splenic Involvement in Lymphoma. S.FRED BRUNK,* HRAIR P. GULESSERIAN,* ROBERTL. GIVLER,*AND ROBERTT. GUTHRIE,* Iowa City, Iowa (introduced byWilliam R. Wilson**).

Attempted cure of lymphoma by radiation therapy requiresaccurate staging. This study was designed to determine the

incidence of hepatic and splenic involvement in lymphomas.130 patients with lymphomas had exploratory laparotomy,splenectomy, and biopsy of liver, bone marrow, and lymphnodes. 88 patients had Hodgkin's disease whereas 42 patientshad non-Hodgkin's lymphoma. Splenic involvement occurredin 4 of 23 patients with clinical stage I Hodgkin's disease,in 13 of 27 with stage II disease, 22 of 35 with stage IIIdisease, and in 1 of 3 patients with stage IV disease. Theliver was not involved in stage I or II Hodgkin's patients;in stage III, 11 of 35 patients had involvement, and in stageIV, 1 of 3 patients had hepatic involvement. In non-Hodg-kin's lymphoma none of the 12 patients with stage I or IIdisease had hepatic or splenic involvement. In stage III, 18of 28 had involvement of the spleen and 7 of 28 had involve-ment of the liver. In stage IV, 1 of 2 patients had involve-ment of the liver and spleen. 32% of Hodgkin's patients and16% of non-Hodgkin's lymphoma patients had their stagechanged as a result of exploratory laparotomy. Whetherstaging by laparotomy is of benefit to patients with stage Ior II disease is uncertain. If the spleen is not removed sur-gically in patients with stage I or II disease, the spleenshould be irradiated along with the nodes in the upper ab-domen. In stage III disease, exploratory laparotomy providesvaluable information about areas of involvement and permitsmore carefully planned radiation therapy. (Research sup-ported by grants from ACS, Iowa Division, Inc., and NIHgrant 5TO1 HE5577-10.)

52. Evidence for Multiphasic Release of Postheparin Lip-olytic Activity. JOHN D. BRUNZELL,* NEALE D. SMITH,*DANIEL PORTE, JR., ANDEDWIN L. BIERMAN, Seattle, Wash.

A group of hypertriglyceridemic subjects with untreatedfasting hyperglycemia has been characterized by normal post-heparin lipolytic activity (PHLA) after a standard dose ofheparin (380 U/m2), but an inability to maintain PHLAlevels during a maximal heparin infusion (4560 U/m2 perhr). PHLA levels during the early phase (0-150') of theinfusion were similar to those of nondiabetic hypertriglyceri-demic subjects (60' peak: diabetic: 0.647±0.206, n = 13; non-diabetic: 0.621±0.198, n = 17; ,uEq FFA/ml per min, X+SD).After 150' there was a dissociation of PHLA levels betweenthe two groups with a lower steady-state PHLA attainedin diabetics (per cent decrease from 60': diabetics: 27±17;nondiabetics: 12±11; P <0.02). The per cent decrease wasdirectly related to fasting glucose in diabetics (r = 0.77,P <0.001). Diabetics recently treated with insulin (1-3 wk,n = 10) had no change in the degree of PHLA decrease(37%±+10) or triglyceride levels, despite improved glucoselevels. When treated for longer periods of time, their TGlevels approached normal. In four such subjects the PHLAresponse was then identical with that of the nondiabetics(per cent decrease: 12+6), suggesting that deficit in laterphase of PHLA release was due to insulin deficiency. Toinvestigate the difference between PHLA released early andlate, paired heparin infusions were given to dogs (n = 3)with and without pretreatment with cycloheximide. No con-sistent difference in PHLA was noted between control andcycloheximide-treated dogs during the early phase of heparininfusion (0-150'), but during the later phase, all cyclohexi-mide-treated dogs reached a lower steady-state PHLA com-

16 a

pared with controls (mean per cent decrease from control:37%). Thus, there appear to be two phases of PHLA re-lease: (a) a late phase presumably dependent on continuedprotein synthesis which is impaired in untreated hypertrigly-ceridemic, hyperglycemic subjects, and (b) an early phasewhich is normal in these subjects and independent of newprotein synthesis. (Research supported by VA and NIH.)

53. Antibody Response to Gonococcal Pili in Patientswith Gonorrhea. T. M. BUCHANAN,* J. SWANSON,*AND E.C. GOTSCHLICH, New York.

Gonococcal pili are filamentous structures present on thesurface of colonial variants Ti and T2 of Neisseria gonor-rhoeae. Variants T1 and T2 have produced experimentalinfections in human volunteers, whereas colonial variants T3and T4 devoid of pili have not. Using chick embryos as ananimal model, we found the same correlation between colonialtype and infectivity. To delineate their immunochemistry andfunction, gonococcal pili were purified by repeated precipita-tion with 0.1 M MgC12 and dissolution in 0.1 M Tris bufferpH 8 containing 0.01 M NaNa. This product contained onlypili when examined by electron microscopy and migrated asa single band in urea polyacrylamide gel electrophoresis.Fluorescein conjugated rabbit antibody against purified pilistained the surface of variants T1 and T2 in all of 18 strainstested. No fluorescence was observed with variants T3 andT4 of gonococci, or with strains of pilated meningococci,nonpathogenic Neisseria spp., and Escherichia coli. An anti-gen-binding assay using 'I-labeled gonococcal pili was de-veloped to detect anti-pili antibody. Sera were tested from12 persons with no history of gonorrhea and from 41 patientswith the following conditions: gonorrhea (28), nasopharyn-geal carriage of meningococci (5), meningococcal meningitis(3), nonspecific urethritis (2), and acute illnesses other thangonorrhea (3). The mean antigen binding by serum fromthe patients with gonorrhea was significantly higher than themean for the controls (P < 0.001) over a wide range ofdilutions. The amount of antigen bound by serum from 22 of28 patients with gonorrhea was higher than the highestamount bound by any of the controls (P <0.001). This testmay prove useful for diagnosing gonorrhea and detectingasymptomatic gonococcal infections. (Research supported bygrants from the NIH.)

54. Renal Tubule Calcium Reabsorption after Parathy-roidectomy. JOHN BUERKERT,* DANIEL MARCUS,* AND REXJAMISON,* Stanford, Calif. (introduced by John Leut-scher**).

Previous studies have shown that parathyroid extract(PTE) enhances calcium reabsorption in the distal nephron.Uncertainty about the nature of PTE caused us to examineparathyroid influence by studying tubule function in 15normal rats and 17 rats 24-48 hr after parathyroidectomy.Samples were obtained by micropuncture: calcium was mea-sured by helium-glow photometry and inulin-14C by scintilla-tion counting. After parathyroidectomy, plasma ultrafilterablecalcium decreased from 3.02±0.06 to 2.00±0.06 mEq/liter,GFR remained unchanged, and urinary excretion of calciumand sodium increased significantly. In normal rats, the proxi-

mal tubule fluid-to-ultrafilterable (TF/UF) calcium was1.06±0.024 (significantly > 1, P = 0.025); TF/UF calciumin the distal tubule was 0.31+0.038. Fractional calcium re-absorption by the proximal tubule was 53.7+2.8% and in thedistal tubule, 93.7±0.05%. After parathyroidectomy, TF/UFcalcium was unchanged in the proximal but below normalin the distal tubule, 0.21+0.02, (P < 0.025). Unexpectedly,fractional calcium reabsorption by the proximal tubule wasstrikingly reduced to 29.3±5.4% (P < 0.001). Sodium re-absorption was similarly diminished from 57.1±0.22% to32.7±0.49% (P < 0.005), although dietary intake had notincreased. Since the filtered load of sodium, FN., unlike Fc.,did not decrease, the calcium-sodium-linked reabsorptivemechanism in the proximal tubule was partially uncoupled.The increment in calcium delivered to Henle's loop in para-thyroidectomized rats was completely reabsorbed, as indi-cated by normal fractional reabsorption in the distal tubule,95.7±0.6%o. These experiments complement previous work bysuggesting the enhanced urinary calcium excretion whichoccurs in the absence of parathyroid hormone reflects dimin-ished calcium reabsorption by the collecting tubule. How-ever they also reveal profound suppression in absolute re-absorption of calcium and sodium by the proximal tubule,and probable enhancement in absolute reabsorption of cal-cium by Henle's loop. (Research supported by grants fromAHAand NIH.)

55. Inhibition of Renin Secretion by Propranolol: aSpecific Treatment for Renal Hypertension? FRITZ R.BUIHLER,* LESLIE BAER,* EDWIN D. VAUGHAN,* HANS R.BRUNNER,* ANDJOHNH. LARAGH, New York.

Renin secretion may be stimulated through an autonomicpathway involving intrarenal beta-receptors. Propranolol hasbeen shown to suppress renin secretion. In this study, pro-pranolol was administered to 32 patients (28 with essentialand 4 with renovascular hypertension), characterized intolow (n = 7), normal (n = 20), and high renin (n = 6) cate-gories. Their mean arterial pressures were 154+18, 134±16and 151+20 mmHg (mean ±SD). 40 to 540 mg propranololwere given daily in 15 studies under metabolic ward condi-tions for 4-6 days and in 19 other studies for up to 10 months.In all studies, propranolol lowered plasma renin activitymarkedly and often to well below the normal range. Thegreatest fall was observed in high renin patients (84%o) ascompared with 60 and 47%o in the other groups. Mean pres-sure fell by 22%o in high as compared with only 12% in nor-mal and 7% in low renin patients. The mean pressure droppedby more than 15 mmHg in 5 of 6 high renin patients asopposed to 11 of 20, and 3 of 7 in the other groups. Irrespec-tive of control renin (two high, two normal) large fall inpressure (mean=23 mmHg) occurred in all renovascularhypertensive patients. Propranolol did not consistently pro-duce commensurate reductions in aldosterone. The data indi-cate a hyperkalemic effect of the drug, operating directly toraise aldosterone and suppress renin secretion. The resultssuggest a relationship between pre-existing renin levels andthe hypotensive effect of propranolol. It may constitute adiagnostic indicator and an effective treatment for hyper-tensive disorders which involve increased renin recretion.Lack of response suggests low renin or a prerenal type of

17 a

hypertension. (Research supported by USPHSGrant l-P17-HL-14148-01.)

56. Enterobacterial CommonAntigen-Induced Lympho-cyte Reactivity in Inflammatory Bowel Disease. DAVIDM. BULL* AND THOMAS F. IGNACZAK,* Columbus, Ohio(introduced by Norton J. Greenberger).

Certain ubiquitous environmental antigens possess antigeniccross-reactivity with host constituents, possibly leading toautoimmune tissue damage. Using hemagglutination-inhibi-tion, we demonstrated an antigenic relationship between en-terobacterial common antigen (ECA) and colonic mucosalextracts. Using ECA as antigen in the macrophage migra-tion inhibition test, we then sought the presence of cellularimmunity to this bacterial component in the inflammatorybowel diseases (IBD), ulcerative colitis (UC), and regionalenteritis (RE). ECA from Salmonella typhimurium was in-cubated in TC199 with purified lymphocytes from IBD pa-tients, patients with other intestinal diseases, and healthysubjects. Supernates from these cultures were placed inmigration chambers with capillary tubes packed with guineapig peritoneal macrophages. Migration inhibitory factor(MIF) activity was expressed as the migration index (MI)where MI = migration in antigen-stimulated supernate/mi-gration in unstimulated supernate. In 15 patients with moder-ately severe IBD (8 UC, 7 RE) the MI was 0.48+0.09(range 0.32-0.61). In eight healthy subjects and in seven ofeight patients with other intestinal diseases, the MI was 0.93+0.10 (range 0.73-1.09). Control lymphocyte cultures notstimulated by ECA, from three patients with severe, life-threatening IBD (two UC, one RE) yielded inhibitory super-nate, suggesting in vivo induction of an inhibitor, possiblyMIF. Migration in TC199 with ECA or in unstimulatedsupernates with ECA added during the migration phase wasnot inhibited. ECA-stimulated IBD lymphocytes did notelaborate soluble cytotoxic factors capable of damaging la-beled-Cr5' chicken erythrocytes. We conclude that ECA-in-duced cellular autoimmunity may induce or perpetuate IBD,with soluble lymphotoxins being less likely as a mechanismof tissue damage than one involving direct cell-cell contact.

57. Disturbance of Lymphocyte Circulation by Granulo-matous Infection. WARD E. BULLOCK,* Lexington, Ky.(introduced by J. W. Hollingsworth).

In this study, the circulation of thoracic duct lymphocytes(TDL) from normal Lewis rats was traced through lym-phoid tissues of rats infected with Mycobacterium leprae-murium. Results suggest impaired TDL recirculation inrats with granulomatous infection, possibly due to entry-block at lymph nodes and splenic white pulp. A mean of1.3 X 106 uridine-OH-labeled TDL/g body weight was givenintravenously to four matched pairs of infected and normalrats. Lymphocytes were then collected from thoracic ductfistulas (TDF) every 8 hr for 3 days. The mean peak(16-24 hr) lymphocyte output (LO) of infected rats wasdecreased (17.7 X 10'/hr) as compared with normals (22.0X 106/hr) and was below normal at all other time intervals.By radioautography, the ratio of labeled lymphocyte output(LLO) from infected vs. normal rats at 24 hr was 0.56.Four pairs of rats were depleted of lymphocytes by TDF

3 days before intravenous injection of labeled TDL. Lymphwas collected every 4 hr thereafter for 48 hr. The meanLO of normals was 5.4 X 106/hr at 0 hr with a peak of17.4 X 106/hr at 16 hr. LO of infected rats at 0 hr was4.4 X 106/hr and increased only to 7.1 X 106/hr at 16 hr. Theratio of LLO from infected vs. normal rats at 16 hr was0.41. Two pairs of lymphocyte-depleted rats were sacrificedat 12 and 24 hr after injection of "1chromium-labeled TDL.Mean radioactivity/100 mg of infected vs. normal spleen at12 and 24 hr was 28.8 and 38.6%, respectively. Radioactivityof infected lymph nodes was 41.1 and 35.4% of normal atthese times. These findings demonstrate a mechanism bywhich granulomatous involvement of lymphoid tissue mayinterfere with the delayed immune response. (Research sup-ported by NIH Grant AI-10094.)

58. Stimulatory Effects of Induced Phagocytosis on Iso-lated Thyroid Cell Function. G. BURKE,* K. KOWALSKI,*D. BABIARZ,* AND S. SATO,* Chicago, Ill. (introduced byEric Reiss**).

Stimulation of endocytosis is a very early effect ofthyrotropin (TSH) on thyroid. However, the relationshipbetween endocytosis and the many other TSH effects onthyroid is not clearly defined. Since phagocytosis in isolatedthyroid cells (ITC) is a model for in vivo endocytosis ofcolloid, we induced phagocytosis in ITC by incubating themat 370C with 0.109 p-diameter polystyrene microbeads;phagocytosis was confirmed in each experiment by electronmicroscopy and/or spectrophotometric analysis of dioxanecell extracts. ITC incubated with 50-100 /A-diameter poly-styrene macrobeads (too large to ingest) served as con-trols. Microbead-induced phagocytosis in ITC was consist-ently accompanied by increases in: (a) cyclic AMP-`Cformation from adenine-8-"C (66%) ; (b) iodide-"'3I trapping(40%); (c) protein and RNA synthesis (30%o); (d)phospholipogenesis (50%); (e) a-aminoisobutyric acid-- _Cuptake (15%). In all instances, save the last, the magnitudeof increase was related to the concentration of microbeadsadded. Polystyrene macrobeads did not influence ITC func-tion in any of these experiments. Aminotriazole, 5 X 10-' M,a peroxidase inhibitor, blocked the stimulatory effect ofmicrobead-induced phagocytosis on phospholipogenesis only.These studies indicate that in ITC, phagocytosis, per se,may alter membrane-bound adenyl cyclase activity. Theresultant increase in cyclic AMP may be a triggeringmechanism for subsequent metabolic changes occurring dur-ing phagocytosis. Since these changes mimic those inducedby TSH, it is suggested that a variety of thyrotropineffects on thyroid may be secondary to stimulation of col-loid resorption. Although increased cyclic AMP formationis believed to precede, and be requisite to, all other TSHeffects on thyroid, it would appear that in this in vitro ITCsystem induction of phagocytosis may be the primary event.(Research supported by NIH Grant AM 11136.)

59. Effect of High Doses of Methylprednisolone on Im-munity in Man. WILLIAM T. BUTLER, ROGERD. RoSSEN,*AND EVAN M. HERSH,* Houston, Tex.

To study the effects of methylprednisolone (MP) onimmune mechanisms in the absence of other immunosup-

18 a

pressive agents or immunologically mediated diseases, wegave 10 normal adult male volunteers 96 mg of MP dailyfor 5 days. IgG concentrations began to decrease shortlyafter MP treatment in nine of the volunteers, reached a lowvalue between 2 and 4 wk, and then began to rise gradually.Mean concentration of IgG before MP was 12.4 mg/ml;by 4 wk, the mean had fallen by 21.8% to 9.7 mg/ml. Infour cases, the IgG decreased by more than 30%. No sig-nificant changes occurred in 10 untreated normal controlsstudied simultaneously. Plasma survival of IgG-'"I wasnormal during MP treatment; a slightly increased catabolicrate occurred for 1-2 wk after MP but was not sufficientto account for the decreased serum IgG. Despite this de-creased IgG synthesis, antibody responses to keyhole limpethemocyanin and diphtheria toxoid given at the start of MPwere normal. Circulating lymphocytes decreased slightlyduring MP treatment; however, the blastogenic response oflymphocytes in vitro was markedly suppressed to phyto-hemagglutinin and to antigens. For example, the mean re-sponse to streptolysin-O before MP was 15,500 cpm/10'cells; 48 hr after starting MP the mean was 1200 cpm/10'cells. Recovery of responsiveness began by the 4th treatmentday; normal responses reappeared within 2 days after stop-ping MP. A short course of high doses of MP has pro-found effects on the immune system; suppression of lympho-cyte reactivity appears early during treatment whereas thenet effect of decreased immunoglobulin synthesis is delayedin appearance and persists for prolonged periods.

60. A Defect in Lipid Mobilization from Adipose Tissuein Diphtheritic Guinea Pigs. DAVID R. CHALLONERANDMARILYNN McKEE,* Indianapolis, Ind.

During metabolic studies on the effects of diphtheria toxinon myocardium we noted that, despite a loss in body weight,the adipose tissue mass in diphtheritic guinea pigs appearedgrossly to be as abundant as that in control animals whichhad gained weight on ad lib. feeding. To investigate this 35,250-g guinea pigs were injected with 2 LD5o'S of diphtheriatoxin and placed on a modified paired feeding regimen sothat the food intake in diphtheritic and control groups wasequal. Over 3 days diphtheritic animals went from 265+6to 218+4 g while control animals went from 272±6 to237+4 g. Despite this, the epididymal fat pad weights were:diphtheritic, 285±+15 vs. control, 220±+12 mg, P = 0.002; andfat pad weight (milligrams)/final body weight (grams) =diphtheritic, 1.32+0.006 vs. control, 0.93±0.04, P <0.001.Greater differences were found on ad lib. feeding. Fat padsfrom control animals yielded 2962±+181 cells/mg of fat paddigested vs. 1790±+182 for diphtheritic animals, P = 0.002,demonstrating that the fat cells were larger in diphtheriticanimals. Blood glucose values remained unchanged whileblood FFA levels decreased to 65%o of those in control ani-mals, P = 0.002. Fat cells were isolated, incubated in vitrowith increasing doses of epinephrine, and glycerol releaseassayed. The response of diphtheritic cells was not differentfrom normal cells. These results suggest a defect in lipidmobilization in vivo in diphtheritic guinea pigs despite acatabolic state. The decrease in blood FFA despite a pre-viously demonstrated decrease in FFA oxidation by isolatedhearts and whole animals also indicates a defect of mobiliza-

tion into the blood FFA pool. The defect does not appearto be in the lipolytic enzymes of the fat cell, suggesting thepossibility of a defect in sympathetic nervous control (neu-ropathy?). This defect in lipid mobilization in the face ofcaloric need could be of pathophysiologic significance in thediphtheritic syndrome and could be subject to therapeuticmanipulation. Whether primary or secondary, the carcassweight loss in diphtheritic animals is at the expense ofcomponents other than fat, most likely muscle. (Supportfrom the NIH, HE 06308 and AM 11805, and the IndianaHeart Association is gratefully acknowledged.)

61. Alteratiops in the Adenylate Cyclase-Cyclic 3'5'-Adenosine Monophosphate (cAMP) System in RatHepatoma and Human Carcinoma of the Colon. REUBENCHAYOTH,* SHELDONEPSTEIN,* AND JAMES B. FIELD, Pitts-burgh, Pa.

Since cAMP may be involved in malignant transforma-tion of cells, the adenylate cyclase system was evaluated inhuman colon carcinoma and ethionine-induced rat hepatoma.Each rat, when possible, provided tissue from malignantnodules, premalignant nodules, and normal liver. In humans,tissue was obtained from the carcinoma and adjacent nor-mal mucosa. Tissues were assayed for cAMP, adenylatecyclase, and phosphodiesterase activities and adenine-3H in-corporation into cAMP-3H. Hepatoma cAMP was sig-nificantly higher than normal liver or premalignant nodules(15.2+2.7, 3.2±0.3, and 3.3±0.2 nmoles/g protein, respec-tively). Malignant nodules incorporated more adenine-'Hinto cAMP-3H (180±-8% of normal liver) than premalig-nant nodules (106±12%). Malignant nodule adenylate cy-clase (395+70 cmp/mg protein) was also significantlyhigher than normal liver (226+15). Phosphodiesterase wassimilar in all three tissues. Although cAMP in colon carci-noma (3.3+0.6 nmoles/g protein) and normal mucosa(4.7+1.3) were similar before incubation, after 20 minincubation in Krebs-Ringer bicarbonate buffer, normal mu-cosa cAMP (40±+15) was significantly higher than in car-cinoma (7±2). Adenylate cyclase in carcinoma and adenine-3H incorporation were significantly reduced to 38 and 40%oof that in normal mucosa, respectively. A benign polyp ina patient with carcinoma gave results similar to normalmucosa. Prostaglandin A1, B1, E1, and Fia increased cAMPand adenine-3H incorporation into cAMP-3H in slices ofcarcinoma and normal mucosa with a greater increase innormals. These studies indicate that differences in theadenylate cyclase system exist between malignant and nor-mal tissues but they are sometimes in opposite directions.(Research supported by grant from NIH.)

62. Increased Number of Stem Cells in Blood of Pa-tients with Myelofibrosis. PAUL A. CHERVENICK,* Pitts-burgh, Pa. (introduced by Dane R. Boggs).

Colonies of granulocytes and mononuclear cells can begrown in vitro from blood and marrow of normal indi-viduals and from patients with various diseases. This reportdescribes the growth of increased numbers of colonies fromthe blood of patients with myelofibrosis. Blood leukocytesfrom five patients with idiopathic myelofibrosis were sus-pended in 1.4% methylcellulose, McCoys 5A culture medium,

19 a

and 15% fetal calf serum. Colony growth was stimulatedby a feeder layer of peripheral leukocytes. After incubationfor 18 days at 370C in 7.5% C02, colonies were countedand removed for morphological identification. Increased num-bers of colonies grew from four of five patients and rangedfrom 80 to 560 per 106 nucleated blood cells plated. Normalsubjects gave rise to 1-22 colonies, and blood from 15patients with nonmalignant hematologic diseases gave riseto 1-50 colonies per 106 nucleated cells. Colony size wassimilar to that observed from normal individuals (50-1500cells) and consistently larger than colonies from blood ofpatients with acute myeloblastic leukemia (maximum 300-400 cells) and most patients with chronic myelocyticleukemia. Morphologically colonies consisted of eithereosinophils, neutrophils, monocytes, or macrophages. Bloodleukocyte counts ranged between 3,600 and 32,000/mm8.Granulocytes capable of proliferation (myeloblasts, pro-myelocytes, myelocytes) ranged between 3 and 37%o. No di-rect correlation was observed between the blood leukocytecount or the number of potentially proliferating granulo-cytes and the number of colonies. This study indicates thatincreased numbers of stem cells circulate in the blood ofpatients with myelofibrosis. Colonies arising from thesepatients have growth characteristics similar to those fromnormal individuals and are larger than colonies from leu-kemic cells. This suggests that idiopathic myelofibrosis maynot be intimately related to myeloid leukemia.

63. Heavy Incorporation of Mevalonic Acid-2-14C intoCholesterol Precursors by Human Platelets In Vitro.PHIN COHENAND ARIE DERKSEN,* Boston, Mass.

Pure human platelet suspensions in TES or Krebs-Ringerbuffer, or autologous plasma, were incubated with mevalonicacid-2-14C. Lipids were extracted and subjected to thin-layer chromatography (TLC) on silica H (hexanes: ether:acetic acid/55: 45: 1) and AgNOB-impregnated silica H(chloroform: acetone/99: 1) along with authentic referencesof commercially available sterols and quinones. As shown byradioautography of TLC separations, quinones were notlabeled, but 1.5-2% of total radioactivity was incorporatedinto eight well-defined bands. More than half of the uptakeappeared in lanosterol and squalene, both identified by TLCand, after purification, by GLC-mass spectrometry. Lano-sterol, and two other compounds, putatively sterols, wererecovered in extracts of digitonin precipitates, as shown byAgNOs TLC-radioautography. However, only trace amountsof label appeared in the cholesterol region in the two TLCsystems. Approximately 10% of incorporated radioactivitymigrated with the fatty acids, suggesting prenoic acid label-ing. Other compounds, including one heavily labeled withRF immediately subjacent to lanosterol or unimpregnatedsilica H, are presently unidentified. Labeling of squalene andthe compound subjacent to lanosterol was considerably in-creased in an oxygen-poor environment. Disruption ofplatelets by sonication, freeze thawing, or detergents, sharplyreduced the number and intensity of all labeled bands. Bycontrast, exposing platelets to hypoosmolar media stronglystimulated uptake of the label, particularly into lanosterol'ssubjacent neighbor, suggesting an architectural requirementfor optimal mevalonic acid incorporation. All known stimu-

lants for cholesterol biosynthesis-oxygen, energy substrates,cations, redox intermediates-in various combinations, didnot increase labeling in the cholesterol region, but did affectlabeling of some other compounds. (Supported by grantHE-13584 from NIH and grant DADA-17-70-C-0083 fromthe U. S. Army.)

64. Relative Importance of Preload and Afterload asDeterminants of Left Ventricular Performance in AcuteMyocardial Infarction. JAY N. COHN, JOSEPH A. FRAN-CIOSA,* MARTIN I. BRODER,* NABIL GuIHA,* AND CONSTAN-TINOS J. LIMAS,* Washington, D. C.

Acute myocardial infarction (AMI) is characterized byincreased left ventricular filling pressure (LVFP), reducedstroke volume (SV), and low, normal, or elevated aorticpressure (AP). SV may be influenced by changes in dias-tolic filling (preload) and AP, which is an importantdeterminant of left ventricular systolic wall tension (after-load). Hemodynamic effects of primary reduction in preload(tourniquets or phlebotomy) and primary reduction inafterload (nitroprusside infusion) was studied in 30 patientswith AMI. During acute reduction in preload LVFP fellfrom an average of 23.6 to 15.6 mmHg (P <0.01) whileSV was reduced from 32.6 to 29.2 ml/M2 (P <0.01) andheart rate (HR) was unchanged (87 to 84 beats/min).Mean AP fell from 100.0 to 91.5 mmHg (P < 0.01). Dur-ing nitroprusside infusion LVFP fell from an average of23.8 to 12.4 mmHg (P <0.01) while SV rose from 24.5to 28.4 ml/M2 (P < 0.02) and HRwas unchanged (95.4 to96.6 beats/min). Mean AP fell from 102.9 to 88.4 mmHg(P < 0.01). Therefore primary decrease in venous returnand primary vasodilatation both reduced LVFP and AP,but SV rose only with the latter. These data indicate thatpharmacological vasodilatation is hemodynamically morefavorable than reduction in venous return in patients withAMI. Furthermore, greater reduction in LV wall tensionduring ejection may indicate lower myocardial oxygen con-sumption during vasodilator infusion. (Supported in part byNIH grant HE 09785.)

65. Mechanism of Steroid Hormone Action: Proof ofthe Messenger RNAHypothesis. JOHN COMSTOCK,*GARYROSENFELD,* ANTHONY MEANS,* AND BERT O'MALLEY,Nashville, Tenn.

Steroid hormones stimulate growth and protein synthesisin target tissues. It has been suggested that these effectsare mediated by messenger RNA (mRNA). However, muchcontroversy has centered on whether steroids act primarilyat the transcriptional or translational level due to lack ofdefinitive evidence on the existence of these hypotheticalmRNA's in animal cells. In the chick model system, estrogencauses quantitative and qualitative changes in oviduct RNAand subsequent synthesis of cell specific proteins such asovalbumin. We now report the isolation of ovalbuminmRNAfrom oviduct. The 12S-17S RNA was extractedfrom polysomes of estrogen-stimulated chick oviduct andtested for mRNAactivity in a cell-free system containingribosomes and translational factors from rabbit reticulo-cytes. Before addition of the chick RNA, valine-14C is in-corporated almost entirely into hemoglobin. Addition of chick

20 a

RNA results in appearance of a 14C protein peak coincidentwith authentic ovalbumin in SDS-acrylamide gel electro-phoresis. Synthesis of ovalbumin-14C showed linear depend-ence on exogenous oviduct mRNA. Oviduct ribosomal RNAor RNA from liver demonstrated no capacity to direct invitro synthesis of ovalbumin-1'C. Hen oviduct containedlarge quantities of ovalbumin mRNA, but RNA from im-mature chicks contained no mRNAfor ovalbumin; how-ever, mRNAcould be induced in immature chicks withestrogen. Finally, estrogen withdrawal resulted in disappear-ance of chick ovalbumin mRNA. Estrogen readministrationinduced marked accumulation (> 10-fold) of ovalbuminmRNA. These data represent an initial isolation and trans-lation of specific mRNAfrom eucaryotic hormone-dependenttissue and document its presence to be directly dependent onprior hormonal stimulation. It also substantiates the hy-pothesis that steroid hormones act in target cells to induceprotein synthesis by first promoting accumulation of mRNA.

66. The Role of Sialic Acid in the Phagocytic Activityof Blood Neutrophils. ANDREAS CONSTANTOPOULOS*ANDVICTOR A. NAJJAR,** Boston, Mass.

Stimulation of the phagocytic activity of blood neutro-phils has been obtained by an autologous cytophilic 'y-globulin fraction that binds specifically to polymorphonu-clear leukocytes (1967. Biochemistry. 6: 3386). The wholestimulatory activity resides in a small peptide, tuftsin. Tuft-sin is cleaved off the parent -y-globulin molecule by aspecific enzyme leukokininase present on the outer surfaceof the cell membrane (1970. Nature [London]. 228: 672).It has been identified as Thr-Lys-Pro-Arg and synthesized.Its activity in vitro is expressed in hormone-like quantitiesand is composed in the process. It appears to be synthesizedin the spleen since it is completely absent in splenectomizeddogs and humans. Sialic acid is indispensible for its bio-logical activity. Treatment of the cells with neuraminidasefrom either Vibrio cholerae or C. perfringens purified byaffinity chromatography (1971. Biochem. Biophys. Res.Commun. 44: 178) inhibits completely the stimulatory effectexerted by tuftsin and its parent molecule, the specific leuko-philic y-globulin. Treatment of the cells with influenza virusneuraminidase resulted only in a partial inhibition. The latterenzyme is specific for the 2-3' and two bacterial enzymescleave both 2-3' and 2-6' linkages. It can be assumed there-fore, that membrane-bound sialic acid of neutrophils containsboth linkages and is concerned with receptor site binding oftuftsin, whether free as the tetrapeptide or bound to theparent carrier specific 'y-globulin. (Supported by USPHSgrant AI-09116.)

67. Definition of the Essential Red Cell Membrane. RICH-ARD A. CoopER,* Philadelphia, Pa. (introduced by A. S.Relman**).

Red cell membrane free-cholesterol (FC) exchanges withplasma lipoprotein FC. It increases under certain pathologicconditions in vivo, particularly liver disease, and decreaseswhen lipoprotein FC is esterified in vitro. Yet no simplerelationship exists between plasma FC concentration andmembrane FC content. However, FC exists with phospho-lipids (PL) and protein (Prot) in membranes and plasma

lipoproteins. Interrelationships between these constituentswere therefore studied. Membrane FC/PL mole ratios wereincreased to 0.92-1.41 in 22 patients with severe liver disease(14 with "target" and 8 with "spur" cells), but were 0.87+0.02 in 13 normals. FC/PL and FC/Prot of low densitylipoprotein (LDL) and high density lipoprotein (HDL)were also increased in patients. Membrane FC/PL corre-lated well with the FC/PL of whole serum (r=0.78) andof LDL (r=0.86), and it correlated closely with the FC/Prot of LDL (r = 0.94). Lipoproteins and red cells weremore saturated with FC in patients with spur than with tar-get cells; however, these data formed a continuum yieldingsimilar regression lines. Extrapolation of these regressionlines to zero values for the FC of serum and LDL interceptedmembrane FC/PL mole ratios of 0.63, 0.48, and 0.62. In-creases in normal membrane FC/PL induced by incubationin FC-enriched normal serum corresponded to values pre-dicted from these regression analyses. Decreases due to FCesterification in vitro followed a regression line which inter-cepted a membrane FC/PL mole ratio of 0.50 at zero serumFC. Thus, the FC saturation of plasma lipoproteins, as de-fined by their FC/PL or FC/Prot ratios, relates directly tothe FC/PL mole ratio of red cell membranes. Moreover, amembrane FC/PL mole ratio of approximately 0.55 definesthe lipid composition of the "essential membrane"; valuesgreater than this reflect the FC saturation of plasma lipo-proteins.

68. A "Proliferation Inhibitory Factor" (PIF) Producedby Human Blood Lymphocytes. S. R. COOPERBAND,J. R.GREEN,* AND A. M. BADGER,* Boston, Mass.

Circulating lymphocytes are involved in the production ofdelayed hypersensitivity, graft rejection, and immune sur-veillance against the growth of neoplastic cells. It is prob-able that these effects are mediated, at least partly, by anumber of humoral effector factors. We have observed theproduction of a new factor (PIF) by human blood lympho-cytes in culture which acts to inhibit the growth of non-lymphoid cells and may be involved in immune surveillance.PIF is produced by mitogen-(PHA, Concanavalin A) andantigen-(mixed leukocyte cultures, PPD) stimulated cells.It may be produced by purified lymphocytes, and is detect-able 3 hr after stimulation, although it accumulates during2-3 days in culture. The factor upon a variety of humancell types, in culture, to inhibit DNA synthesis, mitotic in-dex, new cell production, and replication of individuallycloned cells without cytotoxicity. Wegenerally assay PIF inunfractionated culture supernatants at dilutions of 1:20-1: 100, but we have observed activity in dilutions of 1: 1000.Different cells demonstrate different sensitivity to PIF-embryonic and tumor cells generally are most sensitive, whileadult primary cells (i.e., kidney) are insensitive. The PIFappears to be species specific, and human PIF is not activeagainst chick or murine cells. It did not appear to inhibitproliferation of autologous lymphocytes stimulated by PHA.Wehave examined the mechanism of action of PIF on clonedHeLa cells which are in synchronous division. The PIF doesnot influence cells when added after the onset of the G1 phaseof cell division and the next cell division occurs normally.When PIF is present during mitosis, cells become arrested in

21 a

the following G1 phase. This effect occurs even when PIF isadded for only a few hours during mitosis and early G1, andthen removed. The PIF is heat stable up to 850C for 30 min,nondialyzable, nonsedimentable at 90,000 g and is destroyedby trypsin. (Research supported by grants from NIH.)

69. Thrombokinetic Studies in Acute Leukemia. DALE H.COWAN,* Cleveland, Ohio (introduced by John W. Harris**).

Thrombocytopenia commonly occurs during the course ofacute leukemia and is generally attributed to decreased plate-let production resulting from marrow invasion by leukemiccells. No studies using quantitative techniques have beenpublished to substantiate this contention. To evaluate throm-bokinetics in acute leukemia, platelet production and survivalwere measured in untreated patients with this disorder. In ninepatients (platelets: 14,000-325,000/,ul) the survival of auto-logous 5"Cr-labeled platelets ranged from 0.5 to 5.9 days, re-covery of labeled platelets was 22-45%, and platelet turnoverwas 13,900-334,800 platelets/Al per day (normal: 8.5+0.2days, 62+3%, 49,200+3600 platelets/Al per day, respectively).In five of the nine, turnover was greater than normal. In allnine patients, labeled platelets disappeared exponentially; innormals the pattern of disappearance is linear. Using Hark-er's method, total megakaryocyte mass in the patients withleukemia ranged from 0.4 to 61.3 X 1010 A3/kg (median 9.3X 1010 A3/kg) (mean of nine normals: 5.1+1.0 X 1010 8/kg).In six patients, the total mass exceeded normal. In five pa-tients, the numbers of megakaryocytes were increased, rang-ing from 35 to 408 X 106/kg (normals: 13.0+4.3 X 106/kg)and cell volumes were decreased, ranging from 1503 to 2637A (normal: 3982+253 Is). These changes are similar tothose seen in megaloblastic anemias. In four of these five andin one other patient, total production, determined fromtotal megakaryocyte mass, exceeded effective production, de-termined from platelet turnover. Serum B12 and erythrocytefolate levels were normal in each. The data indicate thattotal thrombopoiesis is generally not subnormal in patientswith acute leukemia. Rather, it often exceeds normal. Throm-bocytolysis is present and is often uncompensated despite in-creased total thrombopoiesis. One mechanism for the un-compensated thrombocytolysis is ineffective thrombopoiesisoccurring in the absence of cellular folate or B12 deficiency.(Research supported by Grant T-548 from ACS and GrantCA 12399 from NIH.)

70. Communication between Normal and Enzyme-De-ficient Cells in Tissue Culture. RODYP. COX, MARJORIE R.KRAUSS,* M. E. BALIS,* AND JOSEPH DANCIS,* New York.

Correction of certain mutant phenotypes by intimate con-tact with normal cells-i.e. "metabolic cooperation"-is aneasily studied form of cell communication. Metabolic coopera-tion between normal cells and mutant cells deficient in hy-poxanthine-guanine or adenine phosphoribosyl transferase(HGPRTase and APRTase, respectively) appears to be theresult of transfer of the enzyme product, nucleotide, or nu-cleotide derivative, from normal to mutant cells. This processshows selectivity in that mutant derivatives of mouse L cellsare unable to function as recipients of HGPRTaseor APRT-ase products, while hamster and human fibroblasts with theseenzyme deficiencies exhibit correction of the mutant pheno-

type, when in contact with normal donor cells. There is alsoselectivity with respect to substances transferred, since othermutant phenotypes, i.e. G-6-PD deficiency, are not correctedby contact with normal cells. Species specificities do not ap-pear to influence metabolic cooperation, therefore hetero-specific cell mixtures provide an opportunity to cytologicallydistinguish cells and study individual cell interactions. Trans-fer of nucleotide from normal to mutant cells is less depend-ent on energy production than is the incorporation of radio-active purines into cellular material. The nucleotide trans-location mechanism is not susceptible to sulfhydryl blockingagents. (Research supported by grants from NIH.)

71. Partial Gonadal Mosaicism and Genetic Counselingin X-Linked Lethal Recessives. DANIEL CRAMER,* R. J.KRYSCIOR,* E. R. PIERCE,* AND EDMONDA. MURPHY,**Bethesda and Baltimore, Md.

Well-known theory shows that one-third of all X-linkedrecessive lethals are new mutations. Since mutations can arisewherever copying of DNAoccurs, the proportion of affectedova may vary randomly as a branching process. A convenientmodel is perhaps that proposed and solved approximately byLea and Coulson (J. Genet. 1949. 49: 264) for the propaga-tion of mutant bacterial strains on a culture medium. Moreexact methods are here described for finding (a) the individ-ual probabilities and (b) the moment-generating functionand hence the moments of the distribution of the number ofmutant cells. These probabilities are the prior distributionof the individual proportions and are modified in the usualway by the posterior information to give a posterior distribu-tion of proportions of mutant cells. From these it is easy tocompute the risk that the next son will be affected. Biologi-cal data suggest that the ova are a randon sample of about300,000 cells from a total of about 5,000,000. A mutationprobability per mitosis is chosen such that the mean numberof mutant cells accords with population estimates of the mu-tation rate per generation. From these values it is shownthat partial gonadal mosaicism has only a small effect on theclassical risk of recurrence (Hum. Heredity. 1969. 19: 126).The methods are illustrated from actual data.

72. Differential Effects of Exogenous Free Fatty Acids(FFA) on Catecholamine-Induced Changes in Lipolysisand Glycogenolysis in the Perfused Working Rat Heart.M. F. CRASS III* ANDJ. C. SHIPP, Omaha, Nebr.

The purpose of this study was to examine interactions ofcatecholamines and exogenous substrates (glucose, FFA) oncardiac lipolysis and glycogenolysis. Hearts of fed male ratswere labeled in vivo with palmitate-1-"4C, and preperfusedfor 5 min followed by a 30 min recirculated perfusion in aclosed working system at 10 cm H20 left atrial filling pres-sure. Perfusate was oxygenated Krebs-Henseleit bicarbonatebuffer containing 5.5 mmglucose with or without palmitate(3%o albumin). Epinephrine (E) or norepinephrine (NE)was added after 2 min of recirculated perfusion. Tension andrate changes were monitored. Tissue glycogen, content and4C labeling of triglycerides, and "CO2 were determined. In

the absence of added catecholamines, exogenous FFA, butnot glucose, inhibited triglyceride fatty acid oxidation andglycogenolysis. E and NE, 8.2 X 10' to 1.3 X 10' M, pro-

22 a

duced concentration-dependent increases in triglyceride mo-bilization, triglyceride fatty acid (TGFA) oxidation, andglycogenolysis, without affecting content or labeling of tissuephospholipids in hearts perfused with glucose alone. Maxi-mal stimulation with E and NE was observed at 4 X 10-7 M.With added palmitate (0.6 mM), the E-stimulated lipolysisand TGFA oxidation was abolished, though the full glyco-genolytic effect of E persisted. These findings, under condi-tions which approximate those in vivo, indicated that cate-cholamines and exogenous FFA exert different regulatoryeffects on triglyceride and glycogen metabolism of heartmuscle. (Supported by grants from NIH [No. AM-14986]and Nebraska Heart Association.)

73. a-Galactosidase Defect in Fabry's Disease. SpecificInhibition of One Enzyme by Myoinositol. JOHNC. CRAW-HALL AND MARIANNABANFAVLI,* Montreal, Canada.

Fabry's disease has been characterized by a deficiency ofthe enzyme ceramide trihexosidase. a-Galactosidase activity inleukocytes and cultured skin fibroblasts is also reduced toabout 20% of normal control values. This residual activitycould be accounted for either by reduction of the normalenzyme quantity, a structural enzyme variant, or as a re-sidual isoenzyme. Wehave studied these possibilities by useof myoinositol as a specific inhibitor of a-galactosidase using4-methylumbelliferyl-a-galactoside as substrate. Initial ex-periment showed that myoinositol was an effective competi-tive inhibitor of the normally occurring enzyme but was notinhibitory to the enzyme found in mutant cells. It is knownthat there are differences between the properties of mutantand normal cell enzymes in that the former has a higher Kmand is not heat labile compared with the latter. We havecombined this information together in an experimental de-sign to measure myoinositol inhibition of normal cell a-gala-ctosidase before and after heat inactivation at 510C for 75min and the following changes were observed. The specificactivity of the enzyme was rapidly reduced at first but thelast 10%o of activity remained relatively thermostable. TheKmof the enzyme changed from 3 mmto 19 mmand myoinosi-tol which was originally an inhibitor of the enzyme becamenoninhibitory. This data indicated that two a-galactosidaseisoenzymes are present in normal cultured fibroblasts. Themore active of these has the lower Ki, increased heat lability,is inhibited by myoinositol, and is deficient in the cells ob-tained from patients with Fabry's disease. The other, whichis still present in cells obtained from patients with Fabry'sdisease has a higher Ki, minimal heat lability, and is notinhibited by myoinositol.

74. Role of Growth Hormone (GH) in the Alterationsof Carbohydrate Metabolism Induced by Oral Contra-ceptive Agents (OCA). MAYERB. DAVIDSON* AND GERALDB. HOLZMAN,* Los Angeles, Calif. (introduced by J. Brown).

The mechanism whereby OCA cause glucose intoleranceis unknown. Metabolic responses to tolbutamide were uti-lized to study this problem since glucose, GH, insulin, andfree fatty acids (FFA) are all affected by this single agent.13 normal subjects and 8 hypopituitary patients on replace-ment medication were tested before starting and 3 and 6

months after OCA were begun. In normal subjects afterOCA, not only were fasting GH levels elevated, but aftertolbutamide, there was a significantly blunted fall of plasmaglucose, an increased GHresponse, and a heightened reboundof FFA. As expected, plasma cortisol was markedly elevatedby OCA. Although unchanged by OCA, insulin responsesduring the three tests were significantly correlated in thesame subject. Therefore, insofar as the tolbutamide-stimu-lated response reflects release of stored hormone, the quickly,releasable insulin pool remains constant in a single individualand is unaffected by OCA. In contrast to normals, the glu-cose, FFA, and insulin responses to tolbutamide were un-changed by OCA in the hypopituitary patients. Fasting GHlevels were mostly undetectable by our assay and there wereonly minimal hypoglycemia-induced rises of GH in any ofthe tests. Plasma cortisol was also markedly elevated byOCA in these pituitary-deficient patients. Five additionalnormal females on OCAhad an increased GH (mean+SEM)response to a moderate amount of exercise (two flights ofstairs) compared to five women not on OCA (16.4+3.5 vs.5.0±+1.9, P <0.025). This suggests that subjects on OCAare exposed to high levels of GH throughout the day. It isconcluded that estrogen stimulation of GH secretion is themechanism whereby OCAcauses glucose intolerance.

75. Complement Abnormalities and Renal Vascular In-jury. N. K. DAY,* H. GEIGER,* R. MCLEAN,* A. MICHAEL,AND R. A. GooD,** Minneapolis, Minn.

Complement component deficiencies have been associatedwith renal-vascular disease. The presumption has been thaithe C abnormalities and renal-vascular pathology are relatedbecause C components are being utilized in immunologicalprocesses injurious to the glomerulus (e.g., serum sickness).Early we encountered three distinct selective deficiencies ofC components associated with progressive renal disease andwe can now link hereditary deficiency of Clr, Cls, C2 toincreased susceptibility to infection, to renal-vascular dis-ease, and to lupus erythematosus. An additional experiencewith a patient suffering from recurrent hematuria and glo-merulitis after infection seems most provocative in this per-spective. Histologically the patient's renal lesion comprisedfocal glomerulonephritis with hyalinization and glomerulardeposition of #lC. The serum showed depression of hemo-lytic C3, properdin, and conversion of C3 at 0°C; C3PA(Miiller-Eberhard) was reduced in fresh and incubatedserum. Total hemolytic C in fresh serum is high and dropssharply at 0°C. C activation is dependent on Mg'+ and noton Ca++ and appears to involve primarily activation of C3but to a lesser extent C5 and C7. Cl, C4, C2, C6, C8, C9, aswell as Clq and Cls, remain normal by immunochemical and/or functional assay. The vigorous activation of the alternateC pathway observed in vitro at low temperatures is prob-ably occurring in vivo as well (cf. C3PA) to account forthe episodes of renal vascular disease. Taken together, theseseveral clinical associations suggest that the integrity of theclassical C pathway may provide not only a defense againstinfection but a major defense against excessive activation ofthe alternate or properdin pathway which when activatedexcessively may lead to serious renal-vascular injury. (Aidedby NIH.)

23 a

76. Cyanate-s"C: a New Method for Red Cell Survivals.FRANK G. DE FURIA,* JOSEPH H. GRAzIANO,* AND ANTH-ONY CERAMI,* New York (introduced by Zanvil Cohn**).

The reversible binding of 51Cr and the hydrolysis of thephosphate bond formed with 'H- or 8P-labeled diisopropyl-fluorophosphate (DFP) has led to difficulties in determiningthe in vivo survival time of erythrocytes. The recent findingthat low concentrations of cyanate (-NCO) react specificallywith the amino terminal valine of the hemoglobin moleculein an irreversible carbamylation reaction has led us to de-velop a cyanate-"C method for determining red cell survival.30 ml of blood anticoagulated with acid citrate dextrosesolution B was divided into two portions and incubated for1 hr at 37'C with either 25 AtCi of Na25CrO4 (SA 30.2mCi/g) (Chromitope) or KN'CO (SA 80 mCi/g). Thecells were washed three times, recombined, suspended insaline, and returned to the donor animal. The blood sampletaken 24 hr after reinfusion was considered to be day 0for both methods. The average apparent half-survival time(tj) by 51Cr and 'C methods, respectively, were: fouradult female rhesus monkeys, 15.6 and 19.3 days; four adultfemale beagles, 19.2 and 27.5 days; an anemic Basenji dog(hereditary pyruvate kinase deficiency), 1.5 and 3.4 days.In contrast 'Cr data, the 14C data yielded a straight linewhen plotted on linear coordinates, further emphasizing theirreversible nature of the carbamylation. Since -NCO reactsselectively with RBC's in vivo, intravenous administration ofthe label is also feasible. 100 ,uCi of cyanate was administeredintravenously to an adult female beagle and the anemicBasenji with a resultant tj of 22.3 and 5.2 days, respec-tively. The greater tj of the Basenji observed when cya-nate-"C was administered in vivo probably reflects theadvantage of labeling fragile RBC in vivo. (Supported inpart by NIH Contract Grant 71-2371; USPHS GrantAM-14691, and grants from the Rockefeller Foundationand the Children's Blood Foundation.)

77. The Prognostic Implication of Glomerular Depositsin Systemic Lupus Erythematosus. MARTIN DILLARD,*ISABELLE DUJOVNE,* CONRADL. PIRANI,* AND VICTOR E.POLLAK, Chicago, Ill.

In patients with systemic lupus erythematosus (SLE) itis now thought that glomerular lesions result from deposi-tion of circulating immune complexes. Deposits are assumedto occur mainly in the subepithelial position, and y-globulin,with anti-DNA activity, has been eluted presumably fromthis site. To assess the relationship and character of de-posits, and severity and activity of renal disease, 40 renalbiopsies from 24 patients were investigated. The extent andsite of deposits was estimated semiquantitatively (0-4+) inlow power electron micrographs of glomeruli. The plasmaproteins in the deposits was assessed in 0.5 I frozen substi-tuted sections, using monospecific antisera against heavychains (,y, a, ,u), light chains (K, X), 61C globulin, fibrino-gen, albumin, and transferrin. Later, sections were stainedwith periodic acid-Schiff to facilitate precise localization.Deposits were not distributed randomly in the glomeruli.There was a positive relationship between: (a) the extentof subendothelial and mesangial deposits, (b) the extent ofsubepithelial and intramembranous deposits. There was a

striking inverse relationship between the extent of sub-endothelial and subepithelial deposits. Subepithelial depositscontained IgG, IgM, and 61C globulin. With extensive sub-epithelial and intramembranous deposits, the clinical coursewas benign, and serum f1C globulin levels were normal.Serial studies revealed persistence of these deposits foryears. Subendothelial deposits contained all plasma proteinsincluding immune globulins, fi globulin, fibrinogen, albumin,and transferrin. In the presence of subendothelial depositsserum 81C globulin levels were low. Extensive subendo-thelial deposits occurred only with histologically activelesions, and were predictive of death within weeks. Sub-endothelial deposits disappeared with treatment in survivors.These studies indicate an important predictive role forelectron microscopy and immunohistologic observations inpatients with lupus nephritis. (Supported by NIH GrantAM-10314.)

78. Sensitivity of Methods for the Detection of Anti-DNAAntibodies. CAROLEDORSCH*AND EUGENEBARNETT,Los Angeles, Calif.

The relative sensitivity of four methods for the detectionof antinuclear antibodies has been examined. 41 heat-inacti-vated sera were selected on the basis of positive indirectimmunofluorescent antinuclear antibody tests (ANA). Thediagnoses included systemic lupus erythematosus (SLE)(29), rheumatoid arthritis (6), and miscellaneous diseases(6). 23 sera had precipitating antibodies to double strand(ds) and/or heat-denatured, single strand (ss) calf thymusDNA by counterimmunoelectrophoresis (CIE) in 1%o aga-rose in barbital buffer, pH 8.2. 8 of these 23 sera alsoprecipitated with DNA by Ouchterlony double immuno-diffusion in 0.4% agarose in phosphate-buffered saline(PBS), pH 7.4. 32 of the 41 -sera showed positive DNAbinding (> 25%) by the ammonium sulfate precipitationtechnique of Farr, using ds-calf thymus DNA-labeled withactinomycin D-8H. These included 21 of the 23 sera con-taining precipitins by CIE. Of the 32 positive-binding sera,all 10 with valtes > 60% had precipitating antibodies inCIE. 22 sera had DNA-binding values between 25%o and60%; only 11 of these were positive by CIE, while 11contained no demonstrable precipitins to ss- or ds-DNA.To explain differences in precipitability with DNA of serawith comparable DNA-binding values, differences in immu-noglobulin class, quantity, affinity, and specificity wereevaluated. CIE represents a method for detecting substancesprecipitating with DNAwhich is more rapid, requires lessserum, and is more sensitive than the more commonlyemployed double immunodiffusion in PBS. Neither of thesemethods detects DNA antibodies in as many sera as doprimary binding assays with radio-labeled DNA. Neitherhas the same clinical discriminatory value as ANA testswhich, when negative, essentially exclude the diagnosis ofSLE. (Supported in part by USPHSGrant GM15759.)

79. In Vitro Synthesis of DNAComponents of HumanGenes for Globin: Use as a Probe for Messenger RNA(mRNA). L. W. Dow,* D. KACIAN,* M. TERADA,* S.METAFORA,* S. SPIEGELMAN,* P. A. MARKS, AND A. BANK,New York.

24 a

Purification of human mRNAfrom polyribosomes ofhemolytic anemia reticulocytes (retics) and purified RNA-dependent DNA polymerase (reverse transcriptase) fromavian myeloblastosis virus permitted the in vitro synthesisof DNA-8H. This DNA-'H product has a size correspond-ing to 8.3S (neutral sucrose gradient), and hybridizes rap-idly with human globin mRNA, but not with ribosomal orviral RNA. The DNA-8H-human globin mRNAhybridbands at 1.55 g/cm! in CS2SO4, the expected density for atrue DNA-RNA hybrid in which both polynucleotides areof about the same length. Purified DNA-'H product bandsat 1.47 g/cm3 in Cs2SO4: the density of a DNA strandwith approximately the mol per cent G+C (50%) contentexpected for globin mRNA. These data suggest that theDNA synthesized in vitro approaches the size predicted fora human globin structural gene and has a base sequencecomplementary to human globin mRNA. The human lOSisolated from polyribosomes of normal retics directs syn-thesis of equal amounts of normal a- and ,B-globin in aKrebs cell-free system. 8-thal retic mRNAdirects synthesisof a-chains fivefold that of p-chains in the Krebs cell-freesystem. The decreased formation of f8-chains in thal couldbe due to deficient or defective mRNAfor 8-chains. Thein vitro synthesized DNA-'H provides a probe to distinguishthese alternatives. The DNA-3H product of normal humanglobin mRNAhybridizes with thal globin mRNA. If thereis decreased mRNAfor fl-globin in thal, higher concen-trations of thal globin mRNAwill be required for satura-tion of the DNAcompared to that with a given concentra-tion of normal globin mRNA. (Supported by NIH, NationalScience Foundation, ACS, and Cooley's Foundation grants.)

80. Ultrastructural Changes Associated with PlateletNucleotide and Calcium Release. MICHAEL J. DROLLER,*Bethesda, Md. (introduced by Robert S. Gordon, Jr.).

Platelet activity is thought to be mediated by the secre-tion, or "release," of calcium and nucleotides from plateletsin response to various aggregating agents. The present studydescribes the morphological events that occur during suchnucleotide/Ca++ release. Washed human platelets were in-cubated at 37°C for 10 min with either thrombin, collagen,ADP, or epinephrine, and fixed in 3% glutaraldehyde and1% OsO4 for electron microscopy. At 37°C, where maximalrelease of nucleotides/Ca++ occurs in response to thrombinand collagen, alpha granules, thought to contain the sub-stances released, almost completely disappear. A complexlamellar-vacuolar system, continuous with the cell mem-brane and probably the pathway for outward movement ofnucleotides, now dominates platelet ultrastructure. Wisps ofelectron-dense material, comprising fibers 50-80 A in di-ameter with an occasional 180-220 A periodicity, are con-spicuous in the lamellar spaces. The wisps are morphologi-cally similar to fibrin precipitated either in vivo or in vitro.They are seen even after only 5 sec of incubation, whenmany alpha granules have not yet been released. Epinephrineand ADP are less effective in stimulating release and itsassociated morphological changes. Prostaglandin E. and theo-phylline, inhibitors of aggregation, inhibit nucleotide/CaZrelease and many of the accompanying ultrastructural events.Release is not observed in platelets incubated at 4°C, or at

37°C without any of these agents; such platelets undergonone of the morphological changes that characterize re-lease. The differences thus observed between the effects ofvarious agents on the platelet release reaction establish apattern of events that undoubtedly play an integral role insubsequent clot formation.

81. Interaction of Glycoprotein Hormones with Agarose-Concanavalin A. MARIA L. DUFAU,* TsuNEo TSURUHARA,*AND KEVIN J. CATT,* Bethesda, Md. (introduced by M. B.Lipsett**).

The ability of agarose-coupled Concanavalin A to bindcarbohydrate and glycoprotein molecules has been applied tochromatography of glycoprotein hormones. The carbohydratebinding activity of Concanavalin A is mainly directed towardD-gluco- and D-mannopyranosyl residues, and interactionwith polysaccharides has been ascribed to the terminala-linked nonreducing portion of the molecule. However,HCGand subunits bearing predominantly sialic acid ter-minal residues were bound strongly by Sepharose-Concana-valin A, and could be displaced by 0.2 M methyl a-D-glucopyranoside. Desialylated HCGwas eluted more slowlythan the intact hormone, and asialo-agalacto-HCG was notreadily dissociable from combination with Concanavalin A.While the binding of unmodified HCGby Concanavalin Ais largely attributable to the substantial mannose contentof the molecule (16 residues/mole), the presence of terminalsialic acid and galactose residues significantly influences thebinding affinity of HCGfor Concanavalin A. Human lutein-izing hormone and follicular-stimulating hormone were alsoabsorbed by Sepharose-Concanavalin A columns, and elutedwith methyl a-D-glucopyranoside. Prior combination withspecific antibody did not prevent HCG-'I binding to Con-canavalin A, and subsequent elution with methyl gluco-pyranoside released the antigen-antibody complex in highyield. Chromatography of HCGon Sepharose-ConcanavalinA was applied to: (a) preparation of "I-labeled HCGwithimproved binding affinity for antibody and gonadal receptorsites, (b) purification and isolation of HCG from crudeurine extracts, (c) comparison of plasma and urinary HCG,revealing an elution profile consistent with partial desialyla-tion of the urinary hormone. These experiments have shownthat the group-specific interactions of glycoprotein hormoneswith agarose-Concanavalin A are of considerable value forpurification and isolation of such hormones and their deriva-tives, and for the preparation of radioiodinated glycoproteinhormones of improved binding characteristics.

82. Experimental Streptococcus viridans Endocarditis:Alternatives to Penicillin in Prophylaxis. DAvI DURACK*AND ROBERTPiTERSDORF, Oxford, England.

Antimicrobials other than penicillin are often prescribedfor patients with valvular heart disease undergoing dentalor other surgical procedures in an attempt to prevent bac-terial endocarditis (BE). If patients cannot tolerate peni-cillin or if bacteremia with resistant organisms is antici-pated, a variety of antibiotics has been recommended. Thisapproach is empiric at best because there is no clinical orexperimental evidence to document its efficacy. The devel-opment of a model in which BE regularly follows one

25 a

intravenous injection of Streptococcus viridans into rabbitswith pre-existing sterile endocardial lesions permitted asystematic examination of this problem. Antibiotics weregiven 30 min before or after an intravenous injection ofapproximately 108 S. viridans into rabbits with sterile endo-carditis in doses calculated to correspond with those usedin man. After 24 hr the animals were killed and the vegeta-tions cultured quantitatively. The organism was fully sensi-tive to all drugs and serum levels greatly exceeded minimuminhibitory concentrations. Persisting infection with S. viri-dans was found despite administration of cephaloridine,cephalexin, erythromycin, tetracycline, and cotrimoxazole.However, vancomycin was successful in sterilizing the vege-tations in all animals. If treatment was delayed for 6 hrafter induction of bacteremia, the vegetations were notsterilized by any drug except vancomycin which was onlypartially successful. A patient under penicillin treatment forstreptococcal endocarditis was studied during the course offull-mouth dental extractions. Early operative blood cultureswere positive for oral microorganisms while cultures subse-quent to infusion of vancomycin were sterile. The resultsindicate that vancomycin is an effective alternative to peni-cillin for preventing BE in susceptible hosts. Moreover,they suggest that this model provides an excellent systemfor the study of antimicrobials and their mechanisms ofaction.

83. Tumor Antigen (Carcinoembryonic Antigen-CEA)in Gastric Adenocarcinoma Masked by a Binding Sub-stance, Likely an Antibody. BARBARAJ. DYCE* AND BER-NARD J. HAVERBACK,** Los Angeles, Calif.

We have found at least two tumor-specific antigens ingastric, pancreatic, colonic, and rectal adenocarcinoma andtheir metastases. It was considered unusual when perchloricacid extract from liver metastases of a gastric adenocarci-noma, purified by Sepharose 4B gel column chromatographyand repeated polyacrylamide gel electrophoresis, containedno detectable CEA using highly specific antiserum with theOuchterlony gel double diffusion technique. Accordingly, anexplanation for this negative finding was sought and spe-cifically the masking of antigen by naturally occurring anti-body was investigated. As CEA is soluble and immuno-globulins or CEA bound to goat antibody are insoluble in0.6 N perchloric acid, the precipitate from this extractionwas treated with 15% NaCl or acidified to pH 3 to disso-ciate the glycoprotein antigen from antibody. Antibody wasthen precipitated by 50% saturated ammonium sulfate.After exhaustive dialysis and concentration, CEA was thenfound in the supernatant in both instances showing twoprecipitin bands with highly specific antiserum. Also, treat-ment of an initial isotonic saline homogenate of the tumorwith 15%o NaCl or controlled acidification to pH 3 withHC1, followed by 50% NH4SO4 precipitation, resulted inthe presence of two antigens in the supernatant. Subsequentpurification of these antigens by Sepharose 4B gel columnchromatography and repeated polyacrylamide gel electropho-resis, showed them to have similar chemical characteristics,as well as lines of' identity with antigens extracted byGold's method from hepatic metastases of colonic adeno-carcinoma. It is clear from these studies that (a) the pres-

ence of CEA in this cancer was masked by a naturallyoccurring substance, likely an antibody, and (b) perchloricacid extraction did not dissociate CEA from this substance.Also, it is possible that CEA may be masked by a naturallyoccurring substance, likely an antibody, in serum givingnegative or falsely low values, and dissociation proceduresshould be included in methodology.

84. Mechanisms of Hyperoxaluria in Regional Enteritis(RE). DAVID EARNEST,* HIBBARD WILLIAMS, AND WILLIAMADMIRAND,* San Francisco, Calif.

The mechanisms of hyperoxaluria and recurrent calciumoxalate nephrolithiasis were investigated in 20 patients withRE (8 ileal disease, 12 ileal resection). This hyperoxaluriahas been attributed to increased supply of the oxalate pre-cursor, glyoxylate, from bacterial degradation of glycine-conjugated bile salts (GBS). Oral taurine (8 g/day) re-sulted in a consistent reduction in oxalate excretion (mean51%) by substitution of taurine for glycine in bile salt con-jugation and thereby decreasing the quantity of GBS in thegut. Similarly, oral tetracycline (3 patients) reduced oxalateexcretion (mean 57%) by decreasing glyoxylate synthesisfrom GBSby intestinal bacteria. Accumulation of glyoxylate,as in primary hyperoxaluria type I, results in increasedurinary excretion of both oxalate and glycolate. In contrast,urinary glycolate excretion was normal in all 20 hyperoxa-luric RE patients. Furthermore, intravenous glyoxylate-14Cadministration (five patients) demonstrated decreased gly-colate synthesis in the presence of increased formation ofoxalate-14C (mean 19.2%, normal 12% of dose). These find-ings suggested increased conversion of glyoxylate to oxalateby LDH-NAD. Indeed, the elevated serum lactate: pyruvateratio (mean 35, normal < 15) in nine hyperoxaluric REpatients supported this concept. The discovery of hyper-oxaluria in three cholecystectomy patients with bile diver-sion by T-tube drainage confirms that this enhanced oxalatesynthesis may be related to increased NADH: NAD-linkedbile salt synthesis which is accelerated in many patients withileal disease. These findings demonstrate that a dual mecha-nism may be responsible for the hyperoxaluria that occursin patients with RE: (a) expansion of the glyoxylate pool,and (b) increased NAD: NADH-linked conversion of gly-oxylate to oxalate. (Research supported by grants fromNIH.)

85. Oxygen Transport after Blood Loss. MILES J. ED-WARDS,* AND BELLE CANON,* Portland, Oreg. (introducedby J. David Bristow).

Westudied the effect of mild blood loss on blood's affinityfor oxygen, designated by its inverse expression, the P50(oxygen pressure, Po2, at 50%o oxygen saturation of hemo-globin). A high P6o would improve oxygen unloading in thetissues. We also correlated the Pro with two of its knowndeterminants; (a) red cell 2,3-diphosphoglycerate (DPG)and (b) the proportion of red cells which are young, indi-cated by the activity of red cell glutamic oxaloacetic trans-aminlase (GOT), an enzyme present only in young red cells.We removed 250}750 ml blood over 1-3 wk from normalyoung adult nonsmoking males, taking blood samples forstudy 1-3 wk later. Hemoglobin concentrations decreased

26 a

(mean 16.0-14.8 g/100 ml) but remained within the normalrange. The P50 increased significantly (mean 28.2-30.1 torr)and correlated with increased red cell GOT, indicating thepresence of greater numbers of young red cells. DPGdid notchange significantly (mean 13.1-13.0 A.M/g hemoglobin). Theincreased P50 which followed blood removal is presumablythe result of increased erythropoietic activity, thereby in-creasing the relative numbers of young red cells in blood.Young red cells have a high P50. Wehave calculated that theimproved oxygen unloading afforded by the increased P50more than compensated for the decreased hemoglobin con-centration so that cardiac output could decrease and stillmaintain an unchanged normal mixed venous Po2. This mayhelp explain the previously reported decrease in frequencyof angina pain in patients with angina pectoris after phle-botomy (Burch and dePasquale. 1963. Arch. Intern. Med.111: 687). (Research supported in part by NIH Grant HE-12249.)

86. Effect of Water Loading on Pulmonary Extravascu-lar Water Volume (PEWV) with and without Norepin-ephrine Infusion. JAMES H. ELLIS* AND JOHN F. MURRAY,San Francisco, Calif.

We tested the hypothesis that water-loading animals withperipheral vasoconstriction would cause increases in PEWVowing to redistribution into the lungs of the infused volume.PEWVwas measured by the multiple indicator dilution tech-nique in 16 anesthetized (pentobarbital) dogs using albumin-"I, THO, and indocyanine green. Initial measurements ofPEWVwere made in all animals and were repeated im-mediately before and after water loading (20 ml/kg over 10min). Vasoconstriction was induced in nine dogs by intra-venous norepinephrine infusion (0.1-0.4 ,ug/kg per min) thatwas continued during water loading. Seven animals werewater loaded without norepinephrine administration. Controlanimals had a mean PEWVof 4.64±0.85 (SD) ml/kg ini-tially, 4.89±0.69 ml/kg before and 5.22± 1.16 ml/kg afterwater infusion. The increase in PEWVafter water infusionwas not statistically significant (P > 0.20). Cardiac indexincreased 56%o from 160+30.2 to 249+39.5 ml/kg per min(P < 0.001) after water infusion; mean pulmonary arteryand left atrial pressures increased 5.7±+1.7 mmHg and 2.8+0.4 mmHg, respectively. Animals given norepinephrinehad a PEWVof 4.70+0.75 ml/kg initially, 4.94+0.92 ml/kgbefore, and 5.34+0.92 ml/kg after water infusion. The in-crease in lung water after water loading was statisticallysignificant (P <0.01). Water loading produced an 87% in-crease in cardiac index from 166+31.6 to 311+88.5 ml/kg permin (P < 0.001) without significant differences in pulmon-ary artery or left atrial pressures compared to those of con-trols. We conclude that virtually none of an acutely ad-ministered water load is detectable in the PEWVof controlanimals, but in the presence of vasoconstriction, lung waterincreases significantly. These finding may have clinical rele-vance to fluid administration in critically ill patients. (Sup-ported by research grants HE-06285, HE-05705, and HE-14201-SCOR from NIH.)

87. Ouabain-Calcium Interaction in Myocardial Micro-somal Membranes. MARKL. ENTMAN,* JULIUS C. ALLEN,*

ANDARNOLDSCHWARTZ,* Houston, Tex. (introduced by Ru-bin Bressler).

Cardiac glycosides are thought to exert their cardiotoniceffect by increasing myocardial calcium concentration insystole. It has been suggested that this action depends ondigitalis binding to a specific binding site on the membrane,Na+, K+-ATPase. A microsomal membrane fraction was iso-lated from dog heart which had the ability to bind calciumand also contained ouabain-inhibitable Na+, K+-ATPase (12-20, ,umoles/mg per hr). Several interactions between cal-cium and ouabain were observed. (a) Ouabain (10' M)stimulated ATP-induced calcium exchange (29.9 vs. 21.1nmoles/mg) but not ATP-dependent calcium binding (38.2vs. 36.3). (b) Calcium (10-100 piM) stimulated specificouabain-3H binding (50-100%) to the preparation at pH 6.8(which is optimal for calcium binding). The basic charac-teristics of this binding were the same as those describedfor "purified" Na+, K+-ATPase in the presence of standardligand conditions (ATP, Mg'+, and Na'). (c) Endogenouscalcium (10 uM) did not affect the N+, K+-ATPase but itprevented its inhibition by ouabain (104-10-` M). Removalof the calcium by EGTA chelation restored the ability ofouabain to inhibit the Na+, K+-ATPase activity. The co-existence of Na+, K+-ATPase and calcium-binding activityin the preparation was required for these interactions; i.e.,ouabain does not affect calcium exchange in fractions lack-ing Na+, K+-ATPase and calcium does not alter the ouabaininhibition of "purified" Na+, K+-ATPase (the latter lackscalcium-binding activity). This represents in vitro evidencethat cardiac glycosides may act by binding to Na+, K+-ATP-ase and effecting augmented myocardial calcium flux. Whilethe exact cellular location of the membrane preparation isunknown, the presence of sarcolemma is probable. (Researchsupport by grants from NIH and American Heart Associa-tion.)

88. The Induction of Interferon by Vaccinia Antigen InVitro: Lymphocyte-Macrophage Interaction and Effectof Reimmunization with Vaccinia Virus. Lois B. EPSTEIN,*DAVID A. STEVENS,* AND THOMASC. MERIGAN, San Fran-cisco and Stanford, Calif.

Previous studies demonstrated that the nonspecific mito-gen, phytohemagglutinin (PHA), induces interferon (IF) incombined lymphocyte-macrophage cultures from normal in-dividuals. Purified protein derivative (PPD) also inducesIF in similar cultures but only when derived from sensitizeddonors. In both instances IF production was associated withtransformation of small lymphocytes to blastlike cells. Inview of these results and the observation that live virus in-duces IF from leukocytes in vitro, the present studies wereundertaken to determine whether a viral antigen (inactivatedvaccinia virus) could induce transformation and IF produc-tion in combined cultures. Cultures were prepared fromdonors who had been vaccinated with live vaccinia virusfrom 2 to 30 yr previously. The cultures were harvested at5-9 days after addition of heat-killed vaccinia virus (VH).Although there was significant lymphocyte transformation,no IF was detected in cultures from five of six donors.These five donors were reimmunized with live vaccinia virusand combined cultures prepared from 4 to 22 wk later. In

27 a

all, significant IF titers were detected, with the maximum(11-44 U) occurring at 4 wk. Despite this increase in IFafter reimmunization, no significant rise in lymphocyte trans-formation was noted. As was the case with PHA and PPD,the presence of autochthonous macrophages greatly enhancedthe production of IF in VH-treated cultures over that ob-served with cultures of lymphocytes alone. In addition, VH-induced IF had the same physical-chemical characteristics asdid the PHA- and PPD-induced IF. (Research supported byNIH Grants CA 11067-03, AI 05629-09.)

89. The Effect of Fetal Thyroidectomy on Thyroid Ki-netics in Fetal and Maternal Sheep. ALLEN ERENBERG,*KEIICHIRO OMORI,* WILLIAM OH,* AND DELBERTA. FISHER,Torrance, Calif.

Thyroid hormone kinetics and turnover have recently beenreported in fetal sheep. The present study assesses the effectof fetal thyroidectomy on thyronine (T4) metabolism. Fetalsheep were thyroidectomized at 90-110 days' gestation. Afterplacement of exteriorized catheters, tracer doses of T4-"Iand T4-"'I were injected into the fetus and mother, re-spectively, 4-30 days post-thyroidectomy. Serial blood sam-ples were drawn for 72 hr during maternal perchlorate ad-ministration. Serum butanol extracts were prepared for dualisotope counting. From this data the maternal and fetalvolumes of T4 distribution, half-times of T4 turnover, andmean fractional T4 clearance rates were estimated to be(12.6 and 0.97 liters, 1.22 and 0.99 days, and 0.50 and 0.77,respectively), values similar to those in nonthyroidectomizedanimals. Mean fetal serum T4 levels fell from a prethyroi-dectomy value of 12.2 jug/100 ml to 1.7 ,ug/100 ml on day 3post-thyroidectomy; maternal levels did not change. Meanmaternal and fetal T4 turnover values were 13.5 and 8.5,ug/kg per day (normal 5.5 and 40), respectively. Fetal serumthyroid-stimulating hormone (TSH) values were 300-1500,uU/ml 20-29 days post-thyroidectomy, confirming fetal hy-pothyroidism. Placental transfer of T4 was bidirectional with< 1 ,ug/day being transferred in the maternal to fetal direc-tion. The high fetal to maternal TSH and maternal to fetalT4 gradients indicate that the fetal pituitary-thyroid systemis autonomous and that placental T4 transport is inadequatefor fetal needs. (Research supported by grant HD 04270-04from NIH.)

90. An In Vitro Model of Gluten-Sensitive Enteropathy(GSE): Evidence that Gluten Is Not Directly Toxic toGastrointestinal (GI) Epithelium. Z. MYRONFALCHUK,*CLEMENTINE SESSOMS,* AND WARRENSTROBER,* Bethesda,Md. (introduced by J. E. Rall).

The toxicity of gluten in GSEmay be direct or indirect.To study this question an in vitro model of GSEwas sought.Jejunal biopsy specimens from patients with GSE and con-trol individuals were maintained in organ culture for 48 hr.Alkaline phosphatase (p-tase) activities of specimens wereobtained to estimate their functional state. P-tase activitiesin nine normal controls, and in nine disease controls withlow base line p-tase levels, showed minimal increases duringculture (mean: 1.6-fold). In contrast, specimens from sevenGSE patients in relapse with initially low p-tase values,showed significantly greater increases in p-tase activity

(mean: sevenfold) (P < 0.001) during culture with finalvalues usually coming to normal. However, when specimensfrom patients in relapse were cultured in the presence of apeptic-tryptic digest of gluten (PT), no increase of p-taseactivity was noted. This inhibition of increase of p-tase ac-tivity can be considered an in vitro model of GSE. Wenowturned to the question of the mechanism of gluten toxicity.Specimens from individual GSE patients obtained beforeand after in vivo gluten challenge were studied. In fourpatients before challenge, two- to threefold increases ofnearly normal base line p-tase activities were unaffected bythe presence of PT in the culture medium; in no case didactivities fall. In contrast, after in vivo gluten challenge,cultures showed fivefold increases in activity during culturewhich could be abolished by the presence of PT. Thus, PTtoxicity in vitro is expressed only after in vivo exposureto gluten, suggesting that gluten is not directly toxic to GImucosa, but must first activate an endogenous effectormechanism.

91. Interbacterial Transfer of R Factor during HumanShigellosis. W. EDMUNDFARRAR, JR.,* MARGENEEIDSON,*PATRICIA GUERRY,* STANLEY FALKOW,* LEWIS B. DRUSIN,*AND RICHARD B. ROBERTS,* Atlanta, Ga., Charleston, S. C.,Washington, D. C., and New York (introduced by JosephC. Ross).

Documentation of in vivo transfer of R factors in thehuman intestine, under conditions in which exogenous super-infection by resistant strains can be excluded, is rare. Wehave recently observed a probable instance of such in vivotransfer. A newborn infant acquired shigellosis during anursery outbreak caused by a strain of Shigella sonnei re-sistant to streptomycin (Sm), tetracycline (Tc), andampicillin (Am), designated strain R-1. After therapywith oral and intramuscular kanamycin (Km), he excretedorganisms with the resistance pattern SmTcAmKm(strainR-2). Strain R-1 transferred all three resistance deter-minants during mixed cultivation with Escherichia coli K-12,regardless of which drug was used to select recipients,indicating that all three determinants were components ofa single R factor. When strain R-2 was grown with E. coliK-12, most resistant recipients exhibited resistance patternsof SmTcAmor Km alone, indicating that R-2 possessedat least two factors. SmTcAmKm-resistant E. coli exhibitedtwo distinct molecular species of extrachromosomal DNA,of 70 million and 43 million molecular weight. SmTcAm-resistant cells had only the 70 million molecular weightplasmid DNA, whereas cells resistant to Kmonly possessedonly plasmid DNA of 43 molecular weight. These findingssuggest that R-1 acquired a second R factor conferring re-sistance to kanamycin from another organism in the patient'sgastrointestinal tract. To our knowledge this is the first re-port of acquisition of a second R factor by an already multi-resistant organism during the course of a clinical infection.(Research supported by grants from CDC, National ScienceFoundation, and U. S. Army.)

- -,.

92. Echocardiographic Detection of Dyskinetic and Aki-netic Segments of the Left Ventricle. HARVEY FEIGEN-

28 a

BAUM, BETTY C. CORYA,* SONIA CHANG,* LEE L. KONECKE,*AND JOHN C. FISCHER,* Indianapolis, Ind.

Echocardiography provides a noninvasive means of visual-izing internal cardiac structures. The purpose of this studywas to see whether or not echocardiography could detectabnormally moving segments of the left ventricle. Echo-cardiograms were done on 25 normal individuals and on20 patients with angiographically proven akinetic or dys-kinetic left ventricular segments. The ultrasonic transducerwas directed so as to record echoes from the true posteriorleft ventricular wall (PLV), the interventricular septum(IVS), the inferior left ventricular wall (ILV) near theposterior papillary muscle, and the anterior left ventricularwall (ALV). In the normal individuals, the PLV and ILVmoved anteriorally about 1 cm during ventricular systole.The IVS and the ALV moved posteriorally about 0.5 cmduring ventricular systole. In five patients with angio-graphically proven akinetic or dyskinetic septums, the IVSshowed diminished, absent, or paradoxical motion while thePLV motion was increased. In two patients with angio-graphic dyskinetic anterior walls, the ALV demonstratedparadoxical motion. Three patients had diaphragmatic myo-cardial infarctions, and the ILV motion was reduced. Inall three patients the IVS motion exceeded 0.5 cm. Fivepatients had akinetic or dyskinetic segments involving theapex or lateral walls. Although these areas were notvisualized on the echocardiogram, indirect evidence wassuggested by exaggerated motion of the IVS and/or thePLV. In five patients the echocardiograms were technicallyunsatisfactory. These results indicate that echocardiographycan detect akinetic and dyskinetic segments of the left ven-tricle both directly, by recording diminished echo motionof the involved areas, and indirectly, by noting exaggeratedmotion in the remaining noninfarcted muscle. Unfortunately,satisfactory echograms could not be obtained on all patients.(Supported by NIH grants HE-09815-06, HE-6308, HTS5363, HE-5749, and the Indiana Heart Association.)

93. Evidence for the Existence of Two Classes of Gluco-corticoid Receptors in Rat Kidney. DAVID FELDMAN,*JOHN W. FUNDER,* AND ISIDORE S. EDELMAN, San Fran-cisco, Calif.

Previous studies documented the properties of high affinitycytoplasmic aldosterone receptors (type 1) in the rat kidney.We recently demonstrated the presence of high affinity cyto-plasmic glucocorticoid receptors (type 2) which bind dexa-methasone (DM) somewhat better than corticosterone (B).Wenow present evidence for the existence of a third classof renal cytoplasmic receptors (type 3) which have a highaffinity for B and very low affinities for both DM andaldosterone. The relative affinities of the type 3 receptorswere: B > cortisol > deoxycorticosterone > progesterone >aldosterone > DM= spirolactone. By Scatchard analysis, theequilibrium dissociation constant of the type 3-B complexwas 3 X 10' M (250C). In density gradients this complexsedimented at 4S or in low Ca' at 8S. In kidney slices,type 3 receptors are differentially distributed, the papillahaving twice the concentration of the cortex. Specific nuclearuptake occurs in both zones of the kidney even when type 1and 2 receptors are blocked, and is proportional to the type

3 receptor concentration. The nuclear complex thus gen-erated sediments at 3S. The type 3 receptors are distinguish-able from plasma transcortin by virtue of differences inrates of dissociation, kinetics of binding in vivo, ion de-pendence of sedimentation coefficient, and nuclear uptake.The evidence indicates, therefore, that the mammalian kid-ney contains one mineralocorticoid and two glucocorticoidreceptors and we propose that each receptor regulates dis-tinct functions. Current efforts are directed towards ascer-taining the separate physiological roles of these receptors.(Supported by NIH Grants HL-06285 and HL-05725.)

94. Gluconeogenesis in Human Diabetes: Evidence of aPrimary Abnormality in Hepatic rather than PeripheralProcesses. PHILIP FELIG,* JOHN WAHREN,*EROL CERASI,*AND ROLF LUFT,* New Haven, Conn. and Stockholm,Sweden (introduced by L. R. Freedman).

Increased gluconeogenesis has long been recognized insevere ketotic diabetes. The mechanism of glucose overpro-duction and whether augmented gluconeogenesis occurs inmilder diabetes has not been established. The present studywas undertaken to characterize the gluconeogenic pattern innonketotic diabetics, and to determine if an abnormality inintrahepatic processes or augmented peripheral release andavailability of glucose precursors is the primary metabolicdefect responsible for augmented gluconeogenesis. Splanchnicand leg exchange of individual amino acids, lactate, pyru-vate, and glucose was determined by arterial, hepatic venous,and femoral venous catheterization in seven nonketotic dia-betics (24 hr after insulin withdrawal) and in 29 healthycontrols. Although splanchnic glucose output did not differsignificantly between groups (diabetics, 3.0±0.2 mg/kg permin; controls, 2.6±0.5 mg/kg per min), splanchnic uptakeof alanine, other glycogenic amino acids, lactate, andpyruvate was 50-150% greater in the diabetics. Totaluptake of glycogenic substrates could account for 32%of hepatic glucose output in the diabetics but for only16%o in the controls. The greater uptake of glucose pre-cursors in the diabetics was due to a two- to threefold in-crement in the splanchnic fractional extraction of thesesubstrates. In marked contrast, the arterial levels of alanineand other glycogenic amino acids were reduced by 20% inthe diabetics, while their output from the leg was equivalentto that in controls. Conclusions: In nonketotic diabeticstotal splanchnic glucose output is not elevated but therelative contribution from gluconeogenesis is increased two-fold. The augmented gluconeogenesis is a consequence ofincreased hepatic extraction of glucose precursors resultingsecondarily in depletion of circulating substrate. These re-sults suggest that the liver rather than the periphery isthe primary site of the gluconeogenic abnormality in diabetes.

95. Possible Mechanism of Compensated Hemolysis inHereditary Spherocytosis. Louis A. FERNANDEZ* ANDALLAN J. ERSLEV,** Philadelphia, Pa.

Compensated hemolysis is frequently observed in patientswith hereditary spherocytosis (HS) and characterized byan accelerated erythropoiesis despite a normal hemoglobin.Since tissue hypoxia is the usual driving force for erythro-

29 a

poiesis it has been suggested that the oxygen affinity of redcells is high in HS causing tissue hypoxia. The mean Ps,of 15 normals (Hgb: 15.1, Retic: 1.6, MCHC: 34.1) was27.8 mmHg and the mean P50 of 12 HS patients (Hgb:13.3, Retic: 9.9, MCHC: 37.9) was 28.4 mmHg. Althoughthe 2,3-diphosphoglycerate (2,3-DPG) of the red cells waslower in HS (2.65 mM/liter) than in normals (4.66 mM/liter), the effect of 2,3-DPG on the P5m was apparentlycancelled out, possibly by the effect of a high MCHC. Inany case, it appears that compensated hemolysis cannot beexplained on a high oxygen affinity of red cells in circulat-ing blood. However, it may be more relevant to considerthe oxygen affinity of blood actually perfusing the oxygensensor responsible for the regulation of red cell production.This oxygen sensor is believed to be located in the kidneyand probably in the medulla which is relatively hypoxic.Since medullary blood is hyperosmolar, the P60 was measuredin normal and HS red cells before and after exposure tohyperosmolar plasma (Na, 200 mEq/liter, Cl, 181 mEq/liter; and urea, 520 mOsmolar/liter). The P50 of red cellsfrom eight normals showed a decrease of 1.3 mmHg, whilethe P50 of red cells from six patients with HS showed adecrease of 4.7 mmHg. Consequently, the oxygen affinityof red cells from patients with HS is higher than that ofnormals in a hyperosmolar environment and, if the oxygensensor is located in the renal medulla, could explain theexistence of compensated hemolysis in patients with HS.

96. The Effect of Prostaglandin (PGA1) on AdrenalFunction in Man. MARSHALP. FICHMAN,* GLENNLITTEN-BERG,* JANNIE WOO,* AND RICHARD HORTON, Los Angeles,Calif.

Prostaglandins stimulate steroidogenesis in vitro. PGA1,the only known prostaglandin not rapidly metabolized bythe lung, was administered to 15 normal and 5 anephricpatients. Aldosterone was measured by radioimmunoassay,cortisol by competitive binding, and renin (PRA) by bio-assay. In 13 supine normal subjects given a subdepressordose (<0.6 ,ug/kg per min) by constant infusion for 1 hr,PGA1 strikingly increased plasma aldosterone from 4.2±2.0(SD) to 13.7+11.4 ng/100 ml (P <0.01). However, plasmacortisol did not increase significantly (P > 0.5). Surpris-ingly, plasma renin (215±158) also did not increase (285±233 mnug/100 ml, P > 0.4). No change in serum Na and Kwas observed. 4-hr subdepressor infusions to normals furtherincreased aldosterone, and were associated with fever andmarked rises in cortisol without change in renin. No con-sistent changes in thyroid-stimulating hormone or growthhormone was observed. Dexamethasone (0.5 mgX 6 hr X 8)eliminated the fever and cortisol response, but did not pre-vent the rise in aldosterone. Depressor doses of PGA,(5 2.5 /Ag/kg per min) further increased aldosterone withvariable increases in plasma renin in normals. 1-hr infusionsof Z 0.6 gg/kg per min of PGA1 in anephric patients in-creased plasma aldosterone in t without changing cortisol,but in contrast to normals lowered blood pressure. Thesestudies indicate that PGA, increased aldosterone independentof alterations in ACTH, renin, and serum electrolytes.PGA1 may be an important regulator of aldosterone secre-tion in man. (Research supported by NIH.)

97. Altered Disease in Rats due to Mutants of ReovirusType 3. BERNARD N. FIELDS* AND CEDRIC S. RAINE,*Bronx, N. Y. (introduced by M. D. Scharff).

Temperature-sensitive mutants of reovirus type 3 havebeen isolated and characterized genetically and biochemi-cally. In contrast to the wild-type virus which produces anacute necrotizing encephalitis in rats infected by intracerebralinoculation, certain mutants induce a slowly progressivecommunicating hydrocephalus. Other mutants induce noovert disease. The biochemical defects in those mutantswhich induce hydrocephalus resides in the inability to as-semble outer coat peptides. The assembled viral outer pro-tein shell thus appears to be responsible for the lytic cellresponse and acute encephalitis due to wild reovirus type 3.Since pseudomyxovirus viral factories have been found incertain diseases such as subacute sclerosing panencephalitisand systemic lupus erythematosis, the use of well-charac-terized viral mutants could provide a model system forstudying pathogenesis of certain slow virus infections andthe possible virus etiology of autoimmune diseases. It ispossible that mutations in genes coding for outer viral coatpeptides are responsible for an altered virus-cell interactionin these diseases. (Supported by grants from the NIH,National Science Foundation, and ACS.)

98. Rapid Serum Bactericidal Assay by 50Cr Release.JOSHUA FIERER* AND A. I. BRAUDE,** San Diego, Calif.

Complement-mediated destruction of Gram-negative bac-teria is a major defense mechanism. We have devised arapid, sensitive assay for serum bactericidal activity thatdirectly measures the effect of serum on the bacterial cellwall. Bacteria grown in broth were resuspended in sodiumchromate (Rachromate) for I hr at 37°C. After washingin cold saline, labeled bacteria were added to serum in aratio of 1: 10. Portions were removed, filtered throughMillipore filters (0.45 1L), and then counted in a gammacounter. Filterable radioactivity was expressed as per centof total radioactivity of each sample after subtraction ofactivity released by saline control. The 51Cr is associatedwith endotoxin as demonstrated by radioautography. Re-lease of 'Cr by normal serum began almost immediatelyand proceeded exponentially for 7 min until 70-80% of thelabel was filterable. This was accompanied by a 2 log fallin bacterial counts. Heated, hydrazine-treated, and EDTA-treated serum released only 5%o of counts. Normal serumreleased only 28% of 'Cr from a serum-resistant Escherichiacoli. This assay is a direct measure of complement-mediateddisruption of the cell wall, the process that leads to bacterialdeath and thus permits the demonstration of subtle defectsin bactericidal activity. For example, diluting serum 1: 4caused a 40% fall in 'Cr release but only 12%o decrease inkilling ability. This bactericidal assay is rapid and repro-ducible and avoids the errors inherent in colony counting.It is a direct measure of the action of serum on bacterialcell walls. (Supported by NIH Fellowship lF03 A143980-01and NIH Grant A.I. 10108-01.)

99. Human Growth Hormone: Evidence for Its Conver-sion or Secretion of a Second Factor. S. EDWIN FINE-BERG* AND THOMASJ. MERIMEE, Boston, Mass.

30 a

Human growth hormone (HGH) is secreted throughoutthe life-span of normal individuals, but its physiologic roleis unknown in the adult. To establish its function in theadult, acute metabolic effects were compared after intra-venous infusions and arterial perfusions. The previouslyreported actions of HGHafter intravenous 'administrationwere not seen when physiologic or pharmacologic concen-trations were perfused across the tissues of the human fore-arm. (a) A 30 min intravenous infusion of HGHresultedin a mean plasma concentration of HGHof 79.7+6.4 ng/ml;low dose and high dose HGH perfusions achieved meanplasma concentrations of 29.7±2.6 ng/ml and 231.9+8.4 ng/ml. The intravenous infusion caused an early decrease of freefatty acid (FFA) concentration of 21.0±3.6% (P < 0.05)and at 2 hr, an increase to 146.0±7.3% of the mean controlconcentration. Neither intra-arterial dose had an early effecton FFA and only the high dose perfusion caused increasedlipolysis at 2 hr. (b) Systemic glucose concentrations werelowered 9.5±1.9%o (P < 0.05) by intravenous HGH; bothlow and high dose arterial perfusions failed completely to in-crease glucose uptake and actually caused a small decrease inglucose uptake across muscle. (c) K+ uptake was stimulatedacutely by low and high dose perfusions across peripheralmuscle. Change in uptake per 100 ml of forearm tissue was,for low dose, 9.0±3.04 Amoles; and for high dose, 11.0+3.8

nmoles. The data demonstrate that the acute insulin-likeeffects of systemically administered HGH do not resultfrom actions of HGHon peripheral tissues, regardless ofthe dose. The data are consistent with one of two inter-pretations: (a) HGHmust be converted to a more activeform, or (b) HGH stimulates the secretion of a secondfactor.

100. Interferon Induction and Development of TissueViral Resistance in Chickens. MARTIN S. FINKELSTEIN,*New York (introduced by A. Taranta).

Although interferon-inducing agents have been found ef-fective in preventing experimental viral illnesses, their roleis often ambiguous since they may also stimulate the im-mune and reticulo-endothelial systems, cause fever, inactivateviruses directly, interfere with viral replication by noninter-feron means, and produce various toxic effects. These studiesin chickens employed the following: two different interferoninducers (poly rI/rC and statolon), administered by differ-ent routes to stimulate interferon in the trachea or in theserum; infection in vitro of extirpated trachea and kidneysto measure tissue anti-viral resistance of a respiratory andvisceral organ; and three different viruses with varyingdegrees of sensitivity to interferon to determine the spe-cificity of resistance. Interferon titers were measured in theserum and in the tracheal scrapings, in order to relate tissueresistance with the amount of interferon present at the timeof sacrifice and infection. It was found that: (a) interferon-like protection can be confined to specific organs, dependingupon the interferon inducer used and its route of administra-tion. (b) There is significant correlation between the degreeof tissue protection with the amounts of interferon presentin the serum or tracheal scrapings (particularly when largeamounts of interferon are present), but the presence of smallamounts of interferon, or the absence of detectable interferon

frequently is not indicative of protection or a lack of tissueprotection. And (c) the technique of testing the specificityand degree of viral resistance of extirpated organs providesan important index of the efficacy of interferon inducers ad-ministered in vivo in the protection against a specific infec-tion. (Supported by USPHSGrant No. AI 09977-01 Al.)

101. Role of Pulmonary Surfactant in Pathogenesis ofEmphysema in Smokers. THEODOREN. FINLEY, San Fran-cisco, Calif.

The major hypotheses on the pathogenesis of emphysema,(a) alveolar overdistention due to obstruction and (b) weak-ness or tearing of the alveolar wall, make no mention of theacellular surface-active layer, pulmonary surfactant. Thislayer tends to prevent either alveolar collapse or overexpan-sion. From the considerations below a weakness and tearingof alveolar walls could be the direct result of airway obstruc-tion and a breakdown in the surfactant lining of the alveolidue to smoking. The principal component of pulmonary sur-factant is dipalmitoyl lecithin (DPL). A normal humanlung at 7 liter lung volume, with alveolar surface area of70 m2, requires 210 mg of DPL for a packed monolayer(DPL area per molecule 35 A'). Animal lavage studiesyield about 0.5 mg of "alveolar" DPL per gram of lung.A 700 g normal human lung from a smoker or nonsmokerwould be expected to contain 350 mg of alveolar DPL. Frombronchopulmonary lavage data in eight nonsmokers, about70% of alveolar DPL is free (noncellular) giving a valueof 250 mg DPL. This would be more than enough to forma packed monolayer at all lung volumes. However, from lav-age data in eight cigarette smokers. only 20%o or 70 mg ofalveolar DPL would be free. At all but the smallest lungvolumes a DPL monolayer would not be tightly packed.During expansion of the alveolar surface a break in the DPLmonolayer would occur in some alveoli. Factors that lead tooverexpansion of such alveoli, such as airway obstruction,could cause a break in the protective surface film in smokers.In these alveoli direct contact to cell walls of noxious agentsin cigarette smoke, proteolytic enzymes, bacteria, etc., wouldbe possible. This could result in weakness and tearing ofalveolar walls, i.e., emphysema.

102. Changes in Brain Amines after Portal Flow Diver-sion and Acute Hepatic Coma. JOSEF E. FISCHER,* J. How-ARD JAMES,* AND Ross BALDESSARINI,* Boston, Mass. (in-troduced by W. Gerald Austen).

Previous work from this laboratory has suggested that theaccumulation of false neurotransmitter amines may explainsome of the pathophysiology of hepatic failure including someof the cardiovascular, renal, and central nervous systemsymptomatology. In the following experiments the origin ofsome of these amines is investigated and further changes inneurotransmitters are documented. Phenylethylamines, suchas tyramine or alphamethyltyramine, were given by gastrictube to rats 2 months after portacaval shunt and their sham-operated controls. Animals were killed at 3 hr and brainsand hearts were assayed for total and 8-hydroxylated amines.The brains and hearts of the shunted animals contained 2-5times the 8-hydroxylated derivatives as that of the controls.

31 a

(Tyramine: sham 1.83±0.17, shunted 4.97±0.18 ng Octo-pamine/g heart.) (a-Methyltyramine sham 0.74±0.19, shunted1.76±0.32 nCi a-methyloctopamine/brain, P < 0.02; for heartsham 7.8±+1.4, shunted 38.8±4.7 nCi a-methyloctopamine P<0.001.) These results strongly suggest that the increasedaccumulation of amines perhaps have their origin in the gut.Acute hepatic coma was prepared in rats by hepatic arteryligation 24 hr after end-to-side portacaval shunt. Animalswere killed when moribund and brains and hearts were as-sayed for serotonin, norepinephrine, and dopamine. Whiledopamine was unchanged, norepinephrine was markedly de-creased in the brains of the comatose rats: sham 318.1±+13.9(n=21), coma 231.5±21.5 ng/g (n=16) P<0.001. Sero-tonin was markedly increased: sham 398+0.20, coma 517±0.30 ng/g (n = 10) P < 0.002. These findings further sug-gest that the recently reported efficacy of L-DOPA in thearousal of patients in hepatic coma may be due not only toits effects on restoring normal levels of the adrenergic trans-mitters, dopamine and norepinephrine, but also perhapssecondary to its recognized property of diminishing the ac-cumulation of serotonin. To date L-DOPA has been admin-istered to 11 patients in deep, unresponsive hepatic coma,with prompt awakening in 6, delayed response in 2, and nonoticeable effect in 3. The characteristics of patients whoshowed a response as well as the cardiovascular changessecondary to L-DOPA administration will be discussed.(Supported by a Grant No. AMNS 15347-01 from NIH.)

103. Extracellular Products of Human Glomerular Cellsin Tissue Culture. ALFRED J. FisH,* DAVID M. BROWN,*ROBERTL. VERNIER,** ANDALFREDF. MICHAEL, Minneapolis,Minn.

Although sites of biosynthesis are unknown, glomerularbasement membrane (GBM) contains a collagen moiety witha high hydroxylysine/hydroxyproline ratio, a glucose-galac-tose moiety, and a heteropolysaccharide containing sialic acid;a polyanionic material is found on glomerular epithelial cells.Using autopsy kidneys from infants, sterile decapsulatedglomeruli were isolated and explanted in tissue culture fromwhich morphologically distinct "epithelial" cells surviving 13-15 passages were derived. By electron microscopy glomerularcells, when compared to fibroblasts, had larger rounded nu-clei, multiple nucleoli, and more generous cytoplasm. Theextracellular matrix between epithelial cells contains granu-lar membranous material distinct from the fibrillar productsadjacent to fibroblasts. Fluorescein-labeled antiserum to hu-man GBMstained the extracellular material produced byboth glomerular cells (GC) and fibroblasts (FB). Absorp-tion of the anti-GBM serum with (a) purified GBM in-hibited the staining of membranous material in GC, FB, andthe GBMof normal human kidney (NHK) sections; (b)lyophilized washed FB inhibited the staining of FB mem-branes, partially inhibited that of GC, but did not inhibitNHK staining; and (c) lyophilized GC and FB inhibitedstaining of both FB and GC, but not NHK. These studiesindicate that the membranous material from glomerular cellsshare antigenic determinants with GBM. Glomerular cellpreparations contain 3-hydroxyproline and convert proline-'Hto hydroxyproline-8H. No polyanionic material was detected-in extracellular membranes with alcian blue or colloidal iron

(pH 1.9), although positive staining was seen in extracellu-lar globules. Wehypothesize that human glomerular cells intissue culture are capable of synthesizing some component(s)of GBMand the glomerular polyanion. (Supported by US-PHS, Minnesota Heart Association, the Minnesota MedicalFoundation, Twin Cities Diabetes Association, and theAmerican Heart Association.)

104. Nucleotide Sequence Analysis of Human Hemoglo-bin Messenger RNA. BERNARDG. FORGET* AND CHARLESA. MAROTTA,* Boston, Mass., and New Haven, Conn. (in-troduced by Sherman M. Weissman).

Hemoglobin messenger RNA (mRNA) has previouslybeen demonstrated in the 9S RNA region of total reticulo-cyte RNA fractionated by means of sucrose gradient cen-trifugation. We have isolated 9S RNA, free of all otherRNA, by polyacrylamide-gel electrophoresis of total reticu-locyte RNA, and the gel-purified 9S RNA was shown tostimulate amino acid incorporation by a cell-free proteinsynthesizing system. 9S RNA and other RNA componentswere digested with T1 ribonuclease and labeled by transferof the terminal phosphate of -y-'P-labeled ATP by meansof polynucleotide kinase prepared from phage T4 infectedEscherichia coli cells which were deficient in their ribonu-clease content. Two-dimensional oligonucleotide maps suit-able for further sequence analysis were obtained from lessthan 0.2 Ag of RNA; thus the method could be applied tothe analysis of small quantities of mRNAisolated from re-ticulocytes of patients with sickle-cell anemia and otherhemoglobinopathies. Rabbit reticulocyte 9S RNA could beseparated into two bands by polyacrylamide-gel electrophore-sis. Each component was shown to given nearly identicaloligonucleotide maps which were similar to the fingerprintobtained from the single 9S RNAof reticulocytes from pa-tients with sickle-cell anemia. By contrast the patterns givenby 28S and 18S RNA were very different. This approachwas shown to be applicable to the analysis of the mRNAofthalassemic cells and provides the first direct chemical ap-proach to fine structure analysis of the hemoglobin genes inman. (Research supported by grants from the NIH andACS.)

105. Mechanism of Action of Arabinosyl Cytosine (ara-C): Assignment of Replicative Function to a HumanDNAPolymerase. R. M. Fox,* G. V. RAMAREDDY,* ANDM. GOULIAN, La Jolla, Calif.

Recently, multiple DNA polymerases have been foundin both animal and microbial cells; however, their precisefunctions remain largely unknown. Studies employing ara-binosyl cytosine triphosphate (ara-CTP) to distinguish bio-logical roles of microbial enzymes in repair or replication ofDNA, led to a similar investigation with human enzymes.Two DNA polymerases were purified from a permanentdiploid line of cultured human lymphocytes. Similar to otherresults with animal cells, the two enzymes were distinctivein cellular location, behavior on DEAE-cellulose, optimafor pH and [Mg++], and response to SH inhibitors. One ofthe polymerases of nuclear origin was unaffected by ara-CTP, in contrast to the marked inhibition of a second poly-

32 a

merase of predominantly cytoplasmic location. Ara-C spe-cifically inhibits DNA replication of intact cells, and itsprincipal form in the cell is known to be the triphosphate.Therefore, the inhibition of the "cytoplasmic" DNA poly-merase by ara-CTP permits the tentative assignment ofreplicative function to that enzyme, despite its paradoxicallocation. Additional support for this interpretation comesfrom studies with human peripheral blood lymphocytes, inwhich phytohemagglutinin-stimulated DNA synthesis, likenormal replication, is inhibited by ara-C. Correlated withstimulation by phytohemagglutinin is an increase in the levelof "cytoplasmic" enzyme but not "nuclear" enzyme. Theseresults combined with studies on several microbial enzymesidentify a class of ara-CTP-inhibitable DNA polymerases,which appear to have an essential role in replication. Thisalso clarifies older observations, made before recognition ofthe multiplicity of DNApolymerases, that ara-CTP inhibitsmammalian but not microbial DNApolymerases. (Supportedby USPHS [5-RO1-CA11705], ACS [P-567], and NHMRC[Australia].)

106. Islet Phospholipogenesis and Glucose-StimulatedInsulin Secretion. NORBERT FREINKEL AND NASSER CO-HANIM,* Chicago, Ill.

Glucose-stimulated insulin secretion may be subdividedinto: effects of glucose or glucose metabolites upon islet"receptor" . increased Ca` flux .-4 microtubular realign-ments and emiocytosis. To assess associated membrane events,we have correlated insulin secretion by isolated rat isletswith concurrent incorporation of orthophosphate-'P into thecholine (PC), ethanolamine (PE), and inositol (PI) phos-phoglycerides. Incubation with 50 mg/100 ml glucose didnot stimulate insulin release although 'P incorporation in-creased as incubation was prolonged (PC = PE> PI). In-creasing glucose to 100 mg/100 ml stimulated insulin releaseminimally coincident with qualitatively different increasesin the relative rates of phospholipogenesis (PE > PC= PI).At 200-300 mg/100 ml glucose, labeling of PC and PEplateaued; PI increased in parallel with increasing insulinsecretion at 300-500 mg/100 ml glucose. To document thateffects upon PI and PE were linked to "stimulated" insulinsecretion, responses to 50 vs. 300 mg/100 ml glucose werecorrelated in the following situations. Islets from 48-hrfasted rats exhibited less stimulated insulin secretion thanfed animals coincident with reduced PE and PI labeling butunchanged PC. Islets from 48-hr fasted, 24-hr refed ratsdisplayed greater stimulated insulin secretion coincident withgreater increases in PE and PI than PC. Islets "pulse-labeled" with 3P exhibited greater turnover of PI than PCduring "chase" with 'P and 300 mg/100 ml glucose. Con-trariwise, incubation without Ca++, or with tubulo-activeagents (colchicine) blocked stimulated insulin secretion butdid not affect PI or PE selectively. Thus, the selective effectsupon PE and PI may occur before the Ca++ and microtubu-lar aspects of the secretory sequence. Weconclude that "trig-gering" insulin secretion in islets with glucose elicits thesame changes in selected acidic phosphoglycerides that char-acterize secretory stimulation in other structures and mayconnote some primary perturbation of a membrane receptor.(Supported by USPHSAM-10699 and AM-05071.)

107. Effect of Differentiated Myeloid Cells on Hemato-poietic Stem Cell (HSC) Proliferation. W. FRIED, W. H.KNOSPE, AND F. E. TROBAUGH,JR., Chicago, Ill.

Previous studies showed that HSC transplanted into micegiven cytoxan 4 days before X-irradiation regenerate fasterthan those in mice which were only irradiated. The factors,responsible for this enhancement of HSC replication, werestudied. Five mg of cytoxan was injected into CF1 mice andcontrols received saline. 1-10 days afterward they received1000 R X-irradiation and 106 marrow cells. HSC in theirfemoral marrow were assayed at various times afterward bythe spleen colony method of Till and McCulloch. Injectionof cytoxan 4 days before X-irradiation significantly en-hanced regeneration of transplanted HSC, whereas injection7 days before resulted in a significant delay of HSC regen-eration. The number of HSC and nucleated marrow cellswere determined, and differential cell counts were done 1 hrafter X-irradiation on femoral marrow cells of mice whichreceived saline or cytoxan 4 or 7 days before X-irradiation.The nucleated cell count, number of HSC, and number ofdifferentiated cells of mice given cytoxan 4 days beforeradiation were less than in controls which received salineonly. In those treated 7 days preradiation the number ofHSC remained less than control, the nucleated cell countwas no different from control, but differentiated cell countsrevealed 85% of the marrow cells to be differentiated mye-loid cells (bands and segs) as compared to < 50% in thecontrol group. These results suggest the conclusion that theproliferation of hematopoietic stem cells is modulated by cell:cell interactions with mature myeloid elements in the milieu.(Research supported by grants from NIH, ACS, and Leu-kemia Research Foundation.)

108. Hemodynamic Identity of Progressing Essential(Man) and Spontaneous (Rat) Hypertensive Heart Dis-ease. EDWARDD. FROHLICH* AND MARCA. PFEFFER,* Okla-homa City, Okla. (introduced by Eugene D. Jacobson).

Left ventricular functional characteristics of progressinghypertensive heart disease have not been shown clinicallyor experimentally, although studies in essential hypertensiveman demonstrated group hemodynamic differences even be-fore cardiac failure. Because of ethical considerations inpursuing these studies in uncomplicated essential hyperten-sion with developing left ventricular hypertrophy (LVH),20 young (Y; aver. 24 wk) and 10 old (O; aver. 77 wk)spontaneously hypertensive rates (SHR) were comparedwith the same number of age- and sex-matched normal Wis-tar controls (NCR). Measurements were made under etheranesthesia using electromagnetic flow probes (ascendingaorta) and superior vena cava (SVC) and aortic catheters.Aortic maximum flow acceleration (Accel) and peak velocity

Heartrate CO LVER Accel Peak-Q LV: body

MAP (beats/ (ml/ (ml/ (ml/ (ml/ wtRats (mmHg) min) min) sec) sec2) sec) (X10-3)

NCRY 85 342 95 3.36 250 5.28 1.97SHRY 115* 397* 90 2.98* 251 5.22 2.81*NCRo 85 270 71 2.90 245 4.75 1.98SHRo 114* 356* 65 2.27* 174* 3.75* 3.55** Indicates at least P < 0.01 significance; CO = cardiac output; LVER= left ventricular ejection rate.)

33 a

(Peak-Q) were used as indices of left ventricular contrac-tility. There was no evidence of pulmonary or hepatic con-gestion, and SVC pressures were normal. Thus, the charac-teristics of the young and old SHR were very similar toessential hypertensive man with left atrial hypertrophy (earlyLVH) and gross LVH, respectively. The data also demon-strate that left ventricular function is significantly and pro-gressively impaired in LVH resulting from systemic hyper-tension, even before development of cardiac decompensation.

109. Serum Precipitins in the Diagnosis of DisseminatedCandidiasis. J. DAVID GAINES* AND JACK S. REMINGTON,**Palo Alto, Calif.

Although the problem of disseminated candidiasis (DC)is well recognized, less than 50%o of cases are diagnosedpremortem. A rapid, reliable means of diagnosis of infectionwith Candida is needed to allow for early instigation oftherapy. A prospective study was performed to define theusefulness of the Candida precipitin test for diagnosis ofDC in a university and community hospital population. Pre-cipitins were found in 26 (27%) of 95 patients with sus-pected DC. In 21 (80%) of the 26 the diagnosis of DCwas subsequently established. The most common underlyingcondition in these cases was multiple surgical procedures.Five patients with positive precipitins were lost to follow-up.Two patients with DC (one with terminal leukemia andone with multiple abdominal surgery) had no precipitins.Two leukemics with candidemia just before death had noprecipitins and no evidence of DC at postmortem. Cross-reactions due to other fungal infections were not observed;seven of the precipitin negative patients were found to haveaspergillosis and one, coccidioidomycosis. There were nopositive reactions among a general outpatient populationused as controls. Precipitin lines in conventional immuno-diffusion methods often did not appear for 48-72 hr. Devel-opment of a method which employs a combination of elec-trophoresis and immunodiffusion provided results in allcases within 2 hr, was simple to perform, and agreed per-fectly with the Ouchterlony method. This method is easilyadaptable for use in clinical laboratories.

110. First Step in Digitalis Action: Direct Study ofOuabain Binding to the Plasma Membrane. JERRY GARD-NER,* THOMASCONLON,* AND CATHERINE FRANTZ,* Be-thesda, Md. (introduced by Leonard Laster**).

The initial step in the mechanism of action of cardiacglycosides is thought to involve an interaction with theplasma membrane. To characterize this interaction we haveinvestigated the binding of ouabain-3H to the plasma mem-brane of intact human erythrocytes. We have found thatouabain binding is reversible, does not depend on cellularenergy metabolism, exhibits a high degree of chemical spe-cificity, and can be detected at ouabain concentrations aslow as 1 X 10-10 mole/liter. Ouabain binds to a single classof glycoside binding sites via a saturable reaction. There are1200 glycoside binding sites per cell and the ouabain con-centration at which half-maximal binding occurs (KB) is1.5 X 10' mole/liter. Both monovalent and divalent cationsact at sites which are functionally distinct from the glyco-side binding site and their effect is to alter the affinity of

the glycoside site for ouabain. Furthermore, the site at whichmonovalent cations exert their effects on ouabain binding isfunctionally distinct from that site at which divalent cationsexert their effects. Na, Li, and Cs increase ouabain binding(KB = 5.0 X 10-1, 9.7 X 10', and 2.8 X 108 mole/liter, re-spectively), K and Rb decrease ouabain binding (KB = 2.6X10-7 and 1.8 X 10' mole/liter, respectively), and Ca, Ba,and Mg increase ouabain binding (KB = 8.3 X 10', 4.5 X 10',and 1.7 X 10' mole/liter, respectively). In addition to alter-ing the affinity of the glycoside site, divalent cations alsoregulate the affinity of the monovalent site for monovalentcations. These observations indicate that the initial actionof cardiac glycoside is association with a membrane "re-ceptor complex" which has three functionally distinct sites:a monovalent cation site, a divalent cation site, and a glyco-side binding site. The particular cation-determined func-tional configuration of the membrane complex, in turn,determines the affinity of the glycoside binding site forouabain.

111. Nephron Structure and Function in Precystic Kid-neys of Rats Fed Diphenylamine. KENNETHD. GARDNER,JR.,* AND SIDNEY SOLOMON,* Honolulu, Hawaii (intro-duced by Roy H. Maffly).

To document pathophysical factors responsible for cystdevelopment in affected mammalian kidneys, standard micro-dissection and micropuncture techniques were used to ex-amine anatomical and functional properties of precysticnephrons in rats fed 1% diphenylamine for 5-9 months.Latex castings disclosed focal narrowing and dilatation ofproximal and distal nephrons. Pressures in vivo in 30 visiblydilated proximal tubules in 10 precystic kidneys averaged(+SE) 26.2±1.6 cm water. Pressures in 24 relatively un-dilated tubules in the same kidneys averaged 19.6±0.8 cmwater. The difference between means was significant (P <0.001). When individual proximal tubules were injected withaqueous lissamine green and tritiated inulin, two excretorypatterns were observed. One was characterized by promptexcretion of green highly radioactive urine from the injectedleft kidney. It followed injection of 6/19 tubules in sevenprecystic kidneys and 12/12 free-flowing tubules in sevencontrol rats. The second was characterized by the absenceof dye-tinged urine and the excretion of less highly radio-active urine by both kidneys. It followed injection of 13/19precystic nephrons and was produced intentionally in 6/6normal nephrons by preliminary partial occlusion with oil,thus implicating obstruction as a cause for the two patternsof dye-isotope excretion that were observed in precystickidneys. Dye-isotope excretion correlated with intratubularpressure (r = - 0.698; P < 0.001) but not with urinaryflow within the range observed (1.64 - 15.98 ,ul/min).These results delineate two populations of nephrons in theprecystic kidneys of diphenylamine-fed rats and support aconcept that obstruction and increased intratubular pressureparticipate in cyst formation in the susceptible mammaliankidney. (Supported by NIH Grant HE 13137.)

112. Induction of Nasal Interferon by Rhinoviruses anda Topical Drug. BIENVENIDO G. GATMAITAN,* EDITH D.STANLEY,* AND GEORGEGEE JACKSON,** Chicago, Ill.

34 a

Rhinovirus (RV) 21, passage 3, induced nasal interferon(IF) when administered intranasally to volunteers. Virusshedding from the nose preceded interferon production,which was measurable on day 2, reached peak titer on days4-6, and sharply declined thereafter. In antibody-free sub-jects the pattern of virus shedding was not altered by theappearance or titer of IF. In three subjects with preexistingserum antibodies, virus excretion was inversely related topeak IF titers. Persons in whom nasal IF was measurablehad an ameliorated illness compared to those with infectionwho failed to produce IF. The severity of the symptoms,however, was not related to the titer of IF produced. Apassage 14, "attenuated" strain of RV 29 induced similarIF responses. A chemical inducer (NN-dioctadecyl-NN'-bis(2 hydroxyethyl) propanediamine), administered by intra-nasal spray, induced nasal IF, similar to virus administra-tion. Application of the drug 24 hr before RV challenge didnot prevent infection. The inducer and virus were additivein elevating the IF titer in nasal secretion. Thus, IF whichappears in nasal secretion after RV infection, or a chemicalinducer may diminish symptoms, but the IF titers reacheddid not prevent RV infection or change the patterns ofvirus shedding.

113. Hormone-Receptor Interactions in Circulating Cells:Studies in Normal and Pathologic States in Man. JAMESR. GAVIN,* JUANITA A. ARCHER,* MAXINE A. LESNIAK,*PHILLIP GORDEN,* AND JESSE ROTH, Bethesda, Md.

Since polypeptide hormones exert their effects on cellreceptors, the accessibility of the circulating cell makes thistissue desirable for studies in man. We find that circulatingand cultured human lymphocytes have receptors for insulinthat have the same affinity and biologic specificity for insulinas the receptors studied in liver and fat cells of rats. Todetermine whether the number of receptors and/or theaffinity of the receptor for insulin is the same or differentin normal individuals, acromegaly, obesity, and hypopitui-tarism, the maximal binding of insulin-"I (Bma.) and thehalf-maximal displacement of insulin-"2I (J B...) by un-labeled insulin have been determined. In the 8 studies innormals and 15 patient studies the mean range Bmax equals2.45-3.53%o and the mean range i B. equals 6.5-9.4 ng/ml.The apparent affinity constant for the normals and patientsequals 109 M1 (limit of sensitivity = 1-2 ng/ml) ; no statis-tical differences could be demonstrated between the normalsand the patient groups under these experimental conditions.In the cultured cell, however, an additional order of insulinreceptor sites with an apparent affinity constant of 1010 M1has been detected, increasing the sensitivity to 0.1 ng/ml.With this method the purified insulin component of circulat-ing insulin has the same biologic specificity as pancreaticporcine insulin, whereas the proinsulin-like component hasa similar biologic specificity to porcine proinsulin. Specificgrowth hormone (HGH) receptors are also present inhuman cultured lymphocytes. HGH-'5I is bound to the cul-tured lymphocytes and displaced only by biologically activeHGH. The ability of an HGHpreparation to displace HGH-'I from the lymphocyte receptor is directly proportional tothe biological activity of the preparation. The sensitivity ofthis method is between 5 and 10 ng/ml of HGH. When

plasma from an acromegalic patient is filtered on G-100Sephadex (superfine), the major peak of immunoreactiveHGHdisplaces HGH-1'I from the cultured lymphocytes ina manner indistinguishable from a purified pituitary HGH.These data indicate that it is now possible to study directlyin man hormone receptors and the interaction of the hor-mone receptor with its circulating polypeptide hormone.

114. Complement Components C8 and C9 in Sera andUrine of Patients with Chronic MembranoproliferativeNephritis (CMPGN). H. GEIGER,* N. DAY,* AND R. A.GOOD,** Minneapolis, Minn. (introduced by Robert Ul-strom) .

In chronic membranoproliferative nephritis (CMPGN)activation of the complement system, particularly throughthe alternative pathway, has been shown. Studies in CMPGNregarding the complement system focus mainly on C3 andits conversion. No information is available about C8 andC9 in sera of patients with CMPGN. To date, determina-tions of C8 and C9 in the urines of patients with renaldiseases have not been reported. In this study hemolyticC8 and C9 were measured in sera and urines of 12 patientswith CMPGN. In serum C8 was found to be significantlydecreased in 8 patients. Only 3 patients showed significantlylow C9 titers, whereas in 2 patients C9 was found to beelevated. The C8 and C9 titers did not always correlatewith hemolytic C3. Hemolytic C9 could be measured inurine of all CMPGNpatients; in 11 of 12 patients C8 wasdetectable in urine. Hemolytic activity of C8 and C9 inurine could be inhibited by EDTA, suramin, and heat inac-tivation (56°C, 30 min). When Clearance C8/Clearance C9index was compared with Clearance IgG/Clearance Trans-ferrin index low selectivity for the IgG/transferrin index(>0.12) was found in 11 of 12 patients; whereas in 10 of12 patients the C8/C9 index revealed low selectivity (>0.05) when compared to cases of idiopathic nephrotic syn-drome. Low selectivity using complement index was alsofound, in cases of chronic glomerulonephritis (5) andnephrotic syndrome (3) of unknown etiology, congenitalnephrosis (2), lupus erythematosus (2), and kidney trans-plantations (3). In summary, hemolytic C8 and C9 can bemeasured regularly in unconcentrated urine of patients withproteinuria. Determination of C8/C9 index provides a re-liable tool to detect low selectivity in cases with proteinuria.(Aided by NIH.)

115. Functional Mapping of the Hypothalamus: RNASynthesis in Response to Hypertonic Saline. JACK M.GEORGE,* Columbus, Ohio (introduced by James V. War-ren**).

As a new noninvasive approach to localization of hypo-thalamic function, we have performed radioautography on9-,um serial sections of hypothalamus using rapid incorpora-tion of cytidine-8H into RNA as an index of polypeptideand protein synthesis. Two groups of seven mice each weregiven water (control) or 3% saline for 2 days and theneach mouse was injected with 0.15 ,uCi cytidine-'H intra-peritoneally. The mice were killed 90 min later and radio-autographs prepared. Treatment of histologic sections withribonuclease before coating with photographic emulsion pre-

35 a

vented the appearance of > 95% of silver grains. Controlmice had 0.6 silver grains/gm' (mean) ±0.2 (SE) in thesupraoptic nucleus compared to 2.4 grains/flm2 ±0.2 (P <0.001) in the saline-treated mice. This result agreed witha previous pilot experiment involving four mice. Hypertonicsaline is known to stimulate synthesis of vasopressin in thesupraoptic nucleus. There was also a significant differencein cytidine-3H incorporation in the ventromedial area adja-cent to the third ventricle: control 4.0 grains/gm2 +0.3;saline 6.6 grains/gm2 ±0.8 (P < 0.01). This area may beinvolved in reception and transmission of the hypertonicstimulus. There were no significant differences between con-trol and saline groups in nine other hypothalamic areascounted. We conclude that dehydration is associated withincreased incorporation of cytidine-3H into RNA in preciseareas of the hypothalamus. Functional mapping of hypo-thalamic response to stimuli is possible. (Supported in partby Bremer Foundation.)

116. Role of Complement in Escherichia coli Bacteremiain the Squirrel Monkey. DAVID N. GILBERT,* JACK A.BARNETT,* AND JAY P. SANFORD, Dallas, Tex.

Endotoxins are capable of activating the terminal com-ponents of the complement sequence in vitro and in vivo.Serum complement levels fall after intravenous administra-tion of endotoxin. Complement activation might be eitherprimary or coincidental in the pathophysiology of Gram-negative bacillary bacteremia. Availability of a nontoxicprotein fraction of cobra venom which inactivates C3 madefeasible in vivo studies on the influence of complement deple-tion on experimental Escherichia coli bacteremia. Squirrelmonkeys (27 controls, 31 cobra factor [CoF]-treated ani-mals) were depleted of complement (means CH50 levels:control 226 U, CoF-treated 31 U; means C3 levels: con-trols 187 U, CoF-treated 36 U) and lethal inocula of viableor heat-killed 'P-labeled E. coli (4 X 109 organisms/kg)were administered intravenously. Striking neutropenia oc-curred rapidly in controls (the average leukocyte count de-creased from 19,000/mm3 to 3500/mm3 with 1 min) while therate of occurrence of neutropenia was 20-30 times slowerin CoF-treated animals. This blunted response was paral-leled by a decreased sequestration of neutrophils in the pul-monary circulation during the initial minutes postchallenge.In complement-depleted monkeys, the initial rate of clearanceof either viable or 'P-labeled E. coli was slower and re-sultant levels of bacteremia and endotoxemia (measured byactinomycin D-mouse assay) were significantly higher thanin controls. The late defect in bacterial clearance was cor-rected by administration of fresh monkey serum. Theseobservations are consistent with the following mechanism:phagocytosis of intravascular bacteria and detoxification ofendotoxin is more efficiently performed by marginated neu-trophils and margination is complement mediated. Thus,complement-mediated neutrophilic leukocyte function is animportant host defense mechanism in Gram-negative bacil-lary bacteremia. (Research supported by grants fromNIAID, NIH, USAR and D Comm., DOD.)

117. Decrease in the Hemolytic Anemia of Sickle CellDisease after Administration of Sodium Cyanate. PETER

N. GILLETTE,* CHARLES M. PETERSON,* JAMES M. MAN-NING,* AND ANTHONYCERAMI,* New York (introduced byAttallah Kappas).

Studies in these laboratories have recently shown that cya-nate can inhibit the sickling phenomenon in vitro by spe-cifically carbamylating amino-terminal residues of hemoglo-bin S (Proc. Nat. Acad. Sci. U.S.A. 1971. 68: 1180). Thisobservation has led to the initiation of clinical studies de-signed to evaluate the possible use of cyanate in the treatmentof this disease. In a group of patients with sickle cell diseasecyanate treatment in vitro of a small aliquot of erythrocyteswas associated with an increase towards normal in the ap-parent 50% chromium-51 survival time when these cells werereturned to the patient (Proc. Nat. Acad. Sci. U.S.A. 1971.68: 2791). If cyanate decreased sickling, then chronic ad-ministration would be expected to diminish the hemolyticanemia. Eight of ten patients with sickle cell disease receiv-ing cyanate orally over 2-4 months showed a mean increasein hemoglobin from 8.5 to 10.7 g/100 ml and increased hema-tocrit from 26 to 31%. The maximum increase in blood in-dices was 40% of control values in one patient. Mean totalserum bilirubin declined from 2.1 to 1.4 mg/100ml. Carbamy-lation (as moles of cyanate per mole of hemoglobin tetramer)achieved from a given oral dose was variable. The two in-dividuals who did not respond to oral cyanate did show ap-propriate responses when treated intravenously. Toxic sideeffects observed were dose related, reversible, and well toler-ated. These results demonstrate the hematologic efficacy ofcyanate as a possible treatment for sickle cell disease andestablish its short-term safety in humans. (Research sup-ported by grants from NIH, National Science Foundation,the NF-MOD.)

118. Effect of Immunoregulatory Alpha-Globulin on An-tigen and Mitogen Binding by Lymphocytes. A. H. GLAS-GOW,* S. R. COOPERBAND,K. SCHMID,* AND J. A. MANNICK,Boston, Mass.

We, and others, have demonstrated an immunosuppressivefactor in plasma which may be isolated with the a-globulins.We have hypothesized that this factor serves a regulatoryfunction in vivo and have called this factor immunoregula-tory a-globulin (IRA). Wehave previously found that IRAinhibits graft rejection and primary and secondary antibodyresponses. In vitro studies have demonstrated that IRA in-hibits blast transformation and proliferation induced by bothantigens and mitogens, without cytotoxicity. The inhibitionis easily removed by washing the lymphocytes, and IRA isineffective if added after exposure to activating agent. Thesestudies suggest that IRA acts upon the early events involvedin antigen and mitogen activation. We have therefore ex-amined the earliest event in activation-antigen and mitogenbinding. Antigen binding by human blood or murine spleenlymphocytes was studied by: (a) rosette formation withsheep erythrocytes, alone, and as carrier for diptheria ortetanus toxoid or tuberculin, and (b) uptake of free BSA-l"I.Mitogen binding was studied with 'I-labeled PHAand Con-canavalin A. IRA at 1-3 mg/ml inhibited the binding of allantigens: 80%o inhibition of rosette formation, and 60% in-hibition of free antigen uptake. IRA did not interfere withantigen binding by preformed antibody (measured as agglu-

36 a

tinin titer). There was little inhibition of PHAbinding and aminimal inhibition of Conconavalin A uptake, although lym-phocytes cultured with IRA did not show the normal in-creased uptake of Conconavalin A which occurred withtransformation. These studies suggest that IRA interfereswith the earliest events of antigen uptake at the surface oflymphocytes, but not by altering the antigen or immuno-globulin. Since it does not alter mitogen uptake it must exertits antiproliferative action metabolically, and at a membranesite near the antigen-binding site, but distant from the mito-gen-binding areas. (Research supported by grants from NIHand U. S. Dept. of the Army.)

119. Importance of Cell-Cell Interaction in the Prolifera-tion of Human Hematopoietic Cells in Liquid Culture.DAVID W. GOLDE* AND MARTIN J. CLINE, San Francisco,Calif.

To examine the role of cellular interaction in the pro-liferation and differentiation of human hematopoietic tissue,normal and neoplastic human marrow cells were cultured inliquid suspension utilizing an in vitro diffusion chamber. Theapparatus consists of a small glass bulb in which the cellsuspension is separated from a large volume of medium bydialysis membrane. This system permits intimate cell con-tact while allowing for adequate nutrient supply and me-tabolic waste removal. At intervals, total and differentialcounts were performed and radioautographs prepared afterincubation with thymidine-'H. Under appropriate conditions,granulopoiesis and mononuclear cell proliferation persistedfor several weeks in cultures of normal human bone marrow.Proliferation depended on cellular proximity provided by ahigh cell density within the chamber. This situation obviatedthe need for exogeneous stimulating factors required in semi-solid systems. Growth of mononuclear cells on the dialysismembrane was a concomitant of all normal marrow culturesand may be necessary for the growth and differentiation ofother cell lines. Maturation of early red cell precursors toerythrocytes was noted in cultures up to 1 wk old, but noproliferation of erythroid elements was observed. Leukemiccells from untreated patients had a high proliferative capacityand labeling index in cultures 1-2 wk old. In one cultureinitiated with leukemic myeloblasts only mature polys wereobserved at 25 days. It is concluded that proliferation anddifferentiation of human hematopoietic cells can be accom-plished in liquid culture, can proceed in the absence of exo-genous stimulating factors, and are dependent on cell-cellinteraction. (Research supported by ACS Grant No. C I-60.)

120. The Biochemical Basis of Androgen Resistance inthe Testicular Feminization Syndrome in the Mouse.JOSEPH L. GOLDSTEIN* ANDJEAN D. WILSON, Seattle, Wash.and Dallas, Tex.

Previous work has demonstrated that male pseudoherma-phroditism in the X-linked testicular feminization syndrome(Tfm) in the mouse, and probably in man, results from re-sistance to testosterone action during androgen-mediatedsexual differentiation of male embryos; this androgen re-sistance is believed to result from diminished androgen bind-ing to nuclear receptor sites in target tissues. To characterizethe mechanisms by which this defect might arise, the binding

of dihydrotestosterone-1,2-8H has been examined in extractsof the submandibular gland, a tissue in which androgen-mediated synthesis of several proteins is deficient in Tfmmice, utilizing equilibrium dialysis, gel filtration on Sepha-dex G-100, and sucrose density sedimentation analysis. Toensure that secondary testosterone deficiency was not re-sponsible for the effects observed, control, castrated, andtestosterone-treated groups were studied. Both by gel filtra-tion and by density gradient, specific dihydrotestosteronebinding in the cytosol was localized in a discrete proteinpeak of approximate 3S size and mol wt less than 60,000,easily separable from the larger androgen-binding protein ofmouse serum. In castrated animals no difference in the con-centration of binding protein was demonstrated between male,female, and Tfm animals. In testosterone-treated male andfemale animals the cytosol-binding protein was markedly de-creased, a finding compatible with the concept that thecytosol-binding protein-androgen complex is consumed byconversion to less soluble nuclear binding proteins. In con-trast, the cytosol-binding protein of Tfm animals increasedafter testosterone treatment so that binding in Tfm cytosolwas 5-10 times greater than that of cytosol from testo-sterone-treated male animals. These findings suggest thatdeficient nuclear binding of androgen in the Tfm mouse re-sults from inability to transfer cytosol-binding protein tonuclear receptor sites. This abnormality could result frommutation in cytosol protein not involving androgen binding,from abnormality in nuclear receptor complex, or from adefective component in the transfer process.

121. Receptors in Plasma: New Parameter of Target Or-gan Disease. THEODOREGOODFRIEND*ANDFREj FYHRQUIST,*Madison, Wis. (introduced by George Rowe).

Plasma from some severely ill patients contained macro-molecules with many of the known characteristics of recep-tors for angiotensin. These characteristics include high af-finity, specificity, and saturability of hormone binding, andinhibition by chemicals such as diethylstilbesterol and SKF525-A. The molecular weight of the binding macromoleculeswas approximately 800,000, and they migrated in the 6,-globulin region on electrophoresis. Assay involved simple gelfiltration of plasma mixed with labeled hormone. Maximumbinding capacity of positive plasmas was 0.01 pmole/ml. Afew plasmas also contained macromolecules of differentmolecular weight which bound labeled insulin or glucagon.Of 2000 plasmas tested, 106 were positive for the angioten-sin binders. One of the conditions associated with positiveplasma was surgical relief of arterial occlusion (four pa-tients). This was mimicked experimentally by occlusion andrelease of limb circulation in anesthetized dogs. After therelease, dog plasmas transiently contained binders for angio-tensin which resembled those from patients. Other diseasesassociated with plasma angiotensin binders included glomer-ulonephritis (6), septicemia (3), and chemotherapy for avariety of tumors (23). Relative specificity of binders forcongeners of angiotensin was characteristic of the pathology.For example, postocclusion human and dog plasmas boundoctapeptide 20 times better than heptapeptide, while glomer-ulonephritis plasmas bound octapeptide 5 times better thanheptapeptide. There was no pattern of blood pressure or

37 a

chemistry, including enzymes, that differentiated patientswith binders from normals or others with similar diagnoses.The appearance of receptors in plasma may be a parameterof target cell damage, with the binding characteristics indi-cating the location of the diseased organs. Their presence inblood does not appear to interfere with hormone action, buttheir absence from targets may.

122. Evidence for Participation of the Central NervousSystem (CNS) in the Transition from the Fed to theFasted State. CHARLES J. GOODNER, DONNA KOERKER,*PERTTI TOIVOLA,* CHARLES C. GALE,* AND JOHN ENSINCK,Seattle, Wash.

The role of the endocrine pancreas in the adaptation tofasting has been emphasized but the central nervous system(CNS) may also participate via the autonomic and neuro-endocrine systems. Wehave examined this adaptation in con-scious baboons after conditiong to two isocaloric feedingschedules: food ad lib. between 8 a.m. and 8 p.m., or foodonce per day at 4 p.m. Blood was sampled and urine collectedat 4-hr intervals from 8 a.m. to 12 p.m. during 2 days' feed-ing, 3 days' fasting, and 2 days' refeeding. The metabolicchanges of fasting were heralded by a 20%o fall in glucoseoccurring at the time of the first missed meal. This occurred8 hr earlier in the ad lib.-conditioned than in the once perday-conditioned animals (at 8 a.m. vs. 4 p.m., respectively).Insulin fell progressively reaching basal concentrations after36 hr of fasting; thereafter glucose and insulin remainedlow and relatively constant until refeeding. FFA, glycerol,and glucagon rose but, in contrast to glucose and insulin,displayed a circadian pattern in parallel with urinary NEand E. Basal growth hormone did not change with fasting.In other studies, ganglionic or 8-adrenergic blockade, after36 hr of fasting, produced a 50% decline in FFA, glycerol,and insulin. The effect of conditioning to different feedingschedules, the cyclic pattern of lipolysis and glucagon secre-tion, and the results of blocking studies indicate an importantrole for the CNS in the adaptation to fasting. Wehave pre-viously demonstrated that CNS glucoreceptors can affectlipolysis. The sequence of changes in early fasting suggeststhat a fall in glucose may be the signal for both the CNSand pancreatic control systems to initiate the adaptation tofasting. (Research supported by grants from NIH.)

123. Segment Power as an Index of Left VentricularFunction in Man. L. GOULD,* W. KENNEDY,* G. HAMIL-TON,* M. FRIMER,* AND H. T. DODGE, Seattle, Wash.

LV wall segment power (Ps) was calculated at every 0.01sec during the cardiac cycle as the product of wall stress andnormalized velocity of wall contraction in radial, equatorial,and meridianal directions, in contrast to merely the circum-ferential direction as in the past. Primary data were obtainedfrom biplane angiocardiograms (12/sec) and catheter-man-ometer systems having uniform amplitude responses to 20-25 cycles/sec. 73 patients with various degrees of myocardialdysfunction (nonvalvular) were studied, including 9 patientswith normal hearts and coronary arteriograms as controls.Ps values ranged from 715±198 g-cm/cm3-sec in normals(ejection fraction, 60±6%o) to 234+79 g-cm/cm' in pri-mary cardiomyopathy (ejection fraction, 19+6%). Values

for Ps correlated with other indices of ventricular perform-ance with the following r values: ejection fraction +0.85,maximum external pump power/EDV +0.91, systolic work/LV mass + 0.94, EDV/m2- 0.73, ED equatorial stress- 0.4, ED pressure - 0.40, maximum normalized velocity ofcircumferential shortening + 0.90, maximum rate of ejection+ 0.54, maximum dp/dt +0.56, maximum dp/dt/P + 0.44. Inpatients with localized contraction abnormalities ejectionfraction was significantly reduced, whereas Ps was only mod-estly diminished; in patients with diffuse generalized LVdysfunction Ps was reduced to the same extent as ejectionfraction. The results suggest that Ps may also permit quan-titative assessment of LV function in the presence of local-ized contraction abnormalities. Ps includes more variablesthan any other index of ventricular performance and may beinterpreted in terms of muscle mechanics (force-velocity rela-tionships) without assuming any particular model for thecontractile characteristics of muscle. Thus, it may have someadvantages over other measurements used to characterizeLV performance. (Supported by NIH Grant 118303.)

124. Defective Regulation of Protein Kinase Activity inThree Liver-Derived Tissue Culture Lines. DARYL K.GRANNER,Iowa City, Iowa (introduced by Stanley G. Koren-man).

The only proven mechanism of adenyl cyclase (AC)-medi-ated effects is through activation of protein kinase (K). c-AMPbinds to a protein (R) which inhibits the activity of K;cAMP binding dissociates R from K thereby activating thelatter. This model has been tested in three hepatoma celllines (HTC, RLC, and H4-II-E) because they all containsubnormal amounts of AC, and in at least one (HTC cells)tyrosine aminotransferase is not induced by added cAMP,whereas it is in liver. All three cell lines have markedly de-creased amounts of cytosol R as compared to liver (L, 2.33±0.15 pmoles cAMP bound per mg protein; H4-II-E, 0.58±0.02; RLC, 0.46±0.03; and HTC, 0.32±0.02). As predictedfrom the model: (a) cytosol K is higher in all these cell linesthan liver (L, 9.51±0.82 pmoles 'P transferred per mg pro-tein; HTC, 47.0+2.3; RLC, 95.2+15); (b) cAMP causes asmaller increase in K activity when added to crude or par-tially purified extracts of HTC (+ 217%) or RLC cells(+ 183%) than liver (+ 793%); (c) other cyclic nucleo-tides (cIMP, cGMP, cCMP, dbcAMP) stimulate K activityin proportion to their ability to inhibit cAMPbinding to R;(d) K activity of HTC cells can be reduced to the livervalue by addition of liver R, and in this reconstituted systemcAMP is effective. We conclude that these multiple defectsin this important pathway afford a unique opportunity tostudy the regulation of related metabolic processes. (Sup-ported by USPHSGrant No. 12191.)

125. Prophylactic Polymyxin B for the Prevention ofColonization by Nosocomial Gram-Negative Bacilli of theUpper Respiratory Tract of High-Risk Patients. SHELDONGREENFIELD,* DANIEL TERES,* LEONARDS. BUSHNELL,* JOHNHEDLEY-WHYTE,* AND DAVID S. FEINGOLD, Boston, Mass.

A prospective study was designed to try to prevent hos-pital-acquired pneumonia with Gram-negative bacilli in high-risk patients. Since the etiologic organism in pneumonia most

38 a

often enters the lung by aspiration from the upper respira-tory flora, the first goal of the study was to prevent coloniza-tion of the upper respiratory passageways. We report herethe success of prophylactic polymyxin B given by aerosal,one method employed in the study, in preventing colonizationOf the respiratory tract with Gram-negative bacilli. 43 pa-tients on admission to the Respiratory-Surgical IntensiveCare Unit (RSICU) were randomized into control and thepolymyxin-treated groups. Treatment consisted of 2.5 mg/kgper day of polymyxin B by aerosol in six divided doses intothe posterior pharynx and also trachea if an endotrachealtube or tracheostomy was present. The two groups weresimilar with respect to age, concomitant administration ofantibiotics, and number with tracheostomies. The patients inthe control group spent an average of 7.1 days in the RSICUwith a total of 62 days on mechanical ventilation; comparablefigures for the polymyxin group were 9.0 days and 120 days.Of the 19 control patients, 13 became colonized with Gram-negative bacilli, whereas of the 24 patients treated withpolymyxin aerosol, 4 became colonized (P < 0.01). No ad-verse reactions to the antibiotic were recognized. We con-clude that polymyxin B given prophylactically by aerosolcan reduce the incidence of colonization of the upper res-piratory tract by nosocomial Gram-negative bacilli in high-risk patients.

126. Mechanisms of Endotoxin Tolerance: Effect of Ex-change Transfusion. SHELDONE. GREISMAN, Baltimore, Md.

Following a single intravenous injection of a lethal doseof bacterial endotoxin, most of the toxin is rapidly re-moved by the reticuloendothelial system (RES); apprecia-ble amounts, however, continue to circulate for hours. It hasbeen proposed that enhanced clearance of this residual cir-culating toxin by the RES accounts for endotoxin tolerance.The present studies were designed to critically test thisthesis by determining whether early removal of the residualcirculating endotoxin by exchange transfusion would sig-nificantly reduce mortality. Femoral artery cannulation wasperformed in healthy, unanesthetized New Zealand albinorabbits (2.0-2.5 kg). A control group of randomly selectedanimals then received 2.5 mg E. coli endotoxin intravenouslyfollowed by 4000 U heparin 20 min later (equivalent toamounts received by test animals during exchange trans-fusion). The test group received 2.5 mg E. coli endotoxinintravenously, and 20 min later exchange transfusion per-formed as follows: 10-ml aliquots blood were rapidly with-drawn via the femoral artery cannula and 10-ml fresh hepa-rinized aliquots of blood (filtered to remove microemboli)were rapidly returned via the same cannula. Volume of ex-change transfusion equaled 14%o body weight and accom-plished exchange of approximately 80%o of the original bloodvolume. Control exchange transfusions performed before en-dotoxin injections did not enhance mortality. Mortality innonexchanged animals was 83%o (19 deaths/23 total); mor-tality in exchanged animals was 70% (16 deaths/23 total).It is concluded that mortality after sudden massive endo-toxemia cannot be reduced significantly by removal 20 minlater of a major portion of the residual circulating toxin.The findings are consistent with the concept that toleranceto endotoxin is dependent not upon enhanced clearance of

toxin from the blood by the RES, but upon diminished sts-ceptibility of the cellular elements of the RES to the in-jurious effects of the toxin.

127. Defective Bile Salt Synthesis as Cause of Choles-terol Gallstones. ScoTr M. GRUNDY, ALLAN L. METZGER,ANDRONALDADLER, Phoenix, Ariz.

Cholesterol solubilization in bile depends largely on bilesalts. Conversion of cholesterol into bile salts is thus crucialfor removal of cholesterol from the body. This conversion isunder feedback regulation by bile salts in the enterohepaticcirculation. In the present study, we have found a defect inregulation of bile salt formation that is probably the majorcause of cholesterol gallstones in the American Indianpopulation. Hepatic secretion rates for bile salts were de-termined in 16 Indian and 7 Caucasian women with and with-out gallstones, respectively. Our method employed duodenalintubation with 3-lumen tubes; markers and liquid formulawere infused continuously for 18-24 hr. Secretion rates weredetermined by marker dilution principles. Average rates ofbile salt secretion in Indian women with stones were mark-edly reduced, as compared to normal Caucasian women (470±35 vs. 1504±271 mg/70 kg per hr; P <0.001). These re-duced secretions could have been due to abnormalities ineither synthesis or reabsorption of bile salts. These twoparameters were compared in six Indian women with stonesand six normal Caucasian women. Fecal bile salt excretionswere determined for 2-6 wk. Average excretions for the twogroups were not significantly different. Since bile salts in theenterohepatic circulation were relatively depleted in Indians,even normal fecal excretions indicated reduced fractional re-absorption. However, previous studies have shown that Cau-casians can rapidly replenish bile salt pools in the presenceof sizeable intestinal losses; therefore, we conclude thathomeostatic regulation of bile salt synthesis is defective inIndian women. The resulting decrease in hepatic secretionof bile salts exceeds limits required for solubilizing choles-terol in bile; hence, cholesterol gallstone formation is almostinevitable.

128. Mechanism of Action of an Escherichia coli En-terotoxin. R. L. GUERRAN-r,* N. F. PIERCE,* U. GANGULY,*W. B. GREENOUGHIII,* AND C. K. WALLACE,* Baltimore,Md. (introduced by C. C. J. Carpenter).

Enterotoxigenic Escherichia coli can cause acute waterydiarrhea. We compared the effects of E. coli enterotoxin(ECT) and cholera toxin (CT) upon jejunal function. ECTwas the lyophilized, dialyzed, cell-free supernatant of liquidculture of a jejunal isolate from a man with severe diar-rhea. Net water fluxes were determined in ligated segmentsof canine jejunum using a dilution marker (PSP). Iso-tonic ECT (500 Aig/cm) induced maximum net secretion(8.2±+1.4 Al/cm - min, mean ±SE) within 10 min. 15 and75 min after ECT removal net absorption had returned,being 4.0+2.2 and 4.4+1.1 /Al/cm - min, respectively. IsotonicCT induced maximum net secretion (14.6±2.1 il/cm * min)only after 150 min. Jejunal mucosal adenyl cyclase activityincreased 193+49%o after 10 min ECT exposure (P < 0.01),59+31% (P > 0.10) 75 min after ECT removal, and 246+60% (P < 0.01) 150 min after CT exposure. ECT perfusion

39 a

500 gig/min) did not increase jejunal clearance of 'Cr-labeled serum albumin, or diminish enhanced sodium ab-sorption when the perfusion solution contained 60 mmoles/liter glucose. ECT perfusion during sustained maximumsecretory response to CT did not increase the net secretoryrate. Choleragenoid, a natural toxoid of CT, was employedto compare mucosal binding of ECT and CT. Prior ex-posure of rabbit jejunum to choleragenoid (0.2 ;tg/cm)blocked response to CT whereas 7 ,ug/cm choleragenoiddid not alter ECT response. We conclude that ECT in-duces prompt fluid loss and mucosal adenyl cyclase activa-tion, which are not sustained after ECT removal. Failureof choleragenoid to inhibit ECT indicates different mucosalbinding for CT and ECT. This may contribute to the strik-ing difference in time course of their effects. (Supportedby NIH Grant AI-07625 and Career Development Awardto N. F. Pierce.)

129. Very Low Density Lipoprotein Metabolism inAbetalipoproteinemia. CHRISTIAN GULBRANDSEN,* VIR-GINIA EVANS,* ALEXANDER NICHOLS,* AND ROBERT S.LEES,* Cambridge, Mass. and Berkeley, Calif. (introducedby Richard J. Wurtman).

Wehave shown that human plasma very low density lipo-proteins (VLDL) are converted to low density lipoproteins(LDL) by the squirrel monkey. To determine whether suchconversion occurs in man, and to localize further the meta-bolic lesion in abetalipoproteinemia, we infused isolatedhuman VLDL into each of two boys with that disease.VLDL, obtained from a patient with type IV hyperlipopro-teinemia after 3 days of fat-free feeding, was injected intra-venously at a dose equivalent to 11 mg protein/kg bodyweight in one subject and 3 mg/kg in the other. Serialplasmas from the recipients were analyzed for VLDL andLDL by analytical ultracentrifugation. In addition, VLDLprotein was analyzed by the Lowry technique and LDLprotein by radial immunoassay after preparative ultracentrifu-gation at density 1.006. Both patients showed sequentialdegradation of VLDL with gradual downward shift of themajor Sf peak with time and produced lipoproteins whichwere immunologically, electrophoretically, and ultracentrifu-gally LDL. In the subject who received the larger amount,the infused VLDL had a half-life of 6 hr; peak LDL pro-tein concentration of 2.62 mg/dl occurred at 12 hr. In theother patient the infused protein had a half-life of 1.5 hr;peak LDL protein concentration of 2.42 mg/dl occurred at4 hr. VLDL incubation in the recipients' plasma in vitroat 370C for 72 hr produced negligible amounts of LDL pro-tein. The data show (a) that patients with abetalipopro-teinemia can catabolize VLDL, (b) that sequential VLDLdegradation occurs in them, and (c) that LDL is a productof VLDL catabolism in humans. (Supported by NIH,American Heart Association, and National Dairy Council.)

130. Failure of Cyclic GMPto Stimulate Short-CircuitCurrent. P. F. GULYASSYAND G. A. PORTER,* San Fran-cisco, Calif., and Portland, Ore.

An increase in short-circuit current (SCC) across toadbladders was observed in response to cyclic GMPby Bour-goignie et al. (Science [Washington]. 1969. 165: 1360). We

recently observed marked variations in the purity of fourdifferent commercial preparations of cyclic GMP. CyclicGMPfrom Boehringer-Mannheim Corp. (BMC) and Nu-tritional Biochemicals (NB) were found to be pure bycolumn, paper, and thin-layer chromatography and had lessthan 1% noncyclic nucleotide as shown by release of inor-ganic phosphate with Crotalus atrox venom. The snakevenom released 10-11% of the phosphate in "cyclic GMP"obtained from Calbiochem (CB) and P-L Biochemicals,Inc. (PL) and anion exchange chromtography showed six toseven -impurities (by O.D. 260). We compared the effectsof pure and impure cyclic GMPon short-circuit current.In six paired cross-over studies mean per cent change inSCC (compared to control period) was +6.9±8.4 SE (NS)for pure cyclic GMP (BMC) and +23.5±6.8 (P<0.025)for impure cyclic GMP (CB). On subsequent cross-overre-addition of the serosal media, responses were + 42.3±10.9(P < 0.025) to cyclic GMP (BMC) and + 80.9±+10.0 (P< 0.001) to cyclic GMP (CB). In another set of eightpaired studies potassium cyclic GMP(CB) gave - 2.3±2.2(NS) when compared to paired KCl controls on first addi-tion and + 17.9±6.6 (P < 0.05) on re-addition. Pure sodiumcyclic GMP (NB) gave - 0.9±+1.2 (NS) compared topaired NaCl controls on initial addition in eight studies.Thus, the rise in SCC seen by Bourgoignie et al. with"cyclic GMP" (CB) may have been due to an impurityor a derivative formed during incubation. (Supported bygrants from NIH.)

131. Primary Hyperparathyroidism: ImmunochemicalCharacterization of Parathyroid Hormone in the Circu-lation. J. F. HABENER,* G. V. SEGRE,* D. POWELL,* P.DEE,* AND J. T. POTTS, JR., Boston, Mass.

We recently reported that parathyroid hormone (PTH)is secreted from the parathyroids in vivo as a polypeptideof mol wt 9500 and, after secretion, it is cleaved and ahormonal fragment of mol wt 7500 is detected in the circu-lation. Using radioimmunoassays that specifically detect thecarboxyl-terminal (sequence 53-84) and amino-terminal (se-quence 14-19) regions of the PTH molecule, we have mea-sured immunoreactive PTH in plasma, and in fractions fromgel filtration of plasma obtained from the general circula-tion, parathyroid effluent blood, and extracts of adenomasin patients with primary hyperparathyroidism. Results showthat C-reactive PTH in the general circulation measures6-20 times higher than N-reactive PTH, whereas in theeffluent blood and in extracts of adenomas, C- and N-reactive PTH measure the same. Gel filtration reveals that85-95% of the immunoreactive PTH in the general circu-lation consists of a late-eluting hormonal fragment of molwt 7500 that reacts in the C-assay and not in the N-assay.The 5-15% of immunoreactive PTH in the plasma, thatreacts in the N-assay, elutes coincident with intact, nativePTH. Since a continuous peptide sequence consisting of atleast the first 21 amino acids from the N-terminus is re-quired for biological activity, we conclude that much ofcirculating immunoreactive PTH is biologically inactive.Our recent observations show that PTH is biosynthesizedas a prohormone of mol wt 11,500, is cleaved, and is storedin and secreted from the gland as a molecule of mol wt 9500.

40 a

Thus, it appears that the metabolism of PTH is complexand involves at least two specific cleavages; the first occurswithin the gland, in preparation for secretion of activehormone, and the second occurs after secretion and repre-sents inactivation of the hormone.

132. Growth Hormone-Releasing Factor in AcromegalicPlasma. Reduction during Successful Medical Manage-ment. THAD C. HAGEN,* A. M. LAWRENCE,* AND LIDIAKIRSTEINS,* Chicago, Ill. (introduced by Henry T.Ricketts**).

Lack of absolute functional autonomy of growth hormonesecretion in patients with active acromegaly has argued forcontinuing hypothalamic modulation of pituitary growthhormone release in this disease. This laboratory has demon-strated a negative feedback of growth hormone upon itsfurther secretion in normal man; this exists also in acro-megaly but threshold for feedback is set higher. Theseobservations suggest acromegaly may be caused by chronichyperstimulation of somatotrophs by hypothalamic growthhormone-releasing factor. We have reported that plasmafrom patients with active acromegaly enhances in vitro mon-key pituitary growth hormone release 4- to 5-fold over thatobtained with normal plasma. This effect is heat stableand dialyzable. In patients with active acromegaly, medicaltherapies designed to suppress hypothalamic secretory activ-ity have resulted in clinical benefit and lowering of higlhbasal growth hormone levels. Pretreatment plasmas greatlyenhanced monkey pituitary growth hormone release; ac-tivity was substantially reduced in plasmas obtained duringsuccessful treatment of acromegaly. Cessation of therapyresulted in a marked rebound phenomenon in one patient.Finally, in one patient rendered panhypopituitary after 9'Srhypophysectomy, plasma growth hormone-releasing activitygreatly exceeded that of pretreatment samples, a situationreminiscent of the postadrenalectomy syndrome in Cushing'sdisease. Weconclude that growth hormone-releasing activityof acromegalic plasma is substantially higher than that ofnormal plasma; since this activity is heat stable and di-alyzable in both normal and acromegalic plasmas, differencesin growth hormone-releasing potential appears quantitativerather than qualitative. Since successful medical therapy ofacromegaly, aimed at suppressing the hypothalamus, is asso-ciated with reduction of this factor in acromegalic plasma,it is concluded that this activity, present in the peripheralcirculation, is most likely of hypothalamic origin.

133. Increased Stem Cell Response to ErythropoietinInduced by Androgens. WILLIAM N. HAIT,* Dov GOR-SHEIN,* JOANNE H. JEPSON,* EMMANUELC. BESA,* ANDFRANK H. GARDNER, Philadelphia, Pa.

The ability of androgens to increase plasma erythropoietin(ESF) levels, as demonstrated in the rodent, is an inade-quate explanation of the observed erythropoietic response inman. Human anemias, attributed to a shortage of erythroidprecursors, display extraordinarily high ESF titers. ESFtiters often decrease during androgen therapy as recogniz-able erythroid elements in the marrow increase. The follow-ing experiments demonstrate the ability of 19-nortestosteronedecanoate (19-ND) to enhance erythropoiesis by increasing

the pool of erythroid-committed precursors responsive toESF. 19-ND was unable to induce measurable ESF levelsin mice maintained in 60% ovygen environment (HOX).Mice made polycythemic (PCT) by prolonged hypoxia(HCT > 55) were subsequently maintained in HOX. PCT-HOXanimals receiving, respectively, (a) 1.25 mg 19-NDfollowed by 0.5 U ESF 24 hr later, (b) appropriate vehicles,(c) ESF, and (d) 19-ND were compared to identical PCTgroups maintained in HOX. 'Fe incorporation into RBCof PCT-HOX animals receiving vehicle (1.2±0.4%) wassignificantly lower than their PCT counterparts (P < 0.10)indicating further suppression of endogenous ESF by HOX.PCT-HOX animals also exhibited lower (P < 0.05) 'Feincorporation after 19-ND injection. The administration ofa single dose of 19-ND only 24 hr before ESF injectionenhanced erythropoiesis in both systems when compared toESF alone (P < 0.01). These observations imply that: (a)19-ND cannot induce ESF production in HOX, and that(b) it rapidly triggered GO or G1 pluripotential stem cellsinto an erythroid-specific G1 cell cycle responsive to ESF,and that (c) further differentiation of such erythroid-com-mitted cells requires ESF. (This research supported by agrant from the John A. Hartford Foundation.)

134. Control of Breathing in Uremia: Enhanced Ven-tilatory Response to CO2 after Hemodialysis. ROBERTHAMILTON,* PAUL EPSTEIN,* LEE HENDERSON,* NORMANEDELMAN,* AND ALFRED FISHMAN,** Philadelphia, Pa.

The hyperventilation that persists in uremic patients afterhemodialysis remains unexplained. To test whether this isdue to increased sensitivity of the respiratory chemorecep-tors, we studied the ventilatory response to C02 (AV/APco2)in eight patients with uremia. Despite a mean increase inplasma HC03- of 4.86 mmoles/liter, the AV'/APco2 increasedfrom a predialysis value (mean+sE) of 3.160.71 liter/minper mmHg to 5.59±1.25 immediately after dialysis (P <0.005) and decreased to 3.45±0.94 24 hr later. To deter-mine whether these changes were due to (a) dialysis of atoxin depressing ventilation, (b) extracorporeal circulationper se, (c) acid-base changes in the CSF, or (d) an effectof the dialysis disequilibrium syndrome, we produced chronicrenal failure (without metabolic acidosis) in three goatsby f renal infarction. In 51 studies of AV/APco2 in theseunanesthetized goats, uremic animals showed a significantincrease in AV/APco. immediately after hemodialysis (P<0.05), as in uremic patients; nonuremic goats did notshow this response. The AV/APc0o of the goats was slightlygreater after production of uremia (P < 0.05), thus refutingthe thesis of toxic depression. Changes in AV/APco2 wereunrelated to changes in acid-base status of plasma or CSF.However, adding urea to the dialysate, to match the con-centration- in the plasma, prevented the postdialysis increasein AV/APco2 in uremic goats. The data suggest that post-dialysis hyperventilation is due to increased respiratory sen-sitivity to C02. This increase in respiratory sensitivity isnot related to dialysis of a depressant toxin, extracorporealcirculation per se or changes in CSF acid-base status. It isa consequence of the dialysis disequilibrium syndrome. (Re-search supported by NIH Grants HE-08805, HE-00340,5T01 AM-05634, and 5M01 FR-00040.)

41 a

135. Plasma Activator of Plasminogen: Cause of a Fa-milial Bleeding Diathesis. JAMES W. HAMPTON,* FRED B.OLDHAM,* DAVID BANNERJEE,* EKREM KALMAZ,* ANDROBERT DELANEY,* Oklahoma City, Okla. (introduced byColin M. MacLeod**).

12 members of both sexes in four generations of a familywere demonstrated to have enhanced plasma activator ofplasminogen which could be inhibited in vitro or in vivoby epsilon amino caproic acid. Plasma activator activity wasmeasured by a semiquantitative caseinolytic assay modifiedfrom the method of Johnson and Tse. In all 12 membersthe plasma activator was elevated (15-45 streptokinase-likeunits with no activity in normals even after prolonged exer-cise). Plasma activator was confirmed by a fibrin platemethod which utilized test plasma, a partially purified humanplasminogen, and a fibrin substrate which was free of plas-minogen or activator activity. 9 of the 12 patients had ahistory of excessive bleeding and showed serum fibrin degra-dation products, low fibrinogen, short euglobulin lysis times,and reduced antiplasmin activity, which could be interpretedto indicate in vivo activation of the fibrinolytic system. Lowfactor VIII activity in one affected member had promptedthe incorrect diagnosis of classic hemophilia. In four af-fected members plasma activator increased 2- to 7-fold withexercise. The activator is present in the supernatant ofCohn fraction I-0 and in the 0.4-0.5% ammonium sulfateprecipitate. Chromatography on Sephadex G-100 indicates apattern similar to tissue activator. Urokinase antiserumgave no precipitin reaction on double immunodiffusion. Thenature of the plasma activator has previously been subjectonly to conjecture and its presence has not been describedin association with familial in vivo fibrinolysis. (Researchsupported by NIH Grants HE 12316, 2 TO1 AM 05107,and 5-K3-HL-14,227.)

136. In Vitro Synthesis of Porphyrins by Skin. L.HARBER, D. BICKERs,* A. RIFKIND,* AND A. KAPPAS, NewYork.

Although the pathway of porphyrin biosynthesis has beenextensively studied in erythropoietic and hepatic tissue, nosimilar studies have been reported in skin. This investigationwas designed to demonstrate the in vitro synthesis of por-phyrins in skin using y-aminolevulinic acid (ALA) as asubstrate. Induction of the mitochondrial enzyme, 'y-amino-levulinic acid synthetase (ALA-S), was also attempted usingallylisopropylacetamide (AIA) and the 5p8-steroid, etio-cholanolone (Etio). Multiple biopsies were obtained fromfive foreskins removed during routine circumcision of 3- to5-day-old infants. Each biopsy specimen, 13-15 mg wet wt,was incubated in tissue culture. ALA ranging from 104-102M, was added to each incubating specimen. Epidermal por-phyrin production was qualitatively noted by fluorescentmicroscopy in 4-6 hr and quantitated by spectrofluorimetry.A dose-related production of porphyrin ranging from 3.71to 171.4 ulumoles was detected in the biopsy specimens incu-bated with ALA. Porphyrin production after ALA incuba-tion was also noted using the above technique in the skinsof chick embryos, rats, guinea pigs, and adult humans.Intense activity was also seen in plucked human hair folli-cles. Control incubation of ALA without skin, and skin

incubated without ALA, did not lead to porphyrin produc-tion. Similarly, no porphyrin production was observed inskin incubated with ALA precursors, nor was porphyrin-ogenesis induced by AIA or Etio. These studies demon-strate the presence of enzymes necessary to convert ALAto porphyrins in skin. Minimal or no ALA-S activity wasnoted in any of the skin specimens studied. (Research sup-ported by USPHSgrants and an NIH Training Grant.)

137. Plasminogen Kinetics in Man. LAURENCEA. HARKER,GOTTFRIED SCHMER,* AND SHERRILL J. SLICHTER,* Seattle,Wash.

Fibrinolytic activity in vivo was evaluated by measure-ments of plasminogen turnover in 16 normal subjects and in37 patients. Plasminogen for 1"I labeling was prepared fromsingle-donor normal plasma by one-step EACA-agarose af-finity chromatography that involved plasminogen replace-ment chromatography. The protein was biochemically homo-geneous with caseinolytic activity of 30 R-C units per mgand about 85,000 mol wt. Plasminogen turnover was de-termined from concentration divided by tj disappearance ofits 'I label. Plasminogen levels were measured by caseino-lytic, affinity chromatographic, and immunologic techniques.In vivo platelet and fibrinogen turnovers were measuredsimultaneously. Plasminogen-'"I disappeared logarithmicallyin vivo with a half-time of 30.9±2.1 hr. The mean plasmino-gen turnover was 0.11±0.01 mg/ml per day. In four normalindividuals undergoing strenuous exercise intermittently over24 hr, the tj was shortened to 17.2±+1.2 hr, and turnover in-creased twofold (0.21 mg/ml per day). In five patients in-fused with urokinase the tj was reduced to 3.3+1.2 hr. On-going thromboembolism in nine patients was associated withincreased plasminogen destruction (ti 16.3+2.3 hr) and aturnover of 0.34+0.08 mg/ml per day. Parallel changes wereobserved in platelet and fibrinogen kinetics. Other patientswith ongoing "intravascular clot coagulation" (malignancyin 11, malaria in 2, bacteremia in 3) rapidly consumedplasminogen (mean tj 7.7±+1.8 hr, turnover 0.65 mg/ml perday) at rates comparable with the increase in platelet andfibrinogen consumption. In these patients heparin therapyreduced both fibrinogen and plasminogen consumption to nor-mal values, while EACA corrected plasminogen consump-tion to normal without diminishing the destruction of fibrino-gen. On the basis of these data it is concluded that fibrino-lysis is a highly reactive system which responds to a widevariety of physiological and pathological stimuli, and thatfibrinolytic activation appears to be quantitatively coupledwith the rate of fibrin formation. (Research supported byGrants HE-11775 and HE-06242 from the NIH.)

138. A Unique Mechanism for Collagenolysis Associatedwith Invasive Neoplasm in Rabbits. EDWARDD. HARRIS,JR.,* CHARLES S. FAULKNERII,* AND SUMNERWOOD, JR.,*Hanover, N. H., and Rahway, N. J. (introduced by ThomasP. Almy**).

Malignant tumors invade and destroy normal connectivetissues, presumably by enzymatic action. We have partiallycharacterized collagenolytic activity associated with the as-citic form of rabbit V2 carcinoma in New Zealand white

42 a

rabbits. 30 X 106 carcinoma cells were injected into thighmuscle of rabbits. Tumors grew rapidly; in 3-4 wk animalswere sacrificed, tumors excised, and explants incubated at370C in modified Eagle's medium without serum. Rabbittumor collagenolytic activity (RTCA), assayed using nativecollagen substrate, was concentrated from pooled culturemedium. The RTCA was partially purified by chromatog-raphy on Bio-Gel A 0.5 and found to be active at pH 7-8,inhibited by chelating agents, and without caseinolytic ac-tivity. Against collagen in solution at 10'-270C the RTCAdecreased specific viscosity (flsp) to 13-20%, while synovialcollagenase yielded products retaining 40-45% of initial 7sp.Reaction products electrophoresed on acrylamide revealedbands not seen as products of other collagenase action. Ex-amination of SLS aggregates of the reaction products byelectron microscopy have indicated that the collagen moleculeis cleaved by RTCAat a site 57% from the C-terminus, andthat the preferential site of cleavage by other mammalianenzymes-25% from the C-terminus-was spared by RTCA.The discovery of a tumor-associated enzyme with a quali-tatively different mechanism of action from functionallysimilar enzymes in nonmalignant tissues is unusual and mayhave significance in further study of mechanisms of tumorinvasion. (Research supported by grants from NIH, ArthritisFoundation, Easter Seal Foundation, and by the Merck In-stitute for Therapeutic Research.)

139. Mechanism of Action of Type II AntiarrhythmicDrugs. LURA A. HARRISON,* JOHN WITTIG,* AND ANDREWWALLACE, Durham, N. C.

The purpose of these experiments was to test the feasi-bility that the unique properties of the Purkinje-muscle(PM) junction might serve as a useful model for testingthe actions of antiarrhythmic agents. Previous studies haveindicated that action potential (AP) duration is greatest indistal Purkinje fibers (3-4 mm from the PM junction).Premature beats initiated proximal or distal to the area ofmaximal AP duration may propagate with decrement orblock. These studies were performed on the right bundlebranch and distal Purkinje fibers removed from the dogheart, utilizing standard microelectrode techniques. Actionpotential duration was 215±5 msec in proximal right bundle,315±10 msec in distal Purkinje fibers, and 170±7 msec inmuscle. Premature beats with a coupling interval comparableto AP duration at the area of maximal duration conductedwith decrement. Earlier beats were confined to the fibersproximal to the area of maximal AP duration. Lidocaine(2.34 ,tg/ml) shortened AP duration by 100±7 msec at thearea of maximal AP duration, but had little effect on proxi-mal right bundle or muscle. Decremental conduction andblock were abolished by lidocaine. Norepinephrine (NE) 1X 10' M prolonged AP duration, increased decremental con-duction, and prolonged the functional refractory period ofdistal fibers. Propranolol 0.25 - 1.0 mg/liter shortened APduration at the area of maximal AP duration, and abolisheddecremental conduction and block of premature beats. Theseobservations demonstrate that the unique properties of thePM junction may contribute to the genesis of arrhythmiasand that this is the site of maximal effect of type II anti-arrhythmic agents. The action of these agents on the PM

junction is dose dependent and at therapeutic concentrationsthey exert effects which very probably contribute to theirantiarrhythmic effectiveness. (Research supported by USPHS.)

140. Glucose Homeostasis in the Newborn Rat: Role ofGluconeogenic Substrates and Ketones. MOREY HAY-MOND,* IRENE KARL,* ANTHONY PAGLIARA,* AND DAVIDKIPNIS, St. Louis, Mo.

Fasting in the newborn rat results in hypoglycemia whichmay be due to a functional defect(s) in the gluconeogenicenzymic system or to deficient substrate availability. Term ratswere litter mated into fasted (maintained at 300C, constanthumidity) and fed (6 animals per mother) groups. Blood wasobtained by decapitation at timed intervals for the micro-fluorometric determination of glucose, p-hydroxybutyrate,acetoacetate, lactate, pyruvate, glutamate, alanine, and glu-tamine. Fasting glucose increased from 77+3 mg/100 ml(mean ±SEM) to 100+3 mg/100 ml by 6 hr and then fell to26+1 mg/100 ml by 18 hr. Acetoacetate increased slightly(73±8 to 179±+14 ,uM) while f8-hydroxybutyrate remained un-changed (111±7 AM). Lactate (7.2±0.5 mM) and pyruvate(0.198+0.022 mM) elevated at birth, fell throughout the fast(18 hr, 1.2±0.06 and 0.047+0.01 mm, respectively). Alaninealso decreased from 867±42 to 275+9 ,UM. Glutamine initiallydecreased (1141+53 to 824±39 ,uM) and then increased to1036+49 AM. Intraperitoneal administration of alanine (500mg/kg) in fasted animals, rapidly increased blood glucose(A 25 mg/100 ml). Fed animals maintained glucose > 80 mg/100 ml at all times. Lactate, pyruvate, and glutamate weresimilar to fasted animals except lactate at 18 hr was higher(2.4±0.1 vs. 1.23±0.06 mM) and glutamate higher (356+11vs. 275+9.13 ,m). Alanine decreased at 2 hr (867±42 to598+34 ,UM) and then remained constant. Glutamine levelsremained unchanged. B-Hydroxybutyrate (111+7 to 872+5,UM) and acetoacetate (73+8 to 1010±38 ,UM) demonstratedsignificant changes throughout. These studies demonstratethat glucose homeostasis in the fasted newborn rat is notlimited by functional hepatic gluconeogenic enzyme deficien-cies, but rather appear to be due to deficiencies of endogenousgluconeogenic and fatty acid substrates. In fed animals,ketone body formation derived from mother's colostrum(22% fat) appears to be an important source of energywhich may spare glucose. (Research supported by USPHSGrant AM1921 from the NIH.)

141. Hyperlipidemia in Coronary Heart Disease: Lipo-protein Characteristics of a Classification Based on Ge-netic Analysis. W. R. HAZZARD,* J. L. GOLDSTEIN,* H. G.SCHROTT,* A. G. MOTULSKY,** AND E. L. BIERMAN, Seattle,Wash.

One current classification of the hyperlipidemias (pheno-types I-V) is based on the level of plasma triglyceride(TG), content of cholesterol (C) in low density lipopro-teins (LDL), and electrophoretic mobility of isolated lipo-protein (LP) fractions. An alternative classification hasemerged from a genetic analysis based on the distributionand segregation of plasma C and TG in more than 2000relatives of 130 consecutive hyperlipidemic survivors of myo-cardial infarction. Three genetically distinct disorders were

43 a

identified by this approach: hypertriglyceridemia (HTG),hypercholesterolemia (HG), and combined HTG+ HC(Goldstein, Hazzard, et al., Clin. Res., April 1972). In anattempt to relate these genetic disorders to LP phenotypes,isolated LP fractions from index cases of 44 families werecharacterized LDL-C was lowest in HTG (154±50 mg/100ml [mean -SD], n = 16), intermediate in combined HTG+HC (198±47, n=20), and highest in HC (266±52,n = 8). Since close correlations were noted in all geneticgroups between total C and LDL-C (r = 0.93) and totalTG and VLDL-TG (r = 0.93), whole plasma C and TGaccurately reflected LP levels. However, because of theoverlap in individual LDL-C values of index cases with thedifferent genetic disorders, LP phenotyping (using suggested95% normal limits for LDL-C) agreed with the geneticdiagnosis in only 50% of cases-those at the extremes ofHTG (pure type IV) and HC (pure type Ha). Index caseswith combined HTG+ HC showed three different LP pheno-types; IV (40%,), IIb (40%), and Ha (20%). Thus thesestudies demonstrate that the specific genetic disorder in theindividual patient with hyperlipidemia and coronary heartdisease cannot be reliably diagnosed from determination ofhis LP phenotype. They also suggest that the simple mea-surement of plasma TG and C in relatives is likely to yieldmore meaningful information regarding the genetic natureof his lipid disorder than detailed analysis of his lipopro-teins. (This work was supported by grants from NIH.)

142. Histidine-Dependent Zinc Loss, Hypogeusia, An-orexia, and Hyposmia. R. I. HENKIN, H. R. KEISER,* ANDD. BRONZERT,* Bethesda, Md.

Zinc metabolism, gustation, and olfaction were studied infive patients with scleroderma and in three normal volun-teers on and off histidine, given for 2-day periods, in dosesof 8, 16, 32, 48, and 64 g (patients) and 8, 16, and 32 g(volunteers). In patients, serum zinc (mean+l SEM, sg/100ml) decreased directly with histidine administration (baseline 82+8; 32 g, 71±6; 64 g, 50±3) whereas urinary zincexcretion (mean±l SEM in /Lg/24 hr) increased (base line402+25; 32 g, 2552+513; 64 g, 8490±408). In volunteers,serum and urinary zinc exhibited similar changes. Adminis-tration of '-32 g histidine produced elevations of 5- to80-fold in thresholds for four taste qualities (NaCl, sucrose,HCl, and urea) and each patient reported anorexia anddecreased taste acuity (hypogeusia) ; thresholds for threevapors (pyridine, nitrobenzene, and thiophene) increased100-fold above normal, indicating decreased olfactory acuity(hyposmia). Similar, although less severe, changes occurredin volunteers. With withdrawal of histidine a gradual re-turn toward normal occurred in serum and urinary zinc,hypogeusia, anorexia, and hyposmia. Zn+, 100 mg givenorally during administration of 64 g histidine, returnedhypogeusia and hyposmia to normal within 8 hr. Zinc inserum is complexed primarily with macromolecular species(az-macroglobulin and albumin) but also with micromolecu-lar species (histidine and cysteine). In vitro, increased histi-dine concentrations strip Zn++ from Zn++-albumin complexesforming Zn++-histidine complexes. The present studies sug-gest that this ligand shift can occur in vivo and is associatedwith hyperzincuria, subsequent hypozincemia, and depletion

of zinc from receptor sites in gustatory and olfactorysystems.

143. The Metabolism of the Steroid and Glycine Moietyof Glycine-Conjugated Bile Acids in Man. GERSHONW.HEPNER,* ALAN F. HOFMANN, AND PAUL J. THOMAS,*Rochester, Minn.

Bile acids are deconjugated and dehydroxylated duringenterohepatic cycling in man. To measure the magnitude ofthese bacterial biotransformations. cholyl-'H glycine-1-GC,chenodeoxycholyl-3H glycine-1-"C, and deoxycholyl-'H gly-cine-l-"C were synthesized and administered to 12 healthyvolunteers. Bile samples were collected daily for 7 days andthe specific activity decay curve of each moiety was deter-mined. Breath excretion of "C was assessed by interval"CO2 specific activity determinations. The daily fractionalturnover rate of the glycine moiety was about 3 times greaterthan that of the steroid moiety in the three conjugatedbile acids. Assuming six enterohepatic cycles daily, about20%o of the glycine-conjugated pool was deconjugated percycle; the greater part of the steroid moiety thus liberatedwas reabsorbed. "C excretion correlated highly (r = 0.95)with daily fractional turnover of the glycine moiety for allthree bile acids; hence the rate of deconjugation can beeasily assessed quantitatively by interval "CO2 specific ac-tivity determination. The daily fractional turnover of thesteroid moiety of the two dihydroxy bile acids was nearlyidentical and less than that of the cholyl moiety, indicatingthat dihydroxy bile acids are absorbed more efficiently(99%o per cycle) than the trihydroxy acid (95% per cycle).Input of deoxycholic acid from the intestine was about 0.3mmoles/day or one-quarter of all the cholic acid lost fromthe enterohepatic circulation. These data describe (a) a sim-ple method for determining the metabolism of the twomoieties of a conjugated steroid participating in entero-hepatic cycling, (b) a simple method for quantifying bileacid deconjugation in healthy man, and (c) new insightsinto the intestinal conservation of the bile acid nucleus.(Supported by NIH Grant AM6908.)

144. The Assimilation or Production of Gaseous Nitro-gen by Tissues. J. H. HERRON,* H. A. SALTZMAN, B. A.HILLS,* AND J. A. KYLSTRA,* Durham, N. C.

The classic assumption that gaseous nitrogen (N2) ismetabolically inert has been challenged by recent observa-tions of uptake or elimination in the steady state by menbreathing air. Since accurate measurement of a small dif-ference between large volumes of inhaled and exhaled N2is difficult, gas exchange was evaluated in six resting menequilibrated with a normoxic gas (Po, of 150 mmHg)containing only 02%o N2, thereby reducing the error ofmeasurement by about two orders of magnitude. The re-spired concentrations of 02, CO2, and N2 were measuredon a gas chromatograph. Respired volumes were measuredin a Tissot gasometer. The measured volumetric differencesbetween exhaled and inhaled N2 were small, averaging 1.2±2.53 (SD) ml/min STPD when fasting and 1.7±2.99 to2.9±2.17 ml/min STPD 1-5 hr after ingesting protein.These results indicate that, under the conditions tested, anychemical assimilation or production of gaseous nitrogen by

44 a

tissues is insignificant relative to the standard error ofmeasurement (1 ml/min STPD). (Research supported bygrants from NIH (HL07896) and ONR.)

145. The Kinetics of Hepatic Ferritin. CHAIM HERSHKO,*JAMES D. COOK,* AND CLEMENTA. FINCH, Seattle, Wash.

The kinetics of iron storage and release by two liver cellpopulations was measured in the rat after intravenous injec-tion of purified ferritin labeled with 'Fe. 97-99% of solubleferritin was taken up by the hepatocytes, while the com-plex, ferritin-antiferritin, was localized exclusively in theKupffer (RE) cells. Release of radioiron from the Kupffercells was complete in 5 days. Turnover in the hepatocyteswas much slower (2-3%o/day), corresponding to the dailyexchange of iron between plasma and liver as measured byplasma iron turnover. Chemical determinations combinedwith cell differentiating studies showed that over 95%o ofhepatic ferritin iron was localized within the parenchymalcells of the liver. A fivefold increase in the release of radio-iron from hepatocytes occurred when erythropoiesis wasstimulated by phlebotomy, and the liver ferritin concentra-tion fell from 97 to 28 ,ug/g tissue in 5 days. Radioiron lossfrom the parenchymal pool was also induced by the chelat-ing agent desferrioxamine, whereas RE cell radioiron wasinaccessible for chelation. Desferrioxamine was shown tointeract selectively with parenchymal ferritin '9Fe and notwith other radioiron complexes processed by the hepatocyte.Thus, hepatic ferritin is located almost completely in thehepatocyte; this ferritin may be mobilized by chelate actionor in response to physiologic needs. (Research supported byGrant HE-06242 from the NIH.)

146. Elevated Cadmium in Emphysematous Lungs. Rus-SELL N. HIRST, JR.,* H. MITCHELL PERRY, JR.,** MARCOSG. CRUZ,* AND JOHN A. PIERCE,** St. Louis, Mo.

Several lines of evidence suggest a relationship betweencadmium and emphysema. Each cigarette contains about 1 ,tgof cadmium. Occupational exposure to cadmium has beenimplicated in some human emphysema. Inhalation of cad-mium aerosols by dogs and intratracheal instillation of cad-mium into hamsters has induced emphysema. The livers ofpatients dying with emphysema contain significantly morecadmium than the livers of other patients. The present studyreports levels of cadmium in emphysematous lungs. Tissuesfor metal analysis were obtained at autopsy and storedfrozen. One lung was prepared by the Gough technique andgraded for emphysema after the 1970 method of Thurlbeckand associates using a scale of 0-100. 12 subjects withemphysema scores greater than 60 were compared to 12subjects with scores less than 20. Cadmium, zinc, lead,sodium, potassium, and calcium concentrations were deter-mined by atomic absorption spectroscopy in 1-g aliquots ofkidney, liver, and lung after ashing at 450°C for 48 hr.Mean values for cadmium were 82±60 ,ug/g tissue ash inthe emphysematous lungs, and 20±+13 ,ug/g in the controllungs (P < 0.005). Mean cadmium levels also were greaterin the other tissues from emphysematous as compared tocontrol patients (liver 191 vs. 63 ,ug, P < 0.025; kidney 2846vs. 1745 ,ug, P < 0.10). Zinc, lead, sodium, potassium, andcalcium did not differ significantly in the lungs, livers, or

kidneys between the emphysematous and control patients.The finding of an elevated cadmium level in emphysematouslungs merits further study. The source of the cadmium, therate of clearance, and the chemical state all are important.The role of cadmium in the pathogenesis of emphysema inman remains uncertain.

147. Role of the Frank-Starling Mechanism in StrenuousExercise. LAWRENCED. HORWITZ,* JAMES M. ATKINS,*STEPHEN J. LESHIN,* AND STANLEY A. DUNBAR,* Dallas,Tex. (introduced by Jere H. Mitchell).

In six dogs running on a level treadmill, measurementswere made of left ventricular pressure with a solid-statetransducer, aortic flow with an electromagnetic probe, andleft ventricular internal transverse diameter with sonocardi-ometer crystals. The animals ran for 3-min periods at 3-4mph (mild exercise), 6-8 mph (moderate exercise), and 10-13 mph (severe exercise). Heart rate increased from astanding control of 104±7 (meantsEM) beats/min to 183+14 in mild, 222±13 in moderate, and 259±15 in severe ex-ercise. Stroke volume increased from 33-4 cc to 35-4 cc inmild, and 37±3 cc in severe exercise. Mean left ventricularend-diastolic diameter was 30.6 mminitially, 30.8 mmin mild,and 31.9 mmin severe exercise; corresponding end-systolicdiameters were 23.4, 22.9, and 23.1 mm. Mean left ventricularend-diastolic pressure rose 4 mmHg in mild and 12 mmHgin severe exercise, while peak systolic pressure rose 26 and75 mmHg, respectively. Therefore, despite extremely highheart rates a small increase in stroke volume occurs withexercise (P < 0.05 at both mild and severe levels). In mildexertion the increase in stroke volume is due primarily toincreased cardiac muscle fiber shortening resulting in a lowerend-systolic diameter, whereas in severe exertion there isfurther augmentation in stroke volume due to an enlargedend-diastolic diameter-evidence that the Frank-Starlingmechanism is operative. (Research supported by Grant HE14355 from the NIH.)

148. Colfibrate-Induced Antidiuresis. JOAN HOWANITZ,*MARCIA VAN GEMERT,* MYRONMILLER,* AND ARNOLDM.MOSES, Syracuse, N. Y.

Five patients with diabetes insipidus, who were able torelease some antidiuretic hormone (ADH) on dehydration,developed an antidiuresis during treatment with clofibrate 500mg q.i.d. while on random fluid intake (urine osmolality 141vs. 340 mOsm/kg). The antidiuresis was similar to thatwhich results from ADHwith no change in sodium, potas-sium, solute, or creatinine excretion. Clofibrate diminishedthe ability of these patients to excrete a maximally diluteurine after water loading (minimum Uosm 52 vs. 123 mOsm/kg). When oral water loads were given to 14 normal subjectstreated with clofibrate, there was also impaired ability to ex-crete the water and to dilute the urine maximally (minimumU..m 68 vs. 87 mOsm/kg). In these normal subjects im-munoassayable urinary ADHexcretion during water loadingincreased from a mean of < 1 AU/min before treatment to32.3 ,U/min during clofibrate treatment. Similarly, in fourpatients with diabetes insipidus on ad lib. fluid intake, urinaryADH excretion increased from 1.3 AU/min to 2.9 AU/minduring clofibrate treatment. The antidiuretic action of the

45 a

drug in patients with diabetes insipidus was partially over-come by water loading or ethanol administration, both ofwhich are capable of inhibiting ADH release. To furtherstudy the mechanism of action, 40 mg of clofibrate was in-jected into rats with total ADH deficiency (Brattleborostrain) and no antidiuresis occurred. When clofibrate wasinjected into these rats along with exogenous ADH, therewas no augmentation of the action of the ADH. These com-bined clinical and animal data support the conclusion thatclofibrate has no intrinsic antidiuretic action and exerts itsantidiuretic effect by causing release of endogenous ADHfrom the neurohypophysis. (Supported by VA ResearchFunds.)

149. Influence of Renal Transplantation on Aminonucleo-side Nephrotic Syndrome. JOHN R. HOYER,* JEAN RATTE,*AND ALFRED F. MICHAEL, Minneapolis, Minn.

Although pathogenesis of the idiopathic nephrotic syn-drome is unknown, our recent demonstration of recurrentdisease after renal transplantation suggests importance ofhumoral mechanisms. Exchange of renal grafts between ratswith aminonucleoside of puromycin (PA) nephrosis and nor-mal rats was performed to analyze humoral and renal fac-tors. All results reported are urinary protein (mean+ 1 SDmg/24 hours) of groups of three or more grafts on specifiedposttransplant day. Renal transplantation of kidneys fromPA-injected donors (150 mg/kg intravenously) to bilaterallynephrectomized normal rats revealed the following. Kidneysfrom rats with massive proteinuria (495±50 mg) transfer thedisease: 427±66 mg on day 3 (normal transplanted controls31±+14 mg). Kidneys transplanted from groups of PA rats1, 4, 24, and 48 hr after injection (during the 4-5 day induc-tion period before proteinuria) developed massive proteinuriaat the expected time; mean proteinuria in each group wasgreater than 500 mg on day 7. One kidney transplanted 15min after injection excreted 677 mg on day 7. A normal kid-ney's influence on proteinuria was studied in two ways. (a)4 hr after injection of donors with PA (100 mg/kg), kidneyswere transplanted to normal unilaterally nephrectomized(UnNe) rats. Proteinuria was not increased in the first 9days (33±24 mg). After removal of the normal kidney, pro-teinuria developed (351±92 mg 3 days later). (b) Trans-plantation of a normal kidney to UnNe rats given PA 4 hrpreviously resulted in decreased proteinuria (57+16 mg onday 7) compared to UnNe Pa rats (657+101 mg). Thesestudies demonstrate: (a) the nephrotic syndrome can betransferred by transplantation of the nephrotic kidney; (b)crucial biochemical events leading to massive proteinuria aftera latent period of several days occur in the kidney withinminutes after PA injection; and (c) presence of a normalkidney reduces protein excretion of a "nephrotic" kidney.

150. Culture of Human Endothelial Cells Derived fromUmbilical Cord Veins. ERIC A. JAFFE,* RALPH L. NACH-MAN, ANDCARL G. BECKER,* New York.

Recent immunofluorescence studies have shown that humanplatelets and human endothelial cells contain immunologicallysimilar smooth muscle actomyosins which are absent in ma-ture connective tissue. Clinical and experimental observations

suggest the possibility that platelets interact with endothelialcells in maintaining normal blood vessel integrity. To studythese relationships, we have attempted to grow human endo-thelial cells in culture. Human umbilical veins were incubatedwith buffered collagenase solutions. The enzymatically re-moved intimal cells were cultured in medium 199. Compari-son of stained histologic sections of umbilical veins obtainedbefore and after enzyme treatment revealed a selective lossof the endothelial cell monolayer lining the vessel's interior.In tissue culture, on plain and collagen-covered cover slips,these cells grew as a monolayer of closely opposed polygonalcells, 50 X 70 / with prominent intercellular interdigitations.Comparison of these cells with standard fibroblast cell linesgrown in culture revealed gross morphologic differences.Thrombosthenin, the smooth muscle actomyosin, was ex-tracted from human platelets and partially purified. Fluores-ceinated antithrombosthenin (F-AT) brilliantly stained hu-man endothelial cells and vascular and uterine smooth musclein tissue sections. The cultured endothelial cells obtainedfrom the umbilical veins stained brilliantly with F-AT, whilestandard fibroblasts cultures stained only weakly. Mixed cellagglutination reaction studies suggested the presence of ABOblood group antigens on the umbilical vein endothelial cellsand their absence in mature fibroblasts and smooth musclecells in blood vessel media. These results suggest that thecells cultured from umbilical veins are endothelial cells. Cul-ture of these cells offers an opportunity to clarify the inter-action of platelets and vascular endothelium using an in vitromodel of the blood vessel wall.

151. The Effect of Pericardial Tamponade on CoronaryHemodynamics in the Awake Dog. M. M. JARMAKANI,*PHILIP A. McHALE,* AND JOSEPH C. GREENFIELD, JR., Dur-ham, N. C.

In order to evaluate the effects of pericardial tamponadeon coronary hemodynamics, phasic coronary blood flow wasmeasured in chronic dog preparations during changes in epi-cardial pressure induced by infusion of saline into the peri-cardial space. Statham electromagnetic flow probes wereimplanted on the ascending aorta (Ao) and left circumflexcoronary artery (LCCA) in eight adult mongrel dogs weigh-ing 25-35 kg. Pressures were measured in the aorta (AP),pericardium (PCP), left ventricle (LV), through suitablecatheters. All studies were done 7-15 days after the surgicalprocedure. Tamponade was induced with saline infusions of25-cc increments. Maximal pericardial pressure ranged from25-55 mmHg. Pericardial pressures were maximal duringdiastole at low PCP, but were maximal during systole witha high PCP. Total coronary blood flow (CBF) decreasedwith PCP>30 mmHg. Systolic CBF decreased graduallywith increasing pericardial pressure and correlated withcoronary intramural pressure (AP-PCP). At maximumPCP systolic coronary flow was retrograde in five of eightexperiments. Hyperemic coronary flow was abolished at peri-cardial pressures > 25 mm Hg. These data suggest thatduring pericardial tamponade coronary artery intramuralpressure is a major determinant of coronary flow. Intra-myocardial pressure also may play a significant role in regu-lating coronary blood flow. (This research supported byNIH grants HE09711 and HE10179.)

46 a

152. Seizures, H202 Formation, Lipid Peroxides in Brainduring Exposure to Oxygen under High Pressure(OPH). S. A. JERRETT,* D. JEFFERSON,* AND C. E. MENGEL,Columbia, Mo.

Our previous studies showed that OHP caused H202 toform in RBC's causing lysis. The present study was under-taken to determine: (a) if H202 occurs in brain during OHPand (b) any relation to lipid peroxidation and seizures.Chow-fed (CF), tocopherol-deficient (TD), and tocopherol-supplemented (TS) mice were exposed to 100% 02 at 45psia for 60 min. Before OHPmice were given aminotriazole(AT) 2100 mg/kg intraperitoneally. At room air, brain H202of TD mice was higher (34 U) than that of CF mice (27U) (P < 0.01). OHPCF mice showed 41 vs. 71 U for OHPTD mice (P < 0.005). Tocopherol supplementation preventedH202 formation. Lipid peroxides paralleled H202 formation.Seizures occurred in 100% of TD, 50% of CF, and 0% ofTS mice. To examine relationships between lipid peroxida-tion, H202 formed, and seizures. TD rats were given AT andexposed to 100% 02 at 60 psia for 5, 15, and 35 min. Thefirst significant increase in lipid peroxides was at 5 min 0.62/Amole (OHP rats) vs. 0.55 /Amole (rats at room air) (P< 0.025). H202 in OHPrats (79 U) was first greater thanrats at room air (52 U) at 15 min (P < 0.01). At thesetimes there was no clinical evidence of seizures. These data:show that a close relationship exists between H202 forma-tion and lipid peroxidation in brain, both of which precede(clinically evident seizure activity, during OHP. (Researchsupported by grants from the Department of Naval Researchand National Aeronautics and Space Administration, andNational Institutes of Health.)

153. The Role of Acetaldehyde in the Myocardial De-pressant Effects of Ethanol. M. J. JESRANI,* V. SETHI,*M. I. KAHN,* H. A. OLDEWURTEL,* AND T. J. REGAN,**Newark, N. J.

Ethanol has been shown to acutely depress left ventricular(LV) function but the contribution of its first metabolite,acetaldehyde (ACH), is unknown. Using doses to effectblood levels associated with moderate ethanol ingestion, 15intact anesthetized dogs were infused with ACH in saline50 ,ug/kg per min for 2 hr. Peak blood concentrations wereless than 1 Ag/ml by gas chromatography. LV dp/dt maxi-mumdecreased from 2015 to 1775 mmHg/sec (P <0.02).Stroke volume declined despite a significant rise of end-diastolic pressure and volume, measured by indicator dilu-tion. Aortic pressure and heart rate were unchanged. To testthe role of its first metabolite, ethanol was administered in-travenously (1.5 g/kg) over 2 hr to groups A (n = 8) andB (n = 5). The latter were pretreated with pyrazole 6mmoles/kg orally to inhibit the production of ACH. GroupA had a rise of LVEDP from 5.0±0.4 to 9.8±0.6 mmHgand stroke output declined from 14.7±1.4 to 7.6±1.1 ml/min,which persisted for 3 hr without significant change in heartrate or arterial pressure. With similar control hemodynamicvalues, in group B animals the infusion of ethanol producedno significant change in left ventricular end-diastolic pres-sure, stroke output, or dp/dt maximum throughout the ob-servation period when ethanol oxidation was inhibited. Thus,ACH at blood levels found after moderate ethanol intake

has a depressant action on the left ventricle and appears tohave a major role in the myocardial response to ethanol.

154. Fetal Gene Expression in Human Tumor Cells.OLIVER W. JONES,* ANDREWTAYLOR,* AND MARYA. STAF-FORD,* La Jolla, Calif. (introduced by William L. Nyhan).

It has recently been postulated that viral information, in-cluding that portion of the viral genome responsible fortransforming normal cells to tumor cells, is vertically trans-mitted in vertebrates (R. J. Huebner and G. J. Todaro. Proc.Nat. Acad. Sci. U.S.A. 1969. 64: 1087). The postulate alsostates that genes for certain tumor viruses, which later inlife may act as determinants of cancer, may be importantalso as gene determinants during embryogenesis. The viral in-formation manifest through fetal gene expression should bedetectable during a certain period of fetal development. De-repression of the fetal gene activity should occur in tumor cellsand in cells undergoing transformation with oncogenic viruses.Wehave discovered a thymidine kinase variant in normal hu-man fetal tissue that may represent one form of virogene ex-pression. Thymidine kinase has been partially purified fromhuman fetal and postnatal liver, spleen, and fibroblasts. Prop-erties of fetal enzyme different from those in postnatal tissueinclude molecular size, electrophoretic characteristics, pHoptima, heat stability, utilization of phosphate donors, in-hibition by dCTP, and intracellular compartmentalization.Fetal form of thymidine kinase dominates during a briefperiod of fetal development and is not detectable in postnatalcells. The enzyme is detectable in human cancer tissue butnot in benigh tumors. Fibroblast cell lines cultured from hu-man donors were transformed with SV40 DNA virus. Theform of thymidine kinase produced in transformed cellschanges from postnatal to fetal. Thus, a human fetal genefor thymidine kinase may fulfill criteria of virus expression,vertically transmitted as part of the natural genetic apparatusin human cells. (Supported by grants from the AmericanCancer Society and National Institutes of Health.)

155. Survival of Toxoplasma gondii within PhagocyticVacuoles: Evidence for Absence of Lysosomal Fusion.THOMASC. JONES* ANDJAMES G. HIRSCH,** New York.

Observations have been made on the interaction betweenthe obligate intracellular protozoan, Toxoplasma gondii, andseveral types of mammalian cells in tissue culture. Parasitesfreshly isolated from the peritoneal cavity of infected micewere viable (> 95% excluded trypan blue) and were highlyinfectious in vivo for mice and in vitro for Hela cells and formouse macrophages and fibroblasts. The mechanism of en-try of toxoplasmas into the various cell types was phago-cytosis, as revealed by electron microscopic observationsduring the early minutes after toxoplasma-cell contact. Allor nearly all parasites that entered Hela cells divided. Inmacrophages, however, two populations were s-een: approxi-mately 50% of the organisms remained morphologically nor-mal and multiplied after a lag period of several hours, where-as the other half of the toxoplasmas became morphologicallyabnormal within 1/2 to 3 hr and had undergone autolysisby 6 hr. The toxoplasmas which appeared to be degeneratingwere seen in typical phagocytic vacuoles. These vacuoleswere converted to phagolysosomes by fusion with primary

47 a

lysosomes (Golgi vesicles) and with secondary lysosomes(dense granules), as demonstrated by electron microscopicstudies using cells prelabeled with thorotrast or using cyto-chemical acid phosphatase reactions. In sharp contrast, well-preserved or dividing toxoplasmas in all three cell typeswere within vacuoles with an unusual vacuolar membranesurrounded by a layer of rough endoplasmic reticulum or bymitochondria. In macrophages these phagocytic vacuoles re-mained negative for acid phosphatase and for thorotrastmarkers of lysosomal fusion. These studies thus suggest thatsurvival of toxoplasmas within cells may be related to theabsence of fusion of lysosomes with parasite-containing vacu-oles. (Research supported by a grant from NIH.)

156. Renal Enterobacterial Infection as the Cause ofEquine Infectious Anemia. ROBERT K. JORGENs* ANDKATHLEENE. ROBERTS,** Owego, N. Y.

More evidence has accrued to assign renal infection as thecausative factor of several degenerative diseases of man (J.Clin. Invest. 1970, 1971. Abstracts.) Treatment universallyreversed the clinical situation. Comparative studies on equinesrevealed that one of their degenerative diseases, equine infec-tious anemia, was causally related to renal enterobacterialinfection. These studies revealed that this disease is neitherinfectious nor solely equine. Horses were successfully treatedand the renal lesion as well as the anemia reversed by avaccine identical with that used in humans.

157. Immunologic Studies of Human BronchoalveolarCells and Secretions. P. F. JURGENSEN,* G. N. OLSEN,* J.E. JOHNSON,* E. W. SWENSON,* E. M. AYOUB, AND R H.WALDMAN,* Gainesville, Fla. (introduced by L. E. Cluff**).

The cellular and humoral immune responses to immuniza-tion with influenza virus vaccine either subcutaneously or byaerosol were studied. Bronchoalveolar (BA) lavage, nasalwashing, and venipuncture were carried out on 21 volunteersbefore and after immunization. Cell-mediated immunity wasassessed by studying the inhibition of macrophage migration(IMM) using the BA lymphocytes, and determining the f-N-acetylglucosaminidase (BNAG) level of the alveolar mac-rophages as a measure of lysosomal acid hydrolase activity.Immunoglobulin and antibody levels were measured on thesupernatants. The results indicate that the IgG: IgA ratiosare different in the various body fluids studied, being about5: 1 in serum, 1: 2 in nasal secretions, and 3: 1 in BA wash-ings. Influenza-neutralizing antibody titers in BA fluid andin nasal washings were highest after aerosol immunization,with the mean titer rising 5-fold in the BA fluid of volun-teers immunized by aerosol, and 3-fold in those immunizedsubcutaneously. The serum antibody titers, however, rose 4-fold and 25-fold respectively. After aerosol immunization,BA lyphocytes gave 30% inhibition in the IMM test, whilecirculating lymphocytes gave 5% inhibition. After subcutane-ous immunization, a reverse pattern was seen, with BAlymphocytes giving 16% inhibition, while circulating lympho-cytes gave 28% inhibition. The level of BNAGin alveolarmacrophages rose significantly after aerosol, but not aftersubcutaneous immunization. These data confirm and extendthe results of several studies in recent years which indicatethat the lower respiratory tract is a relatively independent

immunologic organ. (This work was supported by NIHGrant AI10295-01 and by the Irwin Strasburger Founda-tion.)

158. Prolonged Steady-State Protein and Secretory Im-munoglobulin A Synthesis and Secretion by IntestinalMucosa. MARTIN F. KAGNOFF*AND ROBERTM. DONALDSON,JR., Boston, Mass.

Investigations into intestinal synthesis and secretion ofimmunoglobulins are currently hampered since the limitedviability of available in vitro preparations allow steady-stateconditions for only 1-2 hr. Wehave examined de novo syn-thesis and secretion of protein and secretory immunoglobu-lin A (sIgA) by rabbit jejunal mucosal biopsies culturedin an organ culture system. Incorporation of leucine-14C intotissue protein by these biopsies was linear with time, andthe rate of protein synthesis was the same at the beginningand end of 24 hr of incubation. After a 3-6 hr lag, biopsiessteadily secreted radiolabeled protein into the incubationmedium. In "pulse-chase" experiments, the kinetics of pro-tein turnover were linear and radiolabeled protein wassteadily secreted during the 6 hr "chase." Cyclohexamideinhibited protein synthesis by more than 99%. RadiolabeledsIgA secretion by single biopsies was assayed after 3-24 hrincubation using anti-rabbit sIgA covalently linked to asolid phase immunoadsorbant. Increases in specific bindingto the bromoacetylcellulose-antibody immunoadsorbant werelinear with time between 6 and 24 hr. After 24 hr, com-ponents of SIgA composed 13-22% of the total radiolabeledprotein secreted by the intestinal mucosa. To document thatsIgA-"C was present in the incubation medium as an intactmolecule containing secretory component (SC), sIgA-14Cwas purified from pooled concentrated medium by sequentialcolumn chromatography on Sephadex G-200, DEAE cellu-lose, and Sepharose 4B. This purified material uniformlygave a single radioactive precipitin arc by immunodiffusionand immunoelectrophoresis with specific antibodies againstsIgA, SC, and a-chain. These results demonstrate that, whenincubated in an organ culture system, intestinal mucosalbiopsies synthesize and secrete protein and sIgA understeady-state conditions for prolonged periods. (Researchsupported by NIH Grants AM05025 and AM 11867.)

159. A Possible Mechanism of Connective Tissue Defectin Homocystinuria. ANDREWH. KANG,* Boston, Mass.(introduced by Donald B. Martin).

Homocystinuria is an inherited disorder of methioninemetabolism characterized biochemically by homocysteinemia,homocystinemia, and homocystinuria. Clinically, the affectedpatients manifest wide-spread deformities and malfunctionsof connective tissue including joint laxity, kyphoscoliosis,genu valgum, severe osteoporosis, ectopia lentis, and vas-cular abnormalities. Although the basic defect has beenshown to be the deficient activity of the enzyme, cysta-thionine synthetase, the mechanism for connective tissuealterations has not been elucidated. In view of the clinicalsimilarities to the many of the features observed in osteo-lathyrism, the possible influence of the metabolites on theproperties of collagen was investigated. Purified rat andchick skin tropocollagen, rich in cross-link precursor alde-

48 a

hydes, solubilized in 0.05 M Tris-0.16 M NaCl, was incu-bated at 370C in the presence of homocysteine, homocystine,or methionine to form insoluble native-type fibrils, and afterthe fibril was completed, each sample was examined forreversibility of insolubility upon cooling to 4C. The resultsobtained clearly showed that the collagen solution contain-ing homocysteine fails to form insoluble fibrils. Furthermore,much less of the reducible amino acid residues, 8-hydroxyly-sinonorleucine, the aldol condensation product, and the post-histidine compound, which have previously been shown tobe involved in collagen cross-linking, are formed in thepreparations containing homocysteine as compared with thecontrol and the samples containing methionine or homo-cystine. It is concluded that homocysteine interferes with theformation of the reducible intermolecular cross-links whichhelp stabilize the collagen supramolecular complex. These invitro results were tentatively confirmed by analyses of theskin biopsy samples from two patients with vitamin B.unresponsive homocystinuria, which showed similarly de-creased content of the cross-link-related compounds. (Sup-ported by grants from NIH and National Science Foun-dation.)

160. Evidence of a Genetic Defect in Pyruvate Oxidationin Three Patients with Friedreich's Ataxia. R. A. P.KARK,* J. P. BLASS,* AND W. K. ENGEL,* Los Angeles,Calif. and Bethesda, Md. (introduced by C. M. Pearson.)

17 patients with various spinocerebellar degenerations werescreened for abnormalities of oxidative metabolism. Muscleslices from 9 patients oxidized 2-14C-pyruvate to 14CO0 atrates 6 standard errors below the mean for 17 normal anddisease-control patients (1.23+0.13 umoles hr' g1 protein;mean-sEM) as did muscle from 6 of 16 patients with motorneuropathies. There was no relation between pyruvate oxi-dation and the mild-to-moderate neuropathic changes evidentin muscle by histochemical study. Three of the nine patients(from two families) whose clinical diagnoses were Fried-reich's ataxia were investigated further. Blood pyruvatelevels were elevated in the basal state. Skin fibroblasts fromthese three were studied after some 25 serial passages intissue culture. The test fibroblasts oxidized pyruvate to MCO2more slowly than did cells from nine normal and disease-controls when oxidation was stimulated by 8 /AM dinitro-phenol. Results were similar with 1-1'C-pyruvate (9.7-14.9vs. 19.3-31.3 Mmoles hr1 g' protein for controls) and with2-14C-pyruvate (2.52-4.14 vs. 4.86-10.62 for controls). Therates of oxidation of 1-14C-palmitate and of 1-14C-glutamatewere within the lower range of the controls. The lower ratesof oxidation after multiple serial passage in culture probablyreflect genetically determined deficiencies in the fibroblastssince material present in the original sample would havebeen diluted by a factor of 10. or more. Some patients with"Friedreich's ataxia" appear to have a biochemical defectwhich is genetically determined, which is as yet unidentified,but which is reflected in reduced oxidation of pyruvate.(Supported by NICHD Grant HD-05061.)

161. Effect of Hypothyroidism on Sodium Reabsorp-tion (TN.) and Renal Na-K-ATPase. ADRIAN I. KATZ*

AND MARSHALLD. LINDHEIMER,* Chicago, Ill. (introducedby Leslie J. DeGroot).

Decreased Na-K-ATPase in crude homogenates of severalorgans from thyroidectomized (Tx) rats has been reportedand correlated with reduced oxygen consumption and in-creased sodium content of corresponding tissue slices. Thepresent study was designed to evaluate further the effect ofhypothyroidism on Na-K-ATPase in the kidney, and todetermine whether its decrease is the direct effect of thyroidhormone deficiency or represents only an adaptive responseto decreased reabsorptive sodium load. Age- and weight-matched Tx (Free Thyroxine Index = 1.4±0.1 SEM) andcontrol (FTI = 7.3±0.6) rats were studied simultaneously.TN. was lower in Tx rats than in controls (110.4+9.2 vs.152.3+9.0 gEq/min per 100 g, P < 0.005) and was accom-panied by quantitatively similar differences in renal micro-somal Na-K-ATPase (63.6±4.0 vs. 89.2±7.9 gmoles Pil/mgprotein per hr, P < 0.01). This decrement was greater inouter medulla than cortex and appeared specific, since en-zymes not involved in sodium transport (Mg++-ATPase,glucose-6-phosphatase, 5'-nucleotidase) remained unchanged.KmNa (22 mEq/liter) was similar in Tx and controls.Triiodothyronine (10-' to 10-' mole/liter) had no effect invitro on microsomal Na-K-ATPase of either Tx or controlanimals. In order to increase filtered and reabsorbed sodiumper kidney, Tx rats were uninephrectomized. 3 wk later TN.in the remaining kidney increased markedly, exceeding thatcalculated per kidney of sham-operated controls. Despitepersistence of the hypothyroid state, Na-K-ATPase activityalso increased significantly when compared to sham-operatedTx animals (P < 0.005), and was similar to that of intactcontrols. The decrease in renal Na-K-ATPase in thyroiddeficiency appears to be mediated primarily through de-creased tubular sodium transport, since it can be reversedby increasing TN. despite continuing hypothyroidism. (Re-search supported by a grant from NIH.)

162. Metabolism of Proinsulin, Insulin, and C-Peptide.ADRIAN I. KATZ* AND ARTHURH. RUBENSTEIN,* Chicago,Ill. (introduced by Richard L. Landau**).

Proinsulin comprises a significant proportion of circulatingbasal immunoreactive insulin, and its concentration variesindependently of insulin after beta cell stimulation. How-ever, the relative contribution of secretion and metabolismin determining proinsulin levels is uncertain. Similarly, themetabolic fate of C-peptide is unknown. Urinary clearance,renal extraction, and metabolic clearance rates (MCR) ofbovine proinsulin, insulin, and C-peptide were measured inseparate experiments in rats. Each peptide was infused atdifferent rates to obtain a wide range of steady-state plasmalevels. Concentrations in arterial and renal venous bloodand in urine were measured by radioimmunoassay. Renalhemodynamics were evaluated by measuring GFR (inulinclearance) and renal plasma flow (PAH clearance andextraction). Urinary clearance of both insulin (7.2±0.6 ALl!min) and proinsulin (3.4+0.6 tl/min) was low, neitherexceeding 0.2% of simultaneously measured GFR. However,the kidney extracted significant amounts of both peptides(insulin, 40.2±2.0%; proinsulin, 34.9±1.57o). Insulin MCRwas much higher than that of proinsulin (16.4±0.4 ml/min

49 a

vs. 6.7±0.3 ml/min; P < 0.001); both were independent ofplasma concentration. Renal extraction contributed 53.9±6.2%o of proinsulin MCR, but only 25.2±3.6 of insulin MCR(P < 0.005). C-peptide was handled similarly by the kidney(12.0+2.4 ;LI/min urinary clearance; 41.3±3.5%o extraction);its MCRwas slow, approaching that of proinsulin. We con-clude the following. (a) Renal handling of proinsulin,insulin, and C-peptide is similar, being characterized byhigh extraction rates and very low urinary clearances. (b)Proinsulin and C-peptide MCRare comparable, approximat-ing 40%, that of insulin. This difference is probably due tosubstantial extraction of insulin, but not of the other twopeptides, by the liver. (c) The kidney is a major site fordegradation of all three peptides, contributing 25%o ofinsulin MCRand more than 50%o of the MCRof proinsulinand C-peptide.

163. Episodic Secretion of Aldosterone in Supine Man.FRED H. KATZ,* PEGGYROMFH,* AND JUDITH A. SMITH,*Denver, Colo. (introduced by Karl E. Sussman).

Frequent blood sampling has shown that cortisol is se-creted in bursts occurring during only a portion of the dayin response to bursts of ACTH. Since aldosterone is alsoknown to vary in "steady-state" subjects and to respond toACTH, we have measured plasma aldosterone, renin activ-ity, and cortisol by sensitive radioimmunoassay and com-petitive protein-binding methods every 30 min in healthysupine individuals over 24 hr. The subjects were maintainedon 130 mEq sodium-70 mEq potassium diets for 4 daysand acclimated to the metabolic ward. Blood was drawnfrom an indwelling needle. In the first two subjects studiedplasma aldosterone varied from 0 to 296 pg/ml. These menhad six and eight secretory bursts of cortisol and four andsix of aldosterone, respectively. The aldosterone peaks in theprewaking hours were synchronous with those of cortisol.The lesser aldosterone peaks that occurred later in the dayappeared to lag somewhat behind those of cortisol. Activecortisol secretion occurred 34 and 46% of the time andaldosterone secretion 21 and 31%o of the time. Aldosteronesecretion rates in these supine subjects, calculated by themethod of Weitzman et al., were 41 and 101 ,ug/24 hr.Plasma renin activity also varied within the normal supinerange. Secretion of renin seemed concordant with the othertwo hormones during the major secretory episodes but largerenin peaks later in the day were not followed by corre-sponding aldosterone peaks. These results demonstrate epi-sodic basal secretion of aldosterone for the first time andsuggest that the ACTH-cortisol system has an importantrole in this periodicity. (Supported by grants from ThePopulation Council, NIH, and Veterans Administration.)

164. A Soluble Receptor for IF-B.1 from Guinea Pigand Human Ileum. MAX KATZ* AND BERNARDA. COOPER,Montreal, Canada.

Absorption of vitamin B12 in many species requires in-trinsic factor (IF) which binds B12 and attaches specificallyto a receptor in the intestine. We have solubilized a factorfrom guinea pig and human ileal mucosa which specificallybinds IF-BE2 complex. Mucosa from guinea pig ileum wassonified, treated with Triton X-100, and 100,000 g's super-

natant was dialyzed and lyophilized. Portions of this extractwere incubated with 'CoB,2 and gastric juice from subjectsand from a patient described previously, whose gastric juicecontained a nonfunctional IF unable to bind to guinea pigmucosal homogenate (GPMH). Filtration on SephadexG-200 revealed some radioactivity excluded from the gelin the case of the normal, but not the abnormal, gastricjuice. Fractionation on Sepharose 4-B of normal IF-7CoB,2and extract revealed several peaks of radioactivity withmolecular sizes ranging from 400,000 to 1,600,000. Theseprocedures were applied to terminal ileal mucosa obtainedfrom a single human subject. A material was present in thelyophilized extract of the 100,000 g supernatant which boundIF-fCoBm2 and was excluded from Sephadex G-200. Disso-ciation of '7CoB,2 from human IF in the presence of excess'CoB1, was faster (ti 150 min) at 37°C than when the"7CoB=-IF was attached to GPMH. No dissociation of thelatter was observed over 360 min. The macromolecular fac-tor described above, which bound only the normal IF-B12complex, probably is soluble receptor for IF-B,2. WhenIF-B1, complex was bound to receptor on GPMH, the rateof dissociation of 'CoB1, was much less than when not sobound. (Research supported by grant from MRCof Canada.)

165. Site of Association of Protein, Lipid, and Carbohy-drate Components of Low Density Lipoprotein in Intes-tinal Epithelial Cells. JACQUES I. KESSLER, Montreal,Canada.

The cellular equivalent of microsomes is the site of syn-thesis of low density lipoprotein (LDL) by intestinal epi-thelium (Kessler. 1970. J. Biol. Chem. 245: 5281). The siteof association of the protein, lipid, and carbohydrate com-ponents of LDL is unknown. To elucidate this, corn oilcontaining palmitate-"C was fed to rats and a reasonablypure suspension of lipid-loaded intestinal epithelial cells wasobtained by ultracentrifugal flotation. The cells were incu-bated with 'H-labeled amino acids and glucosamine-j4C,with and without puromycin. Medium LDL of solubilizedrough (RSM) and smooth surface membranes (SSM)was isolated by immunoprecipitation and protein, lipid,and hexosamine radioactivities assayed. Aliquots of RSMand SSM were also incubated in fortified media contain-ing 'H-labeled amino acids and glucosamine-14C. Lipidradioactivity occurred in medium LDL at 10 min and re-mained reasonably constant for 120 min. Protein and hexos-amine radioactivity in medium LDL occurred at 40 min,increased linearly up to 80 min, and remained constantthereafter. LDL protein radioactivity of RSMwas higherthan that of lipid and hexosamine and at 30 min all radio-activities approached those of medium LDL. LDL radio-activities of SSM paralleled those of medium LDL. Incontrast to RSM, protein radioactivity of SSMapproachedthat of medium LDL after 80 min. Puromycin markedlydecreased the radioactivities of RSM, SSM, and mediumLDL. Inhibition of lipid incorporation was not manifestuntil 40 min. Protein radioactivity of incubated RSMwashigher than that of lipid and hexosamine. The opnosite wasobtained with SSM. Radioactivities of medium LDL weregreater when RSMand SSMwas incubated together. Theseresults indicate that the aproprotein of LDL is synthesized

50 a

by the RSM and subsequently associates with lipid andcarbohydrate in the SSM before release.

166. Collagenase-Sensitive Adrenocorticotropic Hormone(ACTH) Receptor in Isolated Adrenal Cells. ABBAS E.KiTABCHI* AND DONALDB. WILSON,* Memphis, Tenn. (in-troduced by Gene H. Stollerman**).

We have previously shown that isolated adrenal cellsprepared by a trypsin digestion method (IACT) producecorticosterone in response to cyclic AMP (c-AMP) andmicrounit quantities of ACTH. Adrenal cells prepared bycollagenase digestion method (IACC), however, exhibit aloss of ACTH response without alteration of c-AMP-induced steroidogenesis. In order to study the mechanismof action of ACTH in these two cell preparations, adenylcyclase (AC) was measured in IACT and IACC prepara-tions in response to ACTHand NaF by measuring conver-sion of ATP-a&P to cyclic 3'5'-AMP-'P. Since neitherpreparation contained an appreciable amount of cyclic nu-cleotide phosphodiesterase, AC was measured in these cellsin the absence of methyl xanthines. ACTHand NaF stimu-lated AC in IACT preparations in a dose-related fashion.With 20 MAU of ACTH and 20 mmNaF, AC activity in-creased from 44 to 152 and 136 pmoles/10 min per adrenal,respectively. AC activity in IACC, however, was below 10pmoles in the control and experimental conditions. Lack ofeffect of ACTHin IACC was shown not to be due to resid-ual collagenase in the preparation. We conclude that (a)loss of ACTH response in IACC has been demonstratedwith concomitant diminution of AC activity; (b) trypsin-digested cells exhibited ACTH concentration-dependent ACactivity; and thus, (c) collagenase treatment impairs ACTHbinding and/or ACTH-activated AC in isolated adrenal cells.(Supported by grants from Veterans Administration andNIH.)

167. Metabolism of Thyroid Hormones by PhagocytizingHuman Leukocytes. S. J. KLEBANOFFAND W. L. GREEN,*Seattle, Wash.

Myeloperoxidase (MPO), H202, and an oxidizable co-factor such as iodide, bromide, chloride, or thiocyanate ionsform an antimicrobial system which may be operative inintact leukocytes. When iodide is the halide employed inthe isolated system, iodination of the microorganisms occurs.Iodide also is converted to organic form by intact phago-cytizing leukocytes. Thyroxine (T4) and triiodothyronine(T3) can replace iodide as the cofactor in the isolated MPO-mediated antimicrobial system. We have, therefore, studiedthe metabolism of "I-labeled T4 and T3 in vitro by intacthuman leukocytes and by purified MPO. Resting leukocytesmetabolize T4 and Ts very slowly. Addition of phagocytiz-able particles, such as preopsonized zymozan or Lactobacillusacidophilus, stimulates hormone metabolism about 15-fold.The chief products are iodide and material which remainsat the origin on paper chromatography. The latter consists,in part at least, of iodinated protein. MPOand H202 alsorapidly degrade T4 and T3, forming iodide and origin mate-rial. Degradation by intact phagocytizing leukocytes or byMPO and H202 is inhibited by propylthiouracil, azide,cyanide, and ascorbic acid. Methimazole also inhibited the

formation of iodinated origin material; however, the rateof deiodination of To or Ta is increased by this agent. Leu-kocytes of two patients with chronic granulomatous diseasedegraded T4 and T8 much more slowly than normal leuko-cytes, whereas intermediate rates were obtained with leu-kocytes from a patient with hereditary MPO deficiency.We conclude that the metabolism of T4 and To by intactphagocytizing leukocytes may be the major source of iodidefor the cell and may be responsible for the accelerated thy-roid hormone turnover observed during bacterial infections.(Research supported by USPHS Grants AI 07763 andAM-1000.)

168. Countercurrent Multiplication System-New Model.JUHA P. KOKKO* AND FLOYD C. RECTOR, Dallas, Tex.

The present study reports a new model for countercur-rent multiplication in which both the descending limb ofHenle (DLH) and thin ALH operate as purely passiveequilibrating segments. This model is based in part ontransport characteristics obtained by perfusing isolated seg-ments of rabbit DLH in vitro. Water and solute transportwas examined by imposing an osmotic gradient across theDLH by addition to the bath either (a) 112 mmoles/literurea, or (b) combination of 204 mOsm/liter NaCl and 119mOsm/liter urea. The rise in osmolal and volume markerratios of collected to perfusion fluids were not statisticallydifferent. In both cases simultaneous urea influx was calcu-lated from the appearance rate of the urea-"C added to thebath. The collected urea concentration increased by 5.2+2.5mmoles/liter in (a) and 7.8+3.0 mmoles/liter in (b). Underboth conditions, intraluminal [NaCl] rose significantly abovethe ambient surroundings. For the passive equilibrationmodel to be operative the following additional membranecharacteristics should be present: thin ALH, more permeableto NaCl than urea; thick ALH, active NaCl transport;outer medullary and cortical collecting duct (CD), ureaimpermeable; inner medullary and papillary CD, urea per-meable. The salient feature of the proposed model is thatthe energy generated by active outward NaCl transport bythick ALH (expressed as high urea concentration in outermedullary CD by virtue of water abstraction) is transmittedto the papilla (by way of urea diffusing down its concen-tration gradient). In turn, papillary interstitial urea ab-stracts water out of DLH generating high intraluminalNaCl concentrations which allows the entire system tooperate by passive diffusion of NaCl out of the thin ALH.By this model, the observed medullary concentration gra-dients are developed without invalidating the mass balanceequations. (Research supported by grants from NIH andTexas Heart Association.)

169. Evidence That the Inhibition of DNA Replicationin Tumor Cells Is Not a Specific Function of CyclicAMP. J. KOWAL, Cleveland, Ohio.

Cyclic AMP has been reported to inhibit replication incultures of neoplastic cells and induce reversion towards a"normal" morphology. Up to 90%o inhibition of thymidine-'H(TdR-'H) incorporation into the DNA of a clonal line ofACTH-responsive mouse adrenal tumor cells by cyclic AMPand ACTH has been described (Masui and Garren. 1971.

51 a

Proc. Nat. Acad. Sci. 68: 3206). The effect of ACTHwasattributed to its stimulation of intracellular cyclic AMPlevels. We have investigated this in clonal lines and pri-mary tumor cultures. For most experiments, cultures intheir growth phase were incubated with ACTH or othercompounds from 4 to 6 hr and pulsed with TdR-'H duringthe final 1-2 hr. Maximal steroidogenesis ( 10-fold) canbe achieved with 2 mU/ml of ACTH. A decrease in TdR-'Hincorporation into DNAwas not perceived until 20 times thislevel was reached. At 400 mU/ml, incorporation of TdR-'Hwas decreased only to 75% of control cells. This effect wasreversed by replenishing the cells with fresh medium whenadding the TdR-'H. 10 mmcyclic AMPmaximally stimu-lates steroidogenesis; TdR-'H incorporation is inhibited 30-40%o. Adenosine stimulates steroidogenesis in the clonal line,but has no effect on primary cultures. However, adenosineinhibited TdR-'H incorporation > 95% in both. Although10 mMcyclic CMPhas the same steroidogenic activity as10 mmcyclic AMP, TdR-'H incorporation was inhibited< 10%o by cyclic CMP. Cytidine inhibited TdR-'H incor-poration > 60%. However, cyclic CMPis not metabolizedby phosphodiesterase. The data suggest that this inhibi-tion is not a property of cyclic AMP per se and may bedue to adenosine or a common metabolite of adenosine andcyclic AMP as well as other nucleosides. The inhibitioncan be dissociated from the stimulation of steroidogenesisand intracellular cyclic AMP production. Possible reasonsfor the remarkably high concentrations of ACTH requiredfor this effect are still under investigation.

170. Decreased Lysyl-Protocollagen Hydroxylase Activ-ity in Fibroblasts from a Family with a Newly Rec-ognized Disorder: Hydroxylysine-Deficient Collagen.STEPHENM. KRANE, SHELDONR. PINNELL,* AND RICHARDW. ERBE,* Boston, Mass.

We have recently described two sisters, ages 9 and 12,with scoliosis, recurrent joint dislocations, and hyperextensi-ble skin and joints. Parents and older sister appeared nor-mal. Analysis of dermal collagen (normal by electron mi-crosopy) revealed marked decrease in hydroxylysine (about5%o normal) with normal content of hydroxyproline andother amino acids. Hydroxylysine-deficient dermal collagenshowed increased solubility in denaturing solvents. Hydroxy-lysine was also reduced in fascia and bone, and urinaryratios of hydroxylysine/hydroxyproline were low. We rea-soned that the defect might be due to deficient lysyl-protocollagen hydroxylase. In the present study skin fibro-blasts cultured from family members and controls wereharvested 48 hr after reaching confluence. Cell lysates wereassayed for lysyl-protocollagen hydroxylase using chicktibial 14C- or 'H-labeled lysine-protocollagen as substrateby measuring 14C- or 'H-labeled hydroxylysine formed. Inaddition, some lysates were also assayed for prolyl-proto-collagen hydroxylase using '4C-labeled proline-protocollagenby measuring hydroxyproline-'4C formed. Whereas in theaffected sister, lysyl-protocollagen hydroxylase in fibroblastlysates averaged 11 and 12%o of controls respectively, theirlevels of prolyl-protocollagen hydroxylase were not reduced,consistent with the specificity of the defect in collagenhydroxylation. Preliminary data reveal no alteration in

apparent affinity for either the lysine-protocollagen or a-ketoglutarate substrates in the mutant cell. Fibroblast lysyl-protocollagen hydroxylase was within the range of controlsin the unaffected sister and about half control in the mother.This novel disorder is the first inborn error of human col-lagen metabolism in which both the deficient product (hy-droxylysine) and enzyme (lysyl-protocollagen hydroxylase)are identified. (Supported by NIH and Gebbie Foundationgrants.)

171. Selective Measurement of Lipoprotein Lipase andHepatic Triglyceride Lipase in Postheparin Plasma.RONALDKRAUSS,* ROBERTLEVY, HERBERTWINDMUELLER,*LAURA MILLER,* AND DONALD FREDRICKSON,** Bethesda,Md.

We have now obtained additional evidence that tri-glyceride lipase activity released into plasma by heparincontains not only lipoprotein lipase (LPL) but a differentlipase of hepatic origin (HTGL). Using rat and humantissues in an assay with triolein-"C (7.5 mg/ml), we foundthat preincubation with protamine sulfate inhibited morethan 90% of adipose tissue LPL, but less than 10% ofHTGL. With rat plasma as enzyme source, protamine in-hibited 43% of the heparin-released triglyceride lipase activ-ity. In completely hepatectomized rats plasma lipolytic ac-tivity was inhibited 91%o by protamine. With varying degreesof subtotal hepatectomy, protamine-inhibitable activity wasconstant, while residual activity was directly proportionalto the amount of liver remaining. Thus in this assay post-heparin lipolytic activity that is protamine inhibitable pro-vides an accurate measure of LPL and residual activityrepresents HTGL. This selective assay was then used toreevaluate clinical states previously associated with abnor-malities in total postheparin lipolytic activity. In 26 normalsLPL activity was 2.7±1.0 U (AM FFA/ml per hr), andHTGL was 11.4±3.1. In 10 patients with type I hyper-lipoproteinemia, LPL was 0.30+0.15 and HTGL 8.9±3.3,whereas in 12 type V patients with comparable triglyceridelevels, both activities were normal. In three normals low-fat diets caused an average 66% reduction of LPL with a16%o increase of HTGL. Oral contraceptives given to twonormals and progesterone to one type V patient increasedLPL an average of 67%; HTGL increased only 7%. Inthree hypothyroid patients, LPL was normal (1.8-4.7), butHTGLwas low (1.6-5.2). These data point to the necessityof distinguishing LPL and HTGL in the clinical assessmentof plasma postheparin lipolytic activity.

172. The Endocrinologically Normal Enlarged Sella.DOROTHYT. KRIEGER, New York.

Previous reports have indicated abnormalities of growthhormone responsiveness in most patients with radiologicalevidence of increased sellar size, even where tests of gonadal,thyroid, and adrenal function have been normal. In review-ing subjects with demonstrable sellar enlargement (limitedto almost exclusively intrasellar involvement) studied byour laboratory, eight subjects were encountered with normalgrowth hormone responsiveness to insulin hypoglycemia. Itwas then realized that these subjects were unique in thatthey were all female, moderately to massively obese, all but

52 a

one had had at least one pregnancy, and their ages rangedfrom 22 to 61 yr. None had clinical evidence of endo-crinopathy and all manifested age-appropriate urinary go-nadotropin titers, normal basal levels of urinary 17-hydroxy-corticosteroid, 17-ketosteroid, and plasma protein-boundiodine, and normal cortisol responses to hypoglycemia. Visualfields were normal in all instances save for two patientswho demonstrated suggestive defects when tested with asmall red object. None had evidence of increased intra-cranial pressure. Electroencephalograms and brain scanswere normal in all. The sellar enlargement in all but oneinstance was unassociated with evidence of clinoid erosion.Cerebral angiography or pneumoencephalography were notwarranted in all instances in view of the absence of symp-toms necessitating such studies for diagnostic or therapeuticreasons. Where such studies were performed there was noevidence of suprasellar extension. The presumptive etiologyof the sellar enlargement in these cases, based on radio-logical criteria, was that of chromophobe adenoma. In theabsence of pneumoencephalographic studies in all instances,the possibility of an "empty sella" syndrome cannot be ex-cluded. The present findings indicate that evidence of markedsellar enlargement is compatible with a normal endocrinestatus according to currently employed parameters of evalu-ating pituitary function.

173. An Enzymatic Explanation for Increased HepaticTriglyceride Formation in Rats Fed High Sugar Diets.ROBERTG. LAMB* ANDHAROLDJ. FALLON, Chapel Hill, N. C.

Serum triglyceride (TG) levels are increased by highcarbohydrate intake in man and animals. Previous studiesperformed in vitro and in vivo have shown an increasedhepatic TG formation in rats fed high glucose and highfructose diets. The enzymatic basis for the increased TGformation was studied using liver microsomal preparationsfrom rats fed 75%o fructose or glucose for 3 days. Synthesisof diglyceride (DG) and TG from 'C-labeled glycerol-3-P,palmitate, ATP, and CoA was increased 2- to 5-fold bythese diets. The ratio of neutral lipid to phosphatidic acid(PA) formed was increased 2- to 3-fold. The increasedDG and TG formation also was observed in studies withwhole homogenate. Measurement of individual enzymatic re-actions in the pathway suggest PA phosphatase is the rate-limiting step and is increased 2- to 3-fold by high sugardiets. Rat liver supernate (100,000 g) stimulates DG andTG formation by microsomes. The stimulating factor isprecipitated by 40%o (NH4)2SO4 and has PA phosphataseactivity. High sugar intake increases the stimulation by thisfactor 3- to 4-fold. The increase in PA phosphatase activityand TG formation was observed after 24 hr and was maxi-mumafter 3 days for 75%o glucose and 11 days for 75%ofructose. The results suggest that high sugar intake resultsin accelerated hepatic TG formation as a consequence ofincreased PA phosphatase activity. These changes may ac-count for the rise in serum TG observed with high carbo-hydrate intake in animals. (This research supported byNIH Grant No. AM-09000.)

174. Reduction of Postperfusion Cytomegalovirus Infec-tions after the Use of Leukocyte-Depleted Blood. DAVID

J. LANG,* PAUL A. EBERT,* BRADLEY M. RODGERS,* H.PRESTONBOGGESS,* AND ROBERTS. RIxSE,* Durham, N. C.(introduced by Samuel L. Katz.)

After extracorporeal perfusion 50-60% of patients de-velop virologic and/or serologic evidence of cytomegalovirus(CMV) infection. It has been proposed by one of us(D. L. J.) that after perfusion, donor and/or recipientleukocytes may be stimulated to divide, as in a mixed lym-phocyte culture, activating latent CMV associated withleukocytes. To test this hypothesis, alternate patients wereperfused with leukocyte-depleted or whole blood (controls).Preoperative leukocyte counts performed on treated pumpbloods averaged 42% of the controls. Virologic and sero-logic studies were performed preoperatively and 2, 10, 30,and 90 days after surgery. There were 10 patients in eachgroup. Four of six controls who were CMVantibody-nega-tive preoperatively, seroconverted after perfusion. Virus wasrecovered from the blood of three and the urine of two ofthese patients. One of four controls who were seropositivepreoperatively developed a significant titer rise. In contrast,only one of eight patients perfused with leukocyte-poorblood who were seronegative before surgery developed aviremia and became antibody positive; another patient mani-fested a viruria but remained antibody-negative. In the lattercase the preoperative pump blood manifested a high leuko-cyte count; the only instance in which the leukocyte concen-tration in the experimental perfusate was in the range ofthe controls. Two patients with preexisting CMVantibodyexhibited no titer change and yielded no virus. These find-ings support the proposed hypothesis and suggest a meansfor reducing if not eliminating transfusion-associated CMVinfections. (Research supported by NIH RCDASK04-HD-19472-02.)

175. Antibodies Arising during Hepatitis That Reactwith Antigens in Normal Stools. ROGER F. LANGE, N.RAPHAEL SHULMANN,** CARLA S. KNEPP, AND CECEIL N.COLEMAN,Bethesda, Md.

Six of 30 patients with acute viral hepatitis (HAA posi-tive or negative) and 6 of 25 patients with high titer anti-HAA after multiple transfusions developed 7S serum anti-bodies against normal stool antigens detected by immuno-diffusion and counter immunoelectrophoresis. Of 66 patientswith cirrhosis and other forms of chronic liver disease, onlytwo had anti-stool antibodies, and of 151 normal individuals,none had anti-stool antibodies. Stool antigen was preparedby extracting homogenized stool with saline and filteringthrough a 450 nm pore size. Stools from 43 of 57 normalindividuals and 22 of 30 patients with hepatitis or otherliver disease contained the stool antigen. Different stool ex-tracts produced precipitin lines of identity or partial identity.The antigen was not detected in serum, urine, bile, extractsof human pancreas, stomach, intestine, or liver or in extractsof bacterial cultures of antigen positive stool. The antigenhas chemical characteristics of protein, is destroyed by tryp-sin, and the molecular weight by Sephadex G-200 filtrationvaries from 60,000 to greater than 300,000. No virus-likeparticles were seen in antigen-antibody precipitates by elec-tron microscopy and there were no antigenic determinants incommon with HAA or its subtypes AY or AD. All precipi-

53 a

tins were soluble in saline, hence could not be analyzed his-tochemically. These findings indicate that antigenic sub-stances normally present in the bowel can gain access toantibody-forming tissue during liver disease, particularlyacute viral hepatitis. This new antigen-antibody system, inaddition to HAA, must be considered in the pathogenesis ofhyperglobulinemia, "auto-immune" phenomena, immune-com-plex disease, and immunologic cross-reactions associated withliver disease.

176. Coronary and Intensive Care Training for Nurses:A Two Year Study of Four Projects of a Regional Medi-cal Program. STEPHENB. LANGFELD* AND RONALDE. MIL-LER,* Haverford, Pa. (introduced by John K. Clark**).

The Greater Delaware Valley Regional Medical Program,recognizing the critical role of trained nursing personnel inachieving whatever benefits result from coronary care units,have implemented six programs for training nurses in coro-nary and intensive care. This paper presents and analyzesdata covering four programs which have been in operationfor 2 yr. Of the 361 trainees surveyed by questionnaires, datahave been obtained from 295. 68% are currently employedin special care units and 82% of these are working full time.66% of the nurses currently working are responsible for fiveor less patients. Training has permitted an increase in theorganizational level of responsibility for 55% of the nurseswho were previously working in special care units and for78%o of those previously employed in some other capacity.There has been an increase in the number of nurses per-forming each of a list of specified procedures with a markedshift toward authority for doing so devolving upon the nurseherself. Even though 77% considered their training adequate,45%o of these have sought additional training. Participants'grades at the end of the course showed significant improve-ment and appeared to be reliable predictors of subsequentperformance. (Study supported by grant from HSMHAofUSPHSto Greater Delaware Valley RMP.)

177. Observations on Instantaneous Flow across the In-competent Mitral Valve In Vivo. SHLOMOLANIADO,* HY-MAN MILLER,* EDWARDYELLIN,* AND ROBERT W. M.FRATER,* New York (introduced by Henry D. Lauson**).

Simultaneous mitral valve flow (measured by a toroidalelectromagnetic flow probe at the mitral ring), left atrial,left ventricular and aortic pressures, electrocardiogram, andaortic flow were recorded in 15 dogs during normal sinusrhythm, irregular junctional rhythm, ectopic beats, and ven-tricular pacing. Cutting chordae tendinae created valvularinsufficiency. The major fraction of regurgitant flow oc-curred in early ventricular systole: a rapid fall in velocityof flow was noted after aortic valve opening, presumably be-cause a second outlet was available. With irregular ventricu-lar rates, the ratio of regurgitant volume to total strokevolume for anv beat was determined by the preceding cyclelength and the dynamics of flow of the previous beat. Thelatter factor was more significant in short cycles. The dis-tribution of ventricular ejection in one beat determined in-stantaneous aortic diastolic pressure and thus the impedanceto forward flow the subsequent beat faced. Changes in ab-solute regurgitant volumes from beat to beat during arrhyth-

mias were minimal, explaining the clinical observation thatthe murmur of mitral regurgitation is uniform in intensityin premature beats and during arrhythmias. "Double regurgi-tation" (diastolic as well as systolic) was noted when re-laxation of left ventricular muscle was rapid and abbrevi-ated, limiting the duration of the period of maximal increasein compliance of the ventricle. Inflow across the mitral valvein mid-diastole created a buildup of pressure in the left ven-tricle resulting ultimately in a brief mid-diastolic period ofpositive ventriculo-atrial pressure gradient accompanied bymodest regurgitant flow. In certain cases of mitral regurgita-tion, it is speculated that the third heart sound may arisefrom an extra closure of the mitral valve in diastole.

178. The Cardiac Beta-Adrenergic Receptor: Solubiliza-tion, Characterization, and Purification by Affinity Chro-matography. ROBERTLEFKOWITZ,* EDGARHABER, GEOFFREYSHARP,* AND DONALDO'HARA,* Boston, Mass.

Inotropic aqd chronotropic effects of catecholamines (CA)are mediated by beta-adrenergic receptors (BAR) with con-sequent stimulation of adenyl cyclase. Receptors and enzymeare found in membrane-bound particulate fractions of tissue.This has retarded detailed study of the individual com-ponents of this system. The 78,000 g particulate fraction ofventricular myocardium contains both the BAR (i.e., CAare specifically bound in the order of their potency as beta-adrenergic agonists) and an adenyl cylase which is stimu-lated by these CA with the same rank order of potency. Tosolubilize the components, particles were treated with de-oxycholate or Lubrol P.X. The BAR and cyclase werefound in the supernatant after 1 hr centrifigation at 100,000 g.Affinity and specificity of catecholamine binding in the solu-ble supernatant was defined by the capacity of drugs to dis-place bound norepinephrine-'H. The order of potency ofbeta-adrenergic drugs was isoproterenol > epinephrine ' nor-epinephrine> dopamine> Dopa. Propranolol inhibited bind-ing, whereas metabolites and alpha-active drugs were < 0.1%as active. KD for norepinephrine was 4 X 10' mmole/liter.PCMB abolished binding, suggesting participation of freeSH groups. When filtered on Sepharose 4B, norepinephrine-binding activity eluted at volumes corresponding to molecularweights of 40,000 and 160,000. An affinity chromatographycolumn (ACC) was constructed by covalently linking nor-epinephrine to agarose via an extended hydrocarbon side-chain. When aliquots of supernatant were passed over theACC, 90-100% of the BAR was adsorbed. Protein concen-tration (Lowry) of the effluent was not measurably differentfrom supernatant. Receptors could then be quantitativelyeluted with epinephrine at pH 3.8, thus achieving a strikingdegree of purification in a single step. These purified recep-tors retain the specificity of the original particulate prep-arations. (Supported by NIH No. HL14150.)

179. Ultrastructural and Functional Correlates in CardiacCells Grown in Tissue Culture. MARIANNE J. LEGATO,*New York (introduced by Nicholas P. Christy**).

This electron microscopic study of 2- to 6-day-old ratventricular cells grown in tissue culture contributed to ourunderstanding of the development and function of intercellu-lar connections, the mechanism of mitochondrial prolifera-

54 a

tion, and the manner in which myofilaments are generatedand assembled into sarcomeric units. It is apparent thatneither nexal linkages between cells or a transverse tubularsystem are present even in the oldest (6 days) cells of thepreparation, although they resemble the adult working ven-

tricular myofiber in almost all other respects. Since the cellsbeat in synchrony at this stage at rates as high as 180, itmay be assumed that the T system is not necessary for ex-

citation-contraction coupling, and that the nexus is not theonly low-resistance junction through which intercellulartransmission of impulses occurs. Z substance is abundant inthe earliest cells and precedes the development of the firstmyofilaments. It is the Z substance which serves as thesource for or template upon which myofilament assembly andsarcomeric construction occur.

180. Passive Hemagglutinating Antibodies (PHA) inCerebospinal Fluids (CSF) of Patients with Herpes-virus hominis (HVH) Encephalitis. A. MARTIN LERNER,CARL B. LAUTER,* DAVID C. NOLAN,* AND MARY JANESHIPPEY,* Detroit, Mich.

In order to assess incidence and spectrum of disease ofHerpesvirus hominis (HVH) infections of the central nerv-

ous system (CNS), an accurate means of diagnosis is needed.Such a test is also needed to initiate and evaluate antiviraltherapy. Wehave modified a type-specific microindirect pas-sive hemagglutinating antibody (PHA) procedure for HVHand studied responses in rabbit and human sera and in hu-man cerebrospinal fluids (CSF). PHA are 2.6-14 times to4-22 times more sensitive than neutralizing antibodies toHVH, types 1 or 2, respectively. PHA to HVH were as-

sayed in spinal fluids obtained early during illnesses of seven

patients with encephalitis caused by type 1 virus (meanCSF-PHA= 65) and in two infants with congenital dis-seminated type 2 virus infection (mean CSF-PHA= 768).Specific CSF-PHA were found in eight of the nine patientswith HVH encephalitis. The antibodies were IgM in con-

genital cases, but IgG in the others. PHA in CSF of pa-tients with HVH encephalitis were similar in patients inwhom the diagnosis was made by isolation of viruses at brainbiopsy, and in others in whom craniotomies were not done.PHA in CSF do not correlate with degree of meningeal in-jury as indicated by pleocytoses or protein. Kinetics of risesin PHA in CSF and sera suggest that the anti-herpesvirusantibodies in the CNS are locally produced. In all but one

patient in whom specimens were available, type-specific PHAwere 1/16 or greater by the 7th day of illness. PHA to HVHwere virtually absent in CSF from 35 control patients withvarious diseases requiring lumbar puncture (mean CSF-PHA, H. hominis type 1 or 2, < 10). PHA to HVHin CSFof patients with encephalitis are a rapid, specific, inclusive,and noninvasive means of diagnosis.

181. Exercise Testing in Patients with Occluded Aorto-Coronary Saphenous Vein Bypass Grafts. STEPHEN J.LESHIN,* LAWRENCED. HORWITZ,* ROGERR. ECKER,* GUN-NAR BLOMQVIST,* W. L. SUGG,* AND CHARLESB. MULLINS,*Dallas, Tex. (introduced by Norman M. Kaplan).

Multistaged exercise tests are used to objectively quantifyfunctional improvement after coronary bypass graft surgery.

55

However, the correlation between functional improvementand graft patency is unknown. This relationship was ex-amined in seven patients with nonfunctioning grafts at post-operative cardiac catheterization. Four patients had all graftseither occluded or nonfunctional. Three patients had oneopen, functioning graft and one graft which was either oc-cluded or nonfunctional. Nonfunctional grafts could be visu-alized angiographically but no dye appeared to move intodistal coronary vessels. Angiography and graded bicycle ex-ercise were performed preoperatively and repeated an aver-age of 8.5 months after surgery (range 4.5-13.0 months).The clinical histories and results of exercise tests correlatedwell. In four patients angina was partially or completelyrelieved and exercise performance improved (average preopmaximal work load 260 kpm vs. 490 kpm postop). In onepatient angina disappeared, but congestive heart failurecaused severe clinical limitation and the maximal exercisework load was unchanged. Two patients continued to haveangina; in one of the exercise test was unchanged and inone the maximal work load decreased from 300 to 150 kpm.Of the four patients who showed improvement by exercisetesting, three had no functioning grafts and the other hadone of two grafts open. Myocardial infarction during sur-gery or postoperatively was documented in two of the fourclinically improved patients. Thus, the results of exercisetesting cannot be used as the sole index of graft patency.Angiographic confirmation of graft patency is necessarybefore clinical improvement can be attributed to improvedmyocardial perfusion. (Research supported by NIH GrantHE06296.)

182. Restoration of Catecholamine Responsiveness ofSolubilized Myocardial Adenylate Cyclase by Phosphati-dylinositol. GERALD S. LEVEY,* Miami, Fla. (introduced byWilliam J. Harrington).

Catecholamines serve critical roles in cardiac function.They augment heart rate and contractility and activate myo-cardial adenylate cyclase by a beta adrenergic receptormechanism. This receptor is thought to be closely related toor identical with the membrane-bound adenylate cyclase. Wehave described the preparation of a solubilized myocardialadenylate cyclase which is unresponsive to catecholaminestimulation. Since membrane lipids play an important rolein maintaining the functional integrity of the hormone-re-ceptor complex, we examined the effect of several phospho-lipids on restoring the responsiveness of the solubilized ade-nylate cyclase to three catecholamines, norepinephrine, epine-phrine, and isoproterenol. The addition of phosphatidylinosi-tol restored catecholamine activation of the solubilized adeny-late cyclase, whereas phosphatidylserine and phosphatidyl-ethanolamine did not. Half-maximal activation was achievedwith norepinephrine 8 X 10' mole/liter, epinephrine 1 X 1OVmole/liter, and isoproterenol 1 X 10' mole/liter, concentra-tions approximately 100-fold lower than those required inparticulate preparations and similar to the sensitivity ob-served in intact physiologic preparations. Maximal concen-trations of the catecholamines produced an increase in cyclic3',5'-AMP accumulation of approximately 150%, similar tothat obtained in particulate preparations. Catecholamineactivation of the solubilized adenylate cyclase in the presence

of phosphatidylinositol was abolished by the beta-adrenergicblocking agent DL-propranolol. Whennorepinephrine was pres-ent at 2 X 10' mole/liter, its effect was abolished by 1 X 101 MDL-propranolol, a concentration of blocker that was withouteffect in the presence of 2 X 10' M norepinephrine. Theminimally effective concentration of phosphatidylinositol forrestoring norepinephrine responsiveness (0.05 ,gg/50 ,ug en-zyme protein) is comparable to that calculated to be presentin heart muscle. These results provide further evidence forthe importance of membrane lipids in hormone-sensitiveadenylate cyclase systems and define a role for phosphatidy-linositol as a possible molecular component of the cardiacbeta-adrenergic receptor. (Research supported by NIH Grant1 RO1 HE 13715-01.)

183. A Nonimmunoglobulin Protein-the Major Com-ponent of Some Human Amyloid Fibrils. MARKLEvIN,*EDWARDC. FRANKLIN, BLAS FRANGIONE, AND MORDECHAIPRAS,* New York.

Recent studies have demonstrated the identity of the majorsubunit of many amyloid fibrils to the variable region oflight chains. Starting with purified amyloid fibrils from fivepatients with secondary amyloidosis or FMF, we have iso-lated a homogeneous protein which appears to be similar inall, and whose partial amino acid sequence is unlike anyknown immunoglobulin. In order to separate carbohydratefrom the protein component of amyloid, fibrils from subjectswith secondary amyloidosis and FMF were treated with0.05 M HC1. More than 50% of the material was extractedas a homogeneous protein with a mol wt of ± 6000, whichbound Congo red and could be made to assume a fibrillarappearance. Partial amino sequence studies of two of theseproteins were identical. The amino acid sequence of the first50 residues derived from three cyanogen bromide fragmentsof one of these was unrelated to any known immunoglobulin.After complete reduction and alkylation of two amyloidfibrils, this component was recovered in the second peak onSephadex G-100 filtration while the first peak consisted pri-marily of material which resembled light chains on peptidemaps. Although similar material could not be extracted insignificant amounts from tissues of two patients with mul-tiple myeloma and macroglobulinemia, the possible existenceof small amounts of a similar component could not be ex-cluded with certainty. The finding of a homogeneous com-ponent not related to immunoglobulins as the major com-ponent of some forms of amyloid raises questions about therole of immunoglobulins in the pathogenesis of amyloid andthe possibility of chemically defining different types of amy-loid. (Research supported by grants from NIH and A.F.)

184. Inhibition of Insulin Secretion by Diphenylhydan-toin (DPH) In Vitro and In Vivio: Selective Effects inDiabetes and Obesity. SEYMOURR. LEvIN,* JAMES W.REED,* KING-NIEN CHING,* JOHN W. DAvIS,* ROBERTBLUM,* AND PETER H. FORSHAM,** San Francisco, Calif.

DPH (25 /Ag/ml) inhibits insulin secretory responses toglucose and arginine in vitro in the perfused rat pancreas.Whereas only the late phase of glucose-induced insulin secre-tion is inhibited by DPH, both early and late arginine-

induced secretion are inhibited by the drug. In vivo testswere (a) 100 g oral glucose tolerance (OGTT), and (b)intravenous arginine (ARG) 30 min after intravenous glu-cose (GL). Tests were done on separate days, before andafter 4 days of DPH ingestion by eight lean nondiabeticnormals (NL), seven lean early diabetics (D), six obesenondiabetics (OND), and four obese diabetics (OD). DPHserum levels attained were similar to those in epilepticstaking the drug (mean 15.1 ,ug/ml). In NL, mean per centchange in total insulin response in OGTT after DPH wast 30%. In contrast, in the other groups, these values (and Pvalue versus NL) were: D 4 30% (P < 0.02), ONDI 33%(P < 0.01), OD J 53% (P < 0.005). In NL, mean per centchange, after DPH, in total insulin response to intravenousARGafter intravenous GL was I 11%. Alterations seen inthe other groups (and P values versus NL) were: D J 59%(P < 0.005), ONDJ 53% (P < 0.05), OD ! 31% (NS).With OGTT these insulin changes were accompanied byminimal changes in glucose levels in NL, glucose rises inD and in the obese. Thus, nonobese diabetics and obese sub-.jects appear more sensitive than normals to DPH-inducedreduction of insulin secretion. In the diabetics, this may re-flect an effect on p-cells with limited capacity. In obesity,this may represent increased suppressibility of a highly ac-tivated secretory mechanism. (Research supported by grantsfrom NIH and L. J. Skaggs Foundation.)

185. Influence of Water Diuresis on Ampicillin Treat-ment of Enterococcal Pyelonephritis. SANDRA P. LEvI-SON* AND DONALDKAYE, Philadelphia, Pa.

Treatment with ampicillin for 2 wk has been shown toreduce but not eradicate bacteria from kidneys of rats withenterococcal pyelonephritis. Water diuresis produced byoffering rats 5% dextrose in tap water (D/W) has beenshown to have a variable effect in decreasing renal titersof enterococci. In the present studies, the effect of ampi-cillin (40 mg intramuscularly twice a day) in combinationwith water diuresis was determined on renal titers of entero-cocci after intravenous inoculation. Ampicillin injectionswith or without diuresis were started 4 or 21 days afterinitiation of infection and continued for 7 or 14 days. Incomparison to controls (saline injections in rats drinkingtap water), diuresis plus saline injections did not lowerrenal titers of enterococci (log median titers 5.0). Injectionof ampicillin in nondiuresing rats somewhat reduced renaltiters of enterococci after both 7 and 14 days of treatmentstarted 4 or 21 days after initiation of infection. Diuresisplus ampicillin reduced renal titers significantly as comparedto nondiuresing rats receiving ampicillin. Log median titersin diuresing rats receiving ampicillin starting 4 days afterinfection were 1.0 and 1.8, respectively, after 7 and 14 daysof therapy as compared with 5.0 and 4.3 in nondiuresingrats receiving ampicillin. In diuresing rats receiving ampi-cillin starting 21 days after infection, log median titers were1.0 after both 7 and 14 days of therapy as compared with4.2 and 2.5 in nondiuresing rats receiving ampicillin. Thesestudies demonstrate that diuresis resulting from administra-tion of D/W plus ampicillin starting 4 or 21 days afterintravenous injection of enterococci reduces renal titers morethan ampicillin or diuresis alone.

6