Vol. 44, No. 1INFECTION AND IMMUNITY, Apr. 1984, p. 91-960019-9567/84/04091-06/02.00/0Copyright © 1984, American Society for Microbiology

Toxic Interactions of Benzyl Alcohol with Bacterial EndotoxinTHOMAS A. CEBULA,1* ANTOINE N. EL-HAGE,1 AND VICTOR J. FERRANS2

Pharmaceutical Research and Testing, National Center for Drugs and Biologics, Food and Drug Administration,Washington, D.C. 20204,1 and National Heart, Lung, and Blood Institute, Bethesda, Maryland 202052

Received 6 September 1983/Accepted 16 January 1984

Acute toxic interactions of intravenously administered benzyl alcohol and Escherichia coli 055:B5(Boivin preparation) endotoxin were examined in rodents. Lethality studies in male CD-1 mice demonstrat-ed that these agents were more toxic when administered in combination than when either was administeredalone. Prophylactic treatment with diazepam (5 mg/kg intraperitoneally) protected against lethality inducedby either the combination or the endotoxin yet offered little, if any, protection against the lethal effects ofbenzyl alcohol. Similar treatments with naloxone (5 mg/kg intraperitoneally) failed to protect against eitherendotoxin-induced or benzyl alcohol-induced lethality, but they significantly protected against the lethaleffects of the combination. Although hexobarbital-induced sleeping time was prolonged in endotoxin-treated mice (but was normal in benzyl alcohol-treated mice), a more protracted effect on sleeping time wasobserved in mice treated with both benzyl alcohol and endotoxin. Moreover, male Wistar rats treated withbenzyl alcohol (40 mg) showed no evidence of hepatic lesions, but rats treated in combination with sublethaldoses of the alcohol (40 mg) and the endotoxin (0.4 mg) developed hepatic lesions which were more severe

than those observed in rats treated with endotoxin (0.4 mg) alone. A correlation between altered bloodchemistry values and severity of hepatic lesions was demonstrated. These data show in vivo toxicinteractions between benzyl alcohol and bacterial endotoxin. In addition, our results indicate that the toxiceffects induced by the benzyl alcohol-endotoxin combination are due to an enhancement of the lethalproperties of bacterial endotoxin.

Although the acute toxic effects of benzyl alcohol are welldocumented (17, 18, 23, 29, 35, 39), this compound has beenutilized for the last 30 years as an effective bacteriostaticagent in formulations intended for intravenous (i.v.) therapy.Although the antibacterial concentration is much lower thanthe concentration deemed dangerous to human adults (35),little is known about the toxicity of benzyl alcohol inneonates. Recent reports (13, 25; J. J. Gershanik, B.Boecler, W. George, A. Sola, M. Leitner, and C. Kapadia,Clin. Res 29:895a, 1981) have associated benzyl alcohol withthe deaths of 16 premature infants. These infants, who werefitted with i.v. catheters, had received relatively largeamounts of the preservative (99 to 405 mg/kg per day), sinceboth the saline used in i.v. washes and the water used toreconstitute medications contained 0.9% benzyl alcohol.Because benzyl alcohol, benzoic acid, and hippuric acidwere found in the sera of these infants, it was postulated thatthe detoxifying mechanism of the liver, needed to handle thelarge insults of benzyl alcohol administered, had not fullydeveloped in these neonates (13, 25, 38). However, thecomplicated medical histories of the patients have not beenruled out as contributing factors in these fatal toxic episodes.Although symptoms in animals given toxic doses of benzylalcohol (17, 29, 35) mimic those of the affected infants (13,25, 38), a majority of these symptoms are also common toendotoxin shock (21, 49, 58). Furthermore, because iatro-genic infections are more frequent in catheterized patients(49) and since endotoxic shock is generally not as readilydiagnosed in neonates as in adults (19, 49), the role ofbacterial endotoxins in the "benzyl alcohol deaths" shouldnot be disregarded.The present studies, employing a rodent model, were

* Corresponding author.

initiated to assess whether some of the toxic propertiesascribed to benzyl alcohol might be due to in vivo interac-tions of the preservative with bacterial endotoxin.

MATERIALS AND METHODS

Chemicals. Benzyl alcohol (Fisher Certified; Fisher Scien-tific Co., Pittsburgh, Pa.), sodium benzoate (USP foodgrade; J.T. Baker Chemical Co., Phillipsburg, N.J.), benzal-dehyde (Tenneco Chemicals Inc., New York, N.Y.), diaze-pam (Hoffmann-La Roche, Inc., Nutley, N.J.), naloxone(ENDO Laboratories, Inc., New York, N.Y.), and hexobar-bital (K & K Laboratories, Inc., New York, N.Y.) wereused throughout these studies. The bacterial endotoxin(lipopolysaccharide), a Boivin preparation of Escherichiacoli 055:B5, was obtained from Difco Laboratories, Detroit,Mich. Endotoxin solutions were quantitated by the Limulusamoebocyte lysate gel-clot assay, with a lysate (Pyrotell,Associates of Cape Cod, Inc., Woods Hole, Mass.) with asensitivity of 0.025 ng/ml. Benzyl alcohol was analyzed byhigh-pressure liquid chromatography with a Waters high-performance liquid chromatograph fitted with a Waters C-18column (4 mm by 30 cm), using 0.025 M sodium phosphate-methanol (65:35; pH 4) as the mobile phase. Analysesshowed the benzyl alcohol to be >99.9% pure; benzaldehydewas present as a minor contaminant (<0.05%).

General procedure. Benzyl alcohol and endotoxin wereinjected into CD-1 mice and Wistar rats (Charles RiverBreeding Laboratories, Inc., Wilmington, Mass.) via the tailvein at various doses contained in a volume of 0.2 to 0.8 mlof pyrogen-free water (Travenol water for irrigation). Con-trol groups received the water vehicle via the same route.The combination was administered to animals in two ways.Either the compounds were premixed and administered in asingle injection, or benzyl alcohol and endotoxin wereinjected sequentially within 1 min of each other. Both

91

on May 1, 2016 by guest

http://iai.asm.org/

Dow

nloaded from

92 CEBULA, EL-HAGE, AND FERRANS

methods yielded similar results. In the mice, dilation of theveins was induced by crowding, and in the rats, dilation wasinduced by application of xylene to the tails. After injec-tions, the animals were returned to cages and given freeaccess to food and water.

Lethality in mice. CD-1 male mice, weighing 20 to 24 g,were assigned to groups of at least eight animals each.Various doses of either benzyl alcohol or endotoxin wereadministered to determine minimal toxic doses of the indi-vidual agents; survivors were observed for 72 h. Based onthese data, further studies were conducted to probe toxicinteractions of the preservative and endotoxin.

Sleeping time in mice. In separate experiments at variousintervals after benzyl alcohol or endotoxin treatments orboth, hexobarbital (80 mg/kg) was administered intraperito-neally (i.p.) to CD-1 mice, and sleeping time was monitored.Sleeping time is the duration (in minutes) from the time ofinjection until the reappearance of the righting reflex, whichis defined as the capacity of the mouse to right itself from aprone position three consecutive times within 1 min. Dataare presented as the mean + standard error of the mean foreach group.Rat studies. Male Wistar rats, weighing 360 to 410 g, were

injected with benzyl alcohol (40 mg per rat) or endotoxin(0.40 mg per rat) or both, marked for identification, andreturned to cages (one rat per cage). Animals were killed bydecapitation 24 h after treatment, and blood samples werecollected. Complete serum chemistry determinations wereperformed by Vet Path Laboratories, Rockville, Md. Theheart, kidney, liver, and spleen were excised and fixed in10% buffered Formalin. The tissues were embedded inparaffin blocks, and sections were prepared and stained withhematoxylin and eosin for light microscopic examination.The frequency and severity of the tissue lesions wereassessed and graded on a scale of 0 to 4+, in which 0 was nodamage, 1+ was involvement of only an occasional cell, 4+was severe involvement of 25% or more of the cells in thevisual field, and 2+ and 3+ were intermediate degrees ofinvolvement. Histological samples were evaluated withoutprior knowledge of the clinical chemistry results or of thetreatments given the various groups of animals.

Statistical analyses. Results are reported as meansstandard error of the mean. As appropriate, Student's t andx tests were used for analyses, and a P value of <0.05 wasconsidered statistically significant. Fifty percent lethal doseswere determined by the method of Litchfield and Wilcoxon(37).

RESULTS

Lethality in mice. Various doses of benzyl alcohol wereadministered i.v. to male CD-1 mice to determine theminimal toxic dose of the compound (Table 1). The treatedanimals were observed for 72 h, and mortality ratios wererecorded. The mice treated with toxic doses of benzylalcohol died within 3 min of treatment, with tremors, gaspingrespiration, and convulsions usually presaging these deaths.That the induced lethality is due to the alcohol, rather than

the trace of benzaldehyde (0.037%) present in these prepara-

tions, is emphasized by the fact that doses of the aldehyde up

to 1.2 mg (i.e., about 1,000 times the amount contaminatingthe doses of benzyl alcohol injected) were not toxic. This ofcourse does not preclude metabolic conversion of the alco-hol to benzaldehyde, although benzoic acid, not benzalde-hyde, is thought to be the major metabolite of benzyl alcohol(18, 25). However, the rapidity with which the injections

TABLE 1. Lethal effects of endotoxin and benzyl alcohol andtheir combination in male CD-1 micea

Mortality ratiob (no. of deaths/total no. ofmice)

Treatment Dose (mg/mouse) +0.05 mg +0.15 mg

Alone of endotox- of endotox-in in

Endotoxin 0.05 0/80.10 1/160.20 6/160.40 8/180.60 25/270.80 8/8

Benzyl 0.5 0/8 0/8 1/5alcohol 0.9 0/8 0/8 6/8

1.8 0/26 1/8 8/82.7 3/18 4/8 15/153.6 20/24 8/8 NDc5.4 16/16 ND ND

a Results are based on at least three experiments in whichoverlapping doses were used to establish the composite table.

b Mortality at 72 h. At endotoxin doses of 0.6 mg and above, themajority of deaths occurred by 24 h, but a portion (10 to 40%depending upon the experiment) died within 1 h and usually withinminutes of the injection.

c ND, not determined.

caused death (s3 min) and the lack of toxicity of benzalde-hyde (up to 1.2 mg; zero of eight mice) and sodium benzoate(up to 12 mg; zero of eight mice) suggest that the toxicity isdue to the parent compound rather than to a metabolite.Deaths induced by endotoxin were more protracted, oc-

curring over a 72-h period. Endotoxin treatment resulted inlethargic behavior, diarrhea, and, at higher doses, tremorsand convulsions. As calculated from the endotoxin data,0.05 and 0.15 mg of endotoxin would be expected to be lethalin 0 and 15% of the treated mice, respectively. However,when the mice were treated simultaneously with these dosesof endotoxin and various doses of benzyl alcohol, mortalitywas increased (Table 1). Qualitatively similar results wereobtained in ICR female mice when a fixed nontoxic dose ofbenzyl alcohol and various doses of endotoxin were adminis-tered simultaneously (data not shown). Deaths resultingfrom toxic doses of the combination were temporally indis-tinguishable from deaths induced by endotoxin alone.Attempts to ameliorate the toxic effects of benzyl alcohol

and endotoxin with prophylactic treatments of either diaze-pam or naloxone are shown in Table 2. Although a lethaldose of endotoxin (0.6 mg) induced convulsions and death in80% of the mice within 24 h, in mice administered diazepam(5 mg/kg i.p.) 30 min before endotoxin treatment, only oneanimal convulsed, and overall mortality was 25%. In micesimilarly treated with naloxone (5 mg/kg i.p.), this dosage ofendotoxin caused 68% mortality. Moreover, both diazepamand naloxone offered significant protection against the lethaleffects induced by the simultaneous administration of benzylalcohol and endotoxin; a combination of benzyl alcohol (1.8mg) and lipopolysaccharide (0.20 mg) was lethal in 65% ofthe control mice, but this combination was toxic to only 5%of the mice pretreated with either diazepam or naloxone(Table 2).

In contrast, pretreatment of the mice with either diazepamor naloxone had little, if any, effect on benzyl alcohol-

INFECT. IMMUN.

on May 1, 2016 by guest

http://iai.asm.org/

Dow

nloaded from

BENZYL ALCOHOL AND ENDOTOXIN INTERACTIONS 93

TABLE 2. Effects of diazepam and naloxone on benzyl alcohol-and endotoxin-induced lethality in CD-1 mice'

Dose (mg/ Pretreat- MortalityTreatment mouse) mentb ratiocBenzyl alcohol 1.8 None 0/24

3.6 None 19/283.6 Diazepam 15/283.6 Naloxone 17/28

Endotoxin 0.20 None 6/160.60 None 16/200.60 Diazepam 5/20d0.60 Naloxone 17/25

Combination: benzyl 1.8 + 0.20 None 13/20alcohol + endotoxin Diazepam 1/20d

Naloxone 1/20da Data represent the results of three experiments for benzyl

alcohol and two experiments each for endotoxin and the combina-tion.

b Mice were given diazepam or naloxone (5 mg/kg i.p.) 30 minbefore treatment with benzyl alcohol, endotoxin, or the combina-tion.

c Mortality at 72 h. Numbers indicate number of deaths/totalnumber of mice.

d p < 0.05 versus respective control by x2.

induced mortality (Table 2), although diazepam did preventconvulsions induced by the preservative.

Hexobarbital-induced sleeping time. The effects of suble-thal doses of benzyl alcohol, endotoxin, and their combina-tion on hexobarbital-induced sleeping time in CD-1 micewere examined. For these experiments, various groups of 10males each (weighing 20 to 24 g) were treated i.v. withbenzyl alcohol (1.8 mg) or endotoxin (0.075 mg) or both, andafter 6 and 24 h, hexobarbital (80 mg/kg i.p.) was injected,and sleeping time was monitored.From separate experiments it was determined that shortly

after the treatments (30 min), sleeping times of the groupstreated with benzyl alcohol (24 ± 2 min) or endotoxin (27 ± 3min) were normal relative to controls (23 ± 1 min). Thesleeping time of the mice given the combination was similar-ly unremarkable (26 ± 2 min). However, 6 h after injectionsof the various agents, hexobarbital sleeping time was pro-longed in groups given either endotoxin (51 ± 3 min) or acombination of endotoxin and benzyl alcohol (56 ± 2 min),yet it was normal in the benzyl alcohol-treated group (27 ± 1min) relative to controls (28 ± 2 min).Although benzyl alcohol (1.8 mg) when administered alone

had no direct effect on hexobarbital-induced sleeping time at24 h (22 ± 2 min) relative to controls (23 ± 2 min), incombination it appeared to exacerbate the effects of endo-toxin. That is, even after 24 h, when sleeping time hadreturned to normal in the endotoxin-treated group (23 ± 3min), mice administered both benzyl alcohol and endotoxinstill had abnormal sleeping times (55 ± 4 min).

Toxicity in rats. Blood chemistry and histopathologicalevaluations were performed on rats treated with sublethaldoses of benzyl alcohol (40 mg) or endotoxin (0.40 mg) orboth; results are summarized in Table 3 and Fig. 1. Allanimals (44 of 44) survived various treatments and werekilled 24 h in posttreatment to obtain serum and tissuesamples. Diffuse lymphoid necrosis of the spleen was ob-served in the endotoxin-treated rats (eight of eight) but not inthe control (zero of four) or benzyl alcohol-treated (zero of

TABLE 3. H-epatotoxic effects of benzyl alcohol and endotoxin inWistar rats

Treatmenta No. of Hepatotoxic scorebrats 0 1+ 2+ 3+ 4+

Control 8 8 0 0 0 0Benzyl alcohol 12 12 0 0 0 0Endotoxin 12 3 7 1 1 0Combination 12 2 1 1 7 1

a Rats received (i.v.) either benzyl alcohol (40 mg), endotoxin(0.40 mg), or both. The water vehicle (1.2 ml i.v.) was used forcontrols. Data are the results of three experiments (see the legend toFig. 1).

b Values under hepatotoxic score indicate the number of ratshaving hepatic lesions. Lesions were assessed as described in thetext.

601-

0

0

z0

40F

20

+ 00a75i1;

16001- 800 M

1200 -

302:5Z

600

800 -

400

C, 6F ,N0°o

-p 0-4 -

& I c

400 -

2C

I2 Io i

00

a;.

800 F

600 -

0

4L 40

20

0 _

0

FIG. 1. Blood chemistry parameters of rats treated with benzylalcohol, endotoxin or both. Serum samples were obtained from rats24 h after treatment. Data, from three separate experiments, repre-sent the mean ± the standard error of the mean of 12 rats each forgroups receiving benzyl alcohol (40 mg), endotoxin (0.40 mg), andthe combination, and for 8 rats serving as controls (1.2 ml of thewater vehicle). Blood urea nitrogen (BUN), cholesterol, and triglyc-eride concentrations and aspartate aminotransferase (SGOT), ala-nine aminotransferase (SGPT), and alkaline phosphatase (A.P.)activities were significantly elevated (P < 0.05; Student's t test) ingroups administered endotoxin alone or in combination with benzylalcohol relative to controls or benzyl alcohol-treated animals.

VOL. 44, 1984

on May 1, 2016 by guest

http://iai.asm.org/

Dow

nloaded from

94 CEBULA, EL-HAGE, AND FERRANSd!4 i*.a; -' '

r, ..?',, ai.

a'~~~~~~~~~a''^_'

Ia a' a' 'a

a''a"' a' a' a a'a' a~ a' 'a A

a .5 a'

a''a'''aaa a aaa

9a'a''If '

four) rats. The kidneys and hearts of all groups werehistologically unremarkable except in three animals receiv-ing endotoxin, in which myocardial necrosis was apparent.For these tissues, no differences between groups receivinglipopolysaccharide alone or in combination with benzylalcohol could be discerned.Most striking, however, was the incidence and severity of

hepatic injuries observed in this study (Table 3). Hepaticlesions were noted in 10 of 12 rats treated with the benzylalcohol-endotoxin combination. Liver necrosis was minimalin one (1+), moderate in one (2+), and severe in eight (3+ to4+) of these animals. Although 9 of 12 rats given lipopoly-saccharide alone had lesions, the lesions were much lesssevere than those in animals given the combination; i.e.,lesions were minimal in seven (1+), moderate in one (2+),and severe in only one (3+). In contrast, no lesions wereobserved in either control (water vehicle-treated) or benzylalcohol-treated animals. Liver lesions were characterized bymarked multifocal acute necrosis and infiltration of polymor-phonuclear leukocytes (Fig. 2).Serum values for glucose, bilirubin, electrolytes, creati-

nine, albumin, globulin, and protein in all groups wereessentially unchanged from control values (data not shown).However, activities of aspartate aminotransferase, alanineaminotransferase, and alkaline phosphatase and concentra-tions of blood urea nitrogen, cholesterol, and triglycerideswere significantly elevated in rats treated with endotoxineither alone or in combination, but they were normal inbenzyl alcohol-treated rats. Although serum values for ani-mals treated with the combination were markedly higherthan those of the endotoxin-treated group, they were notsignificantly different because of the large variations amongindividual animals within this group (Fig. 1).

DISCUSSIONThe present studies show a novel in vivo interaction

between benzyl alcohol and bacterial endotoxin. Dosages ofthe two compounds that alone would be nontoxic togetherproduced acute effects, as evidenced by increased mortalityin mice (Table 1) or frequency and severity of hepatic lesionsin rats (Table 3). Our data are consistent with the interpreta-tion that benzyl alcohol accentuates the toxic properties ofendotoxin. This conclusion is based on the findings thatdiazepam, which protects against the lethal effects inducedby the combination, also protects against endotoxin-inducedlethality yet has virtually no effect on benzyl alcohol-inducedmortality (Table 2).

Since both benzyl alcohol (18, 25) and endotoxin (40, 47,53, 57) are ultimately cleared from the circulation by theliver, these compounds might be expected to affect adverse-ly the metabolism of hexobarbital; an effective monitor ofaltered drug metabolism is prolongation of hexobarbitalsleeping time (12, 31, 52). Although endotoxin alone affectedhexobarbital sleeping time, the effect was somewhat tran-sient, such that by 24 h posttreatment, sleeping time hadreturned to normal. The evanescent nature of low doses ofendotoxin, as measured by increased alanine aminotransfer-ase activity and altered sulfobromophthalein clearance, has

FIG. 2. Hepatic lesions of rats treated with benzyl alcohol,endotoxin, or both. Light micrographs of liver section were obtainedfrom animals 24 h after treatment with the water vehicle (A), benzylalcohol (B), endotoxin (C), and a combination of benzyl alcohol andendotoxin (D). Original magnification, x80.

INFECT. IMMUN.

on May 1, 2016 by guest

http://iai.asm.org/

Dow

nloaded from

BENZYL ALCOHOL AND ENDOTOXIN INTERACTIONS 95

previously been reported (45). When benzyl alcohol wasinjected with endotoxin, it prolonged sleeping time, and theeffect of the combination was more protracted than that ofendotoxin alone. These data further suggest that benzylalcohol enhances the action of endotoxin, perhaps by inhibit-ing the detoxification and or clearance of endotoxin from thecirculation.Endotoxin is normally rapidly cleared by the reticuloendo-

thelial system (RES) of liver and spleen (32, 40, 47, 48, 53,57), with RES function primarily concentrated in the Kupffercells of the liver (32, 47). Agents such as saccharated ironoxide (8), Thorotrast (5, 6, 9), ethanol (4), and trypan blue(5), which depress RES function, potentiate the toxic effectsof endotoxin. It is tempting to speculate that benzyl alcoholalso interferes with RES function, since it affects lipidbilayer structure (20) and influences the integrity of bothmacrophage (15) and hepatocyte (27) membranes, resultingin the stimulation (2, 56) or inhibition (43, 46) of variousmembrane-bound enzyme activities. Furthermore, benzylalcohol inhibits clearance by hepatocytes of asialo-glycopro-teins by acting as a selective inhibitor of fusion of lysosomesand endocytic vesicles (55). In light of the Ca2+ involvementin endotoxicosis (16, 24, 30, 44), it is important to note thatbenzyl alcohol stimulates Ca2+ uptake (10) and activatesCa2'-dependent enzymes (2, 56).Animals failing to efficiently clear toxic doses of endotoxin

show a myriad of responses, ranging from fever to shock anddeath; these diverse biological effects have been extensivelydetailed in recent reviews (1, 7, 11, 26, 34, 41). Both direct(12, 42) and indirect effects of endotoxin, i.e., toxin-inducedactivation of macrophages and attendant release of endoge-nous mediators (see references 7 and 54), contribute to thecomplex sequelae of endotoxicosis.

Clinically, renal failure is the most common complicationof endotoxin shock (1, 50). Our finding that blood ureanitrogen concentrations were markedly elevated in endotox-in-treated animals (Fig. 1) without apparent lesions is conso-nant with that of others (1, 8, 45). This increase presumablyreflects the effects of endotoxin in reducing both the glomer-ular filtration pressure and the excretory function by renaltubules (8).The effects of endotoxin on the liver are manifested by

both the abnormal blood chemistry values, i.e., elevatedaspartate aminotransferase and alanine aminotransferaseactivities (Fig. 1; 7, 45, 47), and the histopathological find-ings (Fig. 2; 28, 32) characterized by marked multifocalnecrosis and infiltration of polymorphonuclear leukocytes.Although the present studies do not address the mechanismfor hepatic injury, others have postulated an interaction ofendotoxin with Kupffer cells (4, 32, 47). It has not beenadequately resolved whether the resulting hepatocyte necro-sis is due to obstruction of hepatic microcirculation (33) or tothe action of cytoxic factors released from activated Kupffercells on parenchymal cells (3).Our studies in mice show that both naloxone, an opiate

antagonist, and diazepam, a gamma-aminobutyric acid an-

tagonist, protect against lethality induced by the benzylalcohol-endotoxin combination, suggesting that endorphins(14, 22, 33) and catecholamines (26) may mechanisticallycontribute to the observed lethality. However, since nalox-one and diazepam share other properties, such as theirability to modulate Ca>2 levels in vivo (36, 50, 51), it is notpossible at present to distinguish whether naloxone anddiazepam exert their protective effects by a common mecha-nism, i.e., modulation of Ca2+ influx, or by different mecha-nisms.

The present studies, providing evidence of an importantinteraction of benzyl alcohol and bacterial endotoxin, arenoteworthy in light of recent reports linking the alcohol withthe deaths of 16 premature infants (13, 25, 38). The symp-toms described in the benzyl alcohol poisonings (25, 38) arestrikingly similar to those of endotoxin shock (1, 21, 49).Although bacteremia and sepsis appear to have been ruledout, at least in the studies by Gershanik et al. (25), endotoxe-mia as a contributing factor in these deaths has not. If thesedata can be extrapolated to humans, they suggest that theneonatal deaths ascribed to benzyl alcohol poisoning mayhave been the result of benzyl alcohol-endotoxin toxicinteractions.

ACKNOWLEDGMENTS

We thank John R. Hohmann for the high-pressure liquid chroma-tography analyses of benzyl alcohol and related compounds andJohn C. Atkinson for computer assistance.

LITERATURE CITED

1. Agarwal, M. K., and G. Lazar. 1977. Metabolic basis ofendotoxicosis. Microbios 20:183-214.

2. Ahkong, Q. F., G. M. Botham, A. W. Woodward, and J. A.Lucy. 1980. Calcium-activated thiol-proteinase activity in thefusion of rat erythrocytes induced by benzyl alcohol. Biochem.J. 192:829-836.

3. Alexander, P., and R. Evans. 1971. Endotoxin and doublestranded RNA render macrophages cytotoxic. Nature (London)232:76-78.

4. AlU, M. V., and J. P. Nolan. 1967. Alcohol induced depression ofreticuloendothelial function in the rat. J. Lab. Clin. Med.70:295-301.

5. Beeson, P. B. 1947. Tolerance to bacterial pyrogens. II. Role ofthe reticulo-endothelial system. J. Exp. Med. 86:39-44.

6. Bennett, I. L. 1952. Further investigation of effect of colloidalmaterials upon Schwartzman phenomenon. Proc. Soc. Exp.Biol. Med. 81:248-250.

7. Berry, L. J. 1977. Bacterial toxins. CRC Crit. Rev. Toxicol.5:239-318.

8. Berry, L. J., and D. S. Smythe. 1961. Effects of bacterialendotoxins on metabolism. IV. Renal function and adrenocorti-cal activity as factors in the nitrogen excretion assay forendotoxin. J. Exp. Med. 114:761-778.

9. Berry, L. J., and D. S. Smythe. 1965. Some metabolic aspects oftolerance to bacterial endotoxin. J. Bacteriol. 90:970-977.

10. Blow, A. M. J., G. M. Botham, and J. A. Lucy. 1979. Calciumions and cell fusion. Biochem. J. 182:555-563.

11. Bradley, S. G. 1979. Cellular and molecular mechanisms ofaction of bacterial endotoxins. Annu. Rev. Microbiol. 33:67-94.

12. Bradley, S. G. 1981. Direct action of bacterial endotoxin on

cells, mitochondria, and lysosomes. Prog. Clin. Biol. Res. 62:3-14.

13. Brown, W. J., N. R. M. Buist, H. T. C. Gipson, R. K. Huston,and N. G. Kennaway. 1982. Fatal benzyl alcohol poisoning in a

neonatal intensive care unit. Lancet i:1250.14. Carr, D. B., R. Bergland, A. Hamilton, H. Blume, N. Kasting,

M. Arnold, J. B. Martin, and M. Rosenblatt. 1982. Endotoxin-stimulated opioid peptide secretion: two secretory pools andfeedback control in vivo. Science 217:845-848.

15. Cherenkevich, S. N., J. M. Vanderkooi, and A. Holian. 1982.The lipid integrity of membranes of guinea pig alveolar macro-

phages studied by nanosecond fluorescence decay of 1,6-diphe-nyl-1,3,5-hexatriene: the influence of temperature and benzylalcohol. Arch. Biochem. Biophys. 214:305-310.

16. Connor, J., J. Fine, K. Kusano, M. J. McCrea, I. Parnas, andC. L. Prosser. 1973. Potentiation by endotoxin of responsesassociated with increases in calcium conductance. Proc. Natl.Acad. Sci. U.S.A. 70:3301-3304.

17. DeLand, F. H. 1973. Intrathecal toxicity studies with benzyl

VOL. 44, 1984

on May 1, 2016 by guest

http://iai.asm.org/

Dow

nloaded from

96 CEBULA, EL-HAGE, AND FERRANS

alcohol. Toxicol. Appl. Pharmacol. 25:153-156.18. Diack, S. L., and H. B. Lewis. 1928. Studies in the synthesis of

hippuric acid in the animal organism. VII. A comparison of therate of elimination of hippuric acid after the ingestion of sodiumbenzoate, benzyl alcohol, and benzyl esters of succinic acid. J.Biol. Chem. 77:89-95.

19. Di Luzio, N. R., and T. J. Friedmann. 1973. Bacterial endotox-ins in the environment. Nature (London) 244:49-51.

20. Ebihara, L., J. E. Hall, R. C. MacDonald, T. J. McIntosh, andS. A. Simon. 1979. Effects of benzyl alcohol on lipid bilayers. Acomparison of bilayer systems. Biophys. J. 28:185-196.

21. Elin, R. J., and S. M. Wolff. 1976. Biology of endotoxin. Annu.Rev. Med. 27:127-141.

22. Faden, A. I., and J. W. Holaday. 1980. Experimental endotoxinshock: the pathophysiologic function of endorphins and treat-ment with opiate antagonists. J. Infect. Dis. 142:229-238.

23. Feasby, T. E., A. F. Hahn, and J. J. Gilbert. 1983. Neurotoxicityof bacteriostatic water. N. Engl. J. Med. 308:966-967.

24. Fritz, M. M., and D. 0. R. Keppler. 1982. Liver calcium andendotoxin action. Naturwissenschaften 69:147-149.

25. Gershanik, J., B. Boecler, H. Ensley, S. McCloskey, and W.George. 1982. The gasping syndrome and benzyl alcohol poison-ing. N. Engl. J. Med. 307:1384-1388.

26. Gilbert, R. P. 1960. Mechanisms of the hemodynamic effects ofendotoxin. Physiol. Rev. 40:245-279.

27. Gordon, L. M., R. D. Sauerheber, J. A. Esgate, I. Dipple, R. J.Marchmont, and M. D. Houslay. 1980. The increase in bilayerfluidity of rat liver plasma membranes achieved by the localanesthetic benzyl alcohol affects the activity of intrinsic mem-brane enzymes. J. Biol. Chem. 255:4519-4527.

28. Gorodischer, R., J. Krasner, J. J. McDevitt, J. P. Nolan, andS. J. Yaffe. 1976. Hepatic microsomal drug metabolism afteradministration of endotoxin in rats. Biochem. Pharmacol.25:351-353.

29. Gruber, C. M. 1923. The pharmacology of benzyl alcohol and itsesters. II. Some of the effects of benzyl alcohol, benzyl benzo-ate and benzyl acetate when injected intravenously upon therespiratory and circulatory systems. J. Lab. Clin. Med. 9:92-112.

30. Hagmann, W., and D. Keppler. 1982. Leukotriene antagonistsprevent endotoxin lethality. Naturwissenschaften 69:594-595.

31. Higuchi, T., T. Nakamura, and H. Uchino. 1977. Effects ofphenobarbital and endotoxin on the lethality and metabolism of6-mercaptopurine in male BALB/c mice. Cancer Res. 37:3668-3671.

32. Hirata, K., A. Kaneko, K. Ogawa, H. Hayasaka, and T. Onoe.1980. Effect of endotoxin on rat liver. Analysis of acid phospha-tase isozymes in the liver of normal and endotoxin-treated rats.Lab. Invest. 43:165-171.

33. Holaday, J. W. 1983. Cardiovascular consequences of endoge-nous opiate antagonism. Biochem. Pharmacol. 32:573-585.

34. Kabir, S., D. L. Rosenstreich, and S. E. Mergenhagen. 1978.Bacterial endotoxins and cell membranes, p. 59-87. In J.Jeljaszewicz and T. Wadstrom (ed.), Bacterial toxins and cellmembranes. Academic Press, Inc., New York.

35. Kimura, E. T., T. D. Darby, R. A. Krause, and H. D. Brondyk.1971. Parenteral toxicity studies with benzyl alcohol. Toxicol.Appl. Pharmacol. 18:60-68.

36. Lee, C. H., and B. A. Berkowitz. 1976. Stereoselective andcalcium-dependent contractile effects of narcotic antagonistanalgesics in the vascular smooth muscle of rat. J. Pharmacol.Exp. Ther. 198:347-356.

37. Litchfield, J. T., Jr., and F. Wilcoxon. 1949. A simplifiedmethod of evaluating dose-effect experiments. J. Pharmacol.Exp. Ther. 96:99-113.

38. Lovejoy, F. H. 1982. Fatal benzyl alcohol poisoning in neonatalintensive care units: a concern for pediatricians. Am. J. Dis.Child. 136:974-975.

39. Mcht, D. I. 1918. A pharmacological and therapeutic study ofbenzyl alcohol as a local anesthetic. J. Pharmacol. Exp. Ther.11:263-279.

40. Mathison, J. C., and R. J. Ulevitch. 1979. The clearance, tissuedistribution, and cellular localization of intravenously injectedlipopolysaccharide in rabbits. J. Immunol. 123:2133-2143.

41. Mathison, J. C., and R. J. Ulevich. 1983. Mediators involved inthe expression of endotoxic activity. Surv. Synth. Pathol. Res.1:34-48.

42. McGivney, A., and S. G. Bradley. 1979. Action of bacterialendotoxin and lipid A on mitochondrial enzyme activities ofcells in culture and subcellular fractions. Infect. Immun.25:664-671.

43. Mitranic, M. M., J. M. Boggs, and M. A. Moscarello. 1982. Theeffect of linoleic acid and benzyl alcohol on the activity ofglycosyltransferases of rat liver Golgi membranes and somesoluble glycosyltransferases. Biochim. Biophys. Acta 693:75-84.

44. Morrison, D. C., Z. G. Oades, and D. DiPietro. 1981. Endotoxin-initiated membrane changes in rabbit platelets. Prog. Clin. Biol.Res. 62:47-64.

45. Munson, A. E., D. C. Drummond, A. C. Adams, and S. G.Bradley. 1976. Enhanced toxicity for mice of combinations ofbacterial lipopolysaccharide and vincristine. Antimicrob.Agents Chemother. 9:840-847.

46. Needham, L., A. D. Whetton, and M. D. Houslay. 1982. Thelocal anaesthetic and bilayer fluidising agent, benzyl alcohol,decreases the thermostability of the integral membrane proteinadenylate cyclase. FEBS Lett. 140:85-88.

47. Nolan, J. P. 1981. Endotoxin, reticuloendothelial function, andliver injury. Hepatology 1:458-465.

48. Noyes, H. E., C. R. Mclnturf, and G. J. Blahuta. 1959. Studieson distribution of Escherichia coli endotoxin in mice. Proc. Soc.Exp. Biol. Med. 100:65-68.

49. Petersdorf, R. G., and D. C. Dale. 1980. Gram-negative bacter-emia and septic shock, p. 561-566. In K. J. Isselbacher, R. D.Adams, E. Braunwald, R. G. Petersdorf, and J. D. Wilson (ed.),Harrison's principles of internal medicine, 9th ed. McGraw-HillBook Co., New York.

50. Peyton, J. C., and J. L. Borowitz. 1979. Chlorodiazepoxide andtheophylline alter calcium levels in subcellular fractions of ratbrain cortex. Proc. Soc. Exp. Biol. Med. 161:178-182.

51. Polc, P., E. P. Bonetti, L. Pieri, R. Cumin, R. M. Angioi, H.Mohler, and W. E. Haefly. 1981. Caffeine antagonizes severalcentral effects of diazepam. Life Sci. 28:2265-2275.

52. Rose, W. C., S. G. Bradley, and I. P. Lee. 1972. Enhancedtoxicity for mice of vincristine and other chemotherapeuticagents with Salmonella typhosa endotoxin and Pseudomonasaeruginosa. Antimicrob. Agents Chemother. 1:489-495.

53. Ruiter, D. J., J. van der Meulen, A. Brouwer, M. J. R. Hummel,B. J. Mauw, J. C. M. van der Ploeg, and E. Wisse. 1981. Uptakeby liver cells of endotoxin following its intravenous injection.Lab. Invest. 45:38-45.

54. Schlessinger, D. (ed.). 1980. Microbiology-1980, p. 3-167.American Society for Microbiology, Washington, D.C.

55. Tolleshang, H., and T. Berg. 1982. Evidence for the selectiveinhibition of fusion between endocytic vesicles and lysosomesby benzyl alcohol. Biochem. Pharmacol. 31:593-595.

56. Voorheis, H. P., and B. R. Martin. 1982. Local anaestheticsincluding benzyl alcohol activate the adenylate cyclase in Try-panosoma brucei by a calcium-dependent mechanism. Eur. J.Biochem. 123:371-376.

57. Willerson, J. T., R. L. Trelstad, T. Pincus, S. B. Levy, and S. M.Wolff. 1970. Subcellular localization of Salmonella enteritidisendotoxin in liver and spleen of mice and rats. Infect. Immun.1:440-445.

58. Wolff, S. M. 1973. Biological effects of bacterial endotoxins inman. J. Infect. Dis. 128:s259-s264.

INFECT. IMMUN.

on May 1, 2016 by guest

http://iai.asm.org/

Dow

nloaded from