HEMATOGENOUSPYELONEPHRITIS IN RATS. I. ITS PATHO-GENESIS WHENPRODUCEDBY A SIMPLE NEW

METHOD1'2

By ABRAHAMI. BRAUDE, ALVIN P. SHAPIRO, AND JENNIE SIEMIENSKI

(From the Department of Internal Medicine, The University of Texas, Southwestern MedicalSchool, Dallas, Texas)

(Submitted for publication January 7, 1955; accepted June 8, 1955)

The study of experimental pyelonephritis hasbeen generally restricted to that form of the dis-ease which develops after ligation of the ureter(1-3). When pyelonephritis is complicated byureteral ligation, however, a clear distinction can-not be made between the effects of urinary ob-struction and those of infection. We have at-tempted, therefore, to establish experimentalpyelonephritis in the absence of hydronephrosisand have succeeded by subjecting rats to a simplenew technique requiring only two steps: 1) In-jection of bacteria into the blood and 2) Mas-sage of the kidneys through the intact abdominalwall. This method has provided an experimentalmodel which appears to reproduce closely thatform of human pyelonephritis occurring in pa-tients with no manifest hydronephrosis. In addi-tion, it dispenses with the cumbersome surgicalprocedures previously required for the productionof experimental pyelonephritis.

This report describes the pathogenesis of pyelo-nephritis in rats when renal massage is combinedwith the inoculation of strains of Escherichia coliisolated from human infections.

METHODS

I. The effect of renal massage on the incidence of pyelo-nephritis in rats inoculated uith E. coli

Sprague-Dawley rats weighing 150 to 200 gm. weredivided into two groups of 15 each, and inoculated intra-cardially with 0.5 ml. of an 18-hour tryptose broth cul-ture of E. coli, or approximately 10' living bacterial cellsas determined by plate counts. This strain of E. colihad been isolated from the blood of a patient a few daysbefore the experiment and had been subcultured onlytwice. Immediately after the bacterial inoculation, onegroup of 15 rats were subjected to bilateral renal mas-sage. The kidneys of each rat were grasped by thumb

1 Aided by research grant from Public Health Service,H-1955.

2 Presented in part at the 27th Annual Meeting of TheCentral Society for Clinical Research, October 30, 1954.

and forefinger through the intact abdominal wall (Fig-ure 1) and gently but firmly massaged during ether anes-thesia for exactly 5 minutes. In order to maintain auniform degree of pressure on the massaged kidneys,each period of 5 minutes was divided between two of usinstead of confining the total period of massage for eachanimal to one of several persons. The 15 rats in thesecond group were anesthetized under ether for 5 min-utes but their kidneys were not massaged. The kidneysof a third group were massaged for 5 minutes underether anesthesia but the rats were not inoculated withbacteria. This third group also served as controls forthe experiments in section 3b and are described there.

Within 48 hours, 4 rats in each group had died afterthe inoculation of E. coli. The survivors recovered fullyand appeared in excellent condition. At the end of twoweeks, the rats were sacrificed by exposure to etherand exsanguination. Approximately 1.0 ml. of bloodwas divided between trypticase soy broth and blood agarfor culture and 0.5 ml. was used for determining ureanitrogen. The kidneys were examined grossly and theneach divided into several representative parts for cul-ture and microscopic examination. For culture, the kid-ney was ground to a pulp with sterile sand, suspendedin water, and all of the remaining tissue cultured onblood agar, EMB agar, and tryptose broth. The kid-neys were fixed in 10 per cent formalin and the sectionsstained with hematoxylin and eosin and with Schiff'speriodic acid stain.

The procedures described in this experiment for renalmassage, bacterial inoculation, and postmortem examina-tion of tissues was that used in all experiments whichfollow, unless otherwise stated.

II. Effect of massage on localization of E. coli in kidneyas measured by bacterial counts

From a group of 9 rats inoculated intracardially withE. coli (same strain as in I), 3 each were sacrificed at 4minutes, 1 hour, and 2 hours after unilateral renal mas-sage had been completed. The massage was applied tothe right kidney of 4 rats and to the left kidney of the 5other rats. Immediately before sacrifice by excessive ex-posure to ether anesthesia, blood was drawn from theheart for plate counts. Each kidney was weighed,ground to a pulp, and suspended in 9.0 ml. of sterile wa-ter. Bacterial plate counts were then made on eachaqueous suspension of kidney tissue. The entire proce-dure was carried out aseptically.

1489

ABRAHAMI. BRAUDE, ALVIN P. SHAPIRO, AND JENNIE SIEMIENSKI

...

FIG. 1. TECHNIQUE FOR RENAL MASSAGEILLUSTRATEDIN A SACRIFICED RAT

An incision was made to allow visualization of thekidney as it is grasped between the thumb and forefinger.

An additional procedure was carried out with the kid-neys of one rat in each group of 3 sacrificed at the speci-fied time intervals. Before suspensions of ground tissuewere prepared for plate counts, these kidneys were di-vided in half and the cortex and medulla from one halfwere separated. Then each separated portion (intacthalf, cortex, medulla) of the six kidneys was weighed,ground, and suspended in water.

III. Serial studies of pathologic changes in massagedkidneys after inoculation of E. coli

a. Changes produced by a straiit of high virulence.From a group of 22 rats, 5 each were sacrificed at 2, 24,and 48 hours and the remaining 7 at 7 days after inocu-lation of approximately 108 E. coli and bilateral renalmassage. Each rat was inoculated with either of twostrains of E. coli which had been isolated from the urineof patients with pyelonephritis. The two strains ap-peared to be of equal pathogenicity. Cultures were madeof the urine, heart blood, and kidneys; the urinary sedi-ment and the kidneys were examined microscopically;and urea nitrogen was measured on samples of heartblood drawn immediately before sacrifice.

b. Changes produced by a strain of low virulence.Rats were divided into 5 groups of 10 each. One groupof 10 received bilateral and another group unilateral(right kidney only) renal massage immediately aftereach animal had been inoculated intracardially with ap-proximately 108 E. coli. This strain had been isolatedfrom the urine of a patient with pyelonephritis and itsvirulence was reduced by repeated subculture. Of thethree remaining groups of 10 each which served as unin-fected controls, one group received bilateral renal mas-sage, a second unilateral renal massage, and the thirdno massage. Two rats from each of the five groupswere sacrificed at 1, 11, 17, 24, and 42 days after inocu-lation and the same examinations were carried out as inIIIa. In addition, blood pressures were measured justbefore sacrifice by the tail-plethysmographic method ofWilliams, Harrison, and Grollman (4).IV. Serial studies on utrine obtained from living rats

by cystotomy

The experiment in IIIb was repeated but instead ofsacrificing the animals, urine was obtained from them bycystotomy during anesthesia. The urines from 2 or morerats in each group were obtained at 2, 3, 5, 6, 8, 10 and15 weeks after inoculation and then cultured and ex-amined microscopically; but no rat was subj ected tocystotomy more than once for the first 6 weeks. After6 weeks, each of the animals examined was subjectedto its second cystotomy. In addition, each of the inocu-lated rats had received two additional intracardiac in-jections of the same strain of E. coli as follows.

7 weeks-Intracardiac injection of 103 E. coli with norenal massage. This was done to determine whether theinitial renal infection had increased the vulnerability ofthe kidneys so that pyelonephritis could now be estab-lished with relatively small numbers of E. coli.

9 weeks-Intracardiac injection of 106 E. coli accom-panied by a second bilateral or unilateral renal massage.This was done to determine whether prior renal infectionhad rendered the kidneys relatively immune to reinfection.

Cystotomies were done aseptically on rats anesthetizedwith ether. The bladder was exposed by a small su-prapubic incision and urine withdrawn by syringe througha 26 gauge needle. The abdominal wall was repairedwith a few silk sutures. At the time of second cys-totomies, complete healing of the original cystotomy wasusually found.

Before each cystotomy, cardiac blood was drawn fromeach anesthetized rat for culture and urea nitrogen de-termination. Blood pressures were measured everyother week.

RESULTS

I. The effect of renal massage on the incidence ofpyelonephritis in rats inoculated with E. coli

The results summarized in Table I disclose that95 per cent of the massaged kidneys were infectedin contrast to only 23 per cent of the control kid-

1490

EXPERIMENTAL HEMATOGENOUSPYELONEPHRITIS

TABLE I *

EFFECT OF RENAL MASSAGEON THE DEVELOPMENTOF PYELONEPHRITISAFTER INTRACARDIAC INJECTION OF E. COLI IN RATS

TOTAL E. COLI ISOLATED FROM GROSS MICROSCOPICNUMBER PATHOLOGIC EVIDENCE OF

OF RATS BLOOD KIDNEY CHANGES PYELONEPHRITIS

NO MASSAGE 1I 0 5 0 0

BOTH KIDNEYS II 0 21 14 20MASSAGED

* Four rats in each group of 15 died within 48 hours after inoculation ofE. coli.

iieys after inoculating rats with E. coli. Evenmore striking was the observation that althoughno pathologic changes were found in any controlkidneys, they were so extensive in the massagedgroup that gross evidence of pyelonephritis was

found in 64 per cent and microscopic evidence in91 per cent. Massaged kidneys of rats receivingno inoculation of E. coli were invariably sterile andonly rarely displayed any sign of trauma fromrenal massage. Such trauma in non-infected mas-

A B

FIG. 2. MICROSCOPICCHANGESIN PYELONEPHRITIc KIDNEY INFECTED WITH VIRULENT STRAIN OF E. Ccli AND

SACRIFICED AT 2 WEEKSA. WEDGESHAPEDAREA OF INTERSTITIAL INFLAMMATION EXTENDINGFROMCORTEXTO MEDULLA (X 100)B. DILATED TUBULES ARE FILLED WITH POLYMORPHONUCLEARCELLS, EVEN THOUGHCELLS COMPRISING THE IN-

FLAMMATORYEXUDATEIN THE INTERSTITIAL TISSUES ARE VIRTUALLY ALL MONONUCLEAR(X 700)Only polymorphonuclears are sufficiently motile to move into the tubules in large numbers.

1491

..X t.

1492

IMET

FIG. 3. CoMpRESPONDING NOPHOTOGRAPHEDMINUTE PERIOD

ABRAHAMI. BRAUDE, ALVIN P. SHAPIRO, AND JENNIE SIEMIENSKI

opposite side which were not massaged (Figure3). This difference in weight between the mas-saged and non-massaged kidneys disappearedwithin two hours as the hyperemic appearancesubsided. The number of bacteria per gram ofmassaged kidney remained about 10 to 100 timesgreater than that of the control and exhibited abacterial concentration of 10 to 100 times greaterthan that of the blood. The higher bacterial countin the massaged kidney resulted from an initialheavy localization of E. coli in the cortex alone asillustrated by the 100 fold increase in bacterial count

2 3 4 5 6 observed in the cortex of massaged over controlc

r kidney. Later, the bacterial counts in the medullaof the massaged kidney also increased over that

?ARISON OF MASSAGEDKIDNEY AND COR- of the control. These results are given in detailIN-MASSAGED KIDNEY FROM SAME RAT in Figure 4 for the 3 rats whose kidneys wereIMMEDIATELY AFTER COMPLETION OF 5- separated into cortex and medulla and in TableOF MASSAGE II for the remaining 6 rats.

Note marked hyperemia and enlargement of massagedkidney.

saged kidneys was never manifested by more thana minute hemorrhage or very small areas of tubu-lar necrosis.

The kidneys which developed pyelonephritiswere usually enlarged and sometimes showed smallabscesses beneath the capsule. On section, therewere frequently diffuse areas of suppuration visi-ble. The microscopic appearances of these kid-neys were identical to that seen in human pyelo-nephritis (5). There were wedge shaped areasof interstitial inflammation with its base in thecortex and the apex extending down into themedulla. The process usually spread up to theglomerulus but did not invade it. Although thegreat majority of cells comprising the inflamma-tory reaction in the tissues at two weeks weremononuclear, those in the tubules were almost allpolymorphonuclear. Many areas containing di-lated tubules filled with pus cells were found inassociation with the interstitial inflammation.The microscopic changes are illustrated in Figure2.

II. Effect of massage on localization of E. coli inkidneys as measured by bacterial counts

Immediately after massage, the kidneys ap-peared to contain an increased amount of bloodand were much heavier than the kidneys on the

III. Serial studies of pathologic changes in mas-saged kidneys after inoculation of E. coli

a. After inoculation of a highly virulent strainof E. coli no inflammatory cells appeared in thekidneys or urine during the first 2 hours eventhough cultures of all kidneys and urines yieldedlarge numbers of E. coli. At 24 hours pyuria wasobserved only in the 2 rats in which the early mi-croscopic signs of pyelonephritis were also found.

,s 6

StoU)

55)

0:3

0-j.

Effect of Renal Massageon Bacterial Counts ofKidney After IntracardiacInjection of 108 Viable

E. coli

2.22 gm. MASSAGED

.56- KIDNEY1.48

1.63\BLOOD

.51- CONTROL

KIDNEY(Not massaged )

O 1 a

Effect of Renal Massageon Distribution of Bac-teria in Kidney Cortexand Medulla After Intra-cardiac Injection of 108

Viable E. coli

MASSAGEDKIDNEY

. . Cortex)

Medulla

CONTR CortIex

Medul lc

CONTROLKIDNE)

0Hou rs

1 2

FIG. 4. EFFECT OF RENAL MASSAGEON BACTERIALCOUNTSOF KIDNEYS AFTER INTRACARDIAC INOCULATIONOF E. Coli

EXPERIMENTALHEMATOGENOUSPYELONEPHRITIS

TABLE II

EFFECT Of RENAL MASSAGEON BACTERIAL COUNTSOF KIDNEY AFTER INTRACARDIAC INACTION.OF 108 VIABLE E. COLI

TIME ELAPSED IDENTIFICATION TOTAL NO. WEIGHT OF NO. BACTERIA NO. BACTERIA(MINUTES) NUMBER MASSAGED BACTERIA KIDNEY PER GM. PER Ml.

AFTER INOC. OF RAT PER KIDNEY (m..) OF KIDNEY OF BLOOD

No L 85,400 1.0154 84,138

4 tYes R 740,733 1.2821 577,800 2.330,666

Yes L 254,866 1.3204 193,0802

No R 29,500 1.1294 25,243 2,620,666No L 59,133 1.0081 58,547

3

o60 Yes R 327,866 1.1782 278,325 15,225Yes L 145,400 1.2113 120,165

4No R 8,550 1.1695 7.307 11,580

_ No L 35,533 1.1671 30,370

120Yes R 449,733 1.2057 371,680 1,173

6 Yes L 72,700 1.2619 57,700

No R 18,400 1.2996 14,154 3,013

Ls Left K I dneyRz Right Kidney

There were no pus cells in the urine of the otherrats examined at 24 hours, and there was no pyelo-nephritis. Similarly, at 48 hours, the finding ofpyuria was limited to those rats whose kidneysdisplayed the typical gross and microscopic find-ings of acute pyelonephritis. Yet the urine cul-ture from all rats yielded E. coli during the firstand second days. By one week, pyelonephritis

was present in 13 kidneys of all 7 rats and pyuriawas present in the urine of all rats. The resultsare summarized in Table III.

The earliest evidence of pyelonephritis was

found at 24 hours when neutrophils began to at-tack bacterial masses situated between the tu-bules. By 48 hours, interstitial abscesses com-

posed of dense accumulations of neutrophils, were

TABLE III *

CORRELATION OF CULTURES AND CELLULAR CHANGES IN EXPERIMENTAL PYELONEPHRITIS

TIME ELAPSED AFTER INTRACARDIAC INJECTION VIRULENT STRAIN 108 E. COLI

2 hours day 2 days 7 days

U B RK LK U B RK LK U B RK LK U B RK LK

Rat + +1H + + H+ + + *f -H* + * 0 ++1+ -4+1+ 4+1+ + ++H+ ++ P PN-H4-j- PNI444 P+ PN* PN*

Rat + +"+ 1Hl*+H + + +H+ +H+ + + 441 **1 P+ PN* +2 Pi+ ± PN++f+ PN

Rat ++f+ +H+ +*+ + + +H+1 +H + * 44H* +* +H4 444* 4H+3 ~~~~~~~~~~~~~~~~~~~~~P444PN-+H+ PN-f4

Rat +4+144+4+H4+ 4-+ +H+ +H+f + + -4++ +1+ 4- 444+ +P+ PN + pf PN+H+ PN + p+P44+ PN 44-

Rat + +H+f +H+*f+ +++H+1+H+ -H - ++ 4+ +1- 41+ +P+ PN + P-f PN+H+ PN44+* P * PN*+H4 PN**f

Rat 4 o +4 H6 F * -f PN4H+ PN4

Rat + 444+ 4+7 C J | P44- PN+H+ PN 4+

P * PyuriaPN = Pyelonephritis

* Cultures were made of blood (B), urine (U), right kidney (RK),(P = pyuria; PN= pyelonephritis; 0 = sterile.)

and left kidney (LK).

1493

ABRAHAMI. BRAUDE, ALVIN P. SHAPIRO, AND JENNIE SIEMIENSKI

d~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-1

% g4_

-~ ~ ~ ~~-

V.

A.-~

V E4t,

8C<

M w~~~~~~~~ Sr w1. *~~~~~~~~~~~~~k

# 'i .' ... #~~~~~~~~~~~~~~~~~r

ZF0e~~~- A

FIG. 5. MICROscoPIc APPEARANCE OF INTERSTITIALSUPPURATIONIN PYELONEPHRITIC KIDNEY OF RAT SAC-RIFICED 48 HOURSAFTER INOCULATION OF E. Coli

observed grossly and microscopically in the cortexand medulla (Figure 5). At one week, however,many lymphocytes and plasma cells began to par-

ticipate in the pyelonephritic process. These mo-

nonuclears, accompanied by an equal number ofneutrophils, comprised a dense infiltrate whichusually extended from cortex to medulla in a

wedge shaped fashion. The glomeruli were usu-

ally spared although the tubules were frequentlydestroyed or filled with large clumps of neutro-phils.

The values for B.U.N. and blood pressure are

given in Table IIIA. The B.U.N. ranged from26 to 46. A few of these values, when comparedwith those of normal controls given in Table VA,are slightly elevated. The blood pressures on 5rats taken at 1 week were all normal.

b. After inoculation of a strain of low virulence,the infection is characterized by very few signsof inflammation at any time. The infection was

self-limited to about 3 weeks. Despite repeatedlypositive cultures of kidney tissue, no renal inflam-mation was found during the first 17 days and nopyuria developed. In addition, urine cultures weresometimes negative even though the kidneys wereinfected with E. coli. It is particularly note-worthy, however, that microscopic evidence ofpyelonephritis sometimes appeared after the infec-tion had disappeared. This late inflammation,which was entirely mononuclear in character, wasnot accompanied by pyuria. The results are sum-marized in Table IV. The systolic blood pressureof one animal was 130 mm. Hg at 42 days afterbilateral renal massage and inoculation E. coli.The blood pressures of all other animals receiv-ing renal massage were 115 or less. The bloodpressures in normal control rats were 120 or less.The B.U.N. ranged from 24 to 31 in normals andfrom 20 to 33 in the animals receiving renal mas-sage with inoculation of E. coli. No cultural orpathologic sign of pyelonephritis appeared in anyof the control rats.

IV. Serial studies on urines obtained from liv-ing rats by cystotomy

Table V summarizes the results of serial exami-nations of urines obtained by cystotomy fromrats after renal massage and inoculation of astrain of E. coli of low virulence. As in SectionILIb, pyuria was uncommon and renal infectioncaused by this strain appeared to be self-limited;in these animals it subsided after about 6 weeks.After subsidence of infection, reinoculation with

TABLE IIIA *

TIME ELAPSED AFTER INTRACARDIACINJECT ION 108 E. COLI

RAT NUMBER 2 HRS. I DAY 2 DAYS 7 DAYS B.P.

44 ND 34 26 ND

2 36 ND ND 34 ND

3 32 28 30 46 1104 32 ND 38 32 115

5 . 34 46 31 42 100

6 34 42 100

7 38 105

Blood urea nitrogens of rats sacrificedserially. Blood pressures just beforesacrifice are also given for those examinedat 7 days. (ND =No determination)

* These results were obtained on rats listed in Table III.

14-94

EXPERIMENTALHEMATOGENOUSPYELONEPHRITIS

TABLE IV *

RESULTS OF CULTURE OF URINE (U) BL000D B) RIGHT KIDNEY (2R) AID LEFT KIC CEY(LK)UP6N SACRI IC~ OF RATS

DAYS ELAPSED AFTER INTRACARDIAC INOCULATION OF IO8 E. COLI

I I I 17 24 42

U S RK LK U B RK LK U B RK LK U 8 RK LK U B RK LK

KASSAGE Rat O + + + 0 0 + + 00 0 0 0 0 0 0 00 0 0BOTH I

KIDNEYS Rat + + + + + 0 + 0 00 + 0 00 0 0 0 0 0 021PR PU

MASSAGE Rat + + + + 0 + + 0 00 0 00000 0 0 0 0 0RIGHT 3

KIDNEYONLY Rat O + + + O O + O O O O 000 + 0 00 0 0

4 PN PN

PN a PyelonephritisNo Pyuris observed in any animal.

* + = isolation of E. coli; 0 = sterile. Isolation of E. coli from left kid-neys, 1 day after inoculation of rats receiving renal massage only on rightside, was attributed to the presence of E. coli in the blood contained in thosekidneys.

a relatively small number of E. coli without renalmassage did not lead to recurrent signs of renalinfection; but bacteriuria was reestablished at 10weeks by renal massage in conjunction with theinoculation of larger numbers of E. coli. Thelatter results suggest that upon subsidence ofpyelonephritis, the kidneys were neither undulysusceptible to reinfection by small bacterial inocu-lums, nor immune to reinfection by the usual in-oculums. No sign of urinary infection appearedat any time in urines obtained from control rats.The highest blood pressure obtained in any animalwas 120 mm. Hg. The B.U.N. in rats subjectedto renal massage and inoculation of E. coli rangedfrom 25 to 35, while the B.U.N. of normal con-

trols varied from 26 to 34. These results aregiven in Table VA.

DISCUSSION

The pathologic appearance of the kidneys ofrats, infected with virulent strains of E. coli, wasalmost identical to that seen in human pyelone-phritis caused by E. coli. In both cases, the in-flammation is interstitial and extends in a wedgeshaped fashion from its base in the cortex downto the medulla (5). In rats, as in man, it was alsofound that urinary findings may be normal de-spite extensive infection in the interstitial tissueof the kidney (6). The acute pyelonephritis in

TABLE. V *

RESULTS OF URINE CULTURE FOR E. COLI OBTAINED BY CYSTOTONYWITHOUT SACRIFICING RATS

WEEKSELAPSED AFTER FIRST INTRACARDIAC INJECTIONOF 108 E. COLI

2 3 5 6 7 8 9 10 15

RAT .j. +00 0[[MASSAGE I °P Intra Intra _ O

BOTH - cardiac cardiacKIDNEYS RAT -11 + 0 0 injeclion 0 inJec ion 0 +

ONLYRAT ~~~+ ~ in E. Coli omE. Coli

MASSAGE RAT 0 0 0 0 and 0RIGHT 3 NO kidneysKIDNEY MASSAGEH massaged

ONLYRAT++. +[01 P _

P = P uriaAll blood culture sterile.BUN and Blood pressure normal on all animals.

* The designations "Rat 1, Rat 2, Rat 3, Rat 4" are used only to identifyrats within the group examined on a given day; no rat was subjected tocystotomy more than once for the first 6 weeks. In addition to those listed,urines were also sterile from 2 rats receiving bilateral renal massage at 6weeks and from 2 rats receiving unilateral renal massage at 15 weeks.

1495

ABRAHAMI. BRAUDE, ALVIN P. SHAPIRO, AND JENNIE SIEMIENSKI

TABLE. VA *

WEEKSELAPSED AFTER FIRSTINTRACARDIAC INJECTION OF 108 E. COLI

Blood urea nitrogens and blood pressures of rats examined periodicallyby cystotomy.

* The B.U.N. is the upper value, and the blood pressure the lower value given ateach time interval for each group of rats. The values given are averages for allrats examined in a group. (N.D. = no determination.) These results were ob-tained on rats listed in Table V and on the controls.

rats and in man also resemble each other in thathypertension and usually uremia are absent inboth cases. These similarities between pyelone-phritis in humans and in these rats suggest thatthe experimental disease may be used as a modelfor studying certain aspects of human pyelone-phritis.

Renal massage appears to favor the develop-ment of hematogenous pyelonephritis in bactere-mic rats by causing increased numbers of bacteriato localize in the kidney. It is not clear from thepresent data whether this heavy localization ofE. coli resulted from renal damage, urinary ob-

struction, or from the delivery of more bacteriathrough increased renal flow of infected blood.Although no obstruction of the urinary tract andonly rare renal damage were evident at post-mortem examination of the massaged kidneys,these abnormalities might have been detectable iffunction studies had been carried out. Likewise,in the absence of hemodynamic studies, it cannotbe established whether the hyperemic appearanceof the massaged kidneys represents increasedblood flow or simple engorgement and stasis. Itis also possible that the relatively poor lymphaticsupply of the kidney was damaged and not ade-quate to remove the excessive numbers of bac-teria which could then establish infection. Pierce(7) has demonstrated that in contrast to skin, in-

testine, or uterus, there are very few renal lym-phatics present to take over as emergency unitsto handle any additional load over that normallypresent. In the present experiments, bacteriawhich lodged in the kidneys repnained there eventhough they were of such low virulence that theyinduced no necrosis or inflammatory reaction.This would account for the persistence of E. coliin the kidney and urine for several weeks in theabsence of inflammatory cells.

Once the experimental infection was established,its manifestations varied with the virulence ofthe organism. Highly virulent strains led to sup-puration and pyuria. Strains of lower virulenceproduced a mild self-limited infection with some-times little or no sign of renal inflammation. Uponanalysis of these two types of infectious processesin the experimental animals, several points becomeevident which help explain certain clinical findingsin human pyelonephritis which are not well under-stood. For example, it was found in suppurativeexperimental pyelonephritis that, although thecellular reaction in the tissues was predominantlymononuclear by two weeks, virtually all the cellsin the tubules were polymorphonuclears. Thisdifference is accounted for by the great motility ofpolymorphonuclears which enables them to moveout of the inflamed areas into the tubules. Thisexperimental observation is useful in explaining

mnoc. with E. Colil 2 3 5 6 0 1at 0, 7, & 9 wks. 2 3 5 6 8 10- 5Massage both 25 29 27 30 26 35 ND

Yeskidneys 117 III 112 112 107 109 110Massage right 28 25 30 25 33 ND ND

Yeskidney 116 107 109 112 110 107 112No massage 30 31 27 29 30 31 ND

Yesof kidneys 112 106 106 104 103 106 109Massage both 29 29 26 31 27 31 ND

Nokidneys 108 109 107 113 109 115 115No massage 34 29 28 32 ND 26 ND

No(normal controls)I 109 104 1106 1110 1109 108 107

1496

EXPERIMENTALHEMATOGENOUSPYELONEPHRITIS

why the polymorphonuclear is almost the onlyleukocyte found in the urinary sediment in humanpyelonephritis and why the urinary sediment inchronic pyelonephritis frequently contains no cellsafter the renal inflammation becomes lymphocyticin character.

During the experimental renal infection bystrains of low virulence, there was a differentreason for the absence of pyuria. Here the cellu-lar reaction in the kidney itself was so slight, ifpresent at all, that there was often no significantfocus of polymorphonuclears to supply the tubules.In such an infection, only motile bacteria werepresent in the urine and even these were oftenabsent. Whether a comparable form of acutenon-suppurative pyelonephritis occurs in humanbeings can only be conjectured, however, becausepathologic and cultural examinations of such kid-neys would be almost impossible to secure. Yetthere is abundant evidence to suggest that a clini-cal entity of this type exists. The syndrome offever, chills, sweats, and bacteriuria occurring inthe absence of flank pain, costovertebral angletenderness, and pyuria is not uncommonly pro-duced by E. coli. The course of this disease isgenerally self-limited to a few days or weeks,but may recur for many years and lead to chronicpyelonephritis (6). In this connection, it is ofinterest that in the experimental infection, therewas not only a lack of immunity against subse-quent attacks of E. coli pyelonephritis, but alsoevidence which suggested that the pyelonephriticchanges may first appear as the infection vanishes(Table III). Thus it may be possible for re-peated infections by the same bacterial species tooccur and initiate inflammatory reactions whichperpetuate themselves after the bacteria disappear.

These phenomena and others relating to chronicpyelonephritis are under investigation in ratswhose renal infections have persisted for over 6months. It is anticipated that these studies nowin progress may illuminate some of the problemsrelated to chronic human pyelonephritis.

CONCLUSIONS

Renal massage appears to favor development ofhematogenous pyelonephritis in bacteremic ratsby causing large numbers of bacteria to localize inthe kidney. The severity of the pyelonephritisdepends on both renal massage and bacterial viru-lence. Renal infection (proved by positive cul-tures of kidneys) may exist in the absence ofpyuria and bacteriuria.

REFERENCES

1. Spitznagel, J. K., and Schroeder, H. A., Experimentalpyelonephritis and hypertension in rats. Proc.Soc. Exper. Biol. & Med., 1951, 77, 762.

2. Mallory, G. K., Crane, A. R., and Edwards, J. E.,Pathology of acute and of healed experimentalpyelonephritis. Arch. Path., 1940, 30, 330.

3. Lepper, E. H., The production of coliform infectionin the urinary tract of rabbits. J. Path. & Bact.,1921, 24, 192.

4. Williams, J. R., Jr., Harrison, T. R., and Grollman,A., A simple method for determining the systolicblood pressure of the unanesthetized rat. J. Clin.Invest., 1939, 18, 373.

5. Bell, E. T., A Textbook of Pathology, 5th ed., Phila-delphia, Lea & Febiger, 1944.

6. Weiss, S., and Parker, F., Jr., Pyelonephritis: Itsrelation to vascular lesions and to arterial hyper-tension. Medicine, 1939, 18, 221.

7. Pierce, E. C., 2nd, Renal lymphatics. Anat. Rec., 1944,90, 315.

SPECIAL NOTICE TO SUBSCRIBERS

Post Offices will no longer forward the Journal when you move.

Please notify The Journal of Clinical Investigation, BusinessOffice, 622 West 168th Street, NewYork 32, N. Y. at once whenyou have a change of address, and do not omit the zone number ifthere is one.

1497