Kidney International, Vol. 6 (1974), p. 24—37

Medullary plasma flow and intravascular leukocyteaccumulation in acute renal failure

KIM SOLEZ, ELIZABETH C. KRAMER, JENNIFER A. Fox and ROBERT H. HEPTINSTALL

Department of Pathology, The Johns Hopkins University School of Medicine and Hospital, Baltimore, Maryland

Medullary plasma flow and intravascular leukocyte accumula-tion in acute renal failure. The accumulation of nucleated cells inthe vasa recta in acute renal failure has been attributed to intra-vascular hematopoiesis caused by reduced medullary bloodflow. Medullary plasma flow was determined in rats with HgCI2-induced acute renal failure by infusing '251-albumin iv. for 15or 30 sec and measuring radioactivity in plasma and the renalpapillae. Measurements made 40 to 48 hr after i.v. injection ofHgCI2 revealed a significantly raised medullary plasma flow at atime when there were prominent collections of nucleated cellsin the vasa recta. Vascular tracer studies using India ink demon-strated that blood flow was maintained in vasa recta whichcontained nucleated cells. Electron microscopic and histo-chemical studies showed that these cells were lymphocytes,neutrophils and monocytes, present in approximately the sameproportions as in the peripheral blood. It is suggested that thenucleated cells in the vasa recta are not produced within thekidney, but are brought there via the circulating blood. Theiraccumulation within the vasa recta is unexplained but is un-related to a reduction in medullary blood flow.

Debit médullaire et accumulation intravasculaire de leucocytesdana l'insuffisance rénale aiguë. L'accumulation de cellulesnucldées dans les vasa recta au cours de l'insuffisance renal aiguëa etc attribuée a l'hématopoIEse intravasculaire déterminde parle diminution du debit sanguin médullaire. Le debit médullairea ete dCterminé chez des rats atteints d'insuffisance renale aigudinduite par HgC12 au moyen d'une perfusion intraveineused'albumine 1251 durant 15 ou 30 seconds et de la mesure de Iaradioactivité du plasma et de la papille rénale. Les détermina-tions faites 40 a 48 heures aprés l'injection intraveineuse deHgC12 montrent une augmentation significative du debitmedullaire contemporainc de l'existence de quantitds impor-tantes de cellules nucléées dans les vasa recta. Les etudes vascu-laires a l'aide de l'encre de chine comme traceur dCmontrent quele debit sanguin est maintenu dans les vasa recta qui contiennentdes cellules nucleCes. La microscopie électronique et les etudeshistochimiques montrent que ces cellules sont des lymphocytes,des neutrophiles et des monocytes, approximativement dans lesmêmes proportions que dans le sang périphérique. Ii est suggéréque les cellules nucléées des vasa recta ne sont pas produites dansle rein mais y sont apportées par Ia circulation sanguine. Leuraccumulation a l'intérieur des vasa recta n'est pas expliquée maiselle n'est pas Ia consequence d'une diminution du debit sanguinmedullaire.

Received for publication December 4, 1973;and in revised form February 4, 1974.© 1974, by the International Society of Nephrology.

24

It is well established that renal cortical blood flow ismarkedly reduced in acute renal failure [1—5] but thesituation with regard to medullary blood flow is lessclear. Inert gas washout techniques for measuringrenal blood flow have demonstrated an apparentincrease in medullary blood flow in acute renal failure[1, 4, 5] but this has been attributed to a merging of thecortical and medullary components of the washoutcurves rather than to a true increase in medullaryblood flow [1, 5].

One very common morphologic feature of acuterenal failure which has been interpreted as suggestingthat medullary blood flow is decreased is the presenceof large numbers of lymphocytes, neutrophils andother nucleated cells in the vasa recta of the renalmedulla. In an autopsy series, this lesion was observedin 63 of 66 unselected cases of acute renal failure ofmore than one day's duration and was sometimes theonly definite histologic lesion present [6], a fact notedpreviously by Finckh, Jeremy and Whyte [7]. Baker[8] attributed the accumulation of nucleated cells in thevasa recta to intravascular hematopioesis. Otherinvestigators [9—12] postulated that the nucleated cellswere released into the circulating blood by the bonemarrow, spleen and lymph nodes and then becamelodged in the vasa recta due to hemodynamic factors.Both Baker [8] and Remmele [12] considered thislesion to be the result of decreased medullary bloodflow.

Recent review articles discussing the circulation ofthe kidney [13—15] have emphasized the technicaland interpretive difficulties inherent in the measure-ment of medullary blood flow using methods currentlyavailable. In an effort to overcome these difficulties, wehave developed a simple method of determiningmedullary plasma flow in the rat based on the radio-active albumin infusion methods of Lilienfield, Mag-

Medullary plasma flow in acute renal failure 25

anzini and Bauer [16] and Ganguli and Tobian [17].This paper describes the use of this method in theinvestigation of renal medullary hemodynamics inacute renal failure induced by mercuric chloride, andthe relationship of medullary plasma flow to theaccumulation of nucleated cells in the vasa recta. Inaddition, medullary plasma flow was measured aftersaline loading and in shock.

Methods

Female Holtzman rats weighing 180 to 220 g werehoused in pairs, or singly in metabolism cages, andfed a routine diet of standard laboratory chow(Purina, Ralston Purina Co., St. Louis, Mo) and un-limited tap water.

Determination of medullary plasma flow.1 In prelim-inary experiments using the method of Lilienfield et al[16] as modified by Ganguli and Tobian [17] for therat, we observed a great deal of apparent variation inmedullary blood flow among control animals, con-firming the experience of the original investigators whodescribed these methods. We were able to reduce thisvariability by infusing the radioactive albumin intra-venously instead of intraarterially. Further modifica-tion of the necessary surgical techniques resulted ina method which gave reproducible results.

The method is based on the fact that the transit timefor circulating albumin in the renal papilla of thenormal rat is about 40 sec (see Results section). Ifradioactive albumin is allowed to circulate throughthe kidney for 30 sec, the accumulation rate ofpapillary radioactivity during this period is propor-tional to plasma flow rate. This determination ofmedullary plasma flow does not take into accountglomerular filtration. However, since the glomerularfiltrate is largely reabsorbed before the papillaryportion of the descending loop of Henle is reached, itcontributes directly to medullary plasma flow. Ifglomerular filtration is not taken into account, there isless than a error in the determination of medullaryplasma flow (see Discussion section).

To determine medullary plasma flow, the rats wereanesthetized with sodium pentobarbital (50 mg/kg)i.p. and placed on a warmed platform to maintainnormal body temperature (37.5 to 38.8°C). The trachea,left jugular vein and right carotid artery were cannu-lated. Carotid artery pressure was monitored using apressure transducer (Sanborn Co., Waltham, MA).Except for the group of animals with hemorrhagicshock, only experiments in which mean blood pressure

1 The method described measures inner medullary (papillary)plasma flow, referred to in this paper as "medullary plasmaflow".

remained above 90 mm Hg were accepted. A midlineabdominal incision was made and loose, nonocclusive1—0 silk ties were placed around the renal pedicles.Using a syringe pump (Sage Instruments, Cambridge,MA) a solution of '251-albumin (Squibb and Sons,Princeton, NJ) (1 tCi/ml) in 0.9% NaCI, to which asmall amount of sodium fluorescein had been added,was infused into thejugular vein at a rate of0.37 ml/minunder UV light until 30 sec after the appearance ofyellow fluorescence at the kidney hilus. Jugular vein-to-kidney circulation time proved to be 2.5 to 3.0 sec inall experimental groups except for the shock animals,in which it averaged 4.0 to 5.0 sec. At the end of theinfusion period, the ties around the renal pedicleswere pulled tight. A continuous sample of arterialblood (total volume, 0.3 to 1.0 ml) was collected at aconstant rate through the carotid cannula,, for the final30 sec of the infusion period. This procedure resultedin the radioactive albumin circulating through thekidneys for a 30-sec period which coincided with theblood collection period. For the HgCl2-treated group,in which papillary transit time was less than 30 sec, a15-sec period was used as well.

The kidneys were removed with the pedicle tiesintact and cut into slightly uneven "halves" from poleto pole so that all of the papilla remained in just onehalf. In the kidneys of all experimental groups, thepapilla could be readily distinguished from the outermedulla. The papillae were cut out using smalldissecting scissors, wiped lightly against nonabsor-bent paper to remove adhering blood and placed intocapped, preweighed scintillation counting tubes.2The remaining portions of the kidneys were fixed inbuffered 10% formaldehyde solution and prepared forhistologic examination, which confirmed that thepapilla had been removed and the outer medulla leftintact. The hemocrit value of the blood sample wasdetermined and an aliquot of the blood sample wasplaced in a scintillation counting tube. After weighingthe papillae, the radioactivity of the blood sample andof the papillae was measured in a gamma wellscintillation counter (Baird-Atomic Inc., Cambridge,MA) and expressed as cpm/ml of plasma and cpm/100 g of papilla, respectively. (Sample height did notexceed 1 cm, thus minimizing geometric error ingamma counting [18].) By dividing papillary radio-activity by plasma radioactivity, medullary plasma

2 In a departure from the method of Lilienfield et al [16], wedid not quick-freeze the kidneys before sectioning them, sincepreliminary experiments with control animals produced repro-ducible results whether the kidneys were used in the fresh or inthe frozen state (31.34± 8.34 (SD) mI/min/lOO g for 12 frozenkidneys vs. 32.22±5.16 (SD) mI/mm/bOg for 22 unfrozenkidneys).

26 Solez et a!

flow was determined, expressed in units of ml plasma/min/100 g of papilla.

Determination of papillary transit time for albumin.To determine the approximate transit time for circu-lating albumin in the renal medulla, 20 normal rats and27 rats with HgCI2-induced acute renal failure wereprepared as already described for medullary bloodflow determination. After '251-albumin had beenallowed to circulate through the kidney for 15,30,45 or60 sec, the renal circulation was interrupted and theapparent volume of distribution of albumin in the renalpapilla was determined by comparing papillary radio-activity to average plasma radioactivity. The time atwhich volume of distribution leveled off was taken torepresent transit time for albumin.

Induction of acute renal failure. With the animalunder ether anesthesia, after a small lateral neckincision had been made, mercuric chloride solution(2 mg/mI of saline solution) was injected into thejugular vein of rats in a dose of 5 mg/kg. The incisionwas closed with wound clips and the animals placed inmetabolism cages for urine collection under mineraloil. Urine albumin concentration was determinedusing Albustix (Ames Co., Elkhart, IN) and urineoutput for the 24-hr period preceding determination ofmedullary plasma blood flow was measured. The samedeterminations were also carried out on six controlrats. Medullary plasma flow was determined at 40 to48 hr after mercury injection, a time at which therewere numerous nucleated cells in the vasa recta. Serumurea nitrogen and serum creatinine concentrationswere determined on blood obtained at the time of thisprocedure using an autoanalyser (Technicon Corp.,Tarrytown, NY) as previously described [19].

Test for glomerular or tubular albumin leak. Todetermine whether substantial amounts of radioactivealbumin might be entering the papilla via the glomeru-lar filtrate during the 30-sec infusion, the ureters of sixrats injected with mercuric chloride, 5 mg/kg 42 to 48hr before, were catheterized with PE 10 tubing(Intramedic, Clay Adams, Parsippany, NJ). After i.v.injection of 0.18 Ci of 1251-albumin (the amountused for medullary plasma flow determinations),urine flow rate and radioactivity of the urine samplesobtained during the next four hours were determined.

Determination of papillary water content. The watercontent of the papillae of eight control rats and eightrats with acute renal failure 48 hr after HgCl injectionwas determined by comparing wet weight to dry weightafter drying the tissue to constant weight in an oven at80CC.

Induction of hemorrhagic shock. Thirteen normal ratswere prepared for medullary plasma flow determina-tions and samples of 3 to 4 ml of blood were removed

via the carotid cannula to produce mean bloodpressures of 30 to 70 mm Hg after a 30-mm equilibra-tion period. Medullary plasma flow was then deter-mined as already described.

Saline loading. Each of eight normal rats receivedan i.v. infusion of 0.9% NaCI equal to l0% of itsbody wt at a constant infusion rate over a period ofone hour, after which medullary plasma flow wasdetermined.

Vascular tracer studies. In 13 rats, acute renal failurewas induced with mercuric chloride (5 mg/kg) asalready described. Forty-eight hours after mercuricchloride injection, a midline abdominal incision wasmade and loose ties were placed around the renalpedicles. One milliliter of India ink diluted 1: 1 with0.9% NaC1 was injected into the inferior vena cava.The ties around the renal pedicles were pulled tight 5,10, 15, 20, 25, 30 or 35 sec after ink injection (N=4 for15- and 20-sec groups; N= 1 for other groups). Bothkidneys of each rat were removed and prepared forhistologic examination. An indentical procedure wascarried out in seven normal rats.

White blood cell and differential counts, histochemicalstudies. In five control rats and seven rats with renalfailure 42 to 48 hr after the injection of mercuricchloride (5 mg/kg), white blood counts were deter-mined in duplicate on heparinized blood samplesusing a hemocytometer (American Optical Co.,Buffalo, NY). Differential counts were done in tripli-cate (300 cells counted for each animal) using smearsprepared with Wright's stain. Smears of peripheralblood and 4 t frozen sections of the kidneys from sixrats with acute renal failure 48 hr after HgCI2 treat-ment were stained with Giemsa for differential counts.In an additional similar group of six rats with acute renalfailure, blood smears and kidney frozen sections werestained for acid phosphatase using a modification of thediazo method of Barka and Anderson [20]. This wasdone to distinguish between monocytes (acid phos-phatase positive) and other large mononuclear cells. Ineach animal a differential count was performed on 300white blood cells in the kidney vasa recta and 300white blood cells in the blood smear.

Preparation of tissues for light and electron micro-scopy. Paraffin-embedded kidney tissue which hadbeen fixed in I 0% buffered formaldehyde solution wascut at 4 t and stained with hematoxylin-eosin, periodicacid-Schiff and Giemsa stains. Tissue taken for elec-tron microscopy from six rats with acute renal failureand six control rats, was fixed in 3% glutaraldehyde in0.1 M cacodylate buffer, postfixed in 2% osmiumtetroxide and embedded in epoxy resin (Epon) usingmethods previously described for this laboratory [21].Sections were stained with lead citrate and uranyl

Medullary plasma flow in acute renal failure 27

acetate and examined in an electron microscope(AEI-801, AEI Scientific, Falls Church, VA).

Statistical treatment. Student's t test was used toevaluate the significance of differences betweenmeans in the experimental groups [22].

Results

Albumin transit time. Studies of the apparent volumeof distribution of infused 1251-albumin in the renalpapillae of 20 normal rats (40 kidneys) over timeshowed a linear increase which leveled off about 40sec after the beginning of the circulation of radio-active albumin through the kidney (Fig. 1). This sug-gested that the normal papillary transit time foralbumin is 40 sec, a result consistent with the earlierwork of Kramer, Thurau and Deetjen [23, 24] demon-strating an Evans blue transit time of 45 sec in thepapillae of the dog. In the HgC12-treated group,transit time was less certain since volume of distributiondid not increase in a linear fashion (Fig. 2), possiblybecause of increased intrarenal heterogeneity ofblood flow. Using the intersection of the slope of theinitial rise with slope of the plateau at 45 to 60 sec asthe transit time, a transit time of 24 sec was obtainedfor the HgCl2-injected group. (A transit time of 30 secis obtained if the average slope during the first 30 secis used instead of the slope of the initial rise.) Becauseof this short transit time, the results of the 15-secinfusions were used to determine medullary plasmaflow in this group. The maximum volume of distribu-tion was approximately the same in the HgC12-treatedgroup as in the control group.

Medullary plasma flow. Medullary plasma flow,determined as already described using a 30-sec

26

/

—— I

15 30 45Time, see

Fig. 1. Volume of distribution of radioactive albumin in thepapillae of 20 control rats at 15, 30, 45 and 60 sec after theinfused '251-albumin first entered the renal arteries. The apparentcirculation time for albumin is 40 sec. Volume of distributionincreases linearly for the first 30 sec, suggesting that littleradioactive albumin leaves the papilla during this time.

15 30 45 60Time, see

Fig. 2. Volume of distribution of radioactive albumin in thepapillae of 24 rats with HgCl2-induced acute renal failure atdifferent times after the infused '251-albumin first entered therenal arteries. Apparent circulation time is more rapid than incontrol rats and rate of accumulation of labeled albumin isgreater, whether determined at 15 or 30 sec.

infusion time, was found to average 32.22± 1.10(sEM) ml/min/100 g of papilla in 11 control rats (Fig. 3),a value comparable to that reported by Lilienfield et a![16] (25.44 ml/min/lOO g) and by Ganguli and Tobian[17] (38.0 ml/min/100 g). The average absolutedifferences between the values for the two kidneys of asingle animal was 4.70 1.42 (sEM) ml/min/100 g.

- BPJ 50—70 mm Hg

BP30-49 mm Hg

Shock Salineloading

Fig. 3. Medullary plasma flow in experimental and control groupsdetermined by using a 30-sec infusion of radioactive albumin.Lines connect the values for the left and right kidneys of indivi-dual animals. Medullary plasma flow is significantly elevated inthe rats with HgCI2-induced acute renal failure even thoughmedullary plasma flow is probably underestimated using the30-sec infusion period in this group (see Fig. 4).

38

34

30

26

2218

14

10

6

2

0

/// . —

-—// V//7—// ://

.7/

/////

22

18

14

//

64

6056

52

,-., 44

4036

028

Control60

HgCI2(40—48 hr)

28 Solez et a!

Medullary plasma flow was significantly elevated(P<0.00l) in eight saline-loaded rats (53.16± 1.94[SEMI ml/min/100 g), and significantly reduced (P<0.001) in eight rats with severe hemorrhagic hypo-tension and mean blood pressures of 30 to 49 mm Hg(11.16± 1.56 [sEM] mi/mm/bOg). Five rats with lesssevere hypotension (mean blood pressure, 50 to 70mm Hg) had nearly normal medullary plasma flow(29.00 1.00 [SEMI ml/min/l00 g). In seven rats withacute renal failure 40 to 48 hr after mercuric chlorideinjection, meduliary plasma flow, determined using a15-sec infusion time (Fig. 4), was significantly elevated(P< 0.001) expressed in terms of papillary weight(58.35 6.08 [SEMI ml/min/100 g. (Using a 30-sec in-fusion period [Fig. 3], a meduilary plasma flow in tenrats with renal failure was underestimated as 40.501.06 [sEM] ml/min/lOO g, a result still significantlydifferent from the control value [P<0.01J.) Sincepapillary weight was increased approximately 9%in these rats (0.01888 0.0010 [SEMI g vs. 0.01728

100

969288

84

80

76

72

68

64

:: 60E 56

52

o 4844

40

36

32028

2420

16

12

8

40

HgCI2(40—48 hr)

Fig. 4. Medullary plasma flow in rats with HgCI2-induced acuterenal failure, determined by using a 15-sec infusion of radioactivealbumin. Data from control rats using a 30-sec infusion is givenfor comparison.

0.0008 [SEMI g for controls), they had an even more ele-vated total papillary plasma flow (11.02 1.20 l.tl/min/papilla) compared to controls (5.57±0.19 [sEM] 1/min/papilla). Water content of the papillae of controlrats was 76.8± 1.4 (sEM)% as compared to 85.4±0.8(sEM)% in rats with acute renal failure, confirming thatthe weight difference was due to edema.

At the time that medullary plasma flow was deter-mined in the mercuric chloride-treated rats, theanimals were virtually anuric and had serum ureanitrogen concentrations ranging from 276 to 312 mg/100 ml (mean, 298 10 vs. 13.7 2.2 for controls) andserum creatinine concentrations ranging from 6.2 to7.4mg/lOOml (mean, 6.80±0.16 vs. 0.60±0.07 forcontrols). Stools were black, semi-liquid and guaiacpositive. Hematocrit values averaged 36.4 1.2 (sEM)%as compared with control values of 46.7 0.7 (sEM)%.The rats were lethargic and anorexic, but had normalblood pressures.

No evidence of increased glomerular leakiness toalbumin in rats with mercuric chloride-induced acuterenal failure was found. Urine samples collected in the24-hr period before medullary blood flow determina-tion contained 2 to 3 albumin (100 to 300 mg/100 ml) as compared with trace to 1 + albumin in con-trols (15 to 30 mg/100 ml). However, urine output wasso markedly reduced in the rats with acute renalfailure (1.02±0.35 [SEMI ml/24 hr vs. 11.33± 1.87[sEM] ml/24 hr for controls) that total albuminexcretion was not increased. In six rats which receivedan i.v. injection of '251-albumin 48 hr after acute renalfailure was induced with mercuric chloride (5 mg/kg),the urinary excretion rate of radioactivity was so low(0.3 to 0.8 cpm/30 sec) that it could not have accountedfor more than a 0.25 increase in apparent medullaryblood flow. It therefore seems unlikely that radio-active albumin brought to the papilla via the glomeru-lar filtrate constituted an important source of error inthe determination of medullary plasma flow in ratswith nephrotoxic acute renal failure.

Vascular tracer studies. In animals with acute renalfailure 48 hr after the injection of mercuric chloride,India ink was found in vessels of the renal cortex andouter medulla five seconds after injection into theinferior vena cava. At ten sec it could be seen in a fewof the arterial vasa recta of the papilla. By 15 sec mostof the arterial vasa recta of the papilla containedIndia ink and this material could be seen in the venousvasa recta of the outer medulla which contained largenumbers of nucleated cells. At 20 sec, India ink ap-peared for the first time in the venous vasa recta of thepapilla which contained many nucleated cells (Fig. 5).By 30 sec after injection, the vasa recta were diffuselyfilled by india ink.

//

Control

-

.'1' :

.t

Medullary plasma flow in acute renal failure 29

Fig. 5. Vasa recta in the papilla of a HgCl2-treated rat, 20 sec after i.v. injection of India ink. Leukocytes intermingled with India inkcan be seen (arrows), At the upper right is a vas rectum containing leukocytes which has not yet filled with India ink (hematoxylin-eosin, )< 650).

Fig. 6. Numerous leukoc.vtes (arrowheads) in the vasa recta of the outer medulla adjacent to damaged tubules containing numerous densecasts 44 hr after HgCl2 injection (Giemsa, x 370).

In control animals, the same sequence was observedbut apparent circulation time was somewhat slower,with India ink first appearing in the arterial vasa rectaof the papilla at 20 sec and in the venous vasa recta at30 sec.

Two conclusions could be drawn from the vasculartracer studies: 1) There was maintenance of bloodflow through the same vasa recta which containedaccumulations of nucleated cells. 2) The nucleatedcells were abundant in the venous (ascending) vasarecta,3 and relatively scarce in the arterial (descending)vasa recta, a finding consistent with the serial sectionstudies of Baker [8]. The India ink flow rate in themedulla appeared to be more rapid in rats withmercuric chloride-induced acute renal failure than in

Venous and arterial vasa recta were distinguished using thecriteria of Moffat [30], Lever and Kriz [31], and Newstead andMunkacsi [32].

controls, but it is recognized that transit time cannotbe accurately measured with this technique.

White blood count and differential, histochemicalstudies. The average white blood count in rats withacute renal failure 40 to 48 hr after treatment withmercuric chloride was significantly elevated (18,192298 cells/mm3 vs. 9,435 669 cells/mm3 for controls).As demonstrated in Table 1, this leukocytosis was dueprimarily to a seven-fold increase in the number ofneutrophils which constituted l0% of the bloodleukocytes in control animals and 40% in rats withrenal failure. A slight increase in the numbers oflymphocytes and monocytes was also observed. Astudy of Giemsa-stained blood smears and kidneyfrozen sections from six rats with acute renal failuredemonstrated that the leukocyte population in the vasarecta was very similar to that in the peripheral blood forall cell types except monocytes, which made up I .9%of the leukocytes in the peripheral blood and 5.5% of

30 Solez et al

Table 1. Total white blood cell and differential counts in control rats and rats with acute renal failurea

Totalleukocytes

Neutrophils Monocytes Lymphocytes Eosinophils

Acute renal failure 18,192±298/mm3

7,313±373/mm3

(40.2±2.0%)

49161/mm3

(2.7±0.3%)

10,206±320/mm3

(56.1 1.8%)

182±36/mm3

(1.0±0.2%)

Control 9,435±669/mm3 997±66/mm3(10.6 0.7%)

283±85/mm3(3.0 0.9%)

7,982± 179/mm3(84.6 I .9%)

174±47/mm3(1.8 0.5%)

Values indicate 15EM.

Table 2. Differential white blood cell counts on peripheral and vasa recta blood in rats withacute renal failure

Neutrophils Monocytes Lymphocytes Eosinophils

Peripheral 34.0±2.3% 1.9 0.2% 63.8 2.1% 0.2±0.1%Vasa recta 33.4±2.0% 5.5±0.5% 60.4±2.1% 0.3±0.1%

a Different groups of animals were used to derive the data in Tables 1 and 2.b Values indicate I SEM.

the leukocytes in the vasa recta (Table 2). Acidphosphatase stains of blood smears and kidney frozensections from six additional rats with HgCI2-inducedacute renal failure confirmed this difference in the pro-portion of monocytes. In the vasa recta, 6.4± 1.4(sEM)% of the leukocytes were acid phosphatasepositive cells morphologically consistent with mono-cytes; in the peripheral blood only 2.3 0.5 (sEM)% ofthe leukocytes were of this type. In both the vasarecta and the peripheral blood, a small proportion ofthe leukocytes (1.0±0.3 [SEM]% and 1.4±0.6 [sEM]%)were morphologically consistent with monocytes butacid phosphatase negative. Nucleated red blood cellsmade up less than 0.2% of the nucleated cells in theperipheral blood of animals with acute renal failure andwere not observed in the vasa recta.

Histology. In pilot studies, the first unequivocalaccumulation of nucleated cells in the vasa recta, withthe number of leukocytes exceeding the number of redcells in some vessels, was found at 36 hr after HgC12injection. By 40 to 48 hr, the time of which the studiesdescribed in this paper were carried out, the lesion wasmuch more prominent (Fig. 6).

Electron microscopic findings confirmed that theleukocytes in the vasa recta were of three types (Fig.7): neutrophils, lymphocytes and monocytes. Thesegmentation of the neutrophils was variable. Somehad two rather coarse lobes, while most had multiplelobes. Leukocyte accumulation occurred predomin-antly in the ascending vasa recta, which could be

distinguished from the descending vasa recta by theirlarger size and the presence of endothelial fenestrae3(Fig. 8). Leukocyte-to-leukocyte contacts were fre-quently observed. In many cases thin cytoplasmicprocesses were seen extending from a leukocyte to-ward an endothelial cell in a way that suggested thatthe surface coats of the two cells were in contact. Incontrast to this very loose contact between leukocytesand the endothelium of the ascending vasa recta,several of the few leukocytes that were observed in thedescending vasa recta could be seen to be more firmlyattached to the wall of the vessel and surrounded byplatelets in areas of apparent endothelial discontinuity(Fig. 9). Leukocyte accumulations were observed inthe ascending vasa recta of both the outer and theinner medulla. No cells were seen which were morpho-logically consistent with normoblasts or white bloodcell precursors, and no mitoses were observed. Nochanges were observed in endothelial cells which mighthave suggested transformation into leukocytes.

The other light microscopic renal morphologicchanges were similar to those described in mercury-induced acute renal failure by other authors [25—29].There were obvious necroses of the mid and terminalportions of the proximal tubules with numerous denseeosinophilic casts. Many tubules appeared dilated, andthere was extensive interstitial edema. Collections ofleukocytes were frequently found in the venous vasarecta adjacent to necrotic tubules. However, no inter-stitial inflammatory infiltrates were seen.

:'

(4

7?

Medullary plasma flow in acute renal failure 31

Fig. 7. Monocytes (M), lymphocytes (L) and neutrophils (N) in an ascending vas rectum of a rat 46 hr after HgC12 injection. The leuko-cytes appear to be making loose contact with the endothelium and with each other (lead citrate and uranyl acetate, x 7,000).

a.

-.

t -

____

____

A

I'/

I if I

L1

I 34

4,

32 Solez et al

Fig. 8. 4 neutrophil (N) and a monocyte (M) in an ascending vas rectum with characteristic fenestrae (F) 46 hr after HgCI2 injection.Cytoplasmic projections of the leukocytes extend toward the endothelium (arrows) and make contact with other leukocytes (arrowhead).There is no flattening of the neutrophils against the vessel wall and the endothelium appears intact throughout (lead citrate and uranylacetate, x 17,700).

'S

Medullary plasma flow in acute renal failure 33

Fig. 9. A lymphocyte (L) with accompanying platelets (P) attachedto the wall of a descending vas rectum 40 hr after HgC12 injection.Note the characteristic fenestrae of the adjoining ascending vasrectum (F) (lead citrate and uranyl acetate, x 11,700).

Discussion

It is difficult to find a reliable means of determiningrenal medullary plasma flow [13—15, 33, 34] and it istherefore necessary to examine critically the way inwhich this was done in the present study. The methodused is a simple one and several lines of evidence sug-gest that the results generated by this method arevalid.

First, the values obtained for medullary plasma flowin saline-loaded animals, in hemorrhagic shock and incontrol animals using this method are consistent withthe data obtained by other investigators using differ-ent methods [13—17, 23, 24, 35, 36]. Medullary bloodflow seems to be well maintained in moderate hypo-tension (mean blood pressure, 50 to 70 mm Hg) con-firming that medullary blood flow is autoregulated, assuggested by Carriere et al [37] and Ganguli andTobian [17]. In more severe hypotension (mean bloodpressure, 30 to 49 mm Hg) medullary blood flow ismarkedly reduced, in keeping with the findings ofKramer [38] and Aukland and Wolgast [39].

Second, within each of the experimental groups,there is a relatively narrow range of values for medul-lary plasma flow, so that the differences between

groups are easily demonstrated. The standard devia-tion of the mean in our control group (5.16 ml/min/100 g) is considerably less than that of Ganguli andTobian (11.40 ml/min/l00 g) [17] and Lilienfield et al(8.02 ml/min/l00 g) [16]. There is excellent agreementbetween values for the right and left kidney in thesame animal.

Third, since the method involves weighing thepapilla, alterations in tissue volume can be correctedfor and do not represent an uncontrolled variable as isthe case in several other methods [13, 40]. Both totalplasma flow and plasma flow per unit weight may becalculated.

Fourth, small differences among animals in theconcentration of radioactive indicator in arterial bloodare automatically corrected for by using a carotidarterial blood sample from which one may accuratelyestimate the average concentration of radioactiveindicator in blood perfusing the kidney during theinfusion period. Miscalculation of jugular vein torenal artery circulation time could therefore introducelittle error into the final calculation of medullaryplasma flow. The use of an intravenous infusionassures good mixing of radioactive indicator withplasma, which is often inadequate with intraarterialinfusions [41].

The method used does not reflect differences inmedullary plasma flow brought about by differences inthe amount of plasma water filtered by the glomerulus.Rather, it assumes that the concentration of albuminin plasma entering the papilla is the same as that insystemic arterial blood. In the normal animal,filtration fraction is about 0.33 [42]. All but approxi-mately 1 5% of the glomerular filtrate is reabsorbedbefore the papillary portion of the 1oop of Hnle isreached [43] and, thus, contributes to medullary plasmaflow. Assuming that filtration fraction in juxta-medullary nephrons is the same as that for the kidneyas a whole, calculated medullary plasma flow would beapproximately 5% greater than actual medullaryplasma flow in the normal rat. There is evidence that inhemorrhagic hypotension [44] and acute renal failure[45, 46] filtration fraction is markedly reduced, and itis likely that calculated medullary plasma flowapproaches actual medullary plasma flow in thesesituations. This presumably would also be the case insaline-loaded rats, in which filtration fraction in deepnephrons has been reported to be massively decreased[47].

In the normal animal there is extensive leakage ofalbumin into the interstitial space where it equilibrateswith the large extravascular albumin pool [48—50]. Thefenestrated ascending vasa recta are quite permeableto albumin and other small molecular weight proteins.

34 So!ez et a!

Shimamura and Morrison [51] have recently demon-strated extensive leakage of ferritin and horseradishperoxidase out of the descending vasa recta 30 secafter intravenous injection in normal rats. There is noevidence that the conduit vessels bringing blood to themedulla are similarly permeable to albumin. Itdeserves to be emphasized that extensive leakage ofalbumin into the papillary interstitium from theascending vasa recta interferes in no way with thedetermination of medullary plasma flow described inthis paper. It makes no difference whether the radio-active albumin brought to the medulla via the vasarecta is intravascular or extravascular at the timepapillary radioactivity is measured. The fact that themaximum volume of distribution for radioactivealbumin is similar in rats with acute renal failure andin control rats (Figs. 1 and 2) suggests that leakageof albumin out of the vasa recta is not significantlygreater in acute renal failure than it is in the normalanimal.

The method used to determine medullary plasmaflow is based on two major assumptions: 1) that radio-active albumin reaches the papilla only via medullaryblood flow, and 2) that little radioactive albuminleaves the papilla during the infusion period. There isno evidence that substantial amounts of radioactivealbumin are brought to the papilla via the glomerularfiltrate or other routes and, thus, it seems likely thatthe first assumption is valid. The fact that the volumeof distribution of radioactive albumin increaseslinearly for the first 30 sec in control animals (and, thus,that medullary plasma flow determined at 15 sec is thesame as that determined at 30 sec; see Fig. 1) suggeststhat the second assumption is upheld in this group. Ifsubstantial amounts of radioactive albumin left thepapilla during the infusion period, then the plot of thevolume of distribution during the first 30 sec would beexpected to be nonlinear, with a decreasing slope. Inthe group of animals with acute renal failure, it seemsunlikely that substantial amounts of radioactivealbumin left the papilla prior to the time at whichmedullary plasma flow was determined (15 sec; seeFig. 2), but if this were the case, medullary plasmaflow would be underestimated in this group.

The method used in the present study to measuremedullary plasma flow seems to be both more reliableand subject to fewer interpretive difficulties than theother methods which have been described in theliterature [13—17, 23, 24, 33]. It is likely that thistechnique will be found useful in the study of medul-lary hemodynamics in other disease states besidesacute renal failure.

The accumulation of nucleated cells in the vasarecta is a conspicuous feature of acute renal failure.

Baker [8] and Remmele [12] attributed this pheno-menon to decreased medullary blood flow with localhematopoiesis occurring during shock. The presentstudy, and the earlier work of Jaenike and Schnee-berger [52] and Rosen et al [53] demonstrate that thislesion is characteristic of acute renal failure in theabsence of shock. The absence of red blood cell andwhite blood cell precursors, and the similarity of theleukocyte population in the vasa recta to that in theperipheral blood, argue against local production ofleukocytes in the vasa recta and strongly suggest thatthe nucleated cells in the vasa recta are brought therevia the circulating blood. It is unclear whether or notthe accumulation of nucleated cells in the vasa recta isrelated to the polymorphonuclear leukocytosis obser-ved in acute renal failure [54].

The accumulation of nucleated cells in the vasarecta occurs in the presence of elevated medullaryplasma flow, a seemingly paradoxical situation.Findings of India ink vascular tracer studies confirmthat blood flow is maintained in the vessels containinglarge numbers of nucleated cells. The lesion persistsdespite increased blood flow, probably because thenucleated cells are loosely attached to the walls of thevasa recta and to each other. Recently, we have obser-ved that in rats with unilateral mercuric chloride-induced acute renal failure (with one kidney protectedby ureteral ligation [55]), there are prominent collec-tions of nucleated cells in the vasa recta in the affectedkidney but not in the protected kidney. This is truedespite the fact that ureteral obstruction is associatedwith a profound reduction in medullary plasma flow(K. Solez, to be published). Thus, it appears clear thatthat this lesion is not caused by reduced medullaryplasma flow.

The accumulation of leukocytes in renal medullaryvessels is a local phenomenon rather than a systemicone. We have not observed a similar phenomenon inthe vessels of the lungs, liver or intestinal villi ofanimals with nephrotoxic acute renal failure.

The accumulation of leukocytes in the vasa recta inacute renal failure is similar to the margination ofleukocytes which occurs as a part of the acute inflam-matory response. However, margination results in thepreferential accumulation of neutrophils [56], whilein acute renal failure neutrophils and lymphocytesaccumulate in the vasa recta in the same relative pro-portions as in the peripheral blood. Flattening ofleukocytes against the endothelial surface and sub-sequent diapedesis is not observed in the vasa recta inacute renal failure. Since the rate of movement ofmarginated leukocytes is proportional to blood flowrate, and since only slowly moving leukocytes emi-grate out of vessels [57], the elevated medullary blood

Medullary plasma flow in acute renal failure 35

flow in acute renal failure may be one of the factorswhich prevents leukocyte emigration from the vasarecta. There is no evidence that the leukocytes in thevasa recta are immobilized, and it seems likely that theleukocytes are in constant motion rolling along theendothelial surface.

It is possible that the accumulation of leukocytes inthe vasa recta results from the accumulation in therenal medulla of a chemotactic substance in a situationin which the endothelial and hemodynamic changeswhich favor firm sticking and emigration of leukocytesare not present. Preliminary results in this laboratorysuggest that synthesis by the renal medulla of prosta-glandin E (a potent chemotactic substance [58]) isincreased in acute renal failure, a finding also reportedby Torres, Strong, Romero and Wilson (personalcommunication). There is presently insufficient evi-dence to connect this finding with the phenomenon ofleukocyte accumulation in the vasa recta. It has beensuggested recently that transport of prostaglandinsfrom the renal medulla to the cortex via the venousvasa recta may be an important factor in the autoregu-lation of renal blood flow [59].

In 1947, Trueta et al [60] suggested that there was ashunting of blood away from the cortex through themedulla in acute renal failure. The presence or absenceof vascular pathways that would permit an aglomerularshunting of blood has been extensively debated [61—63]. Vascular channels between the afferent and effer-ent arterioles bypassing the glomeruli have been de-scribed by Ljungqvist [64] and Ljungqvist and Wager-mark [65] but their presence has recently beendenied by Spinelli et al [66]. Lameire, Schaper andRingoir [67] concluded on the basis of studies usingradioactive microspheres that there was a 30%aglomerular shunting of blood through the medulla inmercuric chloride-induced acute renal failure in therat. Using similar techniques, Stein et al [68] wereunable to demonstrate such a shunt pathway in thedog.

The increased medullary plasma flow in nephrotoxicacute renal failure demonstrated in the present studycould be the result of increased blood flow to juxta-medullary nephrons or of a medullary shunt. Thesetwo possibilities cannot be differentiated withoutfurther study. Investigations of renal blood flow distri-bution in acute renal failure which have used inert gaswashout techniques have shown an apparent increasein inner cortical blood flow [1, 2, 4, 5], but theinterpretation of these results is hazardous due to lackof definite anatomic localization of the washout curvecomponents in acute renal failure [1, 5] and othertechnical limitations of these techniques [14, 37, 69].

In preliminary studies, we have recently found that

medullary plasma flow is increased in glycerol-inducedacute renal failure just as it is in the mercuric chloridemodel. If the increased medullary plasma flow demon-strated in the present study of nephrotoxic acute renalfailure also occurs in oliguric states associated withless prominent renal structural damage, it may contri-bute to the low urine osmolarity observed in thesestates by washing out the medullary osmotic gradient[24]. Decreased outer cortical perfusion has come tobe regarded as the "final common pathway" [70] ofacute renal failure. Whether the increased medullaryplasma flow demonstrated in this study will be foundto be similarly universal in other types of acute renalfailure remains an unanswered question pendingfurther studies.

Acknowledgments

This work was presented in part at the 57th AnnualMeeting of the Federation of American Societies forExperimental Biology in Atlantic City, New Jersey,April 16, 1973. This study was supported by PublicHealth Service grants HL-07835 and GM-00415.Patricia Bulluck typed the manuscript and MarilynMiller provided technical assistance.

Reprint requests to Dr. Kim Solez, Department of Pathology,The Johns Hopkins Hospital, Baltimore, Maryland 21205, U.S.A.

References

1. HOLLENBERO NK, EPSTEIN M, ROSEN SM, BASCH RR, OKENDE, MERRILL JP: Acute oliguric renal failure in man:Evidence for preferential renal cortical ischemia. Medicine47:455—474, 1968

2. HOLLENBERG NK, ADAMS DF, OKEN DE, ABRAMS HL,MERRILL JP: Acute renal failure due to nephrotoxins: Renalhemodynamics and angiographic studies in man. N Engl JMed 282:1329—1334, 1970

3. AYER G, GRANDCHAMP S, WYLER T, TRUNIGER B: Intra-renal hemodynamics in glycerol-induced myohemoglo-binuric acute renal failure in the rat. Circ Res 29:128—135,1971

4. FLAMENBAUM W, MCNEIL JS, KOTCHEN TA, SALADINO AJ:Experimental acute renal failure induced by uranyl nitrate inthe dog. Circ Res 31:682—698, 1972

5. HOLLENBERG NK, SANDOR T, CONROY M, ADAMS DF,SOLOMON HS, ABRAMS HL, MERRILL JP: Xenon transitthrough the oliguric human kidney: Analysis by maximumlikelihood. Kidney mt 3:177—185, 1973

6, SOLEZ K, KRAMER EC, HEPTINSTALL RH: The pathology ofacute renal failure (ARF): Leukocyte accumulation in thevasa recta. Am JPathol 74:3la, 1974

7. FINCKH ES, JEREMY D, WHYTE HM: Structural renaldamage and its relation to clinical features in acute oliguricrenal failure. Q J Med 31:429-446, 1962

8. BAKER SB DE C: Intravascular haemopoiesis in the renalmedulla in shock. J Pathol Bacteriol 75:421-428, 1958

9. REMMELE W, BLEYL H, LOFFLER H: Das Vorkommen

36 Solez et a!

unreifer Blutzellen im Leukozytenkonzentrat nach Kreis-laufkollaps. Schweiz Med Wochenschr 94:1336—1338, 1964

10. REMMELE W, LEDER LD,GAUL W: Herkunft und Bedeutungder intravaselen Zellansammiungen in der sog. Schockniere.Verh Dtsch Ges Pathol 68:170—184, 1964

11. REMMELE W, BLEYL H, CARSTENSEN H: Untersuchung Uberdie Cytologie des Blutes im Kreislaufshock: I. Die Cytologiedes Zellkonzentrates bei Schockpatienten und bei Kanin-chen mit Entblutungsschock. Z Ges Exp Med 147:333—345,

196812. REMMELE W: Zur pathologischen Anatomie des Kreislauf-

shocks beim Menschen: Ill. Ansammiungen von Blut- undKnochenmarkszellen in den Markgefassen der Niere. K/inWochenschr 46:803—809, 1968

13. LADEFOGED J, MUNCK 0: Distribution of blood flow in thekidney, in Kidney Hormones, edited by FISCHER JW, NewYork, Academic Press mc, 1971, p. 31

14. THURAU K, LEVINE DZ: The renal circulation, in The Kidney,edited by ROUILLER CL, MULLER AF, New York, AcademicPress mc, 1971, vol. 11!, p. 1

15. GRUNFELD JP, RAPHAEL JC, BANKIR L, KLEINKNECHT D,BARGER AC: Intrarenal distribution of blood flow, inAdvances in Nephrology, edited by HAMBURGER J, CROSNIER

J, MAXWELL MH, Chicago, Yearbook Medical Publishers,1971, vol. I, p. 125

16. LILIENFIELD LS, MAGANZINI HC, BAUER MH: Blood flow inthe renal medulla. Circ Res 9:614—617, 1961

17. GANGULI MD, TOBIAN L: Renal medullary plasma flow inhypertension. Fed Proc 31:394, 1972

18. KATZ MA, BLANTZ RC: Geometric error in tissue gammacounting. JApplPhysiol32:533—534, 1972

19. SOLEZ K, JAENIKE JR, RICHTER GW: Polyuria, hypertension,and altered renal function in experimental microembolicrenal disease: Relation to hypertensive diseases in man.Lab Invest27:243—253, 1972

20. BARKA T, ANDERSON PJ: Histochemical methods for acidphosphatase using hexazonium pararosanilin as a coupler.JHistochem Cytochem 10:741—753, 1962

21. HILL GS, WYLLIE RG, MILLER M, HEPTINSTALL RH:Experimental papillary necrosis of the kidney: 11. Electronmicroscopic and histochemical studies. Am JPathol 68:213—228, 1972

22. DixoN WJ, MASSEY FJ: Introduction to Statistical Analysis.New York, McGraw-Hill Book Co mc, 1969

23. KRAMER K, THURAU K, DEETJEN P: Haernodynamik desNierenmarks: I. Capillarie Passagezet, Durchblutung,Gewebshaematokrit, und 02-Verbrauch des Nierenmarks insitu. Arch Ges Physiol 270:251—269, 1960

24, THURAU K: Renal hemodynamics. Am J A'Ied 36:698—7 19,1964

25. EDWARDS JG: The renal tubule (nephron) as affected bymercury. AmJPatholl8:1011—l027, 1942

26. R0DIN AR, CROWSON CN: Mercury nephrotoxicity in the

rat. AmJPathol41:297—313, 1962

27. HARBER MH, JENNINGS RB: Renal response of the rat to

mercury: The effect of sex and sex hormones. Arch Pathol

79:218—222, 196528. TAYLOR NS: Histochemical studies of nephrotoxicity with

sublethal doses of mercury in rats. Am JPathol 46:1—21,1965

29. CUPPAGE FE, TATE A: Repair of the nephron followinginjury with mercuric chloride. Am J Pathol 51:405—429,

1967

30. MOFFAT DB: The fine structure of the blood vessels of therenal medulla with particular reference to the control of

the medullary circulation. J Ultrastruct Res 19: 532—545,1967

31. LEVER AF, KRIz W: Counter-current exchange between thevasa recta and the 1oop of Henle. Lancet 1:1057—1060, 1966

32. NEWSTEAD J, MUNKACSI I: Electron microscopic observa-tions on the juxtamedullary efferent arterioles and arteriolaerectae in kidneys of rats. Z Zellforsch Mikrosk Anat 97:

465—490, 196833. HARSING L, KALLAY K, BARTHA J, DEBRECZENI L: Effect of

hyperosmotic albumin and hypertonic saline infusion onrenal medullary blood flow. Acta Physiol Acad Sci Hung 36:227—237, 1969

34. STEIN JH, BOONJARERN S, WiLsoN CB, FERRIS TF: Altera-tions in intrarenal blood flow distribution: Methods ofmeasurement and relationship to sodium balance. Circ Res21:1—61, 1973

35. STEIN JH, BOONJARERN S, MAUK RC, FERRIS TF: Mech-anism of redistribution of renal cortical blood flow during

hemorrhagic hypotension in the dog. J C/in Invest 52:39-47,

1973

36. WALLIN JD, BLANTZ RC, KATZ MA, ANDREUCCI yE,RECTOR FC JR, SELDIN DW: Effect of saline diuresis onintrarenal blood flow in the rat. Am J Physiol 221 : 1297—1304, 1971

37. CARRIERE S, THORBURN GD, O'MORCHOE CCC, BARGER

CA: Intrarenal distribution of blood flow in dogs during

hemorrhagic hypotension. Circ Res 19:167—179, 196638. KRAMER K: Das akute Nierenversagen im Schock, in

Schock, Pathogenese und Tlmerapie, edited by BOCK KD,Berlin, Springer-Verlag, 1962

39. AUKLAND K, WOLGAST M: Effect of hemorrhage and re-transfusion on intrarenal distribution of blood flow in

dogs. J C/in Invest 47:488—501, 196840. WALLIN JD, RECTOR FC, SELDIN DW: Measurements of

intrarenal plasma flow with antiglomerular basement-

membraneantibody. AmJPhysiol 221:1621—1628, 197141. BRAND PH, COHEN JJ: Effect of renal arterial infusion rate

on distribution of radioisotopes in kidneys. J Appl Physiol33:627—634, 1972

42. DAUGHARTY TM, UEKL IF, NICHOLAS DP, BRENNER BM:Comparative renal effects of isoncotic and colloid freevolume expansion in the rat. Am J Physiol 222:225—235,1972

43. PITTS RF: Physiology of the Kidney and Body Fluids.Chicago, Yearbook Medical Publishers, 1968, p. 124

44. SELKURT EE: Primate kidney function in hemorrhagicshock. AmJPhysiol2l7:915—961, 1969

45. REUBI FC, VORBURGER C, TUCKMAN J: Renal distributionvolumes of indocyanin green, [1Cr] EDTA, and 24Na inman during acute renal failure after shock. J C/in Invest 52:223—325, 1973

46. JAENIKE JR: The renal lesion associated with hemoglo-binemia: A study of the pathogenesis of the excretory

defect in the rat. J Cl/n Invest 46:378—387, 196747. BARRAT U, WALLIN JD, RECTOR FC, SELDIN DW: Influ-

ence of volume expansion on single-nephron filtration rateand plasma flow in the rat. Am J Physiol 224:643—650,1973

48. SLOTKOFF LM, LILIENFIELD LS: Extravascular renal albu-min. AmJPhysiol2l2:400—406, 1967

49. WILDE WS, VORBURGER C: Albumin multiplier in kidneyvasa recta analyzed by microspectrophometry of T-1824.Am JPhysiol 213:1233—1243, 1967

50. VENKATCHALEM MA, KARNOVSKY MS: Extravascularprotein in the kidney. Lab Invest 27:435—444, 1972

Medullary plasma flow in acute renal failure 37

51. SHIMAMURA T, MORRISON AB: Vascular permeability of therenal medullary vessels in the mouse and rat. Am J Pathol71:155—163, 1973

52. JAENEKE JR, SCHNEEBERGER EE: The renal lesion associatedwith hemoglobinemia: II. Its structural characteristics in therat. JExp Med 123:537—545, 1966

53. ROSEN S, MAILLOUX LU, LAWSON NL, TESCHAN PE: Acuterenal failure in the rat: Ii. Light and electron microscopicobservations. Lab Invest 18:444—459, 1968

54. SIMON D, LUKE RG: Polymorphonuclear leukocytes andlymphopenia in acute renal failure and metabolic acidosis inthe rat. C/in Sci Mo! Med 45:397—402, 1973

55. SZABO S: Prevention of mercuric chloride intoxication bytemporary unilateral ureteral ligation. Res Exp Med (Ben)148:321—326, 1972

56. FLOREY HW: Genera/Pathology. Philadelphia,WB SaundersCo, 1970, pp. 22—39

57. ATFIERTON A, BORN GUR: Relationship between the velo-city of rolling granulocytes and that of the blood flow invenules. JPhysiol 233:157—165, 1973

58. KALEY F, WEINER R: Effect of prostaglandin E1 on leuko-cyte migration. Nature [New Biol] 234:114—115, 1971

59. HERBACZYNSKA-CEDRO K, VANE JR: Contribution of intra-renal generation of prostaglandin to autoregulation of renalblood flow in the dog. Circ Res 33:428—436, 1973

60. TRUETA J, BARCLAY AE, DANIEL PM, FRANKLIN KJ,PRICHARD MML: Studies of the Renal Circulation. Oxford,Blackwell, 1947

61. MOFFAT DB, FOURMAN J: The vascular pattern of the ratkidney. JAnat 97:543—553, 1963

62. PLAKKE RK, PFEIFFER EW: Blood vessels of the mammalianrenal medulla. Science 146:1683—1685, 1964

63. KRIZ W: Der architektonische und funk tionelle Aufbau derRattenniere. Z Zellforsch Mikrosk Anat 82:495—535, 1967

64. LJUNGQVIST A: Structure of the arterio-glomerular units indifferent zones of the kidney. Nephron 1:329—337, 1964

65. LJUNGQVIST A, WAGERMARK J: The adrenergic innervationof intrarenal glomerular and extra-glomerular circulatoryroutes. Nephron 7:218—229, 1970

66. SI'INELLI FR, WIRZ H, BRUCHER C, PEI-ILING G: Non-existence of shunts between afferent and efferent arterioles ofjuxtamedullary glomeruli in dog and rat kidneys. Nephron 9:123—128, 1972

67. LAMEIRE N, SCHAPER W, RINGOIR S: Circulation intra-rénale dans l'intoxication mercurielle du rat étudie parmicrosphères marguées. J Urol Nephrol 78:771—775, 1972

68. STEIN JH, BAEHLER RW, Cox JW, FERRIS TF: Renal hemo-dynamic alterations during drug-induced renal vasocon-striction and acute hemorrhagic hypotension, in Conferenceon Acute Renal Failure, edited by FRIEDMAN EA, ELIoHouHE, Washington, DC, Government Printing Office, DHEW(NIH) Pub. No. 74—608, in press

69. MOWAT P, Luu AN, MAXWELL MH: Limitations of 133Xewashout technique in estimation of renal blood flow. Am JPhysiol 223:682—688, 1972

70. HOLLENBERG NK: Vascular factors in the pathogenesis ofacute renal failure in man, in Conference on Acute RenalFailure, edited by FRIEDMAN EA, EuoHou HE, Washington,DC, Government Printing Office, DREW (NIH) Pub. No.74—608, in press