ANNALS O F CLINICAL AND LABORATORY SCIEN CE, Vol. 11, No. 3 Copyright © 1981, Institute for Clinical Science, Inc.

New Concepts in the Pathogenesis of Acute Tubular Necrosis Associated with Sepsis*

ALAN V. RICHMAN, M.D., ELIZABETH G. OKULSKI, M.D.,and JOHN U. BALIS, M.D.

Department o f Pathlogy, University o f South Florida, College o f Medicine,

Tampa, FL 33612 and

James A. Haley Veterans Hospital, Tampa, FL 33612

ABSTRACT

The kidneys of rhesus monkeys, infused either with a single bolus of endotoxin (10 mg per kg) or continuously at the rate of 10 mg per kg per hour for periods of up to 22 hours, have been examined by light and electron microscopy. Monkeys infused continuously with Ringer’s lactate were used as controls. Only minor morphologic changes were observed in those ani­mals receiving a bolus of endotoxin. In the animals continuously infused, sequestration of neutrophils and monocytes was observed in the peritubular capillaries and, to a lesser extent, in the glomeruli. These changes were associated with phagocytosis of endotoxin, occasional fibrinous deposits, and extensive endothelial damage with focal capillary disruption. In the advanced stages, interstitial edema and early necrosis of tubular epithelium were observed. Our data indicate that endothelial damage and associated events relating to the sequestration of phagocytic leukocytes involve the peritubular capillaries primarily and that this process plays a role in the genesis of acute tubular necrosis associated with endotoxemia. In prelimi­nary studies involving the study of kidneys from patients dying with documented Gram negative sepsis and acute renal failure, sequestered nucleated cells have been observed in the peritubular capillaries of the renal cortex and upper medulla. This suggests that similar patterns of endotoxin mediated vascular injury may be occurring in human sepsis.

Introduction

There is considerable uncertainty re­garding the pathogenesis of renal lesions

* Supported in part by MRS of the Veterans Ad­ministration and USAMRDC Contract DAMD 17- 78-C-8078.

in septic shock, although acute renal fail­ure is a frequent complication of medical or surgical sepsis.32 Previously a primate model for prolonged endotoxemia5 was introduced by us. This model closely simulates human septic shock which is usually the result of prolonged sepsis. In

2110091-7370/81/0500-0211 $01.50 © Institute for Clinical Science, Inc.

212 RICHMAN, OKULSKI, AND BALIS

our model, rhesus monkeys are continu- th ree percen t buffered paraform alin,ously infused with endotoxin for periods post-fixed in one percen t osm iumextending up to 22 hours. This model has tetroxide and embedded in Epon 812.been utilized to examine critically the Semi-thin sections were stained with Ter-morphologic changes in the kidneys of ry’s methylene blue stain. Ultra-thin sec-monkeys with prolonged endotoxemia. tions were stained with uranyl acetate andOur studies indicate that the peritubular lead citrate and examined with an elec-capillaries are a major site for endotoxin tron microscope.! Sections from at leastinduced vascular damage and that this in- ̂ th ree rep resen ta tive blocks w ere jury plays a role in the development of examined ultrastructurally.acute tubular injury during shock as­sociated with endotoxemia. Results

Materials and Methods

Detailed methods for the continuous in­fusion of endotoxin into rhesus monkeys have previously been reported.5 In brief, male rhesus monkeys weighing from three to four kilograms were continuously infused through the inferior vena cava with E. coli endotoxin* at a rate of 10 mil­ligrams per kilogram per hour with the use of a perfusion pump. Arterial pressure was measured through a catheter in the aorta. The experimental group consisted of 11 animals which were killed by exsanguina­tion at 6,8,10,13,17 and 22 hours after the onset of endotoxin injection. In addition, three monkeys received intravenously single bolus endotoxin injections (10 mg per kg) and were killed at three hours for study. Controls included three monkeys continuously infused with Ringer’s lac­tate solution for up to 17 hours.

L i g h t M i c r o s c o p y

Tissue was fixed in 10 percent buffered formalin for light microscopy. Paraffin em bedded sections were stained with hematoxylin and eosin, periodic acid- Schiff, and phosphotungstic acid- hematoxylin stains.

E l e c t r o n M i c r o s c o p y

A total of eight to 10 tissue blocks from each animal were immersion fixed in

Monkeys receiving a single bolus en­dotoxin injection developed severe sys­temic hypotension with arterial pressure levels rapidly falling at three hours to 30 to 40 percent of preinfusion values. In mon­keys with continuous endotoxemia, a biphasic arterial pressure response was observed. T here was an in itia l m ild hypotensive phase during the first two hours, probably rep resen ting an a t­tenuated “bolus” effect. A stable period of approximately eight to 10 hours followed with blood pressure maintained at 70 to 90 percent of preinfusion levels. After this stable period, there was a gradual but pro­gressive hypotensive phase. This final phase of progressive hypotension was as­sociated with disseminated intravascular coagulation as evidenced by steadily de­creasing levels of platelets, fibrinogen, Factor VIII, and Factor XII.5

Only minor morphologic and ultra- structural changes were seen in the kid­neys of animals receiving bolus endo­toxin. In many of the small arteries and arterioles, marked vascular spasm was noted. Occasional m arginating poly­morphonuclear (PMN) leukocytes were noted within the glomerular capillaries, averaging less than one per glomerulus. Similar marginating cells were also ob­served in a small number of peritubular capillaries.

Structural changes were noted in the m icrovasculature of all of the con-

* Difco 026 : B6, control 574514. f Phillips 301.

PA TH O G EN ESIS O F ACU TE TU BU LA R NECROSIS IN SEPSIS 213

tin u o u sly in fu sed m onkeys, and th ese changes p rogressed throughout the p e ­riod of observation. N um erous seq u es­te red PM N leukocytes and m ononuclear leukocy tes w ere id en tified w ith in th e peritubular capillaries (figures 1, 2A and 2B) associated w ith prom inent stasis of erythrocytes (figure 1). M em branous par­ticles having the characteristic un it struc­ture of particu late endotoxin4,5 w ere id en ­tified w ith in digestive vacuoles of both the PM N leukocytes and m onocytes (fig­ure 2B). O ccasional tactoids of fibrin and platelets w ere observed adjacent to the m arginating leukocytes, bu t no occlusive throm bi w ere evident. A th ird cell type was observed in the endotoxin exposed anim als, includ ing those anim als receiv­ing a single bolus injection. This cell type

show ed high nuclear/cytoplasm ic ratios w ith nuclei dem onstrating condensation of the nuclear chrom atin at the nuclear m em brane and prom inent nucleoli. T he cytoplasm contained few m itochondria, abundant polyribosom al aggregates, and occasional p a ra lle l lam e llae o f rough endoplasm ic reticulum . T hese cells w ere in terp re ted as transform ed lym phocytes.

C onsistent and progressive endo thelia l changes w ere m anifest in the peritu b u lar capillaries of all m onkeys with prolonged endotoxem ia. T hese included focal endo­thelia l sw elling a lternating w ith en d o ­th e lia l cy to p lasm ic a tte n u a tio n , focal en d o th e lia l gaps (figures 2A and 2B), sloughing of in tact endothelial cells into the m icrocirculation, and fragm entation of endothelial cytoplasm w ith focal disrup-

F i g u r e 1. Kidney of rhesus monkey after 17 hours of continuous endotoxin infusion. The peritubular capillaries are engorged with red blood cells. There are many sequestered leukocytes in the peritubular capillaries. Epon embedded, Terry’s methylene blue x 700.

2 1 4 RICHM AN, OKULSKI, AND BALIS

F i g u r e 2A a n d F i g u r e 2B (h i g h e r m a g n i f i c a t i o n o f F i g u r e 2A). Kidney after 10 hours of continuous infusion. Within this peritubular capillary is a monocyte and neutrophil. Digestive vacuoles (V) containing particulate endotoxin are present in the monocyte. There are focal endothelial defects (arrows) with moderate interstitial edema. 2A = x 3610, 2B = X 13050.

PATHOGENESIS O F ACUTE TUBULAR NECROSIS IN SEPSIS 215

tion of the peritubular capillaries in the later stages of prolonged endotoxemia. In­terstitial edema accompanied the endo­thelial disruption (figures 2A and 3) and was associated with focal degenerative and necrobiotic alterations of the tubular epithelial cells (figure 3).

Marginating leukocytes with ingested endotoxin were also identified in the glom erular capillaries. E ndo thelia l changes w ere dem onstrab le in the glomeruli, but they were less intense than those observed in the peritubular capil­laries. In none of the experimental ani­mals was there evidence of glomerular endothelia l disruption, and capillary basal laminas were uniformly intact.

Within the micro vasculature of the con­trol monkeys continuously infused with Ringer’s lactate, rare marginating leuko­cytes were noted. Endothelial alterations, however, were not observed, and the structure of the glomeruli, interstitium, and tubules was unremarkable.

These observations prompted an in­quiry as to whether or not similar altera­tions in the microcirculation might be oc­curring in human sepsis. In preliminary studies, the kidneys were examined of pa­tients dying with documented Gram nega­tive bacteremia and acute renal failure. Excluded from study were those kidneys with extensive autolysis or acute pyelo­nephritis. Nucleated cells, predominantly mononuclear, were observed in the capil­laries of the renal cortex and upper medulla in the kidneys of seven of 13 pa­tients (figure 4).

DiscussionEarly studies by Thomas and Good31

into the pathogenesis of the endotoxin in­duced generalized Shwartzm an p h e­nomenon in rabbits suggested that the glom erulus rep resen ted the prim ary target of injury and that tubular necrosis was secondary to occlusion of glomerular capillaries by thrombotic lesions. McKay et a l18,20 have em phasized the role of intravascular coagulation in the patho­genesis of the generalized Shwartzman

phenom enon in rabbits. As m arked species differences exist in the biological response to endotoxemia, it is difficult to extrapolate data to primates from that ob­tained from lower animals. There have been relatively few studies concerned with renal alterations associated with en­dotoxin shock in primates. Cavanagh et al6 studied baboons who received a single bolus injection of endotoxin. A precipi­tous drop in renal blood flow was observed occurring well before any demonstrable changes in either aortic pressure or car­diac output and suggested that in the ba­boon the kidney represented a major target organ in endotoxin shock. No mor­phologic or ultrastructural renal observa­tions were discussed by these authors. Coalson et al8 observed diffuse glomeru­lar thrombi in baboons infused over a three hour period with liveE. coli. In con­trast, these same authors did not see glomerular thrombi in baboons infused with endotoxin.9 Proximal tubular necro­sis, however, was seen in baboons receiv­ing either live E. coli or endotoxin.14

Two previous studies of rhesus mon­keys with endotoxin shock have em­phasized the lack of glomerular thrombi. McKay et al19 studied five rhesus monkeys who received a single bolus of endotoxin (10 mg per kg). Although no occlusive glomerular thrombi were observed, small fibrin tactoids and platelets were identifi­able within the glomerular capillaries. In a recent study by Coalson et a l10 rhesus monkeys were infused with live E. coli over a 30 minute period and observed for periods up to 27 hours. Although these workers docum ented progressive in ­creases in blood urea n itrogen and creatinine concentration, neither glo­merular thrombi nor glomerular endo­thelial changes were observed. Mild dila­tation of the proximal convoluted tubule was noted along with increased eosino- philia of the tubular epithelial cells.

The consistent association of endo­thelial injury with sequestered phago­cytic leukocytes suggests a pathogenic role for these cells. In the lung, the rapid

RICHM AN, OKULSKI, AND BALIS

F ig u r e 3. Kidney after 22 hours of continuous endotoxin infusion. The tubules are separated by an edematous interstitium. There is focal necrosis of tubular epithelium (arrows). Epon embedded, Terry’s m ethylene blue, x 370.

sequestra tion , deg ranu lation , and frag­m e n ta tio n o f le u k o c y te s w ith in th e pulm onary vascular b ed have b een as­sociated w ith vascular injury in endotoxin shock.4,5' 7 A variety of biologic activities has b ee n a ttr ib u te d to th e re le a se of lysosomal factors from leukocytes, includ­ing increases in vascular perm eability , disruption o f vascular structures, prom o­tion of coagulation, and triggering of kinin form ation.16,21,33,34 T he vascular injury ob­served by us may w ell be re la ted to one or m ore o f th e b io lo g ic a c tiv it ie s o f lysosomes.

Kinin re lease following the sequestra­tion of PM N leukocytes could play an in­itiating role in the developm ent of endo­thelial injury. N ies e t a l23 show ed that endotoxin-leukocyte in teraction can gen­erate kinins before the dem onstrab le re ­lease of lysosomal enzym es. Sacks e t a l27

dem onstrated tha t endothelial dam age in c u ltu re d h u m a n u m b ilic a l cord ce lls could be provoked by exposure to both granulocytes and activated com plem ent. T h eir data suggested that this dam age was m ediated by toxic oxygen radicals and not lysosom al pro teases. T hus, the data of N ies et al and Sacks e t al suggest that en ­dothelial injury may occur through the participation of leukocytes bu t in d ep en ­d e n tly o f th e le u k o c y tic ly so so m al proteases.

A lte rn a tiv e ly , th e re is ex p e rim en ta l ev id e n c e th a t en d o to x in -in d u ce d e n ­dothelial injury may not require leuko­cytes. G aynor e t a l11,12 exam ined the endo thelium o f rabbits in jected w ith a sing le su b le th a l dose o f endotoxin . In th ese rabbits, m arked increases in en ­d o th e lia l ce ll d e so x y rib o n u c le ic ac id (DNA) synthesis w ere observed reflecting

PA TH O G EN ESIS O F ACU TE TUBULAR NECROSIS IN SEPSIS 217

F ig u r e 4. Kidney from a 48 year old male dying from Gram negative sepsis and acute renal failure following massive small bowel infarction. This capillary from the upper m edulla is filled with leukocytes, predominantly monocytes. Hematoxylin and eosin, x 800.

en d o th e lia l reg en e ra tio n fo llow ing in ­jury. D esquam ation of endothelial cells into the circulation was also observed. N either the induction o f m arked neu tro ­pen ia through nitrogen m ustard infusion, nor prior anticoagulation w ith heparin p reven ted this endotoxin induced endo­thelial injury.11,13 T he data of Gaynor e t al suggest a d irec t toxic affect of endotoxin upon endo the lia l cells, an effect in d e ­p en d en t of neu troph ils and intravascu- lar coagulation.

O f in terest is the observation o f m ono­nuclear cells w ith the structural appear­ance of transform ed lym phocytes. These cells w ere observed early in the course of continuous endotoxin infusion, and their consisten t p resen ce im plies a possib le role in endothelial injury. Endotoxin is a po ten t B-cell activator28 and also interacts w ith T-cells.2,17'22 T reatm en t of m acro­

phages w ith endotoxin can re n d e r the m acrophages cytotoxic to tum or ce lls .1 T his activation of m acrophages does not have im m unologic specificity and does n o t re q u ire th e p a r tic ip a tio n o f B or T -lym phocytes.1-26 T hus, in our experi­m ental m odel m acrophages could theo re t­ically be ren d ered cytotoxic to endo thelia l cells by a com bination o f direct effects and through ind irec t effects via lym phokines e laborated by activated lymphocytes.

W hile only prelim inary, the observa­tion o f se q u e s te re d leukocy tes in th e m icrocirculation of patients w ith acute ren a l fa ilu re an d G ram n eg a tiv e b ac ­terem ia suggests that sim ilar patterns of endotoxin m ediated vascular injury may be occurring in hum an sepsis. Previous authors have described nucleated cells in the vasa recta in kidneys of patients with acute renal failure of unspecified e tio l­

218 RICHMAN, OKULSKI, AND BALIS

ogy.3,29 Some have a ttr ib u ted this phenom enon to local in travascular hematopoiesis secondary to decreased medullary blood flow during shock.3,24 Others, however, have observed this le­sion in acute renal failure in the absence of shock.15,25 Similarly, in our model, sequestration of leukocytes and endo­thelial injury occurred well before the onset of clinical shock. In a recent study involving rats with mercuric chloride in­duced acute renal failure, Solez et al30 ob­served nucleated cells in the vasa recta and demonstrated that these were leuko­cytes brought there via the circulating blood. The selective accum ulation of leukocytes in the microcirculation in a va­riety of pathologic conditions leading to acute renal failure suggests that this phenomenon may represent a common pathway for renal injury.

Conclusion

Our studies indicate that in primates with prolonged endotoxemia, endothelial damage and associated events relating to the sequestration of phagocytic leuko­cytes involve the peritubular capillaries primarily. This “ inflammatory” process plays a role in the development of acute tubu lar necrosis ; associated w ith endotoxemia.

Acknowledgm ents

The authors are grateful to Mr. George Kasnic and Mr. James Patterson for technical assistance.

References1. A le x a n d e r , P. and E v a n s , R.: Endotoxin and

doub le stranded RNA ren d er m acrophages cytotoxic. Nature New Biol. 232:76-78, 1971.

2. A l l i s o n , A. C. and D a v ie s , A. J. S.: Require­m ent of thym us-dependent lym phocytes for potentiation by adjuvants of antibody forma­tion. Nature 233:330-332, 1971.

3. B a k e r , S. B. DE C.: Intravascular haemopoiesis in the renal m edulla in shock. J. Path. Bacteriol. 75:421-428, 1958.

4. B a lis , J. U., G e r r e r , L. I., R a p p a p o r t , E . S., and N e v i l l e , W. E .: M echanism s o f b lood vas­

cular reactions of the prim ate lung to acute endotoxem ia. Exp. Mol. Path. 21:123-137, 1974.

5. Ba l is , J. U., Ra p p a p o r t , E. S., G e r r e r , L., F a r e e d , J., B u d d in g h , F ., and M e s s m o r e , H. L.: A primate m odel for prolonged endotoxin shock. Blood-vascular reactions and effects of glucocorticoid treatm ent. Lab. Invest. 38:511- 523, 1978.

6. C a v a n a g h , D. and R a o , P. S.: Endotoxin shock in the subhum an primate. I. Hemodynamic and biochem ical changes. Arch. Surg. 99:107-112, 1969.

7. C o a l so n , J. J., H in s h a w , L. B., and G u e n t e r ,C. A.: The pulmonary ultrastructure in septic shock. Exp. Molec. Path. 12:84-103, 1970.

8. Co a l so n , J. J., H in s h a w , L. B., G u e n t e r , C.A., Be r r e l l , E. L., and G r e e n f ie l d , L. J.: Pathophysiologic responses of the subhuman prim ate in experim ental septic shock. Lab. In ­vest. 32:561-569, 1975.

9. Co a l so n , J. J., Ben ja m in , B., Ar c h e r , L. T., Be l l e r , B., Gil l ia m , C. L., T a y l o r , F. B., and H in s h a w , L. B.: Prolonged shock in the baboon subjected to infusion ofE. coli endotoxin. Circ. Shock 5:423-437, 1978.

10. Co a l so n , J. J., Ar c h e r , L. T., H a l l , N. K., Ke r n , J. D., Ben ja m in , B. A., Be l l e r -T o d d ,B., and H in s h a w , L. B.: Prolonged shock in the monkey following live E. coli organism infu­sion. Circ. Shock 6:343-355, 1979.

11. Ga y n o r , E., Bo u v ie r , C., and Sp a e t , T. H.: Vascular lesions: Possible pathogenetic basis of the generalized Shwartzman reaction. Science 170:986-988, 1970.

12. G a y n o r , E.: Increased mitotic activity in rabbit en d o the lium after endotoxin . Lab. Invest. 24:318-320, 1971.

13. G a y n o r , E .: T he ro le o f g ranulocytes in en d o to x in -in d u ced vascu la r in jury . B lood 41 :797-808, 1973.

14. H in s h a w , L. B., Ben ja m in , B., H o l m e s , D. D., Be l l e r , B., Ar c h e r , L. T., C o a l so n , J. J., and Wh it s e t t , T.: Responses of the baboon to live E scherichia coli organism s and endotoxin. Surg. Gynecol. Obstet. 145:1-11, 1977.

15. Ja e n ik e , J. R. and Sc h n e e b e r g e r , E. E.: The renal lesion associated w ith hemoglobinemia.II. Its structural characteristics in the rat. J. Exp. Med. 223:537-545, 1966.

16. J a n o f f , A.: M ediators o f tissue damage in hum an polym orphonuclear neutrophils. Semi­nars Hemat. 3:96-130, 1970.

17. Ko e n ig , S., H o f f m a n , M. K., and T h o m a s , L.: Induction of phenotypic lym phocyte differ­entiation in LPS unresponsive mice by an LPS- induced serum-factor and by lipid A-associated protein. J. Immunol. 178:1910-1911, 1977.

18. M c K a y , D. G., M A RGARETTEN, W ., and CSAVOSSY, I.: An electron microscope study of the effects of bacterial endotoxin on the blood- vascular system. Lab. Invest. 15:1815-1829,1966.

19. M c Ka y , D. G ., M a r g a r e t t e n , W., and CSAVOSSY, I.: An electron microscopic study of

PATHOGENESIS OF ACUTE TUBULAR NECROSIS IN SEPSIS 219

endotoxin shock in rhesus m onkeys. Surg. Gynecol. Obstet. 225:825-832, 1967.

20. M c Ka y , D. G.: V essel w all and throm bo- genesis-endotoxin. Thromb. Diath. Haemorrh. 29:11-26, 1973.

21. Mo v a t , H. Z., Ud aka , K., and Ta k e u c h i, Y.: Polym orphonuclear leukocyte lysosomes and vascular injury. Throm b. D iath. Haem orrh. (Suppl.) 40:211-224, 1970.

22. N e w b u r g e r , P. E., H a m aoka , T., and Ka t z ,D. H.: Potentiation of helper T cell function in IgE antibody responses by bacterial lipopoly- saccharide (LPS). J. Im munol. 113:824-829,1974.

23. N ie s , A. S. and Me l m o n , K. L.: Variation in endotoxin-induced kinin production and effect betw een the rabbit and rhesus monkey. Amer. J. Physiol. 225:230-233, 1973.

24. Re m m e l e , W.: Z ur pathologischen Anatomie des Kreislaufshocks beim Menschen: III. An­sam m lungen von Blut-und Knochenmarkszel­len in den M arkgefassen d e r N iere. Klin. W ochenschr. 46:803-809, 1968.

25. Ro s e n , S., M a il l o u x , L. U., La w s o n , N. L., and T e s c h a n , P. E .: A cute renal fa ilu re in th e rat. II . L ig h t and e lectron m icroscopic observa­tions. Lab. Invest. 28:444-459, 1968.

26. R u c o , L. P., M e l z e r , M. S., a n d R o s e n - STREICH, D. L.: M acrophage ac tiv a tio n for tu m o r cy to to x ic ity : co n tro l o f m ac ro p h ag e tum oric ida l capacity by th e LPS gene. J. Im ­m unol. 121 :543-548, 1978.

27. Sa c k s , T., Mo l d o w , C. F., C r a d d o c k , P. R., Bo w e r s , T. K., and Ja c o b , H. S.: E n d o th e lia l dam age provoked by toxic oxygen radicals re ­lea se d from co m p lem en t tr ig g e red g ran u lo ­cytes. Prog. C lin. Biol. Res. 2 1 :719-724, 1978.

28. Sk id m o r e , B. J., C h il l e r , J. M., M o r r iso n , D. C ., and We ig l e , W. O.: Immunologic proper­ties of bacterial lipopolysaccharide (LPS): cor­relation betw een the mitogenic, adjuvant and im munogenic activities. J. Immunol. 114:770- 775, 1975.

29. So l e z , K., Kr a m e r , E. C., and H e p t in s t a l l , R. H.: The pathology of acute renal failure (ARF): Leukocyte accum ulation in the vasa recta. Amer. J. Path. 74:31a, 1974.

30. So l e z , K., Kr a m e r , E. C., F o x , J. A., and H e p ­t in s t a l l , R. H.: M edullary p lasm a flow and in travascu lar leukocyte accum ula tion in acute ren al failure. K idney In tern a t. 6:24-37, 1974.

31. T h o m a s , L. and G ood, R. A.: Studies on the generalized Shwartzman reaction I. General observations concerning the phenom enon. J. Exp. Med. 96:605-625, 1952.

32. Wa r d l e , E. N.: Endotoxin and acute renal fail­ure. N ephron 24:321, 1975.

33. W e i SSM A N, G.: Lysosomes. New Engl. J. Med. 237:1084-1090, 1143-1149, 1965.

34. W eissm a n , G., Z u r ie r , R. B., and H o f f s t e in , S.: Leukocytes as secretory organs of inflamma­tion. Agents Actions 3 :370-379, 1973.