Biology of acute renal failure: Therapeutic implications

Kidney International, Vol. 52 (1997), pp. 1102-1115

NEPHROLOGY FORUM

Biology of acute renal failure: Therapeutic implications Principal discussant: WILFRED LIEBERTHAL Boston Medical Center, Boston, Massachusetts, USA

furosemide (320 mg intravenously) again produced no response in the urine output. No more diuretics were administered.

Editors

By the sixth day after surgery, the patient was drowsy and less alert than

usual. Physical examination disclosed asterixis. His respirations were 24/mm; his jugular veins were distended to 6 cm; and chest examination revealed bibasilar crackles. His cardiac examination was normal. Periph-

JORDAN J. COHEN JOHN T. HARRINGTON

NIcoLAos E. MADIAS

eral edema was present. His weight had increased 7 kg since admission. A chest radiograph showed evidence of mild congestive heart failure. His blood urea nitrogen had risen to 140 mg/dl and his serum creatinine to 6.5 mg/dl. The serum electrolytes were: sodium, 130 mEq/liter; potassium, 6.2 mEq/liter; chloride, 95 mEq/liter; and bicarbonate, 15 mEq/liter. His urine output during the previous five days had been approximately 300 cc/day. A subclavian catheter was placed and hemodialysis begun. During the next 14 days, he received seven hemodialysis treatments. His

Managing Editor CHERYL J. ZUSMAN

Tufts University School of Medicine

blood urea nitrogen and serum creatinine fell 15 mg/dl and 0.5 mg/dl respectively during a 24-hour period between the last two hemodialysis treatments. Hemodialysis was discontinued. Over the next few days, the blood urea nitrogen fell to 25 mg/dl and the serum creatinine to 1.0 mg/dl, and the patient was discharged from the hospital. Patient 2. A 36-year-old man with a 20-year history of insulin-dependent

CASE PRESENTATIONS Patient I. A 46-year-old black man was brought to the Boston Medical Center emergency room after being stabbed multiple times in the abdomen. Physical examination on arrival revealed a systolic blood pressure of 60mm Hg and a pulse rate of 160 beats/mm. Three deep stab wounds were present in the abdomen. There was diffuse abdominal tenderness and guarding. An upright abdominal radiograph revealed free air under the diaphragm. A Foley catheter was passed and 100 ml of urine was collected. The urinalysis was normal. Blood urea nitrogen on admission was 15 mg/dl and serum creatinine was 0.8 mgldl. The patient received two liters of normal saline intravenously over 30 minutes and then five units of packed red cells over the next three hours. His blood pressure steadily increased and was 110/60 mm Hg five hours after admission to the emergency room. He passed no urine for the first few hours after admission but then began to pass 10 to 15 ml of urine per hour. When he was hemodynamically stable, a laparotomy was performed.

Three lacerations of the small bowel were repaired and the spleen removed because of a severe splenic laceration. He received additional saline and three more units of blood during surgery. His blood pressure during surgery was stable at about 115/70 mm Hg. He continued to pass about 10 to 15 ml of urine during surgery. The first day after surgery, the patient had a blood pressure of 120/80 mm Hg and a pulse rate of 70 heats/mm. His blood urea nitrogen was 40 mg/dl and his serum creatinine was 1.6 mg/dl. The liver enzymes and serum amylase were normal. His urine output over the previous 12 hours had been 200 ml. Urinalysis revealed no blood or protein. 1he sediment

diabetes mellitus presented to Boston Medical Center with worsening angina. The patient had a history of coronary artery disease presenting

initially as an uncomplicated myocardial infarction three years before this admission. He also had severe peripheral vascular disease necessitating a right aorto-femoral bypass procedure four years previously. In addition, he had a history of severe diabetic retinopathy and chronic renal dysfunction (baseline serum creatinine of 3.0 mg/dl) and nephrotic-range proteinuria presumed to be secondary to diabetic glomerulosclerosis. After the myocardial infarction, the patient had been asymptomatic and

without chest pain. However, one week before admission he began to complain of severe central chest pressure whenever he exerted himself, and which became progressively more severe, frequent, and prolonged during the course of the week before admission. Physical examination revealed a blood pressure of 160/93 mm Hg, a pulse rate of 78 heats/mm (regular), and 18 respirations per minute. No jugular venous distention or chest or cardiac abnormalities were evident. Abdominal examination was

normal. The patient had bruits over the left carotid and left femoral arteries. No pedal pulses were palpable bilaterally. Laboratory data on admission included a blood urea nitrogen of 45 mg/dl and serum creatinine of 3.3 mgldl. Total blood count and serum electrolytes were normal. The electrocardiogram disclosed evidence of an old inferior myocardial infarction but no other changes or abnormalities. Urinalysis revealed 4+ albumin and 2+ glucose, with 2 to 3 red cells/highpower field and many oval fat bodies in the urinary sediment. A coronary angiogram was scheduled. The patient received one liter of 0.45% saline prior to the procedure. The intravenous infusion of 0.45%

saline was continued during and after the angiogram at a rate of 120 contained many renal tubular epithelial cells and darkly pigmented nil/hour. During the angiogram, the patient received 155 ml of a nonionic ('muddy brown") casts. He was given furosemide (160 mg intravenously) contrast medium. A critical stenosis of the left anterior descending with no response in his urine output. Three hours later, another dose of

coronary artery was identified; a successful angioplasty was performed without complications. The patient received no diuretics before or after the angiogram.

The Nephrology Forum is funded in part by grants from Hoechst Marion Roussel, Incorporated; Amgen, tncorporated; Merck & Co., Incorporated; Dialysis Clinic, Incorporated; and R & D Laboratories.

In the first 24 hours after angiography, he excreted only 200 ml of urine, and the intravenous infusion was discontinued when the patient developed hihasilar crackles. At that point, he was given 320 mg of furosenude. A diuresis of 150 ml ensued over the next three hours. The patient's urine

© 1997 by the International Society of Nephrology

third day after angiography, his serum creatinine rose to 4.6 mg/dl and 5.8

output then fell again to approximately 15 mI/hour. On the second and

1102

Nephrology Forum: Acute renal failure

mg/dl, respectively. The urine sediment, which was examined 071 multiple

occasions, showed oval fat bodies and 5 to 10 renal tubular epithelial cells/high-power field but no casts. On the third day post angiography, the urine output increased to 750 ml. On the fourth day following angiography, the serum creatinine fell to 4.6

mg/dl. Over the next three days, the urine output increased steadily to approximately 2 liters/day, and the serum creatinine fell to baseline (3.5 mg/dl) before the patient was discharged.

DISCUSSION

1103

was based on work in the isolated kidney perfused by a red-cellfree perfusate which, without being subjected to any injurious insult, rapidly develops spontaneous necrosis of cells of the mTAL with relative sparing of all other nephron segments. This same morphologic pattern of injury has been consistently reported by many other groups using the same red-cell-free, isolated perfused kidney model [9, 10]. But the relevance of this model to the in-vivo

situation was subsequently questioned by studies in isolated kidneys perfused in the presence of red cells. These investigations

DR. WILFRED LIEBERTIIA!. (Renal Section, Boston Medical Center, and Associate Professor of Medicine, Boston University School of

have clearly demonstrated that the extensive necrosis of the

Medicine, Boston, Massachusetts, USA): These two cases are examples of acute renal failure resulting from an ischemic insult (Patient 1) and radiocontrast-medium-induced renal injury (Patient 2). This Nephrology Forum will provide an overview of the recent advances in our understanding of the pathophysiology of acute renal failure (ARF) that results from acute ischemic or nephrotoxic insults to the kidney. I will demonstrate how basic research in this area has led to the potential for the development of an array of novel forms of treatment for humans with ARF.

kidney is related to, and depends on, the absence of red cells in the perfusate. When uninjured isolated kidneys are perfused in the presence of red cells, all nephron segments, including the mTAL segment, remain morphologically intact [9, 101. Furthermore, when the isolated erythrocyte-perfused kidney is subjected

Pathophysiology

The term conventionally used in clinical practice to refer to ARF that results from ischemic or toxic injury to the kidney is acute tubular necrosis (ATN) [1]. This term accurately defines the population of cells that are the major target of injury in this form of ARF. A number of studies, predominantly using micropuncture

techniques and published more than 15 years ago, showed that renal tubular epithelial cell injury compromises glomerular filtration rate (GFR) by obstructing tubules and allowing "backleak" of glomerular filtrate [1—4]. The term "backleak" refers to the unregulated passage of solute and water from the tubular lumen into the interstitium and ultimately back into renal venous capillaries and renal veins. Subsequent studies by Myers and colleagues have provided evidence that both intratuhular obstruction as well as tubular backleak are important factors leading to the

mTAL segment that occurs spontaneously in the isolated perfused

to hypoxic [10] or ischemic injury [11—13], the pattern of injury is

indistinguishable from that following ischemia in experimental models in vivo, with the most severe injury affecting the straight segment of the proximal tubule. Recent studies demonstrating that the straight segment of the proximal tubule is the site of maximal mitotic activity after injury also suggest that lethal injury is most prominent in this segment [14]. Thus, available evidence indicates that the cells of the S3 segment of the proximal tubule sustain the most severe injury following an ischemic or a toxic renal insult. Hemodynamic abnormalities: The role of persistent medullary hypoxia in ATN A reduction in total renal blood flow to 40% to 50% of normal

has been consistently reported both in experimental animal models of ARF as well as during the maintenance phase of ATN in humans [11. For many years, the importance of these hemody-

namic changes in mediating the maintenance phase remained uncertain. Many investigators concluded that the hemodynamic alterations were relatively unimportant, because the observed reduction in GFR associated with ischemic renal failure in reduction in renal blood flow is insufficient to account for the humans [5]. In most cases of ATN, obstruction is due to the profound reduction in GFR. Furthermore, augmentation of total formation of intratubular casts of proteinaceous material that renal flow by volume expansion or administration of vasodilators such as dopamine and prostaglandin analogues does not substanoften contain renal tubular epithelial cells. Numerous experimental studies of ischemic and toxic ARF in tially improve GFR [1]. However, it has become clear that vivo have consistently demonstrated that the straight portion of hemodynamic abnormalities do play an important role in ATN by the proximal tubule (the S3 segment) is the most susceptible to causing persistent regional disturbances in renal blood flow and ischemic and toxic injury [6, 71. The cells in this part of the oxygen supply that predominantly affect the outer medulla of the nephron lose the normal tubular brush border and undergo kidney [15—17]. The medulla of the normally perftised kidney is constantly in a S2 segments) are less severely damaged; they mainly lose brush markedly hypoxic state (P02 of 10—20 mm Hg), whereas the border and show less necrosis than do cells in the straight S3 oxygen tension of the cortex is about 50 mm Hg [15, 18]. The segment [61. The degree to which cells of the medullary thick hypoxic state of the normal medulla is due to countercurrent ascending limb (mTAL) segment sustain injury and necrosis diffusion of oxygen from descending (arterial) to ascending (vevaries in different studies and experimental models, but in general nous) vasa rectae and is therefore a necessary consequence of the is far less severe than that suffered by the straight segment of the concentrating mechanism of the kidney [15, 18]. Oxygen utilizaproximal tubule [6, 7]. Tubular cells in the inner medulla and tion that is needed for active solute transport by the straight those lining the collecting duct appear to sustain the least injury proximal tubule segments and mTAL segments also contributes to the low oxygen tensions in this region of the kidney [15]. Followi,6]. Despite the results of these studies from the late '70s and early ing an ischemic or toxic insult to the kidney, oxygen tension in the '80s, controversy has arisen over the last decade regarding the medulla falls even further because of a persistent reduction in issue of which segments of the nephron are most severely injured blood flow [15, 16]. Thus, the hemodynamic abnormalities assoin ARF. In the early 1980s, Epstein's group hypothesized that the ciated with ARF likely contribute to renal dysfunction by mainmTAL segment is the major site of injury in ATN [8]. This theory taining the tubular segments within the outer medulla in a extensive cell necrosis. Proximal tubular cells in the cortex (SI and

1104

Nephrology Forum: Acute renal failure

ET released into plasma -___

NO-induced vasodilation

ET

Reduced NO-mediated1 Marked reduction in antagonism of ET NO preduction action

VASOCONSTRICTION

• ETA receptor

0 ET8 receptor

persistent state of severe hypoxia, thereby adding to, as well as prolonging, the initial ischemic or toxic insult. This persistent abnormality in medullary perfusion appears to be due to at least two independent factors: intrarenal vasoconstriction, and the physical congestion of the medullary vasculature by red cells, platelets, and leukocytes. Let us look at the mechanisms underlying each of these factors separately. Intrarenal vasoconstriction. Although renal tubular epithelial cells are clearly the principal target of injury in ATN, injury to the vascular endothelium contributes to the intrarenal vasoconstriction associated with acute renal injury. Endothelial dysfunction

leads to an imbalance in endothelin (ET) and endotheliumderived nitric oxide (EDNO) production in ATN.

The endothelins are a family of peptides of 21 amino acids produced by the endothelium [191. The three isoforms of ET (ETI, ET2, and ET3) all arc produced in renal tissue [19]. This dicusssion will focus on the ETI isoform, the only one clearly known to play a role in ARF [19]. Not only is it the most potent vasoconstrictor yet described, ETI also has greater vasoconstrictor

Fig. 1. The normal balance of endothelial production of endothelin and nitric oxide (EDNO) within the kidney (panel A) is reversed following acute renal injury (panel B).

activity in the renal vasculature than in any other vascular bed [19]. Also, ET1 has unusually long-lasting vasoconstrictor activity [19].

Understanding the mechanisms involved in the production and action of ET provides many opportunities for therapeutic blockade of this system. The mature ET peptide is produced from its precursor, prepro El', by a two-step process. Prcpro El' is first converted to "big ET." Big ET, which has little vasoconstrictor activity, is then converted to the mature (21 amino acid) ET by a

specific endopeptidase called endothelin-converting enzyme (ECE) [19]. Two receptors for endothelin (ETA and ETB) reside in the kidney [19]. The ETA receptor, present on vascular smooth muscle, mediates most of the vasoconstrictor effects of ET1 (Fig. 1). The ETB receptor is expressed predominantly on endothelial cells, where it stimulates the release of nitric oxide from endothe-

hal cells (Fig. 1) [10]. The ETB receptor generally has been viewed as a mechanism for offsetting the vasoconstrictive effects of ETI. However, the ETB receptor is also expressed to some extent

in some smooth muscle cells, where it can also mediate vasocon-

striction (Fig. 1) [19]. Thus, ET production and its resulting

Nephrology Forum: Acute renal failure

1105

vasoconstriction potentially can be blocked at multiple points, un [26] as well as endothelial vasodilators including prostaglanincluding inhibition of ET1 gene transcription, blockade of the dins and EDNO [26, 391. Brezis and associates have suggested conversion of prepro ET to big ET, inhibition of the ECE and, that EDNO's response to radiocontrast medium is a protective mechanism against radiocontrast-medium-induced production finally, ET receptor blockade [19]. Endothelin-receptor blockers (either specific ETA receptor of ET and that the predisposition of patients with underlying antagonists or antagonists that block both ETA and ETB recep- chronic renal failure, diabetes, and vascular disease to radiotors) improve renal function and ameliorate the morphologic contrast-medium-induced ARF [1] may be due to the presence severity of tubular injuly associated with experimental ischemic of chronic underlying endothelial injury, which blunts or preARF [20—22]. Also, ET receptor antagonists reverse intrarenal vents EDNO release in response to the administration of this vasoconstriction and improve renal function following a number nephrotoxin [15, 39]. of nephrotoxic insults that commonly cause ARF, including acute cyclosporine administration [23, 24], rhabdomyolysis [25], and use of radiocontrast medium [26]. In addition to ET receptor block-

ade, other potential therapeutic modalities inhibiting the ET

Our understanding of the role of NO in ARF has been

expanded, and at the same time complicated, by recent studies implicating NO produced by ischemic tubular cells as a mediator of cellular injury [40, 41]. The NO produced by tubular cells is

system in ARF include specific inhibitors of ET gene transcription and ECE inhibitors. It is of interest that atrial natriuretic peptide (ANP) inhibits transcription of the ETI gene and reduces synthesis of the ETI peptide [27]. This effect might contribute to ANP's well-documented protective effect in ischemic and toxic forms of experimental ARF [1, 13].

generated by an inducible nitric oxide synthase (iNOS), an

How does endothelin produce such severe and long-lasting

sition to ATN might be related in part to increased tubular

vasoconstriction in ARF [28, 29]? Kon's group has demonstrated that cyclosporine nephrotoxicity is associated with increases in ET1 gene expression, ET peptide production, and expression of the ETB receptor, all of which can persist for hours to days after the insult [28]. The same group also has demonstrated that ETB receptors on endothelial cells mediate the autoinduction for ETI synthesis (Fig. 1) [29]. On the basis of these interesting findings, Kon and associates have suggested that renal injuries associated

generation of NO [42]. In summary, in patients with ARF, opposing abnormalities in NO production within the endothelial and tubular compartments of the kidney contribute to renal injury. Reduced endothelium-

with ARF induce synthesis of endogenous ET1, which then perpetuates its own production after the initiating injury has

enzyme distinct from cNOS. Ischemia increases the activity of iNOS. The resultant increase in production of NO within the renal tubular cells likely contributes to cytotoxicity, possibly by combin-

ing with superoxide to form the oxidant peroxinitrite [40, 41]. Sepsis also increases iNOS activity, and septic patients' predispo-

derived NO production causes vasoconstriction and worsens ischemia; increased NO production by tubular cells exacerbates the injurious effects of ischemia on these cells. This compartmentalization of NO likely explains why neither NO donors (such as nitroprusside) [13] nor nonspecific NO inhibitors have a protective effect in ARF [41]. Any therapeutic intervention designed to

modulate NO in ARF requires selective alterations of NO production within endothelium and/or renal tubular cells. Goligorto increasing ET release, simultaneously reduces the release of sky's group has led the way in this challenging area by showing EDNO produced by the constitutive nitric oxide synthase (eNOS) that antiscnse oligonucleotides to iNOS both ameliorate injury of (Fig. 1) [1]. This reduction in EDNO production has been renal cells in culture that have been exposed to oxidative stress described in response to many renal insults that result in ATN, [41] and improve renal function in an isehemic model of ARF in resolved (Fig. 1) [28, 29]. Evidence suggests that sublethal endothelial injury, in addition

including ischemia [12, 30], cyclosporine toxicity [31], and pigment

(myoglobin and hemoglobin)-induced ARF [32]. Some evidence indicates that renal cNOS activity recovers within 24 to 48 hours after an ischemic insult [33], but the reduction in cNOS activity and EDNO production following cyclosporine toxicity and pigment-induced renal failure is more persistent (Fig. 1) [31, 34]. A deficiency in NO production in ARF can contribute to vasoconstriction in a number of ways. First, the constitutive production of NO normally plays an important roic in regulating vascular tone by maintaining basal systemic and renal arterial vasodilation [35]. Lack of EDNO also can be detrimental by increasing ET production and vasoconstrictor activity; NO downregulates the production and activity of endothelin in vitro both by inhibiting gene transcription and production of ET [36], as well as by modulating the effect of ET on its receptor (Fig. 1) [37]. We recently reported the results of rat studies demonstrating the importance of EDNO as an inhibitor of ET production in vivo [38]. Thus, many causes of experimental ARF, including ischemia, hemoglobinuria, myoglobinuria, and cyclosporine, are associated with the dual and additive effects of an increase in endothelin-

induced vasoconstriction and a reduction in EDNO-induced vasodilation (Fig. 1). Radiocontrast medium is the only nephrotoxin known to stimulate endothelial production of both endothe-

vivo [43]. Medullaty congestion. In addition to vasoconstriction, medullary

isehemia due to congestion and physical obstruction of the capillaries of the outer medulla by red cells, platelets, and leukocytes also contributes to the persistent reduction in medullary perfusion associated with ARF [16, 17, 44—46]. Although some investigators have stressed the importance of leukocytes in contributing to medullary congestion [44, 46, 471, the role of leukocytes in the pathogenesis of ARF has remained controversial until recently [1, 4]. One of the many benefits of the explosion of

knowledge regarding the structure and function of leukocyte adhesion molecules [481 has been the clear evidence that ischemic injury to many organs, including the kidney, stimulates the release

of inflammatory mediators that in turn activate the adhesion molecules on leukocytes and the expression of their ligands on endothelia! cells [48]. These events lead to increased leukocyte-

endothelial adhesion and endothelial injury, analogous to the events associated with a classic inflammatory reaction [48]. Antibodies that neutralize the binding activity of either the adhesion

molecules on leukocytes [49] or on endothelial cells [50—52] substantially ameliorate the functional and histologic injury associated with ischemic ARF. The conclusions of these studies using

antibodies have been substantially strengthened by the finding

1106

Nephrology Forum: Acute renal failure

Table 1. Potential responses of renal tubular cells to renal injury Sublethal injury Cellular dysfunction Loss of polarity Loss of tight junction gate function Loss of cell-substrate adhesion Exfoliation of viable cells from tubular basement membrane Aberrant renal tubular cell-cell adhesion Altered gene expression Cellular dedifferentiation Recovery of cell function Lethal injury Necrosis Apoptosis

profound reduction in GFR characteristic of this disease is a well-recognized paradox. One of the main themes of this Nephrology Forum is that renal tubular cells can respond to renal injury in many different ways (Table 1) [58, 59]. Many tubular segments, including the distal convoluted tubule and collecting duct, prob-

ably escape injury altogether. Tubular cells that are lethally injured can die by apoptosis, a process distinct from necrosis [601. Finally, accumulating evidence suggests that many renal tubular cells are sublethally injured and, while appearing intact morphologically, can still contribute to the tubular leakiness and obstruction associated with ARF (Table 1) [61, 62]. Sublethal injury. Sublethal injury markedly impairs the function

of renal tubular cells. Injury to the cytoskeleton contributes to that knockout mice lacking the gene for ICAM-1, the endothelial

ligand for leukocyte adhesion molecules, also are protected

many of these functional disturbances [61, 62]. The actin cytoskeleton plays an important role in epithelial cell function; this subject

is discussed in greater detail in two excellent reviews [63, 64]. Briefly, however, the actin cytoskeleton is important in many aspects of normal cell function and structure, including mainteinjury remain incompletely defined, but considerable evidence nance of normal brush-border structure, cell polarity, tight juncindicates that leukocytes contribute to the persistent hemody- tion integrity [65] and normal adhesion of epithelial cells to the namic impairment associated with ischemic injury [531 both by tubular basement membrane of the nephron [63, 64]. The alterreleasing vasoconstrictors (such as leukotrienes and thrombox- ations in the actin cytoskeleton that occur in the absence of ane) [47] as well as by causing endothelial injury, which leads to irreversible injury and cell death contribute importantly to cell the congestion of the capillaries of the outer medulla by red cells, dysfunction by causing abnormalities in each of these cell funcleukocytes, and platelets [16, 17, 44—46, 54]. Furthermore, the tions. imbalance I have described in endothelin and EDNO production The actin cytoskeleton is present as a layer of microfilaments associated with renal injury likely contributes to the leukocyte- below the apical plasma membrane, where it forms the terminal induced damage. The constitutive release of EDNO into plasma web. Bundles of actin filaments extend from the terminal web into inhibits activation of leukocytes [55] while endothelin promotes the microvilli and contribute to the architectural stability of the the adhesion of leukocytes to endothelial cells [54]. Thus, the brush border [66]. Kellerman and Bogusky have reported that protective effects of endothelin blockade are likely due not only to ischemic injury to the kidney in vivo rapidly disrupts the actin against acute ischemic renal injury [53]. The mechanisms by which leukocytes promote ischemic renal

reduced endothelin-mediated vasoconstriction, but also to blockade of endothelin-induced leukocyte activation and amelioration of medullary congestion [541.

cytoskeleton of the microvilli and redistributes F-actin microfilaments from the microvilli into the cytosol [67]. These findings provide an explanation at a molecular level for the disruption of Leukocyte-induced injury also probably contributes to the the proximal tubular brush border noted many years ago both in exacerbation of acute renal damage in patients with established experimental models of ischemic injury [6] as well as in humans ATN undergoing hemodialysis. "Incompatible" dialysis mem- with ATN [58]. branes made of cellulose acetate activate leukocytes [56]. A The important concept that sublethal injury to tubular cells is reduction in leukocyte-induced renal injury probably accounts for associated with profound alterations in the function of the the important observation by Hakim et al [57] that patients with nephron was initially suggested by Molitoris, who demonstrated ARF are more likely to survive and recover renal function when that sublethal injury leads to loss of the normal polarity of renal dialyzed with a relatively biocompatible dialysis membrane than epithelial cell monolayers [61, 65, 66, 68]. Loss of cell polarity with a bioincompatible membrane. Thus, the development of occurs when the "fence" function of the tight junction fails [63]. effective antibodies or soluble antagonists directed against the Molitoris and colleagues have shown that the Na/K-ATPase, panoply of adhesion molecules associated with leukocyte-induced which is normally restricted to the basolateral cell membrane, injury [481 may provide novel and potentially additive forms of redistributes following sublethal injury and results in the presence treatment to prevent or ameliorate ARF. of functional Nat/K' -ATPase molecules on the apical domain of the cell [61, 65, 66, 681. This event follows disruption of the actin Ischemic and toxic injury to tubular cells cytoskelelon, which normally "anchors" the Na /K -ATPasc

Although the term ATN accurately defines the population of molecules to the basolateral cell membrane. Loss of polarity of the Na/K -ATPase profoundly disturbs the normal unidirecof the multiplicity of responses that renal tubular epithelial cells tional transport of salt and water of the renal epithelial cell [61, demonstrate following an ischemic or toxic insult (Table 1). 65, 66, 681. "Acute tubular necrosis" implies that the predominant response Further investigation by Molitoris and other investigators has of tubular cells to injury is cell death by necrosis. Necrosis of some added to our understanding of the ways in which sublethal injury tubular cells clearly occurs, but morphologic studies of ATN in can alter the function of renal epithelial cells [62, 68—741. In humans over the last 2 to 3 decades have revealed that areas of addition to its "fence" function, the tight junction also has a frank necrosis are focal and relatively sparse [58, 59]. In fact, the "gate" function, which refers to the barrier to paracellular movemarked disparity between the relatively subtle histologic lesions ment of water and solute across all normal epithelia. The "gate" seen in biopsy specimens from humans with ATN and the function of renal tubular cells is also markedly impaired by cells most severely injured in this disease, it fails to provide a sense

1107

Nephrology Fomm: Acute renal failure

sublethal injury [62, 68]. This loss of gate function prevents the

renal epithelium from acting as a barrier to free movement of solute and water across the tubular epithelium. Thus, severe

Normal renal epithelium

"backleak" of glomerular filtrate can occur in the absence of lethal injury and without any apparent morphologic evidence of tubular cell injury [62]. The loss of both "fence" and "gate" functions of

the tight junction is reversible if the cell ATP content recovers before lethal cell damage ensues [62]. in addition to loss of tight junction integrity, sublethally injured renal cells undergo a loss of cell-matrix adhesion. Epithelial cells

Sublethal injury

are normally attached to the underlying matrix by the 131 integrin family of transmembrane proteins normally localized to the basal

aspect of the cell. This protein has an intracellular domain that connects to the actin cytoskeleton via a complex of proteins called the adhesion plaque [64]. The extracellular domain of the protein attaches to receptors present on matrix proteins, such as fibronec-

tin and collagen, which constitute the tubular basement membrane [64]. One well-defined matrix receptor for the 131 integrin is a tripeptide sequence of arginine-glycine-asparagine (abbreviated

as RGD) present on many matrix proteins. The loss of polarity associated with sublethal injury to renal cells causes redistribution

of proteins other than the Na 7K-ATPase that include the /31 integrins. As a result, /3 iritegrins also become expressed on the apical domain of sublethally injured renal cells (Fig. 2) [69, 70]. The redistribution of the /31 integrin probably explains, at least in part, the observed impairment of cell-substrate adhesion in sublethally injured renal cells in culture [62, 69, 71, 72]. The disruption of cell-substratum attachment demonstrated in vitro following sublethal injury is relevant to the pathophysiology of human ATN. Detachment of renal epithelial cells from basement membrane has been documented by morphologic studies of renal biopsy specimens taken from humans with ATN [59]. Acute renal failure also is associated with the loss of viable tubular cells from the nephron and their excretion in the urine [59]. Thus, in addition to lethal cell injury and loss of tight junction function, the detachment of viable cells from the renal epithelium provides yet another mechanism for backleak of glomerular filtrate. Goligorsky and associates have expanded our understanding of

Presence of an excess of free RGD

•i•cj c.r .

S

2. Sublethal injury to renal tubular cells leads to exfoliation of

Fig. the likely functional consequences of sublethal injury to renal viable renal tubular cells and aberrant 1 integrin-mediated cell-cell tubular cells by hypothesizing that sublethal injury leads to adhesion. (Adapted from Ref. 71 with permission.) intratubular obstruction as well as to backleak of glomerular filtrate [70]. They hypothesized that loss of f31 integrin polarity leads to aberrant binding of tubular cells to one another so that other, as-yet-unidentified mechanisms of action in ARF, this

exfoliated tubular cells attach to one another as well as to beneficial effect is probably mediated, at least in part, by preventnon-exfoliated, sublethally injured cells (Fig. 2) within the tubular lumen [70, 71]. The hypothesis, described schematically in Figure 2, is based on data from renal cells in culture, which suggest that

ing intratubular obstruction (Fig. 2) [74]. Further studies are needed to clarify the mechanisms by which

RGD mediates its protective effect. Also, the possibility that the /31 intcgrins expressed on the surface of sublethally injured adhesion molecules other than the /31 integrin (such as clusterin cells remain functionally capable of binding to their RGD recep- or cadherin, which mediate cell-cell adhesion) or matrix receptors tor [69, 70, 72]. On the basis of this hypothesis, Goligorsky's group other than RGD contribute to the pathogenesis of ARF needs to theorized further that administration of an excess of soluble be investigated. The therapeutic potential of multiple soluble RGD-containing molecules would saturate the extracellular bind- factors that interfere with tubular cast formation by binding to ing site of all exposed 131 integrins within the lumen of the adhesion molecules and/or to their matrix receptors remains an nephron and thus prevent or ameliorate renal tubular cell-cell exciting area of basic investigation. adhesion and intratubular obstruction (Fig. 2) [71]. To test this Lethal cell injury: Necrosis. The morphology and biology of hypothesis in vivo, they examined the effect of an intravenous necrosis of tubular cells following injury have been discussed in infusion of soluble, RGD-containing peptides into rats with detail elsewhere [4, 75]. 1 will discuss necrosis only briefly in this established ischemic ARF and found that the RGD-containing Nephrology Forum to emphasize the important differences hepeptides ameliorated the reduction in GFR associated with twecn necrosis and apoptosis. Cell necrosis is induced by rapid ischemic injury [71, 73]. While RGD-containing peptides can have and overwhelming depletion of cellular energy stores and by

1108

Nephroloi Forum: Acute renal failure Table 2. Mechanisms of cell necrosis Severe depletion of cellular ATP stores Redsiced activity of membrane transport pumps Cell swelling

Increase in intracellular free calcium Activation of phospholipases Activation of proteases Glycine depletion Plasma and subcellular membrane injury

irreversible injury to the lipid bilayer of the plasma membrane and subcellular organelles [4, 75, 761. These events disrupt membrane integrity and transport activity and induce cell swelling as a result of the entry of sodium and water into the dying cell [4, 75]. Loss

of plasma membrane integrity later allows cytosolic contents containing proteolytic enzymes to leak into the extracellular space. The resultant injury to and inflammation of surrounding

Table 3. Potential causes of apoptosis of renal tubular cells in ARF Physiologic activators

TNFa

TGF

Loss of "survival" factors Deficiency of renal growth factors Impaired cell-matrix adhesion Loss of cell-cell adhesion Cytotoxic agents Hypoxia; ischemia; anoxia Reactive oxygen species Increased intracellular free calcium Hyperthermia Nutrient deprivation Irradiation Pharmacologic agents Chemotherapeutic agents Cyclosporine; rapamycin Irradiation

tissues is easily observed histologically. In necrotic cell death, many derangements in cellular metabo-

ischemia or to toxins at concentrations below those required to lism occur concurrently and in an uncoordinated but additive induce rapid metabolic collapse and necrosis [60, 79]. It seems as manner. These alterations ultimately cause the rapid demise of if lethally injured cells not rapidly killed by necrosis detect the the cell. Some of the biochemical events that are the putative injurious signal and initiate the apoptotic pathway. We have causes of necrosis are listed in Table 2 [4, 75]. Because the events demonstrated that cisplatin, a common cause of ATN, induces that lead to necrosis are chaotic and unregulated and involve necrosis of renal tubular cells in culture when these cells are multiple and diverse metabolic pathways, attempts at developing exposed to high concentrations (--l mm) of the drug. In contrast, therapeutic strategies to prevent necrosis after an ischemic or much lower concentrations (—50 /.tm) of cisplatin produce the toxic event have been only partially effective at best. classic morphologic features of apoptotic cell death [781. Our Apoptosis. In direct contrast to necrosis, apoptosis is a carefully study confirms the observation made by others in many different choreographed and genetically directed form of cell death that cell types that the mechanism of renal tubular cell death depends requires energy [60, 76]. The morphologic characteristics of on the severity of the injury [79]. apoptosis and necrosis are quite distinct and remarkably constant Other known triggers of apoptosis might be important in ARF. among different cell types [76, 77]. In contrast to necrosis, cells One important cause of apoptosis is a relative deficiency of

dying by apoptosis become progressively smaller as cytosolic volume decreases and the nucleus condenses [771. This process

essential "survival factors" that are necessary not only for cell growth, but also for the maintenance of cell viability [80]. The

can be identified by light or fluorescent microscopy with the appropriate staining techniques: the normal chromatin structure

apoptotic program therefore can be regarded as a "default pathway," which is always ready to be activated unless it is

of the nucleus is lost and the nucleus appears as a round,

constantly suppressed by survival factors. Survival factors include

featureless mass [60, 78]. The most reliable way to recognize soluble growth factors as well as signals transmitted into the cell apoptosis is by transmission electron microscopy, in which the from interactions between cells (cell-cell adhesion) or between intact plasma membrane and subcellular organelles as well as cells and their underlying matrix (cell-substrate adhesion) [80]. nuclear condensation are specific features [77]. Epithelial cells depend on normal cell-cell and cell-matrix Another feature unique to apoptosis is the fragmentation of the adherence to remain viable. Loss of signals received from these condensed nuclei into multiple pieces. The apoptotic cell eventu- normal interactions also leads to apoptosis. Thus the loss of ally breaks apart into many "apoptotic bodies," pieces of nuclear normal cell-cell and cell-matrix adhesion associated with cell material surrounded by a rim of cytoplasm and enclosed by an injury [621 might lead to apoptosis in ARF. The soluble growth intact plasma membrane [77, 781. Apoptotic cells and bodies are factors necessary for survival depend on the cell type. Renal phagocytosed by resident macrophages, as well as by surrounding tubular epithelial cells in culture die by apoptosis in the absence epithelial cells and fibrohiasts. This clearance mechanism is of renotrophic hormones such as epidermal growth factor (EGF) extremely efficient, and apoptotic cells and bodies disappear and insulin-like growth factor-i (IGF-1) [81]. Perhaps the defirapidly from the tissue without any trace of surrounding tissue ciency of renal EGF production that occurs following acute renal injury or inflammation. For this reason, apoptosis is usually injury [82, 83] contributes to tubular cell apoptosis in ARF. inconspicuous and difficult to detect in tissue sections, even when Another inducer of apoptosis in epithelial cells, TNFa, might responsible for extensive cell loss [60, 71. contribute to ARF in septic states associated with an increased Many factors can cause apoptosis (Table 3). Jerrold Levine and production of this cytokine [60]. I have reviewed these factors in a recent review article [60]. The The complex mechanisms involved in what has become known realization that both apoptosis and necrosis can be induced by the as the "execution" phase of apoptosis are still unfolding. One of same noxious stimuli has resulted in much interest in the role of the most intruiging enigmas in the cellular biology of apoptosis is apoptosis in different forms of renal injury, including ATN [60]. how such a bewildering diversity of precipitating factors can result

Apoptosis is a common response in cells exposed to modest

in a genetically controlled form of cell death with the same

1109

Nephrology Forum: Acute renal failure

morphologic manifestations. One biochemical event characteristic

of apoptosis is the activation of an endonuclease that causes fragmentation of DNA [76]. In apoptosis, the endonuclease acts at internucleosomal sites, generating fragments of DNA in integer multiples of nucleosomal size (200 base pairs). This fragmentation produces the "ladder" pattern seen when DNA from apoptotic cells is subjected to agarose electrophoresis. In contrast, during

[Acbvation of an iniflator Protease]

necrosis, DNA typically is degraded into small fragments of varying lengths, resulting in a "smear" pattern on DNA electroof a cascade of cytop lasmic and nuclear proteasesj cells or tissue sections the DNA fragmentation associated with apoptosis is by end-labeling the free ends of the DNA generated by endonuclease activity with a fluorescently labeled nucleotide k via the enzyme terminal deoxynucleotidyl transferase (TUNEL) [Cleavage of multiple substrates in both cytosol and nucleus] method [84]. A ladder pattern on DNA electrophoresis and/or positive TUNEL staining strongly supports apoptosis. However, one must recognize that the ladder pattern of DNA fragmentation 1 is not completely reliable for distinguishing apoptotic from ne-

phoresis [60]. Another convenient method for identifying in intact

L

.

crotic cell death. Apoptosis, defined morphologically, can occur in

the absence of internucleosomal DNA cleavage, and DNA laddering or positive TUNEL staining has been demonstrated in renal epithelial cells that undergo necrosis in response to hypoxiareoxygenation injury [60, 85]. Thus, the biochemical feature of endonuclease activation, although characteristic of most cells undergoing apoptosis, has been dissociated in some situations from the classic morphologic changes that define apoptotic cell death [60, 77]. Furthermore, the importance of endonuclease activation as an initiating event in the execution phase of apoptosis has been questioned [78]. ytoplasmic rather than nuclear events (such as endonuclease activation) probably play the initiating and primary role in mediating apoptotic cell death [86, 87]. A more proximal (upstream) event in the execution phase of apoptosis involves the activation of a cascade of proteases (Fig. 3) [86, 87]. A growing number of studies support a role for members of the interleukin-1 converting enzyme (ICE) family of proteases in mediating apoptosis. All members of the ICE-related family cleave substrates after an aspartate residue. These enzymes are produced as inactive pro-

enzymes that are activated by cleavage at critical aspartate

Morphologic and biochemical features of apoptosis

• Nucleosomal DNA cleavage • Nuclear condensation • Nuclear fragmentation

• Cytoskeletal changes • Cell membrane budding • Apoptotic bodies

• PHAGOCYTOSIS Fig. 3. The execution phase of apoptosis is mediated by activation of an ICE-like protease, which leads in turn to activation of a cascade of related

proteases that ultimately cleave multiple cytoplasmic and nuclear substrates.

recent description of inhibitors of ICE-like proteases that can prevent apoptosis induced by a number of stimuli has reinforced the concept that activation of proteases is a proximal and important event in apoptosis (Fig. 3) [90, 91]. These studies also suggest that the development of inhibitors of apoptosis have therapeutic potential in conditions such as ARF in which apoptosis contributes to loss of renal tubular epithelial cells.

residues that themselves conform to the substrate consensus for ICE family proteases [86, 87]. It is currently believed that one or a few ICE-like proteases are initially activated following an apoptotic "trigger." These proteases then activate additional ICE-like protcases in successive and expanding hierarchies (Fig.

The Bcl-2 family of proteins is important in regulating the execution phase of apoptosis [92]. The Bcl-2 protein as well as other members of this family (such as BAG-i and Bcl-XL) protect cells against apoptosis. In contrast, other Bcl-2-like proteins (such as Bad, Bak, and Bcl-X) promote cell death. Thus, for any given 3) [86, 87]. This proteolytic cascade ultimately leads to the apoptotic stimulus, the balance between death and survival apactivation of proteascs that target multiple proteins within the pears to be determined by the ratio of apoptosis-promoting and cells. The proteolytic cleavage of these cytoplasmic and nuclear -suppressing Bcl-2 family proteins [92]. Although the mechanism substrates leads to the demise of the cell by apoptosis (Fig. 3) [86, by which members of the Bcl-2 family regulate apoptosis is not 87]. Nuclear substrates of ICE-like proteases known to be cleaved in

known, it is clear that the susceptibility of cells to apoptosis can be

altered by changing the level of expression of individual family apoptosis include the DNA repair enzyme poly(ADP-rihose) member proteins. Alteration in the expression of various mempolymerase (PARP) and lamin, the nuclear envelope protein bers of the Bcl-2 family of proteins provides another therapeutic lamin [87]. Cleavage of lamin might contribute to the nuclear strategy by which cells can be pharmacologically manipulated to condensation and fragmentation characteristic of apoptosis [87, either prevent or facilitate apoptosis. 881. Protease activation of endonucleases also might lead to the Despite the difficulty in identifying apoptotic cell death in tissue DNA fragmentation associatcd with apoptosis [87]. ytoplasmic sections, evidence has emerged suggesting that apoptosis contribsubstrates for ICE-like proteases also include cytoskeletal pro- utes to cell death following ischemia and/or reperfusion injury in teins such as actin and fodrin [86, 87, 89]. The extent to which vivo in the kidney [93, 94] as well as in other organs such as the cleavage of cytoskeletal proteins induces the phenotypic changes heart [95]. However, this area remains controversial because sonic of apoptosis (such as cell membrane "budding") is unknown. The investigators have been unable to demonstrate any morphologic

1110

Nephrology Forum: Acute renal failure

evidence of apoptosis following ischemia and/or reperfusion injury to kidneys in vivo [96]. We have examined the mechanisms of

tion of neighboring cells in the S3 segment of the proximal tubule, where irreversible cell injury has occurred [82].

The molecules that mediate the proliferative response after cell death in proximal tubular cells in culture and have demonstrated apoptosis in response to multiple stimuli, including renal renal injury have not been defined, and the role of known growth factor withdrawal [811, chemical anoxia [97], and cytotoxic

proximal tubular cell mitogens—such as EGF, IGF-1, transforming

agents such as cisplatin [78]. Clearly, in view of the potential for pharmacologic interventions that could inhibit apoptosis, studies to establish the mechanisms of apoptosis in renal tubular cells and the extent to which apoptosis contributes to the loss of these cells following ischemic and toxic renal injury have become extremely important.

growth factor (TGF-a), and hepatocyte growth factor (HGF)— remains uncertain [98]. Further clarification of the responses to

Recovery from ATN

sublethal injury at a genetic level may ultimately provide the tools to alter this response in ways that accelerate repair and regeneration in patients with ATN. This line of research has stimulated investigation of a number of renotrophic hormones as therapy of ARF. I will conclude by briefly reviewing this issue. Renal growth factors, including EGF [101], IGF-1 [102], and HGF [103], already have proved effective in facilitating recovery of renal function in experimental animals when administered after an ischemic or toxic insult to the kidney [104]. The mechanisms responsible for this beneficial effect are still being actively inves-

A striking characteristic of ischemic and toxic ATN is that the injured and dysfunctional kidneys usually recover normal structure and function. This dramatic process depends on a complex series of events, including recovery of sublethally injured cells as tigated, but it is known that these growth factors stimulate well as the removal of intratubular cellular debris and casts. regeneration of cells following injury; EGF, IGF, and HGF are all Finally, an essential part of the recovery process is replacement of known mitogens for proximal tubular cells [104]. Also, the exprestubular cells that are lost via exfoliation of viable cells from the sion of the receptors for renal cell growth factors is increased after tubular epithelium or by cell death from necrosis or apoptosis. acute renal injury in rats [104]. These observations explain the This process requires replication of surviving, viable renal cells in robust proliferative response of tubular cells following administration of these mitogens after injury. areas of cell loss [82]. Growth factors also might inhibit apoptosis in ARF [60]. As I Under normal circumstances, renal tubular cells in vivo are quiescent and do not readily divide in response to growth factors. have already mentioned, there may be a relative deficiency of However, following an ischemic or nephrotoxic insult, surviving renal growth factors such as EGF in ARF [83]. The administration renal tubular cells re-enter the cell cycle and replicate, thus of growth factors might inhibit apoptosis by replacing necessary replacing irreversibly injured cells [14, 82]. These cells possibly survival factors that inhibit activation of the "default" pathway of undergo partial dedifferentiation that allows them to undergo apoptosis. While theoretically feasible, the extent to which adminmitosis (Table 1) [14, 82]. The complex series of events that istration of growth factors promotes recovery of ARF by inhibitmediate this process is still being elucidated. What is known is that ing apoptosis remains to be established. To summarize, I have emphasized in my discussion how basic cell regeneration begins very soon after the noxious insult [82]. Ischemic and nephrotoxic renal injury both induce alterations in research, by elucidating the molecular and cellular events respongene expression similar to those induced by growth factors in cells sible for ARF, has led to the promise of new therapies for this in culture [82]. Transient expression of genes such as c-fos and common disease. The next few years should prove an exciting time egr-1 [14, 981, as well as KC and JE [99], occurs almost immedi- for clinical researchers interested in examining the efficacy of ately after renal ischemia and precedes increased protein or DNA these agents, either alone or in various combinations, in their synthesis. Some of these genes (such as the "early immediate patients with ATN. response genes" egr-1 and c-fos) code for DNA-binding transcription factors that likely initiate the transcription of the many other QUESTIONS AND ANSWERS genes necessary for cellular division [14, 82]. Other genes such as DR. NIcoLAos E. MADIAS (Chief Division of Nephrology, New KG and JE code for proteins with chemotactic effects that can attract into areas of injury monocytes and other neutrophils that England Medical Center, Boston, Massachusetis): You have discussed a long list of contributory factors in the pathogenesis of participate in the reparative response [82]. It is interesting that most studies have localized the expression acute renal failure and suggested a number of strategies for of these genes to the mTAL segment of the nephron; other intervention. In your judgment, which avenues would be the most nephron segments show little evidence of altered gene expression rewarding to pursue for further research in terms of prevention [14, 99, 100]. A complex system of biochemical pathways must and amelioration of established acute renal failure? DR. LIEHFRTHAI.: The route taken in clinical studies to examine remain intact for mTAL cells to perceive an injurious stimulus and to modulate transcription of the appropriate genes in re- the efficacy of various experimental agents shown to ameliorate sponse to this "stress" signal [82, 100]. For these reasons, the ARF in animals will largely be determined by the process of FDA studies demonstrating activation of gene expression in mTAL approval. Two of the agents mentioned in my discussion, atrial cells add to the evidence that I summarized earlier that argues natriuretic peptide (ANP) and the growth factor IGF-1, already against the mTAL segment being a major site of lethal cell injury are being examined in humans with ATN. Antibodies to some of in ARF [1, 6, 10, 11]. Furthermore, the close anatomic proximity the leukocyte adhesion molecules have been tested in phase-I of the S3 segments of the proximal tubule and mTAL segments human trials and will soon be available for studies in ARF. The within the outer medulla suggests the intriguing possibility that testing of these antibodies will be complicated by the large variety growth factors and cytokines released by mTAL cells following of adhesion molecules involved in the process of leukocyte renal injury act in a paracrine fashion to stimulate the regenera- adhesion and activation as well as by the experimental data

Nephrology Fomm: Acute renal failure

suggesting that blockade of some adhesion molecules (for example, ICAM) is more effective than inhibition of others (such as the selectins). Also, agents that block either the production of endothelin (ECE inhibitors) or endothelin action (ET receptor antag-

1111

DR. LIEBERTHAL: The red-cell-free kidney represents a very poorly oxygenated model. Whereas the cell-free perfusate has a

very high oxygen tension, the total oxygen content, which is

dependent on the oxygen dissolved in the aqueous phase, is very onists) will be available soon for experimental use in humans. low [8—111. In addition, loss of oxygen from the medulla by Ultimately, combinations of these agents might be the most countercurrent diffusion of oxygen is substantially exacerbated by

effective form of treatment in ARF. The other strategies I the lack of red cells [8—11]. Futhermore, in the uninjured isolated mentioned (such as disintegrins and ICE-like protease inhibitors) are still in the early phase of experimental investigation. DR. MADIAS: That leads me to my second question. Could you tell us a bit about the experience with growth factors in animals and, possibly, in humans with acute renal failure? Also, what are the systemic effects of these compounds, and what is their impact on the body's protein metabolism in acute renal failure? DR. LIEBERTIIAI: A potentially important systemic consequence

perfused kidney, glomerular filtration is preserved, so the delivery

of solute to the tubules remains substantial. It is this unusual combination of extremely poor oxygen delivery to the medulla with the continuing high energy demands for sodium transport that explains the susceptibility of the mTAL segments to necrosis in this model [8—li]. This situation is unlikely ever to occur in ATN in vivo because, by definition, GFR is severely reduced in this situation and red cells are always present [9—il].

of IGF-1 administration is its potent anabolic effect, which has DR. RONALD PERRONE (Division of Alephrology, New England been well documented in humans as well as in animals without Medical Center): You described an impressive array of experimenacute renal failure [102, 104]. The net catabolic rate is often tal data derived primarily from rat models and cell culture. Are markedly elevated in acute renal failure, a metabolic complication there areas of human renal function or physiology, biochemistry, that can profoundly increase nutritional and dialysis require- etc., in which clear differences exist that we might want to target ments, impair wound healing, and increase overall morbidity and differently? Do we know enough about all the pathways involved mortality rates in patients with ATN [1]. Studies in animals have to be able to extrapolate data from rats to humans? demonstrated that IGF-1 reduces the rate of protein degradation, DR. LIEBERTHAL: As you imply, experimental studies suggest stimulates protein synthesis, and substantially ameliorates the that the different processes that play a role in renal failure marked loss of body weight associated with acute renal failure (hemodynamic factors, tubular obstruction, tubular backleak) [102, 104]. The improved metabolic milieu resulting from the contribute to a loss of GFR to varying extents in different models relatively greater GFR in IGF-1-treated animals probably con- of ARF. However, we know little about the relative role played by tributes to the associated reduction in catabolic rate. However, these different pathophysiologic factors in humans with ATN. IGF-1 appears to have anabolic effects that are independent of the Thus we are as yet unable to target different therapies for patients uremic state [1021. Furthermore, other growth factors such as with different causes of ARF. DR. MADIAS: You didn't mention the renin-angiotensin system EGF do not appear to have effects on catabolic rate and body weight in animals with acute renal failure equivalent to those of in your talk, yet this system is activated in human and experimenIGF-1 [1041. Thus, it is possible that IGF-1 will turn out to be tal acute renal failure, and angiotensin II has direct effects on particularly useful in the treatment of humans with ATN by renal cells. Could you briefly discuss the role of the reninreducing protein breakdown as well as facilitating recovery from angiotensin system in the pathophysiology of acute renal failure, even if to exclude it? renal injury. DR. LIEBERTHAL: That is an important question. For lack of DR. JOHN T. HARRINGTON (Dean, Tufts University School of Medicine, Boston, Massachusetts): You said that backleak was one

time, I have not discussed the potential role of the renin-

of the mechanisms that underlie acute tubular necrosis. What's the best evidence in humans for the existence of that phenomenon?

angiotensin system in ARF, but the vast body of literature on this subject has been carefully reviewed elsewhere [1051. It is clear that

renin production by the kidney is increased in experimental

models of ARF. However, studies that have examined the effects of glomerular filtrate plays an important role in the pathophysi- of either angiotensin-converting-enzyme inhibitors or angiotensin ology of ATN in humans has been provided by Myers and Moran IT-receptor blockade have produced variable results, with amelio[5]. They compared inulin clearance with the clearance of a ration of ARF in a few studies and no beneficial effect reported in neutral dextran molecule, which is also freely filtered at the many others [105]. The role of the renin-angiotensin system in glomerulus but is of larger molecular weight than inulin. Neither humans is also unresolved [105]. It remains possible that inmolecule is reabsorbed or secreted by the renal tubules, so the creased release of angiotensin II, most likely in a paracrine or clearances of these two molecules by the normal kidney are autocrine fashion within the juxtaglomerular apparatus, contribcomparable (that is, fractional excretion of the dextran equals utes to renal failure by inducing afferent arteriolar and/or mesanunity). In contrast, in patients with ischemic ATN, dextran gial cell constriction [1, 4, 5, 105]. DR. ANDREW J. KING (Division of Nephrolo, New England clearance is substantially higher than that of inulin (fractional excretion of the dextran exceeds unity). This indicates that inulin, Medical Center): I am impressed with the animal data suggesting because of its smaller size, is able to passively "backleak" from the that endothelin production increases following challenge with cyclosporine or endotoxins and in the setting of acute tubular tubular fluid at a faster rate than the dextran. DR. HARRINGTON: You discussed the Epstein model of acute necrosis [23—26, 106]. The data in humans are relatively sparse. tubular necrosis, in which an absence of red cells in the perfusate My recollection is that plasma endothelin levels are unchanged or causes necrosis in the medullary thick ascending limb. I had mildly elevated in patients with acute renal failure. Could you cite assumed that this model was a well-oxygenated model. If that is any other evidence in humans that endothelin plays a role in the the case, why would the absence of red cells he so critical? pathogenesis of ATN? DR. LIEBERTFIAL: The most convincing evidence that backleak

1112

Nephrology Fonim: Acute renal failure

DR. LIEBERTHAL: I agree with your point that plasma endothelin

levels do not reflect the importance of endothelin in the pathophysiology of ARF. However, it is important to appreciate that endothelin is a paracrine hormone secreted to a large extent through the basolateral aspect of endothelial cells and therefore in the direction of vascular smooth muscle cells rather than the plasma. For this reason, plasma levels are a relatively insensitive marker of endothelin's contribution to any disease process. Futhermore, if the abnormality of endothelin secretion is localized to the vascular bed of one organ (for example, the kidney in ARF), any increase in plasma endothelin levels from that organ will be diluted in a peripheral plasma sample. Peripheral plasma endothelin levels likely are increased only when the abnormality in endothelin production is severe and present in multiple vascular beds. This might explain the very high plasma levels of endothelin in some patients, for example, those with hepatorenal syndrome.

diabetic patients with normal renal function have basically similar risk for acute renal failure as do non-diabetics? DR. LIEnERTHAL: There are no definitive answers to your

questions. The pathophysiology of radiocontrast nephropathy is

complex and likely involves many interacting factors. In my discussion I mentioned the animal studies [39] that have led to the hypothesis that the susceptibility of patients with CRF to radiocontrast nephropathy is due, in part, to dysfunction of endothelial cells within the renal vasculature, which accentuates the renovascular imbalance of endothelin and NO in these patients. Patients with normal renal function, including diabetics, are more likely to have relatively normal renovascular endothelial function. DR. HARRINGTON: Let me ask a broader question. What differ-

ence does this all make? We all know that for the most part patients don't die of acute renal failure; they die of sepsis or their

original disease. It seems to me that the same percentage of

To come back to your question, direct evidence for an involve- patients, approximately two-thirds, die now following an episode ment of endothelin in disease in humans will have to await the use of ARF as did 20 years ago. How do we affect that larger issue? of endothelin antagonists in clinical studies. DR. LIEnERTHAL: In my opinion, I feel your question reflects an DR. KING: Would you please comment on the possible role of a overly pessimistic view. I agree that there is a category of patients disturbance in tubuloglomerular feedback due to impaired nitric with ARF who are so debilitated by age, immune deficiency, oxide production in the pathogenesis of renal vasoconstriction in and/or multi-organ failure that their survival is unlikely to be patients with ATN? altered by any intervention that ameliorates the concurrent presDR. LIEBERTRAL: Nitric oxide, produced within the macula ence of ARF. However, I also believe that for many patients, any densa by the neural NO synthase (nNOS), plays a role in intervention that either prevents ARF or shortens the maintemodulating afferent arteriolar tone and intraglomerular pressure nance phase of established ARF is likely to improve morbidity via the tubuloglomerular feedback mechanism. No studies have and reduce mortality. The beneficial effect of the use of more directly examined the role of NO production at this site in ARF, biocompatible dialysis membranes on the survival of patients with however. DR. MknIAs: Would you comment on the following two inter-

ventional modalities used in acute renal failure? As you know, "low-dose" dopamine infusion is used extensively without much evidence of benefit but with considerable concern about the associated risks. Also, calcium-channel blockers have been reported to exert protective effects in experimental and clinical acute renal failure. DR. LIEBERTHAL: I agree entirely with your remarks regarding

dopamine. There is no evidence that it improves GFR, and there is some concern that it is deleterious to critically ill patients. I therefore do not recommend the use of dopamine for patients

with ATN. An excellent review on this subject was recently published [107].

In response to the second part of your question on calciumchannel blockers (CCB5), substantial experimental data indicate

that these drugs have potential as therapeutic agents in the prevention and treatment of ARF. Also, a few convincing studies in patients receiving cadaveric renal transplants suggest that CCBs administered prophylactically (before and after transplantation) are helpful in preventing or ameliorating ischemic ATN immedi-

ately following transplantation. I believe that additional controlled human studies examining the efficacy of CCBs in the prevention and treatment of ATN in other situations are still warranted.

ATN receiving hemodialysis [57] is just one indication that patients with ATN are likely to benefit from any intervention that reduces the need for dialysis altogether.

DR. MADIA5: Do growth factors influence the cytoskeletal abnormalities of cells that have sustained sublethal injury? DR. LIEnERTHAL: Yes. Substantial evidence has shown that growth factors play a central role in initiating and modulating the cytoskeletal events associated with the proliferative response. It is quite feasible that the effect of growth factors in ARF is mediated, at least in part, by promoting repair of the cytoskeleton.

DR. MAmAs: Did you say that apoptosis-related genes are upregulated in models of acute renal failure? DR. LIEnERTHAL: I know of no direct evidence as yet that expression of genes related to the apoptotic process (such as genes coding for the BCL2- or ICE-like protease families of proteins) is altered in acute renal failure. However, as I discussed in some detail, the expression of genes involved in the proliferative and regenerative phase of ATN is upregulated following an ischemic insult to the kidney. DR. KING: An injured kidney loses its ability to regulate renal blood flow. You raised the possibility that hemodialysis adds insult to injury. Should we be taking any special approach to hemodialysis for patients with ATN? For instance, should we delay it? DR. LIEBERTHAL: The notion that repeated episodes of hypotension during hemodialysis exacerbate ARF by causing repeated

DR. MARK E. WILLIAMs (Renal Unit, Joslin Diabetes Center, acute ischemic injury remains an unproven hypothesis that is Boston): You mentioned that diabetes and pre-existing renal based on two indirect sources. First, Conger demonstrated that impairment are risk factors for radiocontrast-medium-induced the renal autoregulatory response is impaired in rats with experARF [1]. Also, you cited evidence that a dysfunctional renal imental ARF [30] and has suggested, on the basis of these

vascular endothelium might be responsible for the increased risk findings, that the acutely injured kidney is more susceptible than of ARF in diabetic patients [39]. What factors might be respon- the normal kidney to repeated episodes of acute ischemic damage sible for such susceptibility in chronic renal insufficiency? Why do during hypotensive episodes. Also, Solez reported that renal

Nephrology Forum: Acute renal failure

tissue from patients with ATN showing areas of regeneration together with fresh lesions of acute ischeniia suggests that persistent renal failure in ATN is associated with repeated episodes of renal ischemia [58]. However, in answer to your question, I do not believe that we should delay dialysis, when indicated, in patients with ARF. The important question that remains to be answered

by the appropriate clinical studies is whether continous venovenous hemodialysis or other slow continuous therapies will prove

superior to conventional hemodialysis in treating patients with ARF.

1113

18. LEVY M, IMPERIAL E: Oxygen shunting in renal cortical and medullary capillaries. Am J Physiol 200:159—162, 1961 19. KOHAN D: Endothelins in the kidney: physiology and pathophysiology. Am J Kidney Dis 22:493—510, 1993 20. CHAN L, CHITFINANDANA A, SHAPIRO J, SHANLEY P, SCHRIER R:

Effect of an endothelin-receptor antagonist on ischemic acute renal failure. Am J Physiol 266:F135—F138, 1994 21. GELLAI M, JuciuS M, FLETCHER T, DEWOLF R, NAMBI P: Reversal of

postischemic acute renal failure with a selective eodothelin A

receptor antagonist in the rat. J Clin Invest 93:900—906, 1994 22. CtDZEL J: Role of endothelin during reperfusion after isehemia in isolated perfused rat kidney. J Cardiovasc Pharmacol 24:867— 874, 1994

Reprint requests to Dr. W. Lieberthal, Renal Section, Boston Medical Center, 88 East Newton Street, Boston, Massachusetts 02118, USA

NOTE ADDED IN PROOF The proteins belonging to the ICE-like family of proteases have recently been named "caspases" [1081.

REFERENCES 1. BRADY H, BRENNER B, LIEBERTHAI. W: Acute Renal Failure (5th ed), edited by BRENNER BM, Philadelphia, Saunders, 1996, pp 1200—1 252 2. BURKE TJ, CRONIN RE, DucuiN KL, PETERSON LN, SCHRIER RW:

lschemia and tubule obstruction during acute renal failure in dogs: mannitol in protection. Am J Physiol 238:F305—F314, 1980 3. Dortouou JF, VENKATACHALAM MA, BERNARD DB, LEVINSKY NG:

Tubular leakage and obstruction in acute ischemic renal failure. Kidney Tnt 13:208—222, 1978

4. BONVENTRE J: Nephrology Forum: Mechanisms of ischcmic acute renal failure. Kidney mt 43:1160—1178, 1993 5. MYERS BD, MORAN SM: Hemodynamically mediated acute renal failure. N Engl J Med 314:97—105, 1986 6. VENKATACHALAM M, BERNARD D, D0NOH0E J, LEVINSKY N: Isch-

cmic damage and repair in the rat proximal tubule. Differences among SI, S2 and S3 segments. Kidney mt 14:31—49, 1978 7. TORHORST J, Di ROUGEMENT D, BRtJNNER F, TUIEL G: Morphology

of the renal medulla in isehemic acute renal failure in the rat.

Nephron 31:296—300, 1982

8. BREZIS M, ROSEN S, SlivA P, EPSTEIN F: Renal isehemia: a new perspective. Kidney mt 26:375—383, 1984 9. LIEBERTHAL W, STEPHENS G, WOLF E, RENNKE H, VASILEVSKY M,

VALERI C, LEVINSKY N: Effect of erythrocytes on the function and morphology of the isolated perfused kidney. Renal Physiolory 20:14— 24, 1987 10. ENDRE Z, RATCLIFFE P, TANGE J, FERGUSON DJP, RADDA G, LEDINGHAM J: Erythrocytes alter the pattern of renal hypoxic injury:

Predominance of proximal tubular injury with moderate hypoxia.

23. Fooo A, HELLINGS S, INAGAMI T, K0N V: Endothelin receptor antagonism is protective in vivo in acute cyclosporine toxicity. Kidney mt 42:770—774, 1992 24. KIVLIGHN S, GABEL R, SIEGL K: Effects of BQ123 on renal function and acute cyclosporine-induced renal dysfunction. Kidney lot 45:131— 136, 1994 25. KARAM H, BRUNEVAL P, CLOZEL J-P, LOFFLER B-M, BARIETY J, CLOZEL M: Role of endothelin in acute renal failure due to rhabdomyolysis in rats. J Pharmacol Exp Ther 274:481—486, 1995 26. CANTLEY L, SPOKES K, CLARK B, MCMAHON E, CARTER J, EPSTEIN

F: Role of endothelin and prostaglandins in radiocontrast-induced renal artery constriction. Kidney mt 44:1217—1223, 1993 27. EM0RI T, HIRATA Y, SEGUCHI TI, KANNO K, MARUMO F: Cellular

mechanisms of natriuretic peptide-induced inhibition of endotheho-i biosynthesis in rat endothelial cells. Endocrinology 133:2474— 2480, 1993 28. IWASAKA 5, HOMMA T, KON V: Site specific regulation in the kidney of endothelin and its receptor subtypes. Kidney mnt 45:592—597, 1994 29. IWA5AKA S, HOMMA T, MATSUDA Y, KON V: Endothelin receptor

subtype B mediates autoinduction of endothelin-1 in rat mesangial cells. J Biol Chem 270:6997—7003, 1995 30. CONGER J, ROBINEITE J, SCHRIER R: Smooth muscle calcium and

endothelium-derived relaxing factor in the abnormal vascular responses of acute renal failure. J Clin Invest 82:532—537, 1988 31. BOSALLER C, FORTERMANN U, HETEL R, CRISTOPH 0, RESCHKE V,

FLECK E: yclosporine A inhibits endothelium-dependent vasodilation and vascular prostacyclin production. Eur J Pharmacol 165:165— 169, 1989 32. MOTTERLINI R, VANDERGRIFF K, WINSLOW R: Hemoglobin-nitric oxide interactions and its implications. Transfusion Med Rev 10:77— 84, 1996 33. CONGEK J, ROBINE1TE J, VIILAR A, RAIJ L, SCHULTZ P: Increased nitric oxide synthase activity despite lack of response to endothelium-

dependent vasodilators in postischemic acute renal failure in rats. J Clin Invest 96:631—638, 1995 34. DE NIc0LA L, THOMSON 5, WEAD L, BROWN M, GABBAI F: Arginine

feeding modifies cyclosporine ncphrotoxicity in rats. J Clin Invest

11. LIEBERTHAL W, RENNKE H, SANDOCK K, VALERI C, LEVINSKY N:

92:1859—1865, 1993 35. BAYLIS C, ENGELS K, SAMSELL L, HARTON P: Renal effects of acute

Tschcmia in the isolated-perfused rat kidney: protective effect of

endothelial-derived relaxing factor blockade are not mediated by

Clin Sci (London) 76:19—29, 1989

hypothermia. Renal Physiol Biochem 11:60—69, 1988 12. LIEHERTHAL W, WOLF EF, RENNKE HG, VALERI CR, LEVINSKY NO:

angiotensin II. Am J Physiol 264:F74—F78, 1993 36. KOUREMBANAS S, MCOIJILLAN L, LEUNG G, FALLER D: Nitric oxide

Renal ischemia and reperfusion impair endothelium-dependent vas-

regulates the expression of vasoconstrictors and growth factors by

cular relaxation. Am J Physiol 256:F894—F900, 1989 13. LIEBERTHAL W, SHERIDAN A, VALERI C: Protective effect of atrial

vascular endothehium under both normoxia and hypoxia. J Clin Invest

natriuretic factor and mannitol following renal ischemia. Am J

92:99—104, 1993 37. GOLIGORSKY M, TSUKAHARA H, MAGAZINE H, ANDERSEN T, MALIK

Physiol 258:F1266—F1272, 1990 14. WITGALL R, BROWN D, SCHWARTZ C, BONVENTRE J: Localization of

A, BAHOtJ W: Termination of endothelin signalling: role of nitric oxide. J Cell Physiol 158:485-494, 1994

proliferating cell nuclear antigen, vimentin, c-fos and clusterin in postischemic kidney. Evidence for a heterogeneous genetic response among nephron segments and a large pool of mitotically active and

38. THOMPSON A, VALERI C, LIEBERTIIAL W: Endothehin receptor .A

dedifferentiated cells. J Clin Invest 93:2175—2188, 1994 15. BREZI5 M, ROSEN S: Hypoxia of the renal medulla—its implications for disease. N Engi J Med 332:647—655, 1995 16. HELLBERG 0, KALLSKOG 0, WOLGAST M: Red cell trapping and postischemic renal blood flow. Differenees between the cortex, outer and inner medulla. Kidney mt 40:625—631, 1991 17. MASON J, TOIo10RST 1, WCISCH J: Role of the medullary perfusion

defect in the pathogenesis of isehemic renal failure. Kidney mt 26:283—293, 1984

blockade alters hemodynamic response to nitric oxide inhibition in the rat. Am J Physiol 269:H743—H748, 1995 39. AGMON Y, PELEG H, GREENFELD Z, Rosi S, BREZIS M: Nitric oxide

and prostanoids protect the renal outer medulla from radioeontrast toxicity in the rat. J Clin Invest 94:1069—1075, 1994 40. Yu L, GENGARO P, NIEDERBERGER M, BURKE T, SCHRIER R: Nitric

oxide: A mediator in rat tubular hypoxia/reoxygenation injury. Proc NatlAcad Sci USA 9 1:1691—1695, 1994 41. PERESLENI T, N0IRI E, BAHOU W, GOLIGORsKY M: Antisense

oligodeoxynueleotides to inducible NO synthase rescue epithelial cells from oxidative stress injury. Am J Physiol 270:F97l—F977, 1996

1114

Nephrology Forum: Acute renal failure

42. SZABO C, SOUTHAN 0, THIEMERMANN C: Beneficial effects and improved survival in rodent models of septic shock with S-methylisothiourea, a potent and selective inhibitor of inducible nitric oxide synthase. Proc Nati Acad Sci USA 91:12472—12476, 1994 43. N0IRI A, PERESLENI T, MILLER F, GOLIGORSKY M: In vivo targeting

of inducible NO synthase with oligonucleotides protects rat kidney

66. FISH

intravascular leukocyte accumulation in acute renal failure.

Kidney

mt 6:24—37,

45. MASON J,

WELSCI-I

obstruction

MOLITORIS

B: Alterations in epithelial polarity and the

Renal Failure, edited by LAZARUS J, BRENNER B, New York,

against ischemia. J Clin Invest 97:2377—2383, 1996 44. SOLEZ K, KRAMER E, Fox J, HEPTINSTALI. R: Medullary plasma flow

and

E,

pathogenesis of disease states. N EngI J Med 330:1580—1588, 1994 67. KELLERMAN P, BOGUSKY R: Microfilament disruption occurs very early in ischemic proximal tubule cell injury. Kidney mt 42:896—902, 1992 68. M0LIT0RIS B: Cellular basis of ischemic acute renal failure, in Acute

Churchill Livingstone, 1993, pp

1—32

69. GAILIT J, COLFLESH D, RABINER

1974 J, TORHORST J: The contribution of vascular

1,

SIMONE J, GOLIGORSKY M:

and dysfunction of integrins in cultured renal epitheexposed to oxidative stress. Am J Physiol 264:F149—F157,

Redistribution

hal

to the functional defect that follows renal ischemia.

cells

1993

Kidney mt 31:65—71, 1987 46. HELLBERG P, KALLSKOG 0, OJTEG G, WOLGAST M: Peritubular capillary permeability and intravascular RBC aggregation after ischcmia: effect of neutrophils. Am J Physiol 258:F1018—F1025, 1990 47. KLAUSNER J, PATERSON I, GOLDMAN G, KOBZIK L, RODZEN C, LAWRENCE R, VALERI C, SHEPRO D, HECHTMAN H: Postischemic

70.

GOLIGORSKY M, LIEBERTHAL W, RACUSEN L, SIMON E: Integrin

71.

receptors in renal tubular epithelium: new insights into pathophysiology of acute renal failure. Am J Physiol 264:F1—F8, 1993 NOIRI E, ROMANOV V, FOREST T, GAILIT J, DIBONA G, MILLER F,

Physiol 256:F794—F802, 1989 48. BRADY H: Leukocyte adhesion molecules and kidney diseases. Kidney mt 45:1285—1230, 1994 49. RABB H, MEDNIOLA C, DIETZ J, SABA S, ISSEKUTZ T, ABANILLA F,

72. LIEBERTHAL W, MCKENNEY J, KIEFER C, SNYDER L, KROSHIAN V, SJAASIAD M: integrin-mediated adhesion between renal tubular

Role of CD11a and CD11b in ischcmic

KESSLER H, GOLIGORSKY M: Cyclic RGD peptides ameliorate isch-

SOM P, OSTER Z, GOLIGORSKY M: Pathophysiology of renal tubular

obstruction: Therapeutic role of synthetic RGD peptides in acute renal failure. Kidney mt 48:1375—1385, 1995

renal injury is mediated by neutrophils and leukotrienes. Am J

BONVENTRE J, RAMJREZ G:

acute

J Physiol 267:F1052—F1058, 1994 50. KELLY K, WILLIAMS W, COLVIN R, BONVENTRE J: Antibody to renal failure in rats. Am

intercellular adhesion molecule-i protects against ischemic injury. Proc NatlAcad Sci USA 91:812—816, 1994 51. RABB H: Cell adhesion molecules and the kidney. Am J Kidney Dis 23:155—165,

1994

52. LINAS S, WHITrENBURG D, PARSONS P, REPINE J: Isehemia increases

neutrophil retention and worsens acute renal failure: role of oxygen metabolites and ICAM 1. Kidney Int 48:1584—1591, 1995 53. KELLY K, WILLIAMS W, COLVIN R, MEEHAN S, SPRINGER T, GurrlERREZ-RAMOS J-C, BONVENTRE J: intercellular adhesion molecule-

cells following anoxic injury. JAm Soc Nephrol 8:175—183, 1997 73. Norm E, GAILIT J, SHETH D, MAGAZINE H, GURRATH M, MULLER C, emie acute renal failure in rats. Kidney Int 46:1050—1058, 1994 74. GOLIGORSKY M, DIBONA G: Pathogenetic role of Arg-Gly-Asp recognizing integrins in acute renal failure. Proc Soc Nati Acad Sci USA 90:5700—5704, 1993 75. WEINBERG JM: The cell biology of ischemic renal injury. Kidney Int 39:476—500, 1991 76. HOCKENBERY D: Defining apoptosis. Am J Physiol 146:16—19, 1995 77. KERR J, WYLLIE A, CURRIE A: Apoptosis: a basic biologic phenom-

enon with wide ranging implications in tissue kinetics. Br J Cancer

26:239—257, 1972 78. LIEBERTHAL W, TRIACA V, LEVINE J: Mechanisms of death induced

1-deficient mice are protected against ischemic renal injury. J Clin Invest 97:1056—1063, 1996 54. ESPINOZA G, FARRE A, CERNADAS M, MANZARBEITIA F, TAN D, DIGuINI E, MOSQUERA J, MONTON M, MILLAS 1, HERNADO L, CASADO S, CARAMELO C: Role of endothelin in the pathophysiology

of renal-ischemia reperfusion in normal rabbits. Kidney

mt 50:776—

782, 1996 55.

by cisplatin in proximal tubular epithelial cells: apoptosis versus necrosis. Am J Physiol 270:F700—F708, 1996 79. DIVE C, HICKMAN

J:

Drug-target interactions: only the first step in

the

committment to a programmed cell death? BrJ 196, 1991

80.

RAFF M:

Social

Cancer 64:192—

controls on cell survival and cell death. Nature

1992 81. LIEBERTHAL W, TRIACA V, BRADY B, LEVINE J: Apoptosis of renal 356:397—400,

KUBES P, KANWAR S, NIu X-F, GABOURY J: Nitric oxide synthesis

J, ZAOUI P, HAKIM R: Modulation of granulocyte LAM-1 and Mac-I during dialysis: a prospective randomized control

epithelial cells: role of serum deprivation and oxidant injury (abstract). JAm Soc Nephrol 5:902, 1994 82. SAFIRSTEIN R: Gene expression in nephrotoxic and isehemic acute renal failure. JAm Soc Nephrol 4:1387—1395, 1994

trial.

83. PRICE P, MEGYESI J, SAGGI S, SAFIRSTEIN R: Regulation of transcrip-

inhibition induces leukocyte adhesion via superoxide and mast cells. FASEB J 7:1293—1299, 1993 56. HIMMELFARB

Kidney mt 41:388—395,

1992

57. HAKIM R, WINGARD R, PARKER R: Effect of the dialysis membrane

in the treatment of patients with acute renal failure. N EngI J Med 331:1338—1342, 1994 58. SOLEZ K, MOREL-MAROGER L, SRAER JD: The morphology of "acute

tubular necrosis" in man: analysis of 57 renal biopsies and a comparison with the glycerol model. Medicine 58:362—376, 1979 FIvuSI-1 B, LI U-L, SLATNIK I, SOLEZ K: Dissociation of tubular cell detachment and tubular cell death in clinical and experimental "acute tubular necrosis." Lab Invest 64:546—556, 1991 60. LIEBERTHAL W, LIVINE J: Mechanisms of apoptosis and its potential role ill renal tubular epithelial cell injury. Am J Physiol 271:F477— F488, 1996 61. MoIJT0RIS BA: lschemia-induced loss of epithelial polarity: potential role of the actin cytoskeleton. Am J Physiol 260:F769—F778, 1991 62. KROSHIAN VM, ShERIDAN A, LIEBERTHAL W: Functional and cy-

59. RACUSEN L,

toskeletal

changes induced by sublethal injury in proximal tubular

epithelial

cells. Am

J Physiol 266:F21—F30, 1994

B: Structure, biochemistry and assembly of epithelial junctions. Am J Physiol 253:C749—C758, 1987 64. BURRIDGE K, FATH K, KElLY T, NUCKOLI.S C, TURNER C: Focal adhesions: transmemhranc junction between the extracellular matrix and the cytoskeleton. Annu Rev Cell Biol 4:485—525, 1988 63. GUMBINER tight

65. M0uT0ISIs B: Na/K-ATPase that redistributes to apical membrane during ATP depletion remains functional. Am J Physiol 265:F693—F697, 1993

tion by the rat EGF gene promoter in normal and isehemic murine kidney cells. Am J Physiol 268:F664—F670, 1995 84. GAVRIELI Y, SHERMAN Y, BEN-SASSON 5:

Identification of pro-

grammed cell death in situ via specific labeling of nuclear DNA fragmentation.

J Cell Biol 119:493—501, 1992

85. UEDA N, WALKER P, Hsu S, SHAH 5: Activation of a 15-kDa endonuclease in hypoxia/reoxygenation injury without morphologic features of apoptosis. Proc Nat! Acad Sci USA 92:7202—7206, 1995 86. MARTIN S, GREEN D: Protease activation during apoptosis: death by a thousand cuts? Cell 82:349—352, 1995 87. PATEL T, GORES U, KAUIMANN S: The role of proteases during apoptosis. IASEB 10:587—597, 1996 88. MARTIN S, GREEN D, COTTER T: Dicing with death: dissecting the components of the apoptosis machinery. Trends Biochem Sci 19:26—

30, 1994 89. WHYTE M:

ICE/CED-3 proteases in apoptosis.

Trends

Cell Biol

1996 90. FEARNHEAD H, DINSDAI.E D, COHEN C: An interleukin-1f3-convert6:245—248,

ing enzyme-like protease is a common mediator of apoptosis in thymoeytcs. FEllS Lert 375:283—288, 1995 91. JACOLI50N M, WElL M, RAFF M: Role of Ced-3/ICE-family proteases in staurospaurine-induccd programmed cell death. J Cell Biol 133: 1041—1051, 1996

92. OLTAVI Z, KORSMEYER S: Checkpoints of dueling dimers foil death wishes. Cell 79:189—192, 1994

Nephrology Fontm: Acute renal failure 93. SCI-IUMER M, COLOMBEL M, SAWCHUK T, GOBE G, CONNOR J,

O'TooLu K, OLLSON C, WISE G, BUT-rYAN R: Morphologic, biochemical, and molecular evidence of apoptosis during the reperfu-

sion phase after brief periods of renal isehemia. Am J Pathol 140:831—838, 1992

1115

cAMP response elements and to an AP-1 binding site from the c-jun promoter. J Biol Chem 270:30084—30092, 1995 101. HUMES H, CIESLINSKI D, COIMBRI T, MESSANA J, GALVAO C: Epidermal growth factor accelerates renal generation and repair and recovery of renal failure. J Clin Invest 84:1757—1761, 1989

94. SHIMIZU A, YAMANAKA N: Apoptosis and cell desquamation in

102. DING H, KOPPLE J, HIRSCHBERG R: Recombinant human insulin-like

repair process of ischemic tubular necrosis. VirchowsArch [B]64:171— 180, 1993 95. GOTrLIEB R, BURLESON K, KLONER R, BABIOR B, ENGLER R: Reperfusion injury induces apoptosis in rabbit cardiomyocytcs. J Clin Invest 94:1621—1628, 1994

103. KAWAIDA K, MArsuMoTo K, SHIMAZU H, NAKAMURA T: Hepato-

96. TWATA M, MYERSON D, TOROK-STORB B, ZAGER R: An evaluation of

renal tubular DNA laddering in response to oxygen deprivation and oxidant injury. JAm Soc Nephrol 5:1307—1313, 1994 97. LIEBERTHAL W, TRIACA V, LEVINE J: Mitochondrial inhibitors induce apoptosis in renal tubular epithelial cells (abstract). JAm Soc Nephrol 3:983, 1995 98. SAFIRSTEIN R, PRICE P, SAGGI S, HARRIS R: Changes in gene expression after temporary renal ischemia. Kidney mt 37:1515—1521,

growth factor-i accelerates recovery and reduces catabolism in rats with ischemie acute renal failure. J Clin Invest 91:2281—2287, 1993

cyte growth factor prevents acute renal failure and accelerates renal regeneration in mice. Proc NatlAcad Sci USA 91:4357—4361, 1994 104. HAMMERMAN M, MILLER S: Therapeutic use of growth factors in renal failure. JAm Soc Nephrol 5:1—11, 1994 105. KING A: Endothelins: Multifunctional peptides with potent vasoactive properties (chap 38), in Hypertension, Pathophysiology, Diagnosis, and Management (vol 1), edited by LARAGH JH, BRENNER BM, New York, Raven, 1995, pp 631—672 106. COLLINS D, KLOTMAN P: Renin-Angiotensin System and Arachidonic

Acid Metabolites in Acute Renal Failure (3rd ed). New York,

1990 99. SAFIRSTEIN R, MEGYESI J, SAGGI 5, PRICE P, PooN M, ROLLINS B,

Churchill-Livingston, 1993, pp 69—106 107. DENTON M, CHERTOW G, BRADY H: "Renal-dose" dopamine for the

TAUBMAN M: Expression of cytokine-like genes JE and KC is increased during renal ischemia. Am J Physiol 261:F1095—F1101,

treatment of acute renal failure: scientific rationale, experimental

1991

100. MoRoo H, BONVENTRE J, Porvtno C, KYRIAKIS J, FORCE T:

Isehemia and reperfusion enhance ATF-2 and c-Jun binding to

studies and clinical trials. Kidney mt 49:4—14, 1990 108. ALNEMRI ES, LIvINGsroN DJ, NICHOLSON DW, SALVESEN G, TIIORNBERRY NA, WONG WW, YUAN J: Human ICE/CED-3 protease nomenclature. Cell 87:171, 1996