Acute kidney injury (AKI) is defined as a sudden decrease in the glomerular filtration rate (GFR) occurring over a period of hours to days. The Acute Dialysis Quality Initiative (ADQI) has developed the RIFLE classification of AKI that divides AKI into the following stages: (1) risk, (2) injury, (3) failure, (4) loss of function, and (5) end-stage kidney disease.¹,²,³ The RIFLE criteria have been validated in multiple studies (i.e., as the RIFLE class increases so does mortality).¹,²,³ The term AKI, in general, replaces the term acute renal failure (ARF) and ARF is restricted to patients that have AKI and need renal replacement therapy. AKI has also replaced the term acute tubular necrosis (ATN). The term ATN emerged from the early observation on renal biopsy that necrosis of some renal tubular epithelial cells may occur in humans with acute renal injury.⁴ Tubular epithelial casts (muddy brown casts) are excreted in the urine of these patients. However, it is now known that the tubular necrosis is quite patchy and alone could not account for GFR less than 10 mL/min/1.73m², the functional hallmark of clinically significant AKI.⁴ Moreover, a percentage, ranging from 30% to 70%, of the urinary tubular epithelial cells has been shown to be viable by culture and the exclusion of vital dyes.⁵,⁶ This observation is somewhat surprising because, generally, cells, which are separated from their extracellular matrix, undergo apoptosis.⁷,⁸,⁹,¹⁰ Emerging results suggest that adhesion molecules (e.g., cadherins and integrins) may allow a cell-to-cell or a cell-to-matrix adhesion that not only avoids apoptosis, but also may contribute to intratubular obstruction.¹¹,¹²,¹³,¹⁴ Intraluminal tubular casts on a renal biopsy are a hallmark of clinical AKI, and earlier nephron dissection studies by Jean Oliver demonstrated a preferential location of these casts in the medullary collecting duct.¹⁵ This location is of particular relevance to the overall low GFR in AKI because thousands of nephrons drain into a single medullary collecting duct.

There is debate about whether the pathophysiology of clinical AKI is primarily tubular or vascular. In fact, initial rat micropuncture studies were unable to consistently detect an elevation in tubular pressures in experimental ARF and thus the term vasomotor nephropathy replaced the ATN term for the clinical syndrome.¹⁶,¹⁷,¹⁸ Experimental results have emerged that both vascular and tubular factors are involved in the pathogenesis of clinical AKI. Most recently, the role of endothelial injury and dysfunction¹⁹ in promoting an inflammatory response in AKI²⁰ has received prominence. Thus, this chapter will discuss the potential tubular and vascular factors, as well as inflammatory processes, involved in the pathogenesis of ischemic AKI (Fig. 29.1). Toxic AKI will be discussed in another chapter. However, it must be emphasized that AKI in humans is frequently multifactorial. Both ischemic and toxic insults combine in a synergistic fashion to cause clinical AKI.

The understanding of the pathogenesis of ischemic AKI is of considerable importance for several reasons. First, this clinical syndrome is quite frequent, occurring in 5% to 10% of hospitalized patients and in 30% to 40% of intensive care unit (ICU) patients.²¹,²²,²³ The incidence is likely to increase in the future because of the use of newer nephrotoxic drugs and the performance of more complex procedures in older patients.²⁴ Second, ischemic AKI has a very high mortality particularly when requiring dialysis to treat the resultant uremic syndrome. Overall mortality averages from 40% to 50%; however, the patient in the ICU with ischemic AKI may have a mortality in excess of 80%, particularly if the patient has multiorgan failure.²⁵,²⁶,²⁷,²⁸,²⁹,³⁰ It is widely quoted that the mortality of AKI has only improved slightly in the last 40 years.³¹ However, a study³² suggests there has been improvement in AKI mortality between the late 1970s and the early 1990s. Third, there is considerable evidence that a functional component of the renal failure exists. Specifically, a histologic examination of the kidney from patients with clinical AKI exhibits normal glomeruli, occasional tubular necrosis, some intraluminal casts, and modest interstitial edema.⁴ There is virtually no evidence for irreversible tissue damage and the morphologic changes alone fail to support the presence of a GFR less than 10 mL/min/1.73m².

Available models to study the pathophysiology of renal cell ischemia are listed in Table 29.1.³³ An understanding of these models will allow a better interpretation of the multiple studies discussed in this chapter.

Proximal and distal tubular cells in culture have been widely used to study tubular injury. These cells change from their normal dependence on oxidative mitochondrial metabolism to glycolysis under culture conditions.³⁴ As a result, these cultured tubules become less susceptible to oxygen deprivation. Thus, exposure to drugs like antimycin-A, ionomycin, or a combination to induce “chemical” ATP depletion and subsequent necrosis or apoptosis is used. Cultured cells also undergo considerable structural change, which includes simplification of both their apical and basolateral compartments.³⁵ The presence of necrosis rather than apoptosis in these cells may be related to the level of ATP depletion. In cultured mouse proximal tubules subjected to ATP depletion below 15% of control values, the cells died of necrosis; whereas in ATP depletion to 25% to 70% of control values all the cells died of apoptosis.³⁶ Therefore, although cultured tubule cells are the least complex model and allow for an understanding of the mechanisms involved, the therapeutic implications for in vivo AKI are limited.

Freshly isolated rat or rabbit proximal tubules in suspension are also widely used to study proximal tubular injury.³⁷,³⁸,³⁹,⁴⁰,⁴¹,⁴²,⁴³,⁴⁴,⁴⁵ The method of isolation of tubules is by collagenase digestion and Percoll centrifugation. Hypoxia is achieved by gassing the suspension with 95 %N₂/5%CO₂ for up to 15 minutes, thereby reducing the pO₂ to approximately 30 mm Hg. Lactic dehydrogenase (LDH) release into the suspension medium is measured as an index of lethal membrane injury.⁴⁶,⁴⁷ The tubules are preincubated with cytoprotective agents and enzyme inhibitors before the induction of hypoxia, and the effect of these agents on cell membrane injury can be determined. The presence of necrosis rather than apoptosis during short periods of hypoxia (15 to 30 minutes) in this model was demonstrated using DNA-specific dyes, such as Hoechst 33342 and propidium iodide.⁴⁸ Another study has also demonstrated that during hypoxia there is endonuclease activation without morphologic features of apoptosis in the same model of rat proximal tubules.⁴⁹ Freshly isolated tubules are valuable for both structural and metabolic investigations because they retain the biochemical properties of the in vivo state and a high degree of structural integrity, and are highly polarized and fully differentiated.³³ However, the tubules are highly sensitive to ATP depletion, and severe hypoxia or anoxia results in necrosis of more than 50% of the cells after 30 minutes. Isolation methods also expose the tubules to repeated 4°C exposure and collagenase.

In whole animal studies—usually rats, rabbits, or mice —a clamp model of ischemic AKI is used.⁵⁰,⁵¹,⁵² Ischemic AKI is generally induced by (1) clamping of both the right and left renal pedicles or renal arteries or (2) unilateral renal pedicle or artery clamp preceded by contralateral nephrectomy. The renal vessels are clamped for varying periods of time, generally from 30 to 60 minutes, followed by varying periods of reperfusion. This results in a reversible model of AKI in which the blood urea nitrogen (BUN) and serum creatinine reach a peak at 24 to 48 hours reperfusion and
then gradually normalize over the next 7 days.⁵⁰,⁵¹,⁵²,⁵³ However, renal vessel clamping in rats results in extensive necrosis of proximal tubules. This necrosis is much more extensive than is seen in humans with ischemic AKI. Nevertheless, although animal models of ischemic AKI are complex with many experimental limitations, they provide important leads for future therapeutic clinical interventions.

In the isolated perfused kidney model, the kidney is removed from the animal. The perfusate usually consists of a Krebs-Henseleit buffer with albumin. Urine is collected by cannulation of the ureter. This model allows for the study of factors independent of changes in systemic hemodynamics and neural activity. Further advantages include the study of specific circulatory factors or pharmacologic agents that are added to the perfusate. These agents are thus delivered directly to the kidney. The disadvantages of the model are (1) the absence of red blood cells in the perfusate impairs oxygen delivery to the medullary thick ascending limb (mTAL) and (2) perfusate flow greatly exceeds normal in vivo values. The isolated perfused kidney is regarded as a model of selective hypoxia to the medullary thick ascending limb.⁵⁴

Studies in patients with AKI, although having important experimental limitations, have the most direct therapeutic value. An analysis of urine cytology represents a noninvasive method for potentially defining the cause of AKI.⁴ Meyers and colleagues⁵⁵,⁵⁶,⁵⁷ have examined patients with ischemic AKI postrenal transplantation by obtaining biopsies of these allografts at the time of transplantation. Biomarkers of kidney injury would greatly facilitate the early detection and the precise diagnosis of AKI.

AKI is usually diagnosed by recording increases in serum creatinine and decreased urine output over several days. However, serum creatinine is not a good marker of renal function in AKI because its concentration can be affected by factors not related to renal function such as the volume of distribution, muscle mass, and creatinine secretion.²⁸ When the kidney is injured and the true GFR suddenly drops, but there is a slow increase in serum creatinine over days. This new steady state, which may take up to 7 days, is reached when creatinine generation equals creatinine excretion. In contrast to serum troponin in myocardial infarction, an increase in serum creatinine lags and may not be directly related to tubular injury in AKI. Recent studies have examined urine and serum biomarkers of kidney injury that have the potential to facilitate the diagnosis of AKI.

Recently described molecules such as the cytokine interleukin (IL)-18,⁵⁸ kidney injury molecule-1 (KIM-1),⁵⁹ cysteine-rich protein 61 (Cry61),⁶⁰ neutrophil gelatinase-associated lipocalin (NGAL),⁶¹ and sodium/hydrogen exchanger isoform 3 (NHE3)⁶² have demonstrated compelling results as markers of AKI at the preclinical level. Studies have been initiated to explore these molecules in human AKI.

Because of the crucial importance of early therapy in the management of AKI, markers are being explored for early diagnosis. NGAL was investigated as an early biomarker for AKI following cardiopulmonary bypass in 45 patients.⁶³ Urine and serum were collected at baseline and at frequent intervals for 5 days following a cardiopulmonary bypass. All patients who developed AKI (defined as a 50% increase in serum creatinine) displayed a significant increase in serum and urine NGAL very early after the cardiopulmonary bypass compared to patients without AKI. These results show that NGAL may be a sensitive, early urinary, and serum biomarker for AKI.

In a nested case-control study within the adult respiratory distress syndrome (ARDS) network trial, urinary IL-18 was investigated as an early marker of AKI.⁶⁴ Median urine IL-18 levels were significantly higher in AKI cases (defined as a 50% increase in serum creatinine) as compared to controls. On multivariable analysis, urine IL-18 values predicted the development of AKI 24 and 48 hours later after adjusting for demographics, sepsis, acute physiology and chronic health evaluation (APACHE) III score, serum creatinine, and urine output. After controlling for other parameters, a rise in urine IL-18 by 25 pg per milliliter was associated with an increased odds ratio of AKI by 19% for the next 24 hours. Urine IL-18 performed as a diagnostic test with an area under the receiver operator characteristic curve of 73%. The conclusion of this study is that urinary IL-18 levels can be used for the early diagnosis of AKI.

A recent study demonstrated that serum cystatin C appears to increase 24 to 48 hours before serum creatinine in patients with AKI.⁶⁵ However, cystatin C is a marker of GFR, or a functional marker, and is not a structural or biochemical marker of renal tubular injury.

There are multiple promising serum and urinary biomarkers (eg, IL-18, neutrophil-gelatinase-associated lipocalin [NGAL], kidney injury molecule-1 [KIM-1], cystatin C, liver fatty acid-binding protein [L-FABP]), which detect AKI before the rise in serum creatinine and predict outcomes in patients with AKI.⁶⁶ Prospective studies to determine the use of these biomarkers in larger populations have been initiated. In this regard, a National Institutes of Health (NIH)-funded clinical consortium consisting of investigators from nine academic centers called TRIBE-AKI (Translational Research Investigating Biomarkers in Early Acute Kidney Injury) has been established. Currently, the consortium is performing a prospective multicenter observational cohort study of 1,800 patients receiving cardiac surgery to determine whether urine IL-18, urine NGAL, and serum cystatin C are biomarkers for the early diagnosis and long-term outcomes of AKI. Ultimately, disease control studies to determine the impact of a biomarker screening on AKI morbidity and mortality are desirable. In this regard, a prospective study in over 500 ICU patients in New Zealand determined whether erythropoietin therapy decreases the incidence of AKI, as determined by serum creatinine and serum cystatin C, and lowers levels of urinary IL-18, NGAL, and KIM-1.⁶⁷ In this
study, early intervention with erythropoietin based on urine biomarker levels did not affect the outcome of AKI.⁶⁷

It has been proposed that GFR in ischemic AKI is really not as low as measured because glomerular filtrate leaks across the damaged epithelial monolayer and/or the tubular basement membranes. In some toxic experimental models of ischemic AKI, diffuse tubular necrosis and basement membrane damage has been associated with evidence for back-leak of glomerular filtrate. The term “back-leak” of glomerular filtrate refers to the unregulated passage of salt and water from the tubular lumen into the interstitium and later back into the renal venous capillaries and renal veins.⁶⁸ However, the level of epithelial and basement membrane damage with these experimental toxic models (e.g., cisplatin, mercuric chloride) is virtually never observed in human AKI. Myers et al.⁶⁹ have performed human studies using solute sieving curves in search of tubular back-leak of glomerular filtrate. They found dextran sieving curves could sometimes exceed inulin sieving curves, thus providing evidence in support of the back-leak of solutes (i.e., dextran), which are normally unable to cross intact tubular epithelial basement membranes. However, even when accepting the validity of this method for documenting tubular back-leak of filtrate, the calculated amount would only account for a decrease in renal function of 8% to 10%. Thus, although tubular back-leak of glomerular filtrate might occasionally occur in patients with severe ischemic AKI, it is unlikely to be a dominant pathogenic factor.

The tight junction of polarized tubular epithelial cells is the most apical component of the junctional complex and serves as an important permeability barrier.⁷⁰ Tight junctions also control cell polarity.⁷¹ However, the tight junction in the proximal tubule is relatively “leaky” with as much as one-third of proximal sodium reabsorption occurring via the paracellular route. The tight junctional complex is a dynamic and regulated structure. Some of its protein components have been identified and include the transmembrane protein occludin. Nontransmembrane proteins on the cytosolic leaflet include zona occludens (ZO)-1, ZO-2, cingulin, 7H6, and several unidentified phosphoproteins. Interactions of some of these proteins with the actin cytoskeleton is a major determinant of the tight junction structure and may also play a role in the regulation of tight junction assembly.⁷⁰ The integrity of the tight junction is disrupted during ischemic injury and must be reestablished for recovery.⁷² There is in vitro experimental evidence in cell culture studies for an impaired tight junction between tubular epithelial cells undergoing chemical anoxia.⁷³ Ruthenium red, which normally is impermeable to tight junctions, has been shown to enter the ZO after chemical hypoxia and renal ischemia.⁷³,⁷⁴ Energy depletion abolishes the gate function of the tight junction, as determined by the dramatic decrease in transepithelial resistance, but it leaves the fence function intact, as determined by the maintenance of lipid polarity.⁷⁵ In an ATP depletion-repletion model in Madin-Darby canine kidney cells, tight junction proteins such as ZO-1 reversibly form large complexes and associate with cytoskeletal proteins.⁷⁶ A model has been proposed in which a key, potentially regulated step in the generation of the ischemic epithelial cell phenotype is the interaction between tight junction proteins and fodrin and/or other cytoskeletal proteins.⁷⁶ Intracellular calcium plays a role in tight junction reassembly after ATP depletion.⁷² In this study, the role of intracellular calcium in tight junction reassembly after ATP depletion-repletion was studied using the cell-permeable calcium chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N’,N’-tetra-acetic acid-AM (BAPTA-AM). Lowering intracellular calcium during ATP depletion was associated with a significant inhibition of the reestablishment of the permeability barrier following ATP repletion as measured by transepithelial electrical resistance and mannitol flux, marked alterations in the subcellular localization of occludin by immunofluorescent analysis, and decreased solubility of ZO-1, and other tight junction proteins by the Triton X-100 extraction assay. This suggested that lowering intracellular calcium potentiates the interaction of tight junction proteins with the cytoskeleton.

Studies have also shown the importance of small GTPases (Rho, Rac, Cdc-42) in the integrity of the ZO.⁷⁷ ATP depletion with chemical hypoxia has been demonstrated to inactivate these GTPases and thereby contribute to increased paracellular back-leak of filtrate.⁷¹,⁷⁸ Expression of constitutively active ras homolog gene family, member A (RhoA) GTPase in ATP-depleted Madin Darby canine kidney (MDCK) cells prevents tight junction disassembly.⁷¹ Cyanide-induced chemical hypoxia increases the kinase activity of c-Src and causes its translocation to cell-cell junctions where it binds to and phosphorylates β-catenin and p120, suggesting that this may contribute to the loss of epithelial barrier function.⁷⁹ It is likely that the effects on tight junction integrity seen in ATP depletion are due, at least in part, to the inhibition of Na-K-ATPase.⁸⁰ The relevance of these cell culture observations to clinical ischemic AKI in patients remains to be proven.

Proximal tubular injury, whether it be sublethal reversible dysfunction, necrosis, or apoptosis, has been extensively studied. Mechanisms of proximal tubular injury that will be discussed in this section are outlined in Table 29.2. The study of proximal tubular injury is of special relevance in order to explain the decreased tubular sodium reabsorption that occurs with a postrenal ischemia.

In the normal kidney, Na⁺ is vectorially transported from the proximal tubule lumen across the apical membrane microvilli into the tubular epithelial cells and then across the basolateral membrane into the interstitium and the peritubular circulation.⁸¹ Na⁺ influx into the polarized proximal tubular epithelial cells across the apical membrane is passive
down the Na⁺ gradient via the H⁺/Na⁺ exchanger and various Na⁺ cotransporters. The Na⁺ gradient is maintained by an active transport via the Na⁺/K⁺-ATPase at the basolateral membrane of the proximal tubular cells (Fig. 29.2A).

The earliest signs of clinical AKI are urinary muddy brown casts and an increased fractional excretion of sodium (FE_Na).⁸²,⁸³ The proximal tubule is the most frequent morphologic site of injury in ischemic AKI in both humans and animals.⁸⁴ The S₃ segment of the proximal tubule is particularly prone to ischemic injury, perhaps because of its location in the outer medulla, which is relatively hypoxic compared to the renal cortex.⁸⁴ The proximal tubule nephron site is also associated with impaired vectorial sodium transport. The earliest morphologic changes with ischemic injury include invagination and sloughing of the brush border membrane into the lumen, an abnormality that is compatible with impairment and a loss of apical sodium antiporters and cotransporters responsible for sodium entry into the proximal tubular epithelium.⁸⁵,⁸⁶,⁸⁷ The tubules lose their polarity.⁸³ In vitro studies have shown that ATP depletion leads to dephosphorylation and inactivation of the actin binding protein, ezrin, and activation of the actin depolarizing protein in the proximal tubule membrane.⁸⁸,⁸⁹ This leads to a disruption of the microvillar actin and a loss of the brush border membrane.⁹⁰ A loss of polarity of the proximal tubule cells during chemical anoxia and ischemia has also been shown with the translocation of the Na-K-ATPase to the apical membrane (Fig. 29.2B).⁹¹,⁹²,⁹³ The translocated Na-K-ATPase remains functional.⁹⁴ In MDCK cells exposed to ATP depletion, there is a loss of polarity of Na-K-ATPase and a dissociation of the membrane-cytoskeleton complex at the spectrin-ankyrin interface.⁹⁵ Thus, sodium transport back
across the apical membrane and into the proximal lumen, as well as decreased proximal tubular sodium reabsorption, has been proposed in response to hypoxia and ischemia.

Studies in cadaveric transplanted kidneys with both prompt and delayed graft functions have been compared relative to the cellular location of the actin binding proteins, ankyrin and spectrin, and Na-K-ATPase using selective antibodies. In kidneys with delayed graft function, approximately 50% of the ankyrin, spectrin, and Na-K-ATPase were translocated from the basolateral membrane to the cytoplasm (Fig. 29.3). Those kidneys with prompt graft function had only minimal translocation of these proteins from the basolateral membrane (Fig. 29.3).⁵⁵ These observations therefore have contributed to our understanding of reversible and sublethal tubular dysfunction in ischemic kidneys in vivo (Fig. 29.3).

Because proximal tubular cells in a culture convert from primarily oxidative to glycolytic metabolism and alter their phenotype,³⁴,³⁵ confirmation of experimental results in other systems is advisable. The use of freshly isolated proximal tubules to study the response to hypoxia has also been enlightening. These tubules, in general, maintain their phenotype and oxidative metabolism but are more sensitive to hypoxia, as assessed by LDH release, than in vivo tubules.⁹⁶,⁹⁷,⁹⁸,⁹⁹ Hypoxia for 15 to 30 minutes causes a reproducible release of LDH, and the cells die by necrosis.⁴⁸,⁴⁹ The central role of intracellular calcium and various protective maneuvers against hypoxic injury have been demonstrated in these isolated proximal tubules. However, before discussing the role of intracellular calcium in proximal tubular injury, we shall briefly consider adenine nucleotides. It is unquestioned that the first effect of ischemia, hypoxia, or mitochondrial inhibition in most in vitro and in vivo models is to compromise adenine nucleotide metabolism. A decreased production of ATP precedes the increase in intracellular calcium.