Pathophysiology of Acute Kidney Injury

Classifications and Definitions

Acute Kidney Injury (AKI) is a heterogenous syndrome defined by a rapid decline in glomerular filtration rate (GFR) which may lead to accumulation of metabolic waste products and disturbances in fluid, electrolyte and acid-base handling. In the last few years, the previous terminology of Acute Renal Failure (ARF) has largely been abandoned and the new term AKI has been widely adopted in the medical literature. In this chapter, the term AKI will be used to describe the entire spectrum of disorders that share the same physiologic characteristics but may be pathophysiologically distinct.

In 2002, based on the recommendations of the Acute Dialysis Quality Initiatives (ADQI) the staging of AKI was proposed. The RIFLE classification of AKI is divided into three levels of renal dysfunction namely, “ r isk”, “ i njury” and “ f ailure” based on either GFR or urine output criteria whichever is more severe. The GFR criteria assesses renal dysfunction by using either changes in serum creatinine or the percentage decline in GFR from baseline GFR. The parameters used have a high sensitivity, are prognostic based on severity of stage, but represent a retrospective analysis. Two other very important clinical outcomes are also incorporated into the RIFLE criteria: “ l oss” of kidney function as defined by persistent AKI with the need for renal replacement therapy (RRT) for more than four weeks, and “ e nd” stage kidney disease, defined as the need for RRT for greater than three months. These parameters are more strict and therefore have a higher specificity.

Recently, the Acute Kidney Injury Network (AKIN) proposed a modification of the RIFLE criteria with the addition of a>= 0.3 mg/dl increase in serum creatinine to the criteria that define Risk. Because of the linear relationship between GFR and 1/serum creatinine, it should be kept in mind that a small rise in serum creatinine, in the normal or near normal values, corresponds to relatively large decline in GFR. For example, in a steady state, a serum creatinine of 0.8 mg/dl in a young 50 kg female corresponds to a GFR of approximately 100 ml/min, while a serum creatinine of 1.6 mg/dl would reflect in a GFR of 50 ml/min, a decline of nearly 50% in GFR. In the non-steady state of AKI, the serum creatinine value should be considered very carefully, as this is not an equilibrium state. Thus an overestimation of the GFR is made if standard steady state calculations are made using formulas such as Cockroft-Gault or MDRD if the serum creatinine is rising. Hence during the evolution of acute renal dysfunction, serum creatinine underestimates the degree of renal dysfunction, and conversely it underestimates the degree of renal recovery as function recovers. Having said that, the change from baseline and perhaps the rate of rise of creatinine are still potentially useful in the clinical evaluation of GFR decline. Even creatinine clearance (when collection is accurate) suffers from the same limitation as mentioned above, and causes an overestimation of GFR. Determining an accurate GFR in clinical settings is rarely accomplished and perhaps more important would be the determination of whether the GFR is improving or not. Hence one should always remember that to detect AKI, it is essential to assess the change in creatinine , and that across all patients, a single creatinine valuenever corresponds to a given GFR .

The production rate and volume of distribution of creatinine also affect the plasma creatinine value, while other factors such as medications (e.g. cimetidine) can inhibit distal tubular creatinine secretion and cause a rise in serum creatinine without being secondary to AKI. Abnormalities in liver function, decreased muscle mass and aging decrease the production of creatinine, whereas fever, immobilization, glucocorticoids or muscle trauma increase its production.

Urine output is sensitive to changes in renal hemodynamics, but extremely insensitive to define or differentiate AKI. Oliguria has been defined as urine output of <400 ml/day or <0.3 ml/kg/hour, and has been found to be associated with a higher mortality as compared to non-oliguric ARF. However, severe AKI can exist in non-oliguric states as well, whereas obstructive causes could present with a fluctuating levels of urine output. The RIFLE criteria does employ urine output as markers of risk, injury or failure, though concordance between serum creatinine and urine output criteria has not been established, even with regards to mortality risk.

Acute Kidney Injury: Incidence and Risk Factors

Etiology

Frequently the cause for AKI is multifactorial, involving more than one insult. However, traditionally the etiology of AKI has been categorized anatomically into pre-renal, renal and post-renal ( Figure 76.1 ).

Prerenal Acute Kidney Injury

The most common cause of AKI is prerenal azotemia and it accounts for about 40–55% of the cases. It results from kidney hypoperfusion due to a reduced effective arterial volume (EAV). The effective arterial volme is the volume of blood effectively perfusing organs. Conditions causing hypovolemia resulting in a reduced EAV include hemorrhage (traumatic, gastrointestinal, surgical), GI losses (vomiting, diarrhea, nasogastric suction), kidney losses (over-diuresis, diabetes insipidus) and third spacing (pancreatitis, hypoalbuminemia). In addition cardiogenic shock, septic shock, cirrhosis, nephrotic syndrome and anaphylaxis all are pathophysiologic conditions that decrease effective circulating volume, independent of the volume status, resulting in reduced renal blood flow. Pre-renal AKI reverses rapidly once renal perfusion is restored because the renal parenchyma remains uninjured. However, when hypoperfusion is severe, it may result in ischemia leading to acute tubular necrosis.

Pre-renal azotemia has also been divided into volume responsive and volume non-responsive. In volume responsive pre-renal azotemia correction of the patients volume status results in increased kidney perfusion and resolution of the disorder. In volume non-responsive forms, additional intravenous volume is of no help in restoring kidney perfusion and function. Disease processes such as severe congestive heart failure and sepsis may not respond to intravenous fluids as markedly reduced cardiac output or a reduction in total vascular resistance, respectively, prevent volume mediated improvement in kidney perfusion. Therefore it is essential to understand the mechanism mediating pre-renal azotemia in order to correct it.

Hypovolemia causes a decrease in mean arterial pressure which activates baroreceptors and initiates a cascade of neural and humoral responses. This leads to the activation of sympathetic nervous system that causes increased production of catecholamines especially norepinephrine. The other major consequence is the activation of the renin-angiotgensin-aldosterone (RAAS) system that causes production of angiotensin II (ATII), a very potent vasoconstrictor. There is also an increased release of anti-diuretic hormone (ADH) mediated both by hypovolemia and a rise in extracellular osmolality, that retains water, as well as influencing urea back-diffusion into the papillary interstitium.

In response to volume depletion or states of decreased EAV there is increased intra-renal ATII activity. This increases proximal tubule Na ⁺ absorption through a complex effect in the glomerulus by preferentially increasing the efferent arterioral resistance. Thus the glomerular hydrostatic pressure is increased and preserves GFR. With severe volume depletion there is greater ATII activity leading to afferent arteriolar constriction, that reduces both renal plasma flow and the filtration fraction. ATII has also been shown to have direct effects on transport in proximal tubule through receptors located in the proximal tubule. It has also been postulated that the proximal tubule can locally produce ATII. Hence under conditions of volume depletion, ATII stimulates a larger fraction of the transport, whereas volume expansion will blunt this response.

There is also significantly increased renal sympathetic nerve activity in pre-renal azotemia. Studies have shown that in volume depleted states adrenergic activity independently constricts the afferent arteriole as well as changing the efferent arteriolar resistance through ATII. Renal nerve activity is linked to renin release through β-adrenergic receptors on renin-containing cells while α-a adrenergic influences primarily the vascular resistances within the kidney. In contrast α-2 adrenergic agonists primarily decrease the glomerular ultrafiltration coefficient via ATII. Although vasodilation might be expected as a result of acute removal of adrenergic activity, a transient increase in ATII is actually seen, along with constancy in GFR and renal blood flow. Even after sub-acute renal denervation renal vascular sensitivity increased to ATII as a result of major up-regulation of ATII receptors. Hence complex effects on the renin-angiotensin activity occur within the kidney secondary to renal adrenergic activity during pre-renal azotemia.

All these systems work together and stimulate vasoconstriction in musculocutaneous and splanchnic circulations, inhibit salt loss through sweat, stimulate thirst and retain salt and water to maintain blood pressure and preserve cardiac and cerebral perfusion. Various compensatory mechanisms preserve glomerular perfusion. Autoregulation is achieved by stretch receptors in afferent arterioles that cause vasodilation in arterioles in response to reduced perfusion pressure. Under physiologic conditions autoregulation works until a mean systemic arterial blood pressure of 75–80 mm Hg. Below this, the glomerular ultrafiltration pressure and GRF decline. Kidney production of prostaglandins, kallikrein and kinins as well as nitric oxide is increased contributing to the vasodilation. NSAIDs, by inhibiting prostaglandin production, worsen kidney perfusion in patients with hypoperfusion. Selective efferent arteriolar constriction, which is a result of ATII, helps preserve the intraglomerular pressure andGFR. ACE inhibitors inhibit synthesis of angiotensin II and so disturb this delicate balance in patients with severe reductions in EAV such as severe CHF or bilateral renal artery stenosis and can worsen prerenal azotemia. On the other hand, very high levels of angiotensin II seen in circulatory shock causes constriction of both afferent and efferent arterioles which negates its protective effect.

Although these compensatory mechanisms are protective against acute renal failure, they are overwhelmed in states of severe hypoperfusion. Renovascular disease, hypertensive nephrosclerosis, diabetic nephropathy as well as older age predispose patients to kidney hypoperfusion at lesser degrees of hypotension.

Post-Renal Acute Kidney Injury

Post-renal AKI occurs from either ureteric obstruction or bladder/urethral obstruction. AKI from ureteric obstruction requires that the blockage occur either bilaterally at any level of the ureters, or unilaterally in a patient with a solitary functioning kidney or CKD. Ureteric obstruction can be either intraluminal or external. Bilateral ureteric calculi, blood clots, and sloughed renal papillae can obstruct the lumen, while external compression from tumor or hemorrhage can block the ureters as well. Fibrosis of the ureters intrinsically or of the retroperitoneum can narrow the lumen to the point of complete luminal obstruction. The most common cause for post-renal azotemia is structural or functional obstruction of the bladder neck. Prostatic conditions, therapy with anti-cholinergic agents and a neurogenic bladder can all cause post-renal AKI. Relief of the obstruction usually causes prompt return of GFR if the duration of obstruction has not been excessive. The rate and magnitude of functional recovery is dependent on the extent and duration of the obstruction.

AKI resulting from obstruction usually accounts for less 5% of cases, although in certain settings, e.g., transplant, it can be as high as 6–10%. Clinically the patient can present with pain and oliguria, though these are non-specific. Because of the ease of ultrasonography, its diagnosis is usually straightforward, although a volume depleted patient or a patient with severe reduction in GFR may not show hydronephrosis on radiological assessment. Since initially during the course of the disease GFR is not affected, volume repletion can help with the diagnosis by increasing GFR and urine production into the ureter leading to dilation of the ureter proximal to the obstruction and enhancing ultrasound visualization. Early diagnosis and prompt relief of obstruction remain key goals in preventing long-term parenchymal damage as the shorter the period of obstruction the better the chances for recovery and long-term outcomes. The pathophysiology and treatment of obstructive uropathy are discussed extensively in another chapter.

Intrinsic or Intra-Renal Acute Kidney Injury

It is helpful to divide the causes of intrinsic renal azotemia among categories that delineate the site of the initiating injury. Thus the most useful classification is as follows: (1) Vascular—(a) large renal vessels and (b) renal microvasculature; (2) Glomerular; (3) Tubular and (4) Interstitial. AKI secondary to vasculitides and rapidly progressive glomerulonephritides are discussed elsewhere in the text. Below we shall focus on AKI pathophysiology secondary to tubular and interstitial diseases before going on to discuss ATN in detail.

Experimental Models of ARF

Despite a variety of animal and cell culture models of AKI, there remains a need to develop in vivo experimental models of ischemic AKI more closely mimicking clinical human AKI for the development of effective therapies. Some of the important principles in studying the pathophysiology of AKI in various models include the importance of measuring parameters at multiple appropriate time points and the ability to control physiological functions known to affect kidney function (e.g., temperature, blood pressure, anesthesia, fluid status etc.). A limitation in many experimental models is the lack of co-morbidities such as aged animals, chronic kidney disease, multi-organ failure, pre-existing vascular changes or multiple renal insults, which quite often co-exist in human AKI. We will briefly describe the pros and cons of using these experimental models ( Table 76.4 ).

The warm ischemia-reperfusion renal clamp model is one of the most widely used experimental models in rats and mice because of its simplicity and reproducibility. In rats the inflammatory response, tubular injury and repair, and medullary congestion are similar and probably comparable to human ischemic ATN. However, in human AKI, isolated ischemia is seen rarely seen and neither is there usually complete cessation of blood flow to the kidneys. In this model, important mediators of injury suchreactive oxygen species (ROS) and perioxynitrite species may have a different or delayed role as compared to low oxygen states in hypoperfusion models. Total blood flow cessation also prevents the degradative products of the ischemic kidney from being eliminated. Other factors playing a role in the pathophysiology of AKI such as inflammatory mediators released from gut ischemia, endothelium, smooth vascular muscle cells need to be taken into consideration in any experimental model. Release of bowel proteins into the circulation can act as inflammatory mediators and increase the susceptibility to AKI. The S3 segment of the proximal tubule is almost completely necrosed in such models, a finding not seen very frequently in human ARF. In contrast to animal models, human AKI histological biopsy data are lacking at early time points from the onset of insult. This has made comparison between animal models and human AKI of limited value. Lastly, drug delivery is prevented in total occlusion models, which actually may be of significant value during the peak ischemic insult.

The cold ischemia-warm reperfusion model resembles AKI in human transplants but this model is inadequately studied and difficult experimentally. In the isolated perfused kidney model, the kidney is perfused in ex vivo using perfusates either with or without erythrocytes, and employs either ischemic (stopping perfusate) or hypoxic (reduced oxygen tension of erythrocytes) to induce functional impairment. The morphological patterns are different in erythrocyte free and erythrocyte rich perfusates. The latter system is more comparable with what is observed histologically in animal models. Additionally, limitations include exclusion of various inflammatory mediators, neuro-endocrine hemodynamic regulation, and systemic cytokine and growth factor interactions known to be present and play a pathophysiologic role in animal models.

Cardiac arrest is a common scenario leading to human ARF. Rabb et al. have described whole body ischemia reperfusion injury model induced by 10 minutes of cardiac arrest, followed by cardiac compression resuscitation, ventilation, epinephrine and fluids, which that lead to a significant rise in SCr and renal tubular injury at 24 hours. One of the unique advantages of this model is the cross talk between vital organs such as the brain, heart, lung and the renal hemodynamics. A hypoperfusion model of AKI using partial aortic clamping was first described by Zager et al. may be more representative of human AKI reflecting a state of reduced blood flow to the kidney with systolic blood pressure around 20 mm Hg, resulting in reproducible AKI. This was also recently adapted and refined in a study where a novel compound, soluble thrombomodulin, was used to minimize ischemic injury in a partial aortic clamp AKI model.

Toxic models of renal failure employ various known toxins, such as radiocontrast, gentamicin, cisplatin, glycerol and pigments including myoglobin and hemoglobin. Septic models to study AKI include cecal ligation and puncture, endotoxin infusion andbacterial infusion into the peritoneal cavity. The endotoxin model which is simple, inexpensive and suitable to study new pharmacological agents, has certain drawbacks as well. There is variability amongst sources and types of lipopolysaccharide (LPS) endotoxin, rate and method of administration, and it is usually of short duration due to the high mortality associated with the doses required to induce AKI. It also tends to be a vasoconstrictive model and does not recapitulate the hemodynamics nor inflammation of human sepsis. In the cecal ligation and puncture model (CLP), there is considerable similarity with sepsis in humans with acute lung injury, metabolic derangement and systemic vasodilation, accompanied by increased cardiac output initally. However there is some variability depending on the mode and size of cecal perforation. Star et al. have developed a new sepsis model keeping under consideration the following facts: (1) animals should received the same supportive therapy that is standard for ICU patients (i.e., fluid resuscitation and antibiotics); (2) age, chronic co-morbid conditions and genetic heterogeneity vary. Complex animal models of human sepsis that introduce these disease-modifying factors are likely more relevant and may be more pharmacologically relevantthan simple animal models. The zebra fish model developed by Bonventre et al. has the advantages of markedly improved accessibility of the kidney, feasibility of knock-down and upregulation of genes and a short phenotypic readout time, while at the same time possessing the complexity of an organism to study renal injury. These properties may make it a useful inexpensive tool to screen therapeutic agents in the future.

Experimental models of hypoxic acute kidney damage differ morphologically in the distribution of tubular cell injury and tubular segment types differ in their capacity to undergo anaerobic metabolism, mount hypoxia-adaptive responses mediated by hypoxia-inducible factors (HIFs). Hence it is important to keep them in mind the potential pitfalls when evaluating experimental studies or therapeutic interventions using these models. The lack of ability to demonstrate effectiveness of an agent in humans which has been shown to be efficacious in animal models, does not necessarily reflect a flaw with the model. Most often, the agent is administered very late in the course of the human disease, and the patient heterogeneity of the population makes it even more difficult to establish true efficacy.

Role of Biomarkers in AKI

Changes in serum creatinine and/or urine output to diagnose AKI may not be able to identify the early stages of intrinsic kidney injury. Early identification and subsequent early pharmacologic intervention may improve outcomes in AKI. In order to facilitate the early diagnosis of intrinsic injury, multiple biomarkers of tubular injury have been evaluated. Biomarkers for AKI include N-acetyl- B -D-glucosaminidase (NAG), kidney injury molecule 1 (KIM-1), neutrophil gelatinase-associated lipocalin (NGAL) and interlukin 18 (IL-18), among others. In addition, serum cystatin C has been proposed as more sensitive and timely than serum creatinine for detecting changes in GFR, and urinary cystatin C has been proposed as a marker of tubular injury. Although not utilized yet for routine clinical use, these biomarkers have the potential to provide an early diagnosis of intrinsic AKI and the ability to differentiate pre-renal AKI from intrinsic tubular damage, as well as to provide prognostic information of an episode of AKI. One biomarker or a panel of these biomarkers may eventually provide the necessary early diagnosis to allow therapies to limit kidney damage and promote recovery of kidney function. Please refer to Chapter 75 for a detailed explanation on the role of serum and urinary biomarkers in AKI.

Morphological Changes of AKI

Acute Tubular Necrosis (ATN) is the most common form of AKI, and this process of renal tubular injury encompasses more than just cell death followed by repair. It is easier to understand the entire spectrum of injury if one looks at the different compartments involved and the phases they go through ( Fig. 76.2 ).

Tubular Epithelial Cell Injury

Although data regarding which nephron segments in humans with ATN are more severely affected is sparse due to lack of biopsies early in the course of ATN, experimental animal models provide sufficient information to help understand and delineate the mechanisms of ATN by histological analysis. In ATN the most severe tubular injury takes place within the outer medulla of the kidney, and involves the S3 segment of the proximal tubule (pars recta) and the medullary thick ascending limb (MTAL) of the distal nephron. The S3 segment has limited capacity to undergo anaerobic glycolysis. Secondly, due to its unique primarily venous capillary regional blood flow, there is marked hypoperfusion and congestion in this medullary region post injury that persists even though cortical blood flow may have returned to near normal levels after ischemic injury. Endothelial cell injury and dysfunction are primarily responsible for this phenomenon, known now as the “extension phase” of AKI. The proximal tubule S ₁ and S ₂ segments are most commonly involved in toxic nephropathy due to their high endocytic rates leading to increased cellular uptake of the toxin.

The apical brush border of proximal tubule cells (PTC) is damaged early resulting in microvilli disruption and detachment from the cell surface forming membrane bound “blebs” released into the tubular lumen. Loss of microvillar surface leads to ineffective enzymatic activity, endocytosis, channel and transporter density resulting in diminished effective transcellular absorption. Patchy detachment and subsequent loss of tubular cells exposing areas of denuded tubular basement and focal areas of proximal tubular dilatation along with the presence of distal tubular casts is also a major pathological findings in ATN. Because the surviving adjacent cells tend to spread out and become flattened, in an attempt to completely or partially cover the denuded epithelium, the appearance is often of a flattened and pauci-cellular epithelium.

Sloughed off tubular cells are also present in the tubular lumen, where they can be overtly necrotic or viable . These cells along with brush border vesicle remnants and cellular debris combine with Tamm-Hosrfall glycoprotein (THP) and form the classical “muddy-brown granular” casts, that have the potential to obstruct the tubular lumen. Cast formation may be potentiated by relative stasis of tubular fluid flow because of the reduction in GFR. On biopsy, these casts may not be captured since they exist within the medulla and actual site of obstruction may be a short segment. However, dilation of tubules proximal due to obstruction is often seen as long as the GFR and regional tubular epithelium are maintained. Eventually decompression of these tubules will occur because of decreased GFR, damage to proximal tubular cells, and persistent reabsorption of tubular fluid in uninjured areas. The injury to the proximal tubule cells in humans is also seen in experimental models of ischemic AKI ( Fig. 76.3 ).

Apoptotic features are more commonly seen in both proximal and distal tubule cells as compared to necrosis which itself is inconspicuous and restricted to the highly susceptible outer medullary regions. Apart from the proximal tubular cells, the other major epithelial cells of the nephron are those of the medullary thick ascending limb located distally. Apoptotic changes have been detected in human AKI, as shown in distal nephron segments in nephrotoxic acute tubular necrosis. Distal tubular cell apoptosis also occurs in donor biopsies before engraftment, which was predictive of delayed graft function due to acute tubular necrosis. In an ex vivo model of hypoxic AKI, inhibition with FG-4497 (specific prolyl-hydroxylase inhibitor which leads to activation of HIF-Aplha) in the isolated perfused kidney led to decreased selective outer medullary distal tubular injury. The course of a tubular cell alterations may take different paths depending on the type and extent of injury as discussed later.

Glomerulus

The glomerular tuft collapses in ischemic injury, and some investigators have described the diameter of ischemic endothelial cell fenestrae on average to be larger than that of untreated kidneys. Other human biopsy studies have documented enlargement of juxtaglomerular apparatus during the oligoanuric phase, and thickening and coarsening of foot processes. But these findings have not been confirmed and there still exists a paucity of data on glomerular changes in human ATN in different stages. Glomerular epithelial cell injury in ischemic, septic or nephrotoxic injury is not classically seen although some studies have shown thickening and coarsening of foot processes and recently Wagner et al. have shown podocyte specific molecular and cellular changes.

Epithelial Cytoskeletal Abnormalities

Cytoskeletal Alterations

Cellular structure and function are mediated by an interactive and dynamic role of the actin cytoskeleton including but not limited to proximal tubule brush border microvilli structure and function, cell polarity, endocytosis, signal transduction, cell motility, movement of organelles, exocytosis, cellular division and migration, barrier function of the junctional complex, cell-matrix adhesion and signal transduction. Actin is present in globular form (G-actin) that can self-assemble into filamentous (F-actin) to form helical microfilaments. In conjunction with actin-binding proteins, guanosine triphosphatases (GTPases) and adenosine triphophate (ATP), the dynamic process of actin assembly and disassembly is accomplished. The actin cytoskeleton is present as a layer of microfilaments below the apical plasma membrane, forming the terminal web. The architectural integrity of brush border microvilli is dependent upon extensions of actin filaments from the terminal web to the tip of the individual microvilli.

Under physiologic conditions most actin monomers are ATP-bound, while the bulk of actin found within actin filaments is ADP-bound. The unique binding sites of actin for nucleotides and divalent cations allow conformational changes, apart from its hydrolytic properties that are activated by polymerization. The G-actin ADP complex released on depolymerization (disassembly) undergoes nucleotide exchange with abundant cytosolic ATP, and is then stored as a high energy G-actin ATP intermediate by thymosin-sequestering proteins until needed for polymerization. Four groups of actin-binding protein mediate different effects on the actin cytoskeleton including actin sequestering, capping, severing and nucleation. These effects in turn coordinate the continual remodeling of the cytoskeleton and give it the ability to respond to various internal and external stimuli.

Ischemic insults to proximal tubule cells induce a rapid and severe degeneration of the microvillar F-actin core, which in turn mediates the plasma membrane finger-like microvillar structural morphology changes including loss of the apical membrane through blebbing. The degeneration of this F-actin core occurs as a result of ATP depletion leading to depolymerization of microvillar actin. In addition ezrin, an actin binding phosphorylated protein, becomes dephosphorylated during ischemia and the attachment between the microvillar F-actin core and the overlying plasma membrane is lost ( Fig. 76.4 ). The apical membrane is then either exfoliated as “blebs” into the tubule lumen or internalized with the capability of being recycled during cellular recovery. Furthermore, the concentration of F-actin in the cell increases with the formation of large cytosolic aggregates in the perinuclear region and also near the junctional complexes and basolateral membrane. ATP-G-actin levels decrease rapidly during ischemia. Since ADP G-actin cannot be sequestered by thymosin its concentration exceeds a critical threshold resulting in the polymerization in an unregulated fashion.

The actin binding protein family of cofilin, also known as actin depolymerizing factor (ADF), has been shown by Molitoris and colleagues to be a critical mediator in F-actin severing during ischemic injury. In proximal tubule cells (PTC) ADF/cofilin, which when phosphorylated is inactive and does not bind actin, gets rapidly dephosphorylated and therefore activated by renal ischemia. This leads to relocalization from the cytoplasm to the surface membrane, as well as in shed membrane-bound vesicles seen in PTC lumen. The mechanism by which disruption of F-actin structure occurs also involves the role of another family of actin binding protein called tropomyosin. Under physiologic conditions tropomyosin binds to and stabilizes the F-actin microfilament core in the terminal web, and protects the filaments from ADF/cofilin induced severing and depolymerization. Ashworth et al. have demonstrated that after ischemic injury there is dissociation of tropomyosin from the microfilament core providing access to microfilaments in the terminal web for F-actin binding, severing and depolymerizing actions of ADF/cofilin proteins.

Alterations in the activity of Rho family of GTPases also contributes to changes in actin cytoskeleton associated with ischemia. First, chemical ATP depletion was shown to cause Rho GTPase inactivation. Secondly, GTP depletion during ischemia could also inactivate Rho GTPase function. These two findings, coupled with the finding that cells transfected with a constitutively active form of RhoA during chemical ATP depletion are protected against actin depolymerization, provide evidence of ischemia-induced RhoGTPase inactivation. An additional cytoskeletal component important for cellular polarity and protein trafficking is microtubules. Wald et al. have shown that in reperfused rat proximal tubules non-centrosomal microtubule organizing centers (MTOCs) were fully detached from the cytoskeleton and scattered throughout the cytoplasm at three days after reperfusion when brush borders membranes were mostly reassembled with normal F-actin distribution. At that time microtubules were also fully reassembled but lacked their normal apicobasal orientation, hence demonstrating that the reestablishment of the submembrane F-actin does not seem to be sufficient for a full polarization of the cells. Microtubule formation also occurs by continuous assembly and disassembly of α and β tubulin heterodimer with an intricate polymerization process. Studies indicate that during ischemia α and β tubulin do not participate in microtubule polymerization and their localizations are also different. The fact that GTP levels are depleted by 90% after 30 minutes kidney ischemia in rats supports this assumption of impaired microtubule polymerization.

Epithelial cells are characterized by an asymmetrical distribution of proteins and lipids in the apical and basolateral membrane resulting in surface membrane polarity of these cells. In ischemic ATN this polarity is abolished, but has the potential for re-establishment during recovery. Molitoris et al. have shown evidence to suggest the Na ⁺ K ⁺ -ATPase pump that normally resides in the basolateral membrane of proximal tubule cells, under conditions of chemical anoxia, is redistributed to the apical membrane. This redistribution, which can occur as early as 10 min after ischemia, is another consequence of the disruption of the actin cytoskeleton, which normally maintains the attachment of the Na ⁺ K ⁺ -ATPase to the basolateral membrane. Furthermore, both ankyrin and fodrin dissociated from F-actin and each other during ATP depletion. These data were confirmed and indicate wide spread actin cytoskeleton alterations during ATP depletion lead to altered protein-protein interactions. Other nephron segments such as the distal tubule cells, and TAL do not show similar apical redistribution of the Na ⁺ K ⁺ -ATPase. This redistribution results in functional consequences reflected in the loss of unidirectional transport of salt and water across the epithelial cell, resulting in one mechanism of the high fractional excretion of Na ⁺ seen in patients with ATN ( Fig. 76.5 ).

Junctional Defects and Permeability Alterations

Cell-cell junctional complexes actively participate in the establishment and maintenance of cell polarity, paracellular transport, cytoskeletal interactions, and rearrangements in cellular shape. Ischemia also induces functional changes in the epithelial junctional complexes, which are comprised of at least three structures: adherens junctions (also known as zonula adherens (ZA)), tight junctions (zonula occludens (ZO)) and desmosomes. Tight junctions are located directly apical to the adherens junction, are composed of a growing list of proteins such as occludin, claudin, protein kinase C (PKC), ZO-1 etc., with multiple functions such as adhesion, permeability, structural integrity and paracellular transport of solutes. The actin present in the cortical belt, and in the terminal web, is also linked to the tight junction. Adherens junction, which are located directly below the tight junction, form strong cell–cell adhesion complexes and are composed of proteins such as cadherins and catenins are associated with numerous other junctional and cytoplasmic proteins. They are responsible for adhesion of adjacent cells, regulation of adhesion, and are also linked to the actin cytoskeleton.

In vivo and in vitro studies indicate that in early ischemic injury there is an “opening” of the tight junctions as the proteins such as ZO-1 and cingulin become insoluble during ATP depletion, and associate into macromolecular intracellular complexes. This leads to increased permeability of the tight junctions in sub-lethal injury resulting in back-leakage of glomerular filtrate, which is an important factor in the reduction of GFR as discussed later. If ATP is repleted before lethal injury the permeability defect resolves. PKC signaling regulates both tight junction and adherens junction assembly. Evidence suggest that tight and adherens junction proteins including ZO-1, ZO-2, ZO-3, occludin, vinculin, and p100–p120 are affected by this kinase signaling pathway. It is also likely that disruption of the actin cytoskeleton and reduction in activity of Rho-GTPases also contribute to the changes in the tight junction during ischemia. The loss of adherens integrity is in part caused by activation of c-Src which translocates to the adherens junction and tyrosine phosphorylates components such as β-catenin. Nigam et al. have demonstrated the role of tyrosine kinases and phosphatases in the disassociation of adherens junction. The importance of these animal model findings is highlighted by findings in human allografts with ATN where Kwon et al. have shown the same features noted in experimental models of loss of cell polarity and tight junctions.

Epithelial cells also lose their attachment to the underlying extracellular matrix, the mechanism for which has been elucidated as at least being partially due to loss of polarity and redistribution from the basal membrane to the apical membrane of β1-integrins. Integrins are transmembrane proteins normally responsible for the anchoring of epithelial cell to the matrix through actin cytoskeleton and actin-binding proteins. In vitro studies of MDCK cells have shown that adherence of these cells to a collagen I substratum is mediated by peripheral actin filaments and adhesion complexes regulated by myosin light chain kinases and adhesion complexes controlled by RhoA. The detachment and loss of tubular cells into the lumen also contributes to the back-leakage of the glomerular filtrate, and at the same time the β1-integrins and E-cadherins might even play a role in mediating the aggregation of these exfoliated cells worsening intraluminal cast formation.

Although the glomerular injury is not as prominent in AKI, Wagner et al. have demonstrated in an in vivo rat model that renal ischemia induces podocyte effacement with loss of slit diaphragm and proteinuria owing to rapid loss of interactions between the tight junction proteins Neph1 and ZO-1. Cell culture models using human podocytes further showed that ATP depletion resulted in rapid loss of Neph1 and ZO-1 binding, and redistribution of Neph1 and ZO-1 proteins from the cell membrane to cytoplasm; ATP recovery increased phosphorylation of Neph1 and restored Neph1 and ZO-1 binding and their localization at the cell membrane.

Tubular Obstruction

Tubular obstruction has been noted in ischemic as well as toxic models of injury. Renal tubular epithelial cells can be seen in the urine of patients with AKI, and can be either alive, apoptotic or necrotic. Micropuncture studies, done over 20 years ago, demonstrated elevation of intratubular pressure early after reperfusion following renal artery occlusion. This is characteristically evident as tubular dilatation, with cast deposition in the distal nephron causing luminal obstruction and back pressure. Although intratubular pressures tend to fall towards normal after 24 hours, the presence of persistent obstruction can be revealed by extracellular volume expansion, which again elevates intratubule pressures. An obstructing cast in the collecting duct could potentially impair the function of multiple nephron units, as many nephrons drain into a single collecting duct.

The term “back-leak” generally implies the passive movement of GFR into the interstitium from the tubular lumen, eventually being recirculated to the systemic vasculature through the venous network. Studies have revealed that if radiolabeled compounds are microinjected into renal tubules after ischemic injury, they can be detected in the contra-lateral kidney. Human studies by Myers et al. provided evidence of transtubular leakage of GFR after ischemic renal failure as well as tubular obstruction, leading to a reduction in measured or effective GFR. The presence of areas of open PTC tight junction or denuded basement membrane in electron-microscopy biopsy specimens provides a logical morphological explanation of back-leakage. However, understanding the mechanism responsible for tight junction dysfunction or detachment of tubular cells is key to defining the event of cast formation and tubular obstruction. The integrin superfamily of proteins, located on the basal aspect of the cell, is responsible for the complex cell-matrix adhesion events. The β chains coupled with α chains form β1 integrins, which interact with the actin cytoskeleton and actin-binding proteins such as α-actinin, vinculin and talin. The extracellular domain of the β1 integrins attaches to receptors of proteins such as collagen and fibronectin, which are abundant and constitute the tubular basement membrane. The tri-peptide sequence of arginine-glycine-asparagine (abbreviated as RGD), is a well define receptor for β1 integrin on the extracellular matrix. The loss of polarity causes redistribution of β1 integrins, which become expressed in the apical domain of sublethally injured cell. It was hence hypothesized that administraton of an excess of soluble RGD containing molecules would saturate the extracellular binding site of all exposed β1 integrins within the lumen of the nephron, and thus prevent luminal renal tubular cell-cell adhesion, and this prevent intratubular obstruction. Both intravenous and direct intrarenal infusion of RGD peptides resulted in amelioration of ischemic AKI. It is also possible that the detached tubular cells adhere to Tamm-Horsfall protein (THP) in the distal tubule by RGD sequence peptides. By using dual labeled RGD peptide sequences, it was also discovered that RGD peptides also mapped to intimal surface of vessels in ischemic kidneys. Further studies utilizing RGD peptides in ARF could provide key answers to questions of vascular and epithelial injury in ATN. Recent studies also show a role of the sphingosine-1 phosphate receptor (S1PR) in maintaining structural integrity after AKI. Okusa et al. have shown that S1PRs in the proximal tubule are necessary for stress-induced cell survival, and S1P ₁ R agonists are renoprotective via direct effects on tubular cells.