Keywords

1 Introduction

Acute kidney injury (AKI) is a highly prevalent and devastating clinical problem, one for which new treatments and prophylactic therapies are urgently required. In order to understand and devise novel interventions, a complete understanding of the changes seen in renal physiology in response to acute insults is required. This chapter will discuss the alterations seen in cell polarity, the cell–cell signalling in the tubular compartment and the function of the enzyme heme oxygenase-1 (HO-1) in experimental AKI. Furthermore, the role of acute inflammation in AKI will be considered, including the roles played by cytokine release, inflammatory leukocytes, toll-like receptors (TLRs) and complement activation on disease initiation and progression (Fig. 5.1). Much of our understanding of this area is based on experimental models of acute kidney injury in rodents such as cisplatin nephrotoxicity and renal ischaemia/reperfusion injury. While these models provide valuable information of putative pathophysiological changes within the injured kidney, it must be borne in mind that in the clinical setting, patients with AKI are usually advanced in age, with multiple comorbidities and renal insults, and where human data exists, this will also be discussed.

Fig. 5.1
figure 1

Acute kidney injury. This diagram illustrates the various topics covered in this chapter and their approximate timings and roles in the initiation, propagation and subsequent repair of the injured kidney

Full size image

2 Cell Polarity in AKI

Tubular cell polarity is of fundamental importance to normal renal function. The maintenance of luminal and basolateral membrane specificity is central to providing vectorial and selective molecule transport. During tubulogenesis, epithelia coordinate the polarity of individual cells in relation to their neighbouring cells and matrix to create spatially and functionally distinct membranes. This coordination occurs largely through orientating signals from cell–cell and cell–matrix interactions.

The essential structures in determining cell polarity include specific surface membrane domains, a number of junctional complexes (tight junctions, gap junctions, adherens junctions and desmosomes) and the actin cytoskeleton [1].

2.1 Acute Injury

There is a loss of cell polarity in acute renal injury [2,3,4,5,6]. Much of our current understanding stems from in vivo ischaemia/reperfusion injury (IRI) models and in vitro models of cultured tubular cells, which allow variable degrees of injury and ATP depletion [7, 8]. An important response to tubular injury is the redistribution of key structural proteins between the apical surface and basolateral membrane. Both actin and the actin-binding protein, villin, have been shown to migrate from the apical to the basolateral plasma membrane within 1 h of reperfusion following ischaemic injury [9]. The tubular apical cytoskeleton is exquisitely sensitive to hypoxia, and assembly of actin and intracellular microfilaments diminishes rapidly as ATP levels are reduced [9,10,11,12]. Furthermore, in IRI models, fragmentation of the microtubule network itself has been described [13]. Human studies are consistent with the observations in rodents. In human transplant allograft studies, redistributed cytoskeletal proteins are seen in injured kidneys [14].

Ischaemic injury and ATP depletion in vitro and in vivo also disrupt the epithelial tight junctions [15, 16]. Depleting ATP causes tight junctions to aggregate intracellularly and to form insoluble particulate complexes [17, 18]. β[beta]1-integrins, which mediate epithelial matrix and cell–cell adhesion, migrate to the apical surface during ischaemic injury [19]. The now apically expressed β[beta]1-integrins remain functional and may contribute to abnormal epithelial cell adhesion within the tubule following desquamation [20]. These free tubular cells may adhere to each other and to exocytosed cytoplasmic contents and result in tubular blockage, raising intratubular pressure and resulting in increased “backleak” (see below) of glomerular filtrate from tubular lumen into peritubular capillaries [21]. Loss of these junctional proteins and resultant cytoskeletal change disrupts the molecular anchors, leading to exfoliation of tubular cells into the lumen [22,23,24,25].

In humans, posttransplant allograft biopsies of injured kidneys demonstrate redistribution of the Na+/K+-ATPase from the basolateral membrane to both the apical membrane and cytoplasm [3, 26,27,28]. The Na+/K+-ATPase is thought to be crucial in tight junction formation in addition to its functions in sodium reabsorption [29]. The loss of epithelial cells and Na+/K+-ATPase reduces overall tubular reabsorptive capacity. When the increased sodium load is sensed at the macula densa, this triggers tubuloglomerular feedback, reducing blood flow to the glomerulus and further decreasing effective GFR. The loss of epithelial continuity leads to tubular filtrate translocating paracellularly into the renal interstitium and renal venous system. This phenomenon results in reduced effective glomerular filtration and is known as “backleak ” [14, 22, 30, 31].

Recovery of polarity following ischaemic injury requires reassembly of the tight junction catenin components. While milder injury may allow reuse of preformed catenins, more severe ischaemia may require synthesis of new proteins in the endoplasmic reticulum [15]. In some models of injury, reorganisation of the tubular microstructure can be observed as early as 24-h post-reperfusion [13].

3 Cell Signalling in AKI

3.1 Arachidonic Acid Metabolites

While three pathways of arachidonic acid metabolism exist, cyclooxygenase (COX)-generated compounds are believed to be the most important for renal tubule–tubule signalling [32]. It is worth noting that other arachidonic acid metabolites (e.g. epoxyeicosatrienoic acids or 20-hydroxyeicosatetraenoic acid derived from the P450 system) can alter epithelial sodium transport, glomerular haemodynamics and vascular reactivity and have anti-inflammatory effects [33, 34]. Both COX1 and COX2 enzymes produce PG compounds in the normal and diseased kidney, with ischaemia and vasoconstriction both reported to produce acute increases in levels of vasodilatory PGE2 and PGI2 [35]. Levels of COX2 increase in response to renal insults, ageing, diabetes [36], heart failure [37] and lithium treatment [38], with COX2 inhibition reported to be protective [39]. The metabolites of COX1 and COX2 have important roles within the kidney, regulating blood flow, affecting the release of renin and mediating NaCl excretion [40]. Supporting this, agonists of the PGE2 receptor have lessened renal injury after experimental mercuric chloride-induced AKI [41], while a clinical trial of the PGI2 analog Iloprost lessened contrast-induced AKI in patients [42]. While these functions fulfil important physiologic roles, increased levels of COX metabolites are seen in disease states and likely contribute to inflammatory injury. COX2-derived prostanoids are renal vasodilators [43], and the inhibition of this effect by non-steroidal anti-inflammatory drugs likely contributes to their nephrotoxicity [44].

Three studies of 20-HETE antagonism or supplementation have been published with conflicting results [45,46,47], with the most recent papers agreeing that inhibition of 20-HETE improves vascular tone and renal function after murine experimental IRI [45, 46]. Resolvins, protectins and maresins produced via the degradation of omega-3 fatty acids and docosahexaenoic acid have all been reported to promote the resolution of the inflammation following the acute phase of renal injury [48], with their administration to mice mitigating the adverse effects of experimental IRI [49].

3.2 Adenosine Triphosphate (ATP)

While ATP is generated within mitochondria and is predominantly intracellular, it is recognised also to be a paracrine signal that acts among renal tubules [50, 51]. At one-thousandth of the intracellular concentration, extracellular ATP (eATP) mediates autocrine and paracrine signalling between renal tubular cells. ATP ligates two families of purinergic receptors—ionotropic (P2X) and metabotropic (P2Y) [52]. When released by dying renal cells in AKI, ATP functions as a danger-associated molecular pattern (DAMP) and elicits immunostimulatory responses in leukocytes [53].

In the mouse, P2X7 receptor inhibition protects against IRI, with protection also seen in the unilateral ureteric obstruction (UUO) model of renal injury and fibrosis in rats [54, 55].

3.3 Nitric Oxide

Nitric oxide (NO) is generated in tissues by nitric oxide synthase (NOS) with all three NOS enzymes present within the tubular cells of the kidney [56]. NOS3 is expressed in cells of the proximal convoluted tubule (PCT), thick ascending limb (TAL) and collecting duct (CD) and NOS2 within PCT, TAL, distal convoluted tubule (DCT) and CD cells, while NOS1 is found at low levels in the TAL and CD. Tubular epithelial NO exerts autocrine and paracrine effects promoting natriuresis and diuresis and conveys signals to the adjacent vasculature [57]. Physiologic renal NO release has beneficial effects on renal haemodynamics and function. Renal NO levels decline with advancing stages of CKD [58].

Studies of experimental IRI demonstrate that pre-induction of NO ameliorates disease severity, while co-treatment with NO inhibitors abolishes protection [59]. Potentiating tubular NO production merits investigation as a translational protective therapy for acute and chronic renal injury [60].

3.4 Dopamine

Renal dopamine is generated within the cells of the PCT and secreted through the apical and basolateral membranes exhibiting local paracrine effects and hormonal actions via the blood stream on distal nephron segments [61]. Dopamine signals via five known receptors, broadly grouped into D1-like (D1 and D5 receptors) and D2-like (D2, D3 and D4) groups. In the kidney, the D1 receptor family is expressed throughout the nephron, juxtaglomerular apparatus and vasculature [62].

In experimental IRI in animals, dopamine has been associated with improved urine output and renal function. However, attempts to translate this into clinical practice have been unsuccessful, with no beneficial effect on rates of death/AKI or dialysis after cardiac surgery seen after dopamine infusion in multiple small studies (reviewed in [63]).

3.5 Angiotensin II

Angiotensin II (ATII) is synthesised and released from the proximal tubule and impacts renal water and electrolyte uptake in addition to other sides of production and effects on the vasculature. ATII stimulates sodium uptake by the cells of the PCT, TAL and the CD [64]. In health, ATII augments salt and water uptake via the tubular AT1 receptor with the renal AT1R, a driver of systemic hypertension [65]. Via hormonal or paracrine signalling, ATII also induces proliferation, hypertrophy, inflammation and matrix production by tubular cells [66]—all features common to progressive renal disease. Additionally, ATII mediates tubule–tubule crosstalk indirectly via interstitial pericytes and fibrocytes [67]. Recent work has shown that in murine experimental IRI, ATII levels increase in the aftermath of injury, although its vasoconstrictive actions appear to be antagonised by the presence of reactive oxygen species [68].

3.6 Bradykinin

The nine-amino acid peptide bradykinin has effects on the heart, kidney and systemic blood vessels [69]. Bradykinin is synthesised in the kidney by the TAL and CD with secretion occurring across apical and basolateral membranes [70]. Bradykinin’s effects occur predominantly at a local tissue level, influencing blood pressure via release of NO and prostaglandins [69]. Through interaction with bradykinin B1 and B2 receptors, bradykinin promotes diuresis and natriuresis.

Bradykinin B1 receptors are expressed by differentiating renal tubules [71] and mediate potentially pro-inflammatory effects, with B1 receptor KO mice normotensive and protected from inflammation and AKI [72]. Both kinin receptor inhibition and kinin B1 and B2 receptor knockout mice are protected against cisplatin-induced AKI [73, 74], with B1 receptor antagonists also reducing fibrosis after experimental UUO presumably due to inhibition of bradykinin’s pro-inflammatory actions, including promotion of migration of immune cells to injured tissue [75].

4 Heme Oxygenase-1

Although the constitutively expressed enzymes heme oxygenase-2 and oxygenase-3 provide some basal metabolic activity, the predominant route of mammalian heme metabolism is via the inducible enzyme heme oxygenase-1 (HO-1). While removing a source of oxidative stress and generating free iron, the other products of heme metabolism, biliverdin and carbon monoxide (CO), are all recognised to possess immunomodulatory, antiapoptotic and in the case of CO vasoactive properties [76,77,78,79]. HO-1 induction is also coupled to increased availability of ferritin, leading to prompt conjugation and removal of free iron—removing another source of potential oxidative stress.

4.1 Heme Oxygenase and the Kidney

Our current knowledge of the role of HO-1 in renal disease is largely based on experience of animal models of kidney disease, with additional insights from limited studies performed in human renal biopsy material . Such human data would support HO-1 as being a component of the response of the tubulo-interstitial compartment to injury, with tubular induction of HO-1 in response to injury correlating with reduced markers of oxidative stress [80]. These findings are consistent with the first reported case of human HO-1 deficiency, which was characterised by systemic inflammation, haemolysis and nephropathy with progressive tubulo-interstitial inflammation [81]. Consistent with this, the HO-1 −/− mouse demonstrates exaggerated iron accumulation in the kidney under physiologic conditions with heightened susceptibility to AKI following models of IRI, rhabdomyolysis and cisplatin-induced AKI [82,83,84].

As the principal route of free heme conjugation in the event of tissue injury, HO-1 is induced in a range of organs in response to IRI. Such upregulation has been demonstrated in the rodent kidney in response to experimental IRI [85], human renal allografts (maximal in organs subject to delayed graft function) [86] and cardiac surgery (maximal in patients with AKI) [87]. Renal HO-1 can be induced by diverse noxious stimuli including hypoxia, LPS administration and bile duct ligation, all of which result in protection against a subsequent, more severe experimental IRI insult—implicating HO-1 induction as a potential mediator of the phenomenon of ischaemic preconditioning [88].

In general, chemical upregulation of HO-1 prior to the induction or renal injury results in functional and structural protection against IRI, while inhibition of HO-1 activity results in an abolition of protected phenotype and often an augmented pattern of injury. Such approaches are not without caveats, as the widely used HO-1 inducer, hemin, is itself a pro-oxidant and the HO-1 inhibiting protoporphyrin compounds almost ubiquitously impact the activity of inducible nitric oxide synthase [89].

The mechanisms underlying the protected phenotypes reproducibly seen by HO-1 induction are likely multifactorial (Fig. 5.2). Three broad areas have been identified as downstream targets of HO-1: the maintenance of renal blood flow, the promotion of cell survival and the modulation of immune phenotype in response to renal injury.

Fig. 5.2
figure 2

Current working hypothesis of putative actions of HO-1 in acute kidney injury. HO-1 and the downstream metabolites of heme exhibit pleiotropic effects on the inflammatory response, the cell survival and the microcirculation, all of which are of potential importance in mediating its potential protective effects after IRI (dashed arrows)

Full size image

4.2 Renal Blood Flow and the Microcirculation

Chemical inhibition of HO activity results in reduced medullary blood flow , supporting a role for HO-1 in the maintenance of medullary perfusion [90]. HO-1 induction in renal transplantation in rats increased capillary flow and diameter of intrarenal vessels when assessed by intravital microscopy [91]. Carbon monoxide has important effects on the circulation—via its potent vasodilatory effects and the inhibition of platelet aggregation [92]. Micropuncture studies have demonstrated that HO-1 induction via stannous mesoporphyrin resulted in the abolition of tubuloglomerular feedback-induced afferent arteriolar vasoconstriction, with the effect reproducible by the administration of either a CORM or exogenous biliverdin—implicating both heme metabolites in the mediation of this effect [93]. Studies combining inhaled CO with infused bilirubin in rat renal transplantation demonstrated synergistic effects on both graft survival and on GFR and blood flow rates [94].

4.3 Cell Apoptosis and Survival

A consistent pattern of reduced cell death is present throughout the studies of HO-1 upregulation, while HO-1 inhibition/genetic depletion results in increased levels of apoptosis/necrosis and impaired autophagy, jeopardising cell survival after AKI [95]. The pleiotropic effects of HO-1 make the dissection of the exact mechanism of this survival benefit problematic. It remains possible that protection is related to a secondary phenomenon due to improved perfusion and reduced immune activation. Mitochondrial dysfunction is a feature of IRI, and mice with transgenic overexpression of HO-1 within these organelles are protected from experimental IRI [96]. In vitro evidence from studies of tubular epithelial cell culture has implicated HO-1 as promoting cell survival by induction of the cyclin-dependent kinase inhibitor p21 [97].

4.4 Effects of HO-1 on Immune Phenotype

It is now widely accepted that HO-1 may act as a ‘molecular brake’ on the activation, recruitment and amplification of immune responses (reviewed in [76]). Overexpression of HO-1 results in reduced expression of leukocyte adhesion molecules and reduced activity of the NF-κ[kappa]B pathway. Constitutively HO-1-deficient animals exhibit increased levels of monocyte chemoattractant protein (MCP-1) [82]. HO-1 has been shown to be a target antigen for CD8+ regulatory T cells, resulting in modulation of cellular immune responses [98]. Given the recognised role of lymphocyte populations in determining susceptibility to IRI, this adds an additional ‘extra-enzymatic’ arm to the immunomodulatory properties of HO-1.

Studies in aged mice demonstrated reduced levels of renal HO-1 after IRI, with dosing of the HO-1 inducer heme arginate protective. Of note, this effect was lost when HO-1-expressing macrophages were pharmacologically depleted—implicating the myeloid cell as an important target of HO-1’s effects [99]. Further studies have borne out the importance of myeloid cell HO-1 expression, with animals where HO-1 deficiency is restricted to myeloid cells, exhibiting delayed recovery and increased fibrosis after IRI [100]. Studies in HO-1 −/− mice injected with a bacterial artificial chromosome containing functional human HO-1 demonstrate protection from subsequent IRI, again validating the potential role of this protein in human AKI [101].

4.5 Candidate Pharmacological Inducers of HO-1

Given the accumulating experimental evidence for the beneficial effects of HO-1 in renal injury, induction of HO-1 in the clinical setting is under active investigation. While statins have been shown to induce HO-1 in vitro [102], data in recent large prospective clinical trials have not shown alterations in rates of acute kidney injury [103]. The lipid-lowering agent probucol has HO-1-inducing activity in animal models [104]; however its clinical utility has been limited by its undesired effect of lowering HDL levels. The compound heme arginate (HA) has HO-1-inducing properties [105], is stable at injectable pH [106] and is licenced and available for the treatment of acute porphyria [107]. As such, HA represents a promising translational agent for HO-1 induction in man, with clinical studies currently ongoing [108].

5 Cytokine Release in AKI

The release of cytokines by injured/necrotic renal cells or resident leukocytes is a well-recognised component of early AKI. Subsequent signals produced by recruited leukocytes contribute to further inflammation and tissue damage [109]. A wide range of cytokines and chemokines has been detected in experimental animal models of AKI, with chemokine release occurring in both intrinsic renal cells and recruited leukocytes in response to reactive oxygen species, cytokine release, nuclear factor κ[kappa]B (NF-κ[kappa]B) activation and ligation of TLRs (summarised in Table 5.1) [110].

Table 5.1 Cytokine/chemokine release and receptor expression in experimental AKI
Full size table

The exact roles for each cytokine have been harder to define. An example of this is the case of IL-18 where studies in septic AKI and cisplatin-induced AKI show different outcomes to those seen with IRI [111,112,113] when the effects of IL-18 blockade have been investigated. These differences may reflect the different pathways of importance in the various experimental models of kidney injury or indeed the relative efficacy and off-target effects of blocking sera vs. genetic ablation in the models used. It is also worth noting that despite various successful interventions targeting cytokine signalling in rodent models of AKI, no therapies have yet proved successful in clinical trials, reinforcing the limitations of current in vivo experimentation in modelling the multifaceted aetiologies of AKI in man.

5.1 Factors Released by Injured/Necrotic Cells in Early AKI

Hypoxia and reoxygenation in the kidney can induce apoptosis or regulated necrosis (‘necroptosis’) via diverse mechanisms including consequent to mitochondrial injury [114]. While apoptosis is considered a non-inflammatory mode of cell death, necrotic death results in the release of damage-associated molecular patterns (DAMPs) , including alarmins and cytokines in the early aftermath of acute renal injury, all of which are capable of inducing further regulated cell death in the parenchymal cell pool [115, 116]. As indicated previously, release of free heme moieties in the context of cell stress acts as an additional source of oxidative damage to neighbouring cells, with the potential for further propagation of cellular injury and death if it is not metabolised via the induction of protective HO-1 [117].

Both the tumour necrosis factor (TNF-α[alpha]) and interferon-γ[gamma] have recently been shown to be capable of inducing necroptosis [118], while TNF-α[alpha] and interleukin-18 can induce neutrophil cell death via neutrophil extracellular traps (NETosis) [119]. The release of extracellular histones (a DAMP) has also been shown in experimental murine systems to be capable of inducing necrotic cell death within the kidney [119].

5.2 Cytokine Production in Senescent or G2/M-Arrested Tubular Cells

Paracrine signalling is a recognised feature of senescent tubular cells expressing p16INK4a [120]. These cells accumulate with ageing and renal injury. They produce TGF-β[beta],HGF, IGF-1 and VEGF and promote fibrogenesis and further senescence and are associated with impaired tubular proliferation. Studies in both progeroid and wild-type aged mice demonstrate that depletion of p16INK4a+ve senescent cells delays ageing and increases healthy lifespan [121, 122].

Transgenic mice expressing the simian diphtheria toxin receptor allow the study of the effects of selective, repeated tubular injury on the kidney signalling, function and scarring. Repeated tubular injury results in tubular, vascular and glomerular loss with increased fibrosis [123]. The response of tubular cells to acute injury, including expression of TGF-β[beta] acting in a paracrine and autocrine way, has been implicated as important in determining fibrotic outcomes and eventual glomerulosclerosis [124, 125].

Our group has demonstrated that models of severe and progressive renal injury (including severe IRI) show accumulation of tubular cells in the G2/M phase of the cell cycle [126]. These cells adopt a pro-fibrotic profile in vivo and in vitro, secreting growth factors including connective tissue growth factor (CTGF) and TGF-β[beta]1 in addition to increased collagen 4 α[alpha] 1 and 1 α[alpha] 1 mRNA [126] (Fig. 5.3). Thus, cell cycle arrest can induce paracrine signalling from tubular cells themselves, which can be a key component of the fibrotic response to renal injury, a hypothesis supported by recent reports from other laboratories [127,128,129]. Pharmacological inhibition of G2/M-arrested cells reduced fibrosis, whereas increases in the G2/M-arrested proportion of cells in the cell cycle exacerbated fibrosis [127,128,129].

Fig. 5.3
figure 3

Paracrine effects of the secretory phenotype seen in G2/M-arrested tubular epithelial cells. The proportion of tubular cells in the G2/M phase of the cell cycle increases in response to severe or sustained renal injuries. These cells produce factors, including CTGF and TGF-β[beta]1. These exert paracrine effects on both neighbouring tubular cells as well as on the phenotypes and numbers of interstitial cells, endothelial cells and macrophages. There is deposition of collagen and other matrix material within the kidney

Full size image

6 Inflammation in AKI

There is a growing body of evidence suggesting that the disruption of renal architecture and function seen in acute tubular necrosis may have a more prominent inflammatory aetiology than initially believed [130]. Renal IRI is associated with the production of numerous proinflammatory cytokines and chemokines from the kidney, which are chemotactic for leukocytes, including macrophage inflammatory protein 1 α[alpha] (MIP-1α[alpha]), monocyte chemotactic protein-1 (MCP-1), regulated upon activation normal T-cell expressed and secreted (RANTES), interleukins 1 and 6, the CXC cytokine KC and tumour necrosis factor-α[alpha] (TNF-α[alpha]) [131,132,133,134,135,136]. Several types of leukocytes have been implicated as potentially pathogenic and are discussed below.

6.1 Neutrophils

The neutrophil has been identified as infiltrating early into the post-ischaemic kidney [137] and is often measured as a surrogate of tissue injury severity [110]. However studies employing antineutrophil strategies have not shown consistent protection. Part of the explanation for differences may relate to the species studied (e.g. rat vs. mouse) and incomplete neutrophil depletion strategies. It is possible that the volume of neutrophil recruitment is a marker of injury severity rather than the driver of the process. While blocking cytokines or integrins to reduce neutrophil ingress have demonstrated functional protection in animal models [138, 139], two different depleting antibody approaches failed to protect the post-ischaemic kidney from injury despite alterations in neutrophil number [140, 141]. Additional recent work studying experimental cisplatin nephropathy showed no change in renal injury levels despite 90% neutrophil depletion [142]. Transplanted kidneys from aged animals demonstrate higher levels of tubular injury, despite neutrophil influx equivalent to younger kidneys [143], and a multicentre human renal transplant trial showed no benefit from neutrophil inhibition via ICAM-1 blockade [144]. As such any role for the neutrophil as the major effector of renal IRI remains unproven and controversial [145].

6.2 Lymphocytes

A number of studies have presented results demonstrating the involvement of the lymphocytes , components of the adaptive immune system, in the initiation of IRI. Several studies have demonstrated profound functional protection from IRI in rodent models with either genetic or pharmacologic depletion of T lymphocytes, [146,147,148,149,150], with a similar magnitude of protection seen in μ[mu]MT mice lacking B lymphocytes [151].

The actual effector mechanisms underlying the pathogenic role of lymphocytes in IRI remains unclear, although IFN γ[gamma] production by CD4+ cells has been implicated [146]. Wild-type serum, but not cell transfer into μ[mu]MT animals, restores the injury phenotype, suggesting a soluble factor originating in B lymphocytes [151]. The complexity of the system is emphasised by the conflicting data relating to the RAG-1 knockout (KO) mouse. Although the mouse is deficient in both T and B lymphocytes, there is no resultant protection from IRI [152]. Interestingly, injury is ameliorated by adoptive transfer of wild-type lymphocytes [147].

Studies have reported the key role played by T regulatory lymphocytes (Tregs) in determining the recovery from ischaemic renal injury [153]. Tregs are present in augmented numbers within the post-ischaemic kidney at both 3 and 10 days post-injury. Furthermore, when anti-CD25 antibodies were employed to deplete Tregs in vivo in the recovery phase of IRI, there were a persistence of structural injury and a reduction of tubular proliferation in both cortical and medullary compartments.

6.3 Macrophages

The macrophage (MΦ[phi]) is well characterised as a key effector cell in several models of renal inflammation [154,155,156]. Furthermore, in a model of lymphocyte-mediated injury such as renal transplantation, it has been demonstrated that targeting MΦ[phi] for depletion enhances allograft survival [155]. It is known that MΦ[phi] enters the kidney in the aftermath of IRI [137], and recent studies have demonstrated that ablation of renal MΦ[phi] can improve outcome in models of renal IRI [157, 158], glomerulonephritis [159] and fibrosis [160].

The role of the MΦ[phi] in renal IRI remains incompletely understood. One report of their depletion in rats showed improved long-term recovery after IRI [161]. This stands in contrast to evidence from other renal models in mice showing that depletion of MΦ[phi] prevented repair in mice after IRI [162]. Thus MΦ[phi] represents key contributors to the successful resolution of renal inflammation but if persisting can contribute to renal fibrosis (reviewed in [163, 164]).

Macrophages have conventionally been described as either classically (‘M1’) or alternatively (‘M2’) activated in response to stimuli. Classically activated MΦ[phi] was thought to arise in response to interferon-gamma and LPS stimulation, produce NO and had been demonstrated to be pathogenic to renal cells both in vitro and in vivo [165,166,167,168]. Alternatively activated macrophages were thought to be induced by IL-4, IL-10 and IL-13 and stimulate cellular proliferation and collagen deposition [169, 170]. However, more complete characterisation of the various functional states of tissue MΦ[phi] has confounded attempts to subdivide these cells into two discrete-activated subgroups. A recently proposed revision to this classification scheme again suggests a more fluid ‘spectrum’ of activation states based on cytokine profile [171] (summarised in Fig. 5.4). Briefly, this variability allows for macrophages to respond to various environmental and cytokine cues to adopt features of ‘classical activation’, ‘regulatory’ and ‘wound healing’ MΦ[phi] subsets, with provision for further alteration in the expressed characteristics on the basis of subsequent interactions.

Fig. 5.4
figure 4

This diagram summarises the proposed classification of macrophage phenotype proposed by Mosser and Edwards. Solid black arrows indicate cellular/cytokine stimuli for each polarisation, and dotted lines indicate the plasticity which exists between all activation states. Adapted from Mosser DM, Edwards JP. Exploring the full spectrum of macrophage activation. Nat Rev Immunol. 2008 Dec;8(12):958–69. PubMed PMID: 19029990. Pubmed Central PMCID: 2724991. Epub 2008/11/26. eng.; used with permission

Full size image

There is circumstantial evidence supporting a role for classically activated MΦ[phi] in the evolution of renal IRI. T-cell-derived IFN-γ[gamma] is of documented importance in IRI pathogenesis [146]—a cytokine associated with classical activation of MΦ[phi]. Classically activated MΦ[phi] has been shown to be pathogenic to renal cells both in vitro and in vivo [165,166,167,168]. It is recognised that MΦ[phi] phenotype is modified within a hypoxic microenvironment, such as a tumour [172] or an ischaemic organ [173].

Previous work has demonstrated the pathogenic significance of MΦ[phi]-derived nitric oxide (NO) production as a death signal for resident renal cells both in vitro and in vivo [166, 167]. NO is implicated in the evolution of renal injury [174], and it is interesting that inhibition of NO generation has also been shown to be protective in models of IRI [175,176,177]. The leukocyte Fc receptor is recognised as an important mechanism for MΦ[phi] interaction with deposited immunoglobulins, an important step in the initiation and progression of renal inflammation [178, 179]. Activating FcR knockout mice exhibit reduced infarct size in models of cerebral ischaemia [180].

Until recently, a widely accepted view of the resident mononuclear phagocyte system held MΦ[phi] and dendritic cells (DC) as clearly distinct in terms of cellular function while occupying overlapping anatomical sites in peripheral tissues and the reticuloendothelial system. In practice, distinguishing between macrophages and dendritic cells has relied on the use of cell surface markers thought to be specific to either cell. The progressive refinement and increasing number of available markers have served to complicate rather than simplify our understanding of the renal mononuclear phagocyte system.

Increasing the characterisation of both the surface markers expressed by MΦ[phi] and DC, and further characterisation of the functional capabilities and phenotypic plasticity of both cells, has led to an ongoing blurring of the previously accepted ‘unique’ characteristics of each lineage [181]. It is now recognised that activated macrophages can demonstrate antigen presentation and co-stimulatory capacity [182], while DCs are capable of NO production and cytotoxicity [183, 184]. This has led to the emergence of an alternative conceptual view where a common myeloid progenitor gives rise to a spectrum of mature cells within tissues adapted to diverse roles with overlapping functions, rather than two functionally distinct lineages [185, 186].

7 Toll-Like Receptors in the Kidney and AKI

7.1 Toll-Like Receptors in the Kidney

The toll gene was first described in Drosophila melanogaster in 1985 [187]. Toll-like receptors (TLRs) have been identified in humans and multiple vertebrate and invertebrate species in subsequent years [188, 189]. In humans and experimental mice , TLRs are expressed by cells of the innate immune system, including macrophages, dendritic cells and natural killer cells, along with some epithelial, endothelial and mesenchymal cell populations. TLRs comprise a range of receptors to various pathogen-associated molecular patterns, which when triggered activate the innate immune response. In the kidney, epithelial and mesenchymal cells express TLRs 1 to 4 and 6, with similar expression patterns seen in both mice and humans [190].

7.2 TLR Activation in Response to Acute Kidney Injury

It has been demonstrated at both a transcriptional and protein level that TLR2 and TLR4 are upregulated in the rodent kidney in the aftermath of experimental IRI [191] and can be found at markedly increased levels at 5-day post-injury [192]. In the context of the acutely injured kidney, damage-associated molecular patterns (DAMPs) and alarmins released from necrotic renal cells are capable of ligating TLRs, activating MyD88-dependent and MyD88-independent signalling leading to NF-κ[kappa]B expression and a rapid activation and recruitment of innate immune cells [53].

Supporting the importance of TLR2 signalling in the evolution of IRI, studies using various approaches to block TLR2 signalling in one or more of the kidney and leukocyte populations demonstrated that renal TLR2 deficiency was sufficient to protect the kidney from both inflammation and injury [193]. Similarly, TLR −/− mice demonstrated reduced numbers of renal neutrophils and chemokines in the aftermath of injury and a reduced degree of renal injury and dysfunction after IRI [194], with chimeric mice suggesting an equal contribution from renal and leukocyte-derived TLR4 activation. TLRs are also triggered by both DAMPs and PAMPs in sepsis-induced AKI, with LPS triggering inflammatory activation via TLR4, a situation where TLR4−/− mice exhibit marked protection [189]. Haemolytic uraemic syndrome with associated diarrhoea (D + HUS) is a significant cause of AKI, particularly in the paediatric population. There is experimental evidence that the recognised nephrotoxicity of Shiga toxin in this disease is augmented by the presence of bacterial LPS [195] with studies in man demonstrating much higher TLR4 positivity in neutrophils in acute D + HUS patients [196].

7.3 Ligands Responsible for TLR Activation in the Kidney

The endogenous ligands activating inflammatory signalling via TLR ligation remain incompletely understood. Potential TLR4 ligands biglycan, HMGB1 and hyaluronan are all increased in the aftermath of experimental IRI [197]. Studies using chimeric animals with TLR4 deficiencies in either kidneys or leukocytes suggest that HMGB1 is released from injured intrinsic renal cells and activates leukocytes via TLR4 ligation [198]. Supporting this, blockade of HMGB1 via antibody administration protects against experimental IRI [199].

TLR signalling via DAMPs such as HMGB1 results in inflammatory macrophage activation and immune-mediated tissue damage in the early ‘amplification’ phase of kidney injury [53]. Subsequent macrophage phenotypic switching from pro-inflammatory M1 to a pro-repair M2 polarisation state has been shown to be a vital part of successful renal repair after experimental injury [162]. These pro-repair macrophages secrete IL-22 via a mechanism requiring TLR4 ligation [200]—suggesting that these pathways could have dual roles of importance in both injury and repair phases.

8 Complement Activation in the Kidney in AKI

8.1 The Complement System in Health and Disease

The complement pathway is an ancient, highly conserved component of innate immunity [201]. It is a complex system, which can be activated via three cascades, namely, the classical, alternate and lectin pathways . Complement activity is regulated by over 40 genes, but despite this complexity, each pathway ultimately converges on C3 and C5 cleaving them to form soluble C3a, C3b, C5a and C5b.

A key downstream effect of C3a and C5a is the generation of a pro-inflammatory and chemotactic environment. C3a and C5a bind to G protein-coupled anaphylatoxin receptors C3aR and C5aR1 which increase production of co-stimulatory molecules and interleukin 6 [202]. C3b or C4b may opsonise cell surfaces for macrophages and lead to the production of macrophage inflammatory protein-2 and keratinocyte-derived chemokine [203].

The second major effect of C5 cleavage is the ability of C5b to catalyse the formation of the C5b-9 membrane attack complex. The membrane attack complex can insert itself into cell membranes leading to cell lysis and contributing to an inflammatory milieu by activating intracellular pathways and synthesis of pro-inflammatory eicosanoids [204].

8.2 Activation of Complement in Acute Kidney Injury

The complement system is a dynamic equilibrium where there is a balance between continuous complement activation and regulation . Within the human kidney, there is continuous C3 activation, as demonstrated by C3d staining along the basement membranes of Bowman’s capsule and renal tubules in normal kidneys [205]. This activation is regulated by a variety of proteins including factor H, decay-accelerating factor, membrane cofactor protein, C4-binding protein and complement receptor 1-related gene/protein y [206].

Complement plays an important role in various disease processes , which affect the kidney and lead to acute injury. Complement activity is central in atypical haemolytic uraemic syndrome, ANCA-associated vasculitis and anti-glomerular basement membrane disease [207]. It is unclear whether the role of complement in these specific processes is novel or reflects the broader role of complement in AKI. Thus IRI is used experimentally as a generic injury model from which to draw broader conclusions.

In AKI, the renal microenvironment shifts the balance towards complement activation . Following ischaemia reperfusion, there is upregulation of tubular cell production of C3, which activates local complement [208,209,210]. Compounding this increased activation, ischaemic tubules lose their regulatory surface molecules [211, 212]. Other factors contributing to complement activation may include an acidic environment, circulating pathogens, vascular congestion, exposed intracellular antigens and microparticles [207, 213,214,215,216].

During AKI, complement activation occurs primarily through the alternative pathway [152, 217, 218]. Kidneys from patients with acute tubular necrosis had increased C3d deposition along a greater number of tubules than normal, but C4d was absent supporting the central role of alternative pathway activation [205]. In murine experimental IRI, in contrast to the heart, intestine and lung, tubular epithelial cells were the main structures damaged by complement-mediated attack, and the renal vasculature was spared [219]. The formation of the membrane attack complex may also contribute to the damaging effect of complement on the kidney [219].

Inhibiting complement remains an attractive avenue of potential intervention with inhibition reducing IRI-induced renal damage in murine models [205, 212, 220]. Difficulties in translating complement inhibition to human studies include the challenges of tissue penetration to the level of the tubule, the uncertainties surrounding timing and intensity of complement inhibition and the potential for serious adverse effects, most crucially infection [207].