Acute Kidney Injury and Maladaptive Tubular Repair Leading to Renal Fibrosis

Abstract

Purpose of Review

Despite improvements in acute kidney injury (AKI) detection, therapeutic options to halt the progression of AKI to chronic kidney disease (CKD) remain limited. In this review, we focus on recent discoveries related to the pathophysiology of the AKI to CKD continuum, particularly involving the renal tubular epithelial cells, and also discuss related ongoing clinical trials. While our focus is on injured renal tubular epithelial cells as initiators of the cascade of events resulting in paracrine effects on other cells of the kidney, the summation of maladaptive responses from various kidney cell types ultimately leads to fibrosis and dysfunction characteristic of chronic kidney disease.

Recent Findings

Recent findings that we will focus on include, but are not limited to, characterizations of 1) the association between cell cycle arrest and cellular senescence in renal tubular epithelial cells and its contribution to renal fibrosis; 2) chronic inflammation with persistent cytokine production and lymphocyte infiltration among unrepaired renal tubules; 3) mitochondrial dysfunction and a unique role of cytosolic mtDNA in fibrogenesis; 4) prolyl hydroxylase domain (PHD) proteins as potential therapeutic targets, and 5) novel mechanisms involving the Hippo/YAP/TAZ pathway.

Summary

Potential therapeutic options to address CKD progression will be informed by a better understanding of fibrogenic pathways. Recent advances suggest additional drug targets in the various pathways leading to fibrosis.

Introduction

According to the recently released 2019 executive summary of US Renal Data System (USRDS), the prevalence of chronic kidney disease (CKD) among Medicare patients has steadily climbed to 14.5% (1), a historical peak in the past decade, despite the increasing patient awareness of kidney disease, novel biomarker development, and governmental supports for education and research. In the clinical setting, we have, over the last decade, recognized the strong association between acute kidney injury (AKI) and the development of CKD or end-stage kidney disease (ESKD) (2). This association between injury and maladaptive repair is also reproducible in laboratories by multiple in vitro or in vivo models. However, due to the complexity among the many kidney cell types and lack of optimal CKD murine models (3), the pathophysiology of AKI to CKD transition has not yet been adequately elucidated. Given the importance as sites of injury and the fact that they occupy a large fraction of the cell mass in the kidney, proximal tubules have been the main focus of research on AKI to CKD transition,. Our laboratory has delineated differences between adaptive and maladaptive repair after kidney injury and described that maladaptive repair of proximal tubules after AKI can contribute to progressive renal interstitial fibrosis secondary to cell cycle arrest, profibrotic cytokine secretion, pericyte activation with myelofibroblast generation, inflammation, loss of peritubular capillaries, and production of extracellular matrix (4). In this brief review, we aim to summarize important findings during the past one year and shed some light on future research directions in mechanistic insight into fibrogenic repair that often follows kidney injury.

Cell cycle arrest and Cellular Senescence

Activation of DNA damage response (DDR) signaling is crucial for the reparative process in renal proximal epithelial cells after AKI. When not fully repaired, proximal tubules (PT) undergo cell cycle arrest at G2/M, likely a protective mechanism related to the maintenance of genomic stability (5, 6). This cell cycle arrest at G2/M, if it persists, however, leads to a profibrotic secretory phenotype and ultimately fibrosis and irreversible damage (7). Ataxia telangiectasia mutated and Rad3-related (ATR), an upstream enzyme in DDR working in concert with other sensor kinases, detects DNA strand breaks and further phosphorylates downstream checkpoint proteins. Although we have reported that inhibition of ATM reduces profibrotic factors in immortalized PT cell lines (7), the role of ATR in PT remained largely unknown until recently. Our laboratory generated PT-specific-Atr-knockout mice (ATR^RPTC−/−) and these mice demonstrated a more severe fibrotic phenotype in various AKI models (cisplatin, ischemia-reperfusion injury (IRI), and unilateral ureter obstruction (UUO)), suggesting that ATR activation after AKI protects against maladaptive tubular repair resulting in less fibrosis (8). Of note, there are clinical trials using ATR inhibitors synergistically with other chemotherapy for advanced cancers (e.g. NCT02723864, NCT04095273). Given the high risks of developing AKI among oncology patients, close monitoring of renal outcomes of patients receiving ATR inhibitors might be warranted.

The cellular senescence associated with cell cycle arrest after either severe injury or repeated bouts of mild or moderate injuries with increasing DNA damage over time can be considered an accelerated aging phenotype (4). Our laboratory recently reported that rapamycin (TOR)-autophagy spatial coupling compartment (TASCC), a complex promoting senescence-associated secretory phenotype (SASP) initially described in Ras-induced senescence (RIS), is present in G2/M arrested PT in murine AKI models and human CKD (9). We identified Cyclin G1 (CG1) is a key factor for TASCC formation, and reported that both deletion of CG1 globally or deletion of Raptor (a major component of TASCC complex) specifically in PT, significantly ameliorated the renal fibrosis. p16^Ink4a, a marker of cellular senescence, was upregulated in both biopsies of human diabetic kidney disease (DKD) samples and transplant renal grafts with poor function (10). Liu et al recently demonstrated that senescence occurs early, within 2–3 days, after acute injury (11). This group reported that this early senescence response is mediated by epithelial toll-like and interleukin 1 receptors (TLR, IL-1R) in a cell autonomous process. One attractive potential therapeutic intervention to reduce renal fibrosis is to target senescent cells after AKI. To date, this approach has resulted in conflicting results. Depletion of p16^Ink4a (+) senescent cells in folic acid (FA)-induced murine AKI model reduced fibrosis without reducing injury, whereas a reduction in senescent cells after injury by disrupting the interaction between FOXO4 and p53, demonstrated no or partial protective effects on renal fibrosis, consistent with a prior study using the UUO model (12). Other groups recently reported that injection of pLVX-shp16INK4A (p16-shR) plasmid after unilateral IRI successfully abrogated fibrosis exacerbated by Wnt9a overexpression (13), and preliminary studies from a phase II clinical trial using senolytic agents in DKD patients are promising (14).

Inflammation

A good deal of recent attention has focused on the role of inflammation and innate immunity in the processes involved in AKI to CKD transition. After kidney injuries, activation of TLRs in PT by proinflammatory chemokines and danger-associated molecular patterns (DAMPs) can trigger an excessive inflammatory response. In folic acid (FA)-induced AKI (11) and a mouse model of nephronophthisis (NPHP) (15), a model of a group of autosomal recessive diseases causing end-stage kidney failure, an increase of interleukin-1 (IL-1) was found to activate TLR/IL-1R on PT and facilitated interstitial fibrosis. Deletion of Myd88, a downstream effector of TLR/IL-1R pathway upstream of NF-kB, reduced kidney damage and fibrosis. These findings were consistent with our prior findings that IL-1 is important in fibrosis development using a novel human kidney organoid model for fibrosis (16). Interestingly, IL-1 was also found to modulate renal damage in the presence of immunoglobulin free light chains (FLCs), a pathogenetic component of multiple myeloma or lymphoproliferative diseases (17). Ying et al. identified a specific V_L domain of immunoglobulin light chain activates ROS production and the STAT1 pathway in renal PT cells, and subsequently induces IL-1β and TGF-β production leading to renal interstitial fibrosis. The authors also treated human kidney HK-2 PT cells with ruxolitinib, a tyrosine kinase inhibitor that targets JAK/STAT, and demonstrated a decrease in IL-1β/TGF-β production after exposure to immunoglobulin light chains. Therefore, combining FLC reduction therapy and JAK inhibitors might be an appealing treatment for patients developing severe renal impairment in the future. In patients with established CKD and high-risk for atherosclerosis, inhibition of IL-1β also demonstrated reduced major cardiovascular events based on the recently completed CANTOS (Canakinumab Anti-Inflammatory Thrombosis Outcomes Study) trial (18).

Activation of other inflammatory cytokine receptors with ligands such as tumor-necrosis factor (TNF) and interferons (IFNs) was associated with necroptosis, a form of regulated cell death (RCD) (19). Necroptosis is associated with activation of the “necrosome” consisting of phosphorylated receptor-interacting serine-threonine protein kinase (RIPK) and mixed lineage kinase domain-like (MLKL). This results in permeabilization of the cell membrane and subsequent cell death. Although it has been shown that inhibiting necroptosis improved renal outcomes in murine AKI models (20), two recent studies further established its effects on AKI to CKD transition (21, 22). After IRI, the upregulation of RIPK3 and MLKL in PT created crosstalk between PT and macrophages, where NLRP3 inflammasome activation and IL-1β secretion from macrophages increased necrosome formation in PT (21). As such, this uncontrolled inflammatory response after AKI promotes interstitial fibrosis and CKD. There are also data from a recent anti-fibrotic drug development program showing that reducing NLRP3 inflammasome activating signaling pathways resulted in less fibrosis in a UUO model (23). Necroptosis also plays a role in collecting ducts. Collecting duct (CD) principal cells (PC)-specific deficiency of Integrin-Linked Kinase (ILK), a scaffold protein on focal adhesions, led to increased tubular necroptosis and subsequent renal fibrosis (22). Further studies using kidney cell-specific knockout of necroptosis-related genes are required to further characterize the effects on renal fibrosis development.

Innate and Adaptive Immunity

Rapid immune responses resulted from AKI can be roughly divided into two main categories: innate (involving e.g. macrophages, complements, or neutrophils) and adaptive (lymphocytes) immunity. In some inherited or acquired complement-mediated disorders, unchecked complement activation can cause persistent inflammation and eventually irreversible damage, and the recent success using complement inhibitors to treat complement-mediated kidney diseases has shed some light on potential therapeutic effects on AKI to CKD progression. Although kidneys are not the primary site of complement production, an upregulated expression of complement pathways in tubular epithelial cells and their subsequent activation has been shown after AKI (24). In IRI (25) and DKD (db/db mice) murine models (26), blockade of C5aR or genetically deletion of C5aR significantly reduced profibrotic cytokines and interstitial fibrosis. Others also showed that inhibiting C1r serine protease (27), the initiator of complement activation in the classical pathway, or C3a/C3aR in the alternative pathways (28), mitigated interstitial fibrosis in various murine models.

Adaptive immunity likely is an important contributor to AKI to CKD transition. Using RNA-sequencing analyses at different time points in a murine IRI model, a clear signal related to adaptive immune responses was identified after weeks to months of IRI (29). A similar observation was recently reported by Chang-Panesso et al. using lineage-tracing (kidney injury molecule-1 (KIM-1)(+) cells after injury) and the translating ribosome affinity purification (TRAP) method, revealing a significant immune system-related phenotype at the RNA level after 7 or 14 days of IRI (30). For instance, two specific types of T cells, effector γδ T cells, a subgroup of T cells recognizing antigens more broadly as compared to classical αβ T cells, and mucosal-associated invariant T (MAIT) cells characterized by CD3⁺ TCR Vα7.2⁺ CD161^hi, have been known to be highly associated with chronic inflammatory conditions. In human kidney biopsies, effector γδ T cells and MAIT cells were significantly elevated in biopsies with interstitial fibrosis, suggesting these two T cell types that function as a bridge from innate to adaptive immunity were associated with CKD (31, 32). A higher expression of CD69, an activation marker of T cells, was upregulated in fibrotic kidneys along with other proinflammatory cytokines and colocalized with aquaporin-1 (+) tubular cells (32). Mehrotra et al. (33) also identified a certain subset of renal CD4⁺ T cells expressing Oral1, a store-operated calcium entry (SOCE) channel promoting IL-17 secretion, which plays an important role in AKI to CKD transition.

Renal tubular production of vascular endothelial growth factor (VEGF-C and VEGF-D) was recently implicated in lymphangiogenesis after AKI (34), which might explain at least in part the prolonged effects of AKI related to adaptive immunity (35). Lymphangiogenesis was observed in CKD kidney biopsies with intrarenal immune cell expansion, and blocking lymphangiogenesis in murine kidneys reduced renal fibrosis after UUO (36). However, conflicting data were presented at the American Society of Nephrology Kidney Week and published as an abstract demonstrating that inhibition of lymphangiogenesis exacerbates cisplatin-induced AKI model (37). Furthermore, how renal tubular epithelial cells, particularly PT cells, interact with infiltrated lymphocytes remains largely unknown. Our laboratory has previously described that KIM-1 expression on the PT cell apical membrane mediates phagocytosis (38) and more recently further identified a novel mechanism whereby PT-KIM-1 is involved in antigen presentation, facilitated by major histocompatibility complex (MHC). This leads to suppression of CD4+ T cells activation but an increase of regulatory T cells (Treg) recruitment and reinforces the important immunomodulatory role of PT (39). Taken together, these data suggest a better characterization of subtypes of lymphocytes in kidneys and the timing of their infiltration is urgently needed.

Mitochondria Dysfunction

PT mitochondrial dysfunction has been implicated in increased susceptibility to injury and fibrotic sequelae. There has been particular attention to maintaining mitochondrial biogenesis and integrity during injury and repair (40). Dynamin-related protein 1 (DRP1), a GTPase regulating mitochondria fission processes, causes mitochondrial fragmentation and apoptosis under stress. PT-specific deletion of Drp1 either before or after IRI promoted epithelial recovery with decreased fibrosis (41), suggesting that the loss of healthy mitochondria in PT accelerates fibrosis development. Wei et al. reported that upregulation of microRNA-688 (miR-688) decreased mitochondria fragmentation after AKI via suppressing mitochondrial protein 18 kDa (MTP18) and thereby had renoprotective effects. However, the authors did not look into the effects of miR-688 on renal fibrosis development (42).

Most recently, two groups concurrently introduced the hypothesis that the release of mitochondria DNA (mtDNA) in the context of AKI induces immune responses and AKI progression. Mitochondrial transcription factor A (TFAM), an essential protein bound to mtDNA, was found to be reduced in kidneys obtained from patients with CKD and murine fibrosis models. Downregulation of TFAM in PT was previously attributed to miR-709 overexpression in AKI models but its role in AKI to CKD transition remained unknown (43). In tubule-specific TFAM knockout mice (Ksp-Cre/Tfam^flox/flox), Chung et al. (44) demonstrated renal tubules had not only severe mitochondrial loss but a cytosolic translocation of mtDNA and activation of cGAS-stimulator of interferon genes (STING) DNA sensing pathway, increased cytokine production, macrophage infiltration, and fibrosis. Crossing Ksp-Cre/Tfam^flox/flox with STING^−/− mice or pharmacologically inhibiting the STING pathway both attenuated renal dysfunction and fibrosis. Similarly, in a cisplatin-induced AKI model, mtDNA dislocation facilitated by BAX-induced mitochondrial membrane permeabilization also triggered the STING pathway and subsequent inflammatory responses (45).

Energy to satisfy the high metabolic demands of PT is mainly supplied by mitochondria fatty acid oxidation (FAO) and subsequent ATP production. Derangement of mitochondrial lipid metabolism has been implicated in CKD progression (46). In diabetes, it has been established that increased circulating free fatty acids and triglycerides with incomplete free fatty acid oxidation (46), promotes mitochondrial ROS production due to imbalance of proteins regulating fatty acid transport and oxidation (47). Kruger et al. (48) genetically deleted carnitine acetyltransferase (CrAT) in mice (PT-CrAT mice), a key enzyme exporting partially oxidized fatty acid out of mitochondria, to recapitulate the substrate overload seen in DKD in proximal tubules. These mice developed substantial tubular apoptosis and fibrosis, which was hastened by high-fat-diet (HFD). Interestingly, the author also observed secondary glomerulosclerosis in PT-CrAT mice, which supported our previously proposed tubule-centered, rather than glomeruli-centered, DKD progression pathobiology (40). On the other hand, insufficient FAO within PT can also exacerbate renal fibrosis. Reduced phosphorylation of acetyl-CoA carboxylase (ACC) has been found in PT cells from murine UUO kidneys with interstitial fibrosis (49). Unphosphorylated ACC resulted in increased malonyl CoA that blocks the carnitine palmitoyl transferase (Cpt1), an mitochondrial surface enzyme important for movement of fatty acids into mitochondria. Hence, phosphorylation of ACC enhances mitochondrial FAO and protects against fibrosis after injury. Several studies have suggested anti-fibrogenic effects of metformin, and this might be mediated by metformin-induced phosphorylation of ACC by AMP-activating protein kinase (AMPK), leading to active FAO metabolism to support renal tubular regeneration (49).

Hypoxia

PT cells utilize ATP generated by the mitochondrial respiratory chain for filtrate reabsorption facilitated by the Na-K-ATPase, and therefore, are highly oxygen-dependent. Unsurprisingly, lack of oxygen delivery during AKI can lead to mitochondrial dysfunction, paradoxical ROS overproduction, and inflammation (50). Endothelial inflammation and dysfunction lead to peritubular capillary rarefaction which further increases tissue hypoxia and accelerates this vicious cycle. Moreover, a recent paper revealed a link between hypoxia and cell cycle arrest after AKI in renal tubular epithelial cells. Liu et al. reported that with hypoxia, a significant increase of miR-493 suppressed the expression of Stathmin (STMN)-1, a cell cycle regulator, and induced G2/M cell cycle arrest and profibrotic cytokine release in vitro. After administering adeno-associated virus (AAV) to silence miR-493 in the murine UUO model, the author observed an enhanced expression of STMN-1 and less renal fibrosis. The effects of cell cycle distribution after AAV exposure were not reported (51).

Hypoxia-induced factors (HIFs), a family of key transcription factors stabilized by escaping degradation under hypoxia, facilitate various downstream pathways in the kidneys including VEGF and erythropoietin (EPO) upregulation. Active HIFs are heterodimers composed of a varied α subunit (HIF-1/2/3α) and a shared β subunit, which is also designated as HIF-β or Aryl hydrocarbon Receptor Nuclear Translocator (ARNT). Under normoxia, HIF-1α is hydroxylated by prolyl hydroxylase domain (PHD) proteins, subsequently bound to VHL E3 ubiquitin ligase and eventually degraded in proteasomes (52). Whether HIF-1α is protective after AKI, however, has been an ongoing debate due to its profibrotic profile via pathways such as G2/M cell cycle arrest (53, 54) in multiple murine CKD/DKD models. Li et al. (55) recently demonstrated that under hypoxia, HIF-1α binds to FoxO3, a stress-responsive transcription factor previously shown to upregulate autophagy after UUO, and subsequently inhibits FoxO3 prolyl hydroxylation and degradation by the ubiquitin-proteasome system (UPS). Tubular deletion of FoxO3 (Pax8-rtTAx FoxO3^fl/fl) exacerbated interstitial fibrosis after IRI with reduced autophagy and more severe oxidative injuries, suggesting a renoprotective role for FoxO3 after AKI. Notably, both HIF-1α and FoxO3 undergo prolyl hydroxylation under hypoxia but the hydroxylation may be mediated by different isoforms of PHDs. There are currently several PHD inhibitors undergoing clinical trials to treat anemia in patients with ESKD or CKD (56). Further studies would be essential to relate potential beneficial effects on hemoglobin levels with potential effects on AKI to CKD transition and will shed light on the ongoing controversy relating to possible safety concerns with HIF-1α stabilizing agents use. Besides its role of forming a heterodimer with HIF-1α under hypoxia, ARNT (HIF-1β) was recently uncovered as a new therapeutic target to attenuate renal fibrosis by its homodimerization (57). Intrigued by an early report (58), Tampe et al. discovered that low doses of FK506, a frequently prescribed transplant medication, can disrupt FKBP12/YYI complex and subsequently upregulate ARNT transcriptionally. Increased homodimeric ARNT, in turn, activated ALK transcription, a pivotal protein in the canonical BMP pathway, to promote tubular regeneration after injury. Both a low dose of FK506 and a small molecule targeting the FKBP12/YYI complex ameliorated the fibrotic phenotype after UUO in mice. Clinically, the BMP pathway activation has been identified as a key mechanism causing pulmonary arterial hypertension, and the use of FK506 in patients with such conditions has shown early promising results (NCT01647945). Therefore, treatment with low-dose FK506 might be a potential approach to prevent CKD progression in the future.

Hippo/YAP/TAZ Signaling

Secretion of profibrotic cytokines such as TGF-β and connective tissue growth factor (CTGF) after AKI has been well-known to contribute to the progression of renal fibrosis via paracrine effects on adjacent fibroblasts. Anorga et al. (59) recently reported that injury resulted in elevated TGF-β and this, in turn, upregulated the yes-associated protein (YAP)/transcriptional coactivator with PDZ-binding (TAZ) pathway. The YAP/TAZ pathway is essential for abnormal tissue repair and facilitation of tumorigenesis. There was also enhanced CTGF expression via SMAD3 activation in renal epithelial cells. Activation of the TAP/TAZ pathway was also shown to accelerate activation of Gli1+ myofibroblasts (60). Interestingly, overexpressing TAZ alone in HK-2 cells recapitulated G2/M cell cycle arrest, dedifferentiation, and promoted profibrotic cytokine production, suggesting a unique role of Hippo-TAZ pathway independent of TGF-β signaling. The blockade of Src-mediated YAP activation also reduced fibrosis in the UUO model (61).

Conclusions

In this review, we have attempted to highlight recent data relating to the mechanisms of AKI to CKD transition, and in particular, data related to some ongoing clinical applications and potential therapeutic options. Because of space constraints we could not include many other contributions and for this we apologize to the authors who have reported important non-included studies. Currently, collaborations such as the NIDDK-funded ReBuilding a Kidney (RBK) or Kidney Precision Medicine Project (KPMP) consortia have been launched to interrogate existing data systemically and obtain new data with human samples with the hope of complementing individual investigator-based studies to close important knowledge gaps. Importantly, the human tissue model of 3D human organoids has been applied to the study of various human kidney diseases such as polycystic kidney disease (PKD) (62). More recently these techniques have been applied to fibrotic diseases generated from genetic causes including mucin-1 kidney disease (MCD) (63) and cytokine (IL-1β) mediated processes (16), demonstrating the potential of using human kidney organoids to study CKD pathophysiology. Using kidney organoids derived from human inducible stem cells (iPSCs) of CKD patients may also provide personalized treatments and understand unforeseen adverse drug reaction by either in vitro or murine models. (64). With increased recognition of the growing impact of CKD on the world’s adult population (65), a sense of urgency is clearly apparent and better understanding of the effects of AKI on CKD and its progression is central to finding new solutions to combat kidney disease progression. In our opinion next-generation therapies for kidney diseases will soon lead to a brighter future for our patients who have undergone kidney injury.

Scheme of novel pathways in AKI to CKD progression. (1) Cell cycle arrest and senescence: in overt renal tubular injuries, cellular senescence/TASCC formation is present in proximal tubular cells, resulting in a profibrotic cell cycle arrest at the G2/M phase. Deletion of DDR protein such as ATR in proximal tubules exacerbates maladaptive repair. (2) Inflammation: inflammatory responses after AKI upregulates cytokine/chemokine release including IL-1 and TNF-α, which subsequently activates downstream MyD88-mediated NF-κB production and necrosome formation. Complement activation, increased lymphangiogenesis and persistent infiltration of immune cells, also have been implicated in AKI to CKD transition. (3) Hypoxia: reduced O₂ delivery and vascular rarefaction contribute to hypoxia-mediated tubular injuries. Increased miR-493 is associated with G2/M arrest, and the deletion of FoxO3 leads to increased autophagy and oxidative stress. (4) Hippo/YAP/TAZ pathway: secretion of TGF-β after AKI increases the nuclear accumulation of TAZ, and activates myofibroblast in concert with CTGF, leading to maladaptive repair and interstitial fibrosis. (5) Mitochondrial dysfunction: dysregulated fatty acid oxidation (FAO) leads to reduced ATP generation and excessive oxidative stress. Lack of phosphorylated ACC increased malonyl-CoA formation, which inhibits free fatty acid utilization by blocking CPT1 transporter. Translocation of mtDNA during tubular injury triggers the STING pathway, enhancing cytokine production and inflammation. ATR, Ataxia Telangiectasia mutated and Rad3-related; TASCC, rapamycin (TOR)-autophagy spatial coupling compartment; RIPK1, receptor-interacting serine-threonine protein kinase; MLKL, mixed lineage kinase domain-like; STMN-1, Stathmin-1; TGF-β, transforming growth factor-beta; CTGF, connective tissue growth factor; YAP, yes-associated protein; TAZ, transcriptional coactivator with PDZ-binding motif; ACC, acetyl-CoA carboxylase; CPT, carnitine palmitoyl transferase; STING, cGAS-stimulator of interferon genes.