KIDNEY REPAIR AND REGENERATION: PERSPECTIVES OF THE NIDDK (RE)BUILDING A KIDNEY CONSORTIUM

Acute Kidney Injury (AKI) impacts ~13.3 million individuals and causes ~1.7 million deaths per year globally ¹. Numerous injury pathways contribute to AKI including cell cycle arrest, senescence, inflammation, mitochondrial dysfunction, endothelial injury and dysfunction and can lead to chronic inflammation and fibrosis. However, factors enabling productive repair vs. non-productive, persistent injury states remain less understood. The (Re)Building a Kidney consortium (RBK) is an NIDDK consortium focused on both endogenous kidney repair mechanisms and the generation of new kidney tissue ². This short review provides an update on RBK studies of endogenous nephron repair, addressing the questions 1) What is productive nephron repair? 2) What are the cellular sources and drivers of repair? and 3) How do RBK studies promote development of therapeutics? Also, we provide a guide to RBK’s open access datahub for accessing, downloading, and further analyzing data sets.

WHAT IS PRODUCTIVE VS. NON-PRODUCTIVE NEPHRON REPAIR?

Mouse models show that nephron tubules have an intrinsic regenerative potential ³ however upon repeated injury, cells can be diverted to a non-productive senescent state. RBK research groups focus on ischemia reperfusion injury to induce acute kidney injury since it is the most common cause of AKI and, as an experimental model in mice, can be performed with variable ischemic durations to produce mild or severe tubule injury and model the AKI-to-CKD progression. Whole kidney RNA-sequencing characterization after murine ischemia reperfusion following recovery from IRI for twelve months provided clues about repair processes and the AKI-CKD transition ⁴ (Figure 1). Principle component analysis uncovered a cascade of gene expression patterns that mapped to tubular injury/repair, fibrosis, and innate and adaptive immunity processes. Specific genes, like Krt20, a kidney injury marker, tracked progressive proximal tubular injury while genes related to cell cycle and wound repair were transiently upregulated after IRI. Sox9, a transcription factor linked to proximal tubular repair and unresolved injury, remained persistently upregulated over the course of recovery. This study highlighted co-regulated gene clusters associated with the progression of AKI to CKD and revealed that the AKI-CKD transition can occur after a single ischemic episode ⁴. In subsequent studies, single-nucleus RNA sequencing (snRNA-seq) identified distinct, late-stage injury, proximal tubule cell states (Vcam1⁺/Ccl2⁺ with upregulation of NF-κB-, TNF-, and AP-1 signaling pathways) associated with the AKI-CKD transition and a senescence-associated secretory phenotype (SASP) ^{5, 6}. G2-M phase cell cycle arrest and activation of Cyclin G1 pathways in injured tubules has been associated with the formation of target of rapamycin-autophagy spatial coupling compartments (TASCCs) and cytokine production ^{7, 8} although no evidence of G2-M phase arrest was found in proinflammatory subpopulations of injured PTCs in single nucleus cell RNA-seq studies ⁵. Proliferative repair itself may impose an energy burden on dividing tubule cells that correlates with decreased expression of mitochondrial genes and increased susceptibility to cell injury ⁹. While the proximal tubule is the nephron segment most acutely sensitive to AKI, single cell and spatial transcriptomics also reveal gene expression changes in more distal segments (TAL, DCT) for the injury markers Umod and Lcn2 ^{6, 10}. Non-tubule cell types, including peritubular capillary endothelial cells and vascular pericytes are also central to recovery from injury. Recent work has shown that failure to reconstitute PT capillary networks may be a better predictor of CKD with reduced GFR than renal fibrosis in multiple AKI models (IRI, cisplatin and rhabdomyolysis) ¹¹.

Figure 1. Rodent models have revealed important insights of how maladaptive repair of AKI leads to CKD.

After injury, cells may become arrested in G2/M phase, leading to a maladaptive or failed repair state characterized by activation of NfKB signaling and expression of failed repair state markers NfKb1, Vcam1, Dcdc2a, Sema5a, and Ccl2. Maladaptive repair is also associated with endothelial-pericyte dissociation, microvascular rarefaction, and collagen I deposition by myofibroblasts derived from the activated pericytes. Progression to fibrotic scarring is associated with a Senescence Associated Secretory Phenotype (SASP), continued release of pro-inflammatory growth factors and cytokines and abundant expression of Col1a1 and Acta2. ^{4–6, 10}

Analysis of the PTC enhancer and super-enhancer landscape has also identified transcription factors associated with kidney repair, including HNF4A, GR, STAT3, STAT5 and BET proteins. Inhibition of bromodomain-containing protein 4 function with the BET inhibitor JQ1 impaired AKI recovery and increased mortality while reducing fibrosis, revealing an important role of bromodomain-containing proteins in multiple aspects of repair ¹². Translational profiling (TRAP) studies revealed a strong upregulation of FoxM1, which can act as an oncogene, after kidney injury ¹³. FoxM1 siRNA knockdown reduced epidermal growth factor-induced proliferation in cultured human PTCs, supporting a role in PTC proliferation.

RNA seq data sets for many of these experiments and others are available for download at the RBK Datahub site ¹⁴. Figure 2 demonstrates how RBK datahub RNA-sequencing datasets from a unilateral ureteral obstruction model allow for the visualization of annotated clusters. For instance, this data confirms that the chemokine Ccl2 is primarily expressed in dedifferentiated proximal tubule cells while Ccr2 and Ccr4 receptors are expressed in macrophages and activated fibroblasts, suggesting signaling relationships between these cell types that drive fibrosis. This data is part of a larger set of single-nucleus RNA sequencing and single-cell RNA sequencing data from healthy and fibrotic (UUO) mice kidneys and is accessible for further analyses at the RBK datahub (Figure 3)¹⁵.

Figure 2. Leveraging the data hub to reveal intercellular communication networks.

A. Annotated clusters from a unilateral ureteral obstruction kidney analyzed by snRNA-seq. B,C. The chemokine Ccl2 is exclusively expressed in dedifferentiated proximal tubules which are known to upregulate CCl2 after injury. The Ccl2 receptors Ccr2 and Ccr4 are expressed in macrophages and activated fibroblasts, respectively, suggesting that dedifferentiated proximal tubule signals via Ccr2 to these two interstitial cell types. (https://doi.org/10.25548/14-4KG6)

Figure 3. Demonstrating how to access the RBK data for analysis.

The data shown in Figure 2 is available for access from the RBK’s data repository (www.rebuilidingakidney.org). To access it, navigate to the DOI: https://doi.org/10.25548/14-4KG6 (A) Once on the Collection page, click the “View” icon under the Sequencing Study Collection. This will navigate you to the Study detail page. (B) On the Study Detail page, one can access details on the experiment(s), the study analysis files (e.g., T-SNE images) & replicate-level files (including the raw fastq or bam files) available for download. To learn more about using the RBK data repository, visit our documentation at https://www.doi.org/10.25548/01-A001 (or click Help in the menu on www.rebuildingakidney.org). The documentation contains instructions on how to use the RBK data browser including instructive videos on how to find data. To access other types of RBK data, including sequencing, imaging, cell lines, metabolomics, antibody validations, protocols, and more, visit www.rebuildingakidney.org.

WHAT CAN WE LEARN FROM NON-MAMMALIAN MODEL SYSTEMS?

Zebrafish have been increasingly used as an AKI model. The zebrafish kidney provides insights into two modes of renal regeneration: 1) tubule repair and regeneration and 2) stem cell-mediated, new nephron formation (Figure 4). AKI studies of intrinsic tubule cell repair in the larval zebrafish pronephros have been useful for modeling mammalian AKI and performing small molecule screening studies ^{16, 17}. Gentamicin injection causes acute renal failure in larvae with histological and functional changes consistent with aminoglycoside toxicity in mammals ¹⁸. Similar to mammalian kidneys, injured larval pronephric kidneys express kim-1/havcr1 and have similar pathophysiological responses during tubular injury including loss of cell polarity, reactivation of pax2 expression and induction of vimentin, indicating a conserved epithelial-to-mesenchymal transition during tubular injury ¹⁹. Additionally, the larval AKI model can be utilized to screen for therapeutic candidates and compound efficacy translates well to mammalian AKI models ¹⁶. These therapeutic studies allow for rapid first-pass screening without requiring studies of compound pharmacokinetics as is needed in mammalian models ¹⁶.

Figure 4. The zebrafish larval and adult kidneys offer two different models of kidney repair after AKI.

(A) Zebrafish larval (pronephric) kidneys share many genes and critical pathways with the human kidney. A healthy pronephric proximal tubule is stained basolaterally by Na+/K+ ATPase and apically with the monoclonal 3G8. Upon injury, KIM-1 is present at the apical surface, with loss of basolateral Na+/K+ polarity and cell sloughing. During the repair process, reactivated cells actively migrate and stain positive for PAX2 in the nucleus. As cells proliferate, they stain positive for EdU and are found to stain positive for Vimentin, indicating mesenchymal dedifferentiation. (B) The adult zebrafish is a suitable model to study neonephrogenesis. Unlike the larva, adult zebrafish kidneys have a population of six2a+,eya4+, osr1+ fzd9b+ cells ²³ adjacent to the distal nephron that are associated with neonephrogenesis. The adult zebrafish kidney serves as a suitable model to understand stem cell-mediated kidney repair and how this may be engineered in the human kidney.

In the adult zebrafish kidney, overall body growth and acute kidney injury induced the generation of new nephrons from adult stem cells in a process termed neo-nephrogenesis ^20–22. New nephrons connect to pre-existing distal tubules, replicating mesonephric development ^{20, 21}. Adult kidney stem cells have been identified by two methods: transplantation of cells marked by Tg(lhx1a:EGFP) expression ²⁰ and scRNAseq clustering of adult kidney cells that express the nephrogenic mesenchyme markers six2a, eya4, and osr1 ²³. Upon injury, these cells expand in number, migrate, and aggregate as domed clusters on existing tubules to form nephron precursors (Figure 4) ^{20, 21} in a process requiring FGF (chemotaxis, cell polarization ²²) and Wnt signaling (nephron patterning, cell proliferation ²¹). While mammalian kidneys lack adult stem cell mediated repair, both fish and mammals activate growth factor expression after injury ²⁴ which has been associated with both pro-regenerative and pro-fibrotic, context-specific effects ²⁵. Nonetheless, the zebrafish model indicates that a pro-regenerative growth factor environment along with responsive nephrogenic cells can lead to new nephron formation in an adult kidney.

WHAT ARE THE CELLULAR SOURCES OF REPAIR IN TUBULAR AND GLOMERULAR INJURY?

Mouse studies showed that tubular repair following AKI is largely, if not exclusively, mediated through replication of surviving proximal tubule cells with no strong evidence for contribution by a reparative cell type ^{26, 27}. Tubule repair occurs via a Pik3c3-mediated mechanism that involves sensing of amino acid availability and initiating hypertrophic growth of PTCs by activation of the MTORC1-S6K1-rpS6 signaling pathway ²⁸. Further, lineage analysis of SLC34A1 expressing PTCs suggest fully differentiated cells re-enter mitosis to directly replace PTCs after AKI ²⁹. In contrast, analysis of FUCCI transgenic mice, which visualize the cell cycle state, suggested that after injury, proximal tubule cells undergo endocycle progression to S phase, failing to divide, while a small subset of Pax2+/Havcr1− (Kim1−) tubular cells undergo clonal expansion for repair ³⁰. Countering this suggestion, analysis of DNA content in Havcr1+ PTCs indicated that the bulk of cells have a 2N DNA content, arguing against extensive endocycling ³¹. Recognizing alternative interpretations of these results, several RBK groups combined genetic strategies with cell profiling to characterize the AKI response and outcomes in the mouse kidney. Kusaba et al. created a CreERT2 knock in mouse at the Kim-1/Havcr1 locus and performed clonal analysis of injured Kim-1+ cells ³. Interestingly, Kim-1/Havcr1+ cells undergo clonal expansion during repair, counter to a model in which repair is directed by a minor Pax2+/Kim-1− intratubular progenitor type. Moreover, the acutely injured, genetically labeled clones co-expressed Kim1, Vimentin, Sox8, and KI67, indicating a dedifferentiated and proliferative state ¹³.

In glomeruli, terminally differentiated podocytes cannot self-renew ³², whereas endothelial, mesangial, and parietal epithelial cells (PECs) can. Although there is currently no known stem/progenitor cell for glomerular endothelial cells, cells of renin and NG2 lineage can partially replace mesangial cells and parietal epithelial cells ³³. A decrease in podocyte cell number drives the magnitude of glomerular scarring and proteinuria and even small increases in podocyte number can limit scarring ³² which might be considered a clinical goal. Candidate podocyte progenitors include cells of renin lineage (CoRL) and parietal epithelial cells (PECs). In addition to their well-known endocrine function, CoRL are also multi-potent stem cells. Lineage tracing studies show that CoRL can differentiate to a podocyte fate, a process that can be augmented with ACE-inhibitors and ARBs ³⁴. PECs are another adult podocyte progenitor source in disease that can differentiate to a podocyte fate, which can be augmented by several therapies ³⁵. PECs are specialized cells lining the glomerular capsule and exist in different subpopulation. Even though not conclusively shown, PECs appear to constitute their own cellular niche, which can self-renew and be amplified upon stress or disease ^{36, 37}. RBK consortium studies are focused on better understanding the mechanisms of podocyte regeneration by CoRL and PEC progenitors and developing tools to aid in this investigation.

INJURY-INDUCED ECM CHANGES AND TARGETING THERAPEUTICS

In AKI, renal ischemia triggers a cascade of cellular injury and extracellular matrix (ECM) production. By isolating the ECM binding domains of naturally occurring growth and hemostatic factors, native and fibrotic kidney ECM may be targeted. Collagen binding peptides isolated from decorin by the Hubbell group demonstrated 2x greater biodistribution to healthy kidneys via IV injection and 5x greater biodistribution to a fibrotic kidney (7 days post UUO) ³⁸. Moreover, blocking antibodies against three different sub-types of integrins led to de-differentiation of myofibroblasts and the reversal of fibrosis in the UUO kidney ³⁸. Continuing studies examining how kidney injury and disease states change the ability of various ECM-binding peptides to target the kidney provide an opportunity to exploit CKD related pathology for therapeutic advantage.

RBK transcriptomic analysis of AKI is also directing approaches to therapy. The “failed repair” cell state associated with AKI in mice after IRI has now been defined by unique molecular markers and activated signaling pathways ^5–7. Elimination of FR-PTCs using senolytic approaches has strong potential to benefit patient outcomes post AKI ³⁹. Small molecule screens using the zebrafish model of larval AKI has also identified lead therapeutic compounds that have been validated in mouse models of AKI ^{16, 19}. Ongoing RBK studies of sex differences in susceptibility to CKD hold the potential to better define tailored treatments for AKI. Additional studies in progress aim to increase glomerular regeneration in vivo using cell-targeted delivery of novel peptides and small molecules to drive expansion and differentiation of parietal epithelial cell podocyte progenitors ³⁶.

CONCLUSION

Kidney cell types all show varying potential for endogenous repair and tissue replacement. Countering these beneficial responses are non-productive, proinflammatory cell states and limited availability of replacement progenitor cells. Using multiple models of injury, multi-omics approaches, and an open platform for data sharing, the RBK consortium aims to accelerate the pace of discovery and bring us closer to successfully repairing and regenerating injured kidneys.