Interleukin-1β Activates a MYC-Dependent Metabolic Switch in Kidney Stromal Cells Necessary for Progressive Tubulointerstitial Fibrosis

Abstract

Background Kidney injury is characterized by persisting inflammation and fibrosis, yet mechanisms by which inflammatory signals drive fibrogenesis remain poorly defined.

Methods RNA sequencing of fibrotic kidneys from patients with CKD identified a metabolic gene signature comprising loss of mitochondrial and oxidative phosphorylation gene expression with a concomitant increase in regulators and enzymes of glycolysis under the control of PGC1α and MYC transcription factors, respectively. We modeled this metabolic switch in vivo, in experimental murine models of kidney injury, and in vitro in human kidney stromal cells (SCs) and human kidney organoids.

Results In mice, MYC and the target genes thereof became activated in resident SCs early after kidney injury, suggesting that acute innate immune signals regulate this transcriptional switch. In vitro, stimulation of purified human kidney SCs and human kidney organoids with IL-1β recapitulated the molecular events observed in vivo, inducing functional metabolic derangement characterized by increased MYC-dependent glycolysis, the latter proving necessary to drive proliferation and matrix production. MYC interacted directly with sequestosome 1/p62, which is involved in proteasomal degradation, and modulation of p62 expression caused inverse effects on MYC expression. IL-1β stimulated autophagy flux, causing degradation of p62 and accumulation of MYC. Inhibition of the IL-1R signal transducer kinase IRAK4 in vivo or inhibition of MYC in vivo as well as in human kidney organoids in vitro abrogated fibrosis and reduced tubular injury.

Conclusions Our findings define a connection between IL-1β and metabolic switch in fibrosis initiation and progression and highlight IL-1β and MYC as potential therapeutic targets in tubulointerstitial diseases.

Kidney disorders caused by both acute and chronic injury affect over 10% of the human population worldwide, imposing major costs on health care systems. Kidney injury initiates an acute inflammatory process,¹ during which innate immune cells recruited from the circulation release proinflammatory cytokines.^2,3 Among the most potent proinflammatory cytokines, IL-1 functions as an apical proinflammatory mediator, triggering cell stress and driving the synthesis and expression of several other secondary inflammatory mediators by the kidney epithelium and the interstitial stroma.^2,4,5

The progression of kidney injury is marked by pathologic fibrogenic cell activity, commonly leading to the accumulation of excess fibrotic tissue that replaces the parenchyma and resulting in scarring and loss of renal function.⁶ Although the signals that activate fibrogenesis have not been fully elucidated, fibrogenic cell proliferation temporally overlaps with the acute inflammatory phase triggered by nephron injury.² More specifically, several lines of evidence suggest that IL-1, which signals through the IL-1 receptor, and a signal transducing complex known as the Myddosome (comprising MYD88, IRAK1, 2, and 4), regulates kidney fibrosis. Global ablation of IL-1 receptor type I (IL1R1) resulted in significant reduction of interstitial stromal expansion and fibrosis early after acute kidney damage, with fibrogenesis increasing by day 14 postinjury, indicating a role for IL-1β in the initiation of permanent fibrosis.⁷ Similarly, kidneys from Il1r1 knockout (KO) mice present less fibrosis on transplantation into IL1r1 wild type (WT) mice.⁸ In those studies, genetic ablation of IL-1 signaling was global (i.e., in all cell types), and therefore, it is not clear whether impaired fibrosis is the indirect result of reduced epithelial cell damage and inflammation. More recently, a role for inflammatory signals driving fibrosis directly has been supported by experiments showing that ablation, or inhibition of IL-1 receptor–associated kinase 4 (IRAK4) signaling in mouse kidney stromal cells (SCs) results in reduced interstitial fibrosis after ischemic injury.⁵

The initiation of kidney fibrosis involves the activation of otherwise quiescent SCs in a process that encompasses cell growth, proliferation, and differentiation, three cellular processes associated with increased bioenergetic and biosynthetic demand. In the injured kidney microenvironment, populations of kidney resident SCs (also referred to as mesenchymal SCs), which include perivascular mesenchymal progenitor cells, pericytes, and interstitial fibroblasts, expand and differentiate into myofibroblasts that synthesize extracellular matrix and contribute to the inflammatory process through secretion of chemokines and cytokines such as IL-6.^5,9–12 Although the molecular metabolic mechanisms driving SC activity are not well understood, in other cell types, including T cells, the expansion phase that follows activation involves specific metabolic programs that support the accumulation of cell biomass during the initial growth and rapid proliferation.^13,14

Here, we identified a novel mechanism by which IL-1β dependent signaling drives kidney SC activity by inducing a proliferative and metabolic gene program dependent on the transcription factor MYC. By this mechanism, IL1β contributes to renal fibrosis.

METHODS

Human Kidney Transcriptome and Renal Function Datasets

Datasets from two different patient cohorts were used in this study. A dataset from a cohort of 38 patients diagnosed with various forms of CKD was collected under a sponsored research agreement at University of Pennsylvania. The dataset included eGFR and expression profiles derived from RNA-seq for each individual.¹⁵ The second dataset comprised previously reported RNA expression profiles from tumor nephrectomy specimens from six patients.¹⁶ As stated in the original study, the samples were generated from portions of non-neoplastic tissue within 15 minutes of surgical removal.¹⁶ Using next generation sequencing on the Illumina HiSeq 2000 platform producing 50-bp paired end reads, total reads were mapped using the aligner STAR 9. Genes were quantified with RSEM 10, and differential expression was determined by DESeq2. Differentially expressed genes (DEGs) were identified by comparing patients with no CKD with those with CKD in one cohort or comparing low fibrosis area on tissues sections (<10%) with high fibrosis area (>20%) in the other cohort. A promoter analysis of DEGs was performed using gene set enrichment analysis (Broad Institute) to identify transcription factors likely to be controlling the pattern of DEGs. Metabolism genes were defined from Ingenuity Pathway Analysis (Qiagen Bioinformatics). Trichrome-stained slides from formalin-preserved tissue were scored by an experienced kidney pathologist for the degree of interstitial fibrosis.¹⁶

Human Cell Purification, Culture, and CRISPR Mutation

SCs were purified from adult human kidneys obtained from nephrectomy specimens collected at Boston University or through a sponsored research agreement with New England Medical Center using methods previously described.^17,18 A single-cell preparation, excluding capsule and glomeruli, was generated by digestion with a collagenase cocktail (Liberase; Roche). Epithelial cells were depleted using anti-EpCAM and Lotus-lectin–conjugated beads in magnetic columns (Dynabeads). The remaining sample was used to prospectively purify PDGFRβ+ SCs by flow cytometry using directly conjugated anti-CD45 (clone HI30; Biolegend), anti-PECAM (clone WM59; BD Pharmingen), and anti-PDGFRβ (clone 18A2; eBioscience). Only samples with >98% purity were used in the experiments. SCs (P3–P5) were cultured in six-well gelatin-coated plates at a density of 4×10⁵ cells per 1 ml for 24 hours in pericyte medium (Lonza). For cytokine stimulation experiments, cells were washed five times with PBS and cultured overnight in serum-free medium. IL-1β was purchased from Sigma-Aldrich. The BRD4 inhibitor (+)-JQ1 (Selleckchem) and the MYC/MAX inhibitor 10058-F4 (Sigma-Aldrich) were added at the indicated concentrations. SCs deficient in sequestosome-1/p62 (SQSTM1) were generated using CRISPR/Cas9-mediated double-strand break and non-homologous end joining in the first exon. A combined pool of three plasmids each coding for GFP, Cas9, and a single unique sgRNA for the SQSTM1 gene (sc-400099; SCBT) was transfected into purified PDGFRβ+ SCs using the Neon electroporation transfection system (Invitrogen); 2.2 million cells were electroporated with 20 μg of total plasmid pool DNA and plated on a 0.2% gelatin-coated 100-mm dish. Cells were allowed to expand for 48 hours, and then, they were sorted by FACS using GFP as a positive marker for transfection. Absence of SQSTM1 was confirmed by western blot. In some studies, purified mouse PDGFRβ+, NG2+ SCs were treated with IL1β or vehicle for 16 hours, and purified mRNA was analyzed by Illumina micorarray to identify DEGs, as described.¹⁷

Human Kidney Biopsy Sample Preparation and Immunohistochemistry

Human tissues were fixed in periodate-lysine-paraformaldehyde (PLP) fixative for 2 hours, then in 20% Sucrose PBS for 24 hours. Tissue samples were mounted in OCT medium (Tissue-Tek) and sectioned using a cryostat. Primary antibodies against the following proteins were used for immunolabeling: MYC and pyruvate kinase isoform M2 (PKM2; both from Cell Signaling). Sections were incubated for 1 hour in biotinylated donkey anti-rabbit antibody (1:500; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA), washed, and then incubated in avidin-biotin solution (Vectastain; Vector Laboratories, Burlingame, CA) for 1 hour as described previously.¹⁹

Generation of Human Kidney Organoids and IL-1β Stimulation

Kidney organoids were generated from H9 human embryonic stem cells by the previously published protocol.^20,21 H9 was maintained in StemFit (Ajinomoto) supplemented with 10 ng/ml FGF2 (Peprotech). When cells reached 40%–50% confluency, differentiation was initiated with 8 μM CHIR (TOCRIS) as previously reported.²⁰ After 4 days of CHIR treatment, cells were differentiated with Activin for 3 days. Then, the medium was changed to one supplemented with FGF9, and cells were cultured for 1–2 days. After that, cells were dissociated with Accutase (Stem Cell Technologies) and plated onto 96-well plates with U-shaped bottoms (Corning) at 100,000 cells per well in a differentiation medium supplemented with 3 μM CHIR and 10 ng/ml FGF9. After 2 days, the medium was replaced with one supplemented with only 10 ng/ml FGF9, and cells were cultured for 3–4 days. FGF9 was removed after that, and kidney organoids further developed without additional growth factors. Organoids were allowed to mature until day 51 when IL-1β (Sigma-Aldrich) was added to the media. Additionally, 500 nM (+)-JQ1 was added to the media for the indicated incubation time. Organoids were collected and fixed in 4% PFA for 30 minutes followed by cryopreservation in 20% sucrose overnight. Seven-micron organoid sections were prepared by cryosectioning and incubated with antibodies for immunofluorescence. Images were captured by confocal microscopy using a Nikon C1 microscope running EZ-C1 software.

Animal Experiments

The AKI models used 8- to 10-week-old C57Bl/6 male WT mice (Charles River Laboratories). Col1a1-GFP mice²² were used to detect interstitial SCs in AKI. Unilateral ureteral obstruction (UUO) surgery was performed at 8–10 weeks of age as previously described.²⁰ Briefly, after flank incision, the left ureter was occluded at the level of the lower pole with two 4.0 silk ties. Kidney ischemia reperfusion injury (IRI) surgical procedures were performed as previously described.⁹ Daily doses of 50 mg/kg BRD4 inhibitor (+)-JQ1 were delivered by oral gavage. Mice were euthanized using CO₂. Analysis of the SC translatome from healthy and UUO kidneys by Translated Ribosomal Affinity Purification (TRAP) using Col1a1-L10A-GFP transgenic mice was previously reported,²³ and datasets were deposited to GEO DataSets. Microarray data were normalized by Robust Multichip Average using GeneSpring software (Agilent Technologies).²⁴ Microarrays have been deposited in the GEO DataSets (Geo IDGSM1219324 through GSM1219338). All experiments were performed under protocols approved by the Institutional Animal Care and Use Committee at Biogen.

Animal Tissue Preparation and Immunofluorescence

Mouse tissues were collected and stained as previously described.²² Briefly, tissues for cryosectioning (7 μm) were fixed in PLP solution for 2 hours, and then, they were washed in 20% sucrose solution overnight. Primary antibodies used in this study were against the following proteins: MYC, PDGFRβ (Cell Signaling), Ki67 (EMD Millipore), ACTA2/αSMA (Sigma-Aldrich), F4/80 (ThermoFisher), and KIM-1 (1:200; Bonventre Laboratory; R9).²⁵ Tissue sections were mounted in ProLong Gold Antifade Reagent with 4′,6-diamidino-2-phenylindole (Thermo Fisher Scientific). Images were captured by a confocal (Intelligent Imaging Innovation or Zeiss) or standard fluorescent microscope (Nikon) as previously described.²²

Quantitative PCR

Total RNA was extracted from cells or whole tissue using the RNeasy Plus Mini Kit (Qiagen). Purity was determined by A260–A280. cDNA was synthesized using oligo(dT) and random primers (iScript Reverse Transcription Supermix; Biorad). Quantitative PCR was performed using the QuantStudio 7 Flex Real-Time PCR System (Life Technologies) using the/react-text TaqMan react-text: 3672 Gene Expression Assays (Life Technologies). The specific primer pairs used in quantitative PCR are listed in Supplemental Table 1 (Life Technologies).

Western Blot

SCs, 1×10⁵ per well were seeded on 0.2% gelatin-coated 24-well plates and cultured overnight, then serum starved for 4 hours and stimulated with IL-1β for the duration of the experiment. SCs were lysed in RIPA buffer containing Halt Protease and Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific) on ice. After protein quantification, SDS-PAGE and western blotting were performed as described.²⁶ Antibodies against the following proteins: MYC, PKM2, p-P70S6k, P70S6K, p-S6, S6, p62, LC3 (all from Cell Signaling), and PGC1α (Santa Cruz) were used. Immunoprecipitation was performed as described previously using anti-human p62 (rabbit; Abcam) and anti-MYC (rabbit; Abcam) antibodies. For immunoblotting, anti-human p62 (mouse; R&D) and anti-MYC (mouse; Fisher Scientific) antibodies were used. Primary antibodies were detected with peroxidase-conjugated anti-rabbit or anti-mouse IgG (1:3000) and visualized with SuperSignal West Femto Substrate (Thermo Fischer Scientific). Blots were quantified using ImageJ.

Cell Metabolism Experiments

Human kidney SCs stimulated with IL-1β for 24 or 48 hours in the presence or absence of MYC inhibitors, then were trypsinized and plated in XF96 Cell Culture Microplates (Seahorse Bioscience) at a cellular density of 30,000 cells per well, and analyzed using the XF Extracellular Flux Analyzer. For oxygen consumption rate determination, cells were incubated in base assay medium (D5030; Sigma-Aldrich) supplemented with 2 mM glutamine, 10 mM glucose, and 1 mM pyruvate for 1 hour before measurements using the XF Cell Mito Stress Kit (Seahorse Bioscience). Concentrations of oligomycin and FCCP used in the experiment were adjusted for human kidney SCs. For glycolytic metabolism measurements, cells were incubated in basal media before injections using the Glycolytic Test Kit (Seahorse Bioscience). Concentration of 2-deoxyglucose was adjusted to SCs. Raw data were normalized cell numbers using CyQUANT (Thermo Fischer).

Statistical Analyses

Statistical evaluation was carried out using the one-way ANOVA followed by the Tukey post-test using GraphPad Prism (GraphPad Software). A P value <0.05 was considered to be significant. Error bars indicate SEM.

RESULTS

Identification of Metabolic Gene Signatures Associated with CKD Progression and Tubulointerstitial Fibrosis

Although previous studies have focused primarily on the effect of kidney damage on oxidative metabolism,^27,28 the subsequent metabolic changes that occur during progression of kidney disease are less well understood. To identify the metabolic changes associated with severe fibrosis, we studied the metabolic gene profiles from a cohort of patient kidney biopsies.¹⁶ The samples were separated into two groups by fibrosis extent (one group with <10% fibrosis range [i.e., normal kidney] and a second group with >80% fibrosis [i.e., severe fibrosis]).¹⁶ Differential gene transcriptional analysis and gene ontology and function clustering revealed that multiple components of the electron transport chain are significantly downregulated in severely fibrotic human kidneys (Figure 1A). Analysis of metabolic pathways excluding oxidative phosphorylation (OxPhos) indicated the upregulation of multiple components of the aerobic glycolysis pathway, including genes encoding fructose bi-phosphatase-3 (PFKFB3), hexokinase 1 and 2 (HK1 and HK2), lactate dehydrogenase A (LDHA), solute carrier 2A3 (SLC2A3), and pyruvate kinase muscle (PKM) (Figure 1B) and the transcriptional regulator MYC (Figure 1C). HK2, LDHA, and SLC2A3 are well described MYC transcriptional targets, suggesting a central role for MYC in metabolic gene reprogramming of fibroblasts and their precursors. However, expression of HIF1A, a transcription factor regulated by hypoxia that also participates in the regulation glycolysis enzymes, did not vary with fibrosis (Supplemental Figure 1A). To further investigate the progression of the metabolic shift in kidney disease, we performed a computational analysis of transcription factors that could drive differential gene expression identified in a cohort of 38 patients with CKD compared with matched patients who had normal kidney function. This revealed MYC and its coactivator MAZ as factors controlling transcriptional changes in CKD (Figure 1D). In this set of patient samples with incident functional clinical parameters, we used supervised correlation to analyze transcript levels with respective eGFR. Among the highest ranked genes on the basis of correlation P value, we found expression of transcripts encoding the master mitochondrial biogenesis regulator Pparγ coactivator-1α (PPARGC1A or PGC1α) and the inner mitochondrial membrane respiratory complex II succinate dehydrogenase (ubiquinone) iron-sulfur subunit B to directly correlate with eGFR (Figure 1E and Supplemental Figure 1B, respectively). Conversely, expressions of the glycolysis master regulator oncogene MYC and PKM inversely correlated with eGFR with a high level of significance (Figure 1E and Supplemental Figure 1B, respectively). We next studied the anatomic distribution of MYC and PKM isoform 2 (PKM2) proteins to identify the predominant localization in CKD biopsies. Although MYC was virtually absent in healthy kidneys, protein levels were upregulated in diseased tubules and widely present in the nuclei of interstitial cells with fibroblast morphology (Supplemental Figure 1C). The distribution of PKM2 indicated a shift in protein expression, from predominant localization in distal tubules in healthy kidneys to fibroblast location in disease (Supplemental Figure 1C). In addition to correlating with organ function in patients, PPARGC1A expression inversely correlates with Collagen I (COL1A1), whereas MYC expression positively correlates with Collagen I, a major component of the extracellular matrix and a well established indicator of fibrogenesis (Figure 1F).

Figure 1.: RNA-seq analysis of human fibrotic CKD identifies a metabolic switch from oxidative phosphorylation to glycolysis, and MYC is a candidate transcription factor driving these transcriptional changes. (A) Heat map representing expression of genes involved in OxPhos and mitochondrial biogenesis in normal versus fibrotic human kidneys (red high, green low). (B) Expression (measured using fragments per kilobase million [FPKM]) of genes involved in glycolysis, glutamine, and aldehyde metabolism. Values on the bar graph represent mean±SD. *P<0.05; **P<0.01 (green arrows from left to right highlight LDHA, SLC2A3, PKM, and HK2). (C) MYC expression in normal versus fibrotic human kidneys. Values on the bar graph represent mean±SD. **P<0.01. (D) Graph showing analysis of transcription factor binding to DEG promoters, comparing normal kidney with CKD kidney. MYC is highlighted in red. (E) Linear regression analysis of eGFR versus PPARGC1A and MYC expression from a cohort of 38 healthy individuals and patients with CKD. (F) Linear regression analysis of PPARGC1A and MYC versus COL1A1 expression from the cohort of patients.

Collectively, these results identify a glycolytic gene signature in fibrotic kidneys and reveal a significant correlation between the decline in human kidney function, tubulointerstitial fibrosis, and a metabolic gene shift from OxPhos to glycolysis.

An OxPhos to Glycolysis Metabolic Gene Switch Occurs in Kidney Resident SCs during the Transition from Experimental AKI to CKD

The presence of MYC and PKM2 in the interstitial cells of diseased human kidneys prompted us to analyze a potential relationship between the metabolic gene patterns in experimental kidney fibrosis. To this end, we performed UUO in mice, a well characterized model of kidney injury that results in progressive disease and fibrosis (Figure 2A).²⁹ Whole-kidney protein level analysis showed that UUO causes the downregulation of PGC1α and the upregulation of MYC, which occurs concomitantly with upregulation of αSMA/ACTA2, a marker of fibrosis, in whole-tissue analysis (Figure 2A). A shift from OxPhos, involving loss of mitochondria capacity and upregulation of specific glycolysis enzymes, has been reported to occur in tubules undergoing atrophy after kidney injury.^30,31 Because the human tissue–based studies in Figure 1 could partially be explained by changing populations of cells in the diseased kidney compared with healthy kidney, we evaluated metabolic gene changes specifically in myofibroblasts and their SC precursors.²² We compared the in vivo transcriptional profile of myofibroblasts after UUO with their precursors in healthy tissue using the TRAP method from Col1a1-L10aGFP mice (Figure 2B).²⁴ This is a well-validated approach to isolate translated RNA from a particular cell type in vivo; in this case, we evaluated the translated RNA in SCs specifically.²⁴ In these transgenic Col1a1-L10aGFP mice, MYC is found predominantly in Collagen I–positive SCs only after onset of injury (Figure 2C). Analysis of mitochondrial, OxPhos, and glycolysis transcripts actively translated in Collagen I–positive SCs confirmed the metabolic switch and in particular, the upregulation of glycolysis enzymes regulated by MYC in SCs early after injury (Figure 2D).

Figure 2.: Animal models of kidney fibrosis demonstrate a molecular switch from oxidative phosphorylation to glycolysis in SCs. (A) Expression of PGC1α, MYC and αSMA proteins in day 2 (D2) and D7 unilateral ureteral obstruction (UUO) kidneys from WT mice. Values on the bar graph represent mean±SD (t test). *P<0.05. (B) Schematic representing the purification of ribosome-bound translated mRNAs selectively from SCs in kidneys from Col1a1-L10aGFP transgenic mice subjected to UUO. (C) Localization of MYC in the cortical region of UUO kidneys from Col1a1-L10aGFP transgenic mice in which GFP expression labels Col1a1-expressing cells. DAPI, 4′,6-diamidino-2-phenylindole, highlights nuclei. (D) Differential expression of SC specific translated mRNAs pulled down from the kidneys of Col1a1-L10aGFP mice. Values on the bar graph represent mean±SD (one-way ANOVA). *P<0.05 versus UUO D0.

Thus, an important part of the OxPhos to glycolysis shift occurs in myofibroblasts, and it is associated with the activation of a MYC-driven program in tubulointerstitial SCs after injury.

An IL1 receptor/IRAK4/MYC Axis Regulates Kidney SC Activity after Kidney Injury In Vivo

Kidney injury triggers a prompt acute inflammatory response marked by elevated production of IL-1β by cells including monocyte/macrophages and PMNs that infiltrate the injured tissue, as well as by resident SCs.⁴ We analyzed the tubulointerstitial SC translatome in injured Col1a1-L10aGFP mice for indications of IL-1 receptor signaling by looking at previously validated downstream IL-1 response genes.³² We identified an IL-1 response signature in SCs defined by significant differential expression for Il1b, Il6, Ccl2, Bcl3, Cxcl10, Cxcl1, and Cxcl2 after injury (Figure 3A). The prominence of IL-1 receptor signaling was also confirmed in CKD patients (Figure 1). Linear correlation analysis revealed that expression of IL1B inversely correlates with eGFR (Supplemental Figure 2B) and positively correlates with MYC expression (Figure 3B) with high statistical significance. Because the molecular mechanisms driven by IL-1 dependent signaling in kidney resident SCs have not been elucidated, we next stimulated human PDGFRβ+ SCs isolated from human kidney nephrectomies with IL1β SCs have been fate-mapped in mice and shown to differentiate into myofibroblasts in vivo after kidney injury.¹¹ Stimulation of SCs with IL-1β resulted in accumulation of MYC along with PKM2, SLC2A3, and LDHA, proteins, critical for uptake and glycolytic metabolism of glucose (Figure 3C). IL-1β–driven MYC accumulation was impaired by the IRAK4 small molecule inhibitor BIIB-IRAK4i, indicating that MYC upregulation occurs downstream of MYD88, IRAK 1,2,4 signaling complex recruitment to the IL-1 receptor (Figure 3D).⁵

Figure 3.: IL-1/IL-1 receptor–associated kinase 4 (IRAK4) signaling regulates MYC expression and proliferation in kidney resident SCs. (A) Differential expression of an IL-1β-stimulated gene signature in kidney resident SCs following UUO, identified in purified translated mRNA from Col1a1-L10A-GFP mouse kidneys. Values on the bar graph represent mean±SD (one-way ANOVA). *P<0.05 versus UUO D0; **P<0.01 versus UUO D0. (B) Linear regression analysis of IL1β versus MYC expression (measured in FPKM) from the cohort of 38 healthy individuals and patients with CKD. (C) Representative western blot of protein extracts from human kidney SCs incubated with IL-1β for 48 hours. (D) Western blot of MYC in human kidney SCs incubated with IL-1β in the presence of 100 nM of the small molecule IRAK4 inhibitor, BIIB-IRAK4i for 48 hours. (E) Experimental schema of IRI in mice followed by treatment with BIIB-IRAK4i for 7 days. Western blot and quantification (right hand graphs) of PGC1α and MYC levels in the kidneys of mice D7 after unilateral-IRI and treated with either 75 mg/kg IRAK4 inhibitor BIIB-IRAK4i daily or vehicle. Values on the bar graph represent mean±SD (one-way ANOVA). ^a P<0.05 versus sham; ^b P<0.05 versus UIRI + Inhibitor. (F and G) Representative immunofluorescence detection of (F) MYC and (G) Ki67 in PDGFRβ+ SCs in UIRI kidneys of mice treated with either the IRAK4 inhibitor or vehicle. Graphs on right show quantification. DAPI labels nuclei.

These in vitro data suggested that a signaling axis involving IL-1 receptor/IRAK4/MYC is important for kidney resident SC metabolism and proliferation. To test this hypothesis, we studied the effect of inhibiting IRAK4 on MYC expression and SC proliferation in the context of kidney injury. Daily oral delivery of the inhibitor (BIIB-IRAK4i) at 75 mg/kg impaired the significant downregulation of PGC1α and upregulation of MYC, 7 days after ischemic kidney injury (Figure 3, E and F). IRAK4 inhibition also resulted in reduced proliferation of PDGFRβ+ SCs in vivo, a result that may contribute to the recently described antifibrotic effect of this molecule (Figure 3G).⁵

Together, these in vivo findings support the in vitro results showing a role for IL-1β and IRAK4 signaling in MYC activation and further clarify a mechanism contributing to kidney SC activation in kidney injury. The results also validate previous data indicating that IRAK4 signaling makes an important contribution to fibrogenic cell activity in vivo.⁵

IL-1β Induces an MYC-Dependent Proliferative SC Metabolism Regulated by SQSTM1/p62

The in vitro and in vivo data pointed to a molecular model in which IL-1β induces a MYC-driven metabolic program via regulation of MYC-target genes (Figure 4A). To test this hypothesis, we incubated SCs with the MYC inhibitor (+)-JQ1, which was previously shown to inhibit MYC expression in vivo and in vitro.³³ Upregulation of MYC and HK2 induced by IL-1β was significantly reversed by JQ1, whereas upregulation of LDHA and SLC2A3 trended toward reversal (Figure 4B). A similar effect was observed upon coincubation of IL-1β with the small molecule 10058-F4 (Supplemental Figure 3A), which was previously shown to inhibit the formation of the transcriptionally active dimer MYC/MAX and inhibit MYC-regulated gene expression.³⁴ To test the changes in metabolism of SCs treated with IL-1β, we next measured the extracellular acidification rate (ECAR), a direct measure of glycolysis. Exposure of SCs to IL-1β for 48 hours resulted in increased basal (cycles 18–30) and maximum (cycles 30–48) glycolytic capacity (Figure 4C). This stimulatory effect of IL1β on SC glycolysis was blocked by JQ1 (Figure 4C). Similar results were found when a distinct MYC inhibitor 10058-F4 was applied (Supplemental Figure 3B). In addition to stimulating glycolysis, MYC drives proliferation. Recent studies have shown that MYC enables a metabolic program linking glycolysis with cell proliferation.^13,35,36 Unbiased analysis of the transcriptome of purified PDGFRβ+ NG2+ SCs from mouse kidney cultured in vitro showed that IL-1β significantly upregulated the expression of genes encoding cell cycle regulators, including Ccnd1, Cdk6 Cdk7, and Myc, simultaneously with glycolysis components Hk2, Ldha, and Slc2a1 (Figure 4D). We next tested the effect of inhibiting MYC on human SC proliferation in vitro. Analysis of EdU incorporation in cultured human SCs showed that IL-1β stimulation induces SC proliferation, and this effect was reverted by both JQ1 and 10058-F4 (Figure 4E, Supplemental Figure 3C).

Figure 4.: IL-1β induces MYC-dependent glycolytic and proliferative program in SCs through autophagic degradation of p62. (A) Schematic depicting the model for IL-1β activation of an MYC-driven transcriptional program in kidney SCs. (B) Fold change in transcript levels of representative glycolysis factors as well as MYC transcripts induced by IL-1β in the presence or absence of the BET inhibitor (+)-JQ1. Values on the bar graph represent mean±SD (one-way ANOVA). *P<0.05 versus serum free (SF) medium; ****P<0.05 versus IL-1β + JQ1. (C) Extracellular acidification rate (ECAR) of human kidney PDGFRβ+ SCs incubated with IL-1β for 48 hours in the presence or absence of (+)-JQ1. Points on the curves represent mean±SD (t test; n=3 points). *P<0.05. (D) Heat map showing fold increase of transcripts for cell cycle regulators and glycolysis factors in mouse kidney PDGFRβ+ SCs incubated with IL-1β for 16 hours. (More Red = higher fold increase). (E) Fluorescence images detecting EdU incorporation into nuclei of human kidney PDGFRβ+ SCs stimulated with IL-1β for 48 hours in the presence or absence of the BET inhibitor JQ1. EdU was applied during the first 4 hours of the experiment. **P<0.01, ***P<0.001. (F) Representative western blot detecting SQSTM1/p62 and LC3 II/I levels in human kidney SCs stimulated with IL-1β for 48 hours. (G) Representative western blots detecting PGC1α and mΤΟR signaling in Control and SQSTM1/p62 KO human kidney PDGFRβ+ SCs. (H) Representative western blot detecting MYC in Control and SQSTM1/p62 KO human kidney PDGFRβ+ SCs. (I) Nuclear accumulation of MYC in Control and SQSTM1/p62 KO human kidney PDGFRβ+ SCs. (J) Graphs of quantitative PCR for SLC2A3 and LDHA transcripts in Control and SQSTM1/p62 KO SCs. Values are mean±SD (t test). P<0.01. (K) Representative immunoblots from WT human kidney SC protein extracts. (Left panel) SQSTM1/p62 immunoprecipitated. (Right panel) Detection of MYC in SQSTM1/p62 immunoprecipitates (left and center lanes) and input. In the center lane, SQSTM1/p62 was immunoprecipitated from the extracts of SCs incubated with the proteasome inhibitor MG132 for 12 hours. DAPI labels nuclei.

Previous reports indicate that IL-1 receptor signaling regulates autophagy by inducing autophagosome formation in innate immune cells.³⁷ To analyze the effect of IL-1β stimulation on SC autophagy, we incubated human SCs with different doses of IL-1β. Stimulation with IL-1β induced autophagy, which was indicated by an increased LC3II-to-LC3I ratio and degradation of the cargo protein SQSTM1/p62 (Figure 4F). Analysis of autophagy gene expression human kidneys with severe fibrosis compared with normal kidneys (Figure 1) showed significantly reduced mRNA levels of SQSTM1, concomitant with increased expression of several components of autophagy (Supplemental Figure 4A). Similarly, downregulation of SQSTM1 was observed in the transcriptional profile mouse SCs in vivo after UUO in the Col1a1-L10A-GFP transgenic mice (Supplemental Figure 4B). Given that p62 binds polyubiquitin chains and may deliver ubiquitinated proteins for proteasomal degradation,³⁸ and because MYC levels are tightly controlled by the ubiquitin-proteasome pathway,^39–41 we hypothesized that p62 may directly regulate MYC levels. To test the direct effect of p62 on MYC levels, we used CRISPR/Cas9-mediated gene editing to delete the transcription start site and exon 1 of the human p62 gene in human SCs, which resulted in permanent deletion of p62. Similar to previous reports, we found that loss of p62 resulted in reduced levels of PGC1α,⁴² abrogation of mTOR signaling (Figure 4G),⁴³ and reduced mitochondrial OxPhos capacity (Supplemental Figure 5A).⁴² Loss of p62 resulted in increased levels of MYC (Figure 4H), with most of the accumulation occurring in the nuclei of the cultured human SCs (Figure 4I), and increased expression of MYC-regulated glycolysis genes SLC2A3 and LDHA (Figure 4J). Conversely, overexpression of p62 using a lentiviral transgenic construct resulted in lower MYC levels, consistent with p62 regulating MYC (Supplemental Figure 5B). To evaluate whether p62 and MYC interacted, we performed coimmunoprecipitation experiments and immunoprecipitated MYC with p62. In addition, when the proteasome was blocked with the proteasome inhibitor MG132, we pulled down MYC at multiple sizes, consistent with polyubiquitinated MYC (Figure 4K).

The findings point to a major role for IL-1β in metabolic regulation of SC metabolism. This mechanism relies on activation of autophagy and degradation of p62, resulting in loss of mitochondrial capacity and accumulation of MYC, for the activation of a glycolysis gene program.

MYC Inhibition during the Inflammatory Phase Post-AKI Reduces Tubulointerstitial Fibrosis

To study the effect of JQ1 on SC proliferation after kidney injury, we administered JQ1 (50 mg/kg) or vehicle by oral gavage daily to cohorts of mice and performed UUO surgery (Figure 5A); (+)-JQ1 treatment resulted in significantly reduced proliferation in both PDGFRβ+ SCs and other cells 3 days postinjury onset (Figure 5A).

Figure 5.: Inhibition of MYC prevents fibrosis in acute murine kidney injury in vivo. (A) Detection of Ki67 in PDGFRβ+, Col1a1-GFP+ SCs by immunofluorescence in kidneys of Col1a1-GFP mice treated with either the BET inhibitor (+)-JQ1 or vehicle 3 days after UUO surgery. Values on the bar graph represent mean±SD (one-way ANOVA). ***P<0.001. (B) Detection of αSMA/ACTA2+ SC-derived myofibroblasts in kidneys of mice treated with either the BET inhibitor (+)-JQ1 or vehicle 10 days after UUO surgery (one-way ANOVA). Values on the bar graph represent mean±SD. ***P<0.001. (C) Quantitative PCR measuring selected transcript levels in RNA extracted from whole kidney from mice treated with either the BET inhibitor (+)-JQ1 or vehicle 10 days after UUO surgery. Values on the bar graph represent mean±SD (one-way ANOVA). *P<0.05; **P<0.01. (D) Representative images showing tubulointerstitial fibrosis by Masson Trichrome (upper) and picrosirius red (lower) of D10 after UUO or sham surgery kidney sections from mice treated with (+)-JQ1 on day 10 postsurgery. (E) Detection and quantification (right hand panels) of the tubular damage marker KIM-1 and the macrophage marker F4/80 by immunofluorescence in the kidneys of mice treated with either the BET inhibitor (+)-JQ1 or vehicle on day 10 after UUO surgery. Values on the bar graph represent mean±SD (one-way ANOVA). ***P<0.001. DAPI labels nuclei.

In the longer term, JQ1 treatment resulted in reduced fibrosis as indicated by reduced accumulation of pathologic, SC-derived αSMA+ myofibroblasts (Figure 5B), lower expression of Col1a1, Tgfb and Ctgf (Figure 5C), and reduced collagen deposition in the tubulointerstitial compartment (Figure 5D). Treatment with (+)-JQ1 resulted in lower levels of KIM-1, a sensitive marker of proximal tubule injury in epithelial cells, and reduced accumulation of F4/80+ macrophages (Figure 5E). Taken together, those results indicated that MYC inhibition impairs SC expansion and differentiation in the context of kidney injury, while reducing the subsequent accumulation of epithelial injury and inflammation.

IL-1β Stimulation Is Sufficient for the Initiation of Proximal Tubule Damage in Three-Dimensional Human Kidney Organoids

Delivery of (+)-JQ1 in vivo resulted in reduced macrophage infiltration in injured mouse kidneys, suggesting that the renoprotective effect on fibrosis and proximal tubule damage could be secondary to the anti-inflammatory activity of the compound.⁴⁴ To study the effect of IL-1β and MYC inhibition by (+)-JQ1 on kidney SCs in a 3-D environment, we took advantage of the human kidney organoid culture system.²¹ This system offers unique advantages. First, it allows for study of multiple human kidney cell types simultaneously. Second, because it is devoid of immune cells, it constitutes an ideal platform to study the effect of specific inflammatory cytokines in the absence of the complex inflammatory signal milieu found in vivo. Human kidney organoids were generated by directed differentiation of human pluripotent stem cells, which were sequentially differentiated into posterior intermediate mesoderm progenitors, metanephric mesenchyme, pretubular aggregates, and finally, nephrons²¹ (Figure 6A). The specific effect of IL-1β was tested by incubating fully mature day 51 human kidney organoids for 48 and 96 hours in the presence of the cytokine. IL-1β induced organoid hypertrophy, which was indicated by 1.5- and 1.98-fold increases in diameter at 48 and 96 hours, respectively (Figure 6B). Concomitantly, the cytokine induced proximal tubule cell atrophy, with noticeable loss of LTL binding at 96 hours, indicating loss of proximal tubule cell brush border (Figure 6C). Proximal tubule injury was further confirmed by the expression of KIM-1, which was not detected in the tubules of control organoids (Figure 6C). Progressive tubule injury induced by IL-1β involved tubule dilation and epithelial cell cycle arrest, which was indicated by upregulation of p21^WAF1/Cip1 in proximal tubules (Figure 6D).

Figure 6.: IL-1β causes tubular damage and fibrosis in human kidney organoids. (A) Schematic depicting differentiation stages during the directed derivation of human kidney organoids from human pluripotent stem cells in vitro. (B, upper left panel) Schematic indicating collection time points during the incubation of day 51 kidney organoids with IL-1β. (B, lower left panel) Representative bright-field photomicrographs of organoids incubated with IL-1β for 48 and 96 hours. Values on the bar graph in the right panel represent mean±SD of organoid diameter (one-way ANOVA; n=3 organoids per treatment per time point). *P<0.05 versus time-matching control; **P<0.01 versus time-matching control. (C) Detection of KIM-1 in proximal tubules identified by expression of LTL in kidney organoids treated with either vehicle or IL-1β for 96 hours. (D) Detection of p21^WAF1/Cip1 expression in proximal tubules of kidney organoids treated with either vehicle or IL-1β for 96 hours. Arrows indicate p21^WAF1/Cip1+ tubular epithelial cells. Values on the bar graph represent mean±SD (t test; n=3 organoids per treatment per time point). **P<0.01. (E) Immunofluorescence showing the progression of Collagen I accumulation in kidney organoids treated with IL-1β. Values on the bar graph represent mean±SD (t test; n=3 organoids per treatment per time point). *P<0.05; **P<0.01. (F) Detection of COL1A1, fibronectin I (FN1), and ACTA2 by quantitative PCR. Values on the bar graph represent mean±SD (t test; n=3 organoids per treatment per time point). DAPI labels nuclei. *P<0.05.

Concomitantly organoids exposed to IL-1β showed progressive thickening of the Collagen-I+ tubular basement membrane and significant increase in Collagen I deposition in the interstitial space at 48 and 96 h, indicative of a fibrotic process (Figure 6E). Increased fibrogenesis was independently confirmed by measuring transcripts for COL1A1, FN1, and ACTA2 all of which were significantly elevated in the organoids treated with IL-1β (Figure 6F).

IL-1β Stimulation of Human Kidney Organoids Triggers MYC-Driven Activation of SCs and Recapitulates Human Tubulointerstitial Disease

To test the effect of (+)-JQ1 on IL-1β–induced damage, we incubated day 51 kidney organoids with IL-1β in the presence or absence of the inhibitor. Coincubation with (+)-JQ1 significantly inhibited IL-1β–induced organoid increase in size (Figure 7A). At 48 hours IL-1β significantly induced PDGFRβ+ SC proliferation, which was significantly abrogated by coincubation with (+)-JQ1 (Figure 7B, Supplemental Figure 6A). Interestingly, although IL-1β did not increase overall epithelial cell proliferation in the tubules compared with baseline levels in control organoids, expression of Ki67 in organoids incubated with IL-1β alone was prominently confined to dilated (i.e., damaged) tubules (Figure 7B, Supplemental Figure 6A). This pattern of distribution for Ki67 in tubules and interstitium resembles what is observed in the early response after AKI. Of note also is that IL-1β stimulation did not elicit podocyte proliferation in the organoid glomeruli (Supplemental Figure 6, B and C).

Figure 7.: (+)-JQ1 inhibits MYC expression and reduces tubulointerstitial fibrosis in human kidney organoids. (A, left panel) Schematic illustrating sample collection time points for day 51 kidney organoids treated with IL-1β plus vehicle or (+)-JQ1. (A, center panel) Representative bright-field photomicrographs. Values on the bar graph in A, right panel represent mean±SD of organoid diameter (one-way ANOVA; n=3 organoids per treatment). **P<0.01 versus Ctl. (B and C) Detection of (B) Ki67 and (C) MYC in PDGFRβ+ SCs by immunofluorescence in the tubulointerstitial space of kidney organoids incubated with IL-1β with or without (+)-JQ1 for 48 hours. Values on the bar graph represent mean±SD (one-way ANOVA; n=3 organoids per treatment). **P<0.01 versus Ctl. (D) αSMA/ACTA2 expression in interstitial PDGFRβ+ SCs determined by immunofluorescence in kidney organoids incubated with IL-1β with or without (+)-JQ1 for 96 hours. Values on the bar graph represent mean±SD (one-way ANOVA; n=3 organoids per treatment). DAPI labels nuclei. *P<0.05; **P<0.01; ***P<0.001.

Next, we examined MYC distribution in treated kidney organoids. IL-1β stimulation increased MYC accumulation in the nuclei of PDGFRβ+ SCs (Figure 7C). MYC also accumulated in epithelial cells of damaged tubules, albeit mostly in the cytoplasm (Figure 7C). Of note, a similar pattern of MYC accumulation is observed in the kidneys of patients with chronic kidney conditions and in experimentally damaged murine kidneys (Figure 2C and Supplemental Figure 1C, respectively). In the presence of (+)-JQ1, MYC accumulation was reduced close to basal levels (Figure 7C). Next, we assessed the effect of IL-1β stimulation on tubulointerstitial fibrosis. Incubation with IL-1β resulted in the significant upregulation of αSMA/ACTA2 in PDGFRβ+ SCs (Figure 7D), indicating that those cells differentiated into myofibroblasts. αSMA/ACTA2 expression was significantly abrogated in organoids treated with IL-1β in the presence of (+)-JQ1 (Figure 7D), suggesting an antifibrogenic effect in addition to the antiproliferative effect, for this compound.

Therefore, the human organoid experiments revealed that IL-1β can drive tubulointerstitial disease by inducing proximal tubule damage and SC activation and fibrosis. Inhibition of BET activity led to reduced MYC accumulation and proliferation and fibrogenic differentiation of SCs, validating the renoprotective effect of (+)-JQ1, in the absence of leukocytes.

DISCUSSION

Here, we report a novel metabolic mechanism for the initiation and progression of tubulointerstitial fibrosis driven by IL-1β. Our experiments show that IL-1β activates kidney resident SCs, inducing glycolytic metabolism that supports SC proliferation and deposition of fibrotic matrix, through a mechanism that involves stabilization of the oncoprotein MYC. These results are in keeping with early studies reporting increased proliferative activity in kidney fibroblasts stimulated by IL-1βin vitro^26,45 and offer an additional mechanism for previous findings of reduced fibrosis in the kidneys of mice carrying a deletion in the IL-1 receptor that are subjected to IRI.⁷

Persistent IL-1β stimulation in the context of kidney injury leads to degradation of p62 and stabilization of MYC for the activation of a glycolytic-proliferative transcriptional program that drives tubulointerstitial SC activity and fibrogenesis.

Evidence of feedback between inflammation and fibrosis was provided by recent reports indicating that kidney SCs also contribute to the inflammatory response after AKI through activation of a fibroblast inflammasome.⁵ Combined with findings presented in this study, two important conclusions emerge. First, kidney fibrosis could be self-perpetuating and independent from bone marrow–derived inflammatory cells. Second, kidney SCs and their derivatives—namely fibroblasts and myofibroblasts—actively drive kidney inflammation and contribute to parenchyma damage. Although further work is required to support these conclusions, from a therapeutic point of view, it seems that targeting kidney fibrogenic cell-autonomous mechanisms will have antifibrotic, anti-inflammatory, and nephron-preserving effects.

Inflammatory mechanisms play an important role in the transition from AKI to CKD and the progressive loss of kidney function that characterizes CKD. This idea is supported by recent data showing that inhibition of the IL-1 receptor signaling molecule IRAK4 results in reduced IRI⁵ and our results showing that treatment with BIIB-IRAK4i results in reduced SC proliferation and fibrosis. Despite the caveat that IRAK4 is a transducer for multiple Toll-like receptor agonists, the body of data points to IL-1 receptor signaling as a potential therapeutic target in the progression from AKI to CKD. Our efforts are focused on testing this hypothesis using new therapeutic compounds that inhibit IL-1β with high efficacy.^46,47

Importantly, this study also represents a pioneering step toward using human kidney organoids to understand the inflammatory process in the kidney. By studying the effects of IL-1β individually, we obtained a full assessment of the mechanisms driven by this cytokine in multiple human kidney cell types. The results obtained with kidney organoids not only confirmed the mechanisms discovered using two-dimensional human SC cultures, namely that IL-1β drives MYC accumulation, a metabolic switch and proliferation in SCs, but further provided decisive evidence for the role of resident PDGFRβ+ SCs in the generation of myofibroblasts in kidney fibrosis, a process previously debated.¹² In addition, because this system allows the assessment of multiple kidney cell types, our findings extend beyond SCs, showing that IL-1β induces upregulation of p21^WAF1/Cip1 in epithelial cell and tubular damage. Importantly, using this system, we were able to confirm the renoprotective effect of the BET inhibitor (+)-JQ1, highlighting the therapeutic importance of BET proteins and MYC in kidney disease. In this multicellular three-dimensional culture system, we showed that IL-1β alone was sufficient to drive the tissue injury and tissue enlargement over a 4-day period. This enlargement was essentially attributable to SC expansion, because no significant IL-1β–dependent proliferation of tubules or glomerular cells occurred and tubules were interspersed by expanding stroma.

Our results on the regulatory role of p62 in human SC activation provide a novel mechanism for MYC regulation. A scaffold protein with various roles in multiple cellular processes, including autophagy, protein degradation, inflammation, metabolic signaling, oxidative stress response; loss of p62 leads to increased SC proliferation and contributes to the SC glycolytic switch. Our results are in keeping with the previous observation that silencing of p62 expression results in increased SC proliferation and higher IL-6 expression,⁴⁸ but they are in conflict with the observation of lower MYC levels in p62KO fibroblasts from tumor stromas.⁴⁸ It seems plausible, however, that the interaction between p62 and MYC is altered by molecular modifications on both proteins, namely ubiquitination or phosphorylation of MYC, phosphorylation of p62, or the state of cell differentiation, as well as variations in the signals that regulate the homeostasis of both proteins.

In summary, our study reveals a novel mechanism for the reprogramming of renal SCs by IL-1β toward an aerobic glycolysis and the support of fibrogenic activity during kidney injury and fibrosis.

DISCLOSURES

All employees of Biogen have stock in the company. I.A.L. and J.S.D. have a patent for structures and use of IL-1 receptor–associated kinase 4 inhibitors. J.V.B. is a coinventor on KIM-1 patents assigned to Partners Healthcare and is a cofounder of Goldfinch Bio. J.S.D. has served on the scientific advisory boards for Promedior Inc. and Regulus Therapeutics, has stock options with Promedior Inc., is a cofounder of Muregen LLC, and has patents for the use of agents to treat kidney disease.