Estrogen-Related Receptor Agonism Reverses Mitochondrial Dysfunction and Inflammation in the Aging Kidney

Abstract

A gradual decline in renal function occurs even in healthy aging individuals. In addition to aging, per se, concurrent metabolic syndrome and hypertension, which are common in the aging population, can induce mitochondrial dysfunction and inflammation, which collectively contribute to age-related kidney dysfunction and disease. This study examined the role of the nuclear hormone receptors, the estrogen-related receptors (ERRs), in regulation of age-related mitochondrial dysfunction and inflammation. The ERRs were decreased in both aging human and mouse kidneys and were preserved in aging mice with lifelong caloric restriction (CR). A pan-ERR agonist, SLU-PP-332, was used to treat 21-month–old mice for 8 weeks. In addition, 21-month–old mice were treated with a stimulator of interferon genes (STING) inhibitor, C-176, for 3 weeks. Remarkably, similar to CR, an 8-week treatment with a pan-ERR agonist reversed the age-related increases in albuminuria, podocyte loss, mitochondrial dysfunction, and inflammatory cytokines, via the cyclic GMP-AMP synthase–STING and STAT3 signaling pathways. A 3-week treatment of 21-month–old mice with a STING inhibitor reversed the increases in inflammatory cytokines and the senescence marker, p21/cyclin dependent kinase inhibitor 1A (Cdkn1a), but also unexpectedly reversed the age-related decreases in PPARG coactivator (PGC)-1α, ERRα, mitochondrial complexes, and medium chain acyl coenzyme A dehydrogenase (MCAD) expression. These studies identified ERRs as CR mimetics and as important modulators of age-related mitochondrial dysfunction and inflammation. These findings highlight novel druggable pathways that can be further evaluated to prevent progression of age-related kidney disease.

The fastest growing segment of the US population with impaired kidney function is the group aged ≥65 years. This population is expected to double in the next 20 years, whereas the number worldwide is expected to triple from 743 million in 2009 to 2 billion in 2050. This will result in a marked increase in the elderly population with chronic kidney disease. This increase may be further amplified by other age-related comorbidities, including metabolic syndrome and hypertension, that accelerate age-related decline in renal function.¹ Thus, there is a growing need for prevention and treatment strategies for age-related kidney disease.

A gradual age-related decline in renal function occurs even in healthy aging individuals.² Progressive glomerular, vascular, and tubulointerstitial sclerosis is evident on renal tissue examination of healthy kidney donors with increasing age.³ In addition to aging, per se, metabolic syndrome and hypertension can induce mitochondrial dysfunction and inflammation, as well as endoplasmic reticulum stress, oxidative stress, altered lipid metabolism, and elevation of profibrotic growth factors in the kidney, which collectively contribute to age-related kidney disease.²

There is variation in the rate of decline in renal function as a function of sex, race, and burden of comorbidities.4, 5, 6 Interestingly, examination of processes leading to renal sclerosis suggests a role for possible modifiable systemic metabolic and hormonal factors that can ameliorate the rate of sclerosis. With the population of older individuals increasing, identifying preventable or treatable aspects of age-related nephropathy becomes critical.

There is increasing evidence that mitochondrial biogenesis, mitochondrial function, mitochondrial unfolded protein response, mitochondrial dynamics, and mitophagy are impaired in aging, and these alterations contribute to the pathogenesis of the age-related diseases.7, 8, 9 In this regard, current studies are concentrated on modulating these molecular mechanisms to improve mitochondrial function.

Caloric restriction (CR) plays a prominent role in preventing age-related complications. CR prevents the age-related decline in renal function and renal lipid accumulation via, at least in part, inhibition of the increased activity of the sterol regulatory element binding proteins.^10,11 CR is also an important modulator of mitochondrial function. CR prevents age-related mitochondrial dysfunction in the kidney by increasing mitochondrial/nuclear DNA ratio, mitochondrial complex activity, including fatty acid β-oxidation, and expression of the mitochondrial transcription factor nuclear respiratory factor 1 (NRF1), the protein kinase 5′ AMP-activated protein kinase (AMPK), the deacetylases sirtuin 1 and 3, and mitochondrial isocitrate dehydrogenase expression.¹² In addition, CR also prevents the age-related decrease in mitochondrial number in the renal tubules. CR increases expression of the bile acid–regulated nuclear receptor farnesoid X receptor and the Takeda G-protein–coupled receptor. Treatment of 22-month–old mice, fed ad lib for 2 months with the dual farnesoid X receptor–Takeda G-protein–coupled receptor agonist INT-767, reversed most of the age-related impairments in mitochondrial function and the progression of renal disease.¹² More importantly, INT-767 and CR each increased expression of PPARG coactivator (PGC)-1α, estrogen-related receptor (ERR) α, and ERRγ, which are important regulators of mitochondrial biogenesis and function.

The ERRs, ERRα (NR3B1 and ESSRA genes), ERRβ (NR3B2 and ESRRB genes), and ERRγ (NR3B3 and ESRRG genes), are members of the nuclear receptor superfamily. Although one report suggests a role for cholesterol,¹³ there are no confirmed endogenous ligands for these orphan receptors. More importantly, the ERRs do not bind natural estrogens, and they do not directly participate in classic estrogen signaling pathways or biological processes.¹⁴ ERRα and ERRγ are strongly activated by their coactivators, PGC-1α and PGC-1β, respectively.¹⁵ In contrast, receptor interacting protein 140 (RIP140) and nuclear receptor corepressor 1 (NCoR1) are important corepressors of ERR activity.^16,17

ERRα and ERRγ regulate the transcription of genes involved in mitochondrial biogenesis, oxidative phosphorylation, tricarboxylic acid cycle, fatty acid oxidation, and glucose metabolism.¹⁴ However, in addition to overlapping gene activation, there is also ample evidence that ERRα and ERRγ have differential and opposing effects, which can be due to interactions with corepressors, coactivators, posttranslational modification, or differential cell expression.¹⁴ Opposing effects for ERRα and ERRγ are seen in breast cancer,¹⁸ regulation of gluconeogenesis in the liver,^18,19 skeletal muscle function,^18,19 macrophage function,^19,20 and regulation of lactate dehydrogenase A related to anaerobic glycolysis.¹⁴

ERRα and ERRγ are highly expressed in the mouse and human kidney.^21,22

However, the roles of ERRα and ERRγ in modulating age-related impairment of mitochondrial function and age-related inflammation (inflammaging) are not known. The objective of this study was to determine whether a pan-ERR agonist, SLU-PP-332, could improve mitochondrial function and inflammation in the aging mouse kidney by activating ERRα, ERRβ, and ERRγ.

Materials and Methods

Mice

Studies were performed in 4- and 21-month–old male C57BL/6 male mice obtained from the National Institute on Aging aging rodent colony. Mice received either 3% dimethyl sulfoxide vehicle or the pan-ERR agonist SLU-PP-332²³ at a dose of 25 mg/kg body weight per day, administered intraperitoneally. Mice were dosed for 8 weeks, following which they underwent euthanasia. Kidneys were harvested and processed for i) histology, ii) transmission electron microscopy, iii) isolation of nuclei, iv) isolation of mitochondria, and v) biochemical studies, as detailed below.

Another cohort of similarly aged male C57BL/6 mice received either 3% dimethyl sulfoxide vehicle or the stimulator of interferon genes (STING) inhibitor C-176 (Focus Biomolecules, Plymouth Meeting, PA) at a dose of 1 mg/kg body weight per day for 3 weeks via i.p. injection. The 24-month–old caloric restricted mice versus age-matched 24-month–old ad lib fed male mice were obtained from the National Institute on Aging rodent colony.

Immunohistochemistry

Formalin-fixed, paraffin-embedded human tissue sections were subjected to antigen retrieval with EDTA buffer in high-pressure heated water bath, and staining was performed using either mouse monoclonal ERRα (1:2500; Abcam, Cambridge, MA) or ERRγ (1:400; Abcam) antibodies for 90 minutes or pyruvate dehydrogenase (PDH) E2/E3bp (1:1000; Abcam) antibody for 45 minutes. UnoVue horseradish peroxidase secondary antibody detection reagent (Diagnostic BioSystems, Pleasanton, CA) was applied, followed by diaminobenzidine chromogen. Imaging was done with Nanozoomer (Hamamatsu Photonics, Hamamatsu, Japan).

Cell Culture

Primary human proximal tubule epithelial cells (catalog number PCS-400-010) were purchased from ATCC (Manassas, VA). Cells were cultured in Renal Epithelial Cell Basal Medium (ATCC; catalog number PCS-400-030) supplemented with Renal Epithelial Cell Growth Kit (ATCC; catalog number PCS-4000-040) at 37°C in 5% CO₂. Cells were cultured to 70% to 80% confluence and then treated with vehicle (4 mmol/L HCl with 0.1% bovine serum albumin) or 10 ng/mL transforming growth factor (TGF)-β1 (catalog number 7754-BH-005; R&D Systems, Minneapolis, MN) or 10 ng/mL tumor necrosis factor (TNF)-α (catalog number 210-TA-020; R&D Systems) for 24 hours. Cells were then harvested and analyzed for gene expression.

Transmission Electron Microscopy

Kidney cortex tissue (1 mm³) was fixed for 48 hours at 4°C in 2.5% glutaraldehyde and 1% paraformaldehyde in 0.1 mol/L cacodylate buffer (pH 7.4) and washed with cacodylate buffer three times. The tissues were fixed with 1% OsO₄ for 2 hours, washed again with 0.1 mol/L cacodylate buffer three times, washed with water, and placed in 1% uranyl acetate for 1 hour. The tissues were subsequently serially dehydrated in ethanol and propylene oxide and embedded in EMBed 812 resin (Electron Microscopy Sciences, Hatfield, PA). Thin sections, approximately 80 nm, were obtained by using the Leica ultracut-UCT ultramicrotome (Leica, Deerfield, IL) and placed onto 300 mesh copper grids and stained with saturated uranyl acetate in 50% methanol and then with lead citrate. The grids were viewed with a JEM-1200EXII electron microscope (JEOL Ltd, Tokyo, Japan) at 80 kV, and images were recorded on the XR611M, mid-mounted, 10.5 megapixel, charge-coupled device camera (Advanced Microscopy Techniques Corp., Danvers, MA). Mitochondrial morphology was assessed with ImageJ software version 1.53 (NIH, Bethesda, MD; https://imagej.nih.gov, last accessed August 14, 2022) by manually tracing all mitochondria that were completely within the field of view in six random images from each mouse (n = 3 to 4 each group). The area, perimeter, and minimum Feret diameter of each mitochondrion were measured.

RNA Extraction and Real-Time Quantitative PCR

Total RNA was isolated from the kidneys using Qiagen RNeasy minikit (Valencia, CA), and cDNA was synthesized using reverse transcript reagents from Thermo Fisher Scientific (Waltham, MA). Quantitative real-time PCR was performed as previously described,²⁴ and expression levels of target genes were normalized to 18S level. Primer sequences are listed in Table 1.

Table 1.

List of Primers

Name	Forward primer	Reverse primer
Esrrα	5′-CAGGGAGGGAAGGGATGG-3′	5′-ATGAGGAGAGGAGCGAAGG-3′
Esrrb	5′-GCACCTGGGCTCTAGTTGC-3′	5′-TACAGTCCTCGTAGCTCTTGC-3′
Esrrg	5′-AAGATCGACACATTGATTCCAGC-3′	5′-CATGGTTGAACTGTAACTCCCAC-3′
Ppargc1a	5′-GTCAGAGTGGATTGGAGTTG-3′	5′-AAGTCATTCACATCAAGTTCAG-3′
Ppargc1b	5′-TCCTGTAAAAGCCCGGAGTAT-3′	5′-GCTCTGGTAGGGGCAGTGA-3′
Pdk4	5′-AGGGAGGTCGAGCTGTTCTC-3′	5′-GGAGTGTTCACTAAGCGGTCA-3′
Acadm	5′-AACACTTACTATGCCTCGATTGCA-3′	5′-CCATAGCCTCCGAAAATCTGAA-3′
Tfam1	5′-AACACCCAGATGCAAAACTTTCA-3′	5′-GACTTGGAGTTAGCTGCTCTTT-3′
mtDNA	5′-ATAACCGAGTCGTTCTGCCAAT-3′	5′-TTTCAGAGCATTGGCCATAGAA-3′
Sdhc	5′-GCTGCGTTCTTGCTGAGACA-3′	5′-ATCTCCTCCTTAGCTGTGGTT-3′
Atp5b	5′-GGTTCATCCTGCCAGAGACTA-3′	5′-AATCCCTCATCGAACTGGACG-3′
Pdhb	5′-AGGAGGGAATTGAATGTGAGGT-3′	5′-ACTGGCTTCTATGGCTTCGAT-3′
Mdh1	5′-TTCTGGACGGTGTCCTGATG-3′	5′-TTTCACATTGGCTTTCAGTAGGT-3′
Idh3b	5′-TGGAGAGGTCTCGGAACATCT-3′	5′-AGCCTTGAACACTTCCTTGAC-3′
Sucla2	5′-ACCCTTTCGCTGCATGAATAC-3′	5′-CCTGTGCCTTTATCACAACATCC-3′
Cpt1a	5′-CTCCGCCTGAGCCATGAAG-3′	5′-CACCAGTGATGATGCCATTCT-3′
Ndufb8	5′-TGTTGCCGGGGTCATATCCTA-3′	5′-AGCATCGGGTAGTCGCCATA-3′
Opa1	5′-CGACTTTGCCGAGGATAGCTT-3′	5′-CGTTGTGAACACACTGCTCTTG-3′
Miga2	5′-GGAGGACTGAGGGTATGTCCA-3′	5′-CAAGGGCTGTGGCAAAAAGA-3′
Pld6	5′-ACCTGCACCGAGGCTTTAC-3′	5′-CATGTAGTCGCAGTCAGTGATG-3′
Il1b	5′-GCAACTGTTCCTGAACTCAACT-3′	5′-ATCTTTTGGGGTCCGTCAACT-3′
Icam1	5′-GTGATGCTCAGGTATCCATCCA-3′	5′-CACAGTTCTCAAAGCACAGCG-3′
Stat3	5′-AGCTGGACACACGCTACCT-3′	5′-AGGAATCGGCTATATTGCTGGT-3′
Uqcrb	5′-GGCCGATCTGCTGTTTCAG-3′	5′-CATCTCGCATTAACCCCAGTT-3′
Cox6a2	5′-CTGCTCCCTTAACTGCTGGAT-3′	5′-GATTGTGGAAAAGCGTGTGGT-3′
Tmem173	5′-GGTCACCGCTCCAAATATGTAG-3′	5′-CAGTAGTCCAAGTTCGTGCGA-3′
Cdkn1a	5′-CCTGGTGATGTCCGACCTG-3′	5′-CCATGAGCGCATCGCAATC-3′

Bulk RNA Sequencing

Approximately 300 to 500 ng of kidney RNA was used to generate barcoded RNA libraries using the Ion AmpliSeq Transcriptome Mouse Gene Expression Panel, Chef-Ready Kit (Thermo Fisher Scientific). Library quantification was performed using the Ion Library Quantitation Kit (Thermo Fisher Scientific). Sequencing was done on an Ion Proton with signal processing and base calling using Ion Torrent Suite (Thermo Fisher Scientific). Raw reads were mapped to AmpliSeq-supported mm10 transcriptome. Normalized read counts per million were generated using the RNA-seq Analysis plugin (Ion Torrent Community, Thermo Fisher Scientific). The gene expression table was converted to natural logarithm scale and quantile normalized. Genes with expression levels of less than four across all samples were filtered out. Selection of the differentially expressed genes was performed using the t-test. The significance of differentiation was defined as P < 0.05. Enrichment analysis of the Gene Ontology biological processes and pathways was prepared with DAVID²⁵ and the PANTHER online tool.²⁶ Heat map visualization was performed using the Heatmapper tool.²⁷ The raw data have been deposited in National Center for Biotechnology Information Gene Expression Omnibus database (NIH/National Center for Biotechnology Information; https://www.ncbi.nlm.nih.gov/gds/?term=PRJNA642362; accession number: PRJNA642362).

Single-Nuclei RNA Sequencing

Mouse kidney single nuclei were isolated^28,29 and counted using the EVE Automated Cell Counter (NanoEnTek; VWR, Radnor, PA). The resulting mixture was provided to the Genomics and Epigenomics Shared Resource at Georgetown University, and further processed by the Chromium Controller (10x Genomics, Pleasanton, CA) using Single Cell 3′ GEM Kit version 3, Single Cell 3′ Library Kit version 3, i7 multiplex kit, and Single Cell 3′ Gel Bead Kit version3 (10x Genomics), according to the manufacturer's instructions, with modifications for single nuclei. Libraries were sequenced on the Illumina Novaseq S4 System (Illumina, San Diego, CA) to an average depth of >300 mol/L reads passing filter per sample. Sequencing data were processed by CellRanger pipeline (10x Genomics). Briefly, reads were aligned against the mouse mm10 genome reference using the STAR algorithm.³⁰ Barcodes and unique molecular identifiers were filtered and corrected. Only confidently mapped, non-PCR duplicates with valid barcodes and unique molecular modifiers were used to generate a gene-barcode matrix for further analysis. The expression matrix was further investigated using the Loupe Cell Browser (10x Genomics).

Proteomics

Kidney tissue (200 mg) was homogenized and lysed by 8 mol/L urea in 20 mmol/L HEPES (pH = 8.0) buffer with protease and phosphatase inhibitors using Tissue Lyser II (Qiagen). Samples were reduced and alkylated, followed by digestion with LysC (Fujifilm Wako Pure Chemical, Osaka, Japan) in the ratio of 1:100 (enzyme/protein ratio, w/w) at 37°C for 3 hours. Subsequently, the proteins were further digested with trypsin (Promega, Fitchburg, WI) in the ratio of 1:50 (enzyme/protein ratio, w/w) at 37°C overnight after diluting the urea concentration from 8 to 2 mol/L. The resulting tryptic peptides were desalted and frozen dry overnight. The dry peptides were labeled with 11-plex tandem mass tag, according to the manufacturer's instruction. The tagged peptides were concatenated and fractionated with basic-pH reverse-phase high-performance liquid chromatography, collected in a 96-well plate, and combined for a total of 12 fractions before desalting and subsequent liquid chromatography−tandem mass spectrometry processing on an Orbitrap Q-Exactive HF (Thermo Fisher Scientific) mass spectrometer interfaced with an Ultimate 3000 nanoflow liquid chromatography system.³¹ Each fraction was separated on a reverse phase C₁₈ nano-column (25 cm × 75 μm; 2-μm particles) with a linear gradient 4% to approximately 45% solvent B (0.1% trifluoroacetic acid in acetonitrile). Data-dependent mode was applied to analyze the top 15 most abundant peaks in one acquisition cycle.

Mass spectrometry raw files were mapped against Uniprot mouse database version 20170207 using the MaxQuant software package version 1.5.3.30 with the Andromeda search engine.³² Corrected intensities of the reporter ions from tandem mass tag labels were obtained from the MaxQuant search. The normalized relative abundance of a protein in each sample was calculated as the ratio of a protein abundance in a sample/abundance of the same protein in a pool (channel 11; pooled from 20 samples) of the corresponding batch. After performing principle component (PC) analysis on the protein relative abundance matrix, the authors noticed that the samples were still separated by batches on the PC1 to PC2 plane. Thus, the authors further corrected this batch effect by applying the experimental Bayes batch correction method.³³ After batch correction, PC analysis revealed clear separation of samples into biological groups on the PC1 to PC2 plane (Supplemental Figure S1). All further analyses were performed on batch-corrected protein abundance matrix. Selection of the differentially abundant proteins was done using Wilks theorem-based likelihood ratio test³⁴ and t-test statistic. The significance of differentiation was defined as likelihood ratio statistic P < 0.001 and t-test statistic P < 0.05.

Proteomics data are available via ProteomeXchange with identifier PXD020051 (European Molecular Biology Laboratory–European Bioinformatics Institute, https://www.ebi.ac.uk/pride/archive/projects/PXD020051, last accessed August 4, 2023).

Mitochondrial Isolation

Kidney mitochondria were isolated using MITOISO1 (catalog number CS0760; Sigma, St. Louis, MO), according to manufacturer’s instructions.

Mitochondrial Biogenesis

Mitochondrial (Cytb) and nuclear (H19) DNA was quantified by real-time quantitative PCR. Primer sequences used are listed in Table 1.

Mitochondrial Respiration

Basal respiration, ATP turnover, proton leak, maximal respiration, and spare respiratory capacity were measured using the Seahorse XF96 Analyzer (Agilent, Santa Clara, CA) on equally loaded freshly isolated kidney mitochondria. The authors also measured mitochondrial complex I, II, III, IV, and V protein abundance by Native Blue Gel Electrophoresis (Thermo Fisher Scientific) with equally loaded mitochondrial fractions.

Multi-Omics Data Analysis and Integration Bioinformatics Methods

Multi-omics (transcriptomics and proteomics) measurements, reflecting the same biological processes in the samples, were performed by two-way orthogonal partial least square integration³⁵ to identify networks of associated genes and proteins. This analysis identified one orthogonal component (V2) that reflected the expected separation of samples into three distinct groups: young and pan-ERR agonist-treated young mice, old mice, and old pan-ERR agonist-treated mice. The authors identified a subset of genes and proteins with V2 highest loadings (threshold was set to 0.04 or <–0.04) for further analysis.

Western Blot Analysis

Western blot analysis was performed as previously described.^24,36, 37, 38 Equal amounts of total protein were separated by SDS-PAGE gels and transferred onto polyvinylidene difluoride membranes. After horseradish peroxidase–conjugated secondary antibodies, the immune complexes were detected by chemiluminescence captured on Azure C300 digital imager (Dublin, CA), and the densitometry was performed with ImageJ software. Primary antibodies used for Western blot analysis are listed in Table 2.

Table 2.

List of Antibodies

Name	Host	Source	Catalog no.
OPA1	Mouse	BD Biosciences (Franklin Lakes, NJ)	612606
MFN2	Rabbit	Millipore (Burlington, MA)	978-715-4321
DRP1	Mouse	Novus Biologicals (Centennial, CO)	H00010059-M01
p-DRP1 (s616)	Rabbit	Cell Signaling (Danvers, MA)	3455S
Stat3	Rabbit	Cell Signaling	4904T
p-Stat3 (Y705)	Rabbit	Cell Signaling	9145T
PDH e2/e3	Mouse	Abcam	Ab110333

MFN2, mitofusin 2; OPA1, OPA1 mitochondrial dynamin like GTPase; PDH, pyruvate dehydrogenase; p-DRP1, phosphorylated dynamin related protein 1 (DRP1); p-Stat3, phosphorylated Stat3.

Cytokine Arrays

Cytokines in kidney lysates (200 μg total protein pooled from four samples with equal amount of protein) were detected with the Proteome Profiler Array (ARY028; R&D Systems), according to manufacturer's instructions.

Study Approval

Animal studies and relative protocols were approved by the Animal Care and Use Committee at the Georgetown University. All animal experimentation was conducted in accordance with the NIH’s Guide for Care and Use of Laboratory Animals.³⁹

Statistical Analysis

Results are presented as the means ± SEM for at least three independent experiments. Following the Grubbs outlier test, the data were analyzed by analysis of variance and Newman-Keuls tests for multiple comparisons or by t-test for unpaired data between two groups (Prism 6; GraphPad, San Diego, CA). Statistical significance was accepted at the P < 0.05 level.

Results

ERRα and ERRγ Expression Is Decreased in the Aging Human Kidney

Decreased expression of ERRα and ERRγ in the aging mouse kidney is reversed by the dual farnesoid X receptor–Takeda G-protein–coupled receptor agonist INT-767 or CR.¹² Furthermore, increased ERR expression correlates with increased mitochondrial biogenesis and function in the treated aging kidneys.¹² In light of the role of ERR in mitochondrial biogenesis, the current study determined whether decreased expression also occurs in the aging human kidney. Immunohistochemistry was performed with human kidney sections from young (<20-year–old) and old (>70-year–old) individuals (Supplemental Table S1). The results indicate that both ERRα and ERRγ were expressed in renal tubules and that their expression levels were markedly decreased in aging human kidney (Figure 1A).

Figure 1.

Estrogen-related receptor (ERR) α, ERRγ, and pyruvate dehydrogenase (PDH) expression is decreased in the aging human kidney. A: ERRα and ERRγ immunohistochemistry of kidney sections from young and old human subjects. Insets: Magnified images of boxed areas. B: PDH immunohistochemistry of kidney sections from young and old human subjects. C: Renal transforming growth factor (TGF)-β mRNA expression in young and old mice. mRNA expression of the ERRs, PPARG coactivator (PGC)1α, medium chain acyl coenzyme A dehydrogenase (MCAD), superoxide dismutase 2 (SOD2), and uncoupling protein 2 (UCP2) in TGF-β1–treated human primary proximal tubule cells. D: Renal tumor necrosis factor (TNF)-α protein expression in young and old mice. mRNA expression of the ERRs, PGC1α, MCAD, SOD2, and UCP2 in TNF-α treatment of human primary proximal tubule cells. N = 3 for each group (A–D). ∗P < 0.05, ∗∗∗P < 0.001. Scale bar = 100 μm (A and B). AU, arbitrary unit.

Because ERRs are important modulators of mitochondrial biogenesis, staining of the human kidney sections for mitochondrial PDH e2/e3 was conducted, revealing a marked decrease in PDH immunostaining in the aging human kidney samples (Figure 1B).

TGF-β and TNF-α Mediate Decreases in ERRγ in Cultured Human Proximal Tubular Epithelial Cells

TGF-β and TNF-α expression levels are increased in the aging kidney (Figure 1, C and D). Therefore, the study determined whether TGF-β and TNF-α mediate, at least in part, the decrease in expression of ERR seen in the kidney. In cultured human primary proximal tubular cells, TGF-β significantly decreased expression level of ERRγ mRNA, whereas ERRα was unaffected. Expression of potential mitochondrial ERR targets PGC1α, MCAD (encoding medium-chain acyl-CoA dehydrogenase), superoxide dismutase 2 (SOD2), and uncoupling protein 2 (UCP2) were also decreased in TGF-β–treated cells (Figure 1C). Similarly, TNF-α also decreased PGC1α, ERRγ, SOD2, and UCP2 expression (Figure 1D).

ERRα and ERRγ RNA Distribution in the Mouse Kidney

Single-nuclei RNA sequencing was performed to determine where ERRα and ERRγ mRNAs were expressed in the mouse kidney.^28,29 With 3000 to 5000 nuclei sequenced at 100,000 read depth, 12 expression clusters were identified and assigned to major cell types in the mouse kidney (Figure 2). ERRα was expressed in most of the cell types within the kidney, most prominently in the proximal tubules, intercalated cells, and podocytes, and in many cases decreased with aging. Conversely ERRγ was detected in fewer cells and mainly in the proximal tubule and intercalated cells of young mice (Table 3). Compared with young kidneys, the S1/S2 segments of aging proximal tubules showed a decline in both ERRα and ERRγ expression, whereas ERRγ podocyte positivity increased (Table 4).

Figure 2.

Single-nuclei RNA sequencing of young and old kidneys. A: The t-distributed stochastic neighbor embedding of young and old mouse kidneys, with 100,000 read depth and 3000 to 5000 nuclei sequenced. A total of 12 clusters were identified and assigned to major cell types known in the mouse kidney. B: Estrogen-related receptor (ERR) α expression (purple) in young and old kidney as a function of total cell count. Top panels: The highest expression seen in proximal tubules, intercalated cells, and podocytes. ERRγ expression (purple) in young and old kidney as a function of total cell count. Bottom panels: The highest expression seen in proximal tubules and intercalated cells expressed most strongly. AL, ascending limb; CD-PC, collecting duct–principal cell; DCT, distal convoluted tubule; DL, descending limb; EC: endothelial cell; IC: intercalated cell; LH, loop of Henle; Mφ: macrophage; MC, mesangial cell; Pod: podocyte; PT, proximal tubule.

Table 3.

Percentage of ERRα-Positive Cells in Each Cluster Resolved from Single-Nuclei RNA Sequencing

Variable	Young, %	Old, %
Pod	27	14
MC	8	7
EC	4	5
Mφ	4	4
DCT	15	10
LH-AL	21	17
LH-DL	1	8
IC-B	41	25
IC-A	27	12
CD-PC	12	13
PT-S3	32	28
PT-S1/S2	30	19

AL, ascending limb; CD-PC, collecting duct–principal cell; DCT, distal convoluted tubule; DL, descending limb; EC, endothelial cell; ERR, estrogen-related receptor; IC, intercalated cell; LH, loop of Henle; Mφ, macrophage; MC, mesangial cell; Pod, podocyte; PT, proximal tubule.

Table 4.

Percentage of ERRγ-Positive Cells in Each Cluster Resolved from Single-Nuclei RNA Sequencing

Variable	Young, %	Old, %
Pod	0	12
MC	1	1
EC	1	1
Mφ	0	0
DCT	4	4
LH-AL	3	6
LH-DL	3	3
IC-B	0	7
IC-A	8	0
CD-PC	1	4
PT-S3	7	6
PT-S1/S2	6	3

AL, ascending limb; CD-PC, collecting duct–principal cell; DCT, distal convoluted tubule; DL, descending limb; EC: endothelial cell; ERR, estrogen-related receptor; IC: intercalated cell; LH, loop of Henle; Mφ: macrophage; MC: mesangial cell; Pod: podocyte; PT, proximal tubule.

Pan-ERR Agonist Treatment Improves the Age-Related Kidney Injury

Aging mice were treated with a recently available pan-ERR agonist (SLU-PP-332).²³ Two-month treatment of 21-month–old mice significantly improved the age-related albuminuria and decreased kidney weight (Figure 3A). The decrease in albuminuria is likely related to improved podocyte function as protein expression of the podocyte marker, NPHS2 (podocin), was increased by the treatment (Figure 3B). In addition, mRNA expression of profibrotic markers (TGF-β, PAI-1, and Col IV), monocyte/macrophage marker (F4/80), and tubular injury marker neutrophil gelatinase-associated lipocalin (NGAL) was decreased with SLU-PP-332 treatment (Figure 3C). Finally, cytokine profiling was used to indicate an increase in the kidney injury-related proteins, Ngal, kidney injury marker-1, osteopontin, and CC-motif ligand chemokine-21 in the aging kidney, which were partially normalized by pan-ERR agonist treatment (Figure 3D).

Figure 3.

Pan–estrogen-related receptor (ERR) agonist improves age-related renal injury. A: Albuminuria and kidney weight (normalized by body weight) in young and old mice, with or without SLU-PP-332 (ERR) treatment. B: Left panels: NPHS2 (podocin) immunohistochemistry of kidney sections, labeling podocytes, in young and old mice, with or without SLU-PP-332 treatment. Right panel: The mean intensity of NPHS2 staining per glomeruli (glom) is also shown. C: Real-time quantitative PCR mRNA expression of kidney injury markers transforming growth factor (TGF)-β, plasminogen activator inhibitor 1 (PAI-1), collagen 4 A1 (Col4a1), F4/80, and neutrophil gelatinase-associated lipocalin (NGAL) in young and old mice, with and without SLU-PP-332 treatment. D: Cytokine array. Four major spots that correspond to kidney injury markers Ngal, kidney injury marker (Kim)-1, osteopontin, and CC-motif ligand chemokine-21 (CCL21) are highlighted (left panels), with relative changes in protein level, as assessed by densitometry (right panels). Each bar graph represented one sample pooled from four animals per group. N = 5 to 6 for each group (A, B, left panels, and C). ∗P < 0.05, ∗∗P < 0.01. Scale bar = 50 μm (B). ACR, albumin creatinine ratio.

Pan-ERR Agonist Treatment Modulates Mitochondrial Metabolism and Inflammation in the Aging Kidney

ERRα, ERRβ, and ERRγ mRNA abundance was examined to assess the actions of the pan-ERR agonist. The expression of all ERRs was decreased in the kidneys of aging mice, and treatment with the ERR pan-agonist induced significant increases in their expression to levels observed in the young mice (Figure 4A).

Figure 4.

RNA sequencing and proteomics of kidney from old mice treated with vehicle or the pan–estrogen-related receptor (ERR) agonist. A: ERRα, ERRβ, and ERRγ mRNA expression in young and old mouse kidneys, with and without SLU-PP-332 treatment. B: Heat map showing expression patterns of genes differentially expressed in kidneys of old mice treated with vehicle compared with kidneys of old mice treated with SLU-PP-332 treatment. The heat map indicates up-regulation (green), down-regulation (red), and unaltered gene expression (black). The columns represent individual samples. C: Functional pathway enrichment analysis of differentially expressed proteins in kidneys of old mice treated with vehicle compared with kidneys of old mice treated with pan-ERR agonist. The y axis shows significantly enriched pathways. The x axis indicates P value of enrichment of the given pathway. D: Heat map showing expression patterns of proteins differentially expressed in kidneys of old mice treated with vehicle compared with kidneys of old mice treated with pan-ERR agonist. The heat map indicates up-regulation (green), down-regulation (red), and unaltered gene expression (black). The columns represent individual samples. E: Functional pathway enrichment analysis of differentially expressed proteins in old mice treated with vehicle compared with kidneys of old mice treated with pan-ERR agonist. The y axis shows significantly enriched pathways. The x axis indicates P value of enrichment of the given pathway. F: Functional pathway enrichment analysis of subset of genes and proteins identified with two-way orthogonal partial least square (O2PLS) analysis as up-regulated in kidneys of old mice compared with kidneys of young mice and that were down-regulated by ERR treatment in old mice. The y axis shows significantly enriched pathways. The x axis indicates P value of enrichment of the given pathway. G: Functional pathway enrichment analysis of subset of genes and proteins identified with O2PLS analysis as down-regulated in kidneys of old mice compared with kidneys of young mice and that were up-regulated by ERR treatment in old mice. The y axis shows significantly enriched pathways. The x axis indicates P value of enrichment of the given pathway. N = 5 to 6 for each group (A). ∗P < 0.05. TNF, tumor necrosis factor.

Next, bulk mRNA sequencing analysis was used to determine pathways affected by aging and restored by pan-ERR agonist. The molecular changes in the aging kidney were investigated first. The kidneys of aging mice had 448 up-regulated and 463 down-regulated genes as compared with those of young controls (Supplemental Figure S2A and Supplemental Table S2). Pathway enrichment analysis of up-regulated genes revealed that inflammation-related pathways were highly significant. The inflammation-related pathways included inflammation mediated by chemokine and cytokine signaling pathway, epidermal growth factor receptor signaling pathway, toll receptor signaling pathway, and TGF-β signaling pathway, which were all previously reported as up-regulated in aging kidney^40,41 (Supplemental Figure S2B).

In addition, apoptosis and cell cycle regulating pathways were also enriched with up-regulated genes in the aging kidney (Supplemental Figure S2B). Cell cycle dysregulation is a known phenomenon in the aging and diseased kidney and may be indicative of cellular senescence.⁴² Therefore, RNA-sequencing data were anayzed in a supervised manner to assess senescence-associated genes. In the aging kidney, 29 senescence-associated genes were up-regulated and 10 were down-regulated (Supplemental Figure S2C and Supplemental Table S3). In particular, p53 (Trp53), p65 (RelA), and p21 (Cdkn1a), which are major indicators of senescent cells, were highly expressed in aging kidney. The down-regulated genes in aging kidney were enriched in mitochondrial and metabolic processes (Supplemental Figure S2D). Dysregulated mitochondrial functioning has been widely studied in aging kidneys.⁴³ Recently, several studies have shown that damaged mitochondria may trigger inflammation via the cyclic GMP-AMP synthase (cGAS)–STING pathway in kidney disease.^44,45 transcription factor A mitochondrial (TFAM) is a key regulator of mitochondrial gene expression⁴⁶ and is crucial for maintaining mitochondrial DNA (mtDNA) structure, transcription, and replication.⁴⁷ Deletion of Tfam leads to mtDNA escape into the cytoplasm and activation of the innate immune pathway through cGAS-STING activation.⁴⁵ The present mRNA sequencing data indicate that Tfam mRNA levels were decreased in aging kidney (Supplemental Figure S2E). Overall, a dysregulation of mitochondrial and immune processes was observed in the aging kidney.

The RNA-sequencing data were further examined to illuminate the effect of pan-ERR agonist treatment. Pan-ERR agonist treatment mitigated several of the above-mentioned pathways (Figure 4, B and C, and Supplemental Table S4).

Mass spectrometry–based proteomics was performed in addition to mRNA sequencing. Because the concordance of proteomics and RNA transcriptomics is known to be weak,^48,49 the proteomics analysis was expected to reveal additional processes disrupted in aging and attenuated by pan-ERR agonist treatment. At the protein level, mitochondria-related, oxidation-reduction processes, peroxisome, and metabolic pathways were found to be dysregulated in aging kidney, which correlated with pathways found at the mRNA level. Furthermore, several additional dysregulated pathways were identified by proteomics analysis in aging kidney. These additional processes included calcium-binding region, blood coagulation, and biotin/lipoyl attachment processes (Supplemental Figure S2, F and G, and Supplemental Table S5). The processes of mitochondrion, immune-related pathways, transport regulation, focal adhesion, and fatty acid oxidation were improved by pan-ERR agonist treatment (Figure 4, D and E, and Supplemental Table S6).

Combining several layers of -omics data yields a clearer understanding of complex biological phenomena than evaluation of each layer separately.^50,51 Two-way orthogonal partial least squares⁵² indicated that one of the components (V2) reflected the expected relationships between the biological groups under investigation. Specifically, the V2 component profile was associated with the pan-ERR agonist effect of alleviating the dysregulation of genes and proteins in aging kidneys (Supplemental Figure S3). Genes and proteins whose expression correlates with V2 component profile were identified as an interrelated molecular network associated with overall changes in aging and pan-ERR agonist treatment (see Materials and Methods) (Supplemental Table S7). The subset of genes and proteins up-regulated in aging kidney and down-regulated by pan-ERR agonist treatment were enriched with immune-related processes, such as antigen processing and presentation, complement and coagulation cascades, Ig-like C1-type domain protein activation, and the TNF signaling pathway (Figure 4F). The subset of genes and proteins down-regulated in aging kidney and enhanced by pan-ERR agonist treatment was enriched by processes associated with metabolism, such as lipid metabolic processes, fatty acid metabolism, and catalytic activity (Figure 4G).

Taken together, the transcriptomic and proteomic analyses, as well as the integration of the transcriptome-proteome associations, point to down-regulation of mitochondrial metabolism and up-regulation of inflammatory processes in the aging kidney. These analyses also showed that pan-ERR agonist treatment was effective in attenuating these age-related dysregulations. Thus, the present analyses correlate with previous studies to show that the hallmarks of aging kidneys are decreased mitochondrial function and increased inflammation.^53,54

Pan-ERR Agonist Treatment Restores Mitochondrial Function in Aging Kidneys

The canonical function of ERR is to induce mitochondrial biogenesis. The pan-ERR agonist increased expression of the master mitochondrial biogenesis regulators PGC1α and PGC1β in aging kidneys. The expression of the mitochondrial transcription factor, Tfam1, was decreased in the aging kidney and was increased by the pan-ERR agonist. As a result, the mitochondrial DNA/nuclear DNA ratio was increased in aging kidneys following treatment, indicative of increased mitochondrial biogenesis (Figure 5A). The increased mitochondrial biogenesis was accompanied by increased expression of genes related to the mitochondrial electron transport chain (ETC) complexes, such as complex I subunit Ndufb8, complex II subunit Sdhc, complex III subunit Uqcrb, complex IV subunit Cox6a2, and complex V subunit Atp5b (Figure 5B). Native blue gel analysis further showed increased quantities of assembled mitochondrial complexes after the treatment (Supplemental Figure S4). This is consistent with the increased gene expression of enzymes in tricarboxylic acid cycle, such as Pdhb, Mdh1, Idh3b, and Sucla2 (Figure 5C).

Figure 5.

Pan–estrogen-related receptor (ERR) agonist treatment restores mitochondrial function in aging kidneys. A: mRNA expression of PGC1α, PGC1β, and transcription factor A mitochondrial (Tfam1), coregulators of ERRs and mediators of mitochondrial biogenesis, and mitochondria/nuclear DNA ratios, in young and old mouse kidneys, with and without SLU-PP-332 treatment. B: mRNA expression levels of subunits of the mitochondrial electron chain complex (ETC). C: mRNA expression levels of tricarboxylic acid (TCA) cycle enzymes. D: Interrelationship of the TCA cycle and ETC. E: Maximum respiration capacity in mitochondria isolated from the kidneys. F: mRNA expression of the fatty acid β-oxidation enzymes, Cpt1a and Mcad. N = 5 to 6 for each group (A–C, E, and F). ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗∗P < 0.0001.

The interrelationship between the ETC and the tricarboxylic acid cycle is illustrated in Figure 5D. The changes in ETC complexes and the tricarboxylic acid cycle by the pan-ERR agonist resulted in increased maximum respiration capacity in mitochondria isolated in the treated aging kidneys (Figure 5E). In addition, mRNAs encoding enzymes that mediate mitochondrial fatty acid β-oxidation, including Cpt1a and Mcad, were up-regulated by the pan-ERR agonist (Figure 5F), suggesting that ERR agonism promotes mitochondrial fatty acid β-oxidation.

Pan-ERR Agonist Treatment Alters Mitochondrial Dynamics in Aging Kidneys

Transmission electron microscopy showed alterations in the mitochondria of aging kidneys, including decreases in the area, perimeter, and minimum Feret diameter, with these parameters being restored to levels seen in young kidneys on treatment with the pan-ERR agonist (Figure 6A).

Figure 6.

Pan–estrogen-related receptor (ERR) agonist treatment alters mitochondrial dynamics in aging kidneys. A: Transmission electron microscopy of alterations in the mitochondria in young and old mouse kidneys, with and without SLU-PP-332 treatment. a: Normally distributed and structured mitochondria in young mice. b: Mitochondria in normal structure, perpendicularly oriented to the basolateral plasma membrane in young mice treated with SLU-PP-332. c: Chaotically distributed damaged and degraded mitochondria with cristae condensation (black arrow) or loss (red arrow) in old mice. Electron-dense lipofuscin granules are abundant in the cytoplasm (blue arrow). d: The structure of mitochondria preserved in old mice treated with SLU-PP-332. e–g: Quantification of area, perimeter, and minimum Feret diameter of mitochondria. B: Mitofusin-2 (Mfn2) protein abundance in young and old mouse kidneys, with and without SLU-PP-332 treatment (left panels), and protein levels of OPA1 mitochondrial dynamin like GTPase (Opa1) in mouse kidneys (right panels). C: Changes in mitoguardin 2 and mitoPLD (phospholipase D) mRNA levels in the mouse kidneys. D: Changes in dynamin related protein 1 (Drp1) and phosphorylated Drp1 (p-Drp1) protein in mouse kidneys. N = 3 to 4 for each group (A, e–g); N = 5 to 6 for each group in mRNA-level analysis (D); N = 4 for each group in protein analysis (D). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001. Original magnification, ×5500 (A).

Because these mitochondrial changes are reminiscent of alterations in mitochondrial fusion and fission, the expression of proteins that regulate mitochondrial fusion and fission was measured. Mitofusin 2 is found in the outer membrane that surrounds mitochondria and participates in mitochondrial fusion.⁵⁵ There was a significant decrease in mitofusin 2 in the aging kidney, and the ERR pan agonist increased the mitofusin 2 protein abundance in both the young and the old kidneys, fully restoring levels in the old kidneys to levels of the untreated young kidneys (Figure 6B). There were changes in expression of other nuclear-encoded, mitochondrial-expressed mRNAs and proteins that would be congruent with restoration of mitochondrial function. The OPA1 mitochondrial dynamin like GTPase (Opa1) protein localizes to the inner mitochondrial membrane and helps regulate mitochondrial stability, energy output, and mitochondrial fusion.⁵⁶ Although there was no significant change in the Opa1 protein level in the aging kidney, on treatment there was a tendency for the protein level to increase in the aging kidneys (Figure 6B). In addition, there were also significant decreases in mRNA levels of mitoguardin 2 (Miga2) and mitoPLD (Pld6), a divergent family member of the phospholipase D (PLD) family, in the aging kidneys that were increased on treatment with the pan-ERR agonist (Figure 6C). Mitoguardin 2 mediates mitochondrial fusion through mitoPLD.⁵⁷

Dynamin related protein 1 (Drp1) is a member of the dynamin superfamily of proteins and is a fundamental component of mitochondrial fission.⁵⁵ There were significant increases in Drp1 and phosphorylated Drp1 protein in the kidneys of aging mice, which were restored to levels seen in young mice following treatment with the pan-ERR agonist (Figure 6D).

Pan-ERR Agonist Treatment Decreases Inflammation in Aging Kidneys

Mitochondria are immunogenic organelles, and mitochondrial dysfunction generates several immunogenic molecules, including mtDNA⁵⁸ and mitochondrial RNA.⁵⁹ The cGAS-STING has been reported as one of the innate immune receptors to be activated by mtDNA leaking into cytosol.^45,60 In the aging kidneys, expression of STING and cGAS mRNAs and proteins was higher, which was significantly reversed by treatment with the pan-ERR agonist (Figure 7A). Similar changes were observed with RNA sensors RIG-I/MDA5/LGP2 and other nucleic acid sensors, such as the toll-like receptors (TLRs)^61,62 (Figure 7B). Expression of mRNA-encoding components of the NF-κB signaling pathway (Rel, RelB, Nfkb2, and p65 protein) was increased in the aging kidney, and reduced by pan-ERR agonist treatment (Figure 7C). However, expression of interferon α and β genes was not detected in aging kidney samples, and no interferon γ protein expression was detected by the cytokine array (Figure 3D).

Figure 7.

Pan–estrogen-related receptor (ERR) agonist treatment decreases inflammation in aging kidneys. A: Changes in cyclic GMP-AMP synthase (cGAS; left panels) and stimulator of interferon genes (STING; right panels) mRNA and protein levels in young and old mouse kidneys, with and without SLU-PP-332 treatment. B: RIG-I–like receptor [retinoic acid inducible gene 1 (RIG-I), melanoma differentiation associated protein 5 (MDA5), and laboratory of genetics and physiology 2 (LGP2; alias RIG-I-like receptor 3) and toll-like receptor (TLR; 3, 7, and 9) mRNA levels in mouse kidneys. C:Rel, Relb, and Nfkb2 mRNA, and total p65 protein expression in the kidneys of mice. D: Changes in Stat3 mRNA, and phosphorylated Tyr705-STAT3 and total STAT3 protein expression. E: Changes in cellular senescence and senescence-associated secretory phenotype markers. N = 5 to 6 for each group in mRNA-level analysis (A and D); N = 4 for each group in protein analysis (A); N = 5 to 6 for each group (B and E); N = 3 to 4 for each group in protein analysis (D). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001. ICAM1, intercellular adhesion molecular 1; NS, nonsignificant; Timp1, TIMP metallopeptisase inhibitor 1; TNF-α, tumor necrosis factor-α.

Potential STING-activated interferon signaling pathway had corresponding changes in STAT3 (Figure 7D). Significant increases in both total STAT3 and p-Tyr705-STAT3, a marker for STAT3 activation, were observed. These increases were substantially suppressed on treatment with the pan-ERR agonist. In addition to the activation of this inflammatory cascade, expression of senescence marker cyclin dependent kinase inhibitor 1A (Cdkn1a)/p21 was induced in the aging kidney and reduced by the treatment with the pan-ERR agonist (Figure 7E). Another senescence marker, p16^Ink4a, was increased in the aging kidneys but was unaffected by the treatment (Figure 7E). Both of these cellular senescence markers were down-regulated in aging kidneys with life-long CR (Supplemental Figure S5).

To determine whether any senescence-associated secretory phenotype factors were regulated by the pan-ERR agonist treatment in the aging kidney, RNA-sequencing was explored. Real-time PCR verified that RNAs encoding the proinflammatory cytokines IL-1β and TNF-α, the chemokine receptor, CCR5, the metallopeptidase inhibitor, TIMP1, and the cell adhesion molecule, ICAM1, were increased in the aging kidney, and treatment with the pan-ERR agonist decreased their expression (Figure 7E).

STING Inhibition Decreases Inflammation and Increases Mitochondrial Gene Expression in Aging Kidneys

To determine whether the activation of nucleic acid sensors, such as STING, per se, mediates age-related increase in inflammation, aging mice were treated with the known STING inhibitor C-176.⁶³ C-176 decreased IL-1β, STAT3, phosphorylated STAT3, and the senescence marker p21/Cdkn1a expression in the aging kidney (Figure 8A). Unexpectedly, the expression of the master regulators for mitochondria biogenesis, PGC1α and PGC1β, along with ERRα, was increased with treatment in aging kidney (Figure 8B). Genes involved in mitochondrial ETC complex, Ndufs1, Cox6a2, and ATP5e, and the fatty acid oxidation gene, Mcad, were also found up-regulated by the STING inhibitor (Figure 8B).

Figure 8.

Treatment of aging mice with stimulator of interferon genes (STING) inhibitor (SI) C-176. A: Effect of the SI, C-176, on IL-1β, Stat3, phosphorylated Stat3 (p-Stat3), and p21 expression in mouse kidneys. B: Effect of STING inhibition on expression of PGC1α, PGC1β, estrogen-related receptor (ERR) α, mitochondrial NADH-ubiquinone oxidoreductase 75 kDa subunit (Ndufs1) (complex I), mitochondrial cytochrome c oxidase subunit 6A2 (Cox6a2) (complex IV), Atp5e (complex V), and MCAD in mouse kidneys. N = 4 for each group (A and B). ∗P < 0.05, ∗∗P < 0.01.

Interestingly, in proximal tubular epithelial cells, TNF-α, the proinflammatory cytokine that is increased in the aging kidney, induced impairments in genes that mediate mitochondrial biogenesis and the ETC complex (Figure 1D). These results suggest that in aging, inflammation and mitochondrial dysfunction may regulate each other to amplify age-related kidney disease.

Discussion

Here, the study identified the nuclear hormone receptors, the estrogen-related receptors ERRα, ERRβ, and ERRγ, as important modulators of age-related mitochondrial dysfunction and inflammation. ERRα, ERRβ, and ERRγ expression is decreased in the aging kidney, and lifelong CR results in increases in expression of ERRα, ERRβ, and ERRγ in the kidney. In parallel, CR also prevents age-related mitochondrial dysfunction and inflammation.^12,64,65 ERRs therefore act as potential CR mimetics in the kidney. Remarkably, treatment of 21-month–old mice with the pan-ERR agonist for only 8 weeks reversed the age-related increases in albuminuria and podocyte loss, mitochondrial dysfunction, and inflammation. These effects were comparable with those achieved with lifelong CR, which is known to protect against age-related comorbidities, including loss of renal function.^10,11,64

Recent evidence indicates that mitochondrial dysfunction is one of the mediators of cellular senescence, and the associated senescence-associated secretory phenotype includes proinflammatory cytokines and profibrotic growth factors.^66,67 This process may also be involved in the age-related inflammation, termed inflammaging or senoinflammation, which is also prevented by CR.^68,69

The RNA-sequencing data indicated that apoptosis and cell cycle regulating pathways were also enriched in the aging kidney. Further analysis of these data in a supervised manner identified 29 senescence-associated genes up-regulated and 10 genes down-regulated in aging kidneys. In particular, p53 (Trp53), p65 (Rela), and p21 (Cdkn1a), which are the major indicators of senescent cells, were highly expressed in aging kidney.

The down-regulated genes in aging kidney were enriched in mitochondrial and metabolic processes based on the mRNA sequencing data. Dysregulated mitochondrial functioning has been widely studied in aging kidneys.⁴³ RNA-sequencing data suggest that Tfam mRNA levels are decreased in aging kidney, which were verified herein by real-time quantitative PCR. TFAM is a key regulator of mitochondrial gene expression⁴⁶ and is crucial for maintaining mtDNA structure, transcription, and replication.⁴⁷ Deletion of Tfam leads to mtDNA escape into the cytoplasm and activation of the innate immune pathway through cGAS-STING activation, which plays key roles in immunity, inflammation, senescence, and cancer.^45,60,70,71 Recently, other mechanisms have been found to trigger the release of mitochondrial DNA and RNA into cytosol and activate inflammation.^72,73 In addition to the recent identification of the importance of this signaling pathway in mouse models of acute kidney injury, chronic kidney disease, and fibrosis,^44,45 the current study also shows increased expression of STING in aging kidneys, and its down-regulation following treatment with the pan-ERR agonist.

In addition to cGAS-STING signaling, other nucleic acid sensors, including retinoic acid inducible gene 1 (RIG-I), melanoma differentiation associated protein 5 (MDA5), TLR3, TLR5, and TLR7, were found to be regulated in aging kidneys in the current study. To determine whether STING activation, per se, mediates aging-related inflammation, aging mice were treated with a STING inhibitor. STING inhibition only moderately decreased inflammation in aging kidney, indicating the importance of other nucleic acid sensors also found to be regulated in aging kidneys. Interestingly, STING inhibition increased mitochondrial gene expression, suggesting a model in which mitochondrial injury triggers inflammation, thereby compounding mitochondrial dysfunction in aging kidney. Similarly, in proximal tubular epithelial cells, TNF-α, the proinflammatory cytokine that is increased in the aging kidney, induced impairments in genes that mediate mitochondrial biogenesis and the ETC complex. These results suggest that in aging, inflammation and mitochondrial dysfunction may regulate each other to amplify age-related kidney disease. These findings are consistent with reports of mitochondrial damage in acute kidney injury induced by lipopolysaccharide, an endotoxin that activates innate immunity (via TLR4) to induce circulating cytokines.74, 75, 76

Overall, the study found a dysregulation of mitochondrial and immune processes in aging kidney. Pan-ERR agonist treatment mitigated several of the above-mentioned pathways. In summary, this study identified the ERRs as beneficial modulators of mitochondrial dysfunction and inflammation in the aging kidney. Although potential systemic effects of treatment cannot be ruled out as the drug is given via the circulation, data presented have shown the significant renal-specific effect on ERR activation.

Author Contributions

M.L. and X.X.W. conceived and led the study; X.X.W. performed experiments; K.M. assisted with the animal studies and biochemical analysis; Y.Q. and U.G. performed proteomics experiments and assisted with the analysis; J.P., A.T., E.T., and L.B. performed the multi-omics processing and integration analyses; P.D. and A.Z.R. performed immunohistochemistry and analysis; A.E.L. performed human primary cell culture work; P.M.Z. and J.B.K. performed the electron microscopy; B.A.J., K.B., N.S., and A.Z.R. analyzed data; B.E. and T.P.B. provided the reagent pan–estrogen-related receptor agonist; and X.X.W., C.A., and M.L. wrote the manuscript with editorial input from all authors.

Disclosure Statement

None declared.

Supplemental Data

Supplemental Figure S1

Two-way orthogonal partial least square component (V2) profile separates the samples into three biologically meaningful groups: kidneys of young mice treated with vehicle or with pan–estrogen-related receptor (ERR) agonist, kidneys of old mice, and kidneys of old mice treated with pan-ERR agonist.

Supplemental Figure S2

RNA-sequencing and proteomics analysis of kidneys from old mice compared with kidneys from young mice. A: Heat map showing expression patterns of genes differentially expressed in kidneys of old mice compared with kidneys of young mice. The heat map indicates up-regulation (green), down-regulation (red), and unaltered gene expression (black). The columns represent individual samples. B: Functional pathway enrichment analysis of genes up-regulated in kidneys of old mice compared with kidneys of young mice. The y axis shows significantly enriched pathways. The x axis indicates P value of enrichment of the given pathway. C: Heat map showing expression patterns of senescence-related genes differentially expressed in kidneys of old mice compared with kidneys of young mice. The heat map indicates up-regulation (green), down-regulation (red), and unaltered gene expression (black). The columns represent individual samples. D: Functional pathway enrichment analysis of genes down-regulated in kidneys of old mice compared with kidneys of young mice. The y axis shows significantly enriched pathways. The x axis indicates P value of enrichment of the given pathway. E: Transcription factor A mitochondrial (Tfam) mRNA expression profile in young and old kidneys. Expression values are presented in fragments per kilobase of transcript per million mapped reads units. F: Heat map showing expression patterns of proteins differentially expressed in kidneys of old mice compared with kidneys of young mice. The heat map indicates up-regulation (green), down-regulation (red), and unaltered gene expression (black). The columns represent individual samples. G: Functional pathway enrichment analysis of differentially expressed proteins in kidneys of old mice compared with kidneys of young mice. The y axis shows significantly enriched pathways. The x axis indicates P value of enrichment of the given pathway. CCKR, gastrin and cholecystokinin receptors mediated signaling network; EGF, epidermal growth factor; PI3, phosphorus triiodide; TGF-β, transforming growth factor-β.

Supplemental Figure S3

Principle component analysis (PCA) of proteomics data before batch correction procedure and after implementing experimental Bayes batch correction method. After batch correction, the samples on the PC1 to PC2 plane are separated into biologically meaningful groups. ERR, estrogen-related receptor.

Supplemental Figure S4

Native blue gel indicates the increased level of assembled complex I, II, III, IV, and V in the kidneys of old mice after treatment with the pan–estrogen-related receptor (ERR) agonist. N = 4 for each group. ∗P < 0.05.

Supplemental Figure S5

Cellular senescence markers p21 and p16 mRNA levels were increased in the aging kidney. However, in aging kidneys with lifelong caloric restriction (CR), p21 and p16 mRNA expression was down-regulated compared with that in ad lib aging kidneys. N = 4 to 5 samples per group. ∗P < 0.05. ERR, estrogen-related receptor.