Gprc5b Modulates Inflammatory Response in Glomerular Diseases

BASIC RESEARCH www.jasn.org

GPRC5b Modulates Inﬂammatory Response in Glomerular Diseases via NF-kB Pathway

Sonia Zambrano,1 Katja Möller-Hackbarth,1 Xidan Li,1 Patricia Q. Rodriguez,1 Emmanuelle Charrin,1 Angelina Schwarz,1 Jenny Nyström,2 Annika Östman Wernerson,3 Mark Lal,4 and Jaakko Patrakka1

1Karolinska Insitutet/AstraZeneca Integrated Cardio Metabolic Center, Department of Laboratory Medicine, Karolinska Institutet at Karolinska University Hospital Huddinge, Stockholm, Sweden; 2Department of Physiology, Institute of Neuroscience and Physiology, University of Gothenburg, Gothenburg, Sweden; 3Division of Renal Medicine, Department of Clinical Sciences, Intervention and Technology, Karolinska Institutet, Stockholm, Sweden; and 4Division of Bioscience, Department of Cardiovascular, Renal and Metabolic Diseases, Innovative Medicines Biotech Unit, AstraZeneca, Gothenburg, Sweden

ABSTRACT Background Inflammatory processes play an important role in the pathogenesis of glomerulopathies. Finding novel ways to suppress glomerular inflammation may offer a new way to stop disease progression. However, the molecular mechanisms that initiate and drive inflammation in the glomerulus are still poorly understood. Methods We performed large-scale gene expression profiling of glomerulus-associated G protein– coupled receptors (GPCRs) to identify new potential therapeutic targets for glomerulopathies. The expression of Gprc5b in disease was analyzed using quantitative PCR and immunofluorescence, and by analyzing published microarray data sets. In vivo studies were carried out in a podocyte-specificGprc5b knockout mouse line. Mechanistic studies were performed in cultured human podocytes. Results We identified an orphan GPCR, Gprc5b, as a novel gene highly enriched in podocytes that was significantly upregulated in common human glomerulopathies, including diabetic nephropathy, IgA nephropathy, and lupus nephritis. Similar upregulation of Gprc5b was detected in LPS-induced nephropathy in mice. Studies in podocyte-specific Gprc5b knockout mice showed that Gprc5b was not essential for normal development of the glomerular filtration barrier. However, knockout mice were partially protected from LPS-induced proteinuria and recruitment of inflammatory cells. Mechanistically, RNA sequencing in Gprc5b knockouts mice and experiments in cultured human podocytes showed that Gpr5cb regulated inflammatory response in podocytes via NF-kB signaling. Conclusions GPRC5b is a novel podocyte-specific receptor that regulates inflammatory response in the glomerulus by modulating the NF-kB signaling pathway. Upregulation of Gprc5b in human glomerulopathies suggests that it may play a role in their pathogenesis.

JASN 30: 1573–1586, 2019. doi: https://doi.org/10.1681/ASN.2019010089

Glomerular disease processes are the most common Received January 28, 2019. Accepted May 9, 2019. cause of ESRD. Glomerular injury can be caused by Published online ahead of print. Publication date available at systemic diseases, such as in diabetic nephropathy www.jasn.org. (DN) and lupus nephritis (LN), or more glomerulus- ﬁ Correspondence: Dr. Jaakko Patrakka, KI/AZ Integrated Cardio speci c disorders, such as IgA nephropathy (IgAN). Metabolic Center, Department of Laboratory Medicine, Karolinska Patients with glomerulopathies are treated with Institute at Karolinska University Hospital Huddinge, Blickagången 6, angiotensin-converting enzyme inhibitors and/or 141 57 Huddinge, Sweden. Email: [email protected] angiotensin receptor II blockers to diminish proteinuria. Copyright © 2019 by the American Society of Nephrology

JASN 30: 1573–1586, 2019 ISSN : 1046-6673/3009-1573 1573 BASIC RESEARCH www.jasn.org

When disease progresses rapidly, steroids and other immu- Significance Statement nosuppressive drugs can be used. These pharmaceutical treatment options do not specifically target kidney Mechanisms regulating inflammatory response in glomeruli are tissue, and currently we lack molecular targets in the poorly understood. In this study, the authors identify a novel – glomerulus to develop more kidney-directed therapeutic G protein coupled receptor (GPCR) that is highly enriched in podocytes, called Gprc5b. Gprc5b is upregulated in common 1 options. human glomerular diseases, such as IgA nephropathy, lupus ne- As podocyte loss is a key event in the progression of glo- phritis, and diabetic nephropathy. Studies in knockout animals merulopathies, many studies have concentrated on analyzing and human podocytes grown in culture show that Gprc5b pro- molecular mechanisms of podocyte death.2 However, podo- motes glomerular inflammation via regulation of NF-kB pathway cyte loss is often a late event in the disease and it is likely that in podocytes. Upregulation of Gprc5b in human diseases sug- fi gests that this mechanism may play an important role in the targeting earlier changes is a more ef cient approach. Glomer- pathogenesis of common glomerulopathies. ular inflammation and recruitment of inflammatory cells is occurring early in glomerulopathies.3 Involvement of inflammation in the progression of glomerulopathies, such as LN and Transgenic Mouse Lines IgAN, is obvious. Similarly, although traditionally DN has been Gene targeting of Gprc5b gene was performed by The European considered a noninflammatory glomerular disease, global Conditional Mouse Mutagenesis Program (www.eucomm.org). transcript profiling studies have shown inflammatory signal- The construct was targeted to surround the exon 2 of the tm1a(EUCOMM)Wtsi ing and recruitment of immune cells in diabetic glomeruli.4 GPRC5b gene, allele Gprc5b . The mice were The inflammatory signature in glomerulopathies and in a mixed C57bl/6 and 129Sv background. We crossed these fl diabetic glomeruli includes activation of key transcriptional mice with an FLP-deleter line to generate a oxed mouse line, fl Gprc5b Gprc5b-fl Gprc5b-fl regulator NF-kB and JAK/STAT pathways, which are known ox ( ). was crossed with a podocin-cre Gprc5b fi Gprc5b- to drive cytokine production and influx of inflammatory line to inactivate speci cally in podocytes ( cKO). cells.527 The role of podocytes in this process is poorly The genotyping was done by PCR using genomic DNA understood. extracted from ear biopsy specimens. Primers for genotyp- G protein–coupled receptors (GPCRs) are a large family of ing were as follows: CAS-R1-Term: tcgtggtatcgttatgcgcc; Gprc5b Gprc5b proteins that participate in a wide variety of biologic processes. -87598-F: gctggaaggtttctccctct; -87598-R: They have been successfully exploited by the pharmaceutical aagagacaaccaccagacagg. tm14(CAG-td-Tomato)Hze/J industry as 20%–30% of the market share of clinically used The Gt(ROSA)26Sor mice were drugs target members of this family.8,9 To identify potential crossed with a podocin-cre line to activate the td-Tomato ex- fi new pharmaceutical targets in the glomerulus, we recently pression speci cally in podocytes. Breeding and genotyping performed a high-throughput expressional profiling of glo- were done according to standard procedures. merular GPCRs.10 In this study, we follow up on our previous All animal studies were carried out in Preclinical Laboratory study and describe an orphan GPCR, Gprc5b, as a novel, (Karolinska Institutet) and were approved by the Ethical highly podocyte-enriched molecule. We show that Gprc5b Committee on Research Animal Care. expression is induced in common human glomerulopathies andthatitmodulatesinflammatory response in podocytes LPS-Induced Proteinuria Mouse Model via NF-kB pathway. We speculate that Gprc5b could be a Ten-week-old mice were treated by a single intraperitoneal novel molecular target to suppress inflammatory response injection of 13 mg/g body wt of LPS (L9641; Sigma). A total in glomerular disease processes. of 18 littermate control controls and 19 Gprc5b-cKO animals were included. Urine was collected before the injection and at 6, 12, 24, 36, and 56 hours after the injection. Albumin and METHODS creatinine values were measured using Albuwell kit (catalog number 1011; Exocell) and Quantichrome creatinine assay kit Human Material (catalog number DICT-500; BioAssay Systems), respectively. Control human kidney tissue was from kidneys that were removed because of renal carcinoma or from renal biopsy spec- Immunofluorescence imens taken from healthy, living, related donor kidneys at We used frozen human and mouse samples fixed in cold ace- Karolinska University Hospital (Stockholm, Sweden). Only tone and blocked with normal goat serum (G9023; Sigma). The kidney tissue that was histologically normal was used. Renal primary antibodies, incubated overnight at 4°C, were: Gprc5b biopsy samples from patients with DN were collected at (catalog number HPA015247, 1:500; Atlas Antibodies), CD31 Karolinska University Hospital (Stockholm, Sweden) and (catalog number 303106, 1:2000; Biolegend), PDGFRb (catalog Sahlgrenska University Hospital (Gothenburg, Sweden). number MAB1263, 1:1000; R&D System), CD45 (catalog num- The local ethical committees at both sites approved the study ber ab64100, 1:250; Abcam), mouse nephrin (catalog number (approval no. 2010/579-31/1, Stockholm, Sweden; approval BP5030, 1:200; Acris). The human nephrin antibody has been no. 413-09, Gothenburg, Sweden). described previously.11

1574 JASN JASN 30: 1573–1586, 2019 www.jasn.org BASIC RESEARCH

Histologic Analysis and Electron Microscopy Signaling Technology), b-actin (catalog number 8227, For histologic analyses, kidney samples were fixed in 4% para- 1:2000; Abcam). formaldehyde followed by dehydration and embedding in paraffin. At least eight controls and eight Gprc5b-cKO animals Podocytes in Cell Culture were analyzed in each experiment. Hematoxylin and eosin and Human podocytes were cultured as previously described.15 We periodic acid–Schiff staining were done using standard proto- developed stable podocyte cell lines expressing Gprc5b by cols. Transmission electron microscopy (FEI Tecnai Spirit transfecting the cells using Lipofectamine 2000 (Thermo BioTWIN) was performed on renal cortex samples fixed Fisher Scientific). The cells were transfected with p6595 with glutaraldehyde, following standard procedures. MSCV-IP-N-HA-GPRC5B (catalog number 34894; Addgene) or with an empty vector. Clones were obtained by limited di- Quantitative and Conventional PCR lution and expanded under puromycin (P8833; Sigma) selec- Glomeruli were isolated from human and mouse kidneys as tion. The overexpression of Gprc5b in different clones was published previously.12,13 All the quantitative PCRs were per- confirmed by Western blotting. Both control and GPRC5b formed following the standard methods and using SYBR overexpressing cells were deprived overnight with 0.1% Green (catalog number 1725271; Bio-Rad), and the conven- FBS medium and treated with 1 mM of LPS (L9641; Sigma), tional PCR was performed using HotStarTaq polymerase 25 ng/ml of EGF (AF10015A; Peprotech), or 100 ng/ml of (catalog number 03643; Qiagen). The primers used were as Wnt3 (P27467; R&D Systems). To study the role of GPRC5b follows: mGprc5b-F: atgcgggagacagcatttga, mGprc5b-R: gga- in Wnt and EGFR pathways, we used the inhibitors IWP-2 cagccatttcagtccct; huGprc5b-F: cgtggcatcagagagaaaga, (20 nM sc-252928; Santa Cruz Biotechnology) and AG1478 huGprc5b-R: cccagggtccccaggaggaa; Gprc5b exon2–3_F: (5 nM 65852; Merck Millipore), respectively. The inhibitors tgcgggagacagcatttgat, Gprc5b exon2–3_R: gcggagcagttggga- were added 30 minutes before the treatment. tagtc; Gprc5b exon3-_4F: ggactatcccaactgctccg, Gprc5b exon3-_4R: agtttgcaggacgattccgt; Nephrin_F: gagagccccatt- Dual-Luciferase Assay caaaggct, Nephrin_R: agaaggagctcacggtttcg; Coro2b_F: aatg- Weused the Dual-Luciferase Reporter Assay System (Promega) gaaccccttcatcgac, Coro2b_R: agttggcctcctgcagaaca; wt1_f: to measure NF-kB activation using standard procedures. A gtagccccgactcttgtacg, wt1_r: gtcctggtgtgggtcttcag; Aquaporin3_R: vector carrying the Renilla (Renilla reniformis) luciferase re- gctggccggtcgtgaagact, hAquaporin3_F: tgtttcgggccccaatggca; porter gene was used as an internal control. To determinate mccl2_R: ttctttgggacacctgctgct, mccl2_F: ttaacgccccact- the changes in NF-kB activity, we transfected the cells with ex- cacctgc; huccl2_R: gtgtctggggaaagctagggg, huccl2_F: pression vectors encoding human TLR4, human MD-2, and gaccccaagcagaagtggg; IL-6_F: tacatcctcgacggcatctc, IL-6_R: ELAM-1 luciferase reporter plasmid monitoring NF-kB activity. gctacatttgccgaagagcc; IL-10_F: aactgagacatcagggtggc, IL-10_R: aaggtttctcaaggggctgg; hu28S-F: ttgaaaatccgggggagag, hu28S-R: Next-Generation RNA Sequencing of Mouse Glomeruli acattgttccaacatgccag; mGAPDH-F: tgttcctacccccaatgtgt, RNA isolated from glomeruli was measured using the Agilent mGAPDH-R: tgtgagggagatgctcagtg.Thevalueswerecalcu- BioAnalyzer 2100. All samples passed the quality criteria by lated using 28S or GAPDH as housekeeping gene, and using RNA integrity number .8.0, 28S/18S.1.0. RNA samples were 2DD the 2 CTmethod.14 then prepared and sequenced by the commercial service of BGI Tech. Western Blotting All raw sequence reads available in FastQ format was Western blotting was performed using standard procedures. mapped to the mouse genome (mm10), human genome We used Clarity western ECL substrate (catalog number (hg38), and crab-eating macaque (Macaca fascicularis 1705061; Bio-Rad) and the ChemiDoc touch imaging system v5.0.92) using Tophat2 with Bowtie2 option,16,17 where adap- (Bio-Rad). Antibodies used were: anti-Gprc5b (Human Pro- tor sequences were removed using trim_galore before read tein Atlas, 1:500), anti-P65 (catalog number D14E12, 1:1000; mapping. Binary Alignment Map files containing the align- Cell Signaling Technology), anti-pP65 (Ser536) (catalog num- ment results were sorted according to the mapping position. ber 3033, 1:1000; Cell Signaling Technology), anti-FYN (cat- Raw read counts for each gene were calculated using featur- alog number ab184276, 1:1000; Abcam), anti-pFYN (Y530) eCounts from Subread package.18 (catalog number ab182661, 1:1000; Abcam), pAKT (catalog DEseq2 was used to perform the analysis of differential gene number 9271, 1:1000; Cell Signaling Technology), pERK (cat- expression, where genes with raw counts as input.19 The dif- alog number 4370, 1:1000; Cell Signaling Technology), EGFR ferentially expressed genes were identified by adjusted P value (catalog number 4267, 1:1000; Cell Signaling Technology), for multiple testing using Benjamini–Hochberg correction pEGFR (catalog number 2234, 1:1000; Cell Signaling Technol- with false discovery rate values ,0.1. ogy), pSmad2 (catalog number 8828, 1:1000; Cell Signaling Technology), active b-catenin (catalog number 05–665, 1:400; Statistical Analyses Millipore), caspase3 (catalog number 966, 1:1000; Cell Signal- Statistical difference between two groups was analyzed by t test. ing Technology), pSTAT3 (catalog number 9131, 1:1000; Cell Data were analyzed using the Prism software (GraphPad,

JASN 30: 1573–1586, 2019 Gprc5b Modulates Inﬂammation in Glomeruli 1575 BASIC RESEARCH www.jasn.org

San Diego, CA) and was recorded as the mean6SEM, with genes analyzed were strongly downregulated (Figure 2B). P,0.05 defined as significant. This downregulation is probably due to loss of podocytes, as glomeruli in this study were collected from patients with advanced disease (mean GFR 31 mL/min per 1.73 m2). In the RESULTS data set generated from isolated DN glomeruli by Ju et al.5 Gprc5b expression was significantly upregulated, whereas GPRC5b Is Highly Enriched in Podocytes other podocyte genes were either downregulated or not We identified Gprc5b through a large-scale transcript pro- significantly changed (Figure 2C). Taken together, Gprc5b filing where we compared the expression of 378 GPCRs in expression seems to be induced in podocytes during the human glomerulus to rest of cortex tissue (ROK).10 In that progression of DN. study, Gprc5b was one of the highly glomerular-enriched To see whether this upregulation was specifictoDNora GPCRs. To validate this, we performed Western blotting shared molecular mechanism in glomerulopathies, we inves- and conventional PCR on samples obtained from human tigated the expression of Gprc5b in previously published tran- glomerulus and ROK, and in mouse podocytes isolated scriptome data on isolated glomeruli in LN and IgAN. In the from a transgenic mouse line expressing td-tomato in two LN data sets analyzed,5,21 Gprc5b was significantly in- podocytes. Western blotting showed a strong band, sized duced in the glomerulus, whereas other podocyte markers approximately 40 kDa in human glomeruli (Figure 1A), were downregulated (Figure 2, D and E). In the three whereas this protein was not detected in the ROK fraction. IgAN data sets analyzed,5,6,22 Gprc5b was similarly either The size of the protein was in line with published data.20 By significantly induced or unaltered in the glomerulus, PCR, only the glomerular fraction generated a product when whereas other podocyte markers were either downregulated Gprc5b-specific primers were used (Figure 1B). The purity of or unaltered (Figure 2, F–H). the human glomerular fraction was verified by analyzing podocin To exclude the possibility that Gprc5b was expressed in and nephrin expression, whereas aquaporin 3 was used as a tu- glomerulopathies by some other cell type(s) than podocytes, bular marker (Figure 1, A and B). In mice, podocytes isolated we performed immunofluorescence staining in biopsies col- from dt-tomato animals showed an enrichment of Gprc5b when lected from five patients with DN. No significant expression compared with the rest of the glomerulus (Figure 1C). Nephrin was detected outside podocytes as demonstrated by double was used to validate the purity of podocytes. labeling with nephrin (Figure 2I). To summarize, Gprc5b ex- The location of Gprc5b in human glomeruli was evalu- pression seems to be induced in podocytes in three common ated by double-immunofluorescence staining of Gprc5b human glomerulopathies, DN, LN, and IgAN. with nephrin, PDGFRb,andCD31,markersforpodocytes, mesangial cells, and endothelial cells, respectively. Gprc5b GPRC5b Is Not Essential for the Normal Development colocalized with the podocyte foot process marker nephrin, of the Glomerulus Filtration Barrier whereas no overlap was detected between PDGFRb and To understand the biologic role of Gprc5b in the kidney, we CD31 (Figure 1, D–I). To localize Gprc5b in foot processes, generateda podocyte-specificknockoutmouseline(Gprc5b-cKO). we performed immunoelectron microscopic analysis. The We deleted exon 2 of the mouse Gprc5b gene by a strategy signal for Gprc5b was restricted to podocytes and enriched in the summarized in Figure 3A. To validate the genotype of our apical membrane of foot processes (Figure 1J). Taken together, mice we used primers that amplified a 361 bp wild-type our data indicate that Gprc5b is a novel highly podocyte- band, whereas the knockout allele generated 226 and 457 bp enriched molecule localized at the apical plasma membrane. products (Figure 3B). The success of the knockout strategy was tested using PCR in glomeruli isolated from Gprc5b Expression Is Induced in DN and Other Gprc5b-cKO and control mice. We designed specific primers Common Glomerulopathies for the regions between the exons 2–3 and 3–4. In Gprc5b-cKO To determine whether Gprc5b is involved in the pathogenesis mice the band for exons 2–3 is missing, but not the band for of human glomerular diseases, we analyzed its expression in exons 3–4, indicating that exon 2 is depleted but Gprc5b different glomerulopathies. First, we performed qPCR on glo- mRNA is still generated (Figure 3C). To evaluate whether meruli isolated from DN and healthy controls. In diabetic this mRNA is stable and able to generate a protein, we per- glomeruli, podocyte-specific genes Nphs1, Wt1, and Coro2b formed Western blotting for GPRC5b in isolated glomer- showed a significant downregulation, which was probably sec- uli. The Gprc5b-cKO mice showed no protein expression ondary to a loss of podocytes (Figure 2A). In contrast to this, (Figure 3D), indicating the successful inactivation of Gprc5b Gprc5b levels were not significantly altered, suggesting that gene in podocytes. Gprc5b expression was induced in remaining podocytes. The depletion of Gprc5b in podocytes did not have any Tovalidate our finding, we analyzed publicly available tran- notable effect on the development of the glomerulus, as shown scriptome data on isolated glomeruli in human DN. In the data by standard light microscopy (Figure 3E). The ultrastructure set by Woroniecka et al.,7 Gprc5b expression was not signifi- of the filtration barrier as shown by electron microscopy cantly changed in DN glomeruli, although other podocyte was also unaffected, including the presence of regularly

1576 JASN JASN 30: 1573–1586, 2019 www.jasn.org BASIC RESEARCH

A BCinterdigitating foot processes and normal slit diaphragms (Figure 3F). Functionally, the filtration barrier seemed intact as no albumin- huGlom huROK mPod mROG uria was detected in animals aged between huROK huGlom GPRC5b GPRC5b 1 week and 1 year (data not shown). Nephrin Nephrin Aquaporin 3 GAPDH LPS-Induced Glomerulopathy in 28S Mouse Phenocopies Upregulation of Gprc5b in Human Glomerulopathies D Next, we wanted to challenge Gprc5b-cKO GPRC5b 40KDa mice with a disease model that would copy the molecular phenotype detected in human glomerulopathies, i.e., induction of Gprc5b expression in glomeruli. We analyzed previously published23227 and in- house transcriptome data sets generated Podocin 42KDa from isolated glomeruli. We detected no GAPDH 37KDa induction of Gprc5b in DN models db/db and OVE26 (Supplemental Figure 1). Sim- ilarly, no significant change in Gprc5b E GPRC5b F nephrin G Merge expression was detected in adriamycin- induced or other nephropathy models. In contrast to this, LPS-induced glomerulopathy, characterized by rapidly developing transient proteinuria, showed over five- fold induction of Gprc5b expression in isolated glomeruli (Supplemental Figure 1).23 Similarly, quantitative PCR analysis of H GPRC5b/PDGFRβ I GPRC5b/CD31 Gprc5b in glomeruli isolated from streptozotocin-induced diabetic mice showed a

conventional PCR. Nephrin and aquaporin 3 were used as glomerular and rest of kidney fraction markers, respectively, whereas 28S gene was used as a loading control. (C) The PCR for Gprc5b in FACS-isolated mouse podocytes(mPod)andrest ofglomerulus(mROG) shows an enrichment in podocytes. Nephrin was used to validate the purity of podocyte J fractionsandGAPDHasaloadingcontrol. (D) Immunofluorescence staining for GPRC5b (green) in human kidney cortex shows strong immunoreactivity in glomeruli and no significant signal in extraglomerular areas. (E–G) Double staining of Gprc5b (green) and podocyte foot process marker nephrin (red) shows nearly complete colocalization (yellow). DAPI (blue) was used as a nucleus marker. (H and I) Double labeling with the mesangial marker PDGFRb (red) or with the endothelial marker CD31 (red) does not show significant over- Figure 1. GPRC5b is enriched in human and mouse podocytes. (A) In human kidney, lapping reactivity. (J) Immunoelectron micros- Western blotting for Gprc5b detects a 40 kDa protein only in the glomerulus (huGlom) copy shows gold label for Gprc5b (arrowheads) and not in rest of the kidney (huROK). Podocin was used to show the purity of the glo- on podocyte plasma membrane. Magnifications: merular fraction and GAPDH as a loading control. (B) Gprc5b transcript shows strong 340 in (D), 3200 in (E–G), 3400 in (H and I). enrichment in the glomerulus when compared with the rest of the kidney as detected by Scale bar, 250 nm in (J).

JASN 30: 1573–1586, 2019 Gprc5b Modulates Inﬂammation in Glomeruli 1577 BASIC RESEARCH www.jasn.org

A B DN (Woroniecka KI et al.) 2,0 control disease 2,5 control disease

2 * 1,5 * * * * 1,5 1,0 1

Relative gene 0,5 0,5 expression (qPCR) Normalized signal value 0,0 0 GPRC5b NPHS1 WT1 Coro2b GPRC5B NPHS1 WT1 CORO2B

CDDN (Ju W. et al.) LN (Berthier CC et al.) control disease control disease 15 12 * * * * * * 10 * 8

5 4 Normalized signal value Normalized signal value 0 0 GPRC5B NPHS1 WT1 CORO2B GPRC5B NPHS1 WT1 CORO2B

EFLN (ERCB) IgAN (Liu P et al.)

15 control disease 15 control disease * 10 * 10 * * * 5 5

Normalized signal value 0 Normalized signal value 0 GPRC5B NPHS1 WT1 CORO2B GPRC5B NPHS1 WT1 CORO2B

GHIgAN (Ju W et al.) IgAN (Berthier CC et al.)

15 control disease 15 control disease * * * * * 10 * 10

5 5 Normalized signal value 0 Normalized signal value 0 GPRC5B NPHS1 WT1 CORO2B GPRC5B NPHS1 WT1 CORO2B

Figure 2. Glomerular Gprc5b expression is induced in common human glomerulopathies. (A) Quantitative PCR analysis of Gprc5b in isolated glomeruli shows that the expression is not signiﬁcantly altered in DN, whereas podocyte genes Nphs1 and Wt1 are downregulated. *P,0.001. (B) Analysis of microarray data generated by Woroniecka et al.7 in glomeruli isolated from DN and control patients. The expression of Gprc5b is unchanged whereas other podocyte genes are strongly downregulated in DN. *P,0.001. (C) In the data set generated by Ju et al.,5 the expression of Gprc5b is upregulated in DN glomeruli, whereas other podocyte genes are either downregulated or unchanged. *P,0.001. (D) In the data set generated by Berthier et al.,21 the expression of Gprc5b is induced in LN

1578 JASN JASN 30: 1573–1586, 2019 www.jasn.org BASIC RESEARCH

I GPRC5b Nephrin Merge NHK

GPRC5b Nephrin Merge DN

Figure 2. Continued. significant induction of Gprc5b expression (Supplemental effacement that was similar in both animal groups (Supplemental Figure 1). However, as Gprcb-cKO mice have B6 background Figure 2B). that is resistant to streptozotocin induced kidney injury, we To understand the molecular mechanisms of how Gprc5b challenged the Gprcb-cKO mice with LPS. deficiency in podocytes diminishes albuminuria after LPS injection, we performed RNA sequencing in glomeruli iso- Gprc5b Deficiency in Podocytes Delays Response to lated at 24 hours, the time point when the levels of albumin- LPS by Modulating the Activation of NF-kB Pathway uria were maximally different between the groups. A total of As albuminuria levels in LPS-induced nephropathy are chang- 4440 genes were differentially expressed between knockout ing rapidly, we used a rather large animal cohort in our and control glomeruli at 24 hours (Figure 4B, Supplemental study, including a total of 18 control and 19 Gprc5b-cKO Appendix 1). Pathway analysis revealed 164 pathways to be mice. The control mice, as described previously,23 significantly affected. Among the top affected pathways we developed a peak of albuminuria 24 hours after LPS injec- saw signaling events that associate with inflammatory re- tion (Figure 4A). Interestingly, the Gprc5b-cKO animals sponse, including “Cytokine-cytokine receptor interaction,” showed a significant delay in the response, presenting the “Chemokine signalling pathway,” and “NF-kB signalling path- peak of albuminuria at 36 hours, 12 hours after the control way” (Supplemental Figure 3). As Gprc5b has been linked to mice (P,0.05). Both controls and Gprc5b-cKOs recovered NF-kB signaling in adipose tissue,11 we analyzed this pathway completely from albuminuria at 56 hours after the LPS further. The gene expression of all five NF-kB family members injection (Figure 4A). Morphologic analysis in animals eu- were significantly downregulated in Gprc5b-cKO glomeruli thanized 36 hours after the injection demonstrated no compared with controls (Figure 4C). This was observed to- obvious changes at light microscopic level (Supplemental gether with a downregulation of well known downstream tar- Figure 2A), whereas electron microscopy showed foot process gets of NF-kB activation, such as TNF-a, IL-6, IL-1B, and

glomeruli, whereas other podocyte genes are either downregulated or not changed. *P,0.001. (E) In the data set generated by European Renal cDNA Bank (ERCB), the expression of Gprc5b is induced in LN glomeruli, whereas other podocyte genes are either downregulated or not changed. *P,0.001. (F) In the data set generated by Liu et al.,22 the expression of Gprc5b is upregulated in IgAN glomeruli, whereas other podocyte genes are not changed. *P,0.001. (G) In the data set generated by Ju et al.5 the expression of Gprc5b is induced in IgAN glomeruli, whereas other podocyte genes are downregulated or unchanged. *P,0.001. (H) In the data set generated by Berthier et al.,21 the expression of Gprc5b is unchanged in IgAN glomeruli, whereas other podocyte genes are downregulated. *P,0.001. (I) Immuno- ﬂuorescence staining for Gprc5b (green) and nephrin (red) shows colocalization in human DN glomeruli. No staining for Gprc5b is detected in other cell types than podocytes. DAPI (blue) was used as a nucleus marker. Magniﬁcations: 3200.

JASN 30: 1573–1586, 2019 Gprc5b Modulates Inﬂammation in Glomeruli 1579 BASIC RESEARCH www.jasn.org

AB+/+ -/- +/- ICAM1 (Figure 4D). Taken together, the re- 457 bp sults indicate that control mice present a 1 lacZ neo 23 361 bp strong activation of NF-kB in glomeruli 226 bp 24 hours after LPS injection, whereas C Gprc5b-cKO mice are protected from this FLP activation. Of note, to validate that LPS induced NF-kB signaling pathway in wild-type animals, we compared RNA se-

Control GPRC5b-cKO quencing data at 0 and 24 hours after the 1 2 3 Exons 2-3 injection of LPS in control glomeruli. RNA Exons 3-4 sequencing data showed a robust activation of NF-kB signaling pathway as both compo- GAPDH Pod-cre Exon nents of this pathway and down-stream Lox P D effectors were induced (Supplemental Figure 4, Supplemental Appendix 2). FRT 13 Gprc5b Modulates the Recruitment

GPRC5b-cKO Control GPRC5b-cKO of Inflammatory Cells to Glomeruli GPRC5b 40KDa To analyze downstream effects of suppressed NF-kB activation in Gprc5b-cKO, we analyzed whether this affected the pro- tubulin 42KDa duction of chemotactic molecules that play a E key role in recruitment of inflammatory cells. A large number of the chemoattrac- tants were downregulated in Gprc5b-cKO glomeruli when compared with controls, as shown by RNA sequencing data (Figure 5A). The downregulation of ccl2 was validated using quantitative PCR (Figure 5B). To analyze whether this influenced the recruitment of inflammatory cells to glomeruli, we counted the number of CD45-positive glomerular cells in LPS-treated animals.

control GPRC5b-cKO F heterozygote (+/2), and Gprc5b-cKO (2/2) mice. The wild-type band is 361 bp and bands for the floxed allele are 226 and 457 bp, respectively. (C) PCR in mouse glomeruli isolated from Gprc5b-cKO and control mice using specific primers for exons 2–3 and exons 3–4. In Gprc5b-cKO glomeruli no product is amplified when using primers for exons 2–3. (D) Western blotting for Gprc5b in mouse glomerulus shows a band around 40 kDa in control mice, whereas this protein is missing in glomeruli isolated from Gprc5b-cKO mice. Tubulin was used as a loading control. (E) No significant differences were observed in control GPRC5b-cKO light microscopic examination of Gprc5b-cKO kidneys when compared with control kidneys. Figure 3. Podocytes and the glomerulus mature normally in podocyte-specific (F) In electron microscopic analysis, no mor- knockout (KO) animals. (A) Strategy to generate podocyte-specificKOmicefor phologic abnormalities were detected in the Gprc5b. Mice carrying a targeting cassette for exon 2 of Gprc5b gene was crossed glomerular filtration barrier in Gprc5b-cKO mice. with a FLP-deleter line to generate a floxed allele, followed by crossing with podocin- Magnifications: 3200 in (E). Scale bar, 250 nm crelinetoremoveexon2specifically in podocytes. (B) Genotyping of wild-type (+/+), in (F).

1580 JASN JASN 30: 1573–1586, 2019 www.jasn.org BASIC RESEARCH

ABLPS 24h 200 35 WT (n=18) * 30 150 GPRC5b-cKO (n=19) 25

20 100 15 ACR (µg/mg)

–log10(p value) 10 50 5

0 0 0h 6h 12h 24h 36h 56h -6-4-20246 log2FoldChange

C LPS 24h D LPS 24h 60 300 * 50 * 250 * * * 40 200

30 * 150 * 20 * 100 Relative gene expression Relative gene expression 10 50 * * * 0 0 WT KO WT KO WT KO WT KO WT KO WT KO WT KO WT KO WT KO WT KO WT KO NFkB1 NFkB2 Rela Relb Rel TNF IL-6 IL-1B ICAM1 NFKBIA NFKBIB (p50) (p52/P100) (p65)

Figure 4. Gprc5b-cKO mice show a delay in response to LPS challenge and significant transcriptional changes in NF-kB pathway. (A) Urinary albumin-to-creatinine ratio (ACR) at different time points after the LPS injection. In Gprc5b-cKO mice the peak of albuminuria is delayed by 12 hours when compared with control mice. Values are expressed as mean 6SD; *P,0.05. (B) RNA sequencing of isolated glomeruli was performed 24 hours after LPS injection. Volcano plot shows upregulated (red) and downregulated (blue) genes. Differentially expressed genes are defined as genes with P,0.001 and false discovery rate ,0.1, when compared Gprc5b-cKO versus controls. (C) The expression of five family members of NF-kBweresignificantly downregulated in Gprc5b-cKO mice when compared with control glomeruli as shown by RNA sequencing. (D) The expression of downstream targets of NF-kB activation TNF-a, IL-6, IL-1B, and ICAM1 were robustly downregulated in Gprc5b mice glomeruli. *P,0.05.

Interestingly, we detected less CD45-positive cells in Gprc5b-cKO treatment with LPS induced an increase in the phosphorylation glomeruli (Figure 5, C and D), indicating that Gprc5b expres- levels of P65, one of the family members of NF-kB, at 4 and 24 sion in podocytes mediated the recruitment of immune cells to hours (Figure 6, A and B). In contrast, the cells overexpressing glomerular tufts under pathologic stimuli. Gprc5b showed a constitutive activation of P65, which was further exacerbated by LPS treatment. The levels of total Overexpression of Gprc5b Causes a Constitutive P65 did not differ between two groups or different conditions Activation of NF-kB (Figure 6, A and B). No signiﬁcant differences were detected in To validate the results obtained in mice, we carried out in vitro pEGFR/tEGFR and pb-catenin/tb-catenin ratios (Figure 6, A, experiments using immortalized human podocytes.16 As po- C, and D). Tovalidate that Gprc5b also modulates downstream docytes in culture often lose the expression of genes enriched targets of NF-kB signaling, we measured levels of the proin- in podocytes,28 it was not surprising that Gprc5b was not ex- ﬂammatory cytokines ccl2, IL6, and M-CSF1 by quantitative pressed in this cell line (Figure 6A). Therefore, we generated a PCR. A strong elevation of these known NF-kB effectors was stable human podocyte cell line that overexpressed Gprc5b detected in Gprc5b-expressing podocytes (Figure 6E). and mimicked the in vivo situation by challenging the cells The results were further validated using a luciferase re- with 10 mM LPS for 4 and 24 hours. In control cells, the porter for NF-kB activation. In the luciferase assay, the cells

JASN 30: 1573–1586, 2019 Gprc5b Modulates Inﬂammation in Glomeruli 1581 BASIC RESEARCH www.jasn.org

ABControls Gprc5b -cKOs CcI2 CxcI1 3 40 ** CxcI16 CxcI10 2.5 CxcI5 30 Cx3cI1 Cxcr2 2 CcI7 20 CcI3 1.5 CcI5 CcI12 1 10 CxcI2 Relative ccl2 expression CxcI3 0.5 Ccr1 CcI17 0 Ccr5 0 control GPRC5b-cKO CcI6 CxcI11 CcI9 CcI20 CxcI14 C D 5 control GPRC5b-cKO * 4

2 Nephrin / CD45 of CD45 positive cells

o 1 n

0 control GPRC5b-cKO

Figure 5. Expression of chemokines and recruitment of inflammatory cells is suppressed in Gprc5b-cKO glomeruli. (A) The expression of 21 chemokines is significantly downregulated in Gprc5b-cKO glomeruli as shown by a heat map generated from RNA sequencing data. (B) Validation of the downregulation of ccl2 expression by quantitative PCR. GAPDH was used as a housekeeping gene. **P,0.01. (C and D) Immunofluorescence analysis of CD45-positive cells in glomeruli 36 hours after LPS treatment shows significantly less immune cells in Gprc5b-cKO glomeruli when compared with controls. Magnifications: 3200. *P,0.05. overexpressing Gprc5b presented significantly more NF-kB caspase 3 were detected in Gprc5b-expressing cells (Supple- activation when compared with control cells (Figure 6F). mental Figure 5), we observed a significant elevation in total Similarly, after LPS treatment, Gprc5 cells showed more re- and phosphorylated forms of EGFR and b-catenin (Figure 6A). porter activity than control cells (Figure 6F). Taken together, As this suggested that the activation of NF-kB might be down- the in vitro experiments supported our results in knockout stream of EGFR and/or b-catenin, we tested this hypothesis by mice that Gprc5b modulates the activation of NF-kB signaling treating Gprc5b-overexpressing cells with IWP-2 (b-catenin pathway in podocytes. inhibitor) and AG1478 (EGFR inhibitor). EGF treatment caused an activation of EGFR in both control and Gprc5b cells Activation of NF-kB by Gprc5b Is Not Mediated by and, as expected, AG1478 prevented this activation (Supple- EGFR– or b-Catenin–Dependent Signaling Pathways mental Figure 5, B and C). However, the activation of EGFR As several molecular pathways can activate NF-kB in podocytes,29 did not happen simultaneously with the phosphorylation of we analyzed the phosphorylation levels of several molecular P65 in control or in Gprc5b cells, and the treatment of Gprc5b triggers upstream of NF-kB activation, such as EGFR, ERK, cells with only AG1478 did not reduce the phosphorylation STAT3, Smad2, FYN, and b-catenin. Whereas no difference levels of P65. Thus, EGFR did not mediate activation of P65 in intheactivationofFYN,STAT3,Smad2,ERK,AKT,and this cell model. Similarly, for the b-catenin pathway, after

1582 JASN JASN 30: 1573–1586, 2019 www.jasn.org BASIC RESEARCH

A B pP65/tP65 3 *** ** Control GPRC5b 2 ***

LPS 0h 4h 24h 0h 4h 24h 1 Fold change GPRC5b 40 KDa 0 LPS 0h 4h 24h 0h 4h 24h tP65 65 KDa C pEGFR/tEGFR 1,5 pP65 65 KDa 1

tEGFR 175 KDa 0,5 Fold change pEGFR 175 KDa 0 LPS 0h 4h 24h 0h 4h 24h

(total)β-catenin 92 KDa D pβ-catenin/tβ- catenin 1,5 (active)β-catenin 92 KDa 1

Actin 42 KDa 0,5 Fold change 0 LPS 0h 4h 24h 0h 4h 24h E F 900 *** * 14 * 700 ccl2 *** * control *** 500 IL6 12 GPRC5b 300 M-CSF1 * 10 100 * # *** 25 * 8

20 *** * 6 * 15 4 Relative gene expression Relative luciferase activity 10 * * 2 5 * * *

0 0 luc luc+Tl4+Md2 Luc luc+Tl4+Md2 control GPRC5b LPS --++ LPS 0h 4h 24h 0h 4h 24h

Figure 6. Overexpression of Gprc5b in cultured human podocytes causes a constitutive activation of NF-kB. (A) In cultured podocytes, a stable overexpression of Gprc5b promotes the activation of NF-kB after treatment with LPS as shown by increased phosphorylation of P65 (pP65). In the presence of Gprc5b, even untreated cells show P65 activation. No change in total levels of P65 was detected between different samples. Similarly, Gprc5b expression promotes the activation of EGFR and b-catenin as detected by increased phosphorylation of the proteins. The total amount of these proteins was also increased. Actin was used as a loading control. (B–D) Ratios of pP65/tP65, pEGFR/tEGFR, and pb-catenin/tb-catenin. (E) Quantitative PCR analysis of ccl2, IL-6, and M-CSF1 demonstrates the induction of these NF-kB downstream targets in Gprc5b expressing cells. *P,0.001 compared with control 0 hours; ***P,0.001 compared with GPRC5b at 0 hours. (F) Activation of NF-kB using a luciferase reporter system in cultured podocytes with and without 4 hours LPS exposure. Expression of Gprc5b results in overactivation of NF-kB in both untreated and LPS treated cells. Values for all the experiments are expressed as mean6SD of three independent experiments. *P,0.05; **P,0.01; ***P,0.001. Luc, luciferase reporter of NF-kB pathway; Md2, cofactor of Tlr4; Tl4, toll like receptor 4. activation of the pathway with Wnt3, inhibition with IWP-2 Figure 5, D and E). To summarize, Gprc5b-mediated activation did not affect the phosphorylation status of P65, although it of NF-kB pathway in podocytes does not seem to be mediated strongly suppressed b-catenin activation (Supplemental by EGFR or b-catenin.

JASN 30: 1573–1586, 2019 Gprc5b Modulates Inﬂammation in Glomeruli 1583 BASIC RESEARCH www.jasn.org

DISCUSSION induction of Gprc5b expression. LPS injection causes endotox- emia with an increase in serum creatinine levels, and is used Wedescribe a novel, highly podocyte-enriched GPCR, Gprc5b, as a model of AKI.36 However, it is also a well established model the expression of which is upregulated in common human to study glomerular disease processes as it induces albumin- glomerulopathies. Studies in podocyte-specificknockout uria and podocyte damage.23 Obviously, studies in clinically mice show that the induction of Gprc5b expression modulates more relevant mouse models are indicated to explore further inflammatory response in the glomerulus. We suggest that this the involvement of Gprc5b in glomerulopathies. orphan GPCR can offer a new way to suppress inflammation Gprc5b belongs to “Retinoic Acid-Inducible G-protein- in a glomerulus-targeted fashion. coupled receptors” that comprise four orphan receptors RNA sequencing of knockout glomeruli revealed the (Gprc5a–d). Recently, we showed that Gprc5a was also highly NF-kB signaling pathway as a downstream target of Gprc5b. enriched in podocytes, where it contributed to the progres- This association was validated in cell culture, where overex- sion of DN by modulating EGFR signaling.10 In cell culture, pression of Gprc5b in podocytes promoted the activation of we have detected that Gprc5a inhibits NF-kB signaling (Ma X, NF-kB. The NF-kB pathway has been shown to be activated in Patrakka J et al. unpublished data). Thus, Gprc5a and Gprc5b glomeruli isolated from patients with DN, IgAN, and LN.527 seem to have opposite effects in podocytes. It is interesting As we detected induction of Gprc5b in these common glomer- that both of these two structurally closely related receptors ulopathies, we speculate that Gprc5b regulates the activation of are highly enriched in podocytes. To explore the functional NF-kB signaling in human glomerular disorders and thus relation of Gprc5a to Gprc5b, we are generating a double- plays a pathogenic role in disease progression. Data generated knockout mouse line in which both receptors are inactivated in animal models suggest that inhibiting NF-kB signaling can in podocytes. have protective effects in the glomerulus. In passive Heymann We used cultured podocytes to validate the connection be- nephritis rats, activation of NF-kB is detected in podocytes.30 tween Gprc5b and NF-kB. Overexpression studies showed the Treatment with the NF-kB inhibitor pyrrolidine dithiocarba- activation of NF-kB signaling pathway using multiple read- mate suppressed this activation and reduced albuminuria lev- outs. Moreover, we detected an activation of both EGFR and els.30 In a rat crescentic GN model, inhibition of NF-kB with b-catenin, two known upstream effectors of NF-kB. This decoy oligonucleotides had similar beneficial effects by reduc- prompted us to test whether Gprc5b-mediated activation of ing proteinuria and histologic damage.31 In podocytes, the im- NF-kB was mediated by one of these pathways. However, stud- portance of NF-kB pathway has been shown in a transgenic ies using EGFR– and b-catenin–specific inhibitors failed to mouse line in which the suppression of NF-kB exclusively in verify this link. As we did not detect significant changes in podocytes ameliorated the development of glomerulopathy.32 EGFR and b-catenin pathways in our RNA sequencing anal- As abnormal NF-kB activation is a central event in the ysis in vivo, we believe that these pathways are not centrally pathogenesis of many human diseases, therapeutic targeting involved in mediating Gprc5b signaling in podocytes. of the pathway has been investigated intensively for over two To summarize, we identified Gprc5b as a highly podocyte- decades.33 However, these efforts have failed to provide specific enriched GPCR that modulates inflammatory response via inhibitors for clinical use, mainly because of dose-limiting tox- NF-kB pathway. Because of its induction in common human icities caused by global suppression of NF-kB. Recently, efforts glomerulopathies, we suggest GPRC5b as a new attractive have been concentrated on finding ways to manipulate the therapeutic target to suppress glomerular inflammation. signaling pathway in a more cell-/tissue-selective fashion. More studies are warranted to understand the molecular con- In the kidney, Gprc5b, being a highly podocyte-enriched nection between Gprc5b and NF-kB activation in podocytes. molecule, could offer a therapeutic window to target NF-kB in a cell-specificfashion. Gprc5b has previously been investigated in pancreas, brain, ACKNOWLEDGMENTS and adipose tissue. In the pancreas Gprc5b was suggested to be a negative modulator of insulin secretion,34 whereas in Dr. Zambrano was responsible for the design of the study, the majority the spinal cord the downregulation of Gprc5b in neurons of animal and cell culture work, the analysis of experimental data, and is associated with neuropathic pain.35 In adipose tissue, con- wrote the first draft of the manuscript. Dr. Möller-Hackbarth was re- ventional Gprc5b knockout mice showed resistance to in- sponsible for animal work, analysis of experimental data, and manu- flammation induced by a high-fat diet, which is a key event script revision. Dr. Li was responsible for computational analysis of in the development of insulin resistance.20 Interestingly, this the RNA sequencing experiment and manuscript revision. Dr. Charrin protective effect seems to be due to lack of NK-kBpathway was responsible for animal work, analysis of experimental data, and activation in adipocytes, similarly to the situation observed manuscript revision. Ms. Schwarz was responsible for animal work, in our podocyte-specific knockout animals. analysis of experimental data, and manuscript revision. Dr. Rodriguez One limitation of our study was the lack of available mouse was responsible for the design of the study, generation of the knock- models as LPS-induced glomerulopathy was the only one that out mouse line, and manuscript revision. Dr. Nyström provided tissue molecularly phenocopied the situation in patients with the material and was responsible for manuscript revision. Dr. Wernerson

1584 JASN JASN 30: 1573–1586, 2019 www.jasn.org BASIC RESEARCH provided tissue material and was responsible for manuscript revision. 5. Ju W, Greene CS, Eichinger F, Nair V, Hodgin JB, Bitzer M, et al.: De- Dr. Lal was responsible for analysis of experimental data and manu- fining cell-type specificity at the transcriptional level in human disease. – script revision. Dr. Patrakka was responsible for development and Genome Res 23: 1862 1873, 2013 6. Ju W, Nair V, Smith S, Zhu L, Shedden K, Song PXK, et al.: ERCB, design of the study, analysis of experimental data, manuscript C-PROBE, NEPTUNE, and PKU-IgAN Consortium: Tissue transcriptome- revision, and overall planning and execution of the study. driven identification of epidermal growth factor as a chronic kidney disease biomarker. Sci Transl Med 7: 316ra193, 2015 7. Woroniecka KI, Park AS, Mohtat D, Thomas DB, Pullman JM, Susztak K: DISCLOSURES Transcriptome analysis of human diabetic kidney disease. Diabetes 60: 2354–2369, 2011 Dr. Lal is employed by AstraZeneca. Nyström has a consulting agreement 8. Wacker D, Stevens RC, Roth BL: How ligands illuminate GPCR molec- with AstraZeneca. Dr. Patrakka’s research is supported by AstraZeneca. All ular pharmacology. Cell 170: 414–427, 2017 of the remaining authors have nothing to disclose. 9. Roth BL, Irwin JJ, Shoichet BK: Discovery of new GPCR ligands to illuminate new biology. Nat Chem Biol 13: 1143–1151, 2017 10. Ma X, Schwarz A, Sevilla SZ, Levin A, Hultenby K, Wernerson A, et al.: FUNDING Depletion of Gprc5a promotes development of diabetic nephropathy. JAmSocNephrol29: 1679–1689, 2018 11. Ruotsalainen V, Patrakka J, Tissari P, Reponen P, Hess M, Kestilä M, Dr. Patrakka’s laboratory was supported by grants from Karolinska Institutet/ et al.: Role of nephrin in cell junction formation in human nephro- AstraZeneca ICMC, Swedish Diabetes Foundation, Marianne and Marcus genesis. Am J Pathol 157: 1905–1916, 2000 Wallenberg Foundation, Westman Foundation, Swedish Kidney Foundation, 12. Takemoto M, He L, Norlin J, Patrakka J, Xiao Z, Petrova T, et al.: Large- and through the regional agreement on medical training and clinical research scale identification of genes implicated in kidney glomerulus devel- (ALF) between Stockholm County Council and the Karolinska Institute. opment and function. EMBO J 25: 1160–1174, 2006 Dr. Nyström was supported by The IngaBritt and Arne Lundberg Research Foun- 13. Yamamoto T: Isolation and Enrichment of Glomeruli Using Sieving dation, the Swedish Research Council, the Swedish Kidney Foundation, and Techniques. In: Renal and Urinary Proteomics: Methods and Protocols, through the regional agreement on medical training and clinical research edited by Thongboonkerd V, Germany, Wiley-VCH Verlag GmbH & Co, (ALF) between Västra Götaland County and Sahlgrenska University Hospital. 2009, pp 1–7 14. Livak KJ, Schmittgen TD: Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. SUPPLEMENTAL MATERIAL Methods 25: 402–408, 2001 15. Saleem MA, O’Hare MJ, Reiser J, Coward RJ, Inward CD, Farren T, et al.: A This article contains the following supplemental material conditionally immortalized human podocyte cell line demonstrating – online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ nephrin and podocin expression. J Am Soc Nephrol 13: 630 638, 2002 16. Langmead B, Salzberg SL: Fast gapped-read alignment with Bowtie 2. ASN.2019010089/-/DCSupplemental. Nat Methods 9: 357–359, 2012 Supplemental Figure 1. Gprc5b expression in various mouse 17. Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg SL: To- models of glomerulopathy. pHat2: Accurate alignment of transcriptomes in the presence of in- Supplemental Figure 2. Histological evaluation of kidneys 36 hours sertions, deletions and gene fusions. Genome Biol 14: R36, 2013 fi after LPS injection. 18. Liao Y, Smyth GK, Shi W: featureCounts: An ef cient general purpose program for assigning sequence reads to genomic features. Bio- Supplemental Figure 3. Table of the most significantly affected informatics 30: 923–930, 2014 pathways in Gprc5b-cKO glomeruli when compared with control 19. Love MI, Huber W, Anders S: Moderated estimation of fold change and glomeruli collected 24 hours after the LPS injection. dispersion for RNA-seq data with DESeq2. Genome Biol 15: 550, 2014 Supplemental Figure 4. Comparison of glomerular transcriptomes 20. Kim YJ, Sano T, Nabetani T, Asano Y, Hirabayashi Y: GPRC5B activates between 0 and 24 hours after LPS injection in control animals. obesity-associated inflammatory signaling in adipocytes. Sci Signal 5: Supplemental Figure 5. Analysis of different pathways upstream of ra85, 2012 k 21. Berthier CC, Bethunaickan R, Gonzalez-Rivera T, Nair V, Ramanujam M, NF- B activation in Gprc5b-expressing cells. Zhang W, et al.: Cross-species transcriptional network analysis defines Supplemental Appendix 1. RNA sequencing data in Gprc5b-cKO shared inflammatory responses in murine and human lupus nephritis. and control glomeruli 24 hours after the injection of LPS. JImmunol189: 988–1001, 2012 Supplemental Appendix 2. RNA sequencing data in wild-type 22. Liu P, Lassén E, Nair V, Berthier CC, Suguro M, Sihlbom C, et al.: fi glomeruli 0 and 24 hours after the injection of LPS. Transcriptomic and proteomic pro ling provides insight into mesangial cell function in IgA nephropathy. J Am Soc Nephrol 28: 2961–2972, 2017 23. Sun Y, He L, Takemoto M, Patrakka J, Pikkarainen T, Genové G, et al.: REFERENCES Glomerular transcriptome changes associated with lipopolysaccha- ride-induced proteinuria. Am J Nephrol 29: 558–570, 2009 1. Lal MA, Patrakka J: Understanding podocyte biology to develop novel 24. Doné SC, Takemoto M, He L, Sun Y, Hultenby K, Betsholtz C, et al.: kidney therapeutics. Front Endocrinol (Lausanne) 9: 409, 2018 Nephrin is involved in podocyte maturation but not survival during 2. Tharaux PL, Huber TB: How many ways can a podocyte die? Semin glomerular development. Kidney Int 73: 697–704, 2008 Nephrol 32: 394–404, 2012 25. Ding M, Cui S, Li C, Jothy S, Haase V, Steer BM, et al.: Loss of the tumor 3. Holdsworth SR, Tipping PG: Leukocytes in glomerular injury. Semin suppressor Vhlh leads to upregulation of Cxcr4 and rapidly progressive Immunopathol 29: 355–374, 2007 glomerulonephritis in mice. Nat Med 12: 1081–1087, 2006 4. Reidy K, Kang HM, Hostetter T, Susztak K: Molecular mechanisms of 26. Reiniger N, Lau K, McCalla D, Eby B, Cheng B, Lu Y, et al.: Deletion of the diabetic kidney disease. JClinInvest124: 2333–2340, 2014 receptor for advanced glycation end products reduces glomerulosclerosis

JASN 30: 1573–1586, 2019 Gprc5b Modulates Inﬂammation in Glomeruli 1585 BASIC RESEARCH www.jasn.org

and preserves renal function in the diabetic OVE26 mouse. Diabetes 32. Brähler S, Ising C, Hagmann H, Rasmus M, Hoehne M, Kurschat C, et al.: 59: 2043–2054, 2010 Intrinsic proinﬂammatory signaling in podocytes contributes to podo- 27. Ratelade J, Arrondel C, Hamard G, Garbay S, Harvey S, Biebuyck N, et al.: cyte damage and prolonged proteinuria. Am J Physiol Renal Physiol – A murine model of Denys-Drash syndrome reveals novel transcriptional 303: F1473 F1485, 2012 k b targets of WT1 in podocytes. Hum Mol Genet 19: 1–15, 2010 33. Freitas RHCN, Fraga CAM: NF- B-IKK pathway as a target for drug 28. Shankland SJ, Pippin JW, Reiser J, Mundel P: Podocytes in culture: development: Realities, challenges and perspectives. Curr Drug Targets 19: 1933–1942, 2018 Past, present, and future. Kidney Int 72: 26–36, 2007 34. Soni A, Amisten S, Rorsman P, Salehi A: GPRC5B a putative glutamate- 29. Rangan G, Wang Y, Harris D: NF-kappaB signalling in chronic kidney receptor candidate is negative modulator of insulin secretion. Biochem disease. Front Biosci 14: 3496–3522, 2009 Biophys Res Commun 441: 643–648, 2013 30. Mudge SJ, Paizis K, Auwardt RB, Thomas RJ, Power DA: Activation of 35. Chung HJ, Kim JD, Kim KH, Jeong NY: G protein-coupled receptor, nuclear factor-kappa B by podocytes in the autologous phase of pas- family C, group 5 (GPRC5B) downregulation in spinal cord neurons is – sive Heymann nephritis. Kidney Int 59: 923 931, 2001 involved in neuropathic pain. Korean J Anesthesiol 66: 230–236, 2014 31. Tomita N, Morishita R, Tomita S, Gibbons GH, Zhang L, Horiuchi M, 36. Colbert JF, Ford JA, Haeger SM, Yang Y, Dailey KL, Allison KC, et al.: et al.: Transcription factor decoy for NFkappaB inhibits TNF-alpha- Amodel-speciﬁc role of microRNA-223 as a mediator of kidney induced cytokine and adhesion molecule expression in vivo. Gene Ther injury during experimental sepsis. Am J Physiol Renal Physiol 313: 7: 1326–1332, 2000 F553–F559, 2017

1586 JASN JASN 30: 1573–1586, 2019

1 Supplemental Table of Contents

2 Supplementary Figure 1. Gprc5b expression in various mouse models of

3 glomerulopathy.

4 Supplementary Figure 2. Histological evaluation of kidneys 36h after LPS-injection.

5 Supplementary Figure 3. Table of the most significantly affected pathways in Gprc5b-

6 cKO glomeruli when compared to control glomeruli collected 24h after the LPS

7 injection.

8 Supplementary Figure 4. Comparison of glomerular transcriptomes between 0 and

9 24h after LPS injection in control animals.

10 Supplementary Figure 5. Analysis of different pathways up-stream of NF-kB

11 activation in Gprc5b-expressing cells.

12 Supplementary Spreadsheet 1. RNAseq data in Gprc5b-cKO and control glomeruli

13 24h after the injection of LPS.

14 Supplementary Spreadsheet 2. RNAseq data in wild type glomeruli 0h and 24h after

15 the injection of LPS.

17 Supplementary Figure 1. Gprc5b expression in various mouse models of

18 glomerulopathy. (A) Microarray data extracted from previously published data sets

19 (23-27) and an in-house data sets (db/db in 129 male mice, adriamycin nephropathy in

20 8-week old BALB/c male mice). In LPS-induced glomerular damage a strong up-

21 regulation of Gprc5b is detected whereas other models do not show significant changes

22 or no consistent induction. NPH=nephrin KO, DMS=diffuse mesangial sclerosis

23 (Denysh-Drash syndrome), VHL=von-hippel-lindau nephropathy; OVE26=Ove26

24 mice overexpressing calmodulin in pancreatic β-cells, ADM=Adriamycin-induced

25 nephropathy. (B) qPCR analysis of Gprc5b expression in glomeruli isolated from

26 streptozotocin-induced nephropathy (STZ, 8-week old FVB male mice) shows

27 significant up-regulation of Gprc5b.

29 Supplementary Figure 2. Histological evaluation of kidneys 36h after LPS-

30 injection. (A) Histological analysis (PAS staining) of kidneys 36h after the LPS

31 injection shows apparently normal morphology in Gprc5b-cKO and control samples.

32 (B) Both Gprc5b-cKO and control mice exhibit podocyte foot process effacement 36h

33 after the injection, but no significant difference between groups is detected by counting

34 numbers of slits / GBM length. Magnifications: x40 in A; Scale bar: B = 250nm.

36 Supplementary Figure 3. Table of the most significantly affected pathways in

37 Gprc5b-cKO glomeruli when compared to control glomeruli collected 24h after

38 the LPS injection. A total of 164 significantly affected pathways was identified, here

39 we show the 40 most significantly affected pathways (P<0.001 and FDR < 0.1).

41 Supplementary Figure 4. Comparison of glomerular transcriptomes between 0

42 and 24h after LPS injection in control animals. (A) The expression of five family

43 members of NF-kB are significantly up-regulated after 24h when compared to

44 glomeruli collected from uninjected (0h) glomeruli as shown by RNA sequencing. (B)

45 The expression of down-stream targets of NF-kB activation, TNF-a, IL-6, IL-1B,

46 ICAM1, NF-kB inhibitor alpha (NF-kBIA) and NF-kB inbitor beta (NF-kBIB) are up-

47 regulated 24h after LPS-injection. * p<0.05.

49 Supplementary Figure 5. Analysis of different pathways up-stream of NF-kB

50 activation in Gprc5b-expressing cells. (A) No significant differences were detected

51 in the levels of activation of FYN, STAT3, Smad2, ERK, AKT and caspase 3 between

52 Gprc5b over-expressing and control cells as shown by Western blotting for activated

53 forms of these proteins. Actin was used as a loading control. (B-C) Inhibition of EGF-

54 activation with AG1478 does not affect the activation of P65 as detected by blotting for

55 the phosphorylated form of the protein. (D-E) Inhibition of β-catenin pathway with

56 IWP-2 does not affect the activation of P65 as detected by blotting for the

57 phosphorylated form of the protein.

3 Supplementary figure 1 5

A B - 7 7 6 q=4*10 6 * 5 5

4 4

Fold change 3 3

q=0.006 NS Gprc5b expression 2 NS 2 NS NS NS NS NS NS 1 NS 1 Relative 0 0 control STZ LPS VHL NPH DMS 4 days 7 days OVE26 14 days 2 months 4 months 6 months

db/db ADM Supplementary Figure 2

A control GPRC5b-cKO

B control GPRC5b-cKO 2,5 2,0 1,5 1,0 0,5 Number of slits 0,0 WT KO Supplementary Figure 3

Matched Pathway Ids Descriptions P values FDR genes path:mmu00030 Pentose phosphate pathway - Mus musculus 29 of 31 0.00 0.01 path:mmu00040 Pentose and glucuronate interconversions - Mus musculus 32 of 36 0.00 0.01 path:mmu00053 Ascorbate and aldarate metabolism - Mus musculus 23 of 27 0.00 0.01 path:mmu00140 Steroid hormone biosynthesis - Mus musculus 57 of 87 0.00 0.01 path:mmu00190 Oxidative phosphorylation - Mus musculus 133 of 139 0.00 0.01 path:mmu00230 Purine metabolism - Mus musculus 173 of 179 0.00 0.01 path:mmu00240 Pyrimidine metabolism - Mus musculus 102 of 104 0.00 0.01 path:mmu00562 Inositol phosphate metabolism - Mus musculus 69 of 70 0.00 0.01 path:mmu00591 Linoleic acid metabolism - Mus musculus 33 of 50 0.00 0.01 path:mmu00630 Glyoxylate and dicarboxylate metabolism - Mus musculus 29 of 30 0.00 0.01 path:mmu00640 Propanoate metabolism - Mus musculus 31 of 31 0.00 0.01 path:mmu00830 Retinol metabolism - Mus musculus 63 of 89 0.00 0.01 path:mmu00980 Metabolism of xenobiotics by cytochrome P450 - Mus musculus 58 of 65 0.00 0.01 path:mmu00983 Drug metabolism - other enzymes - Mus musculus 45 of 51 0.00 0.01 path:mmu03008 Ribosome biogenesis in eukaryotes - Mus musculus 78 of 83 0.00 0.01 path:mmu03010 Ribosome - Mus musculus 136 of 148 0.00 0.01 path:mmu03013 RNA transport - Mus musculus 156 of 170 0.00 0.01 path:mmu03015 mRNA surveillance pathway - Mus musculus 88 of 96 0.00 0.01 path:mmu03018 RNA degradation - Mus musculus 73 of 83 0.00 0.01 path:mmu03020 RNA polymerase - Mus musculus 30 of 30 0.00 0.01 path:mmu03030 DNA replication - Mus musculus 35 of 35 0.00 0.01 path:mmu03040 Spliceosome - Mus musculus 129 of 135 0.00 0.01 path:mmu03050 Proteasome - Mus musculus 45 of 45 0.00 0.01 path:mmu03060 Protein export - Mus musculus 27 of 31 0.00 0.01 path:mmu03410 Base excision repair - Mus musculus 35 of 35 0.00 0.01 path:mmu04060 Cytokine-cytokine receptor interaction - Mus musculus 232 of 265 0.00 0.01 path:mmu04062 Chemokine signaling pathway - Mus musculus 186 of 198 0.00 0.01 path:mmu04064 NF-kappa B signaling pathway - Mus musculus 94 of 104 0.00 0.01 path:mmu04080 Neuroactive ligand-receptor interaction - Mus musculus 207 of 285 0.00 0.01 path:mmu04110 Cell cycle - Mus musculus 123 of 125 0.00 0.01 path:mmu04120 Ubiquitin mediated proteolysis - Mus musculus 139 of 145 0.00 0.01 path:mmu04141 Protein processing in endoplasmic reticulum - Mus musculus 162 of 167 0.00 0.01 path:mmu04142 Lysosome - Mus musculus 119 of 124 0.00 0.01 path:mmu04144 Endocytosis - Mus musculus 272 of 294 0.00 0.01 Supplementary Figure 4

A B Supplementary figure 5

A Control GPRC5b LPS (1μM) 0h 4h 24h 0h 4h 24h pFYN 55 KDa

pSTAT3 80 KDa

pSmad2 42 KDa

pERK 44 KDa

pAKT 60 KDa

Caspase 3 17 KDa

Actin 42 KDa B C

Control GPRC5b 200 pP65 EGF (100ng/μl) - + - + - + - + 150 AG1478 (40 μM) - - + + - - + + 100 pP65 50 pEGFR 0 % protein expression Control GPRC5b actin EGF - + - + - + - + AG1478 - - + + - - + + D Control GPRC5b Wnt3 (100ng/ml) - - + + - - + + IWP-2 (5 μM) - + - + - + - + pP65 Active β-catenin

actin

E Active β-catenin pP65 600 1500 500 400 1000 300 200 500 100 0 0 Control GPRC5b % protein expression Control GPRC5b % protein expression Wnt3 - - + + - - + + Wnt3 - - + + - - + +

control IWP-2 - + - + - + - + control control IWP-2 - + control - + - + - + GPRC5b GPRC5b GPRC5b GPRC5b