BASIC RESEARCH www.jasn.org

Single-Cell Transcriptome Profiling of the Identifies Key Cell Types and Reactions to Injury

Jun-Jae Chung ,1 Leonard Goldstein ,2 Ying-Jiun J. Chen,2 Jiyeon Lee ,1 Joshua D. Webster,3 Merone Roose-Girma,2 Sharad C. Paudyal,4 Zora Modrusan,2 Anwesha Dey,5 and Andrey S. Shaw1

Due to the number of contributing authors, the affiliations are listed at the end of this article.

ABSTRACT Background The glomerulus is a specialized bed that is involved in production and BP control. Glomerular injury is a major cause of CKD, which is epidemic and without therapeutic options. Single-cell transcriptomics has radically improved our ability to characterize complex organs, such as the kidney. Cells of the glomerulus, however, have been largely underrepresented in previous single-cell kidney studies due to their paucity and intractability. Methods Single-cell RNA sequencing comprehensively characterized the types of cells in the glomerulus from healthy mice and from four different disease models (nephrotoxic serum , diabetes, doxo- rubicin toxicity, and CD2AP deficiency). Results Allcelltypesintheglomeruluswereidentified using unsupervised clustering analysis. Novel marker and signatures of mesangial cells, vascular cells of the afferent and efferent arteri- oles, parietal epithelial cells, and three types of endothelial cells were identified. Analysis of the disease models revealed cell type–specific and injury type–specific responses in the glomerulus, including acute activation of the Hippo pathway in after nephrotoxic immune injury. Conditional deletion of YAP or TAZ resulted in more severe and prolonged in response to injury, as well as worse . Conclusions Generation of comprehensive high-resolution, single-cell transcriptomic profiles of the glo- merulus from healthy and injured mice provides resources to identify novel disease-related genes and pathways.

JASN 31: ccc–ccc, 2020. doi: https://doi.org/10.1681/ASN.2020020220

The glomerulus, the site of filtration in the kidney, mechanisms are used to restore function after injury is a capillary bed composed of endothelial cells, po- or how acute glomerular injury progresses to chronic docytes, and mesangial cells, as well as less abun- dant cell types, such as the parietal epithelial cells (PECs) and vascular smooth muscle cells (SMCs) Received March 13, 2020. Accepted June 7, 2020. (Supplemental Figure 1). Loss of glomerular func- Published online ahead of print. Publication date available at tion is the most common cause of CKD, a major www.jasn.org. health care problem affecting approximately 15% Present address: Dr. Sharad C. Paudyal, Department of Radiation 1 of the population. Glomerular injury is caused by Oncology, Dana-Farber Institute, Brigham and Women’s factors such as diabetes and hypertension, as well as Hospital, Harvard Medical School, Boston, Massachusetts. by immune injury. Glomeruli are particularly sus- Correspondence: Dr. Andrey S. Shaw, Department of Research ceptible to injury because podocytes are largely un- Biology, Genentech, 1 DNA Way, MS93b, South San Francisco, able to regenerate and, therefore, tissue damage is CA 94080. Email: [email protected] considered irreversible. It is not known what reparative Copyright © 2020 by the American Society of Nephrology

JASN 31: ccc–ccc, 2020 ISSN : 1046-6673/3110-ccc 1 BASIC RESEARCH www.jasn.org

fibrosis. Despite their critical roles in kidney function and disease, Significance Statement cells of the glomerulus have largely been underrepresented in pre- vious kidney single-cell RNA sequencing (scRNA-seq) studies due Single-cell transcriptomics techniques have revolutionized the to their paucity and difficulty of isolation. ability to characterize cells from heterogeneous organs like the Here, we performed scRNA-seq using purified glomeruli to kidney. Although glomerular disorders are an important cause of CKD, a thorough characterization of the cells in the glomerulus has characterize all of the cell types and we analyzed their reaction remained challenging due to the technical difficulties of isolating to four common types of kidney injury: immune, metabolic, undamaged cells, especially from glomeruli of diseased animals. toxic, and genetic injury. Our work, which includes sequenc- This study provides a comprehensive single-cell atlas, based on ing of approximately 75,000 glomerular cells, provides a com- approximately 75,000 cells, from glomeruli of healthy mice and prehensive transcriptional signature of all cell types in the mice injured in four ways, including all cell types present. The data set will be a valuable resource for generating precise tools to in- glomerulus, including those that have not been well charac- terrogate specific glomerular cell types and in identifying genes terized previously, such as mesangial cells, PECs, juxtaglomer- involved in the pathogenesis of glomerular diseases. ular (JG) cells, and arteriolar SMCs. Results from four disease models provide new insights into the glomerular response to acute injury and its progression to CKD. C57BL/6J mice were given a single intraperitoneal injection of 20 mg/kg doxorubicin hydrochloride (Pfizer). All animal proce- dures were conducted under protocols approved by the Institu- METHODS tional Animal Care and Use Committee at Genentech, and were performed in accordance with the Guide for the Care and Use of Reagents Laboratory Animals. The reagents used were as follows: Dynabeads M-450 Tosylac- Wwtr1 Yap1 tivated (Thermo Fisher Scientific), Liberase TM (Sigma-Al- Generation of and CKO Mice drich), DNase I (Sigma-Aldrich), trypsin (Thermo Fisher Homologous recombination and mouse embryonic stem (ES) 3–5 fi Scientific), Dispase II (Roche Applied Science), Collagenase cell technology was used to generate genetically modi ed D (Roche Applied Science), paraformaldehlyde (Electron Mi- mouse strains with a Wwtr1 or Yap1 CKO. For Wwtr1,agene 9 croscopy Sciences), and OCT (Sakura Finetek). targeting vector was constructed with a 1415-bp arm of 5 homology corresponding to GRCm38/mm10 3 – 9 Antibodies (chr3): 57,577,558 57,576,144 and a 2066-bp arm of 3 ho- – Anti-FHL2, anti-SERPINE2, anti-RGS2, anti-ADAMTS5, mology corresponding to chr3: 57,574,633 57,572,568. The fl 1 anti–calponin 1, and anti–a smooth muscle actin antibodies 1510-bp region anked by loxP sites ( 1 2) corresponds – were purchased from Abcam. Anti-PKCa antibody was pur- to chr3: 57,576,143 57,574,634. For Yap1, a gene-targeting 9 chased from Thermo Fisher Scientific. Anti-PDGFRb (APB5) vector was constructed with a 1990-bp arm of 5 homology – antibody was purchased from eBioscience. Anti-YAP/TAZ corresponding to GRCm38/mm10 chr9: 8,003,842 8,001,853 (D24E4) antibody was purchased from Cell Signaling. Anti– and a 2045-bp arm of 39 homology corresponding to chr9: b-actin (AC-15) and anti-vinculin (hVIN-1) antibodies were 8,001,304–7,999,260. The 548-bp region flanked by loxP sites purchased from Sigma-Aldrich. Phycoerythrin-conjugated (exon 2) corresponding to chr9: 8,001,852–8,001,305. anti-CCL2 (2H5) and phycoerythrin/Cy7-conjugated The final vector was confirmed by DNA sequencing, line- anti–TNF-a (MP6_XT22) antibodies were purchased from arized, and used to target C2 (C57BL/6N) ES cells using stan- Biolegend. Alexa Fluor–conjugated secondary antibodies dard methods (G418-positive and ganciclovir-negative selec- were purchased from Thermo Fisher Scientific. tion).6 C57BL/6N C2 ES cells7 were electroporated with 20 mg of linearized targeting vector DNA and cultured under Mice drug selection essentially as described. Positive clones were C57BL/6J and BTBR ob/ob (BTBR.Cg-Lepob/WiscJ) mice were identified using long-range PCR followed by sequence purchased from Jackson Laboratory. CD2AP-deficient mice confirmation. have been described previously.2 Generation of Wwtr1 Correctly targeted ES cells were subjected to karyotyping. (TAZ) and Yap1 (YAP) conditional knockout (CKO) strains Euploid gene-targeted ES cell clones were treated with Adeno- and CCL2-YPet reporter strain is described below. All animals FLP to remove PGK neomycin, ES cell clones were tested to were bred and housed at Genentech under specificpathogen- identify clones with no copies of the PGK neomycin cassette, free conditions with free access to chow and water and a and the correct sequence of the targeted was verified. The 12-hour day/night cycle. Only male mice were used. For the neph- presence of the Y chromosome was verified before microin- rotoxic serum nephritis model, age-matched C57BL/6J mice were jection into albino C57BL/6N embryos. Germline transmis- injected intravenously with 100 ml (for scRNA-seq analysis) or sion was obtained after crossing resulting chimeras with 2.3 ml/kg body wt (approximately 60 ml, for YAP/TAZ experi- C57BL/6N females. Genomic DNA from born pups was ments) of sheep anti-rat glomeruli serum (Probetex). For the screened by long-range PCR to verify the desired gene- doxorubicin nephropathy model, age- and weight-matched targeted structure before mouse colony expansion. Genotyping

2 JASN JASN 31: ccc–ccc,2020 www.jasn.org BASIC RESEARCH primers used to identify germline transmission were the follow- glomeruli and washed four to five times with HBSS until ing: Wwtr1.CKO primers were (1) TGGTCACAAGCGTTA samples were .98% pure by visual inspection under a AGC, (2) TGGTTCAAGCCTGTTAAATCA, and (3)CCTACT- microscope. CACCTGGCTGT; expected amplicon sizes were 255 bp for wild For preparation of single-cell suspensions, the purified glo- type, 289 bp for floxed, 440 bp for knockout. The Yap1.CKO meruli were resuspended in 1.25 ml of digestion buffer (0.5% primers were (1) TTGAGTTATGTAGGATGAGCATTA, (2) trypsin, 2.0 U/ml Dispase II, 2 U/ml Collagenase D, 10 U/ml GTATGTCACGGCAACCAA, and (3) TGACCAACCCTAAAG DNase I in prewarmed PBS without calcium and magnesium AGAGA; expected amplicon sizes were 246 bp for wild type, ions) and incubated at 37°C for 40–60 minutes with constant 280 bp for floxed, 320 bp for knockout. agitation (800 RPM) in a ThermoMixer. (We tested the use of cold active protease from Bacillus licheniformis,butitwasnot Generation of CCL2-Ypet Reporter Mice effective at dissociating glomeruli.) During the incubation, A bacterial artificial chromosome (BAC) clone harboring the samples were triturated by pipetting every 5 minutes for the mouse CCL2 gene (RP23-328G11) was obtained from the first 30 minutes. After 30 minutes, the glomeruli were me- BACPAC Resources Center at Children’s Hospital Oakland chanically sheared by passing through a 27 1/2 gauge needle Research Institute. The BAC was modified by inserting twice. The samples were then incubated for an additional 10– YPet-bGHpA (bovine growth hormone polyadenylation sig- 30 minutes until .98% of the cells had been dissociated into nal) into exon 1 of the CCL2 using the Counter Selection single cells by visual inspection under a microscope. Shearing BAC Modification Kit (Gene Bridges, Heidelberg, Germany). of the glomeruli with a syringe needle at earlier time points The 12.8-kb segment of the modified BAC including the YPet- resulted in selective loss of podocytes. Longer incubation in- bGHpA modified CCL2 gene and the 5-kb upstream and 5-kb creased the number of cells isolated, but decreased cell viabil- downstream regions was subcloned using homologous re- ity. The digestion was stopped by adding 10 ml ice-cold PBS combination into a linearized pBluescript II SK plasmid (Agi- with 10% FBS. The samples were then placed in a magnetic lent) that contained homology arms with flanking XhoI sites. separator to remove magnetic beads. The supernatant was The plasmid construct was amplified in Escherichia coli and collected, passed through a 40-mm cell strainer to remove re- purified using the EndoFree Plasmid Maxi Kit (Qiagen). The sidual cell aggregates, centrifuged at 300 3 g for 5 minutes at purified plasmid was digested with XhoIandthefragment 4°C, and resuspended in 1 ml of PBS with 0.1% BSA. Cell containing the CCL2 reporter construct was gel purified for density and viability of single-cell suspensions were deter- injection. C57BL/6 mice were injected by the Mouse Genetics mined using the Vi-CELL XR Cell Counter (Beckman Coul- Core at Washington University (St. Louis, MO). Genotyping ter). The method typically generated 0.53106–13106 cells per primers used to identify founders carrying the transgene were mouse with .90% viability. the following: (1) GCATCGACTTCAAGGAGGAC and (2) GTCAGGAACTCCAGCAGCAC, with expected amplicon Sample Processing, Library Preparation, and size of 295bp. Sequencing Sample processing for scRNA-seq was done using Chromium Preparation of Single Cells from Glomeruli Single Cell 39 Library and Gel Bead Kit version 2 according to Single-cell suspensions of glomerular cells were prepared from the manufacturer’s instructions (103 Genomics). Total cell biologic replicates for each sample using a previously de- density was used to calculate the volume of single-cell suspen- scribedmethodwithmodifications.8 First, glomeruli were iso- sion needed in the reverse master mix, aiming lated using magnetic beads. Kidneys were removed from mice to achieve approximately 6000 cells per sample. cDNAs and en bloc with the abdominal aorta attached and transferred to libraries were prepared according to the manufacturer’sin- petri dishes filled with ice-cold HBSS without calcium and structions (103 Genomics). cDNA amplification and prepa- magnesium ions. After cutting open the abdominal aorta ration of indexed libraries were performed using 12 and 14 along its length, each kidney was perfused directly through cycles of PCR, respectively. Libraries were profiled using the the renal arteries with 1 ml ice-cold HBSS containing mag- Bioanalyzer High Sensitivity DNA Kit (Agilent) and quantified netic beads (200 ml Dynabeads M-450 rinsed and resuspended using the Kapa Library Quantification Kit (Kapa Biosystems). in 10 ml HBSS). The kidneys were then minced into small Each library was sequenced in one lane of the HiSeq 4000 pieces (,1mm3) with a razor blade and incubated in 3 ml (Illumina) to achieve approximately 300 million reads follow- prewarmed digestion buffer (1.5 U/ml Liberase TM, 100 U/ml ing the manufacturer’s sequencing specifications (103 DNase I in HBSS) at 37°C for 20 minutes with constant agi- Genomics). tation (800 RPM) in a ThermoMixer (Eppendorf). All sub- sequent steps were performed on ice or at 4°C unless specified Sequencing Data Quality Control and Preprocessing otherwise. The digested kidneys were passed through a scRNA-seq data were processed with a custom pipeline. 100-mm cell strainer twice. The samples were washed twice Briefly, reads were demultiplexed based on perfect matches with ice-cold HBSS and resuspended in 1 ml HBSS. The sam- to expected cell bar codes. Transcript reads were aligned to ples were then placed in a magnetic separator to collect the the mouse reference genome (mm10) using GSNAP.9 Per-gene

JASN 31: ccc–ccc, 2020 scRNA-Seq Analysis of Mouse Glomeruli 3 BASIC RESEARCH www.jasn.org transcript counts were determined based on the number of Figure 1 and subsequent figures. To identify genes specificto unique molecular identifiers for mapped reads overlapping PECs and mesangial cells, we first identified genes highly en- in the sense orientation, allowing for one mismatch riched in each cell type compared with all other cell types when collapsing unique molecular identifier sequences. To (mean fold change, .3; FDR-adjusted P,0.01). The list was be considered for downstream analysis, cells were required then further refined by removing the genes with expression to exceed a minimum number of detected transcripts, where levels greater than or equal to LSMean 2 in any other cell type. a sample-specific cutoff was set to 0.1 times the total transcript Hierarchic clustering was performed using Ward linkage and count for cells at 30 (the 99th percentile for 3000 cells). Euclidean distance and shown as z-scores of normalized ex- Cells with .5% mitochondrial gene counts and genes de- pression levels. tected in ,0.1% of cells were excluded from further analysis. Genes in the protocadherin g family (Pcdhga1-12, Pcdhgb1-8, Mapping of Disease-Associated Genes Pcdhgc1-5) were removed from further analysis due to multi- For analysis of cell type–specific expression of FSGS disease– mapping. Doublets were removed using a combination of and CKD-associated genes, we generated pseudo-bulk data for high-count thresholding and gating for simultaneous expres- each cell type by calculating the mean expression levels for sion of two or more cell type–specific marker genes. Per-gene each gene. Heatmaps were generated to depict z-scores of nor- transcript counts were normalized by dividing by the total malized expression levels. transcript count for a given cell and multiplying by a scale factor of 10,000. Normalized counts were transformed using Pathway Analysis alog2(x 1 1) transformation. Pathway analyses were performed using the biologic interpre- tation function in Partek Flow. Lists of differentially expressed Dimensionality Reduction, Clustering Analysis, and (DE) genes (gene-specific analysis mean fold change, .1.5; T-Distributed Neighbor Embedding/Uniform Manifold FDR-adjusted P,0.01) were used as input to identify enriched Approximation and Projection Visualization terms or KEGG pathways. Partek Flow software (version 8.0.19.0707) was used for anal- ysis of single-cell data. Principal component (PC) analysis Immunofluorescence Microscopy (PCA) was performed on the highly variable genes, and PCs For immunofluorescence, kidneys were perfusion fixed by with eigenvalues $1 were chosen (maximum 100 PCs) to be transcardial perfusion with fixation buffer (4% paraformal- used for downstream clustering analysis and t-distributed dehlyde in PBS), immersion fixed in fixation buffer for an neighbor embedding or Uniform Manifold Approximation additional hour, and immersed in 30% sucrose in PBS at 4° and Projection (UMAP) visualization. To prevent clustering C overnight. The kidneys were frozen in OCT and cut into 6- artifacts due to tissue dissociation–induced stress, the to 15-mm sections with a cryostat (Leica Biosystems). The 140 dissociation-induced genes reported by van den Brink sections were blocked and permeabilized with Image-iT FX et al.10 were excluded from PCA and downstream clustering (Thermo Fisher Scientific) or 10% normal goat serum or nor- analyses. Clusters were identified using the Louvain clustering mal donkey serum (Jackson ImmunoResearch) in PBS with algorithm. For identification of subclusters within cell types, 0.3% Triton X-100 (Sigma-Aldrich). The slides were incu- cells belonging to each cell type were separated from the total bated with primary antibodies in PBS with 0.3% Triton X- population and reanalyzed as a separate sample. 100 and 1% BSA (Sigma-Aldrich) overnight at 4°C. After three washes with PBS, the slides were incubated with fluorescent Identification of Marker Genes and Differentially dye–conjugated secondary antibodies at room temperature Expressed Genes for 1 hour. After three washes with PBS, the slides were moun- Differential analysis was performed using the ted with ProLong Gold Antifade Mountant with 49,6-diami- Partek GSA (gene-specific analysis) function, which identifies dino-2-phenylindole (Thermo Fisher Scientific) and cured and uses the most suitable statistical model for each transcript overnight. Images were obtained using a Nikon A1R confocal (details can be found at https://documentation.partek.com/ microscope (Nikon). display/FLOWDOC/Gene-specific1Analysis). Marker genes for each cluster were identified by comparing cells in a specific Measurement of Albuminuria cluster with all remaining cells (mean fold change, .1.5; false- Spot urine samples were collected at the indicated time points. discovery rate [FDR]–adjusted P,0.01). Clusters with less Urinary (Bethyl Laboratories) and creatinine (Bio- than five marker genes were considered overclustering arti- Assay Systems) levels were quantified by ELISA according to facts. Expression of canonical marker genes (Supplemental the manufacturers’ instructions. Table 1) was used to identify and assign cell types (podocytes, mesangial cells, endothelial cells, vascular SMCs, immune Histopathology Analysis cells, PECs, tubular epithelial cells [TECs]) to clusters. PECs Kidney tissues were fixed for 24 hours at ambient temperature were identified by expression of Cldn1 and Pax8 in addition to in 10% neutral buffered formalin (VWR), and then processed marker genes. TECs were excluded from analysis in and embedded using a Tissue-Tek VIP processor (Sakura).

4 JASN JASN 31: ccc–ccc,2020 www.jasn.org BASIC RESEARCH

A B C 2 Ehd3 4 podocyte endothelial mesangial SMC immune PEC 4 Fbln2 Nphs1 Pdpn Flt1 Pecam1 Gata3 Pdgfrb Acta2 Ptprc Pax8 2 1 1 Mgp 2 7 3 t-SNE2 4 Trpv4 6 5 3 6 Bmx 5 7

t-SNE1 D E 1 4 Acta2 Tagln Myh11 Ren1Akr1b7 Rgs5 Rergl Map3k7cl Plvap Prkca Art3 Nt5e 6 2 podocyte (1) endothelial (2/4) mesangial (3) 7 PC2 SMC/JG (5) 5 immune (6) PEC (7) 3 TEC

PC1 Adra1a Ren1 Cnn1 G hi F AA H β lo DAPI PDGFR EA calponin 1 α-SMA Adra1b Agtr1a Agtr1b Cygb

AA EA

PKCα ADAMTS5

Figure 1. Unbiased clustering of scRNA-seq data reveals the major cell types of the glomerulus. (A) T-distributed neighbor embedding (t-SNE) representation of 5287 glomerular cells from a healthy C57BL/6J mouse. Labels indicate clusters identified by unsupervised clustering analysis. (B) Violin plots of marker gene expression in each cluster shown in (A). (C) Violin plots showing expression of glomerular capillary endothelial cell (Ehd3) or arteriolar endothelial cell (Fbln2, Mgp, Trpv4, Bmx)markergenesinclusters2and4(D). PCA plot of normal glomerular cells. Cluster labels are identical to (A). Proximity of clusters 3 and 5 indicate high degree of similarity between the two clusters. (E) Violin plots showing expression levels of genes specific to mesangial cells or SMCs/JG cells. The numbers in parentheses indicate cluster number. (F) Immunofluorescence staining of kidney sections shows mesangial-specific expression of the identified marker genes. PDGFRb, which stains mesangial cells in the glomerulus (dotted circle) and stromal cells outside the glo- merulus, was used as reference. Scale bars, 20 mm. (G) Expression levels of the indicated genes are shown in a t-SNE plot of SMCs/JG cells. The two subclusters corresponding to AA SMCs and EA SMCs are outlined. (H) Immunofluorescence staining of kidney sections shows specific staining of calponin 1 in the vascular SMCs of the AA. Scale bar, 10 mm. Normalized expression levels are shown in the violin plots. DAPI, 49,6-diamidino-2-phenylindole; a-SMA, a-smooth muscle actin.

Sections were mounted on Superfrost Plus glass slides (Ri- fibrosis; 1, one to three glomeruli or tubules with increased sur- chard-Allan) and stained with Periodic acid–Schiff (PAS) or rounding fibrosis and/or sclerosis; 2, four to ten glomeruli or Masson trichrome using an automated stainer (DAKO) ac- tubules with increased surrounding fibrosis and/or sclerosis; 3, cording to the manufacturer’s instructions. Severity of glo- 11–20 glomeruli or tubules with increased surrounding fibrosis merular injury was visually scored in a blinded fashion on and/or sclerosis; and 4, .20 glomeruli or tubules with increased PAS-stained slides using a subjective, semiquantitative, five- surrounding fibrosis and/or sclerosis. PAS- and trichrome- point scale as follows: 0, within normal limits; 1, mild, segmen- stained slides were imaged with a Nanozoomer 2.0-HT auto- tal mesangial expansion with or without increased cellularity; mated slide scanning platform (Hamamatsu) at 2003 final 2, moderate, segmental mesangial expansion frequently asso- magnification. ciated with increased cellularity; 3, global membranoprolifer- ative GN; and 4, glomerulosclerosis. The average score for Immunoblotting each mouse was calculated from 20 glomeruli. Severity of Kidney or glomeruli isolated from mice were lysed in ice-cold fibrosis was visually scored in a blinded fashion on trichrome- TNET lysis buffer (50 mM Tris-hydrochloride [pH 7.4], stained slides using a subjective, semiquantitative, five-point 150 mM sodium chloride, 1 mM EDTA, 1% Triton X-100, scale as follows: 0, no appreciable increase in periglomerular cOmplete Protease Inhibitor Cocktail; Roche Applied Science)

JASN 31: ccc–ccc, 2020 scRNA-Seq Analysis of Mouse Glomeruli 5 BASIC RESEARCH www.jasn.org for 15 minutes on ice. The content of the lysates was Ccl2 forward, CAAGAAGGAATGGGTCCAGA; Ccl2 reverse, quantified with the Pierce BCA Protein Assay Kit (Thermo GCTGAAGACCTTAGGGCAGA; Ppia forward, AGCATA Fisher Scientific). Equivalent amounts of each sample were CAGGTCCTGGCATC; and Ppia reverse, CCATCCAGCCAT separated by SDS-PAGE, transferred to nitrocellulose mem- TCAGTCTT. branes, and immunoblotted for YAP/TAZ, b-actin, or vincu- To measure CCL2 secretion from peritoneal macrophages, lin. The blots were visualized by infrared imaging on a LI-COR cells were stimulated with varying doses of LPS Odyssey system (LI-COR). (0.1–100 ng/ml) for 24 hours. The amount of CCL2 in the supernatant was measured using a mouse CCL2 ELISA Real-Time Quantitative PCR Analysis according to the manufacturer’s instructions (R&D systems). RNA was isolated using TRIzol Reagent (Invitrogen), and For in vivo validation of the CCL2-YPet reporter mice, the cDNA was generated using the High-Capacity cDNA Reverse animals were either maintained on a normal chow diet (Pico- Transcription Kit (Applied Biosystems). Real-time quantita- lab rodent diet 20; LabDiet) or a high-fat diet (60 kcal% fat, tive PCR (qPCR) was performed on the ABI 7500 (Applied D12492; Research Diets) for 10 weeks. For two-photon imag- Biosystems) using the TaqMan Universal PCR Master Mix ing of the epididymal fat pads, animals were injected intrave- (Applied Biosystems) and the following TaqMan Gene Expres- nously with QTRACKER Qdot 655 (Thermo Fisher Scientific) sion Assays (Thermo Fisher Scientific): Angptl2 10 minutes before euthanasia to label the vasculature. The fat (Mm00507897_m1), Cebpb (Mm00843434_s1), Co- pads were removed from euthanized mice, rinsed in PBS, and l15a1(Mm00456551_m1), Csrp1 (Mm00456002_g1), Ctgf mounted for imaging. Images were collected using a custom- (Mm01192933_g1), Ddn (Mm02020418_s1), Ednrb ized Leica SP8 Two-Photon Microscope (Leica Microsystems) (Mm00432989_m1), H2-q7 (Mm00843895_s1), Lgals1 equipped with a 253 and 0.95 numerical aperture water (Mm00839408_g1), Lifr (Mm00442942_m1), Mafb immersion objective and a Mai Tai HP DeepSee Laser (Spec- (Mm00627481_s1), Mest (Mm00485003_m1), Nphs1 tra-Physics). Induction of YPet expression in fat cells and (Mm01176615_g1), Rcan2 (Mm00472671_m1), Rpl13a infiltrating macrophages in the mice fed a high-fat diet con- (Mm05910660_g1), Sema3g (Mm01219781_m1), S100a6 firmed the CCL2 reporter mouse is a valid tool to detect (Mm00771682_g1), Thbd (Mm00437014_s1), Tpm1 CCL2 expression in vivo.11 (Mm00445895_g1), and Vegfa (Mm00437306_m1).

Validation of CCL2-YPet Reporter Mice RESULTS For in vitro validation of the CCL2-YPet reporter mice, YPet expression was measured in isolated peritoneal macrophages scRNA-Seq Analysis of Healthy Glomeruli after stimulation with LPS and polyinosinic:polycytidylic acid Because glomerular cells constitute only a minute fraction (poly[I:C]). First, mice were injected with 1 ml of 3% Brewer (,1%) of the kidney, we first isolated glomeruli from thioglyocollate (Sigma-Aldrich) medium into the peritoneal C57BL/6J mice using a magnetic beads–based method8 that cavity. Peritoneal macrophages were isolated from the mice was optimized to consistently yield approximately 0.53106 4 days after injection by flushing the peritoneal cavity with ice- cells per mouse with .90% viability. We began by sequencing cold PBS with 10% FBS and cultured overnight in DMEM with .6000 glomerular cells from a healthy adult male mouse. We 10% FBS before stimulation. For FACS analysis of YPet ex- analyzed 5488 cells that remained after quality control and pression, cells were treated with 100 ng/ml LPS (Sigma-Al- removal of doublets. drich) and 50 mg/ml poly(I:C) (InvivoGen) for 24 hours. Unsupervised clustering analysis after removing For FACS analysis of intracellular CCL2 and TNF-a, cells dissociation-induced genes10 identified seven distinct clusters were pretreated with 2 mM monensin (BioLegend) and of cells (Figure 1A, Supplemental Figure 2, A and B, 5 mg/ml brefeldin A (BioLegend) for 30 minutes, then treated Supplemental Table 2). The cell types of each cluster were with 200 ng/ml LPS and 50 mg/ml poly(I:C) for 3 hours. Cells determined using expression of known marker genes were harvested, stained for intracellular CCL2 and TNF-a (Figure 1B, Supplemental Table 1). Podocytes, endothelial using the Cytofix/Cytoperm kit (BD biosciences), and ana- cells, and mesangial cells were present in similar proportions lyzed by flow cytometry using the LSRII or LSRFortessa (BD and together comprised .90% of the cells (Supplemental biosciences). Table 3), demonstrating the purity of the isolated glomeruli For qPCR analysis, peritoneal macrophages treated with and the efficiency of our tissue-disassociation method.12 The 100 ng/ml LPS and 50 mg/ml poly(I:C) for 24 hours were remaining cells corresponded to vascular SMCs/JG cells (clus- sorted into YPet-positive and YPet-negative cells using the ter 5), immune cells (cluster 6; Supplemental Figure 2C), and FACSAria II (BD biosciences). RNA was isolated from the PECs (cluster 7) (Figure 1B). Small numbers of contaminating sorted cells and used to generate cDNA. qPCR was performed TECs were excluded from further analysis. The top 50 genes using the SYBR Green PCR Master Mix (Applied Biosystems) enriched in each cell type are listed in Supplemental Table 4. and the following primers: YPet forward, ACGACGGCAACT Podocytes and mesangial cells were each grouped into a ACAAGACC; YPet reverse, GTCCTCCTTGAAGTCGATGC; single cluster, indicating they are relatively homogenous in

6 JASN JASN 31: ccc–ccc,2020 www.jasn.org BASIC RESEARCH homeostatic conditions. Despite their close similarity to podo- for renin, suggesting these cells represent SMCs from the AAs cytes, we were able to identify genes highly specific for PECs, (Figure 1G).25 The other cluster, which expressed Adra1b which have been implicated as potential glomerular stem cells (a-adrenergic 1B), had higher expression of angioten- (Supplemental Figure 2D).13 Analysis of endothelial cell sub- sin receptors (Agtr1a, Agtr1b), suggesting these cells are EA clusters showed high levels of Ehd3 in cluster 2, whereas Fbln2, SMCs. We identified Cnn1 (calponin 1) and Cygb (cytoglobin) Mgp, Trpv4, and Bmx were preferentially expressed in cluster 4 as novel markers for AA and EA SMCs, respectively. Immuno- (Figure 1C). Based on previous studies, we concluded that fluorescence staining confirmed specific expression of calponin cluster 2 represents capillary endothelial cells and that cluster 1 in the AA SMCs (Figure 1H). The identification of markers to 4 represents arteriolar endothelial cells from the afferent arte- distinguish between AA and EA SMCs is significant because riole (AA) and efferent (EA). We also identified a differential targeting of these cells to control glomerular novel subpopulation of endothelial cells that highly express flow is the basis of kidney protective antihypertensive drugs.26 the Notch ligand Dlk1 and the endothelin receptor (Ednrb) (Supplemental Figure 2E). The identity of these cells is Mapping of Disease Susceptibility Genes unclear, but DLK1 is known to inhibit angiogenesis.17 Studies over the last couple of decades have led to the identi- fication of susceptibility genes for FSGS, a glomerular disease Identification of Mesangial Cell and SMC Markers that is considered to be a podocyte-specific disease.27 Because Mesangial cells function as the central structural support of our data set contains all of the glomerular cell types at high the glomerulus, interacting with both endothelial cells and resolution, we were interested in mapping expression patterns podocytes.18 They are important producers of extracellular of FSGS susceptibility genes on our data. Surprisingly, we matrix, regulate capillary blood flow, phagocytose extracellu- found that only 20 out of 50 susceptibility genes were exclu- lar debris, and contribute to the homeostasis of the glomeru- sively expressed in podocytes within the glomerulus lus by secreting growth factors. Marker genes commonly used (Supplemental Figure 3A). Of the remaining genes, some to identify mesangial cells, such as Pdgfrb and Gata3, are not were more highly expressed in other cell types, such as me- specific to mesangial cells because these genes are also ex- sangial cells (Supplemental Figure 3B), and some were not pressed by vascular SMCs, as well as stromal cells outside of detected in any glomerular cell type. These findings suggest the glomerulus. a-Smooth muscle actin (Acta2)isalsooften these genes may contribute to FSGS susceptibility through used as a mesangial marker, but it is not highly expressed in novel podocyte-independent pathways. uninjured adult mesangial cells and is more specificto Genome-wide association studies have identified a large SMCs.21 Such lack of a precise genetic signature for mesangial number of potential CKD susceptibility genes which are mainly cells has been an obstacle to understanding the specificfunc- thought to be expressed in TECs.28 We confirmed that many of tion of mesangial cells in the kidney. the genes are indeed highly expressed in TECs, but we also noted Clusters 3 and 5 both expressed Gata3 and Pdgfrb (Figure 1, that some of these genes were expressed in glomerular cells A and B) and were closely related transcriptionally, as indi- (Supplemental Figure 3C). For example, Dach1 and Veg fa were cated by their proximity in the PCA plot (Figure 1D). Expres- relatively specific to podocytes, whereas Adamts5 and Pik3r1 had sion of Acta2, Tagln, Myh11, and JG cell markers Ren1 (renin), specific expression in mesangial cells (Supplemental Figure 3D). Akr1b7,andRgs5 revealed that cluster 5 includes the vascular Wenoted specific expression of some CKD genes in SMCs (Jph2, SMCs and JG cells22 (Figure 1E, Supplemental Table 5). Clus- Prkag2) (Supplemental Figure 3D). The implication that ter 3, therefore, represents bona fide mesangial cells. SMCs may play a role in CKD is intriguing. To define a precise signature for mesangial cells, we first searched for genes that were highly expressed in cluster 3 com- Glomerular Response to Immune-Mediated Injury pared with all other cells (Supplemental Table 6). For a subset Glomerular deposition of circulating immune complexes or of these genes, we confirmed mesangial expression at the pro- local formation of -antibody complexes is the cause of tein level using immunofluorescence staining (Figure 1F, various forms of GN.29 We induced GN in mice using neph- Supplemental Figure 2F). Some of the identified genes— rotoxic serum, a surrogate mouse model for the type of glo- such as Serpine2, Fhl2, Des (desmin), and Pdgfrb—were also merular injury that occurs in or Goodpasture expressed at low levels in SMCs/JG cells (Supplemental disease (Supplemental Figure 4A).30 Injection of the nephro- Figure 2G), so we identified genes that were exclusively ex- toxic serum caused an acute inflammatory response and a pressed in each of the two cell types. We found specific expres- transient proteinuria that peaked 1 day after injection and sion of Plvap, Prkca, Art3,andNt5e in mesangial cells; whereas gradually decreased to near baseline levels after 7 days SMCs/JG cells specifically expressed Acta2, Myh11, Rergl, (Supplemental Figure 4B). After 4–6weeks,significant scar- Map3k7cl, Ren1,andAkr1b7 (Figure 1E). Our results suggest ring of glomeruli (glomerulosclerosis) and fibrosis of periglo- SMCs have been misidentified as mesangial cells in previously merular regions was evident, indicating a continued injury published scRNA-seq studies.23,24 response (Supplemental Figure 4C). Analysis of SMCs revealed two subpopulations. One cluster We analyzed purified glomeruli on day 1 (peak proteinuria) expressed Adra1a (a-adrenergic receptor 1A) and was enriched and day 5 (proteinuria largely resolved) after administration of

JASN 31: ccc–ccc, 2020 scRNA-Seq Analysis of Mouse Glomeruli 7 BASIC RESEARCH www.jasn.org nephrotoxic serum (Supplemental Figure 5, A–F). Podocytes, (Figure 2B). Mesangial cells also showed changes in cytoskel- mesangial cells, and endothelial cells from nephritic mice at etal regulation and cell adhesion, in addition to regulation of day 1 each formed clusters that were clearly distinct from their BP and immune responses (Supplemental Figure 5I). control counterparts, indicating a major shift in gene expres- On day 5, when proteinuria was nearly resolved, heatmaps sion for these three major cell types (Figure 2A). In contrast, of DE genes and PCA revealed the changes observed in podo- PECs and SMCs from control and nephritic mice formed cytes at day 1 had largely normalized with a significant reduc- overlapping clusters, suggesting the changes that were seen tion in DE genes (Figure 2, C and D, Supplemental Table 7). In are not a batch artifact. This conclusion was supported by contrast, the transcriptional profiles of mesangial and endo- using a method of random downsampling of cells (see thelial cells continued to change at day 5 to a state further Supplemental Figure 5G), as well as by qPCR validation of distant from the control cells with an increase in the number DE genes (see below). of DE genes (Figure 2, C and D, Supplemental Table 7). We We identified 263, 85, and 165 upregulated genes and 312, also identified subclusters of dividing mesangial and endothe- 56, and 189 downregulated genes in nephritic podocytes, me- lial cells (Supplemental Figure 5J). A subcluster of endothelial sangial cells, endothelial cells, respectively (Supplemental cells emerged on day 5 that had high expression of proangio- Table 7). Changes in expression levels were confirmed by genic factors (Apln, Pgf) and the plasminogen activator inhib- qPCR for a subset of these genes (Supplemental Figure 5H). itor 1 gene (Serpine1) (Supplemental Figure 5K). Pathway analysis showed that podocytes induced programs in There were increased immune cells after injury with in- 1 cytoskeletal regulation, cell adhesion, and inflammatory re- creased numbers of macrophages (F4/80 ) and dendritic cells 1 1 sponse (Supplemental Figure 5I). As expected, pathways re- (DC-SIGN or XCR1 ) on day 1 (Figure 2E, Supplemental lated to podocyte differentiation were downregulated, because Figure 5L). The number of myeloid cells was further increased several podocyte-specific genes were decreased after injury on day 5, demonstrating that inflammation was persisting

A control C podocytes mesangial endothelial nephritis d1 control nephritis d1 podocytes nephritis d5

PEC 4 SMC z-score -4 top 100 variable genes

podocytes mesangial endothelial

t-SNE2 D control nephritis d1 nephritis d5 endothelial PC2

immune mesangial PC1 t-SNE1

E immune cells B control macrophages control nephritis d1 cDC1 nephritis d1 cDC2 Nphs1 Nphs2 Synpo Wt1 Cdkn1c Mafb nephritis d5 B cells t-SNE2 T cells NK cells t-SNE1

Figure 2. scRNA-seq analysis of nephritic mice reveal cell type–specific responses to injury. (A) T-distributed neighbor embedding (t-SNE) plot of control and nephritic mice at day 1. (B) Violin plots show the downregulation of podocyte-specific genes after neph- rotoxic injury. Normalized expression levels are shown. (C) Heatmaps showing the top 100 most variable genes in each cell type. Each column represents a cell, each row represents a gene. (D) PCA plot of podocytes, mesangial cells, and endothelial cells from control and nephritic mice. (E) t-SNE plots of immune cells show increased immune infiltration after nephrotoxic injury. Cell types shown in the right panel were determined by expression of marker genes shown in Supplemental Figure 5L.

8 JASN JASN 31: ccc–ccc,2020 www.jasn.org BASIC RESEARCH even in the absence of proteinuria. Ccl2, a chemokine respon- and CD2AP-deficient mice2 to model glomerular injuries caused sible for monocyte recruitment, was detected in mesangial by metabolic disorder, drug toxicity, and podocyte-specificge- cells and immune cells at day 1 and was expressed by more netic disease, respectively. cells on day 5 (Supplemental Figure 6A). Although CCL2 has -deficient BTBR ob/ob mice, a model of type 2 di- widely been implicated in glomerular diseases, the source has abetes, become hyperglycemic around 6–8 weeks of age, ex- largely been assumed to be from immune cells. We confirmed hibited renal injury with progressive proteinuria at around that CCL2 is produced by mesangial cells by generating a 10 weeks of age, and developed glomerular lesions that closely CCL2 reporter mouse (Supplemental Figure 6, B–E). After resemble those observed in human diabetic by antibody injury, CCL2 reporter expression increased progres- 20–22 weeks of age.41 We therefore analyzed glomerular cells sively over several days (Supplemental Figure 6, F and G). Flow at 12 and 21 weeks of age (Supplemental Figure 8A). Com- cytometry showed .80% of the CCL2-expressing cells were parison of cells from diabetic (ob/ob) and control (ob/1)mice mesangial cells (Supplemental Figure 6H). This suggests in- by PCA showed changes in podocytes and mesangial cells, but jured mesangial cells are an important driver of inflammation. did not show major changes in endothelial cell gene expres- sion, a surprising finding given that diabetes is considered to Hippo Pathway Is Involved in the Podocyte Response be a vascular disease (Figure 4A).43 In addition, the gene ex- to Injury pression profiles of podocytes and mesangial cells did not To validate the podocyte injury expression signature, we fo- change significantly between 12 weeks and 21 weeks of age, cused our analysis on identifying pathways in podocytes in- suggesting the changes induced by diabetes are chronic duced after injury. Pathway analysis indicated that several (Figure 4A). Pathway analysis showed changes to glucose signaling pathways, including Hippo, TGF-b,NF-kB, and and lipid metabolism pathways in both podocytes and mesan- FoxO pathways were upregulated in injured podocytes gial cells (Supplemental Figure 8B). Cell proliferation path- (Supplemental Figure 5I). Because the Hippo pathway is acti- ways were induced in mesangial cells, and apoptotic pathways vated by disruption of cell junctions and because podocyte were induced in podocytes. Consistent with this, the number injury involves significant junctional reorganization (foot pro- of mesangial cells were increased (41.5% of total cells com- cess effacement), we focused on the role of the Hippo signaling pared with 11.9% in control mice) whereas podocytes were pathway.34,35 decreased (4.4% compared with 6.5% in control mice) at 21 The Hippo pathway is mediated by the coactivators YAP weeks. Consistent with the expanded matrix that is a hallmark and TAZ which, when activated, are stabilized and facilitate of diabetic kidney disease, there was increased expression of transcription of target genes. We saw significantly increased matrix and matrix-modifying in both podocytes and protein levels of YAP and TAZ in glomerular lysates from ne- mesangial cells (Figure 4, B and C). phritic mice (Figure 3A), and known targets of YAP and TAZ Doxorubicin is a model of glomerular injury due to its toxicity —including Ctgf, Cyr61,andAxl—were among the highest to podocytes and glomerular endothelial cells.42,44 In mice, doxo- induced genes in podocytes after injury (Figure 3B). rubicin induces proteinuria that begins after 7–14 days and pro- We generated mice that allowed for conditional deletion of gressive kidney deterioration that occurs after several weeks. 2 2 2 2 either TAZ (Wwtr1 / )orYAP(Yap1 / ) (Supplemental Analysis of glomeruli 14 days after doxorubicin injection revealed Figure 7A). In contrast to a previous study that deleted YAP reduced podocyte numbers (Supplemental Figure 8C) and during development,40 we found that deletion of YAP or TAZ downregulation of podocyte marker genes (Supplemental in adult mice did not, by itself, cause any renal dysfunction Figure 8D). signaling was increased in podocytes, demon- (Supplemental Figure 7B). After nephrotoxic serum injection, strated by increased Cdkn1a and Gadd45g, which was likely sec- 2 2 2 2 however, Wwtr1 / and Yap1 / mice displayed significantly ondary to reactive oxygen species–mediated DNA damage higher levels of proteinuria at day 1 compared with wild-type (Supplemental Figure 8, E and F).45 In support of elevated oxi- mice and they also exhibited delayed resolution (Figure 3C). dative stress, the metallothionein genes Mt-1 and Mt-2 were Kidney sections examined at 5 weeks after nephrotoxic serum among the highest upregulated genes in both mesangial and en- injection showed increased injury for both knockout mice dothelial cells (Supplemental Figure 8, G and H). Distinct to this compared with wild-type controls (Figure 3D). These effects model was a significant increase in the number of neutrophils in 2 2 were less severe in the Wwtr1 / mice compared with the glomerulus after injury (Figure 4D, Supplemental Figure 8I). 2 2 Yap1 / mice, which may be due to incomplete deletion of Expression analysis showed that mesangial production of Cxcl1,a TAZ (Supplemental Figure 7A). This suggests Hippo pathway key neutrophil chemokine, is the likely mechanism (Figure 4E). signaling is important for podocyte recovery after immune Lastly, we analyzed mice lacking CD2AP, which is a scaf- injury and validates the use of our database in identifying folding protein expressed in podocytes and associated with important reparative pathways. sporadic nephrotic syndrome and FSGS in humans.2 CD2AP-deficient mice develop proteinuria around 2–3 weeks Injury Response in Different Glomerular Injury Models of age and die around 6–7 weeks of age due to renal failure. We Weperformed scRNA-seq on several other models of glomerular purified glomeruli from 3-week-old animals, soon after the on- injury. We used BTBR ob/ob mice,41 doxorubicin treatment,42 set of proteinuria. The number of podocytes was significantly

JASN 31: ccc–ccc, 2020 scRNA-Seq Analysis of Mouse Glomeruli 9 BASIC RESEARCH www.jasn.org

A control nephritis B TAZ Ctgf Cyr61 Axl Ankrd1 Tagln Crim1 F3 Gadd45b control YAP nephritis d1

Vinculin

C D Glomerular Score Fibrosis Score ****** ** * 80 5 **** 10000 ** * ** ** WT WT 4 -/- 1000 * 60 Wwtr1 -/- Wwtr1 3 Yap1-/- 100 40 Yap1-/- 2 10 20 1 1 urinary ACR 0 0 0.1 -20 -1 0.01 01235 days post-injection saline saline saline saline saline saline

nephrotoxicserum nephrotoxicserum nephrotoxicserum nephrotoxicserum nephrotoxicserum nephrotoxicserum

Figure 3. The Hippo pathway is required for podocyte response to injury. (A) Immunoblots showing increased protein levels of TAZ and YAP in glomerular lysates 2 days after nephrotoxic injury. Blots are representative of two independent experiments. Vinculin was used as loading control. (B) YAP/TAZ target genes are induced in podocytes after nephrotoxic injury. Normalized expression levels are 2 2 2 2 shown as violin plots. (C) Deletion of TAZ (Wwtr1 / )orYAP(Yap1 / ) exacerbates proteinuria. Albumin-creatinine ratios (ACR) were measured from spot urine collected from mice at the indicated time points after nephrotoxic serum injection. (D) Kidney sections were stained with PAS and Masson trichrome stain 5 weeks after injection with nephrotoxic serum or saline and scored blindly for GN and fibrosis. Scoring criteria are described in the methods. Results were analyzed by multiple t tests using the two-stage linear step-up method (*q,0.05, **q,0.01, ****q,0.0001). WT, wild type. reduced in the CD2AP-deficient mice compared with age- For this reason, we used scRNA-seq to characterize the cells matched, wild-type mice (31.5% of total in wild type, 10.7% of the glomerulus and monitor the changes that occur after of total in knockout) (Supplemental Figure 8J). This was asso- several different types of injury. The quality of the data gener- ciated with upregulation of apoptotic pathways in podocytes ated by single-cell sequencing is greatly affected by the condi- (Supplemental Figure 8K). Consistent with expanded mesangial tion of the prepared cells. Our optimized methods allowed the matrix seen in CD2AP-deficient kidneys, podocytes and mesan- preparation of .0.53106 cells from the glomerulus of a single gial cells both showed increased expression of matrix proteins mouse with approximately 95% viability in ,3 hours. The low (Supplemental Figure 8, L and M). In contrast, there was very mitochondrial read content; high gene-detection sensitivity; little change detected in endothelial cells or immune cells and the balanced sampling of podocytes, mesangial cells, and (Figure 4F). endothelial cells all attest to the quality of our cell preparations (Supplemental Table 8). Using these methods, we were able to generate a comprehensive and detailed single-cell data set fo- DISCUSSION cused on the glomerulus that allowed for in-depth analysis. The distinct clustering of the AA and EA SMCs, which are a minor CKD is a major health care problem affecting approximately population of cells, and the subsequent identification of their 15% of the population, but treatment options are limited be- respective markers, affirms the high resolution of the data set. yond dialysis and kidney transplantation.1 With the recogni- We are not the first to use scRNA-seq to examine the glo- tion that many kidney diseases have a hereditary component merulus. Karaiskos et al.23 sequenced approximately 13,000 andwiththeidentification of susceptibility genes for kidney glomerular cells, but required 32 mice to generate this number diseases, such as FSGS, IgA nephropathy, and CKD, there of cells. The cell preparation in this study is dominated by has been renewed interest in the molecular basis of kidney podocytes, suggesting issues with tissue dissociation. Compar- diseases. A more precise understanding of glomerular cell in- ison with our data suggest that the population of cells that were jury at the molecular level has the potential to revolutionize identified as mesangial cells are instead mostly SMCs/JG cells diagnosis and lead to new therapies. as evidenced by the expression of Acta2, Myh11,andAkr1b7.

10 JASN JASN 31: ccc–ccc,2020 www.jasn.org BASIC RESEARCH

A podocyte mesangial endothelial B Col4a5 Col4a2 Col4a1 Lama5 Loxl2 ob/+ ob/ob wk12 ob/ob wk12 PC2

C Col6a3 Lamb1 Col4a2 Thbs1 Timp3 PC1

D E

control Cxcl1 neutrophils doxorubicin G lymphocytes Nephrtis Diabetes Doxorubicin monocyte/ 15 2.0 30 macrophages 1.5 10 20 1.0 F podocyte mesangial endothelial 5 10 +/+ 0.5 Ren1-positive

Cd2ap mesangial cell

-/- % Cd2ap 0 0.0 0 d0 d1 d5

PC2 ob/+ control

ob/obob/ob wk12 wk21 doxorubucin PC1

Figure 4. Glomerular cells display distinct responses to different types of injury. (A) PCA plots of podocytes, mesangial cells, and endothelial cells from control and diabetic mice. (B and C) Violin plots show increased expression of extracellular matrix and matrix- modifying genes in (B) podocytes and (C) mesangial cells from diabetic mice. (D) T-distributed neighbor embedding (t-SNE) plot of immune cells shows increased numbers of neutrophils and monocytes/macrophages in doxorubicin-treated mice. Immune cell types were determined by expression of marker genes shown in Supplemental Figure 8I. (E) Violin plot shows increased expression of Cxcl1 in mesangial cells from doxorubicin-treated mice. (F) PCA plots of podocytes, mesangial cells, and endothelial cells from wild-type and 2 2 Cd2ap / mice. (G) Graphs showing the fraction of mesangial cells expressing Ren1 after injury caused by nephrotoxic serum, di- abetes, and doxorubicin.

In another study that analyzed glomerular cells from a model which use mitochondrial read content (indicative of damaged of type 1 diabetes (streptozotocin-injected eNOS knockout or dead cells) as a metric, are complicated by the high abun- mice), Fu et al.24 sequenced a total of 644 cells from control dance of mitochondrial transcripts in the renal parenchyma and diabetic mice. This study was limited by the small number when analyzing the whole kidney. This is not a problem when of cells, the authors were unable to distinguish mesangial cells using purified glomerular cells. A table comparing our data from SMCs/JG cells and they were unable to obtain sufficient with previous studies is shown in Supplemental Table 8. numbers of podocytes from diabetic mice for analysis. Ouranalysissuggeststhatdifficulties in identifying mesan- Other studies have identified glomerular cells using single- gial cell–specific genes is likely due to their close similarity cell approaches of the whole kidney. In general, these ap- with SMCs. For example, a recent study defined a mesangial proaches identified relatively few glomerular cells, partly due gene signature using bulk analysis of cells isolated based on to their paucity, but also because tissue-dissociation methods expression of Meis1, a developmental gene of mesangial optimized for the whole kidney are not ideal for glomeruli. cells.49,50 We found that several genes identified in this study Park et al.28 only identified 78 podocytes from 43,745 total are not specific to mesangial cells, and in some cases are spe- cells, whereas Ransick et al.46 found 24 podocytes from 31,265 cifictoSMCs(Myh11, Rergl, Pln,andOlfr558) (Supplemental total cells. It is also difficult to clearly identify mesangial cells Figure 9A). Potential explanations for these discrepancies in these data sets because of the similarity between mesangial could be the expression of Meis1 in cells other than mesangial cells and stromal mesenchymal cells. Wu et al.47 and Lake cells in adult glomeruli (Supplemental Figure 9B) or impuri- et al.48, using single-nucleus RNA-seq of whole mouse and ties in the bulk cell preparations. Our gene signatures suggest human kidneys, identified severalfold more glomerular cells, that several previous single-cell studies have either misidenti- but the gene detection sensitivity was three- to fivefold lower fied SMCs for mesangial cells or conflated these two cell types than in our data set. In addition, the quality control steps, together.23,24 The bona fide mesangial cell markers identified

JASN 31: ccc–ccc, 2020 scRNA-Seq Analysis of Mouse Glomeruli 11 BASIC RESEARCH www.jasn.org in this study should provide more precise tools to interrogate simultaneous monitoring of their responses to injury. the function of mesangial cells in vivo. Whereas we were able to identify and validate the Hippo path- Based on their location adjacent to endothelial cells, mesan- way as critical for podocyte repair, other pathways like the gial cells are considered a specialized .18 They function in TGF-b, Wnt, and Notch pathways are implicated by our stabilizing the vasculature, synthesizing components of the base- data and likely to play important roles. By generating compre- ment membrane, and participating in controlling capillary vas- hensive snapshots of the altered genetic landscapes in multiple cular tone. However, focus over the last 10 years in glomerular injury models, this work provides data to test the link between disease pathogenesis has shifted to the podocyte because mesan- injury type, glomerular cell response, inflammation, and his- gial cells are considered to play a secondary and bystander role.18 tology. Although the exact mechanisms of this process require With the recognition that stromal cells, including fibroblasts and further investigation, our results provide insight into the un- , play critical roles in promoting inflammation in tu- derlying pathophysiologic pathways and potential novel ther- mors51 as well as in and fibrosis,52 by analogy, it is likely apeutic approaches for glomerular diseases. that the mesangial cell is playing a similar role. The four injury models we analyzed displayed a wide spec- trum of outcomes with varying rates and patterns of disease DATA SHARING STATEMENT progression and cell types affected, with little overlap in the transcriptional responses to each type of injury (Supplemental The scRNA-seq data generated in this study have been depos- Figure 10, A and B). However, in all of the injury models, me- ited in the National Center for Biotechnology Information sangial cells showed persistent induction of genes involved in Gene Expression Omnibus (GSE146912) and can be ac- wound healing (Supplemental Table 9) with distinct expression cessed at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi? patterns of matrix (Col4a1, Col4a3, Col6a3, Col8a1) and chemo- acc5GSE146912 or the Broad Institute Single Cell Portal kine genes (Ccl2, Cxcl1, Cxcl13, Cx3cl1) in each injury model. We at https://singlecell.broadinstitute.org/single_cell/study/ also noted induction of Ren1 expression in a subset of mesangial SCP975. Mice generated for this study are available from cells in all models except CD2AP deficiency (Figure 4G). In di- the corresponding author on reasonable request. All other abetic mice, the number of Ren1-expressing mesangial cells unique reagents used in this study are available from the increased twofold from week 12 to week 21 and, in corresponding author on reasonable request. doxorubicin-treated mice, .25% of the mesangial cells ex- pressed Ren1. Collectively, these results suggest mesangial cells may shape the character of the inflammation and wound- healing programs in response to distinct types of injuries. It ACKNOWLEDGMENTS also suggests that a persistent mesangial reaction may drive the chronic decline of kidney function in many diseases. We thank our Genentech colleagues in the Research Biology, Mo- fi We noted some novel ndings in each of the injury models. lecular Biology, Laboratory Animal Resources, and Pathology De- The lack of an endothelial reaction in the ob/ob mice was un- partments for their support of this study. We thank Emil Unanue, expected and suggests that diabetes does not by itself induce Susan Gurley, and Ariel Gomez for helpful discussions, as well as transcriptional changes in endothelial cells. Given the preva- Yulei Wang, Melanie Desbois, and Milena Duerrbaum for help with lence of hypertension in patients with diabetes, hypertension myofibroblast analysis. 53 might be important in injuring endothelial cells. The fact Dr. Ying-Jiun J. Chen and Dr. Zora Modrusan prepared cDNA that BTBR ob/ob mice do not develop hypertension could ex- libraries and performed next-generation sequencing; Dr. Jun-Jae plain why studies using a different model of diabetic nephrop- Chung prepared glomerular cells with assistance from Dr. Jiyeon Lee athy, which does become hypertensive, did show transcriptional 2 2 and performed in vivo and in vitro experiments; Dr. Jun-Jae Chung 24 / changes to endothelial cells. The Cd2ap mice differed from and Dr. Andrey Shaw conceptualized the project, designed experi- the other three models by the lack of Ren1 expression in injured ments, analyzed and interpreted data, and wrote the manuscript with mesangial cells. We speculate this may be due to the young age input from all authors; Dr. Anwesha Dey and Dr. Meron Roose- of the mice. This is an intriguing possibility to explore because Girma generated the Wwtr1.cko and Yap1.cko mice; Dr. Leonard children that have nephrotic syndrome often outgrow it by Goldstein contributed to processing and analysis of the scRNA-seq fi their teen years without brotic sequelae. It should be noted data; Dr. Sharad C. Paudyal generated the CCL2-YPet reporter mice; that there may be differences in tissue dissociation between and Dr. Joshua D. Webster performed histologic analyses. different mouse strains and, therefore, cross-strain compari- sons in cell numbers between C57BL/6J and BTBR mice are not possible. DISCLOSURES Overall, our data illustrate the power of using scRNA-seq to monitor the complex tissue injury process during disease. Fo- Y. Chen, J. Chung, A. Dey, L. Goldstein, J. Lee, Z. Modrusan, M. Roose- cusing on the glomerulus allowed detailed characterization of Girma, A. Shaw, and J. Webster report employment at Genentech. All remain- multiple cell types that exist in small numbers and ing authors have nothing to disclose.

12 JASN JASN 31: ccc–ccc,2020 www.jasn.org BASIC RESEARCH

FUNDING 4. Smithies O, Gregg RG, Boggs SS, Koralewski MA, Kucherlapati RS: Insertion of DNA sequences into the human chromosomal b-globin locus by homologous recombination. Nature 317: 230–234, 1985 This work was supported by Genentech. 5. Thomas KR, Folger KR, Capecchi MR: High frequency targeting of genes to specific sites in the mammalian genome. Cell 44: 419–428, 1986 6. Newman RJ, Roose-Girma M, Warming S: Efficient conditional knock- SUPPLEMENTAL MATERIAL out targeting vector construction using co-selection BAC recombin- eering (CoSBR). Nucleic Acids Res 43: e124, 2015 This article contains the following supplemental material online at 7. Gertsenstein M, Nutter LMJ, Reid T, Pereira M, Stanford WL, Rossant J, http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2020020220/-/ et al.: Efficient generation of germ line transmitting chimeras from DCSupplemental. C57BL/6N ES cells by aggregation with outbred host embryos. PLoS One 5: e11260, 2010 Supplemental Figure 1. Diagram of the glomerular structure. 8. Takemoto M, Asker N, Gerhardt H, Lundkvist A, Johansson BR, Saito Y, Supplemental Figure 2. Unbiased clustering reveals the major cell et al.: A new method for large scale isolation of kidney glomeruli from types of the glomerulus. mice. Am J Pathol 161: 799–805, 2002 Supplemental Figure 3. Mapping of FSGS and CKD susceptibility 9. Wu TD, Nacu S: Fast and SNP-tolerant detection of complex variants – gene expression to glomerular cell types. and splicing in short reads. Bioinformatics 26: 873 881, 2010 10. van den Brink SC, Sage F, Vértesy Á, Spanjaard B, Peterson-Maduro J, Supplemental Figure 4. Nephrotoxic serum nephritis in C57BL/ Baron CS, et al.: Single-cell sequencing reveals dissociation-induced 6J mice. gene expression in tissue subpopulations. Nat Methods 14: 935–936, Supplemental Figure 5. scRNA-Seq analysis of glomerular cells 2017 from nephritic mice. 11. Kanda H, Tateya S, Tamori Y, Kotani K, Hiasa K, Kitazawa R, et al.: MCP- Supplemental Figure 6. Induction of Ccl2 expression in nephro- 1 contributes to macrophage infiltration into adipose tissue, insulin resistance, and hepatic steatosis in obesity. JClinInvest116: toxic serum nephritis. 1494–1505, 2006 Supplemental Figure 7. Conditional deletion of TAZ and YAP. 12. Bertram JF, Soosaipillai MC, Ricardo SD, Ryan GB: Total numbers of Supplemental Figure 8. scRNA-Seq analysis of glomerular cells glomeruli and individual glomerular cell types in the normal rat kidney. from diabetic, doxorubicin-treated, and CD2AP-deficient mice. Cell Tissue Res 270: 37–45, 1992 Supplemental Figure 9. Validation of previously reported me- 13. Miesen L, Steenbergen E, Smeets B: Parietal cells-new perspectives in – sangial cell-specific genes. glomerular disease. Cell Tissue Res 369: 237 244, 2017 14. Vanlandewijck M, He L, Mäe MA, Andrae J, Ando K, Del Gaudio F, Supplemental Figure 10. Glomerular cells show distinct responses et al.: A molecular atlas of cell types and zonation in the vascu- to different types of injury. lature. Nature 554: 475–480, 2018 Supplemental Table 1. Genes differentially expressed between two 15. He L, Vanlandewijck M, Mäe MA, Andrae J, Ando K, Del Gaudio F, podocyte sub-clusters (2 and 4) in Supplemental Figure 2A. et al.: Single-cell RNA sequencing of mouse brain and lung vascular and Supplemental Table 2. List of canonical cell type markers used to vessel-associated cell types. Sci Data 5: 180160, 2018 16. Patrakka J, Xiao Z, Nukui M, Takemoto M, He L, Oddsson A, et al.: determine the identity of cell clusters. Expression and subcellular distribution of novel glomerulus-associated Supplemental Table 3. Composition of glomerular cells from a proteins dendrin, ehd3, sh2d4a, plekhh2, and 2310066E14Rik. JAm C57BL/6J mouse analyzed by scRNA-Seq. Soc Nephrol 18: 689–697, 2007 Supplemental Table 4. Top 50 cell type marker genes. 17. Rodríguez P, Higueras MA, González-Rajal A, Alfranca A, Fierro- Supplemental Table 5. Genes differentially expressed between Fernández M, García-Fernández RA, et al.: The non-canonical NOTCH mesangial cells (cluster 3) and SMCs/JG cells (cluster 5). ligand DLK1 exhibits a novel vascular role as a strong inhibitor of an- giogenesis. Cardiovasc Res 93: 232–241, 2012 Supplemental Table 6. Genes enriched in mesangial cells. 18. Schlöndorff D, Banas B: The mesangial cell revisited: No cell is an is- Supplemental Table 7. Number of differentially expressed genes in land. JAmSocNephrol20: 1179–1187, 2009 nephritic mice. 19. Alpers CE, Seifert RA, Hudkins KL, Johnson RJ, Bowen-Pope DF: PDGF- Supplemental Table 8. Comparison of kidney single cell datasets. receptor localizes to mesangial, parietal epithelial, and interstitial cells – Supplemental Table 9. Genes commonly induced in mesangial in human and primate kidneys. Kidney Int 43: 286 294, 1993 20. Chen Y-T, Chang F-C, Wu C-F, Chou Y-H, Hsu H-L, Chiang W-C, et al.: cells after different injury types. -derived growth factor receptor signaling activates pericyte- myofibroblast transition in obstructive and post-ischemic kidney fi- brosis. Kidney Int 80: 1170–1181, 2011 REFERENCES 21. Uhlén M, Fagerberg L, Hallström BM, Lindskog C, Oksvold P, Mardinoglu A, et al.: Proteomics. Tissue-based map of the human 1. United States Renal Data System: USRDS Annual Data Report: Epide- proteome. Science 347: 1260419, 2015 miology of Kidney Disease in the United States, Bethesda, MD, Na- 22. Brunskill EW, Sequeira-Lopez ML, Pentz ES, Lin E, Yu J, Aronow BJ, tional Institutes of Health, National Institute of Diabetes and Digestive et al.: Genes that confer the identity of the renin cell. J Am Soc Nephrol and Kidney Diseases, 2018 22: 2213–2225, 2011 2. Shih NY, Li J, Karpitskii V, Nguyen A, Dustin ML, Kanagawa O, et al.: 23. Karaiskos N, Rahmatollahi M, Boltengagen A, Liu H, Hoehne M, Congenital nephrotic syndrome in mice lacking CD2-associated pro- Rinschen M, et al.: A single-cell transcriptome atlas of the mouse glo- tein. Science 286: 312–315, 1999 merulus. J Am Soc Nephrol 29: 2060–2068, 2018 3. Kuehn MR, Bradley A, Robertson EJ, Evans MJ: A potential animal 24. Fu J, Akat KM, Sun Z, Zhang W, Schlondorff D, Liu Z, et al.: Single-cell model for Lesch-Nyhan syndrome through introduction of HPRT mu- RNA profiling of glomerular cells shows dynamic changes in experi- tations into mice. Nature 326: 295–298, 1987 mental diabetic kidney disease. J Am Soc Nephrol 30: 533–545, 2019

JASN 31: ccc–ccc, 2020 scRNA-Seq Analysis of Mouse Glomeruli 13 BASIC RESEARCH www.jasn.org

25. Gomez RA, Sequeira-Lopez MLS: Renin cells in homeostasis, re- 40. Schwartzman M, Reginensi A, Wong JS, Basgen JM, Meliambro K, generation and immune defence mechanisms. Nat Rev Nephrol 14: Nicholas SB, et al.: Podocyte-specific deletion of yes-associated pro- 231–245, 2018 tein causes FSGS and progressive renal failure. JAmSocNephrol27: 26. Digne-Malcolm H, Frise MC, Dorrington KL: How do antihypertensive 216–226, 2016 drugs work? Insights from studies of the renal regulation of arterial 41. Hudkins KL, Pichaiwong W, Wietecha T, Kowalewska J, Banas MC, . Front Physiol 7: 320, 2016 Spencer MW, et al.: BTBR Ob/Ob mutant mice model progressive di- 27. Rosenberg AZ, Kopp JB: Focal segmental glomerulosclerosis. Clin abetic nephropathy. J Am Soc Nephrol 21: 1533–1542, 2010 JAmSocNephrol12: 502–517, 2017 42. Lee VWS, Harris DCH: Adriamycin nephropathy: A model of focal 28. Park J, Shrestha R, Qiu C, Kondo A, Huang S, Werth M, et al.: Single-cell segmental glomerulosclerosis. Nephrology (Carlton) 16: 30–38, transcriptomics of the mouse kidney reveals potential cellular targets of 2011 kidney disease. Science 360: 758–763, 2018 43. Barrett EJ, Liu Z, Khamaisi M, King GL, Klein R, Klein BEK, et al.: Di- 29. Jennette JC, D’Agati VD, Olson JL, Silva FG: Heptinstall’s Pathology of abetic microvascular disease: An endocrine society scientific state- the Kidney, Alphen aan den Rijn, The Netherlands, Wolters Kluwer, ment. J Clin Endocrinol Metab 102: 4343–4410, 2017 2014 44. Jeansson M, Björck K, Tenstad O, Haraldsson B: Adriamycin alters 30. Borza D-B, Hudson BG: Of mice and men: Murine models of anti-GBM glomerular to induce proteinuria. JAmSocNephrol20: antibody nephritis. Kidney Int 61: 1905–1906, 2002 114–122, 2009 31. Fujinaka H, Yamamoto T, Takeya M, Feng L, Kawasaki K, Yaoita E, et al.: 45. Barbey MM, Fels LM, Soose M, Poelstra K, Gwinner W, Bakker W, et al.: Suppression of anti-glomerular nephritis by ad- Adriamycin affects glomerular renal function: Evidence for the in- ministration of anti-monocyte chemoattractant protein-1 antibody in volvement of oxygen radicals. Free Radic Res Commun 7: 195–203, WKY rats. J Am Soc Nephrol 8: 1174–1178, 1997 1989 32. Tesch GH, Schwarting A, Kinoshita K, Lan HY, Rollins BJ, Kelley VR: 46. Ransick A, Lindström NO, Liu J, Zhu Q, Guo JJ, Alvarado GF, et al.: Monocyte chemoattractant protein-1 promotes macrophage- Single-cell profiling reveals sex, lineage, and regional diversity in the mediated tubular injury, but not glomerular injury, in nephrotoxic se- mouse kidney. Dev Cell 51: 399–413.e7, 2019 rum nephritis. J Clin Invest 103: 73–80, 1999 47. Wu H, Kirita Y, Donnelly EL, Humphreys BD: Advantages of single- 33. Eardley KS, Zehnder D, Quinkler M, Lepenies J, Bates RL, Savage CO, nucleus over single-cell RNA sequencing of adult kidney: Rare cell et al.: The relationship between albuminuria, MCP-1/CCL2, and in- types and novel cell states revealed in fibrosis. J Am Soc Nephrol 30: terstitial macrophages in chronic kidney disease. Kidney Int 69: 23–32, 2019 1189–1197, 2006 48. Lake BB, Chen S, Hoshi M, Plongthongkum N, Salamon D, Knoten A, 34.MengZ,MoroishiT,GuanK-L:MechanismsofHippopathwayregu- et al.: A single-nucleus RNA-sequencing pipeline to decipher the mo- lation. Genes Dev 30: 1–17, 2016 lecular anatomy and pathophysiology of human kidneys. Nat Commun 35. Halder G, Dupont S, Piccolo S: Transduction of mechanical and cyto- 10: 2832, 2019 skeletal cues by YAP and TAZ. Nat Rev Mol Cell Biol 13: 591–600, 2012 49. Brunskill EW, Potter SS: Changes in the gene expression programs of 36. Ma S, Meng Z, Chen R, Guan K-L: The Hippo pathway: Biology and renal mesangial cells during . BMC Nephrol 13: pathophysiology. Annu Rev Biochem 88: 577–604, 2019 70, 2012 37. Yu F-X, Zhao B, Panupinthu N, Jewell JL, Lian I, Wang LH, et al.: Reg- 50. Brunskill EW, Aronow BJ, Georgas K, Rumballe B, Valerius MT, Aronow ulation of the Hippo-YAP pathway by G-protein-coupled receptor J, et al.: Atlas of gene expression in the developing kidney at micro- signaling. Cell 150: 780–791, 2012 anatomic resolution. [published correction appears in Dev Cell 16: 482, 38. Wang Y, Xu X, Maglic D, Dill MT, Mojumdar K, Ng PK-S, et al.; Cancer 2009]. Dev Cell 15: 781–791, 2008 Genome Atlas Research Network: Comprehensive molecular charac- 51. Turley SJ, Cremasco V, Astarita JL: Immunological hallmarks of stromal terization of the Hippo signaling pathway in cancer. Cell Rep 25: cells in the tumour microenvironment. Nat Rev Immunol 15: 669–682, 1304–1317.e5, 2018 2015 39. Schütte U, Bisht S, Heukamp LC, Kebschull M, Florin A, Haarmann J, 52. O’Dwyer DN, Ashley SL, Moore BB: Influences of innate immunity, et al.: Hippo signaling mediates proliferation, invasiveness, and met- autophagy, and fibroblast activation in the pathogenesis of lung fi- astatic potential of clear cell renal cell carcinoma. Transl Oncol 7: brosis. Am J Physiol Lung Cell Mol Physiol 311: L590 –L601, 2016 309–321, 2014 53. Mogensen CE: Diabetes and hypertension. Lancet 1: 388–389, 1979

AFFILIATIONS

1Department of Research Biology, Genentech, South San Francisco, California 2Department of Molecular Biology, Genentech, South San Francisco, California 3Department of Pathology, Genentech, South San Francisco, California 4Department of Pathology and Immunology, Washington University, St. Louis, Missouri 5Department of Molecular Oncology, Genentech, South San Francisco, California

14 JASN JASN 31: ccc–ccc,2020 Supplemental material Table of Contents

Supplemental Figure 1. Diagram of the glomerular structure

Supplemental Figure 2. Unbiased clustering reveals the major cell types of the glomerulus

Supplemental Figure 3. Mapping of FSGS and CKD susceptibility gene expression to glomerular cell types

Supplemental Figure 4. Nephrotoxic serum nephritis in C57BL/6J mice

Supplemental Figure 5. scRNA-Seq analysis of glomerular cells from nephritic mice

Supplemental Figure 6. Induction of Ccl2 expression in nephrotoxic serum nephritis

Supplemental Figure 7. Conditional deletion of TAZ and YAP

Supplemental Figure 8. scRNA-Seq analysis of glomerular cells from diabetic, doxorubicin- treated, and CD2AP-deficient mice

Supplemental Figure 9. Validation of previously reported mesangial cell-specific genes

Supplemental Figure 10. Glomerular cells show distinct responses to different types of injury

Supplemental Table 1. Genes differentially expressed between two podocyte sub-clusters (2 and 4) in Supplemental Figure 2A

Supplemental Table 2. List of canonical cell type markers used to determine the identity of cell clusters

Supplemental Table 3. Composition of glomerular cells from a C57BL/6J mouse analyzed by scRNA-Seq

Supplemental Table 4. Top 50 cell type marker genes

Supplemental Table 5. Genes differentially expressed between mesangial cells (cluster 3) and

SMCs/JG cells (cluster 5) Supplemental Table 6. Genes enriched in mesangial cells

Supplemental Table 7. Number of differentially expressed genes in nephritic mice

Supplemental Table 8. Comparison of kidney single cell datasets

Supplemental Table 9. Genes commonly induced in mesangial cells after different injury types

Supplemental Figure 1

Supplemental Figure 1. Diagram of the glomerular structure. The glomerulus is a capillary tuft that is supplied by the afferent arteriole and drained by the . Mesangial cells provide structural support for the glomerular , which are lined by specialized fenestrated endothelial cells that allow water and soluble molecules to pass. Blood is filtered through the glomerular basement membrane (GBM) and between the branched foot processes of podocytes. The primary filtrate is collected in the Bowman’s capsule, which is lined with a layer of parietal epithelial cells, then further processed in the downstream tubules to produce urine.

The juxtaglomerular apparatus (JGA) is composed of the that senses changes in salt levels in the tubules and sends signals to regulate constriction of the afferent by vascular smooth muscle cells (SMCs) and secretion of renin by juxtaglomerular (JG) cells.

3

Supplemental Figure 2

Supplemental Figure 2. Unbiased clustering reveals the major cell types of the glomerulus. A. t-SNE representation of 5,488 glomerular cells isolated from a C57BL/6J mouse. Labels indicate clusters identified by unsupervised clustering analysis before removal of dissociation-induced genes and tubular epithelial cells. Analysis of the two podocyte sub-clusters (2 and 4) showed that the sub-clustering was driven by immediate early genes known to be induced by enzymatic tissue dissociation (Supplemental Table 1). B. Violin plots of marker gene expression in each cluster shown in A. C. Expression levels of immune cell marker genes are shown in a t-SNE plot of immune cells. Majority of the cells express the monocyte/macrophage markers Csf1r,

Adgre1, Itgam, and Ly6c2. D. Violin plots showing expression levels of novel PEC marker genes. E. Dlk1 and Ednrb are highly expressed in a sub-cluster of glomerular endothelial cells

4

(outlined in black). Expression levels are shown in a t-SNE plot of endothelial cells. Capillary endothelial cells are outlined in red, arteriolar endothelial cells are outlined in blue. F.

Immunofluorescence staining of kidney sections shows mesangial expression of the identified marker genes. PDGFRb, which stains mesangial cells in the glomerulus (dotted circle) and stromal cells outside the glomerulus, was used as reference. Scale bars, 20 µm. G. Examples of

‘mesangial-enriched’ genes that are also expressed in SMC/JG cells. Normalized expression levels are shown in the violin plots. SMC: smooth muscle cells, JG: juxtaglomerular cells, PEC: parietal epithelial cells, TEC: tubular epithelial cells.

5

Supplemental Figure 3

Supplemental Figure 3. Mapping of FSGS and CKD susceptibility gene expression to glomerular cell types. A, C. Heatmaps showing the expression levels of (A) FSGS genes and (C)

CKD susceptibility genes in each cell type. Each column represents a cell type, each row represents a gene. Results are depicted in z-score. P: podocytes, M: mesangial cells, E: endothelial cells, S: smooth muscle cells, T: tubular epithelial cells, I: immune cells. B, D. Violin

6

plots showing expression of (B) FSGS and (D) CKD susceptibility genes. Normalized expression levels are shown in the violin plots.

7

Supplemental Figure 4

Supplemental Figure 4. Nephrotoxic serum nephritis in C57BL/6J mice. A.

Immunofluorescence staining of kidney sections shows deposition of the nephrotoxic antibodies

(sheep IgG) in the glomerulus. Glomerular deposition of mouse IgG was not observed at any time point. Scale bar, 200 µm. B. Albumin/creatinine ratios measured from spot urine collected at the indicated time points after nephrotoxic serum injection. n=5. C. Representative microscopy of PAS and Masson’s trichrome stained kidney sections from control and nephrotoxic serum-injected mice show glomerulosclerosis (arrowheads) and periglomerular

8

fibrosis (arrows) in the nephritic mice. Kidneys were analyzed 5 weeks after injection. Scale bars, 50 µm.

9

Supplemental Figure 5

10

Supplemental Figure 5. scRNA-Seq analysis of glomerular cells from nephritic mice. A-F.

Unsupervised clustering analysis reveals the major glomerular cell types in cells isolated from mice at (A-C) 1 day or (D-F) 5 days after nephrotoxic serum injection. t-SNE plots were generated from cells isolated at each time point and are labeled to show (A, D) replicate samples,

(B, E) clusters identified by clustering analyses, and (C, F) cell types. G. Down-sampling analysis of nephritic glomerular cells. Since podocytes, mesangial cells, and endothelial cells comprise ~90% of the cells in the dataset, they have a larger effect on the outcome of clustering analyses compared to the remaining cell types. To reduce this bias, clustering analysis was performed after podocytes, mesangial cells, and endothelial cells were randomly down-sampled to match the number of the remaining cell types. The results are shown as a UMAP plot.

Nephritic podocytes, mesangial cells, endothelial cells still form distinct clusters from control cells, while PECs, SMCs, and tubular epithelial cells from normal and nephritic mice are grouped as single clusters. This supports the conclusion that the distinct clustering of nephritic podocytes, mesangial cells, and endothelial cells in Figure 2A is not due to batch effect. H. qPCR validation of DE genes from freshly isolated glomeruli. A subset of genes identified to be up-regulated (blue) or down-regulated (red) from scRNA-Seq analysis were analyzed by real- time qPCR using cDNA prepared from isolated glomeruli. qPCR results (normalized to Rpl13a levels) are plotted against pseudo-bulk (podocyte, mesangial, endothelial) expression levels calculated from scRNA-Seq data and shown as fold change nephritis vs. control. I. Pathways up-regulated (blue) or down-regulated (red) after nephrotoxic injury are shown. J. Expression levels of genes reveal sub-clusters of proliferating mesangial cells (black dotted circle) and endothelial cells (red dotted circle) 5 days after nephrotoxic injury. t-SNE plot is identical to

(E). K. Expression levels of Apln, Pgf, and Serpine1 in a sub-cluster of endothelial cells

11

(outlined) found at day 5. t-SNE plot is identical to (E). L. Expression levels of immune cell marker genes overlaid on the t-SNE plot shown in Figure 2E.

12

Supplemental Figure 6

Supplemental Figure 6. Induction of Ccl2 expression in nephrotoxic serum nephritis. A. t-

SNE plot of mesangial cells, myofibroblasts (MF), and immune cells from control and nephritic

13

glomeruli shows increased number of Ccl2-expressing cells after nephrotoxic injury. Ccl2- positive cells are labeled in color and Ccl2-negative cells are in gray. B. Induction of the YPet reporter in peritoneal macrophages isolated from CCL2-reporter mice upon in vitro stimulation with LPS and poly(I:C). Results represent three individual experiments analyzed by two-tailed

Student’s t test (** P<0.01, **** P<0.0001). C. qPCR analysis shows similar fold-induction of

YPet and Ccl2 transcript levels in peritoneal macrophages isolated from CCL2-reporter mice after in vitro LPS stimulation. Results are representative of 2 individual experiments. D.

Expression of YPet reporter correlates well with CCL2 production/secretion. Peritoneal macrophages isolated from CCL2-reporter mice were treated in vitro with increasing concentrations of LPS. Cells were analyzed by FACS to measure YPet flurorescence; the culture supernatant was used to measure secreted CCL2 protein levels. Pearson’s correlation R2=0.9738,

P=0.0018. E. Images of adipose tissue from control and HFD-fed CCL2-reporter mice obtained using 2-photon microscopy. YPet expression is only detected in peri-vascular cells in normal chow-fed mice. In the HFD-fed mice, reporter expression is induced in fat cells (arrowheads) and macrophages (arrows). F. Images of glomeruli from control and nephritic CCL2-reporter mice obtained by 2-photon microscopy show progressive increase in YPet expression after nephrotoxic injury. G. Glomerular cells were isolated from CCL2-reporter mice at the indicated time points after nephrotoxic serum injection and analyzed by flow cytometry to determine the percentage of YPet-expressing cells. Three mice from two individual experiments were analyzed by multiple t tests using the two-stage linear step-up method (*q <0.05, ***q<0.001). H. CCL2 is expressed in mesangial cells. Glomerular cells isolated from CCL2-reporter mice were stained for the mesangial cell marker PDGFRb and analyzed by flow cytometry. >80% of YPet-positive cells are PDGFRb-positive mesangial cells.

14

Supplemental Figure 7

Supplemental Figure 7. Conditional deletion of TAZ and YAP. A. Wild-type

(Rosa26.CreERT2, Wwtr1+/+, Yap1+/+), TAZ conditional knockout (Rosa26.CreERT2,

Wwtr1flox/flox, Yap1+/+), and YAP conditional knockout (Rosa26.CreERT2, Wwtr1+/+, Yap1flox/flox) mice were given 5 daily injections of tamoxifen. Glomerular lysates were prepared one week after the final injection and immunoblotted for TAZ and YAP. b-actin was used as loading control. B. Deletion of TAZ or YAP does not cause proteinuria. WT, Wwtr1-/-, and Yap1-/- mice were injected with sterile saline. Albumin/creatinine ratios (ACR) were measured from spot urine collected at the indicated time points.

15

Supplemental Figure 8

16

Supplemental Figure 8. scRNA-Seq analysis of glomerular cells from diabetic, doxorubicin- treated, and CD2AP-deficient mice. A. t-SNE plot of glomerular cells from control (ob/+) and diabetic mice (ob/ob). B. Pathways up-regulated in podocytes and mesangial cells in diabetic mice are shown. C. t-SNE plot of glomerular cells from doxorubicin-treated mice. D. Violin plots show down-regulation of podocyte-specific genes in podocytes after doxorubicin treatment.

E. Pathways up-regulated (blue) or down-regulated (red) in podocytes after doxorubicin treatment are shown. F. Violin plots show induction of p53 signaling target genes in podocytes after doxorubicin treatment. G, H. Violin plots show induction of Mt1 and Mt2 in (G) mesangial cells and (H) endothelial cells after doxorubicin treatment. I. Expression of immune cell markers in control and doxorubicin-treated mice. t-SNE plot is identical to Figure 4D. J. t-SNE plot of glomerular cells from 3 week-old wild-type and Cd2ap-/- mice. K. Pathways up-regulated in podocytes and mesangial cells in Cd2ap-/- mice are shown. L, M. Violin plots show increased expression of matrix genes in (L) podocytes and (M) mesangial cells. Normalized expression levels are shown in violin plots.

17

Supplemental Figure 9

Supplemental Figure 9. Validation of previously reported mesangial cell-specific genes. A.

Violin plots showing expression of previously reported mesangial-specific genes. Normalized expression levels are shown. B. Expression level of Meis1 overlaid on a t-SNE plot of glomerular cells (identical to Figure 1A). Meis1 is broadly expressed in non-mesangial cells, including vascular SMCs, endothelial cells, and podocytes. P: podocytes, M: mesangial cells, E: endothelial cells, SMC: vascular smooth muscle cells, PEC: parietal epithelial cells, I: immune cells.

18

Supplemental Figure 10

Supplemental Figure 10. Glomerular cells show distinct responses to different types of injury.

A, B. Venn diagrams showing the number of (A) up-regulated and (B) down-regulated genes commonly altered in different injury models in podocytes, mesangial cells, and endothelial cells.

19

Supplemental Table 1

podocyte mesangial endothelial PEC SMC immune TEC Nphs1 Pdgfrb Flt1 Cldn1 Acta2 Ptprc Fxyd2 Nphs2 Gata3 Tie1 Pax8 Myh11 Lyz1 Slc12a1 Pdpn Des Pecam1 Tagln Csf1r Slc12a3 Wt1 Itga8 Kdr Itgam Slc14a2 Mafb Emcn Cd3d Aqp1 Synpo Ms4a1 Aqp2 Cdkn1c Umod Ptpro PEC: parietal epithelial cell, SMC: smooth muscle cell, TEC: tubular epithelial cell

5 Supplemental Table 1. List of canonical cell type markers used to determine the identity of cell clusters.

20

Supplemental Table 2

Fold change Mean Mean Dissociation-induced Gene symbol Adj p-value (4 vs. 2) (cluster 4) (cluster 2) gene Fos 0.00E+00 6.65 51.53 7.75 Y Cyr61 3.24E-213 5.29 25.21 4.77 Y Ier2 2.12E-279 5.02 13.92 2.77 Y Atf3 5.48E-125 4.69 7.59 1.62 Y Junb 1.49E-244 4.54 19.51 4.30 Y Fosb 4.18E-169 4.26 6.12 1.44 Y Hspa1a 2.75E-234 4.06 97.50 24.02 Y Dusp1 2.15E-204 3.79 18.03 4.76 Y Jun 1.16E-183 3.61 32.85 9.09 Y Egr1 8.96E-126 3.59 7.63 2.13 Y Hspb1 2.76E-107 3.25 15.87 4.89 Y Nr4a1 2.54E-110 2.98 6.09 2.05 Y Hspa1b 9.08E-123 2.97 7.79 2.62 Y Btg2 1.91E-119 2.92 9.61 3.29 Y Hsp25-ps1 1.05E-82 2.81 6.61 2.35 Ubc 6.39E-180 2.74 36.36 13.29 Y Dnajb1 1.76E-70 2.70 5.27 1.95 Y Nfkbia 3.54E-80 2.57 7.02 2.73 Y Slc38a2 1.44E-123 2.45 6.09 2.48 Y Klf6 6.89E-70 2.38 4.75 1.99 Y Hsp90aa1 3.17E-78 2.36 8.08 3.43 Y Zfp36 2.29E-70 2.30 4.32 1.88 Y Btg1 4.64E-71 2.29 5.00 2.18 Y Gadd45b 2.46E-54 2.29 4.17 1.82 Hes1 1.88E-26 2.12 4.06 1.92 Ier3 1.27E-49 2.02 7.42 3.66 Y Ppp1r15a 6.54E-59 1.99 4.13 2.08 Y Socs3 8.01E-28 1.95 2.91 1.50 Y Jund 2.93E-62 1.93 24.14 12.49 Y Pnrc1 8.25E-54 1.89 4.16 2.20 Y Sertad1 4.97E-47 1.88 4.71 2.51 Phlda1 3.41E-25 1.86 4.02 2.16 Y Hspe1 3.21E-41 1.82 4.09 2.24 Y Txnip 3.83E-44 1.80 5.62 3.12 Ccnl1 1.63E-43 1.80 3.26 1.81 Y Klf2 6.09E-16 1.79 2.97 1.65 Y

21

Adm 1.02E-34 1.79 7.87 4.41 Cited2 1.28E-28 1.70 4.77 2.81 Dnajb4 7.70E-40 1.69 3.66 2.17 Y CT009757.7 8.00E-61 1.67 3.35 2.01 Arrdc3 8.60E-55 1.64 3.42 2.08 Hsph1 1.45E-17 1.61 2.63 1.63 Y Srf 8.00E-61 1.61 2.13 1.32 Y Csrnp1 1.63E-16 1.58 2.01 1.27 Y Hspa8 3.48E-36 1.57 5.83 3.72 Y Rhob 5.09E-21 1.55 3.40 2.19 Y Clk1 6.59E-32 1.54 3.61 2.35 Gm12346 1.04E-19 1.53 1.95 1.27 Ifrd1 2.72E-09 1.50 2.45 1.63 Y

Supplemental Table 2. Genes differentially expressed between two podocyte sub-clusters (2 and 4) in Supplemental Figure 2A. 38 of

49 differentially expressed genes are dissociation-induced genes.

5

22

Supplemental Table 3

Cell type # of cells % of total podocyte 1930 35.17% endothelial cells 1858 33.86% mesangial cells 1197 21.81% tubular epithelial cells 207 3.77% vascular SMCs 109 1.99% immune cells 98 1.79% parietal epithelial cells 89 1.62% total 5488 100.00%

5 Supplemental Table 3. Composition of glomerular cells from a C57BL/6J mouse analyzed by scRNA-Seq.

23

Supplemental Table 4

podocyte mesangial endothelial vascular SMCs immune PECs TECs Cdkn1c Sfrp2 Ly6c1 Acta2 Lyz2 Gpx3 Umod H2-Q6 Ptn Emcn Rgs5 Cd74 Lsp1 Fxyd2 Clic3 Serpine2 Plpp1 Tagln Il1b Cpe Spp1 Nphs2 Adamts5 Pi16 Tpm2 Tyrobp Pcp4 Atp1b1 Enpep Bgn Cd24a Myh11 Fcer1g Igsf5 Slc12a1 H2-Q7 Fhl2 Cyp4b1 Sparcl1 Lgals3 Igfbp6 Defb1 Nphs1 Agtr1a Gimap6 Mustn1 Cd52 Dkk3 Wfdc15b Tcf21 Igfbp4 Egfl7 Rergl Plac8 Igfbp2 Klk1 Dpp4 Gata3 Kdr Akr1b7 Ccl4 Ptgis Apela Mafb Ctgf Gimap4 Map3k7cl Ctss Cst12 mt-Nd1 Rab3b Itga8 Ctla2a Fxyd1 Cd14 Stc2 Ppp1r1a Cldn5 Lhfp Ly6e Pln Clec4e Pamr1 Cdh16 Pth1r Rgs2 Pecam1 Nrip2 Coro1a Lbp mt-Cytb Rhpn1 Sept4 Cd300lg Ppp1r14a Wfdc17 Cacna1g Wfdc2 Wt1 Mfge8 Ehd3 Cox4i2 Ccl3 Plppr4 Ldhb Sema3g Myl9 Ramp2 Lmod1 Lyz1 Mir147 Epcam Thsd7a Plvap Fabp4 Cnn1 Ccl9 Efemp1 Tmem213 Angptl2 Maf Ly6a Olfr558 Plek Ms4a2 Pdzk1ip1 Fgfbp1 Gdf10 AU021092 Gm13861 Ifi27l2a Wnt16 Tmem52b Col4a3 Serpini1 Plk2 Pdlim3 Ms4a6c AA467197 Kng2 Hs3st6 Prkca Srgn Ptgis Slfn2 Col6a1 Mal Cd59a Art3 Plpp3 Atp1a2 Chil3 Vit Kcnj1 Arhgap24 Lmo7 Cdh5 Ramp1 Alox5ap Scrn1 Plet1 Tmem150c Pdgfrb Sgk1 Aoc3 Laptm5 Fam180a Cldn8 Synpo Apoe Flt1 Rgs7bp Cybb Nkain4 Kl Pak1 Dkk2 Cdkn1a Kcna5 Spi1 Gxylt2 Sostdc1 Itgb5 Nt5e Cavin2 Itga7 Tnf Tagln Paqr5 Cmtm7 Gucy1a3 Cd200 Jph2 Ccl6 Col6a2 Cldn4 Eif3m Ndufa4l2 Cd9 Kcnab1 Lsp1 Loxl1 Pcbd1 Htra1 Rasgrp2 Sox17 Heyl Samsn1 Pebp4 Sfrp1 Metrnl Aldh2 Gimap5 Lmo1 Cd83 Sod3 Cdh1 Podxl Gpc3 Cmtm8 Gpc6 Ctsc Ogn 1700011H14Rik Aplp1 Cald1 Hsp25-ps1 Susd5 Lilrb4a Pax8 Slc16a7 Npnt Cpm Tspan7 Ppp1r12b Lst1 Clstn2 Gm45886 Ptpro Slc12a2 Mmrn2 1700120C14Rik Cd44 Cldn1 Cyfip2 Pdpn Vcl Rasip1 Fbxl22 Ifitm6 Prss23 Krt18 Mmp23 Col4a1 AC160962.1 Adamtsl1 Cytip Osr2 Clcnkb

24

Gas1 Sorbs2 Adgrf5 Sh3bgr Hp Proser2 Cldn7 Snx31 Ebf1 Gimap1 Grip2 Rac2 Rspo1 Car2 Ezr Mecom Icam1 Asb2 Mir5107 C3 Wwc1 Itm2b S1pr3 Adgrl4 Ldb3 Ms4a6b Pla2g7 Kcnj16 Ddn Pcp4l1 Lmo2 Thsd4 Ms4a4c Ncam1 Rbm47 Aif1l Gstm1 Esam Kcnk3 Cd68 Msmp Wnk4 Nupr1 Atp1b2 CT010498.1 Prkg2 Tlr2 Sp6 Eps8l2 P3h2 Col4a2 Hbegf Nkx3-1 Napsa Acot1 Tfcp2l1 Vegfa Scarf2 Ramp3 Synpo2 Ptafr Bdh2 Cdkl1 Nes Scn1b Icam2 Dhx58os Bcl2a1d Ccdc3 Aldh1l1 Col4a4 Rasl11b Pde2a Rbm24 C5ar1 Dgkg Cox7a1 Gsn Csrp1 Gm45837 Gpr20 Pld4 Smim5 Elovl7 Shisa3 Pik3r1 Tie1 Cap2 Pou2f2 Adra2c Ppargc1a

Supplemental Table 4. Top 50 cell type marker genes. Genes are listed in order of greatest fold-change compared to all other cell

types.

5

25

Supplemental Table 5

Adj p-value Fold change Mean Mean Gene symbol (5 vs. 3) (5 vs. 3) (cluster 5) (cluster 3) Ren1 1.33E-177 126.85 153.49 1.21 Acta2 0.00E+00 90.80 146.81 1.62 Tagln 0.00E+00 61.45 77.07 1.25 Crip1 0.00E+00 26.30 72.39 2.75 Rgs5 0.00E+00 24.93 68.51 2.75 Myh11 0.00E+00 22.87 23.63 1.03 Sparcl1 0.00E+00 21.13 24.36 1.15 Rergl 0.00E+00 12.23 12.41 1.01 Tpm2 0.00E+00 11.15 48.94 4.39 Sncg 0.00E+00 10.80 13.82 1.28 Map3k7cl 0.00E+00 10.15 10.30 1.01 Pln 0.00E+00 9.82 9.97 1.02 Akr1b7 0.00E+00 9.73 14.00 1.44 Mustn1 0.00E+00 8.58 21.91 2.55 Nrip2 0.00E+00 7.39 7.63 1.03 Cd9 1.11E-194 7.32 8.59 1.17 Fxyd1 0.00E+00 6.18 13.57 2.20 Palld 1.74E-251 6.17 8.31 1.35 Lmod1 0.00E+00 6.16 6.32 1.03 Cnn1 0.00E+00 5.90 5.94 1.01 Lgals1 2.50E-267 5.85 19.83 3.39 Cavin3 1.02E-134 5.75 9.51 1.65 Ppp1r14a 0.00E+00 5.38 6.71 1.25 Olfr558 0.00E+00 5.28 5.32 1.01 Actn1 1.23E-198 5.12 10.01 1.95 Zfhx3 2.11E-172 4.99 8.98 1.80 Chchd10 1.78E-113 4.90 5.64 1.15 Rasd1 2.12E-106 4.63 4.86 1.05 Cavin1 9.47E-127 4.60 10.98 2.39 Cox4i2 0.00E+00 4.59 6.88 1.50 Id2 1.17E-66 4.33 9.62 2.22 Gm13861 0.00E+00 4.28 5.05 1.18 Pdlim3 0.00E+00 4.27 4.31 1.01 Ptgis 0.00E+00 4.24 4.32 1.02 Dstn 5.72E-94 4.24 10.29 2.43 Nrarp 4.62E-156 4.18 5.41 1.29

26

Col18a1 2.76E-73 4.11 5.17 1.26 Atp1a2 0.00E+00 3.95 3.98 1.01 Hspb2 2.47E-257 3.95 4.09 1.04 Uba2 2.42E-102 3.94 8.94 2.27 Nme2 2.04E-86 3.84 8.58 2.23 Tpm1 1.32E-94 3.84 26.22 6.83 Ramp1 0.00E+00 3.83 3.86 1.01 Mef2c 4.99E-114 3.81 11.75 3.08 Filip1l 2.38E-173 3.78 4.76 1.26 Eln 8.62E-84 3.72 3.89 1.05 Aoc3 0.00E+00 3.67 4.07 1.11 Cenpa 1.49E-264 3.63 3.90 1.07 Rgs7bp 0.00E+00 3.62 3.68 1.02 Kcna5 0.00E+00 3.56 3.59 1.01 Mylk 4.71E-70 3.38 27.94 8.28 Ppp1r12a 2.01E-86 3.29 8.27 2.51 Cd200 4.41E-59 3.26 5.13 1.57 Gng11 3.11E-47 3.24 16.39 5.05 Tuba1c 4.56E-78 3.22 4.71 1.46 Mgst3 2.36E-64 3.20 4.71 1.47 Cav1 8.28E-73 3.20 3.44 1.07 Myl9 4.62E-105 3.18 91.43 28.77 Csrp2 6.22E-49 3.17 4.62 1.46 Tsc22d1 5.28E-55 3.05 18.45 6.04 Nbl1 3.19E-69 3.05 3.69 1.21 Cd24a 4.47E-62 3.02 5.85 1.93 Nexn 8.28E-117 3.00 3.98 1.32 Angptl4 2.75E-71 -3.02 1.26 3.81 Pdlim2 2.22E-69 -3.03 1.57 4.77 Coil 4.24E-74 -3.03 1.18 3.59 Gpsm3 8.76E-72 -3.07 1.29 3.94 Bst2 1.91E-55 -3.07 1.66 5.11 Mrc1 2.73E-155 -3.08 1.07 3.30 Cp 2.17E-102 -3.10 1.00 3.10 Gm9844 2.01E-71 -3.11 1.33 4.14 Eva1b 7.90E-79 -3.14 2.64 8.28 Mxd4 1.93E-76 -3.15 2.24 7.06 Srgn 5.86E-91 -3.19 1.41 4.51 Psen2 3.69E-100 -3.21 1.06 3.40 Cpm 5.66E-168 -3.22 1.57 5.04

27

Cxcl13 1.13E-31 -3.22 1.05 3.39 Col4a1 3.01E-92 -3.23 3.05 9.85 Tmsb4x 9.84E-241 -3.25 41.44 134.81 Sgk1 7.74E-72 -3.31 2.37 7.86 Rgl1 4.33E-71 -3.31 1.20 3.97 Marcksl1 5.40E-63 -3.43 1.81 6.21 S100a10 2.22E-71 -3.50 1.46 5.11 Slc12a2 2.98E-168 -3.51 1.38 4.85 Igfbp7 7.31E-183 -3.52 54.72 192.44 Cd53 1.18E-76 -3.53 1.15 4.07 Nupr1 6.62E-72 -3.62 1.78 6.42 Anxa2 4.76E-118 -3.70 2.35 8.68 Agtr1a 0.00E+00 -3.72 4.01 14.91 Ifitm3 5.16E-142 -3.75 8.71 32.65 Tmsb10 5.46E-109 -3.78 1.73 6.53 Gstm1 8.36E-114 -3.78 3.28 12.39 Cited2 1.41E-55 -3.78 2.24 8.49 Nrp1 9.46E-177 -3.79 5.29 20.07 Fhl2 2.23E-304 -3.83 4.23 16.19 Gm13588 7.58E-110 -3.90 1.63 6.34 Gpc3 4.58E-169 -3.95 1.61 6.37 Ets1 3.39E-107 -3.96 1.61 6.36 H2-D1 2.97E-203 -4.00 3.54 14.18 Lmo7 8.47E-232 -4.02 1.98 7.98 Serpine2 1.08E-228 -4.06 8.56 34.77 Igfbp5 5.37E-97 -4.12 7.90 32.53 Tns3 7.06E-121 -4.13 2.41 9.95 Gata3 2.94E-287 -4.14 3.02 12.52 Gm22133 3.38E-210 -4.16 8.30 34.52 Ftl1 2.42E-214 -4.18 8.49 35.49 Ehd3 8.11E-111 -4.20 2.25 9.47 B2m 1.76E-170 -4.29 5.94 25.45 Maf 2.37E-216 -4.33 1.84 7.96 Nt5e 1.97E-271 -4.58 1.24 5.67 Dkk2 1.79E-173 -4.61 4.75 21.86 F2r 3.96E-129 -4.67 2.13 9.96 Sept4 0.00E+00 -4.93 2.28 11.22 Itga8 0.00E+00 -5.59 2.09 11.68 Apoe 9.23E-174 -5.68 2.08 11.83 Art3 2.84E-300 -5.86 1.11 6.48

28

Cd34 9.36E-177 -5.98 1.30 7.75 Ctgf 2.04E-267 -6.17 2.46 15.19 Eng 9.89E-281 -6.21 2.18 13.54 Prkca 5.07E-227 -6.63 1.51 10.00 Plscr2 5.66E-232 -6.88 1.85 12.73 Plvap 3.42E-318 -8.33 1.21 10.09 Adamts5 0.00E+00 -8.56 2.53 21.63 Sfrp2 0.00E+00 -11.63 3.87 45.03 Igfbp4 0.00E+00 -14.36 1.56 22.37 Ptn 0.00E+00 -19.08 2.10 40.00

Supplemental Table 5. Genes differentially expressed between mesangial cells (cluster 3) and SMCs/JG cells (cluster 5)

29

Supplemental Table 6

Fold Gene Adj change Mean Mean Mean Mean Mean Mean Mean Mean symbol p-value (mesangial (mesangial) (rest) (podocyte) (endothelial) (SMC) (PEC) (immune) (TEC) vs. rest) Sfrp2 0.00E+00 22.75 44.97 1.98 1.56 1.52 3.82 1.70 1.63 1.64 Ptn 0.00E+00 20.79 39.94 1.92 1.80 1.78 2.09 2.53 1.67 1.66 Igfbp4 0.00E+00 12.72 22.34 1.76 1.33 2.89 1.57 1.82 1.43 1.50 Serpine2 0.00E+00 12.45 34.77 2.79 1.62 2.00 8.05 1.71 1.80 1.59 Adamts5 0.00E+00 11.36 21.61 1.90 1.33 2.25 2.37 2.62 1.42 1.42 Itga8 0.00E+00 8.41 11.67 1.39 1.25 1.29 2.08 1.25 1.22 1.24 Agtr1a 0.00E+00 8.08 14.91 1.85 1.25 1.25 3.86 2.23 1.21 1.27 Gata3 0.00E+00 7.72 12.51 1.62 1.22 1.27 2.94 1.28 1.20 1.82 Plvap 0.00E+00 7.42 10.07 1.36 1.21 1.91 1.22 1.26 1.38 1.17 Ctgf 0.00E+00 7.38 15.18 2.06 1.77 1.41 2.36 3.78 1.33 1.68 Dkk2 0.00E+00 7.26 21.85 3.01 1.35 8.02 4.50 1.37 1.45 1.38 Sept4 0.00E+00 7.09 11.20 1.58 1.17 1.43 2.24 1.38 1.15 2.11 Fhl2 0.00E+00 6.82 16.20 2.38 1.26 1.31 3.93 4.89 1.39 1.48 Lhfp 0.00E+00 6.55 11.37 1.74 1.24 1.37 3.80 1.58 1.21 1.22 Bgn 0.00E+00 6.46 28.85 4.46 1.64 2.71 12.92 5.99 1.85 1.68 Des 0.00E+00 6.23 15.29 2.45 1.33 1.25 7.67 1.96 1.34 1.17 Mfge8 0.00E+00 6.03 32.16 5.34 3.88 3.66 16.23 3.53 2.13 2.57 Prkca 0.00E+00 5.86 9.99 1.70 1.70 1.87 1.42 1.64 2.20 1.39 Art3 0.00E+00 5.70 6.47 1.14 1.07 1.22 1.11 1.16 1.13 1.12 Rgs2 0.00E+00 5.69 15.20 2.67 1.45 1.95 6.45 1.62 2.80 1.77 Lmo7 0.00E+00 5.20 7.99 1.54 1.14 1.62 1.75 1.34 1.11 2.27 Nt5e 0.00E+00 4.98 5.67 1.14 1.11 1.14 1.21 1.08 1.15 1.14 Gdf10 0.00E+00 4.95 7.12 1.44 1.12 1.16 2.97 1.11 1.15 1.12 Serpini1 0.00E+00 4.39 7.77 1.77 1.14 1.43 4.53 1.07 1.23 1.23 Igfbp5 0.00E+00 4.32 32.48 7.52 2.12 27.86 8.00 2.25 2.31 2.56 Maf 0.00E+00 4.25 7.96 1.87 1.06 1.37 1.83 3.86 1.74 1.38 Slc12a2 0.00E+00 4.06 4.84 1.19 1.07 1.23 1.39 1.12 1.10 1.25 Gucy1a3 0.00E+00 4.01 6.01 1.50 1.13 1.16 3.35 1.17 1.06 1.12 Cpm 0.00E+00 3.94 5.04 1.28 1.12 1.20 1.58 1.07 1.33 1.37 Gpc3 0.00E+00 3.93 6.37 1.62 1.60 1.16 1.59 2.82 1.10 1.43 Cd34 0.00E+00 3.90 7.75 1.98 1.21 5.69 1.27 1.34 1.21 1.19 Col4a1 0.00E+00 3.87 9.83 2.54 1.22 3.53 3.05 3.96 1.39 2.08 F2r 0.00E+00 3.79 9.95 2.63 2.05 5.11 2.12 3.82 1.28 1.39

30

Epas1 0.00E+00 3.75 13.64 3.64 1.81 8.00 6.49 2.04 1.73 1.78 Rasgrp2 0.00E+00 3.69 7.49 2.03 1.11 2.02 4.23 1.21 2.45 1.14 Mecom 0.00E+00 3.68 5.92 1.61 1.06 1.83 2.33 1.10 1.07 2.27 Tns3 0.00E+00 3.62 9.94 2.74 3.77 2.51 2.40 3.25 2.00 2.52 Ebf1 0.00E+00 3.58 10.86 3.04 1.14 4.11 9.02 1.28 1.49 1.18 Pdgfrb 0.00E+00 3.51 6.98 1.99 1.30 1.19 4.34 2.74 1.24 1.12 Nrp1 0.00E+00 3.47 20.06 5.78 2.91 18.91 5.08 4.18 1.79 1.84 Vcl 0.00E+00 3.45 7.17 2.08 1.59 1.72 2.78 2.65 1.54 2.20 Scn1b 0.00E+00 3.43 4.09 1.19 1.05 1.21 1.62 1.06 1.04 1.17 Col4a2 0.00E+00 3.37 7.23 2.15 1.17 2.76 3.10 3.03 1.24 1.56 Scarf2 0.00E+00 3.33 3.91 1.17 1.05 1.07 1.38 1.39 1.04 1.11 Cxcl13 2.92E-166 3.26 3.38 1.04 1.03 1.03 1.05 1.06 1.04 1.02 Eva1b 0.00E+00 3.22 8.28 2.57 2.68 4.23 2.50 2.63 1.35 2.06 Angptl4 0.00E+00 3.18 3.80 1.19 1.16 1.09 1.27 1.39 1.08 1.18 Amotl1 0.00E+00 3.14 5.01 1.60 1.18 2.53 1.96 1.35 1.27 1.28 Pik3r1 0.00E+00 3.12 4.74 1.52 1.32 1.32 1.81 1.46 1.38 1.82 Aldh2 0.00E+00 3.10 10.02 3.23 1.75 2.24 4.69 3.39 4.61 2.71 Isyna1 0.00E+00 3.10 7.07 2.28 2.08 1.89 2.96 2.76 1.74 2.27 Mrc1 0.00E+00 3.06 3.29 1.08 1.06 1.04 1.07 1.03 1.20 1.06 Coil 0.00E+00 3.02 3.58 1.19 1.12 1.35 1.19 1.14 1.22 1.10 Ets1 0.00E+00 3.02 6.36 2.11 1.52 4.36 1.53 1.66 2.23 1.36

Supplemental Table 6. Genes enriched in mesangial cells. Genes with very low expression levels (Mean < 2) in all non-mesangial

cell types are in bold.

5

31

Supplemental Table 7

up-regulated down-regulated day 1 day 5 day1 day 5 podocytes 263 80 312 132 mesangial 85 101 56 99 endothelial 165 206 189 248

5 Supplemental Table 7. Number of differentially expressed genes in nephritic mice (> 1.5 FC vs. control, adjusted p-value < 0.01)

32

Supplemental Table 8

Karaiskos Fu Park Ransick Wu Lake this study et al.23 et al.24 et al.28 et al.46 et al.47 et al.48 species mouse mouse mouse mouse mouse human mouse

purified purified whole whole whole whole whole whole whole purified sample glomeruli glomeruli kidney kidney kidney kidney kidney kidney kidney glomeruli (cells) (cells) (cells) (cells) (cells) (nuclei) (nuclei) (nuclei) (nuclei) (cells)

method Drop-Seq Fluidigm C1 10X 10X Drop-Seq DroNc-Seq Drop-Seq 10X snDrop-Seq 10X detected genes/cell 626 3457 940 1571 812 937 736 757 589 2878* (median) mitochondrial reads % 4.92% 10.91% 19.26% 1.02% 22.77% NA NA NA NA 0.72* (median) # of podocytes 10,364 48 78 24 0 45 75 107 859 11,431 # of total cells 12,954 644 43,745 31,265 3,535 2,772 3,056 2,159 17,659 74,149 % podocyte of 80.01% 7.45% 0.18% 0.08% 0.00% 1.62% 2.45% 4.96% 4.86% 15.42% total cells

skewed cell low cell low detection low detection did not distinguish glomerular/non-glomerular endothelial cells *values population numbers of glomerular of glomerular calculated cells cells from healthy conflated conflated B6 sample. vSMCs and vSMCs and did not did not values are notes mesangial cell mesangial identify identify similar for all cells mesangial mesangial samples low yield cells cells (required 32 mice to obtain ~13,000 cells)

5 Supplementary Table 8. Comparison of kidney single cell datasets.

33

Supplemental Table 9

D, C, O, N D, O, N C, O, N D, C, O D, C, N Spink8 Spink8 Acta2 Spink8 Spink8 Lgals1 F2r P2rx1 Lgals1 Lgals1 Vegfd Lgals1 Tagln Vegfd Vegfd Acta2 Vegfd Lgals1 Acta2 Tpm2 Ccnd1 Acta2 Ankrd1 Ccnd1 Acta2 Ccnd1 Spon2 Ccnd1 Enho Spink8 Vegfd Ccnd1 D: doxorubicin -/- C: Cd2ap O: BTBR ob/ob N: nephrotoxic serum nephritis

5 Supplemental Table 9. List of genes induced in mesangial cells in 3 or more injury models. Genes involved in wound healing response are in bold.

34