SETD2 deficiency accelerates sphingomyelin accumulation and promotes the development of renal cancer

Patients with polycystic kidney disease (PKD) encounter a high risk of clear cell renal cell carcinoma (ccRCC), a malignant tumor with dysregulated lipid metabolism. SET domain–containing 2 (SETD2) has been identified as an important tumor suppressor and an immunosuppressor in ccRCC. However, the role of SETD2 in ccRCC generation in PKD remains largely unexplored. Herein, we perform metabolomics, lipidomics, transcriptomics and proteomics within SETD2 loss induced PKD-ccRCC transition mouse model. Our analyses show that SETD2 loss causes extensive metabolic reprogramming events that eventually results in enhanced sphingomyelin biosynthesis and tumorigenesis. Clinical ccRCC patient specimens further confirm the abnormal metabolic reprogramming and sphingomyelin accumulation. Tumor symptom caused by Setd2 knockout is relieved by myriocin, a selective inhibitor of serine-palmitoyl-transferase and sphingomyelin biosynthesis. Our results reveal that SETD2 deficiency promotes large-scale metabolic reprogramming and sphingomyelin biosynthesis during PKD-ccRCC transition. This study introduces high-quality multi-omics resources and uncovers a regulatory mechanism of SETD2 on lipid metabolism during tumorigenesis.

chromatin biological processes, including transcriptional regulation, DNA damage repair, crosstalk of histone modification, alternating splicing and non-histone targets methylation 12 .Mutations or functional loss of SETD2 leads to dysfunctional tumor proteins and further tumorigenesis, progression, chemotherapy resistance and unfavorable prognosis 13 .As a methyltransferase, SETD2 has a capacity to regulate cellular signaling and metabolic responses through epigenetic regulation or modification of proteins [14][15][16] .However, the molecular mechanism of SETD2 in the transition from PKD to ccRCC remains elusive.
Multi-omics analysis allows for a comprehensive understanding of the molecular mechanisms underlying cancers, and facilitates the discovery of potential therapeutic targets 17 .Herein, we perform systematic multi-omics characterizations on PKD mice (driven by c-MYC) and ccRCC mice (driven by c-MYC and Setd2 knockout) to investigate SETD2 function in tumor metabolism during PDK-ccRCC transition.Specifically, we first collect metabolomics and lipidomics data and find sphingomyelin (SM) is the most increased lipid in SETD2 deficient ccRCC.Furthermore, we adopt transcriptomics and proteomics analyses to cross-verify the metabolic alteration after SETD2 deletion on mRNA and protein levels.Our results show that SETD2 loss promotes metabolic reprogramming and eventually leads to enhanced de novo biosynthesis of SM and ccRCC formation.Further spatial metabolomics with clinical ccRCC patient specimens confirm abnormal in situ SM accumulation in humans.This crucial role of SETD2 in lipid metabolism is further validified with in vivo intervention that tumor symptom caused by Setd2 knockout is mitigated by myriocin, a selective inhibitor for sphingomyelin biosynthesis.Therefore, our findings shed light on roles of the SETD2 in metabolic reprogramming and lipid biosynthesis during the transition from PDK to ccRCC.

Multi-omics analysis on SETD2-loss-induced lipid accumulation and PKD-ccRCC transition
To explore potential role of SETD2 in PKD-ccRCC transition in vivo, we generated a PKD mouse model by overexpressing oncogene c-MYC under the control of the Ksp promoter (hereafter referred as KM mice), and knockout Setd2 gene using the same Ksp promoter in this PKD mice (hereafter referred as KMS mice) to establish a ccRCC mouse model 18 (Fig. 1a).As shown in Fig. 1b, cysts were observed in both KM and KMS mice at the age of 20 weeks.Notably, KMS mice exhibited obvious structural abnormalities characterized by positive stains of Oil-red-O, a lysochrome diazo dye used for the staining of lipids 19 , and Carboxy anhydrase 9 (CA9), a major marker to diagnose ccRCC 20 , whereas KM mice did not show any neoplastic masses or lipids accumulation (Fig. 1b, c).Meanwhile, the life spans of KMS mice were much shorter compared with KM mice (Fig. 1d).These results demonstrate that SETD2 deletion increases the lipid accumulation and the incidence of ccRCC in PKD mice, highlighting the metabolic regulatory function of SETD2 in the transition from PDK to ccRCC.
To develop a comprehensive molecular understanding of SETD2loss-induced lipid accumulation and ccRCC generation, we conducted an integrative analysis on SETD2 wild-type PKD mouse model and SETD2 deficient ccRCC mouse model by incorporating multi-omics techniques (Fig. 1e).Specifically, we analyzed metabolomic and lipidomic data first to identify the accumulated lipids.Then, transcriptomic and proteomic data were utilized for regulatory network analysis on metabolic pathways.Human ccRcc specimens were further examined by spatial metabolomics to confirm the abnormal in situ sphingomyelin accumulation.At last, based on multi-omics data, targeted drug was selected for the therapeutic analysis on SETD2 deficient ccRCC mouse model.

Metabolomic and lipidomic analyses reveal sphingomyelin accumulation in SETD2 deficient ccRCC
To identify the metabolic processes and metabolites influenced by SETD2 deletion during PKD-ccRCC transition, the kidney cysts from KM mice and neoplastic masses from KMS mice were isolated and enriched by laser capture microdissection (LCM).Then liquid chromatographymass spectrometry (LC-MS)-based untargeted metabolomics and lipidomics analysis were performed on these samples (Fig. 2a).Principalcomponent analysis (PCA) discriminated SETD2 deficient neoplastic masses from KM controls based on abundance of detected metabolites (Supplementary Fig. 1a).Thirty-five differential metabolites were identified (VIP > 1) (Supplementary Fig. 1b and Supplementary Data 1) and subjected to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis.Pathway enrichment data revealed altered fatty acids biosynthesis and amino acid metabolism in SETD2 deficient KMS samples (Fig. 2b).Further, we examined the differential metabolites (VIP > 1) and found that sphingomyelin (SM) was the most prominently upregulated metabolite in KMS samples compared to KM samples (Fig. 2c).SM, a phosphosphingolipid accounting for 5-10% of total phospholipids, is characterized by a sphingoid backbone that is linked with a fatty acid, an amide bond and a phosphorylcholine head group 21 .Remarkably, SM is an essential component of mammalian cells, and regulates signal transduction, such as proliferation/survival, migration, and inflammation, in tumorigenesis 22 .

Transcriptomic and proteomic analyses further indicate enhanced metabolic pathways in SETD2 deficient ccRCC
To further characterize biological responses triggered by SETD2 deletion during PKD-ccRCC transition at the mRNA and protein levels, we next conducted global transcriptomics and proteomics studies.Cysts from KM mice and neoplastic masses from KMS mice were collected for RNA-sequencing and LC-MS/MS proteomics (Fig. 3a).Transcriptomics analysis identified a total of 1072 differential genes (FDR < 0.05, FC > 1.5) (Fig. 3b), including 817 downregulated genes (76.21%) and 255 upregulated genes (23.79%) in KMS mice compared to KM mice (Fig. 3c).Proteomics analysis identified a total of 290 differential proteins (FDR < 0.05, FC > 1.5) (Fig. 3b, Supplementary Fig. 2a), including 160 upregulated proteins (44.83%) and 130 downregulated proteins (55.17%) in KMS mice compared to KM mice (Fig. 3c).Of note, a prominent uncoupling of mRNA and protein expression after SETD2 ablation was revealed through analysis of the differential mRNAs and proteins between KMS and KM (Supplementary Fig. 3a, b).
To further clarify the biological processes and pathways affected by SETD2 deficiency, we implement enrichment analysis based on differential mRNAs and proteins.KEGG and Gene Ontology (GO) enrichment analysis on proteomics data indicated distinct alterations in metabolic pathways (Fig. 3e, Supplementary Fig. 2b, c), which were more significant than that on transcriptomics data (Fig. 3d, Supplementary Fig. 3c).We next integrated proteomic and transcriptomic datasets to determine the dysregulation of cellular processes induced by SETD2 deficiency.By adopting gene set enrichment analysis (GSEA) pathway enrichment, we found half pathways were either upregulated or downregulated at both protein and mRNA levels, including upregulated fatty acid metabolic process and downregulated amino acid metabolism (Fig. 3f).These results demonstrate that SETD2 indeed regulates broad metabolic pathways during PKD-ccRCC transition.
In our study, we demonstrated that SETD2 ablation resulted in decreased amino acid metabolism and increased fatty acid biosynthesis during PKD-ccRCC transition.Notably, enhanced levels of amino acid and palmitoyl-CoA are critical to de novo biosynthesis of SM 25 .
Our data showed increased protein levels of TECR, KDSR and LPCAT3, which are essential enzymes in SM biosynthesis, as well as increased palmitic acid, sphingomyelin (d18:1/18:0), sphingomyelin (d40:1) and sphingomyelin (d41:1) in KMS samples compared to KM samples (Fig. 4e, f).These results suggest that SETD2-deficiency-induced metabolic alterations contribute to de    while the deletion of SETD2 promotes de novo synthesis of SM during PKD-ccRCC transition.

SETD2 loss is accompanied by increased sphingomyelin in human ccRCC
To further examine the clinical functional relevance between SETD2 and SM biosynthesis, we analyzed transcriptomic and proteomic signatures in ccRCC patients from The Cancer Genome Atlas (TCGA) database (n = 533) and The Clinical Proteomic Tumor Analysis Consortium (CPTAC) database (n = 606), respectively.Clinically, mutations and low expression of SETD2 were widely observed in ccRCC samples (Supplementary Fig. 4a, b).At the protein level, 17 and 38 proteins were upregulated and downregulated respectively at both mouse and human ccRCC, which account for 60.45% of total differential metabolic proteins (Fig. 5a).Specifically, upregulated proteins, including PFKP, SLC16A1, KDSR, LPCAT3 and CDIPT, as well as downregulated proteins, including HAAO, GOT2, FABP3, IDH3G and GLYCTK caused by SETD2 deletion in mice, were also consistently increased or decreased respectively in human ccRCC (Fig. 5b).GSEA analysis on clinical proteomic data from CPTAC database also showed statistically significant enrichment between SETD2 alteration and metabolic pathways (Supplementary Fig. 4c).Correlation analysis showed positive correlations between SETD2 and the levels of proteins in TCA cycle and amino acid metabolism.Meanwhile, the low level of SETD2 was clinically related to the upregulations of proteins in glycolysis, protein catabolic and lipid biosynthesis (Fig. 5c, Supplementary Fig. 4d).These results indicated a positive correlation of SETD2 loss with metabolic reprogramming in human ccRCC, consistent with our previous findings in SETD2 deficient mice.
To further validate the relevance between SETD2 and SM biosynthesis, we detected levels of SETD2 protein and SM in human ccRCC samples.Immunofluorescent (IF) staining showed reduced protein level of SETD2 in ccRCC tissues compared to adjacent normal tissues.Then MALDI imaging mass spectrometry (MALDI-IMS) was performed to measure the in situ metabolites abundance on paired human ccRCC and normal samples with altered SETD2 levels.Our data showed significantly increased SMs in SETD2-low tumors, including SM (40:7;3O, m/z 791.5757),SM (40:6;3O, m/z 793.5889),SM (40:7;3O, m/z 791.5757) and SM (40:5;3O, m/z 795.6071), demonstrating a clear inverse correlation between SETD2 and SM (Fig. 5d and Supplementary Fig. 4e).Collectively, our results demonstrated that SETD2 loss is significantly associated with upregulated SM in human ccRCC.

Inhibition of sphingomyelin biosynthesis relieves the symptom caused by SETD2 deletion in mice
In light of the observation that SETD2 deficiency gave rise to enhanced SM content, which is strongly related to cancer initiation, growth, and immune evasion 22 , we were prompted to test whether SM synthesis inhibitors can confer protective effects against SETD2-loss-induced PKD-ccRCC transition in vivo.De novo synthesis of SM is initiated by the condensation of serine and palmitoyl-CoA, which is catalyzed by serine palmitoyl transferase (SPT) 26 .Myriocin, an analog of sphingosine that functions as a potent inhibitor of SPT, has been evaluated in vivo as a metabolic modulator with favorable effects in metabolic syndrome 25 and cancers [27][28][29] .In our study, ten-week-old KMS mice were treated with myriocin intraperitoneally (0.3 mg/Kg) every two days for 3 months to block de novo synthesis of SM and subjected to ultrasonic testing every month (Fig. 6a).As depicted in Supplementary Fig. 5a, a significant decrease in SM levels were observed in myriocin-treated KMS mice in comparison to the vehicle-treated KMS mice.Inspiringly, the results of ultrasonic testing before and after treatment showed less neoplastic masses in myriocin-treated mice compared to vehicletreated mice (Fig. 6b).And Myriocin-treated mice exhibited lighter body weight (P = 0.0175) and smaller kidney size (P = 0.0073) compared to vehicle-treated mice (Fig. 6b-d).Morphologically, smaller clear-cell areas were observed in myriocin-treated mice compared to vehicle-treated mice.Further histology analysis showed that myriocintreated mice exhibited attenuated staining of CA9 and less lipid accumulation compared to vehicle-treated mice (Fig. 6e, f).These results intuitively revealed that the ccRCC-related symptoms caused by SETD2 deletion were alleviated and progression from PKD to ccRCC was blocked after the utilization of myriocin.Furthermore, to impede SM synthesis, we also employed lentivirus-mediated knockdown systems targeting Serine Palmitoyltransferase Long Chain Base Subunit 2 (SPTLC2), a pivotal enzyme responsible for initiating the de novo synthesis of SM 26 , in SETD-mut ccRCC cell line CAKI-1.Data showed a notable decrease in SM levels (Supplementary Fig. 5b), accompanied by a reduction in cell proliferation both in vitro and in vivo (Supplementary Fig. 5c, d) upon SPTLC2 knockdown.Collectively, the obtained results indicate that SETD2 loss promotes the transition from PKD to ccRCC in a SM dependent manner.
Taken together, our findings demonstrated that SETD2 loss induces extensive and substantial metabolic reprogramming events, including upregulated glycolysis, protein catabolic and fatty acid biosynthesis, as well as downregulated TCA cycle and amino acids metabolism.Metabolic reprogramming caused by SETD2 deletion ultimately accelerates de novo sphingomyelin (SM) synthesis and tumorigenesis during the transition from PKD to ccRCC (Fig. 7).

Discussion
SETD2 is frequently mutated across cancer types, including lung cancer 30,31 , prostate cancer 32 , intestinal cancer 33 , glioma 34 , gastrointestinal tumors 33,35 , osteosarcoma 36 and leukemia 37,38 .Genomic analyses on hepatocellular carcinoma 39 and lung adenocarcinoma 40 showed that deletions in SETD2 might drive metabolic alterations.Metabolic reprogramming is a hallmark of tumor cells and has been recently regarded as a key contributor in tumor progression 7,41,42 .To better understand the role of the SETD2 in tumor metabolism in RCC progression, we systematically investigated the function of SETD2 in PKD-ccRCC transition based on comprehensive analyses of transcriptomics, proteomics, lipidomics and metabolomics.In our study, the dysregulation of carbohydrate absorption, glycolysis, lipid biosynthesis, sphingolipid metabolic, TCA cycle and amino acid metabolism were observed after Setd2 knockout.The alterations on these metabolic pathways have been reported to promote tumor growth, progression and metastasis 28,43,44 , and are consistent with the notion that ccRCC cells are lipid-and glycogen-laden 7 .Therefore, our results indicate a strong relation between SETD2 and metabolic reprogramming in ccRCC.
In this study, we identified abnormal biosynthesis of SM in SETD2 deficient ccRCC.Functions of SM on regulating signal transduction have been proposed, such as proliferation/survival, migration and inflammation 22 .These functions are mostly linked to the presence of SM in the plasma membrane, the Golgi or the nucleus 21 .Cell surface biochemical changes, notably excessive increase in outer leaflet SM content, are of importance in cancer initiation, growth and immune evasion 45 .Accumulating evidences suggest that lung, colon, prostate, breast and kidney cancers have relations to SM content and their metabolizing enzymes, confirming that the presence of inordinate SM amount on cancer cells is associated with the level of malignancy [46][47][48] .Therefore, our work shed light on a mechanism that SETD2 deficiency would exert a crucial impact on SM-mediated biochemical functions and ultimately facilitate the transition from PKD to ccRCC.
A multigene panel that includes PKD1, PKD2, ALG5, ALG9, DNAJB11, GANAB, HNF1B, IFT140 was reported to be relevant to the genetic cause of PKD [49][50][51][52] .However, it is worth noting that there is no evidence indicating a common occurrence of SETD2 mutations in PKD, suggesting that SETD2 deficiency have a low relation to PKD process.Meanwhile, mutations of the SETD2 gene are common (occurs up to 12%) in ccRCC patients 10 , thus indicating that SETD2 deficiency may indeed play a crucial role in the development of ccRCC but not PKD.Increased CA9, recognized as a target of VHL/HIF signaling 53,54 , was observed subsequent to Setd2 knockout in our mouse model.Clinically, low SETD2 is highly related to reduced VHL in mRNA and protein levels (TCGA-KIRC database, Supplementary Fig. 4f, g), suggesting an association between SETD2 loss and decreased VHL expression in the context of ccRCC.This observation provides a foundation for potential future endeavors in basic research, aimed at validating the regulatory influence of SETD2 on VHL/HIF signaling in ccRCC.As a methyltransferase, SETD2-mediated H3K36me3 is essential in ccRCC, and the number of H3K36me3-positive nucleus reduced ~20% in primary ccRCC 10 .In our investigations, altered H3K36me3 modifications on

Glycolysis Citrate cycle (TCA cycle)
Amino acid metabolism Protein catabolic  metabolic genes were detected upon SETD2 deficiency (ChIP-seq assays targeting H3K36me3, Supplementary Data 3 and Supplementary Fig. 6a), and 31.87% of altered metabolic proteins exhibited changes in H3K36me3 modifications on their gene bodies (Supplementary Data 4 and Supplementary Fig. 6b).Moreover, SETD2 methyltransferase mutant is deficient on regulating sphingomyelin (Supplementary Fig. 6c).This analysis indicates that SETD2-mediated H3K36me3 modification involves in the alteration of metabolic proteins and sphingomyelin accumulation.Additionally, extant literature have also asserted that SETD2 is able to regulate biological processes through modification of non-histone substrates, such as STAT1 14 and αtubulin 15 .It remains to be determined whether SETD2 regulates tumor metabolism by interacting with these proteins and methylating them during PKD-ccRCC transition.
Several ongoing clinical trials have provided preliminary insights into fatty acid and lipid-blocking therapies 55,56 .Of note, in our study, tumor symptom caused by Setd2 knockout was relieved by myriocin.This finding suggested that SETD2 could be a promising target for metabolic therapy.Myriocin, a selective inhibitor of serine-palmitoyltransferase (SPT), can block de novo biosynthesis of SM 57,58 .SPT acts as a metabolically responsive "switch" in the lipidome, and inhibition of SPT by myriocin have been co-opted to hinder tumor growth including HCT116, A549 and liver cancer 27,28 .Moreover, contact inhibition, one of the major mechanisms in immune checkpoint underlying the initiation of tumorigenesis, is also affected by the cell surface lipids including SM 45 .Accordingly, SM would be a potential therapeutic target in the anti-tumor immune therapy in the future.
Collectively, we demonstrated that SETD2 deficiency results in large-scale metabolic reprogramming and de novo SM biosynthesis, thereby accelerating the transition from PKD to ccRCC.Our findings reveal the role of the SETD2 in metabolic reprogramming and lipid metabolism in ccRCC.For clinical translation, pharmaceutical investigation of the cross-talks between SETD2 and lipids metabolism especially SM may provide enlightenments for preventing tumorigenesis in PKD patients and ccRCC therapy.

Mouse strains
Setd2 fl/fl mice were generated as previously reported 18 .MYC R26StopFL/+ mice were mated with Ksp Cre mice to generate Ksp Cre ; MYC R26StopFL/+ (MYC overexpression) mice in C57BL/6 background.Setd2 fl/fl mice were mated with Ksp Cre ; MYC R26StopFL/+ mice to generate Ksp Cre ; MYC R26StopFL/+ Setd2 flox/flox (MYC overexpression; Setd2 knockout) mice.Ksp Cre ; MYC R26StopFL/p mice and Ksp Cre ; MYC R26StopFL/p ; Setd2 flox/flox mice were referred as KM and KMS mice, respectively.All mice were maintained in a specific-pathogen-free (SPF) facility and were housed at a temperature of 25 °C in a humidity-controlled environment with free access to food and water in a 12 h light/dark cycle.All experimental procedures were approved by the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University.Diets for all animal experiments were purchased from SLACOM (Catalogue number: P1101F-25).Our study is not sex-or gender-based, this information has not been collected in our research.

Histology and immunohistochemistry (IHC) staining
Tissues were fixed in 10% buffered formalin and fixed tissues were sectioned for hematoxylin and eosin (H&E) staining.For IHC staining, paraffin-embedded tissues were deparaffinized, rehydrated, and subjected to a heat-induced epitope retrieval step by treatment with 0.01 M sodium citrate (pH 6.0).Endogenous peroxidase activity was blocked with 0.3% (v/v) hydrogen peroxide in distilled water.The sections were then incubated with 0.3% Triton X-100 in PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, pH 7.4) for 15 min, followed by 10% goat serum in PBS for 1 hour.Subsequently, samples were incubated with primary antibodies, diluted at appropriate proportion in 1% goat serum overnight at 37 °C.After three washes in PBS, sections were incubated with an HRP-conjugated secondary antibody for 1 h at room temperature.Sections were counterstained with hematoxylin.Three random immunostaining images of each specimen were captured using a Leica DM2500 microscope and analyzed by Image-Pro Plus 6.0 software.Primary antibody against CA9 (NB100-417) was purchased from Novus Biologic.NB anti-CA9 Antibody is a rabbit polyclonal antibody raised against human and mouse CA9 and is recommended for detection of CA9 by IHC at dilution range 1:200 ~1:500 59,60 .Anti-rabbit IgG HRP-linked antibody (#7074) was purchased from Cell Signaling Technology.

Oil-red-O staining and immunofluorescent assays
Tissues were rinsed in ice-cold PBS, fixed in 4% formalin for 15  were stained with freshly prepared Oil-red-O solution for 10 min, then rinsed with 75% ethanol, and nuclei were stained with haematoxylin before examination.For IHC staining, the cover slides were first treated with 0.5% Triton for 10 min and blocked with 10% goat serum at room temperature for 1 h.Primary antibody incubations were performed overnight at 4 °C.After extensive washing with PBS, secondary antibody was applied to the sections at room temperature for 1 h.Slides were washed with PBS three times and then mounted with Vectashield mounting medium (Vector laboratories, Inc.H-1200).Three random immunostaining images of each specimen were captured using a Leica DM2500 microscope and analyzed by Image-Pro Plus 6.0 software.Primary antibody against SETD2 (LS-C332416) was purchased from LifeSpan Biosciences.LS-Bio anti-SETD2 Antibody is a rabbit polyclonal antibody raised against human and mouse SETD2 and is recommended for detection of SETD2 by immunofluorescence at dilution range 1:50 ~1:200 39,61 .Primary antibody against CA9 (NB100-417) was purchased from Novus Biologic.NB anti-CA9 Antibody is a rabbit polyclonal antibody raised against human and mouse CA9 and is recommended for detection of CA9 by immunofluorescence at dilution range 1:200 ~1:500 59,60 .Anti-rabbit IgG (Alexa Fluor® 488 Conjugate) (#4412) was purchased from Cell Signaling Technology.

Metabolomics and lipidomics profiling
Fresh kidney tissues from mice were rapidly frozen in liquid nitrogen and then stored intact at −80 °C until LCM analysis.18-μm thick sections for LCM were thaw-mounted on MembraneSlide (1.0 PEN, ZEISS).ccRCC and PKD tissues were sectioned with a ZEISS PALM system and collected onto AdhesiveCaps (ZEISS).

RNA-seq and analyses
Kidney tissue from mice were harvested for RNA preparation.The complementary DNA (cDNA) libraries were prepared using the NEB-Next TM Ultra Directional RNA Library Prep Kit, NEBNext Poly (A) mRNA Magnetic Isolation Module, NEBNext Multiplex Oligos according to the manufacturer's instructions.The products were purified and enriched by PCR to create the final cDNA libraries and quantified by Agilent2200.The tagged cDNA libraries were pooled in equal ratio and used for 150 bp paired-end sequencing in a single lane of the Illumina HiSeqXTen.The raw sequencing data are evaluated by fastp (https:// github.com/OpenGene/fastp)including quality distribution of nucleotides, position specific sequencing quality, GC content, the proportion of PCR duplication, kmer frequency etc.These evaluation metrics may help us understand the nature of data deeper before subsequent variant evaluation.Mapping of pair-end reads.Before read mapping, clean reads were obtained from the raw reads by removing the adapter sequences, reads with >5% ambiguous bases (noted as N) and low-quality reads containing more than 20 percent of bases with qualities of <20.The clean reads were then aligned to mouse genome (version: GRCm38 NCBI) using the HISAT2 v2.2.1 (http:// daehwankimlab.github.io/hisat2/) 64.HTseq was used to count gene counts 65 and RPKM method was used to determine the gene expression.RUVSeq was utilized to eliminate unwanted variation from RNA-Seq data 66 .P-value and FDR analysis were subjected to the following criteria: i) Fold Change>1.5 or <0.667; ii) P-value < 0.05, FDR < 0.05.

Proteomics profiling
Tissues were homogenized in RIPA buffer containing protease inhibitor and 100 mM PMSF.Protein lysates were processed by the FASP protocol as described 67 .Proteins were digested with trypsin (Promega) and the peptide solutions were further desalted on a MonoSpin C18 column (GL Sciences).For the proteome library, 80 ug peptides were fractionated using a high pH C18 reverse-phase spin column, the fractionated samples were dried and resuspended for LC-MS/MS measurement.

Chromatin immunoprecipitation sequencing and analyses
Cells were cross-linked with 1% formaldehyde for 10 min at room temperature and quenched with 125 mmol/L glycine.The fragmented chromatin fragments were precleared and then immunoprecipitated with Protein AþG Magnetic beads coupled with anti-H3K36me3 (ab9050) and immunoglobulin G (Santa Cruz Biotechnology).Abcam anti-H3K36me3 antibody (ChIP Grade) is a rabbit polyclonal antibody raised against human and mouse H3K36me3 and is recommended for chromatin immunoprecipitation assay at dilution range 1:200 ~1:500 69,70 .After reverse cross-linking, chromatin immunoprecipitation (ChIP) and input DNA fragments were end-repaired and A-tailed using the NEBNext End Repair/dA-Tailing Module (E7442, NEB) followed by adapter ligation with the NEBNext Ultra Ligation Module (E7445, NEB).The DNA libraries were amplified for 15 cycles and sequenced using Illumina NextSeq 500 with single-end 1 × 75 as the sequencing mode.Raw reads were filtered to obtain high-quality clean reads by removing sequencing adapters, short reads (length <50 bp), and low-quality reads using Cutadapt (v1.9.1) and Trimmomatic (v0.35).Then FastQC is used to ensure high reads quality.The clean reads were mapped to the mouse genome (assembly GRCm38) using the Bowtie2 (v2.2.6) software.Peak detection was performed using the MACS (v2.1.1)peak finding algorithm with 0.01 set as the P value cutoff.Annotation of peak sites to gene features was performed using the ChIPseeker and showed in Supplementary Data 3.

Pathway enrichment analysis
Metabolites with significant differences (VIP > 1) between two groups were used for KEGG enrichment analysis using MetDNA v1.0 (http:// metdna.zhulab.cn/) 62.The GO/KEGG enrichment analysis was executed using clusterProfiler package (version: v.4.5.0) 71 through Hiplot Pro (https://hiplot.com.cn/), a comprehensive web service for biomedical data analysis and visualization.Gene expression data of mRNA and protein levels were used to obtain NES over the active GO/KEGG biological processes using GSEA software (v.4.1) (http://software.broadinstitute.org/gsea/index.jsp) 72.GO/KEGG terms were plotted with NES values for mRNA and protein.GO gene sets belonging to Mus musculus used in this article were integrated from AmiGO database (http://amigo.geneontology.org/amigo/landing).

MALDI imaging mass spectrometry (MALDI-IMS)
Fresh samples of human ccRCC and paired normal tissue used for MALDI-IMS were obtained during surgery at the Ruijin Hospital Affiliated to Shanghai Jiaotong University.All samples were collected with the informed consent of patients and in compliance with the strategy of the Ethics Committees of the Ruijin Hospital, Shanghai Jiaotong University School of Medicine.Our study is not sexor gender-based, this information has not been collected in our research.
Fresh tissues were rapidly frozen in liquid nitrogen and then sectioned on ITO slides (Bruker Daltonics) with a Cryostat (Thermo Fisher Scientific) at −20 °C.15 mg/mL DHB (2,5-dihydroxybenzoic acid) matrix solution was sprayed on the tissue surface via the HTX TM-Sprayer (TMSP-M3, HTX Technologies).The slides were imaged on a MALDI-TOF mass spectrometry (timsTOF flex, Bruker Daltonics).Imaging data were acquired in positive ion mode in the range of 300-1500 m/z at 50 μm spatial resolution with 400 laser shots per position.Mass spectra were visualized using SCiLS Lab software (Bruker Daltonics).

Isolation of primary tubular epithelial cells (PTECs) and cell cultures
The isolation procedures and phenotype identification were performed as previously described 73 .PTECs were maintained in Dulbecco's modified Eagle's medium/F-12 GLUTMAX-1 containing 10% fetal bovine serum (FBS), 100 U/ml of penicillin and 100 μg/ml of streptomycin (all used for cell culturing are from Sigma-Aldrich, St. Louis, MO).All the cells were maintained at 37 °C under humidified air containing 5% CO 2 .Mycoplasma, bacteria, and fungi were tested as negative in these cultures.

SETD2 knockout/inhibition and Sphingomyelin Assay
To investigate the role of SETD2 in sphingomyelins accumulation, lentivirus expressing Cre and SETD2-IN-1 (1 μM) were used to knockout Setd2 gene and inhibit SETD2 function respectively in renal primary tubular epithelial cells (PTECs) from Setd2 flox/flox mice.Lentiviral packaging plasmids pCMV-DR8.8and pMD2.G were co-transfected with the backbone plasmid into 293T cells for virus production.Cells were selected in 2.5 µg/mL puromycin in the culture medium or by fluorescence-activated cell sorting to generate the stable transfections.PTECs were then cultured and harvested for sphingomyelin assay.Sphingomyelin Quantification Colorimetric Assay Kit (ab287856) was used for measuring the levels of sphingomyelins in PTECs.

Plasmids, transfection, and lentivirus
h-SETD2 methyltransferase mutant (c.4871C>G, p.S1624C 74 ) plasmid was a kind gift from Guan Ning Lin (Shanghai Jiao Tong University, Shanghai, China).Human SETE2 cDNA was generated by polymerase chain reaction.The cDNAs of p.S1624C SETD2 and wild-type SETD2 was cloned into pCMV6-Entry vector with HA-tag.ShRNA sequences for SPTLC2 and scramble shRNA were cloned into lentiviral vector pLKO.1-GFP-puro(sh-SPTLC2#1: CTTCGAGATTTCTTGAGGTAT; sh-SPTLC2#2 GCTCATACCAAAGAAATACTT).For transfection, cells were transfected with the jetPRIME transfection reagent (Polyplus) according to the manufacturer's instruction.All the constructs generated were confirmed by DNA sequencing.Lentiviral packaging plasmids pCMV-DR8.8and pMD2.G were co-transfected with the backbone plasmid into 293 T cells for virus production.Cells were selected in 2 µg/mL puromycin in the culture medium to generate the stable transfections.

In vivo xenograft assay
Cell suspensions (1 × 10 6 cells) of sh-SPTLC2#1 and control CAKI-1 cells, in a total volume of 100 μL mixed with matrigel in a 1:1 ratio, were injected subcutaneously into the right flanks of five 4-week-old male BALB/C nude mice (SLAC, Shanghai) respectively.Tumor volume was calculated with the following formula: volume = 4/3 × π × tumor length/2 × tumor width/2 2 .Tumors were collected and photographed 20 days after inoculation.All mice were maintained in a specificpathogen-free (SPF) facility and were housed at a temperature of 25 °C in a humidity-controlled environment with free access to food and water in a 12 h light/dark cycle.Diets for all animal experiments were purchased from SLACOM (Catalogue number: P1101F-25).All animal experiments were carried out following the ethical regulations of the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University.The maximal tumor size of 1500 mm 3 was permitted by the Animal Care Committee.In our study, the maximal tumor size was not exceeded.

Quantification and statistical analysis
Statistical evaluation was conducted using Student's t test.Multiple comparisons were analyzed first by one-way analysis of variance.The Pearson correlation was used to analyze the strength of the association between expression levels of SETD2 and its related proteins in patient samples.A significant difference was defined as P < 0.05.

Fig. 1 |
Fig. 1 | SETD2 loss leads to lipid accumulation and ccRCC generation in PKD mice.a Schematic representation of the establishment of PKD and ccRCC mouse model.b Representative images of H&E, Oil-red staining and longitudinal ultrasonic testing.Aberrant areas were indicated by arrows.Kidneys were indicated by white dotted lines in ultrasonic images.c KMS mice exhibited larger aberrant areas (P = 0.0204), thicker renal tubules (P < 0.0001), larger kidney volumes (P = 0.0360), and more positive staining for Oil red (P < 0.0001) and CA9 (P = 0.0009).d Kaplan-Meier survival curve of indicated mice.e General workflow of multi-omics investigations and data analysis.Scale bars, 50 μm.Statistical comparisons were made using a two-tailed Student t test.Data are represented as mean ± SEM.Source data are provided as a Source Data file.

Fig. 2 |Fig. 3 |3
Fig. 2 | Metabolomics analysis reveals sphingomyelin accumulation in SETD2 deficient ccRCC.a Schematic representation of the metabolomics analyses of KM and KMS mice.b Altered KEGG metabolic pathways in KMS mice compared to KM mice enriched by significantly altered metabolites (VIP > 1).c Differential metabolites between KMS and KM mice (VIP > 1).Top 7 of the most upregulated metabolites are colored in red.d Chemical composition of the mouse renal tubules lipidome using ClassyFire categories to classify the lipids diversity of all annotated lipids across assays.e Relative abundance of representative lipids between KM (n = 5) and KMS (n = 3).Statistical comparisons were made using a two-tailed Student t test.Data are represented as mean ± SEM. f Differential lipids between KMS and KM mice (VIP > 1).Top 2 of the most upregulated lipids are colored in red.Source data are provided as a Source Data file.

Fig. 4 |
Fig.4| SETD2 deficiency leads to metabolic reprogramming and upregulated sphingomyelin biosynthesis.a GSEA of KEGG metabolic pathways enriched in KMS mice compared to KM mice.Upregulated and downregulated pathways are shown in red and blue, respectively.b Heat map and quantitative analysis of altered proteins (FDR < 0.05) in upregulated and downregulated pathways between KMS and KM mice.c, d Reprogrammed metabolic pathways in KMS mice compared to KM mice.e Schema of de novo sphingomyelin biosynthesis with upregulated proteins and metabolites in KMS mice compared to KM mice.The most critical representative proteins and metabolites were highlighted in green and yellow, respectively (FDR < 0.05 and fold change >1.5).f KMS mice showed increased TECR (P < 0.0001), KDSR (P = 0.0049), LPCAT3 (P < 0.0001), Palmitic acid (P = 0.0235), and Sphingomyelin d18:1/18:0 (P = 0.0374), d40:1 (P = 0.0299) and d41:1 (P = 0.0198) compared to KM mice.Statistical comparisons were made using a 2-tailed Student's t test.Data are represented as mean ± SEM.Source data are provided as a Source Data file.

Fig. 5 |Fig. 6 |
Fig. 5 | SETD2 loss associates with upregulated sphingomyelin biosynthesis in human ccRCC.a Sample-averaged, normalized and log-transformed change level for altered protein pairs in human and mouse ccRCC.Up-and down-regulated proteins at both mouse and human ccRCC are shown in red and blue, respectively.b Increased PFKP, SLC16A1, KDSR, LPCAT3 and PFKP proteins, as well as decreased HAAO, GOT2, FABP3, IDH3G and GLYCTK proteins were observed in ccRCC samples compared to adjacent normal kidney tissues (P < 0.0001).c The correlation of key protein abundances and protein level of SETD2 in human ccRCC.d Representative immunofluorescent images showing protein levels of SETD2 in ccRCC patients and representative MALDI-IMS images of ccRCC tissue from the same patients.Statistical comparisons were made using a 2-tailed Student's t test.Data are represented as mean ± SEM.Source data are provided as a Source Data file.

Fig. 7 |
Fig. 7 | SETD2 deficiency accelerates sphingomyelin accumulation and promotes the transition from PKD to ccRCC.Schematic representing the role of SETD2 deficiency in promoting large-scale metabolic reprogramming and sphingomyelin biosynthesis during PKD-ccRCC transition.

PKD Frozen tissue section ccRCC Metebolits & Lipids Laser microdissection Metabolomics & lipidomics KMS KM
novo synthesis of SM.Our study reveals a critical function of SETD2 in metabolic reprogramming,