ARTICLESMPDL3b modulates insulin receptor signaling indiabetic kidney diseaseA. Mitrofanova1,2,3, S.K. Mallela1,2, G.M. Ducasa1,2,4, T.H. Yoo1,2,5, E. Rosenfeld-Gur6, I.D. Zelnik6, J. Molina1,2,J. Varona Santos1,2, M. Ge1,2,4, A. Sloan1,2, J.J. Kim1,2, C. Pedigo1,2,7, J. Bryn1,2, I. Volosenco 1,2,8, C. Faul1,2,9,Y. H. Zeidan2,10,11, C. Garcia Hernandez1,2,10, A.J. Mendez12, I. Leibiger13, G.W. Burke3,12, A.H. Futerman6,L. Barisoni14, Y. Ishimoto 15,16, R. Inagi 15,16, S. Merscher1,2 & A. Fornoni1,2Sphingomyelin phosphodiesterase acid-like 3b (SMPDL3b) is a lipid raft enzyme that reg-ulates plasma membrane (PM) ﬂuidity. Here we report that SMPDL3b excess, as observed inpodocytes in diabetic kidney disease (DKD), impairs insulin receptor isoform B-dependentpro-survival insulin signaling by interfering with insulin receptor isoforms binding to caveolin-1 in the PM. SMPDL3b excess affects the production of active sphingolipids resulting indecreased ceramide-1-phosphate (C1P) content as observed in human podocytes in vitro andin kidney cortexes of diabetic db/db mice in vivo. Podocyte-speciﬁc Smpdl3b deﬁciency in db/db mice is sufﬁcient to restore kidney cortex C1P content and to protect from DKD. Exo-genous administration of C1P restores IR signaling in vitro and prevents established DKDprogression in vivo. Taken together, we identify SMPDL3b as a modulator of insulin signalingand demonstrate that supplementation with exogenous C1P may represent a lipid therapeuticstrategy to treat diabetic complications such as DKD.https://doi.org/10.1038/s41467-019-10584-4 OPEN1 Katz Family Division of Nephrology and Hypertension, Department of Medicine, University of Miami, Miller School of Medicine, Miami 33136 FL, USA.2 Peggy and Harold Katz Family Drug Discovery Center, University of Miami, Miller School of Medicine, Miami 33136 FL, USA. 3 Department of Surgery,University of Miami, Miller School of Medicine, Miami 33136 FL, USA. 4Department of Molecular and Cellular Pharmacology, University of Miami, MillerSchool of Medicine, Miami 33136 FL, USA. 5Department of Internal Medicine, College of Medicine, Yonsei University, Seoul 03722, Korea. 6 Department ofBiological Chemistry, Weizmann Institute of Science, Rehovot 76100, Israel. 7 Department of Internal Medicine, Yale University School of Medicine,New Haven 06510 CT, USA. 8 Lewis Gale Medical Center, Salem 24153 VI, USA. 9Division of Nephrology, Department of Medicine, University of Alabama atBirmingham, Birmingham 35233 AL, USA. 10 Department of Radiation Oncology, University of Miami, Miller School of Medicine, Miami 33136 FL, USA.11 Department of Radiation Oncology, American University of Beirut, Beirut 1107 2020, Lebanon. 12 Diabetes Research Institute, University of Miami, MillerSchool of Medicine, Miami 33136 FL, USA. 13 The Rolf Luft Research Center for Diabetes and Endocrinology, Karolinska Institutet, Stockholm 17176, Sweden.14 Department of Pathology, University of Miami, Miller School of Medicine, Miami 33136 FL, USA. 15 Division of Nephrology and Endocrinology, University ofTokyo Graduate School of Medicine, Tokyo 113-8654, Japan. 16 Division of CKD Pathophysiology, University of Tokyo Graduate School of Medicine, Tokyo113-8654, Japan. Correspondence and requests for materials should be addressed to A.F. (email: afornoni@med.miami.edu)NATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 |www.nature.com/naturecommunications 11234567890():,;Sphingolipids and cholesterol are critical components of theeukaryotic plasma membrane (PM), where they serve asmajor mediators of lipid-signaling regulating multiple cel-lular functions1. We have identiﬁed sphingomyelin phospho-diesterase acid-like 3b (SMPDL3b) as a lipid raft enzyme thatregulates integrin activation, cell migration, and cell survival inpodocytes, which are terminally differentiated and specializedcells of the kidney glomerulus2,3. Others have demonstrated animportant role of SMPDL3b in affecting membrane lipid com-position and ﬂuidity in macrophages4. Although the crystalstructure of SMPDL3b was recently reported5, the speciﬁcity of itsphosphodiesterase activity and the possibility of other lipidmodifying functions of SMPDL3b have not been investigated yet.Ceramide is the centerpiece of the sphingolipid metabolicpathway, where it serves as substrate for the generation ofceramide-1-phosphate (C1P) and sphingosine 1-phosphate (S1P)6. C1P and S1P are bioactive sphingolipids regulating importantprocesses, such as cell survival and inﬂammatory responses. S1Pcan be cleaved by S1P lyase, and S1P excess due to a geneticdeﬁciency of S1P lyase was recently found to be associated withadrenal insufﬁciency, nephrotic syndrome, and ichthyosis7–9. Incontrast, while reduction of biologically active C1P has beenlinked to apoptosis and altered inﬂammatory signaling, little isknown about the metabolic or signaling pathways that are con-trolled by C1P and that regulate C1P content.The overall goal of this study is to determine if SMPDL3b mayaffect the availability of biologically active lipids and alter theassembly of protein complexes in lipid raft domains thus inter-fering with physiological signaling. Among signaling complexeslocated in raft domains that are heavily dependent on the PMlipid composition, we are focused on the insulin receptor (IR)-signaling complex for several reasons: (1) IR signaling is a keymodulator of podocyte function10–12; (2) SMPDL3b is stronglyupregulated in glomeruli of patients with insulin resistance anddiabetic kidney disease (DKD)3; (3) IR exists in two isoformswhich are characterized by distinct afﬁnities for lipid raftdomains13 and which are differentially expressed in various celltypes suggesting distinct functions. We therefore tested thehypothesis that SMPDL3b affects the generation of sphingolipidsinvolved in the regulation of IR signaling.Terminally differentiated human podocytes are utilized as amodel system to study in vitro whether SMPDL3b differentiallyaffects IR isoform A (IRA) or IR isoform B (IRB) complex for-mation with caveolin-1 and IR-dependent signaling. A diabeticdb/db mouse model with podocyte-speciﬁc deletion of Smpdl3b isused to determine if SMPDL3b deﬁciency protects from experi-mental DKD. Here we report and discuss that increased expres-sion of SMPDL3b is associated with deﬁciency of biologicallyactive C1P in vitro and in vivo. Exogenous C1P supplementationis sufﬁcient to restore insulin signaling in SMPDL3b over-expressing (SMP OE) cells in vitro and to improve proteinuria inDKD in vivo. Our ﬁndings shed some light on the complexity ofIR signaling, demonstrate an important role of SMPDL3b in themodulation of insulin signaling, and provide the ﬁrst evidencethat biologically active lipids, such as C1P may represent treat-ment options for complications of diabetes associated with highcardiovascular morbidity and mortality such as DKD.ResultsSMPDL3b affects the levels of C1P. We previously describedSMPDL3b as an enzyme expressed in podocyte lipid rafts thatmight play an important role in the pathogenesis of proteinurickidney diseases2,3. Others showed that SMPDL3b functions as aphosphodiesterase with lipid-modifying properties that negativelyregulates immunity in macrophages4. We ﬁrst aimed atdetermining if SMPDL3b also acts as a phosphodiesterase inhuman podocytes and found that overexpression of SMPDL3b inpodocytes (SMP OE) is associated with increased cellular phos-phodiesterase activity, while knockdown of SMPDL3b (siSMP)resulted in decreased phosphodiesterase activity (Fig. 1a). Inaddition to this, the degree of SMPDL3b expression affected thecontent of active sphingolipids in podocytes. While the totalsphingomyelin (Fig. 1b) and total ceramide (Fig. 1c) levelsremained unchanged, total C1P levels were signiﬁcantlydecreased in SMP OE podocytes and increased in siSMP podo-cytes (Fig. 1d) compared to control podocytes (CTRL). Todetermine which ceramide species mostly contribute to theobserved changes in C1P levels, liquid chromatography massspectrometry (LC–MS) was performed. While human podocytesmostly express C16:0, C22:0, C24:0, and C24:1 ceramide andC16:0, C18:0, and C24:0 C1P species, only C16:0 C1P levels weresigniﬁcantly reduced in SMP OE and increased in siSMP podo-cytes (Supplementary Data 1).In addition to its phosphodiesterase activity, protein sequenceanalysis of SMPDL3b using InterPro protein sequencing analysistool, predicted the presence of a phosphatase-like domain. Inorder to understand if SMPDL3b may contribute to the decreasedC1P content by converting C1P into ceramide, we utilized anin vitro assay in HEK293 cells transfected with full-length humanSMPDL3b and found a signiﬁcantly higher conversion of C1P toceramide upon SMPDL3b overexpression (Fig. 1e). However,in vitro enzyme assay determination of SMPDL3b activity withpure recombinant protein failed to detect phosphatase activity(Supplementary Fig. 1). While this could be due to the utilizationof a soluble non-GPI-anchored protein or to the absence of anessential cofactor, we can only conclude that SMPDL3bcontributes to the conversion of C1P into ceramide.SMPDL3b affects IR signaling. We previously observed thatSMPDL3b is strongly upregulated in glomeruli of patients withinsulin resistance and DKD3. This observation, together with theﬁnding that preservation of insulin signaling is a key to protectpodocytes from injury10,12,14,15, prompted us to determine ifSMPDL3b may affect insulin signaling pathways. mRNAsequence analysis in control and SMP OE podocytes was per-formed. We found differential expression of 1948 genesthat passed FDR correction in SMP OE podocytes compared tocontrol (Supplementary Data 2). Using DAVID pathway-enrichment analysis, we found that 62 signaling pathways areaffected in SMPDL3b overexpression cells (SupplementaryData 3). Among them, differential expression of genes involved inthe regulation of phosphoinositide 3-kinase and protein kinase B(PI3K-AKT), sphingolipid signaling, mammalian target of rapa-mycin (mTOR), and insulin resistance pathways was observed(Supplementary Data 3). In addition, similar to what wasdescribed in macrophages4, differential regulation of genesimportant in innate immunity, involved in the tumor necrosisfactor (TNF), nuclear factor kappa-light-chain-enhancer of acti-vated B-cells (NFκB) and toll-like receptors (TLRs)-signalingpathways, was observed (Supplementary Data 3).We ﬁrst validated mRNAseq ﬁndings at the protein level,demonstrating that SMP OE cells, unlike siSMP or CTRLpodocytes, are unable to phosphorylate protein kinase B (AKT),but become responsive to phosphorylate p70S6 kinase (p70S6K)(Fig. 2a, b) after stimulation with insulin. SMP OE podocytes alsodemonstrated increased phosphorylation of eukaryotic transla-tion initiation factor 4E-binding protein-1 (4EB-P1), a down-stream target of mTOR signaling, while the phosphorylation of4EB-P1 was suppressed in siSMP and CTRL podocytes (Fig. 2a,b). Additionally, kinetics of phosphorylation of AKT, p70S6K, orARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-42 NATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 | www.nature.com/naturecommunications4EB-P1 in response to insulin stimulation demonstrated similardynamics between CTRL and siSMP podocytes, while SMP OEpodocytes showed no changes in AKT phosphorylation over time,but increased phosphorylation of p70S6K and 4EB-P1 (Fig. 2c).We therefore focused on SMP OE podocytes in subsequentexperiments. The modulation of IR signaling by SMPDL3b inpodocytes was not linked to a change in IR expression (Fig. 2d).However, subcellular fractionation experiments demonstrated adecreased presence of IR at the PM in SMP OE podocytescompared to control podocytes (Fig. 2e), which could at leastpartially explain why insulin-stimulated AKT phosphorylation isimpaired. Further experiments using ﬂow cytometry assayvalidated our ﬁndings and showed that SMP OE podocytes haveless IR-positive cells (Fig. 2f). The displacement of IR from thePM was independent of caveolin-1, a lipid raft-associated proteinrequired for proper insulin signaling14–16, as phosphorylated andtotal amounts of caveolin-1 in whole cell lysates (SupplementaryFig. 2a) and in the PM fraction (Supplementary Fig. 2b) remainedunchanged. The opposite regulation of AKT and p70S6Kphosphorylation after insulin stimulation observed in SMP OEpodocytes prompted us to investigate if SMPDL3b maycontribute to the functional diversiﬁcation of IR isoformssignaling, as demonstrated in pancreatic beta cells, where IRA,in which exon 11 is excluded, is a modulator of p70S6Kphosphorylation while IRB is a regulator of AKT phosphoryla-tion13. More detailed analysis of IR expression revealed thathuman podocytes express both IR isoforms17, with IRB being thedominantly expressed isoform in both SMP OE and controlpodocytes (Fig. 2g). As tools to experimentally distinguishbetween endogenous IRA and IRB protein expression are notavailable, additional experiments in transfected HEK293 cellswere performed to determine whether SMPDL3b differentiallyinterferes with IRA/IRB signaling.SMPDL3b competes with binding of IRA and IRB to caveolin-1. Given that we found a differential regulation of the insulin-stimulated PI3K-AKT and mTOR-signaling pathways in SMP OEpodocytes, we tested the hypothesis that SMPDL3b over-expression differentially affects IRA and IRB signaling by inter-fering with the formation of IRA/caveolin-1 or IRB/caveolin-1complexes. Immunoprecipitation (IP) experiments of exogenousproteins co-expressed in HEK293 were performed. We found thatSMPDL3b immunoprecipitates both IR isoforms (Fig. 3a) andcaveolin-1, and the ability to form complexes with IR isoformsand caveolin-1 was not affected by the utilization of theSMPDL3b phosphodiesterase mutant SMPDL3b-H135A(Fig. 3b). Considering that SMPDL3b interacts with IRA, IRB,and caveolin-1, we investigated if SMPDL3b competes with IRisoforms to bind caveolin-1 by competitive IP studies in trans-fected HEK293 cells. We found that SMPDL3b facilitates theinteraction of caveolin-1 and IRA (Fig. 3c, left panel) whileinterfering with the ability of IRB to bind caveolin-1 (Fig. 3c, rightpanel), even when expressed at low levels. To conﬁrm theb cae0 5 10 15 200.00.51.01.5C6-NBD-C1P (μM)C6-NBD-Cer(pmol/min/μg)*******pCMV6SMPDL3b200250300350400450500Total sphingomyelin(pmol ×106 cells)0200400600800Total ceramide(pmol ×106 cells)0102030*Total C1P(pmol ×106 cells)**dCTRLsiSMPSMP OECTRLsiSMPSMP OECTRLsiSMPSMP OECTRLsiSMPSMP OE0.000.050.100.15 *PDE activity (Abs 405 nm) **Fig. 1 SMPDL3b affects the conversion of C1P to ceramide. a Bar graph analysis of phosphodiesterase (PDE) activity in control (CTRL), SMPDL3bknockdown (siSMP) and SMPDL3b overexpressing (SMP OE) human podocytes. n= 3 independent experiments; *P= 0.039, two-tailed t-test. b Liquidchromatography mass spectrometry analysis (LC–MS) of total sphingomyelin content in CTRL, siSMP, and SMP OE human podocytes. n= 3 independentexperiments. P= 0.78, F= 0.26, one-way ANOVA. c LC–MS of total ceramide content in CTRL, siSMP, and SMP OE human podocytes. n= 3 independentexperiments. P= 0.78, F= 0.78, one-way ANOVA. d LC–MS of total ceramide-1-phosphate (C1P) content in CTRL, siSMP, and SMP OE human podocytes.n= 3 independent experiments. *P= 0.045, **P= 0.007, F= 32.58, one-way ANOVA. e C1P phosphatase in vitro assay using increasing amountsC6-NBD-C1P. n= 2 independent experiments in duplicate; *P < 0.05, ***P < 0.001; two-tailed t-test. Error bars represent standard deviation (SD)NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-4 ARTICLENATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 |www.nature.com/naturecommunications 3interaction of SMPDL3b with IRA and IRB ex vivo, endogenousIP experiments using pooled mouse glomeruli were performedand showed that SMPDL3b precipitates both IR isoforms andcaveolin-1 (Fig. 3d). While in the absence of SMPDL3b, caveolin-1 preferentially binds IRB, the presence of SMPDL3b interfereswith the binding of IRB to caveolin-1 and facilitates binding ofIRA to caveolin-1, which is consistent with the observation thatinsulin-stimulated AKT phosphorylation is impaired whilep70S6K phosphorylation is augmented in SMP OE podocytes(Fig. 2a, b). As binding of IR to caveolin-1 is important to pro-mote insulin signaling14,18–20, our data suggest that SMPDL3boverexpression may shift IR signaling from IRB-mediated to IRA-mediated signaling. Next, we tested whether methyl-β-cyclodextrin (CD), which was found to act as a chemicalcaveolin-1 inhibitor21, can restore IRB signaling in SMPDL3b-overexpressing cells. We found that CD increased AKT phos-phorylation (Fig. 3e) without affecting p70S6K phosphorylation(Fig. 3e). Additionally, CD treatment of transfected HEK293 cellsresulted in restored interaction between IRB and caveolin-1 inpresence of SMPDL3b overexpression, while IRA/caveolin-1interaction was compromised (Fig. 3f). This occurred in asso-ciation with increased IR and decreased SMPDL3b localization atthe PM (Fig. 3g). While it is not possible to discriminate with thismodel if CD differentially affects IRA or IRB localization, theobserved increase in AKT phosphorylation in the absence ofchanges in p70S6K phosphorylation suggests that CD primarilyrestores the PM localization of IRB.Podocyte-speciﬁc Smpdl3b-deﬁcient mice have normal kidneys.The generation of Smpdl3b ﬂoxed mice is described in detail inapAKT(Ser473)CTRLtAKTGAPDHIns, nM60 -60 -37 -0 0.1 1 0 0.1 1 0 0.1 1siSMP SMP OEp-p70S6K(Thr389)70 -t-p70S6K70 -p-4E-BP1(Thr37/46)20 -t-4E-BP120 -b0.00.51.01.52.02.5Fold change (pAKT/tAKT)0 0.1 1 0 0.1 1 0 0.1 1 Ins, nM**###CTRL siSMP SMP OE CTRL siSMP SMP OE CTRL siSMP SMP OE01234Fold change(p-p70S6K/t-p70S6K)0 0.1 1 0 0.1 1 0 0.1 1 Ins, nM*******012345Fold change(p4E-BP1/t4E-BP1)0 0.1 1 0 0.1 1 0 0.1 1CTRLsiSMPSMP OEIns, nM$$$$$$c0 5 10 15 20 25 30 350100200300Time, min% phosphorylationAKT (Ser473) ****** ********0 5 10 15 20 25 30 350100200300Time, min% phosphorylationp70S6K (Thr389)############0 5 10 15 20 25 30 350100200300Time, min% phosphorylation4EBP1 (Thr37/46)CTRLsiSMPSMP OE$$$$$$$$$$$$$$dCTRL SMP OE0.91.01.11.2Fold change (IR/GAPDH)CTRLSMP OEIRGAPDH95 -37 -eCTRL SMP OE0.00.51.01.5Fold change(IR/Na-K-ATPase)***Na/K-ATPaseCTRLSMP OECTRLSMP OEIRMEK-1/2PM CYTO9510045fCount094188281375Count073145218291Count0265177102Count04897145194CTRL SMP OE Positive ctrl Negative ctrlCountIRβCTRL SMP OE051015IRβ (%)***gCTRLSMP OEIRB 222 bpIRA 186 bp200 -100 ----ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-44 NATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 | www.nature.com/naturecommunicationsthe “Methods” section. To generate podocyte-speciﬁc Smpdl3b-deﬁcient mice, Smpdl3b-ﬂoxed mice were mated with Podocin-Cre mice22. A schematic representation of the Smpdl3b wildtypeallele and targeted allele is shown in Fig. 4a. Genotyping wasperformed by PCR on genomic DNA isolated from tail biopsies.Sequences of primers used for genotyping are indicated in Sup-plementary Table 1. Two primer sets were used, one speciﬁc forthe Podocin-Cre transgene, yielding a 455 bp PCR product andanother set to detect either the Smpdl3b ﬂoxed (237 bp PCRproduct) or wildtype allele (315 bp PCR product) (SupplementaryFig. 3a). To determine if successful Cre-mediated recombinationhas occurred, PCR on genomic DNA isolated from glomeruli wasalso performed using the primers indicated in SupplementaryTable 1. This PCR will only yield a product (198 bp) if Cre-mediated recombination has occurred (Supplementary Fig. 3b).However, a 1075 bp band was also detected which could beexplained by the presence of mesangial and endothelial cells inglomerular extracts in which Cre-mediated recombination didnot occur. Nevertheless, as expected, the 198 bp PCR product wasonly detected when DNA isolated from glomeruli of heterozygous(Cre+;ﬂ/+) and homozygous (Cre+;ﬂ/ﬂ) mice was used but notin wildtype mice. We then performed qRT-PCR to determineSmpdl3b mRNA expression levels in glomeruli isolated fromwildtype and podocyte-speciﬁc Smpdl3b-deﬁcient mice anddemonstrated a 57% reduction of Smpdl3b expression in Cre+;ﬂ/ﬂ mice (Fig. 4b). As expected, Smpdl3b expression levels intubules remained unchanged (Fig. 4c). Similarly, SMPDL3bprotein expression in glomeruli isolated from wildtype (Cre+;+/+) and podocyte-speciﬁc Smpdl3b-deﬁcient mice showed a55% reduction in Cre+;ﬂ/ﬂ mice (Fig. 4d) when compared towildtype littermates. Finally, SMPDL3b expression levels inpodocyte-speciﬁc Smpdl3b-deﬁcient mice were also found to bereduced in immunoﬂuorescence staining for SMPDL3b. Synap-topodin was used as a podocyte-speciﬁc marker (Fig. 4e).Podocyte-speciﬁc Smpdl3b-deﬁcient mice are viable andfertile. At the age of 28 weeks Smpdl3b-deﬁcient micedemonstrated no changes in body weight (Fig. 4f), kidney tobody weight ratio (Fig. 4g), urine albumin to creatinine ratio(Fig. 4h), blood urine nitrogen (Fig. 4i), serum creatininelevels (Fig. 4j), or mesangial expansion score (Fig. 4k) whencompared to wildtype mice. Suppression of SMPDL3bexpression in podocyte-speciﬁc Smpdl3b deﬁcient mice wasnot sufﬁcient to lead to a change in sphingolipid compositionin kidney cortexes, suggesting that other cell types maycompensate for SMPDL3b deﬁciency in podocytes (Supple-mentary Fig. 3c–e; Supplementary Data 4).Smpdl3b-deﬁcient mice are protected from DKD. As podocyte-speciﬁc Smpdl3b-deﬁcient mice are phenotypically normal, wenext tested if podocyte-speciﬁc deﬁciency of Smpdl3b is sufﬁcientto protect diabetic mice from DKD. To test this hypothesis,double knockout mice carrying a mutation in leptin receptor (db/db) and a podocyte-speciﬁc Smpdl3b deﬁciency (ﬂ/ﬂ) were gen-erated (Supplementary Fig. 4a). For genotyping, genomic DNAisolated from tail biopsies and PCR primers for genotyping asindicated in Supplementary Table 1 were used.Four groups of mice were utilized: (1) wildtype control mice(+/+;+/+), n= 10; (2) podocyte-speciﬁc Smpdl3b deﬁcient mice(ﬂ/ﬂ;+/+), n= 11; (3) diabetic control mice (+/+;db/db), n= 9;and (4) podocyte-speciﬁc Smpdl3b-deﬁcient diabetic mice (ﬂ/ﬂ;db/db), n= 9. No changes were found in body weight (Fig. 5a) orkidney to body weight ratio (Fig. 5b) between +/+;db/db and ﬂ/ﬂ;db/db mice. At 28 weeks of age, ﬂ/ﬂ;db/db mice showedsigniﬁcantly lower urinary albumin to creatinine ratios (Fig. 5c)and mesangial expansion (Fig. 5d) compared to +/+;db/db mice,suggesting a protective effect of the podocyte-speciﬁc deletion ofSmpdl3b in DKD progression in this model. We did not detectany changes in BUN (two-tail t-test; P= 0.114), serum creatinine(two-tail t-test; P= 0.466), glycemia (two-tail t-test; P= 0.067), orliver function (ALT and AST, two-tail t-test; P= 0.634 and P=0.670, respectively) between double knockout mice and diabeticmice (Supplementary Table 2), indicating that podocyte-speciﬁcSmpdl3b deﬁciency does not lead to any renal or hepatic toxicity.Additionally, no changes were found in total cholesterol ortriglyceride levels between +/+;db/db and ﬂ/ﬂ;db/db mice(Supplementary Table 2). Because our in vitro data suggestedan important role of SMPDL3b in the generation of ceramidefrom C1P, detailed lipidomic analysis of sphingolipids in kidneycortexes of the double knockout mice was performed usingLC–MS analysis (Supplementary Data 5). Similar to what wefound in SMP OE podocytes (Fig. 1), kidney cortexes from ﬂ/ﬂ;db/db mice demonstrated no changes in total ceramide (Fig. 5e)or total sphingomyelin (Fig. 5f) in association with a decreasedC1P content (Fig. 5g). Podocyte-speciﬁc deletion of Smpdl3b indiabetic mice was sufﬁcient to restore normal C1P levels inkidney cortex (Fig. 5g). These results support an important role ofSMPDL3b in vivo as an active lipid-modifying enzyme andoutline the importance of podocyte SMPDL3b expression in thepathogenesis of DKD.Additionally, podocyte density and apoptosis were assessed byimmunoﬂuorescent labeling using an anti-Wilms’ Tumor 1(WT1) antibody. WT1-positive nuclei were counted in 20consecutive glomerular cross-sections per animal as reported11.Fig. 2 SMPDL3b regulates the insulin-signaling pathway in human podocytes. a Representative Western blot image of main downstream targets of theinsulin receptor signaling in control (CTRL), SMPDL3b knockdown (siSMP) and SMPDL3b overexpressing (SMP OE) podocytes exposed to increasingconcentration of insulin (0, 0.1, 1 nM): phosphorylated (pAKT) and total (tAKT) AKT; phosphorylated (p-p70S6K) and total (t-p70S6K) p70S6K;phosphorylated (p-4EB-P1) and total (t-4EB-P1) 4EB-P1. b Bar graph analysis of the fold change ratio of pAKT to tAKT (left panel), p-p70S6K to t-p70S6K(central panel) and p-4EB-P1 to t-4EB-P1 (right panel) in CTRL, siSMP, and SMP OE podocytes. n= 3 independent experiments; *P < 0.05 (when comparedAKT phosphorylation level in insulin treated CTRL cells); ##P= 0.002, #P= 0.011, F= 5.67 (when compared AKT phosphorylation level in insulin-treatedsiSMP cells); ***P= 0.0002, ****P < 0.0001, F= 21.47 (when compared p70S6K phosphorylation level in insulin-treated SMP OE cells); $$P= 0.008, $$$$P < 0.0001, F= 10.09 (when compared 4EB-P1 phosphorylation level in insulin-treated SMP OE cells); one-way ANOVA. n= 3 independent experiments.c Kinetics of AKT (left panel), p70S6K (central panel), and 4EB-P1 (right panel) phosphorylation in CTRL, siSMP, and SMP OE podocytes treated with 1 nMof insulin. **P= 0.001, F= 9.31; ##P= 0.0003, F= 12.27; $$P= 0.003, F= 7.84, one-way ANOVA. n= 3. d Western blot and bar graph analysis of insulinreceptor (IR) expression in total lysates of CTRL and SMP OE podocytes; P= 0.448, two-tailed t-test. n= 4 independent experiments. e Western blot andgraph analysis of IR localization at the plasma membrane (PM) in CTRL and SMP OE podocytes. Na/K-ATPase was used as a marker for the PM fractionand MEK-1/2 was used as a marker of the cytosolic fraction (CYTO); ***P= 0.0004, two-tailed t-test. n= 5 independent experiments. f FACS and relatedquantiﬁcation analysis of the insulin receptor expression in CTRL and SMP OE podocytes. Unstained cells were used as a negative control; HEK293 cellstransfected with an insulin receptor overexpressing construct were served as a positive control. A total of 30,000 cells per sample was analyzed; ***P <0.0001, two-tailed t-test. n= 4 independent experiments. g PCR analysis of the insulin receptor isoform A (IRA) and isoform B (IRB) gene expression inCTRL, and SMP OE podocytes. Error bars represent standard deviation (SD)NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-4 ARTICLENATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 |www.nature.com/naturecommunications 5ﬂ/ﬂ;db/db mice demonstrated preserved podocyte number perglomerulus when compared to +/+;db/db mice (Fig. 5h) furthersupporting a protective effect of decreased podocyte SMPDL3bexpression in DKD. Moreover, electron microscopy analysisrevealed signiﬁcant improvement of podocyte foot processeffacement and reduced glomerular basement membrane(GBM) thickness in ﬂ/ﬂ;db/db mice compared to +/+;db/dbmice (Fig. 5i) with a tendency to a decreased foot processes width(Supplementary Fig. 4b). Immunohistochemistry analysis showedthat ﬂ/ﬂ;db/db glomeruli were characterized by increasedphosphorylated AKT compared to +/+;db/db mice (Fig. 5j).These results prompted us to further investigate the link betweenpodocyte-speciﬁc SMPDL3b deﬁciency, C1P availability andinsulin signaling in the protection from DKD.c– + ++ +++ –+ + + + –+ + + + +– – – – +hSMPDL3b-GFPhIRA-FLAGhCav1-GFPEmpty-FLAGInputIPIB: Cav1IB: IRIB: Cav1IB: IRIB: SMPDL3b100497510049dhSMPDL3b-GFPhIRB-FLAGhCav1-GFPEmpty-FLAGIB: Cav1IB: IRIB: Cav1IB: IRIPIB: SMPDL3bInputIP: anti-FLAG100497510049– + ++ +++ –+ + + + –+ + + + +– – – – + IP: anti-IRIB: IRIP: anti-IRIB: SMPDL3bInputEn-IPIgGIP: anti-IRIB: Cav15015010020e– + CDSMP OE***++ Ins– +++pAKT(Ser473)tAKTCD, 5 mMCTRL SMP OE–+ Ins, 1 nM++–+++6060p-p70S6K(Thr389)t-p70S6KGAPDH707037f– – – – – + + + + ++ – – – – + – – – –– + – + – – + – + –+ + + + + + + + + ++ + + – – + + + – –– – + – + – – + – +hCav1-GFPhSMPDL3b-HAhIRA-FLAGhIRB-FLAGEmpty-FLAGCD, 5 mMIB: FLAG100IB: GFP49IB: SMPDL3b55InputIB: FLAGIB: GFP10049IPIP: anti-FLAGgMEK-1/2SMP OECTRL– + – + CD, 5 mMIRSMPDL3bNa/K-ATPaseGAPDHPMSMP OECTRL– + – +CYTO75501001005037– + – + CD*$ #a– – +– + –+ + ++ – – Empty-FLAGhIRB-FLAGhIRA-FLAGhSMPDL3b-GFPIB: SMPDL3bIB: SMPDL3bInputIPIB: FLAGIB: FLAG1007510075b+ + – – – – – –– – + + – – – –– – – – + + – –+ – + – + – + –– + – + – + – +– – – – – – + +hCav1-FLAGhSMPDL3b-HAhSMPDL3b-H135A-HAhIRA-FLAGhIRB-FLAGEmpty-FLAGIP: anti-FLAGIB: IRIB: Cav1IB: SMPDL3bInputIB: IRIB: Cav1IB: SMPDL3b10023551002355IPIP: anti-FLAGIP: anti-FLAGFold change (pAkt/tAkt) 2.01.51.00.50.0CTRL– + CDSMP OE++ Ins– +++CTRLFold change(p-p70S6K/t-p70S6K)2.52.01.51.00.50.0Fold change(IR/Na-K-ATPase)***2.01.51.00.50.02.52.01.51.00.50.0CTRL SMP OE– + – + CDCTRL SMP OEFold change(SMPDL3b/Na-K-ATPase)ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-46 NATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 | www.nature.com/naturecommunicationsC1P supplementation restores IR signaling and protects fromDKD. Given our observation that podocyte-speciﬁc Smpdl3b-deﬁcient diabetic mice are protected from experimental DKD,and that kidney cortex from db/db mice are characterized bydecreased C1P content (Fig. 5g), we hypothesized that exo-genous supplementation of C1P may protect from albuminuriaand podocyte injury in DKD. We therefore tested if exogenousC1P supplementation is sufﬁcient to restore podocyte insulinsignaling in vitro and to protect from DKD in vivo. Control andSMP OE podocytes were pre-treated with 100 µM of syntheticC1P C16:0 for 1 h and then stimulated with 1 nM of humaninsulin for 30 min. We found that treatment of SMP OEpodocytes with C1P is sufﬁcient to restore AKT phosphoryla-tion (Ser473) in response to insulin stimulation (Fig. 6a), whileno changes in phosphorylation of p70S6K were observed(Fig. 6b). Mass spectrometry analysis of ceramide species inkidney cortex of db/db mice demonstrated that db/db miceproduce signiﬁcantly less C16:0 ceramide compared to het-erozygous db/+ littermates (Fig. 6c), similarly to what weshowed in SMP OE podocytes (Supplementary Data 1). Todetermine whether exogenous C1P administration amelioratesfeatures of DKD in these mice, 12-week-old db/db mice withestablished albuminuria were injected intraperitoneally with 30mg/kg C16:0 C1P (n= 6) or 0.9% normal saline (n= 6) dailyfor 28 days. Heterozygous (db/+) littermates injected with 0.9%normal saline served as control (n= 6). C1P administration todiabetic db/db mice with established DKD resulted in sig-niﬁcant improvement of albuminuria when compared to salineinjected db/db control mice (Fig. 6d). Additionally, db/db micetreated with C1P demonstrated signiﬁcantly less mesangialexpansion (Fig. 6e). C1P treatment of db/db mice did notnegatively affect kidney function (BUN and serum creatinine),liver function (serum ALT, AST), or serum lipids (triglycerideand cholesterol) (Supplementary Table 3). Additionally, nochanges in body weight (Supplementary Fig. 5a) or glycemia(Supplementary Fig. 5b) were detected. Pathological analysis atthe C1P injection site by standard immunohistochemistry didnot reveal any detectable microscopic change (SupplementaryFig. 5c). In addition, C1P treatment of db/db mice resulted inincreased podocyte AKT phosphorylation when compared todb/db normal saline-injected mice (Fig. 6f). To investigate ifobserved improvement of the kidney function with exogenousC1P administration in db/db mice is related to the improve-ment of glycemic control, we performed intraperitoneal glucosetolerance test. No changes were found in fasting glycemia(Fig. 6g, left panel) or intraperitoneal glucose tolerance test(Fig. 6g, right panel) between saline injected and C1P-treateddb/db mice.DiscussionOur study identiﬁes SMPDL3b as a sphingolipid-related enzymethat inﬂuences the content of active lipids and acts as a masterregulator of insulin signaling in lipid raft domains of the PM. Ourﬁndings suggest a model by which SMPDL3b may affect podocytefunction and survival, where excessive SMPDL3b expression maycause the displacement of IR from caveolin-1-rich domain in aC1P-dependent manner, resulting in impaired ability to phos-phorylate AKT thus promoting podocyte injury in vitro anddevelopment of DKD in vivo.We ﬁrst conﬁrmed the presence of increased phosphodiester-ase activity in SMP OE podocytes when compared to wildtype,which was expected given the homology of SMPDL3b with acidsphingomyelinase and the recently reported crystallographicstructure5. Unexpectedly, this activity was not associated with areduction of the sphingomyelin content and increased ceramideproduction in podocytes, which prompted us to utilize a large-scale lipidomic approach to identify additional sphingolipidspecies that may be affected by SMPDL3b. Among the differentlipid species, a reduction of C1P species was observed, pre-dominantly C16:0 C1P. Interestingly, the analysis of the aminoacid sequence of SMPDL3b revealed the possible existence of aphosphatase domain. We therefore investigated if SMPDL3b maybe involved in the conversion of synthetic C1P to ceramide.While experiments in transfected HEK293 cells strongly sug-gested that SMPDL3b may convert C1P into ceramide, utilizationof puriﬁed soluble recombinant human SMPDL3b did not con-ﬁrm a direct enzymatic activity of SMPDL3b, consistent with thefact that other GPI-anchored enzymes may have impaired activitywhen lacking their GPI anchor23. These data, in conjunction withour prior report that SMPDL3b overexpression is also associatedwith increased sphingosine-1-phosphate (S1P) levels24, suggestthat SMPDL3b regulates the content of active sphingolipids, suchas S1P and C1P. In nephrology, S1P has been studied extensivelyin the context of acute kidney injury25 and has recently generatedattention when a mutation of the gene encoding for S1P lyase,resulting in accumulation of S1P, was found to be associated withnephrotic syndrome9, similar to what was described in S1P lyase-deﬁcient mice7. On the contrary, much less is known about C1P,as its pro-inﬂammatory or anti-inﬂammatory role is still a matterof debate6,26–28. C1P has been shown to modulate AKT phos-phorylation in skin ﬁbroblasts, hematopoietic cells29, macro-phages30, adipocytes31, consistent with the observation thatseveral sphingolipids are major modulators of insulinsignaling32,33. As ceramide has been shown to be a master reg-ulator of insulin signaling and the primary antagonist of proteinkinase B (AKT) in culturing myotubes and liver34–36, the fact thatSMPDL3b affects overall C1P content without affecting ceramideFig. 3 SMPDL3b interacts with IRA, IRB, and caveolin-1 in a competitive manner. a–c Immunoprecipitation (IP) experiments performed in HEK293 cells co-transfected with human hIRA-FLAG, hIRB-FLAG, hSMPDL3b-GFP, hSMPDL3b-HA hSMPDL3b-H135A-HA, hCav1-FLAG, and hCav1-GFP plasmids. EmptyFLAG vector (Empty-FLAG) was used as a negative control. Antibodies (IB) against FLAG, SMPDL3b, insulin receptor β-subunit (IR), or caveolin-1 (Cav1)were used in Western blot analysis. Each IP experiment was repeated at least three times. a SMPDL3b interacts with both IR isoforms but preferentiallywith IRA. bWildtype SMPDL3b interacts with IRA, IRB, and caveolin-1, a mutation (H135A) in the phosphodiesterase activity domain of SMPDL3b does notaffect these interactions. c Overexpression of SMPDL3b enhances the interaction between caveolin-1 and IRA (left panel) and suppresses the interactionbetween caveolin-1 and IRB (right panel). d Endogenous IP experiments using glomeruli isolated from ﬁve C57BL/6 mice indicate that SMPDL3bimmunoprecipitates insulin receptor and caveolin-1. IgG served as a negative control. Each IP was repeated three times. e Western blot and bar graphanalysis of AKT and p70S6K phosphorylation in control (CTRL) and SMPDL3b overexpressing (SMP OE) podocytes pre-treated with 5mM methyl-β-cyclodextrin (CD) and stimulated with 1 nM insulin; *P= 0.02, **P= 0.001, two-tailed t-test. n= 3 independent experiments. f IP experiments performedin HEK293 cells co-transfected with human hIRA-FLAG, hIRB-FLAG, hCav1-GFP, and hSMPDL3b-HA show that 5 mM CD treatment abrogates IRA/Cav1interaction and restores IRB/Cav1 interaction in presence of SMPDL3b overexpression. g Western blot and bar graph analysis of the insulin receptor (IR)and SMPDL3b localization at the plasma membrane (PM) in CTRL and SMP OE podocytes exposed to 5mM CD for 1 h; *P= 0.049, ***P= 0.0001 (for IR);$P= 0.024, #P= 0.031 (for SMPDL3b); two-tailed t-test. n= 3 independent experiments. Error bars represent standard deviation (SD)NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-4 ARTICLENATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 |www.nature.com/naturecommunications 7aWT locusTargetedalleleEx15′ 3′Ex2 Ex3 Ex4 Ex5 Ex6 Ex7 Ex8FRT loxP5,441 bp 885 bpFloxedalleleEx15′ 3′Ex2 Ex3 Ex4 Ex5 Ex6 Ex7 Ex8FRTloxPCre-mediatedexcisionEx15′ 3′Ex3 Ex4 Ex5 Ex6 Ex7 Ex8FRT loxPEx15′ 3′Ex2 Ex3 Ex4 Ex5 Ex6 Ex7 Ex8WT-F WT-RFRT-F FRT-RLacZ NeoKO-FbmRNA fold change in glomeruli(Smpdl3b/Podocin)cmRNA fold change in tubules(Smpdl3b/GAPDH)dSMPDL3bGAPDHTubulesSMPDL3bGAPDH507537507537Glomerulie SMPDL3b SYNPO OVERLAY f0g0h+ flippaserecombinase+ Crerecombinase***1.51.00.50.0Cre++/+Cre+fl/flCre++/+Cre+fl/flCre++/+Cre+fl/flCre++/+Cre+fl/flCre++/+Cre+fl/flCre++/+Cre+fl/flCre++/+Cre+fl/flCre++/+Cre+fl/fl1.11.00.90.80.7***1.51.00.50.0Fold change in glomeruli(SMPDL3b/GAPDH)1.31.21.11.00.90.8Fold change in tubules(SMPDL3b/GAPDH)Cre+; +/+Cre+; fl/fl604020Body weight (g)0.0090.0080.0070.0060.0050.004KW/BW ratio (g)300200100ACR at 28 weeks(μg/mg)i030 j k0123Cre++/+Cre+fl/flCre++/+Cre+fl/flCre++/+Cre+fl/flBUN (mg/dl) 20100.25Serum creatinine(mg/dl)0.200.150.100.050.00 Cre+; fl/flCre+; +/+Mesangial expansion(score)KO-RFig. 4 Podocyte-speciﬁc Smpdl3b-deﬁcient mice are viable and phenotypically normal. a A schematic representation of the Smpdl3b wildtype allele andtargeted locus. Ex1–8 exons 1–8; primers to detect FRT (green arrowheads), the Smpdl3b ﬂoxed allele (KO, blue arrows) and Smpdl3b wildtype allele (WT,red arrows) were used. b, c RT PCR of Smpdl3b expression in glomeruli (b) and tubules (c) pooled from ﬁve different mice; ***P= 0.003, two-tailed t-test.n= 3 independent experiments. d Western blot (left panel) and bar graph analysis of SMPDL3b expression in glomeruli (middle panel) and tubules (rightpanel) pooled from 2 to 5 mice per group; ***P= 0.003, two-tailed t-test. n= 3 independent experiments. e Immunostaining (40×) of SMPDL3b (green)and synaptopodin (red; SYNPO) expression in Cre+;+/+ and Cre+;ﬂ/ﬂ. Bar scale 30 µm. f–j Phenotypical analysis of Cre+;+/+ (n= 10 animals) and Cre+;ﬂ/ﬂ (n= 6 animals) 28-week-old mice. Cre+;+/+ or Cre+;ﬂ/ﬂ mice showed no changes in body weight (f), P= 0.076, kidney weight to body weight ratio(g), P= 0.715, urine albumin-creatinine ratio (ACR) (h), P= 0.208, blood urine nitrogen (BUN) (i), P= 0.659 and serum creatinine(j), P= 0.352; two-tailed t-test. k PAS staining (20×) of 4 μm kidney sections and bar graph analysis in Cre+;+/+ and Cre+;ﬂ/ﬂ mice. Bar scale 30 µm. P= 0.401, two-tailedt-test. Error bars represent standard deviation (SD)ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-48 NATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 | www.nature.com/naturecommunicationsdoes not exclude the possibility that SMPDL3b inﬂuences theceramide/C1P ratios in PMs of podocytes, thus contributing toinsulin resistance. It is possible that the increased ceramidegeneration in SMPDL3b overexpression cells still plays a key rolein the induction of insulin resistance and that our failure to detecta change in ceramide content is linked to its rapid degradationand/or conversion to other substrates. However, should ceramiderather than C1P deﬁciency be the major driver of insulin resis-tance in our model, then the administration of C1P in diabeteswould result in increased ceramide production and furthera0Body weight (g)** ** ** * ** * **b0 0cd +/+; +/+ e01234Mesangial expansion (score)Total ceramide (pmol/mg protein)f0Total sphingomyelin (pmol/mg protein)g0Total C1P (pmol/mg protein)* h05WT1- stained podocyte numberper glomerulus* *i0246 *0GBM thickness (nm)**j pAKT SYNPO Overlay1006080402010 3020Age (weeks)+/+; +/+fl/fl; +/++/+; db/dbfl/fl; db/db1 × 10–28 × 10–36 × 10–34 × 10–32 × 10–35 × 1044 × 1043 × 1042 × 1041 × 104KW/BW ratio (g)******ACR at week 28 (μg/mg)200015001000500**** ****fl/fl; +/++/+; db/db fl/fl; db/db+/+; +/+ fl/fl; +/++/+; db/db fl/fl; db/db*** ***+/++/+fl/fl+/++/+db/dbfl/fldb/db+/++/+fl/fl+/++/+db/dbfl/fldb/db+/++/+fl/fl+/++/+db/dbfl/fldb/db+/++/++/++/+fl/fl+/++/+db/dbfl/fldb/db+/+db/dbfl/fldb/db+/+db/dbfl/fldb/db+/+db/dbfl/fldb/db+/++/+fl/fl+/++/+db/dbfl/fldb/db+/++/+fl/fl+/++/+db/dbfl/fldb/db+/++/+fl/fl+/++/+db/dbfl/fldb/db40003500300025002000500400300200100**** ****25201510+/+; db/db fl/fl; db/db+/+; +/++/+; db/dbfl/fl; db/db08 × 1046 × 1044 × 1042 × 104pAKT (Ser473)relative fluor (a.u.)Foot processes per μm GBM500400300200100NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-4 ARTICLENATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 |www.nature.com/naturecommunications 9worsening of DKD, and the observed improvement of AKTphosphorylation in vitro and in vivo would remain unexplained.The effect of SMPDL3b on insulin signaling, in association withthe observation that SMPDL3b is a lipid raft-associated enzyme4,prompted us to test the hypothesis that SMPDL3b may interferewith the PM localization and function of IR. We ﬁrst veriﬁed thatgenes involved in the PI3K-AKT signaling pathway were affectedin cells overexpressing SMPDL3b. In SMP OE podocytes, theability of insulin to cause AKT phosphorylation was severelyimpaired, while the ability to phosphorylate p70S6K was aug-mented. These functional data suggested a differential regulationof IR isoform signaling, as it was shown before that in pancreaticbeta cells where IRB signaling is primarily linked to AKT phos-phorylation and dependent on its localization in lipid raftdomains for proper function13. Increased SMPDL3b expressionwas not associated with a change in IR expression but affected IRlocalization to PM. While it is not possible to discriminate if thischange in localization affects IRA or IRB, it seems feasible tosuggest that it primarily affects IRB, which is the predominantlyexpressed isoform in podocytes. We therefore tested if SMPDL3binterferes with the localization of the IR to the PM by interferingwith IR binding to caveolin-1. We utilized HEK293-transfectedcells and found that SMPDL3b can bind to either IRA or IRB orcaveolin-1. Endogenous SMPDL3b binding to IR and caveolin-1was conﬁrmed by IP in glomeruli isolated from kidney cortexes.Binding of SMPDL3b to caveolin-1 or to IR did not require itsphosphodiesterase function, suggesting the existence of anenzymatic regulation of protein complex formation affectinginsulin signaling in lipid raft domains. Since the interactionbetween IR and caveolin-1 has been proven to be essential for IRsignaling37–39, and accumulation of the ganglioside GM3 tocaveolae microdomains causes the dissociation of IR-caveolin-1complexes and increased IR motility in adipocytes15, we deter-mined if SMPDL3b interferes with the ability of IR to interactwith caveolin-1. Indeed, we demonstrated that SMPDL3b caninterfere with the binding of IRB to caveolin-1, while facilitatingIRA binding to caveolin-1, which may be responsible forincreased insulin-stimulated p70S6K phosphorylation. Treatmentwith methyl-beta-cyclodextrin, which was found to act as achemical caveolin-1 inhibitor21, reduced SMPDL3b PM locali-zation and was sufﬁcient to restore IR localization to PM and torestore the ability of insulin to phosphorylate AKT. The ability ofsphingolipids to affect insulin signaling has been studied in sev-eral organs. Sarcolemmal ceramides and sphingomyelins regulateinsulin sensitivity in human skeletal muscles40. In adipocytes, S1Pdeﬁciency can lead to insulin resistance via decreased activity ofthe peroxisome proliferator-activator receptor γ41 or via accu-mulation of GM3, which has been shown to dissociate the IR betasubunit from caveolin-115. Earlier studies also demonstrated thathigh ceramide concentrations are associated with insulin resis-tance in rodent liver, as well as in skeletal muscles from rats andobese patients42,43. Summers et al. (2010) demonstrated thatceramide blocks insulin stimulation of AKT by two independentpathways: by blocking AKT translocation to the PM and bydephosphorylation of AKT by protein phosphatase 2A44. To ourknowledge, however, this is the ﬁrst report of how sphingolipidsmay differentially affect IR isoform signaling by the displacingIRB isoform from caveolin-1 rich domains of the PM.In order to understand the in vivo relevance of SMPDL3b in thecontext of diabetic complications that are linked to impairedinsulin signaling such as DKD, we generated podocyte-speciﬁcSmpdl3b-deﬁcient mice. These mice are phenotypically normal,similar to what we described for SMPDL3b-deﬁcient podocytes2.Generation of db/db mice with a podocyte-speciﬁc deletion ofSmpdl3b resulted in a protection from developing DKD, withreduced albuminuria and mesangial expansion and preservation ofpodocyte numbers. This occurred in association with the restora-tion of the C1P content in kidney cortexes, and restoration ofpodocyte-speciﬁc AKT phosphorylation. C1P treatment was notassociated with renal or liver toxicity or with changes in the con-centration of circulating lipids. In order to understand if theprotection observed in Smpdl3b-deﬁcient mice was linked to thedecreased C1P content, we performed additional in vitro andin vivo experiments with exogenous supplementation of C1P.Similar to what was observed in skin ﬁbroblasts30, exogenous C1Padministration was associated with increased AKT phosphoryla-tion in vitro and with protection from albuminuria and mesangialexpansion in db/db mice in vivo. These ﬁndings open the door forthe use of active lipids as potential new therapeutic agents in theﬁeld of diabetes and associated complications. In fact, modulationof enzymes involved in sphingolipid metabolism, such as acidand neutral ceramidases has been studied as a therapeutic optionin Alzheimer’s disease45,46, cancer research47,48, and diabetes41,46.Interestingly, and similar to what we have observed, an imbalancein C1P and S1P availability is also critical in the progression ofother diseases, such as glioblastoma49, acute lung injury50,51,Parkinson’s disease52, pancreatic cancer cells53, andchronic inﬂammatory processes54,55. Nevertheless, little is knownabout the potential utilization of C1P and S1P analogs as ther-apeutic agents.In conclusion, we demonstrated a mechanism by whichsphingolipids may affect PM signaling (Fig. 6h). We have utilizedIR to demonstrate how active lipid products of SMPDL3b mayinterfere with IR signaling in lipid raft domains and howadministration of deﬁcient lipid species may restore IR signalingand protect from DKD. Our study has limitations, as the tech-nical inability to differentiate endogenous IRA and IRB signalingor to determine the enzymatic activity of recombinant non-GPI-Fig. 5 Podocyte-speciﬁc Smpdl3b-deﬁcient diabetic mice are protected from DKD. Four groups of 28 weeks-old mice were utilized: (1) wildtype mice (+/+;+/+), n= 10 animals; (2) mice with podocyte-speciﬁc Smpdl3b deﬁciency (ﬂ/ﬂ;+/+), n= 11 animals; (3) diabetic mice (+/+;db/db), n= 9 animals; (4)podocyte-speciﬁc Smpdl3b-deﬁcient diabetic mice (ﬂ/ﬂ;db/db), n= 9 animals. a A graph showing body weight changes in all four groups of mice during thestudy. b A bar graph showing kidney to body weight (KW/BW) ratio in all four groups of mice at sacriﬁce time point. ***P < 0.0001, F= 44.79, one-wayANOVA. c A bar graph showing urine albumin to creatinine ratio (ACR) in all four groups of mice at sacriﬁce time point. ****p < 0.0001, F= 42.24, one-way ANOVA. d PAS staining (40×) of 4 μm kidney sections and bar graph analysis in all four groups of mice. Bar scale 30 µm. ****P < 0.001, F= 47.17,one-way ANOVA. e–g LC–MS analysis of total ceramide (e), total sphingomyelin (f) and total ceramide-1-phosphate (g) content in kidney cortexes of allfour groups of mice. P= 0.317, F= 1.38 (total ceramide), P= 0.764, F= 0.388 (total sphingomyelin), one-way ANOVA; *P= 0.023, **P= 0.002 (totalceramide-1-phosphate), two-tailed t-test. n= 3 animals per group. h Representative immunostaining (40×) for anti-Wilms’ Tumor 1 (WT1, green) andDAPI (blue) in 4 µm kidney sections of all four groups of mice. Bar scale 30 µm. *P= 0.016, **P= 0.037, two-tailed t-test. n= 3 animals per group.i Representative TEM and bar graph analysis of foot process effacement (yellow arrow) and glomerular basement membrane (GBM) thickness (yellowasterisk) in +/+;db/db and ﬂ/ﬂ;db/db mice. Bar scale 500 nm. *P= 0.020, **P= 0.001, two-tailed t-test. n= 3 animals per group, n= 10 micrographsanalyzed per mouse. j Representative immunostaining (40×) for pAKT (Ser473) (green) and synaptopodin (red; SYNPO) in 4 µm kidney sections and bargraph analysis demonstrating increased AKT phosphorylation in all four groups of mice. Bar scale 30 µm. **P= 0.001, ****P < 0.0001, F= 37.48, one-wayANOVA. n= 3 animals per group. Error bars represent standard deviation (SD)ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-410 NATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 | www.nature.com/naturecommunicationsanchored SMPDL3b. In addition, while our studies have beenlimited to the kidney, understanding the function of SMPDL3b inother organs and in other disease processes may lead to additionaltherapeutic potential. As pharmacological suppression ofSMPDL3b may render cells susceptible to other forms of injuryand an activation of innate immunity pathways2–4, the alternativeapproach to utilize C1P replacement strategies may result in asafer and more effective strategy to treat DKD. Further studiesaC1PInsp-p70S6K(Thr389)CTRL SMP OE––+––+++––+––+++7070CTRL SMP OEInsC1PpAKT(Ser473)tAKT––+––+++––+––+++6060b– –– –++ ++*******Fold change (pAKT/tAKT)– –– –++ ++*c0Ceramide species(pmol/mg) db/dbd0ACR (μg/mg)db/+ NS db/db NS db/db C1Pe012345Mesangial expansion(score)** *fNSC1PpAKT SYNPO Overlay0pAKT(Ser473)relative fluor (a.u.)*g0AUC of IPGTT (i.u.)*********hSMPDL3bIR/Cav1pAKTPodocyte injuryDKDexC1PSmpdl3b KOCDC1Pt-p70S6K2.01.51.00.50.02.01.51.00.50.0****C1PInsulin– –– –++ ++– –– –++ ++C1PInsulinCTRL SMP OE CTRL SMP OEFold change(p-p70S6K/t-p70S6K)500400400300200 200100C16:1 C16:0 C18:0**** db/m600 **db/dbdb/dbdb/+ NS1.5 × 1055 × 1041 × 105****0Fasting glycemia(mg/dl)500400300200100db/+ NSdb/db NSdb/db C1Pdb/+ NSdb/db NSdb/db C1Pdb/+ NSdb/db NSdb/db C1Pdb/+ NSdb/db NSdb/db C1Pdb/+ NSdb/db NSdb/db C1P80,00060,00040,00020,000NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-4 ARTICLENATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 |www.nature.com/naturecommunications 11investigating C1P administration in vivo will open the door totherapies with bioactive lipids targeting the most prevalent formof kidney disease worldwide.MethodsReagents, antibodies, dilution, and mice. RPMI 1640 media for podocytes cul-ture was obtained from Corning (NY, USA; #10-040-CV). DMEM media with 4.5g/l glucose, L-glutamine and sodium bicarbonate (#11995-073) for HEK293 cellculture was obtained from ThermoFisher Scientiﬁc Inc. (CA, USA). Dynabeads M-270 Epoxy Co-IP Kit (#14321D) was obtained from ThermoFisher Scientiﬁc Inc.Human insulin (#I9278-5ML), CD (#H107-100G), 4-nitrophenol standards(#1048-5G) and bis-p-nitrophenolphosphate (#N3002) for phosphodiesterase assaywere obtained from Sigma-Aldrich (MO, USA). C16:0 C1P (#860533) and C6-NBD-Ceramide (#810209P) were obtained from Avanti Polar Lipid (AL, USA).Recombinant human SMPDL3b protein, MYC/DDK-tagged was ordered fromCreative Biomart (#SMPDL3b-2823H).Anti-caveolin-1 (#3251 for phospho-Cav1(Tyr14), dilution 1:1000, #3267 forCav1 (D46G3) XPTM, dilution 1:1000), anti-IR β subunit (4B8) (#3025, dilution1:100), anti-Na/K-ATPase (#3010, dilution 1:1000), anti-MEK-1/2 (D1A5) (#8727,dilution 1:1,000), anti-AKT (#9271 for phospho-AKT (Ser473) (D9E), dilution1:1000, #9272 for tAKT, dilution 1:1000), anti-p70S6Ke (#9234 for phosphor-p70S6K (Thr389 (108D2)), dilution 1:1000, #9202 for p70S6K, dilution 1:1000),anti-4EBP1 (#2855 for phospho-4E-BP1 (Thr37/46) (236B4), dilution 1:1000,#9644 for 4E-BP1, dilution 1:1000) primary antibodies were purchased from CellSignaling Technology (MA, USA). SMPDL3b antibodies purchased from GenWayBiotech (CA, USA, #GWB-2281D4, dilution 1:1000). IR β subunit antibodies (CT-1) for endogenous IP (#MA5-13778, dilution 1:250) were obtained fromThermoFisher Scientiﬁc Inc., HRP-conjugated secondary antibodies from PromegaCorp. (WI, USA) and used in dilution 1:10,000, anti-rabbit TrueBlot HRP-conjugated secondary antibodies from Rockland (PA, USA, #18-8816-31, dilution1:2,000). For immunoﬂuorescence primary rabbit polyclonal anti-WT1 (WT1; #sc-192, dilution 1:300) and anti-goat polyclonal synaptopodin (P19; #sc-21537,dilution 1:500) antibodies were purchased from Santa Cruz Biotechnology (TX,USA), rabbit polyclonal SMPDL3b antibodies from GenWay (#GWB-2281D4,dilution 1:200) and rabbit monoclonal pAKT (Ser473; #9271, dilution 1:300)antibodies from Cell Signaling Technology. Alexa ﬂuorescence 568 (red; #A10042;#11061, dilution 1:500), 488 (green; # A11059, dilution 1:500) secondary antibodiesand DAPI (#D1306) were obtained from Invitrogen | Thermo Fisher Scientiﬁc(CA, USA). For ﬂow cytometry, human BD Fc Block reagent (#564220) wasobtained from BD Bioscience, Human/Mouse Insulin R/CD220 APC-conjugatedantibodies (#FAB1544A, dilution 1:10) and isotope control IgG APC-conjugatedantibodies (#IC108A, dilution 1:30) were obtained from R&D Systems.Mice in which exon 2 of Smpdl3b is ﬂanked by loxP sites were purchased fromthe International Knockout Mouse Consortium (B6N;B6N-SMPDL3btm1a(EUCOMM)Wtsi/H; #MGI:1916022). Mice carrying a Flp-recombinase transgene(B6.129S4-Gt(ROSA)26Sortm1(FLP1)Dym/RainJ; #009086), mice carrying a Cre-recombinase transgene speciﬁcally expressed in podocytes (B6.Cg-Tg(NPHS2-cre)295Lbh/J; catalog #008205) and Leprdb heterozygous db/+ and homozygous db/dbmice (B6.BKS(D)-Lepr/J; catalog #000697) were purchased from the JacksonLaboratories (USA). 18% protein rodent chow diet was ordered from TeklandGlobal (Envigo, WI, USA; #2018S).Plasmids. A full-length human SMPDL3b with GFP tag (hSMPDL3b-pCMV6-AC-GFP; #RG217688), human caveolin-1 with GFP tag (hCav1-pCMV6-AC-GFP;#RG210274), human caveolin-1 with FLAG tag (hCav1-pCMV6-Myc-DDK;#RC210274), and pCMV6-AC-3DDK empty vector (#PS100057) were ordered fromOriGene (MD, USA). A full-length human IR isoform A with FLAG tag (hIRA-FLAG-pRC/CMV) and a full-length human IR B with FLAG tag (hIRB-FLAG-pRC/CMV) plasmids were a kind gift from our collaborator Dr. Ingo Leibiger13. A full-length human HA-tagged SMPDL3b plasmid (pTO-SP-HA-hSMPDL3b) and HA-tagged SMPDL3b plasmid with mutation in 135 histidine (pTO-SP-HA-hSMPDL3b-H135A) were a kind gift from our collaborator Dr. Leonhard Heinz4.Cell culture. A human podocyte cell lines transfected with a thermosensitive SV40-T construct were cultured on type I collagen-coated ﬂasks in RPMI media sup-plemented with 10% FBS, 1% penicillin/streptomycin, and 1% ITS at 33 °C topromote cells proliferation56. Before thermoswitching to 37 °C to promote cellsdifferentiation, cells were grown to conﬂuence 70–80% and trypsinized andreseeded in fresh ﬂasks at a dilution 200,000 cells per 10 ml of RPMI media withoutITS. The origin of normal human podocytes (control) is a kind gift of Dr. JochenReiser, Rush University, Chicago. SMPDL3b overexpression human podocyteswere developed by Dr. Fornoni and validated in ref. 2.HEK293 cells were obtained from ATCC (Manassas, VA; #CRL-1573TM).HEK293 cells were cultured in DMEM media supplemented with 10% FBS and 1%penicillin/streptomycin at 37 °C. Prior to transfection procedure, cells were grownto conﬂuence 70–80%.Treatment of human podocytes. For treatments, human podocytes were differ-entiated for 14 days and serum starved for 24 h starting day 13. Stimulation withhuman insulin was performed at concentrations of 0.1 and 1 nM for 30 min at 37 °C. For kinetics of the insulin signaling over time, podocytes were treated with 1 nMof insulin for 0, 2, 5, 10, 15, 20, 25, and 30 min. Treatment with CD was performedat a concentration of 5 mM for 1 h at 37 °C. C1P was used for treatment at aconcentration of 100 μm for 1 h at 37 °C. C1P stock solution (ﬁnal concentration1.47 mM) was prepared by sonication of C1P in sterile nanopure water on ice usinga probe sonicator, until a clear dispersion was obtained57. Untreated wildtypepodocytes served as a control.LC–MS analysis. Cell lysates from differentiated control (CTRL) and SMP OEpodocytes or kidney cortexes from mice were used for lipidomics analysis. LC–MSanalysis was performed in Medical University of South Carolina, Lipidomics sharedresource (SC, USA). Cells pellets containing at least 1 × 106 cells per sample in vitroor at least 25 mg of kidney cortex per sample in vivo were subjected to liquidextraction58.Phosphodiesterase enzymatic activity assay. Enzymatic activity in CTRL andSMP OE podocytes was measured according to a previously published protocol4.Brieﬂy, CTRL and SMP OE podocytes were differentiated for 14 days at 37 °C in96-well plates and starved for 24 h. To measure the phosphodiesterase activity,podocytes were incubated with isotonic Tris-buffered saline in the presence of 1mM substrate at 37 °C and 5% CO2 for 1 h. Phosphodiesterase activity in WT andSMP OE podocytes was measured by generation of p-nitrophenol from bis-p-nitrophenolphosphate as absorbance at 405 nm in 100 μl reaction volume.C1P phosphatase assay. C1P phosphatase was assayed in vitro using homo-genates from HEK293 cells transfected with hSMPDL3b-pCMV6-Myc-DDK orpCMV6-AC-3DDK essentially as described59. Brieﬂy, 50 μM C6-NBD-C1P com-plexed to 10 μM BSA was incubated with homogenates (2 μg of protein) in 200 mMNaCl, 2 mM EDTA, 100 mM HEPES, pH 7.4 and a protease inhibitor cocktail(1:200) for 10 min at 37 °C. The reaction was terminated by addition of 3.75volumes of chloroform:methanol (1:2), 1.25 volume of chloroform and 1.25volumes of 2M KCl/0.2 M H3PO4. The lower phase was dried under N2 and lipidssubsequently resolved by TLC on Silica Gel plates using chloroform:methanol:Fig. 6 Exogenous C1P replacement restores the ability of SMPDL3b overexpressing podocytes to phosphorylate AKT and reduces albuminuria in diabeticmice. a, b Western blot and bar graph analysis of phosphorylated AKT (pAKT) to total AKT (tAKT) (a) and phosphorylated p70S6 kinase (p-p70S6K) tototal p70S6 kinase (t-p70S6K) (b) in control (CTRL) and SMPDL3b overexpressing (SMP OE) human podocytes pre-treated with synthetic C1P (100 µM)and treated with insulin (1 nM); *P < 0.05, **P < 0.001, ***P < 0.0001, F= 9.66 (for AKT), F= 3.94 (for p70S6K), one-way ANOVA. n= 3 independentexperiments. c LC–MS analysis of ceramide species in kidney cortexes of control mice (db/+) and diabetic mice (db/db). ****P < 0.0001, F= 27.18, two-way ANOVA. n= 3 animals per group. d For exogenous C1P administration in vivo three groups of 12-week-old mice were utilized: (1) control heterozygousmice intraperitoneally (I.P.) injected with 0.9% normal saline (db/+NS), (2) diabetic mice I.P. injected with 0.9% normal saline (db/db NS), 3) diabeticmice I.P. injected with 30mg/kg C1P for 28 days (db/db C1P). n= 6 animals per group. A bar graph showing albumin–creatinine ratio (ACR) in db/+NS,db/db NS and db/db C1P mice. **P= 0.002, F= 10.1, one-way ANOVA. e Representative PAS staining (20×) in 4 μm kidney sections and bar graphanalysis in db/+NS, db/db NS, and db/db C1P mice. Bar scale 20 µm. *P= 0.043, **P= 0.002, F= 9.47, one-way ANOVA. f Representativeimmunostaining (40×) for pAKT (Ser473) (green) and synaptopodin (red; SYNPO) in 4 µm kidney sections and bar graph analysis in db/+NS, db/db NS,and db/db C1P mice. Bar scale 30 µm. *P= 0.044, ****P= 0.0001, F= 8.03, one-way ANOVA. g Bar graph analysis of fasting glycemia (left panel)and area under curve of glucose tolerance test (right panel) after intraperitoneal injection of 1.5 g/kg D-glucose in db/+NS, db/db NS, and db/db C1P mice.*P= 0.013, **P= 0.008, ***P < 0.0001, F= 9.86 (for fasting glycemia), F= 67.31 (for AUC of IPGTT), one-way ANOVA. Error bars represent standarddeviation (SD). h Proposed mechanism by which increased SMPDL3b expression contributes to the pathogenesis of DKD and possible targeting strategiesto prevent DKD development. (↑) overexpression, (↓) suppression, CD - cyclodextrin, exC1P - exogenous C1PARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-412 NATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 | www.nature.com/naturecommunicationsacetic acid: 15 mM CaCl2 (60:35:2:4, v/v/v/v) as developing solvent. For experi-ments with puriﬁed SMPDL3b, 50 μM C6-NBD-C1P complexed to 10 μM BSA wasincubated with SMPDL3b in concentrations 0, 100, 500, and 1000 ng following thesteps described above.Illumina sequencing RNA data analysis. Preparation and sequencing of RNAlibraries was carried out in the John P. Hussman Institute for Human GenomicsCenter for Genome Technology. Brieﬂy, total RNA was quantiﬁed and qualiﬁedusing the Agilent Bioanalyzer to have an RNA integrity score (RIN) above 7. Foreach of the six samples, 500 ng of total RNA was used as input for the IlluminaTruSeq Stranded Total RNA Library Prep Kit with Ribo-Zero to create ribosomalRNA-depleted sequencing libraries. Each sample had a unique barcode to allow formultiplexing and >35 million raw reads in a 2 × 100 paired end sequencing run onthe Illumina HiSeq3000 were obtained.Raw sequence data were processed by the on-instrument Real Time Analysissoftware (v.2.7.7) to basecall ﬁles. These were converted to de-multiplexed FASTQﬁles with the Illumina supplied scripts in the BCL2FASTQ software (v2.17). Thequality of the reads was determined with FASTQC software for per base sequencequality, duplication rates, and overrepresented k-mers. Illumina adapters weretrimmed from the ends of the reads using Trim Galore! package. Reads werealigned to the human reference genome (hg19) with the STAR aligner (v2.5.0a)60.Gene count quantiﬁcation for total RNA was performed using the GeneCountsfunction within STAR against the GENCODE v19 transcript ﬁle.Gene count data were input into edgeR software61 for differential expressionanalysis. Brieﬂy, gene counts were normalized against total aligned reads for eachsample to generate counts per million (cpm) expression value for each gene in eachsample. Given the relatively small sample size per group for comparison (n= 3 ineach group) the exact test implemented in edgeR was used to determine differentialexpression including a false discovery rate P-value.To discover enriched functional-related gene groups, DAVID bioinformaticsresource 6.8 was used.Isolation of PMs. Cells were collected in homogenization media (15 mM KCl, 1.5mM MgCl2, 10 mM HEPES, 1 mM DTT) and incubated on ice for 5 min. Cellpellets were homogenized ﬁve times using insulin syringes and incubated with 2.5M sucrose solution (250 nM ﬁnal concentration). Cell pellets were submitted toseveral centrifugation steps: to separate nuclear fraction (1000 × g, 5 min, 4 °C); toseparate mitochondrial and endoplasmic reticulum fraction (10,000 × g, 15 min, 4 °C); to separate PM fraction and cytosolic fraction (100,000 × g, 1 h, 8 °C). Resultingmembrane pellets were collected and resuspended in 50 μl of homogenizationmedia followed by Western blot analysis. Cytosolic fractions were concentrated in3–4 times using 3K concentrators, 20,000 × g, 30 min, 4 °C followed by Westernblot analysis. Na/K-ATPase was used as a marker of PM fraction, and MEK-1/2was used as a marker of the cytosolic fraction.Flow cytometry. CTRL or SMP OE human podocytes were collected at day 14 ofdifferentiation and incubated with 0.5% BSA, 0.1% Saponin in 1 × PBS for 10 minat room temperature followed by Fc blocking step for 10 min at room temperatureusing 12.5 µg of human BD Fc Block solution. Incubation with 30 µl of IR APC-conjugated antibodies per sample was performed for 1 h at room temperature indark. HEK293 IR-transfected cells were served as positive control. CTRL humanpodocytes incubated with 10 µl of isotop control IgG APC-conjugated antibodieswere used as a negative control. Fluid sheath running through the cytometer washighly ﬁltered (0.22 μm ﬁlters). FACS experiments were performed in four biolo-gical replicates and 30,000 events per sample were analyzed. Fluorescence exclusionwas used to gate on live cells. Intact cells were identiﬁed by conventional ﬂowcytometry using side scatter area (SSC-A) and forward scatter area (FSC-A). IR-positive cells were determined using negative and positive controls gating (Sup-plementary Fig. 6). Results were read on BD LSR II 15-color ﬂow cytometricanalyzer using PE ﬂuorochrome with laser lens 535 nm and emission ﬁlters 585/15.All data were analyzed using FACSDiva 8.0.2 software.PCR analysis to differentiate IR isoforms. mRNA from human podocytes wasextracted using RNAeasy kit according to the manufacture’s protocol. 250 ng RNAwas reverse transcribed to obtain cDNA using the high-capacity cDNA reversetranscriptase kit according to the manufacturer’s protocol. PCR was performedusing GoTaq green master mix at 58 °C, 35 cycles yielding a 222-bp PCR productscorresponding to IR isoform B and a 186-bp PCR product corresponding to IRisoform A, using primer sequences reported in Supplementary Table 117.Transient transfection and IP. HEK293 cells were grown until 70–80% con-ﬂuence prior any transfection. Transient transfection was performed usingFuGene-6 reagent, OPTI-MEM media, and 1 μg of plasmid and incubated inantibiotic-free DMEM media for 24 h. Then cells were grown in DMEM mediasupplemented with 10% FBS and 1% penicillin/streptomycin for next 24 h andstarved for 2–4 h in DMEM with 0.1% FBS and 1% penicillin/streptomycin priortreatment or cells collection. Cell pellets were collected in Triton buffer (50 mMTris, pH 7.5, 150 mM NaCl, 1% Triton) and were precipitated with anti-FLAG M2Afﬁnity Gel62. After a wash pellets were analyzed by standard SDS gelelectrophoresis. Endogenous Co-IP was performed on isolated glomeruli fromwild-type B6N mice using Dynabeads M-270 Epoxy Co-IP Kit according tomanufacturer’s protocol. Brieﬂy, Dynabeads were coupled with IR antibodies.Proteins from isolated glomeruli were obtained using extraction buffer (100 mMNaCl, 2 mM MgCl2, 1 mM DTT, 1:200 protease inhibitor without EDTA).Antibody-coupled beads were incubated with lysates (1.5 mg of antibody-coupledbeads per 1.5 g of isolated glomeruli) at 4 °C for 30 min, followed by washing stepswith buffers (5 min in 200 μl 1 × LWB and 5 min in 60 μl 1xEB) provided by the kit.Animal studies. All animal studies have complied with all relevant ethical reg-ulations and were performed in accordance with the National Institutes of HealthGuidelines. The study protocol was approved by the Institutional Animal Care andUse Committee of the University of Miami, Miller School of Medicine.Genotyping, housing, and phenotypic analysis of mice. To generate podocyte-speciﬁc Smpdl3b-deﬁcient mice, Cre-loxP technology was used. Mice with aSmpdl3b-targeted allele were crossed to Flp-recombinase carrying mice63 to gen-erate heterozygous Smpdl3b ﬂoxed mice with a deleted FRT-Neo cassette (Fig. 4a,Supplementary Fig. 3). These were then mated with Podocin-Cre Tg/Tg mice22 togenerate heterozygous Podocin-Cre positive Smpdl3b ﬂoxed mice (Cre+;ﬂ/+).Finally, the latter were intercrossed to generate homozygous podocyte-speciﬁcSmpdl3b-deﬁcient mice (Cre+;ﬂ/ﬂ). As Podocin-Cre and Flp-recombinase trans-genic mice were used in a pure C57BL/6 background and ﬂoxed Smpdl3bmice wereused in a C57BL/6N background, wildtype littermates were used as a control toaccount for phenotypical variability due to background differences.For genotyping, tissues from tail biopsies of 3-week-old mice were digested withproteinase K and DNA was isolated using QIAamp DNA Mini kit according to themanufacture’s protocol. Smpdl3b, Flp- and Cre- recombinase alleles were detectedusing PCR (information on primers used for genotyping is available inSupplementary Table 4). Podocin-Cre transgenic mice22 were identiﬁed usingprimers speciﬁc for this transgene, yielding a 455 bp PCR product (SupplementaryFig. 3a). To distinguish wildtype mice from podocyte-speciﬁc Smpdl3b-deﬁcientmice, which are characterized by the presence of a 3′-loxP site downstream of exon2 of the Smpdl3b gene, a genotyping PCR on tail biopsies was established. In thisPCR, a set of four primers to detect the wildtype (WT1 and WT2) and the ﬂoxedallele (KO1 and KO2) giving rise to a 315 bp and 237 bp PCR product, respectively,was used (Supplementary Fig. 3a). To ensure successful Cre-mediatedrecombination, which is only detectable in podocytes, PCR on genomic DNAisolated from glomeruli using the primers 5FRT Fw, Smp_Del Rev andFRT_Del_Fw (Supplementary Table 1) was also performed. This PCR will onlyyield a 198 bp product if Cre-mediated recombination has occurred(Supplementary Fig. 3b). A 1075 bp band can be detected in all other cells of aglomerulus.Starting at 4 weeks of age, weight measurements and morning spot urines werecollected bi-weekly until mice reached 28 weeks of age. Male and female mice wereanalyzed in all study groups. Six to ten mice per group (Cre+;+/+– wildtype mice(n= 10) and Cre+/ﬂ/ﬂ –podocyte-speciﬁc Smpdl3b knockout mice (n= 6)) weresacriﬁced and in-depth phenotypical analysis was performed, including urinaryalbumin-to-creatinine ratio, histological analysis (PAS, H&E), glomeruli isolationand blood analysis.Heterozygous Podocin-Cre positive Smpdl3b-ﬂoxed mice (Cre+;ﬂ/+) werecrossed to Leprdb heterozygous db/+ mice to generate double heterozygous mice(Cre+;ﬂ/+;db/+). The latter were intercrossed to generate double mutant mice(Cre+;ﬂ/ﬂ;db/db).For genotyping, tissues from tail biopsies was digested as described above. Smpdl3b,db, and Cre- recombinase alleles were characterized using PCR (information onprimers used for genotyping is available in Supplementary Table 4).Starting at 4 weeks of age, weight and glycemia level measurements andmorning spot urines were collected bi-weekly until mice reached 28 weeks of age.Ten mice per group were sacriﬁced and in-depth phenotypical analysis wasperformed, including urinary albumin-to-creatinine ratio, histological analysis(PAS, H&E, PSR), transmission electron microscopy analysis (GBM thickness andfoot process effacement quantiﬁcation), immunohistochemistry, glomeruliisolation, and serum analysis.All mice were housed in MicroVENT IVC standard ventilated cages (AllentownInc., PA, USA) maintained at 68 ± 4°F with 12 h dark and 12 h light cycle andprovided water and Tekland Global 18% protein rodent chow diet with no restrictions.Urine samples analysis. Morning urine spot was collected bi-weekly. The urinealbumin content was measured by ELISA and urinary creatinine was measured byan assay based on the Jaffe method, using albumin ELISA kit from Bethyl Lab.(#E90-134; TX, USA) and creatinine kit from Stanbio (#0420-500; TX, USA).Values are expressed as microgram albumin per milligram creatinine.Histology and assessment of mesangial expansion. After perfusion of an animalwith 1 × PBS, the right kidney was removed for histological analysis and the leftkidney was collected for glomeruli extraction. Periodic acid-Schiff (PAS) andhematoxylin–eosin (H&E) staining of parafﬁn-embedded kidney sections (4 μmthick) was performed using a standard protocol. Histological images wereNATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-4 ARTICLENATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 |www.nature.com/naturecommunications 13visualized using a light microscope (Olympus BX 41, Tokyo, Japan) at ×40 mag-niﬁcation and analyzed using Image J software64. Twenty glomeruli per sectionwere analyzed for mesangial expansion by semi-quantitative analysis (scale 0–4)performed by two blinded independent investigators65,66.Transmission electron microscopy (TEM). For TEM, a kidney pole was ﬁxed in2% paraformaldehyde (PFA), 2% glutaraldehyde in 0.1 M phosphate buffer (pH=7.4) for at least one week prior to embedding. TEM was performed as describedpreviously67 with minor modiﬁcations. Samples were examined with a JEM-1011transmission electron microscope (JEOL) in the University of Tokyo. Negatives ofelectron micrographs at ×10,000 magniﬁcation were scanned at 600 d.p.i. Mea-surements of the resulting images was performed using Image J software. Eightimages per mouse were analyzed using the distance between two points on a rulerof a photograph as a measurement scale. GBM thickness was measured in 30different points. The number of podocyte foot processes (FP) along the GBM wascounted by hand. A FP was deﬁned as any connected epithelial segment butting onthe basement membrane, between two neighboring ﬁltration pores or slits. Fromeach photograph, the arithmetic mean of the foot process width (FPW) was cal-culated as follows: FPW= π/4*ΣGBM length/Σfoot process68.Blood samples analysis. Blood samples were analyzed for lipid panel, aspartateaminotransferase (AST), alanine transaminase (ALT), and blood urea nitrogen (BUN)in the Comparative Laboratory Core Facility of the University of Miami. Serumcreatinine was determined by tandem mass spectrometry at the UAB-UCSD O’BreinCore Center (University of Alabama, Birmingham) as previously described69.Immunoﬂuorescence staining and confocal microscopy. For immuno-ﬂuorescence staining of kidneys from mice, fresh cut 4 μm-thick tissue in OCT wasused. No ﬁxation agent was applied. Permeabilization was performed using 0.3%Triton X-100 in 1 × PBS for 10 min at room temperature. A blocking step wasperformed using 5% BSA, 2.5% FBS in PBS for 1 h at room temperature. Primaryantibodies were used for 24 h at room temperature. Alexa-conjugated secondaryantibodies were used for 1 h at room temperature. To detect nucleus, DAPIstaining was applied in a 1:500 dilution for 20 min at room temperature.To assess podocyte number and apoptosis in murine glomeruli,immunoﬂuorescent labeling with anti-WT1 was performed. WT1-positive nuclei werecounted in 20 consecutive glomerular cross-sections per animal by two blindedinvestigators11.Images were acquired by laser scanning confocal microscopy using a Leica SP5Inverted microscope, ×40 wet objective (Leica Microsystems CMS GmbH,Germany). Measuring of cell ﬂuorescence was performed using Image J software70.Exogenous C1P treatment of diabetic mice. Six 12-week-old homozygous db/db orheterozygous db/+ mice received daily intraperitoneal injections of 30mg/kg C1PC16:0 or 0.9% normal saline (NS) for 28 days. To prepare a stock solution (5 mg/ml),C1P C16:0 powder was weighted and dissolved in 0.9% NS and 2.5% DMSO bysonication. Aliquots were stored at −20 °C, then melted and sonicated once againbefore application. Urine collection, body weight, and glycemia measurements weredetermined weekly. Mice were sacriﬁced and in-depth phenotypical analysis wasperformed, including urinary albumin-to-creatinine ratio, histological analysis (PAS,H&E, PSR), glomeruli isolation and blood analysis as described previously. aspartateaminotransferase (IPGTT) was performed 3 weeks after exogenous C1P adminis-tration. After 12 h fasting, blood glucose was measured at baseline, and then at 15, 30,60, 90, and 120min after a glucose bolus (1.5 g/kg).Statistical analysis and study design. Data are expressed as a mean ± standarddeviation (SD). A number of experiments ranging between 3 and 5 was utilized asindicated for each distinct experiment. Minimal group sizes for in vitro and in vivostudies were determined via power calculator using the DSS Researcher’s Toolkit withan α of 0.05. Animals were grouped unblinded, but randomized, and investigatorswere blinded for the quantiﬁcation experiments. GraphPad Prism Outlier calculatorsoftware was used to indicate outliers in each set of data obtained for in vitro andin vivo experiments. Signiﬁcant outliers were excluded from further statistical analysis.Two groups of data were compared using the two-tailed t-test. Three and moregroups of data were compared using the one-way analysis of variance (ANOVA),followed by Tukey’s post-test or two-ways ANOVA with multiple comparisons. P <0.05 was taken to indicate statistical signiﬁcance. Statistical analysis was performedusing the GraphPad Prism, version 7.0 (GraphPad Software Inc.).Uncropped Western blots are depicted in Supplementary Figs. 7–10.Reporting summary. Further information on research design is available inthe Nature Research Reporting Summary linked to this article.Data availabilityThe raw data of lipidomic analysis in human podocytes and kidney cortexes of podocyte-speciﬁc Smpdl3b deﬁcient mice and type 2 diabetic podocyte-speciﬁc Smpdl3b deﬁcientmice are available in the Supplementary Datas 1, 3 and 5, respectively. All sequencingdata supporting the ﬁndings of this study are available in the Supplementary Datas 2 and4 and have been deposited in the National Center for Biotechnology Information GeneExpression Omnibus (GEO) and are accessible through the GEO Series accessionnumber GSE129666. Next publically available sources were used in the study: SMPDL3bpredicted domains (“SMPDL3b”). All other relevant data are available from thecorresponding author upon reasonable request.Received: 20 June 2018 Accepted: 15 May 2019References1. Hannun, Y. A. & Obeid, L. M. Principles of bioactive lipid signalling: lessonsfrom sphingolipids. Nat. Rev. Mol. Cell Biol. 9, 139–150 (2008).2. Fornoni, A. et al. Rituximab targets podocytes in recurrent focal segmentalglomerulosclerosis. Sci. Transl. Med. 3, 85ra46 (2011).3. Yoo, T. H. et al. Sphingomyelinase-like phosphodiesterase 3b expression levelsdetermine podocyte injury phenotypes in glomerular disease. J. Am. Soc.Nephrol. 26, 133–147 (2015).4. Heinz, L. X. et al. The lipid-modifying enzyme SMPDL3B negatively regulatesinnate immunity. Cell Rep. 11, 1919–1928 (2015).5. Gorelik, A., Heinz, L. X., Illes, K., Superti-Furga, G. & Nagar, B. Crystalstructure of the acid sphingomyelinase-like phosphodiesterase SMPDL3Bprovides insights into determinants of substrate speciﬁcity. J. Biol. Chem. 291,24054–24064 (2016).6. Chalfant, C. E. & Spiegel, S. Sphingosine 1-phosphate and ceramide 1-phosphate: expanding roles in cell signaling. J. Cell Sci. 118, 4605–4612 (2005).7. Janecke, A. R. et al. Deﬁciency of the sphingosine-1-phosphate lyase SGPL1 isassociated with congenital nephrotic syndrome and congenital adrenalcalciﬁcations. Hum. Mutat. 38, 365–372 (2017).8. Lovric, S. et al. Mutations in sphingosine-1-phosphate lyase cause nephrosiswith ichthyosis and adrenal insufﬁciency. J. Clin. Investig. 127, 912–928(2017).9. Prasad, R. et al. Sphingosine-1-phosphate lyase mutations cause primaryadrenal insufﬁciency and steroid-resistant nephrotic syndrome. J. Clin.Investig. 127, 942–953 (2017).10. Tejada, T. et al. Failure to phosphorylate AKT in podocytes from mice withearly diabetic nephropathy promotes cell death. Kidney Int. 73, 1385–1393(2008).11. Guzman, J. et al. Podocyte-speciﬁc GLUT4-deﬁcient mice have fewer andlarger podocytes and are protected from diabetic nephropathy. Diabetes 63,701–714 (2014).12. Mima, A. et al. Glomerular-speciﬁc protein kinase C-beta-induced insulinreceptor substrate-1 dysfunction and insulin resistance in rat models ofdiabetes and obesity. Kidney Int. 79, 883–896 (2011).13. Leibiger, B. et al. Selective insulin signaling through A and B insulin receptorsregulates transcription of insulin and glucokinase genes in pancreatic betacells. Mol. Cell 7, 559–570 (2001).14. Otsu, K. et al. Caveolin gene transfer improves glucose metabolism in diabeticmice. Am. J. Physiol. Cell Physiol. 298, C450–C456 (2010).15. Kabayama, K. et al. Dissociation of the insulin receptor and caveolin-1complex by ganglioside GM3 in the state of insulin resistance. Proc. Natl Acad.Sci. USA 104, 13678–13683 (2007).16. Eduardsen, K. et al. Cell volume regulation and signaling in 3T3-L1 pre-adipocytes and adipocytes: on the possible roles of caveolae, insulin receptors,FAK and ERK1/2. Cell. Physiol. Biochem. 28, 1231–1246 (2011).17. Villarreal, R. et al. Nephrin contributes to insulin secretion and affectsmammalian target of rapamycin signaling independently of insulin receptor. J.Am. Soc. Nephrol. 27, 1029–1041 (2016).18. Sedding, D. G. et al. Caveolin-1 facilitates mechanosensitive protein kinase B(Akt) signaling in vitro and in vivo. Circ. Res. 96, 635–642 (2005).19. Cohen, A. W., Combs, T. P., Scherer, P. E. & Lisanti, M. P. Role of caveolinand caveolae in insulin signaling and diabetes. Am. J. Physiol. Endocrinol.Metab. 285, E1151–E1160 (2003).20. Razani, B. et al. Caveolin-1-deﬁcient mice are lean, resistant to diet-inducedobesity, and show hypertriglyceridemia with adipocyte abnormalities. J. Biol.Chem. 277, 8635–8647 (2002).21. Lee, J. A. et al. Methyl-beta-cyclodextrin up-regulates collagen I expression inchronologically-aged skin via its anti-caveolin-1 activity. Oncotarget 6,1942–1953 (2015).22. Moeller, M. J., Sanden, S. K., Sooﬁ, A., Wiggins, R. C. & Holzman, L. B.Podocyte-speciﬁc expression of cre recombinase in transgenic mice. Genesis(Landes Bioscience, Austin, 2000) 35, 39–42 (2003). https://www.ncbi.nlm.nih.gov/books/NBK25446/23. Kinoshita, T. & Fujita, M. Biosynthesis of GPI-anchored proteins: specialemphasis on GPI lipid remodeling. J. Lipid Res. 57, 6–24 (2016).ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-414 NATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 | www.nature.com/naturecommunications24. Ahmad, A. et al. Sphingomyelinase-like phosphodiesterase 3b mediatesradiation-induced damage of renal podocytes. FASEB J. 31, 771–780 (2017).25. Bajwa, A. et al. Sphingosine Kinase 2 deﬁciency attenuates kidney ﬁbrosis viaIFN-gamma. J. Am. Soc. Nephrol. 28, 1145–1161 (2017).26. Gómez-Muñoz, A. G. P., Granado, M. H., Arana, L., & Ouro, A. Ceramide 1-phosphate in cell survival and inﬂammatory signaling. In Shingolipids asSignaling and Regulatory Molecules (ed. Chalfant, C. E. & Poeta, M.) LandesBioscience (2000–2013).27. Gomez-Munoz, A. Ceramide 1-phosphate/ceramide, a switch between life anddeath. Biochim. Biophys. Acta 1758, 2049–2056 (2006).28. Hait, N. C. & Maiti, A. The role of sphingosine-1-phosphate and ceramide-1-phosphate in inﬂammation and cancer. Mediators Inﬂamm. 2017, 4806541(2017).29. Kim, C. H. et al. Conditioning for hematopoietic transplantation activates thecomplement cascade and induces a proteolytic environment in bone marrow:a novel role for bioactive lipids and soluble C5b-C9 as homing factors.Leukemia 26, 106–116 (2012).30. Gangoiti, P. et al. Ceramide 1-phosphate stimulates macrophage proliferationthrough activation of the PI3-kinase/PKB, JNK and ERK1/2 pathways. Cell.Signal. 20, 726–736 (2008).31. Chaurasia, B. et al. Adipocyte ceramides regulate subcutaneous adiposebrowning, inﬂammation, and metabolism. Cell Metab. 24, 820–834 (2016).32. Summers, S. A. Sphingolipids and insulin resistance: the ﬁve Ws. Curr. Opin.Lipidol. 21, 128–135 (2010).33. Meikle, P. J. & Summers, S. A. Sphingolipids and phospholipids in insulinresistance and related metabolic disorders. Nat. Rev. Endocrinol. 13, 79–91 (2017).34. Chavez, J. A. et al. A role for ceramide, but not diacylglycerol, in theantagonism of insulin signal transduction by saturated fatty acids. J. Biol.Chem. 278, 10297–10303 (2003).35. Chavez, J. A., Holland, W. L., Bar, J., Sandhoff, K. & Summers, S. A. Acidceramidase overexpression prevents the inhibitory effects of saturated fattyacids on insulin signaling. J. Biol. Chem. 280, 20148–20153 (2005).36. Holland, W. L. et al. Inhibition of ceramide synthesis amelioratesglucocorticoid-, saturated-fat-, and obesity-induced insulin resistance. CellMetab. 5, 167–179 (2007).37. Yamamoto, M. et al. Caveolin is an activator of insulin receptor signaling. J.Biol. Chem. 273, 26962–26968 (1998).38. Van Krieken, R. & Krepinsky, J. C. Caveolin-1 in the pathogenesis of diabeticnephropathy: potential therapeutic target? Curr. Diabetes Rep. 17, 19 (2017).39. Lee, Y. D., Yoon, S. H., Ji, E. & Kim, H. H. Caveolin-1 regulates osteoclastdifferentiation by suppressing cFms degradation. Exp. Mol. Med. 47, e192 (2015).40. Perreault, L. et al. Intracellular localization of diacylglycerols andsphingolipids inﬂuences insulin sensitivity and mitochondrial function inhuman skeletal muscle. JCI Insight 3, e96805 (2018).41. Lee, S. Y. et al. Adipocyte-speciﬁc deﬁciency of de novo sphingolipid biosynthesisleads to lipodystrophy and insulin resistance. Diabetes 66, 2596–2609 (2017).42. Turinsky, J., O’Sullivan, D. M. & Bayly, B. P. 1,2-Diacylglycerol and ceramidelevels in insulin-resistant tissues of the rat in vivo. J. Biol. Chem. 265,16880–16885 (1990).43. Adams, J. M. 2nd et al. Ceramide content is increased in skeletal muscle fromobese insulin-resistant humans. Diabetes 53, 25–31 (2004).44. Stratford, S., Hoehn, K. L., Liu, F. & Summers, S. A. Regulation of insulinaction by ceramide: dual mechanisms linking ceramide accumulation to theinhibition of Akt/protein kinase B. J. Biol. Chem. 279, 36608–36615 (2004).45. He, X., Huang, Y., Li, B., Gong, C. X. & Schuchman, E. H. Deregulation ofsphingolipid metabolism in Alzheimer’s disease. Neurobiol. Aging 31, 398–408(2010).46. Pralhada Rao, R. et al. Sphingolipid metabolic pathway: an overview of majorroles played in human diseases. J. Lipids 2013, 178910 (2013).47. Ogretmen, B. Sphingolipids in cancer: regulation of pathogenesis and therapy.FEBS Lett. 580, 5467–5476 (2006).48. Ogretmen, B. Sphingolipid metabolism in cancer signalling and therapy. Nat.Rev. Cancer 18, 33–50 (2018).49. Marﬁa, G. et al. Autocrine/paracrine sphingosine-1-phosphate fuels proliferativeand stemness qualities of glioblastoma stem cells. Glia 62, 1968–1981 (2014).50. Natarajan, V. et al. Sphingosine-1-phosphate, FTY720, and sphingosine-1-phosphate receptors in the pathobiology of acute lung injury. Am. J. Respir.Cell Mol. Biol. 49, 6–17 (2013).51. Baudiß, K. C1P Attenuates lipopolysaccharide-induced acute lung injury bypreventing NF-κB activation in neutrophils. J. Immunol. (Baltimore, MD:1950) 196, 2319–2326 (2016).52. Sivasubramanian, M., Kanagaraj, N., Dheen, S. T. & Tay, S. S. Sphingosinekinase 2 and sphingosine-1-phosphate promotes mitochondrial function indopaminergic neurons of mouse model of Parkinson’s disease and in MPP+-treated MN9D cells in vitro. Neuroscience 290, 636–648 (2015).53. Rivera, I. G. et al. Ceramide 1-phosphate regulates cell migration andinvasion of human pancreatic cancer cells. Biochem. Pharmacol. 102, 107–119(2016).54. Mitsutake, S. et al. Ceramide kinase is a mediator of calcium-dependentdegranulation in mast cells. J. Biol. Chem. 279, 17570–17577 (2004).55. Graf, C. et al. Neutropenia with impaired immune response to Streptococcuspneumoniae in ceramide kinase-deﬁcient mice. J. Immunol. (Baltimore, MD:1950) 180, 3457–3466 (2008).56. Saleem, M. A. et al. A conditionally immortalized human podocyte cell linedemonstrating nephrin and podocin expression. J. Am. Soc. Nephrol. 13,630–638 (2002).57. Granado, M. H. et al. Ceramide 1-phosphate (C1P) promotes cell migrationInvolvement of a speciﬁc C1P receptor. Cell. Signal. 21, 405–412 (2009).58. Bielawski, J., Szulc, Z. M., Hannun, Y. A. & Bielawska, A. Simultaneousquantitative analysis of bioactive sphingolipids by high-performance liquidchromatography-tandem mass spectrometry. Methods 39, 82–91 (2006).59. Boudker, O. & Futerman, A. H. Detection and characterization of ceramide-1-phosphate phosphatase activity in rat liver plasma membrane. J. Biol. Chem.268, 22150–22155 (1993).60. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29,15–21 (2013).61. Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductorpackage for differential expression analysis of digital gene expression data.Bioinformatics 26, 139–140 (2010).62. Asanuma, K. et al. Synaptopodin regulates the actin-bundling activity of alpha-actinin in an isoform-speciﬁc manner. J. Clin. Investig. 115, 1188–1198 (2005).63. Farley, F. W., Soriano, P., Steffen, L. S. & Dymecki, S. M. Widespreadrecombinase expression using FLPeR (ﬂipper) mice. Genes (New York, NY:2000) 28, 106–110 (2000).64. Montes, G. S. Structural biology of the ﬁbres of the collagenous and elasticsystems. Cell Biol. Int. 20, 15–27 (1996).65. Crowley, S. D. et al. Glomerular type 1 angiotensin receptors augment kidneyinjury and inﬂammation in murine autoimmune nephritis. J. Clin. Investig.119, 943–953 (2009).66. Wei, C. et al. Circulating urokinase receptor as a cause of focal segmentalglomerulosclerosis. Nat. Med. 17, 952–960 (2011).67. Choudhary, V., Ojha, N., Golden, A. & Prinz, W. A. A conserved family ofproteins facilitates nascent lipid droplet budding from the ER. J. Cell Biol. 211,261–271 (2015).68. van den Berg, J. G., van den Bergh Weerman, M. A., Assmann, K. J., Weening, J.J. & Florquin, S. Podocyte foot process effacement is not correlated with the levelof proteinuria in human glomerulopathies. Kidney Int. 66, 1901–1906 (2004).69. Takahashi, N., Boysen, G., Li, F., Li, Y. & Swenberg, J. A. Tandem massspectrometry measurements of creatinine in mouse plasma and urine fordetermining glomerular ﬁltration rate. Kidney Int. 71, 266–271 (2007).70. McCloy, R. A. et al. Partial inhibition of Cdk1 in G 2 phase overrides the SACand decouples mitotic events. Cell Cycle 13, 1400–1412 (2014).AcknowledgementsA.F. and S.M. are supported by the NIH grants DK104753, DK11759 and CA227493. A.F. issupported by the NIH grants DK090316, U24DK076169, U54DK083912, UM1DK100846,and 1UL1TR000460. A.F., S.M., C.F. are supported by Hoffman-La Roche. G.M.D. issupported by a Predoctoral Fellowship of the American Heart Association(16PRE30200010). JJK is supported by NRSA Post-doctoral fellowship from NIH/NIDDK(F32DK115109). We thank Dr. Chalfant for his suggestions in the design of exogenous C1Pexperiments in vivo. We thank Dr. Heinz the CeMM Research Center for MolecularMedicine of the Austrian Academy of Sciences for a kind gift of HA-tagged human wildtypeand mutant SMPDL3b constructs. We would like to acknowledge the skilled assistance ofthe Flow Cytometry Shared Resource and the Analytical Imaging Shared Resource of theSylvester Comprehensive Cancer Center at the University of Miami, Miller School ofMedicine, for the provision of the sophisticated ﬂuorescence analysis.Author contributionsA.M. performed the in vitro and in vivo experiments, analyzed the data, and wrotemethods and results of the manuscript. S.K.M. prepared material for illumina sequencingRNA data analysis, analyzed presence of enriched functional-related gene groups, per-formed endogenous co-IP experiments, and performed some of the in vivo experiments.G.M.D. performed WT1 histological staining and some in vitro experiments. J.M. per-formed PSR histological staining and analysis and some in vitro experiments. T.H.Y.initialized the characterization of the podocyte-speciﬁc Smpdl3b knockout mice colony.E.R.-G. and I.D.Z. performed C6 ceramide production assay in HEK293 cells and inpodocytes using puriﬁed SMPDL3b protein. J.V.S., C.G.H., C.P., M.G., A.S., J.J.K. per-formed some of the in vitro experiments. J.B. and I.V. performed genotyping proceduresand some of the in vitro experiments. C.F. designed some of the immunoprecipitationexperiments. Y.Z. and A.J.M. designed experiments related to sphingolipid analysis. Y.Z.analyzed some of the mass spec data. I.L. designed and provided human insulin receptorisoforms plasmids. G.W.B. conceived one of the techniques utilized. A.H.F. designed C6ceramide production experiments. L.B. evaluated kidney tissue slides for histopathology.Y.I. and R.I. performed TEM experiments. S.M. and A.F. conceived the project, designedNATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-4 ARTICLENATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 |www.nature.com/naturecommunications 15and supervised the study, analyzed the data, and edited the manuscript. A.F. wroteintroduction and discussion of the manuscript, she is the guarantor of this work and, assuch, had full access to all the data in the study and takes responsibility for the integrityof the data and the accuracy of the data analysis.Additional informationSupplementary Information accompanies this paper at https://doi.org/10.1038/s41467-019-10584-4.Competing interests: G.W.B., A.F., and S.M. are inventors on pending or issued patents(US10,183,038, US10,052,345) aimed to diagnose or treat proteinuric renal diseases. Theystand to gain royalties from their future commercialization. A.F. is Chief Scientiﬁc Ofﬁcerof L&F Health LLC and is consultant for Variant Pharmaceutical. VariantPharmaceuticals, Inc. has licensed worldwide rights to develop and commercializehydroxypropyl-beta-cyclodextrin for treatment of kidney disease from L&F Research. S.M. holds equity interest in a company presently commercializing the form ofcyclodextrin referenced in this paper. The patent associated with the use ofhydroxypropyl-beta-cyclodextrin is published under US10,195,227. A.F. is Chief MedicalOfﬁcer of LipoNexT, LLC. The remaining authors delare no competing interests.Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/Peer review information: Nature Communications thanks the anonymous reviewersfor their contribution to the peer review of this work. Peer reviewer reports are?available.Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional afﬁliations.Open Access This article is licensed under a Creative CommonsAttribution 4.0 International License, which permits use, sharing,adaptation, distribution and reproduction in any medium or format, as long as you giveappropriate credit to the original author(s) and the source, provide a link to the CreativeCommons license, and indicate if changes were made. The images or other third partymaterial in this article are included in the article’s Creative Commons license, unlessindicated otherwise in a credit line to the material. If material is not included in thearticle’s Creative Commons license and your intended use is not permitted by statutoryregulation or exceeds the permitted use, you will need to obtain permission directly fromthe copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.© The Author(s) 2019ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10584-416 NATURE COMMUNICATIONS | (2019) 10:2692 | https://doi.org/10.1038/s41467-019-10584-4 | www.nature.com/naturecommunications

Modulation of sphingosine-1-phosphate and

Abstract

Extracted data

Modulation of sphingosine-1-phosphate and

Abstract

Extracted data

Related items

Related items