Renal Glycosphingolipid Metabolism Is Dysfunctional in Lupus Nephritis

BASIC RESEARCH www.jasn.org

† ‡ † Tamara K. Nowling,* Andrew R. Mather, Thirumagal Thiyagarajan, | | María José Hernández-Corbacho,§ Thomas W. Powers, E. Ellen Jones, Ashley J. Snider,§¶ † | Jim C. Oates,* Richard R. Drake, and Leah J. Siskind**

*Research Service, Ralph H. Johnson Veterans Affairs Medical Center, Charleston, South Carolina; †Department of Medicine, Division of Rheumatology and Immunology and Departments of ‡Drug Discovery and Biomedical Sciences and |Cell and Molecular Pharmacology, Medical University of South Carolina, Charleston, South Carolina; §Department of Medicine, Stony Brook University Medical Center, Stony Brook, New York; ¶Research Service, Northport Veterans Affairs Medical Center, Northport, New York; and **Department of Pharmacology and Toxicology and the James Graham Brown Cancer Center, University of Louisville, Louisville, Kentucky

ABSTRACT Nearly one half of patients with lupus develop glomerulonephritis (GN), which often leads to renal failure. Although nephritis is diagnosed by the presence of proteinuria, the pathology of nephritis can fall into one of five classes defined by different forms of tissue injury, and the mechanisms involved in pathogenesis are not completely understood. Glycosphingolipids are abundant in the kidney, have roles in many cellular functions, and were shown to be involved in other renal diseases. Here, we show dysfunctional glycosphingolipid metabolism in patients with lupus nephritis and MRL/lpr lupus mice. Specifically, we found that glucosylceramide (GlcCer) and lactosylceramide (LacCer) levels are significantly higher in the kidneys of nephritic MRL/lpr lupus mice than the kidneys of non-nephritic lupus mice or healthy controls. This elevation may be, in part, caused by altered transcriptional regulation and/or activity of LacCer synthase (GalT5) and neuraminidase 1, enzymes that mediate glycosphingolipid metabolism. We show increased neuraminidase 1 activity early during the progression of nephritis (before significant elevation of GlcCer and LacCer in the kidney). Elevated levels of urinary LacCer were detected before proteinuria in lupus mice. Notably, LacCer levels were higher in the urine and kidneys of patients with lupus and nephritis than patients with lupus without nephritis or healthy controls. Together, these results show early and significant dysfunction of the glycosphingolipid metabolic pathway in the kidneys of lupus mice and patients with lupus nephritis and suggest that molecules in this pathway may serve as early markers in lupus nephritis.

J Am Soc Nephrol 26: 1402–1413, 2015. doi: 10.1681/ASN.2014050508

SLE is a heterogeneous autoimmune inﬂammatory lesions to bland, membranous lesions, it is diagnosed disease presenting in varying degrees and symptoms by the presence of proteinuria or nephritic urine.10 and affecting many organs, including skin, joint, res- piratory, cardiac, and kidney.1,2 SLE is thought to result Received May 22, 2014. Accepted August 7, 2014. from environmental triggers and predisposing genetic factors.2–5 SLE broadly encompasses accumulation and Published online ahead of print. Publication date available at subsequent deposition of immune complexes in or bind- www.jasn.org. ing of autoantibodies to tissues, causing injury and fur- Correspondence: Dr. Tamara K. Nowling, Department of Med- ther activating an immune response.4,6,7 Approximately icine, Medical University of South Carolina, 96 Jonathan Lucas Street, CSB 912 MSC 637, Charleston, SC 29425-6370, or Dr. Leah 1,8 onehalfofpatientswithSLE develop renal disease. J. Siskind, Department of Pharmacology and Toxicology, University Lupus nephritis (LN) is part of a larger family of GN of Louisville, 505 South Hancock Street, CTRB 203, Louisville, KY diseases that often leads to end stage kidney disease.9 40202. Email: [email protected] or [email protected]. Although LN varies from proliferative, inﬂammatory Copyright © 2015 by the American Society of Nephrology

1402 ISSN : 1046-6673/2606-1402 JAmSocNephrol26: 1402–1413, 2015 www.jasn.org BASIC RESEARCH

Glycosphingolipids are a heterogeneous class of lipids in the autoimmune kidney disease. We tested this hypothesis in the sphingolipid family that plays a role in the regulation of cellular MRL/MpJ-Faslpr/J (MRL/lpr) mouse model of LN and human processes, such as proliferation and inflammation. Highly kidney biopsies and urine samples. We show that renal glyco- abundant in the kidney, glycosphingolipids play a role in a sphingolipid metabolism is dysregulated in lupus-prone mice, variety of kidney diseases, including GN, kidney cancer, ne- resulting in elevated levels of renal and urinary glycosphingolipids. phropathy, and polycystic kidney disease.11 Hexosylceramide Glycosphingolipid accumulation is, at least in part, caused by (HexCer), glucosylceramide (GlcCer), and galactosylceramide transcriptional upregulation of NEU1, with significant increase (GalCer) are generated by the addition of sugar moieties to cer- in NEU activity early in disease that continues to increase as amide. GlcCer is generated from GlcCer synthase-mediated disease progresses. Of translational importance, urinary LacCer addition of glucose to ceramide (Figure 1). After formed, lacto- levels and NEU1 protein levels were significantly higher in pa- sylceramide (LacCer) synthases catalyze the addition of galactose tients with LN than patients with lupus without nephritis and to form LacCer. Complex glycosphingolipids are formed by fur- healthy subjects. LacCer levels were also increased in kidney ther additions to LacCer. For example, gangliosides are formed biopsies of patients with LN compared with controls. To our by addition of sialic acid residues to LacCer. After formed, gly- knowledge, this is the first report of glycosphingolipid dyfunc- cosphingolipids are transported within in the cell to the plasma tion in LN, and data strongly suggest that glycosphingolipids membrane and lysosome. Sialidases (neuraminidases [NEUs]) may serve as early markers of LN. remove sialic acids from gangliosides and proteins. NEUs are localized to the lysosome (NEU1), cytosol (NEU2), plasma membrane (NEU3), and mitochondria (NEU 4), and NEU1 RESULTS can be secreted extracellularly to act on the plasma membrane.12 Thus, LacCer can be regenerated by NEUs breaking Kidney Glycosphingolipids Are Elevated in Nephritic down gangliosides. MRL/lpr Lupus Mice Aberrant glycosphingolipid metabolism occurs in several Sixteen-week-old MRL/lpr lupus mice, a model for classes 3 and autoimmune diseases, including autoimmune diabetes,13 thy- 4 LN, have nephritis characterized by renal immune complex roid disease,14 and various neuropathies.15–19 Thus, we hy- deposition, lymphocyte infiltration, proteinuria, and signifi- pothesized a role for glycosphingolipids in LN, a prototypical cant renal pathology compared with age-matched MRL/MpJ lupus-prone mice, which develop nephritis much later in life.10,20 We analyzed renal cortical glycosphingolipid levels in MRL/lpr and MRL/MpJ mice and found that HexCers and LacCers were 7- to 8-fold higher in 16-week-old MRL/lpr mice compared with non-nephritic age-matched MRL/MpJ mice (Figure 2). The increases in HexCers and LacCers are likely not because of an increase in ceramides, because only C24-ceramide was significantly elevated (Figure 3A). HexCers consist of both GlcCer and GalCer, which cannot be distin- guished by HPLC-mass spectrometry (MS)/MS. Supercritical fluid tandem MS was performed at the Medical University of South Carolina Lipidomics Core Facility to separate GlcCer and GalCer. Most of the GlcCer species are significantly elevated (Figure 3B), whereas GalCer species remained unchanged (Figure 3C), indicating that HexCer levels can be considered representative of GlcCer levels. All LacCer species were elevated in the kidney cortices of MRL/lpr compared Figure 1. Schematic depicting basics of sphingolipid metabolism. with MRL/MpJ mice, with C16-andC24:1-LacCers signifi- Ceramide glycosylation produces mono-HexCers, namely GlcCer cantly elevated (Figure 3D). Matrix-assisted laser desorption/ or GalCer. Galactozation of GlcCer produces LacCer, which is used ionization imaging MS (MALDI-IMS) was used to determine to produce complex glycosphingolipids, including gangliosides. the tissue distributions of HexCer (Figure 3E) and LacCer GlcCer and LacCer levels are governed by both their synthesis from (Figure 3F) in the kidneys of 18-week-old nephritic MRL/lpr ceramides and the breakdown of gangliosides by NEUs. Ceramide mice compared with age-matched healthy C57BL/6 mice and GalCer levels are unaltered, whereas GlcCer and LacCer are fi with representative MS/MS spectra (Figure 3G). These re- signi cantly elevated in lupus mice with nephritis. Gray arrows fi (increased or decreased levels) are indicated next to the glyco- sults con rm that GSLs are elevated compared with normal sphingolipids and enzymes with significant changes in the nephritic healthy controls and show that different species have differ- lupus mice compared with the non-nephritic lupus mice. GBA, ent tissue distributions, with most being elevated in the cor- glucocerebrosidase; GM3, monosialodihexosylganglioside; SM, tex region, where many of the pathologic changes occur sphingomyelin. in LN.

J Am Soc Nephrol 26: 1402–1413, 2015 Glycosphingolipids in Lupus Nephritis 1403 BASIC RESEARCH www.jasn.org

(Figure 5A). We wanted to determine the mechanism respon- sible for increased Neu1 message. Sterol regulatory element- binding proteins (SREBPs) are transcription factors known to regulate global lipid metabolism. They can be activated by glycosphingolipids and induce expression of genes that regulate glycosphingolipid synthesis, including NEU1.21 SREBPs are associated with several kidney diseases,22 but their involve- ment in LN has not been shown. Renal cortical SREBP-1 protein levels were increased in MRL/lpr mice compared with MRL/MpJ control mice (Figure 5B). Both SREBP-1a and SREBP-1c were shown to upregulate Neu1 in myotubes.21 To determine which SREBP-1 isoform may be regulating Neu1 in the lupus kidneys, we measured expression of factors in cholesterol and fatty acid metabolism, pathways reg- ulated by SREBP-1a and SREBP-1c, respectively.22 Fatty acid synthase message levels (Figure 5C) and triglyceride levels (Figure 5D) were significantly increased, whereas 3-hydroxy-3- methyl-glutaryl-CoA reductase message levels (Figure 5E) and cholesterol levels (Figure 5F) were not significantly different in Figure 2. Total kidney cortex mono-HexCers and LacCers are the renal cortices of MRL/lpr compared with MRL/MpJ mice. significantly elevated in MRL/lpr mice. HexCer and LacCer were These results are consistent with an upregulation of SREBP-1c measured by LC-MS/MS in renal cortices from 16-week-old MRL/lpr but not SREBP-1a. Together, these results suggest that accumu- (lupus) and MRL/MpJ (control) mice. n=10–15 mice; error bars, lation of GlcCer and LacCer in LN occurs, in part, by SREBP-1c SEM. ***P#0.001 as determined by one-way ANOVA. LC, liquid induction of Neu1 expression, which breaks down gangliosides chromatography. to yield LacCers.

Kidney Glycosphingolipid Degradation Is Dysregulated Renal Glycosphingolipid Dysfunction Begins Early in in Nephritic MRL/lpr Lupus Mice Disease and Corresponds with Elevated Urinary LacCer GlcCers and LacCerscan accumulate by increased synthesis from Renal injury typically results in excretion of plasma proteins into ceramides and/or increased degradation of complex glycosphin- the urine, with albumin accounting for the majority of urine golipids. To determine the mechanism of kidney GlcCer and proteins during severe glomerular injury. In the MRL/lpr model, LacCer accumulation in MRL/lpr mice, kidney cortex message urinary albumin is elevated between 12 and 24 weeks of age.10 To levels of enzymes that directly regulate glycosphingolipid metab- determine if glycosphingolipid levels are detectable in the urine olism were determined. Message levels of GlcCer synthase, of lupus mice and how early in disease they are elevated with which synthesize GlcCer from ceramide, and glucocerebrosi- respect to albuminuria, urinary albumin and C16-LacCers (the dase, which mediates the breakdown of GlcCer to ceramide, most abundant species in the urine of mice; data not shown) were unchanged in nephritic MRL/lpr compared with non- were quantified in 24-hour urine collections from age-matched nephritic MRL/MpJ mice (Figure 4, A and B). Of three major MRL/lpr and MRL/MpJ mice. As reported previously,23 protein- LacCer synthases (GalT5, GalT6, and GalT7) that can synthe- uria increased at 16–18 weeks, with significant differences at 20 size LacCer from GlcCer, only GalT5 message was altered with weeks (Figure 6A). C16-LacCers increased at 14–16 weeks, with lower expression in nephritic compared with non-nephritic significant differences at 18 weeks (Figure 6B). These results mice (Figure 4C). Additional enzymes that mediate glyco- suggest that C16-LacCers increase before proteinuria and may sphingolipid metabolism include galactosidase (GLA), which be early markers of nephritis. breaks down globotriaosylceramide to LacCer, GM3 synthase, We then measured renal cortex GlcCer/LacCer levels and which synthesizes gangliosides from LacCer, and Neus, which NEU activity earlier in disease to determine when they begin to break down gangliosides to form LacCer. GLA and GM3 syn- increase. There was a slight increase in the major GlcCer and thase levels were not significantly changed (Figure 4, D and E), LacCer species between 11 and 14 weeks (Figure 6, C and D), whereas Neu1 was significantly elevated (Figure 4F) in MRL/lpr which significantly increased from 14 to 18 weeks of age (Figure compared with MRL/MpJ mice. Results in Figures 3 and 4 are 6, C and D). These changes paralleled increases in urinary summarized in the pathway diagram in Figure 1. C16-LacCer (Figure 6B). NEU activity in renal cortices of the same mice was significantly higher in 11-week-old MRL/lpr Increased Neu1 Message Translates to Increased Renal mice compared with the activity in 18-week-old healthy Cortical NEU Activity in Lupus Mice C57BL/6 mice, increasing over time (Figure 6E). NEU activity NEU activity was significantly higher in nephritic MRL/lpr increased significantly in MRL/lpr kidneys between 11 and 18 mice compared with non-nephritic MRL/MpJ mice weeks of age and between 14 and 18 weeks of age (Figure 6E).

1404 Journal of the American Society of Nephrology J Am Soc Nephrol 26: 1402–1413, 2015 www.jasn.org BASIC RESEARCH

Figure 3. GlcCer and LacCer species are elevated in kidney cortex of MRL/lpr lupus mice compared with MRL/MpJ or C57BL/6 control mice. Specific species of (A) ceramide, (B) GlcCer, (C) GalCer, and (D) LacCer were quantified by LC-MS/MS (or SFC-MS/MS for the GlcCer and GalCer) at the MUSC Lipidomics Core Facility in 16-week-old MRL/lpr and MRL/MpJ kidneys. Data are the means6SEM. *P#0.05; ***P#0.001 as determined by t test adjusting for multiple comparisons (Holm–Sidak method). Tissue distribution of (E) HexCers (GalCer+GlcCer) C22, C24, and C24:1 and (F) LacCers C16, C22, and C24:1 in 18-week-old MRL/lpr (LN) and C57BL/6 (B6) kidneys. Images are representative of at least four animals from each strain analyzed. Scale bar, 5 mm. (G) Spectra for the LacCers showing the increase in the MRL/lpr compared with the C57BL/6 kidney. LC, liquid chromatography; MUSC, Medical University of South Carolina; SFC, supercritical fluid chromatography.

These results indicate that NEU activity becomes significantly general increase in LacCer staining outside of the glomeruli (Fig- elevated before accumulation of GlcCer and LacCer and sug- ure 7D, upper panel) and intense staining within the glomeruli gest that dysfunctional glycosphingolipid metabolism occurs (Figure 7E, upper panel) in the biopsies of LN patients. The very early in disease in the MRL/lpr kidney. intense staining corresponds with the increase in cellular expan- sion observed in LN (Figure 7, D and E compared with Figure Urinary and Kidney LacCer Levels Are Elevated in 7C, upper panel). Confocal imaging at higher power shows that Patients with LN the intense staining in the glomeruli in Figure 7E is located in the Todetermine whether glycosphingolipid metabolism is dysfunc- mesangial region (Figure 7F). We then measured NEU activity tional in patients with LN, urinary C16-LacCer levels were and NEU1 protein levels in urine samples from the same patients measured and normalized to creatinine or eGFR to account with lupus analyzed in Figure 7A. We observed measurable NEU for differences in renal function. When normalizing to either activity in only two samples: both patients with class 4 nephritis creatinine or eGFR, C16-LacCer was significantly elevated in the (data not shown). The lack of measurable NEU activity in most urine of patients with LN compared with patients with lupus of the samples may be because of the presence of an inhibitor of without nephritis and healthy control subjects (Figure 7A) (data NEU activity in the urine and/or extended handling of the sam- not shown), whereas levels in the serum were not significantly ples. NEU activity is extremely sensitive to temperature changes. different among the groups (Figure 7B). Although the sample NEU1 protein levels in these same urine samples were increased size was relatively small, the significant differences in the urine in patients with LN compared with patients with lupus without were large, emphasizing that GSL metabolism in the kidney may nephritis and healthy controls (Figure 7G) (data not shown). be important in the pathogenesis of nephritis. We then analyzed The results in Figure 7G are representative of eight samples an- LacCer expression in renal biopsies from patients with LN com- alyzed for each group, with all eight patients with LN expressing pared with controls. A low level of expression throughout the measurable levels of NEU1 protein. Compared with patients tissue of the control biopsies (Figure 7C, upper panel) was ob- with LN, light bands were observed in three patients with lupus served as expected, because glycosphingolipid expression is without nephritis and one healthy control. These results show abundant in normal, healthy kidneys. However, we observed a that renal glycosphingolipid dysfunction occurs in patients with

J Am Soc Nephrol 26: 1402–1413, 2015 Glycosphingolipids in Lupus Nephritis 1405 BASIC RESEARCH www.jasn.org

mice. Similarly, urinary LacCers in nephritic mice were significantly elevated compared with in non-nephritic mice. Moreover, the elevation of LacCer in the urine (16 weeks) occurred before development of significant proteinuria (18 weeks), suggesting that urinary LacCer may be an earlier or more sensitive marker of nephritis. By measuring expression of the enzymes involved glycosphingolipid metabolism, we identified possible mechanisms for accumulation of renal GlcCer and LacCer in nephritic mice. The only significant difference in expression of enzymes in the synthesis pathway was for GalT5, one of the LacCer synthase genes, which was significantly downregulated in nephritic mice. A lack of GalT5 was shown to induce accumulation of GlcCer.25 Therefore, accumulation of renal GlcCer in LN mice may be, in part, because of decreased GalT5 expression. Neu1 expression and NEU activity were significantly elevated in lupus mice with Figure 4. NEU activity and expression are significantly elevated in the kidney cor- nephritis compared with non-nephritic tices of 16-week-old MRL/lpr lupus mice. Message levels were quantified in renal lupus mice. The enzymatic assay does not cortices by real-time RT-PCR and normalized to b-actin for (A) GlcCer synthase, (B) distinguish between NEU isoforms, and glucocerebrosidase, (C) LacCer synthase, (D) GLA, (E) GM3 synthase, and (F) NEU1. thus, we cannot determine the contribution Data are the means6SEMs. *P#0.05 as determined by Mann–Whitney test. GM3, of NEU1 to the overall increase in NEU monosialodihexosylganglioside. activity. If increased NEU activity is a major contributor to the elevation of LacCer and/ or GlcCer, the increase in activity should LN and that elevated levels of LacCer and NEU1 protein in the precede increased GlcCer and LacCers. In MRL/lpr mice, im- urine of patients with LN likely are cause by renal and not sys- mune complex deposition in the kidney begins around 8–10 temic GSL dysfunction. Coupled with results in mice, these re- weeks of age followed by local renal production of MCP-1 and sults suggest that urinary LacCer may serve as an early marker other cytokines and renal IgG and C3 deposition with immune for nephritis and/or a screening tool to identify individuals likely cell infiltration occurring between 14 and 16 weeks of age.10,26 to develop nephritis. NEU activity was significantly increased earlier in disease than the increase in GlcCer and LacCers. The early increase in NEU activity in the kidney may be a response to immune complex DISCUSSION deposition and local expression of cytokines. This dysregula- tion may then contribute to the progression of nephritis by To our knowledge, this is the first report showing glycosphingo- promoting additional local renal cytokine production and leu- lipid metabolic dysfunction in the kidneys of mice and patients kocyte infiltration. Indeed, exogenously added LacCer and with LN. Our studies used the MRL/lpr lupus-prone mouse GlcCer to mesangial cells induce production of several cyto- model and the lupus-prone mouse model from which they kines/chemokines important in leukocyte infiltration in LN were derived, the MRL/MpJ strain. We compared age-matched (L.J. Siskind, unpublished observations). Moreover, the tim- nephritic MRL/lpr with non-nephritic MRL/MpJ mice. We also ing of immune cell infiltration in the MRL/lpr kidney corre- used the nonautoimmune-prone C57Bl/6 strain as a healthy sponds with the significant increase in GlcCer and LacCer control. Both MRL strains develop lupus similar to that observed species from 14 to 18 weeks of age. Our previous studies in in human lupus pathologies, including increased IgG levels, the MRL/lpr model showed that mice with improved disease autoantibody production, proteinuria, and eventually, develop- have significantly decreased levels of the major GlcCer/LacCer ment of end stage kidney disease, with the lpr mice developing species in their T cells at 18 weeks of age.27 We speculate that disease at an accelerated rate because of a mutation of the Fas infiltrating T cells and other immune cells may contribute to gene.24 We showed increased kidney HexCer and LacCer in ne- the increased glycosphingolipid levels in the kidney either di- phritic MRL/lpr mice compared with non-nephritic MRL/MpJ rectly through their expression of GlcCer and LacCer and/or

1406 Journal of the American Society of Nephrology J Am Soc Nephrol 26: 1402–1413, 2015 www.jasn.org BASIC RESEARCH

Our results suggest, for the first time, that activation of SREPBs and accumulation of lipids within the kidney also occur in LN. Of translational significance, data showed that patients with LN exhibited dysfunction of glycosphingolipid metabolism similar to the MRL/lpr nephritic mice. Patients with LN had elevated levels of urinary LacCer and NEU1 but not serum LacCer compared with patients without nephritis and healthy controls. LacCer staining of renal biopsies from patients with LN showed intense staining around the glomeruli in the mesangial region compared with biopsies from control patients. Together, our results suggest that the elevated urine LacCer levels in patients with LN are caused by GSL dysfunction in the kidney. Interestingly, the levels of urinary NEU1 seemed to be higher in patients with class 4 compared with class 3 nephritis. Fu- ture work will be aimed at determining if Figure 5. Increased renal NEU expression and activity in MRL/lpr mice are likely in- there are class-specific differences in urinary creased by an SREBP-1c pathway. (A) NEU activity was measured by an activity assay. and/or renal NEU1 and LacCer levels. SREBP-1 protein levels were measured by (B) Western immunoblot. (C) Fatty acid Together, our resultsindicatethat GlcCers synthase and (E) 3-hydroxy-3-methyl-glutaryl-CoA (HMG-CoA) reductase message and LacCers are elevated in a mouse model levels were measured by real-time RT-PCR, and (D) triglyceride levels and (F) cho- of LN as well as patients with LN. Our data lesterol levels were measured by ELISA in renal cortices from 16-week-old MRL/lpr suggest that increased NEU1 activity may be 6 # , and MRL/MpJ mice. Data are the means SEMs. *P 0.05; ****P 0.001 as de- the mechanism by which these lipids accu- – termined by Mann Whitney test. mulate. Data indicate that NEU1 is upregulated at the transcriptional level and suggest that SREBP-1c may play a role in NEU1 up- indirectly by activating resident renal cells that further elevate regulation and dysfunctional glycosphingolipid metabolism GlcCer and LacCer levels. Our results here analyzing the in LN mice. Data suggest that glycosphingolipid metabolism human serum and urine samples suggest that the elevated is dysfunctional early in disease before renal pathogenesis. lipids are largely caused by renal-specific rather than systemic Because GlcCers and LacCers are present in the urine, they contributions. Future studies will define mechanisms respon- also represent potential biomarkers for diagnosis of LN and may sible for the significant increases in glycosphingolipid metab- be a noninvasive means for distinguishing different classes of olism with respect to the pathologic hallmarks of nephritis nephritis. Future studies are aimed at analyzing longitudinal development. samplesofpatientswithoutnephritiswhosubsequentlydeveloped SREBP-1 is a transcription factor that regulates expression of nephritis to determine if LacCer may serve as an early marker for genes involved in lipid homeostasis. When the mature nuclear nephritisandusinginhibitorsofglycosphingoplipidmetabolismin forms of SREBP-1a and SREBP-1c were overexpressed in human lupus-prone mice to determine if glycosphingolipids are, indeed, skeletal muscle cells, microarray analysis revealed increased ex- mediators of disease. pression of several genes involved in glycosphingolipid metabolism, including Neu1.21 We showed an increase in the mature form of SREBP-1 in the kidneys of nephritic mice; because established CONCISE METHODS targets of SREBP-1c are also upregulated in the kidneys of LN-prone mice, data indicate that the mature fragment of Patient Renal Biopsies and Urine Samples SREBP-1c is, indeed, active. Elevated levels of GlcCers and LacCers Clean catch urine samples for the LacCer and NEU protein level are sufficient to induce SREBPactivation.28 Thus, glycosphingolipids analyses were obtained from frozen, centrifuged, and aliquoted may be up- and downstream of SREBP activation. Aberrant acti- samples that were collected for an unrelated lupus urine biomarker vation of SREBPs and renal lipid accumulation occur in kidney study.32 All volunteers gave documented informed consent to partic- diseases and were linked to increased production of inflammatory ipate in Institutional Review Board-approved protocols. Analyzed cytokines and glomerulosclerosis.29,30 Overexpressing the active samples were from patients (predominantly African Americans) form of SREBP-1 in the kidneys of mice induces proteinuria.31 meeting American College of Rheumatology revised criteria for

J Am Soc Nephrol 26: 1402–1413, 2015 Glycosphingolipids in Lupus Nephritis 1407 BASIC RESEARCH www.jasn.org

Figure 6. Renal glycosphingolipid metabolism is dysregulated early during disease development. Timed urine collection was performed every other week in MRL/lpr and MRL/MpJ mice. (A) Albumin and (B) C16-LacCer were measured. Kidneys were collected from 11-, 14-, and 18-week-old MRL/lpr mice and 18-week-old C57BL/6 (B6) mice. (C) GlcCer and (D) LacCer levels and (E) NEU activity were measured in renal cortices. Data are means6SDs. *P#0.05; ** P#0.01; ***P#0.001; ****P,0.001 as determined by a two-way ANOVA adjusting for multiple comparisons.

SLE33 without nephritis, subjects with active LN at entry,32 and con- urine and did not develop nephritis within 1 year of urine collection. trol subjects without autoimmune or inﬂammatory disease. Spot Normal controls were free of known renal disease or inﬂammatory urine and blood collections and clinical evaluations were performed autoimmune disease but did not have renal biopsies performed. For at entry. Patient demographics and clinical measures ascertained the LacCer immunohistochemistry analyses, frozen OCT-embedded from medical records for subjects (urine samples [Figure 7A] and renal biopsies were obtained from some of the patients with LN in- serum samples [Figure 7B]) are shown in Table 1. Disease activity as dicated above and patients who underwent renal transplants (con- measured by the Systemic Lupus Erythematosus Disease Activity In- trol) without signs of rejection. Renal biopsies were graded by the dex,34 eGFR, and the National Institutes of Health morphologic in- World Health Organization criteria, and those patients with classes 3 dices of renal biopsies (Activity Index and Chronicity Index), which and 4 were deemed to have proliferative LN. Recipients of renal trans- correlate with clinical parameters, are included. Subjects with LN had plant met all of the entry criteria for patients with LN except for the ISN/RPS active class 2, 3, 4, or 5 nephritis by biopsy. Snap-frozen lupus criteria and had no clinical or biopsy evidence of rejection. All renal cortical biopsy tissue used for clinical immunohistochemistry samples were processed using standard quality control measures. was stored at 280°C and later retrieved for research immunostaining as described below. Subjects did not have an active infection, ongoing Animals pregnancy, or .50% of glomeruli with sclerosis on biopsy. Patients C57BL/6 and MRL/MpJ mice were purchased from The Jackson with lupus non-nephritis met four American College of Rheumatol- Laboratory (Bar Harbor, ME). MRL/MpJ-Faslpr/J (MRL/lpr)mice ogy criteria for lupus but had ,100 mg protein/g creatinine in the were bred in house starting with breeders purchased from The

1408 Journal of the American Society of Nephrology J Am Soc Nephrol 26: 1402–1413, 2015 www.jasn.org BASIC RESEARCH

MRL/lpr mice were used at 11–18 weeks of age as indicated. MRL/MpJ mice were used at 16 weeks of age. C57BL/6 mice were used at 18 weeks of age. All mice were maintained on a 12-hour light/12-hour dark cycle and provided food and water ad libitum. Animals were maintained under standard laboratory conditions. All animal procedures were approved by the Veterans Administration Institutional Animal Care and Use Committee and followed the guidelines of the American Veterinary Medical Association.

Kidney Homogenates Homogenates were made from kidneys snap frozen in liquid nitrogen. The cortex was cut from the kidney and homogenized using a Tissue-Tearor (Biospec Products, Inc., Bartles- ville, OK) in 20 mM Tris-HCl buffer containing a Complete Protease Inhibitor Cocktail Tablet (Roche, Indianapolis, IN). Protein concentrations were measured using the Pierce Biotech Micro BCA Protein Assay Kit (Thermo Fisher Scientiﬁc, Waltham, MA).

Lipid Levels Sphingolipids were measured quantitatively in 2 mg kidney cortex homogenates or a 300-ml aliquot obtained from 24-hour urine collections in mice or spot urine collections from human subjects. Quantitation of sphingolipid species was performed by the Lipidomics Core at Med- ical University of South Carolina on a Thermo Finnigan TSQ 7000 Triple-Stage Quadrupole Mass Spectrometer operating in a Multiple Re- action Monitoring (MRM)-positive ionization mode as described.37,38 ESI/MS/MS analyses of Figure 7. Urine and renal LacCer is increased in patients with lupus and nephritis. (A) endogenous GlcCer and GalCer molecular spe- Urine and (B) serum from patients with active proliferative LN before induction therapy cies (C18-Sphingoid Base) were performed on a (SLE with nephritis), patients with SLE without nephritis (SLE), and healthy controls were Thermo Fisher TSQ Quantum Triple Quadru- 6 , analyzed for C16-LacCer. Data are means SDs. ***P 0.001 compared with healthy pole Mass Spectrometer operating in an MRM- controls and P,0.01 compared with SLE as determined by a one-way ANOVA and positive ionization mode using a modified adjusting for multiple comparisons. LacCer staining in renal biopsies from (C) a control version.38 subject, (D) a patient with lupus and class 3 nephritis, and (E) a patient with lupus and Briefly, cell pellets corresponding to about 1– class 4 nephritis. Upper panels are LacCer (FITC) staining, and lower panels are 496- 33106 cells were fortified with synthetic and diamidino-2-phenylindole staining at 320. Images are representative of the five not natural internal standards (ISs: C8-GluCer, control, three class 4, and two class 3 samples analyzed. (F) Confocal image of the sample in E at 363 with FITC, DAPI, and bright-field overlay showing intense LacCer C8-GalCer, C12-GluCer, and C12-GalCer; Avanti staining in the mesangial region. (G) NEU1 protein expression in urine from patients Polar Lipids, Inc., Alabaster, AL) and extracted with lupus and nephritis, patients without nephritis (SLE), and healthy controls. with an ethyl acetate/isopropanol/water (60/30/10 vol/vol) solvent system. After evaporation and reconstitution in 100 ml methanol, samples were Jackson Laboratory. To avoid genetic drift in the MRL/lpr population, injected on the Supercritical Fluid Chromatography/Mass Spectrom- additional MRL/lpr mice were purchased from The Jackson Lab- etry Berger Minigram/TSQ Quantum SFC/MS System and gradient oratory and used every other generation for breeding. Measures of eluted from a Berger Silica 60 A 6-mm particle, 4.63250-mm column disease were performed routinely on our in-house MRL/lpr colony, with a 1.0 mM methanolic ammonium formate/2 mM aqueous am- and they were similar, as was survival, to those published previously.10,35,36 monium formate mobile phase system. Peaks corresponding to the

J Am Soc Nephrol 26: 1402–1413, 2015 Glycosphingolipids in Lupus Nephritis 1409 BASIC RESEARCH www.jasn.org

Table 1. Patient demographics and clinical measures of urine samples analyzed in Figure 7A and serum samples analyzed in Figure 7B Sample Age (yr) Women (%) African American (%) SLEDAI eGFR Activity Chronicity Urine samples Control 37611 (20–50) 86 71 SLE 37616 (18–64) 100 67 4 (2–8) 116642 (77–174) SLE with nephritis 3666(29–46) 86 100 10 (2–16) 73640 (16–109) 6.566.4 (0–19) 2.661.7 (1–6) Serum samples Control 37614 (20262) 90 90 SLE 36611 (18–54) 89 89 5 (0–16) 110636 (49–167) SLE with nephritis 3269(20–48) 90 90 16 (8–26) 65638 (20–118) 8.965.3 (1.5–19) 4.162.2 (1–8) Average age, SLE Disease Activity Index (SLEDAI), activity, and chronicity (6SD with ranges in parentheses) are provided for each group. All other subjects not in the African American column were Caucasian. target analytes and ISs were collected and processed using the Xcali- gelatin. For most samples 1:100–1:6400 was sufﬁcient to obtain read- bur software system. ings within the linear range of the standard curve, whereas some Quantitative analysis was on the basis of the calibration curves samples required additional dilution. Standards (mouse albumin; generated by spiking an artiﬁcial matrix with the known amounts of Sigma-Aldrich, St. Louis, MO) were serially diluted in a similar man- the target analyte synthetic standards and an equal amount of the ISs. ner starting at 1 mg/ml. Standards and samples were incubated at The target analyte/IS peak areas ratios were plotted against analyte room temperature for 60 minutes and washed with PBST. HRP- concentration. The target analyte/IS peak area ratios from the samples conjugated goat anti-mouse albumin (Fitzgerald Industries, Acton, were similarly normalized to their respective ISs and compared with MA) diluted 1:10,000 in PBST/0.25% gelatin was added to each well the calibration curves using a linear regression model. and incubated at room temperature for 60 minutes; then, the wells were The calibration standards (C16-GlcCer, C16-GalCer, C18-GlcCer, washed with PBST. Substrate (100 ml/well 20 mlTMB/980ml0.2M

C18-GalCer, C18:1-GlcCer, C18:1-GalCer, C24-GlcCer, C24-GalCer, HCl, 100 ml 0.1 M citrate, pH 4.0, 50 mlH2O2) was added to each well C24:1-GlcCer, and C24:1-GalCer) were synthesized in the Lipidomics and incubated at room temperature for 35 minutes. Absorbance was Share Resources, Medical University of South Carolina (Charleston, read at OD 380 nm on a Thermo Multiskan Ascent Plate Reader SC). Quantitation of the target GlcCer/GalCers, for which no (Thermo Fisher Scientific), and albumin concentration per 1 ml urine synthetic standards are available, was accomplished using calibration was calculated from the standard curve using the Multiskan Ascent curves of the closest counterpart. software. Three readings within the linear portion of the standard curve Cholesterol and triglyceride levels were measured in kidney for each sample were averaged and multiplied by the total volume of homogenates using the Cholesterol/Cholesterol Ester Quantification urine to determine total albumin excreted in 24 hours. Kit (Biovision, San Francisco, CA) and Triglyceride Quantification Kit (Biovision), respectively, according to the manufacturer’s instruc- Gene Expression tions. Briefly, a two-phase extraction of 400 mg kidney homogenate Reverse-transcribed RNA was used in real-time PCR to measure the was performed in 50 ml PBS; then, 200 ml chloroform:isopropanol: relative message levels of GlcCer synthase, the major transcripts of NP-40 (7:11:0.1 vol/vol) was added, vortexed for 2 minutes, and LacCer synthases (GalT5, GalT6, and GalT7), glucocerebrosidase, GLA, centrifuged. The organic phase was transferred to a new tube, dried ST3 b-galactoside a-2,3-sialyltransferase 5 (ST3GALVor GM3 syn- down under nitrogen gas, and placed in a vacuum for 30 minutes to thase), NEU1 and NEU3, fatty acid synthase, 3-hydroxy-3-methyl- remove all remaining chloroform. The dried lipids were resuspended glutaryl-CoA reductase, and b-actin in RNA isolated from kidney in 200 ml assay buffer and sonicated until homogenous (roughly 1 cortices. Briefly, RNA was isolated from a 20-mg piece of kidney minute); 50 ml(100mg worth of protein homogenate) was used for cortex from individual mice using the Qiagen RNeasy RNA Isolation the assay on the basis of prior determination showing that this Kit (Valencia, CA) according to the manufacturer’s instructions. amount of protein had levels in the linear range of the kit. Levels RNA (1 mg) was reverse-transcribed using the Bio-Rad iScript were measured according to the manufacturer’s instructions. cDNA Synthesis Kit (Hercules, CA) in a 20-mlreaction,and1ml cDNA was used per real-time PCR reaction. Each real-time reaction Proteinuria was performed in triplicate on the MyiQ Bio-Rad Thermocycler Urine was collected from mice over a 24-hour period in metabolic using the iQ SYBR Green Supermix (Bio-Rad), and threshold cycle cages. To inhibit bacterial growth, ampicillin, gentamicin, and number for each reaction was determined using the Real-Time iCycler chloramphenicol were added to the collection tubes. Urinary protein iQ software (Bio-Rad). Relative expression was calculated using the levels were measured by albumin ELISA. Albumin was measured in Bio-Rad Gene Expression Macro software as described previously39 96-well plates coated overnight at 4°C with goat anti-mouse albumin. with b-actin as the reference gene. Relative differences between groups Wells were blocked with PBS/0.05% Tween-20 (PBST) plus 0.25% are presented as fold change. All primers were synthesized by IDT gelatin for 60 minutes at room temperature and washed with PBST; (Iowa City, IA). Primer efficiencies were determined for all primer 2-fold dilution series of each urine sample were made in PBST/0.2% sets to ensure similar efficiencies between each primer set for the gene

1410 Journal of the American Society of Nephrology J Am Soc Nephrol 26: 1402–1413, 2015 www.jasn.org BASIC RESEARCH of interest and the primer set for the reference gene (b-actin), and all (0.2 M in 50% methanol/0.01% TCA) was applied using an Image- were .98%. Primer sequences are shown in Table 2. Prep sprayer (Bruker Daltonics, Billerica, MA). Spectra were acquired across the entire tissue section using a Solarix 7T FTICR Mass Spec- NEU Activity trometer (Bruker Daltonics) in positive ion mode in the mass range Kidney homogenates were prepared, and protein concentration was m/z=200–2000 with a SmartBeam II laser operating at 1000 Hz, a laser measured as described above. NEU enzyme activity was measured in spot size of 25 mm, and a raster width of 50 mm. For each laser spot, 500 50 mg kidney homogenate using the Invitrogen AmplexRed NEU Assay spectra were averaged, and all data were normalized using root means Kit following the manufacturer’s instructions. The kit detects activity of square. Images of glycosphingolipid species were generated using Flex- all four NEU enzymes. NEU activity is measured in arbitrary units, and Imaging 4.0 software (Bruker Daltonics). For identification of individ- differences between groups are presented as fold change. ual glycosphingolipid species, continuous accumulation of selected ions/collision-induced dissociation fragmentation was done with refer- Protein Expression ence standards spotted on an MALDI plate with DHB matrix or directly SREBP and NEU1 protein levels were measured by immunoblot anal- on tissue. Continuous accumulation of selected ions allows for trapping ysis. For SREBP detection, 20 mg kidney homogenate was separated of a specific ion of interest within the quadrupole of the Solarix 7T FT- on a 10% SDS-PAGE Tris-EDTAgel (Bio-Rad) and transferred to PVDF ICR Mass Spectrometer. After accumulation, the selected ion was frag- membranes. The membranes were incubated with either anti–SREBP-1 mented using collision-induced dissociation. Structural assignments or anti–SREBP-2 mouse mAbs (SREBP-1 [2A4] and SREBP-2 [1C6]; were made after detection of specific fragmentation patterns (e.g.,the Santa Cruz Biotechnology, Santa Cruz, CA), washed, and incubated sphingoid base at m/z=264; loss of 162/180 for hexose), compared with with anti-mouse–HRP secondary antibody (Sigma-Aldrich). SREBP an in-house structure database,41 and crossvalidated with the Lipid was detected with the Pierce ECL Kit (Thermo Fisher Scientific). Maps database. For NEU1 detection, each sample was normalized to the concentration of creatinine so that the equivalent of 50mg creatinine was loaded Immunohistochemistry per lane. Proteins were separated on a 10% SDS-PAGE TGX gel (Bio- Human renal biopsies (frozen and embedded in OTC) were sectioned at Rad) and transferred to PVDF membranes. The membranes were in- 10 mm. Slides were thawed at room temperature and fixed with acetone cubated with a rabbit anti-NEU1 (H-300; Santa Cruz Biotechnology), at 280°C. Slides were washed with PBST plus 1% goat serum (Vectas- washed, and incubated with an Alexa Fluor 647 goat anti-rabbit IgG tain Kit; Vector Laboratories, Burlingame, CA) and blocked with 5% antibody highly crossadsorbed secondary (Molecular Probes; Invitro- goat serum in PBST for 60 minutes in a humidified chamber. Slides were gen, Eugene, OR). NEU1 was detected by scanning on the Odyssey washed with PBST/1% goat serum and incubated with a 1:50 dilution of Infrared Imaging System (LI-COR, Lincoln, NE). Replicate blots incu- rabbit anti-CD17 (LacCer; Biorybt, San Francisco, CA) in PBST/1% bated with the secondary alone showed little to no background. goat serum overnight at 4°C in a humidified chamber. After washing with PBST/1% goat serum, the slides were incubated with a 1:200 di- MALDI-MS Imaging of Kidney Sections lution of goat anti-rabbit IgG-FITC secondary (Southern Biotechnology Direct profiling of glycosphingolipid expression in kidney tissue Associates, Birmingham, AL) in PBST/1% goat serum for 60 minutes at sections was accomplished by MALDI-IMS using a dual-source 7T room temperature in a humidified chamber in the dark. Sections were Solarix FT-ICR MS instrument as previously described.40,41 Relative washed with PBST/1% goat serum, a drop of DAPI (Vectashield; Vector concentrations of GlcCer and LacCer species were assessed simulta- Laboratories) was added, and a coverslip was placed over sections fol- neously in the same sections by MALDI-IMS in frozen kidney sec- lowed by a final incubation for 15 minutes at room temperature in the tions. Kidney tissues (10 mm) were prepared on ITO-coated glass dark. Images of sections were taken using a Nikon Eclipse 80i micro- slides, rinsed in water, and desiccated. Dihydroxybenzoic acid matrix scope, DS camera, and software (Nikon Instruments, Melville, NY).

Table 2. Primer sequences for real-time PCR Target Gene Sense Sequence Antisense Sequence GlcCer Synthase (UGCG) 59-CTT TGT TCG GCT TCG TGC TCT TC-39 59-TGC TAT ACG GCT GTT TGT CTG TTG C-39 B-4-GalT-5 59-GGC GGA GAA GAT GAC TTG TG-39 59-CGA TGG TGA GGA ATG GAC TTG-39 B-4-GalT-6 59-ATG ATC GGA ACT ACG GAT GTG-39 59-GCC GTT GAT CTT CCT GAA TTG C-39 B-4-GalT-7 59-TAC TGC TCC GCT CCT CTG-39 59-GCT TGT GTT CTG CCT TCT GAT GTC-39 GM3 synthase (ST3 Gal5) 59-GTG TTG AGG TGG GAG AG- 39 59-TGC CAG AGG TGA GTA AGC-39 Neu1 59-ATG TGC CTG TGC TAC TCC TG-39 59-GGT CTT GAT GTC CTT GAT GG-39 Neu3 59-TTC CAC CTT CCC TTC CTC ATC C- 39 59-GCA ATA AGC ACC GTT ATC AAC CAT C-39 GLA 59-GTT CAG TCC CAT CCT TCA G-39 59-CCC TTG AGC ATC ACC ATC TTC-39 Glucocerebrosidase 59-TCT GGA TGC TTA TGC TAA GTA TGG-39 59-CAC GGG AAA TGA AGT CTC TCT G-39 Fatty acid synthase 59-AGG TGC TGG CTG GAG AAG G-39 59-CGG TCG GTG GCT GTG TAT TC-39 3-Hydroxy-3-methyl-glutaryl-CoA reductase 59-GCT TGC TGT GAG AAT GTG ATC-39 59-GAG CCA TCC AAC AGT CAT CTA TGC-39 b-Actin 59-AGA TTA CTG CTC TGG CTC CTA-39 59-CCT GCT TGC TGA TCC ACA TC-39

J Am Soc Nephrol 26: 1402–1413, 2015 Glycosphingolipids in Lupus Nephritis 1411 BASIC RESEARCH www.jasn.org

Statistical Analyses 5. Moser KL, Kelly JA, Lessard CJ, Harley JB: Recent insights into the Statistical analyses were performed using GraphPad PRISM 6 software genetic basis of systemic lupus erythematosus. Genes Immun 10: 373– (GraphPad Software, Inc., La Jolla, CA). Nonparametric one-way or 379, 2009 6. Tsokos GC, Wong HK, Enyedy EJ, Nambiar MP: Immune cell signaling two-way ANOVA was performed with adjustment for multiple in lupus. Curr Opin Rheumatol 12: 355–363, 2000 comparisons as appropriate. All analyses were unpaired, except for 7. Ramanujam M, Davidson A: Targeting of the immune system in sys- the proteinuria and LacCer longitudinal analyses in urine samples temic lupus erythematosus. Expert Rev Mol Med 10: e2, 2008 collected from the same mice over time presented in Figure 6. Specific 8. Ebert EC, Hagspiel KD: Gastrointestinal and hepatic manifestations of statistical analyses used and P values are indicated in the figures. systemic lupus erythematosus. J Clin Gastroenterol 45: 436–441, 2011 9. Isbel NM: Glomerulonephritis—management in general practice. Aust Fam Physician 34: 907–913, 2005 10. Andrews BS, Eisenberg RA, Theofilopoulos AN, Izui S, Wilson CB, ACKNOWLEDGMENTS McConahey PJ, Murphy ED, Roths JB, Dixon FJ: Spontaneous murine lupus-like syndromes. Clinical and immunopathological manifestations in several strains. JExpMed148: 1198–1215, 1978 The authors thank Dr. Darwin Bell and Stacy Steele for assistance with 11. Mather AR, Siskind LJ: Glycosphingolipids and kidney disease. Adv Exp the confocal images and Dr. Paul Nietert for guidance with the sta- Med Biol 721: 121–138, 2011 tistical analyses. 12. Miyagi T, Tsuiki S: Evidence for sialidase hydrolyzing gangliosides – Support for these studies was provided by National Institutes of GM2andGM1inratliverplasmamembrane.FEBS Lett 206: 223 228, 1986 Health National Institute of Arthritis & Musculoskeletal & Skin Diseases 13. Misasi R, Dionisi S, Farilla L, Carabba B, Lenti L, Di Mario U, Dotta F: Grant P60-AR062755 Multidisciplinary Clinical Research Center Pilot Gangliosides and autoimmune diabetes. Diabetes Metab Rev 13: 163– Project (to T.K.N.) and Patient Resource Core (to J.C.O.), South Caro- 179, 1997 lina Clinical and Translational Research (SCTR) Institute, with an aca- 14. Lacetti P, Tombaccini D, Aloj S, Grollman EF, Kohn LD: Gangliosides, demic home at the Medical University of South Carolina, through the thyrotropin receptor, and autoimmune thyroid disease. Adv Exp Med Biol 174: 355–367, 1984 National Institutes of Health Grant Numbers UL1-RR029882 and 15. Nagai Y, Momoi T, Saito M, Mitsuzawa E, Ohtani S: Ganglioside UL1-TR000062, Veteran’s Administration Merit Review Grants syndrome, a new autoimmune neurologic disorder, experimentally BX000115 (to T.K.N.) and CX000218 (to J.C.O.), National Institutes of induced with brain gangliosides. Neurosci Lett 2: 107–111, 1976 Health NIAMS Grant R01-AR45476-11 (to J.C.O.), the Alliance for 16.MiyamotoK,TakadaK,FurukawaK,FurukawaK,KusunokiS:Rolesof Lupus Research (to J.C.O.), National Institutes of Health National complex gangliosides in the development of experimental autoimmune encephalomyelitis. Glycobiology 18: 408–413, 2008 Cancer Institute Grant R01-CA135087 (to R.R.D.), Department of De- 17. Ponzin D, Menegus AM, Kirschner G, Nunzi MG, Fiori MG, Raine CS: fense Grant W81XWH-10-1-0136 (to R.R.D.), National Institutes of Effects of gangliosides on the expression of autoimmune de- Health National Center for Research Resources Grant P20-RR17677 myelination in the peripheral nervous system. Ann Neurol 30: 678–685, Center of Biomedical Research Excellence in Lipidomics and Patho- 1991 biology Pilot Project (to L.J.S.), National Institutes of Health National 18. Steck AJ, Kappos L: Gangliosides and autoimmune neuropathies: Classification and clinical aspects of autoimmune neuropathies. J Institute of Diabetes and Digestive and Kidney Diseases Grant Neurol Neurosurg Psychiatry 57[Suppl]: 26–28, 1994 R01-DK093462 (to L.J.S.), and the Lipidomics Shared Resource of the 19. Willison HJ: Gangliosides as targets for autoimmune injury to the ner- Hollings Cancer Center at the Medical University of South Carolina vous system. J Neurochem 103[Suppl 1]: 143–149, 2007 supported by Cancer Center Support Grant P30CA138313. 20. Theofilopoulos AN, Dixon FJ: Murine models of systemic lupus erythe- – The contents of this publication do not represent the views of the matosus. Adv Immunol 37: 269 390, 1985 21. Rome S, Lecomte V, Meugnier E, Rieusset J, Debard C, Euthine V, Vidal Department of Veterans Affairs, the US Government, or any of the H, Lefai E: Microarray analyses of SREBP-1a and SREBP-1c target genes other funding entities. identify new regulatory pathways in muscle. Physiol Genomics 34: 327– 337, 2008 22. Eberlé D, Hegarty B, Bossard P, Ferré P, Foufelle F: SREBP transcription DISCLOSURES factors: Master regulators of lipid homeostasis. Biochimie 86: 839–848, None. 2004 23. Wang S, Yang N, Zhang L, Huang B, Tan H, Liang Y, Li Y, Yu X: Jak/STAT signaling is involved in the inflammatory infiltration of the kidneys in MRL/lpr mice. Lupus 19: 1171–1180, 2010 REFERENCES 24. Drappa J, Brot N, Elkon KB: The Fas protein is expressed at high levels on CD4+CD8+ thymocytes and activated mature lymphocytes in nor- 1. Manson JJ, Rahman A: Systemic lupus erythematosus. Orphanet J Rare mal mice but not in the lupus-prone strain, MRL lpr/lpr. Proc Natl Acad Dis 1: 6, 2006 Sci U S A 90: 10340–10344, 1993 2. Perry ES, Davies SV, Laversuch C: Acute systemic lupus erythematosus 25. Nishie T, Hikimochi Y, Zama K, Fukusumi Y, Ito M, Yokoyama H, Naruse on the acute medical take: Are we missing anything? Clin Med 11: 268– C, Ito M, Asano M: Beta4-galactosyltransferase-5 is a lactosylceramide 270, 2011 synthase essential for mouse extra-embryonic development. Glycobiology 3. Joseph A, Brasington R, Kahl L, Ranganathan P, Cheng TP, Atkinson J: 20: 1311–1322, 2010 Immunologic rheumatic disorders. J Allergy Clin Immunol 125[Suppl 2]: 26. Pérez de Lema G, Maier H, Nieto E, Vielhauer V, Luckow B, Mampaso F, S204–S215, 2010 Schlöndorff D: Chemokine expression precedes inflammatory cell in- 4. Crispín JC, Liossis SN, Kis-Toth K, Lieberman LA, Kyttaris VC, Juang YT, filtration and chemokine receptor and cytokine expression during the Tsokos GC: Pathogenesis of human systemic lupus erythematosus: initiation of murine lupus nephritis. JAmSocNephrol12: 1369–1382, Recent advances. Trends Mol Med 16: 47–57, 2010 2001

1412 Journal of the American Society of Nephrology J Am Soc Nephrol 26: 1402–1413, 2015 www.jasn.org BASIC RESEARCH

27. Richard EM, Thiyagarajan T, Bunni MA, Basher F, Roddy PO, Siskind LJ, disease activity index for lupus patients. Arthritis Rheum 35: 630–640, Nietert PJ, Nowling TK: Reducing FLI1 levels in the MRL/lpr lupus 1992 mouse model impacts T cell function by modulating glycosphingolipid 35. Zhang XK, Gallant S, Molano I, Moussa OM, Ruiz P, Spyropoulos DD, metabolism. PLoS ONE 8: e75175, 2013 Watson DK, Gilkeson G: Decreased expression of the Ets family tran- 28. Puri V, Jefferson JR, Singh RD, Wheatley CL, Marks DL, Pagano RE: scription factor Fli-1 markedly prolongs survival and significantly re- Sphingolipid storage induces accumulation of intracellular cholesterol duces renal disease in MRL/lpr mice. JImmunol173: 6481–6489, 2004 by stimulating SREBP-1 cleavage. J Biol Chem 278: 20961–20970, 2003 36. Atkinson C, Qiao F, Song H, Gilkeson GS, Tomlinson S: Low-dose tar- 29. Wang Z, Jiang T, Li J, Proctor G, McManaman JL, Lucia S, Chua S, Levi geted complement inhibition protects against renal disease and other M: Regulation of renal lipid metabolism, lipid accumulation, and glo- manifestations of autoimmune disease in MRL/lpr mice. JImmunol180: merulosclerosis in FVBdb/db mice with type 2 diabetes. Diabetes 54: 1231–1238, 2008 2328–2335, 2005 37. Bielawski J, Szulc ZM, Hannun YA, Bielawska A: Simultaneous quan- 30. Jiang T, Liebman SE, Lucia MS, Li J, Levi M: Role of altered renal lipid titative analysis of bioactive sphingolipids by high-performance liquid metabolism and the sterol regulatory element binding proteins in the chromatography-tandem mass spectrometry. Methods 39: 82–91, pathogenesis of age-related renal disease. Kidney Int 68: 2608–2620, 2006 2005 38. Bielawski J, Pierce JS, Snider J, Rembiesa B, Szulc ZM, Bielawska A: 31. Sun L, Halaihel N, Zhang W, Rogers T, Levi M: Role of sterol regulatory Comprehensive quantitative analysis of bioactive sphingolipids by element-binding protein 1 in regulation of renal lipid metabolism and high-performance liquid chromatography-tandem mass spectrometry. glomerulosclerosis in diabetes mellitus. J Biol Chem 277: 18919– Methods Mol Biol 579: 443–467, 2009 18927, 2002 39. Nowling TK, Fulton JD, Chike-Harris K, Gilkeson GS: Ets factors and a 32. Oates JC, Shaftman SR, Self SE, Gilkeson GS: Association of serum newly identified polymorphism regulate Fli1 promoter activity in lym- nitrate and nitrite levels with longitudinal assessments of disease ac- phocytes. Mol Immunol 45: 1–12, 2008 tivity and damage in systemic lupus erythematosus and lupus nephritis. 40. Jones EE, Powers TW, Neely BA, Cazares LH, Troyer DA, Parker AS, Arthritis Rheum 58: 263–272, 2008 Drake RR: MALDI imaging mass spectrometry profiling of proteins and 33. Hochberg MC: Updating the American College of Rheumatology re- lipids in clear cell renal cell carcinoma. Proteomics 14: 924–935, 2014 vised criteria for the classification of systemic lupus erythematosus. 41. Jones EE, Dworski S, Canals D, Casas J, Fabrias G, Schoenling D, Arthritis Rheum 40: 1725, 1997 Levade T, Denlinger C, Hannun YA, Medin JA, Drake RR: On-tissue 34. Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang CH; The localization of ceramides and other sphingolipids by MALDI mass Committee on Prognosis Studies in SLE: Derivation of the SLEDAI. A spectrometry imaging. Anal Chem 86: 8303–8311, 2014

J Am Soc Nephrol 26: 1402–1413, 2015 Glycosphingolipids in Lupus Nephritis 1413