Alterations in β-Cell Sphingolipid Profile Associated with ER Stress and iPLA2β: Another Contributor to β-Cell Apoptosis in Type 1 Diabetes

Type 1 diabetes (T1D) development, in part, is due to ER stress-induced β-cell apoptosis. Activation of the Ca2+-independent phospholipase A2 beta (iPLA2β) leads to the generation of pro-inflammatory eicosanoids, which contribute to β-cell death and T1D. ER stress induces iPLA2β-mediated generation of pro-apoptotic ceramides via neutral sphingomyelinase (NSMase). To gain a better understanding of the impact of iPLA2β on sphingolipids (SLs), we characterized their profile in β-cells undergoing ER stress. ESI/MS/MS analyses followed by ANOVA/Student’s t-test were used to assess differences in sphingolipids molecular species in Vector (V) control and iPLA2β-overexpressing (OE) INS-1 and Akita (AK, spontaneous model of ER stress) and WT-littermate (AK-WT) β-cells. As expected, iPLA2β induction was greater in the OE and AK cells in comparison with V and WT cells. We report here that ER stress led to elevations in pro-apoptotic and decreases in pro-survival sphingolipids and that the inactivation of iPLA2β restores the sphingolipid species toward those that promote cell survival. In view of our recent finding that the SL profile in macrophages—the initiators of autoimmune responses leading to T1D—is not significantly altered during T1D development, we posit that the iPLA2β-mediated shift in the β-cell sphingolipid profile is an important contributor to β-cell death associated with T1D.


Introduction
Diabetes is a consequence of pancreatic islet β-cell dysfunction and/or reduced peripheral insulin sensitivity, and both type 1 and type 2 diabetes (T1D and T2D) are associated with β-cell apoptosis [1,2]. Therefore, it is important to elucidate the underlying involved mechanisms so that treatment regimens can be developed to protect the β-cells and prevent or delay diabetes development.
In addition to the extrinsic (receptor-mediated) and intrinsic (mitochondrial) apoptotic pathways, ER stress is recognized as an inducer of apoptosis in several disease states, including diabetes [3]. Reports from both experimental models and clinical settings have linked diabetes development with ER stress-induced β-cell apoptosis [4], and studies in our laboratory revealed an important role for the group VIA Ca 2+ -independent phospholipase . RT-qPCRs: 18S served as internal control, and fold-change in mRNA was calculated using the delta CT method (ΔCT) and expressed relative to V controls: (A) ER stress marker GRP78, (B) iPLA2β mRNA (inset, iPLA2β protein); and (C) NSMase2 mRNA. Data are presented as mean ± SEM, n = 3 per group. (* Significantly different from V control, p < 0.05.).
Ceramide molecular species were increased in the OE cells relative to V cells, as reflected by the higher pool of ceramides in the OE cells ( Figure 2A). The induction of ER stress in V cells also resulted in an increase in the ceramide pool (in the VT cells) relative to untreated controls (V cells). As NSMase hydrolyzes sphingomyelins to generate ceramides, we investigated whether the observed increase in the ceramide pool was accompanied by changes in sphingomyelins. Indeed, we found similar decreases in the total pool of sphingomyelins ( Figure 2B) in both models of increased ER stress and iPLA 2 β (OE and VT cells) relative to untreated controls (V cells).
Since ceramides can be converted to various sphingolipids that manifest opposite biological activity, we assessed their abundances in the various cell models. Following glycosylation, ceramides can form monohexosyl ceramides (MHCs) [48]. Here, we report an increase in the MHC pool in both the OE and VT groups relative to the V control cells ( Figure 2C). Ceramides can also be phosphorylated to generate inflammatory C1P, which can activate cPLA 2 α and lead to the hydrolysis of arachidonic acid and PGE 2 formation [49]. Similarly, it can be speculated that C1P activates iPLA 2 β to generate lipid mediators that serve to amplify β-cell apoptosis. However, we report that the C1P pools in OE and VT cells were not significantly different from the corresponding pool in the untreated V group ( Figure 2D). Both non-phosphorylated and phosphorylated So and Sa sphingolipid species were decreased in the VT group relative to V ( Figure 2E); however, they were all higher in the OE group relative to the VT group. Collectively, these findings suggest that in general, the increased expression of iPLA 2 β and NSMase2 ( Figure 1) promotes alterations in the sphingolipids profile to favor the generation of lipids recognized to be detrimental and pro-apoptotic, and away from those that may be protective and anti-apoptotic. Data are presented as mean ± SEM pmols lipid/nmol total sphingolipid). One-Way ANOVA was utilized for statistical analysis. (A-D), * OE group significantly different from V and # VT group significantly different from V; p < 0.05; (E), * p < 0.05, † p < 0.01, ∆ p < 0.005, ∂ p < 0.0005, # p <0.0001, n = 3 per group).
Ceramide molecular species were increased in the OE cells relative to V cells, as reflected by the higher pool of ceramides in the OE cells ( Figure 2A). The induction of ER stress in V cells also resulted in an increase in the ceramide pool (in the VT cells) relative to untreated controls (V cells). As NSMase hydrolyzes sphingomyelins to generate ceramides, we investigated whether the observed increase in the ceramide pool was accompanied by changes in sphingomyelins. Indeed, we found similar decreases in the total pool of sphingomyelins ( Figure 2B) in both models of increased ER stress and iPLA2β (OE and VT cells) relative to untreated controls (V cells).
Since ceramides can be converted to various sphingolipids that manifest opposite biological activity, we assessed their abundances in the various cell models. Following glycosylation, ceramides can form monohexosyl ceramides (MHCs) [48]. Here, we report an increase in the MHC pool in both the OE and VT groups relative to the V control cells ( Figure 2C). Ceramides can also be phosphorylated to generate inflammatory C1P, which Data are presented as mean ± SEM pmols lipid/nmol total sphingolipid). One-Way ANOVA was utilized for statistical analysis. (A-D), * OE group significantly different from V and # VT group significantly different from V; p < 0.05; (E), * p < 0.05, † p < 0.01, ∆ p < 0.005, ∂ p < 0.0005, # p < 0.0001, n = 3 per group). Since the above observations were made in INS-1 cells that were genetically modified to express higher iPLA 2 β or chemically induced (thapsigargin, T) to exhibit ER stress, we sought to examine whether spontaneous ER stress in β-cells promotes similar changes in the sphingolipids profile. The Akita β-cells, due to a spontaneous mutation in the Ins2 gene that leads to the generation and accumulation of pre-proinsulin in the ER, develop spontaneous ER stress [50,51]. We previously demonstrated that this occurs in the Akita β-cells without chemical intervention [52]. Here, immunoblotting and qRT-PCR analyses confirm the induction of GRP78 ( Figure 3A), which is associated with the higher expression of both iPLA 2 β ( Figure 3A) and NSMase2 ( Figure 3C) in the AK, relative to wild-type (WT) cells. These analyses confirm that Akita cells, even under basal conditions, exhibit ER stress and that this is associated with higher iPLA 2 β and NSMase2. to express higher iPLA2β or chemically induced (thapsigargin, T) to exhibit ER stress, we sought to examine whether spontaneous ER stress in β-cells promotes similar changes in the sphingolipids profile. The Akita β-cells, due to a spontaneous mutation in the Ins2 gene that leads to the generation and accumulation of pre-proinsulin in the ER, develop spontaneous ER stress [50,51]. We previously demonstrated that this occurs in the Akita β-cells without chemical intervention [52]. Here, immunoblotting and qRT-PCR analyses confirm the induction of GRP78 ( Figure 3A), which is associated with the higher expression of both iPLA2β ( Figure 3A) and NSMase2 ( Figure 3C) in the AK, relative to wild-type (WT) cells. These analyses confirm that Akita cells, even under basal conditions, exhibit ER stress and that this is associated with higher iPLA2β and NSMase2.

Sphingolipids Profiles in WT and Akita (AK) β-cells
Next, LC-ESI/MS/MS analyses, similar to the ones used with INS-1 cells, were performed to assess the relative abundances of sphingolipids (Appendix A) in the WT and AK cells (Figure 4). We report that associated with the spontaneous ER stress and higher basal iPLA2β levels, we observed increases in the pool of ceramides in the AK cells, relative to WT cells ( Figure 4A). In view of the higher basal NSMase2, we examined for changes in sphingomyelins and found that the pool of sphingomyelin molecular species was decreased in the AK cells, relative to WT cells ( Figure 4B). Further, the pools of MHC molecular species ( Figure 4C) and the pro-apoptotic/inflammatory [53] C1P species (Figure 4D) were higher in the AK cells relative to WT cells. In general, the qualitative changes in AK cells were analogous to those described above in OE and VT cells, relative to untreated controls (V cells), reinforcing the suggestion that both iPLA2β increases and ER stress modify the sphingolipids profile to favor pro-apoptotic species in β-cells.

Sphingolipids Profiles in WT and Akita (AK) β-Cells
Next, LC-ESI/MS/MS analyses, similar to the ones used with INS-1 cells, were performed to assess the relative abundances of sphingolipids (Appendix A) in the WT and AK cells (Figure 4). We report that associated with the spontaneous ER stress and higher basal iPLA 2 β levels, we observed increases in the pool of ceramides in the AK cells, relative to WT cells ( Figure 4A). In view of the higher basal NSMase2, we examined for changes in sphingomyelins and found that the pool of sphingomyelin molecular species was decreased in the AK cells, relative to WT cells ( Figure 4B). Further, the pools of MHC molecular species ( Figure 4C) and the pro-apoptotic/inflammatory [53] C1P species ( Figure 4D) were higher in the AK cells relative to WT cells. In general, the qualitative changes in AK cells were analogous to those described above in OE and VT cells, relative to untreated controls (V cells), reinforcing the suggestion that both iPLA 2 β increases and ER stress modify the sphingolipids profile to favor pro-apoptotic species in β-cells.

Effects of Inactivating iPLA 2 β on the Sphingolipids Profile in Akita (AK) β-Cells
As the changes in sphingolipids described in INS-1 and AK β-cells correlated with increased iPLA 2 β, we examined whether they could be reversed by the inactivation of iPLA 2 β. Further, because basal iPLA 2 β expression is higher in AK cells and they exhibit spontaneous ER stress, we examined the effects of S-BEL in AK cells in the absence and presence of thapsigargin (T). We find that both iPLA 2 β ( Figure 5A) and NSMase2 ( Figure 5B) are induced in the presence of thapsigargin (T), in both WTT and AKT cells, relative to WT and AK control cells, respectively. Molecules 2021, 26, x FOR PEER REVIEW 6 of 17

Effects of Inactivating iPLA2β on the Sphingolipids Profile in Akita (AK) β-cells
As the changes in sphingolipids described in INS-1 and AK β-cells correlated with increased iPLA2β, we examined whether they could be reversed by the inactivation of iPLA2β. Further, because basal iPLA2β expression is higher in AK cells and they exhibit spontaneous ER stress, we examined the effects of S-BEL in AK cells in the absence and presence of thapsigargin (T). We find that both iPLA2β ( Figure 5A) and NSMase2 ( Figure  5B) are induced in the presence of thapsigargin (T), in both WTT and AKT cells, relative to WT and AK control cells, respectively. We next assessed the effects of chemical inhibition of iPLA2β by S-BEL on the sphingolipid profile in AK cells. The cells were pre-treated with S-BEL for 30 min prior to treatment with T. After 8 h, the cells were processed for sphingolipid analyses, as above, and the fold-change in sphingolipid species abundances in AKT cells relative to vehicle-

Effects of Inactivating iPLA2β on the Sphingolipids Profile in Akita (AK) β-cells
As the changes in sphingolipids described in INS-1 and AK β-cells correlated with increased iPLA2β, we examined whether they could be reversed by the inactivation of iPLA2β. Further, because basal iPLA2β expression is higher in AK cells and they exhibit spontaneous ER stress, we examined the effects of S-BEL in AK cells in the absence and presence of thapsigargin (T). We find that both iPLA2β ( Figure 5A) and NSMase2 ( Figure  5B) are induced in the presence of thapsigargin (T), in both WTT and AKT cells, relative to WT and AK control cells, respectively. We next assessed the effects of chemical inhibition of iPLA2β by S-BEL on the sphingolipid profile in AK cells. The cells were pre-treated with S-BEL for 30 min prior to treatment with T. After 8 h, the cells were processed for sphingolipid analyses, as above, and the fold-change in sphingolipid species abundances in AKT cells relative to vehicle- We next assessed the effects of chemical inhibition of iPLA 2 β by S-BEL on the sphingolipid profile in AK cells. The cells were pre-treated with S-BEL for 30 min prior to treatment with T. After 8 h, the cells were processed for sphingolipid analyses, as above, and the fold-change in sphingolipid species abundances in AKT cells relative to vehicletreated cells (AK) was compared with those in AK cells treated with T + S-BEL (AKTB) relative to AKT. As expected, the ceramide molecular species were increased by thapsigargin (AKT/AK > 1, Figure 6A), but this effect was reversed by S-BEL (AKTB/AKT ≤ 1) in all species except 18:1/24:0. Concurrently, thapsigargin-induced decreases in sphingomyelin molecular species (AKT/AK < 1, Figure 6B) were reversed by S-BEL. Similarly, various MHC species increased with thapsigargin ( Figure 6C) were also decreased and reversed by S-BEL (AKTB/AKT ≥ 1). However, the C1P species exhibited a differential regulation, where two species (18:1/16:0 and 18:1/24:0) were modestly decreased following thapsigargin treatment (AKT/AK > 1) but only 18:1/24:0 was rescued by S-BEL (AKTB/AKT > 1) ( Figure 6D). While no differences were evident in Sa or Sa1P among the groups, so was increased in the AK, relative to AKT, and the ratio of So1P/So decreased with S-BEL (AKTB). gin (AKT/AK > 1, Figure 6A), but this effect was reversed by S-BEL (AKTB/AKT ≤ 1) in all species except 18:1/24:0. Concurrently, thapsigargin-induced decreases in sphingomyelin molecular species (AKT/AK < 1, Figure 6B) were reversed by S-BEL. Similarly, various MHC species increased with thapsigargin ( Figure 6C) were also decreased and reversed by S-BEL (AKTB/AKT ≥ 1). However, the C1P species exhibited a differential regulation, where two species (18:1/16:0 and 18:1/24:0) were modestly decreased following thapsigargin treatment (AKT/AK > 1) but only 18:1/24:0 was rescued by S-BEL (AKTB/AKT > 1) ( Figure 6D). While no differences were evident in Sa or Sa1P among the groups, so was increased in the AK, relative to AKT, and the ratio of So1P/So decreased with S-BEL (AKTB).

Discussion
It has long been established that sphingolipids play important and dynamic roles in cellular physiology and that their levels are maintained under tightly regulated homoeostasis processes via a variety of highly specialized enzymes [54][55][56][57]. We previously demonstrated that ER stress-mediated β-cell apoptosis is mediated, in part, through the activation of iPLA2β and that an increased expression of iPLA2β amplifies the susceptibility of β-cells to ER stress and subsequent β-cell apoptosis [4,5,8,16,17]. Increases in iPLA2β and ER stress were also associated with accumulations in ceramides, which contributed to β-cell death by triggering the intrinsic apoptotic pathway [4,18]. These observations provided evidence for the involvement of an iPLA2β/ceramide axis in promoting β-cell dysfunction and apoptosis, which are key contributors to the development of diabetes [1,2]. As reports from various experimental models and clinical settings suggested that ER stress contributes to β-cell apoptosis during the evolution of diabetes [4], we set out to assess the link between ER stress, iPLA2β, and β-cell sphingolipids in the present study.
The roles of ceramides are related to their involvement in cell survival and are mainly considered to be pro-apoptotic [58][59][60][61]. Ceramide generation can occur through multiple pathways: (a) de novo, initiated by the condensation of serine and palmitoyl CoA, which

Discussion
It has long been established that sphingolipids play important and dynamic roles in cellular physiology and that their levels are maintained under tightly regulated homoeostasis processes via a variety of highly specialized enzymes [54][55][56][57]. We previously demonstrated that ER stress-mediated β-cell apoptosis is mediated, in part, through the activation of iPLA 2 β and that an increased expression of iPLA 2 β amplifies the susceptibility of β-cells to ER stress and subsequent β-cell apoptosis [4,5,8,16,17]. Increases in iPLA 2 β and ER stress were also associated with accumulations in ceramides, which contributed to β-cell death by triggering the intrinsic apoptotic pathway [4,18]. These observations provided evidence for the involvement of an iPLA 2 β/ceramide axis in promoting β-cell dysfunction and apoptosis, which are key contributors to the development of diabetes [1,2]. As reports from various experimental models and clinical settings suggested that ER stress contributes to β-cell apoptosis during the evolution of diabetes [4], we set out to assess the link between ER stress, iPLA 2 β, and β-cell sphingolipids in the present study.
The roles of ceramides are related to their involvement in cell survival and are mainly considered to be pro-apoptotic [58][59][60][61]. Ceramide generation can occur through multiple pathways: (a) de novo, initiated by the condensation of serine and palmitoyl CoA, which is catalyzed by the rate-limiting enzyme serine palmitoyl transferase (SPT) [62]; (b) sphingomyelin hydrolysis by sphingomyelinases [48,63]; or (c) salvage pathway involving ceramidase [64]. Ceramides can be further converted to C1P by ceramide kinase or to MHCs by glucosylceramide synthase or ceramide galactosyl transferase [48]. Our work [14][15][16][17][18] indicates that the accumulation of ceramides during β-cell apoptosis due to ER stress occurs through the hydrolysis of sphingomyelins by NSMase2. Therefore, we aimed at determining whether the overall sphingolipid profile is modulated by iPLA 2 β in β-cells undergoing ER stress.
In the present study, MS protocols were applied to INS-1 cells and Akita β-cells to discern the impact of ER stress and iPLA 2 β expression on the abundances of various sphingolipids. The INS-1 cells are widely used in studies of β-cell function and survival and the availability in our laboratory of genetically modified iPLA 2 β-overexpressing (OE) INS-1 cells permitted examination of the effects of higher iPLA 2 β expression in the absence of chemically induced ER stress. In addition, parallel comparisons with thapsigargin-treated empty-vector (V) INS-1 cells enabled assessment of the effects of ER stress induction in the absence of genetic modification of iPLA 2 β expression. The Akita β-cells (AK) provided a model of spontaneous ER stress that also expresses higher basal iPLA 2 β [16], thus precluding consideration of the effects of chemical intervention or genetic modification.
Prior to the lipid analyses, immunoblotting and RT-PCR analyses confirmed (a) increased basal expression of iPLA 2 β, NSmase2, and GRP78 in the OE and AK cells, relative to V and WT cells, respectively and (b) induction of ER stress, iPLA 2 β, NSMase2, and GRP78 following exposure to thapsigargin in VT and AKT cells, relative to V and AK cells, respectively. Subsequently, LC-ESI/MS/MS analyses were used to identify molecular species of ceramides, sphingomyelins, monohexosyl ceramides (MHCs), ceramide-1-phosphate (C1P), sphinganine (Sa), sphingosine (So), and the So phosphate derivatives So1P.
Under basal conditions, comparisons of OE vs. V cells reveal that higher iPLA 2 β expression leads to (a) increases in the pro-apoptotic [65] ceramides, (b) decreases in antiapoptotic [17] sphingomyelins, (c) increases in the primarily apoptotic [48] MHCs, (d) no change in inflammatory and apoptotic [66] C1P, and (e) no change in anti-apoptotic [53] So1P. Comparisons of VT vs. V cells following induction of ER stress, as reflected by increased expression of ER stress marker GRP78, and of iPLA 2 β in the VT cells revealed changes in VT sphingolipids that were similar to those in the OE cells.
Similar comparisons between the WT and AK cells under basal conditions revealed relative changes in the AK cells that mirrored those in the OE cells including (a) increases in ceramides, (b) decreases in sphingomyelins, (c) and increases in the MHCs. However, differences were noted in the C1P molecular species (unchanged in OE but increased in AK) and in the So1P/So ratio (unchanged in OE but increased in AKT). This is an important observation, as we also noted an increase in So1P/So ratio in the non-obese diabetic (NOD) mice, an autoimmune model of T1D [67]. The NOD mice develop insulitis at 4-6 weeks of age and spontaneous T1D starting at 16-18 weeks of age [68]. β-Cell death and the development of T1D in the NOD is preceded by the induction of ER stress in the β-cells [69,70]. Given that So1P has been identified as a lipid that promotes T-cell migration and retention in inflamed tissues [71], it is likely that it could manifest deleterious effects in β-cells also.
Exposure to the ER stressor, thapsigargin, amplified the changes in the sphingolipids in AK cells, suggesting that ER stress and consequential downstream processes impact the sphingolipids profile. Since iPLA 2 β is also further induced, we considered the possibility that its activation contributes to the observed changes in sphingolipids. To assess this, AK cells were pre-treated with S-BEL prior to thapsigargin exposure, and this was found to reverse nearly all changes in the sphingolipids, with the exception of d18:1/16:0 C1P, toward control abundances. These findings are consistent with the involvement of iPLA 2 β activation in the regulation of β-cell sphingolipids.
While our data demonstrate similar modifications in the sphingolipids profile with higher iPLA 2 β and ER stress, differences were noted with respect to specific molecular species that were affected. Such discrepancies are not unexpected considering the inherent differences between the models studied, which include the (a) the high basal but long-term expression of iPLA 2 β in the iPLA 2 β-overexpressing INS-1 cells, which promotes ER stress vs. (b) the transient, short-term, chemically induced (thapsigargin) ER stress, which promotes iPLA 2 β expression in the Vector INS-1 cells vs. (c) the basal spontaneous and prolonged ER stress, which is associated with higher iPLA 2 β expression in the AK cells. In view of these differences, it is plausible that the sphingolipids-generating pathways are regulated dissimilarly, giving rise to variations in specific molecular species changes. Furthermore, recent reports suggest that the fate of different sphingolipids and activities of their converting enzymes [72][73][74][75] may be cell-specific and even compartment-specific [31,48]. As reviewed extensively elsewhere [34,48,72,[76][77][78][79][80][81], the sphingolipids-generating pathways are quite complex, involving inter-conversion between the sphingolipids and occurring in multiple subcellular organelles. In this regard, while iPLA 2 β is localized in the cytosol under basal conditions, ER stress causes a temporal re-distribution of iPLA 2 β to the ER, the mitochondria, and the nucleus [5,82]. The compartmentalized and subcellular localization (i.e., endosome/lysosome, endoplasmic reticulum, secretory granules; see Figure 7) of different pools of sphingolipids [48] raises the likelihood that the sphingolipids in different subcellular organelles could be affected at different rates depending on the duration of ER stress and/or iPLA 2 β induction and subcellular mobilization.
the cytosol under basal conditions, ER stress causes a temporal re-distribution of iPLA2β to the ER, the mitochondria, and the nucleus [5,82]. The compartmentalized and subcellular localization (i.e., endosome/lysosome, endoplasmic reticulum, secretory granules; see Figure 7) of different pools of sphingolipids [48] raises the likelihood that the sphingolipids in different subcellular organelles could be affected at different rates depending on the duration of ER stress and/or iPLA2β induction and subcellular mobilization.
Although our studies provided evidence of significant changes in sphingolipids associated with altered iPLA2β expression and the development of ER stress, several questions remain to be addressed. These include the following: (1) identification of the hexosyl component of MHCs; because this was not discernible from the present analyses, it will be important to determine if the species increased are apoptotic glucosyl-ceramides, (2) in addition to NSMase, what and how are the other enzymes of the sphingolipids generating pathways altered by iPLA2β activation and/or ER stress, (3) which molecular species of the various sphingolipids are most prone to be affected and influenced by duration of elevated iPLA2β or ER stress, (4) what specific sphingolipid changes correlate with β-cell apoptosis, and importantly (5) whether alterations in sphingolipids contribute to β-cell apoptosis in the context of diabetes development These labor-intensive issues and other questions of the involved underlying mechanisms are currently being addressed in our laboratory utilizing autoimmune rodent models of type 1 diabetes.  Although our studies provided evidence of significant changes in sphingolipids associated with altered iPLA 2 β expression and the development of ER stress, several questions remain to be addressed. These include the following: (1) identification of the hexosyl component of MHCs; because this was not discernible from the present analyses, it will be important to determine if the species increased are apoptotic glucosyl-ceramides, (2) in addition to NSMase, what and how are the other enzymes of the sphingolipids generating pathways altered by iPLA 2 β activation and/or ER stress, (3) which molecular species of the various sphingolipids are most prone to be affected and influenced by duration of elevated iPLA 2 β or ER stress, (4) what specific sphingolipid changes correlate with β-cell apoptosis, and importantly (5) whether alterations in sphingolipids contribute to β-cell apoptosis in the context of diabetes development These labor-intensive issues and other questions of the involved underlying mechanisms are currently being addressed in our laboratory utilizing autoimmune rodent models of type 1 diabetes.

Materials
Materials and (sources) were as follows: Akita (AK) and wild-type (WT) β-cells were generously provided by Dr. Akio Koizumi (Kyoto Graduate School of Medicine, Kyoto, Japan); Ultra-Performance Liquid Chromatography was a Shimadzu Nexera 30-AD system, and the mass spectrometer was a 5500 QTRAP from ABSciex
In some experiments, the cells were pre-treated for 30 min with the iPLA 2 β inhibitor, S-BEL (10 µM). Then, the cells were washed prior to treatment with vehicle or thapsigargin. All experimental protocols included DMSO vehicle-only treated triplicate controls.

LC-Electrospray Ionization (ESI)-MS/MS Analyses of Sphingolipids
Cells were plated separately in 10 cm cell culture dishes to provide approximately 1 × 10 6 cells at harvesting. Following an 8 h treatment period with DMSO vehicle or thapsigargin, the cell culture dishes were placed on ice, washed twice with phosphate-buffered saline (PBS), and harvested by scraping in 200 µL of PBS. The cells were immediately frozen and stored at −80 • C. At the time of analyses, the cells were thawed on ice followed by sonication to obtain a homogenous mixture, and lipids were extracted using a modified Bligh and Dyer method and analyzed, as described [84]. Briefly, to 200 µL of the cells in PBS, 1.5 mL of 2:1 methanol/chloroform was added. The samples were spiked with internal standards consisting of 50 pmol each of d17:1 sphingosine, d17:0 sphinganine, d17:1 sphingosine-1-phosphate, d17:0 sphinganine-1-phosphate, d18:1/12:0 ceramide-1-phosphate, d18:1/12:0 sphingomyelin, d18:1/12:0 ceramide, and d18:1/12:0 monohexosyl ceramide. The mixture thus obtained was sonicated to disperse the cell clumps and incubated for 6 h at 48 • C. Following incubation, the samples were sonicated, followed by centrifugation to separate particulates. Then, the extracts were dried under vacuum and reconstituted by sonicating in 500 µL of methanol followed by incubation at 48 • C for 15 min, vortexing, and incubation for an additional 15 min at 48 • C. Then, the samples were centrifuged to separate particulates, and 10 µL was used for analysis. The lipids were separated using a Kinetix C18 column (50 × 2.1 mm, 2.6 µm) on a Nexera ultra-high-performance liquid chromatography system and eluted using a linear gradient (solvent A, 58:41:1 CH 3 OH/water/HCOOH 5 mM ammonium formate; solvent B, 99:1 CH 3 OH/HCOOH 5 mM ammonium formate, 20-100% B in 3.5 min and at 100% B for 4.5 min at a flow rate of 0.4 mL/min at 60 • C). Electrospray ionization (ESI) with tandem mass spectroscopy (MS/MS) on a 5500 QTRAP instrument was utilized for the detection and quantitation of analytes in the positive ion mode. Multiple reaction monitoring transitions utilized for the analytes are listed in Appendix A.

Statistical Analyses
Data were converted to mean ± standard error of the means (SEMs), Student's t-test was applied to determine significant differences (p < 0.05) between two samples, and One-Way and Two-Way ANOVA statistical analyses were applied to determine significant differences (at p < 0.05) between more than two sample groups.

Conclusions
In summary, our data demonstrate for the first time that iPLA 2 β and ER stress in β-cells, in concert, modulate the sphingolipids profile in β-cells and that the consequential redistribution between the various sphingolipids favors apoptosis over the survival of the βcells. To our knowledge, molecular species assessments of ceramide-derived sphingolipids have not been examined in detail in β-cells, and our findings provide the basis for future studies that will address important issues in a cell system where the dysregulation of sphingolipids generation could impact its survival, and as a consequence, could play a critical role in the evolution of diabetes or other inflammatory diseases [85,86].

Conflicts of Interest:
The authors declare no conflict of interest.

Species
Precursor Ion (m/z) Product Ion (m/z) DP CE SCIEX 5500 QTRAP mass spectrometer settings for reverse phase chromatographic separation of various sphingolipids. DP: declustering potential, CE: collision energy, EP: entrance potential, CXP: cell exit potential. Transitions and settings were determined manually via direct infusion. Settings and transitions were chosen based on the best signal obtained during infusion using manual tuning in SCIEX Analyst software.