Inhibition of Ca²⁺-Independent Phospholipase A₂β (iPLA₂β) Ameliorates Islet Infiltration and Incidence of Diabetes in NOD Mice

Type 1 diabetes is a consequence of autoimmune destruction of β-cells, involving activation of cellular immunity leading to infiltration of islets by inflammatory immune cells (1,2). An understudied area is the role lipid signals play in this process. Our studies indicate that the group VIA Ca²⁺-independent phospholipase A₂β (iPLA₂β) is induced under a diabetic milieu and that mitigation of iPLA₂β attenuates β-cell death (3–6). These findings raise the possibility that iPLA₂β contributes to β-cell death and consequential diabetes development.

The iPLA₂β enzyme is a member of the family of PLA₂s (7) that catalyzes hydrolysis of the sn-2 substituent from glycerophospholipid substrates to yield a free fatty acid and a 2-lysophospholipid (8), is widely expressed, and participates in numerous biological processes (9). In the islet, iPLA₂β is predominantly localized in β-cells (4,10), which are enriched in arachidonate-containing phospholipid substrates. Following its hydrolysis by iPLA₂β, arachidonic acid can be metabolized by cyclooxygenase (COX) and lipoxygenase (LOX) to bioactive eicosanoids, many of which are recognized promoters of immune responses (11,12). In view of a described link between iPLA₂β and β-cell apoptosis and evidence of its induction in clinical and experimental diabetes (13–15), we considered the possibility that iPLA₂β inhibition preserves β-cell mass and mitigates diabetes development.

To date, the S-enantiomer of bromoenol lactone (S-BEL), with 10 times greater specificity for iPLA₂β than iPLA₂γ (16), has been used to discern the role of iPLA₂β in biological processes. However, several features of BEL, including irreversible inhibition of iPLA₂s, nonspecific inhibition of serine-based non-iPLA₂ enzymes, and cytotoxicity (9,17), render it less feasible for use as an in vivo inhibitor of iPLA₂β.

Fluoroketone (FK) compounds have been synthesized that exhibit selectivity for iPLA₂s versus sPLA₂s or cPLA₂s (18). We demonstrated that FKGK18 is a reversible inhibitor that is 100-fold more specific for iPLA₂β than iPLA₂γ, does not nonspecifically inhibit other serine proteases, is not cytotoxic, and inhibits β-cell apoptosis (19). These observations suggest that FKGK18 is amenable for in vivo administrations and led us to assess its impact on autoimmune diabetes development. The potential for beneficial effects of inhibiting iPLA₂β in vivo with FK compounds is evidenced by amelioration of ovarian cancer and central nervous system–related disorders, in the absence of systemic toxicity, with an earlier generation FKGK inhibitor (20–23).

We report here that FKGK18 administration to spontaneous diabetes–prone NOD mice decreases diabetes incidence, reduces islet infiltration, improves glucose homeostasis, and preserves β-cell area. Furthermore, FKGK18 inhibits immune cell function and adoptive transfer of diabetes. These findings suggest that iPLA₂β activation promotes immune responses and that iPLA₂β inhibition may be beneficial in ameliorating diabetes.

Islet RNA was prepared from 6–7-week-old female C57BL/6J, NOD, NOD.scid, and NOD.Rag mice and male NOD mice (The Jackson Laboratory, Bar Harbor, ME), and cDNA was prepared for quantitative RT-PCR (RT-qPCR) analyses as previously described (3). Primers were based on mouse sequences for iPLA₂β and internal control 18S (gene IDs 53357 and 19791, respectively).

Female NOD mice were generated and maintained according to University of Alabama at Birmingham Institutional Animal Care and Use Committee policies. Mice were administered intraperitoneally three times per week vehicle (PBS + 5% Tween 80) or FKGK18 [20 mg/kg body weight (20)] from 10 days until euthanized. Body weights and blood glucose levels, measured from tail vein blood samples (2 μL) with the Breeze 2 Blood Glucose Monitoring System (Bayer Healthcare, Mishawaka, IN), were recorded weekly. Diabetes incidence was based on two consecutive blood glucose readings ≥15.3 mmol/L, at which time the mouse was euthanized. By 30 weeks, all mice were euthanized, and blood was collected in BD Microtainer Tubes with serum separator for insulin measurements (ELISA Kit; Mercodia, Uppsala, Sweden).

Paraffin sections (10 μm) of pancreas, heart, liver, and kidney were stained with hematoxylin-eosin (H-E) for histological assessment. Islet images were captured, and total islet and noninfiltrated areas (pixels) were determined.

Submandibular vein blood from mice was obtained between 0 and 72 h following FKGK18 administration. To 100 μL of serum sample, 1 mL of chloroform and 50 μL of 1 mol/L aqueous hydrochloric acid were added and the mixture vortexed for 30 s, centrifuged for 5 min, and the organic layer transferred to a clean tube. The remaining aqueous layer was re-extracted with 1 mL of chloroform by the same procedure. The two organic layers were combined, dried, reconstituted in methanol, and analyzed for FKGK18 by high-resolution mass spectrometry, recorded with a quadrupole time-of-flight Bruker maXis impact instrument using negative ion mode electrospray ionization protocol. Data acquisition was carried out with Bruker Daltonics DataAnalysis 4.1.

Prostaglandin E₂ (PGE₂) metabolite levels in urine, collected from individual mice housed in metabolic cages over an 18-h period in the absence of food but free access to water, were determined by enzyme immunosorbent assay (Cayman Chemical, Ann Arbor, MI) and normalized to urine creatinine levels measured by colorimetric assay (Cell Biolabs, Inc., San Diego, CA).

At 25 weeks of age, overnight-fasted mice were administered glucose 2 g/kg body weight in filter-sterilized dH₂O i.p., and tail vein blood samples (2 μL) were collected over a 2-h period for glucose measurement. The mice had free access to water during this time.

Paraffin sections (10 μm) of pancreata were processed for immunostaining using an antigen retrieval protocol as previously described (24,25). Sections were incubated overnight at 4°C with 1° antibodies guinea pig anti-insulin (1:200) (Life Technologies, Carlsbad, CA), rat anti-CD4 (1:50) (Santa Cruz Biotechnology, Santa Cruz, CA), rat anti-CD8α (1:50) (Santa Cruz), or rabbit anti-B220 (1:100) (BD Biosciences, San Jose, CA) and subsequently with fluorescence-labeled 2° antibodies Texas Red and fluorescein isothiocyanate–conjugated antibodies (1:200) (Santa Cruz and Jackson ImmunoResearch, West Grove, PA) in the dark (2 h, room temperature). Nuclei were stained with Hoechst 2 μg/mL for 10 min, and the ratio of total insulin-stained islet region to H-E–stained pancreas section was used to calculate β-cell area.

Single-cell splenocyte suspensions were prepared from 8–12-week-old NOD mice. CD4⁺ T cells were purified using the BD IMag CD4 T Lymphocyte Enrichment Set - DM (BD Biosciences), and resting B cells were purified using BD IMag B Lymphocyte Enrichment Set - DM (BD Biosciences), per manufacturer’s instructions. CD4⁺ T cells (2.5 × 10⁵/well) were seeded in 96-well plates coated with αCD3 (1 μg/mL) (BioLegend, San Diego, CA) with media containing αCD28 (0.5 μg/mL) (BioLegend). Cytokine concentrations (48 and 72 h) in the supernatant were measured by ELISA (interleukin [IL]-2 and interferon-γ [IFN-γ] [BD Biosciences] and TNF-α [R&D Systems, Minneapolis, MN]) as previously described (26). B cells (2.5 × 10⁵/well) were seeded in 96-well plates for 72 h with media containing 1 μg/mL lipopolysaccharide (LPS) (Life Technologies) and 2 ng/mL IL-4 (R&D Systems) as previously described (27). IgG and IgM antibody production were measured in the supernatant from the B cells by ELISA (SouthernBiotech, Birmingham, AL) as previously described (28).

Immune cells were incubated for 3.5 h at 37°C in wells with 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) (Sigma-Aldrich, 20 μL of 5 mg/mL solution). Media were carefully aspirated and replaced with an equal volume of MTT solvent (4 mmol/L HCl and 0.1% Nonidet P-40 in isopropanol). The plate was agitated on a shaker in the dark for 15 min, and absorbance was read at 590 nm.

Single-cell splenocyte suspensions were prepared from 12-week-old male BDC2.5/NOD mice, as previously described (29). CD4⁺ T cells were purified, activated in six-well plates (5 × 10⁶ cells/well) coated with αCD3 and αCD28 (1 μg/mL each), and expanded with media containing 100 U/mL IL-2 for 72 h. They were further expanded into T-75 flasks (5 × 10⁶ cells/flask) with media containing 100 U/mL IL-2 for 72 h, collected, washed twice with Hanks’ balanced salt solution (HBSS), and diluted to 5 × 10⁶ cells/mL in HBSS. Recipient 6-week-old immunodeficient male NOD.scid mice were divided into control and T cells ± FKGK18–administered groups, and diabetes incidence was monitored for 14 days.

Pancreas section and islet images were captured on an Olympus IX81 microscope using cellSens Dimension software and analyzed using ImageJ software (National Institutes of Health).

Data were converted to mean ± SEM and Student t test was applied to determine significant differences between two groups (P < 0.05). Incidence of diabetes was plotted as a survival curve and analyzed by Mantel-Cox test.

RT-qPCR analyses revealed three- to fivefold higher iPLA₂β mRNA in islets from 6- to 7-week-old spontaneous diabetes-prone female NOD mice relative to age-matched diabetes-resistant female C57BL/6J, NOD.scid, and NOD.Rag and male NOD mice (Fig. 1A). Because NOD mice exhibit insulitis by 4–6 weeks of age and develop spontaneous diabetes by 10–20 weeks of age (30), these findings raise the possibility that an early increase in iPLA₂β contributes to diabetes development.

Female NOD mice were administered vehicle or FKGK18 starting at 10 days of age (Fig. 1B), and FKGK18 bioavailability was determined by high-resolution mass spectrometry. Standard FKGK18 (molecular weight 280 g/mol) generated a predominant parent [M-H]⁻ ion at charge/mass ratio (m/z) 279 (Fig. 1C) and electrospray ionization tandem mass spectrometry (ESI/MS/MS) analyses revealed signature ions at m/z 209.0970, arising from the loss of CHF₃ [M-H-CHF₃]⁻, and m/z 68.9945, representing [CHF₃-H]⁻, from the parent ion (Fig. 1D). Analyses of NOD serum from mice treated with FKGK18 for >15 weeks generated a similar spectrum profile (Fig. 1E) and revealed maximum levels of FKGK18 between 2 and 6 h (∼300 ng/mL), which decreased to 70–280 ng/mL by 20–24 h, 0–30 ng/mL by 44–48 h, and were undetectable by 72 h after the last injection. Because the reversible nature of FKGK18 precludes in vivo iPLA₂β activity measurement, urinary 13,14-dihydro-15-keto metabolite of PGE₂ (PGEM) levels were measured and found to be decreased up to 42 h but returned to vehicle levels between 48 and 66 h (Fig. 1F). Taken together, it was deduced that the in vivo inhibitory effects of FKGK18 20 mg/kg i.p. persist for nearly 48 h.

The FKGK18 regimen did not affect body weight (Fig. 2A) or promote discernible tissue damage (Fig. 2B) compared with vehicle. However, diabetes incidence was significantly reduced with FKGK18 (Fig. 2C). Consistent with literature reports of an ∼80% incidence in female NOD mice (30), seven of nine (78%) vehicle-treated mice became diabetic by 25 weeks. In contrast, only 1 of 11 FKGK18-treated mice became diabetic by 25 weeks. At this age, five mice in each group were euthanized for further analyses. Among the remaining six in the FKGK18 group, one additional mouse became diabetic at 28 weeks. The lower diabetes incidence in the FKGK18 group was associated with twofold higher circulating insulin compared with the vehicle group (Fig. 2D). Intraperitoneal glucose tolerance testing revealed that although basal overnight-fasted blood glucose levels were similar in the vehicle and FKGK18 groups, they remained lower in the FKGK18 group over the test period (Fig. 2E). This was reflected by a 31% reduction in the area under curve of the FKGK18 group (Fig. 2F), indicating better glucose homeostasis relative to the vehicle group.

As expected, significant infiltration was evident in islets from the vehicle group (Fig. 3A, left panels). In contrast, insulitis was markedly reduced in the FKGK18 group and when present, was predominantly in the peri-islet region (Fig. 3A, right panels). Percent islet infiltration, determined by measuring total islet and noninfiltrated areas, revealed 54% infiltration of islets in the vehicle group (128 from 9 mice) but only 35% infiltration of islets in the FKGK18 group (195 from 11 mice) (Fig. 3B). Breakdown of the infiltration spectrum (none to complete [0–100%]) revealed twice as many infiltrate-free islets and one-half as many islets with 75–100% infiltration in the FKGK18 group compared with the vehicle group (Fig. 3C). Consistent with the improved islet integrity, insulin staining area (Fig. 3D) relative to total pancreas section area was sevenfold greater (Fig. 3E) in the FKGK18 group, suggesting greater preservation of residual β-cell mass relative to the vehicle group.

In view of reduced islet infiltration in the FKGK18 group, immunofluorescence analyses were used to estimate abundance in the infiltrate of CD4⁺ and CD8⁺ T cells and B cells, key components of autoimmune diabetes development (2). As shown in Fig. 4A, evidence of CD4⁺ T cells (i and ii, iii and iv), CD8α⁺ T cells (v and vi, vii and viii), and B cells (ix and x, xi and xii) was found in islets from both the vehicle and the FKGK18 groups. However, CD4⁺ T cells and B cells per islet (Fig. 4B and D), but not CD8α⁺ T cells (Fig. 4C), were significantly decreased in the FKGK18 group relative to the vehicle group, suggesting that islet infiltration by certain immune cells is impeded with inactivation of iPLA₂β.

PCR analyses revealed that NOD splenocytes, CD4⁺ T cells, B cells, and bone marrow–derived macrophages all express iPLA₂β mRNA (Fig. 5A), raising the possibility that iPLA₂β activation affects their functionality. Initial studies revealed that FKGK18 exposure for up to 72 h does not compromise cell viability (Supplementary Figs. 1A and 2A) and reduces cytokine generation from mixed splenocytes but not macrophages (Supplementary Figs. 1B–D and 2B–C).

The effects of FKGK18 (10⁻⁷–10⁻⁵ mol/L, 72 h) on CD4⁺ T-cell and B-cell function were assessed next. Exposure of CD4⁺ T cells, activated with αCD3 and αCD28 (31), to FKGK18 had no effect on cell viability (Fig. 5B) or generation of IFN-γ and IL-2 (Fig. 5C), but TNF-α generation was significantly decreased (Fig. 5D) relative to the vehicle group. In contrast, viability of B cells (Fig. 5E), activated by LPS + IL-4, was significantly reduced by FKGK18, paralleling decreases in both IgG and IgM production (Fig. 5F). However, FKGK18 (10⁻⁶ mol/L) caused an ∼20% greater reduction in antibody production compared with its effects on cell viability. Moreover, significant decreases in both were still evident when antibody production was normalized to cell viability (Fig. 5G). These findings suggest that sensitivity of different immune cells to iPLA₂β varies and that in addition to affecting immune cell function, iPLA₂β activation may influence B-cell survival/proliferation.

To verify that the in vivo effects of FKGK18 are due to iPLA₂β inhibition, the effects of FKGK18, S-BEL (specific for iPLA₂β), and R-BEL (specific for iPLA₂γ) were compared. Cell viability with vehicle and FKGK18 exposures was similar between 24 and 72 h, suggesting that proliferation of cells was not compromised by FKGK18 (Fig. 6A). CD4⁺ T cells were then exposed to S-BEL or R-BEL for only 30 min (4,5), unlike FKGK18, which was present for the entire 72 h. Similar to FKGK18, viability of CD4⁺ T cells (Fig. 6B) or IFN-γ (Fig. 6C) and IL-2 (Fig. 6D) production were not affected by R-BEL. However, unlike FKGK18, R-BEL did not inhibit TNF-α generation (Fig. 6E). In contrast, cell viability was significantly reduced by S-BEL (Fig. 6B), and cytokine generation was barely detectable (data not shown). These findings suggest that the in vivo effects of FKGK18 were due to inhibition of iPLA₂β and not iPLA₂γ.

Because CD4⁺ T cells and TNF-α are critical mediators of diabetes development and we found that they are affected by iPLA₂β activation, the effects of inhibiting eicosanoid generation on TNF-α production from CD4⁺ T cells was examined. The cells were exposed to inhibitors of COX (indomethacin), LOX (nordihydroguaiaretic acid [NDGA]), COX and LOX (5,8,11,14-eicosatetraynoic acid [ETYA]), or 12-LOX (cinnamyl-3,4-dihydroxy-α-cyanocinnamate [CDC]). In the absence of compromising cell viability (Fig. 7A), indomethacin, ETYA, and CDC mimicked FKGK18 inhibition of TNF-α production, but NDGA at concentrations <10 μmol/L, which inhibit 5-LOX but not 12-LOX, had no effect (Fig. 7B) relative to vehicle treatment. These findings provide evidence for participation of iPLA₂β-derived lipids generated by COX and 12-LOX in modulating immune cell function. This is supported by decreased PGE₂ generation with FKGK18, analogous to the COX inhibitors (Fig. 7C).

Given this evidence for iPLA₂β-derived lipid involvement in modulating CD4⁺ T-cell function, we examined the ability of FKGK18 to inhibit adoptive transfer of diabetes to immunodeficient NOD.scid mice by CD4⁺ T cells. Consistent with unchanged blood glucose levels (Fig. 8A), there was no evidence of diabetes development in untreated control and vehicle- or FKGK18 alone–treated mice (Fig. 8B). However, diabetes incidence was nearly 90% in mice following administration of CD4⁺ T cells (Fig. 8B). In contrast, the incidence was significantly reduced to ∼40% in mice that were pretreated for 1 week with FKGK18 before CD4⁺ T-cell transfer. In comparison, initiation of FKGK18 treatment at the time of transfer or administration of cells pretreated with FKGK18 for 3 days was without protective effect (Fig. 8C).

Type 1 diabetes is a consequence of autoimmune destruction of β-cells, and there is increasing evidence of iPLA₂β induction in diabetes and for its participation in β-cell apoptosis (3–6,13–15,32). We therefore considered the possibility that inhibition of iPLA₂β mitigates β-cell death and ameliorates diabetes development. Consistent with this, we found that iPLA₂β expression in islets from prediabetic female NOD mice is much higher than in diabetes-resistant mice. FK-based inhibitors with variable specificity for iPLA₂β have proven to be effective in countering central nervous system–related disorders (20–22). Our characterization of FKGK18 (19), which exhibits greater specificity for iPLA₂β over other PLA₂s (7,18,33), revealed FKGK18 to be more specific for iPLA₂β than iPLA₂γ and in contrast to BEL (9,16), to not inhibit other serine proteases.

The NOD mouse is a model of spontaneous type 1 diabetes development, and as expected (30), ∼80% of the female NOD mice developed diabetes by 25 weeks of age. In contrast, FKGK18 administration started at 10 days of age significantly reduced diabetes incidence. Concomitant observations included decreases in islet infiltration by immune cells, higher levels of circulating insulin, preservation of β-cell area, and improved glucose homeostasis. Together, these findings suggest that inhibiting iPLA₂β may be effective in ameliorating diabetes evolution.

Mass spectrometry protocols revealed that FKGK18 levels in the serum were maximum by 2–6 h but near minimal by 48 h, implying that the current regimen (three times per week) would not be expected to lead to significant accumulations of the inhibitor, therefore limiting its toxicity. Consistent with this, mice receiving FKGK18 exhibited similar body weights over 30 weeks, no discernible damage in nonislet tissues, prolonged euglycemia, and consequential survival. Taken together, these observations indicate that the FKGK18 regimen used does not promote systemic toxicity and is well tolerated.

The reversible nature of FKGK18 precludes measurement of iPLA₂β activity following tissue isolation, requiring an alternate measure of iPLA₂β inhibition in vivo. Both S-BEL (34,35) and FKGK18 (19) decrease PGE₂ release from pancreatic islets; however, PGE₂ is metabolized further in vivo and thus is an unreliable measure of PLA₂ activity. We therefore measured urinary PGEM levels and found them to be decreased by ∼25% for up to 42 h following FKGK18 administration but returned to control levels beyond 48 h, consistent with its bioavailability. Given its specificity for iPLA₂β over other PLA₂s (18,19), these observations are taken as a reflection of reversible inhibition of iPLA₂β by FKGK18. Because urine was collected from chronically treated mice, the continued generation of PGEMs suggests that activities of other PLA₂s are not compromised by FKGK18.

Remarkably, FKGK18 dramatically reduced insulitis. Infiltration of islets by immune cells is a precursor event that leads to β-cell destruction (30), and CD4⁺ T cells and B cells, which independently and in concert destroy β-cells (36), were reduced by FKGK18. These findings suggest that FKGK18 may impede immune cell migration into or proliferation within the islet. With evidence of iPLA₂β mRNA expression in immune cells, we considered the possibility that CD4⁺ T-cell and B-cell functionality also is affected by FKGK18. Because of their importance to the development of diabetes (37), we examined whether generation of proinflammatory cytokines by immune cells from NOD mice is altered in the presence of prototypical stimulants. Generation of IFN-γ and IL-2 by splenocytes with concanavalin A and of the chemoattractant TNF-α (38) by CD4⁺ T cells with αCD3/αCD28 was reduced by FKGK18. These differences are likely due to splenocytes being an admixture of leukocytes, with the bulk being B cells and only ∼10% CD4⁺ T cells. In contrast, macrophages exposed to LPS were unaffected by FKGK18, analogous to the inability of zymosan to alter arachidonate release from macrophages treated with iPLA₂β inhibitors (39) or from iPLA₂β-deficient mice (40). The lack of effects in macrophages under the conditions tested is further evidence for the absence of nonspecific effects of FKGK18. Consistent with this, R-BEL, which is specific for the membrane-associated iPLA₂γ (16), did not inhibit TNF-α production from CD4⁺ T cells.

Although their precise role has not yet been clearly defined, B cells are recognized to be critical to the development of type 1 diabetes (41,42), and antibody production by B cells was reduced by FKGK18. Of all the immune cells, only B-cell viability was significantly diminished by FKGK18, suggesting differential sensitivity of these cells to iPLA₂β inhibition. This in combination with decreased infiltrate CD4⁺ T cells, which stimulate B-cell proliferation, may account for the reductions in B cells in the islet infiltrate.

These observations suggest that modulation of immune cell function may be a mechanism by which iPLA₂β-derived lipid signals participate in autoimmune β-cell death. Recognizing the critical role of T cells and TNF-α in diabetes development, the effects of eicosanoid-generating pathway inhibitors on TNF-α production by CD4⁺ T cells were examined. Such analyses revealed that the inhibition of TNF-α production by FKGK18 was mimicked by inhibitors of COX (indomethacin), COX and LOX (ETYA), and 12-LOX (CDC) but unaffected by the LOX inhibitor NDGA (at a concentration range that inhibits 5-LOX but not 12-LOX). These findings suggest that lipid signals generated by COX and 12-LOX, subsequent to iPLA₂β-mediated hydrolysis of arachidonic acid, are likely candidate modulators of immune cell function. Consistent with this, FKGK18 inhibited PGE₂ generation by the CD4⁺ T cells analogous to the COX inhibitors. Contribution of COX and 12-LOX products to the autoimmune responses is supported by reports that PGE₂ release is significantly higher from NOD islets that are infiltrated versus noninfiltrated (43) and that deletion of 12/15-LOX protects against diabetes development (44). Further analyses using specific inhibitors of the eicosanoid-generating pathways are expected to more definitively identify the relevant eicosanoids.

The role for iPLA₂β-derived lipid signals in type 1 diabetes development is strengthened by the findings that adoptive transfer of diabetes to NOD.scid mice, which are immunodeficient and diabetes resistant, by CD4⁺ T cells is mitigated by pretreatment of the mice with FKGK18. However, administration of T cells and FKGK18 simultaneously was ineffective in mitigating diabetes development, suggesting that in vivo, iPLA₂β activation permits immune cell responses and that to counter this effect, iPLA₂β inhibition must occur before islet onslaught by immune cells. Furthermore, no protection from diabetes development could be demonstrated with transfer of FKGK18-pretreated T cells, and this most likely reflects the reversible nature of FKGK18 and its limited bioavailability beyond 3 days.

Although we found that inhibiting iPLA₂β may be beneficial in mitigating type 1 diabetes evolution, its participation in multiple biological processes (9) presents the potential for deleterious consequences with global iPLA₂β inhibition. It has been proposed that iPLA₂β plays a housekeeping role to generate substrates for reincorporation into phospholipids. However, several lines of investigation have suggested that modulation of iPLA₂β does not affect membrane phospholipids in islets or other tissues (45,46); thus, FKGK18 is not expected to alter membrane lipid composition. We and others have described a signal transduction role for iPLA₂β, including its amplification of glucose-stimulated insulin secretion (35,47). That it is counterintuitive to inhibit iPLA₂β to achieve euglycemia, however, is allayed by the observations that 1) iPLA₂β is increased, not decreased, in diabetes-prone mice; 2) fasting blood glucose levels in FKGK18-treated mice are not elevated relative to vehicle-treated NOD mice; 3) circulating levels of insulin are higher in the FKGK18 group; and 4) glucose tolerance is improved in the FKGK18 group in contrast to abnormal glucose tolerance in iPLA₂β-deficient mice (47).

In summary, chronic inhibition of iPLA₂β with FKGK18 preserves β-cell mass and reduces autoimmune type 1 diabetes incidence (Fig. 8D). With evidence for iPLA₂β expression in NOD β-cells and immune cells and for mitigation of immune cell function with iPLA₂β inhibition, we suggest that lipids derived from β-cell or immune cell iPLA₂β activation play critical signaling roles to promote recruitment, infiltration, or function of immune cells and that inhibition of iPLA₂β may be beneficial in ameliorating diabetes.

Acknowledgments. The authors thank the University of Alabama at Birmingham (UAB) Comparative Pathology Laboratory and Trenton R. Schoeb (professor of genetics, UAB) for interpretation of histopathologic changes in nonislet tissues; Scott M. Tanner (postdoctoral fellow, Department of Surgery, UAB), Joseph G. Daft (PhD candidate, Department of Pathology, UAB), and Lindsey E. Padgett (PhD candidate, Department of Microbiology, UAB) for immune cell experimental protocols and advice; Lionel C. Clement (assistant professor of nephrology, UAB) and Sumant S. Chugh (professor of nephrology, UAB) for assistance with the urine collection protocol; and Edward A. Dennis (professor of chemistry, biochemistry, and pharmacology, University of California, San Diego) for constructive comments and advice.

Funding. This work was made possible by funds provided by the National Institutes of Healthhttp://dx.doi.org/10.13039/100000002 (R01-DK-69455), American Diabetes Associationhttp://dx.doi.org/10.13039/100000041 (1-09-RA-147), and the Iacocca Family Foundation and the support of the UAB Comprehensive Diabetes Center Beta Cell Biology Core and the Diabetes Research Center.

Duality of Interest. No conflicts of interest relevant to the article were reported.

Author Contributions. R.N.B. contributed to the design, experimentation, data analyses, and manuscript preparation and editing. Y.G. contributed to the design and experimentation. V.M. and G.K. contributed to the synthesis and characterization of FKGK18, data analyses, and manuscript preparation and editing. M.G.K. contributed to the high-resolution mass spectrometry protocol development and analyses for FKGK18 and mass spectrometry spectra. T.A. contributed to the characterization of FKGK18 and manuscript editing. X.L. contributed to the experimentation, data analyses, and manuscript editing. H.M.T. and S.R. contributed to the design, data analyses, and manuscript preparation and editing. S.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 73rd Scientific Sessions of the American Diabetes Association, Chicago, IL, 21–25 June 2013; the 74th Scientific Sessions of the American Diabetes Association, San Francisco, CA, 13–17 June 2014; the 5th International Conference on Phospholipase A2 Mediated Signaling in Translational Medicine, New Orleans, LA, 20–21 May 2013; the 13th International Conference on Bioactive Lipids in Cancer, Inflammation and Related Diseases, San Juan, Puerto Rico, 3–6 November 2013; and the 48th Annual Southeastern Regional Lipid Conference, Cashiers, NC, 13–15 November 2013. In all these cases, only a very general description of the data were included in abstracts of <250 words.