AbstractExtracellular microRNAs (miRNAs) are a new class of biomarkers for cellular phenotypes and disease, and are bioactive signals within intercellular communication networks. Previously, we reported that miRNAs are secreted from macrophage to high-density lipoproteins (HDL) and delivered to recipient cells to regulate gene expression. Despite the potential importance of HDL-miRNAs, regulation of HDL-miRNA export from cells has not been fully studied. Here, we report that pancreatic islets and beta cells abundantly export miR-375-3p to HDL and this process is inhibited by cellular mechanisms that promote insulin secretion. Small RNA sequencing and PCR approaches were used to quantify beta cell miRNA export to HDL. Strikingly, high glucose conditions were found to inhibit HDL-miR-375-3p export, which was dependent on extracellular calcium. Likewise, stimulation of cAMP was found to repress HDL-miR-375-3p export. Furthermore, we found that beta cell ATP-sensitive potassium channel (KATP) channels are required for HDL-miRNA export as chemical inhibition (tolbutamide) and global genetic knockout (Abcc8鈭?鈭?/sup>) approaches inhibited HDL-miR-375-3p export. This process is not likely associated with cholesterol flux, as gain-of-function and loss-of-function studies for cholesterol transporters failed to alter HDL-miR-375-3p export. In conclusion, results support that pancreatic beta cells export miR-375-3p to HDL and this process is inversely regulated to insulin secretion. IntroductionIslets of Langerhans in the pancreas control systemic energy homeostasis primarily through two cell types: insulin-producing beta cells and glucagon-producing alpha cells. In response to high glucose, beta cells secrete insulin in a process known as glucose-stimulated insulin secretion (GSIS); however, chronic exposure to supraphysiological concentrations of glucose, e.g. conditions of systemic insulin resistance, can result in damage to the beta cell, and the development of Type 2 Diabetes (T2D)1,2. HDL have many beneficial properties in various biological processes that underlie pancreatic beta cell integrity and function, including enhancing GSIS3,4. HDL鈥檚 regulation of beta cell integrity and insulin secretion have been reported to be both dependent and independent of cholesterol transporters - ATP binding cassette transporter A1 (ABCA1) and scavenger receptor BI (SR-BI)4,5. HDL also have a wide-variety of alternative functions in many different cell-types conferred in part through the transport of non-cholesterol cargo, e.g. miRNAs6,7,8, suggesting there may be a relationship between insulin secretion and HDL-miRNAs.miRNAs are small non-coding RNAs (18鈥?2 nts in length) that post-transcriptionally regulate gene expression and are key factors in many (patho)physiologies9. In pancreatic beta cells, miRNAs have emerged as critical regulators of insulin secretion and cellular proliferation10,11. miR-375-3p is highly expressed in pancreatic islets and beta cells12,13, and is a key regulator of beta cell proliferation and function. Extracellular miRNAs are also detected in plasma, where they are protected from RNase digestion by protein and/or lipid complexes8,14,15. Recently, miR-375-3p was also found to be released from beta cells in exosomes16, suggesting that miR-375-3p may have both cellular and extracellular roles. Nevertheless, extracellular miR-375-3p is not limited to just exosomes, as we have previously reported that HDL also transport miR-375-3p8. Furthermore, we have previously demonstrated that HDL can transfer miRNAs, including miR-375-3p, to recipient cells, e.g. hepatocytes8. HDL-transferred miRNAs are also functional in recipient cells, as we have previously demonstrated that HDL-miRNAs delivered to coronary artery endothelial cells regulate key inflammatory genes, thus conferring, in part, HDL鈥檚 anti-inflammatory properties6. Collectively, our previous studies support that HDL transports miRNAs in circulation, that HDL-miRNA signatures are altered in metabolic disease, and that HDL-miRNAs serve as cell-to-cell signals within potential intercellular communication networks. Despite these findings, very little is understood about how cells export miRNAs to HDL.Here, we demonstrate that pancreatic islets and beta cells export miRNAs, specifically miR-375-3p, to HDL and report key insights into processes that regulate export. Remarkably, contrary to exosomes in which miR-375-3p export was increased under cellular conditions that promote with insulin secretion16, we found that beta cell miR-375-3p export to HDL is suppressed by these same mechanisms that promote insulin secretion. Multiple conditions that stimulate insulin secretion from pancreatic beta cells - high glucose, tolbutamide, and persistent membrane depolarization - inhibited miRNA export to HDL. Moreover, we found that high glucose inhibition of HDL-miR-375-3p export required extracellular calcium, which is also critical for insulin secretion, thus further establishing the negative relationship between insulin secretion and HDL-miR-375 export. Nonetheless, despite the importance of cholesterol flux for insulin secretion from beta cells, miRNA export to HDL was found to be independent of cholesterol transporters. Collectively, these findings establish that the pancreatic beta cells contribute to extracellular miRNAs on circulating HDL and that beta cell miRNA export is inversely regulated by the cellular processes that promote insulin secretion.ResultsPancreatic beta cells export miRNAs to HDLHigh-throughput sRNA-seq was used to quantify miRNAs on HDL from humans, and miR-375-3p was identified as a top-ranked HDL-miRNA (Table聽SI). To quantify the concentration of miR-375-3p on HDL, miR-375-3p levels were measured by quantitative PCR (qPCR) using 1鈥塵g of HDL from human and mouse HDL isolated by size-exclusion chromatography (SEC) or density-gradient ultracentrifugation (DGUC) followed by SEC. We detected 105鈥?06 copies/mg HDL (total protein) for miR-375-3p in both mouse and human samples (Fig.聽S1a). Furthermore, stoichiometric analysis of miR-375-3p levels on HDL and in plasma, suggests that HDL-miR-375-3p levels contribute to approximately 3.7鈥?.0% of the entire plasma pool of miR-375-3p (Fig.聽S1a). Moreover, we found that plasma and HDL miR-375-3p levels were strongly correlated. Based on miR-375鈥檚 link to beta cell biology and our previous study demonstrating that miR-375-3p was highly abundant on HDL, we hypothesized that pancreatic islets and beta cells contribute miR-375-3p to HDL. To test this hypothesis, we developed an HDL-miRNA export assay to measure miRNA efflux to HDL. Briefly, primary islets or beta cell lines were incubated with serum-free medium containing 1鈥塵g/mL native HDL (islet-nHDL or INS-1-nHDL) for 24鈥塰, or in cell free conditions (cf-nHDL). We then re-isolated nHDL from the culture media by SEC or apolipoprotein A-I immunoprecipitation (apoA-I-IP). To determine if primary human islets export miRNAs to HDL ex vivo, miRNA levels were quantified (by sRNA-seq) on cf-nHDL and on islet-nHDL, and differential expression analysis identified that 17 miRNAs were detected at levels 1.5 fold on HDL after incubation with islets compared to cf-nHDL (Fig.聽1b and Table聽SII). Strikingly, miR-375-3p was found to be the only miRNA that was detected with a large fold change and at high abundance. Real-time PCR was used to confirm that miR-375-3p levels were significantly increased (donor 1: 135-fold, p鈥?鈥?.0007; donor 2: 3.5-fold, p鈥?鈥?.008; donor 3: 609-fold, p鈥?鈥?.02) on nHDL after incubation with islets from three individual islet donors (Fig.聽1c). Other top exported miRNAs (Table聽SII), let-7d-5p, miR-126-5p, miR-183-5p, miR-223-3p were measured by real-time PCR, and only miR-183-5p was found to be exported to HDL, albeit at much lower levels than miR-375-3p (Fig.聽S1b). To assess the specificity of human islet miRNA export to HDL, miRNA profiles from sRNA sequencing of the human islets from donor 1 and their exported HDL-miRNAs were linked (Fig.聽1d, Tables聽SII and III). Most interestingly, we found that only 17 of the 1491 miRNAs detected in islets were exported to nHDL ( 1.5-fold islet-nHDL vs. cf-nHDL levels), suggesting that miRNA export to nHDL is likely to be specific, and not simply reflecting cellular miRNA concentrations (Fig.聽1d).Figure 1Pancreatic islets export miR-375-3p to HDL. (a) Quantification of miR-375-3p in plasma and HDL from WT mice. Correlation of plasma miR-375-3p levels and HDL-miR-375-3p levels. n鈥?鈥?; Linear regression analysis. (b) MA plot of miRNA changes between cf-nHDL and islet-nHDL. n鈥?鈥? human donor. (c) miR-375-3p levels on cf-nHDL and islet-nHDL from three islet preps from individual donors. Donor 1: n鈥?鈥?, donor 2: n鈥?鈥?, donor: 3, n鈥?鈥?; mean鈥壜扁€?5% CI; Two-tailed t-test. (d) Circos depicting top 100 islet miRNAs from 1 human donor (blue) and exported miRNAs from the same islets. sRNA-seq; dots represent RPM in islets and blue bars depict log2 HDL-miRNA fold change in islet-nHDL compared to cf-nHDL.Full size imageTo determine if pancreatic beta cells export miR-375-3p and other miRNAs to HDL, HDL-miRNA export assays were performed in rat INS-1 832/13 cells and sRNA-seq was used to quantify miRNAs on cf-nHDL and INS-1-nHDL. Strikingly, the levels of 61 miRNAs were increased 1.5-fold on INS-1-nHDL compared to cf-nHDL. (Fig.聽2a and Table聽SIV). Similar to our findings of islet exported miRNAs (Fig.聽1b), miR-375-3p export was robust and distinct from other miRNAs. Real-time PCR was used to confirm that INS-1 cells exported miR-375-3p to nHDL (81.48-fold, p鈥?鈥?.0007) (Fig.聽2b). Furthermore, miR-145-5p was used as a negative control, as we found that miR-145-5p was consistently not exported to INS-1-nHDL by real-time PCR (Fig.聽2b). We quantified additional exported miRNAs by qPCR, miR-16-5p, miR-107-3p, miR-30d-5p, miR-182-5p, miR-21-5p, miR-27b-3p, miR-25-3p, miR-132-3p, miR-22-3p (Fig.聽S2). Similar to our findings in islets, only one miRNA, miR-132-3p was found to be significantly exported (p鈥?鈥?.01) and at much lower levels than miR-375-3p (Fig.聽S2). Based on the striking observation that miR-375-3p is the most abundantly exported miRNA, we next investigated how its export is regulated.Figure 2INS-1 (832/13) beta cells export miR-375-3p to HDL. (a) MA plot of miRNA changes between cf-nHDL and INS-1-nHDL. n鈥?鈥?. (b) HDL miR-145-5p and miR-375-3p levels on cf-nHDL and INS-1-nHDL. n鈥?鈥?; mean鈥壜扁€?5% CI; two-tailed t-test. (c,d) INS-1 cellular levels of (c) mature miR-375-3p and (d) primary-miR-375. n鈥?鈥?; mean鈥壜扁€?5% CI; two-tailed t-test. (e) INS-1 miR-145-5p and miR-375-3p export to nHDL: cf-nHDL and INS-1-nHDL after 1, 4, 24, or 48鈥塰 after incubation with INS-1 cells. n鈥?鈥?; mean鈥壜扁€?5% CI; One-way ANOVA with Bonferroni post-test, alpha鈥?鈥?.05. (f) HDL miR-145-5p and miR-375-3p levels on cf-nHDL and INS-1-nHDL. n鈥?鈥?; mean鈥壜扁€?5% CI; two-tailed t-test. (g) Percentage of INS-1 cells in G1, S and G2 phase determined by flow cytometry quantification of propidium iodide stain. n鈥?鈥?; mean. (h) Export of miR-375-3p to nHDL after cell cycle synchronization and enrichment in G1 (0鈥塰), S (4鈥塰) and G2 (12鈥塰) phase. n鈥?鈥?; mean鈥壜扁€?5% CI; One-way ANOVA with Bonferroni post-test, alpha鈥?鈥?.05.Full size imagemiR-375-3p regulates beta cell development, and its levels have been found to be inversely associated with insulin secretion12,17,18. To determine if beta cell export of miR-375-3p to HDL reduces cellular miR-375-3p levels, mature miR-375-3p levels were quantified by real-time PCR in INS-1 cells treated with nHDL for 24鈥塰. Cellular levels of mature miR-375-3p were not altered by nHDL treatments (Fig.聽2c), suggesting that export of miR-375-3p to HDL is not likely a mechanism for the regulation or depletion of cellular miRNA levels. Nevertheless, miR-375 transcription was found to be inhibited by nHDL treatments, as the levels of the primary miR-375 transcript (pri-miR-375) were significantly decreased (p鈥?鈥?.029) in INS-1 cells treated with nHDL compared to untreated cells (Fig.聽2d). To determine the temporal dynamics of beta cell HDL-miR-375-3p export, INS-1-nHDL was purified at multiple time-points. We found that miR-375-3p levels steadily increased on INS-1-HDL over 24鈥塰 鈥?54-fold at 1鈥塰, 116.3-fold at 4鈥塰, 180.5-fold at 24鈥塰, and 4,874.2鈥揻old at 48鈥塰 compared to cf-nHDL levels - whereas miR-145-5p was not exported to HDL at any time-point (Fig.聽2e). Cellular miR-375-3p levels were unaffected at any time-point, and miR-145-5p levels were considerably decreased at 24鈥塰. (Fig.聽S3). To confirm that beta cell miR-375-3p export to HDL is not limited to rat INS-1 cells, HDL-miRNA export assays were also performed in mouse MIN6 cells, and we found that MIN6 cells also export miR-375-3p (46.19-fold), but not miR-145-5p, to nHDL (Fig.聽2f). These results strongly support that pancreatic islets and INS-1 cells robustly and selectively export miR-375-3p to HDL.Unlike native beta cells in the islets which are not proliferative, INS-1 and MIN6 cells are mitotically active19. We found that in unsynchronized INS-1 cells, approximately 77% of cells are in G1 phase, whereas approximately 6鈥?% of cells are in S and G2 phases (Figs聽2g and S4). We next tested whether INS-1 cells that are enriched in G1 (0鈥塰), S (4鈥塰) or G2 (12鈥塰) phases differentially export miR-375-3p to HDL. Of note, we did not observe any differences in cell cycle states when cells were treated with FBS, serum free, or serum free鈥?鈥塶HDL media for the last 4鈥塰 of the experiments (Figs聽2g and S4). Most importantly, we found that miR-375-3p levels on INS-1-nHDL are not affected by the cell cycle (Fig.聽2h), whereas cellular levels of miR-375-3p were found to be increased at 0鈥塰 (p鈥?lt;鈥?.0001) compared to asynchronous cells (Fig.聽S5). Pri-miR-375 levels were not altered across the time points (Fig.聽S5). Due to the observation that despite the differences in proliferation between islets and INS-1 cells, miR-375-3p is the main miRNA exported from both islets and INS-1 cells at high levels, we sought to characterize whether miR-375-3p export is regulated by insulin secretion.Insulin secretion suppresses beta cell miRNA export to HDLPancreatic beta cells maintain systemic energy homeostasis through glucose sensing and insulin secretion; therefore, we sought to determine if the cellular mechanisms that control insulin secretion (i.e. GSIS) also regulate miRNA export to HDL. To determine the impact of high glucose conditions on beta cell miR-375-3p export to HDL, export assays were completed in INS-1 cells with RPMI media containing normal (3鈥塵M) or high (11鈥塵M) glucose levels. In these assays, we failed to find a difference in HDL-miR-375-3p export between the two conditions (Fig.聽3a). However, RPMI medium is hypocalcemic (0.45鈥塵M Ca(NO3)2) and GSIS requires extracellular calcium for membrane depolarization. Therefore, we tested whether high glucose (11鈥塵M) media supplemented with 1.8鈥塵M CaCl2 alters beta cell miR-375-3p export to HDL. The addition of 1.8鈥塵M CaCl2 to the medium raises the extracellular levels to physiological levels, allowing for a rise in intracellular Ca2+ in response to high glucose stimulation. Remarkably, in the presence of extracellular Ca2+, high glucose conditions suppressed beta cell miR-375-3p export to nHDL (p鈥?鈥?.0242 between INS-1-nHDL 3鈥塵M glucose +1.8鈥塵M CaCl2 and 11鈥塵M glucose +1.8鈥塵M CaCl2) (Fig.聽3a). In response to high glucose, beta cells trigger a series of cellular processes that result in the secretion of insulin, including the closure of ATP-sensitive potassium channels (KATP), depolarization of the plasma membrane, and influx of extracellular calcium through voltage-gated calcium channels20. To determine whether the observed inhibition of HDL-miR-375-3p export by high glucose conditions is regulated by cellular mechanisms that regulate insulin secretion, the role of KATP channels in miRNA export to HDL was investigated. HDL-miRNA export assays were performed in INS-1 cells treated with a chemical inhibitor of the KATP channel (tolbutamide)21, which results in depolarization of the plasma membrane. Strikingly, tolbutamide treatment resulted in a significant decrease in beta cell miR-375-3p export to nHDL (p鈥?鈥?.0.004 INS-1-nHDL vehicle compared to INS-1-nHDL 25鈥壩糓 tolbutamide; p鈥?鈥?.008 INS-1-nHDL vehicle compared to INS-1-nHDL 100鈥壩糓 tolbutamide) (Fig.聽3b), without any changes in cellular miR-375-3p levels (Fig.聽S6). Conversely, diazoxide (Dia), a KATP channel activator, prevented the suppression of miR-375-3p export to HDL observed with Veh鈥?鈥塁aCl2 (p鈥?鈥?.7374 between cf-nHDL and INS-1-nHDL in Veh鈥?鈥塁aCl2; p鈥?鈥?.0145 between cf-nHDL and INS-1-nHDL in Dia鈥?鈥塁aCl2) (Fig.聽3c). Diazoxide also failed to alter cellular levels of mature miR-375-3p in INS-1 cells with or without CaCl2 (Fig.聽S6). In addition, we confirmed that both high glucose conditions and tolbutamide treatments stimulated insulin secretion from INS-1 cells, whereas diazoxide inhibited insulin secretion under high glucose conditions (Fig.聽S6).Figure 3Stimulation of insulin secretion blocks INS-1 export of miR-375-3p to HDL. (a) miR-375-3p levels on cf-nHDL and INS-1-nHDL after INS-1 cell treatment with 3鈥塵M or 11鈥塵M D-glucose in the presence or absence of 1.8鈥塵M CaCl2. n鈥?鈥?鈥?; mean鈥壜扁€?5% CI; One-way ANOVA with Bonferroni post-test, alpha鈥?鈥?.05. (b) HDL-miR-375-3p levels on cf-nHDL and INS-1-nHDL from INS-1 cells treated with tolbutamide for 2鈥塰. n鈥?鈥?鈥?; mean鈥壜扁€?5% CI; One-way ANOVA with Bonferroni post-test, alpha鈥?鈥?.05. (c) miR-375-3p levels on cf-nHDL and INS-1-nHDL after INS-1 cell treatment with vehicle or diazoxide (DIA) in the presence or absence of 1.8鈥塵M CaCl2. n鈥?鈥?; mean鈥壜扁€?5% CI; One-way ANOVA with Bonferroni post-test, alpha鈥?鈥?.05.(d) miR-375-3p levels on cf-nHDL and islet-nHDL from mouse WT (wildtype) or SUR1 KO (Abcc8鈭?鈭?/sup>) mice. n鈥?鈥?; mean鈥壜扁€?5% CI; One-way ANOVA with Bonferroni post-test, alpha鈥?鈥?.05. (e) HDL-miR-375-3p levels on cf-nHDL and INS-1-nHDL from INS-1 cells treated with exendin-4 (ex-4) or IBMX for 3鈥塰. n鈥?鈥?; mean鈥壜扁€?5% CI; One-way ANOVA with Bonferroni post-test, alpha鈥?鈥?.05. (f) Schematic depicting miRNA export to HDL in low glucose (left panel), and suppression of export in high glucose (right panel).Full size imageThe KATP channel is a hetero-octameric protein composed of four SUR1 and four Kir6.2 subunits, and deletion of either channel component results in incomplete trafficking to the plasma membrane and complete loss of channel activity22. To replicate our chemical inhibition findings using an orthogonal approach, we tested whether genetic deletion of a KATP channel subunit in mice (sulfonylurea receptor 1, SUR1 (Abcc8鈭?鈭?/sup>)) impacts pancreatic islet miR-375-3p export to nHDL. Primary islets were isolated from Abcc8鈭?鈭?/sup> mice (Fig.聽S7), and miR-375-3p export to HDL was quantified ex vivo, as described above. In agreement with our in vitro studies, loss of the KATP channel in SUR1 knockout islets, significantly reduced miR-375-3p export to nHDL ex vivo (p鈥?鈥?.0363 between WT and Abcc8鈭?鈭?/sup> islet-n-HDL) (Fig.聽3d). KATP鈥揹eficiency in mice (Abcc8鈭?鈭?/sup>) has been reported to cause calcium-associated changes in the beta cell transcriptome23; therefore, we confirmed that the observed loss of HDL-miR-375-3p export was not merely due to decreased miR-375-3p expression in islets isolated from Abcc8鈭?鈭?/sup> mice compared to WT mice. In fact, miR-375-3p levels were found to be increased, not decreased, in Abcc8鈭?鈭?/sup> islets (Fig.聽S7). These findings suggest that miR-375-3p export to nHDL is inverse to insulin secretion. To further test this association, the role of cAMP was investigated.As a second messenger molecule, cAMP enhances GSIS in beta cells through both protein kinase A dependent and independent mechanisms that regulate the exocytotic machinery and membrane depolarization24,25. A rise in cellular cAMP levels can be triggered by increased intracellular Ca2+ concentrations or through hormone signaling, e.g. glucagon-like-peptide 1 (GLP-1). cAMP also alters gene expression and recently, cAMP has been shown to repress transcription of pri-miR-37526. Therefore, we investigated whether miR-375-3p export to nHDL is regulated by cAMP. Two different compounds were used to stimulate cAMP in INS-1 cells, exendin-4 (ex-4), a GLP1R agonist, and IBMX, a phosphodiesterases inhibitor. As a positive control, the expression of c-fos was quantified, as this gene was previously reported to be increased with ex-4 and IBMX treatments26,27. IBMX treatments, but not ex-4, increased Fos mRNA levels in INS-1 cells (p鈥?lt;鈥?.0001), perhaps due to low level of expression of GLP1R in INS-1 cells28 (Fig.聽S8). We found that pri-miR-375, but not mature miR-375-3p levels were down-regulated in INS-1 cells treated with ex-4 or IBMX in serum-free media鈥?鈥塶HDL (Fig.聽S8). Most interestingly, IBMX, but not ex-4, was found to repress miR-375-3p export to nHDL (p鈥?鈥?.0098) (Fig.聽3e). These results further support a model in which stimulation of GSIS from beta cells, either through glucose, membrane depolarization, or cAMP, inhibit miR-375-3p export to nHDL. Furthermore, these results established an inverse link between beta cell miRNA export to HDL and insulin secretion (Fig.聽3f).Beta cell HDL-miRNA export is independent of cholesterol fluxPreviously, studies have demonstrated that HDL enhances beta cell insulin secretion which requires cholesterol transporters4. Based on these findings, we sought to examine the roles of HDL鈥檚 primary receptor, scavenger receptor BI (SR-BI), and key cholesterol transporters, ATP-binding cassette transporter A1 (ABCA1) and ATPB-binding cassette transporter G1 (ABCG1), in regulating beta cell miRNA export to nHDL. SR-BI is a bidirectional transporter of cholesterol and lipids, and mediates HDL-induced cell signaling29,30. We have previously demonstrated that HDL-miRNA delivery to recipient hepatocytes was dependent upon SR-BI8. SR-BI is also expressed in pancreatic beta cells and could, therefore, directly transport miRNAs to nHDL or indirectly facilitate HDL-induced cell signaling promoting miRNA export. To determine if SR-BI-deficiency in mouse islets aids in trafficking miR-375-3p to nHDL, pancreatic islets were collected from Scarb1鈭?鈭?/sup> (SR-BI KO) and c57Bl/6 wild-type mice (WT), and incubated with nHDL ex vivo (Fig.聽S9). Surprisingly, islets from both SR-BI KO and WT mice were found to export miR-375-3p to nHDL and we found no difference between islet genotype (p鈥?鈥?.6876 between WT and Scarb1鈭?鈭?/sup> islet-nHDL) (Fig.聽4a). Moreover, we tested whether beta cell SR-BI regulates HDL-miR-375-3p export in vitro; siRNAs were used to knockdown SR-BI expression in INS-1 cells, which was confirmed at the mRNA and protein levels by real-time PCR and western blotting, respectively (Fig.聽4b,c). Similar to our findings in primary islets ex vivo, SR-BI was not required for miR-375-3p export to nHDL from INS-1 cells (p鈥?鈥?.134 between mock and Scarb1 siRNA INS-1-nHDL) (Fig.聽4d).Figure 4Beta cell miR-375-3p export to HDL does not require cholesterol transporters. (a) miR-375-3p levels on cf-nHDL and islet-nHDL from mouse WT (wildtype) or SR-BI KO (Scarb1鈭?鈭?/sup>) mice. n鈥?鈥?; mean鈥壜扁€?5% CI; One-way ANOVA with Bonferroni post-test, alpha鈥?鈥?.05. (b) INS-1 cellular levels of Scarb1 mRNA and (c) SR-BI protein (western blotting) after transfection with mock or 50鈥塶M siRNA against Scarb1. Cropped images are of the same blot at different exposures. Full blots are available in Fig.聽S10. n鈥?鈥?2; mean鈥壜扁€?5% CI; One-way ANOVA with Bonferroni post-test, alpha鈥?鈥?.05. (d) miR-375-3p levels on cf-nHDL and INS-1-nHDL after knockdown of Scarb1; cells were transfected with mock or 50鈥塶M Scarb1 siRNA. n鈥?鈥?; mean鈥壜扁€?5% CI; One-way ANOVA with Bonferroni post-test, alpha鈥?鈥?.05. (e) ABCA1 and (f) ABCG1 protein (western blotting) after transfection with mock or 50鈥塶M siRNA against Abca1 and Abcg1, in the presence or absence of TO901317 (LXR agonist) and 9-cis-retinoic acid (RXR agonist). Representative of n鈥?鈥?. Cropped images are of the same blots at different exposures. Full blots are available in Fig.聽S10. (g) miR-145-5p and miR-375-3p levels on cf-nHDL and INS-1-nHDL after INS-1 cell transfection with mock or 50鈥塶M siRNAs against Abca1 and Abcg1 and/or LXR/RXR agonists. n鈥?鈥?; mean鈥壜扁€?5% CI; One-way ANOVA with Bonferroni post-test, alpha鈥?鈥?.05.Full size imageWe next sought to investigate the role of cholesterol transporters ABCA1 and ABCG1 in regulating miRNA export to HDL. ABCA1 and ABCG1 mediate cholesterol and lipid efflux to discoidal nascent HDL and spherical HDL particles, respectively31. ABCA1 is also a key mediator of HDL-induced anti-inflammatory cell signaling. We have previously reported that liver-X-receptor (LXR) activation, which increases ABCA1 and ABCG1 expression, failed to alter miR-223-3p export from macrophages to nHDL8. Nonetheless, ABCA1 and/or ABCG1 might regulate miR-375-3p export to nHDL in pancreatic beta cells; therefore, siRNAs were used to knockdown ABCA1 and ABCG1 expression in INS-1 cells, which was confirmed by loss of mRNA and protein levels (Figs聽4e,f and S9). Due to low basal levels of ABCG1 expression in beta cells, we also studied the effect of transporter over-expression using LXR/RXR agonists which promote the transcription of Abca1 and Abcg1 (TO901317, LXR agonist; 9-cis-retinoic acid, RXR agonist) (Figs聽4e,f and S9). HDL-miRNA export assays were performed in conditions of dual Abca1 and Abcg1 knockdown or over-expression; however, neither silencing, nor over-expression of these cholesterol transporters had any effect on beta cell HDL-miR-375-3p export (Fig.聽4g). Thus, SR-BI, ABCA1, and ABCG1 do not likely regulate HDL-miR-375-3p export from pancreatic beta cells. Combined, these results support a model in which beta cell miR-375-3p export and cholesterol efflux to HDL are mediated by distinct transporters and/or pathways.DiscussionHDL particles聽have many diverse beneficial properties, including anti-inflammatory and anti-oxidant capacities3. In the beta cell, HDL promotes cell integrity and proper function, e.g. protection against cellular toxicity and maintenance of GSIS3. Although most of the beneficial properties of HDL are attributed to cholesterol metabolism, HDL transports many types of non-cholesterol cargo, including proteins, vitamins, bioactive lipids, and non-coding sRNAs, which also likely contribute to HDL鈥檚 beneficial properties7. Previously, we reported that HDL transport miRNAs and deliver them to recipient cells where they regulate target gene expression6,8. Nevertheless, very little is known about the origin of HDL-miRNAs and the mechanisms of cellular miRNA export to HDL. In this study, we demonstrate that pancreatic islets and beta cells export selected miRNAs, e.g. miR-375-3p, to HDL. Using sRNA-seq, we profiled miRNA expression in pancreatic islets and on HDL. Comparisons of miRNAs detected in both HDL and islets showed considerable overlap. Based on this overlap, sRNA-seq was further used to identify the miRNAs that are exported from primary islets and beta cells to HDL ex vivo and in vitro, and we found that miR-375-3p is a highly and the most abundantly exported miRNA to HDL. Results suggest that beta cell miRNA export to HDL is inversely linked to insulin secretion. For example, beta cell miRNA export to HDL likely occurs during cellular conditions in which insulin secretion is low, as HDL-miR-375-3p export was inhibited under conditions that stimulate insulin secretion. Although we hypothesized that cholesterol transporters facilitate miRNA transport across the plasma membrane, we found that inhibition of SR-BI, ABCA1, and ABCG1 failed to alter beta cell miRNA export to HDL. Overall, these observations markedly advance our fundamental understanding of HDL-miRNA secretion.Many of the most abundant miRNAs in pancreatic islets and beta cells were found circulating on HDL, including miR-375-3p. Due to miR-375-3p鈥檚 critical importance to beta cell integrity and function, extracellular circulating miR-375-3p levels have been reported as a biomarker of beta cell function18,32. miR-375-3p has previously been detected in plasma, as well as exosomes and microvesicles, prompting renewed interest in the role of miR-375-3p as an intercellular signaling molecule and for its utility as a biomarker of disease (and pre-disease)16. In patients with T2D, no clear consensus exists as to the regulation of extracellular miR-375-3p levels, as different studies have reported an increase, decrease, or no change in extracellular miR-375-3p levels in T2D human subjects and rodent models of T2D18. Nevertheless, in both humans with Type 1 Diabetes (T1D) and murine models of T1D, miR-375-3p was consistently found to be upregulated in plasma samples33. For this reason, circulating (plasma/serum) miR-375-3p has been proposed as a biomarker of beta cell death32,34. Despite reports that beta cells may release miR-375-3p during cell death and serum-free conditions used in our studies may promote some level of cell death, we did not observe substantial cell death in our cell culture conditions. Moreover, in this study, we found that beta cell miRNA export is selective for miR-375-3p, and that many of the other highly-abundant miRNAs in the beta cells and islets are not exported to HDL, as determined by sRNA-seq.Due to the secretory phenotype of the endocrine beta cell, we sought to determine whether cellular mechanisms that control insulin secretion also regulate miRNA export to HDL. We found that while miR-375-3p was readily exported to HDL in low glucose conditions, export was inhibited at high glucose in the presence of physiologic calcium concentrations. In beta cells, under basal, low glucose conditions, the plasma membrane is hyperpolarized and insulin secretion is low. Increases in glucose concentrations above 5鈥塵M result in glucose metabolism and a rise in the ATP/ADP ratio. ATP closes the KATP channels resulting in the depolarization of the plasma membrane. This activates voltage-gated Ca2+ channels, and leads to increased intracellular [Ca2+] levels which result in insulin granule secretion. Many pharmacologic agents have been developed to modulate this process: sulfonylureas promote insulin secretion by binding to the SUR1 subunit and closing the KATP channel, whereas diazoxide opens the channel22,35. Here, chemical inhibition of the KATP channel, independent of high glucose, significantly inhibited miR-375-3p export to HDL. Conversely, diazoxide in high glucose prevent the suppresion of miR-375-3p export caused by CaCl2. It should also be noted that we were able to readily detect GSIS in INS-1 cells, and tolbutamide also increased insulin secretion in low glucose conditions, whereas diazoxide inhibited insulin secretion in high glucose. In high glucose, tolbutamide did not further stimulate GSIS beyond the effect of high glucose. We further tested the requirement of the KATP channel using islets isolated from SUR1 (Abcc8鈭?鈭?/sup>) knockout mice, and islets lacking the KATP channel had a decreased capacity to export miR-375-3p to HDL. We also found that IBMX, but not ex-4, suppressed miR-375-3p export to nHDL. Of note, IBMX resulted in a strong induction of c-fos, a cAMP responsive gene, whereas ex-4 did not, suggesting that a large increase in cAMP may be required in INS-1 cells to inhibit miR-375-3p export to HDL. Based on these observations, a few models can be proposed. Firstly, the KATP channel itself may be responsible for directly transporting miRNAs across the plasma membrane; however, the size of the channel is likely not large enough to facilitate miRNAs transport across the plasma membrane. Secondly, that KATP channel may indirectly aid the transport of miRNAs across the plasma membrane through interaction with a currently unknown membrane transporter, potentially through an allosteric regulatory mechanism. Thirdly, inhibition (closure) or absence of the KATP channel alters the cell polarization state and ion concentrations leading to insulin secretion, while indirectly affecting miRNA release or retention in the cell. Although our studies cannot distinguish between these three models, our findings suggest that ion dynamics likely are involved, as glucose-induced repression of HDL-miR-375-3p export also required extracellular calcium and occurred with elevated cAMP.Although this is the first study to report critical regulators of beta cell miRNA export to HDL, a recent study found that beta cell secretion of exosomes containing miR-375-3p was increased under conditions that promoted insulin secretion, i.e. high glucose, arginine and KCl16. This observed discrepancy鈥攄ecreased miRNA export to HDL compared to increased exosome miRNA release with GSIS鈥攊s in agreement with our previous study in which we reported that inhibition of the ceramide signaling pathway by GW4869 (attenuates exosome secretion) increased miRNA release to HDL8. Although further work is required to fully define the relationship between insulin secretion, miRNA export to HDL, and exosome secretion, results presented here and our proposed model support that beta cell miRNA export through exosomes and HDL occur through distinct routes under opposing cellular stimuli and regulation.Based on the results presented here and in previous studies from our laboratory8, beta cell-derived miRNAs are likely transported by HDL in circulation, and thus, may regulate target genes in cells that take up HDL-miRNAs. We have previously demonstrated that miR-223-3p, a miRNA detected on HDL at comparable levels to miR-375-3p, is biologically active on HDL and regulates inflammatory target genes in recipient endothelial cells6, thus supporting a potential role for beta cell-derived HDL-miR-375-3p in intercellular gene regulation. Nevertheless, the functional impact of HDL-miR-375-3p on cell-to-cell communication networks remains to be determined. In exosomes, miR-375-3p has been reported to be the 3rd most abundant miRNA in exosomes originating from pancreatic beta cells (following miR-709 and miR-1224)36,and miRNAs that are detected in exosomes at 8鈥?0 times lower concentration than miR-375-3p, e.g. miR-15a, have been shown to have biological function in miRNA-mediated communication originating from beta cells37. For example, an exogenous miRNA, cel-miR-238, when expressed in pancreatic beta cells has been shown to be released into exosomes to promote apoptosis in recipient cells36. Similarly, exosomes released from human islets and beta cells have also been reported to be taken up by recipient dendritic cells38. Furthermore, INS-1 beta cells were demonstrated to release exosomes containing miR-15a, which was increased in high glucose conditions37. These INS-1 cell exosomes were demonstrated to target Akt3 and promote apoptosis in recipient retinal glial cells (Muller cells), thus representing a potential underlying mechanism of diabetic complication resulting from beta cell-originating intercellular communication37. Although these miRNAs have been demonstrated to have physiological function in cell-to-cell communication pathways, they are found in exosomes at concentrations much lower than miR-375-3p. In addition, a recent study by Chevillet et al., found that the levels of miR-375-3p carried in exosomes represent a small percentage (2.7%) of total plasma miR-375-3 levels, thus leaving 97.3% to be associated with other carriers, e.g. lipoproteins and RNA-binding proteins39.Nevertheless, essential questions regarding the biological functions of HDL-bound miR-375-3p remain to be determined. HDL-miR-375-3p is abundantly detected in plasma, and likely regulates cell-to-cell communication in some capacity8. We hypothesize that HDL-miR-375 is transported from the beta cell to recipient cells where miR-375-3p regulates target gene expression; however, at this time the identity to the recipient cells/tissues are unknown. Furthermore, it is currently unknown whether HDL-miR-375-3p are delivered to distinct cells and tissues, and whether HDL-delivered miR-375-3p target different genes under distinct physiological conditions, i.e. fast vs. feeding states. miR-375-3p has been extensively studied in pancreatic beta cells, however, a few studies in other cell-types offer clues to the potential physiological effects of HDL-miR-375-3p. For example, miR-375-3p has been shown to negatively regulate osteogenesis, the process of bone formation, by targeting Wnt signaling proteins40. Wnt signaling is also involved in epithelial cell differentiation in the lung and rheumatoid arthritis (RA) pathogenesis41. miR-375-3p was also shown to target Frizzled 8, a component of canonical Wnt signaling, and inhibit epithelial cell differentiation in response to lung injury41. In fibroblasts, miR-375-3p suppression of Frizzled 8 inhibited disease pathogenesis in a rat model of RA42. In the brain, miR-375-3p is associated with neuroprotection in a rat model of cerebral ischemia/reperfusion43.In summary, this study provides regulatory insight into the cellular process that affect miR-375-3p export to HDL from human and mouse islets and pancreatic beta cells. Results suggest that miR-375-3p export to HDL is inhibited by glucose-stimulation (high glucose), but only in the presence of extracellular (1.8鈥塵M) Ca2+. Furthermore, through multiple different approaches, we demonstrate that miR-375-3p export to HDL occurs only under conditions of low insulin secretion, as demonstrated by modulating KATP channel activity and cAMP levels. Moreover, we found that miR-375-3p export to HDL was independent of cholesterol transporters鈥擜BCA1, ABCG1 and SR-BI. Together, these results support that pancreatic beta cell miR-375-3p export to HDL is inversely regulated by the cellular mechanisms that control insulin secretion.Experimental ProceduresHDL IsolationnHDL used for cell culture studies was isolated from healthy volunteers at Vanderbilt University Medical Center, purchased from Interstate Blood Bank, or obtained from the Vanderbilt blood bank. nHDL was isolated from human plasma by density-gradient ultra-centrifugation (DGUC, 1.061鈥?.21鈥塯/mL), 330,000鈥壝椻€?i>g followed by extensive dialysis at 4鈥壜癈 in 1X PBS. The study conformed to the guidelines set out in the Declaration of Helsinki and pertinent ethical regulations. The protocol was approved by the ethical committee of Vanderbilt University Medical Center (IRB# 151573) and all volunteers gave written informed consent for participation in the study. Pooled mouse plasma (50鈥塵l) from random fed mice was purchased from Equitech-Bio, Inc. HDL was isolated from mouse and human plasma by DGUC, as described above for nHDL. For human HDL-RNA analysis, blood was collected from healthy blood donors (n鈥?鈥?0) at Vanderbilt University Medical Center under active IRB protocol. Immuno-affinity purification of HDL was achieved as previously described8 using goat-anti-human apoA-I antibodies conjugated to Sepharose-4B beads (Academy Biomedical Company). RNA was isolated from 100 ug of HDL protein (Norgen Biotek).Animal ExperimentsAll animal experimentation was approved by and carried out in accordance to the Vanderbilt Institutional Animal Care and Use Committee. All transgenic lines were backcrossed to a C57BL/6 background for more than 10 generations. Female Abcc8鈭?鈭?/i> (Abcc8tm1.1Mgn) and wildtype C57B6J (WT) mice were used at 12 wks old23. Scarb1+/鈭?/sup> mice were obtained from The Jackson Laboratory and bred to generate Scarb1鈭?鈭?/sup> mice. Female Scarb1鈭?鈭?/sup> and WT mice were used at 8 wks old. For correlation of plasma and HDL miR-375, blood was collected from 5 12-week old male WT mice. HDL was isolated using size exclusion chromatography. HDL fractions were pooled and RNA was isolated from 1鈥塵g of HDL protein. RNA was isolated from 50ul of plasma.HDL-miRNA Export AssaysPrimary islets (human and rodent) or INS-1 cells were incubated with 1鈥塵g/mL nHDL (isolated by DGUC) in serum-free media for 1鈥?8鈥塰 and maintained at 37鈥壜癈 with 5% CO2 for 24鈥塰. Control nHDL was added to media and incubated in the absence of cells (cell free, cf) at 37鈥壜癈 with 5% CO2 for 24鈥塰. cf-nHDL and islet-nHDL or INS-1-nHDL were isolated from culture media by immunoprecipitation (IP) using goat-anti-human apoA-I antibodies conjugated to Sepharose-4B beads (Academy Biomedical Company). RNA was isolated from 100鈥壜礸 of HDL total protein (Norgen Biotek).Primary Islet CulturePrimary human islets from three individual human donors were obtained through the Integrated Islet Distribution Program. Primary mouse islets were isolated from the transgenic mice described above, as previously described44. Primary islets (50) were added to each well in RPMI supplemented with 5.6鈥塵M glucose, streptomycin (100鈥壩糶/ml), penicillin (100 U/ml), 10% LPDS (DGUC 1.21鈥塯/ml, bottom fraction), and 1鈥塵g/ml DGUC nHDL and maintained at 37鈥壜癈 with 5% CO2 for 24鈥塰. For a control group, nHDL was added to islet-free media and maintained at 37鈥壜癈 with 5% CO2 for 24鈥塰 in cell-free conditions (cf-nHDL). Cf-nHDL and islet-nHDL were isolated by IP against apoA-I, and total RNA was isolated from 100鈥壜礸 of nHDL. Total RNA was also isolated from mouse聽islets to confirm genotype.Cell CultureMIN6 cells were cultured in DMEM supplemented with streptomycin (100鈥壩糶/ml), penicillin (100 U/ml) and 10% FBS. INS-1 832/13 cells were cultured in RPMI supplemented with streptomycin (100鈥壩糶/ml), penicillin (100 U/ml), 10% FBS, and INS-1 supplement which consists of 10鈥塵M HEPES, 2 mM L-glutamine, 1鈥塵M sodium pyruvate, and 0.05鈥塵M 尾-mercaptoethanol. All cells were maintained at 37鈥壜癈 with 5% CO2. For miRNA export assay, INS1 and MIN6 cells were plated at 2鈥壝椻€?05 cells/mL for 24鈥塰 prior to HDL addition. Export experiments were performed at the following glucose concentrations: MIN6 cells 鈥?5鈥塵M D-glucose, INS1 cells 鈥?11鈥塵M glucose, except for experiment used for HDL-sequencing which was performed at 3鈥塵M glucose. DGUC nHDL was added to cells at 1鈥塵g/ml for 1鈥?4鈥塰. For tolbutamide studies, cells were washed with 1x HBSS for 2鈥塰 (0鈥塵M glucose). Media was then added containing 3鈥塵M glucose and either 25鈥壜礛, 150鈥壜礛 tolbutamide or ethanol vehicle for 1鈥塰. RPMI media was not supplemented with CaCl2 for these experiments. 1鈥塵g/mL nHDL was added for an additional 2鈥塰. For low and high glucose studies, cells were wash with 1x HBSS, and then media containing 3鈥塵M or 11鈥塵M glucose supplemented with 1.8鈥塵M CaCl2鈥?/鈭掆€?鈥塵g/ml nHDL was added for 24鈥塰. For cAMP studies, INS-1 cells were pretreated with vehicle, 100鈥壜礛 IBMX, or 100鈥塶M exendin-4 for 1鈥塰 in complete INS-1 media. Cells were then switched to serum free media鈥?/鈭掆€塶HDL supplemented with 100鈥壜礛 IBMX, or 100鈥塶M exendin-4 for an additional 2鈥塰. nHDL was added to the media from each of the conditions, as described above, and maintained at 37鈥壜癈 with 5% CO2 for 24鈥塰 (cf-nHDL). For diazoxide studies: Media was added containing 11鈥塵M glucose and either DMSO (vehicle) or 200鈥壩糓 diazoxide for 1鈥塰鈥?/鈭掆€?.8 Mm CaCl2. 1鈥塵g/mL nHDL was added for an additional 2鈥塰. Cf-nHDL and INS-1-nHDL were isolated by IP for apoA-I or SEC, and total RNA was isolated from 100鈥?00鈥壜礸 of nHDL.Cell Cycle AssaysSynchronization was performed as previously described19. Briefly, cells were plated in INS-1 RMPI media containing 10% FBS. Asynchronous cells were plated at the same time, but remained in INS-1 media for the remainder of the study. After 24鈥塰, synchronized cells were switched to INS-1 media鈥?鈥?.1% FBS for 56鈥塰, at which point cells were treated with 2鈥壜礸/ml aphidicolin (Sigma-Aldrich) for an additional 12鈥塰. Following aphidicolin treatment, media was changed to INS-1 media with 10% FBS and cells were collected at 0鈥塰, 4鈥塰, and 12鈥塰 for flow cytometry analysis of cell cycle phases and RNA content. To quantify miRNA export, media was changed to serum free media or serum free +1鈥塵g/mL nHDL 4鈥塰 prior to the final time-point for each condition. For analysis of cell cycle phases by flow cytometry, cells were trypsinized and counted, and 5鈥壝椻€?05 cells were washed with 1X PBS and fixed in 80% methanol (Sigma-Aldrich) at 鈭?0鈥壜癈 overnight. Methanol was removed by centrifugation at 2000鈥壝椻€?i>g for 10鈥塵in and cells were washed in 1X PBS twice. Cells were stained with 50鈥壩糶/ml propidium iodide (Invitrogen) in the presence of 50鈥壩糶/ml RNAse A (ThermoFisher) on ice for 2鈥塰. DNA content was analyzed using the 3-laser BD LSRII at the Vanderbilt Flow Cytometry Shared Resource Core.Transfection StudiesCells were plated at 2鈥壝椻€?05 cells/mL for 24鈥塰 prior to transfection (48鈥塰) with DharmaFECT 4 (Dharmacon). Transient transfections (50鈥塶M) were conducted with siRNA against Scarb1 (ON-TARGETplus SMARTpool L-098018-02), against Abca1 (ON-TARGETplus SMARTpool L-098018-02), against Abcg1 (ON-TARGETplus SMARTpool L-098018-02).Insulin Secretion AssaysINS-1 cells were plated onto 24-well plates and were grown to confluency prior to assay. The standard tissue culture medium was switched to medium containing 5鈥塵M glucose for 18鈥塰. Insulin secretion was performed in HEPES balanced salt solution (HBSS) (114鈥塵M NaCl, 4.7鈥塵M KCl, 1.2鈥塵M KH2PO4, 1.16鈥塵M MgSO4, 20鈥塵M HEPES, 2.5鈥塵M CaCl2, 25.5鈥塵M NaHCO3 and 0.2% bovine serum albumin, pH 7.2). Cells were washed with HBSS containing 3鈥塵M glucose, followed by a 2鈥塰 incubation in the same buffer. Insulin secretion was then measured in static incubations of HBSS containing 3鈥塵M or 15鈥塵M glucose containing 200鈥壩糓 tolbutamide or diazoxide for 2鈥塰. Insulin levels were measured using the porcine insulin radioimmunoassay (RIA) (Millipore, PI-12K). Human insulin was used for standard curves.Transcriptomics and qPCRHDL-RNA was isolated from equivalent amounts of HDL total protein and then quantified by RT-PCR or sRNA sequencing. Total RNA was isolated using either Total RNA Purification Kit (Norgen Biotek) or miRNeasy Mini Kit (Qiagen). HDL-RNA was isolated from equal HDL-protein amounts for each experiment. Cell and tissue RNA was diluted to equal concentrations. miRNA reverse transcription was performed using TaqMan MicroRNA Reverse Transcription Kit and predesigned TaqMan assays for each miRNAs Mix (Applied Biosystems). mRNA cDNA was obtained using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). qPCR was performed for 40 cycles using Taqman Universal PCR and predesigned Taqman assays for each miRNA or mRNA (Applied Biosystems). Ct values were normalized to a housekeeping gene, U6 for miRNAs and Ppia for mRNAs. Absolute quantification of miR-375 was performed by comparing Cts for samples of unknown miRNA concentration to a standard curve created with known concentration of miR-375 oligonucleotides, ranging from 3 fM鈥?00 pM (Integrated DNA Technologies). Real-time PCR was performed on the QuantStudio 6 Real-Time PCR System, as per according to manufacturer鈥檚 instructions (Life Technologies).Western BlottingWhole cell lysates were isolated using 150鈥塵M NaCl, 1% NP-40, 0.1% SDS, 100鈥塵M Tris-HCl, pH 7.4, supplemented with protease inhibitors (Roche) and cleared by centrifugation at 4鈥壜癈 for 10鈥塵in at 10,000鈥壝椻€塯. Proteins lysates were separated by 4鈥?2% SDS PAGE and then transferred to nitrocellulose membranes. The membranes were probed with monoclonal rabbit anti-SR-BI (Abcam, ab180383, 1:1,000), mouse monoclonal anti-ABCA1 (Abcam, ab18180, 1:1,000), rabbit polyclonal anti-ABCG1 (Novus Biologics, NB400-132, Lot F3, 1:1,000 dilution), or monoclonal mouse anti-GAPDH (Sigma, G8795, 1:10,000) in TBS-Tween20 containing 5% non-fat dry milk. HRP-conjugated secondary antibodies we used: anti-mouse (Promega, W4028, Lot 0000214819, 1:15,000) and anti-rabbit (Promega, W4018, Lot 0000212738, 1:20,000). Immune complexes were detected with Western Lightning Plus-ECL (Perkin Elmer) or Amersham ECL Prime Western Blotting Detection Reagent (GE Healthcare Life Sciences) chemiluminescent substrate.Small RNA SequencingTotal RNA from human HDL (Table聽SI), human islets export (Fig.聽1b and Table聽SII), rat INS-1 cell export (Fig.聽2a and Table聽SIV) were prepared with TruSeq small RNA library kits (Illumina) and sequenced on the HiSeq2500 sequencer SE50 (Illumina). Total RNA from human islets (Table聽SIII) were prepared with TruSeq sRNA kits (Illumina) and sequenced on the NextSeq500 sequencer SE75 (Illumina). All kits were performed as per manufacturer鈥檚 instruction with added amplification cycles. Prior to sequencing samples were size-selected by Pippin-Prep (Sage Science) to collect cDNA 135鈥?00 nts in length. Libraries were cleaned and concentrated (DNA Clean and Concentrator 5 kit, Zymo), tested for quality (High-Sensitivity DNA chips, 2100 Bioanalyzer, Agilent), and quantified (High-Sensitivity DNA assays, Qubit, Life Technologies). Equal concentrations samples were pooled for multiplex sequencing and concentrated (DNA Clean and Concentrator 5 kit, Zymo).InformaticsmiRNA sequencing data were analyzed by an in-house sRNA-seq data analysis pipeline45. Briefly, Cutadapt was used to remove adapters and reads 16 nts in length were aligned to the rat (rno5) or human (Hg18) genome with 1 mismatch allowance by Bowtie1 (v1.1.2)46 and DEseq2 (v1.18.1)47 was used for differential expression between cf-nHDL and INS-1-nHDL or islet-nHDL. For human HDL and islets, reads were normalized and reported as reads per million total reads. For MAplots, log2 fold change and log2 normalized counts were calculated using DESEq2 v1.20.0. Circos plot was generated with R package Circlize v0.4.3. Briefly, log2 fold change of miRNAs exported to HDL were calculated with DESEq2 v1.20.0 and sorted in descending order. Human islet miRNA counts were normalized by total reads per million, averaged among technical replicates, and sorted in descending abundance. Top 100 miRNAs were plotted. Links represent miRNAs exported to HDL that were also present in human islets.StatisticsComparisons between two groups: Two-tailed t-tests were used. One-way ANOVA with Bonferonni post-correction alpha鈥?鈥?.05 was used to compare between multiple groups. A p鈥?lt;鈥?.05 was considered significant. Scatter plots indicate replicates with mean鈥壜扁€?5% CI. Sequencing datasets are available through the Gene Expression Omnibus (GEO): GSE124559, GSE125135, GSE125136, GSE125137. References1.Robertson, R. P., Harmon, J., Tran, P. O., Tanaka, Y. Takahashi, H. Glucose toxicity in beta-cells: type 2 diabetes, good radicals gone bad, and the glutathione connection. Diabetes 52, 581鈥?87 (2003).CAS聽 Article聽Google Scholar聽 2.LeRoith, D. Beta-cell dysfunction and insulin resistance in type 2 diabetes: role of metabolic and genetic abnormalities. Am J Med 113(Suppl 6A), 3S鈥?1S (2002).CAS聽 Article聽Google Scholar聽 3.von Eckardstein, A. Widmann, C. High-density lipoprotein, beta cells, and diabetes. Cardiovasc Res 103, 384鈥?94, https://doi.org/10.1093/cvr/cvu143 (2014).CAS聽 Article聽Google Scholar聽 4.Fryirs, M. A. et al. Effects of high-density lipoproteins on pancreatic beta-cell insulin secretion. Arteriosclerosis, thrombosis, and vascular biology 30, 1642鈥?648, https://doi.org/10.1161/ATVBAHA.110.207373 (2010).CAS聽 Article聽 PubMed聽Google Scholar聽 5.Kruit, J. K., Brunham, L. R., Verchere, C. B. Hayden, M. R. HDL and LDL cholesterol significantly influence beta-cell function in type 2 diabetes mellitus. Current opinion in lipidology 21, 178鈥?85, https://doi.org/10.1097/MOL.0b013e328339387b (2010).CAS聽 Article聽 PubMed聽Google Scholar聽 6.Tabet, F. et al. HDL-transferred microRNA-223 regulates ICAM-1 expression in endothelial cells. Nature communications 5, 3292, https://doi.org/10.1038/ncomms4292 (2014).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 7.Vickers, K. C. Remaley, A. T. HDL and cholesterol: life after the divorce? J Lipid Res 55, 4鈥?2, https://doi.org/10.1194/jlr.R035964 (2014).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 8.Vickers, K. C., Palmisano, B. T., Shoucri, B. M., Shamburek, R. D. Remaley, A. T. MicroRNAs are transported in plasma and delivered to recipient cells by high-density lipoproteins. Nature cell biology 13, 423鈥?33, https://doi.org/10.1038/ncb2210 (2011).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 9.Bartel, D. P. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116, 281鈥?97, doi:S0092867404000455 [pii] (2004).10.Plaisance, V., Waeber, G., Regazzi, R. Abderrahmani, A. Role of microRNAs in islet beta-cell compensation and failure during diabetes. Journal of diabetes research 2014, 618652, https://doi.org/10.1155/2014/618652 (2014).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 11.Dalgaard, L. T. Eliasson, L. An 鈥榓lpha-beta鈥?of pancreatic islet microribonucleotides. Int J Biochem Cell Biol 88, 208鈥?19, https://doi.org/10.1016/j.biocel.2017.01.009 (2017).CAS聽 Article聽 PubMed聽Google Scholar聽 12.van de Bunt, M. et al. The miRNA profile of human pancreatic islets and beta-cells and relationship to type 2 diabetes pathogenesis. Plos one 8, e55272, https://doi.org/10.1371/journal.pone.0055272 (2013).ADS聽 CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 13.Tattikota, S. G. et al. Argonaute2 mediates compensatory expansion of the pancreatic beta cell. Cell metabolism 19, 122鈥?34, https://doi.org/10.1016/j.cmet.2013.11.015 (2014).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 14.Valadi, H. et al. Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nature cell biology 9, 654鈥?59, https://doi.org/10.1038/ncb1596 (2007).CAS聽 Article聽 PubMed聽Google Scholar聽 15.Arroyo, J. D. et al. Argonaute2 complexes carry a population of circulating microRNAs independent of vesicles in human plasma. Proceedings of the National Academy of Sciences of the United States of America 108, 5003鈥?008, https://doi.org/10.1073/pnas.1019055108 (2011).ADS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 16.Zhang, A. et al. Islet beta cell: An endocrine cell secreting miRNAs. Biochem Biophys Res Commun 495, 1648鈥?654, https://doi.org/10.1016/j.bbrc.2017.12.028 (2018).CAS聽 Article聽 PubMed聽Google Scholar聽 17.Poy, M. N. et al. miR-375 maintains normal pancreatic alpha- and beta-cell mass. Proceedings of the National Academy of Sciences of the United States of America 106, 5813鈥?818, https://doi.org/10.1073/pnas.0810550106 (2009).ADS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 18.Eliasson, L. The small RNA miR-375- a pancreatic islet abundant miRNA with multiple roles in endocrine beta cell function. Molecular and cellular endocrinology, https://doi.org/10.1016/j.mce.2017.02.043 (2017).19.Montemurro, C. et al. Cell cycle-related metabolism and mitochondrial dynamics in a replication-competent pancreatic beta-cell line. Cell Cycle 16, 2086鈥?099, https://doi.org/10.1080/15384101.2017.1361069 (2017).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 20.Komatsu, M., Takei, M., Ishii, H. Sato, Y. Glucose-stimulated insulin secretion: A newer perspective. J Diabetes Investig 4, 511鈥?16, https://doi.org/10.1111/jdi.12094 (2013).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 21.Proks, P., Reimann, F., Green, N., Gribble, F. Ashcroft, F. Sulfonylurea stimulation of insulin secretion. Diabetes 51(Suppl 3), S368鈥?76 (2002).CAS聽 Article聽Google Scholar聽 22.Li, N. et al. Structure of a Pancreatic ATP-Sensitive Potassium Channel. Cell 168, 101鈥?10 e110, https://doi.org/10.1016/j.cell.2016.12.028 (2017).CAS聽 Article聽 PubMed聽Google Scholar聽 23.Stancill, J. S. et al. Chronic beta-Cell Depolarization Impairs beta-Cell Identity by Disrupting a Network of Ca2+ -Regulated Genes. Diabetes 66, 2175鈥?187, https://doi.org/10.2337/db16-1355 (2017).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 24.Seino, S. Shibasaki, T. PKA-dependent and PKA-independent pathways for cAMP-regulated exocytosis. Physiol Rev 85, 1303鈥?342, https://doi.org/10.1152/physrev.00001.2005 (2005).CAS聽 Article聽 PubMed聽Google Scholar聽 25.Seino, S., Takahashi, H., Fujimoto, W. Shibasaki, T. Roles of cAMP signalling in insulin granule exocytosis. Diabetes, obesity metabolism 11(Suppl 4), 180鈥?88, https://doi.org/10.1111/j.1463-1326.2009.01108.x (2009).CAS聽 Article聽Google Scholar聽 26.Keller, D. M., Clark, E. A. Goodman, R. H. Regulation of microRNA-375 by cAMP in pancreatic beta-cells. Molecular endocrinology 26, 989鈥?99, https://doi.org/10.1210/me.2011-1205 (2012).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 27.Susini, S., Van Haasteren, G., Li, S., Prentki, M. Schlegel, W. Essentiality of intron control in the induction of c-fos by glucose and glucoincretin peptides in INS-1 beta-cells. FASEB journal: official publication of the Federation of American Societies for Experimental Biology 14, 128鈥?36 (2000).CAS聽 Article聽Google Scholar聽 28.Jo, S. et al. miR-204 Controls Glucagon-Like Peptide 1 Receptor Expression and Agonist Function. Diabetes 67, 256鈥?64, https://doi.org/10.2337/db17-0506 (2018).CAS聽 Article聽 PubMed聽Google Scholar聽 29.Ji, Y. et al. Scavenger receptor BI promotes high density lipoprotein-mediated cellular cholesterol efflux. J Biol Chem 272, 20982鈥?0985 (1997).CAS聽 Article聽Google Scholar聽 30.Assanasen, C. et al. Cholesterol binding, efflux, and a PDZ-interacting domain of scavenger receptor-BI mediate HDL-initiated signaling. J Clin Invest 115, 969鈥?77, https://doi.org/10.1172/JCI23858 (2005).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 31.Oram, J. F. Vaughan, A. M. ATP-Binding cassette cholesterol transporters and cardiovascular disease. Circ Res 99, 1031鈥?043, https://doi.org/10.1161/01.RES.0000250171.54048.5c (2006).CAS聽 Article聽 PubMed聽Google Scholar聽 32.Song, I., Roels, S., Martens, G. A. Bouwens, L. Circulating microRNA-375 as biomarker of pancreatic beta cell death and protection of beta cell mass by cytoprotective compounds. Plos one 12, e0186480, https://doi.org/10.1371/journal.pone.0186480 (2017).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 33.Assmann, T. S., Recamonde-Mendoza, M., De Souza, B. M. Crispim, D. MicroRNA expression profiles and type 1 diabetes mellitus: systematic review and bioinformatic analysis. Endocr Connect 6, 773鈥?90, https://doi.org/10.1530/EC-17-0248 (2017).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 34.Latreille, M. et al. miR-375 gene dosage in pancreatic beta-cells: implications for regulation of beta-cell mass and biomarker development. Journal of molecular medicine 93, 1159鈥?169, https://doi.org/10.1007/s00109-015-1296-9 (2015).CAS聽 Article聽 PubMed聽Google Scholar聽 35.Ashcroft, F. M. Mechanisms of the glycaemic effects of sulfonylureas. Hormone and metabolic research鈥?鈥塇ormon- und Stoffwechselforschung鈥?鈥塇ormones et metabolisme 28, 456鈥?63, https://doi.org/10.1055/s-2007-979837 (1996).CAS聽 Article聽 PubMed聽Google Scholar聽 36.Guay, C., Menoud, V., Rome, S. Regazzi, R. Horizontal transfer of exosomal microRNAs transduce apoptotic signals between pancreatic beta-cells. Cell Commun Signal 13, 17, https://doi.org/10.1186/s12964-015-0097-7 (2015).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 37.Kamalden, T. A. et al. Exo-miR-15a Transfer from the Pancreas Augments Diabetic Complications by Inducing Oxidative Stress. Antioxidants redox signaling, https://doi.org/10.1089/ars.2016.6844 (2017).38.Cianciaruso, C. et al. Primary Human and Rat beta-Cells Release the Intracellular Autoantigens GAD65, IA-2, and Proinsulin in Exosomes Together With Cytokine-Induced Enhancers of Immunity. Diabetes 66, 460鈥?73, https://doi.org/10.2337/db16-0671 (2017).CAS聽 Article聽 PubMed聽Google Scholar聽 39.Chevillet, J. R. et al. Quantitative and stoichiometric analysis of the microRNA content of exosomes. Proceedings of the National Academy of Sciences of the United States of America 111, 14888鈥?4893, https://doi.org/10.1073/pnas.1408301111 (2014).ADS聽 CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 40.Sun, T. et al. miR-375-3p negatively regulates osteogenesis by targeting and decreasing the expression levels of LRP5 and beta-catenin. Plos one 12, e0171281, https://doi.org/10.1371/journal.pone.0171281 (2017).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 41.Wang, Y. et al. miR-375 regulates rat alveolar epithelial cell trans-differentiation by inhibiting Wnt/beta-catenin pathway. Nucleic acids research 41, 3833鈥?844, https://doi.org/10.1093/nar/gks1460 (2013).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 42.Miao, C. G. et al. miR-375 regulates the canonical Wnt pathway through FZD8 silencing in arthritis synovial fibroblasts. Immunol Lett 164, 1鈥?0, https://doi.org/10.1016/j.imlet.2015.01.003 (2015).CAS聽 Article聽 PubMed聽Google Scholar聽 43.Wang, Y. et al. Downregulated RASD1 and upregulated miR-375 are involved in protective effects of calycosin on cerebral ischemia/reperfusion rats. Journal of the neurological sciences 339, 144鈥?48, https://doi.org/10.1016/j.jns.2014.02.002 (2014).CAS聽 Article聽 PubMed聽Google Scholar聽 44.Brissova, M. et al. Intraislet endothelial cells contribute to revascularization of transplanted pancreatic islets. Diabetes 53, 1318鈥?325 (2004).CAS聽 Article聽Google Scholar聽 45.Allen, R. M. et al. Bioinformatic analysis of endogenous and exogenous small RNAs on lipoproteins. bioRxiv (2018).46.Langmead, B., Trapnell, C., Pop, M. Salzberg, S. L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol 10, R25, https://doi.org/10.1186/gb-2009-10-3-r25 (2009).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 47.Love, M. I., Huber, W. Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq. 2. Genome Biol 15, 550, https://doi.org/10.1186/s13059-014-0550-8 (2014).CAS聽 Article聽 PubMed聽 PubMed Central聽Google Scholar聽 Download referencesAcknowledgementsINS-1 832/13 cells were provided by Hans-Ewald Hohmeier, MD, PhD, Chris B. Newgard, PhD, and Duke University School of Medicine. MIN6 were a gift from Dr. Praveen Sethupathy. The authors would like to acknowledge Robert C. Taylor, Stuart R. Landstreet, Wanying Zhu and Evan Chaudhuri for the purification of HDL supported by Lipoprotein and HDL Function Core (P01HL116263). Alvin Powers, MD, Chunhua Dai, MD, Greg Poffenberger, David Jacobson, PhD, and Prasanna Dadi were instrumental in procuring primary human islets through the Integrated Islet Distribution Program, for which we are grateful. Drs Mark Magnuson, MD and Jennifer Stancill, PhD kindly provided the Abcc8鈭?鈭?/sup> mice. The Abcg1 antibody was provided by Dr. Elizabeth Tarling, PhD. The authors would like to thank Rachel Chandler for her illustrations in Figure 3f, and Dr. Ryan Allen and Carrie Wiese for helpful discussions of the manuscript. This work was supported by the American Heart Association 15PRE25090205 (L.R.S.), 14CSA20660001 (K.C.V.); and the National Institutes of Health T32HL007411 (L.R.S.), 5R25GM062459-10 (L.R.S.), P01HL116263 (M.F.L.), R01HL128996 (K.C.V.) DK20593 (K.C.V.), and K22HL113039 (K.C.V.). Mouse islet isolations were performed by the Islet Procurement Analysis Core which is supported by the Vanderbilt Diabetes Research and Training Center (P60 DK020593). Insulin measurements were performed by the Vanderbilt Hormone and Analytical Services Core which is supported by NIH grants: DK059637 (MMPC) and DK020593 (DRTC). Flow Cytometry experiments were performed in the VMC Flow Cytometry Shared Resource. The VMC Flow Cytometry Shared Resource is supported by the Vanderbilt Ingram Cancer Center (P30 CA68485) and the Vanderbilt Digestive Disease Research Center (DK058404).Author informationAffiliationsDepartment of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, TN, USALeslie R. Sedgeman聽 聽Kasey C. VickersKineMed, Inc., Emeryville, CA, USACarine Beysen聽 聽Scott TurnerDepartment of Medicine, Vanderbilt University Medical Center, Nashville, TN, USAMarisol A. Ramirez Solano,聽Danielle L. Michell,聽MacRae F. Linton聽 聽Kasey C. VickersDepartment of Biostatistics, Vanderbilt University Medical Center, Nashville, TN, USAQuanhu Sheng聽 聽Shilin ZhaoAuthorsLeslie R. SedgemanView author publicationsYou can also search for this author in PubMed聽Google ScholarCarine BeysenView author publicationsYou can also search for this author in PubMed聽Google ScholarMarisol A. Ramirez SolanoView author publicationsYou can also search for this author in PubMed聽Google ScholarDanielle L. MichellView author publicationsYou can also search for this author in PubMed聽Google ScholarQuanhu ShengView author publicationsYou can also search for this author in PubMed聽Google ScholarShilin ZhaoView author publicationsYou can also search for this author in PubMed聽Google ScholarScott TurnerView author publicationsYou can also search for this author in PubMed聽Google ScholarMacRae F. LintonView author publicationsYou can also search for this author in PubMed聽Google ScholarKasey C. VickersView author publicationsYou can also search for this author in PubMed聽Google ScholarContributionsL.R.S. and D.M. performed experiments. L.R.S., M.R.-S. and K.C.V. analyzed data. M.R.-S., Q.S. and S.Z. performed sequencing analysis. C.B. and S.T. provided patient samples. L.R.S., M.F.L. and K.C.V. acquired financial support for this project. L.R.S. and K.C.V. conceived the project and wrote the manuscript with contributions from all authors.Corresponding authorCorrespondence to Kasey C. Vickers.Ethics declarations Competing Interests The authors declare no competing interests. Additional informationPublisher鈥檚 note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary information Supplementary MaterialRights and permissions Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article鈥檚 Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article鈥檚 Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Reprints and PermissionsAbout this articleCite this articleSedgeman, L.R., Beysen, C., Ramirez Solano, M.A. et al. Beta cell secretion of miR-375 to HDL is inversely associated with insulin secretion. Sci Rep 9, 3803 (2019). https://doi.org/10.1038/s41598-019-40338-7Download citationReceived: 09 April 2018Accepted: 08 February 2019Published: 07 March 2019DOI: https://doi.org/10.1038/s41598-019-40338-7 CommentsBy submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate. Sign up for the Nature Briefing newsletter 鈥?what matters in science, free to your inbox daily.