Identification and Functional Implications of Sodium/Myo-Inositol Cotransporter 1 in Pancreatic β-Cells and Type 2 Diabetes

Inositol is a naturally occurring cyclitol found in nine known isoforms, of which myo-inositol (MI) accounts for the largest population of active stereoisomers (1). Inositol is an important constituent of living cells that is known to take part in various cellular processes, including the development of peripheral nerves (2) and osteogenesis (3). Abnormal inositol metabolism has been reported in both type 1 and type 2 diabetes, as reflected by elevated urinary MI and d-chiro-inositol excretion (4–6). Indeed, MI supplementation has been documented to prevent the onset of gestational diabetes (7–9). Such findings have implicated MI in the regulation of glucose homeostasis. In addition to its role as an organic osmolyte in cells (10), MI is also a molecular precursor for the phosphoinositide (PI) second messenger system and perhaps a signaling molecule in its own right (11).

Rat pancreatic islets conduct active transport of MI (12), and exogenous MI has been reported to be necessary for the maintenance of β-cell function in rat islets (13). That is, in the absence of MI, rat islets were unresponsive to glucose stimulation, whereas increases in exogenous MI augmented insulin release after a glucose challenge (13). Pancreatic β-cell function and survival have also been shown to involve PI and its downstream effectors; specifically, glucose-stimulated activation of phospholipase C/protein kinase C (PLC/PKC) signaling has been previously reported to mediate second-phase insulin secretion by mobilizing Ca²⁺ ions stored in the endoplasmic reticulum (14). Meanwhile, activation of phosphoinositide 3-kinase (PI3K)–Akt signaling has been reported to promote β-cell function and survival (15,16). All of these findings suggest that MI may play a role in the intracellular signaling events of pancreatic β-cells.

Extracellular MI is taken up by membrane-bound inositol transporters (17). Two types of sodium-dependent MI cotransporters have been identified in humans, namely sodium/myo-inositol cotransporter 1 and 2 (SMIT1 [SLC5A3] and SMIT2 [SLC5A11], respectively; the latter is also known as sodium–glucose cotransporter [SGLT] 6). Both transporters are members of the solute carrier 5 gene family (18,19), and both transport MI along the concentration gradient of Na⁺ in a stoichiometric ratio of 2 Na⁺ ions for each MI molecule transported (20). SMIT1 shows higher specificity toward MI than SMIT2, which also transports d-glucose and d-chiro-inositol (21). Tissue expression studies have revealed widespread expression of both transporters in many human tissues, with particularly high expression levels in the brain and kidney cells (19,22).

Given that pancreatic islets are known to transport MI extensively (12) and the aforementioned recent findings showing that MI metabolism has been implicated in diabetic pathophysiology, the primary aim of the current study was to investigate the expression and function of inositol transporters in pancreatic islets and β-cells. In addition, we examined the mechanism by which MI may regulate β-cell function.

Male Zucker diabetic fatty (ZDF, fa/fa) rats, their age-matched littermates (Zucker lean [lean, fa/⁺]), and Wistar rats (8–12 weeks old) were supplied by the Laboratory Animal Services Center of The Chinese University of Hong Kong. All rats were housed in a temperature- and humidity-controlled room under a 12-h light/dark cycle with ad libitum rodent chow (Teklad 7001, 4.4%; Harlan Teklad Global Diets) and water. All experimental procedures were approved by the Animal Experimentation Ethics Committee of The Chinese University of Hong Kong (Ref: # 14-025-MIS_5A).

ZDF rats (n = 5 per group) were assigned randomly to the treatment group (ZDF-MI), in which MI (Sigma-Aldrich, St. Louis, MO) supplementation was administered for 1 month in the drinking water (6 g · L⁻¹), or to the non–MI-treated control group (ZDF-Con). Water consumption was monitored daily to calculate mean intake of MI. A mean daily MI intake of 0.57 ± 0.04 mg/d · g (body wt) was recorded. Food and water intake and body weight were measured weekly.

A glucometer (Bayer Corporation, Robinson Township, PA) was used to measure blood glucose in blood samples drawn from the tail vein. Intraperitoneal glucose tolerance tests (IPGTTs) were administered wherein, after a 6-h fast, rats were challenged with glucose (2 g ⋅ kg⁻¹), and blood glucose was measured 0, 15, 30, 60, 90, and 120 min thereafter. For insulin tolerance tests (ITTs), after a 4-h fast, rats were challenged with insulin (1.25 units/kg body wt) by intraperitoneal administration, and blood glucose was measured 0, 15, 30, and 60 min thereafter. Area under the curve (AUC) values for blood glucose were then calculated.

Pancreata were embedded in O.C.T. compound (Sakura, Tokyo, Japan) and were frozen. Cryostat sections (6 µm thick) were cut, mounted, and blocked in 2% BSA (Sigma-Aldrich) with 0.1% Triton X-100 (Sigma-Aldrich) for 30 min at room temperature. For the α- and β-cell analyses, sections were probed with rabbit polyclonal anti-insulin antibody (1:250; Santa Cruz Biotechnology, Santa Cruz, CA) and mouse monoclonal anti-glucagon antibody (1:250; Abcam, Cambridge, MA) and incubated for 2 h at room temperature with Alexa Fluor 568 donkey anti-rabbit antibody (1:250; Life Technologies, Carlsbad, CA) and Alexa Fluor 488 goat anti-mouse antibody (1:250; Life Technologies). For the β-cell proliferation analyses, sections were probed at 4°C with guinea pig polyclonal anti-insulin antibody (1:250; Life Technologies) and rabbit polyclonal anti-Ki67 antibody (1:250; Abcam) and then incubated for 2 h at room temperature with Alexa Fluor 488 goat anti-guinea pig antibody (1:250) and Alexa Fluor 568 donkey anti-rabbit antibody (1:250). Immunolabeling was analyzed by assessing the ratios of areas occupied by the red and green fluorescent signals within each islet by using ImageJ software (National Institutes of Health, Bethesda, MD).

Intact islets were isolated by injecting collagenase P (Roche, Mannheim, Germany) intraductally into harvested pancreata as described previously (23). Isolated islets were cultured overnight before treatment with 5.6 or 28 mmol/L d-glucose (Sigma-Aldrich) in the presence of 0.5 mmol/L phlorizin (Sigma-Aldrich) and/or 500 µg/mL MI for the designated periods of time. For purification of pancreatic β-cells, isolated islets were dispersed, and β-cells were isolated by flow cytometry as previously described (24). β-Cell purity was further assessed by flow cytometry using intracellular staining with Alexa Fluor 647 Mouse Anti-Insulin (BD Biosciences, San Diego, CA) and PE Mouse Anti-Glucagon (BD Biosciences) as previously described (25).

The insulinoma cell line INS-1E cell, a gift from Dr. Pierre Maechler (26), was cultured in a humidified chamber with 5% CO₂ in RPMI 1640 medium (11.2 mmol/L glucose) supplemented with 10% FBS, 1 mmol/L sodium pyruvate, 50 μmol/L 2-mercaptoethanol, 10 mmol/L HEPES, 100 units/mL penicillin, and 100 μg/mL streptomycin (all from Invitrogen, Waltham, MA). Cultures were passaged once weekly by gentle trypsinization. INS-1E cells were treated with 11.2 mmol/L or 28 mmol/L glucose, 0.5 mmol/L phlorizin, and/or 500 µg/mL MI for the experimentally indicated periods of time.

SMIT1 expression was suppressed with small interfering RNAs (siRNAs) for rat SMIT1 (constructed by Life Technologies, Hong Kong). INS-1E cells were transfected with siRNA-SMIT1 oligos or siRNA negative control oligos with Lipofectamine RNAi Max transfection reagent (Invitrogen) for 48 h, according to the manufacturer’s instructions. The sequences of the knockdown and control oligos are listed in Supplementary Table 1. Knockdown efficiency was determined by measuring SMIT1 expression levels in real-time PCRs and immunoblotting (Supplementary Fig. 2).

Glucose-stimulated insulin secretion (GSIS) was performed, as described previously (23), to assess β-cell function. Treated cells or isolated islets were normalized for their basal insulin-secreting status by preincubation in Krebs-Ringer bicarbonate buffer (KRBB) that contains 1.6 mmol/L d-glucose for 1 h. After normalization, the cells and islets were incubated in KRBB with 1.6 mmol/L d-glucose for 1 h, and the buffer was collected. The cells and islets were incubated in fresh KRBB with 16.7 mmol/L d-glucose for 1 h, during which stimulated insulin secretion was measured. For the measurement of intracellular insulin content, treated cells or isolated islets incubated in HCl-ethanol overnight at −20°C were homogenized and further incubated overnight at −20°C. Homogenized samples were centrifuged, and the supernatant was neutralized with 1 mol/L Tris (pH 7.5). ELISAs with a specific rat insulin ELISA kit (Antibody and Immunoassay Services at The University of Hong Kong) were used to quantify the amount of insulin released or intracellular insulin content in the collected buffer samples.

INS-1E cells were incubated for 30 min in a cell incubator in KRBB with 16.7 mmol/L d-glucose and 5 μmol/L Fluo-8 AM (Abcam). After incubation, cells were rinsed twice with PBS and treated with 28 mmol/L d-glucose. Fluorescence signals were monitored continuously for 7 min after glucose administration at 490 and 525 nm excitation and emission wavelength, respectively. Results were expressed as the F-to-F₀ ratio, with F₀ being the basal fluorescence intensity in the unstimulated condition.

The intracellular MI level was quantified by high-performance liquid chromatography tandem mass spectrometry (HPLC/MS), as previously described (27). In brief, treated cells were extracted with 3 volumes of acetonitrile to precipitate proteins. The precipitated proteins were removed by centrifugation, and the samples were diluted in HPLC water before HPLC/MS analysis. Separation of samples was performed in HPLC, with the protocol of an isocratic gradient of 95% HPLC water to 5% acetonitrile. Quantification was done by analyzing the multiple reaction monitoring transitions of 178.8 → 86.4 from the coupled triple quadrupole tandem MS analysis, operating in negative-ion mode. Intracellular MI levels were normalized to total protein levels.

TRIzol reagent (Invitrogen) was used for total RNA extraction from pancreatic islets and cells, according to the manufacturer’s instructions. First-strand cDNA was reverse transcribed with a PrimeScript reverse transcriptase master mix kit (Takara Bio Inc., Shiga, Japan). Gene expression was analyzed by PCR analysis or real-time PCR quantification, wherein cDNA samples were combined with SYBR Green QPCR master mix (Applied Biosystems, Carlsbad, CA) and specific primers (Supplementary Table 1). Relative gene expression was analyzed by the comparative threshold cycle method (2^−ΔΔCT) and normalized to β-actin.

Islet and cell proteins were extracted with the CytoBuster Protein Extraction Reagent (Novagen, Darmstadt, Germany). Extracted proteins were fractionated by 10% SDS-PAGE and transferred to nitrocellulose membranes (Bio-Rad, Munich, Germany), which were blocked with 5% milk and then probed with anti-SMIT1 (Novus, St. Louis, MO), anti-AKT (Cell Signaling Technology, Beverly, MA), anti–phospho-AKT (Ser473; Cell Signaling Technology), anti-ERK1/2 (Cell Signaling Technology), anti–phospho-ERK1/2 (Thr202/Tyr204; Cell Signaling Technology), or anti–β-actin (Santa Cruz Biotechnology) antibodies. Horseradish peroxide–conjugated secondary antibodies were applied to the membranes after washing. Labeled protein bands were visualized on autoradiography films (Fujifilm, Tokyo, Japan) after application of enhanced chemiluminescence detection reagent (GE Healthcare, Piscataway, NJ). The protein bands were analyzed with ImageJ software. Protein bands were normalized to β-actin protein levels.

Lipids were extracted from treated cells as previously described (28). Briefly, cells were scraped with cold 0.5 mol/L trichloroacetic acid and centrifuged. The cellular pellets were washed with 5% trichloroacetic acid/1 mmol/L EDTA. Neutral lipids were removed from the cell pellets by addition of methanol:chloroform (2:1), and acidic lipids were extracted with methanol:chloroform:12 mol/L HCl (80:40:1). The lower organic phases were collected after phase separation with chloroform/0.1 mol/L HCl and then vacuum dried in a vacuum dryer for subsequent assays.

Intracellular phosphatidylinositol 4,5-bisphosphate (PIP₂) levels were quantified with a PIP₂ Mass ELISA Kit (Echelon Bioscience Inc., Salt Lake City, UT), according to the manufacturer’s instructions. Briefly, extracted lipids were dissolved in PBS with 0.25% protein stabilizer (kit supplied). The resuspension was probed with a PIP₂ detector protein (kit supplied) for 1 h and then incubated in a 96-well PIP₂-coated microplate. A peroxidase-linked secondary detection reagent was applied to detect plate-bound PIP₂-detector protein, generating colorimetric signals measured at 450 nm. PIP₂ levels were expressed as pmol/10,000 cells or pmol/number of islets, according to calculations based on the standard curve generated. Intracellular inositol triphosphate (IP₃) levels analyses were performed with a Rat IP₃ ELISA Kit (CUSABIO Biotech, Wuhan, China), according to the manufacturer’s instructions.

Group data are reported as means ± SEMs. Differences between groups were detected by Student unpaired two-tailed t tests or one-way ANOVAs, followed by Tukey post hoc tests, wherein P < 0.05 was considered statistically significant.

Expression of both SMIT1 and SMIT2 was detected in freshly isolated rat pancreatic islets and INS-1E cells by RT-PCR assays (Fig. 1A and B). Quantitative PCR analysis revealed that expression of SMIT1 in INS-1E cells was 300-fold higher than that of SMIT2 (Fig. 1C), with a primer extension efficiency of 96.3% for SMIT1 and 98.1% SMIT2. Immunoblot analysis further demonstrated an 80-kDa SMIT1-specific antibody-reactive protein band in assays of samples from rat pancreatic islets and INS-1E cells (Fig. 1D). Further analyses of the purified rat pancreatic β-cells revealed a twofold higher SMIT1 mRNA expression in the β-cell population than in that of the non–β-cell population (Fig. 1E). A purity of 97.3% was achieved (Fig. 1F), indicating that SMIT1 is specifically expressed at a high level in rat pancreatic β-cells.

MI is known to play an important role in regulating intracellular osmolarity and fluid balance (19). To assess the osmoregulatory role of SMIT1, INS-1E cells were treated with 28 mmol/L mannitol for 6 h. SMIT1 mRNA expression was significantly increased (Fig. 2A). Further immunoblot analysis revealed that the protein expression was also upregulated in INS-1E cells treated with 28 mmol/L mannitol for 12 h (Fig. 2B) relative to the controls. Similarly, exposure of INS-1E cells to acute high-glucose concentrations (28 mmol/L glucose for 6 h) increased SMIT1 mRNA expression by twofold relative to controls (Fig. 2C). In addition, the protein expression level of SMIT1 was also augmented in INS-1E cells treated with 28 mmol/L glucose for 12 h (Fig. 2D). To exclude the effect of insulin on SMIT1 expression under high-glucose conditions, INS-1E cells were further treated with 5 µmol/L GSK1838705A, an insulin receptor blocker. The results revealed that the insulin receptor blocker failed to reduce the augmented SMIT1 expression in high-glucose–treated INS-1E cells (Fig. 2D). Furthermore, treatment of insulin alone on INS-1E cells did not have a significant effect on SMIT1 expression (Fig. 2E).

We further examined the effect of hyperglycemia on SMIT1 expression using the ZDF diabetic rat model. Our results showed that islets from 11- to 12-week-old ZDF diabetic rats expressed SMIT1 mRNA fivefold higher than that observed in islets from lean rats (Fig. 2F), consistent with a response to extracellular hypertonic stress. Immunoblot analysis of islets from 11- to 12-week-old ZDF diabetic rats also revealed an increased SMIT1 expression (Fig. 2G). Surprisingly, we found that SMIT1 mRNA expression level in islets from 7- to 8-week-old ZDF diabetic rats was also upregulated (Fig. 2H).

Given that diabetes is associated with abnormal inositol metabolism (4–6), we hypothesized that chronic exposure to high-glucose and/or chronic hyperglycemia would reduce SMIT1 expression. Our quantitative PCR experiment demonstrated significant decreases in SMIT1 mRNA expression of INS-1E cells after a 48-h or 72-h high-glucose treatment (Fig. 3A). Levels of SMIT1 protein were also downregulated in INS-1E cells in the 72-h high-glucose–treatment group (Fig. 3B). In contrast to the augmented SMIT1 mRNA expression in early-stage diabetic rat islets, we found that the pancreatic islets of more mature ZDF rats (24–36 weeks old) displayed significantly reduced SMIT1 mRNA expression (Fig. 3C). Immunoblot analysis also revealed a reduction in SMIT1 protein expression in mature ZDF rat islets (Fig. 3D). To rule out the influence of age on the downregulation of SMIT1 expression, we compared the expression of SMIT1 mRNA in islets from young lean rats (8 weeks old) and more mature lean rats (24–36 weeks old). The results showed that the respective SMIT1 mRNA expression levels were comparable (Fig. 3E), indicative of age independence.

We next examined whether SMIT1 is involved in the regulation of insulin secretory function. To address this issue, we used phlorizin, a nonspecific inhibitor of SGLT (29), to study SMIT1-mediated GSIS. As a first step, an expression profile of SGLT family in INS-1E was performed to determine the presence of any other SGLT family members, the activities of which could be inhibited by phlorizin. RT-PCR experiments confirmed that INS-1E cells express SMIT1 mRNA and, to a lesser extent, SMIT2 mRNA, but the other transporter family member mRNAs were not detected (Supplementary Fig. 1). SMIT inhibition with phlorizin (0.5 mmol/L) for 48 h impaired GSIS in INS-1E cells cultured in a normal glucose (11.2 mmol/L) condition, and no further impairment of GSIS was observed in cells cultured in high-glucose (28 mmol/L) conditions (Fig. 4A). Subsequent incubation of the treated cells in a phlorizin-free solution with 500 µg/mL MI for 3 h rescued normal GSIS in phlorizin-treated cells (Fig. 4B). MI supplementation produced fold-magnitude changes in the GSIS of the treated cells (see histogram in Fig. 4C). Rat pancreatic islets that were treated with 0.5 mmol/L phlorizin for 48 h also showed impaired GSIS (Fig. 4D and E). Moreover, phlorizin-treated cells exhibited a significant reduction in intracellular insulin content (Fig. 4F). When the insulin secretion level was expressed as the percentage of insulin content, its level was still significantly reduced, indicating that insulin biosynthesis and secretion were both affected (Fig. 4G).

To further confirm the functional role of SMIT1 on GSIS, we also performed siRNA-mediated knockdown of SMIT1 in INS-1E cells. Our results showed that transfection of INS-1E cells with SMIT1-specific siRNA led to reductions in 75% of SMIT1 mRNA expression and 50% of SMIT1 protein expression as well as 50% of the intracellular MI level (Supplementary Fig. 2). This SMIT1 knockdown consistently resulted in impaired GSIS relative to control siRNAs (Fig. 4H and I). Furthermore, SMIT1 knockdown also significantly decreased intracellular insulin content (Fig. 4J) and secretion level expressed as a percentage of insulin content (Fig. 4K).

Given that MI is a major precursor for PI (11), we hypothesized that alterations in SMIT1 activity would modulate intracellular PI levels. Firstly, RT-PCR experiments were performed to demonstrate that INS-1E cells are able to express the major enzymes involved in the conversion of MI to PI (Supplementary Fig. 3). Secondly, our ELISA results confirmed that SMIT1 knockdown with siRNA-SMIT1 reduced intracellular PIP₂ levels in INS-1E cells (Fig. 5A). We next sought to investigate whether the downstream of PI signaling cascades would also be affected by the deficiency of SMIT1. INS-1E cells with SMIT1 knockdown displayed a significant reduction in the intracellular IP₃ level, but the transfected cells were unresponsive toward a high-glucose challenge at 28 mmol/L for 5 min (Fig. 5B). Furthermore, SMIT1-knockdown INS-1E cells failed to exhibit an increase in intracellular [Ca²⁺] in response to 28 mmol/L glucose treatment (Supplementary Fig. 4). Given that SMIT1 knockdown impaired PIP₂ signaling cascades, we next studied its effect on PI3K/phosphatidylinositol 3,4,5-triphosphate (PIP₃)/Akt signaling cascades. AKT phosphorylation was used as the first reporter. SMIT1 knockdown reduced basal AKT phosphorylation and stimulated the phosphorylation level upon acute insulin treatment (100 nmol/L) in INS-1E cells (Fig. 5C). Similarly, ERK phosphorylation was reduced at basal and insulin-stimulated conditions (Fig. 5D). Meanwhile the expression of immediate downstream target genes of the PI3K/Akt signaling pathway was further used as a readout for any potential disturbances over the signaling cascades. Our results demonstrated that SMIT1 knockdown also reduced those transcription of genes that encode proteins responsible for β-cell function and known downstream transcription targets of PI3K/Akt signaling (30,31), namely Pdx1, GLUT2, and Ins2 (Fig. 5E–G). Interestingly, SMIT1 knockdown differentially altered insulin/IGF-I signaling in INS-1E cells. Specifically, we found that the mRNA expression levels of Irs1 and Irs2 were decreased and increased, respectively (Fig. 5H and I).

For the in vivo functionality assessments, we sought to examine whether MI supplementation is beneficial to diabetic conditions. Our results showed that treatment of pancreatic islets isolated from 24- to 36-week-old ZDF rats with 500 µg/mL MI for 48 h resulted in a rescue of GSIS (Fig. 6A and B). For in vivo MI supplementation (0.57 ± 0.04 mg/d · g for 1 month), the treatment did not alter body weight or food or water intake between MI-treated and untreated ZDF rats (Supplementary Table 2). However, the supplementation improved glucose tolerance in 24- to 36-week-old ZDF rats, as reflected by reduced fasting blood glucose level (Fig. 6C) and improved IPGTTs with reduced glucose concentrations (AUC results) relative to untreated age-matched ZDF controls (Fig. 6D and E). However, we did not observe any significant differences in insulin sensitivity between MI-treated and untreated ZDF rats, as evidenced by ITTs (Fig. 6F and G) and HOMA-insulin resistance data (Fig. 6H).

We also proceeded to examine the direct pancreatic islet function of MI-treated rats and found that islets isolated from the MI treatment group exhibited improved GSIS compared with that of untreated ZDF control islets (Fig. 7A and B). Consistently, there was a significant increase in serum insulin concentration in the MI-treated ZDF rats at 15 and 30 min after glucose administration compared with that of the untreated ZDF rats (Fig. 7C), indicating an improvement in islet function, as further reflected by an increased HOMA-B index (Fig. 7D). To investigate the beneficial effects of MI supplementation on islet function, islet morphology was studied by characterizing the β-cell–to–α-cell ratio and β-cell proliferation using immunohistochemical staining. MI supplementation remarkably increased the β-cell–to–α- cell ratio (Fig. 7E–G) and β-cell proliferation in ZDF rat islets, as demonstrated by an increase in Ki-67–positive cells to total β-cells in the MI-treated group (Fig. 7H and I).

To further elucidate the mechanism by which MI supplementation improves islet function, we characterized the intracellular PIP₂ levels in MI-treated and untreated ZDF rat islets. Results showed that ZDF control islets exhibited a remarkable reduction in PIP₂ level compared with that of lean rats, whereas MI supplementation partially restored the decreased intracellular PIP₂ level in ZDF rat islets (Fig. 7J). In addition, MI supplementation significantly elevated the basal phosphorylation level of AKT (Fig. 7K) and ERK (Fig. 7L). Moreover, MI-treated ZDF rat islets displayed an augmented insulin content (Fig. 7M) and percentage of insulin secreted from intracellular insulin content compared with the ZDF control rat islets (Fig. 7N).

In the current study, we report for the first time the expression of Na⁺-dependent MI cotransporters in pancreatic islets and β-cells and provide a plausible molecular explanation for prior observations of active MI transport in pancreatic islets (12,13). The particularly strong expression of SMIT1 observed in this study suggests that there is a dominant role of SMIT1 in β-cell MI transport and that the effects of pharmacological SMIT inhibition with phlorizin can be attributed, at least in large part, to SMIT1. We obtained ex vivo and in vivo data showing that MI administration alone can improve islet function, potentiating insulin secretion in glucose-challenge experiments. Genetic knockdown of SMIT decreased expression of Irs1 and increased expression of Irs2, consistent with a recent report showing that inhibition of insulin receptor substrate signaling elevates Irs2 expression as a form of feedback regulation (32). On the basis of our findings, we have developed a novel molecular model of SMIT1-mediated regulation on PI signaling in pancreatic β-cells via manipulation on the intracellular PIP₂ level. Furthermore, we demonstrated that MI, being a cellular osmolyte and the precursor of PI, is involved in both osmoregulation and intracellular signaling events and regulates insulin secretion in pancreatic β-cells.

Osmoregulation in brain and kidney cells under hypertonic stress is known to require SMITs and MI (33,34), and previous studies have shown SMIT1 being upregulated in response to extracellular hypertonic stress in neural cells (35,36). Although β-cells are also prone to extracellular hypertonicity, their responses to fluid balance with regulatory mechanisms are not well characterized. Our view is that the presently observed upregulation of SMIT1 expression in response to acute hypertonic stress, resulting from transient high-glucose exposure or early-stage hyperglycemia, reflects an adaptive response; it is because MI transporter upregulation enables β-cells to accumulate osmolytes and thereby raises their intracellular osmolarity for the maintenance of cell volume and fluid balance. Such autoregulation of cell volume may be important for β-cell function given that hypotonicity alone has been shown to induce β-cell insulin release (37,38).

Because MI is a PI precursor, the MI transporter SMIT1 may influence intracellular signaling events by manipulating the PI pool. Our finding that knockdown of SMIT1 in INS-1E cells reduced intracellular PIP₂ levels in β-cells is consistent with prior study (i.e., increased SMIT1 activity raised intracellular levels of both MI and PI), thereby altering the behavior of PIP₂-dependent ion channels in ganglion neurons (33). Furthermore, we observed that knockdown of SMIT1 led to impaired downstream PIP₂ signaling, including PIP₂/IP₃/Ca²⁺ and PI3K/Akt signaling, such that a reduction in AKT phosphorylation was observed and the transcription of genes that encode β-cell function regulating proteins was downregulated. Indeed, previous studies have reported that PI, being a major intracellular second messenger signaling molecule, has an important physiological role in β-cell function (39,40). In particular, PIP₂ generates two downstream effectors, diacylglycerol and IP₃; these effectors have been shown to regulate GSIS by promoting the maturation and exocytosis of insulin granules (41,42). PIP₃, a downstream counterpart of PIP₂, has also been shown to be involved in the regulation of GSIS as well as the proliferation and survival of β-cells through PI3K/Akt signaling (40). In light of these findings, our present data introduce a previously unidentified signaling cascade between SMIT1 and PI signaling in β-cells, which we have modeled in Fig. 8.

Interestingly, our data show that SMIT1 expression was downregulated in islets cultured under chronic high-glucose exposure and hyperglycemic conditions but augmented after acute high-glucose treatment. The data are indeed in comparison with different glycemic states of different stages during diabetic development. The ZDF rat is a well-established model for type 2 diabetes (43,44). Our data showed that although young ZDF rats attained normal fasting blood glucose levels at 6–8 weeks of age (stage 1), their random blood glucose levels were significantly elevated (Supplementary Fig. 5); by 9–12 weeks of age through ∼20 weeks (stage 2), they become fully diabetic, with evident hyperglycemia and hyperinsulinemia (Supplementary Fig. 5); and from 20 weeks old onwards (stage 3), they become severely hyperglycemic, yet their plasma insulin concentrations reduced remarkably, as reported previously (43,44). It is surprising to identify that SMIT1 mRNA expression has already been upregulated when young ZDF rats display elevated random blood glucose levels. We postulate that this might be a compensatory mechanism by which the rat pancreatic islets adapt to the steeply rising demand for insulin secretion during the development of insulin resistance at that stage. Such upregulation was also observed in islets from the 11- to 12-week-old ZDF rats, inconsistent with the response toward hypertonic stress in high-glucose–treated INS-1E cells. On the contrary, SMIT1 mRNA expression was significantly downregulated in the islets from the 24- to 34-week-old ZDF rats. This reversal may represent a decompensated osmoregulation of β-cells under chronic high-glucose exposure, which may be relevant for elucidating the pathogenesis of type 2 diabetes.

Although our molecular model is suggestive of SMIT1-mediated regulation of PI signaling cascades, it is also possible that SMIT1 might modulate insulin secretion in β-cells through PIP₂-modulated effects on ion channel activities, similar to the SMIT1 regulation of ion channels in ganglion neurons (33). Our data showed that such regulatory events might also occur in β-cells given that PIP₂/PLC/IP₃ signaling, which regulates insulin secretion through intracellular Ca²⁺ modulation (14), was downregulated upon SMIT1 knockdown. Furthermore, reductions in PIP₂ have been shown to elevate basal β-cell insulin secretion by desensitizing K_ATP channels to intracellular ATP (45). However, to our surprise, basal insulin secretion was not augmented in our SMIT1 knocked-down INS-1E cells. Given prior data implicating the PI phosphorylation cycle in the regulation of PIP₂ turnover in β-cells (46), we postulate that a dramatic reduction in PIP₂ levels might be necessary to desensitize the K_ATP channels and that, in SMIT1 knocked-down cells, de novo synthesis of PIP₂ may compensate partially for such a PIP₂ loss. Molecular studies on the relationships between SMIT1, PIP₂, and K_ATP are necessary to test this hypothesis. Interestingly, the presently reported beneficial effects of MI supplementation on β-cell and islet function in diabetes are consistent with prior clinical studies showing that MI supplementation can prevent gestational diabetes (7–9), as well as with studies showing MI-related improvements in insulin sensitivity in diabetic mice (47,48). To this end, studies of SMIT1 knockout mice (49) could be used to further characterize the in vivo functional role of SMIT1 in rodent models.

In conclusion, given the relatively ubiquitous expression of SMIT1 (19,22), our data provide not only a new perspective on the pathophysiology of type 2 diabetes but also broad physiological significance with respect to molecular aspects of other diseases characterized by altered inositol metabolism.

Funding. This work was fully supported by the General Research Fund of The Research Grants Council of the Hong Kong Special Administrative Region, China (Ref. No.: CUHK14107415), awarded to P.S.L.

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

Author Contributions. S.Y.T.L. designed and performed experiments, analyzed and interpreted data, and drafted the manuscript. S.T.W.C. and D.Z. performed experiments and analyzed data. P.S.L. designed the experiments, analyzed and interpreted data, and revised the manuscript. P.S.L. is the guarantor of this work and, as such, had full access to all of 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 75th Scientific Sessions of the American Diabetes Association, Boston, MA, 5–9 June 2015.