Nitric oxide and insulin secretion
The potential role of NO in insulin secretion has been widely disputed, and the results obtained are highly controversial. It has been reported that NO stimulates (Laychock et al., 1991[70]; Schmidt et al., 1992[110]; Willmott et al., 1995[133]; Ding and Rana, 1998[26]; Spinas et al., 1998[119]; Matsuura et al., 1999[84]; Spinas, 1999[118]; Smukler et al., 2002[116]; Nystrom et al., 2012[95]; Gheibi et al., 2017[40]), inhibits (Panagiotidis et al., 1992[101], 1994[99], 1995[100]; Gross et al., 1995[45]; Akesson and Lundquist, 1999[2]; Akesson et al., 1996[3], 1999[1]; Antoine et al., 1996[5]; Salehi et al., 1996[107], 1998[108]; Henningsson and Lundquist, 1998[51]; Henningsson et al., 1999[48], 2000[49], 2001[50]; Tsuura et al., 1998[127]) or has negligible effect (Jones et al., 1992[61]; Gheibi et al., 2018[41]) on insulin secretion in studies using islets and β-cell lines with various types and concentrations of NOS inhibitors and NO donors (Table 2(Tab. 2); References in Table 2: Akesson et al., 1999[1]; Ding and Rana, 1998[26]; Gheibi et al., 2017[40], 2018[41]; Henningsson et al., 2002[52]; Jiminez-Feltstrom et al., 2005[60]; Laychock et al., 1991[70]; Mezghenna et al., 2011[90]; Nystrom et al., 2012[95]; Panagiotidis et al., 1995[100]; Salehi et al., 1998[108]; Smukler et al, 2002[116]; Spinas et al., 1998[119]; Tsuura et al., 1998[127]). In the following sections, we will discuss in more detail how NO may stimulate or inhibit insulin secretion.

Stimulatory effect of NO on insulin secretion
The initial evidence that NO played a role in the regulation of insulin secretion came from Laychock and colleagues in 1991 (Laychock et al., 1991[70]). They found that sodium nitroprusside by increasing the cGMP level in rat islets, stimulated insulin secretion, while inhibition of NOS decreased glucose- and arginine-induced cGMP release (Laychock et al., 1991[70]). This was further supported by the finding that L-arginine-derived NO increases basal and GSIS in isolated mouse islets (Henningsson and Lundquist, 1998[51]; Henningsson et al., 1999[48]) and the glucose-responsive clonal pancreatic β-cell line HIT-T15 (Schmidt et al., 1992[110]). A concomitant release of insulin and NO is induced by L-arginine in the presence of D-glucose, with the median effective arginine concentrations (EC50) for insulin and NO release equal to 150 µM and 50 µM, respectively, both of which are within the physiological range of circulating L-arginine levels. Interestingly, L-arginine also decreases the EC50 for D-glucose's stimulation of both NO and insulin release (from 15 mM to 5 mM) (Schmidt et al., 1992[110]).
Endogenously produced NO also plays an important role in insulin secretion. Indeed, scavenging of NO with cPTIO (carboxy-2-phenyl-4, 4, 5, 5-tetramethylimidazoline-1-oxyl 3-oxide) in highly glucose-responsive INS-1 cells, is able to significantly reduce the stimulation provided by 15 mM glucose (by ~40 %) (Smukler et al., 2002[116]), highlighting the involvement of endogenously produced NO in secretagogue-induced insulin secretion under physiological conditions. It also seems that the early phase of insulin secretion is NO-dependent as scavenging of endogenous NO or inhibition of NOS with L-NMMA (NG-Monomethyl-L-arginine, monoacetate) in rat pancreatic islets blunts the early insulin peak by 60-65 % and 46 %, respectively (Spinas et al., 1998[119]). This finding may also explain why some studies looking into accumulated insulin release in pancreatic islets argue against a stimulatory effect of NO on insulin release. The mechanism by which NO stimulates insulin secretion is shown in Figure 1(Fig. 1) and also discussed below.

Nitric oxide increases insulin synthesis
Increased insulin synthesis has been reported following NO treatment both in Min6 β-cells as well as in intact pancreatic islets (Campbell et al., 2007[16]). High fat diet STZ-induced diabetic rats were shown to have low islet insulin content; long term nitrite supplementation in these rats increased islet insulin content, indicating increased insulin synthesis (Gheibi et al., 2017[40]). Nitric oxide (NO gas) stimulates the activity of the insulin gene promoter in Min6 β-cells with a maximal 2.5-fold stimulation at 24 h, an effect which is reversed using PI3-kinase inhibitor (wortmannin), indicating that PI3-kinase activity is essential for the effects of NO on insulin gene promoter activity (Campbell et al., 2007[16]). In addition, NO (NO gas) increases endogenous insulin gene expression in Min6 cells and isolated rat islets of Langerhans, promoting the nuclear accumulation of PDX-1 (pancreatic and duodenal homeobox factor-1) and its subsequent binding to the insulin gene promoter (Campbell et al., 2007[16]). PDX-1 plays an important role in β-cells by linking glucose metabolism to events in the β-cell nucleus; in response to elevated glucose levels, PDX-1 is mobilized from its resting position in the cytoplasm into the β-cell nucleus, where it binds and activates the insulin gene promoter (Macfarlane et al., 1999[81]). By contrast, treatment with S-nitrosoglutathione (GSNO), a source of bioavailable NO, has no effect on islet insulin content in human and rat islets (Hadjivassiliou et al., 1998[47]). Similarly, L-NAME (N(ω)-nitro-L-arginine methyl ester) does not affect the attenuation of proinsulin synthesis, or the depletion of islet insulin content induced by palmitate (Bachar et al., 2010[8]).

Nitric oxide increases cGMP levels
Increased cGMP level is a putative mechanism through which NO exerts its action; L-arginine with D-glucose, in the absence or presence of a non-selective phosphodiesterase inhibitor (isobutyl-methylxanthine), increase the level of cGMP in rat pancreatic islets and in HIT-T15 cells (Schmidt et al., 1992[110]). In addition, direct exposure to the NO donor, 3-morpholinosydnonimine (SIN-1), also elevates basal cGMP levels in HIT-T15 cells (Schmidt et al., 1992[110]). The stimulatory effect of hydroxylamine on insulin secretion is abolished by GC inhibition with ODQ (1H-[1, 2, 4]oxadiazolo[4, 3-a]quinoxalin-1-one) while that stimulatory ability of NO is mimicked by the activation of GC using YC-1 (3-(5-hydroxymethyl-3-furyl)-1-benzylindazole), and by the membrane-permeable cGMP analog (8-(4-chlorophenylthio)-cGMP) in INS-1 cells and rat islets (Smukler et al., 2002[116]), again supporting the notion that NO acts principally by stimulating GC. Elevated cGMP levels increase Ca2+ influx at 7 mM but not at 2.8 mM of glucose. This appears to result from activation of VDCC in rat pancreatic β-cells as the [Ca2+]i elevation is abolished by nicardipine, a dihydropyridine class of Ca2+ channel blockers (Matsuura et al., 1999[84]). Interestingly, it is believed that low NO levels act through the NO/sGC/cGMP pathway while high NO levels exert their effects independently of cGMP (Lazo-de-la-Vega-Monroy and Vilches-Flores, 2014[71]). Moreover, the NO/sGC/cGMP pathway participates in other positive effects for β-cells, such as enhancing islet blood flow (Nystrom et al., 2012[95]) and decreasing apoptosis (Tejedo et al., 2001[126]).

Nitric oxide increases intracellular Ca2+ levels
Nitric oxide increases intracellular Ca2+ levels through mobilization of Ca2+ from intracellular stores, such as the endoplasmic reticulum and mitochondria (Laffranchi et al., 1995[67]; Willmott et al., 1995[133]) or through inhibition of KATP channels and subsequent membrane depolarization, leading to opening of VDCCs (Smukler et al., 2002[116]). Application of diazoxide, a specific activator of KATP channels, is able to inhibit hydroxylamine-stimulated insulin release in INS-1 cells (Smukler et al., 2002[116]); in this study, Ca2+ imaging with Fura-2 demonstrated that hydroxylamine stimulates a spiking, oscillatory elevation of [Ca2+]i, similar to that seen with glucose (15 mM) and this [Ca2+]i response occurs simultaneously with membrane depolarization (Smukler et al., 2002[116]).
Nitric oxide (NO gas), by binding to cytochrome c and/or cytochrome oxidase inhibits the mitochondrial respiratory chain, causing decreased mitochondrial membrane potential and mobilization of Ca2+ from mitochondria (Schweizer and Richter, 1994[112]). Addition of low concentrations of NO to INS-1 cells results in a rapid increase in insulin secretion, which is paralleled by decreased mitochondrial membrane potential and also an intermittent rise of cytosolic Ca2+. Furthermore, NO-induced Ca2+ release from the mitochondria and increased insulin secretion are independent of extracellular Ca2+, as chelation of intracellular but not of extracellular Ca2+, decreases NO-induced insulin secretion (Laffranchi et al., 1995[67]). In addition, when intracellular Ca2+ levels are raised in advance, the NO-induced cGMP elevation restores normal intracellular Ca2+ levels via Ca2+ sequestration into the endoplasmic reticulum (Matsuura et al., 1999[84]), suggesting that NO, via elevation of cGMP, not only increases insulin secretion (Ishikawa et al., 2003[56]), but also protects β-cells from excessive Ca2+ increases, which may lead to apoptosis (Kaneko et al., 2003[62]).
Elevated Ca2+ levels also increase NO production in INS-1 cells (Smukler et al., 2002[116]) and it is interesting to note that induction of NO production is an early event in the onset of the insulin secretion process. Although NO production can be detected prior to an increase in intracellular Ca2+ levels (Nunemaker et al., 2007[94]), it is not clear that this happens because the Ca2+ levels required for NOS activation are actually below the threshold for detection by Fura-2. Indeed, this interpretation is compatible with a body of literature which demonstrates requirement of Ca2+/calmodulin binding for the activation of NOS (Spratt et al., 2007[120]). As such, one could speculate that producing NO at such an early point would support the idea of a positive role, rather than a negative one, and that for negative regulation to occur, a certain threshold must be crossed.

Nitric oxide acts through S-nitrosylation
Nitric oxide through S-nitrosylation of glucokinase (at cysteine-371) and syntaxin 4 (at cysteine-141) facilitates GSIS (Rizzo and Piston, 2003[105]; Wiseman et al., 2011[134]; Kruszelnicka, 2014[65]; Seckinger et al., 2018[113]). Using quantitative imaging of multicolor fluorescent proteins fused to glucokinase, it has been demonstrated that the dynamic association of glucokinase with secretory granules is modulated by NO (Rizzo and Piston, 2003[105]). Indeed, insulin is found to stimulate NO production leading to S-nitrosylation of glucokinase in cultured β-cells (βTC3 cells); moreover, inhibition of NOS disrupts glucokinase association with secretory granules and glucokinase conformation (Rizzo and Piston, 2003[105]). It has been demonstrated that elevated glucose and S-nitrosylation, both induce the same high-activity glucokinase conformational state I in βTC3 cells (Seckinger et al., 2018[113]). Ultimately, attachment of a nuclear localization signal sequence to NOS in βTC3 cells drives glucokinase to the nucleus in addition to its normal cytoplasmic and granule targeting (Rizzo and Piston, 2003[105]). These data suggest that the regulation of glucokinase localization and activity in pancreatic β-cells is directly related to NO production and that the association of glucokinase with secretory granules occurs through its interaction with NOS.
The SNARE proteins are post-translationally modified by NO, which in turn impacts vesicle exocytosis; NO-stimulated vesicle exocytosis was found to be mediated by activation of the core complex proteins involved in docking and fusion of the vesicles (Meffert et al., 1996[89]). In experiments using recombinant proteins, NO donors (sodium nitroprusside (SNP), acidified sodium nitrite, S-nitrosoglutathione, S-nitrosocysteine and a saturated solution of NO gas) increase formation of the VAMP/SNAP-25/syntaxin 1a core complex and inhibit the binding of n-sec1 to syntaxin 1a (Meffert et al., 1996[89]); these NO donors lower the EC50 of VAMP binding to SNAP-25/syntaxin (Meffert et al., 1996[89]). In addition, S-nitrosylation of syntaxin 1a (at cysteine-145) seems to be a molecular switch to disrupt Munc18-1 binding to the closed conformation of syntaxin 1a, thereby facilitating its engagement with the membrane fusion machinery (Palmer et al., 2008[98]). Taken together, these activities are predicted to promote vesicle docking/fusion.
At present, only the t-SNARE protein syntaxin 4 has been shown to be specifically S-nitrosylated in pancreatic β-cells; once S-nitrosylated, this event facilitates GSIS (Wiseman et al., 2011[134]; Kruszelnicka, 2014[65]). Interestingly, while syntaxin 1 is a target of S-nitrosylation in neuronal cells and tissues (Pongrac et al., 2007[102]; Palmer et al., 2008[98]), it is not modified in response to glucose in the pancreatic β-cells, demonstrating that a similar complement of exocytotic proteins results in differing functional rate kinetics across different tissues. The cellular content of S-nitrosylated syntaxin 4 has been shown to peak acutely within 5 min of glucose stimulation in both human islets and MIN6 β-cells (Wiseman et al., 2011[134]). Mutation in the gene encoding cysteine-141-syntaxin 4 prevents S-nitrosylation induced by the NO donor (S-nitrosoglutathione), fails to exhibit glucose-induced activation and VAMP2 binding, and consequently fails to potentiate insulin release (Wiseman et al., 2011[134]). In addition, monitoring of single-cell exocytosis, using the styryl dye, FM1 43, demonstrates that NO (hydroxylamine, SNP, and SIN-1) exerts a rapid stimulatory effect on insulin secretion from INS-1 cells (Smukler et al., 2002[116]).

Nitric oxide increases islet blood flow
Increased islet blood flow, which supplies oxygen and nutrients to the islets, is another mechanism by which NO may increase insulin secretion (Nystrom et al., 2012[95]). Nitric oxide has been suggested as an important regulator of islet blood flow; insulin secretion can be rapidly modulated by changes in rat islet microcirculation (Jansson and Hellerstrom, 1983[57]). The islets have a well-developed vascular network with a highly fenestrated endothelium in capillaries, thus allowing rapid and efficient delivery of oxygen and nutrients to the endocrine cells. Interestingly, islet blood flow can be regulated independently from the exocrine part of the pancreas (Bonner-Weir and Orci, 1982[14]; Jansson and Hellerstrom, 1983[57]). Nystrom and colleagues studied the effects of nitrite on pancreatic rat islet blood flow and dynamic changes in insulin secretion and glycemia (Nystrom et al., 2012[95]). They found that sodium nitrite increases islet blood flow, with no such effect on total pancreatic blood flow; the enhancement of islet blood flow was followed by an increase in plasma insulin concentrations. However, glycemia was unchanged (Nystrom et al., 2012[95]), suggesting mobilization of counterregulatory hormones, such as glucagon and cortisol, opposing insulin action. In addition, nitrite-increased islet blood flow was abolished by a GC inhibitor and a NO scavenger and mimicked by a cGMP agonist (Nystrom et al., 2012[95]). This indicates that NO-induced vasodilation is mediated through the GC/cGMP/ PKG pathway. Inhibition of NOS by LNAME, was shown to decrease whole pancreatic and, in particular, islet blood flow in normal and diabetic rats (Svensson et al., 1994[124]). Indeed, to accommodate the elevated demand for insulin delivery into the peripheral circulation, islet capillaries expand through dilation but not by angiogenesis (Dai et al., 2013[24]).
Taken together, emerging evidence has demonstrated that NO in the β-cells, through increasing cGMP and intracellular Ca2+ levels, or via S-nitrosilation of glucokinase and syntaxin 4 as well as vasodilation of islet vasculature, increases insulin secretion.