Mechanisms of statin‐induced new‐onset diabetes

Abstract

Statins, with their lipid‐lowering properties, are a first‐line therapy for the prevention of cardiovascular diseases. Recent evidence, however, suggests that statins can increase the risk of new‐onset diabetes (NOD). The molecular mechanisms of statin‐induced NOD are not precisely known, although some pathophysiologic mechanisms have been suggested. Specific to the beta cell, these mechanisms include alterations in insulin secretion, changes in ion channels, modulation of signaling pathways, and inflammation/oxidative stress. Outwith the beta cell, other suggested mechanisms involve adipocytes, including alterations in adipocyte differentiation and modulation of leptin and adiponectin, and genetic and epigenetic mechanisms, including alterations in microRNA. The evidence supporting these and other mechanisms will be discussed. Greater understanding of the underlying mechanisms linking the onset of diabetes to statin therapy is essential and clinically relevant, as it may enable novel preventative or therapeutic approaches to be instituted and guide the production of a new generation of statins lacking this side effect.

Recent evidence, however, suggests that statins can increase the risk of new‐onset diabetes (NOD; the molecular mechanisms of statin‐induced NOD are not precisely known, although some pathophysiologic mechanisms have been suggested. Greater understanding of the underlying mechanisms linking the onset of diabetes to statin therapy is essential and clinically relevant, as it may enable novel preventative or therapeutic approaches to be instituted and guide the production of a new generation of statins lacking this side effects.

1. INTRODUCTION

Cardiovascular disease (CVD) is the leading cause of mortality worldwide (Organization, 2011). One major risk factor for CVD is hypercholesterolemia, and so this is a very important therapeutic target for CVD treatment and prevention (Mendis, 2010). Statins, 3‐hydroxy‐3‐methyl‐glutaryl coenzyme‐A (HMG‐CoA) reductase inhibitors, therefore have a central role in primary and secondary CVD prevention via reduction of low‐density lipoprotein cholesterol (LDL‐C) (Taylor, Huffman, & Ebrahim, 2013).

Whilst the benefits of statins have been well documented in individuals younger than 75 years old, the value of statins in people aged >75 years is controversial. The Cholesterol Treatment Trialists' (CTT) meta‐analysis calculated an absolute risk reduction of 0.6%/year per 38.7 mg/dl reduction in LDL‐C levels in patients aged >75 years. However, the absolute effect of a 38.7 mg/dl cholesterol lowering on the rate of annual ischemic heart disease mortality is 10‐fold larger in older vs younger patients (Ruscica et al., 2018).

Improvement in endothelial function, stabilization of atherosclerotic plaque, antioxidant, antithrombotic, and anti‐inflammatory actions is a few of the multitude of effects of the lipid‐independent and so‐called pleiotropic actions of statins (Chruściel et al., 2016; Davignon, 2004; Parizadeh et al., 2011; Sahebkar et al., 2015; Sahebkar, Serban, Mikhailidis et al., 2015; Sahebkar et al., 2016; Sahebkar, Serban, Ursoniu et al., 2016; Sahebkar, Ponziani, Goitre, & Bo, 2015; Sahebkar, Pećin Rathouska, Derosa, Maffioli, & Nachtigal, 2016; Serban et al., 2015). However, in addition to these positive effects, some adverse effects of statins have been identified, such as an increase in liver enzymes and myopathy (Bellosta, Paoletti, & Corsini, 2004). Recently, there has been much interest in the potential unforeseen adverse effects of statins, and specifically in promoting the development of type 2 diabetes mellitus (T2DM; Brault, Ray, Gomez, Mantzoros, & Daskalopoulou, 2014b). Much research has shown that statins increase the risk of diabetes as much as 9%, meaning that one case in every 255 patients using statins for 4 years will develop new‐onset diabetes (Sattar et al., 2010). Because of this side effect of statins, researchers are looking for new therapeutic alternatives. For example Ruscica, Gomaraschi et al. (2014) showed that a nutraceutical combination containing red yeast rice (RYR) has a lipid‐lowering activity comparable with pravastatin, a finding consistent with other reports on RYR and its combination with berberine and other nutraceuticals (Banach et al., 2018; Bianconi, Mannarino, Sahebkar, Cosentino, & Pirro, 2018; Cicero et al., 2017; Johnston, Korolenko, Pirro, & Sahebkar, 2017; Pirro et al., 2016; Sahebkar, Serban, Gluba‐Brzózka et al., 2016; Ward, Sahebkar, Banach, & Watts, 2017). In March 2012, the US Food and Drug Administration reported that there was adequate evidence to warn about the risk of diabetes induced by statins (Food & Administration, 2012). The molecular mechanism of statin‐induced new‐onset diabetes mellitus (NODM) is complex and not clearly understood although several pathophysiologic mechanisms have been proposed (Chrysant, 2017). The aim of this review is to summarize the proposed mechanisms of new‐onset diabetes associated with statin therapy.

2. STATINS ACT ON CHOLESTEROL BIOSYNTHESIS PATHWAY

HMG‐CoA reductase reduces HMG CoA to mevalonate. It is an enzyme that is greatly regulated by the amount of cholesterol and also it is a rate‐determining enzyme in the cholesterol biosynthetic pathway. Statins inhibit HMG‐CoA reductase competitively. Statins have an HMG‐like portion along with a hydrophobic group. The HMG‐like portion of the statin reversibly competes and causes inhibition of the HMG binding portion of the active site of HMG‐CoA to binding. Enzymes in peroxisomes metabolize mevalonate to free style diphosphate, and then to squalene. Subsequently, squalene epoxides and oxidosqualene cyclone convert squalene into lanosterol. A cascade of oxidations, demethylation, and reductions then lead to cholesterol formation from lanosterol. Prenylated proteins, haem A, dolichol, and, most importantly, ubiquinone is formed along with squalene before the formation of lanosterol from mevalonate‐5‐PP (Olivier & Krisans, 2000).

3. PHARMACOLOGICAL DIFFERENCES BETWEEN STATINS

Statins are amphiphilic drugs. They need to enter cells, be it directly by membrane interactions in the case of lipophilic agents (simvastatin, fluvastatin, and atorvastatin) or by way of carrier proteins in the case of hydrophilic agents, such as pravastatin. Rosuvastatin has an intermediate behavior. The intestinal absorption of statins is quite variable, that is from 30% and to 85%. Most statins (the exceptions being pravastatin and partially rosuvastatin) undergo a first pass liver metabolism, with the systemic bioavailability being reduced to 5–30% of the administered dose. Most metabolites are pharmacologically active, again with the exception of pravastatin and also of fluvastatin.

Statins are predominantly metabolized by the cytochrome P‐450 (CYP) system. The exceptions, pravastatin, and rosuvastatin are hydrophilic molecules undergoing minimal metabolic handling. Metabolic transformation of statins explains to a minimal extent the variable cholesterol‐lowering activity (Sirtori, 2014).

The IC50s against HMG CoA reductase for each statin generally indicate a reduced range of daily dosage for the more powerful agents, the lower ones being found with pitavastatin. An exception is a cerivastatin, with an active daily dose <1 mg, associated, however, with a considerably increased risk of muscular and generalized toxicity (Simpson, 2001).

Atorvastatin and rosuvastatin have the longest half‐lives (20 hr) versus about 12 hr for simvastatin. All the others, including pitavastatin (half‐life, 10 hr), have a t½ˇ around 4 hr. While a prolonged half‐life may possibly improve activity in noncompliant patients, some studies suggest alternate‐day administration (Sirtori, 2014).

4. PROPOSED MECHANISMS FOR NEW‐ONSET DIABETES INDUCED BY STATINS

4.1. Inhibition of the 3‐hydroxyl‐3‐methylglutaryl‐CoA reductase receptor

Inhibition of the 3‐hydroxyl‐3‐methylglutaryl‐CoA reductase (HMGCR) receptor appears to play a role in the increased risk of new‐onset diabetes induced with statins. In Mendelian randomization analysis, two single‐nucleotide polymorphisms (SNPs) have been considered in the HMGCR gene as proxies of HMGCR inhibition by statins (Swerdlow et al., 2015). Each additional allele was associated with about a 2.3 mg/dl (0.06 mmol/l) lower LDL‐C level and a 0.66 pound (0.3 kg) higher body weight, along with slightly higher waist circumference, plasma glucose, and insulin concentration. These changes were accompanied by a 2–6% higher risk of diabetes (Robinson, 2015).

4.2. Ca⁺² channels in ß‐cells

Intracellular calcium (Ca²⁺) concentration, which is mainly controlled by the opening of voltage‐gated calcium channels, induces insulin secretion in pancreatic ß‐cells. Dysfunction or changes in the levels of these channels can significantly influence glucose homeostasis (Brault et al., 2014b). Yada, Nakata, Shiraishi, & Kakei (1999). demonstrated that simvastatin directly inhibited l‐type Ca²⁺ channels leading to blockage of glucose‐induced Ca²⁺ signaling in rat pancreatic islet ß‐cells. There is no clear mechanism suggested for the link between inhibition of cholesterol synthesis and Ca²⁺ channel dysfunction; however, it was proposed that incorrect sorting of membrane lipid‐raft bound proteins or alteration in the confirmation of the Ca²⁺ channel subunits may have caused these effects (Xia et al., 2008).

4.3. Glucose transporter 4

Glucose transporter 4 (SCL2A4 or GLUT4), a facilitated transporter, is responsible for peripheral insulin‐mediated glucose influx. It mediates glucose uptake in striated muscle cells and adipocytes (Khan & Pessin, 2002). GLUT4 recruitment from its intracellular storage to the plasma membrane requires initiation of the signaling cascade, stimulated by insulin‐receptor tyrosine kinase phosphorylation. Atorvastatin and lovastatin have been shown to induce insulin resistance via reduction of GLUT4 translocation to the cell membrane (Robinson, 2015).

4.4. Caveolin

One of the important plasma membrane proteins is caveolin‐1. GLUT4 is translocated to caveolae‐rich regions of the membrane when stimulated by insulin (Cohen, Combs, Scherer, & Lisanti, 2003; Gustavsson, Parpal, & StrÃ, 1996). Studies have shown that GLUT4 and caveolin‐1 are suppressed by atorvastatin, leading to a decrease in glucose uptake and an increase in insulin resistance in muscle, liver, and adipose tissue (Abel et al., 2001; Nakata, Kotani et al., 2006). This may lead to reduced translocation of GLUT4, and induction of hyperglycemia and hyperinsulinemia, early markers of noninsulin‐dependent diabetes mellitus (Brault et al., 2014b).

4.5. Insulin receptor substrate and insulin signaling

It has been proposed that statins induce new‐onset diabetes by attenuating insulin signal transduction by inhibition of necessary phosphorylation events and alteration in the cellular distribution of small G proteins.

4.5.1. Insulin receptor substrate‐1 and Akt

The insulin receptor substrate (IRS)‐1 phosphorylation pathway is the definitive pathway for insulin signaling. Insulin or insulin‐like growth factor (IGF) binding to the insulin receptor (IR) leads to insulin receptor substrate (IRS‐1) phosphorylation through phosphatidylinositol 3‐kinase (PI3K; Takaguri, Satoh, Itagaki, Tokumitsu, & Ichihara, 2008).

Phosphorylated Akt monitors glucose uptake through the regulation of GLUT4 translocation to the plasma membrane (Takaguri et al., 2008). IRS‐1 and Akt phosphorylation are reduced with atorvastatin therapy in a dose‐dependent manner (Takaguri et al., 2008). Atorvastatin therapy also leads to downregulation of IRS‐1 and IR‐ß levels in the differentiation process in adipocytes (Nakata, Kotani et al., 2006).

4.5.2. RhoA and Rab4

Two small G proteins, RhoA and Rab4, mediate insulin signaling transduction via regulation of IRS‐1 and Akt phosphorylation (Takaguri et al., 2008). Rab4, which is involved in intracellular vesicle transport of GLUT4 to the plasma membrane in response to insulin signaling, is present on GLUT4 vesicles in rat adipocytes (Cormont et al., 1996). Therefore, insulin‐stimulated glucose transport requires Rab4 (Kinsella & Maltese, 1992). Takaguri et al. (2008) have reported that atorvastatin downregulated the GLUT4 membrane expression, in part via induced Rab4 dysfunction.

In 3T3‐L1 adipocytes, RhoA modulates the activities of IRS‐1. Pretreatment with atorvastatin was shown to reduce the active membrane fraction of both RhoA and Rab4 while increasing the inactive cytosolic levels (Takaguri et al., 2008).

4.6. Inhibition of isoprenoid, coenzyme Q10, and dolichol biosynthesis

Statins also inhibit the biosynthetic pathway of cholesterol via blocking HMG‐CoA reductase. The most important products that are affected are dolichol, farnesyl pyrophosphate (FPP), geranylgeranyl pyrophosphate (GGPP), and coenzyme Q10 (CoQ10), also known as ubiquinone. Isoprenoids stimulate upregulation of GLUT4 in adipocytes leading to an increase in glucose uptake (Chamberlain, 2001). Chamberlain (2001) showed that suppression of isoprenoid synthesis plays a key role in lovastatin induced insulin resistance.

Dolichol is an important cofactor in N‐linked glycosylation, which is essential for intracellular membrane‐bound receptor processing (Carlberg et al., 1996). Lovastatin has been shown to significantly decrease dolichol and CoQ10 with a subsequent decrease in cholesterol biosynthesis in rats (Ciosek et al., 1993). CoQ10 is a necessary component of the mitochondrial respiratory chain and is therefore essential for ATP production. A decrease of mitochondrial ATP leads to closure of ATP dependent potassium channels on ß‐cells and thereby inhibits insulin secretion due to Ca²⁺ influx inhibition (Ashcroft et al., 1994). Circulating CoQ10 is transported by LDL; therefore, statin treatment may affect its levels. Lovastatin has been shown to reduce both circulating and tissue CoQ10 in humans, which may inhibit the insulin secretion of pancreatic ß‐cells by reducing the ATP production (Folkers et al., 1990).

4.7. Adiponectin and leptin

Adiponectin, also named adipocyte complement‐related protein, is an adipocyte‐specific 30 kDa secreted protein. Adiponectin has been extensively characterized as an insulin sensitizer. Indeed, in experimental in vivo models, adiponectin overexpression in adipose tissue led to an improvement in systemic insulin sensitivity, an effect that was lost when adiponectin was genetically deleted. Adiponectin suppresses hepatic gluconeogenesis and also exerts antiapoptotic activity on pancreatic beta‐cells, and promotes the reconstitution of beta‐cell mass (Ruscica, Baragetti, Catapano, & Norata, 2017)

Maeda et al (2002) found fatty acid catabolism disruption and increases in serum TNF‐a in adiponectin knockout (KO) mice. Insulin resistance and glucose levels were also increased in these KO mice when challenged with a high fat/high sucrose diet. TNF‐a and adiponectin show opposite effects: TNF‐a attenuates IRS‐1 PI3K activity, whereas adiponectin stimulates it, but only when TNF‐a is absent (Maeda et al., 2002). Because hypoadiponectinemia correlates with insulin resistance and obesity, the inhibitory action on of statins on the adiponectin level is a potential mechanism for new‐onset diabetes (Brault, Ray, Gomez, Mantzoros, & Daskalopoulou, 2014a). Wanders, Plaisance, & Judd (2012) recently summarized the current evidence on statins and the level of adiponectin. At the cellular level and in humans, some studies showed that the use of a statin (atorvastatin and simvastatin) was associated with a reduction in adiponectin levels and the subsequent diminution in insulin sensitivity (Koh et al., 2011)

Indeed, Koh et al. (2008) found that simvastatin therapy for two months at four different dosages (10, 20, 40, and 80 mg daily) in subjects with hypercholesteremia significantly decreased adiponectin independently of the dose, and showed an associated decrease in insulin sensitivity, as assessed via Quantitative Insulin‐Sensitivity Check Index.

Another adipocyte‐secreted hormone is leptin, a cytokine‐like hormone that plays an important regulatory role in energy metabolism by inducing satiety and increasing energy expenditure. Human leptin, a 16 kDa protein of 146 amino acids, circulates in the blood in two forms: the monomeric one (free leptin) and the higher molecular weight form which is bound to the soluble leptin receptor (Ruscica et al., 2017) and has been suggested as the leptin which positively regulates ß‐cell mass (Chetboun et al., 2012). When preadipocytes were incubated with simvastatin in vitro, inhibition of leptin gene expression occurred independently of farnesylation (Maeda & Horiuchi, 2009). It has been proposed that ultimately, the decrease in leptin associated with atorvastatin, rosuvastatin, and simvastatin treatment negatively affects the proliferation of β‐cells and insulin secretion, thus leading to the development of diabetes (Chetboun et al., 2012)

Maeda & Horiuchi (2009) suggested that the simvastatin reduced leptin is a result of the inhibition of PI3K leading to elevating in intracellular cAMP, a protein kinase A (PKA) activator. Perhaps the suppression of leptin secretion was responsible for this reduction in activated PKA (Maeda & Horiuchi, 2009).

4.8. Uncoupling protein 3

Uncoupling protein 3 (UCP3) has been presumed to limit the aggregation of nonesterified fatty acids in mitochondria (Schrauwen, Saris, & Hesselink, 2001). Many studies have shown that skeletal muscle insulin sensitivity decreases as a result of high intramyocellular triglyceride content, not only in diabetic and obese subjects but also in nonobese and nondiabetic subjects (Krssak et al., 1999; Pan et al., 1997). Larsen et al. (2013) recently have studied oral glucose tolerance test and UCP3 in simvastatin treated individuals; they found glucose intolerance and decreases in UCP3 occur concurrently in simvastatin‐treated subjects, and hypothesized that statin‐induced insulin resistance is a result of statin‐induced UCP3 reduction.

4.9. Insulin sensitivity OR insulin secretion

It is uncertain as to whether variations in insulin sensitivity or secretion mediate the effects of statins. Bellia et al. (2012) followed T2DM subjects who received rosuvastatin or simvastatin for 6 months and noted ß‐cell degeneration and a decline in glycemic control without any alteration in insulin sensitivity. The authors proposed that the decline in insulin secretion has greater impact as a mechanism for the development of new‐onset diabetes as opposed to insulin sensitivity. Although certain studies support a stronger effect of insulin secretion over insulin sensitivity, it is probable that both are present to a certain degree before the onset of diabetes (Brault et al., 2014a).

4.10. Lower LDL‐C, a significant increase in body weight, and features of insulin resistance

Increased levels of atherogenic lipoproteins containing apolipoprotein B, mainly low‐density lipoproteins, are a major causal factor in coronary heart disease. Lowering the concentration of LDL‐C reduces the risk of cardiovascular morbidity and mortality, with a direct correlation between the degree of absolute LDL‐C lowering and the extent of event reduction. Statins, HMG‐CoA reductase inhibitors, are the preferred option for lowering LDL‐C and reaching LDL‐C goals (Danchin et al., 2018)

Interesting new information shows that a genetic polymorphism induces HMG‐CoA reductase inactivation, leading to a significant decrease in LDL‐C and an increase in body weight, with the development of insulin resistance (Swerdlow et al., 2015). This information was confirmed in the randomized statin trials and demonstrated that one particular allele was related to a significant increase in NODM. Because statins inhibit HMG‐CoA reductase as their mode of action, this may at least partly explain their diabetogenic effect (Ganda, 2016).

4.11. Effects on lipoprotein particle size

Other researchers have suggested that statins could induce NODM via alteration of lipoprotein particle size. Mackey et al. (2015) and Mora et al. (2010) have shown that statins increased the size of very low‐density lipoprotein cholesterol (VLDL‐C) particles and decreased the size of LDL‐C, HDL‐C, and the lipoprotein insulin resistance (LPIR) score, which all contribute to the development of NODM. In contrast, large LDL‐C, HDL‐C, and small VLDL‐C particle size has an inverse effect on NODM. Dugani et al. (2016) found that rosuvastatin reduced the size of LDL‐C and HDL‐C particles, and the LPIR score.

4.12. Low‐density lipoprotein receptor

Statins reduce blood cholesterol in two ways: they can significantly decrease synthesize of LDL‐C in blood by blocking HMG‐CoA reductase; in addition, statins also increase blood LDL‐C removal through upregulation of LDL receptors (LDLR) in the liver and peripheral tissue (Beltowski, Wojcicka, & Jamroz‐Wisniewska, 2009).

Increased hepatic LDLR has been observed in statin‐treated dogs (Alberts et al., 1980). Though statins decrease LDL cholesterol, they also have side effects for pancreatic islets. The molecular mechanisms behind these side effects involve an increase in LDLR‐mediated uptake of LDL‐C in islets via statin‐induced upregulation of the LDLR expression. The most likely mechanism is that the pancreatic islets are damaged from abnormal cholesterol levels via upregulation of the LDLR expression with consequent elevated LDLR‐mediated uptake of LDL‐C. Indeed, many studies have reported that dysfunction of pancreatic beta cells is due to the accumulation of cholesterol in the pancreatic islets (Brunham, Kruit, Verchere, & Hayden, 2008; Fryirs, Barter, & Rye, 2009).

Proprotein convertase subtilisin/kexin type 9 (PCSK9) is a protein, mainly synthesized and secreted by the liver, which binds to specific proteins and escorts them to endosome/lysosome compartments for degradation. The LDLR has been identified as the main target of PCSK9; mice lacking Pcsk9 exhibit an increased hepatic LDLR expression, resulting in an increased clearance of circulating LDLs and hypocholesterolaemia, whereas PCSK9 overexpression induces a two‐fold increase in plasma cholesterol levels. In summary, PCSK9 critically controls the LDLR expression in pancreas, perhaps contributing to the maintenance of a proper physiological balance to limit cholesterol overload in beta cells. Pcsk9 deficiency, in turn, results in increased LDLR expression and cholesterol ester accumulation in pancreatic islets, which impairs insulin secretion (Da Dalt et al., 2018)

4.13. Genetic polymorphisms

Genetic polymorphisms are another mechanism that may cause statin‐induced insulin resistance (Ruscica, Macchi, Morlotti, Sirtori, & Magni, 2014; Sattar & Taskinen, 2012). Stančáková et al. (2009) found that among 5,327 nondiabetic Finnish men, those who had the following SNIPs had impaired early‐phase insulin release or impaired conversion of proinsulin to insulin: TCF7L2, SLC30A8, HHEX, CDKN2B, CDKALI, MTNR1B, KCNJ11, and IGF2BP2. This study suggests that several possible pathophysiologic mechanisms are responsible for the diabetogenic effects of statins, with all of these mechanisms leading to a reduction in insulin release or insulin sensitivity (Baker, Talati, White, & Coleman, 2010; Yokote, Shimano, Urashima, & Teramoto, 2011).

To illustrate the mechanism by which statins induce diabetes, Swerdlow et al. (2015) used the Mendelian randomization principle. They recognized a SNP, rs17238484‐G, in the HMGCR gene and examined its relation to BMI, plasma insulin levels, plasma glucose levels, and the risk of type II diabetes. They found that lower levels of LDL‐C, higher waist circumference, BMI, plasma insulin levels, and plasma glucose levels show a relationship with every additional allele of the SNP. In addition, the Rs17238484‐G allele was linked to an enhanced risk of type II diabetes (OR per allele 1.02; 95% CI, 1.00–1.05), suggesting that the enhanced risk is at least partially described by the decreased activity of HMG‐CoA reductase. Labos, Brophy, Smith, Sniderman, & Thanassoulis (2018) further supported this hypothesis with a follow‐up study that reanalyzed the data from 25 primary and secondary statin randomized control trials that provided data on cardiovascular events and 12 trials that provided data on incident diabetes.

4.14. microRNAs

Small noncoding RNA molecules, consisting of 21–23 nucleotides, are termed microRNAs (miRs or miRNAs) and their function is to regulate gene expression. Recently, miRNAs have been proposed to be involved in the increased incidence of diabetes because of their contribution to the gene expression alterations observed in dysfunctional beta cells and in insulin‐resistant tissues. miRNAs are essential for pancreatic islet development (Guay, Roggli, Nesca, Jacovetti, & Regazzi, 2011). Studies have reported that a panel of miRNAs, particularly miR‐375, miR‐7, miR‐195, miR‐126, miR‐9, miR‐96, and miR‐34a, have a role in pancreatic development and insulin secretion, and that miR‐7, miR‐139, miR‐145, and miR‐1 have a role in the insulin growth factor‐1 receptor expression (Chakraborty, Priya, & Bandyopadhyay, 2013). For example, miR‐96 causes a decrease of Noc2 expression, a RabGTPase effector that is essential for exocytosis of insulin (Lovis, Gattesco, & Regazzi, 2008).

Several miRNAs that are selectively expressed in pancreatic endocrine cell lines have been identified by Poy et al. (2009). miR‐375 plays an important role in the regulation of glucose‐stimulated insulin secretion (Tang, Tang, & Özcan, 2008) through an effect on pancreatic β‐cell development. Plaisance et al. (2006) found that miR‐9 causes suppression of the expression of the Onecut2 transcription factor and thereby increases the levels of granuphilin, leading to inhibition of insulin secretion because granuphilin causes downregulation of insulin secretion.

The role of miR‐124a in the modulation of the expression of secretory pathway proteins involved in the insulin exocytose machinery has also been reported (Baroukh et al., 2007; Lovis et al., 2008).

In tissues targeted by insulin, miRNAs also regulate insulin signaling. IRS1 has been identified as a target of miRNAs, specifically miR‐145 (Shi, Sepp‐Lorenzino, Prisco, Linsley, & Baserga, 2007). Overexpression of miR‐145 leads to a reduction in IRS‐1 protein levels without affecting IRS1 mRNA levels in human colon cancer cells. On the other hand, miR‐143 KO mice are protected from the insulin resistance associated with obesity. One study emphasized the role of miR‐143 in obesity‐related insulin resistance and suggested that the miR‐143‐ORP8 pathway is a potential target for the treatment of T2D (Jordan et al., 2011). In addition to the above‐mentioned miR‐143, other miRNAs have been recognized that are important for control of differentiation of human and mouse adipocytes (Esau et al., 2004).

Other research has demonstrated that miR‐103/107 suppression leads to improvement of glucose homeostasis and insulin sensitivity, although its overexpression impaired glucose homeostasis in both liver and fat (Kumar, Nath, Prasad, Sharma, & Li, 2012).

A balance between the insulin secretion by beta cells and target tissue sensitivity to hormone action is required for blood glucose homeostasis. Previously, the molecular mechanism of insulin resistance was not fully understood, but miRNAs were thought to play a role in it. Abundant quantities of 4 miRNAs, miR‐122, miR‐148a, miR‐192, and miR‐194, are present in the liver, and were downregulated in Dicer‐null hepatocytes, emphasizing their potential role in glucose and lipid metabolism regulation. More recently, alterations in about 60 miRNAs in drug‐free, lean, type 2 diabetic subjects were detected by Gallagher et al. (2010). Interestingly, the onset of diabetes seemed to be induced after the alterations in miRNAs. miR‐133a/b has been shown to reduce insulin‐induced glucose uptake as a result of a decrease in glucose transporter type 4 expression in cardiomyocytes (Horie et al., 2009).

Recent research has shown that miRNAs mediate the positive pleiotropic effects of statins via modulation of lipid metabolism, enhancement of endothelial function, inhibition of inflammation, improvement of plaque stability, and immune regulation. miRNAs are implicated in statin‐related interindividual variations in therapeutic response, directly via HMG‐CoA reductase, or indirectly through targeting cytochrome P450 3A functionality and PCSK9 biology (Mohajeri et al., 2018)

Also, the effect of statins on miRNAs has been demonstrated in many studies. For example, 12 months of atorvastatin therapy led to a reduction in circulating miR‐122 in the Anglo‐Scandinavian Cardiac Outcomes Trial (n = 155). Circulating miR‐122 was associated with insulin resistance, obesity, metabolic syndrome, type‐2 diabetes, and an adverse lipid profile. A similar response to atorvastatin was observed in mice and cultured murine hepatocytes (Willeit et al., 2016).

One of the miRNAs that has a regulatory function in lipid homeostasis is circulating miR‐30c. Research has shown that miR‐30c is transported in both exosomes and on HDL3, and treatment with pravastatin increased circulating miR‐30c expression significantly, further adding to the pleiotropic effects of statins (Sodi, Eastwood, Caslake, Packard, & Denby, 2017). miR 221/222, is closely linked to the proliferation of endothelial cells (EPC), and atorvastatin decreases EPC numbers and decreases the miR 221/222 levels in patients with CAD, contributing to the beneficial effects of atorvastatin in this disorder (Minami et al., 2009). It has been suggested that c‐miR‐499‐5p may serve as a biomarker of statin‐potentiated muscle damage (Min et al., 2015).

He et al. 2007 showed overexpression of miR‐29a/b/c in the murine insulin‐responsive 3T3‐L1 adipocytes caused inhibition of insulin‐stimulated glucose uptake and Akt activation. Another target of miR‐29 is syntaxin‐1. Syntaxin‐1 is necessary for the fusion of GLUT4 vesicle to the plasma membrane (He et al., 2007). Esau et al. (2004) showed that miR‐143 was upregulated during adipocyte differentiation while inhibition of miR‐143 led to the overexpression of insulin‐sensitive GLUT4 (Esau et al., 2004). In addition, transgenic overexpression of miR‐143 has been reported to reduce insulin‐stimulated Akt activation and affect glucose homeostasis (Kumar et al., 2012; Figure 1, 2, 3).

Figure 1.

The mevalonate pathway. HMGCR catalyzes the conversion of HMG‐CoA to mevalonic acid, a necessary step in the biosynthesis of cholesterol. Statins act by inhibiting HMG‐CoA reductase, the key enzyme of the mevalonate pathway. Mevalonate pathway inhibition by statins causes reductions in other downstream products. The major components which may be involved in new‐onset diabetes are GGPP, FPP, dolichol and coenzyme Q10. FPP: farnesyl pyrophosphate; GGPP: geranylgeranyl pyrophosphate; HMG‐CoA: 3‐hydroxy‐3‐methyl‐glutaryl coenzyme‐A; HMGCR: 3‐hydroxyl‐3‐methylglutaryl‐CoA reductase receptor

Figure 2.

Intracellular actions of statins. Cav1: caveolin‐1; GLUT4: glucose transporter 4; GTP: guanosine triphosphate; IGF: insulin‐like growth factor; IR: insulin receptor; IRS‐1: Insulin receptor substrate; miR: microRNA; PI3K: phosphatidylinositol 3‐kinase; S: statins [Color figure can be viewed at wileyonlinelibrary.com]

Figure 3.

Mechanisms for statin‐induced new‐onset diabetes mellitus [Color figure can be viewed at wileyonlinelibrary.com]

Induction of miRNA‐33a expression by simvastatin and atorvastatin in the liver has been demonstrated by Allen and colleagues, suggesting a possible link between insulin secretion reduction, which ultimately leads to the development of new‐onset diabetes, and statin induce expression of miRNA‐33a (Allen et al., 2012).

As mentioned above, a vast number of miRNAs have been determined to be involved in glucose homeostasis through various mechanisms, and many studies have also demonstrated the effect of statins on miRNA. Accordingly, some miRNAs may potentially be involved in statin‐induced diabetogenic effects (Osipova et al., 2014). However, additional studies are needed to further confirm this causal relationship between statin treatment and miRNAs in new‐onset diabetes (Osipova et al., 2014).

5. CONCLUSION

We here report the most current studies describing possible mechanisms for statin‐induced new‐onset diabetes. In several meta‐analyses and clinical trials, some combination of these effects most probably contributes to the recognized diabetogenic effects of statins (Table 1).

Table 1.

Possible pathophysiologic mechanisms for statin‐induced new‐onset diabetes mellitus

HMG CoA reductase receptor inhibition

Dysfunction or changes in calcium channel in pancreatic ß‐cell ► decreased insulin secretion

GLUT4 translocation to cell membrane ► decreased peripheral insulin‐mediated glucose influx

Decreased caveolin‐1, decreased glucose uptake and increased insulin resistance in adipose tissue as well as in muscle and liver

Decreased insulin signaling ✓ Decreased phosphorylation of IRS‐1and Akt, decreased insulin signaling ✓ Decreased small G receptors, decreased phosphorylation of insulin receptor substrate, decreased insulin signaling

Modulation of β‐Cell inflammation, oxidation, and apoptosis, increased insulin resistance

Decreased downstream metabolites: ✓ Decreased dolichol, impaired IGF, decreased mitochondrial function, decreased insulin secretion ✓ Decreased Coenzyme Q10, impaired IGF, decreased mitochondrial function, decreased insulin secretion, and decreased GLUT4 translocation

Decreased adipocyte maturation/differentiation: Decreased expression PPAR and CCAAT/enhancer‐binding protein, impair maturation of adipocyte Preadipocytes do not secrete an insulin‐sensitizing hormone Mature adipocytes have more GLUT4 translocation to the cell membrane Mature adipocytes secret adiponectin ► insulin resistance Decreased leptin secretion ► negatively regulate ß‐cell mass ► decreased insulin secretion

Decreased uncoupling protein 3 ► aggregations of nonesterified fatty acids in mitochondria ► decreased skeletal muscle insulin sensitivity

Lower LDL‐C, a significant increase in body weight, and features of insulin resistance

Increased the size of VLDL‐C particles and decreased the size of LDL‐C, HDL‐C, and LPIR ► score increased incidence of NODM

Increased LDLR‐mediated uptake of LDL‐C in islets via statins‐induced upregulation of the LDLR expression, dysfunction of pancreatic b cell

Some loss‐of‐function genetic polymorphisms associated with weight gain, increased glucose, decrease in insulin sensitivity or insulin release ► diabetes ✓ Nucleotide polymorphisms in the HMGCR receptor gene ► inhibition of the HMGCR

Change in microRNAs expression, change in posttranscriptional gene expression, and reduce insulin secretion and sensitivity

There is valid clinical data that patients with diabetes mellitus greatly benefit from statin therapy as it reduces the incidence of cardiac events (Brault et al., 2014a), cardiac remodeling, ER stress, nerve cell injury, toxic effects of glucose and ox‐LDL, cardiac and neuronal ischemic/reperfusion injury, MI adverse effects, abdominal aortic aneurysm progression, and anti‐atherosclerotic activity (Mollazadeh et al., 2018). Thus the present data certainly do not recommend stopping statins in subjects with recently diagnosed diabetes mellitus (Brault et al., 2014a).

Because the incidence and mechanisms underlying statin‐induced new‐onset diabetes are still being characterized, and because of the significant benefit of statins in decreasing cardiovascular mortality, statins remain the first choice treatment for prevention of CVD despite the presence of several novel lipid‐lowering agents (Sahebkar & Watts, 2013). However, the risk of new‐onset diabetes has been shown to differ depending upon the particular choice and dose of statin used and, at an individual level, on the patient's risk profile. Mach et al. (2018) demonstrated that statin use is not associated with adverse effects on cognitive function or with clinically significant deterioration of renal function and does not increase the risk of cataract or hemorrhagic stroke in individuals without prior stroke. They concluded that statin treatment is remarkably safe, though there is a modest risk of new‐onset DM with long‐term statin treatment. For instance, individuals who have several risk factors for diabetes could be prescribed a less diabetogenic statin, such as pravastatin. Knowledge of the underlying mechanisms related to the development of diabetes in association with statins is essential and clinically relevant and may provoke novel preventative or therapeutic approaches to this problem and/or guide the production of new generation statins without such side effects (Brault et al., 2014a).

CONFLICTS OF INTEREST

The authors declare that they have no conflicts of interest.