The development of T2DM is the result of interaction between environmental factors (e.g. unhealthy diet, sedentary lifestyle, stress) and a strong hereditary component (43). Currently, several hundreds of genetic variants were associated with T2DM, although mostly with only minor effects on disease development (44).

Numerous studies suggested that m⁶A and m⁶Am demethylase FTO is among the genes whose variants possess the highest genetic risk of T2DM (44). However, this link is still controversial with significant interethnic differences (45, 46). For instance, the common FTO rs9939609 variant was associated with T2DM in white American, Palestinian, Asian Indian, and obese Iraqi populations, but not in Bengalee Hindu, North Indian, nor Saudi populations (47–57). Also, other genetic polymorphisms in the FTO gene were identified as T2DM risk factors. Carriers of the FTO rs17817449 variant in the Czech-Slavonic and obese Iraqi populations were more susceptible to T2DM and chronic diabetic complications (44, 51, 58). In Iranian obese women, FTO variants rs763967273, rs759031579, rs141115189, rs9926289, rs76804286, and rs9939609 were all related to T2DM (59). On the contrary, African-Americans carrying the rs1421085 C allele were found to be protected against diabetes (54). The polymorphisms in FTO gene seem to regulate the expression level of FTO and its enzymatic function. Detrimental effects of high or low expression of FTO were already confirmed in experimental studies. For instance, it has been shown that FTO depletion activates inflammatory response, one of the main pathogenic features in T2DM patients (60).

Besides FTO, variants of IGF2BP2, an m⁶A reader, were also associated with a significant risk of T2DM development, namely variant rs4402960 in Asian Indian Sikhs, Czechs, or Italians, and rs11705701 in the Chinese population (44, 55, 61, 62).

Although further studies are needed to unravel the complex polygenic background of T2DM, it seems to be clear that genetic polymorphisms in genes encoding epitranscriptomic regulators are associated both with T2DM and its complications.

Pancreatic β-cell failure mediated by metabolic stress is the central event in the pathogenesis of T2DM (63). Although the mechanisms underlying β-cell dysfunction are still not fully understood, emerging data suggest an involvement of epigenetic modifications in the adaptation of β-cells to metabolic stress (64).

m⁶A sequencing in dispersed islets from controls and T2DM patients revealed 6,078 differently methylated sites in 4,155 mRNAs and a higher number of sites with decreased levels of m⁶A methylation in T2DM compared to controls. Gene ontology analysis of the m⁶A methylome revealed that the genes affected in T2DM patients are involved in cell-cycle regulation, receptor signaling, insulin secretion, and pancreas development (65). The decreased total m⁶A levels were observed in Langerhans islets of T2DM patients and also in islets of mice fed with a high-fat diet (a model mimicking T2DM phenotype). Similarly, high glucose conditions (state typical for T2DM) also resulted in lower methylation levels in non-diabetic human pancreatic islets as well as in mouse β-cell line (Min6) (66). Gene expression analysis in whole islets collected from healthy humans and patients with T2DM revealed a down-regulation of several m⁶A regulators in diabetic individuals – methyltransferase METTL14, demethylases FTO and ALKBH5, and readers YTHDF1 and YTHDF3. In addition to transcripts, protein levels of methyltransferases METTL3 and METTL14 were also decreased (65). The reduction of FTO gene expression and METTL3/14 protein levels in T2DM human islets was observed also in other studies (67–69). RNA-seq datasets (GSE153855; GSE153855) from T2DM and non-T2DM individuals revealed increased gene expression of readers IGF2BP2-3 and decreased gene expression of writer WTAP and readers YTHDF2-3, YTHDC1, and HNRNPC (70–72). m⁶A reader IGF2BP2 was also up-regulated in β-cells obtained from cadaver pancreases of T2DM patients (73). The current knowledge of diabetic epitranscriptomic changes in human Langerhans islets is summarized in Figure 2 . Overall, it seems that the whole epitranscriptomic machinery is attenuated in human diabetic islets. The only up-regulated genes IGF2BP2-3 have also functions unrelated to epitranscriptomics, which might explain their opposite trend.

In contrast to these results, Bornaque et al. (66) showed that high glucose concentrations in Min6 cells increased mRNA expression of important m⁶A regulators – methyltransferase Mettl3 and demethylases Fto and Alkbh5. Glucose treatment also induced a shift in the subcellular protein localization of METTL3 and ALKBH5 (66). Overexpression of FTO in Min6 cells promoted the production of reactive oxygen species (ROS) and led to NF-κB activation, which resulted in the inhibition of insulin secretion (74). These differences between a specific mouse cell line and heterogeneous human islets might be explained by interspecies variation or islet heterogeneity.

MTC specifically regulates the postnatal functional maturation of β-cells. Mice with deletion of Mettl3/14 in Ngn3⁺ endocrine progenitor cells developed hyperglycemia and hypoinsulinemia 2 weeks after birth. This study also showed that Mettl3/14 deletion silenced the expression of important transcription factors, such as Mafa, Nkx6-1, or Pdx1 (69). Other studies using mouse models with β-cell-specific deletions of MTC subunits (Mettl3, Mettl14, Wtap) also pointed out the importance of MTC in maintaining β-cell function. Deletion of either subunit resulted in decreased m⁶A levels (65, 70, 75). METTL3 deficiency led to β-cell failure and hyperglycemia (75). METTL14-deficient mice exhibited decreased β-cell mass, reduced insulin secretion, and glucose intolerance (65, 76, 77). Deficiency of WTAP was associated with a reduction of METTL3 levels and resulted in severe hyperglycemia and β-cell failure. Overexpression of Mettl3 in β-cells partially prevented the negative effects of WTAP deficiency (70). Comparing Mettl3-βKO and Wtap-βKO mice revealed down-regulation of β-cell-specific transcription factors (such as Mafa, Nkx6-1, Pdx1, Neurod1, or Foxa2) and insulin secretion-related genes (such as Ins1, Ins2, Brsk2, Cacna1c, Doc2b, Ffar1, G6pc2, Gck, Gipr, Hadh, Ica1, Nnat, Park7, Pclo, Selenot, Serp1, Slc30a8, Stxbp51, Sytl4, Trpm2, Ucn3, and Uqcc2) (70). Besides methyltransferases, also β-cell-specific deletion of reader Ythdc1 resulted in β-cell failure and diabetes (71, 78). This was likely due to the decreased gene expression of β-cell-specific transcription factors (such as Mafa, Nkx6-1, Neurod1, and Hmgn3) and insulin-related genes (such as Ins1, Ins2, Gck, G6pc2, Sytl4, Doc2b, Pclo, Cacna1c, Slc30a8, Ffar1, Gipr, Nnat, and Selenot). Transcription factor MAFA decreased dramatically also on protein level in Ythdc1-βKO islets (71). Yang et al. suggested that YTHDC1 may regulate mRNA splicing and export to modulate glucose metabolism in β-cells by interacting with serine/arginine-rich splicing factor 3 (SRSF3) and cleavage and polyadenylation specific factor 6 (CPSF6) (78).

These data indicate that m⁶A/m⁶Am epitranscriptomic machinery vastly affects the biology of pancreatic β-cells and plays a role in the induction of diabetic phenotype. However, the data are still fragmental, and more studies covering more m⁶A/m⁶Am regulators are needed to elucidate the exact role of epitranscriptomic regulations in the diabetic pancreas.

Cardiovascular disease (CVD) is a common comorbidity and a major cause of mortality among people with T2DM. More than 30% of all T2DM patients are affected by CVD (79). Cardiac dysfunction observed in patients with diabetes that occurs in the absence of other cardiovascular risk factors (such as hypertension, coronary artery disease, or valvular disease) is referred to as diabetic cardiomyopathy (DCM) (80). This condition is characterized by cardiac diastolic dysfunction and later by heart failure (HF) and cardiac death. It is estimated that the risk of HF is 2-3 times higher in individuals with T2DM and that approximately 12% of diabetic patients eventually develop severe HF often leading to death (81). The epitranscriptomic modifications, including m⁶A, are known to play various roles in the physiology and pathophysiology of the cardiovascular system (20, 82–85). Recent studies have shown that changes in m⁶A methylation also contribute to HF progression (86–90). However, the role of cardiac m⁶A and m⁶Am machinery is not well-characterized in T2DM.

Altered cardiac m⁶A patterns were detected in db/db mice (model of T2DM and DCM). The differentially methylated transcripts were linked mainly to cardiac fibrosis, myocardial hypertrophy, and myocardial energy metabolism (91). The higher total m⁶A mass in DCM was associated with the down-regulation of demethylase FTO on both gene and protein levels, while levels of METTL3, METTL14, and ALKBH5 were stable (91). Interestingly, mice with T1DM-induced DCM (C57BL/6 mice injected with streptozotocin) exhibited a different dysregulation of epitranscriptomic machinery ( Figure 3 ). Total m⁶A levels in the hearts of these mice were decreased. This was linked with an increase of ALKBH5 in the cardiomyocytes of DCM mice and subsequent activation of the Hippo signaling pathway through a YTHDF2-dependent action (92). These results suggest that the two types of diabetes might affect the epitranscriptomic background of DCM differently. It has been reported already that T1DM and T2DM might affect the heart in a different way and result in dissimilar DCM phenotype. This was explained mainly by the different myocardial insulin action (insulin deficiency in T1DM vs insulin resistance and hyperinsulinemia in T2DM) and thus distinct signaling downstream of the insulin receptor (93). Therefore, the contradictory epitranscriptomic results may be explained by the different phenotype between the two types of diabetes. However, further research is needed to resolve this issue.

Most of the studies dealing with m⁶A/m⁶Am regulations in DCM have been executed on T1DM animal models. Pyroptosis, a type of proinflammatory cell death, is tightly involved in DCM progression. Methyltransferase METTL14 was down-regulated in the hearts of rats with DCM (T1DM-induced) and enhancement of its expression inhibited pyroptosis in myocardial tissues and improved systolic function (increased fractional shortening and ejection fraction) via down-regulation of lncRNA Tincr. The expression of Tincr was regulated in a YTHDF2-dependent manner (94). Peng et al. (95) reported that lncRNA Airn ameliorated diabetes-induced (T1DM) cardiac dysfunction caused by cardiac fibrosis. Their data showed that Airn binds to m⁶A reader IGF2BP2 and protects it from ubiquitin-proteasome-dependent degradation, leading to an m⁶A-dependent stabilization of p53 mRNA by IGF2BP2 and subsequent reduction in cardiac fibrosis (95).

Despite the limited amount of data available, it is becoming evident that epitranscriptomic dysregulations in diabetic cardiac tissue might have a significant effect on the function of the heart. However, the exact role of m⁶A and m⁶Am in DCM induced by each type of T2DM is yet to be deciphered.

Diabetic nephropathy (DN), also known as diabetic kidney disease, is a prevalent microvascular complication of T2DM often leading to end-stage renal disease, a life-threatening condition (96). According to the International Diabetes Federation reports, up to 40% of diabetic patients might develop DN (97).

Xu et al. reported, that human kidney 2 (HK-2) cells stimulated with high glucose decreased total m⁶A methylation level and also methyltransferases METTL3 and METTL14 (98). Interestingly, Jiang et al. observed increased m⁶A modification in diabetic mice which was caused by elevated levels of METTL3. They also found increased METTL3 levels in renal biopsies from DN patients. Further experiments showed that METTL3 exerted pro-inflammatory and pro-apoptotic effects in an IGF2BP2-dependent manner and that targeting METTL3 alleviated the DN injury (99). A negative effect of METTL3 in DN was reported also by Tang et al. (100). METTL14 was also highly expressed in the kidneys of DN patients and HRGEC (high glucose-induced human renal glomerular endothelial cells). METTL14 worsened renal injury and inflammation was reported in db/db mice (101). Lu et al. also reported high levels of METTL14 in renal biopsy samples from patients with glomerulosclerosis and DN. Mice with podocyte-specific METTL14 deletion were then associated with improved glomerular function and alleviated podocyte injury compared to wild-type nephropathic mice (102). Also the third component of the MTC – WTAP – was reported to induce pyroptosis and inflammation in high glucose-treated HK-2 cells (103). Besides the methyltransferases, FTO was described to promote the progression of DN (104). However, several SNPs in the FTO gene were associated with a significantly lower risk of nephropathy in T2DM patients (62). Urine levels of m⁶A were decreased in patients with T2DM and even more with DN (105).

The existing data indicate that m⁶A machinery is affected in DN and that its dysregulation has a negative outcome on the progression of the pathology.

Liver disease ranks among notable causes of death in T2DM patients (106). Non-alcoholic fatty liver disease (NAFLD) is the most common chronic liver disease and is strongly associated with T2DM (107–109). The prevalence of this comorbidity among T2DM patients reaches up to 70% (110). It has been described that NAFLD is promoted by m⁶A modification dysregulation (111–116). Moreover, liver tissues from T2DM patients and mice on HFD showed elevated levels of m⁶A and also METTL3. Hepatocyte-specific knockout of Mettl3 in mice then led to improved insulin sensitivity and decreased fatty acid synthesis (117). Jiang et al. also reported that baicalin – a flavonoid glycoside used in traditional Chinese medicine – suppressed T2DM-induced liver tumor progression in a METT3/m⁶A-dependent manner (118).

Chronic exposure to hyperglycemia affects the microvasculature, eventually leading to diabetic retinopathy (DR), the main cause of blindness in the developed world. It has been described that m⁶A modification is regulated by various risk factors associated with DR, such as inflammation, oxidative stress, angiogenesis, or glucose and lipid metabolism (119). FTO polymorphism (rs8050136) was associated with a higher risk of DR (120). In retinal pigment epithelium (RPE) cells, high-glucose conditions down-regulated the expression of METTL3 on both transcript and protein levels. Further experiments showed that METTL3 overexpression alleviated the cytotoxic effects of high-glucose on RPE cells, while METTL3 depletion had the opposite effect (121). Conversely, diabetic stress-induced up-regulation of METTL3 and subsequent increase of m⁶A levels in human retinal pericytes and also mouse retinas. Specific depletion of METTL3 in pericytes suppressed diabetes-induced pericyte dysfunction and vascular complication in vivo (122). A recent study showed down-regulation of METTL3 in vitreous humor samples from patients with DR, a mouse model of DR, and also high glucose-induced human retinal microvascular endothelial cells (123).

Despite these conflicting data on METTL3 expression, it seems to be clear that epitranscriptomic regulations are affected in DR, but the exact role of m⁶A in the pathogenesis remains to be elucidated in the future.

Dysregulation of autophagy is a contributing factor for delayed wound healing in diabetic skin. YTHDC1, an m⁶A reader, has been described as a modulator of autophagy in diabetic keratinocytes which regulates the mRNA stability of an autophagy receptor (124). Interestingly, YTHDC1 interacted and cooperated with ELAVL1 (ELAV-like RNA binding protein 1), a well-established RNA stabilizer also linked to m⁶A methylation. It has been described previously that loss of m⁶A methylation enhances ELAVL1 RNA binding to increase RNA stability (125).

Decreased m⁶A methylation levels were detected in RNA isolated from the peripheral blood of T2DM patients and also diabetic rats (126, 127). In accordance with these results, significantly higher gene expression of FTO (and not ALKBH5) in peripheral blood from T2DM patients was detected (126). However, Onalan et al. (127) observed an up-regulated expression of both demethylases in venous blood samples from T2DM patients. The increased expression of FTO on both gene and protein levels was later confirmed by another study which pointed out the correlation between high FTO levels and T2DM severity (128). The gene expression of FTO was also up-regulated in white blood cells from T2DM patients compared to healthy individuals and the expression level of FTO was positively correlated with fasting glucose concentration (129). Besides erasers, METTL3 mRNA was down-regulated in serum samples from T2DM patients (121). Progressively higher T2DM risk was associated with low serum IGF2BP3 levels (72).

Taken together, the content of m⁶A or its regulators in the peripheral blood may serve as novel potential biomarkers of T2DM in the future (126).

Metformin is the first-line therapy for the treatment of T2DM, yet its molecular mechanisms of action are not fully understood (130, 131). The main effect of metformin treatment is inhibition of hepatic gluconeogenesis. At the molecular level, several mechanisms have been proposed to explain this phenomenon, such as inhibition of mitochondrial complex I activity, activation of AMPK, or increase in hepatocellular redox state due to inhibition of GPD2 (glycerol-3-phosphate dehydrogenase 2). The secondary effects of metformin treatment include an increase in muscle glucose uptake, a decrease in intestinal glucose absorption, and a change in the composition of the gut microbiome (130).

According to recent studies, metformin also affects epitranscriptomic regulations, including m⁶A machinery. Metformin was shown to reduce m⁶A methylation via the down-regulation of methyltransferase METTL3 in breast cancer cells (132). In hepatocellular carcinoma, metformin treatment was associated with METTL3 inhibition (133). Metformin also attenuated multiple myeloma cell proliferation and encouraged apoptosis by suppressing METTL3-mediated m⁶A methylation of its targets (134). Surprisingly, METTL3 expression was up-regulated after metformin treatment in adenocarcinoma cells (135). YTHDC2, a key m⁶A reader, is an important target of metformin in preventing the progression of vascular smooth muscle cell (VSMC) dysfunction under high glucose, a simulation of VSMC dysfunction caused by T2DM (136). Recently, Liao et al. (137) showed that metformin combats obesity by targeting FTO in an m⁶A-YTHDF2-dependent manner. This study suggests that metformin inhibited the protein expression of FTO, resulting in higher m⁶A methylation in mRNAs of crucial cell cycle regulators. The binding of YTHDF2 to modified transcripts then triggered mRNA decay and subsequent decrease of protein expression. In consequence, the mitotic clonal expansion process was blocked and adipogenesis was inhibited.

This fragmentary information suggests that metformin may both decrease and increase m⁶A methylation and that the target tissue or cell type may be the determining factor. However, in vivo studies focusing on the epitranscriptomic effect of metformin are needed to decipher this phenomenon, as the in vivo and in vitro response may also differ, especially if the primary target of metformin treatment is the liver. Despite these ambiguities, the association between epitranscriptomics and metformin is revealing itself, however, the role of m⁶A modification in the treatment of diabetes remains unclear.