Signaling pathways and intervention for therapy of type 2 diabetes mellitus

Type 2 diabetes mellitus (T2DM), a chronic noncommunicable disease that is diagnosed by aberrant high blood glucose levels, has attracted increasing attention due to its high prevalence and enormous health burden. Currently, there are more than 537 million patients with diabetes, most of which are T2DM, and this number is projected to reach at 783 million by 2045, 1 accompanied by a younger trend in the onset age around the world. 2 , 3 Beyond a straightforward rise in blood glucose, T2DM may cause a series of complications that are linked to the vascular and neural damages triggered by hyperglycemia, such as diabetic nephropathy (DN), diabetic retinopathy (DR), diabetic neuropathy, and cardiovascular disease (CVD). 3 Of note, T2DM is emerging as an increased risk of severe COVID‐19, the recent novel coronavirus pandemic worldwide. For a long duration, T2DM and its complications have been witnessed to impose substantial effects in the quality of life and socioeconomic burden, 4 which inspires the incredible progress in mechanism exploration and drug intervention for T2DM. However, although current antidiabetic drugs, insulin therapy, and lifestyle interventions, for example, metformin administration, carbohydrate restriction, and/or endurance exercise, have warranted decent control of T2DM progression, implementing and maintaining these changes for prolonged periods are still challenging, especially given the pervasiveness of drug side‐effects and the accessibility of calorically dense foods and sedentary lifestyle. To further develop new pharmacological strategies that could potently target pathological mechanisms and avoid side‐effects, it is still urgent and important to unceasingly disclose the mechanistic underpinnings of T2DM. Numerous signaling pathways play essential roles in the development of T2DM and are implicated in its therapy. From a pathological view, insulin resistance (IR) and subsequent insulin deficiency due to pancreatic β cell damage are two main pathological features of T2DM, and their variable combination further contributes to the complexity of T2DM and the diversity in the patients' conditions. 1 In addition, other pathological processes, including chronic inflammation, incretin dysregulation, hyperglucagonemia, lipolysis, central appetite dysregulation, abnormal gastric emptying, gut dysbiosis, and islet amyloid polypeptide (IAPP) deposition, are also regarded as key regulators in the pathophysiology of T2DM. 1 The endeavors throughout the past decades have uncovered that a series of signaling pathways play important roles in controlling these pathological changes. For example, phosphoinositide 3‐kinase (PI3K)/protein kinase B (PKB, also known as AKT) signaling cascade regulates insulin response, 5 and AMP‐activated protein kinase (AMPK) pathway prevents β cell dysfunction. 6 From a therapeutic view, most of current glucose‐lowering drugs have been found to exert their pharmaceutical effects dependently of signaling pathways, more or less. For instance, glucagon‐like peptide‐1 (GLP‐1) receptor agonists bind to their receptors on β cells and promote insulin exocytosis by increasing intracellular Ca 2+ levels via the protein kinase C (PKC)/cyclic adenosine monophosphate (cAMP) signaling pathway. 7 , 8 Recently, four categories of potential hypoglycemic drugs with new mechanisms of action have been proposed, which may stimulate insulin secretion, utilize the incretin axis, maintain hepatic glucose homeostasis, and improve insulin sensitivity, repsectively. 9 Interestingly, these potential drugs target a number of key receptors, vital enzymes, or ion channels, such as glucokinase (GK) activators, G‐protein‐coupled receptor 40 (GPCR40), GLP‐1 receptor (GLP‐1R), and so on, which are involved in numerous signaling pathways associated with T2DM. 9 Therefore, a better understanding of the signaling pathways is of importance for in‐depth dissecting the pathological mechanisms for T2DM and facilitating the development of targeted drugs and interventions against T2DM and its complications. In this review, focusing on the two most important pathological features of T2DM, that is, IR and impaired β cells, align with simplified mention of other pathological changes, we elaborated the roles of signaling pathways in pathological changes of T2DM and its complications on the one hand, and expatiated the intervention mechanisms of important glucose‐lowing drugs on these signaling pathways and emphasized the important targets on the other hand.

T2DM is conventionally featured with two pathological traits: IR and subsequent β cell dysfunction, which are the consequences of the feedback loops between disordered insulin secretion and insulin action (Figure 2 ). FIGURE 2 Pathological features of T2DM. T2DM is characterized by insulin resistance and β cell dysregulation, along with various pathological disturbances, which are coordinately regulated by multiple tissues and organs, including the pancreas, liver, adipose, muscle, gut, and brain. As the major characteristic of T2DM, insulin resistance occurs when the tissues become less sensitive to insulin, leading to increased glucose production and lipogenesis in the liver, enhanced lipolysis and reduced triglyceride synthesis in the white adipose tissue (WAT), and decreased glucose uptake and elevated fatty acid oxidation in the skeletal muscle, and so on. Failure of pancreatic β cells that are responsible for the production and secretion of insulin results in decreased cell number and impair insulin secretion. The vicious cycle of insulin resistance, hyperglycemia, inflammation, and so on, aggravate the onset, development, and progression of T2DM. The online resource inside this figure was quoted or modified from Scienceslide2016 plug‐in. IR, defined as the impairment of insulin sensitivity, is a predictor of T2DM and describes the failure of target organs/tissues to answer insulin stimulation. 20 Previous studies have reported that numerous factors, including obesity, overnutrition, physical inactivity, gastrointestinal microbial disturbance, and family history of T2DM could drive IR. 21 , 22 Insulin is the only one hormone that actively lowers blood glucose by acting on its target organs/tissues, including the hypothalamus, liver, adipose tissue, and skeletal muscle. The hypothalamus is regarded as the major regulator of appetite, and the signaling pathways triggered by insulin, as well as leptin, play key roles in maintaining energy expenditure, glucose homeostasis and insulin sensitivity in peripheral tissues. 23 Insulin stimulation can inspire divergent physiological responses in the peripheral tissues, including increased glycogenesis and de novo lipogenesis (DNL) but decreased gluconeogenesis in the liver, 24 enhanced glucose uptake in the skeletal muscle and adipose tissue, suppressed lipolysis in the adipose tissue, and so on. Dysregulation of these responses, at least a part of them, is therefore to be the major consequence of IR. Additional pathological changes may be also intimately associated with IR. For example, compositional and functional alterations in gut microbiota have been observed in patients with IR and metabolic dysfunction, 25 which may modulate cellular metabolism and energy homeostasis through homeostatic and pathogenic microbiota‐host interactions. In addition, a decrease in circulating adiponectin, an insulin‐sensitizing, anti‐inflammatory, antiatherosclerotic, and hepato‐protective factor predominantly produced by adipocytes, 26 has been shown to extensively promote hepatic and muscular IR. 27 , 28 , 29 , 30 Pancreatic β cells function as the hub for insulin secretion (Figure 3 ), therefore declining β cell function due to dysregulated genetic and external factors is key to T2DM progression. 31 β Cell dysfunction is clearly presented in the patients with hyperglycemia, but whether this trait occurs early or late in the T2DM remains controversial. The current conclusion is more inclined to that β cell function may be weakened early in T2DM progression and gradually decline as glucose tolerance deteriorates. 32 Under normal conditions, β cells are inactive in secreting insulin during fasting period, and postprandial transient hyperglycemia can stimulate β cells to enhance glucose‐stimulated insulin secretion (GSIS), thereby increasing the blood insulin to meet the demands for lowering blood glucose. Generally, β cells are capable of producing sufficient insulin to compensate for IR and maintaining euglycemia. However, chronic exposure to excess circulating nutrients, along with the consequent changed epigenetic factors, could induce a toxic state in the pancreatic islet, resulting in β cell dysfunction and compensatory failure followed by insulin deficiency, eventually causing hyperglycemia and T2DM. 33 FIGURE 3 Multiple signaling pathways regulate β cell function. Elevated glucose enters the β cell via GLUT1/2 transporters to produce ATP through glycolysis in the cytoplasm and oxidative metabolism in the mitochondria, resulting in an increase in the ratio of ATP to ADP. Elevated cytoplasmic ATP leads to the close of ATP‐sensitive potassium channel, membrane depolarization, and subsequent Ca 2+ influx, triggering the release of insulin from insulin granules. Meanwhile, multiple signaling pathways participate in mediating β cell function and insulin secretion, including AMPK, MAPK, WNT, PI3K/AKT, TGFβ, YAP/TAZ, and so on. Substantial evidence supports that metabolic stress

The AMPK pathway is famous for its critical role in sensing cellular energy status (Figure 6 ). 100 AMPK can be canonically activated by the increasing AMP and/or ADP along with declining ATP through the upstream kinase liver kinase B1 (LKB1)‐dependent Thr172 phosphorylation, or noncanonically activated by other stimulations through several recently described pathways that are independent of AMP/ADP, including those related to the lysosome, mitochondrion, Ca 2+ /calmodulin‐dependent protein kinase kinase 2 (CaMKK2) and TGFβ‐activated kinase 1 (TAK1). 101 , 102 , 103 Owing to these mechanisms, AMPK can sense the availability of glucose, glycogen, FAs, Ca 2+ , leptin, and adiponectin, and the damage to lysosomes and nuclear DNA, as well as the stimulation of multiple drugs. 15 , 102 Basically, activated AMPK pathway is able to endorse ATP‐producing catabolic pathways via phosphorylating and activating certain proteins, while curb energy consumption via phosphorylating and inactivating proteins involved in anabolic (biosynthetic) pathways, thereby balancing cellular metabolism and functions. 102 , 104 In this regard, the downstream events of the AMPK pathway encompass: carbohydrate or glucose metabolism, FA and cholesterol metabolism, protein synthesis, the counteracting effects on mTORCs, mitochondrial biogenesis, mitophagy, and autophagy. 100 , 105 Such potent effects of the AMPK pathway in cellular metabolism and functions are prone to endow it with an incredible significance in the peripheral insulin action, glucose uptake, nutrient intake, lipid metabolism, inflammation, insulin secretion, and thus systematic homeostasis of glucose and lipids, warranting the vigorous attention of its therapeutic potential in the area of T2DM. 104 FIGURE 6 The AMPK pathway regulates glycolipid metabolism in T2DM. Under conditions of energy requirement, such as starvation, exercise and OS, the decrease of ATP/AMP ratio activates AMPK directly. Meanwhile, low glucose decreases FBP level, which is sensed by aldolase. Reduced FBP only binds to a small number of aldolases, the remaining which occupy TRPV. Aldolase strongly interacts with TRPV channels and blocks Ca 2+ release, suppressing v‐ATPase activity and stimulating AMPK activity. Metformin interacts with PEN2 and inhibits v‐ATPase activity, thereby activating AMPK. Besides, obesity‐induced hyperinsulinemia activates protein kinases to suppress AMPK activity. Upon activation, AMPK acts on multiple downstream targets to increase ATP generation and decrease ATP consumption by inhibiting anabolic processes and increasing catabolic processes, subsequently regulating a series of downstream targets to affect glucose and lipid metabolism. Degrading of AMPK activity is widely observed in the skeletal muscle from mice and humans with obesity and T2DM, probably due to that high glucose could inhibit the AMPK activity by breaking the active LKB1 complex, 106 activating the protein phosphatase PP2A, 107 upregulating the phosphatase PH domain and leucine rich repeat protein phosphatase 2 (PHLPP2), 108 and inducing ubiquitination degradation of AMPK subunits. 109 , 110 AMPK activation has been found to exert constructive effects on the glucose uptake of skeletal muscle, and regarding this, the deficiency of muscular AMPK activity might compromise the whole‐body glucose homeostasis during T2DM. In detail, the activation of AMPK increases glucose uptake through the mechanisms by which involve the inhibited sequestering of GLUT4 to Golgi apparatus, the enhancement of GLUT4 translocation to plasma membrane, and the upregulation of GLUT4 expression. 103 In the cells that hinge on GLUT1, AMPK also boosts their glucose uptake through promoting the translocation of GLUT1 to the plasma membrane and increasing GLUT1 expression. 111 Since glucose uptake is a key step in the downstream of insulin action, such physiological supportive effect of AMPK in glucose uptake might bypass the PI3K/AKT pathway to link AMPK activation with IR inhibition. 112 In fact, muscle contraction, hypoxia and adiponectin stimulation have been found to initiate GLUT4 translocation via the activation of AMPK in the skeletal muscle, which might counteract muscular IR. 20 Supporting this, multiple drugs or natural products that activate AMPK can improve IR and glucose homeostasis in the mice and patients with obesity or T2DM, including metformin, 15 O304, 113 MK‐8722, 114 Canagliflozin, 115 and PF‐739. 116 Interestingly, although AMPK does not directly regulate the insulin/AKT pathway, the ATK pathway has instead been identified to inhibit AMPK activation, 117 which may complicate the relationship between the AMPK pathway and insulin action. Collectively, identification of accurate alterations, detailed mechanistic cues, and pharmaceutical activators regarding the skeletal muscle‐specific AMPK pathway might confer a basis for the advancement of T2DM mechanistic investigation and drugs intervention. In recent perspectives, h

The WNT pathways are classified into two major groups, β‐catenin‐dependent (canonical) or β‐catenin‐independent (noncanonical) (Figure 7C ). 179 The nuclear translocation of β‐catenin is key to activating canonical WNT pathway, where it binds to the transcription factor T‐cell factor/lymphoid enhancer‐binding factor (TCF/LEF) and consequently controls the transcription of target genes. 179 The noncanonical pathways are subdivided into the WNT/Ca 2+ pathway and WNT/planar cell polarity (PCP) pathway. 180 Activated WNT/Ca 2+ pathway enhances Ca 2+ influx and initiates various signaling pathways that phosphorylate RORα and induce nuclear translocation of NFAT and nemo like kinase. 180 The WNT/PCP pathway works in both dishevelled (DVL)‐independent and ‐dependent ways: the binding of WNT to the receptor like tyrosine kinase (RYK) is capable of activating protein tyrosine kinase (SRC) independently of DVL, while its binding to the ROR1/2‐Fzd complex can activate DVL, a central mediator of WNT/PCP pathway, and then trigger the activation of Ras‐related C3 botulinum toxin substrate 1 (RAC1), Ras homolog gene family (RhoA), and cell division cycle 42 (CDC42), thereby governing the cytoskeleton remodeling and the activation of AP‐1 and NFAT. 181 , 182 Additionally, some novel noncanonical WNT pathways, such as WNT/mTOR, WNT/YAP/TAZ, WNT/LRP5/mTOR/AKT and WNT/Hippo, have been delineated. 182 , 183 It is warranted that aberrant WNT pathways play important roles in multiple pathological processes, including IR and β cell dysfunction, 184 , 185 , 186 and are causative to the progression of T2DM, 185 , 187 yielding a latent therapeutic strategy for treating this disease. 184 , 188 Dysregulated WNT pathways has been implicated in hepatic IR. It has shown that overexpression of β‐catenin is correlated with a rise in fasting glucose concentrations, while specific knock‐out of β‐catenin in the liver is sufficient to improve hepatic insulin sensitivity and decrease blood glucose concentrations in obese mice. 189 Mechanistically, decreased phosphorylation of hepatic IRS1/2 and GSK3β may mediate the inhibitory effect of WNT/β‐catenin pathway in insulin sensitivity. Meanwhile, the abundant FoxO1 nuclear accumulation connects WNT/β‐catenin activation with hepatic glucogenesis. Consistent with these findings, mice with a knockdown of LDL receptor‐related protein 6 (LRP6), a WNT coreceptor, were resistant to HFD‐induced hyperglycemia and hepatic IR, probably due to the enhanced transcription of leptin receptor. 189 Furthermore, aberrant expression of the key effector of WNT/β‐catenin pathway, transcription factor 7 like 2 (TCF7L2), has been witnessed to induce the transcription of gluconeogenic enzymes (e.g., FBP1, PCK1, and G6Pase) and insulin signaling proteins (e.g., IRS1/2 and AKT2). 190 , 191 Intriguingly, a number of gene TCF7L2 variants are correlated with the susceptibility of T2DM, which primes them as effective predictors of T2DM risk. 192 Hepatic activation of WNT/PCP pathway was also observed to trigger the serine phosphorylation of IRS1 via activating the JNK signaling. 193 In addition, intercellular and interorgan communications might contribute to the hepatic consequences of WNT. For example, secreted frizzled‐related protein 4 (sFRP4), an adipokine with elevated expression in obese WAT, can function as a WNT antagonist and promote hepatic DNL and IR, 194 adding an additional layer of complexity to the role and mechanism of WNT pathway in hepatic function and metabolism. The WNT pathways also modulate adipose IR 185 and adipogenesis, 195 notwithstanding the diverse regulatory roles of different WNT pathways. It has been disclosed that the activation of WNT5a/PCP pathway could promote adipose IR via inducing adipose inflammation. 196 Concomitant with this, sFRP5, a protein counteracting WNT5a/PCP activation, is able to suppress inflammation by blocking the JNK pathway, and thus improve glucose and insulin intoleration in obese mice. 197 Additionally, WNT pathways in the brown adipose tissue (BAT) also regulate IR progression. For instance, knockdown of LRP6, a receptor of WNTs, was shown to improve BAT insulin sensitivity through increasing the expression of PGC1α and uncoupling protein 1 (UCP1). 198 In terms of adipogenesis, it appears that different WNT members have distinct functional outcomes. Basically, several WNT members, including WNT3a, WNT6, WNT8, WNT10a, and WNT10b, have been observed to suppress adipogenesis in the β‐catenin‐ or PCP‐dependent ways, 184 , 199 while others, such as WNT4, WNT5a, WNT5b and WNT11, are capable of stimulating adipogenesis. 195 , 200 , 201 Given the complexity of WNT members, it is of interest to dissect their real roles in adipogenesis under diverse physiological and pathological conditions. The WNT10b/β‐catenin pathway is depressed in the skeletal muscle tissues of overweight and prediabetes. Therefore, not surprisingly, activating WNT10b/β‐catenin pathway could improve insulin

The Hippo signaling pathway, mainly comprising macrophage stimulating 1/2 (MST1/2), large tumor suppressor 1/2 (LATS1/2), and their cofactors such as salvador homolog (SAV) and MOB kinase activator 1A/B (Mob1A/B), predominantly controls the activity of YAP/PDZ‐binding motif (TAZ), two closely related mammalian transcriptional coactivators that shuttle between the cytoplasm and nucleus (Figure 7B ). 292 , 293 , 294 Generally, active MST1/2 phosphorylates LATS1/2, which in turn phosphorylates YAP/TAZ and thus inhibits the activity of YAP/TAZ. 292 In contrast, dephosphorylation allows YAP/TAZ to translocate into the nucleus and combine with its partner transcriptional enhanced association domain (TEAD) to reprogram gene expression. 292 However, the upstream regulation of YAP/TAZ is beyond the Hippo pathway, especially when cells are stimulated by those uncanonical signals such as mechanical stimuli, G protein‐coupled receptor (GPCR) ligands, metabolites, and cell stresses. 292 , 293 , 294 Currently, deciphered regulatory loops that involve many kinds of metabolic responses gradually related YAP/TAZ and Hippo pathway to the peripheral glucose metabolism, insulin signaling, the fate of β cells, and thus pathological processes of T2DM. 293 , 294 , 295 Cells may employ the YAP/TAZ as a coordinator to balance glucose energy supply and consumption. 295 Glucose deprivation, reduced glucose uptake and inhibited glycolysis could repress the activity of YAP/TAZ, the formation of the YAP/TEAD complex, and thus suppress the transcription of target genes, 296 , 297 , 298 such as GLUT3, hexokinase 2 (HK2) and phosphofructokinase B3 (PFKB3), as well as to diminish glucose consumption. Mechanistically, deficient glucose metabolism could act through the activation of AMPK and LAST1/2 to inhibit YAP/TAZ activity, and could also depress the phosphofructokinase 1 (PFK1) to prevent the formation of the YAP/TEAD complex. 296 In contrast, high glucose could upregulate HBP‐dependent O‐GlcNAcylation of YAP, which enhances YAP activity by restraining LATS‐dependent phosphorylation and proteasomal degradation. 299 Given that blood glucose level after long‐standing fasting and postprandial during T2DM is associated with dysregulated cellular glucose metabolism, it is possible that the Hippo pathway may modulate T2DM progression via regulating cellular glucose metabolism. YAP is decreased in the skeletal muscle of obese patients and mice with IR. 300 Consistent with this, the YAP/TAZ pathway might promote peripheral insulin signaling and repress glucogenesis, whereas the Hippo pathway works inversely. For example, YAP/TAZ might upregulate the insulin/AKT pathway by potentiating IRS2 transcription in the liver of mice with codeleted PTEN and SAV1. 301 Moreover, the mutually concordant positive effects between YAP/TAZ and mTOR pathways also underpin the connection between YAP/TAZ and the insulin/AKT signaling. 302 On the contrary, the Hippo pathway exerts converse effects on insulin signaling and glucose metabolism. One evidence is that upregulating the upstream Hippo kinase MST3 could exacerbate IR, hyperglycemia and hyperinsulinemia by depressing IRS1 and upregulating the transcription of gluconeogenic regulators and enzymes. 303 YAP can also act through PGC1α to suppress the transcription of hepatic gluconeogenic genes, lower plasma glucose level and improve glucose tolerance. 304 Additionally, glucagon activates LATS and restrains YAP, 305 which in return eradicates upregulation of hepatic gluconeogenic genes and represses gluconeogenesis. 304 The Hippo pathway may also serve as an essential regulator in the capacity of adipose tissue to endure metabolic stress. 306 To deal with the increasing stress during obesity, WAT might rely on YAP/TAZ to resist apoptosis and ameliorate T2DM. 38 , 306 Additionally, it has shown that the impairments of FA oxidation and consequent enhanced adiposity of skeletal muscle in obese patients or prediabetic mice are caused partially by reduced YAP. 300 Intriguingly, the Hippo pathway may also participate in inflammatory responses 300 , 307 and mitochondrial maintenance, bringing about an attractive notion that dysregulation of YAP/TAZ and Hippo pathway might influence the advancement of T2DM in the manners beyond our imagination. The Hippo pathway and YAP/TAZ also modulate pancreatic cell differentiation, proliferation and apoptosis. 293 , 308 Overall, YAP/TAZ could render expansion of early embryonic pancreas epithelium, but hamper endocrinogenesis of pancreatic progenitor cells. 308 In harmony with this, YAP/TAZ in adult pancreatic β cells remains low expression level, 308 indicating that silencing of YAP/TAZ is critical for the maturation of β cells. In contrast, the Hippo pathway 309 tends to induce apoptosis of β cells, as MST1 is activated to empower β cell apoptosis, while its deletion pronouncedly restores β cell function and mass, attenuating diabetic conditions in mice with T2DM.

The TGFβ superfamily contains a number of subfamily proteins, such as bone morphogenetic proteins (BMPs), activins and TGFβs (Figure 7D ). Activation of TGFβ receptors could further activate small mothers against decapentaplegic homolog (SMADs), a class of second messengers, and other signaling pathways, such as MAPK, RHO GTPase, and PI3K/AKT pathways. 322 The TGFβ pathway connects contextual determinants with specific cellular responses by controlling SMAD‐dependent transcription programming and integrating with other pathways. These fundamental mechanisms underlie the function of TGFβ pathway in controlling peripheral insulin signaling and pancreatic β cell biology during T2DM pathogenesis. An increasing body of evidence supports that the TGFβ pathway can affect glucose homeostasis. It is increasingly clear that several members of the TGFβ superfamily, including activin A and B, GDF11, BMP2‐4, and TGFβ1‐3, have emerged as novel regulators in the insulin signaling of adipose tissue, skeletal muscle, and liver. 323 Among which, the bioactivity of the activin/GDF11 is reported to be modulated by the antagonists follistatin (FST) and follistatin like 3 (FSTL3), which could increase fat mass and adipose IR, and therefore disrupt glucose homeostasis. 324 Specifically, overexpression of fstl3 in obese mice is capable of improving muscle insulin sensitivity and reducing fat accumulation, but enhancing hepatic glucagon sensitivity. 325 By contrast, genetic removal of fst enhanced WAT insulin sensitivity and suppressed HGP, thereby ameliorating glucose tolerance in obese mice. 326 BMPs, particularly BMP4, 6, and 7, are also involved in the control of glucose homeostasis. Specifically, BMP7 promotes, while BMP4 decreases insulin sensitivity in the adipose and muscle of T2DM mice, respectively. 327 Mechanistically, BMP7 can activate PDK1 and AKT to boost GLUT4 translocation to the plasma membrane and in turn increase glucose uptake, while BMP4 counts on the activation of PKCθ to induce IR. BMP6 may restore the levels of blood glucose and lipids in T2DM mice via reducing hepatic gluconeogenesis and glucose output. 328 However, the dysregulation of TGFβ pathway in the diabetic peripheral tissues remains incompletely understood and awaits further study. The discoveries about the roles of TGFβ in pancreatic development and functions establish another strong link between the TGFβ pathway and T2DM, 329 regardless of the diversity of TGFβ pathway in controlling the proliferation, apoptosis and differentiation of β cells. To be specific, activin A could inhibit β cell proliferation, but facilitate the generation of β cells from adult stem cells, human embryonic stem cells, ductal cells, human amniotic epithelial cells), 330 and α cells. 331 Activin A and activin B potentiate the dedifferentiation of β cells by depressing and upregulating the transcription of crucial genes associated with β cell maturity and immaturity, respectively. 332 Activation of SMADs is key to the regulation of TGFβ signaling on β cell fate. In detail, SMAD7 is believed to enhance proliferation, 333 while SMAD2 and SMAD3 block proliferation of β cells by regulating the nuclear localization of p27, a cell‐cycle modulator inhibiting cyclin‐dependent kinase (CDK). Otherwise, activation of TGFβ/SMAD3 signaling also has a role in β cell apoptosis. 334 Furthermore, SMAD2, 3, and 7, particularly SAMD7, connect the dedifferentiation and proliferation of β cells together. 335 Recently, the roles of TGFβ pathway in controlling β cell functions are also emerging. For instance, activin A/B, TGFβ1, FST, GDF11, and BMPs, have been demonstrated to enhance GSIS, 336 , 337 , 338 while the addition of FSTL3 is able to reverse these effects in both functional and nonfunctional islets. 338 Mechanistically, these effects may be predominantly attributed to the activation of SMAD2, since its disruption could compromise insulin secretion under high glucose and diabetic conditions. 337 Furthermore, the effects of TGFβ pathway on GSIS may be associated with the modulation of K ATP and calcium channels activity, 337 as well as the altered expression of essential genes governing insulin secretion, such as the GLUT2 and calcium voltage‐gated channel subunit alpha1 D genes. 338

BAs, the major components of bile, are synthesized from cholesterol in liver via the classical pathway regulated by cholesterol‐7α‐hydroxylase (CYP7A1). 371 BAs play important roles in facilitating the absorption of dietary lipids, maintaining cholesterol homeostasis and activating hormones (Figure 9 ). Over the past decades, BAs have been regarded as critical regulators in glucose, lipid and energy metabolism, as well as potential targets for preventing cardiovascular complications and improving glycemic control in T2DM patients. 372 BAs act primarily through activating intracellular ligand‐activated receptors, among which the best‐studied are FXR and G protein‐coupled BA receptor (TGR5). FXR (also known as NR1H4) is a member of nuclear receptor family and expressed in many tissues. 373 FXR located in the intestine and liver, could be activated directly by BAs and forms a heterodimer with retinoid‐X‐receptor (RXR). FXR exerts effects on the expression of downstream genes and the inhibition of BAs synthesis through two ways: (1) inducing the small heterodimer partner (SHP) to inhibit CYP7A1 in the liver; (2) increasing circulating FGF19/15 in the gut to repress CYP7A1. TGR5, belonging to the rhodopsin‐like superfamily of G protein‐coupled receptors, is a cell membrane receptor for secondary BAs and moderately expressed in nearly all tissues. BAs could regulate glucolipid metabolism via binding to FXR and TGR5. 374 For example, activated FXR is able to repress hepatic DNL 375 and gluconeogenesis 376 by inhibiting the expression of SREBP1c, PEPCK, and G6Pase. Moreover, systematic knockout of TGR5 could aggravate IR and glucose intolerance through miR‐26a and the subsequent cAMP/PKA pathway. 377 , 378 Furthermore, both FXR and TGR5 are capable to regulate GSIS in pancreatic β cells through inducing the relocalization of GLUT2 on the membrane. 379 TGR5 is coexpressed with FXR on L cells in the intestine, and the activation of FXR could facilitate the release of GLP‐1, 380 which then enhances insulin secretion and improves glucose homeostasis. Intriguingly, hyocholic acid has been shown to promote GLP‐1 secretion through synchronously activating TGR5 and inhibiting FXR, which exerts profound effects on glucose metabolism in the pig and diabetic mouse models. 381 As well, a recent study showed that soluble dietary fiber oligofructose could activate TGR5 to enhance GLP‐1 activity by stimulating 6α‐hydroxylated BAs, thus improving host glucose metabolism and maintaining glucose homeostasis. 382 All of above suggest that BAs signaling pathways mediated by intestinal GLP‐1 have strong potential for T2DM therapeutic applications. The gut microbiota has been showed to regulate the level and composition of BA metabolites through secreting enzymes to catalyze the dehydroxylation of BAs, thus influencing the function of BAs in regulating glycolipid metabolism. In turn, BAs also interact with and affect the gut microbiota, 383 thus creating a dynamic equilibrium between them. Interestingly, metformin has been reported to decrease the abundance of species of B. fragilis in the intestine to upregulate the bile acid glycoursodeoxycholic acid in the gut, thus improving metabolic dysfunction and hyperglycemia in T2DM patients. 384 Adjustment of BAs homeostasis has been applied to optimize glycaemia parameters and improve pathological conditions in various diabetic animal models and human patients, 385 , 386 which implies potential therapeutic roles for BAs signaling in the treatment of T2DM.

4.12.1 PPARs pathway PPARs, including PPARα, PPARβ/δ, and PPARγ, have been characterized as core lipid sensors that modulate whole‐body energy metabolism (Figure 9 ). 467 PPARα is the predominant PPAR isoform in the liver and plays an important role in regulating fatty acid transport, ketogenesis, and β‐oxidation. 468 In addition to lipid regulation, while, PPARβ/δ could increase the production of GLP‐1 to stimulate insulin secretion and reduce IR. 469 PPARγ mainly functions as a core regulator of adipogenesis, improving insulin sensitivity and glucose metabolism. In terms of mechanisms, deletion of IRF3 can increase the expression of PPARγ, leading to severe IR and glucose intolerance. 470 As opposed to IRF3, TAZ could protect mice from HFD‐induced IR and glucose intolerance through upregulating the expression of PPARγ. 471 Furthermore, PPARγ may induce the enhanced level of FGF1 in the adipose tissue, and considering the aforementioned protective function of FGF1 in insulin sensitivity and glucose metablism, 363 PPARγ/FGF1 axis may be indispensable in maintaining insulin sensitization an metabolic homeostasis. 364

By binding to membrane receptors, IL‐6 leads to the activation of multiple downstream signaling pathways inside target cells, such as PI3K/AKT, MAPK, and janus kinase (JAK)/STAT. 479 IL‐6 is reported to inhibit IRS, AKT2 and ERK phosphorylation directly or indirectly in the liver, leading to impaired insulin signaling pathway and IR. 480 , 481 , 482 , 483 Besides, IL‐6 is able to suppress glucose transport by affecting adiponectin and GLUT4, 484 repress lipogenesis by inhibiting PPARγ, 485 and stimulates lipolysis by activating AMPK in the adipose tissue, 486 which may link the IL‐6 pathway to the progression of IR and T2DM. However, there are also some conflicting results. For instance, in the case of HFD feeding, deletion of IL‐6 and IL‐6R promoted the development of hepatic IR. 487 , 488 In addition, the role of IL‐6 in pancreatic islets appears to be contradictory. Several studies have shown that IL‐6 contributes to the induction of low grade inflammation in the islets, which ultimately results in impaired insulin secretion. 489 , 490 , 491 Paradoxically, acute exposure of IL‐6 does not seem to exert its potential hazardous effects on the islets, 492 or affect the normal functioning of β cells. 480 , 493 , 494 , 495

The NOD‐like receptor (NLR) family pyrin domain‐containing 3 (NLRP3) inflammasome is a cytosolic multiprotein complex composed of the innate immune receptor NLRP3, the adapter apoptosis‐associated speck‐like protein containing card (ASC), and the inflammatory protease caspase‐1. Several endogenous ligands related to metabolic stress, such as lipopolysaccharide (LPS), 508 , 509 palmitic acids, oleic acid, 510 homocysteine and ROS, 511 , 512 have been reported to activate NLRP3 inflammasome in metabolic organs. Activated NLRP3 inflammasome induces the maturation and release of IL‐1β and IL‐18, responding to microbial infection, endogenous danger signals, and environmental stimuli. 513 IL‐1β has been demonstrated to induce IR by reducing IRS1 phosphorylation and expression 514 and inducing IL‐6 release. 515 The deficiency of NLRP3, caspase‐1, or IL‐1β in the liver and adipose tissue can result in reduced inflammasome activation, improved lipid metabolism, enhanced insulin sensitivity, and reduced blood glucose in DIO mouse models. 516 , 517 , 518 Additionally, IL‐1β can influence the activation of MAPK 519 and NF‐κB, 520 which in turn dysregulates the expression of genes involved in β cell death. Likewise, NLRP3 knockout is able to protect β cells from damage, rescue HFD‐induced β cell loss, and increase insulin secretion. 521 , 522

T2DM causes a number of long‐term complications in the neural, macrovascular and microvascular systems, which are instrumental for death and disability of patients with T2DM (Figure 10 ). 10 , 534 These complications encompass the damages to the coronary and cerebrovascular arteries, the kidneys, the eyes, and the nerves. 10 , 534 In the following section, we will revisit the potential signaling pathways, as well as their associated therapeutic targets, underlying T2DM complications. FIGURE 10 Signaling pathways in T2DM complications. Hyperglycemia causes macrovascular and microvascular lesions, contributing to diabetic complications such as diabetic retinopathy, diabetic nephropathy, diabetic neuropathy, and diabetic cardiovascular disease. Targeting the major pathological features of different complications, some signaling pathways have been discovered. These signaling pathways play important roles in promotion or inhibition of disease development, contributing to the treatment of diabetic complications. 5.1 Diabetic nephropathy Diabetic nephropathy is perhaps among the most devastating T2DM complications, which bedevils up to 40% of T2DM patients and leads to kidney failure, CVD, and premature death. 535 The pathological changes of DN mainly include renal fibrosis and podocyte injury, 536 which are linked to various signaling pathways, particularly the TGFβ and mTOR pathways. Dysregulation of TGFβ pathway is widely considered playing a key role in renal fibrosis, which is characterized by excessive deposition of ECM. Correspondingly, targeting TGFβ was also identified to be of therapeutic potential for DN. 537 , 538 Considerable evidence has revealed increased expression and activation of TGFβ pathway components, such as TGFβ1 and SMAD2/3, in the fibrotic kidney. 537 , 538 Consistently, inhibition of TGFβ1 or its downstream pathways substantially restrained renal fibrosis in DN, whereas overexpression of TGFβ1 prompted renal fibrosis, suggesting a profibrotic role of TGFβ pathway in DN. 537 , 538 TGFβ pathways can induce the activation of myofibroblasts, apoptosis of endothelial cells and podocytes, overproduction of ECM, and impairment of ECM degradation via multiple downstream mechanisms. 538 , 539 , 540 , 541 These downstream mechanisms include the imbalance between upregulation of profibrotic SMAD3 and downregulation of antifibrotic SMAD7, interplays with other signaling pathways (e.g., MAPK, WNT/β‐catenin, ERK, PI3K/AKT, mTOR pathways), and the involvement of miRNAs, lncRNAs and epigenetic modifications. 538 Accordingly, numerous approaches blocking TGFβ signaling, including TGFβ neutralizing antibodies, antisense TGFβ oligodeoxynucleotides, soluble human TGFR2, and specific inhibitors to TGFR1 kinases can effectively halt the progression of renal fibrosis and DN. 537 , 538 , 542 However, due to the side‐effects of TGFβ inhibition in autoimmune diseases, direct targeting of TGFβ1 is barely possible to be a viable strategy against renal fibrosis and DN. 538 Alternatively, indirect approaches to antagonizing TGFβ signaling might hence be promising to yield effective therapeutic agents for DN. In addition to TGFβ signaling, other signaling pathways may be of therapeutic potential in renal fibrosis and DN. For example, it has shown that high glucose can trigger the JAK/STAT signaling cascade to stimulate excessive proliferation of glomerular mesangial cells and aggravate DN, 543 and the small‐molecule inhibitor of JAK/STAT, baricitinib, has been demonstrated to decrease albuminuria in patients with T2DM and DN. 544 Podocyte injury is another important pathological feature of DN, in which the mTOR pathway is hyper‐activated due to high concentrations of glucose. 545 , 546 Enhanced mTORC1 signaling may trigger podocyte injury and DN by stimulating podocyte hypertrophy, mesangial expansion, glomerular basement membrane thickening, foot process effacement, podocyte loss, and albuminuria. 546 , 547 In line with this, restoration of mTORC1 activity is a critical therapeutic approach for the prevention of DN. For example, rapamycin, a specific and potent inhibitor of mTOR, markedly ameliorated mesangial expansion and proteinuria. 548 Moreover, the mechanism by which metformin and SGLT2 inhibitors exert kidney‐protective effects may be also related to the mTOR pathway, albeit it is currently unclear. 546

Among diabetes neuropathy, distal symmetric polyneuropathy is very common and defined as a loss of sensory function beginning distally in the lower extremities. 568 Approximately 30–50% of patients with diabetic neuropathy develop neuropathic pain called painful diabetic neuropathy (PDN). 569 PDN is characterized by neuropathic pain, small‐fiber degeneration, and dorsal root ganglion (DRG) nociceptor hyperexcitability. 570 DRGs are important neurons that comprise thermoreceptors, mechanoreceptors, and itch sensors. Nav1.8, an α‐subunits of voltage‐gated sodium channels, is important for the development of abnormal pain sensation. 571 Methylglyoxal, a reactive metabolite increased in diabetes, could posttranslationally modify Nav1.8 and enhance the activity of the nonselective cation channel TRPA1, respectively resulting in Nav1.8 gain of function pain and neuron hyperexcitability. 572 , 573 In addition, excitatory CXCR4/CXCL12 signaling in Nav1.8‐positive DRG neurons plays a critical role in the pathogenesis of mechanical allodynia and small‐fiber degeneration in a mouse model of PDN. 570 In T2DM, Ca 2+ channels have also been implicated in PDN. For example, CaV3.2 activity is enhanced through the glycosylation of extracellular arginine residues, resulting in hyperexcitability pain in DRG neurons. 574

Recent evidence showed a mutual interplay between COVID‐19 and T2DM. 579 On the one hand, COVID‐19 patients with diabetes are more likely to develop severe condition of higher death incidence compared with those without diabetes. 580 , 581 On the other hand, new‐onset hyperglycemia, ketoacidosis, diabetes, and severe metabolic complications of preexisting diabetes have been observed in patients with COVID‐19. 582 , 583 , 584 There is an intricate interaction network existing between diabetes and COVID‐19. Common pathogenetic processes between COVID‐19 and diabetes mellitus discovered by differential gene expressions pattern analysis are related to the mRNA metabolism, subcellular organelle organization, nucleotide synthesis, immune responses, and autophagy process. Other studies revealed that five biomarker genes ( CP , SOCS3 , AGT , PSMB8 , and CFB ) and 4 miRNAs (hsa‐miR‐298, hsa‐miR‐3925‐5p, hsa‐miR‐4691‐3p, and hsa‐miR‐5196‐5p) closely bridge T2DM and COVID‐19. 585 , 586 , 587 Severe COVID‐19 is associated with worse glycemic control, 588 and the reason for this may partly lie in increased expression of ACE2, a key receptor for severe acute respiratory syndrome coronavirus‐2 (SARS‐CoV‐2), in lungs of T2DM patients. 589 Growing evidence suggests that elevated glucose level and glycolysis may directly increase and sustain SARS‐CoV‐2 replication. 590 Meanwhile, diabetes can affect immune response to SARS‐CoV‐2. Hyperglycemia and IR are capable of increasing synthesis of AGEs and proinflammatory cytokines to mediate tissue inflammation. 591 Additionally, hyperglycemia may also impede type I interferon production and signaling 591 to augment inflammatory responses to SARS‐CoV‐2 infection, leading to extreme systemic immune responses. 592 , 593 The precise mechanisms underlying the development of new‐onset T2DM in people with COVID‐19 are still not known, but may involve a number of complex etiologies. On the one hand, SARS‐CoV‐2 could directly infect β cells and lead to β cell dysfunction and metabolic dysregulation. 594 SARS‐CoV‐2 could also bind to NRP1, thus attenuating insulin secretion and inducing β cell apoptosis through the JNK/MAPK pathway. 595 Meanwhile, SARS‐CoV‐2 might downregulate ACE2 activity and subsequently activate macrophages and trigger NF‐κB signaling in the pancreas, leading to excessive synthesis and secretion of inflammatory cytokines, eventually damaging islets and β cells. 596 , 597 On the other hand, IR, resulting from viral infection seems to be another cause of hyperglycemia upon COVID‐19. Infection of SARS‐CoV‐2 may induce an inflammatory state, 598 , 599 in which IL‐6 might play a primary albeit not exclusive role in inducing IR and β cell dysfunction. 579 , 600

Thiazolidinediones (TZDs), a kind of classic glucose‐lowering drugs, are the ligands for PPARγ. PPAR agonists primarily stimulate WAT remodeling and modulate lipid flux to improve insulin signaling and glucose homeostasis. TZDs binding to PPARγ could promote FFAs storage and reduce lipid ectopic accumulation in subcutaneous fat tissue. 609 Rosiglitazone and pioglitazone are the representatives of TZDs. Due to the potential for serious adverse cardiovascular effects, the United States Food and Drug Administration adds a "black box warning" to the rosiglitazone label, 610 while pioglitazone can significantly lower risk of death, myocardial infarction, or stroke among patients with diabetes. 611 In order to make good use of PPARγ, some studies focused on the potential effects of posttranslational modifications (PTMs) of PPARγ on treating T2DM in terms of phosphorylation, acetylation, ubiquitination, SUMOylation, O‐GlcNAcylation, and S‐nitrosylation. PTMs have shed light on selective activation of PPARγ, which shows great potential to circumvent TZDs' side effects while maintaining insulin sensitization. 612

GLP‐1 is an incretin hormone and exerts its action by binding to GPCRs to stimulate insulin secretion through rapid increases of cAMP and intracellular Ca 2+ . 619 In β cells, the binding of GLP‐1 and its receptor leads to activation of adenylate cyclase (AC) and a subsequent increase in cAMP. cAMP activates PKA, thereby inducing the closure of K + channel and the opening of Ca 2+ . The subsequent Ca 2+ influx promotes exocytosis of insulin granules and acute secretion of insulin into the circulation. 620 GLP‐1 also contributes to the control of blood glucose by inhibiting glucagon secretion, gastric emptying, and food ingestion. 619 , 621 GLP‐1R agonists also have a role in the treatment of neurodegenerative diseases, 622 nonalcoholic fatty liver disease 623 and obesity. 624 Given that GLP‐1 is mainly inactivated by dipeptidyl peptidase4 (DPP4) within 3 minutes in the circulation, 625 DPP4 inhibitors are applied to provide prolonged effects in vivo. 626 Notably, recent studies suggested DPP4 as a functional receptor for MERS‐CoV and SARS‐COV‐2, and its inhibitors have been found to inhibit infection with coronavirus. 627 , 628

Alpha‐glucosidase is an intestinal brush border enzyme responsible for the hydrolysis of disaccharides into monosaccharides, which is necessary for carbohydrate absorption. 635 Inhibition of alpha‐glucosidase results in reduced carbohydrates absorption and increased GLP‐1, thus decreasing the postprandial blood glucose. 636 , 637 Alpha‐glucosidase inhibitors that are currently used in clinic include acarbose, voglibose, and miglitol, as well as a variety of natural products such as hypericin, oleanolic acid and ursolic acid. 638

The glucagon receptor (GCGR) is a GPCR. Its activation results in increased glycogenolysis and gluconeogenesis via the cAMP/PKA pathway, 640 while its antagonization improves glucose control in T2DM. Nevertheless, adverse events stopped the development of GCGR antagonists (GRA). Although with new approaches, none of GRA has progressed to phase III clinical trials so far. 641 Additionally, glucagon signaling has been associated with increased energy expenditure. GCGR agonists can increase energy expenditure and reduce hepatic fat by promoting HGP. Although this may cause a spike in blood glucose, which can be effectively offset by GLP‐1R agonists. 642 Thus, GCGR has become a focus as a pharmaceutical target in the context of bi‐ or tri‐modal peptide agonists for the treatment of metabolic diseases. 643

The G protein‐coupled receptor 40 (GPR40), also known as free fatty acid receptor 1 (FFAR1), is highly expressed in the pancreas and enteroendocrine cells in the gastrointestinal tract. When glucose level elevates, GPR40 facilitates insulin secretion through the PLC/IP3/PKC pathway. 650 , 651 GPR40 could also promote incretin secretion, such as GLP‐1 and gastric inhibitory polypeptide. 652 LY2922470 is a GPR40 agonist that has shown an effective and persistent dose‐dependent reduction in glucose levels and a significant increase in insulin and GLP‐1 secretion in preclinical trials. 653 GPR40 agonists might also regulate inflammatory responses and thus play a role in improving the development of diabetes complications. 654 , 655

Several mitochondria‐targeted antioxidants and peptides have been found to reduce oxidative stress and mitochondrial damage, so as to exert therapeutic effect on some chronic diseases such as diabetes and Alzheimer's disease. These antioxidants include SkQ1, MitoQ, and Coenzyme Q. 658 Both SkQ1 and MitoQ are ubiquinone derivatives. Ubiquinone can serve as electron carriers and antioxidants to profoundly prevent lipid peroxidation. 659 Interestingly, Coenzyme Q is a lipid‐soluble molecule, one of the key components of electron transport chain. After being oxidized, Coenzyme Q needs to undergo a self‐reducing process before it can continue to function as an antioxidant. 660 Szeto‐Schiller (SS) peptide is a novel mitochondrial‐targeted peptide. 661 The water‐soluble tetrapeptide SS‐31 concentrates in the inner mitochondrial membrane and may exert radical‐sweeping ability without reliance on mitochondrial membrane potential and energy. 662 Given that administration of mitochondria‐targeted antioxidants may restore mitochondrial function and alleviate symptoms, further examination and analysis of treatment efficacy and safety are suggested to prepare for clinical trials.

Throughout the past decades, the surged prevalence of T2DM and its vicious complications has coined an urgent craving for better understanding the mechanisms underpining the pathogenesis of T2DM and how to manage T2DM efficiently, particularly through drug administration. As a multifactoral disease, T2DM is driven by genetic, epigenetic, and nongenetic mechanisms, among which a substantial fraction are interdependent signaling pathways. A highly possible paradigm for various signaling pathways acting in T2DM, is where they serve as both the causes and consequences of T2DM progression and, function in an interacted manner rather than separately. In particular, environmental factors, together with genetic risks correlated to T2DM, are prone to elicit vibrations of the signaling pathways in the peripheral tissues and pancreatic islets. Interfering with these signaling pathways seems to engender outcomes that are conducive or obstructive to the pathological pertubances of T2DM, especially IR and β cell dysfunction. In terms of IR, modifications in these signaling pathways are likely to rely on their multifaceted interplays with the insulin/AKT to fluctuate the peripheral susceptibility to insulin action, despite the discrepancies in the specific roles of individual pathways. In terms of β cell dysfunction, signaling pathways also converge to regulating insulin secretion and fate of β cells, so intrinsic or acquired traits of these signaling pathways appear to partially dictate the adaptive capacity of β cells in response to metabolic demands. Importantly, many medications that normalize blood glucose levels and prevent or inverse diabetic complications have been identified to establish their pharmaceutical benefits via the mechanisms involving signaling pathways. One prominent example is the extensively used clinical antidiabetic drug, metformin, which acts through the AMPK pathway to powerfully improve metabolic disoders. 15 Nevertheless, a fraction of drugs or compounds targeting signaling pathways have unexpected side‐effects or inadequate effectiveness, owing to additional pathophysiological roles or minor metabolic effects of signaling pathways. For example, TGFβ‐specific therapies have been witnessed to exert inevitable effects on immune system, thus slowering the progress in employing them for treating DN. For these reasons, discerning the functions and mechanisms of signaling pathways in T2DM would help ease the road to well‐managing T2DM with drugs. However, there are several obstacles in studying signaling pathways and interventions of T2DM. Firstly, an unfortunate trend is excessively highlighting the pathological significance of a sole molecule in signaling pathways, which is intensified by the dramatic progress in genetic manipulation technologies. This misfortune of trend in large part relates to concerns that manipulating genes could inevitably cause adaptive or compensatory effects concealing the real effects of certain molecules or pathways in T2DM progression. Luckily, considerable progresses in omics methods, such as, spatially resolved genomics, transcriptomics, and metabonomics, make it available to rectify this trend. In particular, devising ways to assess genomic, proteomic, and metabolic status in prediabetic and diabetic tissues while preserving spatial information and single cell resolution, is of great propensity to delineate the inclusive but accurate molecular mechanisms of T2DM. Furthermore, biomolecular condensates or droplets have been increasingly identified as an interaction basis for the molecules in signaling pathways, 670 which is less emphasized in the past studies of T2DM. Hence, it might be urgent to highlight the roles of liquid condensates or liquid–liquid phase separation in T2DM, which is possibly linked to abnormal protein aggregation. In addition, another important question to purse for the accurate roles of signaling pathways in T2DM, is to create in vitro models that sincerely mimic the real progression of T2DM. Nevertheless, feasible approaches to address this problem might be involved in the organoid technologies, which may largely recapitulate essential features of in vivo organ development and biological function. As such, they offer tractable and faithful in vitro tools for disentangling molecular mechanisms of T2DM and developing regenerative pancreatic islets, as well as confer a promising strategy that compensates for pharmacological therapies in advanced T2DM. Importantly, we anticipate that systematic application of these novel methodologies and notions in basic research and clinical translation of signaling pathways in T2DM would have a promising impact in contributing to the discovery of antidiabetic drugs with enhanced effectiveness and safety. This is possibly because precise interventions usually derive from precise investigations. In summary, remarkable insights over the past few decades have gained vital understandings of signaling pathways relate