Resetting the glucose sensor: dorzagliatin and the evolution of glucokinase activation in diabetes therapy

Abstract

The progressive decline in functional β-cell mass represents a fundamental, unmet challenge in type 2 diabetes (T2DM) management. Dorzagliatin (HMS5552), a fourth-generation, dual-acting, and glucose-dependent glucokinase activator (GKA), offers a distinct mechanistic approach that aims to move beyond symptomatic glycemic control toward potential glucose sensor modulation. This review critically synthesizes the development of dorzagliatin, designed to overcome the hypoglycemia and tachyphylaxis of earlier GKAs. While its unique allosteric mechanism enables synchronized restoration of insulin secretion and hepatic glucose metabolism, we critically evaluate its clinical positioning. Observational data from the DREAM study suggest sustained drug-free glycemic maintenance; however, we emphasize that these findings do not yet meet the rigorous 2021 American Diabetes Association/European Association for the Study of Diabetes consensus definition of “diabetes remission” and require validation through randomized withdrawal trials. Furthermore, we address the critical absence of long-term cardiovascular outcomes trials (CVOTs) and the theoretical concerns regarding hepatic de novo lipogenesis. We also propose a synergistic framework for combining dorzagliatin with sodium-glucose cotransporter-2 inhibitors to counteract compensatory endogenous glucose production, positioning it as a precision-driven component in the evolving T2DM therapeutic landscape. Should future CVOTs demonstrate safety and efficacy, GKAs like dorzagliatin could potentially be integrated into treatment guidelines, possibly as a “sensor-calibrator” for a subset of patients identified through metabolic and genetic phenotyping. However, it must be emphasized that no CVOT data are currently available for dorzagliatin.

Introduction: the unmet need in diabetes therapy and the revival of the glucokinase target

The global epidemic of diabetes presents a formidable and growing public health challenge. By 2045, it is projected that nearly 700 million adults will be living with diabetes, the majority with type 2 diabetes (T2DM).¹ T2DM is characterized by a dual core pathophysiology: progressive dysfunction and failure of pancreatic β cells and concomitant insulin resistance in peripheral tissues.²

The current therapeutic landscape has been transformed by glucagon-like peptide-1 receptor agonists (GLP-1 RAs) and sodium-glucose cotransporter-2 (SGLT2) inhibitors, which offer profound weight loss and cardiorenal protection.³ However, a significant therapeutic gap persists. Approximately 40–50% of patients fail to achieve glycemic targets due to tolerability issues (eg, gastrointestinal adverse events with GLP-1 RAs), contraindications, or cost barriers.^4–6 More fundamentally, while these agents successfully manage comorbidities and can improve β-cell function indirectly (eg, by relieving glucotoxicity) or, in the case of GLP-1 RAs, directly via cyclic adenosine monophosphate (cAMP)-mediated pathways, they do not directly target the primary defect in glucose sensing that precipitates β-cell failure.⁷ Nevertheless, glucose-lowering itself—whether through SGLT2 inhibitors or GLP-1 RAs—may reduce endoplasmic reticulum (ER) stress and allow β-cell rest, contributing to improved functional outcomes. However, most treatments act as symptomatic “glucose controllers” and their effects on halting or reversing the underlying decline in β-cell mass remain uncertain. This therapeutic gap underscores a critical need for agents that address the root cause of dysregulated glucose sensing and secretion. While GLP-1 RAs and SGLT2 inhibitors currently dominate the therapeutic landscape through potent weight loss and cardiorenal mechanisms, their mechanisms of action differ from direct glucokinase (GK) activation. GLP-1 RAs, for example, not only potentiate insulin secretion but also enhance β-cell glucose sensitivity via cAMP signaling.⁸ However, these agents do not directly modulate the glucokinase sensor.⁷ Dorzagliatin offers a distinct mechanistic positioning: it aims to modulate the “thermostat” setting itself,⁹ suggesting a potential role in precision medicine, particularly for non-obese, early-stage patients where β-cell sensory dysfunction may dominate over insulin resistance. Nevertheless, this hypothesis remains to be tested in prospective, phenotype-driven trials.

Central to this homeostasis is GK, the body’s principal “glucose sensor”. The seminal work of Franz Matschinsky and colleagues in the 1980s first established GK as the pancreatic β-cell glucose sensor,^10–12 laying the entire foundation for targeting GK in diabetes therapy.^{13 14}

Complementing these genetic and physiological studies, the GK inhibitor mannoheptulose has long been used as a pharmacological tool to demonstrate that GK activity is essential for glucose-stimulated insulin secretion (GSIS). Classic studies showed that mannoheptulose blocks glucose phosphorylation in β cells and abolishes GSIS, confirming that GK activity is essential for glucose sensing. Recently, however, this picture has been complicated by a paradoxical finding: List and colleagues¹⁵ reported that chronic mannoheptulose exposure under hyperglycemic conditions prevented, rather than exacerbated, the detrimental effects of high glucose on β-cell function. In other words, GK inhibition can protect β cells from glucotoxicity—a finding that challenges the conventional rationale for GK activation. These diverse lines of evidence converge on a unifying theme: GK is the body’s principal glucose sensor, a role rooted in its distinctive kinetics.

This hexokinase isoform exhibits a distinctive kinetic profile with a low affinity for glucose (high Michaelis constant, Km) and positive cooperativity, making its activity exquisitely sensitive to physiological changes in blood glucose concentration. In pancreatic β cells, GK catalyzes the rate-limiting step in glycolysis, coupling glucose metabolism to ATP production, closure of ATP-sensitive potassium channels, membrane depolarization, calcium influx, and ultimately, GSIS. In hepatocytes, GK regulates the balance between glycogen synthesis and gluconeogenesis, thereby controlling hepatic glucose output. This dual-tissue role makes GK a master regulator of postprandial glucose disposal and fasting glucose production.^{16 17}

The therapeutic potential of GK activation was recognized decades ago, but the clinical translation journey has been arduous. Early-generation GKAs, such as piragliatin, MK-0941, and AZD1656, demonstrated potent initial glucose-lowering effects.¹⁸ However, these promising starts were consistently derailed in Phase II trials by a triad of dose-limiting adverse effects: (1) an unacceptable risk of hypoglycemia, as these full agonists disrupted GK’s glucose dependency¹⁹; (2) hypertriglyceridemia and elevated low-density lipoprotein-cholesterol, likely due to unrestrained hepatic lipogenesis²⁰; and (3) tachyphylaxis, a rapid loss of glycemic efficacy within weeks, potentially linked to GK regulatory protein (GKRP)-mediated sequestration in the nucleus or downregulation of GK expression.²¹ These failures were not merely clinical setbacks, they provided a crucial mechanistic lesson: effective and safe GK activation must be contingent on ambient glucose levels, thereby preserving the enzyme’s intrinsic sensor function.²²

The development of dorzagliatin (HMS5552) was conceived as a direct response to these lessons. Engineered as a fourth-generation, glucose-dependent partial allosteric activator, dorzagliatin was designed to cooperate with glucose to enhance GK activity primarily within the hyperglycemic range.¹⁹ Unlike its predecessors which acted as “blunt instruments”, dorzagliatin functions as a “precision calibrator” of the glucose sensor. This review aims to provide not only a comprehensive synthesis of dorzagliatin’s journey but also a critical appraisal of its place in diabetes therapeutics. We specifically address its differentiation from current standard-of-care agents, analyze its mechanism as a form of “pharmacological sensor repair”, evaluate the strength and limitations of the clinical evidence for disease modification, and rigorously examine the scientific rationale and challenges for its application beyond T2DM.

Pharmacological foundation: the mechanism of glucose-dependent GK activation

The promising clinical profile of dorzagliatin has been closely linked to its optimized kinetic interaction with GK, a property that distinguishes it from earlier GKAs that failed in clinical trials. Whether this profile will translate into long-term clinical success—particularly with respect to cardiovascular safety and durable glycemic control—remains to be established.

The importance of preserving GK’s glucose dependency is underscored by classic pharmacological experiments. The GK inhibitor mannoheptulose, which blocks glucose phosphorylation in β cells, has been shown to abolish GSIS while leaving responses to other secretagogues intact.¹⁵ This finding established that GK activity is not merely permissive but essential for GSIS, and it set a key benchmark for GKA development: any therapeutic activator must enhance GK activity without overriding its inherent glucose sensitivity—a lesson learnt from the failure of first-generation GKAs that acted as glucose-independent full agonists.

Enzyme kinetic profile: precision engineering of the sensor

The sigmoidal kinetic response of GK to glucose concentration is the biochemical basis of its sensor function. Early GKAs acted as “full” or “super” agonists, dramatically increasing the maximum velocity (Vmax) and/or drastically reducing the glucose concentration required for half-maximal activity (S0.5), effectively linearizing the activity curve.²³ Structurally, these full agonists locked GK into a “super-active” closed conformation regardless of glucose presence, abolishing the threshold sensitivity and leading to inappropriate activation at normoglycemic levels.¹⁹

In contrast, dorzagliatin functions as a partial agonist with high positive cooperativity. Crystallographic studies reveal that dorzagliatin binds to a discrete allosteric pocket located near the enzyme’s hinge region, approximately 20 Å from the glucose-binding site.¹⁹ Unlike full agonists that lock GK into a catalytically active “closed” conformation, causing constitutive activation, dorzagliatin stabilizes the transition state without freezing the enzyme’s dynamics. It effectively lowers the energy barrier for glucose binding while preserving the essential “super open” to “closed” conformational cycle.^{19 24} This profile ensures that the drug’s effect is amplified by rising glucose concentrations and diminishes as glucose falls, creating an intrinsic gluco-regulatory safety switch (figure 1A,B).

Figure 1. Structural basis and dual-organ mechanism of dorzagliatin. (A) Structural precision: Crystal structure of GK. Dorzagliatin binds to an allosteric pocket ~20 Å from the glucose-binding site. Unlike full agonists that lock GK in an active conformation, dorzagliatin stabilizes the transition state while preserving enzyme dynamics. (B) Kinetic safety window: Comparison of GK activity kinetics. Black dotted line: physiological response. Red line: first-generation full agonists, which linearize activity and cause hypoglycemia. Green curve: dorzagliatin, a glucose-dependent partial agonist that retains sigmoidal kinetics (Hill coefficient >1), enhancing activity primarily in the hyperglycemic range while maintaining a safety window at normoglycemia. (C) Dual-organ mechanism: In pancreatic β cells, dorzagliatin enhances glucose-stimulated insulin secretion by promoting glucose-driven ATP generation and subsequent insulin granule release. In hepatocytes, dorzagliatin preserves GK-GKRP interaction and GK nuclear-cytoplasmic shuttling, allowing nuclear sequestration of GK during fasting. This maintains the physiological restraint on hepatic glucose flux and de novo lipogenesis, thereby potentially mitigating the risk of hepatic steatosis associated with earlier glucokinase activators. GK, glucokinase; GKRP, GK regulatory protein.

Dual-organ pharmacology: restoring coordinated physiology

Pancreatic β-cell action

In the β-cell, dorzagliatin’s primary effect is to lower the threshold for GSIS.²⁵ By allosterically activating GK, dorzagliatin enhances the rate-limiting step of glycolysis, increasing ATP generation, closing ATP-sensitive potassium channels, depolarizing the membrane, and ultimately promoting calcium-dependent insulin granule exocytosis. This action restores the physiological coupling between ambient glucose concentration and insulin secretion, which is impaired in T2DM. Clinical studies have shown that dorzagliatin treatment improves β-cell function indices, including the Disposition Index (DI) and early-phase insulin secretion (IGI30).^{26 27}

Effects on pancreatic α cells and enteroendocrine L cells

Beyond β cells, GK activation in pancreatic α cells plays a pivotal role in suppressing inappropriate glucagon secretion during hyperglycemia, further refining the glucose-sensing network.²⁵ In the gut, activation in intestinal L cells may potentiate glucose-stimulated GLP-1 secretion.²⁸ These extra-β-cell actions may contribute to the overall glucose-lowering effect of dorzagliatin, although their relative clinical significance remains to be quantified.

Hepatic action

In the liver, dorzagliatin enhances the flux of glucose into glycolysis and glycogen synthesis pathways while simultaneously suppressing gluconeogenic gene expression (eg, glucose-6-phosphatase).^{19 21} This dual action—promoting glucose utilization and inhibiting glucose production—effectively reduces hepatic glucose output, contributing significantly to both fasting and postprandial glycemic control.²⁹

Extra-pancreatic actions

GK is also expressed in other glucoregulatory tissues. Beyond pancreatic α cells and intestinal L cells, GK is expressed in the central nervous system (CNS), where it plays a critical role in neuronal glucose sensing.³⁰ GK-expressing neurons are found in hypothalamic regions that regulate energy balance, feeding behavior, and the counter-regulatory response to hypoglycemia.³⁰ Genetic and pharmacological studies have shown that modulating GK activity specifically within the hypothalamic arcuate nucleus can alter insulin secretion and systemic glucose tolerance.³¹ Furthermore, GK-expressing neurons in the paraventricular thalamic nucleus form extensive connections with forebrain areas involved in reward and stress processing.³²

This CNS expression raises important considerations for systemic GK activators. On one hand, a recent Mendelian randomization study suggested that genetically mimicked GK activation may be associated with a reduced risk of cognitive decline, and low-dose dorzagliatin treatment in a diabetic rat model prevented spatial memory impairment.³³ On the other hand, the brain-penetrant properties of certain GKAs have raised theoretical concerns about potential CNS effects, including the possibility of altered appetite regulation or impaired counter-regulatory responses to hypoglycemia.³⁴ Clinical data on earlier GKAs (AZD6370, AZD1656) suggest that the central counter-regulatory response to hypoglycemia remains preserved,³⁴ but whether these findings generalize to dorzagliatin—and whether chronic GK activation in the brain carries any long-term safety risks—remains unknown.

Given that GK exists as neuroendocrine and hepatic isoforms with identical catalytic properties,³¹ systemically administered dorzagliatin would be expected to activate CNS GK. However, no dedicated clinical studies have examined the central effects of dorzagliatin. This represents a significant knowledge gap. Until such data are available, the benefits and risks of CNS GK activation by systemic dorzagliatin remain speculative, and the clinical relevance of central GK activation for the drug’s overall efficacy and safety profile cannot be determined.

This synchronized, glucose-contingent action on the liver and pancreas effectively mimics and amplifies the physiological insulin response.²⁹ This distinct mechanism—targeting the “sensor” rather than the “effector” organs—suggests a theoretical complementarity with insulin sensitizers (like metformin) and incretin-based therapies, offering a rationale for future combination strategies^{9 35} (figure 1C).

Clinical development: from proof-of-concept to disease modification

The clinical development program for dorzagliatin has systematically evaluated its pharmacokinetic, pharmacodynamic, and clinical effects in humans, although several important limitations exist as discussed below. It is important to note that the vast majority of published clinical data, including the pivotal Phase III SEED and DAWN trials, have been generated in Asian (primarily Chinese) populations. The efficacy and safety profile of dorzagliatin in other ethnic groups remain to be established.

Phase I studies: establishing the safety and pharmacokinetic bedrock

Initial first-in-human studies in healthy volunteers (single ascending dose) and patients with T2DM (multiple ascending dose, MAD) established an excellent safety and tolerability profile.^{36 37} No drug-related serious adverse events were reported, and overall adverse event rates were comparable to placebo. Pharmacokinetic analysis revealed a half-life supportive of two times per day dosing.^{36 37} Critically, pharmacodynamic data from the MAD study provided the first human proof-of-concept, demonstrating a dose-dependent reduction in 24-hour weighted mean glucose and glucose area under the curve, confirming its all-day glycemic control potential.^{37 38}

Phase II dose-finding study: efficacy proof and the emergence of β-cell benefit

This pivotal 12-week, randomized, double-blind, placebo-controlled trial in 258 drug-naïve or wash-out T2DM patients provided key evidence for the efficacy of dorzagliatin.²⁶ It identified 75 mg two times per day as the optimal dose, producing a robust placebo-adjusted reduction in glycated hemoglobin (HbA1c) of −1.07%. Importantly, the incidence of hypoglycemia was low and statistically similar to placebo, validating its glucose-dependent safety mechanism.

Beyond glycemic efficacy, this study yielded the first clear clinical signal of dorzagliatin’s disease-modifying potential. It demonstrated significant improvements in β-cell function, quantified by a marked increase in the DI—a composite measure of β-cell function adjusted for insulin sensitivity. Remarkably, this improvement in DI persisted for 1 week after drug discontinuation, suggesting a sustained, positive impact on β-cell physiology that outlasted the drug’s pharmacokinetic presence. This early finding laid the groundwork for the “remission” hypothesis, suggesting a temporary pharmacological intervention could have lasting metabolic benefits.²⁶

Phase III confirmatory trials (SEED and DAWN): consolidation and mechanistic insight

The SEED and DAWN trials, published in Nature Medicine, confirmed these findings in larger, distinct T2DM populations, solidifying dorzagliatin’s clinical profile.^{29 39} The SEED trial evaluated dorzagliatin 75 mg two times per day as monotherapy in 463 drug-naïve patients over 24 weeks.³⁹ It met its primary endpoint, with a placebo-adjusted HbA1c reduction of −1.02% (p<0.0001). The DAWN trial assessed dorzagliatin 75 mg two times per day as an add-on therapy in 767 patients inadequately controlled on metformin.²⁹ It also achieved its primary endpoint, providing an additional placebo-adjusted HbA1c reduction of −0.66% (p<0.0001).

Both trials consistently showed significant improvements in homeostatic model assessment 2 of β-cell function (HOMA2-β), reaffirming the β-cell function benefit. A prespecified pooled analysis of these studies provided deeper mechanistic insight: it revealed that the magnitude of improvement in IGI30 during dorzagliatin treatment was the strongest independent predictor of achieving HbA1c <7.0% at week 24.²⁷ This analysis powerfully connects dorzagliatin’s core pharmacological action (restoring early insulin secretion) directly to its ultimate clinical outcome (glycemic control), building a compelling causal evidence chain that distinguishes it from mere secretagogues. The key efficacy and safety outcomes from these two pivotal trials are summarized in table 1.

Table 1. Summary of key efficacy and safety outcomes from dorzagliatin Phase III trials.

Feature	SEED study (monotherapy)	DAWN study (add-on metformin)
Study design	Randomized, double-blind, placebo-controlled	Randomized, double-blind, placebo-controlled
Patient population	Drug-naïve T2DM	T2DM inadequately controlled on metformin
Sample size (N)	463	767
Duration	24 weeks	24 weeks
Primary endpoint (HbA1c)	−1.02% (vs placebo, p<0.0001)	−0.66% (vs placebo, p<0.0001)
β-cell function (HOMA2-β)	Significant improvement	Significant improvement
2 h-PPG reduction	−2.83 mmol/L	−2.48 mmol/L
Hypoglycemia	<1%	<1%
Lipid profile	Neutral or no significant triglyceride elevation	Neutral or no significant triglyceride elevation
Body weight change	Neutral (vs placebo, p>0.05)	Neutral (vs placebo, p>0.05)
Key conclusion	Effective glycemic control with sensor repair potential	Synergistic effect with metformin; safe combination

P values are versus placebo. Hypoglycemia defined as plasma glucose ≤3.9 mmol/L.

“Sensor repair potential” refers to improvements in β-cell function indices (HOMA2-β, IGI30) observed during treatment.

HbA1c, glycated hemoglobin; HOMA2-β, homeostatic model assessment 2 of β-cell function; 2h-PPG, 2-hour postprandial glucose; T2DM, type 2 diabetes.

The DREAM study: sustained glycemic maintenance versus clinical remission

An interesting observation emerged from the DREAM study, a prospective observational extension of the SEED trial.⁴⁰ Among patients who achieved good glycemic control on dorzagliatin monotherapy and then voluntarily stopped all glucose-lowering medication, a remarkable 65.2% maintained HbA1c <7.0% for 52 weeks post-treatment. However, due to the non-randomized, open-label design and selection bias, these findings require confirmation in randomized withdrawal trials before any definitive conclusions can be drawn.

However, a rigorous reappraisal of these data is mandatory. According to the 2021 American Diabetes Association (ADA)/European Association for the Study of Diabetes (EASD) consensus, “remission” is defined as HbA1c <6.5% for at least 3 months after discontinuing all glucose-lowering medications.⁴¹ The DREAM study, as a non-randomized, open-label observational extension, primarily included “responders” from the SEED trial, introducing significant selection bias. Therefore, the observed 52-week drug-free period should be precisely characterized as “sustained glycemic maintenance” rather than definitive “disease remission”. The mechanisms underlying this sustained effect remain unknown.

It is important to note that dorzagliatin enhances glucose-stimulated insulin secretion, which could theoretically increase, rather than decrease, β-cell workload and ER stress.⁴² Some studies have proposed that chronic GK activation might paradoxically accelerate β-cell failure through metabolic overstimulation.⁴³ Thus, the sustained glycemic maintenance after dorzagliatin withdrawal is an unexplained observation that requires further mechanistic investigation.

Future randomized, double-blind, placebo-controlled withdrawal trials are the gold standard required to confirm whether dorzagliatin can modify the natural history of T2DM. Subsequent analysis identified that the improvement in β-cell function (specifically the DI) achieved during the active treatment phase was the key predictor of this outcome.⁴⁴ However, this finding is observational and requires confirmation in dedicated, randomized withdrawal trials.

Precision medicine: identifying the ideal responders

The clinical efficacy of dorzagliatin exhibits notable inter-individual heterogeneity, highlighting the necessity for a precision medicine framework. Retrospective analyses suggest that patients with higher baseline HOMA2-β levels and a more preserved IGI30 derive the most significant benefit,⁴⁵ indicating that a minimum threshold of functional β-cell mass is required for “sensor repair” to be effective. Furthermore, the genetic landscape of the GK and GKRP genes may serve as predictive biomarkers. For instance, patients with specific GK polymorphisms that mimic the kinetic profile of the maturity-onset diabetes of the young, type 2 (MODY2) may represent a “sweet spot” for GKA therapy.⁴⁶ Identifying these genomic and biochemical signatures will be crucial to move beyond a “one-size-fits-all” approach and position dorzagliatin as a targeted therapy for the right patient at the right metabolic stage (figure 2).

Figure 2. Precision medicine landscape: phenotypic stratification and the “Sweet Spot” for dorzagliatin therapy. Conceptual mapping of T2DM heterogeneity along two axes: insulin resistance (x-axis) and β-cell function (y-axis). Three patient clusters are highlighted: (1) Obese cluster: high insulin resistance (left), typically managed with GLP-1 receptor agonists. (2) Dorzagliatin “Sweet Spot”: central region representing the “non-obese phenotype”—ideal candidates (“super-responders”) with predominant defects in glucose sensing and early-phase insulin secretion, rather than severe insulin resistance. (3) Genomic biomarkers: DNA helix denotes the role of GK polymorphisms in identifying patients with “MODY2-like” kinetics, enabling genotype-guided GKA therapy. Overall, this framework supports a shift from “one-size-fits-all” treatment toward mechanism-based precision selection. GK/GCK, glucokinase; GKA; glucokinase activator; GLP-1, glucokinase activator; HOMA-B, homeostatic model assessment of β-cell function; HOMA-IR, homeostatic model assessment of insulin resistance; MODY2, maturity-onset diabetes of the young, type 2; T2DM, type 2 diabetes.

Safety profile and risk-benefit assessment: a favorable differentiation

Dorzagliatin’s safety profile is a direct clinical manifestation of its glucose-dependent pharmacology, distinguishing it clearly from both earlier GKAs and other insulin secretagogues.

Hypoglycemia

This is the defining safety advantage. Across clinical trials, the incidence of documented hypoglycemia (plasma glucose ≤3.9 mmol/L) with dorzagliatin was low and comparable to placebo,^{26 36–39} starkly contrasting with the significant risk associated with sulfonylureas.

Lipid metabolism

Early GKAs like MK-0941 were plagued by dose-dependent hypertriglyceridemia.²⁰ Dorzagliatin has demonstrated a neutral to modestly favorable effect on lipid profiles in clinical trials,^{26 36–39} with no consistent signal of significant triglyceride elevation.

Body weight

Dorzagliatin exhibits a weight-neutral profile,^{26 27 39–42} a favorable characteristic compared with agents associated with weight gain (eg, sulfonylureas, thiazolidinediones, insulin).

The critical absence of cardiovascular outcomes trial data and long-term hepatic safety

Despite its promising glycemic efficacy, dorzagliatin faces two major hurdles in the current treatment paradigm.

First, and most fundamentally, no cardiovascular outcomes trial (CVOT) has been completed or is currently registered for dorzagliatin. The lack of CVOT data severely limits its recommendation for patients with established atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease.⁴⁷ Current guidelines prioritize agents with proven cardiovascular benefit—such as SGLT2 inhibitors and GLP-1 RAs—for these high-risk populations, as these drugs have demonstrated clear survival benefits.^{3 48} Until CVOT data become available, dorzagliatin cannot be recommended as an alternative to guideline-directed therapies in patients with high cardiovascular risk, and its use should be considered only in individuals without such risk factors.

Second, while short-term (24-week) data suggest lipid neutrality, the long-term metabolic consequence of hepatic GK activation requires scrutiny. Crucially, the safety differentiation of dorzagliatin likely hinges on its preservation of the GK-GKRP axis.⁴⁹ Structural biology indicates that dorzagliatin, unlike early “full” agonists (eg, MK-0941), does not disrupt the inhibitory binding of GKRP to GK.¹⁹ This implies that during fasting or hypoglycemia, GK can still be sequestered in the nucleus by GKRP. By maintaining this physiological nuclear-cytoplasmic shuttling, dorzagliatin retains the intrinsic molecular “brake” on hepatic de novo lipogenesis, thereby mitigating the risk of steatosis.⁵⁰ However, theoretical concerns remain: chronic enhancement of glucose flux could still promote hepatic steatosis in susceptible individuals. Therefore, while the preservation of the GK-GKRP axis provides a mechanistic safeguard, future studies using proton density fat fraction (MRI-PDFF) are essential to definitively exclude the risk of non-alcoholic fatty liver disease progression over multiyear treatment periods.

Expanding the therapeutic frontier: GKA in type 1 diabetes and monogenic diabetes

Rationale in T1DM: targeting the residual hepatic-intestinal axis

The therapeutic rationale in T1DM relies primarily on the extra-pancreatic axis. However, a critical distinction must be made between dual-acting agents like dorzagliatin and hepatoselective GKAs like TTP399.⁵¹ In the absence of functional β cells, the pancreatic GK activation by dorzagliatin may be redundant, whereas its potential to suppress glucagon via α cells²⁵ and enhance hepatic glucose uptake⁵² remains highly relevant.

However, the application of a dual-acting GKA in T1DM warrants a dialectical caution: while enhancing residual β-cell function (in those with detectable C-peptide) is desirable, pharmacological overstimulation in a stressed, autoimmune-targeted environment could theoretically exacerbate ER stress and accelerate β-cell exhaustion.⁵³ This potential “double-edged sword” necessitates careful dose-titration studies to balance functional rescue against metabolic stress. The risk of hypoglycemia in T1DM remains the primary concern for any systemic GKA⁵⁴; thus, the “safety switch” of glucose-dependency must be even more robustly validated in the context of exogenous insulin therapy.

Proof-of-principle from hepatoselective GKA: the TTP399 precedent

The hepatoselective GKA TTP399 provides the most compelling clinical validation of this concept. Granted US Food and Drug Administration Breakthrough Therapy designation for adjunctive T1DM treatment, TTP399’s success in the SimpliciT-1 Phase II trial was landmark. In adults with T1DM on insulin therapy, adjunctive TTP399 (800 mg/day) for 12 weeks resulted in: An additional HbA1c reduction of ~0.5% versus placebo. A reduction in total daily insulin dose by ~8.4%. A 40% reduction in symptomatic hypoglycemic events.⁵⁵

Notably, contrary to theoretical concerns about increased ketogenesis, TTP399 did not increase the risk of diabetic ketoacidosis (DKA). A dedicated insulin withdrawal study suggested it might even mitigate DKA risk, with none of the TTP399-treated participants meeting prespecified DKA criteria compared with 42% of placebo-treated participants.⁵⁶ The proposed mechanism involves enhanced postprandial hepatic glycogen synthesis, creating a “glycogen buffer” that can be mobilized to counter insulin-induced hypoglycemia, while increased hepatic glucose flux may reduce the availability of substrates for ketone body production.⁵⁵

Dorzagliatin in T1DM: the “α-cell paradox” and safety concerns

The promising TTP399 data validate the hepatic GK activation strategy but cannot be directly extrapolated to support dorzagliatin’s use in T1DM. The critical distinction lies in tissue selectivity. TTP399 is a hepatoselective activator, while dorzagliatin is a dual-acting (pancreatic+hepatic) activator.

In the context of T1DM, dorzagliatin’s action on residual pancreatic islet cells introduces theoretical risks. Specifically, the “α-cell paradox” must be considered. Pancreatic α cells also express GK, where it serves to suppress glucagon secretion at high glucose levels.⁹ However, in T1DM, α-cell signaling is often dysregulated due to the loss of paracrine insulin signaling.⁹ Pharmacological activation of GK in α cells could theoretically lead to unpredictable glucagon modulation. Furthermore, stimulating dysfunctional residual β cells might exacerbate ER stress or accelerate immune-mediated destruction.

Therefore, unlike hepatoselective agents which bypass the islet entirely, dorzagliatin’s dual mechanism poses unique challenges in T1DM. Current evidence does not support the widespread use of dorzagliatin in the general T1DM population. Future investigations should prioritize not the broad T1DM population, but carefully defined subgroups with preserved C-peptide levels (eg, latent autoimmune diabetes in adults), where the risk-benefit ratio may be more favorable.⁴⁶ Any such studies must begin with rigorously designed dose-finding trials to balance potential functional rescue against the theoretical risks of ER stress acceleration and immune-mediated β-cell destruction.

Application in GK-MODY: a paradigm of mechanism-driven precision therapy

In contrast to the uncertainty in T1DM, dorzagliatin has demonstrated clear efficacy in GK-MODY, a monogenic form of diabetes caused by heterozygous inactivating mutations in the GK gene.⁵⁷ In these patients, dorzagliatin enhances insulin secretion and significantly improves glucose sensitivity, effectively compensating for the reduced intrinsic activity of the mutant GK enzyme.⁴⁶ This serves as a prime example of mechanism-based, targeted therapy, confirming dorzagliatin’s fundamental action as a GK activator in humans and offering a personalized treatment option for this specific genetic subgroup.^{57 58}

However, it is crucial to recognize that the vast majority of individuals with GK-MODY have a benign, asymptomatic, and non-progressive phenotype characterized by mild fasting hyperglycemia. Such patients rarely develop typical microvascular or macrovascular complications of diabetes and generally do not require pharmacological intervention. Therefore, dorzagliatin should not be routinely offered to all GK-MODY patients; rather, its use should be reserved for the uncommon cases with more pronounced hyperglycemia or specific complications where treatment is deemed necessary.

Homozygous or compound heterozygous loss-of-function mutations in GK cause a much rarer and more severe phenotype—permanent neonatal diabetes mellitus—which requires immediate insulin therapy from birth. The potential role of dorzagliatin in this condition, if any, remains unexplored.

Future directions: synergistic combinations and precision medicine

To maximize the clinical value of dorzagliatin, future research must move beyond monotherapy efficacy and focus on mechanistic synergy with standard-of-care agents and the adoption of modern glycemic metrics.

Therapeutic positioning in the era of GLP-1 RAs and SGLT2 inhibitors: complementary or competitive?

The remarkable success of GLP-1 RAs and SGLT2 inhibitors—driven by proven cardiorenal benefits and weight reduction—has fundamentally reshaped the T2DM treatment landscape.³ This raises a critical question: where does dorzagliatin fit? Rather than competing directly with these agents, dorzagliatin occupies a mechanistically distinct and potentially synergistic niche. While GLP-1 RAs amplify insulin secretion in a glucose-dependent manner, their efficacy remains contingent on the β cell’s intrinsic glucose-sensing capacity—precisely the defect that dorzagliatin addresses. Similarly, SGLT2 inhibitors’ paradoxical increase in endogenous glucose production (EGP) represents a mechanistic gap that hepatic GK activation can directly fill. Thus, dorzagliatin should not be positioned as an alternative to guideline-directed therapies, but rather as a foundational “sensor calibrator” that could be combined with these agents to achieve complementary physiological restoration.⁴⁷ Head-to-head trials comparing dorzagliatin monotherapy with low-dose GLP-1 RAs or SGLT2 inhibitors in early-stage, non-obese patients with T2DM would help clarify its first-line potential in specific phenotypes.

The “twin-engine” strategy: synergizing with SGLT2 inhibitors

Current guidelines strongly favor SGLT2 inhibitors for their cardiorenal benefits. However, SGLT2 inhibitors treatment paradoxically increases EGP by elevating glucagon levels, which can attenuate their glucose-lowering potency.⁵⁹ Dorzagliatin offers a precise mechanistic countermeasure: by activating hepatic GK and potentially inhibiting α-cell glucagon secretion, it suppresses EGP.⁹ Theoretically, this combination targets all three key organs—the kidney (glycosuria), the liver (suppressed EGP), and the pancreas (insulin/glucagon balance)—providing a robust, “twin-engine” approach to glycemic control without increasing hypoglycemia risk.

Amplifying the incretin axis: combination with GLP-1 RAs

While GLP-1 RAs are potent insulin secretagogues, their efficacy relies on the β cell’s responsiveness to glucose. T2DM acts as a state of “sensor blindness,” where β cells fail to efficiently recognize glucose thresholds. Dorzagliatin acts as a “sensor sensitizer” effectively lowering the threshold for GSIS, which may synergistically amplify the insulinotropic effects of GLP-1 RAs. Furthermore, since GLP-1 RAs primarily address postprandial glucose and weight, while dorzagliatin effectively targets fasting plasma glucose via hepatic regulation, their combination could offer comprehensive 24-hour glycemic coverage.

Beyond HbA1c: targeting glycemic variability and time in range

In the era of continuous glucose monitoring, the goal of diabetes therapy has shifted from merely lowering HbA1c to improving the “quality” of glycemic control—specifically, reducing glycemic variability and increasing time in range (TIR).⁶⁰ Since GK acts as the physiological “thermostat,” repairing its function should theoretically result in a more stable glucose profile compared with sulfonylureas, which drive secretion independently of glucose. Future trials should prioritize TIR as a primary endpoint to demonstrate dorzagliatin’s superiority in flattening glucose excursions.

Precision medicine: identifying the “super-responders”

A “one-size-fits-all” approach dilutes the efficacy signal of distinct mechanisms. Future research should prioritize:

Clinical phenotyping

Patients with “early-phase secretory defects” but preserved basal insulin are likely ideal candidates.

Genotypic stratification

Patients with specific GK polymorphisms or mutations (eg, GK-MODY) represent a structurally defined target population where dorzagliatin could serve as a targeted “correction therapy”.^{58 61}

Global access and health economics: the unspoken determinant of clinical impact

Beyond efficacy and safety, the real-world impact of dorzagliatin will be profoundly shaped by health economics and regulatory positioning. As an oral agent with a potentially disease-modifying profile, dorzagliatin could offer significant advantages in healthcare settings where access to injectable therapies (GLP-1 RAs) or expensive agents (SGLT2 inhibitors) is limited. Its projected cost-effectiveness relative to generic medications (eg, metformin, sulfonylureas) will determine its adoption in low- and middle-income countries, where the majority of the global diabetes burden resides. Furthermore, its potential to reduce long-term complications through sustained glycemic maintenance—if confirmed—could justify premium pricing in value-based reimbursement models. However, pricing strategies and health technology assessments have not been publicly discussed, representing a critical gap in the literature. Future studies should model the long-term cost-effectiveness of dorzagliatin-based strategies, particularly in combination with SGLT2 inhibitors, to inform guideline recommendations across diverse economic settings.

Unresolved questions and inherent limitations of dorzagliatin

A critical review must explicitly acknowledge the shortcomings and unknowns. First, as noted above, no CVOT data exist. Second, the generalizability of current findings is limited because nearly all published clinical trials were conducted in Asian (primarily Chinese) populations. Whether the observed efficacy and safety profiles apply to Caucasian, African, or other ethnic groups is unknown. Third, the sustained drug-free glycemic maintenance observed in the DREAM study does not meet the formal ADA/EASD definition of diabetes remission and is subject to selection bias. Fourth, long-term effects on β-cell mass, hepatic steatosis (via de novo lipogenesis), and cognitive function remain uncharacterized. Fifth, and most fundamentally, two serious theoretical concerns about chronic GK activation have been raised in the literature and require careful consideration.

Potential for β-cell exhaustion from chronic GK activation

A recent and provocative hypothesis challenges the conventional view that hyperglycemia itself drives β-cell failure. Instead, it proposes that hyperactivation of glucose metabolism—the very process that GK activation enhances—may be the primary cause of progressive β-cell dysfunction in diabetes.⁴³ According to this model, chronic overstimulation of glucose flux leads to mitochondrial oxidative stress, ER stress, and ultimately β-cell dedifferentiation or apoptosis. If this hypothesis is correct, then conceptually, glucokinase inhibitors might be a better therapeutic strategy than GK activators for preserving long-term β-cell function in T2DM.⁴³

Short-term clinical trials of dorzagliatin²⁷ have shown improvements in β-cell function indices (eg, DI, HOMA2-β), which appear to support the opposite conclusion. However, these surrogate measures do not directly assess β-cell mass or long-term viability. The critical question—whether continuous GK activation over years leads to eventual β-cell exhaustion—remains unanswered. No long-term (≥2 years) study has been published for dorzagliatin or any other GKA. Until such data are available, the possibility that chronic GK activation may paradoxically accelerate β-cell decline cannot be dismissed.

This theoretical concern is supported by direct evidence from a mouse model of congenital hyperinsulinism caused by an activating GK mutation.⁶² In this model, sustained β-cell-specific GK activation initially increased β-cell proliferation and mass (a 62% increase at 1.5 months), but this was followed by a progressive decline, with decreased absolute β-cell mass and impaired glucose tolerance by 8 months of age. The observed increase in DNA damage response and Chop expression suggested that metabolic stress contributed to β-cell death.⁶² These findings demonstrate that chronic GK activation can be beneficial only transiently and that increased β-cell workload ultimately leads to β-cell failure—a critical observation that has not yet been adequately addressed in the clinical development of dorzagliatin.

This hypothesis has now received direct experimental support from complementary human and rodent studies. Lloyd and colleagues showed that chronic exposure of human islet microtissues and INS-1 cells to GK activators (dorzagliatin or MK-0941) impaired insulin secretion and metabolism to the same extent as chronic hyperglycemia alone. Conversely, pharmacological inhibition of GK with mannoheptulose prevented and partially reversed these deleterious effects.⁶³ In a separate study, overnight treatment of db/db islets with mannoheptulose enhanced insulin secretion and restored pulsatile release, paradoxically increasing ATP content and reduced nicotinamide adenine dinucleotide (phosphate) flux.⁴³ These findings converge on the same sobering conclusion: chronic GK overstimulation, whether genetic, pharmacological, or metabolic, may ultimately accelerate β-cell decompensation. This raises the possibility that the short-term improvements in β-cell function indices observed with dorzagliatin may not translate into long-term preservation of β-cell mass or function.

Potential adverse effects on cognitive function

GK is expressed not only in pancreatic islets and liver but also in the CNS, where it plays a role in neuronal glucose sensing (see Dual-organ pharmacology: restoring coordinated physiology). Emerging evidence has raised concerns that systemic GK activation could have unintended effects on cognitive function. A recent study reported that GK activation was associated with detrimental effects on cognitive performance in preclinical models, raising the possibility that chronic treatment with a brain-penetrant GKA could impair memory or executive function.⁶⁴

The relevance of these findings to dorzagliatin is currently unknown. Dorzagliatin’s ability to cross the blood-brain barrier has not been systematically studied, and no clinical trial to date has included cognitive function as a prespecified safety endpoint. This represents a significant knowledge gap. Given the growing recognition of the link between diabetes, hypoglycemia, and cognitive decline, the potential for GKAs to adversely affect cognition—whether directly via CNS GK activation or indirectly via altered glucose metabolism—warrants dedicated investigation in future trials. Until such studies are performed, the cognitive safety profile of dorzagliatin remains undefined.

These concerns, together with the absence of CVOT data and limited ethnic generalizability, underscore that dorzagliatin’s risk-benefit balance is not yet fully established. Long-term, adequately powered studies are urgently needed.

Conclusion

Dorzagliatin represents a notable step forward in the evolution of GKAs, with a pharmacologic profile that differs from earlier, failed compounds. However, several major uncertainties remain. While it offers a unique mechanism to improve β-cell function, its long-term positioning will depend on validating its safety regarding hepatic lipid metabolism and demonstrating hard clinical outcomes. Ultimately, its greatest potential may lie not as a monotherapy competitor, but as a mechanistic partner to SGLT2 inhibitors and GLP-1 RAs, forming a comprehensive strategy for physiological glucose homeostasis. Crucially, the absence of CVOT data and the lack of long-term safety evidence on β-cell function, hepatic steatosis, and cognitive outcomes preclude any definitive recommendation for its widespread use. Future research must prioritize these areas. Furthermore, the predominance of data from Asian populations limits the generalizability of current evidence.

Data availability statement

No data are available.