Effects of Dorzagliatin, a Glucokinase Activator, on α- and β-Cell Function in Individuals With Impaired and Normal Glucose Tolerance

Abstract

Dorzagliatin is a dual-acting allosteric activator of glucokinase (GCK). Dorzagliatin improved second-phase insulin secretion in individuals with type 2 diabetes and heterozygous carriers of GCK mutations. We investigated the effects of dorzagliatin on pancreatic insulin, glucagon, and glucagon-like-peptide 1 (GLP-1) secretion in individuals with impaired glucose tolerance (IGT) and normal glucose tolerance (NGT). In a double-blind, randomized, crossover, single-dose study, 9 participants with IGT and 10 with NGT underwent 2-h 12 mmol/L hyperglycemic clamp following a single dose of dorzagliatin 50 mg or matched placebo. Plasma insulin, C-peptide, glucagon, and total GLP-1 levels were measured at regular intervals. There were no differences in first-phase insulin after the dorzagliatin dose in either group. Dorzagliatin significantly increased second-phase insulin secretion rate and β-cell glucose sensitivity by 1.3-fold compared with placebo in IGT but remained similar in NGT. Dorzagliatin increased basal plasma insulin in the NGT group only. Glucagon (area under the curve_{0–120 min} = 161 ± 58 vs. 234 ± 70 pmol*min/L [mean ± SD]; P = 0.01) was suppressed after dorzagliatin in the NGT group but not the IGT group. Plasma glucagon was positively correlated with total GLP-1 levels. Dorzagliatin did not affect insulin sensitivity in either group. Dorzagliatin has different actions on β- and α-cells depending on glucose tolerance, increasing second-phase insulin secretion in IGT while enhancing glucose-suppression of glucagon secretion in NGT.

Article Highlights

• Dorzagliatin is a dual-acting allosteric glucokinase (GCK) activator that increases β-cell glucose sensitivity and second-phase insulin in GCK monogenic diabetes of the young.

• Actions of dorzagliatin on α- and β-cell function in normal and impaired glucose tolerance are unknown.

• In this study, dorzagliatin increased second-phase insulin in individuals with impaired glucose tolerance while suppressing glucagon in participants with normal glucose tolerance during a hyperglycemic clamp.

• With increasing glucose, plasma glucagon and total GLP-1 levels declined progressively. A modest to moderate positive correlation between glucagon and total GLP-1 was observed under both dorzagliatin and placebo treatments.

Introduction

Glucokinase (GCK), a critical enzyme in glucose homeostasis, is prominently expressed in the pancreas and liver, and also is detected in other glucose-sensing cells (1,2). It catalyzes the phosphorylation of glucose to glucose-6-phosphate, acting as a central rate-limiting enzyme in glucose metabolism. In the liver, GCK is pivotal for orchestrating glycogenesis and glycogenolysis essential for glucose storage and release (3). In pancreatic β-cells, GCK serves as a critical glucose sensor, essential for glucose-stimulated insulin secretion (GSIS). Additionally, results of several animal and in vitro studies suggested a direct role for α-cell GCK sensing in regulating glucose suppression of glucagon secretion (GSGS) (2,4,5).

Dorzagliatin is a novel, first-in-class, dual-acting allosteric glucokinase activator (GKA). It directly binds to a pocket distal to the active site of GCK, increases its affinity for glucose, and lowers the set point for GSIS in pancreatic β-cells (6,7). In type 2 diabetes (T2D), dorzagliatin reduced fasting and postprandial glucose levels in a dose-dependent manner (8,9). In phase 3 randomized controlled trials, dorzagliatin treatment led to a −0.57% and −0.66% placebo-subtracted reduction in HbA_1c in drug-naïve and metformin-treated patients with T2D, respectively, with sustained efficacy at 52 weeks (10,11). Furthermore, in an observational extension study, 65% of patients with early-stage T2D who received dorzagliatin monotherapy had sustained drug-free remission at 52-week follow-up (8). These findings suggest dorzagliatin may have a role in restoration of β-cell function in those with earlier stages of diabetes.

In patients with GCK monogenic diabetes of the young (GCK-MODY), we showed that a single dose of dorzagliatin significantly enhanced the second-phase insulin secretory response (ISR) and improved β-cell glucose sensitivity (β-CGS) by nearly twofold during a 2-h hyperglycemic clamp, compared with placebo (12). In participants with recent-onset T2D, dorzagliatin increased basal and second-phase insulin secretion, albeit to a lesser extent than in GCK-MODY with paradoxical reduced first-phase insulin secretion. Dorzagliatin did not affect glucagon secretion in T2D or GCK-MODY. Based on these findings, we hypothesized that altered β-cell bioenergetics in the presence of wild-type GCK might account for the attenuated response to dorzagliatin in T2D (12).

In this proof-of-concept study, we aimed to investigate the effects of dorzagliatin in individuals with normal glucose tolerance (NGT) and impaired glucose tolerance (IGT) to explore its effect in the full spectrum of glycemia. The primary objective was to evaluate the impact of dorzagliatin on β-cell function, based on the hypothesis that it may increase β-cell function across the range of glucose intolerance. Our secondary aims were to explore dorzagliatin’s potential to enhance GSGS via pancreatic α-cell GCK, considering the importance of glucagon nonsuppression in the etiology of T2D (4,13). Because GLP-1 is produced within α-cells of the pancreatic islets, we also assessed pancreatic GLP-1 secretion in response to intravenous glucose stimulation and the effects of dorzagliatin (14). In this double-blind, placebo-controlled, crossover study, we evaluated the effects of a single dose of dorzagliatin on β- and α-cell function during a 2-h hyperglycemic clamp in individuals with IGT and in participants with NGT.

Research Design and Methods

Study Design

This was a double-blind, randomized, crossover study in 10 participants with IGT and 10 with NGT. All participants received a single oral dose of dorzagliatin 50 mg or matched placebo followed by a 2-h hyperglycemic clamp on two occasions separated by a 2-week washout period (Supplementary Fig. 1). The study was approved by the Joint New Territories East Cluster–The Chinese University of Hong Kong Clinical Research Ethics Committee, Lui Che Woo Clinical Sciences Building, Prince of Wales Hospital Shatin, Hong Kong (approval no. 2021.516–T), and is registered with ClinicalTrials.gov (identifier NCT05468229). All participants provided written informed consent. The study was conducted at the Phase 1 Clinical Trial Centre, The Chinese University of Hong Kong, Prince of Wales Hospital. Participants were recruited between 20 October 2022 and 1 December 2023, and clinical data were collected between 27 October 2022 and 5 December 2023.

Study Participants

Eligible participants were adults aged 18–65 years with a BMI between 18 and 30 kg/m². In the IGT group, additional inclusion criteria included fasting plasma glucose (FPG) levels <7.0 mmol/L, HbA_1c <6.5%, and a 2-h plasma glucose (2h-PG) level between 7.8 and 11.1 mmol/L after a 75-g oral glucose tolerance test (OGTT), with no prior treatment with glucose-lowering medications, including traditional Chinese medicines for glycemic control. For the NGT group, inclusion criteria included FPG <5.6 mmol/L, HbA_1c <5.7%, and a 2-h PG <7.8 mmol/L after a 75-g OGTT, and no history of glucose-lowering drug treatment. Exclusion criteria applied to all participants and included body weight <45 kg, history of stroke or cardiovascular disease within 6 months of recruitment, severe renal dysfunction (estimated glomerular filtration rate <30 mL/min/1.73 m² or renal dialysis), severe hepatic dysfunction (aspartate transaminase and/or alanine transaminase greater than three times the upper limit of normal), history of drug abuse or excessive alcohol intake, diabetes mellitus, dehydration, diarrhea or vomiting at the time of recruitment, severe infection, anemia, pregnancy or lactating, recent blood loss >300 mL, severe hypoglycemia within 3 months prior to screening, and use of strong or moderate CYP3A4 inhibitors or inducers. The NGT and IGT groups were matched for BMI (±5 kg/m²) and sex.

Study Procedures

During the screening visit (visit 1), demographic data, medical and family history, and physical examination data were collected to determine eligibility. Measurement of FPG level and HbA_1c, complete blood testing, renal and liver function tests, lipid profile, and a urinary pregnancy test in women with reproductive potential were performed for assessment of eligibility. A venous blood sample was collected for genotyping of monogenic diabetes.

Study Treatment and Assignment

Each eligible participant was randomly assigned to receive either dorzagliatin 50 mg (HMS5552; Shanghai STA Pharmaceutical Product Co., Ltd.) or matching placebo (Shanghai STA Pharmaceutical Product Co., Ltd.) on two occasions, separated by 14 ± 2 days. A lower dose of dorzagliatin 50 mg was selected based on experience from phase 1 studies, which included healthy control participants with the potential risk of hypoglycemia during the fasting state (15). The treatment sequence was pregenerated randomly by a computer program by personnel independent of the study team. Each participant received the study drug from pack 1 on the first clamp visit (visit 2) and from pack 2 on the second clamp visit (visit 3). Investigators, study personnel, monitors, and participants were masked to the treatment sequence except for emergency unblinding. The treatment sequence was only unblinded after data lock by independent personnel after completion of biochemical analyses and calculations.

Hyperglycemic Clamp

Prior to each hyperglycemic clamp visit (visits 2 and 3), participants fasted for at least 8 h overnight and refrained from smoking, alcohol use, and vigorous exercise for 48 h. Participants received the study drug (50 mg of dorzagliatin or placebo) with 240 mL of water at −30 min before the commencement of hyperglycemic clamp (0 min). This post-dose, pre-clamp basal period was also shortened from 90 to 30 min compared with the previous study, due to potential risks of fasting hypoglycemia among individuals in the IGT and NGT groups (12). Venous blood for glucose, insulin, C-peptide, total GLP-1, and glucagon measurements was collected from a retrograde cannula inserted into the dominant hand or wrist vein placed in an insulated warm box with air heated to 40–50°C for arterialization of venous blood.

Blood samples were collected at −20, −10, and 0 min to define baseline glucose, insulin, C-peptide, and glucagon levels. At time 0, a primed dose of 20% dextrose was given over 15 min via an infusion pump to rapidly increase the level of the arterialized whole-blood glucose to a target of 12 mmol/L. Blood glucose was thereafter maintained at 12 mmol/L for 120 min by titration of dextrose infusion rates according to arterialized blood glucose level measured in duplicate every 5 min using a bedside glucose analyzer (catalog YSI 2300; Yellow Spring Instrument) (coefficient of variation [CV] <2%). Blood samples were obtained to measure insulin, C-peptide, and glucagon levels every 2 min in the first 10 min and thereafter at 20, 40, 80, 100, and 120 min to assess second-phase insulin response. Total GLP-1 level was measured at 0, 10, 20, 40, 80, and 120 min. At the end of the clamp, 20% dextrose infusion was gradually switched off over 30 min, followed by monitoring of blood glucose level for at least 30 min until return to euglycemic levels.

After 14 ± 2 days of washout, participants underwent a second hyperglycemic clamp (visit 3) with either dorzagliatin or placebo administration depending on sequence randomization. After each clamp study, all participants self-monitored blood glucose twice daily for 48 h for reporting of any hypoglycemic event, with phone follow-up (visit 4) 3–7 days after last dosing to ascertain any adverse events.

Biochemical Assays

Blood samples for hormonal assays were collected in chilled EDTA tubes and then centrifuged at 3,000 rpm for 10 min at 4°C. Insulin was measured spectrophotometrically using insulin ELISA kits (Mercodia AB, catalog 10–1113-01; intra-assay CV = 3.3%, interassay CV = 6.1%). C-peptide level was measured spectrophotometrically using C-peptide ELISA kits (Mercodia AB, catalog 10–1136-01; intra-assay CV = 3.2%, interassay CV = 4.8%). Glucagon level was measured spectrophotometrically using glucagon ELISA kits (Mercodia AB, catalog 10–1271-01; intra-assay CV = 8.3%, interassay CV = 9.2%). Total GLP-1 was measured using chemiluminescent immunoassay (total GLP-1 NL-ELISA kits; Mercodia AB, catalog 10–1278-01; intra-assay CV = 10.1%, interassay CV = 13.7%). Absorbance at 450 nm was read on a Multiskan FC Microplate Photometer (Thermo Fisher Scientific).

Genotyping

At the screening visit, 6 mL of peripheral blood sample was collected from each participant and DNA was extracted using a MagNA Pure 96 System (Roche Diagnostics). Variants of monogenic diabetes were identified by next-generation sequencing. A total of 34 genes (Supplementary Methods) related to monogenic diabetes were included in the custom targeted DNA panel. The sequencing regions covered exons and flanking regions located within 25 bp upstream and downstream of each exon. DNA libraries were prepared according to the reference guide of AmpliSeq for Illumina Custom and Community Panels. Pooled libraries were sequenced by the Illumina MiSeq instrument with paired-end run of 151 cycles with FASTQ files as raw sequencing output per read. Sequencing variants that passed quality control were interpreted according to the published guidelines (16–18).

Calculations

Basal glucose was determined as mean glucose levels (area under the curve [AUC]/time) from −20 to 0 min. Steady-state glucose during hyperglycemic clamp was determined as mean glucose levels (AUC/time) from 80 to 120 min. AUC was calculated using the trapezoid rule. Baseline insulin and C-peptide levels were calculated as the average values from −10 to 0 min. Acute (first-phase) insulin (AIRg) and C-peptide (ACPRg) responses to glucose were calculated as the mean incremental response above baseline from samples drawn in the first 10 min of hyperglycemic clamp. Second-phase insulin and C-peptide (SCPRg) responses were calculated as the mean incremental response in the last 40 min of the 2-h clamp. The maximum concentrations (C_max) and time to maximum for insulin and C-peptide were calculated respectively. The insulin sensitivity index (ISI) was calculated as the mean glucose infusion rate in the last 40 min of hyperglycemic clamp divided by average plasma insulin concentration in the same period (19).

Prehepatic ISR (pmol/min/m²) was determined by deconvolution of peripheral C-peptide concentrations using a two-compartment model of C-peptide kinetics and population-based C-peptide kinetic parameters (20). Basal insulin secretion (ISRb) was calculated as mean ISR (AUC/time) from −10 to 0 min. Absolute first-phase insulin secretion (ISR1abs) was calculated as mean ISR (AUC/time) between 0 and 8 min. Absolute second-phase insulin secretion (ISR2abs) was calculated as mean ISR (AUC/time) between 80 and 120 min. Incremental first-phase insulin secretion (ISR1inc) and incremental second-phase insulin secretion (ISR2inc) were defined as ISR1Abs minus ISRb and ISR2abs minus ISRb, respectively. β-CGS (slope of ISR vs. glucose) was calculated as ISR2inc divided by glucose change (steady-state glucose minus basal glucose) (21). The β-CGS ratio was determined by comparing the β-CGS under dorzagliatin treatment with that under placebo. Glucagon and GLP-1 AUCs were determined between 0 and 120 min.

The glucagon secretion rate (GSR) was calculated by applying nonparametric deconvolution to peripheral glucagon concentrations using a population-derived parameters of a glucagon kinetics model (22,23). Basal glucagon secretion (GSRb) was calculated as mean GSR (AUC/time) from −10 to 0 min. Absolute glucagon secretion between 0 and 15 min (GSR1abs), representing the initial 15-min glucagon secretion, was calculated as mean GSR (AUC/time) from 0 to 15 min. GSR2abs, representing the last 40-min glucagon secretion, was calculated as mean GSR (AUC/time) from 80 to 120 min. Incremental glucagon secretion were defined as absolute GSR minus GSRb in the same period, respectively. Glucagon suppression sensitivity indicates the responsiveness of glucagon secretion by α-cell to plasma glucose concentration fluctuations (RGSR) (21), calculated as the decremental GSR divided by incremental glucose in the same period, respectively.

Statistical Analysis and Sample Size Determination

Continuous variables were reported as the mean ± SD. Categorical variables were described as counts and percentages. Baseline group differences were compared by independent-samples t test or χ² test. A paired-samples t test was used to compare differences for normally distributed data and Wilcoxon signed-rank test for nonnormally distributed data between treatment conditions. As a proof-of-concept study, the primary end point was to evaluate effects of dorzagliatin versus placebo on first-phase insulin secretion in NGT and IGT populations. Secondary end points included effects of dorzagliatin on basal and second phase insulin secretion, insulin sensitivity, glucagon, and GLP-1 AUC in the NGT and IGT groups, respectively. As an exploratory analysis, the correlation estimates between glucagon and total GLP-1 were estimated using a validated repeated-measure mixed effects model proposed by Hamlett et al. (24), which incorporates a covariance matrix to explicitly define the correlation structure within a participant’s repeated measurements. The corresponding SAS code used to perform this analysis is provided in (Supplementary Methods).

We selected a sample size of 10 based on our previous study with similar design (12), which demonstrated significant difference in insulin secretion with dorzagliatin versus placebo. We expected similar or larger effect size of dorzagliatin versus placebo in the IGT and NGT groups. The efficacy population comprised all participants in receipt of the study drug who were compliant with inclusion and exclusion criteria, study drug administration, and completed all doses of the study and all sampling procedures. The safety population included all participants who received any amount of the study drug. We reported results in the NGT and IGT groups without formal statistical comparisons. We conducted exploratory subgroup analysis comparing those with isolated IGT and combined impaired fasting glucose (IFG; defined as FPG >5.6 mmol/L at screening) and IGT.

Adjustment for multiplicity was not applied in this proof-of-concept study. Statistical analyses were conducted using R software (version 4.3.1) and SAS software (version 9.4; SAS OnDemand for Academics, SAS Institute Inc.). A P value of <0.05 was considered statistically significant.

Data and Resource Availability

Deidentified data underlying the results reported in this article will be made available within 12 months and for 5 years after publication and upon reasonable written request to the corresponding author. No applicable resources were generated or analyzed during this study.

Results

Participant Characteristics

A total of 38 individuals were screened and 20 participants (n = 10 each in the IGT and NGT groups) met eligibility criteria and were enrolled in the study. One participant in the IGT group (P10) withdrew informed consent after the visit 2 clamp and was excluded from the efficacy analysis due to missing data from visit 3, but their data remained in the safety analysis.

Demographics of 9 IGT and 10 NGT participants included in the efficacy analysis are listed in Table 1. Both groups were similar in sex and BMI distribution; the NGT group had younger mean age. The NGT group had lower FPG (4.8 ± 0.3 vs. 5.2 ± 0.5 mmol/L) and 2h-PG (5.0 ± 0.8 vs. 8.9 ± 1.0 mmol/L) than the IGT group but similar mean HbA_1c. All enrolled participants were negative for mutations in 34 genes related to monogenic diabetes. Of the IGT participants, six had isolated IGT and three had combined IFG plus IGT.

Table 1.

Summary of clinical characteristics of study participants

Characteristic	IGT (n = 9)	NGT (n = 10)	P
Age (years)	55.0 ± 7.5	39.5 ± 11.8	0.01
Sex (female)	5 (55.6)	5 (50.0)	0.81
BMI (kg/m2)	23.9 ± 3.0	24.2 ± 3.6	0.84
HbA1c (%)	5.6 ± 0.6	5.4 ± 0.3	0.37
HbA1c (mmol/mol)	38.1 ± 6.5	36.0 ± 3.1	0.39
FPG (mmol/L)	5.2 ± 0.5	4.8 ± 0.3	0.04
2h-PG (mmol/L)	8.9 ± 1.0	5.0 ± 0.8	<0.01

Data are reported as mean ± SD or n (%) of characteristics. Group differences were compared by independent-samples t test or χ² test.

Glucose

Basal arterialized blood glucose levels were similar following dorzagliatin and placebo administration in both the IGT (5.2 ± 0.5 vs. 5.1 ± 0.4 mmol/L; P = 0.21) and NGT (4.6 ± 0.3 vs. 4.6 ± 0.4 mmol/L; P = 0.86) groups. During the hyperglycemic clamps, steady-state blood glucose levels were similar for the IGT and NGT groups receiving dorzagliatin or placebo treatments (Fig. 1A and E and Table 2).

Figure 1.

Blood glucose levels in IGT (A) and NGT (E) participants during hyperglycemic clamp. Plasma insulin levels in IGT (B) and NGT (F) participants during hyperglycemic clamp. Insulin secretion rates by C-peptide deconvolution in IGT (C) and NGT (G) participants during hyperglycemic clamp. Plots of insulin secretion rates against blood glucose levels at baseline and during clamp steady state in IGT (D) and NGT (H) participants. Blue line, dorzagliatin; red line, placebo. Data in plots are mean ± SE. Conc, concentration.

Table 2.

Comparisons of β-cell function parameters in IGT and NGT groups during hyperglycemic clamp with treatment of dorzagliatin or placebo

Parameter	IGT (n = 9)			NGT (n = 10)
	Placebo	Dorzagliatin	P	Placebo	Dorzagliatin	P
Glucose (mmol/L)
Basal blood glucose	5.1 ± 0.4	5.2 ± 0.5	0.21	4.6 ± 0.4	4.6 ± 0.3	0.86
Steady-state blood glucose	11.7 ± 0.3	11.6 ± 0.6	0.94	11.5 ± 0.3	11.5 ± 0.4	0.81
Insulin (pmol/L)
Basal insulin	46.1 ± 24.7	51.1 ± 22.5	0.51	35.2 ± 17.6	51.8 ± 34.1	0.01
AIRg	117.9 ± 107.7	98.0 ± 68.4	0.30	221.5 ± 206.9	189.1 ± 176.7	0.09
SIRg	291.3 ± 160.5	345.7 ± 144.4	0.11	415.0 ± 296.7	571.1 ± 538.4	0.08
Cmax-In	211.9 ± 171.0	188.7 ± 115.9	0.45	353.5 ± 296.7	342.0 ± 280.1	0.58
Tmax-In (min)	6.0 ± 1.7	6.2 ± 2.3	0.84	4.6 ± 1.0	5.4 ± 1.9	0.34
C-peptide (pmol/L)
Basal CP	582.5 ± 221.9	626.1 ± 226.9	0.37	411.0 ± 135.6	513.4 ± 257.4	0.32
ACPRg	445.9 ± 268.6	392.0 ± 234.3	0.24	844.1 ± 521.8	636.4 ± 452.5	0.06
SCPRg	1,780.0 ± 354.0	2,138.8 ± 522.1	0.01	2,291.0 ± 777.4	2,473.7 ± 1,061.6	0.49
Cmax-CP	1,182.9 ± 493.3	1,175.2 ± 478.1	0.91	1,374.5 ± 770.8	1,395.4 ± 658.0	0.81
Tmax-CP (min)	6.7 ± 1.7	6.4 ± 2.2	0.82	5.4 ± 2.5	6.2 ± 2.2	0.47
ISR (pmol/min/m2)
ISRb	69.8 ± 30.2	81.2 ± 32.3	0.21	51.8 ± 20.4	79.5 ± 64.4	0.38
ISR1abs	307.7 ± 147.2	288.3 ± 136.2	0.38	429.5 ± 272.2	402.5 ± 211.5	0.41
ISR1inc	237.9 ± 135.8	207.1 ± 121.4	0.16	377.6 ± 265.1	323.0 ± 233.9	0.10
ISR2abs	349.1 ± 68.8	425.6 ± 97.8	<0.01	423.0 ± 119.9	471.9 ± 190.6	0.14
ISR2inc	279.4 ± 48.0	344.4 ± 73.6	<0.01	371.2 ± 111.0	392.4 ± 196.8	0.57
β-CGS (pmol/min/m2 per mmol/L)	42.6 ± 9.8	55.6 ± 17.7	0.01	54.0 ± 17.0	57.2 ± 27.7	0.56
ISI and DI
ISI (10 × 10−5*mmol/kg/min per pmol/L)	10.8 ± 7.3	10.8 ± 5.0	0.97	14.0 ± 6.5	16.4 ± 10.1	0.30
DI (10 × 10−3*mmol/kg/min)	9.2 ± 6.4	8.1 ± 4.6	0.50	23.2 ± 16.3	21.3 ± 13.3	0.62

Data are presented as mean ± SD. Paired-samples t test was used for normally distributed data and Wilcoxon signed-rank test was used for nonnormally distributed data. C_max-In, maximum concentration of acute insulin response; C_max-CP, maximum concentration of acute C-peptide response; DI, disposition index (the product of insulin sensitivity index and acute insulin response to glucose); SIRg, second-phase insulin response to glucose; T_max-CP, time point of C_max-CP; T_max-In, time point of C_max-In.

Insulin and C-Peptide

Plasma basal insulin concentrations were higher following dorzagliatin treatment in the NGT group (51.8 ± 34.1 vs. 35.2 ± 17.6 pmol/L; P = 0.01) but did not differ by treatments in the IGT group. The AIRg was not significantly different after dorzagliatin treatment in both groups (Table 2). The plasma C-peptide concentrations are shown in Supplementary Fig. 2. ACPRg did not differ between treatments in NGT and IGT groups, whereas SCPRg was significantly higher following dorzagliatin treatment in the IGT group only (2,138.8 ± 522.1 vs. 1,780.0 ± 354.0 pmol/L; P = 0.01).

Insulin Secretion Rate by Deconvolution

In participants with IGT, ISRb did not differ by treatments, with comparable ISR1inc observed under dorzagliatin and placebo treatments (207.1 ± 121.4 vs. 237.9 ± 135.8 pmol/min/m², respectively; P = 0.16). Dorzagliatin treatment increased second-phase insulin secretion, as indicated by ISR2abs (425.6 ± 97.8 vs. 349.1 ± 68.8 pmol/min/m²; P < 0.01) and ISR2inc (344.4 ± 73.6 vs. 279.4 ± 48.0 pmol/min/m²; P < 0.01) compared with placebo (Fig. 1C and Table 2). β-Cell response to dorzagliatin did not differ between the isolated-IGT and IFG plus IGT subgroups (Supplementary Table 2 and Supplementary Fig. 3).

In the NGT group, no significant differences in ISRb were observed after dorzagliatin treatment compared with placebo. Both ISR1abs (dorzagliatin vs. placebo, respectively: 402.5 ± 211.5 vs. 429.5 ± 272.2 pmol/min/m²; P = 0.41) and ISR1inc (dorzagliatin vs. placebo, respectively: 323.0 ± 233.9 vs. 377.6 ± 265.1 pmol/min/m²; P = 0.10) were similar between treatments. Likewise, second-phase insulin secretion (ISR2abs, dorzagliatin vs. placebo: 471.9 ± 190.6 vs. 423.0 ± 119.9 pmol/min/m²; P = 0.14) was not different with dorzagliatin versus placebo (Fig. 1G and Table 2).

In the IGT group, ISR2abs increased with dorzagliatin (425.6 ± 97.8 pmol/min/m²) and appeared comparable to the NGT group under placebo treatment (423.0 ± 119.9 pmol/min/m²), although no formal statistical comparison was performed.

β-Cell Glucose Sensitivity and Insulin Sensitivity

The plots of ISR against blood glucose in both groups shifted upward following dorzagliatin versus placebo administration. In the IGT group, dorzagliatin enhanced ISR in a glucose dose–dependent manner, as evidenced by a steeper ISR–glucose relationship. This dose-dependent enhancement was not observed in the NGT group (Fig. 1D and H). β-CGS, expressed as slope of the ISR against blood glucose, was 1.3 times higher during dorzagliatin versus placebo administration in the IGT group (55.6 ± 17.7 vs. 42.6 ± 9.8 pmol/min/m² per mmol/L; P = 0.01). No differences were observed in the NGT group between dorzagliatin or placebo (57.2 ± 27.7 vs. 54.0 ± 17.0 pmol/min/m² per mmol/L, respectively; P = 0.56) (Table 2). ISI was similar after dorzagliatin or placebo treatment in the IGT (10.8 ± 5.0 vs. 10.8 ± 7.3 × 10⁻⁵*mmol/kg/min per pmol/L; P = 0.97) and NGT (16.4 ± 10.1 vs. 14.0 ± 6.5 × 10⁻⁵*mmol/kg/min per pmol/L; P = 0.30) groups (Table 2). Although no formal statistical comparison was performed, β-CGS in the IGT group after dorzagliatin treatment (55.6 ± 17.7 pmol/min/m² per mmol/L) appeared comparable to β-CGS in the NGT group with placebo (54.0 ± 17.0 pmol/min/m² per mmol/L).

Glucagon

Baseline plasma glucagon levels did not differ under dorzagliatin and placebo conditions in the IGT group (5.7 ± 2.7 vs. 7.0 ± 3.5 pmol/L, respectively; P = 0.30) and NGT group (6.3 ± 4.2 vs. 6.4 ± 2.1 pmol/L, respectively; P = 0.56) (Table 3). During the hyperglycemic clamp, plasma glucagon levels decreased progressively. We observed a greater suppression of glucagon in response to intravenous glucose in the NGT group after dorzagliatin treatment (AUC_{[0–120 min]} = 161.2 ± 58.1 vs. 234.2 ± 69.7 pmol*min/L, respectively; P = 0.01), with no significant difference between treatments in the IGT group (Fig. 2, Table 3, and Supplementary Table 3). Glucagon response to dorzagliatin did not differ between the isolated-IGT and IFG plus IGT subgroups (Supplementary Table 2 and Supplementary Fig. 3).

Table 3.

Comparisons of α-cell function parameters in IGT and NGT groups during hyperglycemic clamp with treatment of dorzagliatin or placebo

Parameter	IGT (n = 9)			NGT (n = 10)
	Placebo	Dorzagliatin	P	Placebo	Dorzagliatin	P
Glucagon
Basal glucagon (pmol/L)	7.0 ± 3.5	5.7 ± 2.7	0.30	6.4 ± 2.1	6.3 ± 4.2	0.56
AUC all glucagon (pmol × min/L)	440.6 ± 263.7	406.0 ± 187.5	0.56	359.1 ± 105.5	285.2 ± 126.7	0.19
AUC basal glucagon (pmol × min/L)	69.2 ± 35.3	57.0 ± 26.9	0.30	64.3 ± 21.0	62.9 ± 42.0	0.56
AUC[0–120] glucagon (pmol × min/L)	305.3 ± 200.1	291.1 ± 156.3	0.75	234.2 ± 69.7	161.2 ± 58.1	0.01
GSR (pmol/min)
GSRb	9.3 ± 5.2	7.3 ± 3.8	0.25	8.4 ± 3.7	8.6 ± 6.8	0.49
GSR1abs (0–15 min)	4.7 ± 2.7	4.6 ± 3.0	0.92	3.4 ± 1.5	2.2 ± 1.0	<0.01
GSR1inc (0–15 min)	−4.6 ± 3.0	−2.7 ± 3.1	0.24	−4.9 ± 3.3	−6.4 ± 6.0	0.50
GSR2abs (80–120 min)	2.2 ± 1.7	2.2 ± 1.7	0.97	2.0 ± 1.3	1.0 ± 0.4	0.01
GSR2inc (80–120 min)	−7.1 ± 4.2	−5.1 ± 3.7	0.29	−6.4 ± 3.0	−7.6 ± 6.7	0.54
RGSR1 (pmol/min per mmol/L)	−0.8 ± 0.5	−0.6 ± 0.6	0.43	−0.9 ± 0.6	−1.2 ± 1.1	0.50
RGSR2 (pmol/min per mmol/L)	−1.1 ± 0.6	−0.8 ± 0.6	0.44	−0.9 ± 0.5	−1.2 ± 1.1	0.50
GLP-1
AUC total GLP-1 (pmol × min/L)	264.5 ± 114.3	224.4 ± 79.2	0.21	280.9 ± 147.1	210.8 ± 87.3	0.05

Data are presented as mean ± SD. Paired-samples t test was used for normally distributed data and Wilcoxon signed-rank test was used for nonnormally distributed data. GSR1inc, incremental glucagon secretion between 0 and 15 min; GSR2inc, incremental glucagon secretion between 80 and 120 min.

Figure 2.

Plasma glucagon levels in IGT (A) and NGT (D) participants during hyperglycemic clamp. Plasma total GLP-1 levels in IGT (B) and NGT (E) participants during hyperglycemic clamp. GSRs in IGT (C) and NGT (F) participants during hyperglycemic clamp. Blue line, dorzagliatin; red line, placebo. Data in plots are mean ± SE.

Under dorzagliatin conditions, GSRb did not differ, but early glucagon secretion GSR1abs (2.2 ± 1.0 vs. 3.4 ± 1.5 pmol/min; P < 0.01) and GSR2abs (1.0 ± 0.4 vs. 2.0 ± 1.3 pmol/min; P = 0.01) in the last 40 min were lower compared with placebo in the NGT group (Fig. 2 and Table 3). GSR did not differ between treatment conditions in the IGT group (Fig. 2 and Table 3). The slope of GSR versus glucose (RGSR1; 0–15 min) and RGSR2 (80–120 min), as indices for the sensitivity of glucagon secretion suppression to the increment in glucose concentration, did not differ for both treatments in the IGT and NGT groups (Table 3).

GLP-1

Plasma total GLP-1 levels were low but detectable and mildly decreased during the hyperglycemic clamp under both treatment conditions in both groups. There was a reduction in total GLP-1 AUC during dorzagliatin versus placebo in NGT group only (210.8 ± 87.3 vs. 280.9 ± 147.1, respectively; P = 0.05). (Table 3 and Fig. 2). We applied a validated repeated-measure mixed model separately in both the NGT and IGT groups under dorzagliatin and placebo treatments. The estimated correlation parameters (ρ) between GLP-1 and glucagon are summarized in Supplementary Table 5. A moderate positive correlation was observed in the IGT group under both treatments (dorzagliatin: ρ = 0.65; placebo: ρ = 0.64). In the NGT group, a moderate correlation was observed with dorzagliatin treatment (ρ = 0.46) and a lower correlation with placebo (ρ = 0.17). Figure 3 shows the positive relationships between plasma GLP-1 and glucagon during the hyperglycemic clamp in both groups.

Figure 3.

Scatterplot of glucagon and total GLP-1 in IGT (A) and NGT (B) participants during hyperglycemic clamp.

Adverse Events

There were no serious adverse events, deaths, or adverse events leading to treatment discontinuation (Supplementary Table 1). There were three hypoglycemic events in two NGT participants (N01 and N07) and one IGT participant (P09) following dorzagliatin treatment. All hypoglycemic episodes occurred within 24 h after dorzagliatin and 4–10 h after hyperglycemic clamp. All events were mild (self-monitored blood glucose 3.3–3.8 mmol/L) and did not require medical treatment. All three episodes were judged possibly related to dorzagliatin. One participant experienced transient swelling of the eyelids 24 h after placebo, which spontaneously resolved.

Discussion

In this double-blind, single-dose, crossover hyperglycemic study, we observed 1) dorzagliatin significantly increased second-phase insulin secretion rate and improved β-CGS in the IGT group but not the NGT group; 2) dorzagliatin had an effect on the α-cell by facilitating GSGS in the NGT group, and total GLP-1 levels also decreased in parallel; 3) plasma glucagon was positively correlated with total GLP-1 levels; and 4) dorzagliatin had no effect on insulin sensitivity.

Dorzagliatin 50 mg significantly increased second-phase ISR by 1.2-fold in the IGT group. This was greater than that reported in our previous study among participants with recent-onset T2D despite a lower dosage of dorzagliatin (12) but less than the 1.5-fold increase in participants with GCK-MODY (Supplementary Table 4). β-CGS was 1.3 times higher after dorzagliatin 50 mg administration compared with placebo in IGT participants. This is consistent with the known actions of dorzagliatin on GCK by shifting glucose thresholds for insulin secretion (25). However, this finding contrasts with the lack of significant increase in second-phase ISR and β-CGS in the NGT group during within-group analysis. In a phase 1a trial of dorzagliatin in healthy volunteers, a single dose of dorzagliatin 50 mg reduced plasma glucose with increased insulin secretion following an oral glucose load (15). This discrepancy suggested that the effects of dorzagliatin may be more pronounced through incretin-mediated pathways rather than pancreatic β-cell insulin secretion in individuals with normal GCK function and expression.

To our knowledge, our data are the first to support a potential role of α-cell GCK in directly regulating GSGS in vivo in healthy humans, consistent with in vitro and animal studies (4). In the NGT group, dorzagliatin did not affect acute insulin secretion but suppressed glucagon levels. Although we cannot fully exclude the possibility that insulin directly contributed to glucagon suppression, the significant reduction in glucagon in the absence of a notable insulin increase suggests a potential direct effect of dorzagliatin on α-cell GCK function or an indirect systemic mechanism beyond insulin regulation. Other groups demonstrated that manipulation of GCK activity and expression in cultured single rat α-cells altered the glucose threshold for glucagon release (26). In a mutant mouse with an α-cell–specific activating GCK mutation, glucagon suppression was enhanced in vivo and ex vivo with improved glucose metabolism during feeding with high-fat diets (5). Perfusion studies in islets isolated from surgical samples following pancreatectomy in patients with gain-of-function mutations of the GCK gene showed suppressed glucagon secretion with increased sensitivity of GSIS (27). The lack of similar responses in individuals with IGT or T2D might be due to reduced α-cell GCK expression or reduced signaling due to hyperglycemia (4,28) which might attenuate effects of a single dose of dorzagliatin. In a rat model of T2D, dorzagliatin administration did not alter glucagon levels (29). However, we speculate that prolonged administration of dorzagliatin might ameliorate hyperglycemia and restore α-cell GCK expression with enhanced suppression of glucagon.

Because approximately 95% of circulating GLP-1 originates from intestinal L cells after oral nutrient ingestion, total GLP-1 levels were expectedly low during the hyperglycemic clamp, which bypasses enteral stimulation. Nonetheless, total GLP-1 was detectable and declined progressively in both groups, consistent with findings from prior hyperglycemic clamp studies that measured GLP-1 in humans (30,31). In α-cells, glucagon and GLP-1 are cleaved from the preproglucagon in a 8:1 ratio (32). We observed positive correlations between glucagon and GLP-1, suggesting a potential link in their coregulation via the α-cell. However, we did not perform formal statistical testing of the correlation coefficient, due to small sample size and model complexity; thus, our results should be interpreted with caution.

Intriguingly, we found that total GLP-1 showed greater suppression after dorzagliatin administration only in the NGT group. These parallel responses, similar to that for GSGS, raised the possibility that both glucagon and proglucagon peptide secretion in the α-cell might be regulated by GCK. Because we used nonoral glucose stimulation in this study, we expected minimal intestinal GLP-1 secretion. The physiological relevance of pancreatic GLP-1 secretion and its insulinotropic effects under this nonphysiological model of intravenous glucose–induced hyperglycemia remains uncertain. However other studies have highlighted dysregulated α-to-β-cell signaling may contribute to the defects in T2D (33,34). We also cannot exclude the possibility of an indirect effect via insulin on pancreatic GLP-1 secretion.

This study was conducted to complement our previous experiment in GK-MODY and T2D regarding the effects of dorzagliatin in the full spectrum of glycemia. In our previous hyperglycemic clamp experiment with participants with recent-onset T2D, dorzagliatin increased basal insulin secretion by twofold, which was followed by a paradoxical 10-fold decrement in first-phase insulin secretion (12). This paradoxical decrement did not occur in the present study, and no significant differences in ISRb or ISR1inc after dorzagliatin or placebo treatment were observed in either participant group. We hypothesized at the time that promoted glycolysis through increased glucose phosphorylation by dorzagliatin might cause a sustained increase in the ATP-to-ADP ratio, resulting in a higher ISRb in T2D. This was followed by diminished first-phase ISR due to a disequilibrium between exocytosis and recruitment of the readily releasable pool of granules. In this context, glycolytic oscillations and changes in the ATP-to-ADP ratio can alter K_ATP conductance and trigger basal insulin release (35,36). In the IGT group, there were individuals with concomitant IFG, which might suggest reduced glucose sensing and affect basal and early phase insulin response (37). In our exploratory subgroup analysis, we found numerically greater increases in basal CP and ISRb in the IFG plus IGT group than in the isolated-IGT group. However, due to the very small sample size, these changes did not reach statistical significance.

Our study was not specifically designed to assess basal insulin secretion. In addition, we shortened the preclamp dosing interval to 30 min as compared with 90 min in our previous study (12). With this shorter interval, we estimated that approximately 50% C_max of dorazagliatin would be attained. This submaximal drug exposure might not be sufficient to elicit the basal insulin response. Therefore, the findings of dorzagliatin on basal insulin should be interpreted with caution. More studies with a larger sample size specifically designed to compare IFG and IGT participants are necessary to elucidate the distinct metabolic effects of dorzagliatin on basal insulin across the prediabetes spectrum.

The strengths of this study include use of a double-blind, placebo-controlled, crossover hyperglycemic clamp study design to assess insulin secretion and offer a detailed estimation of β-cell function, glucose sensitivity, and GSGS. Due to concern about hypoglycemia, we did not use the same dose or timing of administration as in our previous GCK-MODY and recent-onset T2D study (12). This limited our direct comparison of the effects of dorzagliatin in individuals with NGT, IFG, IGT, GCK-MODY, and T2D. Our study was not designed to investigate the effects of dorzagliatin on incretin hormones via the intestinal route including GLP-1 and glucose-dependent insulinogenic polypeptide. The impact of dorzagliatin on incretin responses and intestinal L-cell GCK activity requires investigation with enteral studies. We are currently conducting studies combining hyperglycemic clamp with OGTT in individuals with IGT (ClinicalTrial.gov identifier NCT06671340). This study will allow us to assess the roles of incretin signaling and α-cell and β-cell interactions through systemic glucose regulation following dorzagliatin administration. Our NGT group was younger than the IGT group; however, despite no significant differences in BMI and sex between groups, the degree of second-phase insulin enhancement with dorzagliatin was smaller in the NGT group. The sample size was small. We only evaluated the effect of dorzagliatin following a single dose and the effects of repeated dosing on long-term and sustained glucose response in NGT and IGT warrant further investigations.

Our findings unveil several avenues for future research. The varying responses to dorzagliatin based on glucose tolerance status call for further investigations. Given the expression of GCK in both α- and β-cells, the enhanced suppression of GLP-1-glucagon response to dorzagliatin in NGT strongly supports the importance of GCK in regulating this coupled response. However, we cannot fully exclude the possibility that the observed GLP-1-glucagon response may be mediated indirectly via insulin, despite no significant enhancement in insulin secretion observed in our study. Downregulation of β-cell and possibly α-cell GCK expression might impair insulin secretion and GSGS (38). In individuals with T2D, GCK expression was reduced by >60% compared with individuals with NGT (39). In diabetic mouse models, 1-month treatment with dorzagliatin restored hepatic GCK gene expression and GCK protein compared with the untreated group (29). Definitive studies are needed to confirm whether chronic treatment with dorzagliatin can restore α- and β-cell GCK expression and address deficits of both impaired pancreatic insulin secretion and nonsuppression of glucagon in people with IGT and T2D.

In conclusion, this proof-of-concept study supports a role of dorzagliatin on pancreatic β- and α-cell function in enhancing second-phase insulin secretion in IGT and facilitating GSGS in NGT. Delving into the long-term effects of dorzagliatin on α- and β-cell function and glucose homeostasis, along with conducting studies involving larger sample sizes and diverse populations with dysglycemia, will be pivotal in advancing our comprehension of the role of GCK activators in glucose management.

This article contains supplementary material online at https://doi.org/10.2337/figshare.29829494.