ABSTRACT
Aims
To investigate the dynamic changes of soluble receptor for advanced glycation end products (sRAGE) and its relationship with β‐cell function and glycemic remission in newly diagnosed type 2 diabetes mellitus (T2DM) after short‐term intensive insulin therapy (SIIT).
Methods
A total of 98 drug‐naive patients with newly diagnosed T2DM underwent a two‐week SIIT and were followed for 16 weeks. Serum sRAGE was measured at baseline, after SIIT, and at the 1‐month follow‐up. HOMA‐β and HOMA‐IR were used to reflect β‐cell function and insulin resistance, respectively. Pearson or Spearman correlation coefficients were used to assess the correlations, and logistic regression was applied to evaluate associations between changes in sRAGE (ΔsRAGE) and 16‐week glycemic remission, adjusting for potential confounders.
Results
sRAGE levels increased significantly after SIIT (P < 0.001) and remained higher at 1 month. The increase in sRAGE after SIIT was positively correlated with the changes in HOMA‐β and negatively correlated with changes in HOMA‐IR. Participants who achieved glycemic remission had greater ΔsRAGE after SIIT than those who did not (273.9 ± 268.2 vs 76.7 ± 257.7 pg/mL, P = 0.001). In multivariate analysis, ΔsRAGE was independently associated with remission (OR = 1.228 per 100 pg/mL, 95% CI 1.006–1.500, P = 0.044), and the AUC of this model was 0.783 (95% CI 0.690–0.876).
Conclusions
Elevation of sRAGE after SIIT is independently associated with short‐term glycemic remission in newly diagnosed T2DM. The increase in sRAGE accompanied by an improvement in β‐cell function, suggests that increased sRAGE may reflect a compensatory response during metabolic recovery.
Keywords: Beta‐Cell Preservation, Early Diabetes Intervention, sRAGE
Short‐term intensive insulin therapy (SIIT) markedly increased circulating sRAGE levels in newly diagnosed type 2 diabetes. The magnitude of sRAGE elevation was associated with improved β‐cell function and a higher likelihood of short‐term glycemic remission.
Abbreviations
- AGEs
advanced glycation end products
- BMI
body mass index
- CI
confidence intervals
- HDL‐c
high‐density lipoprotein cholesterol
- HOMA‐IR
homeostasis model assessment of insulin resistance
- HOMA‐β
homeostasis model assessment of β‐cell function
- LDL‐c
low‐density lipoprotein cholesterol
- OR
odds ratios
- RAGE
receptor for advanced glycation end products
- SIIT
short‐term intensive insulin treatment
- sRAGE
soluble receptors for advanced glycation end products
- T2DM
type 2 diabetes mellitus
INTRODUCTION
Type 2 diabetes (T2DM) is a major global health challenge, and its prevalence continues to rise worldwide. It is estimated that currently, over 400 million people are suffering from T2DM 1 . With the increasing incidence rate, the treatment strategies for T2DM have evolved from glucose control alone to metabolic restoration. Short‐term intensive insulin therapy (SIIT) has been proven to induce drug‐free glycemic remission and improve β‐cell function in newly diagnosed T2DM patients 2 , 3 . Although this concept has been recognized for twenty years, the potential mechanisms responsible for glycemic remission and β‐cell recovery after SIIT are still not fully clear.
Advanced glycation end products (AGEs) are a group of stable covalent compounds formed by non‐enzymatic glycation reactions between reducing sugars and proteins, lipids, or nucleic acids. They bind to their receptor, receptor for advanced glycation end products (RAGE), leading to insulin resistance and β‐cell dysfunction 4 , 5 . In contrast, soluble RAGE (sRAGE) includes the cleaved form (cRAGE) produced by protein hydrolysis of RAGE and the endogenous secreted (esRAGE) isoforms produced by alternative splicing of RAGE. Acting as a decoy receptor, sRAGE can bind to circulating AGEs and reduce oxidative stress and inflammation 5 , 6 . Decreased circulating sRAGE levels have been reported in individuals with newly diagnosed T2DM 7 and are associated with adverse metabolic outcomes and diabetic complications 8 , 9 , 10 , 11 , 12 .
Insulin therapy has been reported to increase sRAGE concentrations 13 , 14 , suggesting that improved metabolic states may be accompanied by alterations in the AGE–RAGE axis. However, whether changes in sRAGE during intensive insulin treatment are related to subsequent glycemic remission has not been clarified. Therefore, in our study, we aimed to (1) explore the dynamic changes in sRAGE during and after SIIT, and (2) investigate whether the increase in sRAGE after SIIT is related to glycemic remission in newly diagnosed T2DM patients. Our findings may provide new insights into the potential involvement of sRAGE in β‐cell functional recovery and metabolic improvement induced by intensive insulin therapy.
MATERIAL AND METHODS
Study participants
Between April 2020 and June 2023, adults aged 18–70 years who were newly diagnosed with T2DM, according to the 1999 World Health Organization standard 15 , were recruited at the First Affiliated Hospital of Sun Yat‐sen University.
Inclusion criteria were: (1) drug‐naive status; (2) HbA1c ≥ 9.0%; and (3) body mass index (BMI) between 20 and 35 kg/m2.
Exclusion criteria included: type 1 or other specific forms of diabetes; current use of medications affecting glucose metabolism; acute diabetic complications (e.g., ketoacidosis, hyperosmolar state); proliferative diabetic retinopathy; significant proteinuria (>300 mg/day or >0.5 g/day); major cardiovascular events within 12 months; uncontrolled hypertension; hepatic dysfunction (ALT ≥2.5 times upper limit or total bilirubin ≥1.5 times upper limit); renal impairment (eGFR ≤ 50 mL/min/1.73 m2); anemia (Hb < 100 g/L); transfusion requirement; or any severe systemic disease.
All participants signed a written informed consent form. This study strictly followed the Declaration of Helsinki and Good Clinical Practice guidelines and was approved by the institutional review board ([2019] 174). This study was from a trial registered on ClinicalTrials.gov (NCT03972982).
Study procedures
All patients were admitted to the hospital following their diagnosis and initiated on a standardized lifestyle intervention. Baseline assessments included anthropometric measurements, medical history, and laboratory testing: liver and kidney function, lipid profile, oral glucose tolerance test (OGTT), HbA1c, glycated albumin (GA), insulin and C‐peptide release tests, AGEs, and sRAGE.
After baseline evaluation, SIIT was initiated using continuous subcutaneous insulin infusion (CSII) with insulin aspart (Novo Nordisk, Denmark) at an initial dose of 0.4–0.5 IU/kg/day. Forty percent of the total dose was delivered as basal insulin evenly over 24 h, and the remaining 60% was equally divided before each meal. Doses were titrated daily to achieve target glucose levels (fasting 4.4–5.6 mmol/L; 2‐h postprandial 4.4–7.6 mmol/L). After achieving the glycemic target, intensive insulin therapy was continued for an additional two weeks.
During hospitalization, participants received a standardized diet and moderate postprandial exercise guidance. Insulin was discontinued after the final dose (prior to dinner), and all post‐treatment measurements were obtained the following morning (≥15 h after the last injection).
Participants were followed at weeks 4, 8, and 16. At 1 month (week 4), all baseline tests were repeated. Participants were required to monitor fasting fingertip glucose at least twice a week; continuous fasting glucose >8.0 mmol/L (confirmed by venous blood glucose) for two consecutive weeks during follow‐up was defined as a failure of remission. The definition of glycemic remission at week 16 was FPG < 7.0 mmol/L and HbA1c < 7.0%.
Measurement and calculation
According to the manufacturer's protocols, we used ELISA kits (Shanghai Enzyme‐linked Biotechnology Co., Ltd., China) to quantify serum AGEs and sRAGE. The intra‐assay and inter‐assay coefficients of variation (CVs) of the ELISA kits were ≤4.5% and ≤6.0%, respectively. The detection ranges of the AGEs and sRAGE assays were 0.3–24 μg/mL and 23.6–2,400 pg/mL, respectively. The antibody used in the AGEs ELISA recognizes multiple AGE epitopes, including Nε‐(carboxymethyl)‐lysine (CML), Nε‐(carboxyethyl)‐lysine (CEL), pentosidine, and pyrraline. Therefore, the measured values reflect the total concentration of AGEs detectable by this antibody in the serum samples.
Changes in parameters relative to baseline at each time point were expressed as Δ (delta) values, for example, ΔsRAGE_after SIIT = sRAGE_after SIIT – sRAGE_baseline.
β‐cell function and insulin resistance were evaluated by the Homeostasis Model Assessment (HOMA): HOMA‐β = 20 × fasting insulin / (FPG – 3.5), HOMA‐IR = (FPG × fasting insulin)/22.5.
Statistical analysis
All analyses were conducted using SPSS v26.0 (IBM Corp., USA) and R v4.3.2 (R Foundation for Statistical Computing, Vienna, Austria). The Shapiro–Wilk test was conducted to assess normality. Normal distribution data were represented by mean ± standard deviation (SD), while non‐normal distribution data were represented by median (interquartile range). Categorical variables were summarized as counts and percentages.
Used repeated‐measures anova or Friedman test to analyze the differences at three time points (pre SIIT, post SIIT, and 1‐month follow‐up), and conducted post‐hoc pairwise comparisons by paired t‐test or Wilcoxon signed‐rank test (Bonferroni‐adjusted P < 0.018). Used independent‐sample t‐test or Mann–Whitney U test to compare baseline characteristics and metabolic parameters between the remission and non‐remission groups.
Pearson or Spearman correlation coefficients were used to evaluate the correlation between biomarkers. Univariate logistic regression was used to determine baseline variables associated with glycemic remission. Variables with P < 0.05 in univariate analysis were included as candidate covariates in the multiple logistic regression model. Multicollinearity was examined using variance inflation factors (VIF), with values <5 considered acceptable. Model validation was performed via ROC curve, calibration curve, decision curve analysis (DCA), and bootstrap resampling.
RESULTS
Study population
Of the 112 participants who received SIIT, 11 were lost to follow‐up, one initiated additional hypoglycemic medication by self during follow‐up, and two had missing post‐treatment sRAGE data. Ultimately, a total of 98 participants (70 males and 28 females) were enrolled in the final analysis (Figure 1). No significant differences in baseline characteristics were observed between the enrolled and excluded groups (P > 0.05).
Figure 1.
Flowchart of participant.
At baseline, the median age and BMI of the participants were 49.0 years (IQR: 35.75–56.25) and 24.73 kg/m2 (IQR: 23.22–26.80), respectively. The median HbA1c level was 11.00% (IQR: 9.68–12.10), and the mean FPG was 11.27 ± 2.73 mmol/L. The median baseline sRAGE was 2123.52 pg/mL (IQR: 1596.94–2417.72).
All baseline characteristics of the participants are presented in Table S1.
Changes in metabolic parameters and sRAGE levels after SIIT
SIIT markedly improved glycemic and metabolic parameters (Table 1 and Figure S1). FPG and HbA1c levels decreased significantly after SIIT and remained lower at the 1‐month follow‐up (both P < 0.001; Figure S1D,E). β‐cell function increased sharply immediately after SIIT and remained elevated at follow‐up, whereas HOMA‐IR transiently decreased and partially rebounded after one month (Figure S1B,C). In addition, lipid profiles showed marked improvement after SIIT, with significant reductions in both triglycerides and total cholesterol (both P < 0.001).
Table 1.
Effects of short‐term intensive insulin therapy on metabolic parameters
| Parameter | Before SIIT | After SIIT | 1‐month Follow‐up | P‐value |
|---|---|---|---|---|
| BMI, kg/m2 | 24.73 (23.22, 26.80) | 24.29 (22.81, 26.87) † | 23.71 (22.20, 26.43) † , ‡ | <0.001 |
| Triglyceride, mmol/L | 1.63 (1.25, 2.48) | 1.16 (0.95, 1.61) † | 1.09 (0.83, 1.64) † | <0.001 |
| Cholesterol, mmol/L | 5.46 ± 1.07 | 4.74 ± 1.21 † | 4.37 ± 1.18 † , ‡ | <0.001 |
| HDL‐C, mmol/L | 1.02 (0.91, 1.22) | 1.05 (0.89, 1.24) | 1.09 (0.95, 1.26) † | 0.016 |
| LDL‐C, mmol/L | 3.63 (3.13, 4.15) | 3.07 (2.41, 3.55) † | 2.74 (2.12, 3.38) † | <0.001 |
| Glycated albumin, % | 26.86 (23.95, 31.71) | 19.16 (17.19, 21.22) † | 15.54 (13.70, 17.71) † , ‡ | <0.001 |
| HbA1c, % | 11.00 (9.68, 12.10) | 9.25 (8.40, 10.10) † | 7.15 (6.50, 7.90) † , ‡ | <0.001 |
| FPG, mmol/L | 11.45 (8.98, 13.10) | 5.45 (4.80, 6.10) † | 6.50 (5.78, 7.33) † , ‡ | <0.001 |
| HOMA‐β | 20.60 (12.13, 34.55) | 53.38 (33.12, 89.97) † | 73.45 (43.33, 122.19) † , ‡ | <0.001 |
| ∆HOMA‐β | – | 28.45 (14.74, 54.01) | 48.67 (26.99, 96.39) | – |
| HOMA‐IR | 3.65 (2.34, 5.37) | 1.40 (0.71, 2.19) † | 3.02 (1.98, 5.34) † , ‡ | <0.001 |
| ∆HOMA‐IR | – | −2.23 (−3.24, −1.22) | −0.32 (−1.71, 0.95) | – |
| AGEs, ug/mL | 22.68 (20.21, 25.93) | 22.96 (19.15, 25.84) | 23.32 (19.98, 26.22) | 0.445 |
| sRAGE, pg/mL | 2123.52 (1596.94, 2417.72) | 2360.69 (1892.89, 2564.71) † | 2148.94 (1794.51, 2480.88) † , ‡ | <0.001 |
| ∆sRAGE, pg/mL | – | 220.68 (77.00, 404.86) | 111.53 (−98.24, 276.55) | – |
Data are presented as mean ± SD for normally distributed continuous variables, median (IQR) for non‐normally distributed variables. Comparisons across the three time points (before SIIT, after SIIT, and 1‐month follow‐up) were assessed using repeated‐measures anova or Friedman tests. Bonferroni‐adjusted P < 0.018 was considered statistically significant for post hoc comparisons.
Compared with the Before SIIT group P < 0.0018.
Compared with the After SIIT group, P < 0. 0018.
BMI, body mass index; FPG, fasting plasma glucose; HbA1c, glycated hemoglobin; HDL‐C, high‐density lipoprotein cholesterol; HOMA‐IR, homeostasis model assessment of insulin resistance; HOMA‐β, homeostasis model assessment of β‐cell function; LDL‐C, low‐density lipoprotein cholesterol.
Circulating sRAGE levels were significantly increased after SIIT (P < 0.001) and remained increased at the 1‐month follow‐up. In contrast, AGEs exhibited no significant changes either after SIIT or at the 1‐month follow‐up (P > 0.05; Figure S1A,F).
Notably, the increase in sRAGE levels was accompanied by improvements in β‐cell function. Correlation analysis further revealed a positive correlation between ΔsRAGE_after SIIT and ΔHOMA‐β_after SIIT (r = 0.3328, P = 0.0008), as well as with ΔHOMA‐β_1‐month (r = 0.5133, P < 0.0001). Conversely, ΔsRAGE after SIIT was negatively correlated with ΔHOMA‐IR_after SIIT (r = −0.2476, P = 0.014; Figure S1G,H).
These findings suggest that the increase in sRAGE during SIIT occurs concurrently with improvements in β‐cell function‐related indices and metabolic parameters.
Comparison of metabolic changes between remission and non‐remission groups
After the 16‐week follow‐up, patients were stratified by glycemic status into the remission group (n = 69) and the non‐remission group (n = 29). Compared to the non‐remission group, the remission group demonstrated a more favorable metabolic profile at all time points (Table 2 and Figure S2).
Table 2.
Comparison of metabolic parameters between remission and non‐remission groups
| Before SIIT | After SIIT | 1‐month follow‐up | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Remission (n = 69) | Non‐remission (n = 29) | P‐value | Remission (n = 69) | Non‐remission (n = 29) | P‐value | Remission (n = 69) | Non‐remission (n = 29) | P‐value | |
| BMI, kg/m2 | 25.07 (23.67, 27.15) | 24.51 (22.70, 26.12) | 0.090 | 24.56 (23.24, 27.57) | 24.08 (22.29, 26.19) | 0.191 | 23.78 (22.38, 26.80) | 23.05 (21.58, 25.82) | 0.283 |
| Alanine transaminase, U/L | 23.00 (17.00, 45.00) | 19.00 (14.00, 31.00) | 0.138 | 21.00 (15.00, 33.50) | 21.00 (18.00, 26.50) | 0.836 | 20.00 (16.00, 27.50) | 21.00 (17.50, 27.500) | 0.803 |
| Serum creatinine, μmol/L | 68.77 ± 11.96 | 61.55 ± 13.41 | 0.010* | 75.81 ± 11.66 | 70.72 ± 14.50 | 0.070 | 75.10 ± 11.65 | 69.00 ± 15.87 | 0.037* |
| Triglyceride, mmol/L | 1.52 (1.15, 2.22) | 2.08 (1.42, 2.86) | 0.030* | 1.12 (0.93, 1.57) | 1.33 (0.96, 2.04) | 0.148 | 1.00 (0.77, 1.50) | 1.32 (0.89, 1.98) | 0.055 |
| Cholesterol, mmol/L | 5.32 ± 1.05 | 5.80 ± 1.08 | 0.044* | 4.58 ± 1.02 | 5.10 ± 1.52 | 0.054 | 4.33 ± 1.15 | 4.46 ± 1.26 | 0.624 |
| HDL‐C, mmol/L | 1.01 (0.88, 1.15) | 1.07 (0.97, 1.37) | 0.050 | 1.05 ± 0.23 | 1.19 ± 0.29 | 0.013* | 1.06 (0.94, 1.26) | 1.10 (0.98, 1.32) | 0.366 |
| LDL‐C, mmol/L | 3.58 (3.02, 4.07) | 3.85 (3.25, 4.26) | 0.089 | 2.97 ± 0.78 | 3.29 ± 1.09 | 0.117 | 2.80 (2.15, 3.47) | 2.38 (1.90, 3.30) | 0.317 |
| Glycated albumin, % | 26.97 ± 4.94 | 29.16 ± 6.32 | 0.069 | 18.67 ± 2.83 | 20.42 ± 3.15 | 0.008* | 15.08 ± 2.09 | 18.17 ± 3.30 | <0.001* |
| HbA1c, % | 10.90 (9.40, 11.85) | 11.50 (10.15, 12.45) | 0.156 | 9.07 ± 1.31 | 9.64 ± 1.18 | 0.042* | 6.94 ± 0.93 | 7.92 ± 1.07 | <0.001* |
| FPG, mmol/L | 10.30 (8.40, 13.00) | 12.40 (11.10, 14.05) | 0.003* | 5.30 (4.70, 5.80) | 6.10 (5.55, 7.40) | <0.001* | 6.00 (5.30, 6.80) | 8.00 (7.05, 9.30) | <0.001* |
| Fasting C‐peptide, nmol/L | 2.46 (2.03, 2.96) | 2.20 (1.80, 2.64) | 0.099 | 2.04 (1.55, 2.39) | 1.90 (1.60, 2.44) | 0.855 | 2.55 (2.08, 3.29) | 2.80 (2.15, 3.45) | 0.297 |
| HOMA‐β | 23.37 (14.27, 39.35) | 14.81 (7.70, 25.24) | 0.006* | 64.00 (36.08, 103.70) | 36.93 (32.81, 53.38) | 0.006* | 78.14 (53.17, 141.19) | 44.27 (31.89, 85.68) | 0.001* |
| ∆HOMA‐β | – | – | – | 32.56 (16.28, 70.54) | 22.12 (8.99, 32.15) | 0.036* | 58.20 (29.23, 105.82) | 35.71 (19.51, 65.65) | 0.008* |
| HOMA‐IR | 3.68 (2.14, 5.55) | 3.62 (2.51, 5.08) | 0.831 | 1.24 (0.64, 1.94) | 1.68 (0.80, 2.58) | 0.194 | 2.79 (1.88, 4.52) | 4.26 (2.57, 6.49) | 0.019* |
| ∆HOMA‐IR | – | – | – | −2.24 (−3.49, −1.31) | −1.81 (−3.00, −0.80) | 0.248 | −0.64 (−2.21, 0.55) | 0.38 (−1.13, 2.26) 2.84 | 0.026* |
| AGEs, ug/mL | 22.31 (20.24, 25.57) | 24.79 (19.67, 27.96) | 0.414 | 22.39 (19.26, 25.14) | 23.43 (18.71, 26.27) | 0.764 | 22.99 (20.11, 25.99) | 23.35 (18.89, 26.68) | 0.654 |
| sRAGE, pg/mL | 2097.99 (1636.51, 2394.50) | 2225.06 (1521.24, 2551.53) | 0.714 | 2408.20 (2021.34, 2568.43) | 2266.98 (1688.49, 2524.07) | 0.182 | 2160.21 (1852.57, 2472.29) | 2107.12 (1524.80, 2608.05) | 0.603 |
| ∆sRAGE, pg/mL | – | – | – | 273.91 ± 268.20 | 76.74 ± 257.74 | 0.001* | 121.54 ± 248.16 | 10.64 ± 248.91 | 0.046* |
Data are presented as mean ± SD for normally distributed continuous variables, median (IQR) for non‐normally distributed variables. Changes in parameters relative to baseline at each time point were expressed as Δ (delta) values.
P < 0.05.
BMI, body mass index; FPG, fasting plasma glucose; HbA1c, glycated hemoglobin; HDL‐C, high‐density lipoprotein cholesterol; HOMA‐IR, homeostasis model assessment of insulin resistance; HOMA‐β, homeostasis model assessment of β‐cell function; LDL‐C, low‐density lipoprotein cholesterol.
At baseline, the remission group had lower FPG and higher HOMA‐β than the non‐remission group (P = 0.003 and 0.006, respectively). After SIIT, both groups showed significant metabolic improvement; however, the remission group achieved lower FPG (P < 0.001), lower HbA1c (P = 0.042), and a greater increase in HOMA‐β (P = 0.006). At the 1‐month follow‐up, the divergence further widened: the remission group maintained lower FPG and HbA1c (both P < 0.001), higher HOMA‐β (P = 0.001), and lower HOMA‐IR (P = 0.019; Figure S2A–E).
Importantly, the rise in ΔsRAGE_after SIIT was much larger in the remission group (273.9 ± 268.2 pg/mL) than in the non‐remission group (76.7 ± 257.7 pg/mL, P = 0.001). Moreover, at the 1‐month follow‐up, ΔsRAGE was still significantly higher in the remission group (P = 0.046). A similar trend was observed for ΔHOMA‐β (P = 0.036 and P = 0.008, respectively), while a significant difference in ΔHOMA‐IR was detected only at the 1‐month follow‐up (P = 0.026; Figure S2F–H).
Association between ΔsRAGE_after SIIT and 16‐week glycemic remission
Binary logistic regression analysis was performed to assess whether the ▵sRAGE_after SIIT was independently associated with glycemic remission. Given the limited sample size, only baseline variables significantly associated with remission (P < 0.05) were included in the multivariable models. In the unadjusted model, each 100 pg/mL increase in ΔsRAGE after SIIT was significantly associated with higher odds of achieving remission (OR = 1.314, 95% CI: 1.102–1.566, P = 0.002). This positive association remained robust after sequential adjustment for potential confounders, including age, sex, and BMI (Model 2), baseline FCP (Model 3), and serum creatinine (Model 4: OR = 1.228, 95% CI: 1.006–1.500, P = 0.044). No multicollinearity was detected among variables (all VIF < 5; Table 3).
Table 3.
Associations between ΔsRAGE and 16‐week glycemic remission in logistic regression models
| Variables | Model 1 OR (95% CI) | P‐value | Model 2 OR (95% CI) | P‐value | Model 3 OR (95% CI) | P‐value | Model 4 OR (95% CI) | P‐value |
|---|---|---|---|---|---|---|---|---|
| ΔsRAGE_after SIIT (per 100 pg/mL) | 1.314 (1.102–1.566) | 0.002 | 1.257 (1.035, 1.526) | 0.021 | 1.234 (1.012, 1.505) | 0.038 | 1.228 (1.006, 1.500) | 0.044 |
| Age, years | – | – | 0.997 (0.957, 1.038) | 0.883 | 0.997 (0.956, 1.04) | 0.898 | 0.991 (0.948, 1.036) | 0.691 |
| Sex, male/female | – | – | 0.497 (0.174, 1.417) | 0.191 | 0.645 (0.213, 1.949) | 0.437 | 1.085 (0.227, 5.175) | 0.919 |
| BMI, kg/m2 | – | – | 1.028 (0.860, 1.229) | 0.763 | 1.003 (0.835, 1.206) | 0.972 | 1.005 (0.837, 1.207) | 0.957 |
| FPG, mmol/L | – | – | – | – | 0.762 (0.620, 0.937) | 0.010 | 0.775 (0.628, 0.955) | 0.017 |
| Serum creatinine, μmol/L | – | – | – | – | – | – | 1.026 (0.971, 1.085) | 0.362 |
Model 1, unadjusted; Model 2, adjusted for age, sex, and BMI; Model 3, further adjusted for baseline FPG; Model 4, further adjusted for baseline serum creatinine. BMI, body mass index; CI, confidence interval; FPG, fasting plasma glucose; OR, odds ratio.
These findings indicate that the increase in sRAGE after SIIT was independently associated with glycemic remission during follow‐up.
Robustness analyses
The unadjusted model (Model 1) showed a significant association between ΔsRAGE and remission (AUC = 0.695, 95% CI: 0.586–0.805, P = 0.002), and the association remained stable after sequential adjustments (Model 4: AUC = 0.783, 95% CI: 0.690–0.876, P < 0.0001; Figure 2).
Figure 2.
ROC curves showing the association between the four logistic regression models and glycemic remission. Model 1: unadjusted; Model 2: adjusted for age, sex, and BMI; Model 3: further adjusted for baseline FPG; Model 4: further adjusted for baseline serum creatinine. CI, confidence interval; OR, odds ratio; ROC, receiver operating characteristic.
Model 4 was further assessed. The calibration curve indicated good concordance between predictions and observations, with a Brier score of 16.8 (95% CI: 12.8–20.9; Figure S3A). DCA further indicated that the model provided additional net clinical benefit across a wide range of threshold probabilities (Figure S3B,C). An optimism‐corrected AUC of 0.783 was obtained from internal validation with 1,000 bootstrap resamples, confirming the model's robustness (Figure S3D).
These findings further support a stable and independent association between the increased sRAGE after SIIT and glycemic remission.
Sensitivity analysis with a stricter definition of glycemic remission (HbA1c <6.5%)
A sensitivity analysis was performed using HbA1c <6.5% as the remission criterion. Based on this definition, 58 patients were classified as remission and 40 as non‐remission. Overall, the key findings remained consistent with the primary analysis (Table S2).
Specifically, ΔsRAGE after SIIT was significantly higher in the remission group than in the non‐remission group (303.31 ± 258.16 vs 88.33 ± 260.77 pg/mL, P < 0.001), and also remained significantly higher at the 1‐month follow‐up (P = 0.042). In unadjusted logistic regression analysis, each 100 pg/mL increase in ΔsRAGE after SIIT was associated with higher odds of achieving remission (OR = 1.373, 95% CI: 1.150–1.639, P < 0.001). This association remained significant after adjustment (Model 4: OR = 1.231, 95% CI: 1.011–1.498, P = 0.039; Table S3).
These sensitivity analyses indicate that the association between the increased sRAGE after SIIT and glycemic remission is robust and not materially affected by the choice of remission definition.
DISCUSSION
Our study found that sRAGE levels increased significantly after SIIT in patients with newly diagnosed T2DM, and the magnitude of the increase was independently associated with glycemic remission. The higher ▵sRAGE after SIIT was accompanied by improvements in β‐cell function. These findings suggest that the increase in sRAGE after SIIT may contribute to creating a metabolic environment favorable for β‐cell recovery and the restoration of normal glucose homeostasis.
After SIIT, sRAGE levels increased, which may be explained by improved glycemic control achieved through exogenous insulin, thereby alleviating the glucotoxic suppression of RAGE shedding or esRAGE transcription. In addition, insulin itself may promote the proteolytic cleavage of membrane‐bound RAGE, generating more soluble isoforms and leading to elevated circulating sRAGE levels 13 , 14 . However, since our study only included patients treated with SIIT and lacked a non‐SIIT control group, it was not possible to establish a causal relationship between SIIT and changes in sRAGE. In contrast, the concentration of AGEs remained unchanged throughout the treatment period, possibly due to their slow metabolic turnover and long half‐life. As irreversible covalent compounds formed through non‐enzymatic glycation and oxidative stress, AGEs are biochemically stable, and short‐term normalization of glucose and metabolism may be insufficient to reverse their accumulation in the body 16 .
The mechanisms underlying glycemic remission after intensive insulin therapy remain incompletely understood. Our findings showed that the increase in sRAGE after SIIT was positively correlated with ΔHOMA‐β and negatively correlated with ΔHOMA‐IR, suggesting that elevated sRAGE levels during SIIT may be accompanied improved insulin secretory capacity and reduced insulin resistance. Therefore, it is plausible that the observed association between ΔsRAGE and glycemic remission is accompanied by changes in β‐cell function. As mentioned earlier, mechanistically, sRAGE acts as a decoy receptor that competitively binds AGEs, preventing their interaction with membrane‐bound RAGE and thus blocking downstream signaling activation. The reduction of oxidative stress and inflammation induced by AGE‐RAGE could reduce β‐cell apoptosis and promote functional recovery 5 , 6 . Previous studies have also demonstrated that increasing sRAGE levels can alleviate oxidative stress and protect cells from damage in diabetes patients 17 , 18 . Obviously, it is biologically plausible that sRAGE may alleviate AGE–RAGE–related oxidative stress and inflammation. Such mechanisms could potentially be related to improved β‐cell function; however, given the extremely limited experimental evidence supporting a causal relationship between sRAGE and the preservation or restoration of β‐cell function. This conclusion is still speculative and requires confirmation in mechanistic studies.
Improving glycemic control increases sRAGE levels has already been previously reported by Devangelio et al. 19 who showed that improvement of metabolic control achieved either by insulin or oral therapy (gliclazide or metformin) significantly raised the levels of sRAGE, both in newly diagnosed diabetic subjects or in poorly controlled patients, and our research also confirmed that metabolic improvement was accompanied by an increase in sRAGE levels after SIIT. Meanwhile, our study also observed that sRAGE levels gradually decreased from the end of SIIT to the 1‐month follow‐up, although they remained higher than baseline. This pattern may reflect that during SIIT, sRAGE increases as a protective factor in response to insulin stimulation 14 and reduced glucose toxicity 13 ; however, as oxidative stress, inflammation, and metabolic dysregulation subside, the compensatory need for sRAGE may decrease 19 . Alternatively, after SIIT, although there is improved glycemic control, the persistent AGEs burden in patients with diabetes might continue to consume available sRAGE, leading to a gradual decline 20 . Whether this reduction represents a favorable or unfavorable prognostic signal requires further investigation. Additionally, previous studies have shown that statins 21 , 22 , thiazolidinediones 23 , 24 , and angiotensin‐converting enzyme‐1 inhibitors 25 can increase circulating sRAGE levels. It would be of interest to explore whether combining intensive insulin therapy with these agents could further enhance or sustain sRAGE levels, potentially leading to improved glycemic outcomes.
This study has several limitations. First, our study lacked a non‐SIIT control group, which limits the ability to determine whether the observed increase in sRAGE is causally related to improvements in β‐cell function or glycemic remission. Therefore, our findings should be interpreted as associative rather than causal. Second, we measured the total sRAGE levels without distinguishing between the esRAGE and cRAGE subtypes, which may have distinct biological functions. Third, although we adjusted for key confounding factors, the limited sample size restricted the inclusion of more covariates, making it impossible to fully exclude the influence of unmeasured or residual confounding factors. Finally, the relatively short follow‐up period limits the evaluation of whether post‐SIIT changes in sRAGE are associated with the durability of glycemic remission. Therefore, our findings mainly reflect short‐term associations, and longer follow‐up studies are needed to clarify their relevance to long‐term metabolic outcomes.
In summary, this study demonstrates that the elevation of sRAGE after SIIT is independently associated with short‐term glycemic remission in patients with newly diagnosed T2DM, and the elevation of sRAGE was accompanied by improvements in β‐cell function, suggesting that increased sRAGE may represent a compensatory biomarker associated with metabolic recovery. These findings provide new insights into the potential role of the AGE–RAGE axis in early diabetes remission. Future large‐scale and long‐term studies are needed to validate these observations and elucidate the mechanistic pathways linking sRAGE to β‐cell preservation and glycemic control.
AUTHOR CONTRIBUTIONS
Qimou Chen: Conceptualization, Methodology, Investigation, Formal analysis, Visualization, Writing—original draft, Writing—review and editing. Fang Wei: Conceptualization, Methodology, Formal analysis, Visualization, Writing—original draft. Jiajing Ma, Xuhui Li and Boyuan Liu: Methodology, Investigation, Writing—original draft; Yanbing Li: Conceptualization, Supervision, Project administration, Funding acquisition, Writing—review and editing.
FUNDING
This study was supported by a grant from the National Key R&D Program of China (2018YFC1314102).
DISCLOSURE
The authors declare no conflict of interest.
Approval of the research protocol: This study complies with the Declaration of Helsinki.
Informed consent: All participants provided written informed consent.
Registry and the registration no. of the study/trial: The trial was approved by the research ethics board of Sun Yat‐sen University (approval number: 2019–174), and this study was from a trial registered on ClinicalTrials.gov (NCT03972982).
Animal studies: N/A.
Supporting information
Table S1. Baseline characteristics of study participants at enrollment.
Table S2. Comparison of Metabolic Parameters Between Remission and Non‐remission Groups. (Remission criteria of HbA1c <6.5%).
Table S3. Associations between ΔsRAGE and 16‐week glycemic remission in logistic regression models. (Remission criteria of HbA1c <6.5%).
Figure S1. Effects of SIIT on glycemic and metabolic parameters.
Figure S2. Comparison of glycemic and metabolic responses between remission and non‐remission groups during and after SIIT.
Figure S3. Validation of Model 4 predicting glycemic remission.
ACKNOWLEDGMENTS
Not applicable.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.
REFERENCES
- 1. Collaborators GBDD . Global, regional, and national burden of diabetes from 1990 to 2021, with projections of prevalence to 2050: A systematic analysis for the global burden of disease study 2021. Lancet 2023; 402: 203–234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Li Y, Xu W, Liao Z, et al. Induction of long‐term glycemic control in newly diagnosed type 2 diabetic patients is associated with improvement of beta‐cell function. Diabetes Care 2004; 27: 2597–2602. [DOI] [PubMed] [Google Scholar]
- 3. Weng J, Li Y, Xu W, et al. Effect of intensive insulin therapy on beta‐cell function and glycaemic control in patients with newly diagnosed type 2 diabetes: A multicentre randomised parallel‐group trial. Lancet 2008; 371: 1753–1760. [DOI] [PubMed] [Google Scholar]
- 4. Prasad K. Low levels of serum soluble receptors for advanced glycation end products, biomarkers for disease state: Myth or reality. Int J Angiol 2014; 23: 11–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Khalid M, Petroianu G, Adem A. Advanced glycation end products and diabetes mellitus: Mechanisms and perspectives. Biomolecules 2022; 12: 542. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Erusalimsky JD. The use of the soluble receptor for advanced glycation‐end products (sRAGE) as a potential biomarker of disease risk and adverse outcomes. Redox Biol 2021; 42: 101958. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Chen Q, Liu L, Ke W, et al. Association between the soluble receptor for advanced glycation end products and diabetes mellitus: Systematic review and meta‐analysis. BMC Endocr Disord 2024; 24: 232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Thomas MC, Woodward M, Neal B, et al. Relationship between levels of advanced glycation end products and their soluble receptor and adverse outcomes in adults with type 2 diabetes. Diabetes Care 2015; 38: 1891–1897. [DOI] [PubMed] [Google Scholar]
- 9. Aubert CE, Michel PL, Gillery P, et al. Association of peripheral neuropathy with circulating advanced glycation end products, soluble receptor for advanced glycation end products and other risk factors in patients with type 2 diabetes. Diabetes Metab Res Rev 2014; 30: 679–685. [DOI] [PubMed] [Google Scholar]
- 10. Colhoun HM, Betteridge DJ, Durrington P, et al. Total soluble and endogenous secretory receptor for advanced glycation end products as predictive biomarkers of coronary heart disease risk in patients with type 2 diabetes: An analysis from the CARDS trial. Diabetes 2011; 60: 2379–2385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Koska J, Saremi A, Howell S, et al. Advanced glycation end products, oxidation products, and incident cardiovascular events in patients with type 2 diabetes. Diabetes Care 2018; 41: 570–576. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Yamagishi S, Nakamura K, Inoue H. Possible participation of advanced glycation end products in the pathogenesis of osteoporosis in diabetic patients. Med Hypotheses 2005; 65: 1013–1015. [DOI] [PubMed] [Google Scholar]
- 13. Sabbatinelli J, Castiglione S, Macri F, et al. Circulating levels of AGEs and soluble RAGE isoforms are associated with all‐cause mortality and development of cardiovascular complications in type 2 diabetes: A retrospective cohort study. Cardiovasc Diabetol 2022; 21: 95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Lam JK, Wang Y, Shiu SW, et al. Effect of insulin on the soluble receptor for advanced glycation end products (RAGE). Diabet Med 2013; 30: 702–709. [DOI] [PubMed] [Google Scholar]
- 15. Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: Diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med 1998; 15: 539–553. [DOI] [PubMed] [Google Scholar]
- 16. Lapolla A, Reitano R, Seraglia R, et al. Evaluation of advanced glycation end products and carbonyl compounds in patients with different conditions of oxidative stress. Mol Nutr Food Res 2005; 49: 685–690. [DOI] [PubMed] [Google Scholar]
- 17. Koutroumani N, Partsalaki I, Lamari F, et al. Protective mechanisms against oxidative stress and angiopathy in young patients with diabetes type 1 (DM1). J Pediatr Endocrinol Metab 2013; 26: 309–317. [DOI] [PubMed] [Google Scholar]
- 18. Reis JS, Veloso CA, Volpe CM, et al. Soluble RAGE and malondialdehyde in type 1 diabetes patients without chronic complications during the course of the disease. Diab Vasc Dis Res 2012; 9: 309–314. [DOI] [PubMed] [Google Scholar]
- 19. Devangelio E, Santilli F, Formoso G, et al. Soluble RAGE in type 2 diabetes: Association with oxidative stress. Free Radic Biol Med 2007; 43: 511–518. [DOI] [PubMed] [Google Scholar]
- 20. Ramasamy R, Yan SF, Schmidt AM. The diverse ligand repertoire of the receptor for advanced glycation endproducts and pathways to the complications of diabetes. Vasc Pharmacol 2012; 57: 160–167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Quade‐Lyssy P, Kanarek AM, Baiersdorfer M, et al. Statins stimulate the production of a soluble form of the receptor for advanced glycation end products. J Lipid Res 2013; 54: 3052–3061. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Tam HL, Shiu SW, Wong Y, et al. Effects of atorvastatin on serum soluble receptors for advanced glycation end‐products in type 2 diabetes. Atherosclerosis 2010; 209: 173–177. [DOI] [PubMed] [Google Scholar]
- 23. Tan KCB, Chow WS, Tso AWK, et al. Thiazolidinedione increases serum soluble receptor for advanced glycation end‐products in type 2 diabetes. Diabetologia 2007; 50: 1819–1825. [DOI] [PubMed] [Google Scholar]
- 24. Oz Gul O, Tuncel E, Yilmaz Y, et al. Comparative effects of pioglitazone and rosiglitazone on plasma levels of soluble receptor for advanced glycation end products in type 2 diabetes mellitus patients. Metabolism 2010; 59: 64–69. [DOI] [PubMed] [Google Scholar]
- 25. Forbes JM, Thorpe SR, Thallas‐Bonke V, et al. Modulation of soluble receptor for advanced glycation end products by angiotensin‐converting enzyme‐1 inhibition in diabetic nephropathy. J Am Soc Nephrol 2005; 16: 2363–2372. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Baseline characteristics of study participants at enrollment.
Table S2. Comparison of Metabolic Parameters Between Remission and Non‐remission Groups. (Remission criteria of HbA1c <6.5%).
Table S3. Associations between ΔsRAGE and 16‐week glycemic remission in logistic regression models. (Remission criteria of HbA1c <6.5%).
Figure S1. Effects of SIIT on glycemic and metabolic parameters.
Figure S2. Comparison of glycemic and metabolic responses between remission and non‐remission groups during and after SIIT.
Figure S3. Validation of Model 4 predicting glycemic remission.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.