CMP was provided from Shengyuan Company (Shandong, China). Casein (CS) was purchased by Hualing Company (Gansu, China). Whey protein (WPC) was purchased from Glanbia Company (Kilkenny, Ireland). ELISA assay kits forinterleukin-6 (IL-6), interleukin-10 (IL-10), interleukin-2 (IL-2), and tumor necrosis factor-alpha (TNF-α) were purchased from Xinbosheng Biotechnology Co., Ltd. (Shenzhen, China). MDA, Superoxide Dismutase (SOD) and GSH-Px detection kits were purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). ROS detection kits were purchased from Shanghai Beibo Biotechnology Co., Ltd. (Shanghai, China). Pepsin (1:10000), trypsin (1:250), streptozotocin (STZ), Periodic acid-schiff (PAS) staining kit, hematoxylin–eosin (H&E) staining kit, Masson’s trichrome staining kit, and immunohistochemical primary antibodies (CD86, CD163) were purchased from Solarbio Technology Co., Ltd. (Beijing, China). Sitagliptin phosphate tablets (SIG) were purchased from Moshadong Pharmaceutical Co., Ltd. (Zhejiang, China).

Simulated gastric and intestinal fluids were prepared for in vitro digestion using a semi-dynamic simulated digestion model (17). To demonstrate the in vitro digestion characteristics of CMP, CS and WPC was used in different ratios for comparison. Therefore, a series of 2% milk protein solutions with varying ratios were prepared as detailed in Table 1. For the OPA assay, 30 μL of each sample was mixed with 240 μL of the OPA reagent. The absorbance was then measured at 340 nm (OD₃₄₀) using a Multiskan™ FC microplate reader (Thermo, Massachusetts, United States) (18). A standard curve was generated using serine, and the degree of hydrolysis was calculated according to Equation 1.

In the formula, Serine NH₂ is the content of serine amino groups per gram of protein, mmol/g. β and α are constants of milk protein. h_tot is the total number of milk protein peptide bonds.

The anti-α-amylase activity was determined by measuring the absorbance at 540 nm (OD₅₄₀) using a Multiskan™ FC microplate reader (19). Two control groups were established, an experimental control without the starch solution and a sample control without the test sample. The α-amylase inhibition rate was then calculated according to Equation 2.

In the formula, the OD value represents absorbance.

The hydroxyl free radical (·OH) scavenging ability was determined by measuring the absorbance of the reaction mixture at 536 nm (OD₅₃₆) using a Multiskan™ FC microplate reader. The reaction mixture without any antioxidants served as the control group (20). Vitamin C (Vc) at a concentration of 0.1 mg/mL was used as the positive control, the ·OH scavenging ability of the samples was calculated according to Equation 3.

In the formula, the OD value represents absorbance.

The total reducing power was assessed by the potassium ferricyanide reduction method, measured at 700 nm (OD₇₀₀) using a Multiskan™ FC microplate reader. The reducing ability was expressed as the absorbance value. Vc at a concentration of 0.1 mg/mL was used as the positive control (21).

This study was approved by the Institutional Ethics Committee of Northwest Minzu University (ethics approval number: xbmu-sm-2024107). Male C57BL/6 J mice (6–8 weeks old, N = 48), weighing 22–25 g, were purchased from the Lanzhou Institute of Veterinary Medicine, Chinese Academy of Agricultural Sciences (Lanzhou, China). The animals were housed in a controlled environment with a constant temperature of 20 ± 5 °C and humidity of 50% ± 5%, maintained on a 12-h light/dark cycle, with free access to standard maintenance feed and pure water. After 1 week of acclimatization, the mice were randomly divided into two groups: a normal diet group (fat content 4.3% w/w, calorie content 10%, Lanzhou Veterinary Research Institute, Chinese Academy of Agricultural Sciences, Lanzhou, China) (NC group, N = 12) and a high-fat diet group (fat content 35% w/w, calorie content 60%, same source) (HFD group, N = 36). At the end of the fourth week, mice in the HFD group were intraperitoneally injected with STZ, dissolved in 0.05 M sterile sodium citrate buffer (pH 4.5), at a dose of 50 mg/kg for three consecutive days. FBG and random blood glucose levels were measured via caudal vein puncture using a handheld blood glucose meter. Mice with FBG ≤ 11.1 mmol/L received additional STZ injections at 30 mg/kg for 2 days. FBG was re-evaluated 3 days and 1 week after the final injection. Mice with FBG > 11.1 mmol/L at both time points were considered successfully modeled for T2DM. The normal control group received intraperitoneal injections of an equal volume of sodium citrate buffer concurrently (22). Following successful modeling, T2DM mice were randomly assigned to three groups (N = 12 per group): the type 2 diabetes mellitus model (DM) group, the SIG group, and the CMP group. Throughout the 8-week intervention, all mice received daily intragastric administration. The NC and DM groups were administered with pure water at a dose of 10 mg/kg/day, whereas the SIG and CMP groups received SIG (10 mg/kg/day) and CMP (200 mg/kg/day), respectively. Body weight and FBG levels were recorded every 2 weeks. The detailed experimental procedure is illustrated in Figure 1.

After 8 weeks of intervention, the animals were fasted for 12 h and anesthetized via intraperitoneal injection of 10% (w/v) pentobarbital sodium. Whole blood was collected from the inferior vena cava into 2 mL centrifuge tubes. The blood samples were allowed to stand at 25 °C for 2 h and then centrifuged at 247 × g for 15 min at 4 °C. The supernatant was transferred to 1.5 mL centrifuge tubes and stored at −80 °C. The mice were sacrificed by cervical dislocation, and the kidneys, liver, ileum, and pancreas were collected. The collected tissues were divided into two portions: one stored at −80 °C for future use, while the other was fixed in a fixative solution for histological analysis. Visceral adipose tissue was collected for flow cytometric analysis.

FBG was measured via tail vein puncture using a blood glucose meter (GLM-77, YASEE Biotech, Qingdao, China). The levels of IL-6, IL-10, IL-2, and TNF-α in serum were detected with ELISA kits. Serum liver function markers [aspartate aminotransferase (AST), alanine aminotransferase (ALT)] and renal function indicators [creatinine (CREA), urea nitrogen (UREA), uric acid (UA)] were measured using a fully automated biochemical analyzer (SAL9000, Mindray, Shenzhen, China). Oxidative stress-related markers in liver homogenates were measured separately according to the MDA, SOD, GSH-Px, and ROS assay kits.

The visceral adipose tissue of mice was digested using a mixed enzyme solution, and M1 and M2 macrophages were identified by flow cytometry. M1 macrophages were labeled with CD86⁺, and M2 macrophages were labeled with CD163⁺; these cells were the quantified using flow cytometry (BriCyte E6, Mindray, Shenzhen, China) (22).

The renal sections were stained with H&E, Masson’s trichrome, and PAS stains, and the pathological changes in renal tissue were observed under a light microscope (23). Liver, pancreas, and ileum sections were examined following H&E staining. Immunohistochemical staining was performed on pancreatic paraffin sections, which were incubated with primary antibodies against CD86 and CD163, respectively, to indicate M1 and M2 macrophage expression (24).

All data were analyzed in triplicate using SPSS software (IBM SPSS Statistics 27, IBM, New York, United States). Intergroup differences were evaluated by the Waller–Duncan post hoc test following one-way analysis of variance (ANOVA). Bars within the same panel labeled with different lowercase letters indicate statistically significant differences (p < 0.05). Data are presented as mean ± standard deviation. Tissue section analysis was performed using ImageJ for Windows 64-bit (USA National Institutes of Health, Maryland, United States). GraphPad Prism software (GraphPad Prism 9.5, GraphPad Inc., California, United States) was used for data visualization and statistical plotting. Pearson correlation coefficients were calculated to assess pairwise associations among variables. The absolute value of the correlation coefficient (|r|) was used to represent correlation strength, with larger |r| values indicating stronger associations. In the figure, edge width is proportional to |r| and is used solely to improve visualization readability.

A semi-dynamic digestion model was used to simulate the digestion of milk proteins, with the results presented in Figure 2. During the simulated gastric phase, the degree of hydrolysis of 100% CMP was significantly higher than that of the 100% CS groups at 30, 60, 90, and 120 min (Figure 2A, P < 0.05). Throughout the intestinal phase, the degree of hydrolysis of the 100% CS and 100% CMP groups was significantly lower than that of the 20% CS + 80% WPC and 100% WPC groups at all time points, while no significant difference was observed between the 100% CS and 100% CMP groups (Figure 2A, P < 0.05). The in vitro α-amylase inhibition rates of 100% CS and 100% CMP hydrolysate were significantly higher than those of the other groups (Figure 2B, P < 0.05). The ·OH free radical scavenging rate of 100% CMP hydrolysate was lower than that of 50% CS + 50% WPC hydrolysate, and there was no significant difference between 100% CMP hydrolysate and the other three groups (Figure 2C). The ferricyanide reducing power of the 100% CMP hydrolysate was significantly stronger than that of the 100% CS hydrolysate, but weaker than that of the 100% WPC hydrolysate (p < 0.05). There was no significant difference between the remaining two groups (Figure 2D). Therefore, CMP was selected for subsequent in vivo experiments due to its excellent digestibility, anti-α-amylase activity, and antioxidant properties.

To explore the effect of CMP on the basic physiological indices of T2DM mice, body weight and FBG levels were monitored throughout 8 weeks intervention period. The results were presented in Figure 3. FBG levels in the NC group stayed within the normal range for the entire intervention period. At the beginning of the intervention (week 0), FBG levels in the DM, SIG, and CMP groups all exceeded 11.1 mmol/L. The DM group remained hyperglycemic throughout the 8-week period. From weeks 2 to 8, FBG levels in both the SIG and CMP groups were significantly lower than in the DM group (p < 0.05), with no difference between them (Figure 3A). Body weight monitoring showed that the NC group exhibited continuous weight gain throughout the intervention, whereas the DM group showed sustained weight loss. CMP and SIG interventions gradually reversed this weight loss within 4 weeks, resulting in no difference compared to the NC group. At 6 weeks, the SIG group exhibited the highest weight gain rate. By 8 weeks, weight changes in all groups slowed, and body weight tended to stabilize (Figure 3B).

To explore whether CMP improves oxidative stress and macrophage polarization in T2DM, oxidative stress-related indicators and macrophage subtypes in visceral adipose tissue and pancreatic tissue were examined following CMP intervention, with the results presented in Figure 4. Compared with the NC group, the levels of oxidative stress markers MDA and ROS in the DM group were significantly increased (Figures 4A,B, P < 0.05). In contrast, the activity of the antioxidant enzyme GSH-Px was significantly decreased (Figure 4C, P < 0.05), while SOD activity was only slightly decreased (Figure 4D, P > 0.05). Compared with the DM group, both CMP and the positive control SIG groups restored the oxidative stress-related indicators to levels comparable with the NC group (Figures 4A–D). Meanwhile, compared with the NC group, the DM group exhibited increased expression of both M1 and M2 macrophages in visceral adipose and pancreatic tissue (Figures 4E–L, P < 0.05). However, the increase in M1 macrophages is greater than that in M2 macrophages, resulting in an elevated M1/M2 ratio at this time. Compared with the DM group, both CMP and SIG showed a similar regulatory trend. Both treatments reduced M1 macrophages expression (Figures 4F,K, P < 0.05) and increased M2 macrophages expression (Figure 4G, P > 0.05; L, p < 0.05). Critically, both interventions effectively lowered the M1/M2 ratio (Figures 4H,M, P < 0.05).

We further investigated whether CMP alleviates systemic chronic inflammatory injury in T2DM mice by assessing serum inflammatory factors, liver and kidney function, and the histology of the liver, ileum, and pancreas. The results are presented in Figures 5, 6. Compared with the NC group, serum levels of the pro-inflammatory cytokines IL-2, IL-6, and TNF-α were markedly increased in the DM group (p < 0.05). CMP and SIG demonstrated a significant downward trend in IL-2 and TNF-α levels, achieving comparable effects with no significant difference from the NC group. IL-6 showed only a downward trend (Figure 5A). Serum biochemical analysis revealed that ALT and AST activities were significantly elevated in the DM group compared to the NC group. However, treatment with CMP and SIG resulted in a significant reduction of these enzyme levels (p < 0.05). And the effect was more pronounced in the SIG group than in the CMP group (Figures 6A,B). The DM group exhibited severe hepatic pathology, including disorganized cell arrangement, inflammatory infiltration, cellular swelling, marked steatosis, and narrowed sinusoids. Following intervention with CMP and SIG, these histopathological changes were visibly improved. Notably, the marked steatosis was significantly alleviated, as confirmed by quantitative analysis (Figures 5B, 6C, P < 0.05). The ileal villi in the DM group were markedly shortened, with a significant increase in adipose vacuolar area at the villus tips. These ileal injuries were significantly ameliorated after CMP and SIG treatments (Figures 5D,E, 6D,E, P < 0.05). The NC group exhibited islets with normal morphology, characterized by clear boundaries, regular shape, neatly arranged cells, deeply stained nuclei, and abundant cytoplasm. In contrast, the DM group displayed significant islet damage, including blurred boundaries, atrophy, cellular disarray, enlarged nuclei, and mild heterogeneity. Treatment with CMP and SIG effectively restored islet morphology and increased their area compared with the DM group (Figures 5F, 6F, P > 0.05). Serum levels of UREA, CREA, and UA were significantly elevated in the DM group compared to the NC group. However, intervention with CMP and SIG led to a significant reduction in these markers (Figures 6G–I, P < 0.05), although the reduction in CREA within the CMP group was not statistically significant. Mice in the DM group exhibited severe renal pathology, characterized by loosely arranged tissue, a disordered structure, marked mesangial expansion, and significantly enlarged glomeruli. However, these pathological alterations were markedly improved following intervention with CMP and SIG (Figures 5G, 6J, P < 0.05). The DM group exhibited significant thickening of the glomerular basement membrane, mesangial matrix expansion, and a pronounced increase in the deposition of PAS-positive substances along the glomerular capillary wall. Following CMP and SIG interventions, these pathological features were significantly ameliorated (Figures 5H, 6K, P < 0.05). Moreover, compared with the NC group, the deposition of blue collagen fibers in the glomeruli and renal tubulointerstitial was markedly increased in the DM group, while CMP and SIG treatments significantly improved this fibrotic change (Figures 5I, 6L, P < 0.05).

Pearson correlation analysis was conducted to examine the relationships among FBG, levels of oxidative stress markers, markers of macrophage polarization, expression of inflammatory cytokine, and indicators of systemic chronic inflammation (Figure 7). FBG levels were positively correlated with oxidative stress factors (MDA, ROS), M1 macrophage levels, and the M1/M2 ratio, while negatively correlated with the activities of antioxidant enzymes (GSH-Px, SOD). Oxidative stress factors were positively correlated with pro-inflammatory cytokines (IL-2, IL-6, TNF-α), whereas antioxidant parameters were positively correlated with the anti-inflammatory cytokine IL-10. Pro-inflammatory cytokines showed positive correlations with indicators of liver and kidney function damage, including ALT, AST, CREA, UREA, and UA. They were also positively correlated with histopathological markers of damage, such as hepatic adipose infiltration, glomerular area, renal glycosylation, and fibrosis. Conversely, anti-inflammatory factor of IL-10 was positively correlated with pancreatic islet area and intestinal villus length.