Diabetic induction in mice differentially alters body weight and body composition depending on genetic background and sex

We first evaluated in phase 1 (Fig. 1A) the body weight in the different mouse models i.e. HFD/STZ- or HFD-induced 129/Sv and C57 mice, and MKR mice. We found that the treatment with HFD alone resulted in significant weight gain only in male C57 mice while the HFD/STZ combination did not induce weight gain, as previously shown (Yin et al. 2020) (Fig. 1B). As expected (Yin et al. 2020), the STZ treatment applied at 10 weeks of age induced a weight loss in both 129Sv and C57 genetic backgrounds, except for the female C57 (Yin et al. 2020), which is partly corrected until 20 weeks of age (Supplementary Fig. 1). Additionally, the MKR mice fed with normal chow did not display any weight modification when they where compared to FVB mice at 20 weeks of age (Fig. 1B).

Fig. 1

The background, sex, and method of T2DM induction significantly influence body weight, body composition, obesity status, and glycemia in mice. A schematic representation of the study design (A). Body weight assessment (B) was conducted in male (M) and female (F) mice induced with HFD/STZ (HFD/STZ) in both 129/Sv and C57 strains, as well as HFD-induced (HFD) mice in both 129/Sv and C57 strains, along with MKR mice from the FVB background, all evaluated at 20 weeks of age. Body composition, determined by NMR, included fat mass, total body fluids, and lean mass (C). The fat-to-lean mass ratio (D) was calculated using NMR data, with dotted lines indicating the thresholds for determining pre-obese (0.466) or obese (0.658) mice. Non-fasting blood glucose was measured using an Accu-check glucometer (E), and dotted lines indicate the hyperglycemia threshold (250 mg/dl). (n = 4 for all groups) All data are shown as mean ± standard deviation. Non-parametric One-Way ANOVA for all genetic backgrounds except MKR, non-parametric Mann–Whitney test; * indicate significance (with P < 0.05 *, < 0.005**) relative to controls (CTRL)

Then, we investigated the mouse body composition (fat, fluid and lean) by NMR spectroscopy (Fig. 1C) and calculated the fat-to-lean mass ratio (FLMR) (Fig. 1D) to evaluate the obesity status of mice, according to Woo et al. (2013). While no change could be recorded in HFD/STZ-induced 129/Sv mice, with an FLMR under the pre-obese threshold (0.421 for males and females) (Fig. 1D), the impact of the HFD/STZ-induction in the C57 strain differed between sex, with an FLMR under the pre-obese threshold (mean ratio = 0.38) in females (Fig. 1D) despite a significant increase in fat mass (Fig. 1C), and an FLMR under the pre-obese threshold (mean ratio = 0.28) in males despite increased levels of fat, fluid and lean (Fig. 1C). In HFD-treated mice, we recorded a significant increase in fat mass (Fig. 1C), with an obese status reached in female 129/Sv and male C57 (mean ratios of 0.71 and 0.83, respectively), and a pre-obese status in male 129/Sv and female C57 (mean ratios of 0.64 and 0.52, respectively) (Fig. 1D). Finally, male and female MKR mice remained under pre-obese threshold (Fig. 1D), despite a significant decrease in fat, fluid and lean levels only in males (Fig. 1C).

Thus, taken together, these results suggested that T2DM-induced mice likely develop either obesity (HFD-induced female 129SV/1 and male C57BL/6 mice), a pre-obese status (HFD-induced male 129SV/1 and female C57BL/6 mice) or no obesity (male and female HFD/STZ-induced and MKR mice), thus giving the opportunity to compare the muscle metabolism in these different situations.

Non-fasting blood glucose is differentially increased in pre-diabetic and diabetic mouse models.

We compared blood glucose levels in different mouse populations following HFD/STZ treatment. All mouse populations exhibited significant hyperglycemia indicative of T2DM (blood glucose levels exceeding 250 mg/dl), though with varying severity (Fig. 1E). Notably, in the 129/Sv strain, males induced by HFD/STZ displayed a 33% higher increase in blood glucose levels compared to females (566 mg/dl vs. 381 mg/dl, respectively). For HFD-induced, only C57 males reached a significant glycemic value indicative of T2DM (246 mg/dl) (Fig. 1E). In the MKR strain, only males became hyperglycemic (437 mg/dl in males, 188 mg/dl in females).

These results show that the different mouse populations differentially reached a T2DM compatible hyperglycaemia.

The skeletal muscle metabolism is differentially altered in the diabetic mouse models

We compared the muscle energy metabolism of relevant mouse populations exhibiting T2DM-compatible hyperglycemia associated with different obesity statuses. Using unbiased metabolomics based on nuclear magnetic resonance spectroscopy (NMR) we compared the concentration of main metabolites from the different energetic pathways of the quadriceps muscle of male (Supplementary Fig. 2A and 2C) and female (Supplementary Fig. 2B and 2D) HFD/STZ-induced 129/Sv (Supplementary Fig. 2A and 2B) and C57 (Supplementary Fig. 2C and 2D), male HFD-induced C57 (Supplementary Fig. 2E), and male MKR mice (Supplementary Fig. 2F), and we expressed the differences in fold change compared to controls (Fig. 2).

Fig. 2

The skeletal muscle metabolism is differentially altered among various diabetic mouse models, influenced by genetic background, sex, and the mode of T2DM induction. Quadriceps muscle metabolomics at 20 weeks, evaluated by NMR, is presented as fold change compared to control mice. Results are shown for male (M) and female (F) HFD/STZ-induced (HFD/STZ) 129/Sv mice (A, B) and C57 mice (C, D), as well as male HFD-induced (HFD) C57 mice (E) and male MKR mice (F). Metabolite categories include AA (amino acids), Lip (lipid pathways), C.H (carbohydrate pathways), and Nuc (nucleotides). (n = 4 for all groups) All data are shown as mean ± standard deviation. Non-parametric Mann–Whitney test for all genetic backgrounds; * indicate significance (with P < 0.05 *, < 0.005**) relative to controls (CTRL)

In HFD/STZ-induced 129/Sv mice, we observed muscular glucose accumulation (Fig. 2A and B) and a decrease in lactate concentrations in HFD/STZ-induced C57 mice (Fig. 2C and D), suggesting a depression in carbohydrate anaerobic catabolism. Additionally, in both strains, we noted a significant increase in 3-hydroxybutyrate, indicating an increased reliance on lipid catabolism for energy needs, particularly in males (Fig. 2A–C).

Interestingly, a sex-dependent alteration in amino acid (AA) homeostasis was found in mouse quadriceps, affecting both essential (EAA) and non-essential (NAA) amino acids (Fig. 2A–F). HFD/STZ-induced 129/Sv males displayed a significant increase in EAA and NAA levels, while females showed low concentrations of EAA and high concentrations of NAA (Fig. 2A, B). In HFD/STZ-induced C57 males, NAA increased, with contrasted alterations in EAA levels, while females displayed decreased EAA levels (Fig. 2C, D). Male HFD-induced C57 mice showed an overall decrease in EAA and NAA concentrations and an increase in lysine (Fig. 2E). In MKR males, AA metabolism appeared less disturbed, with only a tendency to increase EAA and NAA concentrations (Fig. 2F).

Taken together these results strongly suggest that, in the different diabetic mouse models, all the main metabolic pathways, including carbohydrates, lipids, and proteins are perturbed in the fast-twitch muscle i.e. the quadriceps, but with genetic background, gender, and type of induction specificities.

The expression pattern of genes involved in energy metabolism is differentially altered in the muscle of diabetic mouse models, depending on the sex and the type of induction

In order to further investigate diabetes-induced alterations in mouse muscle metabolism, we evaluated by RT-qPCR the mRNA expression patterns of the main components of the carbohydrate, lipid, and amino-acid (AA) metabolic pathways. In terms of carbohydrate metabolism, male HFD/STZ-induced mice exhibited diminished GLUT4 expression alongside an increase in PDK4 (Fig. 3A and E). Notably, male 129/Sv mice displayed reduced levels of PFKM and MTC1 (Fig. 3B and F). In female HFD/STZ-induced mice, especially in C57, there was a decline tendency in GLUT4 (Fig. 3A). Conversely, HFD-induced male C57 mice demonstrated heightened expression of PFKM, GAPDH, PDK2, PDK4, and MCT1, suggesting a potential hindrance in pyruvate transfer to mitochondria (Fig. 3B–F). Thus, while the muscle carbohydrate catabolism was severely depressed in HFD/STZ-induced diabetic males, it was quite normal in females and MKR, and likely enhanced in HFD-induced C57BL/6 male mice.

Fig. 3

The mRNA expression levels of key components in muscle metabolic pathways exhibit selective alterations in distinct T2DM mouse models. Quantification of mRNA gene expression in (i) the carbohydrate pathway, including GLUT4 (A), PFKM (B), GAPDH (C), PDK2 (D), PDK4 (E), and MCT1 (F); (ii) the lipid pathway, including CD36 (G), CPT1 (H), ACC1 (I), ACC2 (J), and DGAT2 (K); (iii) mitochondrial metabolism, including PGC1α (L), and UCP3 (M); and (iv) amino-acid metabolism, including MuRF1 (N), ATF4 (O), ATG1 (P), SLC7A5 (Q), SLC36A1 (R), and SLC3A2 (S), was performed by RT-qPCR. Samples were obtained from male (M) and female (F) HFD/STZ-induced (HFD/STZ) 129/Sv and C57 mice, HFD-induced (HFD) male C57 mice, and male MKR mice from the FVB background at 20 weeks of age. mRNA expression levels were normalized to RPL13 and 26S mRNA. (n = 4 for all groups) All data are shown as mean ± standard deviation. Non-parametric Mann–Whitney test for all genetic backgrounds except C57 males, non-parametric One-Way ANOVA; * indicate significance (with P < 0.05 *, < 0.005**) relative to controls (CTRL)

Turning to lipid metabolism, male HFD/STZ-induced mice displayed an overall increase in lipid uptake, with elevated CD36 in C57, coupled with reduced fatty acid biosynthesis (ACC2 down-regulation in 129/Sv and ACC1 in C57) (Fig. 3G, I and J). Female HFD/STZ-induced mice solely exhibited a decrease in fatty acid biosynthesis (Fig. 3I and J). In HFD-induced C57 males, there was significant up-regulation of CD36, CPT1, ACC2, and DGAT2, alongside ACC1 down-expression (Fig. 3 G, H, J, K and I) I. These data suggest that the muscle lipid metabolism globally adapt in two different ways either coherent with a boost in β-oxidation and inhibition in fatty acid storage, for non-obese mice or a boost in lipid biosynthesis in diabetic obese mice.

Regarding amino acid metabolism, there was a widespread alteration in AA transport gene expression. Reduced expression was observed for SLC7A5 in HFD/STZ-induced male 129/Sv, HFD-induced male C57, and male MKR mice (Fig. 3Q). SLC36A1 demonstrated decreased expression in HFD/STZ-induced male 129/Sv and MKR mice (Fig. 3R). SLC3A2 exhibited decreased expression in HFD/STZ-induced female C57 and male MKR mice, but increased expression in HFD/STZ-induced male C57 and HFD-induced male C57 (Fig. 3S). ATF4 expression increased in HFD/STZ- and HFD-induced male C57 mice, accompanied by a decrease in ATG1 expression (Fig. 3O and P). HFD/STZ-induced female 129/Sv showed increased MuRF1 expression and decreased ATF4 expression (Fig. 3N and O). In MKR muscles, ATF4 and ATG1 expression decreased, indicating a disruption in the balance between protein synthesis and degradation (Fig. 3O and P). These results provide the first lines of evidence that the traffic of free AA, as well as, their use to synthesize proteins, are compromised in diabetic muscles.

Taken together, these data suggested a differential and global adaptation of muscle metabolism to the different diabetic conditions, strongly depending on the gender and the type of induction. This led us to draw three different signatures of muscle metabolism adaptation to T2DM, i.e. non-obese HFD/STZ-induced male and female C57 mice, characterized by a β-oxidation shift with depressed glycolysis, and either (1) increased (males) or (2) decreased (females) free amino acid concentrations, and (3) obese HFD-induced male C57 mice, representing carbohydrate, lipid, and protein hypermetabolism.

Systemic T2DM hallmarks are differentially improved by specific exercise programs

We examined in phase 2 (Fig. 1A) the effects on T2DM metabolic hallmarks of three different exercise paradigms (TP1, 2, and 3, detailed in methods) aimed at specifically targeting the metabolic impairments. To do this, we selected the three mouse populations showing a specific metabolic signature, which in addition shared the same genetic background i.e. C57, to avoid bias.

We first assessed the effectiveness of each exercise program in targeting altered metabolic pathways, using HFD-induced C57 male muscles as a proof-of-concept. As expected, TP2 significantly reduced PDK4 (Fig. 4A) and ACC2 (Fig. 4B) expression, TP3 reduced ACC2, and TP1 had no effect.

Fig. 4

Differential effects of specific exercise programs on body assessment and systemic diabetes hallmarks in various T2DM mouse models. High-intensity TP1, a combination of high and low intensity TP2, and low-intensity TP3 swimming protocols were evaluated in HFD/STZ-induced male (HFD/STZ M) and female (HFD/STZ F) C57, and HFD-induced male (HFD M) C57 mice at 32 weeks of age. Results were compared to relevant untrained diabetic groups (UTG). Quadriceps mRNA gene expression levels of PDK4 (A) and ACC2 (B) in HFD-induced muscles were quantified by RT-qPCR and normalized with RPL13 and 26S mRNA. Body weight (C) was directly recorded using a scale. The fat-to-lean mass ratio (D) was calculated from NMR Minispec LF90 data, with dotted lines indicating thresholds for pre-obese (0.466) or obese (0.658) mice. Random blood glucose (E) was measured using the Accu-check glucometer, with dotted lines indicating the hyperglycemia threshold (250 mg/dl). Glucose tolerance (IPGTT) (F, H, J) and insulin tolerance (IPITT) (L, N, P) tests were conducted, and statistical analysis of the area under the curve (AUC) was determined for IPGTT (G, I, K) and IPITT (M, O, Q). (n = 5 for all groups) All data are shown as mean ± standard deviation. Non-parametric One-Way ANOVA for all groups; * indicate significance (with P < 0.05 *, < 0.005**) relative to untrained (UTG)

We then explored the impact of each exercise on mouse obesity status. If no change in body weight was observed (Fig. 4C; Supplementary Fig. 3A), a significant improvement in FLMR of obese mice subjected to TP1 and TP2 was noticed, reaching a pre-obese or non-obese threshold (mean ratio 0.58 for TP1 and 0.49 for TP2), with no effect for TP3 (mean ratio 0.66) (Fig. 4D; Supplementary Fig. 3B). No changes in body weight or FMLR were observed in non-obese mouse models (Fig. 4C and D; Supplementary Fig. 3A and 2B).

As anticipated, all exercise paradigms significantly reduced non-fasting glycaemia, reaching values below the hyperglycemia threshold (250 mg/dl) in all mouse models (Fig. 4E; Supplementary Fig. 3C). Glucose sensitivity was improved by TP1 and TP2 in non-obese STZ/HFD-induced mice, with optimal effects from TP1 (Fig. 4F–I; Supplementary Fig. 3D–G), and by TP2 and TP3, with TP2 providing optimal effects in obese HFD-induced mice (Fig. 4J and K; Supplementary Fig. 3D and 3E). TP1 efficiently restored insulin sensitivity in non-obese mouse models (Fig. 4L–O; Supplementary Fig. 3H–K), whereas TP2 significantly improved insulin sensitivity only in obese males (Fig. 4P and Q; Supplementary Fig. 3H and 3I).

Thus, all these data show, for the first time, that specific exercise can target specific T2DM-induced metabolic impairment, resulting in optimal benefits for T2DM mouse models.

The sensorimotor performances of the different T2DM mouse models are differentially improved by the specific exercise programs

We finally examined the effects of the different exercise paradigms on T2DM mouse sensorimotor behaviour. Using the maximum speed test to assess the mouse anaerobic performances, we found that, for non-obese T2DM mouse models, only the TP1 program was able to increase the maximum speed from week 20 (mean of 3 L min−1 in males, 3.2 L min−1 in females) to week 32 (similar mean of 4.7 L min−1) (Fig. 5A and B). By contrast, in obese diabetic mice, despite the global positive effects of all exercise programs, it was TP2 that provided the optimal effects on anaerobic performances, with an increase in maximum speed from week 20 (mean 3.8 L min−1) to week 32 (mean 5 L min−1) (Fig. 5C).

Fig. 5

Precision exercise programs are optimal in limiting T2DM hallmarks in the different mouse models. The maximum speed test (A–C) and incremental test (D–F) were assessed from the age of 20 weeks to the age of 32 weeks in HFD/STZ induced male (HFD/STZ M) and female (HFD/STZ F) C57 mice and HFD-induced male (HFD M) C57 mice, submitted to personalized exercises (TP1, TP2 and TP3), and compared to untrained group (UTG). The motor coordination was assessed by (i) the beam walk test, during which the speed (G) and foot slips (H) were measured, (ii) the thermal sensitivity hot plate test (I) and (iii) CMAP amplitude (J) in all trained groups at 32 weeks of age, compared to untrained (UTG). (n = 5 for all groups) All data are shown as mean ± standard deviation. Non-parametric One-Way ANOVA for all groups; * indicate significance (with P < 0.05 *, < 0.005**) relative to untrained (UTG)

Then, using the incremental test, we found that TP1 was optimal in non-obese T2DM mouse models, as it was able to increase the fitness power from week 20 (mean 8 min 26 s in males, and 5 min 17 s in females) to week 32 (mean of 13 min 32 s in males, and 12 min 41 s in females) (Fig. 5D and E) compared to untrained counterparts (mean of 6 min 17 s in males and 6 min 28 s in females). By contrast, TP2 was found optimal in increasing fitness power in obese T2DM mice from week 20 (mean 9 min 45 s) to week 32 (mean 12 min 6 s) (Fig. 5F).

We next investigated the mouse motor coordination, using the beam walk test, and found that TP1 was optimal in improving walking coordination in HFD/STZ-induced male and female C57 mice, notably by limiting the walking speed decline and the number of foot slips (Fig. 5G and H; Supplementary Fig. 3L, and 3M). As expected, TP2 provided the optimal benefits in HFD-induced male C57 mice, with improvement in the walking speed and in the foot slips (Fig. 5G and H).

Finally, using thermal nociception and motor-nerve conduction velocity tests, we found that while T2DM induction proved to significantly decrease thermal sensitivity and motor nerve conduction amplitude in all untrained mouse groups compared to controls (Supplementary Fig. 3N and 3O), exercise differentially limited the sensorimotor alterations with, according to the motor performances, an optimal effect for TP1 in HFD/STZ-induced male and female C57 mice and for TP2 in HFD-induced male C57 mice (Fig. 5I and J).

Taken together, these behavioral data provide the first lines of evidence that the specific exercise-induced improvement of metabolic parameters in T2DM mice is associated with significant benefits in terms of sensorimotor behavior.