Male mice are more sensitive to HFD-induced weight gain, glucose tolerance, and insulin resistance
The final body weight of male mice with 12 weeks of high-fat diet feeding was substantially higher than that of normal diet (ND)-fed male mice (41.5 ± 5.32 g versus 29.0 ± 1.5 g in male controls, P<0.001), while there was no significant difference in the body weight of female mice between HFD and ND (23.8 ± 1.8 g versus 22.1 ± 0.84 g in female controls，P=0.07) (Figure 1A). In addition, the random blood glucose of male mice increased remarkably in response to HFD, in parallel with that the female changed a little (Figure 1B). To evaluate glucose tolerance in HFD mice, the oral glucose tolerance test (OGTT) was conducted at 12 weeks. The result showed the characteristic rapid rise of blood glucose in all mice, peaking within 15-30 min, and approaching the baseline level by 120 min after glucose challenge. However, the blood glucose level of male-HFD mice remained elevated after 120 min compared to other mice, indicating that the HFD-fed male mice were more likely to develop glucose tolerance (Figure 1C). Insulin resistance test (ITT) and fasting serum insulin level results also showed that male-HFD mice were with lower insulin sensitivity and easier to develop hyperinsulinemia than female-HFD mice (Figure 1D, E). Interestingly, there was no obvious changes in food and water intake between males and females (Figure 1F, G). Collectively, these results suggested that male mice are more sensitive to HFD-induced obesity compared to female mice.
NMR spectra and pattern recognition analysis of serum and liver samples
Typical 1H NMR spectra of serum sample is shown in figure 2 A. A series of metabolites were identified, involving amino acid metabolism: leucine, valine, isoleucine, alanine, glutamate, tyrosine, and phenylalanine; lipid metabolism: Low-density lipoprotein/very low-density lipoprotein (LDL/VLDL), and unsaturated fatty acid (poly-UFA); ketone body metabolism: 3-hydroxybutyrate (3-HB) and acetate; energy metabolism: lactate and creatine. To evaluate the differences in metabolic patterns of serum samples, principal component analysis (PCA) was conducted. Results based on PCA show that the male mice with HFD group (M-HFD) was clearly separated from the other three groups (figure 2B).
Figure 2C displayed the representative 1H-NMR spectra in liver samples, containing many metabolites involved in amino acid metabolism (leucine, valine, isoleucine, alanine, glutamate, glutamine, glutathione, aspartate, lysine, glycine, tyrosine, phenylalanine, nicotinurate and betaine), energy metabolism (lactate, succinate, creatine, GTP and fumarate), membrane metabolism (choline and phosphocholine), ketone body metabolism (acetate and 3-hydroxybutyrate), as well as nucleic acid metabolism (uridine, inosine and AMP). Similar to the metabonomics data of serum samples, the M-HFD liver metabolic pattern was significantly distinguished from the other three groups. Moreover, the metabolic pattern of HFD-fed female mice was closer to the normal diet male (M-ND) group (Figure 2D).
Sex difference in dietetic effects on the metabolome in the serum of mice
To further explore the metabolic patterns, partial least squares-discriminate analysis (PLS-DA) and VIP statistics were conducted. The goodness of fit and the predictive capability of the model were shown in Supplementary Table1. A clear discrimination was observed between male and female mice fed with HFD along the PLS1 direction (Figure 3A). The VIP scores indicated that LDL/VLDL played a pivotal role in the separation of both groups (Figure 3B). However, there was no significant difference between males and females with ND, with the value of VIP being less than 2.0 (Figure 3C, D). Additionally, lactate is the major contributor to the separations in female mice with a ND or HFD, but not LDL/VLDL (Figure 3E, F). Interestingly, LDL/VLDL played a pivotal role in the separation of male mice with a ND and HFD (Figure 3G, H). Overall, LDL/VLDL could be the major affected metabolite in the serum of male mice with HFD.
Metabolic analysis of liver samples in male and female mice
Supervised orthogonal partial least squares-discriminant analysis (OPLS-DA) and S-plot were subjected to identify key metabolites that contributed to metabolic pattern changes in liver. Through a pairwise comparison of OPLS-DA, it was seen that either in male or female mice, the metabolic patterns of the liver samples were obviously different between the ND and HFD (Figure 4A, C). Additionally, the S-plots suggests that the separation in male mice was ascribed to the increases of succinate, glutathione, and phosphocholine, as well as decreases in betaine and glycine (Figure 4B). For female mice, the increase of betaine, lactate, choline, phosphocholine, inosine, and the decrease of creatine contributed to the distribution of the two groups (Figure 4D). Figure 4 E and G displayed that the male group was clearly separated from the female group under ND and HFD conditions, respectively. With normal diet, betaine, glycine, choline, alanine, and lactate are the main factors for the differences (Figure 4F). In addition, betaine, inosine, choline, lactate, and glutathione were shown through the loading of S-plot, leading to the separation of M-HFD and F-HFD (Figure 4H). The goodness of fit and the predictive capability of the OPLS-DA were displayed in Supplementary Table 2.
Male mice are more prone to HFD-induced lipid metabolism disorder
Further quantitative analysis and pathway analysis of various metabolites in mouse serum revealed that the content of LDL/VLDL increased dramatically in male mice after HFD (68.81±8.95 in M-HFD versus 9.12±2.55 in M-NC, P<0.001), while the female mice had no change (46.66±16.85 in F-HFD versus 46.77±17.93 in F-NC, P=0.917). More importantly, the LDL/VLDL level of female mice was greater than that of male mice (46.77±17.93 in F-NC versus 9.12±2.55 in M-NC, P<0.001) in respond to ND. However, the LDL/VLDL increased markedly in the serum of male mice with HFD feeding, even higher than that of female mice (68.81±8.95 in M-HFD versus 46.66±16.85 in F-HFD, P<0.05). Additionally, HFD led to the decrease of poly-UFA in the serum of male mice (17.17±1.51 in M-HFD versus 14.37±2.63 in M-NC, P<0.05), while almost no change in female (14.22±5.48 in F-HFD versus 14.53±4.19 in F-NC, P=0.815). Except 3-hydroxybutyrate and lactate, most metabolite levels decreased after HFD-feeding both in male and female mice, but there was no significant difference. The quantitative data analysis of other metabolites was shown in Supplementary Table3 and 4. Additionally, LDL/VLDL and poly-UFA are substances related to lipid metabolism. Together, our results indicate that male mice were more susceptible to HFD-induced lipid metabolism disorders than females (Figure 5).
Changes in metabolites level in liver of mice on different diet conditions
In the liver, the concentrations of metabolites such as amino acid metabolites, energy metabolites, ketone body metabolites, nucleic acid metabolites, and membrane metabolites in female mice were increased after HFD feeding (Figure 5 and Supplementary Table 5, 6). However, most metabolites levels in male mice were reduced, such as short-chain amino acid, alanine, aspartate, lysine, glycine, tyrosine, phenylalanine, lactate, creatine, uridine, inosine, acetate, choline, betaine. In addition, the levels of liver metabolites in male mice were generally higher than those in female with normal diet, and most of them had significant differences. As expected, the levels of most metabolites in male mice reduced after feeding HFD, while the levels in female mice elevated, even close to the levels of the ND-fed male mice group. These results are consistent with the similar metabolic pattern of F-HFD and M-NC group. In conclusion, the enhanced metabolic capacity could be the reason that the female mice did not increase their weight, blood glucose greatly after HFD-feeding.
Betaine could correct HFD-induced abnormal lipid metabolism
To investigate the reason that caused less accumulation of LDL/VLDL in female-HFD mice serum, we analyzed the correlation on LDL/VLDL in serum with liver metabolites. The results of the heatmap showed a significant negative correlation between LDL/VLDL and betaine, especially in mice fed with HFD (Figure 6A). Combined with the analysis of the quantitative data of betaine, it showed that the level of betaine in male mice was higher than that in females under normal dietary conditions (24.49 ± 5.42 in M-NC versus 19.24 ± 2.26 in F-NC, P<0.05). Unexpectedly, after 12 weeks of HFD-feeding, the betaine level in female mice increased, while the male mice decreased (19.09 ± 2.75 in M-HFD and 23.36 ± 4.20 in F-HFD, P<0.05). Thus, we speculated that betaine which is highly expressed in the liver of female mice, could correct the accumulation of LDL/VLDL in serum after HFD.