3.3. Hepatic proteomic and phosphoproteomic analyses revealed insulin- and glucose- independent effects of linagliptin in OSI-906-induced hepatic steatosis
We also performed hepatic quantitative proteomic and phosphoproteomic analyses of the livers from OSI-906- or linagliptin-treated mice (Figure 3a). We identified a total of 1884 proteins, and differentially expressed proteins were analyzed using Progenesis QI software to compare expressions in the OSI-906 vs. vehicle groups and the OSI-906 + Lina vs. OSI-906 groups (Figure 3b).
We focused on proteins with expressions that were significantly altered by the administration of OSI-906 and for which the alterations were reversed by the treatment with linagliptin (Figure 3c and Supplementary Table 2) to address the amelioration of hepatic steatosis by linagliptin in the OSI-906-treated model. Among the proteins that were significantly up-regulated or down-regulated in response to the administration of OSI-906 (Supplementary Tables 2-4), the abundances of perilipin-2 (PLIN2) and cytochrome P450 2b10 (CYP2B10) were reduced by OSI-906 and were restored by treatment with linagliptin. In contrast, the abundance of major urinary protein 20 (MUP20) was increased by OSI-906 and was also restored by treatment with linagliptin. Those results indicate that these molecules might contribute to the process of the linagliptin-induced amelioration of hepatic steatosis. A previous study reported that the expressions of cytochrome P450 enzymes were down-regulated and those of major urinary proteins (MUPs) were up-regulated by treatment with resveratrol, a sirtuin activator, in the liver (Baur et al., 2006). We also found that the protein expression of nicotinamide N-methyltransferase (NNMT) was significantly up-regulated in OSI-906-treated liver. Recently, the contribution of NNMT activation to the development of hepatic steatosis has been reported (Komatsu et al., 2018). Therefore, these findings suggest that sirtuin activity is regulated by OSI-906 administration and might contribute to the improvement in hepatic steatosis (Pissios, 2017). We validated these findings using immunoblotting or gene expression analyses. The protein expressions of PLIN2 and NNMT were increased in OSI-906-treated liver, and these increases were reversed by linagliptin (Figure 4a). In addition, the increased pan acetyl-lysine levels in OSI-906-treated liver were also reversed by linagliptin, implying an enhancement of sirtuin activity by linagliptin in this OSI-906-induced hepatic steatosis model (Figure 4a). The hepatic gene expression ofCyp2b10 was also induced by OSI-906 and was deceased by treatment with linagliptin (Figure 4b). A canonical pathway analysis of differentially expressed proteins in the OSI-906 vs. vehicle groups and the OSI-906 + Lina vs. OSI-906 groups revealed that the administration of OSI-906 activated the cholesterol biosynthesis pathway and inhibited the LXR/RXR pathway, whereas linagliptin inhibited the cholesterol biosynthesis pathway (Figure 5).
In a phosphoproteomic analysis of liver samples from the four groups, more than 3700 sites of phosphorylation were identified. The top increased or decreased phosphopeptides when the OSI-906 vs. vehicle groups or the OSI-906 + Lina vs. OSI-906 groups were compared are shown in Supplementary Tables 5-8.
A canonical pathway analysis of the phosphoproteomics data identified the inhibition of insulin receptor signaling and mTOR signaling in OSI-906-treated liver, consistent with the dual inhibition of insulin/IGF-1 receptors by OSI-906 (Figure 6). Treatment with linagliptin did not seem to affect insulin signaling. An upstream regulator analysis indicated that PPARα and XBP1 were the top 2 predicted transcription activators involved in the process of hepatic steatosis induced by OSI-906, whereas pathways related to SREBF2 and lysophosphatidylcholine were identified in the linagliptin-treated group (Supplementary Table 9).