Alogliptin improves survival and health of mice on a high‐fat diet

Abstract Alogliptin is a commonly prescribed drug treating patients with type 2 diabetes. Here, we show that long‐term intervention with alogliptin (0.03% w/w in diet) improves survival and health of mice on a high‐fat diet. Alogliptin intervention takes beneficial effects associated with longevity, including increased insulin sensitivity, attenuated functionality decline, decreased organ pathology, preserved mitochondrial function, and reduced oxidative stress. Autophagy activation is proposed as an underlying mechanism of these beneficial effects. We conclude that alogliptin intervention could be considered as a potential strategy for extending lifespan and healthspan in obesity and overweight.

that probably is a central process in extending longevity (Johnson, Rabinovitch, & Kaeberlein, 2013). Thus, a reagent served as an activator of autophagy might provide beneficial effects on promoting lifespan.
The broad pleiotropic effects on attenuating multiple age-related diseases of DPP-4 inhibitors and GLP-1 analogues raise the possibility that they may take beneficial effects on extending lifespan. Therefore, we hypothesized that DPP-4 inhibitors play a potential role in extending lifespan. In this study, we explored the potential effects of alogliptin, a highly selective DPP-4 inhibitor, on survival and health of mice on a long-term high-fat diet.
Cohorts of one-year-old male ApoE −/− mice were provided with high-fat diet (60% of calories from fat, HFD) for the remainder of their lives. To the diet, we added the drug in two doses, consisting of 0.01% or 0.03% (wt/wt) alogliptin estimated from previous studies (Moritoh, Takeuchi, Asakawa, Kataoka, & Odaka, 2008;Shah et al., 2011) and found the higher dose was more prominent on increasing survival. Thus, we selected 0.03% (wt/wt) alogliptin as the intervention dose (Supporting Information Figure s1a).
In this dose, the effects of alogliptin intervention on plasma DPP-4 activity and GLP-1 levels were examined in 13 months old mice. Interestingly, plasma DPP-4 activity trended to increase in the HFD group compared with NC (normal chow) group (p > 0.05), while it was significantly inhibited in alogliptin plus HFD (AHF) group when compared with the other two groups (Figure 1b), and consequently preserved a relatively higher level of GLP-1 which decreased in HFD group (Figure 1c).
At about 64 weeks of age, the survival curves of HFD and AHF groups began to separate. With the present age at 112 weeks, 60.0% of mice in HFD group had died (median survival 103 weeks), compared to 45.5% in AHF group and 31.0% in NC group (Figure 1a). Although we could not predict the ultimate mean lifespan extension, comparison of survival curves showed that alogliptin intervention markedly increased lifespan compared to HFD group in the log-rank survival test (χ 2 = 3.950, p = 0.047), but not significantly different from NC group (χ 2 = 1.882, p = 0.170). In addition, sitagliptin, another DPP-4 inhibitor, took a similar longevity-promoting effect (χ 2 = 4.231, p = 0.040; Supporting Information Figure S1b), although the structural similarity between sitagliptin and alogliptin is only 31.3% .
The mean weight of HFD mice steadily gained until 76 weeks of age, after which slowly declined (Figure 1e). From 90 to 107 weeks of age, AHF mice were lighter than HFD mice, consistent with a previous study (Shah et al., 2011) showing alogliptin intervention significantly reduced body weight. But no differences were found in food intake (Figure 1d), total feces mass, or lipid content in the feces (Supporting Information Figure S1c,d). Next, we examined whether the decreased body weight was due to higher energy expenditure.
Metabolic experiments showed heat production of AHF group was significantly higher than HFD group in dark phase ( Figure 1g).

| Attenuated functional decline and improved health
Although lifespan was extended, it was also important to ascertain whether health was improved. Therefore, we performed a series of experiments to measure physiological functions of several vital organs in the old mice.
Firstly, rotarod was performed to test motor coordination, and results indicated that alogliptin intervention improved balance ability of the laboratory mice in AHF group. At 24 months of age, they were even indistinguishable from NC group (Figure 1f). Interestingly, alogliptin intervention also elevated mastication efficiency in AHF mice ( Figure 1h); secondly, high-fat diet elevated systemic blood pressures, especially systolic pressure, and these were attenuated by alogliptin intervention (Figure 1i). Thirdly, it is established that cardiac dysfunction is a significant predictor of death in mice and humans (Dai et al., 2009;Eisenberg et al., 2016). Tibia length-nor- Next, stiffening of large elastic arteries is a strong and independent risk factor of cardiovascular events on senescence (Fleenor et al., 2014;Mitchell et al., 2010). We found alogliptin intervention reduced pulse wave velocity (PWV) on high-fat diet to level similar to NC mice ( Figure 1m). We also examined endothelium-independent vasodilation function of aortas in response to sodium nitroprusside (SNP). Results showed that long-term high-fat diet decreased the vascular vasodilation function, which was remarkably improved by  alogliptin intervention ( Figure 1n); finally, intraperitoneal glucose tolerance test (IPGTT) and insulin tolerance test (ITT) displayed that HFD mice had significantly lower glucose disposal ability than NC mice, while alogliptin intervention exhibited an improvement in glucose and insulin tolerance (Figure 1o-r).

| Decreased organ pathology
Our above animal experiments showed alogliptin intervention improved physiological functions on high-fat diet, and we next investigated whether the structures of organs were also benefited.
Long-term high-fat diet deteriorates insulin sensitivity leading to diabetes and decreases longevity (Baur et al., 2006), and β-cell dysfunction progresses to a reduction in mass is one of pathological fundaments in diabetes (Takeda et al., 2012). In our study, high-fat diet increased glucose and insulin in blood, and alogliptin intervention attenuated these increases (Table 1). Indeed, histological examination of pancreas showed islets from AHF mice exhibited a large insulin-positive cell core encompassed with an orb of α-cells, paralleling NC mice, while the islet architecture of HFD mice was disorganized ( Figure 2a). Measurement of islet β-cell proportion revealed a 16.34% decrease on high-fat diet (p < 0.05), and which was normalized with alogliptin intervention (Supporting Information Figure S3a).
As mentioned above, alogliptin intervention preferred to increase heat production and reduced body weight, we further investigated its effects on body fat distribution. HFD mice had a 2.6-fold increase in abdominal adipose deposition than NC mice, which was attenuated by alogliptin intervention (Figure 2b,c). Moreover, alogliptin intervention decreased TNF-α and IL-6 expression in abdominal visceral adipose tissue (Supporting Information Figure S3b).
Because liver plays a central role in regulating substance and energy metabolism, we then examined the histology of liver in various groups. High-fat diet obviously increased the mass (Supporting Information Figure S3c) and size ( Figure 2d) of liver, and alogliptin intervention prevented these alterations. Hematoxylin and eosin (H&E) and oil red O staining revealed a loss of cellular integrity and an accumulation of small-and-large mixed typed lipid droplets in the liver of HFD mice but not in AHF mice (Figure 2d).
Osteoporosis is a metabolic disorder of skeleton that affects millions of the elder . Thus, we analyzed micro-architectural properties of tibias by microcomputed tomography (micro-CT). As expected, alogliptin intervention remarkably elevated trabecular relative volume (Tb.BV/TV), trabecular number (Tb.N), cortical bone mean density (BMD), and cortical thickness (Ct.Th). All of these parameters declined sharply in HFD group. Conversely, trabecular separation (Tb.Sp) and endosteal perimeter (Es.Pm) were lower in AHF group, these parameters increased in HFD group (Supporting Information Figure S4a,b).
Based on improved PWV and vasodilation function in AHF mice, we subsequently explored the morphology modifications of aortic elastic lamina. H&E (Supporting Information Figure S3d) and elastica van Gieson (VG; Supporting Information Figure S4c) staining showed the elastic lamina of thoracic aortas in HFD mice was straighter and more loose compared with NC samples, while alogliptin intervention, in some degree, preserved the wavy elastic fibers and retarded the loss of elastic density.

| Preserved mitochondrial function
Previous study (Aroor et al., 2015) showed DPP-4 inhibitor improved mitochondrial function, and we found alogliptin intervention reduced body weight, induced energy expenditure, and decreased abdominal adipose deposition. Thus, we wondered whether this drug increased mitochondrial biogenesis and (or) elevated mitochondrial respiration function.
We firstly evaluated the potential effects of alogliptin adminis-

Alogliptin + HFD (AHF)
Glucose-fasted (mM) 6.3 ± 0.9 7.9 ± 2.1 * 7.1 ± 1.7 Glucose-fed (mM) 9.5 ± 1.2 10.9 ± 2.5 10.  Subsequently, we tested this pathway and results showed alogliptin treatment had no influence on the expressions of p-FOXO1 and p-FOXO3a (Supporting Information Figure S6). AMPK is a master energy sensor and mTOR is an evolutionarily conserved regulator of cellular growth and metabolism. In parallel, we confirmed these pathways and found increased AMPK phosphorylation and decreased mTOR phosphorylation in AHF group compared with HFD group (Figure 5a), suggesting alogliptin intervention activated AMPK and inhibited mTOR.

| Upregulated autophagy in vivo
Since AMPK-mTOR signaling pathway is related to longevity and modulates autophagy, autophagy activation also contributes to delaying aging process (Vilchez, Saez, & Dillin, 2014). Thus, we next explored whether alogliptin intervention induced autophagy. Analysis of liver protein levels of microtubule-associated light chain 3B-II (LC3B-II) revealed alogliptin intervention significantly increased relative LC3B-II levels, which were decreased in HFD mice, indicating an accumulation of autophagosome. To distinguish activated autophagy from blocked fusion of autophagosome and lysosome leading to increased LC3B-II, we also measured sequestosome-1 (p62) protein levels. Results showed p62 level was decreased in AHF group, while accumulated in HFD group, indicating autophagy level was higher in AHF group (Figure 5a,b). However, p62 might also degrade via ubiquitin-proteasome pathway leading to decreased p62.
Therefore, we used chloroquine (CQ), a specific autophagy inhibitor blocking lysosomal acidification, to further clarify the autophagy level in AHF mice. We also measured autophagic flux by delivering mRFP- Autophagy is regulated by mTOR-dependent and mTOR-independent signaling pathways (Vilchez et al., 2014). Given that activated AMPK directly inhibits mTOR and thereafter induces autophagy (Rubinsztein, Marino, & Kroemer, 2011). We further explored these signaling pathways (Figure 5a,b) and found AMPK phosphorylation was increased in the liver of AHF mice. Consequently, mTOR and p70-ribosomal S6 protein kinase (p70-S6K) phosphorylation levels decreased, indicating mTOR and p70-S6K were inhibited in AHF mice and subsequently induced autophagy. Regarding mTOR-independent pathways, sirtuin 1 (Sirt1) could also mediate autophagy via deacetylation of several autophagy-related (Atg) proteins and transcription factors (Lee et al., 2008;Vilchez et al., 2014). Here, we showed alogliptin intervention significantly elevated Sirt1 protein level, which decreased on long-term high-fat diet.

GFP
Since AMPK could also directly phosphorylate ULK1(unc-51-like kinase 1) to initiate autophagy, whereas mTOR negatively regulates ULK1 by phosphorylation of Ser757 to inhibit autophagy (Gui et al., 2017;Li et al., 2018). Thus, we explored whether AMPK directly activated autophagy independent of the inhibition of mTOR. Results indicated AMPK-ULK1 pathway did not involve in autophagy activation by alogliptin intervention (Supporting Information Figure S7a,b).  Figure S9).

| Activated autophagy in vitro
To confirm the signaling regulating autophagy in vitro, we also explored mTOR-dependent and mTOR-independent pathways.
Results revealed mTOR ( Figure 6c) and its downstream molecular Mice were injected with adenovirus-encoding mRFP-GFP-LC3 seven days before the end of the study, then the liver cells were isolated and assessed by flow cytometry. (d) Quantitative analysis of (c). MFI, mean fluorescence intensity. NC, normal chow. HFD, high-fat diet. AHF, 0.03% (wt/wt) alogliptin plus HFD. AHQ, AHF plus CQ (chloroquine; 60 mg/kg, i.p. injection 6 hr before sacrificing the mice). Data are expressed as mean ± SD. *p < 0.05 compared with NC, # p < 0.05 compared with HFD, $ p < 0.05 were even close to rapamycin. Simultaneously, we detected increased AMPK phosphorylation by addition of GLP-1, alogliptin, and GLP-1 and alogliptin co-treatment, respectively, which was significantly blocked by Compound C (Figure 6b). Regarding mTOR-independent pathway, we also found GLP-1, alogliptin, and GLP-1 and alogliptin co-treatment increased Sirt1 protein levels, respectively ( Figure 6a). Moreover, GLP-1 and alogliptin co-treatment also had additive effects on elevating expression of Sirt1 and p-AMPK, reducing p-mTOR and p-p70S6K protein levels. In line with primary hepatocytes model, alogliptin administration also activated autophagy in L-02 cells (Supporting Information Figure S8a Figure S7c).

| Autophagy is required for improving health and extending longevity
To determine whether the improved health and extended lifespan of alogliptin intervention on high-fat diet depend on autophagy, we inhibited autophagy by delivering Atg7 siRNA to AHF mice (experiment group, Exp). Firstly, we confirmed knockdown Atg7 in the liver of Exp mice compared with Con (control group) mice (Supporting Information Figure S10a). Atg7 was necessary for autophagosome formation, consequently, we observed significantly increased p62 and decreased LC3B-II proteins in the liver of Exp mice (Supporting Information Figure S10b). Subsequently, we found the attenuated lipid accumulation by alogliptin intervention was disappeared in Exp mice (Supporting Information Figure S10d). In parallel, abdominal adipose deposition reduced in Con mice but not in Exp mice (Supporting Information Figure S10c). Secondly, the Exp mice displayed decreased glucose tolerance (Supporting Information Figure S11a) and insulin sensitivity (Supporting Information Figure S11b) as well as reduced β-cell proportion (Supporting Information Figure S11c).
Thirdly, improved endothelium-independent vasodilation function of aortas was observed in Con mice but disappeared in Exp mice (Supporting Information Figure S12a). In addition, the elastic lamina of thoracic aortas in Exp mice also seemed to be straighter (Supporting Information Figure S12b). Finally, we detected decreased complex activity of electron transport chain (ETC) in mitochondrial respiration in the liver of Exp mice (Supporting Information Figure S13). Notably, the longevity-extending effects of alogliptin intervention markedly diminished when Atg7 knockdown (χ 2 = 5.308, p = 0.021; Supporting Information Figure S14). Collectively, these data indicated alogliptin intervention improved age-related diseases and promoted longevity, at least partly, through activating autophagy.

| DISCUSSION
Our present data for the first time suggested that alogliptin intervention extends longevity and improves health of mice with excess caloric intake. These beneficial effects are mainly related to increased insulin sensitivity, attenuated function decline, decreased organ pathology, inhibited inflammation, preserved mitochondrial function, and reduced oxidative stress. Autophagy activation may be involved in these beneficial effects. Thus, alogliptin intervention is considered as a potential strategy for extending lifespan and healthspan in obesity and overweight.
A growing body of evidence (Aroor et al., 2015;Matsubara et al., 2012;Meng et al., 2016;Moritoh et al., 2008;Zhuge et al., 2016) suggested short-term treatment with DPP-4 inhibitors and GLP-1 analogues attenuated a broad spectrum of age-related diseases. In present study, we proved long-term alogliptin intervention produced similar protective effects on age-associated diseases. Especially, as expected, long-term alogliptin treatment extended longevity and improved healthspan on high-fat diet. Additionally, our study showed sitagliptin, another DPP-4 inhibitor which is structurally distant from alogliptin , took a similar longevity-promoting effect. Based on these data, we speculate the life-extending effect is a class effect of DPP-4 inhibitors.
Our RNA-seq analysis results indicated autophagy may be one of the key mechanisms linked to these beneficial effects of alogliptin intervention. In our study, over-time high-fat diet declined basal level of autophagy, leading to an accumulation of unnecessary and dam- with several previous studies (Bjedov et al., 2010;Harrison et al., 2009;Kenyon, 2010;Pyo et al., 2013;Vilchez et al., 2014) showing either genetical or pharmacological activation of autophagy extended lifespan in devious species. Secondly, it has been proposed autophagy is necessary for maintaining mitochondrial functions (Choi et al., 2017). Moreover, impaired mitochondrial functions increased ROS production and excess ROS led to oxidative stress (Mitchell et al., 2016). In present study, alogliptin increased mitochondrial biogenesis, elevated mitochondrial respiration function and prevented mitochondrial swelling, consequently reduced ROS production. Conversely, the elevated respiration in alogliptin intervention decreased sharply when inhibiting autophagy using Atg7 siRNA. Additionally, in liver, mitochondrial biogenesis is positively regulated by Sirt1 (Baur et al., 2006  In addition, DPP-4 inhibition in vivo also worked on other respects, such as immunoregulation by affecting leukocyte migration (Shah et al., 2011), which might also contribute to improving health and survival. Second, our animal model is obesity and overweight, so it is still not known whether our results are applicable to normal weight animals. Further studies with additional animal models are needed to generalize the effects of alogliptin intervention to normal individuals.
In conclusion, our present data demonstrate that alogliptin intervention is efficacious on improving survival and health of mice on a high-fat diet and we consider these beneficial effects as a comprehensive result. Here, autophagy is proposed as an underlying reason of these beneficial effects. In addition, multiple beneficial effects, including anti-insulin resistance, anti-obesity, cardiovascular protection, and anti-aging, were observed by alogliptin intervention in HFD-fed mice. Because, our animal model in this study is metabolic syndrome, so we consider anti-insulin resistance may be more primary.

| EXPERIMENTAL PROCED URES
More detailed methods are described in the Supporting Information Appendix S1.

| Animals models and diets
Animal procedures conformed to the National Institutes of Health  (Takeda et al., 2012).
To access the effects of autophagy inhibition on longevity and health, another 12-month-old male C57BL/6J mice were randomly assigned to control group (Con, n = 55 mice) and experimental group (Exp, n = 55 mice). The mice were fed with AHF for their reminder lives. During the period, the mice in Exp group were injected with adenovirus-encoding Atg7 siRNA (0.15 ml; 10 10 PFU/ml, Hanbio, China) through the tail vein once every three weeks, and the Con mice were correspondingly injected with adenovirus-encoding scrambled siRNA.
All mice were maintained in a specific-pathogen-free environment (20-22°C) with a 12 hr light/dark cycle and unrestricted access to water and food. Survival curves were plotted using the Kaplan-Meier method, which included all available animals at each time point. The death of mice was recorded on a weekly basis.

| Histology
Tissues were fixed in 4% formaldehyde, then sectioned and stained with H&E or van Gieson (VG). The immunofluorescence staining of pancreatic tissues was performed as previously reported . Livers were frozen at −80°C and stained with oil red O or ROS fluorescent probe-dihydroethidium.

| Mitochondrial mass assays
Primary hepatocytes were isolated from eight-week-old male C57BL/ 6 mice and cultured as previously described (Guo et al., 2017). Mitochondrial mass was stained as previously reported (Baur et al., 2006) using Mitotracke Green FM.

| Other parameters
Serum biochemical markers, rotarod, echocardiography, and et al. are described in the Supporting Information Appendix S1.

| Statistical analysis
Values are expressed as mean ± SD except where noted. Statistical significance was analyzed by one-way ANOVA following with LSD post hoc tests when equal variances were assumed or Tamhane's T2 post hoc tests when equal variances were not assumed for comparisons between two or multiple groups. Homogeneity of variance was tested using Levene's test. The significance values are p < 0.05 (two-sided). Statistical testing was performed using IBM SPSS 22.0 statistics software (IBM Corp, Armonk, NY).

ACKNOWLEDG MENTS
This work was supported by grants from National Natural Science

CONFLI CT OF INTEREST
The authors declare no competing financial interests. access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.