DPP-4 inhibition with linagliptin ameliorates the progression of premature aging in klotho−/− mice

All experiments were approved by the institutional Animal Care and Use Committee of Kumamoto University. Male homozygous mutant klotho−/− mice and C57BL6J mice (wild-type mice) were purchased from Nihon CLEA (Tokyo, Japan). In the present study, according to the instruction of the supplier of these mice, 2 klotho−/− mice and 2 wild-type mice were housed in one cage, since klotho−/− mice are extremely vulnerable to stress and individual housing of klotho−/− mice significantly shortens life span.

Linagliptin was kindly provided by Boehringer Ingelheim (Ingelheim, Germany).

Five-week-old male klotho−/− mice were divided into 2 groups. Group 1 (n = 15) of mice were fed standard diet (MF diet, ORIENRAL YRAST Co., Ltd., Tokyo, Japan) and group 2 (n = 15) of mice were fed the standard diet containing linagliptin (0.083 g/kg diet). This dose of linagliptin is shown to be an appropriate dose for estimation of pharmacological action of linagliptin in vivo, as shown by previous reports [5, 6, 19, 20]. Drug treatment was performed for 4 weeks (from 5 to 9 weeks of age). Survival of each mouse was monitored every day and body weight was monitored every week. Passive avoidance test, monitoring of alopecia, wire-hang test, rotarod test, and oral glucose tolerance test (OGTT) were performed at specified time point as shown in the detailed experimental protocol (Fig. 1). After 4 weeks of the drug treatment, the mice had measured cerebral blood flow (CBF) under 1.5–2% isoflurane, had blood samples taken from the right ventricle, and perfused with phosphate-buffered saline. Subsequently, various organs including cerebrum, cerebellum, left and right ventricle, kidney, subcutaneous and visceral fat, and gastrocnemius muscle were removed and weighed. Then, rostral side from bregma of each cerebrum was immediately frozen. Caudal side from bregma of each cerebrum and kidney was immediately frozen in Tissue-tek OCT embedding medium (Sakura Finetek, Tokyo, Japan). Gastrocnemius muscle was embedded in paraffin. An 8-µm slice was made at 1.43–2.43 mm caudally from the bregma for the following histological evaluations. In addition, 8-µm slice of kidney and 5-µm slice of gastrocnemius muscle were made for the following histological evaluations.

Wild-type mice (C57BL6J) were also divided into 2 groups (n = 10 in each group) and received linagliptin treatment in the same manner as klotho−/− mice.

The passive avoidance test consisted of two sections including (1) training section and (2) memory test section (Muromachi Kikai, Tokyo, Japan), as previously described by Ma et al. [6]. Briefly, in the training section, the mice were given the electric shock (1.6 mA for 3 s) and kept in the dark side of the test box for 60 s, and this step was repeated three times. In the memory test section, the mice were put into the light side of the test box at 1 day after the training section, and recorded the seconds until the mice went into the dark place (for up to 300 s).

Alopecia was scored by visual assessment as follows; whole body of the mice was divided into 4 segments (dorsal and ventral portion of the body, head, and extremities) and each part of obvious alopecia was scored 1. The maximum score is 4 and higher scores indicate severe alopecia.

Wire hang test was carried out to assess motor strength. The mice were placed on wire mesh and allowed to grasp their four limb hanging. Then, the wire mesh was turned upside-down gently and the latency were recorded until the mice fell off to a ground (for up to 60 s).

Rotarod test was done to assess sensorimotor coordination and motor strength as previously described by Hasegawa et al. [21]. Briefly, animals were placed on the rotating spindle (MK-630B, Muromachi Kikai Co., Ltd., Tokyo, Japan) and a training session were performed at a constant race of 4 rotation per minute (RPM) for 1 min. Then, the mice were subjected to the trial on the accelerating spindle (4–40 RPM) and the latency was recorded until the mice fell off from the cylinder. The mean times for three trial of the test were assigned for each animal.

The oral glucose tolerance test was performed as previously described [6, 22]. Briefly, the mice were fasted for 6 h and given glucose (1 mg/g body weight) orally. The blood glucose concentration collected from tail vein was measured at 0, 30, 60, and 120 min after glucose administration using portable glucose meter (Sanwa Kagaku Kenkyusho Co., Ltd., Nagoya, Japan).

The baseline CBF of the mice was measured using a laser speckle blood flow imager (Omega Zone; Omegawave, Tokyo, Japan), as described by Toyama et al. [23].

Serum DPP-4 activity, insulin and glucose concentrations at the end of drug treatment were evaluated. DPP-4 activity was measured using DPP-4-Glo™ Protease Assay (Promega Corporation, Madison, WI, USA) as previously described by Ma et al. [6]. Insulin concentration was measured using a commercial ELISA kit (Morinaga Institute of Biological Science, Inc., Kanagawa, Japan). Non-fasting serum glucose concentration was evaluated using a Glucose C2 test kit (Wako Pure Chemical Industries, Ltd., Osaka, Japan).

To assess the number of hippocampal neuronal cells, we stained two slices which were randomly selected from the brain slices and then bilateral hippocampal CA1 neuronal cells at 200× magnification in each slice were counted. The average number of the cells from the 4 regions was compared between the groups.

The rostral side of right hemisphere was used for western blot analysis according to our previous method [24]. The following primary antibodies were used: anti-phosphorylated Akt and anti-Akt (1:2000, Cell Signaling Technology, Danvers, MA, USA), anti-phosphorylated eNOS (1:1000, BD Transduction Laboratories, Tokyo, Japan), anti-eNOS (1:2000, Cell Signaling Technology), anti-phosphorylated cAMP response element binding protein (CREB) (1:1000, Cell Signaling Technology), and anti-CREB (1:300, Cell signaling Technology) antibodies. The intensity of those bands was quantified using Image J software (National Institutes of Health, Bethesda, MD, USA).

To detect renal calcification, the kidney sections were stained using the von Kossa methods according to the manufacturer’s protocol (Polysciences, Warrington, PA, USA). The deposit area of calcification was calculated by WinROOF version 5.8 analysis software (Mitani Corporation, Fukui, Japan) using 4 pictures at 40× magnification in each section.

Eight-micrometer slices from kidney were stained with Sirius Red F3BA (0.5% wt/vol in saturated aqueouspicric acid, Aldrich Chemical Company, St. Louis, MO, USA) as previously described by Lin et al. [25]. The area of fibrosis was taken a picture at 200× magnification and analyzed by WinROOF version 5.8 analysis software (Mitani Corporation), and expressed as a percentage per region of interest.

The muscle samples with paraffin were stained with hematoxylin and eosin. To evaluate the muscle fiber size, thirty muscle fibers were measured by Image J software (National Institutes of Health) using the method of minimum Feret diameter and the mean diameter were compared between the groups as previously described by Briguet et al. [26].

All data were expressed as the mean ± SEM. The statistical analyses were evaluated using Graphpad prism version 6 (Graphpad Software, San Diego, CA, USA) and Statcel (OMS Publication, Saitama, Japan). The mortality of the animals was analyzed by a standard Kaplan–Meier analysis with a log rank test and χ² analysis. Evaluations collected over time in OGTT were done using one-way repeated measurement ANOVA with a Bonferroni correction. Parametric and non-parametric evaluations were performed by a non-paired t test or Mann–Whitney U test in two groups. Unpaired t test was used only when a normal distribution is confirmed in two groups and similar variances are obtained between two groups. If two groups do not meet these requirements, Mann–Whitney test was used as an appropriate method. Differences of P < 0.05 were considered significant.

As shown in Fig. 2a and Table 1, body weight after 4 weeks of the treatment (at the end of the treatment) was significantly greater in linagliptin group than in control group (11.1 ± 0.3 vs 9.9 ± 0.3 g; P < 0.01). On the other hand, tibia length at the end of the treatment was similar between the groups (15.3 ± 0.1 vs 15.1 ± 0.2 mm), suggesting that linagliptin did not significantly affect the growth of klotho−/− mice (Table 1).

As shown in Fig. 2b, survival rate of klotho−/− mice at the end of the treatment was greater in linagliptin group (93%) compared to control group (67%), although the difference did not reach statistical significance (P = 0.08).

As shown in Fig. 2c, 4 mice of control klotho−/− mice exhibited alopecia 3 weeks after the start of the treatment, while none of linagliptin-treated klotho−/− mice had alopecia (P < 0.05 vs control klotho−/− mice).

As shown in Fig. 3a, latency of klotho−/− mice in passive avoidance test was significantly larger in linagliptin group than in control group (P < 0.05). CBF of klotho−/− mice was larger in linagliptin group than in control group (P < 0.01) (Fig. 3b). Neuronal cell number in hippocampal CA1 region was greater in linagliptin group than in control group (P < 0.05) (Fig. 3c). Linagliptin group had greater cerebral phospho-Akt (P < 0.05), greater cerebral phospho-eNOS (P < 0.05), and greater cerebral phospho-CREB (P < 0.05) levels than control group (Fig. 3d–f).

Blood glucose levels of klotho−/− mice during OGTT were higher in linagliptin group than in control group (P < 0.01) (Fig. 4a), and AUC during OGTT was significantly greater in linagliptin group than in control group (11.1 ± 0.80 vs 7.03 ± 0.46 mg/dl min 10⁻³; P < 0.01) (Fig. 4b).

As shown in Fig. 4c, serum DPP-4 activity of klotho−/− mice was much less in linagliptin group than in control group (P < 0.01). Non-fasting blood glucose levels of klotho−/− mice were greater in linagliptin group than in control group (183.0 ± 22.6 vs 131.3 ± 6.1 mg/dl; P < 0.05), while non-fasting serum insulin levels were not significantly different between the groups (0.40 ± 0.05 vs 0.34 ± 0.28 ng/ml) (Fig. 4d, e).

Falling latency in wire hang test and rotarod test in klotho−/− mice were not significantly different between linagliptin and control groups (Fig. 5a, b). However, as shown in Table 1, gastrocnemius muscle weight of klotho−/− mice was greater in linagliptin group than in control group (3.4 ± 0.1 vs 2.9 ± 0.1 mg/mm tibia length; P < 0.01). In addition, minimum ferret diameter of gastrocnemius muscle was also significantly greater in linagliptin group than in control group (12.5 ± 0.42 vs 10.6 ± 0.63 μm; P < 0.05) (Fig. 5c). As shown in Table 1, kidney weight of klotho−/− mice was greater in linagliptin group than in control group (5.2 ± 0.2 vs 4.6 ± 0.2 mg/mm tibia length; P < 0.01). However, the degree of renal fibrosis and renal calcification of klotho−/− mice were not significantly different between control and linagliptin groups (Fig. 5d, e).

As shown in Table 2, treatment of wild-type mice with linagliptin did not significantly affect body weight, tibia length, or organ weights including kidney weight and gastrocnemius muscle weight.

As shown in Fig. 6, latency of passive avoidance test, CBF, and hippocampal CA1 neuronal cell number in wild-type mice were not significantly affected by linagliptin treatment.

Blood glucose levels and AUC during OGTT (Fig. 7a, b) in wild-type mice were significantly lower in linagliptin group than in control group (P < 0.05 and P < 0.01, respectively). Linagliptin treatment almost completely inhibited serum DPP-4 activity in wild-type mice but did not affect non-fasting blood glucose or serum insulin levels (Fig. 7c–e).

It is established that klotho−/− mice display brain aging characterized by cognitive impairment and hippocampal neurodegeneration [13, 16, 30, 31], although klotho−/− mice have no Alzheimer’s disease-like pathology such as cerebral amyloid plaque and neurofibrillary tangle. In the present study, we obtained the evidence that linagliptin treatment significantly prevented cognitive impairment in klotho−/− mice, as shown by passive avoidance test. Of note, this amelioration of cognitive impairment by linagliptin in klotho−/− mice was associated with the prevention of hippocampal neuronal cell loss and the increase in CBF. The increased CBF in klotho−/− mice by linagliptin seems to be at least in part mediated by the increase in cerebral phospho-eNOS. Therefore, our present work provided the evidence suggesting that DPP-4 inhibition might exert protective effect against brain aging of klotho−/− mice through the suppression of hippocampal neurodegeneration, the improvement of CBF, and the upregulation of brain eNOS. Akt and its downstream molecule, CREB, are well known to be one of the major neuroprotective signaling pathways, and promote hippocampal neuronal cell survival [7, 32‐34]. To address the potential role of Akt and CREB in the brain protective effects of linagliptin in klotho−/− mice, we examined the effect of linagliptin on cerebral Akt and CREB of klotho−/− mice. Interestingly, linagliptin treatment significantly increased cerebral Akt and CREB phosphorylation in klotho−/− mice. Collectively, these results suggest that the inhibition of cognitive impairment with linagliptin in klotho−/− mice might be attributed to the activation of cerebral eNOS, Akt, and CREB by linagliptin. Further study is required to elucidate the precise mechanism underlying the amelioration of cognitive impairment by linagliptin in klotho−/− mice.

Klotho−/− mice are almost indistinguishable from wild-type mice regarding the appearance and the growth up to 3 weeks of age, thereafter they stop growing and body weight gain, develop aging phenotypes, and most of the mice die prematurely at 8–10 weeks of age [12, 15, 35]. However, the mechanism underlying body weight loss in klotho−/− mice is obscure. Furthermore, the specific cause of premature death in klotho−/− mice is unclear since each of these aging phenotypes appears not fatal by itself. In the present study, DPP-4 inhibition with linagliptin significantly slowed body weight loss in klotho−/− mice, and greater body weight of klotho−/− mice treated with linagliptin was associated with greater skeletal muscle weight and greater kidney weight and with the prevention of hair loss. Furthermore, linagliptin treatment tended to prolong survival rate of klotho−/− mice. All these findings suggest that DPP-4 inhibition might exert multiple antiaging effects in klotho−/− mice. However, further detailed investigation is required to define the exact role of DPP-4 in the mechanism of premature aging.

Hypoglycemia is also another aging-related phenotype observed in klotho−/− mice [14]. In the present study, compared with wild-type mice, blood glucose levels of klotho−/− mice were very low and maintained at low levels throughout the OGTT, and this finding is in good agreement with the previous report [14]. Of note are the observations that linagliptin significantly improved hypoglycemia in klotho−/− mice, as shown by the findings of OGTT and significantly increased non-fasting blood glucose levels, while linagliptin did not affect serum insulin levels of klotho−/− mice. Thus, linagliptin treatment significantly ameliorated hypoglycemia in klotho−/− mice, probably independently of insulin. However, the present study did not allow us to elucidate the underlying mechanism of amelioration of hypoglycemia by linagliptin in klotho−/− mice. Furthermore, metabolic cage study could not be technically performed in the present study, since individual housing of klotho−/− mice shortens life span because of much vulnerability to stress.

It has been shown that DPP-4 inhibitor-treated type 2 diabetic patients had lower risks for cardiovascular disease as compared to those for non-DPP-4 inhibitor users, except metformin users [36], and its use did not increase the risk of heart failure compared with sulfonylurea [37]. The initial combination of linagliptin and metformin substantially improved glycemic control without weight gain and with infrequent hypoglycemia [38] and also significantly improved microvascular function in the fasting state [39]. Very importantly, type 2 diabetes is known to significantly accelerate premature aging of systemic organs. Taken together with our present findings on benefit of linagliptin in premature aging mouse, all these findings support the notion that DPP-4 inhibitors may be a promising anti-diabetic agent for prevention of cardiovascular disease in type 2 diabetic elderly patients. In conclusion, we for the first time examined the impact of DPP-4 inhibition with linagliptin on premature aging by using klotho−/− mice, a popular animal model of human premature aging. We found that DPP-4 inhibition with linagliptin ameliorated cognitive impairment and hippocampal neurodegeneration, improved CBF, prevented hair loss, lessened body weight loss, and ameliorated hypoglycemia in klotho−/− mice. Thus, our present work provided the evidence suggesting that DPP-4 inhibition might exert anti-aging effects on multiple organs and provided a novel insight into the potential role of DPP-4 inhibition in premature aging.