Dipeptidyl dipeptidase-4 inhibitor, MK-0626, promotes bone marrow-derived endothelial progenitor cell bioactivities for vascular regeneration in diet-induced obese mice

Metabolic syndrome (MS), overlapping type 2 diabetes, hyperlipidemia, and/or hypertension, based on high-fat diet, poses risk for cardiovascular disease. A critical feature associated with such risk is the functional impairment of endothelial progenitor cells (EPCs). Dipeptidyl dipeptidase-4 inhibitors are known not only to inhibit degradation of incretins to control blood glucose levels, but also to improve EPC bioactivity and induce anti-inflammatory effects in tissues. In the present study, we investigated the effects of such an inhibitor, MK-06266, in ischemia model of MS using diet-induced obese (DIO) mice. EPC bioactivity was examined in MK-0626-administered DIO mice and non-treated control group, using an EPC colony-forming assay and bone marrow cKit+ Sca-1+ lineage-cells, and peripheral blood-mononuclear cells. Our results showed that, in vitro, the effect of MK-0626 treatment on EPC bioactivities and differentiation was superior in comparison with non-treatment. Further, in vivo hindlimb ischemia model experiment indicated that microvascular density and pericyte-recruited arteriole number were increased in MK-0626-administered group, but not control group. Lineage profiling of isolated cells from ischemic tissues disclosed that MK-0626 administration has an inhibitory effect on unproductive inflammation. This occurred via a decrease in the influx of total blood cells and pro-inflammatory cells such as neutrophils, total macrophages, M1, total T-cells, cytotoxic T-cells, and B-cells, with a concomitant increase in number of regeneration-associated cells, such as M2/M ratio and Treg/T-helper. Laser Doppler analysis revealed that at day 14 after ischemic injury, blood perfusion in hindlimb was grater in DIO mice treated with MK-0626, but not in control. In conclusion, the dipeptidyl dipeptidase-4 inhibitor has a positive effect on EPC differentiation in MS model of DIO mice. Following ischemic injury, DPP-4 i sharply reduces recruitment of pro-inflammatory cells into ischemic tissue, and triggers regeneration and reparation process. Thus, DPP-4 i is a promising therapeutic agent for MS treatment.


Introduction
Vascular regeneration is an initial and essential process for organ regeneration.
Endothelial progenitor cells (EPCs) play key roles in vasculogenesis [1] and in regulating angiogenesis [2]. Thus, EPC kinetics and bioactivities are essential for vascular regeneration and organogenesis in regenerative medicine. Type 2 diabetes is associated with the decrease in number and impaired function of circulating EPCs, which in turn has been linked to cardiovascular disease complication [3]. There are several options to stimulate EPC proliferation and biological functions in diabetes that are being actively pursued by researchers [4]. For example, several cytokines such as stromal cell-derived factor-1 (SDF-1), granulocyte colony stimulating factor (G-CSF), and granulocyte macrophage colony stimulating factor (GM-CSF), angiogenic growth factors such as vascular endothelial growth factor (VEGF), and pharmaceutical drugs such as estrogen and statins, have all been reported to augment EPC bioactivities, such as proliferation, differentiation, migration, mobilization, and recruitment of BM-derived EPCs [5,6].
Fadini et al. demonstrated that the DPP-4 i sitagliptin increases the level of circulating EPCs in type 2 diabetic patients, with concomitant up-regulation of SDF-1 [10]. However, at 4 that time, EPC research methodologies were still relatively underdeveloped to obtain clear scientific insights into stem cell biology for vascular regeneration. Recently, our laboratory developed new EPC research methodologies, EPC-CFA (new EPC colony forming assay) and QQ-EPC culture (quality and quantity controlled serum-free EPC culture technique), to identify a variety of EPC phenotypes and functions [11]. These methodologies allow us to precisely evaluate the key factors of EPC kinetics and bioactivities, and ultimately lead us to understand how EPC differentiates in healthy or diseased states to promote new vasculature.
In this study, we investigated whether a sitagliptin analogue, MK-0626, affects EPC kinetics in peripheral blood or bone marrow in diabetic animal models, using the new EPC quality and quantity evaluation methods. Furthermore, the developed cell isolation technique from ischemic muscles was used to define and compare in situ cell phenotype and quantity of hematopoietic cells in tissues, in order to evaluate inflammatory cell kinetics in diabetic animals following MK-0626 administration.

Materials and methods
All studies were performed with the approval of national and institutional ethics committee. The Tokai School of Medicine Animal Care and Use Committee gave local approval for these studies, based on Guide for the Care and Use of Laboratory Animals (National Research Council).  week. Based on previous report, this dose of MK-0626 was predicted to result in continuous blocking of incretins, such as GLP-1 and GIP, and inactivation of DPP-4 [13]. Food intake of the mice was recorded every two days and their body weight (BW) and blood sugar (BS) were measured 9 and 3 days before surgery, and on day 4 and day 11 after surgery. Based on BW at each time point, the volume of MK-0626 solution was adjusted to maintain the same dose in each mouse until the subsequent measurement. The BS was measured using a blood glucose test meter (Gultest Ace R, ARKRAY Factory, Inc. Shiga, Japan) and disposable blood glucose level test sensor (Gultest sensor, Panasonic Healthcare Co., Ltd.).

Reagents
At the end of the experimental period, the mice were anesthetized with pentobarbital and their plasma collected and stored at -80°C.

Cell preparation and culture
Mouse peripheral blood mononuclear cells (mPBMNCs) were collected with 27G-Insulin syringe (TERUMO, Tokyo, Japan) from the apex of the heart under adequate anesthesia using 1.5% to 2.0% isoflurane (Dainippon Sumitomo Pharma Co., Ltd., Osaka, JAPAN). The cells were further isolated by density gradient centrifugation with Histopaque (d=1.083; Sigma-Aldrich Co., St Louis, MO, USA), as previously reported [14].
Preparation of human peripheral blood mononuclear cells (hPBMNCs) was performed after obtaining informed consent from healthy volunteers according the Tokai University institutional the review board. The peripheral blood (10 mL) was drawn by heparinized venous puncture at the forearm. Isolation protocol for hPBMNCs was the same as that for mPBMNCs.
Briefly, cells were cultured with QQ (Quality and Quantity) culture media of Stem Line II

Enrichment of EPCs from bone marrow
Bone marrow (BM) EPCs were isolated from no-fat diet and DIO-mice femurs and tibias as previously described [15]. Nuclear cells were washed with PBS-EDTA and the erythrocytes were removed by hemolyzation lysis buffer. Nuclear cells were initially stained

Endothelial lineage characterization
As previously described [16], after culturing PBMNCs for 7 days in the endothelial cell growth medium (EGM-2

Induction of hindLimb ischemia model
The mice were divided into two groups, Lean and DIO, one week before surgery.
Hindlimb ischemia induction (HLI) was performed under adequate anesthesia by 1.5% to 2% isoflurane to minimize pain, according to the 3Rs rule (replacement, reduction, and refinement). Briefly, the proximal portion of the left common and deep femoral artery with their three branches were successfully ligated with a 6-0 nylon suture (Sigma Rex., Kono manufacturing Co., Ltd. Ichikawa, Japan), and the proximal and distal portions of the saphenous artery were subjected to bipolar electrocautery (MERA N3-14; SENKO MEDICAL INSTRUMENT mfg. Co., Ltd., Tokyo, Japan). The skin was closed with a 4-0 nylon suture (Sigma Rex., Kono manufacturing Co., Ltd. Ichikawa, Japan). To reduce post-9 surgery suffering from pain, pentobarbital (Kyouritu Seiyaku Co., Ltd., Tokyo, Japan) was injected at a dose of 10 mg/kg via i.p. administration.

Laser doppler imaging and blood flow assessment
Baseline laser Doppler perfusion imaging (LDPI; Moor Instrument, Axminster, UK) was performed of animals under anesthesia with 1.5% isoflurane (Dainippon Sumitomo Pharma) and after induction of ischemia at days 0, 7, and 15 to assess blood perfusion ratio in ischemic vs. healthy hindlimb. Acquired data using LDPI were analyzed with moorLDI TM Main software (Laser Doppler Imager ver 5.2; Moor Instruments, Devon, UK). Mice with toe necrosis or limb salvage were included in the study, whereas those with foot necrosis or autoamputation were excluded.

Identification of phenotypes of recruited cells isolated from ischemic tissues
All mice were fed with 5 g HFD per day till enough BS and BW could be retained; then they were divided into two groups, control (received 0.25% methylcellulose only) and MK-0626 group (MK-0626 was administrated per os using sonde 3 days before and 3 days after onset of LHI). At day seven, mouse was sacrificed after anesthesia, and systemically perfused with cold PBS to exclude blood cells. An anterior tibial muscle (ATM) was dissected for further isolation of cells that had infiltrated the ischemic tissue. Our previous immunohistochemistry analysis study showed that ATM is the most sensitive for ischemic injury. In brief, ATM muscle vessels, tendons and nerve fibers were removed under light microscope and minced by using optical fine micro scissors. To effectively liberate skeletal muscle cell types, collagenase type II (500 U/mL) (Worthington Lab.) and collagenase/dispase 10 (1 mg/mL) (Roshe Diagnostics) were used for 1.5 h at 37°C with gentle agitation, as reported elsewhere [17]. After digestion, the tissue was triturated and meshed through a 70-m cell strainer. Finally, cells were washed twice with DMEM (Gibco) and then counted using a hemocytometer. The Fc receptors were blocked with mouse anti-Fc receptor (Biolegend Co.

Statistical analysis
All values are shown as mean ± SE. Mann-Whitney U and Kruskal-Wallis test were used for two and three non-parametric groups with Dunn's multiple comparison test, respectively. For multiple comparisons between groups at different time points, 2-way ANOVA was applied, followed by Tukey's post hoc test. All statistical analyses were performed using GraphPad Prism 7.1 (GraphPad Prism Software Inc., San Diego, CA, USA). P < 0.05 value was considered to indicate statistically significant differences.

EPC differentiation was induced by MK-0626
To evaluate the efficacy of MK-0626 on EPC's colony forming ability and differentiation, PBMNCs were isolated from Lean mice, MK-0626-administered Lean mice, DIO mice, and MK-0626-administered DIO mice, and cultured in semisolid culture media. The number of definitive EPC-CFUs (dEPC-CFUs) from control DIO mice decreased compared to that from control Lean mice, whereas dEPC-CFUs from MK-0626-administered DIO mice were similar in number to those from the Lean littermates (Fig 1A). The frequency of dEPC-CFUs per peripheral blood (PB) volume also sharply decreased in DIO mice compared with that in control mice, while MK-0626-administered mice showed a dramatic improvement in the number of dEPC-CFUs per 1 mL PB (Fig 1B). These data suggest that EPC differentiation in PBMNCs was impaired in DIO mice and MK-0626 treatment effectively recovered the EPC differentiation ability to similar levels seen in Lean mice. We further investigated whether DPP-4 i administration affected bone marrow c-Kit + /Sca-1 + /Lin -(BM-KSL) stem cells in obese and healthy condition. To address this, purified BM-KSL stem cells, isolated from either control (lean) or DIO mice, were evaluated for their EPC colony formation abilities. As shown in Fig. 1C, BM-KSL cells displayed similar EPC-CFUs between all tested groups, indicating that the EPC differentiation ability was not affected by obese condition or MK-0626 administration in BM. However, the calculated number of EPC-CFUs in hemi bone represented stimulated expansion of EPCs in DIO mice, and MK-0626 did not affect the frequency of EPC-CFUs in BM in both DIO and Lean mice (Fig 1D).
To investigate MK-0626 effect on EPC biology, we employed ex-vivo regenerative conditioning in culture with or without MK-0626 on BM-derived KSL cells from DIO or Lean mice. After one week of culturing, colony-forming assay (CFA) was used to evaluate these cells. Conditioning BM-KSL cells with MK-0626 promoted the EPC colony forming potential compared with Lean mice without MK-0626 treatment (Fig 2B). However, EPC-CFA values were higher in DIO mice compared with those in control lean mice, and did not further improve upon MK-0626 administration. Two thirds of the EPCs in treated and non-treated Lean mice were in the pEPC stage, while half of the EPCs in treated and un-treated DIO mice were definitive EPC colonies, which are responsible for vasculogenesis (Fig 2C).

MK-0626 recovered human EPC colony formation and differentiation capability under inflammatory conditions
To verify the effect of MK-0626 on human EPCs, we performed EPC-CFA and EPC culture assay on regenerative conditioned PBMNC. Diet-induced obesity (DIO) develops systemic chronic inflammatory milieu, by increasing secretion of pro-inflammatory cytokines, such as TNFα, IL-6 and IL-1b [18,19]. Based on this, we attempted to assess whether EPC differentiation cascade differs under a normal regenerative condition (QQ culture) or an inflammatory condition induced by TNFα (QQ + TNFα). EPC differentiation was evaluated 13 by EPC-CFA before and after QQ culture of hPBMNCs. Inflammatory conditioning markedly decreased EPC colony forming units (Fig 3Ac and 3B), suggesting that inflammatory conditioning significantly impairs EPC colony formation bioactivity. In contrast, MK-0626 supplementation together with TNFα beneficially recovered the EPC function under inflammatory conditioning (Fig 3Ad and 3B), while MK-0626 treatment alone did not increase EPC activity in regenerative conditioning. This may suggest that under regenerative conditioning healthy EPC activation is at its peak and may not respond to additional stimulation with MK-0626. Furthermore, co-staining showed that UEA-I Lectine and DiI-Ac LDL EPCs were decreased significantly in TNFα-treated DIO cells, while combination of MK-0626 and TNFαrecovered the EPC numbers to the same extent as in the healthy Lean group (Fig 3Ca   through d, and 3D), suggesting that MK-0626 recover EPC function under inflammatory conditions. Together, these results depicted that MK-0626-treatment under inflammatory conditioning favorably enhanced EPC functions such as colony formation and differentiation.

MK-0626 administration promoted blood flow perfusion after HLI in DIO mice
To demonstrate in vivo blood flow perfusion recovery, we induced mouse HLI model in healthy and DIO mice and examined the effects using laser Doppler perfusion imaging. After HLI surgery (left-side limb), DIO mice blood flow perfusion deteriorated sharply (0.23 ± 0.03), in comparison with MK-0626-treated DIO mice. Moreover, at day 13 after HLI surgery, DIO mice treated with MK-0626 showed significant improvement of blood flow (0.46 ± 0.03, P < 0.05), similar to that seen in the Lean groups (0.47 ± 0.06 and 0.53 ± 0.06 in Lean and Lean treated with MK-0626, respectively) (Fig 4A and 4B). This data showed that DPP-4 i accelerates angiogenesis for further enhancement of blood flow and limb salvage in DIOconditioned mice.

MK-0626 administration reduced infiltration of proinflammatory hematopoietic cells lineages
At day 3 after onset of HLI, infiltrated blood cells in the ischemic anterior tibial muscle were successfully isolated and analyzed using flow cytometry for cell subset determination (Fig 5).
Interestingly, DPP-4 i treatment mainly inhibited influx of M1 subset (P < 0.02) along with an upward trend of M2 infiltration from the total macrophages (Fig 6C, 6D, and E). This suggests that DPP-4 i diminished total monocyte/macrophages accumulation in the ischemia-injured tissue, and converted them into the M2 population, which displays strong anti-inflammatory, reparative and angiogenic functions. Interestingly, the influx of stem or progenitor cell (CD117 and Sca-1) was greater in DIO group (P < 0.02) than in the MK-0226-treated DIO group (Fig   6F and 6G).
In summary, MK-0626 administration suppressed the influx of all hematopoietic 15 lineage cells (mainly pro-inflammatory cells) and accelerated regeneration-associated cell (RACs) polarization from pro-inflammatory M1 and CD4 + cells to the antiinflammatory/reparative M2 and T reg cells, thus inhibiting unproductive inflammatory cascades following ischemic injury for further beneficial tissue restoration.

MK-0626 administration enhanced MVD and pericyte recruitment in ischemic tissues
To evaluate MVD and pericyte recruited arterioles, IsolectineB4 and α-SMA costaining was performed. The immunohistology study revealed that in MK-0626-treated animals MVD was dramatically increased in comparison with DIO control (MVD/mm 2 ; 872 ± 56 in MK-0626-treated vs. 528 ± 44 in DIO control, P < 0.001) (Fig 7A, a-d and 7B). To verify effect of MK-0626-treatment on arterial maturation, we stained ischemic hind limb tissues with α-SMA to detect pericyte recruitment. As shown in Fig 7A (e-h) and 7C, the number of α-SMA positive vessels were superior in MK-0226-administered DIO mice (204 ± 81, P < 0.05), compared with that in DIO mice (160 ± 94). These results indicate that MK-0626 administration promoted MVD enhancement along with pericyte recruitment for vascular maturation after ischemic injury.

Discussion
In the present study, we have demonstrated that the DPP-4 inhibitor MK-0626 enhanced vascular development by promoting EPC differentiation and orchestrating regenerative microenvironment of ischemic tissues through reduced influx of proinflammatory cells, such as neutrophils, M1 macrophages, cytotoxic T-cells and B-cells, and instead recruiting regeneration-associated cells, such as M2 macrophages and T reg cells in DIO 16 mice following onset of acute HLI (Fig 8).

Effect of MK-0626 on EPC bioactivity
First, we examined EPC colony forming abilities in PB of animals to evaluate MK-0626-triggered EPC mobilization effects. Definitive EPCs (dEPCs), representing more differentiated colony forming EPCs than primitive EPCs (pEPCs), significantly decreased in Former experiments showed that DIO condition aggravated stromal derived factor-1a (SDF-1a, CXCL12/CXCR4) axis, which is crucial for guiding hematopoietic cells, including EPCs, to the site of injury. SDF-1 is a substrate of DPP-4, and MK-0626 suppress degradation of SDF-1 by inhibiting enzyme activity of DPP-IV, consequently, the SDF-1 increases biological activities in PB and BM [10]. In accordance with this, our finding indicated that diabetic condition specifically impaired function of dEPC mobilization from BM to circulation.
Under such conditions, DPP-4 i might restore the mobilization through preservation of SDF-1 protein by modulating enzyme activity of CD26/DPP-IV.

MK-0626 effect on vascular development
In vivo transplantation experiment demonstrated that while limb perfusion recovery of DIO animal at day 14 was significantly decreased in comparison with healthy Lean mice recovery, MK-0626 treatment abrogated the deterioration in DIO ischemic mice.
Immunohistology analysis revealed that vascular regeneration, such as IB4 stained capillary densities as well as pericyte-triggered arteriole maturation, were superior in MK-0626administered DIO group, likely because the DPP-4 i promotes differentiation of primitive EPC to the definitive EPC to stimulate angiogenesis and arteriogenesis. In acute arterial injury model, short-term inhibition of DPP-4 i enhances endothelial regeneration through inhibition of SDF-1a degradation, and consequently increasing recruitment of circulating endothelial progenitor cells crucial for blood vessel development [20]. DPP-4 i is considered to promote vascular regeneration by two synergistic effects: by inhibiting the degradation of SDF-1a and via anti-inflammatory effects. Such incretin-based therapies with DP-4Ⅳ i, displays antiinflammatory activity through glucose control via DPP-4 [21] Pathological and physiological angiogenesis initiation mainly occur after inflammation [22]. However, on T2DM, chronic inflammation induces excessive inflammatory cytokine secretion, which affects stem cell and EPC differentiation for effective vascular regeneration [23].

Anti-inflammatory effect of MK-0626 to the ischemic tissue
DPP-4 i has been reported as an anti-inflammatory medicine in chronic inflammatory diseases, including T2DM [24], [25]. We assumed this DPP-4 i anti-inflammatory effect might contribute to vascularization and tissue regeneration, especially in situ ischemic tissue through recruitment of immune cells [26,27]. Our hematopoietic cell isolation experiment from 18 ischemic muscles suggested that DPP-4 i had an inhibitory effect on unproductive inflammation by decreasing influx of total blood cell accumulation (by 5-fold), and proinflammatory cells such as neutrophils (by 6.2-fold), total macrophages (by 7.4-fold), M1 (by 7-fold), total T-cells (by 2.2-fold), cytotoxic T-cells (by 1.6-fold), and B-cells (by 4-fold), and by increasing regeneration-associated cells, such as M2/M ratio (by 2-fold) and T reg /T-helper ratio (by 2-fold). Recent research also highlighted that DPP-4 inhibition attenuates obesityrelated inflammation, atherosclerosis, and insulin resistance by regulating M1/M2 macrophage polarization [28,29]. In an ApoE −/− mouse model on high cholesterol diet, long-term treatment with the DPP4 inhibitor sitaglipin significantly reduced atherosclerotic plaque, and this effect was inversely correlated with number of M2 macrophages in the plaque. Blockade of CXCR4/SDF-1 signaling by AMD3100 inhibited aortic M2 accumulation and the therapeutic effect of Sitagliptin [28]. Interestingly, in our study, c-Kit and Sca-1 cells significantly infiltrated into the DIO mice ischemic tissues, in comparison with the MK-0626-treated counterpart. Our ex vivo data revealed that proportion of primitive EPCs was higher than definitive EPCs in PB and BM in DIO mice, indicating that obesity-induced inflammatory milieu decrease stem/progenitor cell differentiation in ischemic tissue due to unproductive inflammation, which supports other studies [23,30,31]. The DPP-4 inhibitors, such as Sitagliptin [32], Vildagliptin [33], Linagliptin [34], Teneligliptin [35], Anagliptin [36], Trelagliptin [37], and Omarigliptin [38] are already in use for clinical treatment of diabetes.
Sitagliptin increases the number of circulating progenitor cells in mouse models [39] and T2DM patients [10,40]. Similar to Sitagliptin, in our study, MK-0626 also increased the number of circulating endothelial progenitor cells in healthy mouse. Many research groups have reported that to increase the number of circulating EPCs in T2DM, EPCs kinetics and the vascular regeneration is important to prevent the disease [41,42]. Some have also reported that DPP-4 i decreases the risk of cardiovascular disease in T2DM [43][44][45][46].
To conclude, our study shows that DPP-4 inhibition has a beneficial effect on vasculogenesis by enhancing EPC differentiation and bioactivity. Moreover, DPP-4 inhibition decreased the recruitment of pro-inflammatory cells into the ischemic injury, along with an increase in regeneration-associated cells, the latter being important in the tissue restoration and regeneration. Further clinical trials on metabolic syndrome need be conducted to prove the therapeutic potential of DPP-4 i in clinical practice as preventive medicine of arteriosclerosis. (A) Laser Doppler imaging was used to analyze blood flow 14 days after ischemia.
The panels show lean mice (Aa), MK-0626-administered lean mice (Ab), DIO mice (Ac), and MK-0626-administered DIO mice (Ad). The region of interest (ROI) for blood flow measurement is shown by the yellow square. (B) The graph presents blood flow ratio of ischemic-to-contralateral hind limb, during the observation period at day 14. Experiments were repeated twice. In the graph, *P < 0.05, **P < 0.01, and ***P < 0.001, determined by Kruskal-Wallis and followed Dun's multiple comparison test. recruitment into the LHI tissue showed an upward trend in the DIO group, in comparison with the MK0626-administered DIO group, although this difference was not statistically significant (P > 0.11) (Fig 6L).

Fig 6. Quantification and characterization of ischemic tissue-infiltrated cells.
(A-L, except E and K) All values are the absolute number of cells that infiltrated per anterior tibial muscle (n = 4-5, each). (E) M2 to total macrophages and (K) Relative ratio of T reg cells to total T cells in CD45 + cells (n = 4-5, each). Data were analyzed by Mann-Whitney U test. group. Experiments were repeated twice with similar results. In the graph, *P < 0.05, **P < 0.01, and ***P < 0.001 as determined by One-Way ANOVA and Dunn's multiple comparison test. Summary of our study.