Angiogenic effects of cell therapy within a biomaterial scaffold in a rat hind limb ischemia model

Critical limb ischemia (CLI) is a life- and limb-threatening condition affecting 1–10% of humans worldwide with peripheral arterial disease. Cellular therapies, such as bone marrow-derived mesenchymal stem cells (MSCs) have been used for the treatment of CLI. However, little information is available regarding the angiogenic potency of MSCs and mast cells (MC) in angiogenesis. The aim of this study was to evaluate the ability of MCs and MSCs to induce angiogenesis in a rat model of ischemic hind limb injury on a background of a tissue engineered hydrogel scaffold. Thirty rats were randomly divided into six control and experimental groups as follows: (a) Control healthy (b) Ischemic positive control with right femoral artery transection, (c) ischemia with hydrogel scaffold, (d) ischemia with hydrogel plus MSC, (e) ischemia with hydrogel plus MC and (f) ischemia with hydrogel plus MSC and MCs. 106 of each cell type, isolated from bone marrow stroma, was injected into the transected artery used to induce hind limb ischemia. The other hind limb served as a non-ischemic control. After 14 days, capillary density, vascular diameter, histomorphometry and immunohistochemistry at the transected location and in gastrocnemius muscles were evaluated. Capillary density and number of blood vessels in the region of the femoral artery transection in animals receiving MSCs and MCs was increased compared to control groups (P < 0.05). Generally the effect of MCs and MSCs was similar although the combined MC/MSC therapy resulted in a reduced, rather than enhanced, effect. In the gastrocnemius muscle, immunohistochemical and histomorphometric observation showed a great ratio of capillaries to muscle fibers in all the cell-receiving groups (P < 0.05). The data indicates that the combination of hydrogel and cell therapy generates a greater angiogenic potential at the ischemic site than cell therapy or hydrogels alone.

www.nature.com/scientificreports/ relationship between angiogenesis and these cells 19,20 . MSCs can differentiate into capillary blood vessels in ischemic areas 21 whilst MCs secrete cytokines and chemokines that enable the migration and differentiation of cells within the ischemic site such as vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), transforming growth factor beta (TGF-β), tumor necrosis factor alpha (TNFα) and interleukin-8 (IL-8) 18 . For example, bFGFs interfere with the migration, proliferation and differentiation of endothelial cells [22][23][24] and enable vascular smooth muscle cells to develop into vessels 25 . In addition, the VEGF family enhance vessel permeability and diameter 24,26 .
In cell therapy, more than 90% of the injected cell suspension is lost and does not engraft 27 . Most studies support the concept that tissue engineering using a three dimensional (3D) porous scaffold improves cell engraftment by controlling cell attachment, mechanical support and the stimulation of new in vivo tissue growth [27][28][29][30] . However, the application of synthetic and chemical materials increases the risk of damage to cells and living tissues. Therefore, the importance of natural scaffolds has been highlighted 27 . A key selection criteria for scaffolds is a 3D scaffold with a proper porosity which does not interfere with angiogenesis 27 . Natural polymers such as alginate and gelatin have no impact on angiogenesis 31 and are cost-effective 29 . The combination of alginate and gelatin has a similar extracellular matrix (ECM) composition to that seen in animals 31 making the combination an appropriate material for tissue engineering.
We hypothesized that applying MSCs or MCs alone or in combination will enhance the development of angiogenesis under ischemic condition. Thus, we evaluated their therapeutic effects as an add-on cell therapy in a rat model of hind limb ischemia (severing of the femoral artery) on a background of tissue modeling and bioengineering using a hydrogel scaffold.

Materials and methods
Study design and animals. Thirty healthy mature male Wistar rats with weight of 200-250 g, were obtained from the animal house, Faculty of Veterinary Medicine, Urmia University, Iran, and randomly divided into six experimental groups (n = 6 per group). Rats were kept in the lab 6 week before the experiment began at an ambient temperature of 23 ± 3 °C, stable air humidity and a natural day/night cycle (14 h light and 10 h darkness). Groups were: (a) Control healthy group; (b) Ischemic control group with ischemia created in the hind limb by femoral artery transection between two ligatures 5 mm apart; (c) Hydrogel scaffold (50 µl) control group; (d) MC injected group; (e) MSC injected group and (f) MCs and MSCs (Mix) group. Hydrogel (alginategelatin) (50 µl) scaffold alone or impregnated with 10 6 of each cell type was applied to the site of ischemia (artery transection). Cells were added to the hydrogels on day 14 after surgery. All procedures were performed based on the guidelines of the Ethics Committee of Urmia University-Iran (Reference No: AECVU-185-2018). The experimental protocols were approved by the Urmia University-Iran ethical committee. The in vivo animal experiments study was carried out in compliance with the ARRIVE guidelines (https:// arriv eguid elines. org).
Surgical procedure. The procedure was carried out based on the guidelines of the Ethics Committee of the International Association for the Study of Pain 32 . Rats were anesthetized by intra-peritoneal administration of ketamine-xylazine (ketamine 5%, 90 mg/kg and xylazine 2%, 5 mg/kg) and then a 5 mm portion of the right femoral artery was ligated and resected to create the hind limb ischemia model. The proximal branches, superficial caudal epigastric, and the muscular branches arteries and veins were also resected 33 .
Histological analysis. Animals were euthanized on day 14 using an overdose of ketamine-xylazine (3 × anesthesia dose) and tissue specimens were taken and fixed in 10% formaldehyde 33 . Paraffin sections (5-7 μm) were prepared using a rotary microtome (Microm, GmbH, Germany). Sections were stained with hematoxylin-eosin (H&E) for histomorphometric studies, Masson's trichrome for collagen distribution, periodic acid-Schiff (PAS) to assess muscle glycogen and with CD34 immunohistochemistry for the analysis of capillary density and vessel diameter at both the transected location and in gastrocnemius muscles. Tissue samples were photographed with a digital camera (Dino-Eye-AM-7023, Taiwan) and analyzed using Dino Capture 2.0 software (Dino-Lite Europe, Naarden, and the Netherlands). In brief, slides were de-paraffinized with xylene and sections rehydrated using an ethanol gradient. The sections were stained with H&E as described before 34 . The number of capillaries and fibers were counted at 5 random 0.0625 mm 2 areas at 400× magnification and their ratios calculated. For histomorphometric evaluation of fibers, cross-sectional muscles were evaluated as described previously 34 .
Rat bone marrow-derived MCs cell isolation. Bone marrow cells were immediately isolated from femur and tibia bones from 6 different animals and MCs derived as described earlier 34,35 . Flow cytometry. MC were characterized by flow cytometry as previously described 35

Rat bone marrow-derived mesenchymal stem cells isolation and characterization. Animals
were euthanized by intra-peritoneal ketamine-xylazine and their hind legs cleanly shaved and prepared aseptically. A skin incision was made and the lateral surface of the femur and tibia exposed. The metaphyseal regions www.nature.com/scientificreports/ of the femur was cut with scissors and the bones flushed using an insulin syringe with 1 ml of endotoxin-free complete media. The collected marrow tissue was placed in DMEM (Dulbecco's Modified Eagles Medium media with antibiotics) and the cells centrifuged for 10 min at 320×g at 4 °C. The cells were cultured at the ratio of 1 × 10 6 /ml in a 25 cm 2 flask in complete media (DMEM, containing 20% FBS and 1% antibiotic mixture of 100 IU/ml penicillin, 100 µg/ml streptomycin) as described previously 36 . Non-adherent cells were removed after 72 h of seeding. The media was changed every 3 or 4 days until the cells reached confluence. Trypsin-EDTA was added for 5 min and the suspension centrifuged at 200g, 4 °C for 5 min to collect the cells. The pellet was re-suspended in buffer (1× D-PBS containing 2% FBS and 0.05% sodium azide) to a final concentration of 5 ~ 10 million cells/ml on ice. Then the cells were aliquot in 100 μl of FACS buffer before addition of antibodies. Cells were stained with CD54-FITC (cat No. 554969), CD90.1-FITC (cat No. 561973), CD45-PE (cat No. 559135) and CD31-PE (cat No. 555027) and then compared with matched isotype control antibodies (Mouse IgG1 (cat No. 551954) for 30 min at 4 °C. All antibodies were purchased from BD Biosciences, USA. Six MSC lines were generated and used in these experiments. For analysis, the cells were washed with FACS buffer twice and then suspend in 100 μl cell FACS buffer. 10,000 events were analyzed by flow cytometry (FACS Calibour, BD, USA) and the data analyzed by FLOJo version 8. Results were calculated as percentages within the indicated gates.
The hydrogel solution was prepared as previously described 35 . Briefly, 75 μl hydrogel was added to 10 6 cells before addition of 7.5 μl horseradish peroxidase (HRP) with gentle mixing. The resultant mixture was then implanted at the surgical site and 15 μl hydrogen peroxide was immediately added to the scaffolds in vivo 40,41 . Immunohistochemical analysis. Tissue section slides were heated at 60 °C for 25 min in oven (Venticell, MMM, Einrichtungen, Germany). Sections were gently washed in washing buffer and then incubated with an anti-CD34 (a marker of endothelial progenitor cells) primary antibody (1:5000, Abcam, Cambridge, MA, USA, cat No. ab81289) for 15 min. Diaminobenzidine-substrate-chromogen was then added for 5 min before slides were washed and counterstained using hematoxylin for 5 s. Sections were then dipped in ammonia solution (0.037 M/L) 10 times, rinsed with distilled water before being cover-slipped. Positive brown staining was observed under a light microscope. The capillary count was conducted by optical microscope and Dinolite lense digital camera (Dino-Eye-AM-7023) and related software at magnification of 400× at the area of 0.0625 mm 2 . Larger vasculatures were counted at magnification of 100× at the area of 0.88 mm 2 at each tissue section. 5 samples from separate animals in each group were examined 35 . Determination of collagen fiber density. Masson trichrome-stained collagen fibers were visualized by light microscopy (Zeiss, Cyber-Shot, Japan). To assess the collagen intensity in the histological sections, the pixel-based intensity of blue-staining, representing the collagen fibers, was assessed at 2530 µm × 2530 µm sections of a photomicrograph by software analysis (Image pro-insight, Media Cybernetics, USA). For this purpose, 20-megapixels images were prepared by the on-board camera (Zeiss, Cyber-Shot, Japan), then, the area (2530 µm × 2530 µm) was measured by MEZZURE software on the photomicrograph, and the means of pixelbased intensities, obtained from 3 images from a section (in total 15 section/each group), were evaluated. Finally, the mean ± SD of intensities was compared between groups 34 .
Statistics. The data were analyzed by GraphpadPrism version 8. All values are expressed as mean ± SEM.
Differences between experimental groups were analyzed using one-way ANOVA. A Bonferroni test was used to identify significant differences between the different pairs of groups. The level of significance was set at P < 0.05.

Differentiation of bone marrow cells into Mast cells and mesenchymal stem cells. MCs were
isolated and cultured over time. Upon plating of bone marrow cells, both adhesive and non-adhesive cells were present. The number of cells present reduced after 5-10 days of culture/passage 3-4 and they took on a heterogeneous morphology. By day 23 the cells took on a mature appearance with granules and a round shape (Fig. 1A). Flow cytometry analysis of harvested cells at day 23 identified > 92% of cells positive for c-kit/CD117 (Fig. 1Ba), FCεRI (Fig. 1Bb) and dual c-kit/CD117-FCεRI (Fig. 1Bc) positive cells. For the development of MSCs, bone marrow cells were cultured in media for up to 72 h after cell seeding and removal of non-adherent cells. Adherent cells at confluence (at third or fourth passage) had the appearance of a single population of MSCs with a spindle-shaped morphology (Fig. 1C). Flow cytometric analysis of MSCs show positive staining against the MSC surface markers CD54 and CD90 but negative for CD45 and the isotype IgG1 control (Fig. 1D).
Preparation and implantation of the hydrogel scaffold. The hydrogel scaffold was prepared through the main component of alginate-gelatin and then the cells encapsulated by hydrogel solution at density of 10 6 cells per 75 μl hydrogel (Fig. 2A). The right femoral artery was ligated and resected to create the hind limb ischemia model (Fig. 2B). Hydrogen peroxide was added to the cells encapsulated by hydrogel solution prior to implantation (Fig. 2C). The resultant mixture was immediately implanted into the resected area of femoral artery (Fig. 2D). www.nature.com/scientificreports/ Determination of ischemia. On day 21 after the operation, evidence of necrosis was macroscopically visible in the foot pads and fingers (arrowed area in Fig. 3A). A combination of H&E, Masson's trichrome staining and IHC for CD34 (arrowed) shows the presence of endothelial cells within capillaries (Fig. 3B). A representative histomorphometric analysis of collagen fiber distribution on tissue vessels is shown (Fig. 3C). There was no significant effect of ischemia or of any treatment on collagen fiber distribution (Fig. 3D).

Histomorphometry of femoral artery area. Qualitative evaluation of angiogenesis in the resected area
of the femoral indicated that application of the hydrogel scaffold alone did not stimulate angiogenesis although this was seen with cell therapy (Fig. 4A). In addition, capillaries were counted at the site of femoral artery resection using an optical microscope (magnification 400×) and a graded lens (1.16 mm square mesh size). Ischemia resulted in a significant decrease in capillary density which was significantly reversed in all three treatment groups with hydrogel (P < 0.05) (Fig. 4B). Hydrogel/MCs and hydrogel/MSCs groups show a significantly enhanced capillary density compared ischemia and scaffold controls (Fig. 4B). Interestingly, the combined MSCs plus MCs with hydrogel reduced the ability of MSCs to enhance capillary density (Fig. 4B). Based on morphology, we classified the results into 3 groups according to cross-sectional diameter (30-50, 50-100 and > 100 μm) at the site of femoral artery transection (P < 0.05) (Fig. 4C). Ischemia induces a significant increase in small vessels and a reduction in the numbers of medium and large vessels (P < 0.05) (Fig. 4C). The effect of MC and MSC therapy was similar with respect to the formation of small and large vessels (Fig. 4C).
In contrast, MSCs did not affect medium sized vessel formation compared to the hydrogen scaffold. In addition, the combination of MCs plus MSCs enhanced small vessel formation compared to MCs and MSCs alone whilst the combination did not significantly enhance large vessel formation (Fig. 4C). The hydrogel scaffold alone reduced the ischemia-induced enhancement of small vessels and further suppressed the numbers of medium vessels (Fig. 4C). MCs and MSCs significantly enhanced the numbers of small and large vessels compared to www.nature.com/scientificreports/ the hydrogel scaffold alone whereas only MCs enhanced medium-sized vessel formation with no effect of MSCs (Fig. 4C). The hydrogel scaffold alone had no significant increase in large vessel formation (Fig. 4C).
Histologic changes in study groups. The numbers of capillaries were assessed by H&E, Masson's trichrome and CD34 staining. A representative CD34+ endothelial cell within a vessel between the gastrocnemius muscle fibers is indicated by an arrow (1500 magnification) (Fig. 5A). The capillary to muscle fiber ratio was significantly reduced in the ischemia group compared to controls (P < 0.05). This was significantly improved by the hydrogel scaffold but this remained significantly reduced compared to the control levels (Fig. 5B). Although, addition of MCs, MSCs and combined MCs plus MSCs did not significantly affect the ratio compared to hydrogel scaffold alone, these were no longer different from the ratio seen in control animals (Fig. 5B).

Discussion
In a femoral artery transection model of CLI the application of hydrogel, MCs and MSCs induce angiogenesis by the development of the capillaries. Ischemia reduced the number of capillaries and large blood vessels but increased the number of small vessels and these changes were associated with reduced capillary to gastrocnemius muscle fiber ratio. Ischaemia-reduced capillary density and number of blood vessels in the region of the femoral artery transection were reversed in animals receiving MSCs and MCs. Furthermore, in the gastrocnemius muscle, immunohistochemical and histomorphometric observation showed a great ratio of capillaries to muscle fibers in all the cell-receiving groups. In contrast, there was little or no effect of the hydrogel scaffold alone on capillary density and number of vessels. This data supports the concept that tissue engineering with a hydrogel scaffold overcomes issues of cell engraftment seen previously 28,35 . Overall, the induction of angiogenesis by cell-based therapy in combination with the hydrogel scaffold was associated with increasing of distribution of collagen fibers and reduction of the capillary: gastrocnemius muscle fiber ratio. Morphometric analysis of capillary angiogenesis showed the development of blood vessels in both MC-and MSC-treated groups. Interestingly, the enhancement of capillary density was seen mainly in the MC-treated group. In addition, the combination of a hydrogel scaffold with MCs showed significantly increases the muscle fiber diameter with an concomitant reduction in capillary-to gastrocnemius muscle fiber ratio. Angiogensis www.nature.com/scientificreports/ following any insult is regulated by the release of factors from endothelial cells 42,43 . One important stimuli is chronic hypoxia which leads to the secretion of growth factors from stromal or parenchymal cells including vascular cells. However, these factors alone were not sufficient to obtain complete repair of the local damage [43][44][45] . Cellular therapies play an important role in angiogenesis 8,9 and MSCs in particular can promote angiogenesis and arteriogenesis through stromal and paracrine actions 10,11 especially during the treatment of injured tissues 14,46,47 . MSCs act, not only through the paracrine secretion of growth factors, but also induce the differentiation of endothelial cells, pericytes and myofibroblasts 45,48-50 . MCs are activated in early phase after the induction of ischemia and mast cell degranulation occurs in the ischemic environment of the hind limb in the ICL model used here. However, the precise mechanism driving degranulation is largely unknown 51 . MCs do not differentiate into other cells, but secrete cytokines and other factors that enable the migration and differentiation of the other cells within the ischemic area 18 .
Emerging evidence indicates the relationship between angiogenesis, MSCs and MCs 17,18 . For example, MCs 17,18,51 and MSCs 14,19 are reported in or near the site of capillary germination which suggests a role of these cells in angiogenesis. MSCs have inhibitory effects on MCs that prevents their migration, degranulation and discharge 48,52 . Thus, it seems that the cross-talk between MSCs and MCs in the development of angiogenesis is very critical.The present study shows that the mean number of capillaries in the MC-treated group was similar Despite these strengths of the study there were some limitations. We did not confirm the identity of the isolated mesenchymal cells following FACs analysis although this has been described by others [53][54][55][56][57] . In addition, it may have been useful to analyse tissue further from the implanted meshes to further understand the pathobiology. In addition, there are more modern acellular composite matrices that may represent better opportunities to improve cell therapy strategies for CLI particularly if they were initially assessed in large animal models compared to the smaller rat model used here.

Conclusions
The results of current study indicated that bioengineered tissue incorporating MCs and MSCs within a hydrogel scaffold stimulates angiogenesis under ischemic conditions and lead to anastomosis in vessels. Both MCs and MSCs may prove important therapeutically in driving angiogenesis under conditions where critical limb ischemia has occured. However, further research in this area is required to determine the optimal combination of scaffold and cells that induce angiogenesis.  Immunohistochemical staining for CD34 to detect endothelial cells. Positive staining is indicated by the dark brown arrowed area between the gastrocnemius muscle fibers. (B) Histogram showing the effect of ischemia and interventions on capillary to gastrocnemius muscle fiber ratio. All values are expressed as the mean ± SEM of results from 5 animals. Asterisks indicate statistically significant differences compared to other samples. *Indicate P value is P < 0.05; **Indicate P value is P < 0.01; *** Indicate P value is P < 0.001.