Histological Study on the Effects of Allogenic Bone Marrow Derived Mesenchymal Stem Cells Along with Local Insulin Therapy on Osteosynthesis in Gap Defects of Radius in Diabetic Rabbit Model

Diabetes mellitus is a metabolic disorder which impairs fracture healing leading to delayed union, mal-union, and non-union. In diabetic animals, bone repair abnormality is insulin dependent as it is reversed by insulin treatment, showing a specific cause and effect relationship between inadequate insulin production and abnormal bone formation (Lu et al., 2003). As insulin promotes bone remodeling and bone formation, exogenous insulin may be used to manage bone homeostasis, osteopathology and accelerate fracture healing in diabetic animals (Paglia et al., 2011).
       
Biodegradable implants like hydroxyapatite filled in bone defects provide scaffold to support the attachment and migration of newly formed bone cells into the osseous defect and help in formation of vascular network through a newly formed bone (Southwood et al., 2005). Mesenchymal stem cell loaded scaffolds provide both osteoinductive stimulants and osteoconductive properties, thus may be used as an alternative for autogenous bone grafting. As delayed healing of fracture is a major problem in diabetic veterinary patients, it was thought that allogenic stem cell therapy might prove helpful in such patients.
       
The aim of present study was to evaluate and compare the histological effects of allogenic bone marrow derived mesenchymal stem cells (BM-MSCs) seeded in hydroxyapatite scaffold along with local insulin therapy and local insulin therapy alone on the osteosynthesis in gap defect of radius in experimentally induced diabetic rabbit model.

The study was conducted in the Department of Surgery and Radiology, G.B.P.U.A and T, Pantnagar during the year 2012, in accordance with guidelines laid down by the Institutional Animal Ethics Committee. Thirty six, clinically healthy, New Zealand White rabbits of either sex aged 6-7 months and weighing 2-2.5 kg were randomly divided into four groups (n=9 animals each): Group A (Healthy control): Hydroxyapatite + working solution of growth media, Group B (Diabetic control): Hydroxyapatite + working solution of growth media, Group C (Treatment group 1): Hydroxyapatite + working solution of growth media + local insulin therapy @ 2 IU/kg b.wt., Group D (Treatment group 2): Hydroxyapatite + cultured allogenic bone marrow derived mesenchymal stem cells + local insulin therapy @ 2 IU/kg b.wt.
       
Working solution of cell growth media was prepared by adding 10% fetal bovine serum (FBS) and 1% antibiotic-antimycotic solution (i.e. mixture of 100 units/ml of Penicillin+ 100 μg/ml of Streptomycin + Nistatin 12.5 U/ml) to “Dulbecco’s Modified Eagle Medium-Low Glucose (DMEM-LG) and stored in refrigerator at 4°C
       
Except group A, diabetes was induced in other groups by administering sterile solution of 5% alloxan monohydrate in normal saline @ 60 mg/kg b.wt. intraperitoneally. Level of serum glucose was estimated at 7^th, 15^th, 20^th, 25^th and 30^th day post alloxan administration; value of 300 mg/dl or more was considered as diabetic (Wang et al., 2010).
       
For bone marrow collection, area of iliac crest on either side was prepared aseptically and 5 ml of bone marrow aspirate; 2.5 ml from either side was collected in 10 ml syringe containing 0.1 ml heparin (3000 U/ml) using 18 G Rosenthal pediatric needle. Bone marrow was collected from non-diabetic rabbits anaesthetized with Xylazine (@ 10 mg/kg b.wt. i.m.) followed ten min later by Ketamine (@ 50 mg/kg b.wt. i.m). Mesenchymal stem cells (MSCs) isolated from marrow aspirate were re-suspended in cell growth media, plated at an average of 1-2×10⁵ cells/cm² and maintained at 37°C under 5% CO₂ and 95% air in CO₂ incubator. On 4^th day, 3-5 ml of growth media was added and non-adherent cells removed. After reaching 70-80% confluency, cells were passaged at lower densities (5000 cells/cm²) into new culture flasks. Culture medium was discarded, cells washed with trypsin-EDTA for 5 min, 3 ml of culture medium was again added and cells centrifuged at 2000 rpm for 5 min. Supernatant was removed, precipitate re-suspended in 10 ml of DMEM and aspirated to obtain a single cell suspension. Cells obtained were replated onto culture flasks at half of their original density, maintained till average cell count of 3.45 × 10⁶/ 200μl was achieved, thereafter used for transplantation.
       
A 5 mm size segmental bone defect was created in right radii of rabbits and filled with hydroxyapatite granules (0.1-0.4 mm diameter) and materials as per experimental protocol. Post-operatively, injection Meloxicam @ 0.5 mg/kg and Enrofloxacin @ 5 mg/kg b. wt. was given for three and five days, respectively. Sutures were removed on 8th postoperative day.
       
Radius-ulna bones collected after sacrificing animals on day 30, 60 and 90 were subjected to histological examination. A 2.5 cm long piece of radius including site of defect and normal bone on both sides was cut, washed with normal saline, fixed in 10% formalin for 48-72 hours (Culling, 1963). Bone section was decalcified and completion of decalcification was assessed by flexibility, transparency and pin penetrability of bone sections. About 4µ thick longitudinal sections of decalcified bone was cut, stained with hematoxylin and eosin (H and E) as per standard procedure (Luna, 1992) and examined for cellular reaction, healing process and fate of grafted hydroxyapatite (Brar et al., 2002). Bone healing was assessed under a protocol with triple blinding according to modified histopathological scoring system (Lane and Sandhu, 1987). Score card for histological parameters in different groups at different time intervals are presented in Table 1. Woven bone formation score was not included in total score as it was formed more during early stages of healing and disappeared gradually.
 

 
Statistical analysis
 
Data was analyzed by one way analysis of variance (ANOVA) as per methods described by Snedecor and Cochran (1994).

Mean (±SD) scores for various histological parameters in different groups at different time intervals and pooled data of various histological parameters irrespective of time interval and groups are presented in Table 2 and 3, respectively.
 

 

 
Day 30
 
In all groups, HA granules were enwrapped within mild to moderate amount of fibrous connective tissue at defect site (Plate 1; Fig A-D). Non-significant difference in degradation of HA between groups was in line with findings of Niemeyer et al., (2010). Atilgan et al., (2007) reported very little resorption of grafted HA at defect site on day 30 in healthy rats, as was reported in our study. Differentiation of MSCs into osteoblast progenitor in canine caused increased osteogenesis and new bone formation (Jang et al., 2008), similar to finding reported in group D. Least osteogenesis in groups B and C might be due to adverse effect of diabetes. More bony union in diabetic animals treated with MSCs compared to demineralized bone matrix (DBM) (Breitbert et al., (2010) was in line with findings of group D animals. Early laying down and mineralization of healing bony tissue due to conversion of osteoblasts into osteocytes and release of Ca ion in matrix led to osteochondral bone union in group D. In other groups, fibrous union indicative of inferior bone quality was recorded. Bone marrow and cortical bone formation was not visible while cancellous bone formation was recorded in groups A, C and D. In group B, scores for cancellous bone formation, periosteal membrane integrity and proportion of woven bone tissue was non-significantly lower than other groups. Diniz et al., (2008) reported similar decrease in cancellous bone volume, cortical width and woven bone tissue in diabetic animals. Delayed fracture healing in diabetes might occur due to infiltration of inflammatory cells and severe inflammatory response (Wetzler et al., 2000), as reported in few animals of group B. (Plate 1; Fig B). On the other hand, absence of inflammatory cells in group D clearly indicated biocompatibility of transplanted allogenic MSCs with HA scaffold and local insulin administration with no immunological activity. Allogenic MSCs transplantation at fracture site result in new bone formation and complications associated with their use in animals have not been recorded (Arinzeh et al., 2003). More osteogenic cells appeared in early stage of healing in stem cell treated group compared to control group (Zhou et al., 2006). In present study, increased osteoblastic activity, more osteoblasts, new lamellar bone containing osteocytes and Howship lacunae with osteoclastic activity were recorded in animals of groups D and A (Plate 1; Fig A-D) compared to groups B and C (Plate 1; Fig A-D).
 

 
Day 60
 
Non-significantly higher scores for degradation of HA, osteogenesis and union of bone edge in group D was in line with findings of Niemeyer et al., (2010). In mesenchymal stem cells (MSCs) treated group, differentiation of MSCs into osteoblasts resulted in increased osteogenesis in rabbits at 8 week interval (Jiang et al., 2010), similar to finding reported in group D. Osteochondral bone union in groups A and D and fibrous to osteochondral in groups B and C indicated superior bony union in group D, which could be due to proposed role of MSCs (Breitbart et al., 2010). Bone marrow and cortical bone formation was still not evident. Cancellous bone formation score was significantly (P<0.05) lower in group B, while periosteal membrane integrity score was significantly (P<0.05) higher in group D. Osteoblastic activity in groups A and D was comparatively lower than groups B and C. Except group B, Haversian system or osteons appeared in all groups (Plate 2; Fig A-D). Similar to our findings, enlarged Haversian tubes and new bone formation in allogenic bone near proximal defect at 8 weeks after surgery was reported by Zhou et al., (2006).Total histological mean score was significantly (P<0.05) higher in groups A and D compared to groups B and C.
 

 
Day 90
 
Compared to day 60, score for degradation of HA increased non-significantly in groups B and C. Significant (P<0.05) increase in score for degree of osteogenesis in groups A, D and C (Table 3) suggested medium to good osteogenesis in groups A and D, medium osteogenesis in group C and weak to medium osteogenesis in group B. Mean score for union of bone gap, periosteal membrane integrity, cancellous bone and bone marrow formation was non-significantly lower in group B. Score for cortical bone formation differed non-significantly between groups while score for woven bone formation was significantly (P<0.05) higher in group B compared to other groups. Lower osteoblastic activity and more haversian system and volkman’s canals in groups D and A (Plate 2; Fig A-D) was in line with findings of Zhu et al., (2010), who reported formation of lamellar structure and Haversian system in the bone gap defects of healthy rabbits filled with HA at 90 day interval. Compared to day 60, total histological mean score increased significantly (P<0.05) in all groups. Total histological score on day 30, 60 and 90 were significantly (P<0.05) lower in group B compared to groups D and A.
  

       
Comparison among groups revealed significantly (P<0.05) higher scores for degradation of HA and cancellous bone formation in groups D and A compared to group B (Table 3). Scores for osteogenesis and union of bone gap were non-significantly higher in group D followed by group A and lowest in groups C and B. Significantly (P<0.05) higher periosteal membrane integrity score in groups D, C and A compared to group B, indicated increased absorption of HA, good osteogenesis and covering of defect by the periosteum in group D. Bone marrow and cortical bone formation scores differed non-significantly between groups. Score for woven bone formation was non-significantly lower in group D compared to other groups (Table 3). Total histological score was significantly (P<0.05) higher in group D and lowest in group B. Perusal of histological observation revealed best healing response in group D followed by healthy control group A and locally insulin treated group C and poorest healing in diabetic control group B. Diabetic state induced by alloxan, delays chondrogenesis and osteogenesis in initial stages of fracture healing (Diniz et al., 2008). In diabetic group, formation of bony callus was delayed but not inhibited and it required long time for remodeling and complete bone consolidation.
       
Significant (P<0.05) improvement in scores for osteogenesis, bone union, marrow formation, cancellous bone formation, periosteal membrane integrity and total score for histological healing were recorded up to day 90 (Plate 3; Fig B). On contrary, score for woven bone formation increased significantly (P<0.05) up to day 60, after which it decreased significantly (P<0.05) to a value lower than at day 30.
       
MSC-implant possess osteogenic and vascularization abilities (Yu et al., (2011); their transplantation at site of fracture not only provides an osteogenic cell source for new bone formation but also secretes growth factors to recruit native cells to migrate to the defect site (Frank et al., 2002). Graft rejection or immune reactions were not reported in any of animals transplanted with allogenic BMMSCs. This is because MSCs lack in MHC II on its cell surface which is responsible for antigenicity of cells (Jung et al., 2009). Better healing in diabetic group treated with BMMSCs might also be attributed to ability of MSCs to suppress and control inflammation caused by diabetes (Honczarenko et al., 2006).
       
Local insulin delivery directly to fracture site restores cell proliferation in diabetic callus and elicits an anabolic effect on various cellular events at fracture site (Irwin et al., 2006), there by enhancing fracture healing in diabetic animals, reflected by higher histological scores in groups C and D compared to group B. Similar improvement in the histomorphometric parameters of allograft incorporation into a rat femur defect after locally insulin application was reported by Dedania et al., (2011). However, local insulin therapy did not completely compensated adverse effects of diabetes on fracture healing as reflected by lower histological score of group C in comparison to groups A and D.
       
Poorest healing response in diabetic control group B, where no treatment was given might have resulted from diabetes induced uncompensated hyperglycemia. Hyperglycemia inhibits osteoblastic differentiation and alters the response of parathyroid hormone that regulates metabolism of phosphorus and calcium resulting in impaired bone healing in diabetic patients (Yu et al., 2011).

Local insulin therapy along with grafting of hydroxyapatite at bone defect site induces faster and better healing in alloxan induced diabetic rabbit model, however, local insulin therapy alone was not sufficient enough to overcome adverse effect of diabetes on fracture healing. Best healing response according to histological evaluation was induced by the combined therapy of allogenic BMMSCs implant along with local insulin group D followed by healthy control group A and locally insulin treated group C and poorest healing was observed in diabetic control group B.

The authors would like to thank the Dean, College of Veterinary and Animal Sciences, the Dean, Post Graduate Studies, Director Research and Director Experiment Station for providing all necessary facilities for the conduct of research work.