Experimental Study And Imaging Evaluation On Therapy For Avascular Necrosis Of Femoral Head By Mesenchymal Stem Cells Transduced With Ad-HGF

Objective1. Rabbit bone marrow mesenchymal stem cells (BMSCs) were transduced with hepatocyte growth factor (HGF) by adenovirus vector. Transduetion efficiency was detected by flow cytometry, transgene expression of HGF in vitro were evaluated by in situ hybridization and immunohistochemical methods.2. Rabbit model of corticosteroid-induced early stage avascular necrosis of femoral head (ANFH) was established and examined with CT perfusion imaging (CTP) and dynamic contrast—enhanced MR imaging (DCE-MRI) to explore the abnormal perfusion of ANFH demonstrated by imaging, the haemodynamic changes of femoral head and the relationship with pathological findings.3. BMSCs transduced with Ad-HGF gene were implanted into the necrotic femoral head through biopsy. The rabbits were examined with CTP and DCE-MRI to investigate the therapeutic effects of BMSCs transduced with Ad-HGF on the improvement of angiogenesis and bone reconstruction in the necrotic femoral head, and the role of CTP and DCE-MRI in the evaluation of gene therapy for ANFH. Materials and Methods1. Preparation of BMSCs transduced with Ad-HGF and in vitro experiment1.1 Preparation of Ad-HGF and Ad-GFP The expression vector of Ad-HGF and Ad-GFP (green fluorescence protein) were constructed by Vector Gene Technology Company Ltd.1.2 Isolation and culture expansion of rabbit BMSCs15ml bone marrow was extracted from rabbits (four weeks old) by bone marrow aspiration. Having centrifugalized and removed the supernatant and fat cells, the cells were plated into DMEM medium supplemented with 10% fetal bovine serum. Non-adherent and loosely attached cells were removed during the medium changes. Original BMSCs were obtained after 2 weeks cultivation. Then BMSCs were digested with pancreatic enzyme and cultivated to the generation-2.1.3 Investigation of transduction efficiency and expression time of BMSCs transduction with Ad-GFPAfter digested with pancreatic enzyme, BMSCs were plated and cultivated in 6 holes plate with a density of 1.5×105/per hole. 50μl Ad-GFP with 11 different MOI (multiple of infection) was added into each hole for BMSCs transduction, with 3 duplicated holes for each MOI. After digested with pancreatic enzyme and washed 3 times with PBS, GFP-positive cells were detected by flow cytometry.After BMSCs transduction with Ad-GFP at MOI=300, GFP-positive cells were detected by flow cytometry at different times.1.4 Investigation of transgene expression of HGF in BMSCs with in situ hybridization and immunohistochemical methodsBMSCs were transduced with Ad-GFP at MOI=300. After digested with pancreatic enzyme, BMSCs were added to silicificated slice, and cultivated. Then BMSCs were detected with in situ hybridization and immunohistochemical methods.2. Establishing rabbit models of steroid-induced ANFH and imaging examinations2.1 Preparation of rabbit models of steroid-induced ANFHTwenty-six adult and healthy rabbits weighing (2.53±0.11) kg were randomly divided into two groups, including 6 normal controls and 20 models. In model groups the rabbits were administered intravenously 10ml/kg of sterile horse serum and repetitively two times at two-week interval. Two weeks after the second serum injection, 7.5mg/kg of Prednisolone was injected intraperitoneally two times per week for four times. Then 20 model rabbits were randomly divided into three groups. 3w group consisted of 4 rabbits were examined with CT and MRI at the end of 3 weeks since steroid injection. Similarly 4w group consisted of 8 rabbits were examined at the end of 4 weeks; 5w group consisted of 8 rabbits at the end of 5 weeks since steroid injection. Normal rabbits and 4 model rabbits were killed for ink perfusion and pathological study at the end of 5 weeks. 2.2 CT examination and data analysisCT scans were obtained on a multislice CT scanner (Lightspeed 16, General Electric Medical System). Plain axial scans were performed on both hip joints. Four target slices altogether 1cm (2.5mm×4) thickness were selected including both femoral heads. Then dynamic enhanced perfusion scans were performed by axial view model with 1-sec scanning time and 1-see interval, totally 90s with 45 scans. 4ml contrast agent was injected at 1.5ml/sec.All the perfusion data was post-processed using the commercially available perfusion 3 software package on AW4.2 workstation. The para-femoral artery was used as artery input. Regions of interest (ROI) were placed in artery, femoral head, proximal femur and muscles. Thus the CT values of each ROI at each scanning times were automatically generated, also with the time-density curve (TDC) and the color maps that representing each perfusion parameters including blood flow (BF), blood volume (BV), mean transit time (MTT) and permeability surface area product (PS). The peak enhancement amplitude (EnA), peak percentage of enhancement (En%) were calculated.2.3 MR imaging and data analysisMR images were obtained with a 1.5T MR scanner (Magnetom Vision Plus, SIEMENS) and a head coil. Coronal T₂-weighted images (TR/TE 4500ms/96ms) with and without fat suppression, and T₁ weighted images (TR/TE 550ms/20ms) were obtained with 3mm slice thickness and 0.3mm gap. Synchronous withâ…£administration of 2ml Gd-DTPA at a rate of 1ml/sec, DCE-MRI was performed with fast gradient echo sequence (TR/TE 80ms/15～60ms, 30°flip angle). Acquisition time for each scan was 11s, totally 20 scans (220s).Bilateral ROIs were chosen in the femoral head, proximal femur and muscles. The signal intensity (SI) of each ROI at each scanning times was automatically generated, as well as the signal intensity-time curve (SI-T) of ROIs. Percentage of enhancement (En%) was calculated and the time-mean En% curve were drawn. Maximum percentage enhancement (En%max) and maximum enhancement rate (EnR) of time-mean En% curve were calculated.3. Experimental study and imaging evaluation on therapy for ANFH by BMSCs transduced with Ad-HGF3.1 Animal groupsAmong 14 rabbits of ANFH (4 were killed for pathological examination, 2 were dead during experiment), 3 rabbits were excluded because of big deviation comparing with mean value of CTP and DCE-MRI. The remaining 11 rabbits were randomly divided into 3 groups: 3 rabbits (6 femurs) as control group; 8 right femurs of 8 rabbits as Ad-HGF-BMSCs gene treatment group; 8 left femurs of 8 rabbits as puncture treatment group.3.2 Methods of treatment and Timing of observationGene treatment group: Having punctured into the femoral head under CT guidance, (1-2×10⁶) BMSCs transduced with Ad-HGF was implanted.Puncture treatment group: punctured only without BMSCs implantation.Control group: natural repa.ir without any treatment.2 and 4 weeks after treatment, animals were examined with CT and MRI. The rabbits were killed at the end of 4w for pathological examination and ink perfusion.4. Ink perfusion and pathological examinationBeing anesthetized, abdominal aorta was exposed and punctured, with saline pressurizely lavage. Then a liquid mixed with ink and dextran by 7:3 proportions was injected until the double lower extremity became black (about 100ml). Both femoral heads were dissected. Then, fixed, decalcified and cut into sections, the structures and the blood vessels of femoral head were observed under the stereo microscope.Rabbits were killed by air embolism. Both femoral heads with epiphysis were dissected. Then fixed, decalcified, paraffin embedded, cut into serial sections, and stained with HE.5. Statistical analysisStatistical analysis were made using SPSS for windows 13.0.All data were recorded as (mean±standard deviation). Percentage of GFP-positive cells in vitro experiment, CT perfusion parameters, En%max and EnR of DCE-MRI among normal and model groups were compared with one-way ANOVA. The statistically significant difference was set at P＜0.05.CT perfusion parameters, En%max and EnR of DCE-MRI among different treatment groups at different timings were compared with general linear model repeated measures. The statistically significant difference was set at P＜0.05.Results1. Preparation of BMSCs transduced with Ad-HGF gene and in vitro experiment1.1 Analysis of transduction efficiency and expression time of BMSCs transduction with Ad-GFP48h after BMSCs transduction with Ad-GFP at different MOI, percentage of GFP-positive cells showed significant difference (P＜0.05). Percentage of GFP-positive cells increased as MOI increased, however there was no significant difference (P＞0.05) when MOI was bigger than 200. When MOI=300, GFP-positive cells were over 98%, so MO I=300 was optimal for BMSCs transduction.After BMSCs transducd with Ad-GFP at MOI=300, the percentage of GFP-positive cells showed significant difference at different timings (P＜0.05). There were GFP-positive cells just 24h after transduction, with peak percentage from 2d to 7d, and then decreased dramatically. There were still GFP-positive cells until 28d.1.2 Transgene expression of HGF in BMSCs by in situ hybridization and immunohistochemical methodsWhen BMSCs were transduced with Ad-GFP at MOI=300, the results of in situ hybridization and immunohistochemistry showed that there were positive hybrids with yellowish brown particles in nuclei and cytoplasms of transduced BMSCs. This proved that Ad-HGF might effectively transduce BMSCs, and that the transduced BMSCs revealed high efficiency of HGF mRNA and gene expression.2. Establishing rabbit model of steroid-induced ANFH and imaging examinations2.1 CT2.1.1 Conventional CTThe changes in femoral head of model rabbits on conventional CT were mild, only demonstrated as blurred margins, small low density dots in subcapital bone marrow and higher density in metaphysis.2.1.2 CTP TDC appearanceThe TDC shape of femoral head in 3w model rabbits was similar to that of the normal group. The ascending segment of 4w model rabbits was slower and mild, with longer peak segment in comparison with that of normal group. The TDC shape of femoral head at 5w model rabbits demonstrated a flat type.2.1.3 Comparison of CTP parametersStatistically significant differences (P＜0.05) existed in mean values of BF, BV, PS among normal and model femoral head. The mean values of BF and BV in normal femoral head were 70.53±9.31ml/(min/100g), 10.28±2.05ml/100g respectively. 4 weeks after steroid injection, mean values of BF and BV began to decrease, showing as 30.65±4.64ml/(min100g), 5.83±1.60ml/100g respectively. 5 weeks after injection, mean values of BF and BV decreased obviously, showing as 16.40±6.17ml/(min100g), 3.37±1.57ml/100g respectively. The EnA and En%showed similar changes as BF and BV, showing as normal value of 39.76±8.92Hu, 4.96±1.25% respectively, mild decreased showing as 20.99±6.76Hu and 2.78±1.28% at 4 weeks, dramatically decreased to 9.51±3.24Hu, 1.19±0.38% at 5 weeks.2.2 MRI 2.2.1 Conventional MRIAvascular femoral head demonstrated joint effusion, dotted or linear high SI on T₂WI, high SI of metaphysis on T₂WI, especially on FS- T₂WI.2.2.2 DCE-MRI SI-T curve appearanceThe SI-T curve of femoral head in 3w model rabbits was similar to that of the normal group (approximately 200). The ascending segment in 4w model rabbits began to decrease (approximately 100). The ascending segment in 5w model rabbits decreased remarkably (approximately 50).2.2.3 The maximum percentage of enhancement and enhancement rateStatistically significant differences (P＜0.05) existed in mean values of En%max and EnR among normal and avascular femoral head. The mean values of En%max and EnR in normal femoral head were 70.58±13.62%, 55.00±10.62 %/min respectively. 4 weeks after steroid injection, mean values of BF and BV in femoral head began to decrease, appearing as 50.22±11.21%, 27.39±6.11%/min respectively. 5 weeks after injection, mean values of En%max and EnR decreased obviously, appearing as 17.03±9.02%, 13.27±7.03 %/min respectively.2.3 Ink perfusion and pathological studyHE stain of femoral head in model group demonstrated trabeculae necrosis with more empty lacunae of osteocytes, bone marrow necrosis with considerable adipocytes and decrease of hematopoietic cells. Ink perfusion of femoral head in model group demonstrated dramatically decreased vascularity, with incompletely perfusion.3. Experimental study and imaging evaluation on therapy forANFH by BMSCs transduced with Ad-HGF3.1CTIn gene treatment group, the TDC shape of femoral head at 2w was similar to that of normal group, with slightly mild ascending segment, approximately 25-30Hu. The ascending segment was decreased, approximately 20Hu at 4w. In puncture treatment group, the TDC shape of femoral head at 2w was similar to that of gene treatment group, with a lower ascending segment. The ascending segment was decreased showing flat shape at 4w. The TDC showed flat shape in control group.Significant differences (P＜0.05) existed in mean values of BF and EnA of femoral head among different treatment groups. The mean values of BF and EnA of femoral head in gene treatment group were the highest, with puncture treatment the next, control group the lowest. Significant differences (P＜0.05) existed in mean values of BF, BV and EnA between two treatment timings in gene treatment, with higher value at 2w. At 4w, there was no significant differences (P＞0.05) in BV and En% between puncture and control group, however, significant differences (P＜0.05) exited in BF, BV, EnA and En% between gene treatment and control group.Significant differences (P＜0.05) existed in PS of femoral head between gene treatment and control group.3.2 MRIAt 2w after treatment, SI-T shape of femoral head both in gene treatment group and puncture treatment group showed increased ascending segment, with the former higher than the latter. At 2w after treatment, the SI-T shape of femoral head in gene treatment group still demonstrated increased ascending segment, yet flat in that of the puncture treatment group.Significant differences (P＜0.05) existed in mean values of En%max and EnR of femoral head among different treatment groups. The mean values of En%max and EnR of femoral head in gene treatment group was the highest, with puncture treatment the next, and control group the lowest. Significant differences (P＜0.05) existed in mean values of En%max and EnR of femoral head between two treatment timings in gene treatment, with higher value at 2w. At 4w after treatment, there was not significant differences (P＞0.05) in En%max between puncture and control group, however, significant differences (P＜0.05) existed both in En%max and EnR between gene treatment and control group.3.3 Ink perfusion and pathological studyGene treatment group: The pathological examinations showed increased vascularity in bone marrow of trabeculae. More osteoblastic cells with new bone formation and reconstruction were found. Increased small vascular net in femoral head was observed in ink perfusion.Puncture treatment group: The pathological findings showed that mildly increased vascularity in bone marrow of trabeculae with fibrous tissue proliferation around the dead trabeculae was seen. Increased small vascular net in femoral head was observed in ink perfusion.Control group: Osteonecrosis was more severe than model group accompanying with partly vascular embolism. Dramatically decreased vascularity was observed in ink perfusion.Conclusions1. GFP adenovirus receptors successfully transduced BMSCs. As MOI=300, GFP-positive cells were over 98%. There were GFP-positive cells just 24h after transduction, with peak percentage from 2d to7d. There were still GFP-positive cells until 28d. These results indicated that adenovirus vector possessed high efficiency transduction and stable gene expression.2. Bone marrow mesenchymal stem cells are ideal cellular vehicles for their easy access, isolation, culture,, as well as in vitro expansion potential.3. Results of in situ hybridization and immunohistochemitry demonstrated that Ad-HGF was capable of effectively transducing BMSCs, and that transduced BMSCs possessed high efficiency of HGF mRNA and HGF gene expression in vitro.4. Horse serum plus corticosteroid injection could induce early stage of ANFH rabbit models. The pathologic changes were trabeculae necrosis with more empty lacunae of osteocytes, and bone marrow necrosis with the aggregation of enlarged adipocytes and reduction of hematopoietic cells. These findings indicated close correlation with that of imaging performance.5. Both CT perfusion and dynamic enhanced MRI could better reflect the haemodynamic changes of femoral head, and could be used as a quantitative method for the detection of blood perfusion in ANFH. Blood perfusion was decreased in avascular femoral head, with decreased BF, BV, and PS values, lower TDC peak enhancement on CTP; as well as decreased time-En% curve, decreased En%max and EnR on DCE-MRI.6. Both CT perfusion and dynamic enhanced MRI could better reflect the haemodynamic changes of femoral head, and could be used as a quantitative method to evaluate the therapeutic changes of ANFH treated with gene therapy. Ad-HGF transduction of BMSCs could improve the blood circulation in avascular femoral head, with increased BF, BV and PS values, higher TDC peak enhancement on CTP; as well as increased time-En% curve, increased En%max and EnR on DCE-MRI.7. The pathological findings and ink perfusion results showed that implantation of BMSCs transduction with Ad-HGF into avascular femoral head could promote angiogenesis, and accelerate necrotic bone repair. These results indicated that Ad-HGF transduction of BMSCs showed a good prospect in gene treatment of ANFH.