OBJECTIVETo investigate the effects of recombinant human granulocyte macrophage colony-stimulating factor (rhGM-CSF) on wound healing and microRNA expression in diabetic rats.

METHODSEighteen male SD rats of clean grade were used to reproduce diabetes model. Four weeks later, a total of 64 full-thickness skin wounds were created on the back of 16 rats with established diabetes, with 4 wounds on each rat. Two symmetrical wounds on either side of the spine were created as a pair according to paired design. Then the wounds were divided into groups A and B according to the random number table and blind method (red and blue tags on the rhGM-CSF or the gel vehicle), with 32 wounds in each group. The ointment with red tag was applied on the wounds of group A and the blue one on group B. The application was conducted once a day, with a thickness of 3 mm, up to post injury day (PID) 14. Gross observation of wound healing was conducted on PID 3, 7, 14. The wound healing rate was determined on PID 3 and 7. On PID 3, 7, 14, tissues from 2, 4, and 8 wounds were harvested from each group respectively for the observation of the histopathological changes with HE staining, and also for analyzing the expression of proliferating cell nuclear antigen (PCNA) and CD31 with immunohistochemical staining (denoted as absorbance value). On PID 7, tissues from 6 wounds in each group were harvested for microarray gene chip to screen the differentially expressed microRNAs. Enrichment analysis of Kyoto encyclopedia of genes and genomes (KEGG) signaling pathway on the differentially expressed microRNAs were performed after the microRNA screening results were validated by real-time fluorescent quantitative RT-PCR. Data were processed with paired t test or two-sample t test.

RESULTS(1) On PID 3, the wound area was significantly decreased, and the wound granulation was significantly proliferated in both groups. On PID 7, the wound area was further decreased, and the wound area was almost filled by granulation in both groups; the conditions in group A were better. On PID 14, all the wounds in group A were almost healed, while a small area of raw wound with incrustation still remained in some wounds of group B. On PID 3 and 7, the wound healing rates of group A were (41 ± 5)% and (75 ± 4)%, significantly higher than those of group B [(31 ± 9)% and (71 ± 4)%, with t values respectively 10.13 and 8.06, P values below 0.001]. (2) On PID 3, the epidermal cells, endothelial cells, and Fbs in the wounds of 2 groups were sparse, with heavy infiltration of inflammatory cells. The above condition in the wounds was better in group A than in group B. On PID 7, the epidermal cells, endothelial cells, and Fbs were gradually well arranged in group A; infiltration of inflammatory cells decreased, and the condition was better than that of group B. On PID 14, the wounds of group A were completely covered by epidermis, while infiltration of inflammatory cells still remained in some wounds of group B. (3) On PID 3, 7, 14, the positive expressions of CD31 and PCNA in group A were respectively 0.275 ± 0.018, 0.345 ± 0.034, 0.305 ± 0.023; 0.406 ± 0.063, 0.223 ± 0.011, 0.045 ± 0.022. They were significantly higher than those of group B (0.222 ± 0.020, 0.229 ± 0.018, 0.197 ± 0.015; 0.324 ± 0.039, 0.162 ± 0.012, 0.018 ± 0.020, with t values from 2.281 to 9.652, P < 0.05 or P < 0.01). (4) According to the microRNAs detection and screening, as compared with group B, 18 microRNAs were up-regulated while 13 were down-regulated in the wounds of group A. (5) The results of real-time fluorescent quantitative RT-PCR had good consistency with the results of microRNAs detection. (6) Enrichment analysis of KEGG signaling pathway showed that among the 31 differentially expressed microRNAs, 4 took part in the MAPK signaling pathway, 3 took part in the Wnt signaling pathway, 1 took part in the TGF-β signaling pathway, 3 took part in the epidermal growth factor receptor signaling pathway, 2 took part in the cell cycle pathway, 5 took part in the axon guidance signaling pathway, 6 took part in the focal adhesion pathway, 3 took part in the regulation of actin cytoskeleton pathway, 1 took part in the extracellular cell matrix receptor pathway, 3 took part in the adherens junction pathway, and 1 took part in the cell adhesion molecules pathway. After disclosing the blind, it showed that the ointment with red tag was the rhGM-CSF gel and the blue one was gel vehicle.

CONCLUSIONSThe rhGM-CSF gel can promote wound healing in diabetic rats, producing significant differential microRNA expression in wounds, and they may be the target at gene post-transcriptional level of rhGM-CSF gel in promoting wound healing.