Expression of metalloproteinases endometrial stromal sarcoma: immunohistochemical study using image analysis.

AIM To investigate the expression of matrix metalloproteinases (MMP), a group of proteolytic enzymes with a central role in extracellular matrix invasion and degradation, in stromal sarcomas. METHODS 11 endometrial stromal sarcomas (four low grade tumours, seven high grade) were stained for MMP-2, MMP-3, and MMP-9 using immunohistochemical stains. The surgical material consisted of nine hysterectomy specimens and two pelvic recurrences. Three hysterectomy specimens, removed for leiomyomas, were studied as controls. Staining area was evaluated using image analysis. RESULTS Age at the time of diagnosis ranged from 21 to 67 years. Four of the 11 patients (three with high grade tumours and one with a low grade tumour) died of the disease, six remained free of disease, and one was lost to follow up. Staining for MMP-2, MMP-3, and MMP-9 was more diffuse in high grade tumours than in low grade tumours and controls. Staining for MMP-3 and MMP-9 was more pronounced in high grade than in low grade tumours (p = 0.04; p = 0.05). Staining for MMP-9 was significantly greater in all stromal sarcomas than in controls (p < 0.001 for high grade tumours v controls; p < 0.01 for low grade tumours v controls). Diffuse staining for MMP-2, exceeding 90% of the tumour area, was observed in three of seven high grade tumours but in no low grade tumours. There was no apparent correlation between staining for any of the three enzymes and survival. CONCLUSIONS Both low and high grade endometrial stromal tumours express matrix metalloproteinases. MMP-3 and MMP-9 are expressed more diffusely in high grade than in low grade tumours. In the individual case, diffuse staining for MMP-2 appears to best characterise the high grade tumours. Thus staining for MMP-2 may aid in differentiating high grade from low grade tumours, and MMP-9 in differentiating normal endometrial stroma from low and high grade endometrial stromal sarcomas. MMP expression does not appear to predict disease outcome in endometrial stromal sarcoma.