The human MHC class I -related chain gene A (MICA), a highly polymorphic MHC-encoded cell-surface glycoprotein, is a stress-induced and transformation related molecule absent from most normal tissues. Expression of MICA has been reported on a variety of mostly epithelial tumors. MICA can activate Natural Killer (NK) cells via interacting with the immunoreceptor NKG2D. Soluble MICA (sMICA), when shedded from tumor by the action of matrix-metalloproteinases (MMPs), can be detected at high levels in the sera of cancer patients. High levels of sMICA can down-regulate NKG2D expression and lead to functional impairment of NK cells thus providing for a mechanism of tumor escape. Gene array expression profiling of bone marrow biopsies revealed high levels of MMP2 and MMP9 in the bone marrow microenvironment, but not in the purified myeloma cells of the same patients. Since we are developing allogeneic NK cell therapy as a novel treatment for high-risk myeloma we decided to study whether the high activity of MMP2 and MMP9 resulted in 1) elevated sMICA levels in myeloma, and 2) affected NKG2D expression by NK cells. We found by ELISA assays that 28% (13/46) of patients with MM at diagnosis contained elevated serum levels of sMICA (median: 265.2 pg/ml; range: 125.9 – 806.5 pg/ml). sMICA levels were low in 16 control healthy donors tested (median: 0.8 pg/ml; range: 0 – 91 pg/ml). Next, we correlated sMICA levels with indicators of tumor burden and prognosticators of outcome in MM. We observed that sMICA levels in the serum of MM patients was not associated with Ras associated oncogene (RAN) expression, presence of abnormal cytogenetics, elevated CRP, elevated b2-microglobulin or elevated serum M-protein levels. There was also no correlation between MICA gene expression in purified MM cells and increased sMICA levels in the serum. We therefore examined several myeloma cell lines and found that high MICA gene expression does not always correlate with MICA expression at the cell surface as detected by flow cytometry. We hypothesize that the lack of correlation between MICA RNA expression and sMICA may be due to variation in translation of the MICA mRNA or failure to transport the MICA protein to the cell surface. Further, we tested the expression of NKG2D on CD3−/CD56+ cells of 5 MM patients with elevated sMICA serum levels and detected no down-regulation of NKG2D on NK cells compared to controls comprising MM patients with normal sMICA levels (n=3) or normal donors (n=5). Recent studies have suggested that tumor-membrane bound MICA may also play a role in the down-regulation of NKG2D on NK cells. We next incubated normal donor NK cells with a) the high membrane bound MICA expressing MM cell line U266 or b) serum from MM patients containing high MICA levels neither of which resulted in down-regulation of NKG2D. We conclude that despite the high expression of MMP2 and MMP9 both soluble or membrane bound MICA do not down-regulate NKG2D on NK cells and hence do not adversely affect NK cell function in MM. One explanation may be that the sMICA levels we found in myeloma patients were in the order of pg/ml, whilst the levels reported in epithelial tumors are 2–3 logs higher.

Author notes

Corresponding author