Previous work has shown that the inv isoform of Mena, an actin binding protein, is associated with invasion at the cellular level and metastasis in the context of the microenvironment in both animal models and humans (Robinson, B.D. et al. Clin Cancer Res 15, 2433-2441 (2009). However, the prognostic value for metastasis of MenaINV itself is unknown because there is no antibody that directly recognizes this isoform. Here we describe a method to assess a surrogate for MenaINV by measuring total Mena and subtracting the levels of the 11a (non-invasive) isoform. Method: Total Mena and Mena11a were measured in two independent retrospective breast cancer cohorts with 20 year follow-up using tissue microarray and quantitative immunofluorescence (AQUA) technology in a previously described multiplexed mode. AQUA scores for each marker were converted into z scores followed by subtraction of Mena 11a (noninvasive form of Mena) from total Mena (invasive and non-invasive) = Mena(inv) surrogate. This was calculated for each patient and correlated with clinical and pathological characteristics as well as disease-free survival in both cohorts.

Results: In the older Yale cohort, Kaplan Meier analysis dividing the Mena(inv) surrogate by quartiles suggested collapse of the top three quartiles followed by comparison to the fourth quartile (log rank p= 0.0003, n=501). The 4th quartile was also significant in node positive (log rank p=0.0047, n=267) and estrogen negative (ER) patient subgroups (log rank p=0.0003, n=234). Cox multivariate analysis showed Mena(inv) was independent of age, tumor size, nuclear grade, nodal status, ER, PR, Her2 (HR=0.636, 95% CI=0.47-0.86, p=0.0038, n=420). The newer Yale cohort showed similar results, but that cohort also had data on local vs. distant recurrence. The relative risk of distant recurrence in this cohort is 2.56 (p=0.011) for patients with high Mena (inv) compared to 1.96 (p=0.055) for any recurrence.

Conclusions: High Mena (inv) surrogate shows prognostic value for poor survival in two independent breast cancer cohorts with some suggestion of preferential prognostic value for distant recurrence.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P3-10-16.