Poster Sessions

CANCER-37

Yujian Zhang

Y. Zhang, L. Xiang, J. Hansen, S. Kawa, M. Onda, M. Ho, R. Hassan, I. Pastan

Immunotoxin SS1P and Taxol Synergy Results from Decreased Shed Mesothelin Levels in Tumor Extra-cellular Space

Background: Antibodies and their derivatives are now used for the treatment of patients with solid tumors, but their behavior in tumor microenvironment is still poorly understood. We have developed a novel nylon-mesh-basket (NMB) method to isolate extra-cellular fluid (ECF) of tumors. This allowed us to analyze the concentration of its components. This method is simple and reliable.We applied it to the study of immunotoxin SS1P. Experiment design and results: SS1P targets mesothelin, a protein highly expressed on the cell surface of mesothelioma, ovarian and pancreatic cancers. SS1P is an immuotoxin composed of the Fv portion of an antibody reacting with mesothelin fused to a toxic protein. Having shown some anti-tumor activity in Phase I trials, SS1P is now being tested in a Phase II setting. Mice studies showed that when SS1P was combined with Taxol to treat KB tumors, striking synergy was observed. To investigate the basis of this synergy, we measured the concentration of shed mesothelin (sMSLN) in KB tumors by isolating tumor ECF and found very high levels of sMSLN. The level greatly exceeded that of SS1P in the tumor and neutralized tumor cell killing activity of SS1P. Taxol treatment reduced sMSLN levels from 23 nM to 8 nM accounting for the synergy. But in Taxol-resistant KB tumors, Taxol did not reduce sMSLN, and synergy was not observed. To identify the presence of SS1P/sMSLN complexes in the ECF of SS1P-treated tumors, we fractionated the ECF on a size exclusion column (TSK). Using a SS1P-specific ELISA, we observed a peak of SS1P in fractions corresponding to the complex position. To understand how reduced sMSLN level led to improved therapeutic effect, we developed FACS-based distribution analysis for SS1P in tumor. Alexa 488-labeled SS1P (20ug) was injected into mice bearing A431/H9 tumors. Taxol treatment increased the percentage of cells labeled with SS1P-Alexa from 35% to 60%. Conclusion: The high sMSLN level in tumor ECF is a significant barrier for SS1P therapy. Pre-treatment with Taxol improved therapeutic effect of SS1P by increasing SS1P distribution in tumor. The increase in distribution resulted from reduced sMSLN level by Taxol. This mechanism can also be applied to other antibody-based combination therapies. NMB method and FACS-based distribution analysis can be used to optimize the combination therapy regimen. They can also be used to study other physiological and pharmacological processes in tumor microenvironment.