M. Okano1, T. Kawaguchi1, M. Oshi1, I. Okano1, E. Katsuta1, K. Takabe1 1Roswell Park Cancer Institute,Breast Surgery,Buffalo, NY, USA
Introduction: Despite massive expenditures in drug development to treat breast cancer, the number of compounds that was proved to be effective in animal models but failed in humans remains high. This is partly because currently used murine models do not mimic human patient cancer biology. Patient-Derived Xenograft (PDX) model, where tumor fragments from the patients are implanted into immunocompromised mice, has emerged as pre-clinical model to address these issues, however, even the optimal site of implantation remain controversial. We hypothesized that patient-derived breast tumor survive and grow better when implanted orthotopically in mammary fat pad (MFP) compared to dorsal subcutaneous space (SQ).
Methods: We xenografted 9 patient breast cancer tumors into NSG mice. 3 tumors (including 1 brain metastatic tumor) were ER(+)HER2(-) and 6 tumors were triple negative (TN). 2 of the TN samples were previously established PDX tumor provided from University of Utah. Using 2-3 mice per each patient sample, 1mm(3) tumor fragments were implanted surgically into both four sites in SQ and four sites in MFP, the bilateral second and forth fat pads. Tumors were resected at the day when the biggest tumor grew up to 1.5cm and they were passaged to another 2-4 mice as the next generation. The size and weight of the tumor were measured. Tumor “take” was defined as tumorigenesis of palpable tumor after implantation regardless of time it took.
Results: PDX tumors were established in 5 out of 9 patient tumors. Except brain metastasis tumor, the average time for tumor take was significantly longer in 1st generation compared to 2nd or 3rd generation (117.3 days vs 60.5 days, p<0.0001). The overall tumor take rate was 51.8% (156/301 implantation site). Take rate from TN tumors was 57.4% (156/272 site), on the other hand, that from ER positive tumors was 0% (0/39 site). Interestingly, tumor take rate was 89.6% (86/96 site) in brain metastatic tumor although it was ER positive tumor. Tumor take rate was significantly better in MFP implantation compared from SQ (61.5%, 99/161 vs 40.1%, 57/140, p<0.001). Tumor weight were significantly heavier in MFP compared to SQ (0.58g vs 0.097g, p<0.0001). With increase of passage, the weight difference between MFP tumor and SQ tumor significantly increased (weight MFP/SQ of 1st generation vs 2nd and 3rd generation were 3.2 vs 5.3, p<0.0001). Even in brain metastasis tumor, take rate was significantly better in MFP implantation than SQ (100%, 48/48 vs 79.2%, 38/48, p=0.0012), and tumor weight were significantly heavier in MFP compared to SQ (0.21g vs 0.50g, p<0.0001). Time to tumor taken was significantly shorter in metastatic tumor compared from the other tumors (51days vs 74days, p=0.018).
Conclusion: Triple negative breast cancers were more successful than ER positive tumors in establishing PDX. Brain metastasis tumor had the best take rate. MPF is better implantation site than SQ to develop breast PDX model.