Recently there has been an important conversation about scientific reproducibility. Dr. Taran Gujral, a new member in the Human Biology Division, published a paper that indirectly brings much-needed nuance to this discussion by studying chemo-sensitivity in pancreas cancer. The article described here highlights Dr. Gujral’s work at Harvard Medical School that will serve as the foundation for his studies here at Fred Hutch. This particular study started from the observation that literature varied widely on the reported sensitivity of cancer cells to a first-line chemotherapy, gemcitabine. In fact, the same pancreas cancer cell line had been reported to be completely insensitive to gemcitabine in one study and highly sensitive in another. A careful examination of many pancreatic cancer cell lines, and many FDA approved chemotherapies revealed that cell-cell contact stimulates specific pathways that can alter drug sensitivities. This study, published in PNAS, shows that Hippo-YAP signaling stimulates both drug efflux and metabolic pathways that may render chemotherapies ineffective, suggesting that Hippo inhibition may be effective in combinatorial therapy.
When scientists question the reproducibility of experiments there are many levels to consider. The most egregious and sensational cases of failed repetition are outright fraud that has resulted in retractions, loss of government funding or medical licenses, and public outcry. While these findings are the most publicized, there are other factors that may contribute to poor reproducibility. The least exciting of these being unreported (and unexpected) subtle differences in experimental protocols. In this study Dr. Gujral found that many pancreas adenocarcinoma cell lines were sensitive to the common, first-line therapeutic, gemcitabine only when cultured very sparsely – meaning cancer cells were not in contact with one another. As cells became more crowed (as little as 50% confluent in a culture dish) they became less responsive to the drug.
This finding was unexpected and thus many control experiments were performed. If crowded cells were placed in a new dish with fewer contacts they immediately regained sensitivity. Also, once cells became crowded the addition of a dozen different growth factors did not alter their response to gemcitabine. Crowded cells had not inhibited cell death pathways, as UV exposure was equally lethal regardless of cell-cell contacts. Density-dependent drug response was also observed with some FDA approved chemotherapies but not all, for example cells became resistant to doxorubicin when crowded, but responded to the anti-mitotic taxanes similarly despite growth environment.
As the initial hypotheses all failed to answer why gemcitabine response changed when cells were contacting each other, Dr. Gujral used reverse phase protein array analysis to interrogate many of the important cell signaling pathways involved in cancer. As expected, as cells grew more densely and contacted each other, the MAPK and AKT pathways were down-regulated and ribosome activity also decreased. One unexpected result was that YAP phosphorylation was increased when cells were densely populated. YAP is a transcription factor that functions within the nucleus to activate cell proliferation programs; YAP is phosphorylated by members of the Hippo signaling pathway, causing its export to the cytoplasm where it no longer drives gene expression. This was observed in the pancreatic cancer cells. At low density there was high YAP signal in the nucleus, but as cells became crowded it was exported into the cytoplasm. As YAP is a transcription factor this suggested a YAP-dependent transcription profile was making cells sensitive to gemcitabine. In this case one player was drug efflux pumps, when a mutant YAP that cannot be exported from the nucleus was expressed the mRNA and protein levels of many drug transporters were stabilized.
To demonstrate the value of these findings in many cancers Dr. Gujral expressed a constitutively active YAP mutant in a pancreatic cancer cell line. Xenograft tumors were generated from the normal YAP and mutant YAP cancer cells and mice were treated with gemcitabine. The over-active YAP tumors showed a greater response to gemcitabine treatment. This was further shown in 20 other patient-derived xenograft models – in these experiments growth inhibition by gemcitabine correlated with protein levels of YAP.
As Dr. Gujral explained, they were not intending to comment on issues of scientific reproducibly, “We thought that some cellular pathways might exist that could tune the resistance to drugs like gemcitabine. The search for such a tuning led us to a previously unknown role of Hippo pathway in mediating sensitivity to several chemotherapeutic drugs including gemcitabine.” Nonetheless it raised the important point that unintentional, confounding factors may be a large source for reproducibility issues. Everyday features of research that go unreported, like cell density, may have unexpected consequences. This research is a spring-board for exciting studies Dr. Gujral is performing here at Fred Hutch, “We are carrying forward this research in several other cancers where Hippo pathway mutations are common such as non-small cell lung cancer and mesothelioma. In fact, our lab - together with Dr. McGarry Houghton – just received funding from the American Lung Association to study how mutations in the Hippo pathway could affect response to chemotherapy in non-small cell lung cancer”. We are looking forward to more exciting research and collaborations from the Gujral Lab.
Gujral TS, Kirschner MW. 2017. Hippo pathway mediates resistance to cytotoxic drugs. Proc. Natl. Acad. Sci. USA, 114(18), E3729-E3738.
The National Institutes of Health and the Human Frontier Science Program provided funding for this research.