Evolutionary selection of pancreatic cancer cells leads to chemoresistance

Researchers at The University of Texas MD Anderson Cancer Center have shown how pretreatment clonal complexity in pancreatic tumor cells leads to chemotherapy resistance, findings that could lead to better-personalized and effective combination therapies.

Tumorigenesis is an evolutionary process. “A tumor is an ecosystem composed of different cell populations, and the fittest survive,” says Alessandro Carugo, Ph.D., research scientist at the Institute for Applied Cancer Science at MD Anderson Cancer Center and co-author of the study published recently in Cell Reports.

To track these different tumor cell populations, the research team used molecular barcoding, attaching an inheritable genomic label, or barcode, to a parent cell that marks all progeny cells.

The researchers analyzed clonal evolution and self-renewal of various populations of pancreatic cancer cells with unique barcodes in vitro and in human tumors that were transplanted into mice. After several passages (transferring the tumor cells to new culture medium or removing a tumor from one mouse and transplanting the tumor cells into another mouse), the researchers noticed something remarkable: the number of barcodes (representing clonal lineages) was drastically reduced in both environments, and the barcoded cells in the mouse models were almost all the same as the barcoded cells in vitro.

“The same population of cells maintains tumor growth in both environments,” explains Andrea Viale, M.D., co-author of the study and assistant professor in the Department of Genomic Medicine at MD Anderson.

The researchers used these shared barcoded cells to create identical tumors, called clonal replica tumors, to investigate the clonal complexity of tumors over time (i.e., the evolution of clonal lineages) across different mice. They treated mice harboring these identical tumors with three different mechanistically unrelated drugs (standard-of-care gemcitabine, MEK1 inhibitor AZD6244, and PI3K/mTOR inhibitor BEZ235).

As often happens in the clinic, gemcitabine-treated tumors at first shrunk, but when pancreatic  cancer  treatment stopped, the tumors grew. The researchers found that gemcitabine was able to eradicate a good portion of the clonal lineages; the clonal complexity of the tumor was reduced, but the chemoresistant lineages survived.

Resistant cells exist before cancer treatment

“This means that resistant cells pre-exist in the tumor. Chemotherapy kills certain clonalities, but resistant clonalities get selected. Before our study, chemoresistance was thought to be the emergence of genetic mutations that allows cancer cells to survive. The surviving cell populations in our study didn’t survive because of mutations but because they already were evolutionarily more fit than the cells killed by gemcitabine” explains Carugo.

The tumors treated with MEK1 inhibitor or PI3K/mTOR inhibitor also lost specific cell populations, but these tumors also relapsed, indicating that other cell populations were selected, changing the clonal complexity of the tumor.

By comparing chemoresistant and chemosensitive clonal lineages, the researchers identified a specific signature (upregulated and downregulated pathways) associated with gemcitabine-resistant clonal lineages.

“Once we refine this signature, clinicians could use this gene panel to stratify patients according to gemcitabine resistance and determine the best cancer treatment combinations,” said Viale. The current signature has about 200 genes, but the researchers hope to refine the signature to around 20 key genes, which will make it cheaper and easier to test patients’ tumor cells.

Tailored therapy to disrupt evolution

If all the clonal lineages in a patient’s tumor could be identified through such signatures, then a tailored combination therapy could attack each clonal lineage, preventing relapse. “New combinations of drugs can act synergistically to reduce the clonal complexity of tumors,” Viale says.

Carugo adds, “This research could lead to truly personalized medicine and unexpected therapy combinations. For example, a new drug may not seem to affect the size of a tumor but may actually eradicate certain clonal lineages and reduce the clonal complexity of the tumor, making it a valuable addition to a combination therapy.”

Theoretically, this research platform could be applied to any type of tumor. “Our platform links genomic and molecular events in a tumor to a real phenotype (the behavior of tumor cells), and we are collaborating with other researchers to determine if this approach will work for other types of cancer, including melanoma, leukemia, and lymphomas,” Viale says.