Tumor heterogeneity is a major problem for developing improved cancer treatments. Although individual therapies may often treat portions of a tumor, differential response of subclones is an important reason for cancer recurrence. So far, precisely identifying subclones and their rates of evolution has been challenging. This is because of the lack of fine longitudinal data from patient tumors, as matched recurrences or metastases are often temporally distant from the original tumor. Patient-derived xenografts (PDXs), i.e. human tumors engrafted and further studied in mice, are a model in which tumors can be dissected and then propagated for controllable time intervals, making them a potentially powerful system for studying changes in tumor subclonal populations. In preliminary studies we have used high-depth sequencing to sensitively detect somatic mutations in PDX fragments. Moreover, we have shown that these mutations change in prevalence as a xenograft grows. In this exploratory study, we propose to test and apply PDXs as an improved system to quantify rates of tumor subclonal population evolution. We will pursue this in two specific aims.
In Aim 1, we will spatially dissect PDX triple negative breast cancer tumors derived from two separate patients, selectively sequence and propagate interlaced fragments from the tumors, and then sequence the propagated fragments after allowing them to evolve in vivo over several months. These experiments will provide rich, cross-validating data that we will analyze with new computational approaches to significantly improve not only identification of subclones within tumors, but also the rates at which subclones mutate and change in prevalence in bulk tumors. To confirm these approaches, we will perform single cell sequencing of hundreds of cells from these tumors.
In Aim 2, we will perform parallel studies on xenografts grown from the same two patient tumors as in Aim 1 but treated with standard-of-care drug therapy. This will allow us to compare how these subclonal populations evolve in conditions similar to a treated patient tumor. If successful, this study will yield a validated, generalizable approach for studying tumor evolution that could be applied to a wide variety of cancers.
Over one million Americans are diagnosed with cancer annually, and cancer recurrence is a main cause of death. Recurrence is due to the re-emergence of tumor cells that were not eradicated during initial treatment. In this proposal, we will explore why certain populations of tumor cells persist and become more prominent after treatment. We will evaluate these issues in mice implanted with cancer cells derived from human cancer patients.
