Cleveland Clinic Researchers Explore Genetic Landscape of Radiotherapeutic Resistance

Priyanka Gopal, MS

Department of Translational

Mohamed E. Abazeed, MD, PhD

Hematology Oncology Research Department of Radiation Oncology

Cleveland Clinic

Cleveland, Ohio

Radiotherapy is highly effective because it can reduce the size of gross tumors or even eradicate them and sterilize microscopic disease noninvasively, thereby preserving organs. Technical advances in this critical tertiary specialty over the last several decades have led to significant improvements in its spatial and dosimetric precision.

However, there has been little progress in the incorporation of biological predictors of response to radiotherapy. In fact, other than average responses for particular tumor lineages, allowing for population-level projections for rates of complete or partial responses or progression, individualized prediction currently is not possible.

In the era of precision medicine, in which molecular taxonomies are increasingly governing the type of drug therapies patients with cancer are receiving, radiation oncologists unfortunately continue to apply their sophisticated therapy bluntly. Standard radiation doses and fractionation schemes are delivered based on the site of anatomical origin of disease and do not take into account the genetic complexity across and within cancer types. An understanding of the interplay between the cancer genome and radiation response has the potential to significantly advance the field of biologically targeted radiotherapy.

Specifically, there is an urgent need to nominate biomarkers that are likely to predict the efficacy of radiotherapy and accelerate their clinical translation. Efforts thus far have been limited in large part because the genetic features regulating tumor cell survival and their frequency across and within individual cancer types had not been studied on a large scale.

In May, the National Cancer Institute (NCI) awarded a $2 million grant to Mohamed E. Abazeed, MD, PhD, to study these questions in his laboratory at Cleveland Clinic in Ohio. The lab is seeking to identify new genetic markers calibrated on the basis of radiation therapy effectiveness and new drug—radiation therapy strategies that more precisely and effectively target lung tumors most resistant to radiation. The goal is to advance a genetically guided radiation strategy that makes tumor cells more sensitive to radiation therapy.

Table. Gene Mutations that Correlate With Radiation Sensitivity¹

Findings from Early Studies

If the hypotheses are correct, the results will demonstrate that radiotherapeutic sensitizers can be selected based on the identity and type of genetic alterations identified in a patient’s cancer, prompting an evolution in the use of radiation from a generic one-size-fits-all approach to one guided by the genetic composition of individual tumors.In 2016, research conducted at Cleveland Clinic revealed new insights into the nature of the response of tumor cells to DNA damage by categorizing genetic determinants of survival after ionizing radiation in 533 human cancer cell lines across 26 cancer types.¹ The response to radiation can be represented as a Gaussian distribution—a high-entropy, multifactorial distribution—that strongly suggests heterogeneity within and across cancer lineages. Overall, 19 top genes were associated with radiosensitivity when mutated (Table). From our screen we identified many genes that, when altered, conferred resistance to radiation. These include NFE2L2, PIK3CA, SMARCA4, ERBB2, BRAF, and the androgen receptor (AR) gene.

One example of the application of this study was in breast cancer. Women with breast cancer undergo breast-conserving surgery or a mastectomy to remove detectable macroscopic disease. But tumor recurrence in the breast or chest wall causes a significant increase in mortality. Radiation is routinely delivered to the intact breast or the chest wall to prevent these frequently lethal recurrences.

Evolving Genetic Changes

We used our profiling effort to identify markers associated with resistance to radiation in breast cancer. We showed that signaling molecules such as ER, ERBB2, JAK/STAT3, and AR were highly associated with resistance. One of the most intriguing and highly associated pathways we identified was related to AR. We showed that AR signaling, in a manner analogous to androgen signaling in prostate cancer, is a major determinant of resisting the effects of radiotherapy. We studied the effectiveness of combining antiandrogen treatments with radiation. We treated orthotopic xenograft models of breast cancers that expressed AR with single agents and combinatorial treatment of radiation and antiandrogens; the experiments showed that tumor control was significantly improved (interactively) with combinatorial treatment compared with single-agent therapy.Recent large-scale sequencing efforts have compared the overall mutational landscapes across multiple cancers but largely disregarded their individual genetic clonal composition, which is likely a critical determinant of patient outcome and a major hurdle to improving care. Tumors are thought to possess distinct microenvironments that can contribute to genetic and phenotypic variability. In many cancer types, this intratumoral heterogeneity confounds our ability to design and select effective therapies.

Using New Models

In our NCI proposal, we posited that the origin of radiotherapeutic resistance and disease relapse is the emergence of treatment-resistant subclones, which represent a portion of the tumor that share distinctive somatic mutations or epigenetic signatures. These distinctive features can potentially influence their evolutionary fitness under selection. That is, treating tumors with chemotherapy and/or radiation therapy kills some tumor cells, leaving behind resistant subclones. Our work will focus on predicting the identity and evolutionary trajectory of these putative resistant subclones.Recent advancements in the design and development of models that more accurately reflect cancer’s genetic composition and its recurrence are poised to help gain more detailed insight into the mechanisms underlying tumor relapse. Our lab’s effort has been, in part, focused on better recapitulating the tumor in its original state using patient-derived xenografts (PDXs). These models are generated by injecting tumors isolated from patients and grown in immunocompromised mice and have been characterized as genetically and histologically comparable to the primary tissue from which they are derived.

If developed in a clinically workable timeframe, PDXs could also be useful in the prediction of tumor response to treatment in humans by testing drugs in mice. For example, our recent manuscript details a co-clinical study of a PDX derived from a patient with metastatic clear cell endometrial cancer. We showed that patient avatars can yield clinically actionable information.² We believe that these models could represent a critical step in directly guiding patients’ treatments and improving clinical trial success.

Finally, we intend to use experimental and computational models to delineate the subclone architecture of tumors before, during, and after therapy in patients and their PDXs. The advantage of this approach is the potential for intervening in the PDX to test our other therapeutic strategies. Together with biological and computational models, we can predict the evolution and velocity of subclonal architecture in the PDXs. We hope to identify and estimate the probability that a low-frequency subclone will survive radiotherapy and lead to resistance and then deploy strategies to depopulate tumors of these resistant subclones before radiotherapy is delivered, thereby ameliorating resistance. Our ongoing study will, in part, focus on studying the clonality of mutations in lung cancer and the role that subclonal events can have on therapeutic resistance and tumor repopulation after initial treatment responses.

References

1. Yard BD, Adams DJ, Chie EK, et al. A genetic basis for the variation in the vulnerability of cancer to DNA damage. Nat Commun. 2016;25:7:11428. doi: 10.1038/ncomms11428.
2. Vargas R, Gopal P, Kuzmishin GB, et al. Case study: patient-derived clear cell adenocarcinoma xenograft model longitudinally predicts treatment response. NPJ Precis Oncol. 2018;2:14. Published July 11, 2018. Accessed August 2, 2018. doi:10.1038/s41698-018-0060-3.