UCLA

ABSTRACT OF THE DISSERTATION

Repair of DNA double strand breaks and radiosensitivity:

modification of DNA repair and radiosensitivity by

microRNA-335 and mtPAP

by

Nathan Thomas Martin

Doctor of Philosophy in Biomedical Physics

University of California, Los Angeles, 2014

Professor Richard A. Gatti, Chair

Biologic responses to ionizing radiation are complex, and numerous cellular signaling cascades are activated with specific temporal kinetics upon exposure. Induction of DNA lesions, especially DNA double strand breaks (DSBs), are thought to be the main mechanism by which ionizing radiation kills cells and the rapid recognition and accurate repair of DSBs is a central determinant of cell survival after irradiation. DNA DSB repair is a complex and coordinated process involving numerous proteins (possibly >1000) that recognize double strand breaks, transmit the damage signals downstream, modify chromatin structure, and localize at break sites leading to the repair of breaks and the maintenance of genomic stability. Deficiencies in DNA DSB repair can cause cellular sensitivity to ionizing radiation or oncogenic transformation. Chapter 1 of this thesis provides an introduction to the effects of ionizing radiation on cells and the chorography of DNA repair mechanisms.

Individuals with genomic instability have been described who are deficient for specific proteins involved in double strand break repair and these patients often develop malignancies at an early age. Due to an inability to adequately repair their DNA in a timely manner, these patients respond severely to radiotherapy and other cytotoxic therapies used to treat cancer. These patients are described under an umbrella syndrome characterized by x-ray sensitivity, cancer predisposition, immunodeficiency, neurologic involvement, and DNA double strand break repair defects (XCIND). Our laboratory has developed radiosensitivity testing in an effort to diagnose these patients and, thereby, avoid severe, and often lethal, reactions to radiotherapy. Radiosensitivity is determined by measuring clonogenic survival in colony forming cell lines derived from lymphocytes isolated from whole blood samples, and a radiosensitive range has been defined by studying numerous radionormal cell lines and radiosensitive cell lines derived from patients with ataxia-telangiectasia, the archetypal radiosensitivity disorder. Chapter 2 summarizes the human radiosensitivity disorders discovered to date and the current state of the field for clinical radiosensitivity testing.

Current radiosensitivity testing takes approximately 90 days to complete because establishment of a colony forming cell line is required. This relatively long turn around is not optimal for patients needing timely radiotherapy intervention for malignancies. Additionally, approximately 5-10% of routine radiotherapy patients have severe reactions to radiotherapy and would benefit from radiation sensitivity screening prior to treatment. Chapter 3 explores two promising assays: 1) post-irradiation measurement of gamma-H2AX foci kinetics and 2) the neutral comet assay, as potential rapid surrogates for the clonogenic survival assay. I confirmed that the neutral comet assay (NCA) is the most promising surrogate assay for the clonogenic survival assay and have adapted the methodology to assay whole blood samples from patients submitted for radiosensitivity testing. This represents a test that could be completed in 2-3 days which is within the timeframe needed for use in the oncology clinic or for suspected XCIND patients requiring faster results.

The gamma-H2AX foci kinetics assay tested in Chapter 3 was not predictive of clonogenic survival, but did identify DNA repair kinetics in a radiosensitive cell line of unknown etiology that were similar to a radiosensitive cell line of known etiology, RNF168 deficiency. I postulated that the radiosensitive cell line of unknown etiology, RS73, had a defect in the same DNA repair pathway as the RNF168 deficient cell line because of their similar kinetics. This `candidate pathway' approach led me to profile 53BP1 and BRCA1 foci kinetics, and I ultimately found a defect in the formation or retention of BRCA1 foci in RS73. However, this approach fell short of identifying the underlying genetic defect responsible for the BRCA1 foci defect and radiosensitivity observed in this cell line due to incomplete knowledge of the BRCA1/53BP1 signaling pathway, at the time of study.

This follow-up study of RS73 was an example of how the `candidate pathway' approach can be hindered by limited current knowledge of specific DNA DSB recognition and repair signaling mechanisms. Further, the `candidate pathway' approach is biased towards identifying defects in known DNA repair genes because the assays used for discovery are designed to test only those DDR mechanisms that are already characterized. Thus, in Chapters 4 and 5, unbiased genome wide methodologies were utilized to profile a panel of radiosensitive cell lines of unknown etiology to identify novel molecules involved in DNA repair and the radiation response; a `candidate gene' approach.

The study presented in Chapter 4 utilized a microRNA microarrays to profile the radiation response of ~1200 microRNA to identify those associated with radiosensitivity. MicroRNA are small, non-coding RNA which can preferentially target specific mRNA through complementary sequences and regulate protein expression by inhibiting translation of these target mRNA. MiR-335 was selected for follow-up study because it was regulated in an ATM dependent manner after irradiation, which suggested it may play a role in the DDR. Interestingly, miR-335 was also overexpressed in two radiosensitive cell lines of unknown etiology, RS7 and RS73, which further suggested a link to radiosensitivity and the DDR. I demonstrated that miR-335 modulates the DDR by targeting CtIP protein levels, a protein involved in end resection and cell checkpoint signaling, leading to disrupted BRCA1 focus formation and radiosensitization. miR-335 was not an obvious candidate for modulating the radiation response and provides an example for using an unbiased, `candidate gene', approach to identify novel molecules and mechanisms in the DDR and in human radiosensitivity.

In Chapter 5, I describe another unbiased `candidate gene' approach which utilized exome sequencing to associate a mutation in the mitochondrial poly-A-polymerase, MTPAP, with radiosensitivity in two siblings from an Amish family. MTPAP was also not an obvious candidate for radiosensitization and was one of many variants identified in the patients studied. I demonstrated a causal link between MTPAP and radiosensitivity by rescuing the radiosensitivity and DNA DSB repair defects observed in the two siblings, RS63-3 and RS63-7, by transfecting WT MTPAP into the patients' cells. Further profiling of the radiation response in these cells indicated that a cellular state of oxidative stress, likely induced by mitochondrial dysfunction, resulted in increased DSBs/Gy induction and increased cell death. Reduced clonogenic survival and the DSB repair defect were abrogated by pre-treating the patients' cells with antioxidants, further indicating that oxidative stress played an important role in the radiosensitive cellular phenotype of these patients. The data presented in Chapter 5 suggested that RS63-3 and RS63-7 are an atypical presentation of XCIND and indicated that the current working model of radiosensitivity may need to be broadened to include genes outside of those directly related to recognizing and repairing DSBs (i.e. classical XCIND presentation).

Chapter 6 summarizes this thesis and presents my concluding thoughts in the context of DNA repair mechanisms, radiosensitivity testing for XCIND and routine oncology patients, and implications for the current working model for human radiosensitivity. I have shown that the neutral comet assay can be rapidly performed on whole blood samples and is an attractive option for radiosensitivity testing in the oncology clinic. The remaining hurdles for translating a functional radiosensitivity assay package to the clinic are discussed and approaches for addressing these hurdles are explored. Unbiased `candidate gene' approaches vs. `candidate pathway' approaches for identifying novel DDR modulating molecules are discussed in the context of the miR-335 overexpressing and MTPAP mutated cells, and molecular tools for following-up candidate gene studies are highlighted. The current working model for XCIND patients assumes that human radiosensitivity stems from defects in DSB recognition or repair genes. The data presented in Chapter 4 and 5 indicate that the current working model for radiosensitization likely needs to be broadened, or a separate category of radiosensitivity patients created, for developing a working model that better describes the spectrum of patients. Chapter 6 discusses the importance and implications of these data for our understanding of the DDR in humans and in the context of the immunologic and neurologic phenotypes observed in many XCIND patients.