Simon N. Powell, MD, PhD (clinical science leader)
Jorge S. Reis-Filho, MD, PhD (basic science leader)
Homologous recombination deficiency is prevalent in breast cancer up to a level of ~25%. Large-scale alterations to the genome have been observed in these tumors, but if double-strand junctions are sequenced in addition, it is possible to categorize these tumors into upstream and downstream defects in the DNA repair pathway. We assert that there are fundamentally different patterns of genome instability for double-strand break repair. One is focused on the function of the BRCA1-BRCA2 pathway, where alterations in function are rather frequent in breast cancers. Although traditionally perceived as equivalent, there is evidence to demonstrate that downstream alterations that are BRCA1-like may have genomic and functional differences from those that are BRCA2-like. Conversely, the upstream defects are focused on sensing DNA damage, which is another way to suppress cancer formation. The DNA damage signaling is less well known but is observed in many types of cancer, including breast cancer.
Thus, the goals are to diagnose the DNA repair defect reliably and with greater resolution. Our hypothesis is that different types of DNA repair defects result in the use of distinct back-up DNA repair mechanisms, which themselves result in specific genomic signatures and sensitivity to different therapeutic agents. Hence, we posit that upstream defects are best targeted by using replication checkpoint inhibitors but that BRCA defective tumors are best treated by targeting the backup pathway, such as PARP inhibitors or new agents beyond PARP inhibitors.
The goal of the first aim is to apply the current genomic landscape tests of HR-deficiency and determine which method predicts most accurately the type of homologous recombination DNA repair defect (i.e. upstream, downstream BRCA1-like, or downstream BRCA2-like). The ultimate goal is to devise a taxonomy based on the genomics features of homologous recombination DNA repair-deficiency in addition to target gene mutations, which will ultimately guide therapeutic options.
The second aim is to generate genetically engineered cell lines to understand the developmental drivers of the genomic landscape changes. In addition, we will use these cells to test new synthetic lethal approaches to target specific subsets of breast cancers with distinct types of homologous recombination DNA repair defects.
The third aim consists of human clinical trials either being conducted at Memorial Sloan Kettering Cancer Center or elsewhere, where we are conducting the trial or leading the analysis of the clinical bio-specimens for correlative study analyses. We will test the use of ATR-inhibitors from a basket trial and retrospectively determine the genomic status of the tumors in responders and non-responders. Similarly, we will study the impact of the PARP inhibitor olaparib in patients who are BRCA1/2 wild-type but harbor a germline and/or somatic genetic alteration affecting homologous recombination DNA repair-related genes. We will extend our studies to also consider the combined effects of radiotherapy in combination with either ATR-inhibitors or PARP-inhibitors.
The ultimate goal of this project is to personalize the treatment of breast cancer patients whose tumors display homologous recombination DNA repair-related defects according to their genetic and genomic features. We seek to substantially improve the outcome of these poor prognosis patients and direct the deployment of therapeutic agents either already approved (e.g. olaparib) or already in clinical trials (e.g. ATR-inhibitors).
Samuel Bakhoum, MD (clinical science leader)
Lewis C. Cantley, PhD (basic science leader)
While considerable progress has been made in treating primary breast cancers, metastatic breast cancers remain a challenge. Metastatic breast cancer cells typically have chromosomal instability (CIN) that involves chromosome-level alterations leading to genomic copy number abnormalities. A major challenge in targeting breast cancers driven by CIN is the lack of known targetable alterations. We recently found that CIN promotes chronic inflammatory signaling in cancer cells. As chromosomes missegregate, they often become encapsulated in micronuclei. Subsequent micronuclear rupture exposes genomic double-stranded DNA to the cytosol. Cytosolic DNA activates anti-viral innate immune pathways, chief among which is cGAS-STING signaling. Under normal circumstances, cGAS-STING activation promotes type I interferon and facilitates cell-mediated immunity. Engagement of STING in normal epithelial cells induces senescence and cell death. We have shown that cancer cells, however, are intrinsically resistant to cGAS-STING activation by virtue of their chronic exposure to cytosolic DNA. Instead, they upregulate alternative pathways downstream of STING, such as NF-kB signaling.
The extent to which cancer cells depend on chronic inflammatory signaling is poorly understood. More importantly, how they subvert innate immune signaling to avoid immune surveillance remains unknown. Our ongoing work reveals that cGAS-STING signaling is sequestered in cancer cells away from the host. Furthermore, human breast tumors upregulate ENPP1, a negative regulator of cGAS-STING signaling. ENPP1 enables immune evasion by degrading cGAMP, the second messenger produced by cGAS, only in the extracellular space. As such, ENPP1 prevents host STING activation in response to tumor-to-host cGAMP transfer. Strikingly, pharmacologic inhibition of STING suppresses metastasis in syngeneic models of melanoma, breast, and colon cancers. We postulate this is because its inhibition in tumor cells outweighs its protective role in the host.
Building on this work, we will expand our pre-clinical testing of STING inhibition in breast cancer probing its efficacy in delaying metastasis and therapeutic resistance (Aim 1). We will then examine whether cGAMP contributes toward the formation of an immune suppressive microenvironment through metabolic breakdown in the extracellular space (Aim 2). Finally, we will develop cGAS-STING-based biomarkers in prospectively collected tumor specimens. We will test whether the status of cGAS-STING signaling and ENPP1 levels can predict response to neoadjuvant chemotherapy and atezolizumab, an immunotherapeutic recently approved for the treatment of metastatic breast cancer (Aim 3).
Our work addresses a clinically unmet need by targeting a subset of breast cancers with CIN and for which there are limited therapeutic options. If successful, it will provide pre-clinical rationale for first-in-human testing of STING inhibitors for the treatment of cancer metastasis as well as the development of novel CIN-related biomarkers to predict therapeutic response.
Sarat Chandarlapaty, MD (clinical science leader)
Reuben S. Harris, PhD (basic science leader)
Although the majority of early stage estrogen receptor (ER)-positive breast cancers are cured through multimodality care, metastatic ER-positive breast cancer remains a lethal disease. Insights into this discrepancy have come through comparative genomic analyses of primary and metastatic tumors. We and others have identified several mutations affecting specific genes that are more prevalent in metastatic cancers than in their primary counterparts, including ESR1, ERBB2, and NF1. These mutations result in resistance to frontline endocrine treatments that are the mainstay systemic therapy in ER-positive breast cancer. Even more striking than such individual mutations, however, has been the finding that certain ‘mutational signatures’ are enriched in metastatic disease as compared with primary breast cancers. These mutational signatures represent the DNA damage and repair processes that shape the cancer genome and can give rise to such mutations and the transformed phenotypes they convey.
A glaring and consistent finding from multiple large-scale sequencing studies has been that the APOBEC mutational signature is both enriched and highly prevalent in ER-positive metastatic disease, comprising the dominant mutational signature for these drug-resistant and ultimately lethal cancers. Our preliminary data confirm that APOBEC activation can promote the development of endocrine resistance in ER-positive cancer models and is associated with characteristic APOBEC-mutational changes in many drug resistance alleles. Together, these results point to the APOBEC mutational process as a key driver in the development and pathogenesis of ER-positive metastatic breast cancer and endocrine therapy resistance.
In this highly collaborative and innovative project, we propose three specific aims to advance the APOBEC mutational process as a biomarker and therapeutic target in breast cancer. (1) We will develop and utilize robust bioinformatic methods to detect the presence and the timing of onset of the APOBEC mutational signature from clinical NGS datasets of both tumor and cell free DNA (cfDNA). We will further ascertain if a promising IHC assay for the A3B enzyme can identify those ER-positive cancers likely to subsequently develop an APOBEC mutational signature. (2) We will determine the mechanisms and kinetics of APOBEC’s contribution to endocrine resistance. We will use isogenic cell line models and patient derived xenografts to dissect the types of resistance patterns that are caused by APOBEC as well their timing and whether the endocrine therapy itself contributes to the induction of APOBEC activity. (3) We will assess both immunologic and synthetic lethal approaches to targeting tumors in which APOBEC activity is induced and determine their capabilities in killing APOBEC-positive cancers.
We anticipate that our findings will uniquely position our team to launch clinical trials testing specific approaches to diagnose APOBEC-positive tumors, to prevent the development of resistance to endocrine therapies, and to target the largest subset of ER-positive endocrine-resistant metastatic breast cancers.