Thrown for a Loop: Awakening BORIS to Evade ALK Inhibition Therapy

Abstract

Mechanisms of acquired resistance to ALK inhibition therapy in neuroblastoma have not yet been elucidated. In a recent issue of Nature, Debruyne et al. demonstrate that resistant MYCN-amplified ALK-mutated neuroblastoma cells overexpress BORIS, resulting in wide-ranging changes in chromatin interaction and transcriptional reprogramming.

The cure rate for patients with neuroblastoma lags significantly behind that of many other common childhood cancers. Amplification of the MYCN oncogene is the most common genomic alteration in neuroblastoma, being detected in about 40% of high-risk cases. These high-level amplification events are seldom subclonal or acquired if not present at the time of diagnosis. MYC family proteins remain undruggable due to lack of enzymatic activity or any deep pocket that can be traditionally targeted by small molecules. However, gain-of-function mutations in the Anaplastic Lymphoma Kinase (ALK) oncogene are the second most common genomic alteration, present in up to 15% of high-risk cases (Bresler et al., 2014). Contrary to MYCN, these mutations can be subclonal nal and/or acquired during standard intensive chemotherapy (Eleveld et al., 2015). The pivotal discovery of activating ALK translocations in non-small-cell lung cancer has driven the development of numerous small-molecule inhibitors of ALK, several of which are now FDA approved for first-line or subsequent therapy.

Over the last decade it has become clear that inhibition of full-length mutated ALK in neuroblastoma presents several therapeutic challenges. Susceptibility of ALK-driven neuroblastomas to crizotinib, a first-generation small-molecule ALK inhibitor, differs according to the underlying mutation, with mutations at two of the three hotspots causing primary resistance. Since crizotinib inhibits kinase activity by competing with cellular ATP for binding to the kinase active site, enhanced ATP-binding affinity reduces its effectiveness (Bresler et al., 2011). Chemotherapy has been shown to sensitize ALK mutant tumors to crizotinib, ultimately leading to a phase 3 trial for patients newly diagnosed with neuroblastoma (NCT03126916). Lorlatinib has been shown to be the only ALK inhibitor currently available that is effective against all common activating mutations found in neuroblastomas (Infarinato et al., 2016) and is currently in phase 1 testing in children with relapsed ALK-driven neuroblastoma (NCT03107988). While these early- and late-phase clinical trials are aimed at optimizing strategies to inhibit mutant ALK, resistance is inevitable, and several substantial challenges remain that will require adopting new paradigms for drug development in order to maximize clinical benefit.

Debruyne and colleagues make significant strides toward elucidating the non-mutational mechanisms of resistance to ALK inhibition therapy (Debruyne et al., 2019). The investigators utilize an in vitro system to model acquired resistance to the tool ALK inhibitor compound TAE684 in a neuroblastoma cell line model with an ALK mutation and MYCN amplification. The authors generate a drug-resistant model by growing the cell line in the continuous presence of increasing concentrations of TAE684 for 1 year. Through epigenomic profiling of the parental cells compared to cells harvested after 6 (intermediate resistance) or 12 (resistant) months, they map transcriptional reprogramming events associated with acquisition of resistance. Loss of MYCN expression occurs first, with concurrent accumulation of H3K27 trimethylation at the MYCN locus and subsequent diminution of genome-wide MYCN occupancy at canonical transcriptional targets. This was followed by an increase in expression of SOX2 and SOX9, followed by overexpression of BORIS, associated with hypomethylation of the BORIS promoter. Loss of BORIS via shRNA depletion restored ALK and MYCN expression, but cells were no longer viable, suggesting a dependence on BORIS in the resistant cell line. Cells were apparently not sequenced, so it is not clear if these changes in the epigenome were accompanied by new mutations in ALK or elsewhere in the genome.

BORIS is encoded by the CTCFL gene, a paralog of CTCF, and is a critical mediator of tertiary chromatin architecture in somatic cells. In this study, overexpression of BORIS resulted in an increase in BORIS-bound DNA, specifically at chromatin loop anchors in DNA regulatory regions enriched with super-enhancer elements. Genes regulated by these regions were consequently overexpressed and comprise a set of genes typically expressed in early neural development. The authors also show that BORIS is overexpressed in a number of resistant cancers, and replication of this work in a resistant Ewing sarcoma cell line confirmed similar BORIS binding profiles. In contrast to Ewing sarcomas, neuroblastomas do not have high basal levels of BORIS, suggesting different mechanisms by which BORIS mediates resistance.

Prior studies of BORIS suggest a role in epithelial-to-mesenchymal transition and cancer “stemness,” via Wnt and Notch downstream signaling pathways. The work done here establishes a complete transcriptional reprogramming in the resistant cancer cells the authors created that are dependent on BORIS regulation of complex chromatin interactions. This is consistent with prior studies demonstrating increased EMT in ALK inhibitor-resistant NSCLCs (Gainor et al., 2016, and implicated a potential important regulatory role of BORIS across a wide range of cancer types. These findings raise further questions about the etiology of acquired resistance to targeted therapy, which likely results from multiple parallel mechanisms. We anticipate that, due to underlying tumor heterogeneity, the existence of rare, stochastic, resistance-conferring genetic alterations within a tumor cell population will be selected for during treatment, and that there will be non-mutational mechanisms of drug resistance, perhaps though epigenetic mechanisms as has been seen in EGFR mutant non-small-cell lung cancer (Hata et al., 2016). Evolution of subclonal mutation, acquired-resistance mutants that abrogate drug efficiency, and rare cell expression of resistance markers that induce transcriptional reprogramming are all mechanisms that have been shown to contribute to cancer therapy resistance (Shaffer et al., 2017). Here, the authors demonstrate that the resistant state is highly defined by BORIS-bound chromatin architecture. It is crucial to build on these studies to further distinguish the evolution of resistance, which likely results from multiple parallel mechanisms, especially within the complex environmental niches that exist in vivo.

The data presented here highlight one potential mechanism of resistance to ALK inhibition therapy in neuroblastoma that will need to be validated in patient samples from ongoing clinical trials. These data are consistent with other recently published data showing the plasticity of neuroblastoma cells and transition of cellular state during tumorigenesis and under the selective pressure of therapy (Boeva et al., 2017; van Groningen et al., 2017). However, it is clear that there are multiple mechanisms by which neuroblastomas may evade potent ALK inhibition therapy. We posit that there are both mutational responses and a non-mutational adaptive response following ALK inhibition therapy. It is becoming increasingly clear that neuroblastomas are genetically heterogenous neoplasms, with substantiall subclonal and regional diversity. This heterogeneity extends to non-genetic mechanisms as well, as transcriptional and epigenomic variability has also been implicated in instigating and sustaining resistant states. The elegant work reported by Debruyne and colleagues highlights the urgent need to access specimens from patients on ongoing clinical trials of ALK inhibitors to verify this mechanism of resistance and define others. This is essential to define potential intervention strategies to prevent acquired resistance, such as the BET inhibition strategy proposed by the investigators, that also must be validated in additional preclinical studies prior to clinical testing. This reveals the complexity of pediatric cancer and how robust translational research in relevant preclinical models is essential to improve patient outcomes.