Epigenetic regulation of integrin ITGA2 and the identification of epigenetic drivers of bone metastasis in breast cancer.

Breast cancer (BC) is globally the second most diagnosed cancer, and it is the commonest cancer of women. In Australia approximately 20,000 BC diagnoses and 3000 BC related deaths occurred in 2021 alone, in terms of cancer deaths second only to melanoma. BC has a strong propensity to metastasize to the bone, and while 5-year survival rates are >90% if BC is caught early as a primary tumour, survival rapidly worsens to less than 14% in those with skeletal metastases. Thus, novel means of counteracting this metastatic progression are greatly needed. Epigenetic changes, chemical modifications around the DNA including DNA methylation, histone modifications, and ncRNAs, help retain the plasticity of metastatic tumour cells and may play a greater role in 'driving' metastatic progression than acquired DNA mutations. Thus, if epigenetic changes drive metastatic progression, it follows that a subset of those epigenetic drivers or ‘epi-drivers’ may guide subsets of metastases such as skeletal metastases. Importantly, epigenetic modifications and mechanisms are potentially reversible, and are providing effective therapeutic targets in selected cancers, with more promising treatments on the horizon. Therefore, this study aimed to characterize and identify epi-drivers of BC bone metastasis, firstly by clarifying the epigenetic regulation of an integrin gene (ITGA2) known to have a role in BC bone metastasis, and secondly by identifying additional and novel epi-drivers in matched BC bone metastases and primary tumour samples.
Integrin ITGA2, which encodes a subunit of the α2β1 integrin collagen receptor, has known functions in bone metastasis, invasion, proliferation, and angiogenesis, and is subject to multiple layers of epigenetic regulation. In BC as well as prostate cancer (PC), which shares a strong propensity for bone metastasis, ITGA2 initially acts as a tumour suppressor gene (TSG) downregulated to promote initial spread and soft tissue metastasis. Secondary upregulation of ITGA2 likely induces bone metastasis, α2β1 engaging the abundant collagen in the bone matrix and guiding invasive pathways. Epigenetic changes likely contribute to ITGA2 regulation in BC bone metastatic progression. The ITGA2 promoter contains a large CpG island (CpGI), and herein it was revealed that in BC, methylation of this CpGI had limited ability to regulate the promoter and it was generally unmethylated in tumour samples and cell lines regardless of expression changes. This is in clear contrast to PC wherein methylation tightly controlled ITGA2 expression. Likewise, up-regulatory histone modifications were relatively uniform in BC cell lines, with classical repressive modifications absent. Instead, a selection of poorly understood histone acetylation modifications possibly regulated ITGA2 in BC along with transcription factors. Several novel transcription factor recognition motifs were identified in the ITGA2 promoter for NF-κB, FOXO3, FOXL2, EGR3, and β-catenin, all known ITGA2 up-regulators. Notably, NF-κB is upregulated by RANKL-RANK signalling, a bone remodelling pathway hijacked by BC cells to promote bone metastasis.
Herein noncoding RNAs were found to regulate ITGA2 in BC. A novel lncRNA, I2ALR, was found to be transcribed in antisense from a promoter nearby to that of ITGA2. Over-expression and knockdown studies revealed that I2ALR negatively regulated ITGA2, and modulated expression of ITGA2 pathways. I2ALR likely downregulated ITGA2 by forming a complex with the ITGA2 mRNA, supported by in-silico analyses. I2ALR was further found to be poly-adenylated via sequencing experiments, and in public datasets its expression correlated with improved survival in several cancer types, including BC. Contrastingly in PC, I2ALR apparently had little role in ITGA2 regulation and had no correlation to survival. Another noncoding RNA, micro-RNA-373-3p, negatively regulated by the oestrogen receptor (ER-α), was also shown herein to regulate ITGA2 expression. MiR-373-3p was proposed to be responsible, in part, for the initial downregulation of ITGA2 in BC progression and soft tissue metastasis, with the loss of ER-α in bone metastases possibly inducing miR-373-3p loss for ITGA2 upregulation. Both I2ALR and miR373 may offer novel therapeutic targets in BC. Overall, this suggests that distinct ITGA2 regulatory mechanisms operate in BC relative to PC. It was likely that epigenetic changes to genes beyond that of ITGA2 alone were responsible for mediating BC bone metastasis. Patient paired primary tumour and metastases have recently proved to be an effective way of identifying novel epigenetic and transcriptional changes that may drive metastatic progression. Therefore, the identification of other novel epi-drivers of BC bone metastases was conducted with EPIC methylation array and RNA-seq on matched BC primary tumour and bone metastasis samples. Bone metastases greatly resembled their matched primary tumours in methylation patterns, and those differentially methylated sites shared across all bone metastases were largely demethylated, indicating continuation in the progression of global hypomethylation and site specific hypermethylation that characterizes primary cancer tumours in bone metastasis. As would be expected from widespread hypomethylation, RNA-seq indicated that most genes differentially expressed in bone metastases relative to primary tumours were upregulated. Many potential epi-drivers were identified and select genes were validated by qPCR, bisulfite sequencing, and examination of publicly available data. HOXB1, a developmental gene with low expression in most tissues was found to have a hypermethylated CpG island in primary tumours, with the 5′ and 3′ flanks of this significantly demethylated in bone metastases, and expression significantly increased in bone metastases. In public datasets HOXB1 was found to be upregulated with increasing BC grade. WNT5A and FLI1 were also validated as putative epi-drivers, both displayed significantly decreased 5′ promoter methylation and increased expression in bone metastases relative to primary tumours. WNT5A and FLI1 have known roles in BC progression and metastasis and (along with HOXB1) are likely epi-drivers of BC bone metastasis. Another putative epi-driver associated genes was the bone metastasis associated RANKL (of RANK-RANKL signalling and ITGA2 regulation), which was de-methylated and up-regulated in BC bone metastases. Overall, these data suggest that a large range of both well characterized and novel BC associated pathways are likely epigenetically dysregulated to promote BC bone metastatic progression. These pathways and their constituent genes, including ITGA2, I2ALR, HOXB1, FLI1, WNT5A, and RANKL, may offer novel points of therapeutic intervention in BC bone metastasis, either through countering epigenetic changes or by targeting transcripts and proteins.