Long Non-Coding RNA Induces De Novo Myelodysplastic Syndrome through Epigenetic Regulation

Long non-coding RNAs (lncRNAs) form the largest part of the mammalian non-coding transcriptome and control gene expression at various levels including chromatin modification, transcriptional and post-transcriptional processing. LncRNAs are implicated in initiation and progression of several cancers. Cancer-associated genomic regions are regions showing high frequency of cancer related abnormalities, such as loss of heterozygosity or amplifications. One such widely studied CAGR is the 8q24.21 genomic region. One SNP of particular importance present at this locus is rs6983267, with the G allele of the SNP conferring increased risk of colorectal, prostate, breast and bladder cancers. CCAT2 is a lncRNA that spans this highly conserved region. CCAT2 has been shown to play an important role in inducing chromosomal instability and supporting cell proliferation and cell cycle arrest. Despite advances in diagnosis of MDS patients, the underlying mechanisms that lead to spontaneous induction of MDS remains poorly understood. Here we attempted to elucidate the role of CCAT2 and its specific alleles (G/T) in regulation of cellular processes that drive spontaneous tumorigenesis using a genetically engineered mouse model. We generated transgenic mice for each CCAT2 allele using random integration approach in C57Bl6/N background, expressing CCAT2 in all tissues of mice.

In this study, we identified that CCAT2 plays an important role in regulation of normal hematopoiesis. Constitutive in vivo overexpression of each CCAT2 transcript in the mice resulted in spontaneous induction of widespread pancytopenias. CCAT2(G/T) BM biopsies displayed severe myeloid or erythroid hyperplasia, and dysplastic megakaryocytic proliferation, along with enhanced proliferation and excessive apoptosis. Interestingly, we identified two distinct phenotypes in CCAT2(G/T) mice with equal prevalence of MDS or mixed MDS/MPN. This suggests that CCAT2 overexpression might affect regulation of hematopoietic stem cells, disturbing their self-renewal or maturation capacity, and subsequently resulting in BM failure. Percentage of HSPCs was significantly reduced in BM of MDS mice, with increased presence of immature erythroid blasts and granulocyte-macrophage progenitors suggesting a block in differentiation. HSPCs of CCAT2(G/T) mice also showed increased frequency of cytogenetic aberrations, including breaks and chromosomal fusions. However, these mice don’t develop sAML, suggesting CCAT2 is critical in initiation of MDS. We further identified significantly higher CCAT2 expression in the MDS patients as compared to healthy volunteers. Patients with sAML had significantly lower expression of CCAT2 as compared to patients with only MDS.

To determine the mechanism by which CCAT2 induces genomic instability and myelodysplasia, we screened for several genes that have been previously reported to induce myelodysplasia as potential targets of CCAT2. Interestingly, EZH2 was downregulated in the BMCs of CCAT2^(G/T) mice compared to WT littermates. EZH2 downregulation was observed in both MDS only and MDS/MPD mice. In CCAT2(G/T) mice, EZH2 and H3K27Me3 reduction was observed in hematopoietic stem  and progenitor cells (HSPCs) as well as lineage positive bulk cells, suggesting that CCAT2 might induce alteration in EZH2 levels in the HSC compartment. Interestingly, we identified miR-26a and miR-26b, that were already reported to target EZH2, were significantly overexpressed only in BM of CCAT2-G mice. These data suggests that CCAT2-G regulates EZH2 expression primarily through regulation of target miRNAs. On the other hand, we identified EZH2 to interact preferentially to CCAT2-T compared to WT or CCAT2-G transcript. These data confirmed that EZH2 preferentially binds to the CCAT2 in an allele-specific manner.

In conclusion, deciphering the role of CCAT2 in spontaneously induced myelodysplasia and cytopenias will help us further characterize the poorly understood MDS/MPN phenotype. CCAT2 mice can serve as a robust model for studying initiation of de novo MDS/MPN that does not progress to secondary AML, and as a pre-clinical model for evaluation of new therapies for MDS. It has high translational potential as CCAT2 can be developed into a diagnostic and prognostic marker, as well as a novel intervention target for MDS therapy.