The long and the short of NEAT1: Roles in paraspeckle formation and cancer

NEAT1 is a well-conserved, highly expressed nuclear RNA that does not code for proteins. It exists in two isoforms of 3.7 and 22.7 kb, of which the longer is the essential architectural component of a nuclear body in the cell, the paraspeckle. In this thesis, we show that the tumor suppressor p53 upregulates NEAT1 transcription leading to the formation of paraspeckles in cultured cells, in mouse skin, and in human tumors of epithelial origin. We used a mouse model of chemically-induced skin carcinogenesis combined with a genetic knockout of Neat1 to demonstrate that this lncRNA contributes to tumor formation. Next, we showed, using in vitro studies, that NEAT1 knockdown sensitizes cancer cells to common chemotherapeutic agents such as doxorubicin and the PARP-inhibitor ABT-888, and that knockdown without external stress already caused an accumulation of DNA damage on its own. In particular, knockdown of the long isoform NEAT1_2, the central building block of the paraspeckle nuclear body, is sufficient to do so, indicating that it may be the paraspeckle which is the functional unit in this process. Knockdown of NEAT1 prevented proper activation of the ATR-Chk1 pathway, decreasing the appropriate DNA repair. This decreased signaling could potentially lead to an increase in collapsed replication forks, causing an increase in S139-phosphorylated Histone 2A.X (ɣ-H2A.X) and an arrest of the cells in the next G1 phase. Finally, higher expression levels of the long isoform in a cohort of cisplatinum-treated ovarian cancer patients could predict poor survival, whereas considering both isoforms together did not have predictive value. The isoform specificity of these latter observations (i.e. knockdown of the long isoform alone was sufficient to induce a growth phenotype) prompted us to follow up on the individual roles of NEAT1_1 and NEAT1_2. We found that despite a common regulation by p53 at the NEAT1 promoter, both isoforms are differentially expressed upon different forms of stress, a process mediated by alternative processing of the NEAT1_1 short isoform at its 3' end. Nutlin-3a, which causes pharmacological activation of p53 and arrests the cells in the G1 phase, led to significant amounts of NEAT1_1 found outside of the paraspeckle nuclear bodies. In contrast, hydroxyurea stalls the cells mid-S phase in addition to activating p53. When treated with this agent, we failed to detect NEAT1_1 outside of nuclear bodies in the majority of the cells. In a non-treated, mixed population, the proportion of cells expressing NEAT1_1 outside of paraspeckles varies greatly. These observations suggested that NEAT1 isoform configurations in the cell could be cell cycle dependent. We subsequently found that NEAT1_1 is the predominant isoform in quiescent (G0) and G1 cells, whereas this isoform disappeared and NEAT1_2 and paraspeckles became primarily expressed from replication (S phase) onwards. Moreover, CRISPR-mediated knockout of the polyadenylation signal revealed that the short isoform does not contribute to the observed replication stress and cell cycle arrest upon knockdown of both. In addition, this isoform does not contribute to NEAT1_2 and paraspeckle function, at least when considering its role of protecting against replication stress. In fact, the data presented here suggest that a long-standing paradigm in paraspeckle research may need to be challenged: from our work it appears that NEAT1_1 does not contribute, at all, to paraspeckle function or formation, and that the hypothesis of NEAT1_1 recruitment into the paraspeckle may be wrong. In effect, in addition to our estimate that NEAT1_1 expression is almost zero in conditions where only NEAT1_2 is observed, we also found a mechanism of NEAT1_1 degradation by the RNA exosome. Here, further studies are needed to establish a cell cycle stage-specific role of the RNA exosome-mediated degradation process. Finally, repeating our short-term skin carcinogenesis experiment in mice lacking both Neat1 isoforms as before, but now also using a knockout mouse model of only Neat1_1 indicated that Neat1_2 competent cells (WT and knock out for Neat1_1) behave in a similar fashion, whereas mice that are Neat1_2 and paraspeckle incompetent show a decrease in epidermal hyperplasia and an increase in DNA damage as shown in the first part of this work. Similarly, phenotypes described by our lab and others affecting the mammary gland and other aspects of mouse fertility could not be recapitulated in mice lacking only Neat1_1, indicating that it is effectively the long isoform alone that confers a survival and growth advantage when cells need to withstand the internal and external stressors. In summary, the work presented in this thesis revealed that cancer cells can hijack the protective function of NEAT1_2/paraspeckles, but, that the more abundant and better conserved NEAT1_1 short isoform does not seem to contribute to this process, and that, although well studied, many aspects of NEAT1 and paraspeckle biology still remain to be understood.status: publishe