br factor it associates with and on
factor it associates with , and on the chromatin landscape, and these factors may be dependent on the tissue of origin, active onco-genes, and mutation spectrum across cancers.
SUV39H1-low tumours showed gene expression signatures of mi-gration and enhanced TGF-β signalling (Fig. 2d, f, Fig. S3a). We also note links between low SUV39H1 and CD66high populations (Figs. S1a, c, 3c); and low SUV39H1 in CD66 populations  may partly explain the migratory phenotypes observed in these cell populations. Further, sorted migrated populations of cervical cancer Amiloride HCL also reflect tran-scriptional signatures of migration and EMT, and these signatures show parallels with transcriptional signatures seen in SUV39H1-low tumours (Fig. 5a-d, f).
Low SUV39H1 abundance also correlates with low levels of its as-sociated histone mark, H3K9me3. Interestingly, one deviation from the
C. Rodrigues, et al.
EMT model observed in migrated populations is the broad depletion of H3K9me3, especially over gene promoter elements (Fig. 6a, b, f), in contrast to the more specific loss of H3K9me3 at mesenchymal gene promoters observed in EMT models . Along with nuclear enlarge-ment observed in migrated populations (Fig. 6b, c, d, right panels), this may suggest a model of enhanced epigenome plasticity in migratory populations - that a depletion of SUV39H1 and H3K9me3 may render genes more susceptible to remodelling and changes in gene expression, in turn allowing cells to switch to a migratory state more efficiently in the presence of migration-inducing cues (Fig. S10). SUV39H1 and H3K9me3 mediated heterochromatin have been demonstrated to play similar barriers to cell fate plasticity in iPSC reprogramming and T-cell lineage commitment [5–9]. A role in cell fate plasticity is further sup-ported by a depletion of SUV39H1 and H3K9me3 in the basal, stem layer in the normal cervix (Fig. 4b). Chromatin-linked cell fate plasti-city is increasingly emerging as a critical regulatory mechanism during cancer progression, including metastasis .
The consistent downregulation of SNAIL1 and SNAIL2 observed in SUV39H1 knocked down cells was non-intuitive, as our initial hy-pothesis predicted an upregulation of these transcription factors (Fig. S7a-c). However, there was an upregulation of a downstream effector of migration, MMP28 [19,20], consistent with the enhanced migratory phenotypes associated with SUV39H1. This data suggests a more complex regulatory mechanism, and it may be that SUV39H1 only ef-fects a smaller subset of the broad transcriptional differences observed in migrated populations (e.g. MMP28 upregulation). Broad H3K9me3 depletion may maintain gene promoters in a poised state, however it is likely that additional regulators are required to shape the overall transcriptional changes observed in migrated populations. However, this subset of changes effected by SUV39H1 is sufficient to yield a phenotype of enhanced migration and invasion with SUV39H1 knock-down (Fig. 1d, Fig. S1e). The data may also possibly suggest differences between migration and EMT, wherein migration associated-down-stream effector genes are altered, but possibly via non-EMT-transcrip-tion factor regulated mechanisms.
Collectively, our findings highlight an association and role for a SUV39H1-low chromatin state in promoting migratory cell populations during advanced cervical carcinoma progression. In subsequent studies, it will be important to gain further insights into mechanisms by which SUV39H1 regulates migratory populations, the regulators of SUV39H1 itself, and the roles of other epigenetic regulators and networks in mi-gratory and metastatic cell populations.
4. Materials and methods
SiHa cells (purchased from ATCC) were cultured in Dulbecco's modified Eagle's Medium (DMEM) with 10% FCS. Cultures were used for assays within 27 passages of acquisition from ATCC. All cultures were routinely tested for mycoplasma using Mycoalert kit (Lonza) and were consistently mycoplasma negative, and examined for retention of described morphology and growth characteristics. For transient over-expression experiments, cells were transfected with an empty vector, pcDNA3, or pLX305 SUV39H1 LV V5 (DNASU plasmid repository) for SUV39H1 overexpression. Lipofectamine 2000 (Invitrogen) was used for transfection, and cells were used for molecular and phenotypic as-says 48 h post transfection. Transient knockdown experiments were performed using esiRNA, as the heterogeneous mixture of these en-doribonuclease derived siRNAs provides greater target specificity than chemically synthesised siRNAs. For these experiments, mission esiRNAs (Sigma) targeting eGFP or SUV39H1 were transfected using lipofecta-mine RNAiMAX (Invitrogen). Cells were used for assays 48 h post transfection. For shRNA knockdown of SUV39H1, shRNAs to SUV39H1 (pSMP-Suv39H1_1) and Luciferase (pSMP-Luc as a control) were ret-rovirally transduced into SiHa cells. Cells were subjected to selection