(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



A single N6-methyladenosine site regulates lncRNA HOTAIR function in breast cancer cells [1]

['Allison M. Porman', 'University Of Colorado Anschutz Medical Campus', 'Biochemistry', 'Molecular Genetics Department', 'Aurora', 'Colorado', 'United States Of America', 'Justin T. Roberts', 'Molecular Biology Graduate Program', 'Emily D. Duncan']

Date: 2022-12 

N6-methyladenosine (m6A) modification of RNA regulates normal and cancer biology, but knowledge of its function on long noncoding RNAs (lncRNAs) remains limited. Here, we reveal that m6A regulates the breast cancer-associated human lncRNA HOTAIR. Mapping m6A in breast cancer cell lines, we identify multiple m6A sites on HOTAIR, with 1 single consistently methylated site (A783) that is critical for HOTAIR-driven proliferation and invasion of triple-negative breast cancer (TNBC) cells. Methylated A783 interacts with the m6A “reader” YTHDC1, promoting chromatin association of HOTAIR, proliferation and invasion of TNBC cells, and gene repression. A783U mutant HOTAIR induces a unique antitumor gene expression profile and displays loss-of-function and antimorph behaviors by impairing and, in some cases, causing opposite gene expression changes induced by wild-type (WT) HOTAIR. Our work demonstrates how modification of 1 base in an lncRNA can elicit a distinct gene regulation mechanism and drive cancer-associated phenotypes.

Funding: This work was supported by the National Institute of General Medical Sciences ( https://www.nigms.nih.gov ) (R35GM119575 and R35GM144358 to A.M.J.; T32GM008730 to J.T.R.); the National Cancer Institute ( https://www.cancer.gov ) (R01CA187733 to J.K.R.; T32CA190216 to A.M.P.; F31CA247343 to J.T.R.; F32CA239436 to M.M.W.; P30CA046934 to the University of Colorado); the National Institute of Dental and Craniofacial Research ( https://www.nidcr.nih.gov ) (K99DE030528 to A.M.P.); the U.S. Department of Defense Congressionally Directed Medical Research Program Breast Cancer Research Program ( https://cdmrp.army.mil/bcrp/ ) (BC170270 to A.M.P.); and the University of Colorado School of Medicine RNA Bioscience Initiative ( https://medschool.cuanschutz.edu/rbi ) (seed grant to A.M.J.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Here, we investigate the function of m6A in HOTAIR-mediated breast cancer phenotypes. We identify 8 prominent m6A sites in HOTAIR and show that a single site (A783) is required for HOTAIR-mediated TNBC cell growth and invasion. We find that YTHDC1, the nuclear m6A reader, interacts with HOTAIR at methylated A783 and this interaction promotes chromatin association and gene repression upstream of PRC2. These results help explain why high HOTAIR is significantly associated with shorter overall patient survival, particularly in the context of high YTHDC1. Mutation of adenosine 783 in HOTAIR to uracil prevents nearly all of the normal gene expression changes that are induced by the WT lncRNA. Surprisingly, at a subset of genes, the A783U mutant induces opposite gene expression changes to WT HOTAIR, reducing molecular and cellular cancer phenotypes in TNBC cells, indicating that the mutant HOTAIR can function as an antimorph. Overall, our results suggest a model where a single site of m6A modification on HOTAIR enables a strong interaction with YTHDC1 for chromatin-mediated repression of its target genes, leading to altered TNBC properties. Collectively, our results demonstrate the potent activity of m6A on lncRNAs and in turn the role they play in cancer biology.

RNA modifications such as m6A have been shown to play critical roles in several human cancers [ 25 ]. In breast cancer, studies have revealed that dysregulation of m6A levels can generate breast cancer stem-like cells and promote metastasis [ 26 – 28 ]. Of the currently designated m6A reader proteins, we have previously shown that hnRNP A2/B1, a proposed non-canonical reader lacking the m6A-binding YTH domain, can interact with HOTAIR to regulate its chromatin and cancer biology mechanisms by promoting HOTAIR interactions with target mRNAs [ 9 , 14 ] ( Fig 1A ). Additionally, a recent proteomics screen found that components of the m6A methyltransferase complex WTAP and RBM15 bind HOTAIR [ 29 ]. This evidence suggests that m6A may play a role in cancers where HOTAIR is overexpressed.

The m6A modification on an RNA is typically recognized by a “reader” protein that binds specifically to methylated adenosine to mediate the functional outcome of m6A deposition. Apart from the YTH family of proteins that contain the YTH domain that directly read m6A, a handful of non-canonical indirect m6A readers have been suggested [ 21 ]. In the case of Xist, the canonical YTH-containing nuclear localized m6A reader YTHDC1 recognizes m6A on Xist to mediate repression of the X chromosome [ 20 , 22 ]. In contrast, m6A on cis-acting chromatin-associated regulatory RNAs leads to their YTHDC1-dependent degradation, preventing transcription of downstream genes [ 23 ]. Collectively, m6A influences the regulatory roles of both mRNA and noncoding RNA via diverse mechanisms [ 24 ].

N6-methyladenosine (m6A) is a reversible RNA modification that can regulate multiple steps of the mRNA life cycle, including processing, decay, and translation [ 18 ]. How m6A regulates lncRNA-mediated processes is less understood. In one example, the lncRNA Xist, a key mediator of X chromosome inactivation, contains multiple m6A sites that contribute to its ability to induce repression of the X chromosome [ 19 , 20 ].

(A) General model for HOTAIR mechanism. HOTAIR is initially recruited to its target loci via RNA–RNA interactions with its mRNA targets that is mediated by hnRNP B1. HOTAIR association with chromatin induces transcriptional interference via an unknown mechanism, promoting heterochromatin formation by PRC2 through H3K27me3. This paper investigates the role of m6A on HOTAIR. (B, C) m6A RNA immunoprecipitation performed with an m6A antibody or IgG control in MCF-7 breast cancer cells (B) or MDA-MB-231 breast cancer cells with transgenic overexpression of HOTAIR (C). An m6A-modified region in EEF1A1 (EEF1A1 m6A) is a positive control, while a distal region in EEF1A1 that is not m6A modified (EEF1A1 distal) serves as a negative control. Three biological replicates are included. (D) Number of HOTAIR transcripts in MDA-MB-231 cells overexpressing WT HOTAIR, A783U HOTAIR, or an Anti-Luciferase control RNA. Experiments include 3 biological replicates each on 3 independently generated clones. (E) m6A sites detected in HOTAIR-expressing cells in 6 experiments (light green to dark green, scale of increasing occurrences). m6A site 783 (dark green, arrow) was detected in every experiment except where it was mutated. (F) m6A RNA immunoprecipitation performed with an m6A antibody or IgG control on MDA-MB-231 cells overexpressing WT HOTAIR, HOTAIR A783U, or an Anti-Luciferase control; 3–5 biological replicates were performed. (G) Doubling time of MDA-MB-231 overexpression cell lines described in (D). Experiments include 3 biological replicates each on 3 independently generated clones. (H) Quantification of Matrigel invasion assays performed with MDA-MB-231 overexpression cell lines described in (D). Four biological replicates were performed. Numerical values in panels 1B–D, 1F–H are included in S2 Data . lncRNA, long noncoding RNA; m6A, N6-methyladenosine.

HOTAIR promotes polycomb repressive complex 2 (PRC2) activity, resulting in heterochromatin [ 6 , 11 – 14 ]. However, HOTAIR initially represses genes upstream of PRC2, though the mechanism is unclear [ 15 ] ( Fig 1A ). HOTAIR also interacts with the repressor lysine-specific demethylase 1 (LSD1) [ 12 , 16 ], although it was recently proposed to inhibit normal function of LSD1 in maintaining epithelial cells [ 17 ]. How HOTAIR specifically accomplishes initial transcriptional repression at its target loci, and how other pathways and cancer contexts influence HOTAIR function, remain elusive.

The human lncRNA HOTAIR is a 2.2 kb spliced and polyadenylated RNA transcribed from the HoxC locus. Originally identified as a developmental regulator acting in trans to repress expression of the HoxD locus [ 5 ], aberrant high levels of HOTAIR are associated with poor survival and increased cancer metastasis in many different cancer types, including breast cancer [ 6 , 7 ]. Exogenous overexpression of HOTAIR in the MDA-MB-231 triple-negative breast cancer (TNBC) cell line results in the repression of hundreds of genes [ 6 ], and it promotes cell invasion, migration, proliferation, and self-renewal capacity in multiple breast cancer cell lines [ 6 , 8 , 9 ]. HOTAIR function is particularly striking in MDA-MB-231 cells, given that this is already a highly invasive breast cancer cell line, and its invasiveness is increased even further by HOTAIR overexpression [ 6 , 9 ]. This is reflective of the prognostic impact of HOTAIR expression in TNBC patients where high HOTAIR expression correlates with poorer overall survival [ 6 , 10 ]. MDA-MB-231 cells express low levels of endogenous HOTAIR, offering an opportunity to study response to HOTAIR transgenic overexpression, which is proposed to mimic the high levels of HOTAIR observed in patients with aggressive TNBC [ 6 ].

Long noncoding RNAs (lncRNAs) are becoming increasingly noted for their roles in transcriptional regulation [ 1 ]. Members of this class of noncoding RNAs are typically longer than 200 nucleotides, transcribed by RNA polymerase II, and processed similarly to mRNAs [ 2 ]. LncRNAs regulate transcription in a variety of ways; they can direct histone-modifying enzymes to their target loci to induce changes in chromatin or can regulate transcription directly by interacting with transcription factors and RNA polymerase II [ 1 ]. Importantly, lncRNAs are often key regulators of epigenetic changes that can drive cancer progression, frequently via their aberrant overexpression [ 3 , 4 ].

Results

HOTAIR contains multiple sites of m6A modification in breast cancer cell lines To investigate the possibility that m6A regulates the function of HOTAIR in a mechanism similar to its regulation of lncRNA Xist, we examined previous genome-wide maps of m6A sites in human cells. Using the CVm6A database [30], we found 3 m6A peaks in HOTAIR in HeLa cells, although the enrichment score for these sites was low (S1A Fig). To evaluate m6A methylation of HOTAIR in relevant breast cancer cells, we performed m6A RNA immunoprecipitation (meRIP) qRT-PCR in MCF-7 cells, which express low levels of endogenous HOTAIR [9]. A significant portion of HOTAIR was recovered upon immunoprecipitation with the anti-m6A antibody (26.6 ± 6.6%, p = 0.00024 versus IgG), similar to an m6A-modified region on the positive control region of EEF1A1 (24.7 ± 2.7%, p = 0.67), and consistently higher than a distal region of EEF1A1 that is not m6A modified (6.4 ± 3.7%, p = 0.0056) (Fig 1B). We further found that m6A modification of HOTAIR is maintained during ectopic expression of HOTAIR in a stable MDA-MB-231 cell line. meRIP in MDA-MB-231 cells expressing transgenic HOTAIR resulted in significant HOTAIR recovery (27.1 ± 5.4%, p = 0.001 versus IgG), similar to the EEF1A1 positive control (24.6 ± 7.0%, p = 0.65) and significantly higher than the EEF1A1 negative control (4.4 ± 2.0%, p = 0.02) (Fig 1C). As mentioned in the Introduction, this transgenic HOTAIR context is a model for high levels of HOTAIR observed in TNBC tumors. These results demonstrate that HOTAIR is m6A modified in 2 distinct breast cancer contexts representative of ER+ and TNBC. To identify single nucleotide sites of m6A modification, we performed a modified m6A eCLIP protocol [31] on polyA-selected RNA from MCF-7 and MDA-MB-231 breast cancer cells (S1B Fig). In MCF-7 cells, we identified 1 m6A site within the HOTAIR transcript at adenosine 783 (S1 Table). Based on this and the RIP result in MCF-7s (Fig 1B), we estimate that A783 is approximately 25% methylated, at a minimum. m6A at adenosine 783 in MDA-MB-231 cells with transgenic HOTAIR was consistently detected with high confidence (S1 Table), along with 7 other sites using our multi-replicate consensus approach [31] (S2 Table). Of note, A783 occurred within a single-stranded region of the HOTAIR secondary structure [16] (S1C Fig) that promotes METTL3/14 dependent methylation [32]. To confirm m6A modification at adenosine 783 in MCF-7 cells, we performed additional meRIP experiments to quantify the level of m6A modification in specific regions of HOTAIR across its transcript, demonstrating that the region containing m6A783 (723–808) had the highest level of HOTAIR recovery (S2A Fig). We performed additional m6A mapping experiments in three breast cancer cell lines that were recently generated from patient-derived xenografts [33]. One of these lines bears A783 m6A methylation as well, and this was the only m6A site identified in HOTAIR in these samples (S3 Table). HOTAIR levels were highest in this cell line (S3 Table and S2B Fig), suggesting that methylation status or detection ability scales with RNA level. To test if HOTAIR is m6A modified by the canonical m6A methyltransferase METTL3/14 complex, we performed shRNA-mediated depletion of METTL3, METTL14, and the adaptor protein WTAP in MCF-7 cells (S2D Fig, left) where A783 is the only methylated residue detected. We observed approximately 3- to 5-fold reduced recovery of HOTAIR in methyltransferase-depleted cells relative to non-targeting controls (p = 0.0063) (S2D Fig, right). Together, these results indicate that m6A methylation of HOTAIR, particularly at A783, occur in breast cancer cells and is dependent on the METTL3/14 complex.

Nucleotide A783 is important for the ability of HOTAIR to promote breast cancer cell proliferation and invasion Given that nucleotide A783 was consistently methylated within HOTAIR in our m6A mapping experiments, in both endogenous and overexpressed contexts, we asked whether this modification had any consequences to HOTAIR function. To directly test the functional role of A783, we mutated the adenosine to uracil at this position (HOTAIRA783U). Both wild-type (WT) and the mutant form of HOTAIR were expressed at similar levels averaging approximately 5,000 to 6,000 copies of HOTAIR per cell (Fig 1D), resembling the high levels of HOTAIR observed in samples from cancer patients [6,10,34]. We then mapped m6A sites in MDA-MB-231 cells overexpressing the HOTAIRA783U mutant as above (S1 Table). While the CLIP-based m6A signature was no longer detected at adenosine 783 when this site was mutated to uracil, we detected m6A modification at 5 of the 7 other multi-replicate consensus sites (S1 and S2 Tables). The results of the m6A mapping experiments in MCF-7 cells and MDA-MB-231 cells with transgenic overexpression of WT or A783U HOTAIR are summarized in Fig 1E. We note that nucleotides 143 and 620 were no longer called with multi-replicate consensus confidence as m6A in the A783U mutant, though m6A143 was only called in WT HOTAIR at our lowest confidence category and m6A620 is called in one of the A783U mutant replicates (S1 Table). Nonetheless, it is possible that methylation at A783 is required for one or both m6A events to occur. However, meRIP on MDA-MB-231 cells expressing HOTAIRA783U resulted in significant HOTAIR recovery (27.6 ± 10.8%, p = 0.00012 versus IgG), similar to WT HOTAIR-expressing cells (31.2 ± 6.1%, p = 2.16 × 10−7 versus IgG) (0.88-fold change, p = 0.49) (Fig 1F). meRIP analysis spanning the HOTAIR transcript in MDA-MB-231 cells expressing WT or A783U mutant HOTAIR also showed similar recovery across the HOTAIR transcript (S1D Fig). Recovery around A783 is unaffected by A783U, providing further evidence that the nearby m6A site at A772 remains methylated, in agreement with our meCLIP analysis (S1 Table). HOTAIRA783U had increased recovery in the region of HOTAIR containing m6A sites 143 and 215 (141–240) (A783U versus WT = 1.6, p = 0.0009) (S2D Fig). Altogether, these results suggest that HOTAIRA783U maintains m6A modification at other sites in HOTAIR. To determine the effect of the A783U mutation on HOTAIR-mediated breast cancer cell growth, we measured the doubling time of MDA-MB-231 cells expressing WT and A783U mutant HOTAIR. As described above, we overexpressed HOTAIR and the HOTAIRA783U mutant in MDA-MB-231 cells and included overexpression of an antisense sequence of luciferase mRNA (Anti-Luc) as a negative control [9]. Similar to previous studies, transgenic overexpression of HOTAIR mediated increased cancer growth and invasion of MDA-MB-231 cells [6] (Fig 1G and 1H). We observed that MDA-MB-231 cells overexpressing WT HOTAIR proliferated more quickly, with a shorter doubling time (26.5 ± 0.83 hours) than cells overexpressing Anti-Luc (28.7 ± 1.2 hours, p = 0.00046) (Figs 1G and S2E). Surprisingly, the single nucleotide mutation of A783U in HOTAIR abolished its ability to enhance MDA-MB-231 cell proliferation; cells expressing HOTAIRA783U proliferated more slowly, with a longer doubling time than those expressing WT HOTAIR (28.4 ± 0.87 hours, p = 0.00026) and grew similarly to cells containing the Anti-Luc control (p = 0.59). To examine the role of A783 of HOTAIR in mediating breast cancer cell invasion, the same MDA-MB-231 cell lines were plated in a Matrigel invasion assay. Overexpression of WT HOTAIR induced a significant increase in number of cells invaded compared to the Anti-Luc control (140.25 ± 23 in WT HOTAIR versus 102.7 ± 16 in control, p = 0.038). In contrast, overexpression of A783U HOTAIR did not lead to an increase in invasion compared to the Anti-Luc control (84.1 ± 22, p = 0.22) and resulted in significantly less cells invaded compared to overexpression of WT HOTAIR (p = 0.012) (Fig 1H). To confirm that the defects of the HOTAIRA783U mutant is due to lack of m6A methylation at that site and not due to an unintended structural alteration, we generated three additional HOTAIR mutants: A782U, A783C, and C784U (S3A Fig). Each of these mutations blocks m6A methylation at A783 either due to mutation of the m6A consensus sequence or by mutation of A783 itself, while having different local sequence changes than A783U. We generated stable MDA-MB-231 cell lines overexpressing each of these constructs and observed similar levels of HOTAIR expression across these lines (S3B Fig). Each mutation caused defects in HOTAIR-induced proliferation (S3C Fig) and invasion (S3D Fig), similarly to A783U. Altogether, these results suggest that m6A modification of adenosine 783 in HOTAIR is key for mediating the increased aggressiveness of TNBC that is promoted in contexts where the lncRNA is overexpressed.

hnRNP B1 is not a direct m6A reader in MCF-7 cells We next sought to address the mechanisms behind HOTAIR m6A783 function. hnRNP A2/B1 has previously been suggested to be a reader of m6A, and the B1 isoform has a high affinity for binding HOTAIR [9,35,36]. However, comparing our previously generated eCLIP results for hnRNP B1 [13] to the m6A eCLIP, both performed in MCF-7 cells, we found that, out of 10,470 m6A sites, only 417 (4%) were identified to contain an hnRNP B1 binding site within 1,000 nucleotides (S4A Fig). Upon mapping hnRNP B1 signal intensity relative to the nearby m6A site, we observed that hnRNP B1 is depleted directly over m6A sites (S4B Fig). These results suggest that hnRNP B1 is not a direct m6A reader, although m6A may indirectly promote its recruitment in some contexts. When comparing hnRNP B1 binding in HOTAIR with m6A sites, B1 binding peaks in MCF-7 cells occur in m6A-free regions of HOTAIR. Conversely, data from in vitro eCLIP analysis of B1 binding to unmodified HOTAIR reveal additional B1 binding peaks in Domain 1 of HOTAIR, one of which occurs near several m6A sites (S4C Fig). Altogether, these data suggest that m6A is not likely to directly recruit hnRNP B1 as a reader, although it could contribute to hnRNP B1 binding.

YTHDC1 interacts with methylated A783 in HOTAIR In light of the results for hnRNP A2/B1 described above, we turned to alternative candidate m6A readers of HOTAIR. YTHDC1 is a nuclear-localized m6A reader that binds m6A sites in noncoding RNAs, including the Xist lncRNA [20]. We reasoned that YTHDC1 was a strong candidate for interaction with HOTAIR, which is an lncRNA that is also primarily nuclear localized. To determine if YTHDC1 interacts with HOTAIR, we performed RNA immunoprecipitation (RIP) qRT-PCR using an antibody to YTHDC1. In both MCF-7 cells expressing endogenous HOTAIR where m6A783 is the only m6A site detected, and MDA-MB-231 cells overexpressing transgenic HOTAIR, a significant portion of HOTAIR RNA was recovered when using antibodies specific against YTHDC1 (2.6 ± 0.007%, p = 0.003 versus IgG; 17.4 ± 9.3%, p = 0.04 versus IgG, respectively) (Fig 2A and 2B). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. YTHDC1 interacts with HOTAIR at m6A783. (A, B) YTHDC1 RIP performed in MCF-7 cells (A) or MDA-MB-231 cells overexpressing transgenic HOTAIR (B) on 3 biological replicates. (C) YTHDC1 RIP performed in MCF-7 cells on 4 biological replicates. RNA recovery was monitored with qPCR probes noted in graph and normalized to recovery observed with qPCR oligos targeting a region of HOTAIR with no detected m6A sites (1819–1923). (D) YTHDC1 RIP performed in MDA-MB-231 cells overexpressing WT HOTAIR or A783U HOTAIR. RNA recovery was monitored with qPCR as in (C) and fraction recovered in mutant compared to WT was determined. Four biological replicates were performed. (E) Schematic of YTHDC1 pulldown experiment with m6A-modified WT and A783U HOTAIR. PP7-tagged domain 2 of WT or A783U HOTAIR was in vitro transcribed and m6A modified with purified METTL3/14. RNA was bound to PP7-Protein A, cellular extract containing FLAG-tagged YTHDC1 was added, and a pulldown was performed with IgG-coupled Dynabeads. Amount of FLAG-YTHDC1 bound was assessed by western blot. (F) Anti-FLAG western blot of pulldown experiment outlined in (E). (G) Quantification of anti-FLAG western blots from 3 replicates. Numerical values in panels 2A–D, 2G are included in S2 Data. RIP, RNA immunoprecipitation; WT, wild-type. https://doi.org/10.1371/journal.pbio.3001885.g002 To further examine interaction of YTHDC1 at m6A783, we performed YTHDC1 RIP experiments in MCF-7 cells, where A783 is the only methylated site detected, using qPCR probes targeting different regions of HOTAIR (see S4C Fig). We found that there was a significant enrichment in recovery of the region containing m6A783 compared to other regions of HOTAIR (4.73 ± 1.6-fold increase versus 3′ region, p = 0.0022 versus IgG, p ≤ 0.02 versus all other regions tested except 328–432 and Xist) (Fig 2C), supporting m6A-dependent association at A783. To determine whether m6A at A783 regulates interactions with YTHDC1, we performed YTHDC1 RIP experiments in MDA-MB-231 cells overexpressing WT or A783U mutant HOTAIR (Figs 2D and S4D). RNA recovery in the YTHDC1 immunoprecipitation in cells overexpressing WT HOTAIR was highest in the region containing A783 (2.05 ± 0.02-fold increase versus 3′ region), but also observed across the HOTAIR transcript, suggesting the potential for other sites of YTHDC1 occupancy on the lncRNA in this context where we have observed additional m6A sites (S1 Table). While the majority of the HOTAIR transcript had similar recovery in WT versus A783U, region 723–808, the region containing the m6A783 site, was most affected by A783U mutation (58 ± 14% reduced recovery in A783U versus WT, p = 0.0003), along with 1 additional region at 1605–1728 (51 ± 7.8% reduced recovery in A783U versus WT, p = 1.6 × 10−5). This change in 1605–1723 is likely due to a secondary interaction of YTHDC1 with this region mediated by m6A783 rather than changes in direct m6A interactions, as we do not observe changes in m6A RIP recovery of this region in HOTAIRA783U (S2D Fig). These results demonstrate a defect in the HOTAIRA783U mutant in its ability to interact with YTHDC1, highlighting a loss of YTHDC1 at the region containing nucleotide 783 when it cannot be methylated. To examine the association of other m6A readers, we performed similar RIP experiments using antibodies to YTHDF1 or YTHDF2 in MCF-7 cells. While a proportion of HOTAIR was recovered with these antibodies, there was no enrichment for the region containing m6A783 (S4E and S4F Fig). This supports a specific role for YTHDC1 as the m6A reader that targets m6A783 of HOTAIR. To further confirm that nucleotide A783 in HOTAIR recruits YTHDC1 via m6A modification, we generated PP7-tagged in vitro transcribed RNA of domain 2 of WT or A783U mutant HOTAIR and performed in vitro m6A methylation with purified METTL3/14 [37] and S-adenosylmethionine as a methyl donor. As noted above, A783 has low structural propensity [16], making it a favorable METTL3/14 substrate. In this purified system, it is possible that other sites can be methylated, even those that do not occur in cells; however, in this context, the only difference between the 2 constructs used is the A783 base substitution, which prevents methylation at this site. The in vitro HOTAIR transcripts were tethered magnetic beads and incubated with FLAG-YTHDC1-containing protein lysates, then the relative recovery of FLAG-YTHDC1 was determined by anti-FLAG western blot (Fig 2E). WT HOTAIR interaction with YTHDC1 was enhanced when the transcript was m6A modified (approximately 3-fold increase, p = 0.04), while A783U HOTAIR interaction with YTHDC1 was not significantly altered by the addition of m6A (approximately 1.3-fold change, p = 0.6) (Fig 2F and 2G). Further supporting a specific role for m6A783, we note the non-canonical “GAACG” sequence at this location was identified as one of the top 10 most abundant m6A-centered 5-mer sequences to interact with YTHDC1 in an in vitro selection study [38]. Altogether, these results suggest that m6A783 of HOTAIR mediates a specific interaction with YTHDC1.

YTHDC1 levels regulate HOTAIR-mediated proliferation of MDA-MB-231 cells and clinical outcomes To test the role of YTHDC1 in HOTAIR’s ability to enhance breast cancer cell proliferation, we stably overexpressed or knocked down YTHDC1 in the context of WT or A783U HOTAIR overexpression in MDA-MB-231 cells (Fig 3A and 3B). We noted that YTHDC1 protein levels tended to be approximately 2-fold higher in cells containing WT HOTAIR compared to A783U mutant HOTAIR (Fig 3B). Although this difference was not significant (p = 0.16), it suggests a potential positive relationship between WT HOTAIR RNA and YTHDC1 protein levels, indicating that high levels of methylated A783 may stabilize a fraction of YTHDC1. Next, we used the MDA-MB-231 cell lines we generated to analyze proliferation as described above. Growth of MDA-MB-231 cells overexpressing WT HOTAIR was not significantly altered by YTHDC1 dosage (0.96-fold change, p = 0.16 for pLX-DC1; 1.08-fold change, p = 0.26 for shDC1, respectively), yet there was a trend towards decreased doubling time with increasing YTHDC1. In contrast, cells with A783U mutant HOTAIR had significant differences in doubling time with overexpression or knockdown of YTHDC1 (Fig 3C). Overexpression of YTHDC1 led to significantly faster growth of MDA-MB-231 cells containing A783U mutant HOTAIR (0.84-fold change in doubling time, p = 0.003), with proliferation rates comparable to cells expressing WT HOTAIR. Knockdown of YTHDC1 in cells containing HOTAIRA783U was particularly potent in reducing the growth rate (approximately 1.2-fold increase in doubling time, p = 0.008). These results suggest that without A783 methylation, the reduced occupancy of YTHDC1 specifically at A783 can be partially compensated by YTHDC1 overexpression and aggravated by knockdown. We suspect this could be mediated by the secondary m6A sites of HOTAIR, permitting some level of compensatory function. However, these results occur in the background of general manipulation of YTHDC1 levels, which likely has additional pleiotropic affects. We address this issue with more specifically targeted experiments using the dCasRX system (see section on “Tethering YTHDC1…”). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. YTHDC1 regulates HOTAIR activity and stability. (A) Western blot results of YTHDC1 protein levels in pLX-DC1 overexpression, shNT control, and shDC1 knockdown MDA-MB-231 cell lines expressing WT or A783U HOTAIR. (B) Quantification of 3 replicates of (A). Protein levels of YTHDC1 were normalized to β-actin levels and are relative to the HOTAIR shNT sample. Four biological replicates were performed. (C) Doubling time of MDA-MB-231 cells containing WT or A783U HOTAIR and overexpression or knockdown of YTHDC1. Three biological replicates on 2 independently generated clones were performed. (D) qRT-PCR was performed on fractionated RNA samples from MDA-MB-231 cells containing overexpression of WT or A783U HOTAIR, and chromatin association was calculated by determining the relative chromatin-associated RNA to input and normalizing to 7SL levels and relative to WT HOTAIR samples. Three biological replicates were performed. (E) Chromatin enrichment was calculated similarly as in (D) in MDA-MB-231 cell lines expressing WT or A783U HOTAIR with knockdown or overexpression of YTHDC1. Values are relative to HOTAIR shNT samples. Three biological replicates each were performed on 2 independently generated clones. (F) qRT-PCR of HOTAIR RNA levels in MDA-MB-231 cell lines overexpressing WT or A783U HOTAIR containing overexpression or knockdown of YTHDC1. Three biological replicates each were performed on 2 independently generated clones. (G) qRT-PCR of HOTAIR RNA levels in MDA-MB-231 cell lines expressing WT, A783U, 6xAU, or 14xAU HOTAIR or an AntiLuc control. Three biological replicates each were performed on 2 independently generated clones. Numerical values in panels 3B–G are included in S2 Data. WT, wild-type. https://doi.org/10.1371/journal.pbio.3001885.g003 To explore how breast cancer outcomes are affected by HOTAIR and YTHDC1 levels, we used Kaplan–Meier plotter [39]. We found that high HOTAIR levels were only significantly associated with decreased survival in the context of high YTHDC1 mRNA (S5A and S5B Fig), suggestive of a role for YTHDC1 in enhancing HOTAIR’s ability to mediate more aggressive cancer. Interestingly, using UALCAN (a tool for analyzing cancer OMICS data) to determine gene expression in normal breast tissue versus breast tumor specimens [40], there appears to be a lack of concordance in the relationship between YTHDC1 mRNA and protein levels in clinical samples: Average YTHDC1 mRNA levels were decreased across a breast tumor panel, while average protein levels were increased (S5C and S5D Fig). However, we reason that most samples with high YTHDC1 mRNA levels were likely to have higher protein levels. These data hint that the cancer phenotypes dependent on HOTAIR association with YTHDC1 may have clinical implications. Discovery of methylated A783 in cells recently derived from a breast tumor (S3 Table) support this as well.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001885

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/