Therapeutic targeting using tumor specific peptides inhibits long non-coding RNA HOTAIR activity in ovarian and breast cancer

Long non-coding RNAs (lncRNAs) play key roles in human diseases, including cancer. Functional studies of the lncRNA HOTAIR (HOX transcript antisense RNA) provide compelling evidence for therapeutic targeting of HOTAIR in cancer, but targeting lncRNAs in vivo has proven to be difficult. In the current study, we describe a peptide nucleic acids (PNA)-based approach to block the ability of HOTAIR to interact with EZH2 and subsequently inhibit HOTAIR-EZH2 activity and resensitize resistant ovarian tumors to platinum. Treatment of HOTAIR-overexpressing ovarian and breast cancer cell lines with PNAs decreased invasion and increased chemotherapy sensitivity. Furthermore, the mechanism of action correlated with reduced nuclear factor-kappaB (NF-κB) activation and decreased expression of NF-κB target genes matrix metalloprotease 9 and interleukin 6. To deliver the anti-lncRNA to the acidic (pH approximately 6) tumor microenvironment, PNAs were conjugated to pH-low insertion peptide (pHLIP). Treatment of mice harboring platinum-resistant ovarian tumor xenografts with pHLIP-PNA constructs suppressed HOTAIR activity, reduced tumor formation and improved survival. This first report on pHLIP-PNA lncRNA targeting solid tumors in vivo suggests a novel cancer therapeutic approach.

tumor-specific targeting of non-coding RNA has-miRNA-155 in lymphoma 14 , providing proof-of-concept for PNA-mediated delivery of noncoding RNAs in human disease.
In this report, we developed a PNA-targeting strategy for HOTAIR serving as a scaffold for polycomb repressive protein complex 2 (PRC2), PRC2 enrichment at specific loci, trimethylation of histone H3 lysine K27 (H3K27me3) by enhancer of zeste 2 (EZH2) and subsequent gene repression 5,15 . We demonstrate that PNA targeting of HOTAIR RNA single stranded regions 16,17 effectively blocks the HOTAIR-PRC2 interaction, inhibits ovarian and breast cancer cell invasion and re-sensitizes to chemotherapy via NF-κB activation and secretion of IL-6 in vitro. The "anti-lncRNA" agent decreased ALDH1A1 activity in ALDH(+) ovarian cancer cells, suggesting HOTAIR inhibition with PNA could reduce the ovarian cancer stem cell population. Conjugating PNAs to pH-low insertion peptide (pHLIP) allowed for PNA-targeting to the in vivo tumor microenvironment, circumventing lower tumor pH levels due to oxidative phosphorylation 18 . In mice harboring platinum-resistant ovarian tumor xenografts, pHLIP-PNA treatment resensitized and reduced tumor growth and prolonged survival. The results represent the first demonstration of PNA-targeting a lncRNA in a solid malignancy in vivo and suggest a novel cancer therapeutic approach for ovarian, breast and other solid cancers.

Results
Inhibiting HOTAIR and EZH2 alters platinum sensitivity and cell behaviors. Inhibiting either HOTAIR 9 or EZH2 19,20 has been reported to reduce tumorigenesis and increase survival in vivo. To examine the effect of inhibiting both HOTAIR and EZH2, we treated a highly platinum-resistant ovarian cancer cell line (A2780_CR5) with dsiRNA targeting HOTAIR and (or) a pharmacological inhibitor of EZH2 (GSK126) and performed survival assays and observed an additive (P < 0.05) on drug sensitivity and survival (Fig. 1A, inhibiting both HOTAIR and EZH2 vs. inhibiting either factor alone). To disrupt the HOTAIR-EZH2 interaction, we targeted the 89-mer minimum interacting region of HOTAIR, which has been recently reported to bind EZH2 16 . By using mFold 21 to validate the predicted secondary structure of this site, we observed a highly predicted, single-stranded region complementing previous such structures (Supplementary Fig. S1A; total of 19 predicted structures) and designed peptide nucleic acids (PNAs) complementary to the single stranded region of the 89-mer domain (Fig. 1B, red lines). Individual PNAs (PNAs 1-5, Supplementary Table S1) were combined with in vitro transcribed, biotinylated full-length HOTAIR (1μM) and recombinant EZH2 ( Supplementary Fig. S1B). Of the five PNAs examined (1 μM each), only PNA3 reduced (approximately 80%) the HOTAIR-EZH2 interaction ( Fig.1C and Supplementary Fig. S1C). As such, PNA4 was used as the control PNA, as it had no effect on HOTAIR-EZH2 interaction. No effect of the other PNAs was observed and importantly none of the PNAs altered the EZH2-ALU (control RNA) interaction ( Fig. 1C and Supplementary Fig. S1D), further demonstrating PNA3specific inhibition of the HOTAIR-EZH2 interaction. The ability of PNA3 to bind in vitro transcribed HOTAIR using gel shift assay. At 1 × 1 −2 μM PNA3, a shifter band was observed ( Supplementary Fig. S1E), whereas no observable band shift was seen with control PNA. PNA3 bound HOTAIR from HEK293 cell lysate ectopically overexpressing full length HOTAIR ( Fig. 1D; 8-fold enrichment of HOTAIR with PNA3 compared to control PNA, determined by qRT-PCR), whereas a no such enrichment was observed using non-specific primer control and primers corresponding to the lncRNA FIRRE. In order to exclude possible off-targets, the sequences of each PNA was aligned to the genome. No significant genes were found to interfere specifically with PNA3, whereas several genes were found to be complementary to PNA5 (Supplementary Table S2).
A positive association between HOTAIR and the master transcription factor NF-κB has been reported (Chu et al., 2011), and we demonstrated NF-κB-mediated transcriptional regulation of HOTAIR induced epigenetic silencing of Iκ-Bα, resulting in a positive feedback loop that ultimately increased NF-κB activation in ovarian cancer 9,23 . We thus performed a cytokine/chemokine screen and then measured HOTAIR levels. Of the cytokines In vitro transcribed and biotinylated HOTAIR or ALU RNA (1 μM) were incubated with PNA1-5 (1 μM) for 1 hr at 25 °C followed by pull-down with streptavidin coated protein A/G plus agarose beads. 50 ng of recombinant EZH2 was added and binding was observed in a polyacrylamide gel. (D) PNA3 or control PNA were biotinylated and incubated with MCF-7 cell lysate for 1 hr followed by a modified ChIRP assay. Graph represents the fold change of HOTAIR as measured by qRT-PCR compared to control non-specific gene and lncRNA FIRRE. (E) A2780_CR5 cells were treated with either PNA3 or control PNA (1 μM) for 24 hrs, and re-plated for clonogenic survival and treated with either CDDP (30 μM) or etoposide (5 μM) for 3 hrs and 24 hrs post-treatment (F) Caspase 3/7 cleavage assay was performed. All western blot data were cropped and acquired under same experimental conditions. Asterisks indicate P < 0.05 (*) or P < 0.01 (**).  examined, HOTAIR expression was increased (P < 0.05) by TNF-α (>15-fold) and TGF-β (5-fold) compared to control (Fig. 3A), in agreement with previous findings 9, 23 . To confirm HOTAIR induction of NF-κB, we used a previously reported luciferase reporter construct containing the E-selectin promoter (861 base pairs containing 3 canonical NF-κB-p65-binding sites as a positive control; Fig. 3B) 24 . Cells were transfected with luciferase constructs and treated with control PNA or PNA3 for 24 hrs and luciferase activity was measured. We observed a 1.4-fold increase (P < 0.05) in luciferase activity by ectopic overexpression of HOTAIR compared to vector control (Fig. 3B), which was decreased (P < 0.05) by PNA3 (Fig. 3B). Furthermore, PNA3 treatment of ovarian and breast cancer cells decreased (P < 0.05) IL-6 secretion into the media (Fig. 3C). Because secreted IL-6 contributes to chemoresistant cancer stem cells by inducing aldehyde dehydrogenase (ALDH1A1) we performed a survival assay with conditioned media (CM) from ovarian cancer cells treated for 24 hours with either PNA3 or control PNA. Increased (P < 0.05) sensitivity to CDDP for cells treated with PNA3 CM vs. control CM was observed, and this effect was rescued with recombinant IL-6,, suggesting that inhibiting IL-6 secretion is a contributing factor to chemosensitivity (Fig. 3D), an effect further supporting claims from our previous work 9 . Furthermore, as the IL6-STAT3 axis regulates ALDH1A1 activity and contributes to ovarian cancer stem cell enrichment 25, 26 , we measured HOTAIR levels in ALDH1A1 positive A2780_CR5 vs. negative cells. An approximate 1600-fold increase in ALDH1A1 expression (Fig. 3E) was detected as well as a 3-fold increase in HOTAIR expression in ALDH1A1 positive ovarian cancer cells relative to negative cells (Fig. 3E). Treatment of A2780_CR5 cells with PNA3 decreased (25%) ALDH1A1 activity (Fig. 3F), suggesting HOTAIR inhibition with PNA could reduce the cancer stem cell population.
Effect of pHLIP-conjugated PNA3 on CDDP sensitivity, tumor formation and survival. An acidic tumor micro-environment (pH~6 vs pH 7), due to increased glycolysis resulting in lactic acidosis (Warburg effect; 27 ), has been reported for solid tumors including breast 28 and ovarian 29 cancers. To target PNAs to the acidic tumor microenvironment, we used pH-low insertion peptides (pHLIPs), which are unstructured peptides in either neutral pH or basic pH and can thus interact with the outer surface of lipids in a reversible manner (Fig. 4A). Based on a previous report of successful pHLIP-PNA targeting a non-coding RNA (microRNA-155) 14 , we conjugated thiolated pHLIP peptide to PNA3 and control PNA (verified using tricine SDS-PAGE gel, Fig. 4B) and examined pHLIP-PNA A2780_CR5 cell entry (normal pH 7.2 vs. acidic pH 6.0 conditions) using immunofluorescence. Signals in cytoplasm, nucleus and cell periphery were observed (Fig. 4B). To examine PNA3 resensitization of A2780_CR5 cells to CDDP, cells were treated with either pHLIP-conjugated-PNA3 or -control PNA under normal or acidic pH and various CDDP concentrations (0, 15, 45 μM) and an MTT survival assay was performed. No change in cell survival was observed between pHLIP-PNA3 and pHLIP-control under normal pH; however in pH 6, pHLIP-PNA3 decreased (P < 0.05) survival (Fig. 4C), indicating HOTAIR targeting and altered CDDP-sensitivity under acidic conditions. To investigate anti-tumor properties of pHLIP-PNA3 in vivo, BALB/c-nu/nu mice were injected with CDDP-resistant A2780_CR5 (2 × 10 6 cells subcutaneously). Once tumors reached ~200 mm 3 , mice were injected intravenously (biweekly for 2 weeks) with pHLIP-PNA3 (1 mg/ kg), pHLIP-control PNA (1 mg/kg) and/or CDDP (2 mg/kg i.p.) (Supplementary Fig. S4F). Tumor volume was reduced (P = 0.02) in mice co-administered pHLIP-PNA3 + CDDP compared to pHLIP-control PNA + CDDP (Fig. 4D), and tumor volume in mice treated with either pHLIP-PNA alone was similar to vehicle-treated mice (Fig. 4D). Survival of mice treated with pHLIP-PNA3 was increased (64% vs. mock; Fig. 4E). Body weight among groups was similar ( Supplementary Fig. S4G), demonstrating that PNAs are non-toxic in vivo.

pHLIP-PNA3-cisplatin combination treatment decreases HOTAIR targets in vivo.
As a positive correlation between tumor growth and the pro-inflammatory cytokine IL-6 has been described 30 (Fig. 3C), and we recently demonstrated that HOTAIR upregulated both IL-6 and MMP-9 in ovarian cancer cells 9 , it was of interest to examine the effect of PNA3 on IL-6 in vivo. In ovarian tumor-bearing mice, blood IL-6 levels increased (P < 0.01) after treatment with CDDP alone or pHLIP-control PNA + CDDP, and IL-6 blood levels were reduced (P < 0.05) after pHLIP-PNA3 + CDDP treatment compared to pHLIP-PNA control + CDDP (Fig. 4F). In addition, tumor expression of IL-6, MMP-9 and ALDH1A1 was decreased (P < 0.05) in mice treated with pHLIP-PNA3 compared to control (Fig. 4G). The increase in IL-6 levels with pHLIP-PNA4 could be due to the activation of pro-inflammatory pathways in the tumors generated by the penetration of pHLIP-PNA4. Taken together, the results demonstrate that pHLIP-PNA3-mediated HOTAIR inhibition reduces ovarian tumor levels of IL-6, MMP-9, and ALDH1A1, increases CDDP sensitivity and improves overall survival.
Platinum-based therapies have been reported to affect the liver and spleen 31 . CDDP treatment increased (P < 0.01) spleen size compared to mock treated (Fig. 4H), but interestingly the combination of pHLIP-PNA3 with CDDP abrogated (P < 0.01) the CDDP-induced increase in spleen size (Fig. 4H). A similar trend for liver size was observed (Fig. 4I) albeit not statistically significant. Furthermore, no apparent histological changes (based on H&E staining) were seen ( Supplementary Fig. S4H), but H&E slides of ovarian tumors from CDDP-treated mice exhibited fewer cells ( Supplementary Fig. S4H), indicating cell death.

Discussion
The lncRNA HOTAIR is frequently overexpressed in solid tumors 32 and correlates with disease progression chemoresistance and poor patient prognosis 9,33 . The oncogenic activity of HOTAIR is dependent upon its interaction with the PRC2 complex, specifically EZH2, an epigenetic modifier frequently perturbed in cancer 34 . In the current study, by inhibiting the activity of the EZH2-binding partner HOTAIR, which is frequently co-expressed in EZH2-overexpressing cancers 35 , we demonstrate a novel and effective strategy for resensitizing resistant ovarian tumors to platinum. We describe a PNA-based approach to block the ability of HOTAIR to interact with EZH2 and subsequently inhibit HOTAIR-EZH2 activity in vitro and in vivo. To initially examine the impact of combined inhibition of HOTAIR and EZH2, we treated ovarian and breast cancer cells with HOTAIR siRNA plus EZH2 catalytic activity inhibitor (GSK126), observing increased chemotherapy sensitivity and reduced cell survival (Fig. 1A). Furthermore, similar to combining genetic knockdown with pharmacologic inhibition, using a PNA to block to the recently described EZH2-interactiing domain of HOTAIR 16 and disrupt the HOTAIR-EZH2 interaction resensitizes cancer cells to clinically relevant cytotoxic chemotherapies (Figs 1E,F, 2, 3 and 4), reduces cell invasion ( Fig. 2A,B) and decreases NF-κB transcriptional activity (Fig. 3B) and IL-6 and MMP-9 expression in vivo (Fig. 4F,G). Similar efficacy of PNA in cell lines with low endogenous HOTAIR such as SKBR-3 and MDA-MB-231 cells could be due to a "dosage effect" where lower HOTAIR levels are more efficiently inhibited by PNA. Our findings on IL-6, a pro-survival cytokine that can transform cells to a "pro-inflammatory cell", also indicate a potential approach for inhibiting IL-6 secretion, tumor progression and chemotherapy resistance development.
Approaches for targeting non-coding RNAs in tumors in vivo include siRNA-mediated knockdown and locked nucleic acids. We show that PNA-pHLIP conjugation is effective in an acidic tumor microenvironment, suggesting that the approach could overcome the impact of Warburg effect, a well-known fundamental aspect of malignant transformation 27 . We further demonstrate that PNA-pHLIP can be safely (based on no change in body weight) and effectively (based on reduced tumor burden) combined with cytotoxic chemotherapy, including platinum-based drugs currently used in the clinic. Importantly, PNA3-pHLIP treatment lowers both tumor and blood levels of IL-6, suggesting that impacting the local (tumor) microenvironment may result in systemic (peripheral) effects, such as reducing inflammation. Although the observed improvement in mouse survival may be considered modest (2 weeks compared to control), the duration represents nearly a two-year increase when converted into human years.
Solid tumors are characteristically associated with an acidic environment as well as reduced oxygen levels, which activates HIF-1α 28 , an oncogene that further promotes tumor growth under low oxygen levels and increase cancer stem cells population 36 . Interestingly, HIF1α was recently shown to regulate HOTAIR expression under hypoxic conditions 37 . As IL-6-STAT3 axis can induce expression of ALDH1A1, a cancer initiating cell marker 38 , it was of interest to determine the effect of PNA treatment on this cell population. We show ALDH1A1(+) ovarian cancer cells display increased HOTAIR expression (Fig. 3E) and PNA3 treatment decreases ALDH1A1 level in vitro (Fig. 3F) and in vivo (Fig. 4G). The observation that PNA3 treatment also decreased IL-6 levels (Figs 3C and 4F,G) indicates that PNA3-targeting enhanced the response of ovarian cancer initiating cells to CDDP (Figs 3E,F and 4G). Our future work will investigate the role of HOTAIR and the lncRNA as a possible therapeutic target in the ovarian cancer stem cell population.
In conclusion, we report for the first time an anti-lncRNA targeting approach in ovarian tumors in vivo. The findings warrant further development of this strategy for targeting oncogenic lncRNAs as a therapeutic strategy in solid tumors.  Table S3) were maintained in EMEM or McCoys media (Invitrogen, Carlsbad, CA). Cell lines were authenticated in 2012 by ATCC and tested for mycoplasma contamination (Manassas, VA). Cisplatin (CDDP) was purchased from Calbiochem (Billerica, MA), and etoposide was purchased from Santa Cruz Biotech (Santa Cruz, CA). LZRS-HOTAIR was a gift from Dr. Howard Chang (Stanford University; Addgene plasmid #26110). Full-length HOTAIR was cloned into pAV5S vector containing a 98-mer aptamer sequence and as a vector control, aptamer cloned into pAV5S was used to account for any possible RNA-dependent signaling effects 40 . Proliferation MTT assays. Cells were grown in 6 cm culture plates until 70% confluence and treated with either PNA3 or control PNA (1 μM final) for 24 hrs. Next day, plates were trypsinized, counted, 2 × 10 3 cells were seeded into 96-well and MTT assay was performed as previously described 9 .
Cell invasion assays. Cells were grown in 6 cm culture plates until 70% confluence and treated with either PNA3 or control PNA (1 μM final) for 24 hr. Next day, 50,000 cells were seeded inside a matrigel invasion chamber insert (Corning Inc., Corning, NY) in serum free media supplemented with 0.1% BSA. Cells were fixed 48hrs later and analyzed 9 .
In vitro transcription RNA. Full length T7-promoter driven HOTAIR and ALU cDNAs were cloned into pcDNA3.1 with a single NHEI restriction site after the transcription stop site. DNA was linearized with NHEI and DNA was in vitro transcribed into RNA according to manufactures protocol (New England Biolabs, Ipswich, MA). The total RNA was purified and DNAseI treated and purified per manufacturers protocol (Qiagen).
Biotinylation, folding, and immunoprecipitation of RNA. Purified RNA (1.67 μM) was 3′-biotinylated according to manufacturers protocol (Thermo Scientific). After biotinylation, RNA was purified and folded. Folded 3′ biotinylated ALU or HOTAIR RNA was incubated with individual peptide nucleic acids (PNAs) (5 μM final) (Supplementary Table S1) PNA Bio (Thousand Oaks, CA) in 10 μL of 1x folding buffer supplemented with RNAse inhibitor (Thermo Scientific) and bovine serum albumin (5 μg BSA) for 30 min at 37 °C. Next, streptavidin HRP antibody (Cell Signaling, Danvers, MA) (Supplementary Table S4) was added to binding buffer supplemented with RNAse inhibitor and incubated at 4 °C for 1 hr. Next, protein A/G plus agarose beads (25 μL; Santa Cruz Biotech) were added and placed into 4 °C rotator for 1 hr. The beads were washed and recombinant polycomb repressive complex 2 (PRC2, 0.1 nM final, Active Motif, cat #31387) was added and incubated for 3hr at 4 °C on a rotator. After incubation beads were washed 3x with 1X IP buffer supplemented with RNAsein and run on BioRad precast polyacrylamide gel.
Clonogenic survival and Caspase 3/7 cleavage assays. 24 hrs after treatment with PNA, cells were washed with 1X PBS and were either not treated or treated with indicated concentrations of CDDP or etoposide. Cleaved Caspase 3/7 activity, indicative of apoptosis, was detected according to manufacturers protocol (Promega, Madison, WI). Percent survival of treated cells was calculated relative to untreated samples.
Aldefluor assay and flow cytometry. ALDH1 enzymatic activity was measured using the Aldefluor assay kit (Stemcell Technologies, Vancouver, Canada) following the manufacturer's instructions and as we have described 41 . ChIPNA assay. Cells were treated with 1 μM of biotinylated PNA3 or control PNA was added. 24 hrs later cells were, trypsinized, and fixed with 4% formalin. The nuclei were isolated as previously described 9 then resuspended in nuclei ChIP lysis buffer The soluble fraction was incubated with anti-streptavidin antibody (Supplementary Table S4) for 2 hrs followed by binding of protein A/G plus agarose beads (Santa Cruz Biotech) for an additional 2 hrs at 4 °C. Beads were washed 3 times with wash buffer at 4 °C and then Proteinase K treated. Nucleic acid was separated with TRIzol and RNA was purified using RNAeasy column (Qiagen). RNA isolate (1 μL) was used per well for qRT-PCR analysis to confirm lncRNA retrieval. LncRNA FIRRE was used as a negative control, LncRNA ANRIL was used as a positive control.

Mouse xenograft experiments.
All animal studies adhered to ethical regulations and protocols approved by the Institutional Animal Care and Use Committee of Indiana University. To assess tumorigenicity of cells, cultured A2780_CR5 cells were re-suspended in 1:1 PBS/matrigel (BD Biosciences) and 2 × 10 6 cells were injected subcutaneously into the left flank of 3-to 4-week-old female nude athymic mice (BALB/c-nu/nu; Harlan, Indianapolis, IN), as described 41,42 . Engrafted mice (n = 6 per group) were inspected three times per week for tumor appearance by visual observation and palpation. Once tumors were ~200 mm 3 , mice were treated with either CDDP (2 mg kg −1 ) or PNA (1 mg kg −1 ) or both CDDP and PNA biweekly for two weeks. Blood samples were collected by puncturing the left lateral saphenous vein with a needle and collected using a capillary tube. Tumor length (l) and width (w) were measured biweekly using digital calipers and tumor volume (v) was calculated as v = ½ × l × w 2 . The investigator measuring tumor size was blinded to the treatment groups. Mice were sacrificed when tumor diameter reached 2 cm or at the end of study.
ELISA and cytokine release assays. Twenty-four hours post-PNA treatment, cells were rinsed with 1X PBS and incubated in serum-free RPMI medium. Total cell counts were determined and ELISA was performed using kits and procedures from R&D systems (Minneapolis, MN; Cytokine release assay,) and eBiosciences (San Diego, CA; IL-6 ELISA).
Luciferase assays. Cells were seeded in 96-well plates (10 4 cells/well) and transfected with pGL3-E-selectin vector (300 ng construct/transfection). Transfection efficiency was normalized with co-transfected PGL4 Renilla plasmid (100 ng). Twenty-four hours after transfection, cells were treated with PNA3 or Control PNA (1μM) for indicated times. Luciferase activity was analyzed using the Dual Luciferase Reporter Assay System (Promega) and a Thermo Scientific Multilabel Plate Reader.
RNA extraction and quantitative RT-PCR (qPCR). RNA was extracted from cell lines and tumors and using RNeasy kit (Qiagen, Venlo, Limburg), cDNA was prepared using MMLV RT system (Promega), and qPCR) was performed with total cDNA and primers for indicated genes and GAPDH or EEF1A as the endogenous control (Supplementary Table S5), using Applied Biosystems 7500 Fast RT-PCR system (Life Technologies, Grand Island, NY) and corresponding software, as we have described 41 . Primers used for can be found in Supplementary Figure S4.
Statistical analysis. All data are presented as mean values ± SD of at least three biological experiments unless otherwise indicated. IC 50 values were determined by Prism 6 (GraphPad Software, San Diego, CA), using logarithm normalized sigmoidal dose curve fitting. The estimate variation within each group was similar therefore student's t-test was used to statistically analyze the significant difference among different groups by using Prism 4.0 (GraphPad Software). For mouse xenograft study, statistical significance was determined using student two-tailed t-test.