Inhibition of lncRNA-NEAT1 sensitizes 5-Fu resistant cervical cancer cells through de-repressing the microRNA-34a/LDHA axis

Abstract Cervical cancer is one of the most diagnosed malignancies among females. The 5-fluorouracil (5-Fu) is a widely used chemotherapeutic agent against diverse cancers. Despite the initially encouraging progresses, a fraction of cervical cancer patients developed 5-Fu resistance. We detected that nuclear-rich transcripts 1 (NEAT1) was significantly up-regulated in cervical cancer tissues and cell lines. Moreover, NEAT1 was positively associated with 5-Fu resistance. Furthermore, expression of NEAT1 was significantly up-regulated in 5-Fu resistant CaSki cervical cancer cells. Knocking down NEAT1 by shRNA dramatically promoted the sensitivity of 5-Fu resistant CaSki cells. We observed a negative correlation between long noncoding RNA (lncRNA)-NEAT1 and miR-34a in cervical cancer patient tissues. Overexpression of miR-34a significantly sensitized 5-Fu resistant cells. Bioinformatics analysis uncovered that NEAT1 functions as a competitive endogenous RNA (ceRNA) of miR-34a in cervical cancer cells via sponging it at multiple sites to suppress expression of miR-34a. This negative association between NEAT1 and miR-34a was further verified in cervical cancer tissues. We found the 5-Fu resistant cells displayed significantly increased glycolysis rate. Overexpression of miR-34a suppressed cellular glycolysis rate and sensitized 5-Fu resistant cells through direct targeting the 3′-untranslated region (UTR) of LDHA, a glycolysis key enzyme. Importantly, knocking down NEAT1 successfully down-regulated LDHA expressions and glycolysis rate of cervical cancer cells by up-regulating miR-34a, a process could be further rescued by miR-34a inhibition. Finally, we demonstrated inhibition of NEAT1 significantly sensitized cervical cancer cells to 5-Fu through the miR-34a/LDHA pathway. In summary, the present study suggests a new molecular mechanism for the NEAT1-mediated 5-Fu resistance via the miR-34a/LDHA-glycolysis axis.


Introduction
Cervical cancer, one of the commonly diagnosed malignancies among females, is the main leading cause of female tumor death around the world (1). The high risks of morbidity and mortality rate of cervical cancer seriously impaired the quality of lives since most cervical cancer patients are diagnosed in the advanced stage (1,2). Although diagnosis and treatment of cervical cancer have been improved, the prognosis of patients still remains poor due to metastasis and chemoresistance (3,4). 5-fluorouracil (5-Fu) is one of systemically adjuvants and palliative chemotherapies against cervical cancers (5). Recently, multiple strategies including the 5-Fu-based combination with other chemotherapeutic agents have been developed to enhance the anti-tumor activity of 5-Fu (6).
Despite the initially encouraging progresses in cervical cancer therapy, a fraction of patients' response to 5-Fu gradually attenuated, resulting from the development of chemoresistance (7).
Consequently, 5-Fu resistance renders it a major challenge for the 5-Fu-based chemotherapy. Thus, investigating the underlying molecular mechanisms of chemoresistance is an essential task for developing novel therapeutic strategies against cervical cancer.
Long noncoding RNAs (lncRNAs), which consist of around 200 nucleotides, is a class of long-non-coding RNAs (8). They were known to play critical roles in various biological cellular processes as well as malignant diseases via regulating target genes expressions (9). Accumulating evidence have characterized lncRNAs to act as essential regulators during the initiation and development of cervical cancer (9). Among them, the LncRNA CRNDE has been reported to promote the growth and metastasis of cervical cancer cells through suppressing PUMA (10).
LncRNA nuclear-rich transcripts 1 (NEAT1) has been found to play an oncogenic role in diverse tumors, such as endometrial cancer (11), breast cancer (12), lung cancer (13), colon cancer (14) and gastric cancer (15). However, the precise roles of lncRNA NEAT1 in 5-Fu resistance and the underlining molecular mechanisms in cervical cancer has not been elucidated. As another class of short, noncoding RNA, microRNAs are capable to bind the 3′-untranslated region (UTR) of their target genes to suppress the corresponding mRNA translation and stability (16). Similar to Downloaded from http://portlandpress.com/bioscirep/article-pdf/doi/10.1042/BSR20200533/905252/bsr-2020-0533.pdf by guest on 03 March 2021 lncRNA, miRNAs play crucial roles in various cancer processes (17). Studies demonstrated miR-34a participates in the initiation and development of various cancers, including cervical cancer (18). However, the roles of the interaction of miR-34a with lncRNA NEAT1 in chemosensitivity of cervical cancer are unmasked.
Cancer cells adapt to tumor microenvironment by majorly depending on aerobic glycolysis for not only energy production but providing essential metabolic intermediates for biosynthesis of macromolecules, a phenomenon called the "Warburg effect" (19). Moreover, this new hallmark of cancer cells has been revealed to be associated with chemoresistance from preclinical and early clinical studies (20). Therefore, effective targeting the dysregulated glycolytic pathway has emerging as a therapeutic approach to overcome chemoresistance of cervical cancers.
In this study, the precise roles of lncRNA-NEAT1 in 5-Fu resistance of cervical cancer will be investigated. We found overexpression of NEAT1 de-sensitized cervical cancer cells to 5-Fu treatment. Molecular mechanism studies revealed that NEAT1 functions as a ceRNA of miR-34a and positively regulates the expression of LDHA, a glycolysis speed limit enzyme. These results indicate effective targeting the NEAT-miR-34a-LDHA axis could contribute to developing new therapeutic agents against 5-Fu resistant cervical cancer. General Hospital, China. All patients did not receive chemo-or radio-therapy before tissues collection. After surgical removal, tissue samples were frozen immediately by liquid nitrogen and stored at-80 °C until use. The clinical histopathological diagnosis of cervical cancer tissues was approved by pathologists. All patients agreed to and gave written informed consent.

Cell culture and reagents
Human normal cervical epithelial cell line, Ect1/E6E7 and cervical cancer cells Hela, C-33A, SiHa and CaSki were obtained from obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). Cells were cultured in RPMI 1640 medium (Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS) (Biorad, Hercules, CA, USA) and 100 IU/ml penicillin G and 100 μg/ml streptomycin (Sigma-Aldrich, Saint Louis, MO, USA) at 37˚C in a 5% CO 2 atmosphere. The establishment of 5-Fu resistant cervical cancer cell line was performed according to previous report (21). Briefly, CaSki parental cells were exposed to 5′-AACGCTTCACGAATTTGCGT-3′. The relative gene expression was calculated using the 2 −ΔΔCt method. Experiments were performed in triplicate and repeated three times.

Luciferase assay
The luciferase assay was performed according to previous reports (23). Briefly, 5x10 4 cervical cancer cells were grown in a 24-well plate to reach 70 % confluence. The Wild type or mutant LDHA 3′-untranslated region containing the predicted miR-34a binding site were ligated into the pmirGLO luciferase reporter vector (Promega, Madison, WI) to generate pmirGLO LDHA-Wt and pmirGLO-LDHA-Mut vectors which were co-transfected with miR-34a precursor or negative control miRNAs by lipofectamine 2000. Forty-eight hours after transfection, luciferase activity was examined using a Dual-Luciferase Assay Kit (Promega, Madison, WI, USA). Experiments were performed in triplicate and repeated three times.

Western blot
Cervical cancer tissue and cells were lysed on ice for 20 mins with RIPA lysis buffer (Beyotime, Shanghai, China) plus 1x protease inhibitor cocktail (Sigma, Shanghai, China). Protein concentrations were determined with a BCA Protein Assay Kit (Pierce; ThermoFisher Scientific).
Equal amount of protein from each sample were separated by 10% SDS-PAGE and transferred onto the nitrocellulose membrane. The membrane was first blocked with 5% non-fat milk in PSBT for 1 hour at room temperature followed by incubating with primary antibodies at 4 o C for overnight. After complete washing by PBST. A horseradish peroxidase (HRP)-conjugated secondary antibody was diluted in PBST buffer for incubating membrane at room temperature for 1 hour. After washing by PBST, protein bands from membrane were visualized by an enhanced chemiluminescence detection system (Beyotime, Shanghai, China). β-actin was an internal control. Experiments were repeated three times. Data were presented as mean ± standard deviation and analyzed by the Prism 6.0 software package (GraphPad Software, Inc.). All experiments were performed for three times independently. The difference between the two groups were analyzed with the Student's t-test, One-way ANOVA method was used for analyzing data from multiple groups. The p-value < 0.05 was considered statistically significant.

lncRNA-NEAT1 is positively associated with cervical cancer
To investigate the roles of lncRNA-NEAT1 in cervical cancer, we initially compared the expressions of NEAT1 in thirty-five paired human cervical cancer tissues and their adjacent normal tissues by qRT-PCR. Results in figure 1A showed that the expression of NEAT1 was significantly elevated in cervical cancer tissues compared with that of normal tissues. To further support the upregulation of NEAT1 in cervical cancer, the expressions of NEAT1 in cervical cancer cell lines, including HeLa, CaSki, SiHa and C33A, as well as a normal cervical epithelial cell line Ect1 were examined. As we expected, the expressions of NEAT1 was significantly upregulated in cervical cancer cells compared with normal cells. Taken together, the above results revealed that NEAT1 has an oncogenic role in cervical cancer.  (Fig. 2B, 2C). To obtain the direct evidence for the NEAT1-promoted 5-Fu resistance, the 5-Fu resistant cell line originating from CaSki cells were established. Cell survival assay demonstrated under 5-Fu treatments, the IC50 of 5-Fu resistant cells was 30.31 uM, which is around 3 times of that in CaSki parental cells (Fig. 2D). Moreover, we observed the lncRNA-NEAT1 expression was significantly increased in 5-Fu resistant CaSki cells (Fig. 2E). Collectively, these results demonstrated that NEAT1 facilitated the 5-Fu resistance of cervical cancer cells.

lncRNA-NEAT1 downregulates miR-34a as a sponge in cervical cancer cells
The above results demonstrated the association between NEAT1 and cervical cancer cells, the underlying mechanisms of the NEAT1-promoted 5-Fu resistance were further investigated.
Increasing evidence revealed that lncRNAs functions as competing endogenous (ce) RNAs through sponging miRNAs (24). To examine whether NEAT1 promotes the 5-Fu resistance of cervical cancer cells by binding and downregulating miRNAs as a ceRNA, we search the potential target miRNAs of NEAT1 from Starbase software http://starbase.sysu.edu.cn/. Interestingly, among multiple candidates, seeding region of miR-34a-5p was suggested to bind with NEAT1 at five sites (Fig. 3A). Since studies have reported that miR-34a could inhibits the tumorigenesis together with lncRNAs (25,26), we focused on miR-34a as ceRNA of NEAT1. Furthermore, we found a significantly invert correlation between lncRNA-NEAT1 and miR-34a in cervical tumor specimen (Fig. 3B). To validate the interactions between NEAT1 and miR-34a, NEAT1 was overexpressed or inhibited in both CaSki1 and SiHa cells. Expectedly, expressions of miR-34a was significantly attenuated by NEAT1 overexpression compared with control cells (Fig. 3C). On the other way, knocking down NEAT1 by shRNA obviously elevated the miR-34a expressions (Fig. 3D). These results demonstrated that NEAT1 sponged miR-9-5p and negatively regulated miR-34a expressions in cervical cancer cells.

miR-34a negatively associates with 5-Fu resistance and inhibits cellular glycolysis rate of cervical cancer cells
We next assessed whether miR-34a was associated with 5-Fu resistance of cervical cancer. The expression levels of miR-34a in cervical cancer tissues and normal adjacent tissues were compared and consistent results showed miR-34a was downregulated in cervical tumor specimen (Fig. 4A). In addition, miR-34a was significantly low expressed in 5-Fu resistant CaSki cells compared with parental cells (Fig. 4B), indicating miR-34a has tumor suppressive functions, present it as a new therapeutic target against chemoresistance in cervical cancer. Moreover, direct evidence showed overexpression of miR-34a sensitized CaSki and SiHa cells to 5-Fu (Fig. 3E, 3F).
Recent studies revealed dysregulated cellular glycolysis contributed to chemoresistance of cancer cells (20). To further investigate the underlying mechanisms for the miR-34a-mediated 5-Fu sensitivity, we examined the cellular glycolysis rate of cervical cancer cells without or with miR-34a overexpression. Consistent results showed overexpression of miR-34a apparently suppressed glucose uptake and lactate product, two readouts for detecting cellular glycolysis rates (Fig 3G, 3H). We also found the glycolysis key enzymes, HK2, PKM2 and LDHA were significantly inhibited by miR-34a (Fig. 4I). Taken together, the above results demonstrated a negative correlation between miR-34a and 5-Fu resistance in cervical cancer cells.

Inhibiting glycolysis sensitizes cervical cancer cells to 5-Fu
To evaluate whether the dysregulated glycolysis of cervical cancer cells directly facilitated 5-Fu resistance, the glucose uptake and lactate product were compared in CaSki parental and 5-Fu resistant cells. As we expect, 5-Fu resistant cells displayed obviously elevated glycolysis rates (Fig. 5A, 5B). Further, we observed CaSki 5-Fu resistant cells were sensitized to 5-Fu by either glycolysis inhibitor treatment (Fig. 5C) or knocking down LDHA by siRNA (Fig. 5D). These data indicating targeting the cellular glycolysis rate could effectively reverse 5-Fu resistance of cervical cancer cells.

miR-34a sensitizes cervical cancer cells to 5-Fu via directly targeting LDHA
Since miRNAs regulate biological functions through binding to 3'UTR of their targeting mRNAs (16). To determine the direct targets of miR-34a, we performed mRNA-miRNA binding analysis by bioinformatics prediction from TargetScan. Interestingly, LDHA, a glycolysis key enzyme which catalyzes pyruvate into lactate, was found as a putative downstream effector (19) (Fig. 6A).
Expectedly, immunohistochemistry experiments showed LDHA protein was significantly upregulated in human cervical tumor tissues (Fig. 6B). Western blot results demonstrated overexpression of miR-34a significantly downregulated protein expressions of LDHA in two cervical cancer cells (Fig. 6C). To identify whether LDHA was the direct target of miR-34a in cervical cancer cells, we constructed luciferase plasmids pGL3-LDHA-WT or pGL3-LDHA-Mut, which contain wild type or miR-34a binding site mutant 3'UTR of LDHA. They were co-transfected with miR-34a-5p precursor or negative control into CaSki and SiHa cells, respectively. Luciferase assays were performed and the activity of the LDHA WT reporter was obviously blocked, while that of the Mut reporter group did not change (Fig. 6D). These results verified that LDHA is a direct target gene of miR-34a-5p in cervical cancer cells. Subsequently, expressions of miR-34a and LDHA mRNAs in cervical cancer tissues displayed a remarkably negative correlation (Fig. 6E). The relative higher miR-34a expression tissues were accompanied with lower LDHA mRNAs (Fig. 6E). The above results demonstrated miR-34a could directly target LDHA in cervical cancer.
To elucidate whether the miR-34a-regulated 5-Fu sensitivity and glycolysis inhibition via targeting LDHA, we performed rescue experiments by transfection of CaSki cells with control miRNAs, miR-34a along or miR-34a plus LDHA overexpression vector. Western blot results indicated co-transfection of miR-34a and LDHA successfully restored the LDHA protein expression (Fig. 7A). Expectedly, CaSki cells with LDHA restoration showed recovery of glucose uptake and lactate product (Fig. 7B, 7C). Moreover, glycolysis key enzymes, HK2 and PKM2 were significantly recovered to normal expressions in miR-34a and LDHA co-transfected cells (Fig. 7D). We assessed whether rescue of LDHA in miR-34a overexpressed cells could overcome the 5-Fu sensitivity. Subsequently, cell viability assay and Annexin V/PITC apoptosis assay consistently demonstrated that CaSki cells with co-transfection of miR-34a and LDHA were re-sensitized to 5-Fu (Fig. 7E, 7F). Moreover, the rescue phenotypes were observed in another cervical cancer cell line, SiHa (Fig. S1A-S1C). Taken together, the rescue results verified the miR-34a-suppressed glycolysis and 5-Fu resistance were through directly targeting LDHA in cervical cancer cells.

Inhibition of NEAT1 sensitizes cervical cancer cells to 5-Fu through promoting miR-34a/LDHA axis
We further characterized whether the NEAT1-promoted 5-Fu resistance was dependent on miR-34a-LDHA axis. CaSki cells were transfected with control shRNA, NEAT1 shRNA, NEAT1 shRNA plus control miRNAs antisense or NEAT1 shRNA plus miR-34a antisense. As shown in figure 8A, S2, knocking down NEAT1 significantly downregulated LDHA and upregulated miR-34a. However, such regulations were further reversed by miR-34a inhibition (Fig. 8A, S2), indicating the NEAT1-promoted LDHA expression was through miR-34a inhibition. Consistently, co-transfection of NEAT1 shRNA and miR-34a inhibitor apparently recovered cellular glycolysis rate (Fig. 8B) and enzyme expressions (8C). Moreover, CaSki cells with co-transfection of NEAT1 shRNA and miR-34a inhibitor showed increased 5-Fu resistance, compared with NEAT1 knockdown alone (Fig. 8D). In summary, the above results elucidated that NEAT1 suppressed miR-34a to promote the glycolysis and 5-Fu resistance in cervical cancer, presenting the NEAT1/miR-34a-5p/LDHA axis as an effective target on regulating 5-Fu sensitivity of cervical cancer cells. Cervical cancer is one of the most common malignancy and the fourth prominent cancer-caused mortality in women (1)(2)(3)(4). Although the cervical cancer patients who received surgical operations, radiotherapy and chemotherapy demonstrated certain anti-cancer effects, prognosis of cervical cancer is still unsatisfied due to the acquired chemoresistance (4). Therefore, understanding the molecular mechanisms underlying the chemoresistance of cervical cancer emerges as an urgent task. In this study, we characterized the 5-Fu resistance cervical cancer cells and found lncRNA-NEAT1 was positively associated with 5-Fu resistance. Moreover, the cellular glycolysis rates were significantly increased in 5-Fu resistant cervical cancer cells, suggesting targeting the NEAT1-mediated glycolysis could effectively overcome chemoresistance.

Discussion
Accumulating evidence have demonstrated that NEAT1 functions as an oncogene in diverse cancers (11)(12)(13)(14)(15)(27)(28)(29). For instance, lncRNA NEAT1 promotes cell proliferation and migration of gastric cancer, cervical cancer, endometrial cancer, and bladder cancer (15,27,28,29). In breast cancer, NEAT is positively correlated with poor survival rate of patients and contributes to paclitaxel resistance in ovarian cancer cells through miR-194/ZEB1 axis (13). Currently, the underlying mechanisms of NEAT1 in regulating 5-Fu resistance of cervical cancer have not been elucidated. We found NEAT1 was significantly upregulated in cervical cancer patients and cells, consistent with previous report (27). Furthermore, by establishing 5-Fu resistant cervical cancer cell line, we observed obviously upregulated NEAT1 in 5-Fu resistant cells compared with parental cells, suggesting NEAT1 is positively associated with 5-Fu resistance of cervical cancer.
It has been acknowledged that lncRNAs and miRNAs take charge of diverse processes of tumors (8,9,16,17). Interestingly, recent studies have demonstrated that lncRNA could interact with microRNAs as competing endogenous RNAs to suppress miRNAs expression (24,30). Our findings firstly illustrated NEAT1 functions as ceRNAs to sponge miR-34a, thereby modulating the de-repression of LDHA, the direct target of miR-34a. We showed consistent results that overexpression of NEAT1 inhibited miR-34a expressions. On the other hand, NEAT1 Downloaded from http://portlandpress.com/bioscirep/article-pdf/doi/10.1042/BSR20200533/905252/bsr-2020-0533.pdf by guest on 03 March 2021 downregulation led to upregulated expression of miR-34a. Importantly, such invert correlation between NEAT1 and miR-34a was further verified in cervical cancer tissues. We identified LDHA as a direct target of miR-34a by luciferase assay. Furthermore, restoration of LDHA in miR-34a overexpressing cells successfully rescued the glycolysis rate, suggesting the miR-34a-mediated glycolysis inhibition was through targeting LDHA.
Proliferating tumor cells display increased lactate production during glucose metabolism even in the presence of sufficient oxygen (19). This phenomenon is known as the Warburg effect, which is recognized as new marker of cancer cells. Moreover, the Warburg effect is tightly related to tumor proliferation, progression and drug resistance (20,31,32). In this study, we focused our attention on the effects of the lncRNA-NEAT1-promoted glycolysis on 5-Fu resistance. In accordance with our expectation, overexpression of NEAT1 effectively de-sensitized cervical cancer cells through glycolysis upregulation. In addition, glycolysis inhibition by either glycolysis inhibitor or LDHA knockdown led to increased 5-Fu sensitivity of cervical cancer cells. Finally, we examined whether the NEAT1-promoted 5-Fu resistance was through miR34a/LDHA pathway. In NEAT1 knockdown cells, the cervical cancer cells showed increased 5-Fu sensitivity. Expectedly, co-transfection of NEAT1 shRNA and miR-34a inhibitor in cervical cancer cells showed recovered glycolysis and 5-Fu resistance, suggesting that NEAT1 suppressed miR-34a to promote the glycolysis and 5-Fu resistance in cervical cancer.
In summary, we propose that the negative association between NEAT1 and miR-34a in cervical cancer cells contributes to the dysregulated cellular glycolysis, resulting in 5-Fu resistance.
Although direct targeting LDHA by miR-34a has been revealed in other cancers, in speaking of lncRNA NEAT1, its function as a ceRNA of miR-34a to further de-suppress LDHA in cervical cancer has not been elucidated. Our future studies will aim to investigate the above molecular pathway in an in vivo animal model. Our findings suggest that targeting the NEAT1-mediated miR-34a/LDHA-glycolysis axis might be regarded as a promising therapeutic strategy against chemoresistant cervical cancer.   GAPDH and U6 were used as internal control for lncRNA and miRNAs, respectively. All data were shown as mean ± S.D. *, p<0.05; **, p<0.01; ***.