TMEM16A alternative splicing coordination in breast cancer
© Ubby et al.; licensee BioMed Central Ltd. 2013
Received: 8 April 2013
Accepted: 11 July 2013
Published: 16 July 2013
Skip to main content
© Ubby et al.; licensee BioMed Central Ltd. 2013
Received: 8 April 2013
Accepted: 11 July 2013
Published: 16 July 2013
TMEM16A, also known as Anoctamin-1, is a calcium-activated chloride channel gene overexpressed in many tumors. The role of TMEM16A in cancer is not completely understood and no data are available regarding the potential tumorigenic properties of the multiple isoforms generated by alternative splicing (AS).
We evaluated TMEM16A AS pattern, isoforms distribution and Splicing Coordination (SC), in normal tissues and breast cancers, through a semi-quantitative PCR-assay that amplifies transcripts across three AS exons, 6b, 13 and 15.
In breast cancer, we did not observe an association either to AS of individual exons or to specific TMEM16A isoforms, and induced expression of the most common isoforms present in tumors in the HEK293 Flp-In Tet-ON system had no effect on cellular proliferation and migration. The analysis of splicing coordination, a mechanism that regulates AS of distant exons, showed a preferential association of exon 6b and 15 in several normal tissues and tumors: isoforms that predominantly include exon 6b tend to exclude exon 15 and vice versa. Interestingly, we found an increase in SC in breast tumors compared to matched normal tissues.
As the different TMEM16A isoforms do not affect proliferation or migration and do not associate with tumors, our results suggest that the resulting channel activities are not directly involved in cell growth and motility. Conversely, the observed increase in SC in breast tumors suggests that the maintenance of the regulatory mechanism that coordinates distant alternative spliced exons in multiple genes other than TMEM16A is necessary for cancer cell viability.
TMEM16A (also known as ANO1, DOG1, ORAOV2 or TAOS2) is a member of the Anoctamin family of membrane proteins, which consists of ten components (known as TMEM16A-K or ANO1-10) that share a highly conserved structure with eight transmembrane domains and cytosolic amino- and carboxy-termini domains. In 2008, three independent groups identified TMEM16A as a bona fide Ca2+-activated Cl- channel (CaCC) essential for a variety of physiological functions including neuronal and cardiac excitation, olfactory and sensory signal transduction, vascular tone and pain perception [1, 2]. TMEM16A is expressed in many tissues that are known to express CaCCs: bronchiolar epithelial cells, pancreatic acinar cells, proximal kidney tubule epithelium, retina, dorsal root ganglion sensory neurons, and sub- mandibular gland [3–5]. In addition to epithelia, TMEM16A is robustly expressed in interstitial cells of Cajal, which are responsible for generating pacemaker activity in smooth muscle of the gut [6–8].
TMEM16A is a sensitive biomarker for the diagnosis of gastrointestinal stromal tumors (GISTs) and it is overexpressed in several cancers including urinary bladder cancer, esophageal cancer, prostate cancer, breast cancer, head and neck squamous cell carcinoma (HNSCC), parathyroid tumors and ovarian tumors [9–13]. However, its role in tumor cell proliferation and migration is not completely understood. In cellular models, TMEM16A has been reported to positively affect cell proliferation and/or migration [9, 12–16]. Conversely, Wang et al.  showed a negative effect of cell proliferation by preventing cell cycle transition from the G0/G1 phase to the S phase via inhibition of cyclin D1 and cyclin E expression . Several mechanisms have been reported that associate changes in Ca2+-dependent channel activities with its tumorigenic potential are unknown: prevention of cell cycle transition from the G0/G1 phase to the S phase via inhibition of cyclin D1 and cyclin E expression , activation of MAPK kinases with promotion of ERK1/2 and cyclin D1 activation  and activation of EGFR and CAMK signalling . In general, similar to other channels such as Ca(2+)-activated K(+) channels and voltage-gated Cl(−) or K(+)channels, TMEM16A might affect cell motility and proliferation though dynamic changes in the cellular volume [18–21].
Alternative splicing of TMEM16A can generate multiple protein isoforms with different electrophysiological properties affecting the voltage and Ca2+-dependence of the channel . The most important AS occurring on TMEM16A transcripts consist in skipping/ inclusion of three exons 6b, 13 and 15. Other minor splicing events affecting exon 1, 10, 14 and 18 have also been reported, leading to the formation of several TMEM16A transcripts [23, 24]. We have previously reported the relative percentage of inclusion of each exon relative to the total amount of TMEM16A transcripts showing that in several normal human tissues, transcripts that predominantly include exon 6b, also tend to lack exon 15 in the same mature transcripts and vice-versa . However, the relative proportion of each isoform that take into account the real association between the AS events on the same transcript is unknown.
In the TMEM16A gene and thus on the nascent transcripts, the three exons, 6b, 13 and 15 are located relatively far apart (exons 6b and 15 are located at 50 kb). Comparison of data from transcript sequencing efforts, EST/cDNA sequences and microarray profiling experiments have provided evidence for AS coordination between multiple exons within a single gene [25–28]. Indeed, distant alternative splicing events on the same gene can be correlated, through a mechanism known as intragenic Splicing Coordination (SC) [27, 29, 30]. The presence of multiple splice variants on the same transcript, as found in TMEM16A, can highly increase its potential to generate multiple transcripts. However, as AS events are generally analyzed separately through PCR amplification of short regions, this correlation is difficult to establish and have not been explored in tumors. The mechanism that regulates the coordinate selection of AS events in the same gene involves both cis-acting regulatory elements present on the nascent transcripts and Polymerase II -dependent transcriptional elongation .
In the present study, using a PCR-assay that amplifies transcripts across the three AS exons, we found that TMEM16A AS of exons 6b and 15 are coordinated in several normal tissues and that this coordination increases in breast tumors. AS of individual exons was not altered in breast tumor and overexpression of the TMEM16A isoforms in a controlled cellular system had no effect on proliferation and migration. These results suggest that TMEM16A expression and splicing does not confer a positive general effect on cell growth and motility. The unexpected increase of SC in breast tumors suggest that cancer cells maintain the regulatory mechanism that coordinates distant alternative spliced exons on the same transcript, a process that, by acting on multiple genes, may be important in tumor cell viability.
In this paper, we have evaluated the potential role in breast cancer of the different TMEM16A isoforms generated through alternative splicing. The analysis of the alternative splicing pattern in vivo in breast tumors and the functional analysis of the isoforms overexpressed in a cellular model strongly suggest that the TMEM16A Ca2+-dependent Cl- channel activities are not directly involved in tumorigenesis. More precisely, the lack of effect on HEK293 growth and motility indicates that, if TMEM16A has a role on cancer progression, it is not due to a general effect on the regulation of cell cycle and migration, as previously proposed [9, 12–15]. However, this does not exclude that TMEM16A has a more cancer-specific role that is relevant to tumor growth in vivo. On the other hand, in TMEM16A transcripts derived from several normal tissues and breast tumors, inclusion/exclusion of the two distant alternatively spliced exon 6b and 15 is not independent but occur in a coordinated manner on a single transcript. As this splicing coordination is significantly increased in breast tumors, the regulatory mechanism that coordinates distant alternative spliced exons on the same transcript is maintained in these cells. We speculate that maintenance of the mechanisms that control SC is necessary for regulation of AS events in other genes directly involved in tumor cell viability.
Several studies have associated TMEM16A expression with cancer [9, 31–33]. However, as tumors may originate from cells that normally express TMEM16A, like GIST, which is thought to derive from the interstitial cells of Cajal [33, 34], it is not clear if this channel is just an associated marker or if it directly promotes tumorigenesis. In addition, the mechanism that associates its Ca2+-dependent Cl- activity to cellular proliferation is unclear, and several studies reported contradictory effects on proliferation and/or migration. In basilar smooth muscle cells, where TMEM16A is abundantly expressed, its downregulation promotes proliferation and accordingly restoration of TMEM16A activity was suggested to be beneficial on hypertension-associated cardiovascular disease such as stroke . On the contrary, other studies reported that TMEM16A promotes cell migration alone  or both proliferation and migration [9, 12, 14, 15]. In xenografts, loss of TMEM16A through siRNA-mediated silencing was reported to inhibit tumor growth [9, 13]. As a result of this positive effect on proliferation, and in contrast to the effect on smooth muscle cells , the therapeutic inhibition of TMEM16A activity was proposed for the treatment of cancers [9, 13]. In this study, the analysis of the different TMEM16A isoforms generated through AS in vivo in breast cancer and in a controlled cellular model, does not support a direct role of the major TMEM16A isoforms in tumorigenesis. The evaluation of the different functional properties of isoforms derived from AS is not an easy task, in particular if referred to migration and proliferation. To avoid clonal variability that can affect transgene expression due to the numbers of copies integrated and site of integration, we used the inducible HEK293 Flp-In Tet-ON system and the results clearly show that the TMEM16A isoforms per se are not sufficient to promote or inhibit proliferation and migration (Figure 4). Even if our result does not support a direct role of TMEM16A, the different isoforms might indirectly affect proliferation and migration in a cell-type specific manner. Accordingly, to affect proliferation and/or migration, TMEM16A expression would require expression of additional factors, and this association could be at the basis of the contradictory effects reported in different cell types. This cell type specificity might be critical for the development of therapeutic strategies, as different pathological conditions and target cells may differently respond to inhibition or restoration of the TMEM16A activity, as previously reported [9, 13, 17]. The HEK293 Flp-In Tet-ON system we have developed here, expressing the different isoforms in a regulated manner, can represent an interesting model to identify, with a high-throughput screening method, those factors that not only promote or inhibit proliferation in a TMEM-dependent manner but also potential AS-specific networks.
An interesting and novel aspect of our study is the identification in several normal tissues of TMEM16A SC and the observation that this coordination increases in breast tumors (Figure 6). In common with ~ 25% of human genes, TMEM16A has more than one alternative splicing possibility and accordingly can generate several isoforms. Through a mechanism known as intragenic splicing coordination, [27, 28, 36] the association between alternative spliced events is not random, as we have found for exon 6b and 15 in TMEM16A. In normal breast, approximately 50% of samples showed SC, which increased to 84% in tumors. Normal mammary glands express TMEM16A[37–39] and the individual hormonal status might have an important effect on its AS pattern and SC thus explaining the variability observed in normal tissues. In several genes, hormones modulate AS [40, 41] but the effect on splicing coordination is unknown. Through histological evaluation, we do not have evidence of the presence of tumor cells in normal samples, but normal glands showed a variable relative abundance of different cell types (like epithelial, stromal and adipocytes) and this might also contribute to the individual variability.
The fact that the expression of the two AS exons 6b and 15 is coordinated in the majority of TMEM16A transcripts derived from tumors (84%), suggest that cancer progression is not associated with a relaxation of this phenomenon. The mechanisms underlying splicing coordination are largely unknown. Fibronectin (FN)  and slo-1 BK potassium channel  are well-studied examples of genes with intragenic AS coordination. In these cases, mutations that affect one AS exon had a profound effect on the other AS events, with both 5’ to 3’ polarity or bi-directionality for FN and Slo-1, respectively. In addition, proper coordination of intragenic alternative splicing has been found essential for normal physiology of the slo-1 gene in vivo in Caenorhabditis elegans. Based on this evidences a mechanism that lead to preferential expression of given alternate exon combination was engaged. Recruitment of splicing factors with direct interactions between the RNA-protein complexes from distinct splice sites, RNA secondary structure determinants and changes in Pol II elongation or processivity have been suggested to be involved [27, 30]. Another interesting hypothesis to explain intragenic AS coordination might involve the formation of chromatin loops. Chromatin loops have been reported to physically link promoters to the end of the gene in order to facilitate Pol II-dependent transcription [42–45]. These loops can also occur in introns, as demonstrated for the 85-Kb long BRCA1 gene in human cell lines and in mouse mammary tissues . Interestingly, these loops change upon estrogen stimulation and during lactational development . Thus it is possible that loops that put in contact distant alternatively spliced regions on the same gene might have a role in splicing coordination and possibly regulated by hormonal status. In this manner, Pol II engaged in transcription at different AS exons might physically communicate with reciprocal influence on the corresponding splicing decisions. This event might be still operative in tumor cells in order to preserve cancer viability. More deep molecular studies are required to unravel the basic mechanism of intragenic splicing coordination and understand their role in cancer.
What could be the advantage of a tumor cell to maintain an intragenic splicing coordination? As none of the TMEM16A isoforms directly affect proliferation or migration (Figure 4), it is possible that during tumorigenesis the cell would have to maintain active and efficient the mechanism involved in splicing coordination. This might not be specifically useful for the TMEM16A expression but for coordinating splicing of other genes directly involved in proliferation or apoptosis such as CD44 gene, Ron gene or FGFR2 gene [46–49].
In conclusion, this study has improved our understanding of TMEM16A splicing coordination with the identification and characterization of a non-random distribution of the mRNA isoforms in normal adult human breast tissues and tumor.
In this context, the maintenance of splicing coordination will be required for preventing a massive transcript alteration that will lead to cell death and thus the study of the basic mechanism involved might be useful for the identification of novel targets for therapeutic intervention.
Surgically resected tissue samples were obtained under the General Surgery, “Presidio Ospedaliero di Gorizia, Italy approved by the local Ethics Committee. Primary breast tumors were obtained with the appropriate informed consent. Normal breast tissues and corresponding breast tumors were separated at the time of excision, and conserved in RNA Later (Ambion Inc) at – 20°C. The total RNA was extracted using TRIreagent solution (Ambion) followed by an additional cleaning step (RNeasy; Qiagen). RNA was quantified using the Nanodrop spectrophotometer instrument (Thermo Scientific) and quality routinely verified on denaturing agarose gels. Normal breast tissues and tumors were classified according to standard histological evaluations, receptor and proliferation markers as reported in Additional file 3: Table S3.
One μg of total RNA derived from ductal and lobular epithelial breast tumors and normal breast samples and one μg of total RNA derived from 20 human tissues (First Choice Human Total RNA Survey Panel; Ambion) was retro-transcribed in standard conditions and amplified by PCR with a set of primers specific for each alternatively spliced exon: for exon 6b, the sense and antisense primers were 803DFAM and 1385R, 5′-[6FAM]CGGAGCACGATTGTCTATGA and 5’- GGGCCATGAAGACAGAGAAG, respectively, for exon 13, 1368DFAM and 1525R, 5′-[6FAM]TCTCTGTCTTCATGGCCCTC and 5’-CTCCAAGACTCTGGCTTCGT, respectively, and for exon 15, 1506DFAM and 1894R, 5′-[6FAM]ACGAAGCCAGAGTCTTGGAG and 5′-GAACCGATCTCTCCATGTCAGCTTCA, respectively. Conditions for PCR were the following: 94°C for 5 min for the initial denaturation; 94°C for 45 s, 58°C for 45 s, and 72°C for 1 min for 35 cycles; and 72°C for 10 min for the final extension. Semi-quantitative analyses of the TMEM16A isoforms were performed by Capillary Electrophoresis (CE) at the BMR genomics, Italy (http://www.bmr-genomics.it/). The relative amount of each mRNA isoform was calculated based on the corrected integrated area of each peak and size was calculated using Rox1000BV marker (Bmr-Genomics). For Spicing Coordination experiments, RNA extracted from human adult normal tissues, normal and tumor breast tissues were amplified using a labelled fluorescent 803DFAM 5′-[6FAM]CGGAGCACGATTGTCTATGA and 1894R primer 5′-GAACCGATCTCTCCATGTCAGCTTCA. Conditions for PCR were the following: 94°C for 3 min for the initial denaturation, 94°C for 45 s, 56°C for 45 s, 72°C for 1 min and 72°C for 10 min for the final extension. 1 μL of the PCR product was dehydrated at the temperature of 60°C for 20 min and sent for the CE analysis. The results are expressed as mean ± S.D. of three independent experiments done in duplicate.
The TMEM16A isoforms 6b-13 + 15-, 6b + 13 + 15+ 6b + 13 + 15-, and 6b + 13-15- were obtained form corresponding plasmids [3, 22]. TMEM16A (6b-13 + 15+) was generated from the TMEM16A (6b + 13 + 15+, and 6b-13 + 15-) constructs by RT-PCR (forward) 5′-TGCGACAAGACCTGCAGCTACTGG-3′ and (reverse) 5′-TGTAGGAATTCAAACTTCAGCAG-3′ and cloned in the PstI and EcoRI of pBS(6b-13 + 15-) to generate pBS(6b-13 + 15+). Subsequently, pBS(6b-13 + 15+) was digested with EcoRI and BamHI and cloned in the corresponding sites of pcDNA3.1 to generate pcDNA3.1(6b-13 + 15+). All the coding sequences for TMEM16A (6b + 13-15-, 6b + 13 + 15-, 6b + 13 + 15+, 6b-13 + 15-, and 6b-13 + 15+) were cloned in the pcDNA5 FRT/TO plasmid (Invitrogen). Stable expression of TMEM16A variants was achieved by Flp-recombinase-mediated recombination in HEK293 Flp-In cells (Invitrogen) followed by hygromycin B selection (InvivoGen). Each TMEM16A-expressing vector that expresses the Flp-recombinase (Invitrogen) was co-transfected with Effectene transfection reagent (Qiagen) and selected with a concentration of 200 μ g/ml hygromycin B. Individual clones were obtained by limited dilution. Induction of TMEM16A isoforms expression was achieved with 0.1 μg/ml tetracycline (Sigma). Cells were grown in DMEM-Glutamax-I media (GIBCO) supplemented with 5% fetal bovine serum (EuroClone).
Cell mobility was assessed using a scratch wound healing assay. HEK293 stable cells were plated (5 × 105 cells) in 6-well plates, coated with poly-D-lysine (1 mg/ml, Sigma). 0.1 μg tetracycline (Sigma) was added one day after plating. Confluent monolayers (24 h after tetracycline induction) were scraped with 200 μl disposable plastic pipette tips. Pictures were taken every 24 h with a confocal microscope (Leica TCS-SL) (4 × amplification). The distances between the wound edges were measured using GraphicConverter software. These experiments were carried out in triplicates.
TMEM16A-inducible HEK293 cells were grown on 6-well plate and cultured in the presence or absence of 0.1 μg/ml tetracycline for 72 hrs to induce TMEM16A isoform expression. The cells were fixed with Bouin solution (Sigma-Aldrich) for 10 minutes at room temperature, washed twice in PBS, and incubated with blocking solution (1% BSA and 1% PBS) for 2 hrs at room temperature. The cells were then stained overnight at 4°C with primary monoclonal antibody against TMEM16A (Abcam) diluted in 0.3% Triton X-100 (Sigma-Aldrich) in 1%PBS-BSA. Cells were washed with PBS and incubated for 2 h with the respective secondary antibodies conjugated to Alexa Fluor-488 (Life Technologies). This was rinsed again in PBS, before viewing cells under a laser-scanning microscope LSM 5 EXCITER (Carl Zeiss MicroImaging).
HEK-293 cells, plated in 96-well microplates, were transiently transfected with a plasmid coding for the halidse-sensitive yellow fluorescent protein (HS-YFP) as previously described . After 24 hours, cells were treated with tetracycline to induce TMEM16A expression. At the time of assay, cells were washed two times with 100 μl PBS (137 mM NaCl, 2.7 mM KCl, 8.1 mM Na2HPO4, 1.5 mM KH2PO4, 1 mM CaCl2, and 0.5 mM MgCl2; pH = 7.4) and incubated for 30 min with 60 μl PBS. Cells were transferred to the stage of a fluorescence microscope equipped with excitation and emission filters for the yellow fluorescent protein. Cell fluorescence in each well was continuously acquired with a CoolSNAP HQ2 camera before and during addition of 165 μl of modified PBS (Cl- replaced by l-; final l- concentration in the well: 100 mM) also containing 1 μM ionomycin. Fluorescence cell traces were normalized to the initial, background-subtracted, fluorescence. TMEM16A activity was calculated from the maximal rate of fluorescence quenching caused by l- influx.
TMEM16A-inducible HEK293 cells, were grown on 96-well plate and cultured in the presence or absence of 0.1 μg/ml tetracycline. After 24 hours, cells were incubated with EdU (Invitrogen). 30 minutes after EdU addition, cells were washed once in PBS, fixed in PFA 4%, for 15 min, permeabilized with 0.5% Triton X-100 in PBS solution for 10 min. EdU detection was carried out using Click-iT EdU Cell Proliferation Assay (Invitrogen).
A statistical analysis was performed on the investigated groups of data using the Student’s t-test and the Fisher’s exact test.
The authors thank Dr. Miguel Mano for technical assistance in cellular proliferation experiments.
This work was supported by research grants from Associazione Italiana per la Ricerca sul Cancro (AIRC) 10387 and from Fondazione Italiana Ricerca Fibrosi Cistica.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.