TAZ maintains telomere length in TNBC cells by mediating Rad51C expression

All cell lines were purchased from the Cell Bank of Chinese Academy of Sciences. MCF7, MDA-MB-453 and MDA-MB-231 breast cancer cells were cultured in High Glucose DMEM (#SH30022.01, Hyclone) supplemented with 10% FBS (#SH30084.03, Hyclone). T47D, HCC1806 and BT549 were cultured in RPMI-1640 medium (#SH30027.01, Hyclone) with 10% FBS. BT549 cells were additionally supplemented with 1 μg/ml insulin (#I9278, sigma). MCF10A was maintained in DMEM F12 medium (#SH30023.01, Hyclone) supplemented with 5% horse serum (#26050088, Invitrogen), 0.5 μg/ml hydrocortisone (#386698, sigma), 10 μg/ml insulin, and 20 ng/ml recombinant human EGF (#100-47, PeproTech). All cells were incubated in 5% CO₂ at 37 °C.

Two different siRNAs specifically targeting TAZ were designed and synthesized by GenePharma company (Shanghai, China). pcDNA 3.1(+)-TAZ and pcDNA 3.1(-)-TAZ-4SA overexpression plasmids were kindly provided by Dr. Jianmin Zhang from the Department of Cancer Genetics & Genomics, Roswell Park Cancer Institute, Buffalo, USA. Transfections of siRNAs or plasmids by using Lipofectamine 2000 (#11668-019, Invitrogene) were performed according to the manufacturer’s protocol. To generate the TAZ knockdown stable transfectants, TAZ-specific lentiviral shRNAs inserted into the GV112 vector (hU6-MCS-CMV-Puromycin) were designed and constructed by GenePharma company (Shanghai, China), and cells were transfected in the presence of 5 μg/ml polybrene according to the manufacturer’s instruction. The sequences of the siRNA and shRNA oligonucleotides are provided in Additional file 1: Table S1.

Cells were lyzed using RIPA buffer (#P0013C, Beyotime, Shanghai, China) supplemented with Protease and Phosphatase Inhibitors (#A32959, Thermo). Proteins were transferred to PVDF membranes after being separated in SDS/PAGE. The primary antibodies were incubated overnight at 4 °C, and the HRP-conjugated secondary antibodies were incubated for 1–2 h at room temperature. Final detection was performed by using ECL Plus Western Blotting Detection Reagents (#WBULS0500, Millipore). Antibodies against TAZ (#4883), β-catenin (#8480), p-ATR (#2853), p-ATM (#5883), p-BRCA1 (#9009), p-CHK1 (#2348), p-CHK2 (Thr1079, #8654) were purchased from Cell Signaling Technology (Beverly, MA, USA). Antibodies against p21 (#sc-6246), hTERT (#sc-393013), Rad51C (#sc-56214) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Antibodies against TRF1 (#ab129177), TRF2 (#ab108997), POT1 (#ab124784), TIN2 (#ab197894), RAP1 (#ab175329) were obtained from Abcam (Cambridge, MA, USA). Antibodies against GAPDH (#HRP-6004), β-actin (#HRP-60008) were obtained from Proteintech Group Inc. (Wuhan, China).

For relative telomere length analysis, genomic DNA was isolated by using the Tissue DNA Purification Kit (#D3396-01, Omega) and used as the template of qRCR reaction to determine the relative telomere length.

For mRNA expression analysis (RT-PCR), total RNAs were isolated by using the RNA Fast 200 isolation kit (#220010, Feijie, Shanghai, China) and reversely transcribed into cDNA by using the cDNA Reverse Transcription Premix (#2641A, Takara) according to the manufacture’s instruction. 1 μl of cDNA was then used for qPCR analyses.

For the analysis of TERRAs expression, RNAs were isolated by using the universal RNA extraction Kit (#9767, Takara). The isolated RNAs were then reversely transcribed by using the complement telomere repeats (CCCTAA)5 as the primers to detect the levels of TERRAs.

qPCR mixtures were prepared by using SYBR qPCR Premix (#RR420L, Takara). qPCR analyses were then performed with the Bio-Rad CFX96TM Real-Time PCR detection system according to the instruction of the qPCR Premix manual. Relative telomere length was normalized to single-copy gene HBG. Fold changes were determined by the ΔΔCt method. mRNA expression was normalized to GAPDH levels and also determined by the ΔΔCt method. All primers are sythesized by Shenggong company (Shanghai, China), and the sequences are listed in Additional file 1: Table S2.

Telomere length was measured by using the Telo TAGGG Telomere Length Assay kit (#12209136001, Roche) according to the manufacturer’s protocol. Briefly, genomic DNA was digested using Hinf I and Rsa I. The terminal restriction fragments (TRFs) were separated by 0.8% agarose gel electrophoresis at 60 V for 4 h and then capillary transferred to the positively charged nylon membrane. DIG-labeled telomere-specific probe was used to hybridization with the TRFs. The DIG-specific antibody coupled to alkaline phosphate was used for chemiluminescence detection.

Cell viability was assessed by the MTT assay. Briefly, equal amount of cells were seeded into a 48-well plate. At specific time points after seeding, 50 μl MTT (5 mg/ml) was added into the culture medium and incubated for 4 h at 37 °C, and then the mediums were removed. The generated formazan crystals were dissolved in 375 μl DMSO, and the optical absorbance values were determined at 490 nm by using a microplate reader (PerkinElmer).

Cells were fixed in cold 70% ethanol at 4 °C overnight. Cells were incubated with 50 μg/mL sodium citrate and 10 μg/mL RNase A in the dark for 30 min. FACS Calibur flow cytometer (BD Biosciences) was used to determine the DNA content. The data were analyzed using the MODFIT software program (Verity Software House).

BrdU assay was performed according to the protocol provided by BioLegend. Briefly, 10 µM BrdU was added to actively dividing cells for 45 min. Cells were fixed in 70% ethanol and incubated with 2N HCl and 0.1M Na₂B₄O₇. Anti-BrdU antibodies (#364103, BioLegend) were then added and incubated for 20 min at room temperature. FACS analysis was performed with a Becton Dickinson Canto instrument (BD Biosciences) to determine the incorporated BrdU levels.

Cell senescence was measured by using a SA-β-gal staining kit (Genemed, Beijing, China). The experiment was performed according to the product manual. Briefly,cells were fixed and stained with SA-β-gal staining solution overnight. Cells were examined under a light microscope. The number of the SA-β-gal-positive blue cells was counted, and the percentage of the SA-β-gal-positive cells versus total cells was calculated.

For IF analysis, cells were seeded on coverslips and fixed in 4% paraformaldehyde for 10 min. 0.2% Triton X-100 was used to permeabilize the membranes. Cells were then incubated with blocking solution for 1 h, followed by incubating with primary antibodies overnight at 4 °C. Antibodies against TRF2 (#ab108997, Abcam), p-BRCA1 (#9009, Cell Signaling Technology), 53BP1 (#NB100-304, Novus), γ-H2AX (#05-636, Merck Millipore Corporation), p53(#sc-126, Santa Cruz Biotechnology) were used as primary antibodies. Alexa fluor 488-labeled or Cy3-labeled secondary antibodies (Zhuangzhi Bio, China) were further incubated for another 2 h. Cells were then counterstained with DAPI and imaged with the laser scanning confocal microscope (Leica).

For IF-FISH, FISH was performed after IF staining described above. Briefly, after secondary antibodies incubation, cells were washed with PBS for 3 times. 4% paraformaldehyde containing 0.1% Triton X-100 was added to re-fix cells. Cells were then incubated with 10 mM glycine for 30 min. The telomeric PNA probes (PANAGENE) were added in a hybridization buffer containing 70% formamide, 1 mg/ml blocking reagent (Roche), 10 mM Tris–HCl pH 7.4, and denatured at 80 °C for 5 m. The probes were then hybridized in a humidified chamber for 2 h at room temperature. After washing the coverslips with wash solution 1 (containing 70% formamide, 0.1% BSA, 10 mM Tris–HCl pH7.4) and wash solution 2 (containing 100 mM Tris–HCl pH7.4, 150 mM NaCl, Tween-20), cells were counterstained with DAPI and imaged with the laser scanning confocal microscope (Leica).

Telomerase activity was measured by using TRAPEZE Telomerase Detection Kit (#S7700, Millipore) according to the manufacturer’s protocol. Briefly, 10⁶ cells were resuspended in 200 μL of 1 × CHAPS lysis buffer and incubated on ice for 30 min. Cell extracts were then diluted in a ratio of 1:5. 2 μl of the diluted cell extracts together with TRAP Reaction Buffer, dNTP Mix, TS Primer, TRAP Primer Mix, and Taq polymerase were prepared and subjected to PCR reaction. The PCR program is designed following the manufacturer’s instruction. The PCR products were separated with 15% non-denaturing PAGE and detected with silver staining.

ChIP was performed following the manufacture’s instruction of MegnaChIP A/G kit (#17-10085, Merk). Briefly, 5 × 10⁶ cells were cross-linked with 1% formaldehyde for 10 min at room temperature, and then quenched with glycine for another 5 min. Cells were lysed and sonicated by using Bioruptor (#UCD-200, Diagenode) with a condition of “30 s ON and 30 s OFF” for 8 cycles at high power setting. Sheared chromatin was incubated with protein A/G magnetic beads and antibodies against p53 (#sc-126X, Santa Cruz Biotechnology) or TEAD4 (ab58310, Abcam) and a mouse or rabbit IgG at 4 °C over night. Genomic DNA was purified after reverse-crosslinking. ChIP DNA was further analyzed by qPCR with primers specific to the p53 binding region of 13q-TERRA promoter or Tead4 binding region of Rad51C promoter. The sequences are listed. 13q-ChIP-Forward: 5′-GACAAGCCCCAGCCCATA-3′, 13q-ChIP-Reverse: 5′-GGGATAGTTAAGGAAGGATCTCCAA-3′, Rad51C-ChIP-Forward: 5′-CCGTTACGCTCACTACTCCC-3′, Rad51C-ChIP-Reverse: 5′-CGGAGACCATAGGACGGAGA-3′.

Statistic analysis was performed using SPSS Version 18.0. All results are presented as mean ± SEM. Shapiro–Wilk test was used for normality test. The data were considered to be normally distributed while P > 0.1, and all data were normally distributed. Statistical analyses were performed with unpaired Student's t-test between two groups and one-way ANOVA followed by Tukey’s multiple-comparisons for multiple groups. Two-way ANOVA followed by Bonferroni post test were used for MTT experiment analysis. All data were derived from at least three independent experimental replicates. P value < 0.05 was considered as statistically significant.

To know the potential roles of TAZ in the regulation of TNBC, we firstly examined the expressional levels of TAZ in both TNBC and non-TNBC. The bioinformatic analyses revealed that the mRNA levels of TAZ are severely higher in both TNBC patients and cell lines than that in non-TNBC (Additional file 1: Fig. S1A and B). These results are consistent to the previous studies that TAZ mRNA and protein expression are preferentially higher in TNBC than in other non-TNBC subclasses [16, 17]. Next, we verified the expressional changes of TAZ between TNBC and non-TNBC by real-time RT-PCR and Western blot analyses in several breast cancer cell lines. In accordance with the data from the above bioinformatic analyses, we found that both the mRNA and protein levels of TAZ are much higher in TNBC cell lines than that in non-TNBC (Fig. 1A, B). Furthermore, copy number variation (CNV) analysis revealed that gain of copy numbers of TAZ gene occurs more frequently in TNBC patients than that in non-TNBC patients (Additional file 1: Fig. S1C). Together, our results indicated that the expressional levels of TAZ are highly upregulated in TNBC. To understand the biological impacts of TAZ in TNBC, we knocked down the expression of TAZ in MDA-MB-231 TNBC cells by distinct shRNAs (Fig. 1C). We found that loss of TAZ significantly prohibits the proliferation of MDA-MB-231 cells (Fig. 1D). The data of cell cycle assay showed that loss of TAZ results in an obvious accumulation of cells in the G0/G1-phase and a decrease of cells in the S-phase. (Fig. 1E). BrdU incorporation assays further confirmed the observation that downregulation of TAZ causes an impaired proliferation activity (Fig. 1F). Additionally, we also found that depletion of TAZ causes significant higher levels of senescent cells (Fig. 1G), upregulation of p21 (Fig. 1C), and increased expressions of senescence-associated cytokines (Fig. 1H), suggesting that an elevated cell senescence is occurred in TAZ-depleted TNBC cells. Taken together, these data suggested that knockdown of TAZ expression inhibits cell proliferation and triggers cell senescence in TNBC cells.

It has been reported that cell senescence is tightly related to shortened telomeres [18]. Together with the findings that loss of TAZ results in cell senescence in TNBC cells (Fig. 1C–H), we therefore suspected that TAZ may affect telomere length in TNBC cells. Intriguingly, telomere-specific qPCR analysis revealed that knockdown of TAZ expression in MDA-MB-231 and BT549 TNBC cells (Fig. 2A, D) leads to a significant decrease of telomere length in both cell lines (Fig. 2B, E). We further confirmed the shortened telomere length in TAZ-depleted cells with Southern analysis of terminal restriction fragments. The terminal restriction fragments were generated by digesting genomic DNA with Hinf I and Rsa I, and detected with telomere specific probe. Knockdown of TAZ causes shorter telomere length characterized by the decrease in TRF length (Fig. 2C, F). We next investigated the effect of TAZ overexpression on telomere length. The plasmids encoding a Flag-tagged wild-type TAZ (Flag-TAZ) and a Flag-tagged TAZ mutant that permanently unphosphorylated (FLAG-TAZ-4SA), which indicates the constitutively activated (CA) form of TAZ, were transfected into BT549 cells (Fig. 2G). Telomere length was then tested, and our results revealed that no obvious changes of the telomere length was observed in both types of TAZ-overexpressing cells (wild-type and the CA mutant) (Fig. 2H). We also overexpressed the wild-type and the CA mutant TAZ in T47D cells which are non-TNBC cells with low TAZ expression and in MCF10A cells which are non-transformed mammary epithelial cells (Fig. 2I, K). No obvious changes of telomere length was observed (Fig. 2J, L). These results suggested that the inhibition of TAZ shortens telomeres in TNBC cells, while the activation of TAZ has no effect on telomere length in both TNBC and non-TNBC cells.

Previous studies indicated that hTERT is essential for the maintenance of telomere length [19]. Together with our findings that loss of TAZ results in significantly shortened telomeres in TNBC cells (Fig. 2), we therefore attempted to test whether TAZ affects the expression of hTERT. To our surprise, not decrease, our results revealed that loss of TAZ results in elevated mRNA and protein levels of hTERT in both MDA-MB-231 and BT549 TNBC cells (Fig. 3A–D). It has been reported that the ~ 2900 bp upstream region containing ~ 300 bp core nucleotides functions as the promoter of hTERT [20, 21]. We therefore cloned these two nucleotides regions into a luciferase reporter construct and performed dual luciferase reporter assays to indicate the effect of TAZ on hTERT promoter activity. In accordance to the effect on hTERT expression, our results indicated that loss of TAZ promotes the luciferase activity suggesting that an elevated hTERT promoter activity occurs under TAZ depletion (Fig. 3E). Overexpression of wild-type TAZ and its CA mutant result in no obvious change of hTERT level in MCF10A non-transformed mammary epithelial cells (Fig. 3F). Terc is an RNA molecule that provides the template for telomerase to add repeated DNA sequence to the ends of telomeres. We therefore tested the levels Terc after TAZ depletion, and no obvious changes was observed in TNBC cells (Fig. 3G, H). In addition, our telomeric repeat amplification protocol (TRAP) assay indicated no significant change of telomerase activity in TAZ-depleted cells (Fig. 3I), further supporting that the changes of hTERT are not the cause of TAZ depletion-induced telomere shortening. Together, these data indicated that knockdown of TAZ upregulates the expression of hTERT in TNBC cells.

Accidentally, we found that knocking down the expression of TAZ with two TAZ-specific siRNAs shows a different effect on hTERT mRNA levels compared with knocking down by using TAZ-specific shRNA. In contrast to the effect of TAZ-specific shRNA (Fig. 3A–D), knockdown of TAZ expression in TNBC cells with TAZ-specific siRNAs transfected for 48 h (Additional file 1: Fig. S2A and B) does not show obvious effect on hTERT mRNA levels (Additional file 1: Fig. S2C and D) and protein levels (Additional file 1: Fig. S2E and F). We therefore tested whether the siRNA-mediated TAZ knockdown can also result in shortened telomere length as detected in TNBC cells with TAZ knockdown by using TAZ-specific shRNAs (Fig. 2). Intriguingly, depletion of TAZ by siRNAs indeed leads to a shortened telomere length phenotype (Additional file 1: Fig. S2G–J), further supporting that hTERT is likely not involved in TAZ-mediated telomere length regulation.

As cells transfected with siRNAs were usually harvested for examination at 48 h after transfection (short-term), while cells transfected with shRNAs were passaged for more than six passages (long-term) when subjected to the related analyses, we therefore speculated that long-term knockdown of TAZ may cause more extreme telomere shortening. We next examined the dynamics of telomere length changes after TAZ knockdown during different passages. As expected, telomere length indeed shows more severe shortening phenotype during prolonged passages (Additional file 1: Fig. S2K), and short time (48 h) depletion of TAZ with shRNAs also shows no obvious effect on hTERT mRNA levels, similar to the effect of siRNAs (Additional file 1: Fig. S2L).

As mentioned above, shelterin complex is tightly involved in the regulation of telomere length by directly binding to telomere DNA to protect telomeres [22]. Therefore, we examined the effect of TAZ depletion by both short-term and long-term approaches on the protein levels of shelterin components. Intriguingly, similar to the effect on hTERT expression, short-term knockdown and long-term knockdown of TAZ expression also showed different impacts on the protein levels of shelterin components. We found that short-term depletion of TAZ showed no obvious effect on the protein levels of the shelterin components (Additional file 1: Fig. S3). However, by contrast, long-term knockdown of TAZ expression resulted in significant decreases of the protein levels of most of the shelterin components except TRF1 and RAP1 (Fig. 4A, B), suggesting that an impaired shelterin complex organization is likely occurred in long-term TAZ-depleted TNBC cells. To further confirm this observation, we performed an IF assay to detect the levels of TRF2. The levels of TRF2 are strongly decreased in long-term TAZ-knocked down TNBC cells (Fig. 4C).

As it has been reported that shelterin complex prevents DDR activation through protecting telomeres[23], and our findings that long-term loss of TAZ downregulates the protein levels of multiple shelterin components (Fig. 4A, B), we therefore examined the activity of the DDR pathway upon knockdown of TAZ expression. As expected, most of the molecular markers of the DDR pathway are upregulated after long-term depletion of TAZ expression in TNBC cells suggesting that knockdown of TAZ expression with long-term approaches activates the DDR pathway (Fig. 4D). Moreover, our IF analysis with phosphorylated BRCA1 antibodies (anti-p-BRCA1) also supports this observation as elevated p-BRCA1 signals were detected after TAZ depletion in BT549 cells with a long-term approach (Fig. 4E). We next checked the telomere-induced DNA damage foci (TIF) in both TAZ-deficient and control cells. We found that the percentage of cells with more than 3 TIFs are elevated in TAZ-depleted cells (Fig. 4F, G). Taken together, these data suggested that long-term depletion of TAZ affects shelterin complex and activates the DDR pathway. Therefore, TAZ positively regulates telomere integrity in TNBC cells.

To explore the potential mechanisms of how TAZ affects telomere length, we determined to test the levels of TERRAs expression, as TERRAs have been reported to regulate telomere length [15, 24]. Therefore, we firstly examined whether loss of TAZ affects the expression levels of different TERRA family members. We detected that depletion of TAZ with both short-term and long-term approaches in TNBC cells all result in upregulated expressions of multiple members of TERRAs (Additional file 1: Fig. S4A–D). These results suggested that loss of TAZ promotes the expressions of TERRA family members, which occurs in both short-term and long-term TAZ depletion conditions, and may contribute to the detected telomere length shortening induced by TAZ depletion.

Rad51C participates in homologous recombination. Loss of Rad51C in both telomerase positive and negative MEF cells causes rapid loss of telomere sequences and breakage of telomere integrity [25]. This Rad51C depletion-induced phenotype in telomere is similar to what we observed in TAZ-knocked down cells. We therefore tested Rad51C levels in TAZ-depleted and control cells. Loss of TAZ results in reduced mRNA and protein levels of Rad51C in both MDA-MB-231 and BT549 TNBC cells (Fig. 5A, B). In addition, knocking down the expression of Rad51C in BT549 cells (Fig. 5C) leads to a significant decrease of telomere length (Fig. 5D) and increased cells with more than 3 telomere-induced DNA damage foci (Fig. 5E). Similar with TAZ overexpression, increased Rad51C expression showed no obvious changes in telomere length (Fig. 5F, G). Importantly, the telomere shortening phenotype caused by TAZ depletion is rescued by Rad51C overexpression in both MDA-MB-231 and BT549 cells (Fig. 5H–J). Together, these data suggested that loss of TAZ inhibits the transcription of Rad51C and further destroys the telomere length homeostasis and telomere integrity. Notably, overexpression of Rad51C in TAZ-depleted cells did not fully restore the telomere length to the levels of control cells (Fig. 5I, J), suggesting that additional mechanisms may also be involved in TAZ-regulated telomere length dynamics. As previously observed, increased levels of TERRAs may be an alternative mechanism causing shortened telomeres in TAZ-deficient cells. TAZ is a transcriptional co-activator that regulates transcription of target genes by binding to transcriptional factors [3]. TEAD family members (TEAD1-4) are the most common TAZ binding partners [26]. We analyzed the TEAD4 chromatin immunoprecipitation followed by high-throughput sequencing (ChIP-seq) datasets with the ENCODE database (https://​www.​encodeproject.​org) to determine whether TAZ-TEAD directly regulate Rad51C transcription. We found that, in different types of cells (A549-lung cancer, H1-human embryonic stem cells, HCT-116-colon cancer, MCF-7-breast cancer, SK-N-SH-neuroblastoma), TEAD4 directly binds to the promoter regions of RAD51C gene (Fig. 6A). However, loss of TAZ expression did not affect the TEAD4 protein levels (Fig. 6B). We next performed ChIP-qPCR analyses in BT549 cells to validate the binding of TEAD4 to RAD51C promoter. We found that TEAD4 is indeed enriched on the promoter region of the Rad51C gene, and the binding is reduced by knocking down the expression of TAZ (Fig. 6C). RAD51C promoter luciferase assays also showed a reduction of luciferase activity in TAZ-depleted cells (Fig. 6D). Together, these results provide evidence that a TAZ-TEAD transcriptional program regulates Rad51 expression in TNBC cells.

From current study, we identified that, in TNBC cells, TAZ maintains telomere length by acting as a transcriptional co-activator of Rad51C. Knockdown of TAZ expression abolishes TAZ-TEAD binding to the promoter of Rad51C, and further impairs Rad51C transcription. Lack of Rad51C causes the telomere shortening and impaired telomere functions. The protein levels of shelterin components is decreased, along with increased telomere-induced DNA damage foci (TIF) and DDR activation. Moreover, high levels of TERRAs in TAZ-deficient cells may also contribute to the observed telomere loss phenotype (Fig. 7).