PMEPA1/TMEPAI isoforms function via its PY and Smad‐interaction motifs for tumorigenic activities of breast cancer cells
Abstract
PMEPA1 (prostate transmembrane protein, androgen‐induced 1)/TMEPAI (transmembrane prostate androgen‐induced protein) is highly expressed in diverse cancers, including breast, lung and prostate cancers. It consists of four isoforms with distinct extracellular regions (isoforms a‐d). The expression and function of these isoforms are still poorly understood. Hence, we aimed to identify the preferentially expressed isoforms in breast cancer cells and analyze possible differences in tumorigenic functions. In this study, we used 5′ Rapid Amplification of cDNA Ends (RACE) and Western blot analyses to identify the mRNA variants and protein isoforms of TMEPAI and found that TMEPAI isoform d as the major isoform expressed by TGF‐β stimulation in breast cancer cells. We then generated CRISPR/Cas9‐mediated TMEPAI knockout (KO) breast cancer cell lines and used a lentiviral expression system to complement each isoform individually. Although there were no clear functional differences between isoforms, double PPxY (PY) motifs and a Smad‐interaction motif (SIM) of TMEPAI were both essential for colony and sphere formation. Collectively, our results provide a novel insight into TMEPAI isoforms in breast cancer cells and showed that coordination between double PY motifs and a SIM of TMEPAI are essential for colony and sphere formation but not for monolayer cell proliferation.
1 INTRODUCTION
Breast cancer is the most frequently diagnosed cancer in women worldwide (Tao et al., 2015), and it requires additional therapeutic strategies for effective treatments (Dai et al., 2015). Conventional therapeutic strategies of surgery in combination with radiation and pharmacotherapy (e.g., hormonal therapy, HER2‐targeted therapy and/or chemotherapy) have been largely used as first‐line treatments (Richie & Swanson, 2003). However, these treatments are of limited success in cases of the triple‐negative breast cancer (TNBC) (Irvin & Carey, 2008) as it does not respond to the current treatments with hormones or molecular targeted drugs (Pal & Mortimer, 2009). Thus, understanding the molecular flux of TNBC progression is urgently required to develop a novel molecular targeted therapy for TNBC patients (De Laurentiis et al., 2010).
Prostate transmembrane protein, androgen‐induced 1 (PMEPA1), also known as transmembrane prostate androgen‐induced protein (TMEPAI) or solid tumor‐associated gene 1 (STAG1), has been researched intensively because of its implication in multiple signaling pathways and human malignancies (Rae, Hooper, Nicol, & Clements, 2001; Xu et al., 2000). A lot of studies have been reported that TMEPAI can be induced by several cell signaling molecules such as androgen, TGF‐β, EGF, Wnt and mutated p53 (Anazawa, Arakawa, Nakagawa, & Nakamura, 2004; Azami, Vo Nguyen, Watanabe, & Kato, 2015; Giannini et al., 2003; Nakano et al., 2010; Xu et al., 2000). Sequence analysis of the PMEPA1/TMEPAI gene indicated that the TMEPAI gene encodes a type 1b transmembrane protein with a short extracellular region and a long cytoplasmic region (Xu et al., 2000). The NCBI database contains five variants of TMEPAI transcripts (variants 1–5) and four isoforms of TMEPAI protein (isoforms a‐d). The alternative first exons give rise to a variable extracellular region in each TMEPAI isoform except for isoform c, which lacks an extracellular region and half of transmembrane domain (Giannini et al., 2003; Rae et al., 2001; Xu et al., 2000). Nevertheless, all isoforms share a common intracellular region containing two PPxY (PY) motifs and a Smad‐interaction motif (SIM) (Giannini et al., 2003; Rae et al., 2001; Watanabe et al., 2010; Xu et al., 2000).
TMEPAI is highly expressed in many types of cancers, including prostate (Li et al., 2008; Xu et al., 2003) renal, stomach, rectal (Rae et al., 2001), ovarian (Giannini et al., 2003), colon (Brunschwig et al., 2003), lung (Vo Nguyen et al., 2014) and breast cancers (Cichon & Radisky, 2014; Singha et al., 2014; Singha, Yeh, Venkatachalam, & Saikumar, 2010). Moreover, Singha et al. (2014) demonstrated that high expression of TMEPAI in TNBC cells is associated with down‐regulation of PTEN and up‐regulation of PI3K/AKT signaling activation, leading to TNBC progression. In accordance with this finding, our previous study also demonstrated that knockdown (KD) of TMEPAI in lung cancer cells reduced their tumorigenic activity as indicated by reduced sphere‐forming ability in vitro and xenograft tumor forming ability in vivo (Vo Nguyen et al., 2014). Taken together, these results suggested important roles of TMEPAI in promoting tumorigenesis of cancer cells. However, the differences in expression and function of TMEPAI isoforms have not been fully addressed. Hence, we aimed to identify which isoforms of TMEPAI are expressed in breast cancer cells and analyze the possible differences among them with regard to tumorigenic functions. In conclusion, our results of 5′ Rapid Amplification of cDNA Ends (RACE) and Western blot analyses indicated that TMEPAI isoform d is the main isoform of TMEPAI in breast cancer cells. Although there were no significant functional differences among TMEPAI isoforms as indicated by complementing each TMEPAI isoform individually in CRISPR/Cas9‐induced TMEPAI knockout (KO) TNBC cells, we found that TMEPAI drives colony and sphere formation via both PY motifs‐ and a SIM‐mediated cellular signaling regulation.
2 RESULTS
2.1 TGF‐β stimulation causes expression of TMEPAI mRNA variant 5 and protein isoform d in breast cancer cells
We evaluated TMEPAI mRNA and protein expression levels in human breast cancer cell lines using quantitative real‐time PCR and Western blot analyses. In these experiments, we used four TNBC cell lines (BT‐549, HCC1395, Hs578T and MDA‐MB‐157) and two non‐TNBC cell lines (BT‐474 and SK‐BR‐3) with or without TGF‐β stimulation. Results showed that TGF‐β increased total TMEPAI expression at both the mRNA and the protein levels in both TNBC and non‐TNBC cell lines (Figure 1a,b). TMEPAI protein expression levels were relatively higher in TNBC cell lines, except for HCC1395 cells. The TMEPAI proteins induced by TGF‐β in these breast cancer cell lines mainly appeared as single similar molecular size bands (Figure 1b). The Smad2 and Smad3 phosphorylation levels were shown because of previous report indicating TMEPAI induction by TGF‐β‐/Smad3‐dependent pathway (Nakano et al., 2010). The results demonstrated that Smad3 phosphorylation levels were correlated with TMEPAI induction under the presence of TGF‐β stimulation in these breast cancer cell lines, although HCC 1395 cells lack both phosphorylation and Smad2/Smad3 protein expression which coincide with low expression of TMEPAI even after TGF‐β stimulation (Figure 1b). These results are consistent that TGF‐β/Smad3 signaling pathway mediates TMEPAI expression (Azami et al., 2015; Nakano et al., 2010). There are five mRNA variants and four protein isoforms of TMEPAI in the NCBI database. All TMEPAI mRNA variants differ in the usage of exon 1 resulting in a distinct N‐terminal sequence for each isoform (Figure S1). We found that TGF‐β stimulation increased the amounts of TMEPAI 5′cDNA in all human breast cancer cell lines used in this experiment. To identify which TMEPAI isoforms are expressed in breast cancer cells, we used a 5′RACE method. After generation of the 5′RACE cDNA, we did 5′RACE PCR to amplify and distinguish each variant of TMEPAI mRNA by PCR product size and sequencing. We found that breast cancer cell lines generate one common major transcriptional variant, TMEPAI mRNA variant 5, in the presence of TGF‐β stimulation (Table 1). As our 5′RACE PCR result might be affected by the amplification efficiency bias, we confirmed this result at the protein level. We carried out Western blot analyses to compare the electrophoretic mobility of endogenous TMEPAI proteins in MDA‐MB‐157 and Hs578T cells with each TMEPAI protein isoform exogenously expressed in COS‐7 cells. After TGF‐β stimulation, these breast cancer cell lines expressed a single band of TMEPAI protein with an electrophoretic mobility corresponding to the band of TMEPAI isoform d (Figure 1c). These data are consistent with the results of 5′RACE experiments leading to the conclusion that breast cancer cell lines generate TMEPAI mRNA variant 5 which is translated to TMEPAI isoform d in response to TGF‐β stimulation.
| Cells | Isoform a | Isoform b | Isoform c | Isoform d | Total number of colonies read | |
|---|---|---|---|---|---|---|
| Variant 1 | Variant 2 | Variant 3 | Variant 4 | Variant 5 | ||
| BT−549 | 0 | 2 | 0 | 0 | 7 | 10 |
| HCC 1,395 | 0 | 1 | 0 | 0 | 6 | 8 |
| Hs578T | 0 | 1 | 0 | 0 | 5 | 8 |
| MDA‐MB−157 | 0 | 0 | 0 | 0 | 5 | 8 |
| BT−474 | 0 | 1 | 0 | 0 | 6 | 8 |
| SK‐BR−3 | 2 | 0 | 0 | 0 | 5 | 8 |
Note
- 5′RACE PCR products are purified from agarose gel and subcloned into pGEM‐T easy vector. Eight to ten random colonies are subjected to DNA sequencing analysis. Sequence results are aligned using nucleotide BLAST in NCBI online software. All breast cancer cell lines mostly express variant 5, which codes isoform d.
2.2 Establishment of TMEPAI knockout in breast cancer cells
We used a CRISPR/Cas9 system to knockout (KO) TMEPAI in breast cancer cells using two TNBC cell lines, Hs578T and BT‐549. A single gRNA targeting exon 2 and one pair of gRNAs to delete exon 4 coding region of the TMEPAI gene were designed to recruit Cas9 which induced double‐strand DNA break (DSB) on the TMEPAI gene target loci (Figure 2a). Protein expression analyses and genomic DNA sequencing were carried out to screen for TMEPAI KO clones. Two KO clones were selected from each cell line: clones #5 and #18 from Hs578T and clones #22 and #29 from BT‐549. We could detect TMEPAI protein in Hs578T and BT‐549 parental cells after TGF‐β stimulation although it was diminished in TMEPAI KO clones (Figure 2b,c). Sequence analysis showed monoallelic or biallelic frame shift mutations in TMEPAI KO clones (Figure S2). Attempts to establish biallelic KO clones using Hs578T failed, suggesting that total loss of TMEPAI is fatal in this cell line. However, the monoallelic KO clones from Hs578T showed a 90% or more decrease in TMEPAI protein and could be used in the evaluation of TMEPAI KO cellular phenotypes. We previously demonstrated that TMEPAI participates in a negative feedback regulation of TGF‐β signaling by binding to Smad2 and Smad3 via its SIM and inhibiting the phosphorylation of Smad2 and Smad3 at the late stage of TGF‐β signaling (Watanabe et al., 2010). To further confirm negative effects of TMEPAI on TGF‐β/Smad signaling using our established TMEPAI KO clones, we evaluated Smad2 phosphorylation levels at 16 hr of TGF‐β stimulation. As a result, the levels of phosphorylated Smad2 in TMEPAI KO clones were increased compared to the parental cells of both Hs578T and BT‐549 cells (Figure 2d,e).
2.3 TMEPAI KO in breast cancer cells show reduced colony‐ and sphere‐forming ability
Cell proliferation, colony formation and sphere formation assays associated with tumorigenic activity were conducted using TMEPAI KO breast cancer cells. Although TMEPAI KO did not significantly affect monolayer cell proliferation in both Hs578T and BT‐549 (Figure 3a,b), TMEPAI KO significantly reduced colony and sphere formation in Hs578T and BT‐549 cells (Figure 3c,d). Collectively, our results indicate that TMEPAI is not essential for monolayer cell proliferation but is crucial to drive colony and sphere formation in breast cancer cells.
2.4 TMEPAI isoform complementation rescues colony‐ and sphere‐forming ability
We hypothesized that various extracellular regions of TMEPAI might be important for its tumorigenic functions, particularly within isoform d. To quantify tumorigenic activity of each isoform, we individually complemented each TMEPAI isoform (a, b or d) into TMEPAI KO breast cancer clones #5 (Hs578T) and #29 (BT‐549) using a lentiviral expression system and performed colony and sphere formation assays. The protein expression levels of each TMEPAI isoform in the TMEPAI KO cells were detected, and complementation verified using Western blot analysis (Figure S3a,b). Isoform c was excluded as protein expression was not comparable to the other three isoforms and was not detected in any of our breast cancer cell lines. The results demonstrated that re‐expression of all TMEPAI isoforms a, b and d significantly rescued colony‐ and sphere‐forming abilities in TMEPAI KO clones compared to vector controls and there was no significant difference among TMEPAI isoforms a, b and d in rescued colony‐ (Figure 4a,b) and sphere‐forming abilities (Figure 4c,d). These data suggest that TMEPAI protein drives colony and sphere formation but there is no functional difference among TMEPAI isoforms at least at high expression levels in these experiments.
2.5 TMEPAI isoforms suppress TGF‐β signaling
We additionally investigated the functional difference of TMEPAI isoforms in TGF‐β signaling pathway. By using (CAGA)12‐luciferase reporter, we found that TMEPAI isoforms a, b and d suppressed the TGF‐β/Smad signaling pathway (Figure 4e). To test whether TMEPAI isoforms suppress the TGF‐β/Smad signaling pathway by inhibition of the R‐Smads phosphorylation, we transfected Smad2 together with constitutively active TβRI, alternatively termed constitutively active ALK5 (ALK5ca), into HEK293 cells. As a result, TMEPAI isoforms a, b and d markedly reduced phosphorylation of Smad2 upon ALK5 activation (Figure 4f). These results suggested that TMEPAI isoforms a, b and d have similar function in inhibition of TGF‐β/Smad signaling pathway.
2.6 Both PY motifs and a SIM of TMEPAI are essential for tumorigenic activities
As previous studies have shown that PY motifs and a SIM are important for TMEPAI functions (Li et al., 2008; Singha et al., 2014; Watanabe et al., 2010; Xu et al., 2003), we expressed TMEPAI isoform a with PY 1 mutation, PY 2 mutation, PY 1&2 mutation and SIM mutation in the same TMEPAI KO clones (Figure S4). The rescue functions of a TMEPAI double PY mutant and a SIM mutant were significantly weaker compared to TMEPAI isoform a wild type and single PY mutants, with almost no recovery in colony formation (Figure 5a,b) and marginal recovery in sphere formation assays (Figure 5c,d). These observations indicate that double PY motifs and a SIM are essential for the colony‐ and sphere‐forming functions of TMEPAI.
2.7 TGF‐β decrease colony‐ and sphere‐forming ability
Our previous study has demonstrated how TMEPAI negatively regulated TGF‐β/Smad signaling via its SIM which could interact with Smad2 and Smad3 (Watanabe et al., 2010). To examine whether our breast cancer cell lines respond to TGF‐β‐mediated growth inhibitory functions, we exogenously added TGF‐β ligand and TGF‐β receptor kinase inhibitor during Hs578T and BT‐549 colony and sphere formation assays. The results showed that the addition of TGF‐β was able to significantly decrease colony‐ (Figure 6a) and sphere‐ (Figure 6b) forming ability compared to the vehicle treated cells, and the effect is partially recovered by TGF‐β receptor kinase inhibitor, SB435412 in sphere‐forming assay (Figure 6b). These results suggested that breast cancer cell lines respond to TGF‐β‐induced sphere growth inhibition partially through TGF‐β type I receptor kinase activity which activates the Smad pathway.
2.8 PI3K/AKT signaling pathway is essential for colony‐ and sphere‐forming ability
TMEPAI has been reported to promote PI3K/AKT signaling activity by degradation of PTEN protein through PY motifs‐mediated NEDD4 interaction in TNBC (Singha et al., 2014). We confirmed that PTEN expression levels were increased in TMEPAI KO cells resulting reduction of AKT phosphorylation on Ser473 (Figure S5). To evaluate whether PI3K/AKT signaling pathway is important for colony and sphere formation in breast cancer cell lines, we tested the effect of PI3K inhibitor, LY294002, on the Hs578T and BT‐549 colony and sphere formation assays. The results showed that LY294002 significantly inhibited colony (Figure 6a) and sphere (Figure 6c) formation compared to the vehicle treated cells, suggesting that PI3K/AKT signaling pathway is required for colony‐ and sphere‐forming ability.
3 DISCUSSION
In this study, we first found that TMEPAI isoform d is the major isoform expressed in breast cancer cells in response to TGF‐β stimulation, but all isoforms a, b and d recovered colony‐ and sphere‐forming activities in TMEPAI KO breast cancer cell lines and did not exhibit significant difference. However, double PY motifs and a SIM, which are commonly located in the cytoplasmic region of all TMEPAI isoforms, were essential for both colony‐ and sphere‐forming abilities induced by TMEPAI in breast cancer cells.
There are only a few reports on TMEPAI isoforms in cancer cells. By using the 5′RACE method, Brunschwig et al. (2003) showed that TMEPAI variant 1, which encodes TMEPAI isoform a, is expressed in four colon cancer cell lines after TGF‐β stimulation. Liu, Zhou, Huang, & Chen (2011) next reported that androgen‐dependent prostate cancer cells preferentially express TMEPAI isoform a. Furthermore, the androgen‐independent type of prostate cancer cells expresses TMEPAI isoforms a and c in response to TGF‐β stimulation (Fournier et al., 2015). In contrast with the previously reported results of colon and prostate cancer cells, we found exclusively large amounts of isoform d in breast cancer cell lines which was consistently up‐regulated by TGF‐β stimulation. TMEPAI was originally identified as an androgen‐inducible protein in the prostate. We subsequently found TMEPAI as a TGF‐β‐inducible protein and identified TCF7L2 binding to a TCF/LEF binding site and Smad3 binding to the Smad binding elements in the intron 1 are essential for the TMEPAI expression induced by TGF‐β (Nakano et al., 2010). Additionally, TMEPAI expression is reported to be induced by multiple signaling such as androgen, TGF‐β, EGF and mutant p53. Thus, different types of cells may express distinct TMEPAI isoforms in response to TGF‐β or other stimuli.
Our current study, as well as two previous reports, suggested that different TMEPAI isoforms have similar functions in tumorigenic activities. Koido, Sakurai, Tsukahara, Tani, & Tomida (2016) described TMEPAI as a novel hypoxia‐inducible gene and indicated that all TMEPAI isoforms are able to enhance HIF‐1 transcriptional activity and the weak effect of isoform c might be simply due to lower protein amounts that was also seen in our results. Another study demonstrated that exogenously expressed TMEPAI isoforms a and b have similar functions to increase cancer stem cell populations in breast cancer cells (Nie et al., 2016). Recent study using RNA seq data from solid tumor samples (prostate, breast, colon and lung tumors) in TCGA data set also suggested that there was no association between TMEPAI isoforms expression and the pathological stage of cancer progression and both of TMEPAI isoforms a and b expression in solid tumor samples was associated with poor prognosis (Sharad et al.., 2019). These results together with our current results suggest that the different TMEPAI isoforms including isoform d are expressed in different type of cancer cells but have similar molecular functions.
TMEPAI‐related publications have been increasing since it was first found in 2000, as TMEPAI might play an important role in tumorigenesis but the molecular functions are not fully understood yet. In our laboratory, we first reported TMEPAI functions as the negative regulator of TGF‐β/Smad signaling pathway through its SIM. Here, we further confirmed that all TMEPAI isoforms a, b and d suppressed TGF‐β/Smad signaling by reducing the Smad2 phosphorylation (Figure 4e,f).
The dual functions of TGF‐β in cancer have been reported for decades (Massague, 2008, 2012; Moses & Barcellos‐Hoff, 2011). By increasing expression of the cyclin‐dependent kinase (CDK) inhibitors (e.g., p21 and p27) and decreasing expression of c‐MYC, TGF‐β inhibits cell proliferation and acts as a tumor suppressor. However, in some tumor cells, TGF‐β‐induced growth inhibitory functions are lost then TGF‐β contributes to tumor invasion and metastasis by inducing epithelial–mesenchymal transition (EMT). The dual roles of TGF‐β in tumor progression are due to TGF‐β‐activated Smad and non‐Smad signaling pathways (Dumont, Bakin, & Arteaga, 2003; Kubiczkova, Sedlarikova, Hajek, & Sevcikova, 2012; Lynch et al., 2001). We found that our breast cancer cell lines, Hs578T and BT‐549, were still responsive to TGF‐β‐mediated growth inhibitory function in colony and sphere formation assays and treatment with the TGF‐β receptor kinase inhibitor, SB435412, partially recovered these effects. The presence of TMEPAI with its SIM might suppress growth inhibition through negative regulation of TGF‐ β/Smad signaling, consequently promoting cancer cell growth in colonies and spheres. Furthermore, we confirmed that re‐expression of TMEPAI SIM mutant failed to fully recover colony‐ and sphere‐forming ability in TMEPAI KO cells (Figure 5). Although TMEPAI protein levels were significantly increased by additional TGF‐β treatment in our experiment (Figure 1), it is reported that most of cancer cells secrete TGF‐β and could maintain TGF‐β activity by autocrine or paracrine signaling, especially when cancer cells were cultured in sphere‐based or in vivo‐based experiments (Ikushima et al., 2009; Tobin et al., 2002). Indeed, we could observe the decreased tumorigenic activities such as sphere and xenograft tumor formation in TMEPAI‐deficient breast cancer cell lines (Figure 3, Abdelaziz, Watanabe, & Kato, 2019) and lung carcinoma cell lines (Vo Nguyen et al., 2014), suggesting endogenous expression of TMEPAI works in these conditions. However, the question remains as to why TMEPAI affects cell proliferation in sphere, colony and in vivo tumor formation but not in monolayer cell culture condition. We hypothesize the autocrine and paracrine secretion of endogenous TGF‐β might be accumulated and more abundant in sphere culture, colony culture or in vivo tumor formation conditions compared to the monolayer culture condition. In monolayer condition, we changed the medium every three days. In contrast, we grow cells for a longer period in the same medium in sphere, colony culture conditions. It may affect the accumulation of endogenous TGF‐β and other stimuli leading to enhanced expression of TMEPAI similarly to in vivo condition.
TMEPAI has been well reported to be involved in PI3K/AKT and androgen signaling pathway through its PY motifs (Li et al., 2008; Singha et al., 2014). TMEPAI PY motifs were reported to interact with the WW domain of NEDD4, causing indirect modulation of PTEN protein degradation, leading to PI3K/AKT signaling activation that contribute to cancer cell proliferation and invasion (Cichon & Radisky, 2014; Singha et al., 2014). In agreement with the previous study, we confirmed that PY motifs were essential for breast cancer colony‐ and sphere‐forming ability as mutations in both PY motifs, but not in each single PY motif, fail to rescue the colony‐ and sphere‐forming ability (Figure 5). This effect could be due to the loss of interaction between TMEPAI and NEDD4 resulting in higher expression of PTEN protein that kept the PI3K/AKT signaling at low levels of activity. Furthermore, our results suggested the importance of PI3K/AKT signaling in breast cancer colony‐ and sphere‐forming ability as treatment with the PI3K inhibitor, LY294002, significantly decreased its colony and sphere formation (Figure 6c,d). However, some studies reported about TMEPAI functions as tumor suppressor in prostate cancer, Li et al. (2008) and Xu et al. (2003) demonstrated that negative feedback regulation of androgen receptor (AR) by TMEPAI results in growth inhibition of prostate cancer. Thus, loss of TMEPAI enhances AR function and promote prostate cancer progression.
In conclusion, TMEPAI works as both tumor suppressor and tumor promoter in different cancer cells. Although the main isoform expressed in prostate cancer cells is different from breast cancer cells, all isoforms of TMEPAI show similar tumorigenic activities in breast cancer cells. TMEPAI has been reported as signaling modulator that suppresses TGF‐β/Smad, androgen and Wnt signaling and promotes PI3K/AKT signaling (Amalia et al., 2019; Li et al., 2008; Singha et al., 2014; Watanabe et al., 2010) Thus, the different roles of TMEPAI seem to depend on the context of the underlying oncogenic pathways activated in cancer cells, such as TGF‐β, PI3K/AKT, Wnt and androgen signaling (Itoh & Itoh, 2018).
In the present study, we propose that TMEPAI contributes to tumor progression through both PY motifs and a SIM (Figure 7). Moreover, TMEPAI PY motifs and a SIM are well conserved among mammals, indicating the importance of these two distinct motifs (Figure S5). Our results highlight TMEPAI functions in promoting tumorigenesis in TNBC cells and could contribute to the development of a novel diagnostic marker and a therapeutic agent for treating TNBC patients.
4 EXPERIMENTAL PROCEDURES
4.1 Cell lines and cell culture
Breast cancer cell lines, HEK293 and COS‐7 cells were obtained from American Type Culture Collection (ATCC) or Riken bioresource research center. Breast cancer cell lines BT‐549 (CVCL_1092), HCC1395 (CVCL_1249), Hs578T (CVCL_0332), MDA‐MB‐157 (CVCL_0618), BT‐474 (CVCL_0179), SK‐BR‐3 (CVCL_0033), COS‐7 (African green monkey kidney cell line transformed by SV40; CVCL_0224) and HEK293 (RCB1637) were obtained from American Type Culture Collection (ATCC) or Riken bioresource research center. MDA‐MB‐157, BT‐474, SK‐BR‐3, COS‐7 and HEK 293 cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Invitrogen) containing 10% fetal bovine serum (FBS, Gibco) and 100 units/ml of penicillin G plus 0.1 mg/ml of streptomycin sulfate (Wako). BT‐549 and Hs578T cells were cultured in DMEM (Invitrogen) containing 10% FBS (Gibco), 100 units/ml of penicillin G plus 0.1 mg/ml of streptomycin sulfate (Wako) and 10 μg/ml of insulin. HCC1395 cells were cultured in RPMI 1,640 medium (Sigma) containing 10% FBS (Gibco) and 100 units/ml of penicillin G plus 0.1 mg/ml of streptomycin sulfate (Wako). In addition, BT‐549 and Hs578T TMEPAI re‐expressed cells were cultured in the presence of 8 μg/ml blasticidin S hydrochloride (Wako). All cells were maintained in a humidified atmosphere with 5% CO2 at 37°C throughout all experiments. Mycoplasma detection was carried out using a mycoplasma detection set (Takara Bio), and all experimental cell lines were negative for Mycoplasma infection.
4.2 Knockout (KO) of all TMEPAI isoforms
Construction of TMEPAI guide RNA expression vectors and establishment of TMEPAI KO cell lines were described previously (Amalia et al., 2019; Wardhani et al., 2017). Knockout (KO) cell lines which are negative for all isoforms of TMEPAI were established from Hs578T and BT‐549 cells using a CRISPR/Cas9 system. Guide RNA (gRNA) sequences targeting the TMEPAI gene on exon 2, intron 3 and exon 4 were designed using the CRISPR Design Online Tool (https://crispr.dbcls.jp). Hs578T TMEPAI KO cell line clones #5 and #18, and BT‐549 TMEPAI KO cell line clone #29 were established using gRNA#3 transcribed from an gRNA cloning vector (Addgene #41824) containing a set of DNA fragments (5′‐TTTCTTGGCTTTATATATCTTGTGGAAAGGACGAAACACCGCAGCCGGCACAGCCAGGGG‐3′ and 5′‐ GACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAACCCCCTGGCTGTGCCGGCTGC‐3′) targeting the TMEPAI gene sequence on exon 2 (5′‐CAGCCGGCACAGCCAGGGG‐3′). BT‐549 TMEPAI KO cell line clone #22 was established using a set of gRNA#4 and gRNA#5; gRNA#4 was transcribed from a vector containing (5′‐ TTTCTTGGCTTTATATATCTTGTGGAAAGGACGAAACACCGCCAGAAGCGATCCTGAGAC‐3′ and 5′‐GACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAACGTCTCAGGATCGCTTCTGGC‐3′) targeting the TMEPAI gene sequence on intron 3 (5′‐CCAGAAGCGATCCTGAGAC‐3′) and gRNA#5 from a vector containing (5′‐TTTCTTGGCTTTATATATCTTGTGGAAAGGACGAAACACCGAGAGAAACTGTATGTGCGA‐3′ and 5′‐GACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAACTCGCACATACAGTTTCTCT‐3′) targeting the TMEPAI gene sequence of 3′ UTR on exon 4 (5′‐AGAGAAACTGTATGTGCGA‐3′).
4.3 Construction of plasmid
Constitutively, active ALK5 (ALK5ca)‐HA and Flag‐Smad2 plasmids constructions were described previously (Itoh, Ericsson, Nishikawa, Heldin, & ten Dijke, 2000; Noda et al., 2006; Watanabe et al., 2010). The (CAGA)12 luciferase reporter and pRL‐CMV Renilla luciferase control reporter plasmids were described previously15. Human TMEPAI cDNA was cloned by the RT‐PCR method. Human TMEPAI isoforms a, b and d, isoform a single PY mutant (Y161A or Y232A), double PY mutant (Y161A, Y232A) and an isoform a SIM mutant (186PPNR to AAAA) were cloned into the pcDNA3.1 vector. All of the TMEPAI mutants were made with a QuickChange site‐directed mutagenesis kit (Stratagene) using TMEPAI isoform a expression plasmid vector as a template.
4.4 Establishment of TMEPAI complemented cell lines
Lentiviral vectors expressing TMEPAI isoforms a, b and d, isoform a single and double PY mutants, and isoform a SIM mutant were made using CSII‐CMV‐MCS‐IRES2‐Bsd vector. Establishment of TMEPAI complemented/re‐expressed cells were described previously (Amalia et al., 2019).
4.5 5′ Rapid amplification of cDNA ends (RACE)‐ready cDNA synthesis and 5′ RACE PCR
Using the extracted total RNA from cells stimulated by TGF‐β for 2 hr, 5′RACE was carried out using the SMARTer RACE cDNA Amplification Kit (Clontech Takara Bio) according to the manufacturer's instructions. 5′ RACE PCR was carried out using Ex Taq polymerase (Takara Bio) according to the manufacturer's instructions. We used a nested PCR strategy to increase specific yield of our cDNA area of interest. For the first step of PCR, we used a universal primer (5′‐CTAATACGACTCACTATAGGGCAAGCAAGCAGCAGTGGTATCAACGCAGAGT‐3′) as the forward primer (containing a complementary sequence to the appended sequence from the 5′RACE ready cDNA Amplification Kit) and E4R (5′‐ GTGGGGGCTCCTCCCCGTC‐3′) as located on TMEPAI exon 4 as the reverse primer. For the second PCR, we used a sequence matching the 3′ half of the universal primer (5′‐ AAGCAGCAGTGGTATCAACGCAGAGT‐3′) as a forward primer, and E3R (5′‐ GTTGCCTGACACTGTGCTCT‐3′) as located on TMEPAI exon 3 as the reverse primer. The PCR products were loaded onto a 2% agarose gel, and bands located at expected product sizes were extracted and purified using a QIAGEN PCR extraction and purification kit (QIAGEN). Then, these products were subcloned into a pGEM‐T easy plasmid vector system (Promega) and used for transformation of E. coli. Plasmid vectors from eight to ten random colonies were subjected to DNA sequencing (DNA Sequencer 3130, Life Technologies, Applied Biosystem).
4.6 RNA extraction, Reverse Transcription‐PCR (RT‐PCR) and quantitative PCR
A total of 1 x 105 cells were seeded and subjected to RNA extraction on the next day. Total RNA was extracted using ISOGEN II reagent (Nippon gene), then precipitated with isopropanol and resolved in DEPC‐treated water. Reverse transcription was carried out using High Capacity RNA‐to‐cDNA Master Mix (Applied Biosystems). Quantitative PCR was performed on a Real‐time PCR system 7500 Fast thermal cycler (Life Technologies, Applied Biosystems) according to the manufacturer's instructions. The primers used for quantitative PCR amplification were as follows: human TMEPAI (forward 5′‐AGAGCACAGTGTCAGGCAAC‐3′, reverse 5′‐GTGCTGCAGGTACGGATAGG‐3′) and human β‐ACTIN (forward 5′‐CGTACAGGTCTTTGCGGATG‐3′, reverse 5′‐GCACTCTTCCAGCCTTCCTT‐3′).
4.7 Western blotting
Cell lysates were then subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS‐PAGE) and electrotransferred to PVDF membrane (GE Healthcare) for immunoblot analysis. The primary antibodies used were rabbit anti‐TMEPAI (15, 1:500), rabbit antiphosphorylated Smad2 (PS2) (Persson et al., 1998, 1:1,000), mouse anti‐Flag (M2, Sigma, 1:1,000), rat anti‐HA (3F10, Roche Applied Science, 1:1,000), mouse antitotal Smad2/3 (BD Bioscience Pharmingen, 1:1,000) and mouse anti‐β‐actin (Sigma, 1:5,000). Detection of bound primary antibody was accomplished with horseradish peroxidase‐conjugated anti‐mouse IgG (GE Healthcare, 1:10,000) or anti‐rabbit IgG (GE Healthcare, 1:10,000) or anti‐rat IgG (GE Healthcare, 1:10,000), and Immuno Star Zeta (Wako) and EZ capture MG (ATTO) were used for the detection of the chemiluminescent reaction.
4.8 Luciferase assay
HEK293 cells were seeded at 1.5 × 105 cells/well in 12‐well plate one day before transfection. Cells were transfected with the (CAGA)12‐firefly luciferase reporter, pRL‐CMV Renilla luciferase control reporter, and TMEPAI expression plasmid using FuGENE6 (Promega). The cell was stimulated with TGF‐ β (1 ng/ml) for 18 hr before lysis. On the 4th day, cells were lysed with lysis buffer (2 mM DTT. 25 mM Tris‐phosphate (pH 7.8), 2 mM trans1‐2‐diaminocyclohexane‐N,N,N’,N’‐tetraacetic acid monohydrate, 1% Triton X100) and mixed with Luciferase Assay Reagent (Promega) and Renilla Luciferase Assay System (Promega). Luciferase activities were determined by luminometer (Berthold) and normalized to respective Renilla luciferase activities.
4.9 Cell proliferation assay
Breast cancer parental and TMEPAI KO cells were cultured for the indicated periods. A total of 5 × 104 cells per well were seeded on 12‐well plates (Corning) at day 0. Cells were collected by trypsinization at the indicated time points and were diluted with trypan blue before cell counting with an Automatic Cell Counter (TMC10, BioRad).
4.10 Colony‐forming assay
Breast cancer parental, TMEPAI KO and TMEPAI complemented cells were seeded at 1 × 103 cells (Hs578T) or 2 × 103 cells (BT‐549) in 10 cm dishes (Corning) and cultured for two weeks. Growing colonies were fixed with 4% formaldehyde for 15 min and stained with 0.01% (w/v) crystal violet (Sigma‐Aldrich) in water for one hour. Colony quantification was accomplished with ImageJ software (National Institutes of Health).
4.11 Sphere formation assay
Breast cancer parental, TMEPAI KO and TMEPAI complemented cells were cultured for two weeks or eight days (for experiments treated with inhibitors) in DMEM/F12 serum‐free medium (Invitrogen) containing 100 units/ml of penicillin G plus 0.1 mg/ml of streptomycin sulfate (Wako) in the presence of 1% methylcellulose (Wako) supplemented with B27 (20 μl/ml; Invitrogen), EGF (20 ng/ml; Sigma‐Aldrich), basic fibroblast growth factor (20 ng/ml; Wako) and insulin (5 μg/ml, Wako) in six‐well plates (Corning) coated with 3% Poly (2‐hydroxyethyl methacrylate) (Sigma) in 95% ethanol for low attachment.
4.12 Statistical analysis
Graphs and curves were drawn using Microsoft Excel software. Data are means ± SD, n = 3 technical replicates from a representative of at least two independent experiments. All statistical analyses of the data were carried out using IBM SPSS Statistic 23.0 software (IBM). Statistical analyses were carried out via ANOVA with Tukey's multiple comparisons test or Kruskal–Wallis with Mann–Whitney test; for data consisting of only two groups, statistical analyses were carried out via independent t test. Probability values of <.05 were considered significant.
4.13 Bioinformatic analysis
The amino acid sequences reported for the TMEPAI proteins were obtained from NCBI. The T‐coffee online software tool was used for multiple amino acid sequence alignments of transmembrane proteins (Chang, Di Tommaso, Taly, & Notredame, 2012).
ACKNOWLEDGMENTS
This research was supported by JSPS KAKENHI Grant Numbers JP18K06682 (YW), JP16K19100 (YW), JP25870093 (YW), JP18H02676 (MK) and JP25293092 (MK). We thank Dr. George Church for kindly providing us gRNA_Cloning Vector (Addgene #41824) and hCas9 expression vector (Addgene #41815). We also thank Dr. Bryan J. Mathis of the Medical English Communication Center (University of Tsukuba) for excellent English editing of this manuscript.
CONFLICTS OF INTEREST
The authors declare that we have no potential conflicts of interest to disclose.
