The calcium channel TRPC6 promotes chemotherapy-induced persistence by regulating integrin α6 mRNA splicing

SUMMARY

Understanding the cell biological mechanisms that enable tumor cells to persist after therapy is necessary to improve the treatment of recurrent disease. Here, we demonstrate that transient receptor potential channel 6 (TRPC6), a channel that mediates calcium entry, contributes to the properties of breast cancer stem cells, including resistance to chemotherapy, and that tumor cells that persist after therapy are dependent on TRPC6. The mechanism involves the ability of TRPC6 to regulate integrin α6 mRNA splicing. Specifically, TRPC6-mediated calcium entry represses the epithelial splicing factor ESRP1 (epithelial splicing regulatory protein 1), which enables expression of the integrin α6B splice variant. TRPC6 and α6B function in tandem to facilitate stemness and persistence by activating TAZ and, consequently, repressing Myc. Therapeutic inhibition of TRPC6 sensitizes triple-negative breast cancer (TNBC) cells and tumors to chemotherapy by targeting the splicing of α6 integrin mRNA and inducing Myc. These data reveal a Ca²⁺-dependent mechanism of chemotherapy-induced persistence, which is amenable to therapy, that involves integrin mRNA splicing.

In brief

Mukhopadhyay et al. report that TRPC6, a cation channel that mediates calcium entry, distinguishes cancer stem cells from other populations in breast cancer and that it sustains their function. Tumor cells that persist after chemotherapy are dependent on TRPC6, and its targeted inhibition sensitizes them to chemotherapy.

Graphical Abstract

INTRODUCTION

Understanding the mechanisms that contribute to intratumor heterogeneity is a challenging biological problem and a key factor in therapy resistance. In this context, cancer stem cells (CSCs) have an essential role because of their ability to differentiate into various morphological and functional cancer cell states and, importantly, because of their resistance to therapy.¹ However, the plasticity of CSCs is a complicated process that involves regulation by the tumor microenvironment and intrinsic metabolic alterations.^2,3 To understand intratumor heterogeneity better and develop strategies to overcome treatment resistance, it is necessary to investigate the mechanisms that sustain CSCs further and exploit them to improve therapy.

The importance of investigating the contribution of CSCs to intratumor heterogeneity and therapy resistance is exemplified by triple-negative breast cancer (TNBC) because it is characterized by extensive intratumor heterogeneity,⁴ a high frequency of CSCs compared to other breast cancer subtypes,⁵ and resistance to standard chemotherapy. Thus, there is an urgent need to understand the nature of TNBC heterogeneity in more detail, especially its resistance to most current therapies, and to develop novel strategies for diminishing the morbid statistics associated with it. One approach to this problem involves investigating mechanisms that distinguish the function of CSCs from other populations, especially in the context of therapy resistance. Our unbiased approach to this problem revealed the potential contribution of calcium channels to the biology of CSCs. This observation piqued our interest because their contribution to TNBC and other cancers merits further investigation for several reasons. Calcium signaling is fundamental to cell proliferation, migration, invasion, and survival,⁶ processes that underlie tumor progression. Numerous studies have highlighted the importance of calcium signaling in the behavior of tumor cells and their response to chemotherapy. More specifically, calcium signaling has been shown to maintain a partial/hybrid epithelial-to-mesenchymal transition (EMT) state,⁷ which is associated with stemness in many cancer models. Another significant finding is that chemotherapy can induce an increase in intracellular Ca²⁺ that causes enrichment of breast CSCs.⁸ Although these and many other studies have highlighted the importance of calcium channels and calcium signaling in cancer, much less is known regarding the contribution of specific calcium channels to sustaining CSC subpopulations present in heterogeneous tumors and about how such channels impact CSC functions, especially the response to therapy.^9–13

In this study, we report that transient receptor potential channel 6 (TRPC6) expression, a non-selective cation channel that mediates calcium entry in TNBC,¹⁴ marks CSCs in TNBC and that its calcium signaling function has a causal role in maintaining stemness and chemoresistance. We also demonstrate that TRPC6 contributes to these functions by regulating mRNA splicing, especially the splicing of the α6 integrin mRNA that contributes to CSC function. Tumor cells that persist after chemotherapy are dependent on TRPC6 and a splice variant of the α6 integrin. Therapeutic targeting of TRPC6 blocks this splicing mechanism and sensitizes TNBC to chemotherapy.

RESULTS

TRPC6 is associated with CSCs in TNBC and has a causal role in stemness

Initially, we analyzed our previously published RNA sequencing (RNA-seq) data that compared the transcriptomes of immortalized mammary epithelial cells (S1 cells) to their transformed counterparts, which are enriched in CSCs.¹⁵ Analysis of these data revealed that the expression of TRPC6 (transient receptor potential cation channel, subfamily C, member 6), is significantly increased in the transformed compared to the immortalized population (Figure 1A). The RNA-seq data were verified by qPCR to quantify the expression of several calcium channels known to promote tumorigenic pathways including ORAI-1,¹⁶ ORAI-3,¹⁶ TRPC6,¹⁷ TRPV6,¹⁸ and TRPM8^19,20 (Figure 1B). To assess whether the expression of TRPC6 is associated with TNBC compared to other subtypes of breast cancer, we performed qPCR on patient-derived organoids (PDOs) and observed that its expression is higher in TNBC relative to estrogen receptor (ER)⁺ organoids (Figure S1A). We also found that TRPC6 is expressed substantially more in TNBC than in non-TNBC (ER⁺) tumor specimens (Figure S1B)

Figure 1. TRPC6 is enriched in CSCs.

(A) Heatmap of differentially expressed genes in the CSC and non-CSC populations from a published model system.¹⁴ TRPC6 is highlighted in the heatmap. False discovery rate (FDR) < 0.05 and |log2FC| (fold change) > 1.

(B) Expression of TRPC6 and other calcium channels was compared in the same CSC vs. non-CSC populations by qPCR.

(C) TRPC6 expression was compared in MDA-MB-231 cells and their TE3 variants.

(D) HMLER cells were sorted into CD104⁻/CD24⁻ (CSC; HMLER^−/−) and CD104⁺/CD24⁺ (non-CSC; HMLER^+/+) populations, and TRPC6 expression was quantified by qPCR.

(E) The CSC (CD44^high/CD24^low) and non-CSC (CD44^low/CD24^high) populations was sorted from a TNBC PDX (left), and TRPC6 expression levels measured were quantified by qPCR.

(F and G) TRPC6 expression was knocked down in either TE3 (F) or CAL-51 (G) cells using shRNA, and expression was rescued with either a wild-type TRPC6 construct (WT) or a pore mutant (G757D) that is deficient in calcium uptake (MUT). These populations were assessed for self-renewal by serial passage of mammospheres (P1, passage 1; P2, passage 2).

(H) CAL-51 control shRNA (shCTRL) or TRPC6 knockdown (shTRPC6) cells were injected into the mammary fat pads of NSG mice in limiting dilution (10⁶, 10⁵, and 10⁴ cells), and the frequency of tumor incidence was determined (right). Tumor incidence was plotted utilizing ELDA in a log plot to estimate the frequency of tumor-initiating cells in each group.

(I) CAL-51 cells (shCTRL, shTRPC6–1, sh+ WT TRPC6, and sh+ mut TRPC6) were injected into the mammary fat pads of NSG mice, and tumor onset in terms of days post-injection was compared among the groups.

The TRPC6 expression data shown in (C) and (D) represent the mean ± SD of a representative experiment performed three times independently. The mammosphere data shown represent the mean ± SD of three independent experiments. For (H), data are presented as log-log plot, and the frequency of tumor-initiating cells is calculated by extreme limiting-dilution analysis. The tumor onset data represent the median days post-injection between the groups. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

We extended our initial findings on TRPC6 to other CSC models. TE3 cells are a variant of MDA-MB-231 cells that were selected for their tumor initiation capacity,²¹ and we observed that TRPC6 expression is significantly higher in TE3 cells than in MDA-MB-231 cells (Figure 1C). We also investigated TRPC6 expression in Ras-transformed, human mammary luminal cells (HMLER),²² which contain a CD104^low/CD24^low subpopulation (HMLER^−/−) that is enriched in stem cell properties¹⁵ and observed that the expression of TRPC6 in the CSC (HMLER^−/−) fraction is significantly higher than in the non-CSC subpopulation (CD104^high/CD24^high or HMLER^+/+) (Figure 1D). We also isolated a CSC population (CD44^high/CD24^low) from a TNBC patient-derived xenograft (PDX) and found that TRPC6 expression is associated with the CSC population compared to the non-CSC population (CD44^low/CD24^high) (Figure 1E). Thus, TRPC6 expression is a marker of CSCs in multiple models.

Given that TRPC6 is a cell surface protein, we postulated that it could be used as a marker to enrich CSCs from heterogeneous tumor cell populations. To test this possibility, we used HCC1806 cells, a heterogeneous TNBC cell line that contains CSC and non-CSC populations based on our previous studies.²³ Two distinct populations of TRPC6^high and TRPC6^low cells (Figure S1D) were identified by fluorescence-activated cell sorting (FACS), and we found that the TRPC6^high population harbors a significantly higher frequency of CSCs than the TRPC6^low population (Figure S1E).

To investigate a causal role for TRPC6 in stemness, we used short hairpin RNAs (shRNAs) to knock down its expression in TE3 and CAL-51 cells (Figure S2A) and observed a significant decrease in self-renewal as assayed by serial mammosphere formation (Figures 1F and 1G). To substantiate these results, we re-expressed either wild-type TRPC6 or a pore mutant TRPC6^G757D (mut TRPC6) (Figure S2B) that is defective in calcium uptake.²⁴ Consistent with our hypothesis that TRPC6-mediated calcium signaling sustains a CSC state, wild-type TRPC6, but not the pore mut, rescued serial mammosphere formation (Figures 1F and 1G). We substantiated these findings using the TRPC6-specific inhibitor BI-749327²⁵ and observed that the inhibitor-treated cells formed significantly fewer mammospheres compared to control cells (Figures S1H and S1I). We also treated the TRPC6^high CSC population sorted from HCC1806 cells with either DMSO or BI-749327 and observed a significant decrease in mammosphere formation ability upon channel inhibition (Figure S1F). Importantly, expression of wildtype (WT) TRPC6, but not the pore mut (Figure S1G), increased mammosphere formation in the TRPC6^low population of HCC1806 cells, providing evidence that TRPC6 is sufficient to drive stemness (Figure S1F). Subsequently, an in vivo assay was done to assess the contribution of TRPC6 to tumor initiation using extreme limiting-dilution analysis (ELDA).²⁶ The data obtained revealed that loss of TRPC6 caused a significant reduction in the frequency of CSCs (Figure 1H). We investigated the role of TRPC6 in facilitating tumor onset in vivo by xenograft implantation of TRPC6 shRNA knockdown CAL-51 cells that had been rescued with either WT or pore mut TRPC6. A significant delay in tumor onset was observed in the TRPC6 knockdown cells compared to the control cells, which was rescued by WT TRPC6 but not by pore mut TRPC6^G757D (Figure 1I). Together, these data implicate a causal role for TRPC6 in stemness.

TRPC6 has a causal role in chemoresistance

Based on the hypothesis that chemotherapy selects for the survival of cells with CSC properties, we sought to assess the functional role of TRPC6 in chemoresistance. For this purpose, we generated TNBC cells (CAL-51) that were resistant to paclitaxel, a drug that is standard of care for TNBC^27,28 (Figure 2A). Following paclitaxel treatment, we quantified TRPC6 mRNA expression and observed that it was ~13-fold higher in the resistant (CAL-51-R) compared to the sensitive population (CAL-51-S) (Figure 2B), an increase that was also observed in TRPC6 protein (Figure 3E). By using a limiting-dilution spheroid assay to determine the frequency of CSCs, we observed that the resistant population (CAL-51-R) contained a significantly higher frequency of CSCs than the sensitive population (CAL-51-S) (Figure 2C). Moreover, inhibition of TRPC6 activity using BI-749327 caused a significant decrease in the frequency of CSCs in the CAL-51-R population (Figure 2D). These results suggested that TRPC6 has a causal role in chemoresistance based on the reports that CSCs are chemoresistant.²⁹ In support of this possibility, we observed that treatment of the CAL-51-R cells with the TRPC6 inhibitor BI-749327 increased their sensitivity to paclitaxel significantly compared to vehicle-control-treated cells (Figure 2E). We verified these findings using a paclitaxel-resistant organoid from a TNBC patient tumor (9883T). Paclitaxel resistance was associated with a significant increase in TRPC6 expression (Figure 2F). Of note, treatment of this resistant organoid with BI-749327 increased its sensitivity to paclitaxel (Figure 2G), indicating that cells that resist paclitaxel treatment are dependent on TRPC6.

Figure 2. TRPC6 has a causal role in chemoresistance.

(A) Characterization of CAL-51 chemotherapy-resistant cells. CAL-51 cells were made resistant to paclitaxel as described in the STAR Methods and then compared to parental cells (termed sensitive) for viability in response to increasing concentrations of paclitaxel.

(B) TRPC6 mRNA expression was quantified in CAL-51-sensitive and -resistant cells.

(C) CAL-51-sensitive and -resistant cells were assayed for the frequency of CSCs using a limiting-dilution mammosphere assay, and the data were analyzed by ELDA.

(D) CAL-51-resistant cells were treated with vehicle control or BI-749327 (10 μM) for 24 h, and the frequency of CSCs was determined using a limiting-dilution mammosphere assay and ELDA.

(E) CAL-51-resistant (CAL-51-R) cells were treated with increasing concentrations of paclitaxel in combination with either DMSO or BI-749327 (10 μM) for 24 h, and cell viability was assessed.

(F) TRPC6 mRNA expression was quantified in chemotherapy-sensitive and -resistant models of a patient-derived organoid (9883T) as described in the STAR Methods.

(G) The organoid models described in (F) were treated with either DMSO, paclitaxel (PTX; 20 nM), BI-749437 (10 μM), or PTX+BI-749327 (BI) for 96 h, and cell viability was measured.

(H) Tumors volumes (in mm³) in mice that had been implanted orthotopically with a human TNBC PDX (PDX HCI028). The mice were divided into 4 groups of 5 mice each. When tumors reached an approximate volume of 100 m³, the mice were treated with either vehicle, PTX (15 mg/kg), BI (15mg/kg), or a combination of PTX and BI. PTX was injected intraperitoneally (i.p.) twice a week, and BI was administered by oral gavage 4 days a week. Day 0 on the x axis indicates the start of the treatments. Tumor volume was measured every 5 days.

For (C) and (D), data are presented as a log-log plot, and the frequency of stem cells is calculated by extreme limiting-dilution analysis. The viability data shown (E and G) represent the mean ± SD of a representative experiment performed three times independently. The tumor volume data shown in (H) are represented as mean ± SEM of the number of mice in the respective groups. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Figure 3. TRPC6-mediated Ca²⁺ entry represses the splicing protein ESRP1.

(A) Heatmap showing genes that are differentially expressed in vehicle- and BI (10 μM)-treated TE3 cells (12 h). FDR < 0.05 and |log2FC| > 1.

(B and C) ESRP1 mRNA (B) and protein (C) expression was assessed in TE3 cells treated with either vehicle control or BI (10 μM) for 24 h.

(D) TE3 and CAL-51 cells were treated with either vehicle control or BI (10 μM) for 24 h, and the expression of the α6A and α6B integrin splice variants was assessed by immunoblotting.

(E) The expression of TRPC6, ESRP1, α6A, and α6B was assessed in the CAL-51-sensitive and CAL-51-R cells by immunoblotting.

(F) The expression of ESRP1, α6A, and α6B in CAL-51-R cells that had been treated with either DMSO or BI (10 μM) for 24 h was assessed by immunoblotting.

(G) CAL-51-R cells were stably transfected with either a control plasmid (vector) or an ESRP1 expression plasmid (ESRP1-HA), and the expression of ESRP1, HA, and α6B was assessed by immunoblotting.

(H) The same cells as in (G) were assayed for their sensitivity to increasing concentrations of PTX.

(I) Parental CAL-51 cells that had been depleted of α6 integrin using CRISPR were stably transfected with either α6A or α6B plasmids. Cell viability in response to increasing concentrations of PTX was measured.

The TRPC6 expression data shown in (B) represent the mean ± SD of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

A critical issue that arose from our data on chemoresistance is whether TRPC6 inhibition could be an effective therapeutic approach in vivo for TNBC. To assess this possibility, we made use of an aggressive PDX model of TNBC that was reported recently (HCI028).³⁰ Treatment of these tumors with either paclitaxel or BI-749327 alone did not have a significant impact on tumor growth. Combined treatment, in contrast, resulted in tumor regression (Figure 2H).

TRPC6 suppresses ESRP1, which enriches for the α6B integrin splice variant

To investigate the mechanism by which TRPC6 sustains CSCs and promotes chemoresistance, we used an unbiased approach and performed RNA-seq on TE3 cells that had been treated with either DMSO or BI-749327 for either 6 or 12 h. Analysis of the data revealed that epithelial splicing regulatory protein 1 (ESRP1) was one of the top mRNAs upregulated in the inhibitor-treated cells (Figure 3A). ESRP1 regulates the splicing of mRNAs associated with an epithelial phenotype, and its loss has been associated with EMT³¹ and stemness³² programs in breast cancer. Our own studies have shown that the expression of ESRP1 is significantly lower in patients with TNBC than in patients with other subtypes of breast cancer and that it has a causal role in controlling stemness.^32,33 Moreover, a similar pattern of ESRP1 expression was observed in organoids derived from patients with TNBC and non-TNBC patients (Figure S3A). We validated the RNA-seq data by qPCR using TNBC cells treated with BI-749327 and confirmed that TRPC6 inhibition increases ESRP1 expression (Figures 3B and S3B). We also observed a significant increase in ESRP1 protein levels compared to control conditions in TE3 cells upon TRPC6 inhibition (Figure 3C).

The ability of TRPC6 to suppress ESRP1 captured our attention because we had previously reported that it regulates splicing of the α6 integrin.³³ There are two splice variants of the α6 integrin cytoplasmic domain: α6A and α6B. We have previously shown that the α6B variant promotes stemness.³³ In contrast, the α6A variant lacks this ability and is associated with non-CSC populations.³³ ESRP1-mediated mRNA splicing generates the α6A variant, and, consequently, loss of ESRP1 results in a predominance of the α6B variant.³³ Consistent with these findings, we observed that inhibition of TRPC6 caused a decrease in α6B and a concomitant increase in α6A protein levels (Figure 3D). To obtain evidence for the role of calcium in TRPC6-mediated repression of ESRP1 and the consequent splicing of α6 integrin, we treated CAL-51 cells with the calcium chelator BAPTA, and we observed an increase in ESRP1 and a concomitant decrease in α6B expression (Figure S2E). Therefore, we hypothesized that TRPC6 suppresses ESRP1 and, consequently, sustains the α6B splice variant that promotes stemness and chemoresistance. To test this hypothesis, we first depleted ESRP1 expression in TRPC6 knockdown cells (TE3 and CAL-51) (Figure S2C) and observed that loss of ESRP1 was sufficient to rescue self-renewal compared to control cells (Figures S3C and S3D). Moreover, expression of the α6B variant in the TRPC6 knockdown cells (Figure S4C) was sufficient to rescue self-renewal (Figures S4A and S4B). Importantly, the exogenously expressed α6B is functional because these transfected cells were able to bind to the α6B ligand laminin 511³⁴ (Figure S4D).

We extended our analysis to the CAL-51-S and CAL-51-R populations described in Figure 2 based on the assumption that cells that persist after chemotherapy have increased TRPC6 and α6B expression but diminished ESRP1 expression. Using immunoblotting to assess protein expression, we verified that the CAL-51-R cells had increased TRPC6 expression compared to the CAL-51-S cells (Figure 3E). We also observed that the CAL-51-R cells had higher expression of α6B and lower levels of ESRP1 and α6A than the CAL-51-S cells (Figure 3E). Inhibition of TRPC6 activity in CAL-51-R cells caused an increase in ESRP1 and the α6A splice variant and a concomitant decrease in α6B (Figure 3F). These data indicated that the TRPC6/ESRP1 axis, which regulates α6 integrin splicing, contributes to the survival of persister cells and that it can be a therapeutic target to enhance chemosensitivity. In support of this possibility, exogenous expression ESRP1 in the CAL-51-R cells decreased α6B expression (Figure 3G) and, importantly, increased their sensitivity to paclitaxel (Figure 3H). Conversely, expression of the α6B variant, but not of α6A, in CAL-51-S cells that has been depleted of integrin α6 using CRISPR (Figure S4E) increased resistance to paclitaxel (Figure 3I).

To gain insight into how TRPC6 represses ESRP1 expression, we used a dataset that highlights transcription factor and gene interactions based on compiled ENCODE chromatin immunoprecipitation (ChIP) data.³⁵ This analysis identified TEAD4 as one such factor. Given that the TEAD family of transcription factors primarily associate with the Hippo effectors YAP and TAZ,³⁶ this observation indicated that TRPC6 may contribute to TAZ activation and the regulation of ESRP1. Indeed, knockdown of TRPC6 or pharmacological inhibition of its activity using BI-749327 caused a significant decrease in TAZ nuclear localization (Figures S5A and S5B) and in the expression of the TAZ target genes CTGF and ANKRD1 (Figures S5C and S5D). Subsequently, we used small interfering RNAs (siRNAs) to diminish TAZ (Figure S2D) and observed a concomitant increase in ESRP1 expression (Figure S5E). We also knocked down the expression of TEADs 1–3-4 using shRNA (Figure S5G) and observed a similar increase in ESRP1 protein levels (Figure S5F). To investigate the TRPC6-mediated repression of ESRP1 by TAZ more rigorously, we expressed constitutively active TAZ (4SA-TAZ) in TRPC6 knockdown cells and observed that active TAZ was sufficient to repress ESRP1 in these cells (Figure S5H).

Given that cell contractility can activate TAZ by a mechanism that involves Rho GTPases³⁷ and previous work demonstrating that TRPC6 can activate RhoA in a calcium-dependent manner,³⁸ we evaluated the contribution of RhoA to the repression of ESRP1 expression. Initially, we assessed RhoA activation in TNBC cells in which TRPC6 expression had been knocked down and then rescued with either WT TRPC6 or the pore mut TRPC6^G757D. The data obtained substantiate the finding the TRPC6 activates RhoA in a calcium-dependent manner (Figure S6A). Subsequently, we assessed a causal role for RhoA-mediated contractility in TAZ activation and ESRP1 repression using the ROCK inhibitor (Y26739). This inhibitor caused a significant decrease in TAZ activation, as assessed by expression of TAZ target genes (Figure S6B), as well as an increase in ESRP1 expression (Figures S6C and S6D). To validate that RhoA activation downstream of TRPC6 causes repression of ESRP1, we transfected a constitutively active RhoA (RhoV14-GST) in TRPC6 knockdown cells and observed that the expression of active RhoA was sufficient to reduce ESRP1 levels (Figure S6E). Together, these data support the hypothesis that TRPC6-mediated RhoA activation contributes to TAZ activation and the consequent repression of ESRP1.

TRPC6-regulated α6 integrin splicing sustains persister cells by repressing Myc

In pursuit of the mechanism by which TRPC6-regulated α6 integrin splicing promotes chemoresistance, we were informed by the report of Dhimolea et al. that repression of Myc has a causal role in sustaining persister cells in breast cancer.³⁹ In support of a role for Myc repression in our mechanism, we observed that Myc expression is significantly lower in CAL-51-R cells compared to CAL-51-S cells (Figure 4A). We also observed that TRPC6 was significantly enriched in the chemoresistant organoid models used by Dhimolea et al. (Figure 4B). Moreover, TRPC6 knockdown in TE3 and CAL-51 cells (Figures 4C and 4D) or inhibition of its function using BI-749327 increased Myc mRNA (Figure 4E) and protein expression (Figure 4F). A critical finding is this regard is that the PDX tumors that were treated with BI-749327 had higher levels of Myc than control tumors, as assessed by immunoblotting of tumor extracts (Figure 4F). To verify a causal role for Myc repression in chemoresistance, we expressed exogeneous Myc in CAL-51-R cells (Figure 4G, right) and observed a significant increase in their sensitivity to paclitaxel (Figure 4G). Conversely, pretreatment of the cells with the Myc inhibitor 10074-G5 increased resistance to paclitaxel (Figure 4H). Inhibition of Myc activity by 10074-G5 is evidenced by the downregulation of its target gene, CDC25A (Figure 4I).

Figure 4. TRPC6-regulated α6 integrin splicing sustains persister cells by repressing Myc.

(A) Myc mRNA expression was quantified by qPCR in CAL-51-sensitive (CAL-51-S) and CAL-51-R cells.

(B) TRPC6 expression (log2fold) was analyzed from a published dataset of chemoresistant (doxorubicin) organoid models that were derived from patient tumors or cell lines (see Dupont et al.³⁷).

(C and D) Myc mRNA expression was quantified in TE3 (C) or CAL-51 (D) cells that had been stably transfected with either a control (shCTRL) or TRPC6 shRNAs (shTRPC6–1, shTRPC6–2).

(E) TE3 cells were treated with either DMSO or BI (10 μM) for 24 h. Myc mRNA expression was quantified by qPCR in TE3 cells.

(F) Myc protein expression by immunoblotting was detected in TE3 cells treated with either vehicle or 10 μM BI for 24 h. Freshly harvested PDX tumors that had been treated with either vehicle or BI (see Figure 2H) were used to extract protein and RNA. Protein lysates were immunoblotted with a Myc antibody. Densitometric values are shown below the immunoblot.

(G) CAL-51-R cells were transfected with either vector alone or a Myc expression vector, and their sensitivity to increasing concentrations of PTX was assessed by quantifying cell viability (left). Immunoblot showing Myc protein levels in CAL-51-R cells transfected with either an empty vector or Myc overexpression plasmid (right).

(H) CAL-51 parental cells that has been pretreated with either DMSO or 10074-G5 (50 μM) for 24 h were assessed for their sensitivity to increasing concentrations of PTX, either alone or in combination with 1007-G5 (50 μM) for an additional 24 h.

(I) Expression of the Myc target gene CDC25A was quantified by qPCR in CAL-51 cells treated with either DMSO or 10074-G5 (50 μM).

The Myc expression levels in (A), (C), (D), and (E) represent the mean ± SD of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

TRPC6 promotes quiescence

Based on the finding that Myc sustains persister cells by inducing quiescence, we sought to determine whether TRPC6 promotes quiescence. For this purpose, we used SETD4 as a functional readout of quiescence because it has been shown to mark quiescent populations in subpopulations of breast cancers.^40,41 We found that TRPC6 knockdown in TE3 cells resulted in a significant decrease in SETD4 expression (Figure 5A). Further, the PDX tumors that were treated with BI-749327 had significantly less SETD4 mRNA expression than control tumors (Figure 5B). To obtain additional evidence that TRPC6-mediated signaling has a causal role in quiescence, we used a reporter consisting of an inactive CDK-binding domain of p27 fused to an mVenus tag.⁴² This reporter has been used to identify p27⁺ quiescent cells in breast cancer.⁴³ Using this reporter, we observed that CAL-51 and HCC-1806 cells treated with the TRPC6 inhibitor exhibited a significant decrease in the p27⁺ population (Figures 5C and 5F), indicating a shift toward a more proliferative phenotype. We also sorted the p27^high and p27^low populations from both CAL-51 and HCC-1806 cells that had been transfected with the p27-mVenus reporter and observed a significant enrichment of TRPC6 and SETD4 expression in the p27^high population compared to the p27^low population (Figures 5D and 5G). In contrast, the p27^low population had higher expression of ESRP1 (Figures 5D and 5G), a result that is consistent with our previous data (see Figure 3). We also observed a significant decrease in mammosphere formation in the p27^high population upon TRPC6 inhibition (Figures 5E and 5H). Based on these data, we hypothesized that the p27^high cells are more resistant to paclitaxel than the p27^low cells. To test this hypothesis, we sorted the p27^high and p27^low populations from CAL-51 cells and observed that the p27^high cells are significantly more resistant to paclitaxel than p27^low cells (Figure 5I).

Figure 5. TRPC6 promotes quiescence.

(A) SETD4 mRNA expression was quantified in CAL-51 cells that had been stably transfected with either a control (shCTRL) or TRPC6 shRNAs (shTRPC6–1,shTRPC6–2).

(B) SETD4 mRNA expression was quantified in a TNBC PDX that had been treated with either vehicle or BI.

(C) CAL-51 cells expressing a p27-mVenus reporter were treated with either vehicle or BI for 24 h. Cells were detached and processed for flow cytometry (mean fluorescence indicated by the number on the top left). Bar graph plotted using mean fluorescence intensity is shown on the right.

(D) Comparison of TRPC6, SETD4, and ESRP1 mRNA expression between the CAL-51 p27^high and p27^low populations.

(E) The p27^high population of CAL-51 cells was treated with either vehicle or BI for 24 h, and their ability to form mammospheres was assessed.

(F) HCC-1806 cells expressing the p27-mVenus reporter were treated with either vehicle or BI for 24 h and processed for flow cytometry. Bar graph of mean fluorescence intensity is shown on the right.

(G) HCC-1806 cells were transfected with a p27-mVenus reporter and sorted into p27^high and p27^low populations. The expression of TRPC6, SETD4, and ESRP1 mRNAs was compared between these populations.

(H) Mammosphere formation was assayed in the p27^high population of HCC-1806 cells that had been treated with either DMSO or BI (10 μM).

(I) The sensitivity of the p27^high and p27^low populations of CAL-51 to increasing concentrations of PTX was assessed. The SETD4 expression data shown represent the mean ± SD of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

TRPC6-regulated α6 integrin splicing represses Myc

Given our data that TRPC6-mediated splicing of the α6 integrin has a casual role in chemoresistance (see Figure 3), we investigated whether the α6B splice variant has a causal role in repressing Myc. Indeed, exogenous expression of the α6B splice variant in TRPC6 knockdown cells (Figure S4C) was sufficient to rescue Myc suppression in both TE3 and CAL-51 cells (Figures 6A and 6B). To investigate the mechanism by which α6B represses Myc, we pursued our previous finding that α6B signaling activates TAZ in breast CSCs³⁴ based on our observation that constitutively active TAZ represses Myc (Figure 6C). This possible mechanism was strengthened by our observation that exogenous expression of α6B in TRPC6 knockdown cells rescues the loss in TAZ activation and causes a suppression in ESRP1 expression (Figure 6D). To implicate TAZ directly in Myc repression, we found that knockdown of TAZ (Figure S2D) caused a significant increase in Myc expression (Figures 6E and 6F). Most importantly, to strengthen this connection we expressed a constitutively active TAZ (4SA) in our TRPC6 knockdown (KD) cells. Indeed, expression of a constitutively active TAZ construct was sufficient to rescue Myc expression in the KD cells to control levels (Figure 6G). Taken together, our data indicate that TRPC6 and the α6B integrin synergize to activate TAZ and, consequently, repress Myc.

Figure 6. TRPC6-regulated α6 integrin splicing contributes to Myc suppression.

(A and B) Either TE3 (A) or CAL-51 (B) TRPC6 knockdown cells (shTRPC6–1, shTRPC6–2) were retransfected with an α6B expression plasmid (α6B-GFP) followed by knockdown of the endogenous α6 integrin (shITGA6). Myc mRNA expression was quantified by qPCR.

(C) Myc expression in the non-CSC and CSC populations described in Figure 1A.

(D) TE3 cells that had been stably transfected with either a control (shCTRL) or TRPC6 shRNAs (shTRPC6–1, shTRPC6–2) were retransfected with an α6B expression plasmid (α6B-GFP) followed by knockdown of the endogenous α6 integrin (shITGA6). Expression of ESRP1, CTGF, and ANKRD1 mRNAs was quantified by qPCR.

(E and F) Myc mRNA expression was quantified in TE3 (E) and CAL-51 (F) cells that had been transfected with siCTRL or siRNA targeting TAZ (siTAZ).

(G) Myc mRNA expression was quantified in CAL-51 cells transfected with either shCTRL or TRPC6 shRNAs (shTRPC6–1, shTRPC6–2) that were re-transfected with a constitutively active TAZ-4SA plasmid (4SA).

The Myc expression data shown in (A), (B), (E), and (F) represent the mean ± SD of two independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

DISCUSSION

In this study, which is summarized in the schematic shown in Figure 7, we demonstrate that a specific cation channel, TRPC6, which is known to regulate calcium entry in TNBC,⁴⁴ is intimately associated with a CSC population and has a causal role in stemness and chemoresistance. The most novel aspect of the study is the mechanism by which TRPC6-mediated calcium entry promotes these functions. Our data revealed that TRPC6 is at the nexus of a signaling network that regulates mRNA splicing. More specifically, we discovered that TRPC6-mediated calcium entry sustains a splice variant of the α6 integrin α6B. TRPC6 and integrin α6B signaling function in tandem to repress Myc and, consequently, sustain a quiescent state that is resistant to chemotherapy. In other terms, tumor cells that resisted chemotherapy in our study were dependent on the TRPC6/α6B signaling axis for their survival. Importantly, we demonstrate that it is possible to sensitize TNBC organoids and PDX tumors to chemotherapy using a specific inhibitor of TRPC6, which disrupts this signaling axis.

Figure 7. Summary schematic.

This schematic depicts the key findings in this study. Treatment of a heterogeneous tumor with chemotherapy results in the survival of persister cells that are resistant to chemotherapy. These cells have high TRPC6 expression levels and stem cell properties. TRPC6 sustains the expression of α6B in this population through suppression of the splicing factor ESRP1. Enrichment of α6B in this population drives chemoresistance through a TAZ-mediated MYC suppression. This suppression of Myc maintains these cells in a quiescent state to survive chemotherapeutic stress.

A major finding in this study is that TRPC6-mediated calcium signaling regulates alternative mRNA splicing. Alternative splicing is a key mechanism for ensuring proper developmental programs that demand vast genetic diversity. Tumors hijack alternative splicing programs that contribute to the intratumor transcriptomic heterogeneity, EMT, and therapy resistance.^45,46 Although previous studies have implicated Ca²⁺ signaling in the regulation of splicing,⁴⁷ the mechanisms reported are distinct from ours.^48–50 These studies focused largely on how calcium-mediated signaling impacts splicing by directly binding RNA elements of specific genes. For example, Ca²⁺/calmodulin-dependent kinase IV (CaMKIV) regulates splicing of the stress axis hormone-regulated (STREX) exon of the calcium-activated potassium channel in neurons by directly binding to the CaMKIV-responsive RNA element.^49,50 In contrast, we demonstrate that TRPC6 regulates the expression of a specific splicing protein, ESRP1. ESRP1 is of interest because it is known to differentially enrich for splice variants that play a crucial role in cell-fate decisions specifically associated with epithelial cell fate.^31,32,51 The focus of our study was the α6 integrin, which exists as two splice variants, α6A and α6B,^52,53 which are regulated by ESRP1. Specifically, the UGC-rich recognition motif on the α6 mRNA binds ESRP1 and, consequently, through exon exclusion, generates α6A.⁵¹ Previous studies by our group established that the α6B variant promotes stemness, a property that is not shared by the α6A variant, and that the relative expression of ESRP1 determines expression of these variants and, consequently, regulates stem cell properties.³³ The novelty of the current study is that we link this integrin splicing mechanism to a specific calcium channel and a calcium-mediated signaling pathway. Although we focused on the splicing of the α6 integrin in this study, ESRP1 is known to regulate the splicing of other mRNAs that contribute to stemness. Of note, ESRP1 enriches for the CD44V splice variant relative to the CD44S variant that has a causal role in stemness.³² Consistent with this observation, we observed that TRPC6-mediated suppression of ESRP1 maintains high levels of CD44S compared to CD44V (Figure S2F). This ability of TRPC6 to orchestrate a differential splicing pattern reveals a novel, Ca²⁺-dependent mechanism to generate transcriptomic diversity that drives stemness in cancer. It is worth noting that these findings may also have relevance for non-cancer diseases that involve TRPC6. Specifically, defects in the TRPC6 gene are associated with stage 5 chronic kidney disease and are often mediated by deregulated expression or activity of adhesion proteins such as α3β1 integrins. These changes alter laminin binding and downstream actin cytoskeletal arrangement, ultimately leading to detachment of podocytes from basement membrane.^54,55

Arguably, the most consequential aspect of our data is the involvement of the TRPC6-mediated splicing mechanism in chemoresistance. Our findings revealed that TNBC cells and organoids that persist after paclitaxel treatment are characterized by high expression of TRPC6 and α6B and low expression of ESRP1. Importantly, we also demonstrated that each of these molecules has a causal role in sustaining the survival of these persister cells. These data are substantiated by our analysis of a publicly available dataset,³⁹ which revealed a correlation between TRPC6 expression and chemoresistance (Figures 2B and 2F). In pursuit of the mechanism involved, we were informed by a seminal study that treatment-persistent tumor cells adopt a state that resembles embryonic diapause characterized by suppressed Myc activity.³⁹ Of particular interest to us was their finding that Myc upregulation sensitizes persister cells to chemotherapy. The data in our study provide a novel mechanism for how Myc can be repressed that involves the ability of the TRPC6/α6B signaling axis to repress Myc by activating TAZ. These data also provide a mechanism for how TRPC6/α6B enable cells to persist after paclitaxel treatment. Most likely, the repression of Myc by TRPC6 maintains cells in a more quiescent state that is less amenable to paclitaxel-mediated killing.

Our study adds significantly to the existing literature on the mechanism by which calcium signaling regulates the function of CSCs. Much of the literature on calcium signaling in cancer has focused on the regulation of transcription mediated by calcium-dependent transcription factors, most notably nuclear factor of an activated T cell (NFAT).⁵⁶ Interestingly, however, our data do not support a role for NFAT in TRPC6-mediated regulation of CSC function (Figure S2G). This observation is consistent with the report that calcineurin/NFAT are dispensable for promoting a partial EMT by Ca²⁺ influx.⁷ Rather, our data reveal a more biophysical mechanism that involves the ability of TRPC6-mediated RhoA activation and the consequent increase in contractility to activate TAZ and repress ESRP1. Although previous studies have demonstrated the ability of TRPC6 to activate RhoA³⁸ and the ability of RhoA-mediated contractility to activate TAZ,³⁷ the regulation of TAZ activation and its impact on mRNA splicing by Ca²⁺ signaling is novel and significant.

In summary, the data we report reveal a novel mechanism that contributes to stemness and chemoresistance in TNBC that involves the regulation of integrin α6 mRNA splicing by TRPC6-mediated Ca²⁺ signaling. The TRPC6/α6B signaling axis enables TNBC cells to persist after paclitaxel treatment by a mechanism that involves the repression of Myc. Our ability to sensitize TNBC organoids in vitro and PDX tumors in vivo to chemotherapy using a specific inhibitor of TRPC6 suggests that this approach could be effective to mitigate the aggressiveness of TNBC, especially for tumors that have become resistant to standard-of-care chemotherapy.

Limitations of the study

The evidence that we provide to implicate TRPC6-mediated calcium entry in stemness and chemoresistance is based on data obtained using a pore mut of TRPC6 (G757D), as well as BAPTA, which chelates calcium. Given that TRPC6 is a non-selective cation channel,¹⁴ more data to implicate TRPC6-mediated calcium entry and the nature of this calcium signaling would strengthen the study. These studies could include the use of a TRPC6 gain-of-function mut (R175Q).⁵⁷ Although we provide compelling data that targeted inhibition of TRPC6 using BI-749327 sensitizes a TNBC PDX to chemotherapy, the use of additional PDX models, as well as other chemotherapeutics, would be beneficial. In addition, it would be important to assess the impact of TRPC6 inhibition on metastasis.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Example: Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dr. Arthur Mercurio (arthur.mercurio@umassmed.edu).

Materials availability

Plasmids and cell lines generated in this study will be freely available upon request.

Data and code availability

• The RNA-Seq data performed on TE3 cells treated with Vehicle or BI-749327 (12 h) has been uploaded to GEO: GSE242592.

• The code we used to analyze our RNA-Seq data is available here: 10.5281/zenodo.8342563 (link: https://zenodo.org/record/8342563).

• Any additional information required to reanalyze the data reported in this work paper is available from the lead contact upon request.”

METHOD DETAILS

Cells and reagents

HMLER cells were provided by R. Weinberg (Massachusetts Institute of Technology). MDA-MB-231 and HCC1806 cells were purchased from the American Type Culture Collection, and the TE3 variant of MDA-MB-231 cells was provided by S. Tavazoie (Rockefeller University). CAL-51 cells were obtained from the Leibniz Institute DSMZ, Germany. The TRPC6 inhibitor BI-749327 was purchased from MedChem Express (Catalog no. HY-111925). The paclitaxel was purchased from Selleckchem (Catalog no. S1150). The NFAT inhibitor was purchased from Cayman Chemicals (Catalog no. 13855). The following Abs that were used for flow cytometry were purchased from BioLegend: CD24–allophycocyanin (APC; ML5, BioLegend), CD44-FITC (IM7, BioLegend) and anti-rabbit FITC (MRM-47, BioLegend), anti-Rabbit-555 (Invitrogen-A32732), TRPC6 (18236–1-AP, Proteintech). For immunoblotting, the following Abs were used: Tubulin [Cell Signaling Technology (3873)], Beta-Actin (3700S, Cell Signaling Technology)TRPC6 (ACC-017, Alomone Labs), GAPDH (14C10) [Cell Signaling Technology 2118S], α6A (1A10) (MAB1356, Millipore), and α6B (6B4) (MAB1358, Millipore), Myc (Y69) (32072, Abcam), ESRP1 (PA5–25833, Thermofisher), HA (3724, Cell Signaling Technology), and GST-HRP (5475, Cell Signaling Technology). For immunofluorescence the following antibodies were used: TRPC6 (18236–1-AP, Proteintech), TAZ (clone M2–616, 560235, BD-Pharmingen).

shRNAs and expression constructs

The following lentiviral shRNAs were obtained from our core facility: TRPC6 (TRCN0000431016, TRCN0000044106). Lentiviral shRNAs specific for ESRP1 (pLV[shRNA] NeoU6>hESRP1[shRNA]) were designed and purchased from Vector Builder (Vector ID: VB220524–1274mnz, and VB220524–1275vdm). TEAD 1/3/4 shRNA were provided by J. Mao (University of Massachusetts Chan Medical School). A TRPC6-HA lentiviral plasmid (pLV[Exp]-Bsd CMV>hTRPC6[NM_004621.6]/HA) was also designed and purchased from Vector-Builder (Vector ID: VB191024–1847whk). A site-directed mutagenesis kit (Q5 Site-Directed Mutagenesis Kit, E0554S) (New England BioLabs) was used to make the desired TRPC6 G757D pore-mutant lentiviral vector based on the backbone (cDNA TRPC6-HA vector). HA-ESRP1 constructs were provided by Dr. Chonghui Cheng (Baylor College of Medicine). pRC-α6A or α6B plasmids were prepared as described (Shaw et al., 1993⁵⁸) and subcloned into the lentiviral vector pCDH. We used the Myc expression plasmid as described in Goel et al., 2016. The quiescent reporter plasmid pCDH-EF1-mVenus-p27K was bought from Addgene (plasmid #176651). Lipofectamine 3000 (Thermo Fisher Scientific) was used for plasmid transfections. SMARTpool siRNAs against TAZ were purchased from Horizon Discovery and used to transfect the cells as per manufacturer’s instructions at a total concentration of 25 nM.

ITGA6 CRISPR knockout system: To generate ITGA6 knockout cells, we used Alt-R CRISPR-Cas9 System (IDT). The following gRNAs were used: (Human sgITGA6–1: CCTTCGGGAGGACAACGTGATC, sgITGA6–2: TTTCCAGTTAATAAGTACCCG). The following reagents were purchased from IDT: Alt-R CRISPR crRNA (2 nmol), CRISPR-Cas9 tracrRNA (Cat. 1072532), and Cas9 Nuclease (Alt-RTM S.p. Cas9 Nuclease 3NLS, Cat. 1074181) and were used to assemble Cas9:crRNA:tracrRNA RNP complex. The RNP complexes were transfected in 1×10⁶ Cal51 cells using Nucleofector Device (Amaxa) with program X-001. After five days in culture, pooled ITGA6 negative cells were sorted using flow cytometry with the ITGA6-APC antibody (Biolegend, Cat. 324208).

Cell-based assays

For mammosphere assays, single cell suspensions of TE3 and HCC1806 cells (10³ cells) or CAL-51 cells (5×10²) were plated on ultralow attachment plates (24 well; Corning Costar; catalog no. CLS3473) in serum-free, mammary epithelial cell growth medium (MEGM) supplemented with EGF (2 ng/mL), bFGF 2 ng/mL, heparin (4 ng/mL), methylcellulose (1%), and B27 supplement diluted 1:50.³² The mammospheres were counted after 5 days using Celigo Imaging Cytometer (Nexcelom) and processed for secondary passage as we described previously.²³ For mammosphere assays using PDXs, tumors were chopped into fragments with scalpel blades and digested with 2 mg/mL of collagenase (Sigma, C0130) and 100 mg of 1 mg/mL hyaluronidase (Sigma, H3506) for 4–6 h at 37°C with shaking. Tissue fragments were vortexed gently every 15–30 min. The single-cell suspension was then passed through a 40-μm cell strainer, centrifuged at 2000 rpm for 5 min and washed three times with PBS. The washed cells were plated as mentioned above for the respective serial passage or limiting dilution mammosphere assays.

For flow cytometry, cells were incubated with primary antibodies and the appropriate secondary antibodies following the manufacturer’s instructions. To assay cell viability, cells were seeded at a density of 5×10³ cells per well. After 24 h, cells were treated with either BI-749327 or paclitaxel as described in the figure legends and viability was assessed using crystal violet staining. The absorbance measured at 600 nm was normalized to the saline solution control. For immunofluorescence microscopy, cells were cultured on glass-bottom dishes (Ibidi #81218–200), fixed with paraformaldehyde (2%) for 20 min and permeabilized with Triton X-100 (0.5%) for 5 min and then blocked in buffer (PBS, 5% normal goat serum, 0.1% Triton X-100, and 2 mM sodium azide) for 30 min. The cells were then incubated with primary antibodies in blocking buffer for 1 h at room temperature, washed with Triton X-100 (0.1%) in PBS and then stained with secondary antibodies for 1 h at room temperature. After washing, the cells were mounted in Vectashield with DAPI (Vector Laboratories, UX-93952–24). Images were captured at x20, x40 and x60 magnification using a confocal microscope (Zeiss).

Biochemical assays

For immunoblotting, adherent cells were washed with PBS and scraped on ice in radioimmunoprecipitation assay (RIPA) buffer, which was supplemented with protease and phosphatase inhibitors (Roche). Laemmli 6X SDS sample buffer (BP-111R, Boston BioProducts) was added to each sample, which were boiled for 10 min prior to SDS-polyacrylamide gel electrophoresis. Immunoblotting primary antibodies were used in the following dilutions: TRPC6 ACC-017 (Alomone Labs) 1:1000, Tubulin and GAPDH 1:2000, ESRP1 (Thermofisher) 1:2000, α6A 1:1000, α6B 1:1000, HA 1:2000. RhoA activity was assessed using a GST fusion protein containing the Rho-binding domain of ROCK (RBD) as previously described^59,62

mRNA quantification was accomplished using an RNA isolation kit (BS88133, Bio Basic Inc.), and complementary DNAs (cDNAs) were produced using an Azura cDNA synthesis kit (AZ-1996, Azura Genomics). Azura View GreenFast qPCR Blue Mix LR was used as the qPCR Master Mix (AZ-1996, Azura Genomics).

RNA sequencing

RNA was extracted from the indicated cells using a Qiagen RNeasy Micro Kit (74004) and sent to Quick Biology for quantification, sequencing, and analysis.

Quick Biology sequencing method

Library for RNA-Seq was prepared according to KAPA Stranded mRNA Hyper prep polyA selected kit with 201–300 bp insert size (KAPA Biosystems, Wilmington, MA) using 250 ng total RNAs as input.

Final library quality and quantity was analyzed by Agilent Technologies 4200 station and Qubit 3.0 (Thermo Fisher Scientific Inc, Waltham, MA) Fluorometer. 150 bp paired-end reads were sequenced on Illumina HiseqX (Illumnia Inc., San Diego, CA).

Each sample had a sequencing depth of 20–30 million. RNASeq analysis was performed with OneStopRNAseq workflow.⁶³ Pairedend reads were aligned to human primary genome hg38, with star_2.5.3a,⁶⁴ annotated with GENCODE GRCh38.p12 annotation release 34.⁶⁵ Aligned exon fragments with mapping quality higher than 20 were counted toward gene expression with feature-Counts_1.5.2.⁶⁶ Differential expression (DE) analysis was performed with DESeq2_1.20.0.⁶⁷ Within DE analysis, ‘ashr’ was used to create log2 Fold Change (LFC) shrinkage for each comparison.⁶⁸ Significant DE genes (DEGs) were filtered with the criteria FDR <0.05. Heatmaps were created with Prism. Gene set enrichment analysis were performed with GSEA.⁶⁹

Patient derived organoids and chemoresistance models

Freshly resected biopsies from de-identified TNBC patients were received from UMass Cancer Center Tumor Bank. The tumor tissue was digested utilizing the tissue dissociation kit (Miltenyi Biotech) and gentleMACS Dissociator. The dissociated tumor was washed 3X with PBS and embedded into reduced growth factor basement membrane extract (Cultrex). These TNBC tumor derived organoids were cultured in organoid media as described previously.⁷⁰ The paclitaxel resistant CAL-51 cells and TNBC organoids were generated by culturing them in increasing concentrations of paclitaxel (1–20nM) for 6 weeks.

Animal studies

For the limiting dilution experiment, CAL-51 cells were suspended in 35 ml of PBS and injected into the mammary fat pad of 6–8-week-old NOD.Cg-Prkdc^scid IL2rg^tm1Wjl (abbreviated as NSG) mice. Tumor occurrence in each group was noted and the results were analyzed using ELDA.²⁶ Mice were monitored for 12–15 weeks, and tumor volume was measured using calipers. Once the tumors reached 1cm³, the mice were euthanized, and the tumors collected for histological analysis. For therapeutic studies, a PDX (HCI028) obtained from the Huntsman Cancer Institute was implanted orthotopically into the mammary fat pad of 6–8-week-old NSG mice. The mice were divided into 4 groups of 5 mice each. When the tumor volume reached 100mm³, mice were treated as described in the legend to Figure 2. All mouse procedures were done under the guidance of the University of Massachusetts Medical School Institutional Animal Care and Use Committee in accordance with the institutional and regulatory guidelines.

QUANTIFICATION AND STATISTICAL ANALYSIS

Student’s t test was used for comparison between two groups. Multiple group comparisons were performed using one-way analysis of variance (ANOVA). Statistical tests were carried out using GraphPad Prism version 9.0, and a p value of less than 0.05 was considered significant. The bars in graphs represent mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
APC anti-human CD24	BioLegend	Cat# 311118; RRID: AB_2072735
FITC anti-mouse/human CD44	BioLegend	Cat# 103006; RRID: AB_312957
Goat anti-Rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor Plus 488	Invitrogen	Cat# A32731; RRID: AB_2633280
Goat anti-Rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor Plus 555	Invitrogen	Cat# A32732; RRID: AB_2633281
Rabbit polyclonal Anti-TRPC6 Antibody	Alomone Labs	Cat# ACC-017; RRID: AB_2040243
Rabbit polyclonal anti-TRPC6	Proteintech	Cat# 18236–1-AP; RRID: AB_10859822
Mouse monoclonal anti-β-Actin (8H10D10)	Cell Signaling Technology	Cat# 3700; RRID: AB_2242334
Rabbit Monoclonal Anti-GAPDH Antibody, Unconjugated, Clone 14C10	Cell Signaling Technology	Cat# 2118; RRID: AB_561053
Mouse Monoclonal Anti-alpha-Tubulin Antibody, Unconjugated, Clone DM1A	Cell Signaling Technology	Cat# 3873; RRID: AB_1904178
Rabbit polyclonal anti-ESRP1 Antibody	Thermo Fisher Scientific	Cat# PA5–25833; RRID: AB_2543333
Mouse monoclonal Anti-Integrin alpha6A, cytoplasmic domain, clone 1A10	Millipore	Cat# MAB1356; RRID: AB_94179
Mouse monoclonal Anti-Integrin alpha6B, cytoplasmic domain, clone 6B4	Millipore	Cat# MAB1358; RRID: AB_94180
Rabbit monoclonal anti-c-Myc antibody [Y69]	Abcam	Cat# ab32072; RRID: AB_731658
Rabbit monoclonal anti-RhoA (67B9) mAb	Cell Signaling Technology	Cat# 2117; RRID: AB_10693922
Rabbit monoclonal anti-HA-Tag (C29F4) mAb	Cell Signaling Technology	Cat# 3724; RRID: AB_1549585
Rabbit monoclonal anti-GST (91G1) mAb (HRP Conjugate)	Cell Signaling Technology	Cat# 5475; RRID: AB_10707323
Mouse monoclonal anti-TAZ	BD-Pharmingen	Cat# 560235; RRID: AB_1645338
Mouse monoclonal APC anti-human CD326 (Ep-CAM)	BioLegend	Cat# 324208; RRID: AB_756082
Bacterial and virus strains
NEB 5-alpha Competent E.coli	New England Biolabs	Cat# C2987
Biological samples
Breast cancer derived Organoids	As described in Walker et al.23	N/A
PDX (HCI028)	Huntsman Cancer Institute	N/A
Chemicals, peptides, and recombinant proteins
Animal-Free Recombinant Human EGF	Peprotech	Cat# AF-100–15-1mg
MEGM bullet kit	lonza	Cat# cc-3150
Recombinant Human FGF-basic	Peprotech	Cat# 100–18B
B-27 supplement (50X)	Gibco	Cat# 17504–001
Insulin	Millipore Sigma	Cat# I1882–100MG
BI-749327	MedChem Express	Cat# HY-111925
Y27632 2HCL	Selleckchem	Cat# S1049
10074-G5	Selleckchem	Cat# S8426
NFAT Inhibitor	Cayman Chemical	Cat# 13855
Critical commercial assays
AzuraQuant cDNA synthesis kit	Azura genomics	Cat# AZ-1995
AzuraView greenfast qPCR blue mix	Azura genomics	Cat# AZ-2401
Tumor Dissociation Kit, human	Miltenyi BioTec	Cat# 130–095-929
Q5® Site-Directed Mutagenesis Kit	New England Biolabs	Cat# E0554S
Qiagen RNeasy Micro Kit	Qiagen	Cat# 74104
Deposited data
Public data: RNAseq from Dhimolea et al.39	GEO	GEO: GSE162285
RNA-Seq: TE3 cells Vehicle, BI treatment	GEO	GEO: GSE242592
Experimental models: Cell lines
MDA-MB-231-TE3	Kindly provided by Dr. S. Tavazoie, Rockefeller University).	N/A
MDA-MB-231	American Type Culture Collection	Cat# HTB-26
CAL-51	Leibniz Institute DSMZ, Germany	ACC-302
HCC-1806	American Type Culture Collection	Cat# CRL-2335
HMLER	Kindly provided by Dr. R. Weinberg, Massachusetts Institute of Technology	N/A
Experimental models: Organisms
NOD.Cg-Prkdcscid IL2rgtm1Wjl (NSG)	Jackson Laboratory	N/A
Experimental model: Strain of bacteria
DH5 α	NEB	Cat# C2987H
Oligonucleotides
h18S fwd 5'-GTCGCTCGCTCCTCTCCTACT-3'; rev 5' TCTGATAAATGCACGCATCCC-3'	IDT	N/A
hTRPM8 fwd, 5'-GTGAAAGCGACTTGGTGAATTTT-3; rev 5'-GTGGCCTTGGAATCTTTGGTAA-3'	IDT	N/A
hTRPV6 fwd 5'-ACTGACCTCGACTCTCTATGAC-3'; rev 5'-GTGGTGATGATAAGTTCCAGCAG-3'	IDT	N/A
hORAII fwd 5'- GACTGGATCGGCCAGAGTTAC-3'; rev 5'-GTCCGGCTGGAGGCTTTAAG-3'	IDT	N/A
hORAI3 fwd 5'-TGGGTCAAGTTTGTGCCCATT-3'; rev 5'-AGCTGGACTAAGGGAGGTAGC-3'	IDT	N/A
hTRPC6 fwd 5' -TTGAGGATGACGCTGATGTG-3'; rev 5-CCAGATTGAAGGGTACAGGAAG-3'	IDT	N/A
hTRPC6 (G757D) primer for sequencing fwd 5' -AAAGAAACTTGACATTTTAGGAAGTCATG-3'; rev 5'-TCTTCATTTATCTTGTTCATCTC-3'	Genewiz	N/A
hESRPI fwd: 5'-CAGAGGCACAAACATCACAT-3'; rev: 5'- AGAAACTGGGCTACCTCATTGG-	IDT	N/A
hCTGF fwd 5' -CAGCATGGACGTTCGTCTG-3'; rev 5'-AACCACGGTTTGGTCCTTGG 3	IDT	N/A
hANKRD1 fwd 5'-AGTAGAGGAACTGGTCACGG-3' rev 5'-TGTTTCTCGCTTTTCCACTGTT-3'	IDT	N/A
hSETD4 fwd 5'-GGAGAACAAGCCGGATCAGAA-3'; rev 5'-AGCAGGCGCTAAGTTTGAATC-3'	IDT	N/A
hMYC fwd: 5'- CTC AAA GCT GGC CAG TAG AA-3' rev 5'- CGT CAC ACG AAC CGA CAA TA-3'	IDT	N/A
hCDC25A fwd: 5'- TCTGGACAGCTCCTCTCGTCAT-3'; rev: 5'- ACTTCCAGGTGGAGACTCCTCT-3'	IDT	N/A
sgITGA6–1: CCTTCGGGAGGACAACGTGATC, sgITGA6–2: TTTCCAGTTAATAAGTACCCG	IDT	N/A
Alt-R CRISPR crRNA (2 nmol)	IDT
Alt-RTM S.p. Cas9 Nuclease 3NLS	IDT	Cat. 1074181
Recombinant DNA
pLV[Exp]-Bsd CMV>hTRPC6[NM_004621.6]/HA	Vector Builder	VB191024–1847whk
pLV[shRNA] NeoU6>hESRP1[shRNA]	Vector Builder	VB220524–1274mnz, and VB220524–1275vdm
HA-ESRP1	Kindly provided by Dr. Chonghui Cheng (Baylor College of Medicine)	N/A
pRC-alpha6A and pRC-alpha6B	As described Shaw et al.58	N/A
GST-RBD plsamid	As described in Ren et al.59	Addgene, Cat #15247
plasmid pCDH-EF1-mVenus-p27K	As described in Correia et al.60	Addgene, Cat #176651
pCDH-puro-cMyc	As described in Cheng et al.61	Addgene, Cat# 46970
Software and algorithms
GraphPad Prism v 9.3.1	GraphPad Software	N/A
FlowJo Software (for MAC) [software application] Version 10.7.1.	Becton, Dickinson, and Company	https://docs.flowjo.com/flowjo/installation/
Fiji by ImageJ	Open Source	https://imagej.net/software/flji/
ELDA: Extreme Limiting Dilution Analysis	Walter Eliza Hall institute of medical research	https://bioinf.wehi.edu.au/software/elda/
Adobe Illustrator	Adobe Creative Cloud	N/A
Affinity DESIGNER 2	Serif	N/A
Biorender		N/A

Highlights.

• TRPC6 sustains quiescent cancer stem cells in TNBC

• TRPC6-mediated Ca²⁺ maintains stemness by regulating α6 integrin splicing that represses Myc

• TNBC cells that persist after chemotherapy are dependent on TRPC6-mediated Myc suppression

• Targeted inhibition of TRPC6 sensitizes TNBC to chemotherapy

INCLUSION AND DIVERSITY

We support inclusive, diverse, and equitable conduct of research.