Transient Receptor Potential Channel TRPC5 Is Essential for P-Glycoprotein Induction in Drug-Resistant Cancer Cells

Transient receptor potential channel TRPC5 is essential for P-glycoprotein induction in drug-resistant cancer cells

Xin Maa,b,c,1, Yanfei Caia,c,1, Dongxu Hea,1, Chang Zoub, Peng Zhangb, Chun Yin Lob, Zhenyu Xub, Franky L. Chanb, Shan Yub, Yun Chena,c, Ruiyu Zhua,c, Jianyong Leia, Jian Jina,c,2, and Xiaoqiang Yaob,2

aSchool of Medicine and Pharmaceutics, Jiangnan University, Wuxi, Jiangsu 214122, China; bSchool of Biomedical Sciences and Li Ka Shing Institute of Health Science, Chinese University of Hong Kong, Hong Kong; and cInstitute of Health Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, China

Edited by Tak W. Mak, The Campbell Family Institute for Breast Cancer Research, Ontario Cancer Institute at Princess Margaret Hospital, University Health Network, Toronto, Canada, and approved August 21, 2012 (received for review February 21, 2012) 2+ An attractive strategy to overcome multidrug resistance in cancer these TRP channels in cancer may involve changes in [Ca ]i (6). chemotherapy is to suppress P-glycoprotein (P-gp), which is a pump However, to date there is still no report on TRP channel in- overproduced in cancer cells to remove cytotoxic drugs from cells. volvement in multidrug resistance and/or P-gp production. In the present study, a Ca2+-permeable channel TRPC5 was found In the present study we explored the role of TRPC channels in P- to be overproduced together with P-gp in adriamycin-resistant gp production and drug resistance. TRPC5 expression was found to breast cancer cell line MCF-7/ADM. Suppressing TRPC5 activity/ex- be substantially up-regulated in the adriamycin-resistant breast can- pression reduced the P-gp induction and caused a remarkable re- cer cell line MCF-7/ADM. Inhibition of TRPC5 caused a marked versal of adriamycin resistance in MCF-7/ADM. In an athymic nude reduction in P-gp expression, leading to a reversal of adriamycin mouse model of adriamycin-resistant human breast tumor, sup- resistance in MCF-7/ADM cells. In animal models, suppressing pressing TRPC5 decreased the growth of tumor xenografts. Nu- TRPC5 activity/expression reversed the adriamycin resistance of clear factor of activated T cells isoform c3 (NFATc3) was the solid tumors that were formed by MCF-7/ADM cell inoculation. transcriptional factor that links the TRPC5 activity to P-gp production. Together, we demonstrated an essential role of TRPC5– Results NFATc3–P-gp signaling cascade in P-gp induction in drug-resistant Up-Regulation of P-gp and TRPC5 in Adriamycin-Resistant Human cancer cells. Breast Cancer Cells MCF-7/ADM. Adriamycin is a frequently used chemotherapeutic drug in the treatment of breast cancer (7). serious problem for cancer chemotherapy is the development Adriamycin binds to DNA and thereby blocks DNA replication Aof multidrug resistance in tumor cells. Many tumor cells and transcription, causing cytotoxicity to tumor cells (8). Adria- overproduce a 170-kDa protein named P-glycoprotein (P-gp) dur- mycin-resistant human breast cancer cells (MCF-7/ADM) were ing treatment. P-gp, coded by mdr1 gene, functions to pump dif- obtained by treating MCF-7/WT cells with stepwise increasing ferent classes of cytotoxic drugs from tumor cells, rendering the concentrations of adriamycin over 8 mo. Western blot analysis tumor cells resistant to multiple chemotherapeutic drugs (1). P-gp was performed to determine the P-gp protein expression level. A ∼ overproduction has been reported in many tumors and in vitro-se- very high level of P-gp expression (molecular mass 170 kDa) was lected, drug-resistant cell lines (1). In clinical medicine, an attractive detected in MCF-7/ADM cells, whereas only a low level of P-gp A approach to overcome the multidrug resistance in cancer chemo- expression was detected in its parental line MCF-7/WT (Fig. 1 ). therapy is to inhibit P-gp activity (1). In this regard, many P-gp Expression of TRPC proteins was examined. TRPC5 expres- ∼ – antagonists (also named P-gp modulators) have been developed sion (molecular mass 110 120 kDa) was substantially higher in B to inhibit P-gp activity. These modulators may be of clinical im- MCF-7/ADM cells than in MCF-7/WT cells (Fig. 1 ). Immu- portance, because their coadministration with chemotherapeutic nostaining showed that TRPC5 was mostly localized at cell surface drugs has the potential to improve drug uptake into the P-gp– and in some puncta or vesicular structures near the cell surface overproducing tumor cells, thereby reversing the multidrug re- (Fig. S1). The expression of several other TRPC isoforms, in- sistance of tumor cells (1). Unfortunately, most of these drugs are cluding TRPC1, TRPC3, TRPC4, and TRPC6, showed no dif- either too toxic or induce intolerable pharmacokinetic interactions ference between MCF-7/ADM and MCF-7/WT cells (Fig. S2). ﬁ TRPC2 is a pseudogene in human, therefore it was not studied. In (1). Their clinical use is limited. It is highly desirable to nd new β agents that are more effective and less toxic. Western blots, -tubulin was used as an internal control to ensure 2+ 2+ that a similar amount of proteins was loaded onto different gel Cytosolic Ca ([Ca ]i) is an important signal for transcrip- A B tional regulation of P-gp (2–4). Research has shown that chelation lanes (Fig. 1 and ). of Ca2+ by 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic Time course study showed that up-regulation of P-gp and acid abrogates the P-gp induction exerted by many drugs, whereas TRPC5 could be found as early as 6 wk after adriamycin treatment 2+ thapsigargin, which increases [Ca ]i by inhibiting the sarcoplas- mic/endoplasmic reticulum Ca2+-ATPase, enhances P-gp production (2, 3). Furthermore, a panel of different Ca2+ channel Author contributions: X.M., F.L.C., J.J., and X.Y. designed research; X.M., Y. Cai, D.H., C.Z., P.Z., C.Y.L., and Z.X. performed research; S.Y., Y. Chen, R.Z., and J.L. contributed new antagonists was found to reduce the P-gp expression (4). reagents/analytic tools; X.M., Y. Cai, D.H., and J.L. analyzed data; and X.M., J.J., and X.Y. TRP channels are a group of cation channels that play key wrote the paper. functional roles in diverse physiological processes, including ther- The authors declare no conﬂict of interest. mosensation, vascular tone regulation, and bone formation (5). This article is a PNAS Direct Submission. Several TRP isoforms, including TRPM8, TRPV6, TRPV1, and 1X.M., Y. Cai, and D.H. contributed equally to this work. TRPM1, have been shown to be involved in cancer pathogenesis (6). 2To whom correspondence may be addressed. E-mail: [email protected] or yao2068@ It has been suggested that TRPM8, TRPV6, and TRPV1 are on- cuhk.edu.hk. cogenic, whereas TRPM1 performs a tumor-suppressing function This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 2+ (6). These TRP channels are all Ca -permeable (5). The role of 1073/pnas.1202989109/-/DCSupplemental.

16282–16287 | PNAS | October 2, 2012 | vol. 109 | no. 40 www.pnas.org/cgi/doi/10.1073/pnas.1202989109 Downloaded by guest on September 25, 2021 used TRPC5 pharmacological antagonist, also reduced the whole- AB E p-gp MCF-7/WT TRPC5 MCF-7/WT cell current (Fig. 1 ). A dominant-negative TRPC5 construct (TRPC5-DN) was used to disrupt the function of endogenous MCF-7/ADM MCF-7: WT ADM MCF-7/ADM MCF-7: WT ADM TRPC5 (12), whereas a TRPC5-specificsiRNAwasusedtosup- 256 6 # 4 # 5 148 press the expression of TRPC5. Transfection of MCF-7/ADM cells 4 3 P-gp 3 TRPC5 2 with TRPC5-DN or TRPC5-siRNA diminished the whole-cell 2 98 1 current (Fig. 1E). As controls, pretreatment of the cells with pre- 148 1 P-gp value 64 tub 0 tub TRPC5 value 0 immune IgG or transfection with an empty plasmid or a scrambled siRNA had little effect on the whole-cell current (Fig. 1E). Immu- C D noblots confirmed that the TRPC5-siRNA could effectively knock MCF-7/ADM MCF-7/WT MCF-7/ADM ± 40 down the expression of TRPC5 proteins by 73% 4% (Fig. S5). MCF-7/WT MCF-7/ADM 2+ Pre IgG Because P-gp could pump Ca -sensitive fluorescence dyes out 40 60 T5E3 of the cells (13), we could not use conventional dye Fura-2 or Fluo-4 # 2+ 2+ 20 to study TRPC5-mediated Ca entry. Instead, a Ca -sensitive 20 40 molecular construct was used (14). It is known that La3+ potentiates Current density (pA/pF) (pA/pF) -100 -50 Current density 2+ -100 -50 TRPC5 activity but inhibits many other Ca -permeable channels 3+ 2+ 500 01 20 μ 500 01 (15). Application of La (100 M) elicited a rise in [Ca ]i in MCF- Voltage (mV) 7/ADM but not in MCF-7/WT cells (Fig. 1F). The La3+-elicited Voltage (mV) -20 0 [Ca2+] rise in MCF-7/ADM cells was inhibited by T5E3 and Current density (pA/pF) density Current i -20 F E F TRPC5-DN, indicating the involvement of TRPC5 (Fig. 1 ). MCF-7/ADM MCF-7/ADM+Pre IgG Effect of TRPC5 Suppression on P-gp Expression. Apossibleroleof MCF-7/ADM+T5E3 TRPC5 in controlling P-gp expression was explored. Inhibition of 60 * * * * MCF-7/WT Vector 3 T5-DN TRPC5 by T5E3 (11) or 2-APB (16) suppressed the P-gp expression 40 2 in a concentration-dependent manner in MCF-7/ADM cells (Fig. 20 2). TRPC5-DN and TRPC5-siRNA also caused a similar reduction 1 in P-gp expression in MCF-7/ADM cells (Fig. 2). In contrast, 0 3+ 3+

(pA/pF,80 mV) La Current density Current La knocking down P-gp had no effect on TRPC5 expression (Fig. S6). l 3 N l l B F1/F0 (GECO1.2) 0 tro 5E tro tro P 0 200 400 600 0 200 400 600 As a control for T5E3, preimmune IgG treatment had no effect on n T 5-D ontrolC on on 2-A Co C C C Time (s) siTRPC5 Time (s) P-gp expression (Fig. 2). Mock transfection with an empty plasmid TRP or a scrambled siRNA also had no effect on P-gp expression (Fig. Fig. 1. Up-regulation of TRPC5 and P-gp in MCF-7/ADM cells. (A and B) 2). siRNAs targeted against TRPC1, TRPC3, TRPC4, TRPC6, Representative immunoblots (Left) and data summary (Right) of P-gp (A) stromal interaction molecule 1 (STIM1), and Orail were also de- and TRPC5 expression (B). Left lanes in immunoblots are samples from MCF- veloped (Fig. S7). Most of these siRNAs had no effect on P-gp 7/WT cells (WT), whereas right lanes are samples from MCF-7/ADM cells expression (Fig. S7). The only exception was Orail-siRNA, which (ADM). Lower in immunoblots is β-tubulin control. In data summary, relative was able to cause a moderate reduction in P-gp expression (∼25%) MEDICAL SCIENCES expression is shown, in which intensity of bands for TRPC5 and P-gp was (Fig. S7). In comparison, suppressing/inhibiting TRPC5 with T5E3, β – divided by corresponding internal control ( -tubulin). (C E) Whole-cell patch TRPC5-DN, or TRPC5-siRNA had a much larger effect on P-gp clamp studies comparing TRPC5 activity under hypotonic conditions be- ∼ – tween MCF-7/WT and MCF-7/ADM cells (C), and effect of T5E3, TRPC5-DN, expression ( 75 80% reduction) (Fig. 2) than that of Orai1-siRNA. TRPC5-siRNA, and 2-APB in MCF-7/ADM cells under hypotonicity (D and E). Presumably, the role of TRPC5 in stimulating/maintaining P-gp 2+ Shown are representative whole-cell current–voltage relationships (Left in C overproduction may be related to Ca entry through TRPC5. and D) and summary of data at +80 mV under different conditions (Right in Normal culture medium contains growth factors, which are known C and E). 2-APB, 60 μM; T5E3 or preimmune IgG, 20 μg/mL; TRPC5-DN or to stimulate TRP channel activity (17). Application of human ﬁ- empty plasmid, 8 μg per well in six-well plates; TRPC5-siRNA or scrambled 2+ broblast growth factor (hFGF) was found to induce a [Ca ]i rise siRNA, 200 nM. Values are means ± SE of three to ﬁve experiments. #P < 0.05 2+ in MCF-7/ADM cells (Fig. S8). TRPC5-siRNA greatly suppressed vs. MCF-7/WT; *P < 0.05 vs. control. (F)[Ca ]i measurement using GECO1.2. 2+ 2+ 3+ the hFGF-induced [Ca ]i rise, supporting a key role of TRPC5 Shown are time courses of [Ca ]i changes in response to La (100 μM) in MCF-7/WT and MCF-7/ADM cells treated with or without T5E3 or TRPC5-DN.

T5E3 (Fig. S3). The up-regulation became higher by week 9 and even A B A B Ctl Ctl higher at 8 mo (Fig. 1 and and Fig. S3). We also established Ctl 10 20 6 Ctl C5-DN Ctl a paclitaxel-resistant MCF-7 cell line (MCF-7/PTX) by treating 256 T5E3 MCF-7/WT cells with stepwise increasing concentrations of pacli- 2-APB siC5 taxel over 10 mo. The expression of P-gp and TRPC5 was also P-gp 3 # # found to be markedly higher in MCF-7/PTX than in MCF-7/WT # # cells (Fig. S4). # # 148 P-gp value #

Functional Presence of TRPC5 in MCF-7/ADM Cells. Functional exis- tub 0 2+ 10 20 4 8 60 120 200 nM tence of TRPC5 was determined by patch clamp and [Ca ]i (g/ml) ( g/w ell) (M) measurement. TRPC5 is activated by hypotonicity (9). In patch clamp recording, the whole-cell current under hypotonicity was Fig. 2. TRPC5 is essential for P-gp production in MCF-7/ADM cells. MCF-7/ much larger in MCF-7/ADM than in MCF-7/WT cells (Fig. 1C). ADM cells were treated overnight with T5E3 [or preimmune IgG as control This current displayed a double-rectifying current–voltage re- (Ctl) in A and B], TRPC5-DN (or vector as control in B), TRPC5-siRNA (or scrambled siRNA as control in B), or 2-APB (or 0.1% DMSO as control in B)at lationship, which is typical of TRPC5 (10). The whole-cell current the indicated concentration. A, Upper: P-gp expression; A, Lower: β-tubulin diminished in MCF-7/ADM cells that were pretreated with control. In data summary (B), relative expression is shown, in which the aTRPC5-speciﬁc blocking antibody, T5E3 (11) (Fig. 1 D and E). intensity of bands for P-gp was divided by corresponding β-tubulin band. Application of 2-aminoethoxydiphenylborate (2-APB), a commonly Values are means ± SE of four experiments. #P < 0.05 vs. control.

Ma et al. PNAS | October 2, 2012 | vol. 109 | no. 40 | 16283 Downloaded by guest on September 25, 2021 (Fig. S8). Orai1-siRNA and TRPC1-siRNA also slightly reduced adriamycin located near cell peripheral regions but not in nu- 2+ A B the hFGF-induced [Ca ]i rise, whereas TRPC3-siRNA, TRPC4- cleus (Fig. 3 and ). This was expected, because P-gp would siRNA, TRPC6-siRNA, and STIM1-siRNA had no effect (Fig. pump adriamycin out of cells in adriamycin-resistant cells (7). 2+ S8). Involvement of Orail in hFGF-induced [Ca ]i rise agrees Interestingly, inhibiting TRPC5 activity or suppressing TRPC5 with its role in P-gp production (Fig. S7), but TRPC1-mediated expression using T5E3, TRPC5-DN, TRPC5-siRNA, or 2-APB Ca2+ inﬂux might not be important for P-gp production (Fig. S7). resulted in adriamycin reaccumulation in the nucleus of MCF-7/ Because TRPC5 played a much more important role than Orail in ADM cells (Fig. 3 B and C). These data agree with the notion determining P-gp expression, we only focused on TRPC5 in that TRPC5 is important for P-gp function. latter study. We next explored the possibility of whether inhibiting TRPC5 could reverse the drug resistance to adriamycin in MCF-7/ADM Effect of TRPC5 Suppression on Subcellular Distribution of Adriamycin cells. Results from MTT assays showed that MCF-7/ADM cells and Adriamycin Resistance. Adriamycin has natural red ﬂuores- were much more resistant to adriamycin-induced cell death (Fig. cence (7). Subcellular distribution of adriamycin was monitored 4) than MCF-7/WT. MCF-7/ADM cells displayed a 630-fold by a confocal laser-scanning microscope as previously reported higher resistance to adriamycin than MCF-7/WT, as determined (7). As expected, adriamycin was found to be mostly accumu- by MTT cytotoxicity assay. MCF-7/ADM cells also displayed lated inside the nucleus in MCF-7/WT cells (Fig. 3A), similar to resistance to many other chemotherapeutic drugs, including the results reported by others (7). In MCF-7/ADM cells, adria- paclitaxel, epirubicin, vincristine, and mitoxantrone (Table S1). mycin accumulation was much lower, with most of residual Administration of T5E3, TRPC5-DN, TRPC5-siRNA, or 2-APB caused a remarkable reversal of adriamycin resistance in MCF-7/ ADM cells, with respective adriamycin IC50 values decreased by 172-fold, 106-fold, 92-fold, and 207-fold, respectively (Fig. 4 A A and B and Fig. S9 A and B). T5E3 or TRPC5-DN also reversed DAPI ADM fluo Merged the resistance to paclitaxel in MTT assay (Fig. S10).

Effect of TRPC5 Antagonists on the Growth of MCF-7/ADM Tumor Xenografts. An animal model of human breast tumor was established by inoculating MCF-7/ADM cells in athymic nude + ADM

MCF-7/WT mice (BALB/cAnNCr-nu/nu) using the method developed by others (18). The tumor continued to grow in size under adriamycin treatment, indicating adriamycin resistance (Fig. 4 C and D and Fig. S9 C and D). Immunostaining showed that TRPC5 and P-gp were abundantly coexpressed in the tumor, as de- E

+ ADM termined by serial sections of the same tumor (Fig. S11 ). In-

MCF-7/ADM jection of T5E3, a lentivirus carrying TRPC5-DN (lenti-TRPC5- B DN), TRPC5-siRNA, or 2-APB at tumor sites substantially reduced the tumor growth (Fig. 4 C and D and Fig. S9 C and D). Immunohistochemistry also conﬁrmed that the P-gp expression

Ctl was substantially reduced in T5E3-, lenti-TRPC5-DN-, TRPC5- siRNA-, or 2-APB-treated tumor xenografts (Fig. S11 A–D). The growth of tumor xenografts was also resistant to paclitaxel, another common cancer chemotherapeutic drug structurally un- related to adriamycin (Fig. S10). Similarly, paclitaxel-resistant

MCF-7/ADM + ADM + MCF-7/ADM tumor growth was markedly reduced by tumor site injection of +T5E3 T5E3, lenti-TRPC5-DN, TRPC5-siRNA, or 2-APB (Fig. S10).

Involvement of NFATc3 in P-gp Induction. Real-time PCR was used C MCF-7/ADM to compare P-gp expression at the mRNA level between MCF-7/ # 14 * * * * ADM and MCF-7/WT cells. The results demonstrated a much 12 10 higher expression level of P-gp mRNA in MCF-7/ADM than in 8 6 MCF-7/WT cells (Fig. 5A). Search in the primary nucleotide 4 fi 2 sequence using the TFSEARCH database identi ed a potential 0 binding site for a Ca2+-dependent transcription factor NFAT l N (Nuclear/Cytosolic) o 3 D (nuclear factor of activated T cells) (Fig. 5B). The binding site is /WT ADM fluorescence ratio fluorescence ADM 7 ntr PC5 ′ mdr1 o T5E R located in the 5 upstream region of transcription initiation C 2-APBControl ControlPC5- iT Controls − − B MCF- TR site at the position 537 to 542 nt (Fig. 5 ). VIVIT, FK506, MCF-7/ADM and cyclosporin A (19, 20) were used to inhibit NFAT activity. Fig. 3. TRPC5 affects subcellular distribution of adriamycin (ADM) in MCF-7/ The results show that all three inhibitors dose-dependently re- ADM cells. Representative fluorescent images of nuclear DAPI staining (Left, duced P-gp protein expression (Fig. 5C). In the absence of blue), adriamycin autofluorescence (Center, red), and merged images (Right) NFAT inhibitors, cellular adriamycin accumulation was lower in MCF-7/WT (A) and MCF-7/ADM (A and B) cells. (B) Representative images and located near cell peripheral regions but not in nucleus in showing the effect of TRPC5 suppression by T5E3 (20 μg/mL, overnight) on MCF-7/ADM cells (Fig. S12). NFAT inhibitors induced a reac- adriamycin redistribution at the subcellular level. Dashed lines in merged cumulation of adriamycin in the nucleus (Fig. S12). In MTT images outline the cell boundary, which could be visualized at higher fi assays, these inhibitors caused a remarkable reversal of adria- magni cation in differential interference contrast mode. (C) Summary data mycin resistance, making MCF-7/ADM cells sensitive to adria- of adriamycin subcellular distribution expressed as nuclear/cytosolic ratio. μ mycin-induced cell death again (Fig. 5D). TRPC5-DN (8 g per well in six-well plates, 48 h); TRPC5-siRNA (200 nM, 48 h); – 2-APB (100 μM, overnight). Adriamycin (10 μg/mL, overnight) was added in There are four isoforms of NFAT, NFATc1 4. Luciferase re- the culture medium 1 d before fluorescence measurement. n =4–6 experi- porter assay was used to determine the specific NFAT isoform(s) ments. #P < 0.05 vs. MCF-7/WT; *P < 0.05 vs. respective control. involved. The 5′ flanking 800-bp fragment, which is located

16284 | www.pnas.org/cgi/doi/10.1073/pnas.1202989109 Ma et al. Downloaded by guest on September 25, 2021 A B Discussion MCF-7/WT MCF-7/WT The major ﬁndings of this study are as follows: (i)TRPC5andP-gp MCF-7/ADM+pre IgG MCF-7/ADM+scr expression were substantially up-regulated in the adriamycin-re- MCF-7/ADM+T5E3 MCF-7/ADM+siTRPC5 sistant breast cancer cell line MCF-7/ADM. (ii) Suppressing TRPC5 activity/expression with T5E3, TRPC5-DN, TRPC5-siRNA, or 2- 120 120 100 100 APB caused a marked reduction in P-gp protein expression in MCF- 80 80 7/ADM cells. These treatments also altered adriamycin distribution 60 60 40 40 at the subcellular level, resulting in adriamycin reaccumulation in 20 20 nucleus. In MTT assay, these treatments, to a great extent, reversed 0 0 the adriamycin resistance of MCF-7/ADM cells. (iii) Injection of Cell viability Cell(%) viability -2 -1 0 1 2 3 4 Cell (%) viability -2 -1 0 1 2 3 4 T5E3, lenti-TRPC5-DN, TRPC5-siRNA, or 2-APB at the tumor ADM conc (log ADM conc (log M) site reversed the adriamycinresistance oftumorxenografts that were C T5E3 D formed by inoculating MCF-7/ADM cells in athymic nude mice, Ctl Ctl siTRPC5 ﬁ iv Low High causing signi cant reduction in tumor size. ( ) Inhibiting NFAT in MCF-7/ADM cells also reduced P-gp protein expression, induced a reaccumulation of adriamycin in nucleus, and reversed the adriamycin resistance in MTT cell death assay. (v) In luciferase reporter Preimmune serum Scrambled siRNA assay, NFATc3 stimulated the transcriptional activity of mdr1 pro- T5E3 (Low doese) siTRPC5 moter, the effect of which was abrogated by deleting an NFAT T5E3 (High dose) )

3 1200

) ′ ﬂ mdr1 3 1200 binding site at the 5 anking region of gene. Together, these 900 results highlight a key functional role of the TRPC5–NFATc3–P-gp 900 600 signaling pathway in the development of multidrug resistance in 600 cancer cells. A schematic illustration is shown in Fig. S14. 300

300 MCF-7/ADM MCF-7/ADM 0 Tumor size (mm size Tumor Tumor size(mm 0 0 5 10 15 20 25 30 0 5 10 15 20 25 30 Days afte treatment Days afte treatment A B mdr1 promoter 700 # Fig. 4. Inhibition of TRPC5 reverses adriamycin resistance in MCF-7/ADM 600 aaagctaatttatctttatattttccatacttattact--//--tactgggat cells by MTT assay and reduces the growth of human breast tumor xeno- 500 -562 NFAT grafts in athymic nude mice. (A and B) MTT assay. MCF-7/ADM cells were 400 μ 300 AAACACTTGTATTACCATTTT--//--ATGGATC treated with T5E3 (20 g/mL, overnight; A) or TRPC5-siRNA (600 nM, 48 h; B), 200

followed by adriamycin incubation at different concentration for 48 h. 100 MCF-7/WT Transcription start Translation start MCF-7/ADM Preimmune IgG and scrambled siRNA were used as respective controls. In 0 6 MTT assay, cell viability was measured and expressed as percentage of no- (ng) mRNA P-gp Ctl VIVIT lue Ctl Ctl

a VIVIT adriamycin control. (C and D) Representative photographs (Upper) of har- MEDICAL SCIENCES C FK506 CsA vested tumors 30 d after T5E3 (C) or TRPC5-siRNA (D) and corresponding Ctl 12 4 # tumor growth curves (Lower) measured at indicated time points. Female 256 # # nude mice bearing xenograft tumors derived from MCF-7/ADM were injec- P-gp 2 # # # ted at the tumor sites with T5E3 [2 μg (low dose) or 4 μg (high dose) or preimmune IgG as control], or TRPC5-siRNA (40 pmole, scrambled siRNA as tub 0 P-gp relative v 1 2 control). Values are means ± SE of three to eight experiments. 5 10 7.5 15 M M M D MCF-7/WT MCF-7/WT MCF-7/WT upstream of mdr1 transcriptional initiation site and contains mdr1 MCF-7/ADM+DMSO MCF-7/ADM+DMSO MCF-7/ADM+DMSO promoter and enhancers, was cloned into the luciferase reporter MCF-7/ADM+FK506 MCF-7/ADM+CsA MCF-7/ADM+VIVIT “ ” mdr1 120

vector to report the transcriptional activity of promoter. We %)

( 100 used an HEK293 cell line that was stably expressing TRPC5, and the 80 cells were incubated with carbachol to stimulate TRPC5 activity. 60 The cells were further transfected with a specificNFATisoform 40 together with the luciferase reporter vector carrying the 5′ flanking 20 mdr1 viability Cell 0 800-bp sequence of gene. The results demonstrated that -2 -1 0 1 2 3 4 -2 -1 0 1 2 3 4 -2 -1 0 1 2 3 4 NFATc3 stimulated the transcriptional activity of mdr1 promoter in these cells, whereas other NFAT isoforms failed to so do (Fig. 6A and Fig. S13). Importantly, deletion of a putative NFAT binding site Fig. 5. NFAT regulates P-gp expression and alters adriamycin resistance in at −537 to −542 nt (ΔTTTTCC) in the 5′ flanking region of mdr1 MCF-7/ADM cells. (A) Comparison of P-gp mRNA level between MCF-7/WT and MCF-7/ADM cells using real-time PCR. A total of 1 × 105 cells were used. gene abrogated the NFATc3-mediated activation of P-gp tran- ′ fl A 2+ (B) Sketch map of 5 upstream anking sequence of mdr1 gene. A putative scription (Fig. 6 ). NFAT is a Ca -dependent transcription factor. NFAT binding site is shown. Transcription start point is labeled as +1. (C and D) 2+ Arisein[Ca ]i is expected to induce NFAT translocation from Effect of FK506, cyclosporin A, or VIVIT treatment on P-gp expression (C)and cytosol to nucleus, where it stimulates gene transcription (21). In- on cell viability by MTT assay (D). In C, MCF-7/ADM cells were treated over- deed, in HEK293 cells that were cotransfected with TRPC5 and night with these agents at the indicated concentration. Shown are repre- GFP-tagged NFATc3, stimulation of TRPC5 with carbachol in- sentative immunoblots (Left) and data summary (Right). In data summary, duced a time-dependent NFATc3 translocation from cytoplasm to shown are relative expression, in which the intensity of bands for P-gp was β nucleus (Fig. 6 B and C). In control cells that were transfected with divided by corresponding -tubulin band. Vehicle (labeled as Ctl) contained 0.1% DMSO. Values are means ± SE of four experiments. #P < 0.05 vs. Ctl. In D, empty plasmid without TRPC5, carbachol application failed to in- μ C MCF-7/ADM cells were treated overnight with FK506 (10 M, Left), cyclosporin duce NFAT nuclear translocation (Fig. 6 ). Together these results A(15μM, Middle), or VIVIT (2 μM, Right), followed by adriamycin incubation established a signaling cascade involving TRPC5, NFATc3 nuclear at different concentrations for 48 h. Data are expressed as percentage of translocation, and P-gp production. no-adriamycin control. Values are means ± SE of four experiments.

Ma et al. PNAS | October 2, 2012 | vol. 109 | no. 40 | 16285 Downloaded by guest on September 25, 2021 A HEK293-TRPC5 of P-gp, although the effect of Orai1-siRNA on P-gp expression Transcription activity (fold) was much smaller than that of TRPC5-siRNA. Vector+ 1 2 3 4 We further examined the effect of TRPC5 inhibition on growth characteristics of human breast tumor xenografts in an athymic

mdr1 luc Vehicle nude mouse model. Treatment of mice bearing human breast tumor xenografts with TRPC5 inhibitor T5E3, lenti-TRPC5-DN, NFATC3+ TRPC5-siRNA, or 2-APB substantially inhibited adriamycin- and mdr1 luc paclitaxel-resistant tumor growth. These results suggest that NFATC3+ inhibiting/suppressing TRPC5 could be a means to reverse multi- Carbachol drug resistance in the animal model of human breast tumor. mdr1 luc # Among the four methods that were used to inhibit/suppress NFATC3+ TRPC5, T5E3, lenti-TRPC5-DN, and TRPC5-siRNA are highly fi fi 7777&& speci c to TRPC5, whereas 2-APB is less speci c. 2-APB has been mdr1 luc * commonly used to inhibit TRPC5 (16), but it can also inhibit other TRPCs, such as TRPC1. In addition, it activates TRPV1-3 (25). B HEK293-TRPC5 C T5E3 is a blocking antibody that was designed to target the per- +NFATC3-GFP meation pore of TRPC5 (11) and has been shown to selectively and TRPC5+NFATC3 2+ Vector+NFATC3 effectively block TRPC5-mediated Ca entry and cation currents 2.5 (11). TRPC5-DN carries mutations within the TRPC5 pore region, 0 min selectively disrupting the function of endogenous TRPC5 (12). 2.0 TRPC5-siRNA could effectively suppress the expression of 10 min 1.5 Carb TRPC5 (Fig. S5). Therefore, the results from T5E3, lenti-TRPC5- 1.0 ## DN, and TRPC5-siRNA were more reliable than those from 2-APB. Whether T5E3 and/or lenti-TRPC5-DN and/or TRPC5- 30 min 0.5 siRNA can be developed into potential therapeutic tools to over- Carbachol NFATC3-GFP ratio (Nuclear/Cytosolic) 0.0 come the multidrug resistance in cancer chemotherapy still needs 0 10 20 30 40 50 60 50 min further investigation. Time (min) NFAT are Ca2+-dependent transcription factors. Their activation by Ca2+ is mediated through the Ca2+-binding proteins Fig. 6. NFATc3 links TRPC5 activity to P-gp production. (A) NFATc3 stimulates calmodulin and calcineurin (21). Previously, Ca2+ entry through the transcriptional activity of mdr1 gene. HEK293 cells stably expressing several other TRPC channels, including TRPC1, TRPC3, and TRPC5 were further cotransfected with NFATc3 (or empty vector) and the TRPC6, has been shown to stimulate NFAT-dependent gene luciferase reporter plasmid carrying 5′ flanking 800-bp sequence of mdr1 Δ transcription (26, 27). A putative NFAT binding site can be found gene (or its TTTTCC version). The cells were incubated with or without 100 ′ fl mdr1 μM carbachol, as indicated. Horizontal axes indicate relative transcriptional in the 5 anking sequence of gene. Thus, we explored the activity in luciferase assay in HEK293 cells overexpressing indicated constructs. possible involvement of NFAT in TRPC5-mediated P-gp pro- Activity in the cells carrying vector + mdr1-luc + TRPC5 was normalized to 1. duction. We found that treatment of MCF-7/ADM cells with Values are means ± SE of six experiments. #P < 0.05 vs. vehicle; *P < 0.05 vs. NFAT inhibitor—VIVIT, FK506, or cyclosporin A—substantially mdr1.(B and C) Representative images (B) and summary data (C) of time- reduced P-gp expression and caused adriamycin reaccumulation dependent migration of GFP-tagged NFATc3 from cytosol to nucleus after in nucleus. These NFAT inhibitors also reversed adriamycin re- 100 μM carbachol challenge. Transfected constructs are indicated. Values are sistance in MTT cell death assay. Several lines of evidence in- ± # < means SE of four experiments. P 0.05 vs. vector + NFATc3. dicate that NFATc3 is particularly important. (i) Overexpression of NFATc3 stimulated the transcriptional activity of mdr1 gene in ii It has been well documented that a rise in [Ca2+] stimulates luciferase reporter assay; ( ) carbachol stimulation of TRPC5 i induced a nuclear translocation of NFATc3; and (iii) deletion P-gp production in cancer cells (3, 4). However, the molecular ′ fl identity of Ca2+-permeable channels that were involved in the of a putative NFAT binding site at the 5 anking region (con- taining promoter and enhancer) of mdr1 gene abrogated the process is unknown. TRPC5 is a Ca2+-permeable nonselective NFATc3-mediated activation of mdr1 gene transcription. These cation channel. The channel plays an important role in neuronal data strongly suggest that Ca2+ entry through TRPC5 stimulates growth cone extension (12, 22), vascular smooth muscle migra- NFATc3 to enhance P-gp production. tion (22), and animal fear behavior (23). The expression level of In conclusion, our data demonstrated that TRPC5-mediated TRPC5 was reported to vary in different tissues/cells (22). Ex- Ca2+ entry, via NFATc3, stimulates P-gp overproduction in cessive expression of TRPC5 channels has been shown to be adriamycin-resistant MCF-7/ADM breast cancer cells. Inhibit- associated with cardiovascular disorders (24). In the present ing/suppressing TRPC5 could reverse the adriamycin resistance study we uncovered another function of TRPC5: its control of P- in these cancer cells and reduce the growth of tumor xenografts gp expression and multidrug resistance in drug-resistant cancer formed by these cancer cells. cells. Three different indexes for multidrug resistance were monitored, including P-gp expression level, adriamycin expulsion Materials and Methods from cells, and adriamycin-induced cancer cell death. In adria- Additional methods are described in SI Materials and Methods. mycin-resistant MCF-7/ADM cells, P-gp expression level was high, adriamycin accumulation was low and outside of the nu- Luciferase Activity Assays. HEK293 cells stably expressing TRPC5 were tran- cleus, and the cells were resistant to adriamycin-induced cell siently cotransfected with 0.2 μg reporter plasmid (mdr1-luc or ΔTTTTCC- death. Inhibition of TRPC5 by T5E3, TRPC5-DN, TRPC5- mdr1-luc), 0.2 μg expression plasmid (NFATc3), and 0.04 μg pRL-CMV, using siRNA, or 2-APB reversed all three of these multidrug resistance Lipofectamine 2000 reagent. After 48 h, cells were harvested for luciferase reporter assay. indexes. These data demonstrated that TRPC5 is a key protein that can control/regulate P-gp expression and multidrug re- Whole-Cell Patch Clamp. Whole-cell current was measured with an EPC-9 patch sistance in the breast cancer cell MCF-7/ADM. TRPC5 may not clamp amplifier as described elsewhere (28). The holding potential was 0 mV, 2+ be the only Ca -permeable channel that could affect P-gp ex- and current–voltage relationship was obtained using a ramp protocol from pression. Orail-siRNA was also found to reduce the expression +100 mV to −100 mV, with 500-ms duration.

16286 | www.pnas.org/cgi/doi/10.1073/pnas.1202989109 Ma et al. Downloaded by guest on September 25, 2021 2+ 2+ [Ca ]i Measurement. ACa -sensitive molecular construct, GECO1.2, was adriamycin (3 mg/kg, i.p., once every 3 d). Animals were also injected once transfected into MCF-7/WT and MCF-7/ADM cells (14). GECO1.2 fluorescence every 3 d at the tumor sites with TRPC5 antagonists; the injection was signals were measured at room temperature (∼23 °C) using an Olympus scheduled 1 d before adriamycin injection. fluorescence imaging system. ACKNOWLEDGMENTS. This work was supported by Hong Kong Research Adriamycin Accumulation and NFATc3-GFP Translocation. Subcellular distribu- Grants Council Grants CUHK479109, CUHK478710, and CUHK478011 (to tion of adriamycin and NFATc3-GFP was determined using a laser scanning X.Y.); Strategic Investment Scheme C (to X.Y.); and a Group Research Grant “ confocal microscope. from the Chinese University of Hong Kong (to X.Y.); Strategic Priority Re- search Program” of the Chinese Academy of Sciences Grant XDA01040000 and Ministry of Science and Technology of China 2011CB966200 (to J.J.); China In Vivo Tumorigenicity Assay. MCF-7/ADM cells were s.c. injected into the National Natural Science Foundation Grant 81100185 (to X.M.) and 81130057 flanks of female nude mice (5 × 106 cells per mouse). When the tumors (to J.J.); and China National Natural Science Foundation Grant 81101667 and reached ∼200 mm3, nude mice bearing xenograft tumors were injected with Jiangsu Province Natural Science Foundation Grant BK2009071 (to R.Z.).

1. Goda K, Bacsó Z, Szabó G (2009) Multidrug resistance through the spectacle of P- 17. Fiorio Pla A, et al. (2005) Canonical transient receptor potential 1 plays a role in basic glycoprotein. Curr Cancer Drug Targets 9:281–297. fibroblast growth factor (bFGF)/FGF receptor-1-induced Ca2+ entry and embryonic rat 2. Shtil AA, Azare J (2005) Redundancy of biological regulation as the basis of emer- neural stem cell proliferation. J Neurosci 25:2687–2701. gence of multidrug resistance. Int Rev Cytol 246:1–29. 18. Yue W, Zhou D, Chen S, Brodie A (1994) A new nude mouse model for post- 3. Riganti C, et al. (2009) Digoxin and ouabain induce P-glycoprotein by activating cal- menopausal breast cancer using MCF-7 cells transfected with the human aromatase modulin kinase II and hypoxia-inducible factor-1alpha in human colon cancer cells. gene. Cancer Res 54:5092–5095. Toxicol Appl Pharmacol 240:385–392. 19. Pu WT, Ma Q, Izumo S (2003) NFAT transcription factors are critical survival factors 4. Komoto C, et al. (2007) Reversal effects of Ca2+ antagonists on multidrug resistance that inhibit cardiomyocyte apoptosis during phenylephrine stimulation in vitro. Circ viadown-regulation of MDR1 mRNA. Kobe J Med Sci 53:355–363. – Res 92:725–731. 5. Venkatachalam K, Montell C (2007) TRP channels. Annu Rev Biochem 76:387 417. 2+ 6. Lehen’kyi V, Prevarskaya N (2011) Oncogenic TRP channels. Adv Exp Med Biol 704: 20. Crabtree GR, Schreiber SL (2009) SnapShot: Ca -calcineurin-NFAT signaling. Cell 138: – 929–945. 210 221, 210, e1. 7. Altan N, Chen Y, Schindler M, Simon SM (1999) Tamoxifen inhibits acidification in cells 21. Hogan PG, Chen L, Nardone J, Rao A (2003) Transcriptional regulation by calcium, independent of the estrogen receptor. Proc Natl Acad Sci USA 96:4432–4437. calcineurin, and NFAT. Genes Dev 17:2205–2232. 8. Zunino F, et al. (1977) The interaction of adriamycin and its beta anomer with DNA. 22. Jiang LH, Gamper N, Beech DJ (2011) Properties and therapeutic potential of transient Biochim Biophys Acta 476:38–46. receptor potential channels with putative roles in adversity: focus on TRPC5, TRPM2 9. Gomis A, Soriano S, Belmonte C, Viana F (2008) Hypoosmotic- and pressure-induced and TRPA1. Curr Drug Targets 12:724–736. membrane stretch activate TRPC5 channels. J Physiol 586:5633–5649. 23. Riccio A, et al. (2009) Essential role for TRPC5 in amygdala function and fear-related 10. Beech DJ (2007) Canonical transient receptor potential 5. Handb Exp Pharmacol 179: behavior. Cell 137:761–772. 109–123. 24. Bush EW, et al. (2006) Canonical transient receptor potential channels promote car- 11. Xu SZ, et al. (2005a) Generation of functional ion-channel tools by E3 targeting. Nat diomyocyte hypertrophy through activation of calcineurin signaling. J Biol Chem 281: – Biotechnol 23:1289 1293. 33487–33496. 12. Greka A, Navarro B, Oancea E, Duggan A, Clapham DE (2003) TRPC5 is a regulator of 25. Hu H, Grandl J, Bandell M, Petrus M, Patapoutian A (2009) Two amino acid residues hippocampal neurite length and growth cone morphology. Nat Neurosci 6:837–845. determine 2-APB sensitivity of the ion channels TRPV3 and TRPV4. Proc Natl Acad Sci 13. Homolyat L, et al. (1993) Fluorescent cellular indicators are extruded by the multidrug USA 106:1626–1631. resistance protein. J Biol Chem 68:21493–21496. 26. Ohba T, et al. (2007) Upregulation of TRPC1 in the development of cardiac hyper- 14. Zhao Y, et al. (2011) An expanded palette of genetically encoded Ca2+ indicators. – Science 333:1888–1891. trophy. J Mol Cell Cardiol 42:498 507. 15. Jung S, Mühle A, Schaefer M, Strotmann R, Schultz G, Plant TD (2003) Lanthanides 27. Poteser M, et al. (2011) PKC-dependent coupling of calcium permeation through potentiate TRPC5 currents by an action at extracellular sites close to the pore mouth. transient receptor potential canonical 3 (TRPC3) to calcineurin signaling in HL-1 my- MEDICAL SCIENCES J Biol Chem 278:3562–3571. ocytes. Proc Natl Acad Sci USA 108:10556–10561. 16. Xu SZ, et al. (2005b) Block of TRPC5 channels by 2-aminoethoxydiphenyl borate: A 28. Shen B, et al. (2008) Epinephrine-induced Ca2+ influx in vascular endothelial cells is differential, extracellular and voltage-dependent effect. Br J Pharmacol 145:405–414. mediated by CNGA2 channels. J Mol Cell Cardiol 45:437–445.

Ma et al. PNAS | October 2, 2012 | vol. 109 | no. 40 | 16287 Downloaded by guest on September 25, 2021