Pharmacological and functional properties of voltagemi independent Ca2+ channels

ION CHANNELS, RECEPTORS AND TRANSPORTERS

Pharmacological and functional properties of TRPM8channels in prostate tumor cells

Maria Valero & Cruz Morenilla-Palao &

Carlos Belmonte & Felix Viana

Received: 5 July 2010 /Revised: 9 October 2010 /Accepted: 11 October 2010 /Published online: 5 November 2010# Springer-Verlag 2010

Abstract Prostate cancer (PC) is a major health problem inadult males. TRPM8, a cationic TRP channel activated bycooling and menthol is upregulated in PC. However, theprecise role of TRPM8 in PC is still unclear. Some studieshypothesized that TRPM8-mediated transmembrane Ca2+

fluxes play a key role in cellular proliferation of PC cells.In contrast, other findings suggest that high TRPM8 levelsmay reduce the metastatic potential of PC cells. A detailedunderstanding of the response of TRPM8 channels topharmacological modulators of their activity is relevant whenconsidering potential therapies, targeting this ion channel totreat PC. We characterized the pharmacological and func-tional properties of native TRPM8 channels in four humanprostate cell lines, PNT1A, LNCaP, DU145, and PC3,commonly used as experimental models of PC. PNT1A is anon-tumoral prostate cell line while the other three corre-spond to different stages of PC. Here, we show that cold- andagonist-evoked [Ca2+]i responses in PC cells are much lesssensitive to well-characterized agonists (menthol and icilin)and antagonists (BCTC, clotrimazole, and DD01050) ofTRPM8 channels, compared to TRPM8 channels in othertissues, suggesting a different molecular composition and/or

spatial organization. In addition, the forced overexpression ofhuman TRPM8 facilitated the trafficking of TRPM8 channelsresiding in the endoplasmic reticulum to the plasmamembrane, leading to a marked potentiation in the efficacyof the different blockers. These results predict that blockersof canonical TRPM8 channels may be less effective inhalting proliferation of PC cells than expected.

Keywords LNCaP. DU145 . PC3 . BCTC . Proliferation .

TRP channel . Transient receptor potential .

Thermosensitivity . Endoplasmic reticulum . Cancer cells .

Calcium influx

Introduction

Prostate cancer (PC) is a major health problem, totallingone fourth of all new cancer cases diagnosed in adult malesin the United States each year, and accounting for about 9%of all cancer-related death rates in the same population [1].Recent studies have focused on the role of TRPM8, a non-selective, calcium permeable, cation channel of the tran-sient receptor potential (TRP) melastatin subfamily of TRPchannels in PC initiation and progression [reviewed by 2],making it a novel molecular target potentially useful in thediagnosis and treatment of the disease. This ion channel isactivated by voltage, cold temperatures, and coolingcompounds, like menthol and icilin [reviewed by 3], andis involved in the detection of moderate cold temperaturesby mammalian somatosensory nerve terminals [4, 5].

Aside from its restricted localization in sensory neurons,the human ortholog of TRPM8, initially termed trp-p8 [6],is also expressed in the testis, bladder urothelium, andprostate tissue [6–8]. Tsavaler and colleagues identifiedTRPM8 transcripts in only one human prostate cell line, the

Electronic supplementary material The online version of this article(doi:10.1007/s00424-010-0895-0) contains supplementary material,which is available to authorized users.

M. Valero :C. Morenilla-Palao : C. Belmonte : F. Viana (*)Instituto de Neurociencias de Alicante,Universidad Miguel Hernández-CSIC,Apartado 18,San Juan de Alicante 03550, Spaine-mail: [email protected]

M. ValeroLaboratorio de Oncología Molecular, CRIB/Facultad deMedicina, Universidad de Castilla la Mancha,02006 Albacete, Spain

Pflugers Arch - Eur J Physiol (2011) 461:99–114DOI 10.1007/s00424-010-0895-0

http://dx.doi.org/10.1007/s00424-010-0895-0

androgen-dependent LNCaP cells, of the four that were tested(PNT1A, LNCaP, DU145, PC3) [6]. In contrast, Zhang andBarritt reported a low level of expression in PC3 cells as well[9]. The expression of TRPM8 in prostate tissue and in LNCaPcells is androgen-dependent, increasing with androgen treat-ment and declining with anti-androgen therapy [9, 10].

The precise physiological function of TRPM8 channelsin normal and cancerous prostate tissue is still unknown.TRPM8 expression upregulates markedly in PC and inother tumors, suggesting an important role in carcinogenesis[6, 11, 12]. This upregulation is lost in the transition toandrogen independence of the tumor. TRPM8 channels arecalcium permeable, and altered calcium transport has beenlinked to cancer [reviewed by 13], suggesting a possibleconnection between altered Ca2+ levels and prostate cancerprogression. Indeed, [Ca2+]i elevations by TRPM8 agonistscan induce proliferation of prostate cancer cells but alsoapoptotic cell death [reviewed by 14] but the particular roleof TRPM8 in this process is still unknown. A recent studyalso showed effects of menthol on prostate cell viability thatwere TRPM8-independent, thus indicating a need toreconsider previous interpretations [15]. PC3 cells artificiallyoverexpressing TRPM8 have reduced motility, suggesting apossible connection between TRPM8 activity and reducedmetastatic potential [16].

The subcellular localization of TRPM8 channels in prostatetissue is also controversial (see below) but clearly distinctfrom the membrane localization typical of a transduction ionchannel such as TRPM8, a critical transducer of cold temper-atures in sensory neurons. According to Thebault et al., thelocalization in LNCaP cells is nearly exclusive to theendoplasmic reticulum (ER) [17]. Moreover, activation ofthe ER channel by cold or menthol produces Ca2+ storedepletion and activates store-operated Ca2+ currents inLNCaP cells. In contrast, the results by Zhang and Barrittsuggested a dual location in the plasma membrane and theER in about equal levels [9]. Finally, results by Mahieu et al.suggest a primarily membrane localization, attributingprevious findings to non-specific labeling of commerciallyavailable antibodies [18]. Other factors influencing thesubcellular localization of the channel are differentiationstatus of the epithelial cells, with a gradual loss of the plasmamembrane localization of TRPM8 during their dedifferenti-ation [14, 19]. A previous study reported the presence of twodifferent splice variants of TRPM8 in prostate cancer cells[20]. In the same line, experiments by Bidaux et al. suggestthat a truncated but functional form of TRPM8 is preferen-tially expressed in the endoplasmic reticulum of PC3 cells[19]. However, little is known about the expression of theseisoforms in different PC cells.

Recently, we performed an extensive pharmacological andfunctional characterization of native TRPM8 channels inneuronal tissue and following transient expression in HEK293

cells [21–24]. We identified several drugs that fully blockedmenthol- and cold-evoked activity of TRPM8 channels. Inthis work, we extended this characterization to nativeTRPM8 channels in a non-tumoral human prostate epithelialcell line (PNT1A) and three commonly used human prostatecancer cell lines: LNCaP, DU145, and PC3. We show thatthe pharmacological and functional properties of TRPM8-mediated responses differ substantially in prostate cellscompared with neuronal tissue. In particular, inhibition ofchannel activity by current TRPM8 blockers is markedlyreduced in prostate tissue. These findings may haveimportant implications when considering possible therapeuticinterventions targeting TRPM8 channels in the prostate.

Materials and methods

Culture of cell lines and transfection

All the prostate tumoral cell lines, LNCaP, DU145, PC3, andthe non-tumoral cell line PNT1Awere obtained fromAmericanType Culture Collection (ATCC). HEK293 cells were obtainedfrom the ECACC (Salisbury, UK). HEK293 and DU145 cellswere cultured in DMEM with 10% fetal bovine serum (FBS).PC3, LNCaP, and PNT1A were cultured in RPMI mediumsupplemented with 10% FBS. Cells were grown at 37°C in ahumidified atmosphere containing 5% CO2. All the lines weremaintained without antibiotics in the media.

For transfection, cells were trypsinized and plated onpoly-L-lysine-coated round coverslips (6 mm diameter) at40,000 cells/coverslip. Cells were transiently transfectedwith plasmid DNA using Lipofectamine 2000 (Invitrogen),1 μg DNA, and 5 μl Lipofectamine per milliliter. Cellswere used 24–48 h post-transfection.

The human TRPM8 in pcDNA3 (a kind gift of Dr. V.Flockerzi) was co-transfected with pGreen Lantern-1 GFP(Life Technologies) at a 1:1 ratio. Transfected cells wereidentified by their green fluorescence emission duringexcitation with 470 nm light.

Neuronal culture

Trigeminal ganglion neurons from neonatal mice werecultured as described previously [25]. In brief, trigeminalganglia were isolated from anesthetized newborn SwissOF1 mice (P1–P5), incubated with 1 mg/ml collagenasetype IA (Sigma) for 45 min at 37°C in 5% CO2, andcultured in medium: 45% DMEM, 45% F-12+10% FBS(Invitrogen), supplemented with 4 mM L-glutamine,200 μg/ml streptomycin, 125 μg/ml penicillin, 17 mMglucose, nerve growth factor (mouse 7S, 100 ng/ml,Sigma). Cells were plated on poly-L-lysine-coated glasscoverslips and used after 1–3 days in culture.

100 Pflugers Arch - Eur J Physiol (2011) 461:99–114

RNA extraction and cDNA synthesis

Total RNA extraction was performed with the Micro-to-Midi Total RNA Purification System (Invitrogen). Theconcentration and quality of the RNA was assessed byNanodrop (Labtech International, Ringmer, UK) andagarose electrophoresis. Thereafter, equal amounts (2 μg)of total RNA was reversed transcribed using Superscript™

II (Invitrogen) in the presence of random hexamers (Sigma)for 1 h at 50°C in a total reaction volume of 20 μl. Theobtained cDNA was always used fresh to further analysis.

Quantitative real-time PCR and data analysis

Real-time quantitative PCR (RT-qPCR) was performed usingthe 7500 Real-Time PCR System (Applied Biosystems).Serial dilutions (2 μl) of each cDNAwere amplified in a three-step cycling program with Platinum® SYBR® Green qPCRSupermix UDG (Invitrogen) and gene-specific primers in afinal volume of 20 μl. We designed primer pairs specific fortwo splice variants of TRPM8 (Fig. 1a). Primer sequences forthe canonical TRPM8, hTRPM8, were sense: 5′-CTGTCATGGACATCCCACTG-3′ and anti-sense: 5′-CTTGGGCAAAACACACAATG-3′. Primer sequences forthe splice variant hTRPM8b [26] were sense: 5′-TTCCTTCCTGTCCACACCATCG-3′ and anti-sense 5′-CAGCTCGTAAAGGATTTCCGCG-3′.

Expression levels of hTRPM8 were normalized toexpression of Sn-RNP polypeptide B (GeneBank® Accessionnumber J04564) amplified in parallel from the same cDNAsamples. Primer sequences for Sn-RNP polypeptide B genewere forward: 5′-AAACAAGCAGAAAGGGAAGA-3′ andreverse: 5′-CAAGTGGAACTCGAGCAATA-3′. Finally, adissociation curve was introduced to verify the specificity ofthe final products. All samples were studied in duplicate.

The primer design for hTRPM8 and Sn-RNP wereperformed with Primer Express software (Applied Biosystems).The ratio of the relative expression of hTRPM8 to Sn-RNPpolypeptide B of each sample was calculated by using the2�ΔΔCT formula.

hTRPM8 immunodetection by Western blotand immunocytochemistry

To prepare crude plasma membranes, prostate cancer cellswere rinsed with PBS and lysed in hypotonic buffer A (2 mMMgCl2, 1 mM EDTA, 1% phenylmethylsulfonyl fluoride,20 mM Hepes at pH 7.4). Samples were centrifuged at9,000×g at 4°C for 10 min to obtain low-speed pelletscontaining the plasma membranes. Protein solubilization wasachieved heating the pellets resuspended in Laemmli buffer.Equal amounts of protein were loaded and resolved by 7.5%SDS-PAGE and transferred to nitrocellulose membrane

(Protran; Schleicher & Schuell). Immunodetection ofhTRPM8 was performed by incubation with anti-hTRPM8(1:500; Abcam catalog number ab3243) and anti-rabbit-HRP(1:2,000; Sigma). The ECL plus Western Blot DetectionSystem (Amersham Pharmacia) was used to develop the signal.

For immunocytochemistry, glass coverslips with seededcells were fixed in 4% fresh paraformaldehyde for 10 min andwashed twice with PBS. After washing, cells were incubatedwith the membrane marker, wheat germ agglutinin (WGA)conjugated to Alexa Fluor 488, 1:200 in PBS for 10 min [27].Cells were washed twice with PBS and permeabilized in0.2% Triton X-100 for 10 min. After wash, cells wereblocked with 3% BSA for 2 h. Primary polyclonal anti-hTRPM8, purified by peptide chromatography [18], wasdiluted 1:1,000 in 3% BSA and incubated overnight at 4°C.The secondary antibody, Alexa Fluor 546 goat anti-rabbit,was added for 2 h at a 1:2,000 dilution. As a control, weomitted the primary antibody. In order to label the nucleus,we used DAPI. Coverslips with stained cells were mountedin a fluorescent antifading medium (Dako). Stained cellswere visualized with a confocal microscope (Leica TCSSP2) equipped with 405, 488, and 543 nm laser lines. Imageanalysis was performed with LCS Lite software (Leica).

Temperature stimulation

Coverslips with cultured cells were placed in a microchamberand continuously perfused (∼0.8 ml/min) with solutionswarmed at 34±1°C. The temperature was adjusted with awater-cooled RDTC-1 Peltier device (ReidDan Electronics)placed directly on the cell field and controlled by a feedbackdevice [28]. Cold sensitivity was investigated with a 50 stemperature drop to 18±1°C.

Fluorescence Ca2+ imaging

Cells were incubated with 5 μM fura-2AM dissolved instandard extracellular solution and 0.02% Pluronic (Invitrogen)for 40 min at 37°C in darkness. In preliminary experiments, wenoticed very poor loading of prostate cells with our standardincubation protocol. Thus, we used sulfinpyrazone (250 μM),an anion transporter inhibitor for 20–30min, known to improvefura-2 loading by inhibiting extrusion [29]. The composition ofthe bath solution was (in mM): 140 NaCl, 3 KCl, 2.4 CaCl2,1.3 MgCl2, 10 HEPES, and 10 glucose, and was adjusted to apH of 7.4 with NaOH.

Fluorescence measurements were made with a Leica DMIRE2 inverted microscope fitted with a 12-bit cooled CCDcamera (Imago QE Sensicam, Till Photonics). Fura-2 wasexcited at 340 and 380 nm with a Polychrome IVmonochromator (Till Photonics) and the emitted fluores-cence was filtered with a 510 nm longpass filter. Calibratedfluorescence ratios (0.5 Hz) were displayed online with Till

Pflugers Arch - Eur J Physiol (2011) 461:99–114 101

Vision software v4.01 (Till Photonics). No correction wasperformed for temperature-dependent changes in the Kd offura-2 to Ca2+. Bath temperature was sampled simulta-neously and threshold temperature values for [Ca2+]ielevation were estimated by linearly interpolating thetemperature at the midpoint between the last baseline pointand the first point at which a rise in [Ca2+]i deviated by atleast four times the standard deviation of the baseline.

Data analysis

Data are presented as means ± SEM. Statistical analyseswere performed with Excel 2001 or Graphpad 4.0.Significance was set at p<0.05.

Reagents

4-(3-Chloro-pyridin-2-yl)-piperazine-1-carboxylic acid (4-tert-butyl-phenyl)-amide (BCTC) was a generous gift from Dr. R.

Jostock (Grünenthal AG, Aachen, Germany). [L-arginyl]-[N-[2,4-dichlorophenethyl]glycyl]-N-(2,4-dichlorophenethyl)glycinamide (DD01050), identified in a previous publication asH-Arg-15-15 C [30], was a gift from Dr. A. Ferrer-Montiel(Universidad Miguel Hernandez, Spain). L-menthol waspurchased from Scharlau Chemie (Barcelona, Spain).Clotrimazole ((1-(o-chloro-α,α-diphenylbenzyl)-imidazole)),thapsigargin, 1,4-dihydroxy-2,5-di-tert-butylbenzene (BHQ),and sulfinpyrazone were from Sigma.

Results

Expression of TRPM8 channels in different humanprostatic cell lines

We used conventional and RT-qPCR techniques tocharacterize the expression levels of TRPM8 mRNA infour different human prostatic cell lines. As shown in

A

CB

0

30

60

90

PC3

DU145

LNCaP

No

rmal

ized

TR

PM

8p

rote

in e

xpre

ssio

n (

x fo

ld)

PNT1A

ED

ATG

ATG

TRPM8b TRPM8

0.50 Kb

0.25 Kb

MW

TRPM8

0

5

10

PC3

PNT1A

LNCaP

DU145

No

rmal

ized

TR

PM

8ex

pre

ssio

n (

x fo

ld)

0

2

4

PC3

PNT1A

LNCaP

DU145

No

rmal

ized

TR

PM

8bex

pre

ssio

n (

x fo

ld)

F

N.T

.hT

RPM

8

IB: TRPM8

IB:Flotillin-1

41

kDa

TRPM8

PNT1

A

LNC

aPD

U14

5

PC3

kDa

100

130

HEK 293

100

130

PC3DU145LNCaPPNT1A-RTH2O

Fig. 1 Expression of trpm8 mRNA and TRPM8 protein in the fourhuman prostate cell lines. a Schematic representation of the humantrpm8 gene (GeneID 79054), with red (forward) and black (reverse)arrowheads indicating relative position of the two sets of primers usedto detect and quantify trpm8 transcripts. b Detection of mRNAexpression of trpm8 in the four cell lines. As negative controls, RNA(lane H20) or retrotranscriptase enzyme (lane −RT) were omitted inthe reverse transcription reaction. The PCR products were analyzed byagarose gel electrophoresis after 40 PCR cycles. Note that expressionof trpm8 mRNA is detectable in all prostate cells lines. c TRPM8protein expression in PNT1A, LNCaP, DU145, and PC3 cell lines.Western blots of membrane extracts prepared from hTRPM8 trans-

fected HEK293 cells and from human prostate cancer cell lines.TRPM8 protein was detected by anti-hTRPM8 antibody and relativeprotein loading was determined by probing with anti-Flotillin 1antibody. d TRPM8 mRNA expression was evaluated by quantitativeRT-PCR in the four prostate cancer cell lines. Expression levels werenormalized to Sn-RNP polypeptide B as described in the “Materialsand methods” section. e TRPM8 splice variant (named TRPM8b)mRNA was evaluated by the same method used in d. f Quantificationof TRPM8 protein expression. Band intensities from three indepen-dent experiments were quantified and normalized to Flotillin 1expression as protein loading control. Levels were quantified relativeto PNT1A expression set at 1.0


Fig. 1b, we could demonstrate the presence of hTRPM8mRNA in the prostate cancer cell lines LNCaP, DU145,and PC3 and the non-tumoral prostate cell line PNT1Ausing conventional RT-PCR with specific primers for thehuman TRPM8 gene (Fig. 1a). The relative expressionlevels of hTRPM8 mRNA in the different cell lines wasfurther analyzed by RT-qPCR. As shown in Fig. 1d, allsamples from tumoral cell lines showed an upregulatedexpression of hTRPM8 mRNA compared with the non-tumoral prostate cell line PNT1A. A previous study [20]reported a splice variant of TRPM8 in human prostatecancer cells, differing from the canonical 1104 aa protein(here identified as TRPM8) at its amino terminus. We useda primer pair specific for this splice variant, here calledTRPM8b (Fig. 1a). Compared to PNT1A, the three PC celllines showed an elevated expression of TRPM8b that wasparticularly prominent in the DU145 cell line (Fig. 1e).The relative expression levels of canonical TRPM8compared to TRPM8b were 9±1.6-, 6±0.8-, and 9.5±1.8-fold higher in LNCap, DU145, and PC3 cells,respectively (n=3).

The ability to detect the expression of TRPM8 mRNA inthe four cell lines prompted us to investigate the expressionof TRPM8 protein. To address this issue, we preparedmembrane fractions of the four cell lines. As shown inFig. 1c, the expression of TRPM8 protein was confirmed incrude membrane extracts for the three cancerous cell linesby Western blot, appearing as a 128-kDa band, whileexpression in PNT1A was very faint. In the samepreparations, as control of relative protein loading, weimmunodetected Flotillin-1, a membrane-bound protein thatcolocalizes with TRPM8 within lipid rafts [31]. HEK293cells transfected with human TRPM8 cDNA also gave riseto a strong band of the same size, which was absent in wild-type (i.e., untransfected) HEK293 cells, confirming thespecificity of the antibody (Fig. 1c). A quantification ofrelative TRPM8 protein expression for the four cell lines isshown in Fig. 1f.

Next, we used confocal microscopy to define the cellularlocalization of the TRPM8 protein in the four prostate celllines [18]. The plasma membrane was co-stained with theAlexa Fluor 488 conjugate of the lectin WGA, and thenucleus was stained with DAPI. As illustrated in a typicalexample of a DU145 cell (Supplementary Fig. 1A–D), thebulk of the TRPM8 signal appeared localized in the cytosolin a reticulate pattern. Most of the TRPM8 label spared thenuclear region. Similar results were obtained for LNcaP(Supplementary Fig. 1E) and PC3 cells (SupplementaryFig. 1F). In PNT1A cells, TRPM8 inmunostaining wasweak or absent (not shown).

These experiments established the expression of TRPM8protein in the three human cancer prostate cell lines. Theresults confirm previous observations of other laboratories

for LNCaP [9, 18, 19] and PC3 [9] cells and provide novelinformation about expression of TRPM8 in DU145. Theexpression of TRPM8 protein in the non-tumoral PNT1Acell line was clearly lower than in the cancer lines, anobservation in line with the known upregulation of TRPM8in prostate cancer [6].

Characteristics of cold-activated [Ca2+]i responses in humanprostatic cell lines

TRPM8 channels are calcium permeable [4] and [Ca2+]ielevations can be used to monitor channel activation byagonists [22]. Mild cooling is the only known endogenousand highly specific activator of TRPM8. Thus, we startedour characterization of TRPM8 function in prostate epithelialcells by examining [Ca2+]i responses to brief pulses of coldtemperature. From a basal temperature of 36°C, cooling to18°C produced a [Ca2+]i elevation in the three cancerous celllines tested. However, as exemplified in Fig. 2a for DU145cells, this [Ca2+]i elevation occurred only in a relativelysmall and variable fraction of the cells tested. In ∼80% of theresponding cells, a second cooling stimulus produced asimilar [Ca2+]i response (Fig. 2b). In the rest, the responsewas labile and disappeared during the second stimulus (notshown). Figure 2c and d summarizes the percentage of cold-sensitive cells and the amplitude of cold-evoked responsesfor each cell line tested under identical conditions.Responses were somewhat more frequent in LNCaP cellscompared to PC3 and DU145. In contrast, responses werevery infrequent in the non-cancerous PNT1A line; usually,only 1–2% of cells per field were activated by coldtemperature (not shown). The amplitude of the [Ca2+]iresponse was also larger in LNCaP compared to PC3 andDU145, and this difference was significant (Fig. 2d). The[Ca2+]i responses to the temperature drop developed slowlyduring cooling (Fig. 2e). For comparison, we appliedidentical cooling protocols to HEK293 cells transfectedwith hTRPM8 (Fig. 2f). Cold-evoked responses weremore frequent (Fig. 2c), had a much larger amplitude(Fig. 2d), and had a faster rise time (Fig. 2g) than in theprostate cells.

Characteristics of [Ca2+]i responses during repeated coolingstimuli in human prostatic cell lines

We investigated the degree of desensitization or tachyphylaxisof cold-evoked responses in LNCaP cells with application ofthree consecutive stimuli at 90 s intervals. As shown inSupplementary Fig. 2, [Ca2+]i signals in responsive cellswere quite stable in amplitude and time course. On average,the amplitude of the second and third response were only 9%and 35% smaller than the first one (n=30, p<0.001,ANOVA). This relative stability of the cold-evoked


responses greatly facilitated the analysis of the pharmaco-logical interventions that we performed.

Responses to menthol and icilin

In contrast to thermal responses, the effects of menthol andicilin, two chemical agonists of TRPM8 channels [4, 32]were much less reproducible in prostate cancer cellscompared to results obtained in sensory neurons [21] orTRPM8-transfected cells [22]. Very few cells responded to500 μM menthol applied at 35°C (Supplementary Fig. 3A).Similarly, few cells responded to the ultrapotent agonisticilin (10 μM; not shown). However, cold-evokedresponses were slightly potentiated by the co-applicationof menthol (Supplementary Fig. 3A) or icilin (not shown).A summary of these results for the different PC cells isshown in Supplementary Fig. 3b.

Effect of Ca2+ store depletion on responses to TRPM8agonists in human prostatic cell lines

There is some debate regarding the subcellular localizationof TRPM8 channels in different PC cells lines [9, 10, 17]and on the ability of chemical agonists to activate thesechannels. Some studies suggest the exclusive presence ofTRPM8 channels within the ER while others indicate a dualpresence at the plasma membrane and within the ER [9]. Toinvestigate the source of the elevated [Ca2+]i during coolingin PC cells, we performed two types of experiments. First,we compared responses to cold under control conditionsand in the absence of external Ca2+ plus 1 mM EGTA. Asshown in Fig. 3a, removing external Ca2+ for 120 s prior tocooling produced a drastic reduction in the amplitude ofcold-evoked responses. On average, cold-evoked [Ca2+]ielevation was reduced by ∼80% in the different PC cell

40

100

160

18

36

60 s

T (

oC

)[C

a2+] i (

nM

)

0

50

100

450

500

550

***

HEK/hTRPM8

DU145PC3

Δ[C

a2+] i (

nM

)

cold-evoked response

LNCaP

***

0

25

50

75

100 ***

(%)

resp

on

ses

to c

old

PNT1ADU145

PC3LNCaP

HEK293

***

A

C

E

C1

C2

C3

500nM [Ca2+]i

0

18oC

36oC

LNCaPPC3

DU1450.01

0.1

1

10

[Ca2+

] i ris

e sl

op

e (n

M/s

)

HEK/hTRPM8

D

G

C1

C2C3

B

F

50

150

250

18

36

LNCaP

20 s

0

1000

2000

18

36

[Ca2+

] i (n

M)

[Ca2+

] i (n

M)

HEK/hTRPM8

T (

oC

)

T (

oC

)

20 s

Fig. 2 Variable response of pros-tate cancer cells to cold tempera-ture. a Transmitted andpseudocolor fluorescence imagesof the [Ca2+]i response to a coldstimulus in DU145 prostaticcancer cells. The calibration baris 20 μm. b Time course of the[Ca2+]i response in threeindividual cells (C1–3) from thesame field as in a. Note thevariability in the responsebetween cells and thereproducibility in individual cellsto repeated stimuli. c Summaryof percentage of cells respondingto a cold stimulus in the fourdifferent prostate cell lines. Forcomparison, the percentage ofresponses in GFP-hTRPM8transfected HEK293 cells is alsoincluded. Note the very lowpercentage of responses inPNT1A cells. ANOVA test***p<0.001. d Mean [Ca2+]ielevation in response to a coldstimulus (18°C) in the differentcell lines. For comparison, theendogenous responses werecompared with responses observedin HEK293 cells transfected withhTRPM8. ANOVA test ***p<0.001 (n>100 cells per line). e, fTime course of [Ca2+]i response inindividual cells during a coolingstimulus. The red line represents alinear fit to the rising phase of theresponse. g Mean rate of [Ca2+]ielevation in different cell linesduring a cooling stimulus (n=25)


lines (Fig. 3b–d). However, albeit smaller, 69% of LNCaPcells, 60% of PC3, and 44% of DU145 cells still retained adetectable [Ca2+]i rise during cooling. This result suggeststhat transmembrane Ca2+ flux is the major, but notexclusive, component of the [Ca2+]i elevation observed inthe response to cooling. Second, to assess the contributionof the ER to the cold-evoked Ca2+ response more directly,we performed experiments after ER Ca2+ store depletion.To this end, cells were stimulated with a cooling ramp incontrol solution and after blockade of SERCA pumps withBHQ (10 μM) applied to the bath. As shown in Fig. 3e,shortly after BHQ application, the [Ca2+]i rose steadily inthe LNCaP and DU145, the two PC cell lines tested, asituation that made comparison with control conditionscomplicated. Nevertheless, ∼75% of LNCaP cells and ∼50%of DU145 cells still responded to the second cold stimulus.For comparison, nearly all (98%) of HEK293 transfected withhTRPM8 gave rise to a robust cold-evoked response afterstore depletion with BHQ (Fig. 3e), suggesting a primarilylocation of TRPM8 at the plasma membrane. Altogether,these results suggest a dual component of the [Ca2+]ielevation in response to agonists in PC cells, a majorfraction flowing through the plasma membrane and a smallercomponent due to Ca2+ mobilization from stores.

During the course of the study, we also noticed that theplasma membrane of PC cells was extremely leaky tocalcium ions compared to other cells (e.g., HEK293 cells).Previous studies in LNCaP cells ascribed this elevatedpermeability to the elevated expression of TRPV6, a highlyCa2+ selective TRP channel [33]. The relative leakiness ofthe plasma membrane to Ca2+ was quantified in LNCaP,PNT1A and untransfected HEK293 cells, in experimentswhere the external Ca2+ was stepped from normal (2.4 mM)to low (0 Ca2+), back to normal, and finally to high(10 mM) Ca2+ levels. As shown in Fig. 4a and b, changesin external Ca2+ produced very large and very rapidvariations in the [Ca2+]i levels in LNCaP cells, significantlyhigher than in the non-cancerous PNT1A cell line orHEK293 cells. This observation complicates the interpreta-tion of the subcellular origin of cold-evoked [Ca2+]iresponses because removal of external Ca2+ may lead to agradual, spontaneous, emptying of the intracellular stores,decreasing the amplitude of responses in the absence ofexternal Ca2+.

The emptying of intracellular Ca2+ stores can alsoincrease the permeability of the plasma membrane tocalcium [34]. To further test the influence of emptycalcium-store conditions on the permeability of the plasma

0

125

250

18

36

0 Ca2+

60 s

[Ca2+

] i (n

M)

T (

oC

)

BA

C D

0

50

100

150

[C

a2+] i (

nM

)

PC3

***0

50

100

150DU145

control 0 Ca2+ washcontrol 0 Ca2+ wash

***

0

50

100

150LNCaP

control 0 Ca2+ wash

***

0

250

500

18

3660 s

DU145

0

450

900

18

36

LNCaP

T (

oC

)

[C

a2+] i (

nM

)T

(oC

)

[C

a2+] i (

nM

)T

(oC

)60 s 0

250

500

18

36

HEK293/hTRPM8

60 s

EBHQ (10 μM) BHQ (10 μM) BHQ (10 μM)

Fig. 3 The [Ca2+]i response ofcancerous cell lines to cold has asignificant ER component. aEffect of external Ca2+ removal,plus 1 mM EGTA, on theresponse of three LNCaP cellsto cold. Note that the cold-evoked response of one cell(black trace) is abolished whilethe response in the other(red trace) is only reduced.Also, note the major effect ofzero Ca2+ on basal [Ca2+]ilevels. b–d Mean [Ca2+]i eleva-tion in different cancerous celllines during a cooling stimulusin presence (i.e., 2.4 mM) orabsence of external Ca2+. Thenumber of cells analyzedfor each cell line was 100,***p<0.001, paired Studentt test. e Effect of calcium storedepletion with BHQ oncold-evoked responses


membrane, we performed the following experiment. Theexternal calcium concentration was switched back and forthbetween 0 and 2.4 mM Ca2+ while the temperature wasmaintained at 36±1°C, and the fluctuations in absolute [Ca2+]ilevels were monitored. This experiment was performed afterincubation in either control solution or 2 μM thapsigargin (TG)for 40–60 min, to deplete intracellular stores. In PNT1A cells,switching between the Ca2+-free and the Ca2+-containingsolution produced only modest changes in [Ca2+]i levels(∼20 nM; Fig. 4e), and these changes remained modest(∼50 nM) even after incubation in TG (Fig. 4c and e). Incontrast, in all PC cells, switching between the Ca2+-free andthe Ca2+-containing solution produced much larger variationsin [Ca2+]i that reflect the higher basal permeability describedpreviously. The same experiment performed after incubation inTG produced a dramatic oscillation in the [Ca2+]i. This result isexemplified in Fig. 4d for DU145 cells, but similar resultswere obtained for all the PC cell lines (Fig. 4f–h).

Altogether, these results suggest that PC cells have anelevated transmembrane permeability to calcium, even

under basal conditions. Thus, rapid leakage of storedcalcium may underestimate the fraction of the response tocooling that originates in the stores.

Effects of overexpression of hTRPM8 on cold- and agonist-evoked [Ca2+]i signals in human prostatic cell lines

Transfection of TRPM8 in PC3 cells induced a prominentoutwardly rectifying menthol-activated current, typical ofmembrane-bound TRPM8 channels, that was completelyabsent in wild-type (i.e., untransfected) PC3 cells [19]. Incontrast, TRPM8 channels overexpressed in LNCaP cellsappear to be retained in the endoplasmic reticulum [17].Thus, we decided to investigate, in a systematic fashion, theresponses of the three PC cell lines to cooling, menthol andicilin, after transfection with hTRPM8. Cells transfectedwith hTRPM8 were identified by the co-transfection withGFP (Fig. 5a, see “Materials and methods” section). Thebasic protocol is shown in Fig. 5b, with DU145 cells in thisparticular example, and consisted of applications of

D

BA

0

50

100

TGcontrol

PNT1A

075

150400

800

control TG

LNCaP

050

100400

800 DU145

TGcontrol0

50

400

800

control TG

PC3

E F G H

0 2.4 2.40 01000

500

0

C

PNT1A in TG

1000

500

0

000 2.4 2.4

DU145 in TG

0

250

500

LNCaPPNT1AHEK293

60 s

2.4010 mM Ca2+

[Ca2+

] i (n

M)

[Ca2+

] i (n

M)

[Ca2+

] i (n

M)

[Ca2+

] i (n

M)

-100

0

100

200

300

10 mM

0 mM

2.4 mM

LNCaPPNT1AHEK293

Ca2+Δ[C

a2+] i

(nM

)

Fig. 4 The plasma membrane of cancer cell lines is extremely Ca2+

permeable. a Effect of different levels of external Ca2+ on the [Ca2+]iin the tumoral LNCaP cell line, the non-tumoral line PNT1A, andHEK293 cells. The traces are representative examples of theresponses observed in each cell line. b Average [Ca2+]i changeproduced in the different cell lines by variations in the externalCa2+concentrations (0, 2.4, and 10 mM). The number of cellsanalyzed was 30 for each cell line. c, d Following incubation in

2 μM thapsigargin (TG), non-cancerous PNT1A (c) and cancerousDU145 (d) were alternated between zero and control (2.4 mM)external Ca2+ solution, and the [Ca2+]i variation was monitored withfura-2. e–h Average [Ca2+]i level observed in zero calcium (openbars) and in 2.4 mM external Ca2+ solution (black bars) in the fourdifferent prostatic cells lines, in control conditions and after incubationin 2 μM thapsigargin (TG). More than 30 cells were analyzed for eachcondition


menthol (500 μM) and icilin (10 μM) at 35°C, followed bya pulse of cold temperature to ∼20°C. Typically, GFP(+)cells (red traces) gave rise to large amplitude [Ca2+]iresponses to the three stimuli (Fig. 5c and SupplementaryTable 1). The probability that a GFP(+) cell responded tothe three stimuli was >69%, and the majority of theremaining cells were unresponsive to any of the threestimuli, suggesting that the efficiency of co-transfection ofhTRPM8 and GFP was not 100%. In contrast, the largemajority of GFP(−) cells did not respond to the threeagonists (not shown).

An unexpected result was obtained when the cold stimuluswas applied in the absence of external Ca2+ to measure thefraction of response dependent on release from intracellular

calcium stores. Under these conditions, the [Ca2+]i elevationproduced by cooling was systematically larger for untrans-fected cells (black traces) compared to hTRPM8-transfectedcells (red traces), although the amplitude of the [Ca2+]i signalwas generally less than 10% of that measured in control(2.4 mM) external Ca2+ (Fig. 5d). As a matter of fact, theresponse evoked from stores was virtually absent in all PCcell lines cells overexpressing hTRPM8 (Fig. 5e–g).

These experiments suggest that overexpression ofhTRPM8 produces two important effects. (1) It leads to aset of functional channels located primarily at the plasmamembrane. (2) It facilitates the trafficking or translocationof endogenous TRPM8 channels residing in the ER to theplasma membrane.

A

0

300

600

18

36

[Ca2+

] i (n

M) zero Ca2+

T (

oC

) 60 s

LNCaP LNCaP/hTRPM80

15

30

Δ[C

a2+] i (

nM

)

***

PC3 PC3/hTRPM80

10

20 ***

DU145 DU145/hTRPM80

10

20 ***

D

36oC

18oC

menthol

icilin

C2

C1

C3 C3

C2

C1

C3

C2

C1

C3

C2

C1

800 nM [Ca2+]i0

0

350

700

18

36

menthol icilin

T (

oC

)

60 s

[Ca2+

] i (n

M)

C1C2C3

E

GF

B

C

0

400

800

****** ***

*****

***

***

MDU145PC3

untransfectedhTRPM8-transfected

LNCaPC I MC I

***

MC I

***

Δ[C

a2+] i (

nM

)

Δ[C

a2+] i (

nM

)

Δ[C

a2+] i (

nM

)

Fig. 5 Responses of prostatecancer cells transfected withhTRPM8 to TRPM8 agonists. aTransmission, GFP fluorescence(excitation at 475 nm) andratiometric florescence [Ca2+]iimages of DU145 cells stimu-lated with menthol (500 μM),icilin (10 μM), and coldtemperature (18°C). Two cells(C1 and C2) showed GFPfluorescence and, presumably,were transfected with hTRPM8.One cell (C3) did not show GFPfluorescence. Calibration bar is15 μm. b Time course of theexperiment shown in a. Note thelarge amplitude of the responsesin transfected cells (red traces)and the smaller response in theuntransfected one (black trace).c Summary of the effect of thethree agonists on PC cellstransfected (red) and untrans-fected (black) with hTRPM8. dEffect of extracellular Ca2+

removal on the [Ca2+]i responseto a cold stimulus inuntransfected (black) andhTRPM8-transfected (red)LNCaP cells. e–g Average[Ca2+]i increment to a coldstimulus in the absence ofexternal Ca2+ in the differentprostate cancer cell lines inuntransfected (black) andtransfected (gray) cells.Note the suppression ofresponses in transfected cells(n=100 cells). **p<0.01,***p<0.001 (unpaired Studentt test)


Pharmacological properties of TRPM8-mediated [Ca2+]isignals in human prostatic cell lines

TRPM8-mediated Ca2+ fluxes have been hypothesized toplay a key role in cellular proliferation of PC cells [9, 19]and the pharmacological manipulation of TRPM8 activitycould be a key target in modulating the growth of PC cells.Thus, we decided to explore the effects of a novel and twopreviously characterized TRPM8 blockers on calciumsignals in the three PC cells.

BCTC is a potent blocker of cold- and menthol-evokedresponses on neuronal and recombinant TRPM8 channels[21–23, 32]. As shown in Fig. 6a, 10 μM BCTC producedan incomplete and reversible suppression of [Ca2+]ielevations to cooling in LNCaP cells. In contrast, in thesame LNCaP cells transfected with hTRPM8, the degree ofinhibition of cold-evoked responses by BCTC was muchhigher, despite larger amplitudes of the [Ca2+]i signals(Fig. 6b). The mean results for untransfected and trans-fected LNCaP cells are shown in Fig. 6c and d, respectively.Intriguingly, the amplitude of the residual [Ca2+]i signal inBCTC was nearly identical in both cases. As summarized inFig. 6e and Table 1, blockade of cold-evoked responses byBCTC was clearly less effective in native (i.e., untransfected)PC cells (black bars) compared with TRPM8-transfectedcells (gray bars).

We performed similar studies with two other blockers ofTRPM8 channels, clotrimazole [23, 24] and DD01050, onthe three PC cells lines. Clotrimazole is particularlyinteresting because of its dual effects on TRP channelactivity; it activates TRPA1 and TRPV1 but blocks TRPM8[24]. The results were comparable to those obtained withBCTC. As shown in Fig. 6f, the inhibition of cold-evokedresponses with clotrimazole was always significantly higherin PC cells co-transfected with hTRPM8.

DD01050 is a novel, orally bioactive, TRP channelblocker [35]. In addition to its blocking effect on TRPV1,we also found a potent blocking effect of DD01050 oncold-evoked responses of mouse trigeminal neurons and ofhuman TRPM8 channels expressed in HEK293 cells(Fig. 7a–c). In contrast, as was the case with the two otherTRPM8 blockers tested (i.e., BCTC and clotrimazole), theinhibition was much less effective on cold-evokedresponses in wild-type PC cells (Fig. 7d), and thisinhibition was strongly potentiated on PC cells co-transfected with hTRPM8 (Fig. 7e). A summary of thedifferential inhibition of cold-evoked responses byDD01050 in different native and hTRPM8-transfected PCcells is shown in Fig. 7f.

Altogether, these results indicate that native cold-evokedresponses in PC cells have a lower sensitivity to blockers ofTRPM8 channels compared to responses observed insensory neurons or in the same cells transfected with

TRPM8, strongly suggesting a different molecular compo-sition and/or subcellular localization of these channels innative PC cells.

To explore this question further, we asked whetherendogenous cold-evoked responses in the absence ofexternal Ca2+ (see Fig. 3a) had different pharmacologicalproperties to TRPM8 blockade, compared to responses incontrol (i.e., 2.4 mM Ca2+). As shown in SupplementaryFig. 4, the amplitude of cold-evoked responses in zero Ca2+

was nearly identical in the absence or presence of 10 μMBCTC. This result indicates that TRPM8 channels locatedin the endoplasmic reticulum are almost insensitive toBCTC applied extracellularly.

Discussion

There is a pressing need for novel therapeutic strategies inthe treatment of metastatic prostate cancer. In this context,the ion channel TRPM8 has emerged as a promising newtherapeutic target and biomarker of different stages of thisdisease [2, 36, 37]. The most important finding in our studyis the discovery that TRPM8-mediated Ca2+ responses,including kinetic properties and the sensitivity to drugs, aredistinct in prostate cancer epithelial cells when compared tocanonical responses of TRPM8 in transfected cells or insensory neural tissue. Furthermore, responses of hTRPM8channels overexpressed in prostate cells are also differentfrom those observed for the native channels in the samecells. These clear differences may be explained by adifferent molecular composition and/or localization of thefunctional channels and have important implications for thedesign of successful therapeutic strategies that targetTRPM8.

Expression of TRPM8 in different prostate cells

Using different biochemical and molecular techniques, wecould demonstrate the presence of TRPM8 in four differentprostate epithelial cell lines: PNT1A, LNCaP, DU145, andPC3. The most striking difference between the four lines,especially at the functional level, was the reduced activityof TRPM8 in the non-tumorous cell line PNT1A comparedwith the three PC cell lines. These results generally agreewith the pioneering work of Tsavaler et al., demonstratingoverexpression of TRPM8 in malign prostate cancer cells[6]. Of the PC cell lines probed with TRPM8 agonists,responses in LNCaP were most robust, correlating with thehigher expression of the classical isoform. This resultagrees with previous studies that always observedTRPM8-mediated responses in this cell line. The higherfunctional expression of TRPM8 in LNCaP cells is consistentwith the following facts. First, TRPM8 expression is regulated


by androgens, and this cell line is the only one of the three PCcells lines that expresses the androgen receptor [9, 10].Second, androgens promote the expression of classicalTRPM8 at the plasma membrane [19]. Albeit smaller,functional responses in DU145 and PC3 were also clear. Incontrast, functional responses in the non-tumorous cell linePNT1A were very rare, despite detection of the protein,albeit at lower levels.

Of the different anti-TRPM8 antibodies we tested(information available upon request), only the non-commercial antibody used by Mahieu et al. [18] provideda reliable staining detectable by immunofluorescence.Staining was mostly located within cytoplasmic granules.However, both the density profiles and the functionalresponses indicate the presence of TRPM8 at the plasmamembrane. Detection of other molecular markers indifferent PC cell lines, including the widely used androgen

receptor, has also been subject to controversy [17, 38]. Thevariability in the ability to detect this key protein in prostatecells has been ascribed to methodological questions andvariable expression levels, a factor that could also influenceTRPM8 detection.

Regardless of the sensitivity of the different methodsemployed, we feel that the functional responses to cold andchemical agonists (menthol and icilin) that we observed andthe reversible inhibition of these responses by all theclassical TRPM8 blockers tested provided unambiguousevidence for expression of functional TRPM8 channels inthe three PC cell lines. The limited functional response inPNT1A cells speaks in favor of a much lower expressionlevel, a result also supported by the quantitative RT-PCRand protein expression results. Functional Ca2+ responses inthe three PC cell lines were caused primarily, but notexclusively, by enhanced influx through the plasma

0

100

200

washcold cold +BCTC

LNCaP ***

0

200

400

washcold

***

cold +BCTC

LNCaP/hTRPM8

0

250

500

18

36 60 s

LNCaP/hTRPM8

60

100

140

180

18

36

LNCaP

60 s

A

C

B

D

E

LNCaP

LNCaP/h

TRPM8PC3

PC3/hTRPM8

DU145

DU451/h

TRPM8

LNCaP

LNCaP/h

TRPM8PC3

PC3/hTRPM8

DU145

DU451/h

TRPM80

20

40

60

80

100

(%)

inh

ibit

ion

** ** ***

0

20

40

60

80

100

(%)

inh

ibit

ion

Clotrimazole (10μM)

***

*** ***

F

T (

o C)

T (

o C)

[Ca2+

] i (n

M)

[Ca2+

] i (n

M)

Fig. 6 TRPM8 inhibitors areless effective in prostate cancercells. a Example of the effect of10 μM BCTC on the [Ca2+]iresponse to a cold stimulus in atypical LNCaP cell. b The sameexperiment in an LNCaP celltransfected with hTRPM8. c, dAverage [Ca2+]i elevation inwild-type and hTRPM8-transfected LNCaP cells duringa cooling stimulus in controlsolution in 10 μM BCTC andafter wash of the drug. Note thelarger inhibition in thetransfected cells despite thelarger initial [Ca2+]i response.***p<0.001, Student paired ttest. (n=50 wild-type and 30transfected cells). e, f Averagepercentage inhibition of thecold-evoked [Ca2+]i responsewith BCTC (10 μM) andclotrimazole (10 μM) in wild-type (black columns) andhTRPM8-transfected(gray columns). **p<0.01,***p<0.001, Student t test(n=80 wild-type and 50transfected cells)


membrane, a result that fully agrees with the observationsby Zhang and Barritt in LNCaP cells [9].

Another clear result was the variability in the response tocold and chemical agonists in individual prostate cells.Phenotypic variation in cell lines is a troublesome but well-recognized phenomenon, also in prostate cell lines [39],with several underlying mechanisms driving this variability:genetic instability, epigenetic silencing by DNA hypermethy-lation, and post-transcriptional modifications. It is not clearwhether the changes in TRPM8 responses that we observedreflect stable differences in protein levels or longer time scalefluctuations [40]. It is worth mentioning that dedifferentia-tion of prostate epithelial cells in culture leads to majorchanges in TRPM8 function [19].

Characteristics of TRPM8-evoked responses in prostatecancer cells

As indicated, we found major differences in the prevalence,kinetics, amplitude, and pharmacological profiles of cold-evoked [Ca2+]i responses in PC cells in comparison withcanonical or classical TRPM8-mediated responses observedin recombinant systems (e.g., HEK293 cells) and mammaliancold-sensitive neurons [4, 5, 21, 41, 42]. In LNCaP cells, themembrane current activated by cold and/or menthol have alsovery different biophysical properties compared to the typicalTRPM8-mediated currents [17]. Physical (i.e., temperature) or

chemical (i.e., menthol and icilin) activators of TRPM8 areless effective on prostate cells compared with neural tissue. Inaddition, drugs that inhibit classical TRPM8 function alsoinhibit, albeit less effectively, cold-evoked response in prostatecells. Moreover, the responses in the absence of external Ca2+,presumably from channels located in the ER, were nearlyinsensitive to the TRPM8 blocker BCTC. These resultssuggest that the lack of accessibility of the drugs to a fractionof the channels (i.e., those located intracellularly) may explainthe lower overall sensitivity to blockade observed in prostatecancer cells. However, it should be pointed out that thedifferential pharmacology may reflect a different molecularcomposition, which leads to intracellular retention of thosechannels. These two hypotheses for explaining the differentialpharmacology of prostate TRPM8 channels are not mutuallyexclusive. It is intriguing that the ratios of canonical TRPM8to the TRPM8b isoform in PC cells are roughly similar to theratios of response amplitudes in the presence or absence ofexternal Ca2+. Recently, Bidaux et al. [19] providedconvincing evidence that the functional and molecularproperties or TRPM8 channels located in the ER are distinctfrom those at the plasma membrane.

Role of TRPM8 channels in prostate cells

The role of TRPM8 in the physiology of the prostate andthe pathophysiology of prostate cancer is still unclear. In

Table 1 Effect of TRPM8 blockers on cold-evoked responses of prostate cancer cell lines

Cellline

Blocker(10 μM)

Experimentalcondition

Δ[Ca2+]i(nM) cold

Δ[Ca2+]i (nM)cold + blocker

Δ[Ca2+]i(nM) cold

Inhibition(%)

Statistic untransfectedversus transfected

LNCaP BCTC Wild-type 163±20 91±15 120±11 63±1 (***) ***+ TRPM8 330±30 50±5 280±20 91±3 (***)

Clotrimazole Wild-type 119±16 46±6 71±5 59±3 (***) ***+ TRPM8 384±96 28±4 274±49 92±7 (***)

DD01050 Wild-type 120±7 84±10 91±9 36±4 (**) **+ TRPM8 381±60 101±10 283±70 76±6 (***)

DU145 BCTC Wild-type 112±7 62±4 82±4 54±5 (***) ***+ TRPM8 269±65 12±6 281±69 97±2 (***)

Clotrimazole Wild-type 125±8 60±7 90±5 55±4 (***) **+ TRPM8 274±40 57±8 213±32 80±3 (***)

DD01050 Wild-type 94±10 73±9 88±9 24±3 (*) ***+ TRPM8 208±30 84±14 168±27 56±4 (***)

PC3 BCTC Wild-type 86±7 37±9 60±10 60±8 (***) *+ TRPM8 374±47 29±4 363±58 90±5 (***)

Clotrimazole Wild-type 64±11 44±10 71±17 32±1 (*) ***+ TRPM8 400±60 13±1 380±35 98±1 (***)

DD01050 Wild-type 86±7 37±9 60±10 60±8 (**) ***+ TRPM8 300±25 44±5 280±59 98±1 (***)

Average ± SEM amplitude (in nM) of [Ca2+ ]i response to a cold stimulus in control conditions, in the presence of the blocker, and after washing. Datafor wild-type cells (in bold) and hTRPM8 transfected cells (in italics). The right-hand column compares the percentage of inhibition in wild-type versushTRPM8-transfected cells. The degree of inhibition was significant for all conditions: *p<0.05, **p<0.01, and ***p<0.001. Cells analyzed rangedbetween 15 and 70 for each condition


prostate tissue, the highest expression of TRPM8 is locatedin the apical region of epithelial cells, suggesting a possiblesecretory function [10]. Different experiments suggest thatbasal TRPM8 activity is required for the survival of theandrogen-sensitive LNCaP cells [reviewed by 2]. However,activation of TRPM8 by menthol can also induce death ofthe same cells, suggesting a complex regulation of TRPM8activity in vivo [9]. In this context, it is also important topoint out effects of TRPM8 agonists (e.g., menthol) oncellular viability of PC cells that are TRPM8-independent[15].

TRPM8 is an ion channel protein permeable to Na+, K+, andCa2+ [4]. Its role in PC may be linked to effects on [Ca2+]ihomeostasis. It is well established that the amplitude andtemporal organization of Ca2+ signals play important roles inthe switch between cell proliferation and apoptosis [reviewedby 43]. Furthermore, the complement of ion channels and theactivity of calcium entry pathways are different in prostatecancer cell lines with different metastatic potential [44].Besides TRPM8, several TRP channels have been detectedin PC cancer cells, including TRPC1, TRPC4, TRPC6, andTRPV6 [45–47]

Due to its dual localization in the plasma membraneand the ER, activation of TRPM8 channels will allowcalcium entry and release from stores, modulatingdifferent pools of intracellular Ca2+. This Ca2+ homeo-stasis is important in many cellular processes includinggene transcription, cellular differentiation, proliferation,and apoptosis [43, 48, 49]. Prolonged [Ca2+]i elevationsby activation of receptor- and store-operated Ca2+ entrypathways can trigger apoptotic responses in many cells[reviewed by 50], including LNCaP cells [49, 51, 52].Specifically, in human prostate cancer, a correlationbetween the mRNA expression of the Ca2+-permeablechannel TRPV6 and the staging of the cancer has beendescribed [53]. In contrast, moderate or oscillatory [Ca2+]ielevations can trigger cellular proliferation [reviewed by43, 54], also in PC cells [48]. Thapsigargin treatment, andsubsequent Ca2+-store depletion, triggered the activationof a calcium influx pathway in prostate cells. Thisactivation was prominent in all tumoral cells lines incomparison with the non-tumoral cell line. A similarpathway has been recently described in LNCaP cells[55].

LNCaP

LNCaP/h

TRPM8

DU145

DU145/h

TRPM8PC3

PC3/hTRPM8

0

25

50

75

100***

***

blo

ckad

e (%

)

***

0

250

500

10

25

40

HEK293/hTRPM8

60 s

CBA

0

200

400

10

25

40

Trigeminal neurons

DD01050 (10 µM)

[DD01050] (µM) T (

o C)

T (

o C)

[Ca2+

] i (n

M)

T (

C)

[Ca2+

] i (n

M)

T ( C

)[C

a2+] i

(nM

)[C

a2+] i

(nM

)

60 s

D

50

150

250

10

25

40

LNCaP

60 s

FE

0

300

600

10

25

4060 s

LNCaP/hTRPM8

0.1 1 100

25

50

75

100

blo

ckad

e (%

)

H2N

NN

NH2

O

O

O

Cl

Cl

Cl

ClHN

H2N NH

DD01050 (10 µM)

DD01050 (10 µM)DD01050 (10 µM)

DD01050 (10 µM)

Fig. 7 The novel TRPM8 inhibitor, DD01050, blocks cold-evokedresponses in different TRPM8-expressing cells. Effect of a 10 μMDD01050 on cold-evoked responses in mouse trigeminal sensoryneurons, b HEK293 cells transfected with hTRPM8, d LNCaP cells,and e LNCaP cells transfected with hTRPM8. Notice the weakerinhibition of DD01050 in the wild-type LNCaP cells compared withneuronal TRPM8 channels or cells transfected with hTRPM8. c Dose–

inhibition curve of cold-evoked responses in HEK293 cells transfectedwith hTRPM8. The solid line represents a fit to the Hill equation thatyielded an IC50=0.86±0.19 μM and a Hill slope of N=1.72±0.57 (n=7–20 cells). f Quantification of the differential blockade of DD01050in different PC cells with and without hTRPM8 transfection. Thedifferences were significant at p<0.001. A minimum of 25 cells wereanalyzed for each condition


Therapeutical implications of TRPM8 expression in PCcells

Interest in development of novel therapeutic agents againstprostate cancer is currently very high; over 200 compoundshave entered clinical development for use in the treatmentof advanced prostate cancer [56]. The selective over-expression of TRPM8 in prostate cancer cells [6] madethis channel a promising novel target. However, as alreadymentioned, it is still unclear whether increasing ordecreasing TRPM8 function is the proper strategy forinhibiting the growth of PC cells. Inhibition of TRPM8expression leads to reduced proliferation of PC cells [2]. Onthe other hand, activation of Ca2+ influx pathways has beenshown to induce apoptosis in LNCaP cells [51]. Thus,activators of TRPM8 channels may find a clinical applica-tion. We see three major implications of our findings for thedevelopment of future therapies targeting TRPM8. (1) Non-permeable drugs against TRPM8, such as humanizedfunctional antibodies, may not be as effective in blockingthe activity of the channel as predicted based on studies inrecombinant systems. A fraction of native TRPM8 channelsin prostate tissue are located intracellularly and appear lessaccessible to drugs that are effective agonists or antagonistsof canonical TRPM8 function [22, 32]. In contrast,silencing of endogenous TRPM8 mRNA (e.g., with anti-sense or specific siRNA) would not suffer from theselimitations and has already been shown to be effective inhalting proliferation of prostate cancer cells [9]. (2) Over-expressing TRPM8 facilitates translocation of endogenousTRPM8 from intracellular compartments to the plasmamembrane, making the channels more accessible topharmacological modulation. Thus, an alternative treatmentstrategy could be based on the activation of traffickingtowards the plasma membrane by pharmacological inter-vention. For example, application of epidermal growthfactors can produce the rapid translocation of TRPC5channels to the plasma membrane [57]. Also, activation ofPKC can produce a rapid delivery of functional TRPV1channels to the plasma membrane [58]. (3) The design ofnon-functional channel subunits as an alternative mecha-nism to alter the functional activity of TRPM8 should beconsidered as a therapeutic alternative.

Finally, in order to calibrate the true clinical relevance ofour findings, future studies should try to verify them onacutely isolated epithelial prostate cancer cells. Althoughour results were qualitatively very similar in the threecancerous cell lines tested, this may not be the case in intacttissue.

Acknowledgments The authors thank E. Quintero, A. Miralles, andA. Pérez-Vegara for their excellent technical assistance and G.Exposito for her help with confocal microscopy analysis. The human

TRPM8 cDNAs was a generous gift of Dr. V. Flockerzi (UniversityHomburg-Saar, Germany). We thank A. Ferrer-Montiel for the gift ofDD01050 and B. Nilius (KULeuven) for the gift of the humanTRPM8-specific antibody. We thank M. Pertusa and L. Pardo forcomments on the manuscript. This work was supported by funds fromthe Spanish MICIIN: projects BFU2007-61855 to F.V and BFU2008-04425 and CONSOLIDER-INGENIO 2010 CSD2007-00023 to C.B.M.V. was the recipient of a predoctoral fellowship of the SpanishGovernment and C.M. holds a postdoctoral fellowship from FIS.

References

1. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ (2009) Cancerstatistics, 2009. CA Cancer J Clin 59:225–249

2. Zhang L, Barritt GJ (2006) TRPM8 in prostate cancer cells: apotential diagnostic and prognostic marker with a secretoryfunction? Endocr Relat Cancer 13:27–38

3. Voets T, Owsianik G, Nilius B (2007) TRPM8. Handb ExpPharmacol 179:329–344

4. McKemy DD, Neuhausser WM, Julius D (2002) Identification ofa cold receptor reveals a general role for TRP channels inthermosensation. Nature 416:52–58

5. Peier AM, Moqrich A, Hergarden AC, Reeve AJ, Andersson DA,Story GM, Earley TJ, Dragoni I, McIntyre P, Bevan S,Patapoutian A (2002) A TRP channel that senses cold stimuliand menthol. Cell 108:705–715

6. Tsavaler L, Shapero MH, Morkowski S, Laus R (2001) Trp-p8, anovel prostate-specific gene, is up-regulated in prostate cancer andother malignancies and shares high homology with transient receptorpotential calcium channel proteins. Cancer Res 61:3760–3769

7. Fuessel S, Sickert D, Meye A, Klenk U, Schmidt U, Schmitz M,Rost AK, Weigle B, Kiessling A, Wirth MP (2003) Multipletumor marker analyses (PSA, hK2, PSCA, trp-p8) in primaryprostate cancers using quantitative RT-PCR. Int J Oncol 23:221–228

8. Stein RJ, Santos S, Nagatomi J, Hayashi Y, Minnery BS, XavierM, Patel AS, Nelson JB, Futrell WJ, Yoshimura N, ChancellorMB, De Miguel F (2004) Cool (TRPM8) and hot (TRPV1)receptors in the bladder and male genital tract. J Urol 172:1175–1178

9. Zhang L, Barritt GJ (2004) Evidence that TRPM8 is an androgen-dependent Ca2+ channel required for the survival of prostatecancer cells. Cancer Res 64:8365–8373

10. Bidaux G, Roudbaraki M, Merle C, Crepin A, Delcourt P,Slomianny C, Thebault S, Bonnal JL, Benahmed M, Cabon F,Mauroy B, Prevarskaya N (2005) Evidence for specific TRPM8expression in human prostate secretory epithelial cells: functionalandrogen receptor requirement. Endocr Relat Cancer 12:367–382

11. Schmidt U, Fuessel S, Koch R, Baretton GB, Lohse A, TomasettiS, Unversucht S, Froehner M, Wirth MP, Meye A (2006)Quantitative multi-gene expression profiling of primary prostatecancer. Prostate 66:1521–1534

12. Yamamura H, Ugawa S, Ueda T, Morita A, Shimada S (2008)TRPM8 activation suppresses cellular viability in human melanoma.Am J Physiol Cell Physiol 295:C296–C301

13. Monteith GR, McAndrew D, Faddy HM, Roberts-Thomson SJ(2007) Calcium and cancer: targeting Ca2+ transport. Nat RevCancer 7:519–530

14. Prevarskaya N, Flourakis M, Bidaux G, Thebault S, Skryma R(2007) Differential role of TRP channels in prostate cancer.Biochem Soc Trans 35:133–135

15. Kim SH, Nam JH, Park EJ, Kim BJ, Kim SJ, So I, Jeon JH (2009)Menthol regulates TRPM8-independent processes in PC-3 prostatecancer cells. Biochim Biophys Acta 1792:33–38


16. Gkika D, Flourakis M, Lemonnier L, Prevarskaya N (2010) PSAreduces prostate cancer cell motility by stimulating TRPM8 activityand plasma membrane expression. Oncogene 29:4611–4616

17. Thebault S, Lemonnier L, Bidaux G, Flourakis M, Bavencoffe A,Gordienko D, Roudbaraki M, Delcourt P, Panchin Y, Shuba Y,Skryma R, Prevarskaya N (2005) Novel role of cold/menthol-sensitive transient receptor potential melastatine family member8 (TRPM8) in the activation of store-operated channels in LNCaPhuman prostate cancer epithelial cells. J Biol Chem 280:39423–39435

18. Mahieu F, Owsianik G, Verbert L, Janssens A, De Smedt H,Nilius B, Voets T (2007) TRPM8-independent menthol-inducedCa2+ release from endoplasmic reticulum and Golgi. J Biol Chem282:3325–3336

19. Bidaux G, Flourakis M, Thebault S, Zholos A, Beck B, Gkika D,Roudbaraki M, Bonnal JL, Mauroy B, Shuba Y, Skryma R,Prevarskaya N (2007) Prostate cell differentiation status determinestransient receptor potential melastatin member 8 channel subcellularlocalization and function. J Clin Invest 117:1647–1657

20. Lis A, Wissenbach U, Philipp SE (2005) Transcriptional regula-tion and processing increase the functional variability of TRPMchannels. Naunyn Schmiedebergs Arch Pharmacol 371:315–324

21. Madrid R, Donovan-Rodriguez T, Meseguer V, Acosta MC,Belmonte C, Viana F (2006) Contribution of TRPM8 channelsto cold transduction in primary sensory neurons and peripheralnerve terminals. J Neurosci 26:12512–12525

22. Malkia A, Madrid R, Meseguer V, de la Peña E, Valero M,Belmonte C, Viana F (2007) Bidirectional shifts of TRPM8channel gating by temperature and chemical agents modulate thecold sensitivity of mammalian thermoreceptors. J Physiol581:155–174

23. Malkia A, Pertusa M, Fernandez-Ballester G, Ferrer-Montiel A,Viana F (2009) Differential role of the menthol-binding residueY745 in the antagonism of thermally gated TRPM8 channels. MolPain 5:62

24. Meseguer V, Karashima Y, Talavera K, D'Hoedt D, Donovan-Rodriguez T, Viana F, Nilius B, Voets T (2008) Transient receptorpotential channels in sensory neurons are targets of theantimycotic agent clotrimazole. J Neurosci 28:576–586

25. Viana F, de la Pena E, Pecson B, Schmidt RF, Belmonte C (2001)Swelling-activated calcium signalling in cultured mouse primarysensory neurons. Eur J Neurosci 13:722–734

26. Bodding M, Wissenbach U, Flockerzi V (2007) Characterisationof TRPM8 as a pharmacophore receptor. Cell Calcium 42:618–628

27. Garcia-Sanz N, Valente P, Gomis A, Fernandez-Carvajal A,Fernandez-Ballester G, Viana F, Belmonte C, Ferrer-Montiel A(2007) A role of the transient receptor potential domain ofvanilloid receptor I in channel gating. J Neurosci 27:11641–11650

28. Reid G, Amuzescu B, Zech E, Flonta M (2001) A system forapplying rapid warming or cooling stimuli to cells during patchclamp recording or ion imaging. J Neurosci Methods 111:1–8

29. Di Virgilio F, Steinberg TH, Silverstein SC (1990) Inhibition ofFura-2 sequestration and secretion with organic anion transportblockers. Cell Calcium 11:57–62

30. Garcia-Martinez C, Fernandez-Carvajal A, Valenzuela B, GomisA, Van Den NW, Ferroni S, Carreno C, Belmonte C, Ferrer-Montiel A (2006) Design and characterization of a noncompetitiveantagonist of the transient receptor potential vanilloid subunit 1channel with in vivo analgesic and anti-inflammatory activity. JPain 7:735–746

31. Morenilla-Palao C, Pertusa M, Meseguer V, Cabedo H, Viana F(2009) Lipid raft segregation modulates TRPM8 channel activity.J Biol Chem 284:9215–9224

32. Behrendt HJ, Germann T, Gillen C, Hatt H, Jostock R (2004)Characterization of the mouse cold-menthol receptor TRPM8 and

vanilloid receptor type-1 VR1 using a fluorometric imaging platereader (FLIPR) assay. Br J Pharmacol 141:737–745

33. Lehen'kyi V, Flourakis M, Skryma R, Prevarskaya N (2007)TRPV6 channel controls prostate cancer cell proliferation via Ca(2+)/NFAT-dependent pathways. Oncogene 26:7380–7385

34. Venkatachalam K, van Rossum DB, Patterson RL, Ma HT, GillDL (2002) The cellular and molecular basis of store-operatedcalcium entry. Nat Cell Biol 4:E263–E272

35. Garcia-Martinez C, Humet M, Planells-Cases R, Gomis A,Caprini M, Viana F, de la Peña E, Sanchez-Baeza F, CarbonellT, De Felipe C, Perez-Paya E, Belmonte C, Messeguer A, Ferrer-Montiel A (2002) Attenuation of thermal nociception andhyperalgesia by VR1 blockers. Proc Natl Acad Sci USA99:2374–2379

36. Prevarskaya N, Zhang L, Barritt G (2007) TRP channels in cancer.Biochim Biophys Acta 1772:937–946

37. Wissenbach U, Niemeyer BA, Flockerzi V (2004) TRP channelsas potential drug targets. Biol Cell 96:47–54

38. Alimirah F, Chen J, Basrawala Z, Xin H, Choubey D (2006) DU-145 and PC-3 human prostate cancer cell lines express androgenreceptor: implications for the androgen receptor functions andregulation. FEBS Lett 580:2294–2300

39. Mitchell S, Abel P, Ware M, Stamp G, Lalani E (2000) Phenotypicand genotypic characterization of commonly used human prostaticcell lines. BJU Int 85:932–944

40. Sigal A, Milo R, Cohen A, Geva-Zatorsky N, Klein Y, Liron Y,Rosenfeld N, Danon T, Perzov N, Alon U (2006) Variability andmemory of protein levels in human cells. Nature 444:643–646

41. Reid G, Flonta ML (2001) Physiology. Cold current in thermor-eceptive neurons. Nature 413:480

42. Voets T, Droogmans G, Wissenbach U, Janssens A, Flockerzi V,Nilius B (2004) The principle of temperature-dependent gating incold- and heat-sensitive TRP channels. Nature 430:748–754

43. Lang F, Foller M, Lang KS, Lang PA, Ritter M, Gulbins E,Vereninov A, Huber SM (2005) Ion channels in cell proliferationand apoptotic cell death. J Membr Biol 205:147–157

44. Ding Y, Robbins J, Fraser SP, Grimes JA, Djamgoz MB (2006)Comparative studies of intracellular Ca2+ in strongly and weaklymetastatic rat prostate cancer cell lines. Int J Biochem Cell Biol38:366–375

45. Fixemer T, Wissenbach U, Flockerzi V, Bonkhoff H (2003)Expression of the Ca2+-selective cation channel TRPV6 inhuman prostate cancer: a novel prognostic marker for tumorprogression. Oncogene 22:7858–7861

46. Flourakis M, Prevarskaya N (2009) Insights into Ca2+ homeostasisof advanced prostate cancer cells. Biochim Biophys Acta1793:1105–1109

47. Wissenbach U, Niemeyer B, Himmerkus N, Fixemer T, BonkhoffH, Flockerzi V (2004) TRPV6 and prostate cancer: cancer growthbeyond the prostate correlates with increased TRPV6 Ca2+channel expression. Biochem Biophys Res Commun 322:1359–1363

48. Vanden Abeele F, Shuba Y, Roudbaraki M, Lemonnier L,Vanoverberghe K, Mariot P, Skryma R, Prevarskaya N (2003)Store-operated Ca2+ channels in prostate cancer epithelial cells:function, regulation, and role in carcinogenesis. Cell Calcium33:357–373

49. Wertz IE, Dixit VM (2000) Characterization of calcium release-activated apoptosis of LNCaP prostate cancer cells. J Biol Chem275:11470–11477

50. Parekh AB, Putney JW Jr (2005) Store-operated calciumchannels. Physiol Rev 85:757–810

51. Gutierrez AA, Arias JM, Garcia L, Mas-Oliva J, Guerrero-Hernandez A (1999) Activation of a Ca2+-permeable cationchannel by two different inducers of apoptosis in a humanprostatic cancer cell line. J Physiol 517(Pt 1):95–107


52. Martikainen P, Isaacs J (1990) Role of calcium in the programmeddeath of rat prostatic glandular cells. Prostate 17:175–187

53. Wissenbach U, Niemeyer BA, Fixemer T, Schneidewind A, Trost C,Cavalie A, Reus K, Meese E, Bonkhoff H, Flockerzi V (2001)Expression of CaT-like, a novel calcium-selective channel, correlateswith the malignancy of prostate cancer. J Biol Chem 276:19461–19468

54. Pandiella A, Magni M, Lovisolo D, Meldolesi J (1989) The effectof epidermal growth factor on membrane potential. Rapidhyperpolarization followed by persistent fluctuations. J Biol Chem264:12914–12921

55. Vanden Abeele F, Lemonnier L, Thebault S, Lepage G, Parys JB,Shuba Y, Skryma R, Prevarskaya N (2004) Two types of store-

operated Ca2+ channels with different activation modes andmolecular origin in LNCaP human prostate cancer epithelial cells.J Biol Chem 279:30326–30337

56. Armstrong AJ, Carducci MA (2006) New drugs in prostate cancer.Curr Opin Urol 16:138–145

57. Bezzerides VJ, Ramsey IS, Kotecha S, Greka A, Clapham DE(2004) Rapid vesicular translocation and insertion of TRPchannels. Nat Cell Biol 6:709–720

58. Morenilla-Palao C, Planells-Cases R, Garcia-Sanz N, Ferrer-Montiel A (2004) Regulated exocytosis contributes to proteinkinase C potentiation of vanilloid receptor activity. J Biol Chem279:25665–25672


Pharmacological and functional properties of voltagemi independent Ca2+ channels

Documents

Transcript of Pharmacological and functional properties of voltagemi independent Ca2+ channels