A panel of expression markers was validated and used to document that, when radical prostatectomy specimens are cultured in low (i.e., <260 μmol/L)–calcium (Ca2+)-serum-free, growth factor–defined (SFD) medium, what grows out are not prostatic cancer cells but basally derived normal transit-amplifying prostatic epithelial cells. The selective outgrowth of the normal transit-amplifying versus prostatic cancer cells is due to the differential effect of low-Ca2+ medium on the structure of Notch-1 and E-cadherin signaling molecules. In low-Ca2+ medium, Notch-1 receptor is conformationally in a constitutively active, cell autonomous form not requiring reciprocal cell-cell (i.e., ligand) interaction for signaling. Such signaling is required for survival of transit-amplifying cells as shown by the death of transit-amplifying cells induced by treatment with a series of chemically distinct γ-secretase inhibitors to prevent Notch-1 signaling. Conversely, in low-Ca2+ medium, E-cadherin is conformationally inactive preventing cell-cell homotypic interaction, but low cell density nonaggregated transit-amplifying cells still survived because Notch-1 is able to signal cell autonomously. In contrast, when medium Ca2+ is raised to >400 μmol/L, Notch-1 conformationally is no longer constitutively active but requires cell-cell contact for reciprocal binding of Jagged-1 ligands and Notch-1 receptors between adjacent transit-amplifying cells to activate their survival signaling. Such cell-cell contact is enhanced by the elevated Ca2+ inducing an E-cadherin conformation allowing homotypic interaction between transit-amplifying cells. Such Ca2+-dependent, E-cadherin-mediated interaction, however, results in cell aggregation, stratification, and inhibition of proliferation of transit-amplifying cells via contact inhibition–induced up-regulation of p27/kip1 protein. In addition, transit-amplifying cells not contacting other cells undergo squamous differentiation into cornified (i.e., 1% SDS insoluble) envelopes and death in the elevated Ca2+ medium. Stratification and contact inhibition induced by elevated Ca2+ are dependent on E-cadherin–mediated homotypic interaction between transit-amplifying cells as shown by their prevention in the presence of a cell-impermanent, E-cadherin neutralizing antibody. In contrast to growth inhibition of normal transit-amplifying cells, supplementation of low-Ca2+-SFD medium with 10% FCS and raising the Ca2+ to >600 μmol/L stimulates the growth of all prostate cancer cell lines tested. Additional results document that, at physiologic levels of Ca2+ (i.e., >600 μmol/L), prostatic cancer cells are not contact inhibited by E-cadherin interactions and Notch-1 signaling is no longer required for survival but instead becomes one of multiple signaling pathways for proliferation of prostatic cancer cells. These characteristic changes are consistent with prostate cancer cells' ability to metastasize to bone, a site of high-Ca2+ levels.

Human prostatic glands are composed of a simple stratified epithelium containing basal and luminal layers separated via basement membrane from a well-developed stromal compartment (1). The homeostatic maintenance of this prostatic epithelium is regulated via a hierarchical stem cell organization (2). In the prostate epithelium, stem cells are rare and are located within the basal layer [i.e., ≤1% of basal cells are stem cells (3)]. Prostate stem cells proliferate rarely to replace (i.e., renew) the fraction of their progeny, which, instead of remaining as uncommitted stem cells, enter a terminal maturation process in which several sequential stages have been identified phenotypically and morphologically (310). The earliest stage is termed transit-amplifying cells, which has a high proliferative potential and is located in the basal layer. These transit-amplifying cells express very low to undetectable levels of androgen receptor (AR) protein and do not express prostatic differentiation marker proteins [e.g., prostate-specific antigen (PSA), human glandular kallikrein-2 (hK2), and prostate-specific membrane antigen (PSMA); refs. 410]. Although this subset of AR-negative transit-amplifying cells does not respond directly to androgen, these cells do require critical levels of androgen-stimulated paracrine growth factors (i.e., andromedins) for their proliferation but not survival (10). These andromedins are produced by the occupancy of the AR by its ligand within prostate stromal cells (1012). These transit-amplifying cells express the dominant-negative, NH2-terminal truncated form of the p53-related, p63 gene (i.e., ΔNp63α isotype) within their nucleus and high levels of “basal-specific” cytokeratins (i.e., keratins 5 and 14), glutathione S-transferase π isoform (GST-Pi), standard form of CD44 (CD44s), transglutaminase type II (TGT-2), and involucrin but only low levels of luminal-specific keratins 8 and 18 (1, 410, 1216). Besides proliferating, these nuclear ΔNp63α-expressing transit-amplifying cells undergo a process of maturation into “basal-intermediate” cells in the basal epithelial compartment (79). This maturation into basal-intermediate cells involves the loss of expression of keratin 14 while maintaining the coexpression of basal-specific keratin 5, luminal-specific keratins 8 and 18, and ΔNp63α coupled with a decrease in their growth fraction (79). The basal-intermediate cells continued to mature with their gain of expression of prostate stem cell antigen protein and AR mRNA but not protein during their migration into the luminal layer to become “luminal-intermediate” cells (7). The luminal-intermediate cells translate AR mRNA and thus express AR protein whose occupancy by androgen induces their maturing into fully differentiated, luminal secretory cells, which are nonproliferative due to their selective expression of the p27/kip1 protein (5, 8). These nonproliferating luminal secretory cells express prostate-specific differentiation marker proteins like PSA, hK2, and PSMA and lack expression of basal markers keratins 5 and 14 and ΔNp63α protein (49). Although the engagement of nuclear AR in these luminal secretory cells does regulate PSA, hK2, and PSMA expression, it does not regulate their survival. Instead, such survival requires adequate levels of the androgen-stimulated stromally derived andromedins (11, 12).

It has been well documented that, when normal human prostate tissue is used for culturing in low (i.e., <260 μmol/L)–calcium (Ca2+)-serum-free, growth factor–defined (SFD) medium, only transit-amplifying cells have a sufficiently high rate of proliferation to allow multiple subpassaging before undergoing proliferative senescence (7, 9, 10, 14). In contrast, the low proliferation rates of prostatic stromal cells, epithelial stem cells, and luminal secretory cells in low-Ca2+-SFD medium result in the elimination of these other cell types during early serial in vitro passages (8, 9, 13, 15). Based on such selection, pure cultures of normal prostatic transit-amplifying cells (i.e., termed PrEC cells) established in and maintained with a low (i.e., 260 ± 14 μmol/L)–Ca2+-SFD medium [i.e., prostate epithelial cell basal medium (PrEBM) complete medium] are commercially available from Clonetics, Inc. (Walkersville, MD). These PrEC cells are p27/kip1, AR, α-methylacyl-CoA racemase (AMACR), PSA, hK2, and PMSA negative while expressing ΔNp63α, TGT-2, GST-Pi, CD44s, involucrin, Notch-1, Jagged-1, and cytokeratins 5 and 14 (i.e., they are transit-amplifying cells; refs. 79, 1419). Although not immortal, these PrEC cells can be propagated for up to 10 serial passages before becoming growth arrested (7, 9, 10).

Based on this success, many groups are using variations of low-Ca2+-SFD medium containing known growth factors [e.g., PrEBM complete medium or keratinocyte serum-free medium (K-SFM) complete medium from Invitrogen (Carlsbad, CA) in an attempt to grow sufficient numbers of prostate cancer cells from surgical material obtained at radical prostatectomy for molecular analysis and to establish new serially passageable prostate cancer cell lines (2030). Such a serum-free approach is favored because FCS contains a variety of undefined, and thus hard to standardize, factors in addition to its potent ability to stimulate growth to unwanted prostate stromal cells (i.e., fibroblasts and smooth muscle cells). Using such a low-Ca2+-SFD medium approach, it has been claimed that pure populations of prostate cancer cells can be grown and propagated from starting prostate cancer tissue (2528). In contrast to these low-Ca2+-SFD medium approaches, all of the presently available, serially propagated human prostate cancer cell lines (i.e., DU-145, PC-3, LNCaP, C4-2B, LAPC-4, CWR22-Rv1, CWR-R1, MDA-PCA-2A, MDA-PCA-2B, DuCap, and VCap) were originally established with and are maintained in 10% FCS containing medium whose Ca2+ is between 650 and 1,860 μmol/L (18). This raises the issue of when using surgical material whether what grows out in low-Ca2+-SFD medium are truly cancer cells as opposed to highly proliferative transit-amplifying cells derived from basal epithelium of normal glands contaminating the starting cancerous tissue. This latter possibility is supported by the demonstration that (a) such low-Ca2+-SFD medium cultures usually are not capable of being propagated beyond 10 passages although they are derived from cancerous tissue (25), (b) they usually lack molecular and karyotypic changes (e.g., 8p+ loss) characteristic of the starting malignant cells obtained directly from the patient without culture (29, 30), and (c) they usually are not tumorigenic when inoculated into nude mice (26, 27, 30). In contrast, all of the malignant prostate cancer cell lines established in high (i.e., >650 μmol/L)–Ca2+-FCS containing medium have karyotypic abnormalities, are spontaneously immortal (i.e., serially passageable), and are tumorigenic when inoculated into nude mice (18). Therefore, the present study was undertaken to test whether it is possible to maintain these definitively malignant prostate cancer cell lines in such low-Ca2+-SFD medium and to resolve the nature of cell lines established from primary prostate cancers in such low-Ca2+, serum-free medium.

Cell culture and reagents. Epithelial cell cultures were established from prostate tissues obtained from radical prostatectomy specimens from seven patients undergoing surgery for localized prostate cancer under an institutional review board–approved protocol as described previously (11). These “in-house”–initiated epithelial cell cultures as well as normal human PrEC obtained from Clonetics were maintained routinely (up to a maximum of 10 passages) in PrEBM [i.e., a modified low-Ca2+ keratinocyte medium] supplemented with bovine pituitary extract (BPE), epidermal growth factor (EGF), insulin, transferrin, hydrocortisone, retinoic acid, epinephrine, triiodothyronine, and gentamicin-amphotericin solution (Clonetics). The total Ca2+ level in this PrEBM complete medium was determined to be 260 ± 14 μmol/L using a colorimetric assay as described previously (31). Normal human prostatic fibroblasts were established and maintained in early-passage culture in RPMI 1640 plus 10% FCS (R-FCS) as described previously (11). Growth of cells was determined using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay as described previously (19).

The LNCaP, PC-3, and DU-145 human prostate cancer cell lines were routinely cultured in 10% FCS (Hyclone, Logan, UT) containing RPMI 1640 (Invitrogen, Carlsbad, CA) whose total Ca2+ was determined to be 650 ± 10 μmol/L. LAPC-4 human prostate cancer cells were grown in 10% FCS plus 1 nmol/L of the synthetic androgen R1881 (Perkin-Elmer, Wellesley, MA) containing Iscove's medium (BioFluid, Rockville, MD) whose total Ca2+ is 1,558 ± 10 μmol/L. MDA-PCA-2B human prostate cancer cells were grown in BRFF-HPC1 (Athena ES, Baltimore, MD) complete medium containing CT, insulin, EGF, BPE, bovine serum albumin, hydrocortisone, and dihydrotestosterone plus 20% FCS whose total Ca2+ is 767 ± 10 μmol/L in tissue culture flasks coated with FNC (i.e., fibronectin and type I collagen) coating mix (Athena ES).

957E/hTERT cells were generously provided by Dr. John S. Rhim (Center for Prostate Disease Research, Department of Surgery, Uniform Services University of Health Sciences, Bethesda, MD; ref. 26). These cells were maintained routinely in K-SFM supplemented with insulin, EGF, BPE, transferrin, hydrocortisone, triiodothyronine, and ethanolamine commercially obtained from Life Technologies (i.e., K-SFM complete medium, Invitrogen) whose total Ca2+ was determined to be 75 ± 2 μmol/L.

Western blotting. Western blots were done on cell lysates equivalent to105 cells per lane as described previously (19). For PSA, hK2, and PSMA, the antibodies used were as described previously (32). Goat polyclonal antibodies specific for Notch-1 (C-20 diluted 1:100) and Jagged-1 (H-114 diluted 1:100) and rabbit polyclonal antibodies specific for GST-Pi (diluted 1:1,000), E-cadherin (H-108, diluted 1:200), and AR (N-20, diluted 1:200) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Goat polyclonal antibodies specific for the valine (1744) cleaved Notch intracellular domain (NICD) form of Notch-1 (diluted 1:1,000) was obtained from Cell Signaling (Beverly, MA).

Reverse transcription-PCR. The RNeasy kit (Qiagen, Inc., Valencia, CA) was used to isolate RNA from cultured cells. RNA (50 ng) was subjected to reverse transcription-PCR (RT-PCR) analysis using TaqMan reverse transcription reagents and SYBR Green PCR Master Mix from Applied Biosystems (Foster City, CA). Primer sequences and conditions are as follows:

Accession no. Primer Primer sequence (5′-3′) Localization Size (bp) Annealing temperature (°C) Cycle no. 
AF404697 v-E6-5′ TTAGAATGTGTGTACTGCAAGC 113-134 163 62 40 
 v-E6-3′ TATAACAATAATGTCTATACTCACT 251-275    
AF486344 v-E7-5′ ATGGAGATACACCTACATTGCA 5-26 156 62 40 
 v-E7-3′ TATTGTAATGGGCTCTGTCCG 140-160    
M99487 PSMA-5′ CTGCTGTGGTGGAGAAACTG 86-105 168 62 35 
 PSMA-3′ CTGTTCCTGCTAAATGTGGTAT 499-520    
NM_000610 CD44-5′ ACAAGTTTTGGTGGCACGCAG 183-203 168 64 35 
 CD44-3′ TATTGAAAGCCTTGCAGAGGTC 329-350    
NM_000526 CK14-5′ ACCTCTCCTCCTCCCAGTTC 1,335-1,354 146 64 32 
 CK14-3′ TCAGTTCTTGGTGCGAAGGAC 1,460-1,480    
NM_000424 CK5-5′ ACCAGGAGCTCATGAACACC 1,521-1,540 150 62 35 
 CK5-3′ CATATCCAGAGGAAACACTGC 1,650-1,670    
NM_000224 CK18-5′ CTGCTGAACATCAAGGTCAAG 1,147-1,168 157 62 32 
 CK18-3′ CTTTGCCATCCACTATCCGG 1,283-1,303    
BC013852 GAPDH-5′ ACCACAGTCCATGCCATCAC 28-47 150 62 32 
 GAPDH-3′ GTTCAGCTCAGGGATGACC 159-177    
NM_001648 PSA-5′ AAAAGCGTGATCTTGCTGGG 249-268 201 60 35 
 PSA-3′ TCACAGCATCCGTGAGCTC 432-450    
BC003551 TGM2-5′ GCATGGTCAACTGCAACGAT 761-780 150 60 32 
 TGM2-3′ GGCCATACTTGACGCGCTG 892-910    
Accession no. Primer Primer sequence (5′-3′) Localization Size (bp) Annealing temperature (°C) Cycle no. 
AF404697 v-E6-5′ TTAGAATGTGTGTACTGCAAGC 113-134 163 62 40 
 v-E6-3′ TATAACAATAATGTCTATACTCACT 251-275    
AF486344 v-E7-5′ ATGGAGATACACCTACATTGCA 5-26 156 62 40 
 v-E7-3′ TATTGTAATGGGCTCTGTCCG 140-160    
M99487 PSMA-5′ CTGCTGTGGTGGAGAAACTG 86-105 168 62 35 
 PSMA-3′ CTGTTCCTGCTAAATGTGGTAT 499-520    
NM_000610 CD44-5′ ACAAGTTTTGGTGGCACGCAG 183-203 168 64 35 
 CD44-3′ TATTGAAAGCCTTGCAGAGGTC 329-350    
NM_000526 CK14-5′ ACCTCTCCTCCTCCCAGTTC 1,335-1,354 146 64 32 
 CK14-3′ TCAGTTCTTGGTGCGAAGGAC 1,460-1,480    
NM_000424 CK5-5′ ACCAGGAGCTCATGAACACC 1,521-1,540 150 62 35 
 CK5-3′ CATATCCAGAGGAAACACTGC 1,650-1,670    
NM_000224 CK18-5′ CTGCTGAACATCAAGGTCAAG 1,147-1,168 157 62 32 
 CK18-3′ CTTTGCCATCCACTATCCGG 1,283-1,303    
BC013852 GAPDH-5′ ACCACAGTCCATGCCATCAC 28-47 150 62 32 
 GAPDH-3′ GTTCAGCTCAGGGATGACC 159-177    
NM_001648 PSA-5′ AAAAGCGTGATCTTGCTGGG 249-268 201 60 35 
 PSA-3′ TCACAGCATCCGTGAGCTC 432-450    
BC003551 TGM2-5′ GCATGGTCAACTGCAACGAT 761-780 150 60 32 
 TGM2-3′ GGCCATACTTGACGCGCTG 892-910    

Immunohistochemistry. Immunohistochemistry was done on cells both grown to 80% confluency in four-well chamber slides (Nalge-Nunc International, Naperville, IL) and as single-cell suspensions. The binding of the primary antibody was detected using the Envision Plus kit from DAKO (Carpinteria, CA). Each of the cell lines was characterized by immunocytochemistry for the expression of (a) AR using a rabbit polyclonal N-20 at 1:500 dilution, (b) p63 using a AB-4 antibody (cocktail; NeoMarker, Fremont, CA) at 1:50 dilution, (c) p27 using kip1 mouse monoclonal antibody (BD Biosciences, San Diego, CA) at 1:50 dilution, (d) cytokeratin 5 using rabbit polyclonal MK5 (Berkeley Antibody Co., Berkeley, CA) at 1:15,000 dilution, (e) cytokeratin 8 using mouse monoclonal C51 (Imgenex, San Ramon, CA) at 1:800 dilution, (f) cytokeratin 14 using mouse monoclonal LL002 (Imgenex) at 1:100 dilution, (g) cytokeratin 18 using mouse monoclonal DC10 (Imgenex) at 1:1,600 dilution, (h) E-cadherin using rabbit polyclonal H-108 antibody at 1:400 dilution, and (i) AMACR using rabbit monoclonal P5045 (Zeta Co., Sierra Madre, CA) at 1:80 dilution.

Additional assays. PSA, hK2, and PSMA expression were assayed as described previously (32). Tumorigenicity in nude mice was assayed as described previously (11). Cytogenetic analysis was via G-banding as described previously (33). SHE78-7 mouse monoclonal anti-E-cadherin neutralizing antibody (2 μg/mL, Zymed Laboratories, South San Francisco, CA) was used to block E-cadherin-dependent homotypic interaction. To inhibit γ-secretase, the peptidyl L-685,458 inhibitor (Bachem, King of Prussia, PA) and the benzodiazepine inhibitor XIX (Calbiochem, La Jolla, CA) were used at 10 μmol/L as described by Shearman et al. (34) and Chrucher et al. (35), respectively. The data obtained were analyzed by one-way ANOVA. Values represent the mean ± SE. P < 0.05 was considered significant.

Phenotypic characteristics of normal transit-amplifying versus malignant prostate cells. A panel of markers, which discriminated between normal human transit-amplifying and malignant prostatic cells, was developed and validated using commercially obtained normal transit-amplifying (i.e., PrEC) cells and a series of eight permanent human prostatic cancer cell lines. The eight cancer lines were selected to encompass the full range of clinical prostate cancers from localized cancer to soft tissue and bone metastases (Table 1). All of these malignant lines are tumorigenic in nude mice. None of these malignant lines express the prostatic basal cell characteristic markers p63 and cytokeratin 14, whereas six of eight lines express the characteristic prostatic luminal cell markers AR, PSA, hK2, and PSMA. Only the DU-145 and PC-3 cells do not express these latter luminal markers, although they and all of the eight malignant lines do express p27/kip1 and the luminal cell markers cytokeratins 8 and 18. All of the malignant cells express AMACR. In contrast, PrEC cells growing in low (i.e., 260 ± 14 μmol/L)–Ca2+-SFD medium (i.e., PrEBM complete medium) are not tumorigenic and express the phenotypic profile characteristics of normal basally derived transit-amplifying cells (i.e., they are p63, GST-Pi, cytokeratin 5, and cytokeratin 14 positive but p27/kip1, AMACR, AR, PSA, hK2, and PSMA negative; Table 1).

Table 1.

Characteristics of the prostatic cell lines used in this study

Cell lineInitial site of originTumorigenic in nude miceExpression of indicated phenotypic marker protein*
p63AMACRp27/kip1Cytokeratin 5Cytokeratin 14Cytokeratin 8Cytokeratin 18AR expression7PSA (ng/107 cells/d)hK2 (ng/107 cells/d)PSMA (pmol/min/mg protein)
PrEC Normal prostate No +++ − − +++ ++ ++ ++ − − − − 
957E/hTERT Primary cancer No +++ − − +++ ++ ++ ++ − − − − 
CWR22-Rv1 Primary cancer Yes − ++ +++ − − +++ +++ +++ (M) 12 ± 3 11 ± 3 23 ± 2 
LAPC-4 Lymph node metastases Yes − +++ ++ − +++ +++ +++ (M) 12 ± 3 11 ± 3 23 ± 2 
MDA-PCA-2B Bone metastases Yes − ++ ++ − +++ +++ ++ (M) 2,155 ± 355 1 ± 1 19 ± 1 
LNCaP-early passage Lymph node metastases Yes − +++ ++ − − +++ +++ ++ (M) 2,155 ± 355 1 ± 1 19 ± 1 
LNCaP-late passage Lymph node metastases Yes − ++ ++ − − +++ +++ ++ (M) 400 ± 35 22 ± 5 165 ± 4 
C4-2B Lymph node metastases Yes − ++ ++ − − +++ +++ ++ (M) 2,100 ± 500 238 ± 54 109 ± 4 
PC-3 Bone metastases Yes − − +++ +++ − − − − 
DU-145 Brain metastases Yes − ++ − +++ +++ − − − − 
Cell lineInitial site of originTumorigenic in nude miceExpression of indicated phenotypic marker protein*
p63AMACRp27/kip1Cytokeratin 5Cytokeratin 14Cytokeratin 8Cytokeratin 18AR expression7PSA (ng/107 cells/d)hK2 (ng/107 cells/d)PSMA (pmol/min/mg protein)
PrEC Normal prostate No +++ − − +++ ++ ++ ++ − − − − 
957E/hTERT Primary cancer No +++ − − +++ ++ ++ ++ − − − − 
CWR22-Rv1 Primary cancer Yes − ++ +++ − − +++ +++ +++ (M) 12 ± 3 11 ± 3 23 ± 2 
LAPC-4 Lymph node metastases Yes − +++ ++ − +++ +++ +++ (M) 12 ± 3 11 ± 3 23 ± 2 
MDA-PCA-2B Bone metastases Yes − ++ ++ − +++ +++ ++ (M) 2,155 ± 355 1 ± 1 19 ± 1 
LNCaP-early passage Lymph node metastases Yes − +++ ++ − − +++ +++ ++ (M) 2,155 ± 355 1 ± 1 19 ± 1 
LNCaP-late passage Lymph node metastases Yes − ++ ++ − − +++ +++ ++ (M) 400 ± 35 22 ± 5 165 ± 4 
C4-2B Lymph node metastases Yes − ++ ++ − − +++ +++ ++ (M) 2,100 ± 500 238 ± 54 109 ± 4 
PC-3 Bone metastases Yes − − +++ +++ − − − − 
DU-145 Brain metastases Yes − ++ − +++ +++ − − − − 
*

+, low; ++, medium; +++, high level of expression; −, not detectable.

M, mutated gene; WT, wild-type.

Phenotypic characteristics of cells grown from surgical material in low-calcium-serum-free, growth factor–defined medium.Table 1 documents that there are characteristic markers that discriminate between normal transit-amplifying and prostatic cancer cells. These characteristic markers were used to determine the cell types that grow out from radical prostatectomy specimens from a series of seven patients with localized prostate cancer using low-Ca2+-SFD medium (i.e., PrEBM complete medium). Although specific specimens from each of these prostatectomies contained histologic cancer, four of the seven specific specimens used to initiate cultures had no histologically detectable cancer. Of the remaining three, one had >50% and the other two had >20% of the area involved with Gleason score 6 cancers in the specimen used for culture. Histologic examination (Fig. 1A) and immunocytochemical staining showed that for the cancers containing specimens the malignant cells express AR (Fig. 1B), PSA (data not shown), PSMA (data not shown), and AMACR (Fig. 1D) but not p63 (Fig. 1C), GST-Pi (data not shown), or CD44s (data not shown). When these cancer-containing specimens were cultured in the low-Ca2+-SFD medium, cells that grew out expressed p63 (Fig. 1E), cytokeratins 5 and 14 (data not shown), and GST-Pi and CD44s (data not shown) but not AMACR (Fig. 1F) AR (data not shown), PSA (data not shown), or PSMA (data not shown). In addition, staining for α-smooth muscle actin and type I collagen was negative, documenting the lack of stromal cell growth in the low-Ca2+-SFD medium. These characteristics were identical to the cells established from the patients whose specimens contained no histologically detectable cancers. Also similar to cells from the patients whose specimens did not contain cancer, cells from the cancers containing specimens could be serially cultured in low-Ca2+-SFD medium only for 8 to 10 passages before becoming proliferatively quiescent. Inoculation of nude mice (n = 5 per specimen) with 0.2 mL Matrigel containing 2 × 106 viable cells from the three cancer-containing specimens passaged twice in low-Ca2+-SFD medium produced no palpable tumors even if followed for >6 months.

Figure 1.

A, histology of the radical prostatectomy specimen that contains >50% Gleason score 6 cancer used for culture in low-Ca2+-SFD medium. Immunocytochemical staining for AR (B), p63 (C), and AMACR (D) in serial sections of specimen from (A). Asterisks, normal prostatic epithelial gland contaminating areas of malignant prostate cells. Immunocytochemical staining of fifth passage cells cultured from specimens from (A) for p63 (E) and AMACR (F).

Figure 1.

A, histology of the radical prostatectomy specimen that contains >50% Gleason score 6 cancer used for culture in low-Ca2+-SFD medium. Immunocytochemical staining for AR (B), p63 (C), and AMACR (D) in serial sections of specimen from (A). Asterisks, normal prostatic epithelial gland contaminating areas of malignant prostate cells. Immunocytochemical staining of fifth passage cells cultured from specimens from (A) for p63 (E) and AMACR (F).

Close modal

Low-calcium-serum-free, growth factor–defined medium selects against malignant prostate cells and normal prostatic fibroblasts. These data document that use of low-Ca2+-SFD medium results in the growth of normal transit-amplifying cells and not prostate fibroblasts or cancer cells also present within the original surgical material. In contrast, all of the currently available permanent human prostate cancer cell lines were established with and are maintained in medium containing ≥650 μmol/L Ca2+ and 10% to 20% FCS (18). Therefore, to test whether the inability to grow malignant cells from the cancer-containing radical prostatectomy specimens is simply a random event or whether such low-Ca2+-SFD medium selected against growth of malignant prostate cells, the cancer lines presented in Table 1 were tested for their ability to grow continuously in two different types of low-Ca2+-SFD medium. The first low-Ca2+-SFD medium was the PrEBM complete medium with defined growth factors (Clonetics) used for growth of commercially obtained PrEC cells, which contains 260 ± 14 μmol/L Ca2+, and the second was K-SFM complete medium with defined growth factors (Invitrogen), which contains 75 ± 2 μmol/L Ca2+.

When FCS is not additionally provided to either of the low-Ca2+-SFD media, all of the cancer cell lines attach poorly to the culture dishes, slow their growth, and begin to die. This process is so dramatic that for some of the lines (i.e., early-passage LNCaP, MDA-PCA-2B, and LAPC-4) all the cells die within two passages. For the other lines (i.e., late-passage LNCaP, C4-2B, PC-3, DU-145, and CWR22-Rv1), such sterilization requires a total of three passages. Likewise, when early-passage human prostatic fibroblasts, which were established initially in R-FCS and which can be passaged serially 10 to 15 times in this high (i.e., 650 μmol/L)–Ca2+ medium, are switched to low-Ca2+ PrEBM complete medium, the cells died within one passage. Thus, none of the human prostatic fibroblasts or cancer cell lines survive long-term without FCS components in low-Ca2+ medium even when they are given insulin, EGF, BPE, transferrin, hydrocortisone, triiodothyronine, and ethanolamine in their medium.

FCS and elevated calcium are required for optimal growth of malignant prostate cells but inhibits growth of normal transit-amplifying cells. The lack of growth of malignant prostate cells in the low-Ca2+-SFD medium could be due to the low-Ca2+-SFD medium either actively suppressing the growth of malignant prostatic cells or passively by its inability to provide critical nutrients and/or attachment growth factors. To resolve this, LNCaP cells from passage 33 were plated and exposed for 1 week to (a) their standard R-FCS medium, (b) low (i.e., 75 μmol/L)–Ca2+ K-SFM complete medium alone, (c) K-SFM complete medium plus added Ca2+ to bring its concentration 650 μmol/L (i.e., equivalent to the level in R-FCS medium), (d) K-SFM complete medium plus 10% FCS (i.e., 367 μmol/L Ca2+), (e) K-SFM complete medium plus 10% dialyzed FCS (i.e., 100 μmol/L Ca2+), or (f) K-SFM complete medium plus 10% FCS and added Ca2+ to bring its total concentration to 650 μmol/L. The results documented that in standard R-FCS medium there were 28,780 ± 2,478 viable cells after 1 week versus only 8,080 ± 680 viable cells (i.e., 28% of R-FCS value) when maintained for 1 week in the low-Ca2+ K-SFM complete medium. Supplementing the K-SFM complete medium to raise the Ca2+ to 650 μmol/L resulted in only a modest increase in cell number (9,650 ± 1,050). In contrast, addition of 10% FCS to the K-SFM complete medium increased the total Ca2+ from 75 ± 2 to 367 ± 5 μmol/L and supplied additional undefined growth factors resulting in a 2.5-fold increase (P < 0.05) in viable cell number at 1 week of treatment (i.e., 24,376 ± 1,229) compared with K-SFM complete medium alone or 85% of the number in standard R-FCS medium. If 10% dialyzed FCS was added to K-SFM complete medium, which supplied additional growth factors but did not increase the Ca2+ (i.e., 100 μmol/L), there was less growth at 1 week (i.e., 15,810 ± 995) compared with when this dialyzed FCS was used but twice as much growth than in unsupplemented K-SFM complete medium. When the K-SFM complete medium was supplemented with both 10% FCS plus sufficient added Ca2+ to reach a level of 650 μmol/L (i.e., equivalent with R-FCS medium), there were 29,400 ± 2,915 viable cells at 1 week (a value equivalent with that of R-FCS medium). These results document that the inability of the low-Ca2+-SFD medium to maintain the growth of malignant prostate cells is due to insufficient levels of Ca2+ coupled with a lack of attachment growth factors.

In contrast to the growth enhancement when prostate cancer cells are exposed to elevated Ca2+ and FCS, when early-passage (i.e., <5) PrEC cells are switched from low (i.e., 260 ± 10 μmol/L)–Ca2+, serum-free PrEBM complete medium to such medium supplemented with either 10% FCS raising the final Ca2+ to 575 ± 5 μmol/L or with only Ca2+ to raise its level to 650 μmol/L as in R-FCS medium, the normal prostatic transit-amplifying cells become growth arrested. Time-lapse videomicroscopy documented that, in low-Ca2+-SFD medium, transit-amplifying cells divide with a doubling time of ∼48 hours. Once a transit-amplifying cell divides, the daughter cells generally do not remain in contact but migrate to fill the culture surface (Fig. 2A). If the cells are allowed to fill the surface (i.e., reach confluence) before subculturing, proliferation ceases due to contact inhibition producing a continuous monolayer (Fig. 2B). If transit-amplifying cultures in low-Ca2+-SFD medium are supplemented with either Ca2+ to raise its level to 650 μmol/L or 10% FCS to raise Ca2+ to 575 μmol/L, the cells rapidly (i.e., within 1 hour) gain an enhanced motility. This enhanced motility allows a subset of cells to contact neighboring cells. Once this contact is made, the cells remain attached producing foci of stratified cells (Fig. 2C). Within these stratified foci, cell proliferation ceases due to contact inhibition as detected by a lack of mitotic figures. Between these stratified foci, there are large areas of unoccupied surface and scattered individual cells (Fig. 2C). These isolated individual transit-amplifying cells eventually stop their motility and undergo squamous differentiation into cornified envelopes and die. This cornification is documented by the observation that these cells are not solubilized by 1% SDS treatment. Instead, such treatment produces cornified envelopes (Fig. 2D). We have documented previously that the death of transit-amplifying cells induced by a variety of agents always results in the cross-linking of involucrin by these cells into cornified envelopes (19). In contrast, death of prostatic cancer cells induced by a variety of agents (19) or by culturing in low cell density in low-Ca2+-SFD medium did not result in squamous differentiation into cornified envelopes.

Figure 2.

Phase-contrast image of PrEC cells in low (i.e., 260 ± 10 μmol/L)–Ca2+ serum-free PrEBM complete medium at 40% to 60% confluence (A) or at confluence (B). C, at ≈40% to 60% confluence, medium supplemented to 650 μmol/L Ca2+ for 3 days. D, such Ca2+-supplemented cultures were exposed to 1% SDS documenting the development of insoluble cornified envelopes.

Figure 2.

Phase-contrast image of PrEC cells in low (i.e., 260 ± 10 μmol/L)–Ca2+ serum-free PrEBM complete medium at 40% to 60% confluence (A) or at confluence (B). C, at ≈40% to 60% confluence, medium supplemented to 650 μmol/L Ca2+ for 3 days. D, such Ca2+-supplemented cultures were exposed to 1% SDS documenting the development of insoluble cornified envelopes.

Close modal

Nature of 957E/hTERT prostate cell line established in low-calcium-serum-free, growth factor–defined medium. Recently, low (i.e., 75 ± 2 μmol/L)–Ca2+-SFD medium containing defined growth factors (i.e., K-SFM complete medium) commercially available from Invitrogen has been used with prostate cancer tissues to establish “prostate cancer lines” (2527). For example, a prostate cancer cell line has been established from localized prostate cancer tissue from a man with a strong family history of prostate cancer using such medium (26). This was accomplished by immortalizing cells derived from the starting cancer tissue with a retrovirus encoding the human telomerase catalytic subunit of human telomerase reverse transcriptase (hTERT) to produce a line denoted as 957E/hTERT (26). Based on the results of the present studies that such low-Ca2+-SFD medium, however, selects for normal transit-amplifying and against malignant prostate cells, the nature of this prostatic cancer line needs validation. Originally, the 957E/hTERT cells had a diploid karyotype, with the only clonal abnormality being a light staining region on the end of chromosome 4q (26). Between in vitro passages 10 and 32, this line underwent trisomy of chromosome 20 (26). On further passage in our laboratory, the line has remained stable and not gained further chromosomal changes. Because such trisomy of chromosome 20 is often characteristic of cells immortalized by E6/E7 of human papilloma virus 16 (36), the 957E/hTERT cells were tested for the presence of E6/E7 by RT-PCR (Fig. 3). Western blots confirmed that these cells were also negative for SV40 large T antigen and Ela viral protein (data not shown). Although these cells were negative for these viral antigens, they were positive for hTERT expression by RT-PCR (data not shown). 957E/hTERT cells are not tumorigenic when inoculated into nude mice as reported previously (26). These data raise the issue of whether by using low (i.e., 75 μmol/L)–Ca2+-SFD medium what was hTERT immortalized was not prostate cancer cells but proliferating transit-amplifying cells derived from the basal epithelium of a normal prostatic gland contaminating the starting cancer tissue.

Figure 3.

RT-PCR for detection of mRNA expression of indicated genes in 957E/hTERT versus LNCaP human prostatic cancer cells.

Figure 3.

RT-PCR for detection of mRNA expression of indicated genes in 957E/hTERT versus LNCaP human prostatic cancer cells.

Close modal

To test this possibility, 957E/hTERT cells growing in the low (i.e., 75 μmol/L)–Ca2+ K-SFM complete medium (Fig. 4A) were evaluated for their expression profile for the discriminatory markers used in Table 1 and their response to adding Ca2+ to the low-Ca2+-SFD medium. These results document that growing 957E/hTERT cells express the basal cell–specific p63 protein within their cell nuclei (Fig. 4B) and cytokeratin 5 (Fig. 4C), cytokeratin 14, and GST-Pi (data not shown) but not p27/kip1, AMACR, AR, PSA, hK2, or PSMA (data not shown). In addition, using RT-PCR, these cells express mRNA for the basal cell characteristic markers CD44, TGM2, cytokeratin 5, and cytokeratin 14 while being negative for the luminal markers AR, PSMA, and PSA (Fig. 3). To test their response to Ca2+, 957E/hTERT cells were plated and exposed to (a) their standard K-SFM complete medium with total Ca2+ (75 ± 2 μmol/L), (b) K-SFM complete medium plus 10% dialyzed FCS with total Ca2+ (100 ± 15 μmol/L), (c) K-SFM complete medium plus 10% undialyzed FCS with total Ca2+ (367 ± 5 μmol/L), (d) K-SFM complete medium supplemented with added Ca2+ to bring level to 650 μmol/L, or (e) R-FCS with total Ca2+ (650 ± 10 μmol/L). After 1 week, there were 47,446 ± 2,150 viable cells in the K-SFM complete medium versus 42,510 ± 3,150 with K-SFM medium supplemented with 10% dialyzed serum. In contrast, these were only 1,200 ± 457 (P < 0.05) viable cells in the K-SFM complete medium plus 10% FCS, 1,050 ± 490 (P < 0.05) viable cells in Ca2+ supplemented K-SFM complete medium, and only 810 ± 57 (P < 0.05) viable cells in the R-FCS medium. Like the situation for the normal transit-amplifying cells, when exponentially proliferating 957E/hTERT cells are switched from their low-Ca2+ K-SFM complete medium to medium containing >360 μmol/L Ca2+, a subset of cells aggregate and stop proliferating, whereas scattered individual cells undergo squamous differentiation-induced death detectable as 1% SDS-insoluble cornified envelopes (Fig. 4D). These combined results document that 957E/hTERT are derived from normal transit-amplifying cells, not prostate cancer cells.

Figure 4.

A, phase-contrast image of 957E/hTERT cells growing on a chamber slide in low (i.e., 75 ± 5 μmol/L)–Ca2+ serum-free K-SFM complete medium. B, immunocytohistochemical detection of nuclear p63 protein expression assayed on a single-cell/agar suspension of 957E/hTERT cells. C, immunocytohistochemical detection of cytokeratin 5 assayed on 957E/hTERT cells grown on a chamber slide in PrEBM complete medium. D, phase-contrast image of 957E/hTERT cells grown in PrEBM complete medium supplemented to 650 μmol/L Ca2+ and exposed to 1% SDS to detect cornified envelopes.

Figure 4.

A, phase-contrast image of 957E/hTERT cells growing on a chamber slide in low (i.e., 75 ± 5 μmol/L)–Ca2+ serum-free K-SFM complete medium. B, immunocytohistochemical detection of nuclear p63 protein expression assayed on a single-cell/agar suspension of 957E/hTERT cells. C, immunocytohistochemical detection of cytokeratin 5 assayed on 957E/hTERT cells grown on a chamber slide in PrEBM complete medium. D, phase-contrast image of 957E/hTERT cells grown in PrEBM complete medium supplemented to 650 μmol/L Ca2+ and exposed to 1% SDS to detect cornified envelopes.

Close modal

Mechanism for the differential response to elevated calcium by normal versus malignant prostate cells. Normal prostatic transit-amplifying cells express the plasma transmembrane receptor Notch-1 and its plasma transmembrane ligand Jagged-1 (7, 17, 37). Notch-1 is a plasma transmembrane receptor whose signaling is associated with nonterminally differentiated, proliferating progenitor cell types (38). Once synthesized, the 300-kDa full-length Notch-1 protein is cleaved by furin-like protease in the Golgi during trafficking of Notch to the plasma membrane producing two subunits, the NH2-terminal 180-kDa extracellular inhibitory fragment (termed Nec) and the plasma transmembrane COOH-terminal 120-kDa intracellular fragment (termed Notch-CTF). These subunits remain associated as a heterodimer to form functionally inactive non-ligand-bound receptor at the plasma membrane (39, 40). The Notch-1 ligand, Jagged-1, is also a plasma transmembrane heterodimeric protein (41) composed of a NH2-terminal extracellular ligand fragment bound to a plasma membrane COOH-terminal intracellular fragment (CTF) produced by proteolysis of the full-length 135-kDa Jagged protein. Reciprocal binding to the extracellular NH2-terminal fragments of the Notch-1 receptor and Jagged-1 ligands between adjacent cells induces the dissociation of the two heterodimeric complexes liberating the CTF of both Notch-1 and Jagged-1. Once liberated, Notch-CTF undergoes cleavage of an extracellular piece to produce a product known as Notch extracellular truncation (NEXT; refs. 39, 40). NEXT is then proteolyzed by γ-secretase to produce a product known as NICD, which contains nuclear localization domains and is translocated to the cell nucleus where it is associated with the nuclear proteins of the RBP-Jκ family to form a transcriptional activator of gene expression (3840).

Both PrEC and 957E/hTERT cells express the full-length, CTF, NEXT, and NICD forms of Notch-1 (Fig. 5A and B) as well as full-length and CTF forms of Jagged-1 (Fig. 5C). Dose-response studies documented that a concentration of 10 μmol/L of either of the structurally unrelated γ-secretase inhibitors L-685,458 or compound XIX completely inhibits the production of the NICD transcription factor needed for nuclear signaling by both PrEC and 957E/hTERT cells (Fig. 5D). When PrEC or 957E/hTERT cells are inoculated and allowed to attach overnight before being exposed to 10 μmol/L of either of the γ-secretase inhibitors, there were no viable cells remaining after 5 days of such exposure. This toxic effect was specific because similar exposure of the prostatic cancer cell lines growing in their standard high-Ca2+ FCS containing medium to 10 μmol/L of either of the γ-secretase inhibitors over a 7-day period produced only a slight inhibition (P < 0.05) of proliferation and no death of any of the cancer lines tested (e.g., percentage growth inhibition was 55 ± 10% for LNCaP, 17 ± 5% for LAPC-4, 20 ± 4% for CWR22-Rv1, 34 ± 6% for PC-3, and 29 ± 6% for DU-145 cells using compound XIX). This inability to induce death of prostatic cancer cells is not due to either lack of Notch-1 and Jagged-1 expression or signaling as documented by the detection of full-length, CTF, and NICD forms of Notch-1 by these malignant cells albeit at much reduced levels (Fig. 5A-C) or the inability of 10 μmol/L γ-secretase XIX inhibitor to block NICD production by the cancer cells (Fig. 5D).

Figure 5.

Western blot of Notch-1 and Jagged-1 expression by normal versus prostatic cancer cells. A, detection of various Notch-1 fragments using a COOH-terminal Notch-1 antibody. FL, full length. B, detection of NICD using an antibody recognizing the valine (1744) cleaved NICD. Note values below the blot are density values normalized to PrEC cells. C, detection of various Jagged-1 fragments. D, inhibition of γ-secretase-induced NICD production. 957E/hTERT, NICD level in control 957E/hTERT cells; 957E/hTERT + Inhibitor, NICD level in 957E/hTERT cells treated with 10 μmol/L γ-secretase inhibitor XIX; LAPC-4, NICD level in control LAPC-4 prostate cancer cells; LAPC-4 + inhibitor, NICD level in LAPC-4 cells treated with 10 μmol/L γ-secretase inhibitor XIX. E, PrEC-SFD, NICD in PrEC cells at low density (i.e., 1,000 cells/cm2) in low (i.e., 260 μmol/L)–Ca2+ PrEBM complete medium; PrEC-SFD + CaCl2, NICD in PrEC cells at low density (1,000 cells/cm2) in PrEBM complete medium with 600 μmol/L Ca2+.

Figure 5.

Western blot of Notch-1 and Jagged-1 expression by normal versus prostatic cancer cells. A, detection of various Notch-1 fragments using a COOH-terminal Notch-1 antibody. FL, full length. B, detection of NICD using an antibody recognizing the valine (1744) cleaved NICD. Note values below the blot are density values normalized to PrEC cells. C, detection of various Jagged-1 fragments. D, inhibition of γ-secretase-induced NICD production. 957E/hTERT, NICD level in control 957E/hTERT cells; 957E/hTERT + Inhibitor, NICD level in 957E/hTERT cells treated with 10 μmol/L γ-secretase inhibitor XIX; LAPC-4, NICD level in control LAPC-4 prostate cancer cells; LAPC-4 + inhibitor, NICD level in LAPC-4 cells treated with 10 μmol/L γ-secretase inhibitor XIX. E, PrEC-SFD, NICD in PrEC cells at low density (i.e., 1,000 cells/cm2) in low (i.e., 260 μmol/L)–Ca2+ PrEBM complete medium; PrEC-SFD + CaCl2, NICD in PrEC cells at low density (1,000 cells/cm2) in PrEBM complete medium with 600 μmol/L Ca2+.

Close modal

These results document that survival of normal prostatic transit-amplifying cells requires unique Notch-1-mediated signaling, which cannot be provided by signaling induced by the added growth factors in the serum-free defined medium. Such Notch-1 signaling is usually not cell autonomous but instead requires direct reciprocal cell-cell (i.e., ligand) interaction (42). This raises the issue of how such Notch-1 signaling can occur when the PrEC and 957E/hTERT cells are routinely plated at a cell density (e.g., 6,000 cells/cm2) that does not allow a high degree of initial cell-cell contact. Cell autonomous, ligand-independent Notch-1 signaling can be induced by lowering the Ca2+ concentration to prevent Nec from initially binding with the CTF form of Notch-1 allowing its proteolysis to liberate NICD inducing nuclear signaling (43). Such low-Ca2+-dependent, cell autonomous (i.e., ligand-independent) Notch-1 signaling could be why normal transit-amplifying cells (i.e., PrEC and 957E/hTERT) can be established selectively at initially low cell density from surgical material when using low-Ca2+-SFD medium. To evaluate this possibility, single-cell suspensions of PrEC or 957/hTERT cells, respectively, were plated in either low-Ca2+-SFD medium or such medium supplemented to raises its Ca2+ to 600 μmol/L into 100-mm tissue culture dishes at very low cell density (1,000 cells/cm2), so that, even by 24 hours, ≥95% of the attached cells are not in contact with other cells. These low-density cultures were harvested at 24 hours and Western blots were done. These results documented that NICD is still produced when cells are not in contact in low-Ca2+-SFD medium but not produced when such low-density cultures are exposed to 600 μmol/L Ca2+ (Fig. 5E).

If these initially low-density PrEC cell cultures are allowed to grow to 50% confluence in low (i.e., 260 μmol/L)–Ca2+ PrEBM complete-SFD medium before being supplemented to raise the Ca2+ to 600 μmol/L, rapid cellular aggregation of subconfluent cultures occurs (Fig. 6A versus Fig. 6B). This results in stratification and cessation of proliferation (i.e., complete lack of mitotic figures) within 24 to 48 hours of exposure of the cell aggregates to elevated Ca2+. This cessation is coincident with a 2.4-fold increase in cellular E-cadherin content in the cells, exposure to 600 μmol/L Ca2+, and a 17-fold increase in the p27/kip1 (Fig. 6D). Immunocytochemical staining for E-cadherin showed weak cytoplasmic expression in PrEC cells growing in low-Ca2+ medium but strong expression at the plasma membrane junctions between stratified cells in culture supplemented to 600 μmol/L Ca2+. If subconfluent cultures of PrEC cells are cotreated with 2 μg/mL of the SHE78-7 mouse monoclonal anti-E-cadherin neutralizing antibody when the cultures are supplemented to raise Ca2+ to 600 μmol/L, stratification of the PrEC cells is completely inhibited (Fig. 6C). Likewise, the cells cotreated with neutralizing antibody continue to proliferate as detected by both the presence of mitotic figures and an increase in total cell number. Thus, antibody-treated Ca2+-supplemented cultures continue to proliferate beyond day 7, which is the time when non-antibody-treated cultures not supplemented with Ca2+ (i.e., control cells) become confluent and stopped proliferating (i.e., contact inhibited). In the low-Ca2+ medium, such contact inhibition is associated with an enhanced (i.e., 30-fold) expression of p27/kip1 protein but no increases in E-cadherin (Fig. 6D). Thus, by day 10, the antibody-treated cultures with 600 μmol/L Ca2+ had twice as many cells as in low-Ca2 contact-inhibited control cultures (i.e., 89,000 ± 695 versus 54,890 ± 4,540; P < 0.05) and were still proliferating.

Figure 6.

A, subconfluent cultures of PrEC cells growing in low (i.e., <260 μmol/L)–Ca2+-SFD medium. B, subconfluent cultures of PrEC cells exposed for 8 hours to SFD medium whose Ca2+ was raised to 600 μmol/L. C, subconfluent cultures exposed simultaneously with 2 μg/mL E-cadherin neutralizing antibody and to SFD medium with 600 μmol/L Ca2+. D, Western blots of E-cadherin and p27/kip1 protein expression content. PrEC control, cells growing in low (i.e., <270 μmol/L)–Ca2+ PrEBM complete medium; PrEC contact inhibited, cells grown to confluence in low-Ca2+ PrEBM complete medium; PrEC 600 μmol/L CaCl2, cells growing in PrEBM medium supplemented to 600 μmol/L Ca2+ for 24 hours; 957E/h control, 957E/hTERT cells growing in low (i.e., >75 μmol/L)–Ca2+ K-SFD complete medium; 957E/h contact inhibited, cells grown to confluence in low-Ca2+ K-SFD complete medium; 957E/h 10% FCS, cells growing in K-SFD complete medium supplemented with 10% FCS, which raises Ca2+ to 375 μmol/L; 957E/h 10% dialyzed FCS, 957E/hTERT cells in K-SFD complete medium supplemented with 10% dialyzed FCS whose Ca2+ is 100 μmol/L. Number below each lane, relative level of expression normalized to PrEC or 957E/hTERT controls, respectively.

Figure 6.

A, subconfluent cultures of PrEC cells growing in low (i.e., <260 μmol/L)–Ca2+-SFD medium. B, subconfluent cultures of PrEC cells exposed for 8 hours to SFD medium whose Ca2+ was raised to 600 μmol/L. C, subconfluent cultures exposed simultaneously with 2 μg/mL E-cadherin neutralizing antibody and to SFD medium with 600 μmol/L Ca2+. D, Western blots of E-cadherin and p27/kip1 protein expression content. PrEC control, cells growing in low (i.e., <270 μmol/L)–Ca2+ PrEBM complete medium; PrEC contact inhibited, cells grown to confluence in low-Ca2+ PrEBM complete medium; PrEC 600 μmol/L CaCl2, cells growing in PrEBM medium supplemented to 600 μmol/L Ca2+ for 24 hours; 957E/h control, 957E/hTERT cells growing in low (i.e., >75 μmol/L)–Ca2+ K-SFD complete medium; 957E/h contact inhibited, cells grown to confluence in low-Ca2+ K-SFD complete medium; 957E/h 10% FCS, cells growing in K-SFD complete medium supplemented with 10% FCS, which raises Ca2+ to 375 μmol/L; 957E/h 10% dialyzed FCS, 957E/hTERT cells in K-SFD complete medium supplemented with 10% dialyzed FCS whose Ca2+ is 100 μmol/L. Number below each lane, relative level of expression normalized to PrEC or 957E/hTERT controls, respectively.

Close modal

Similar experiments were done using 957E/hTERT cells in low (i.e., 75 μmol/L)–Ca2+-KSFD complete medium supplemented with 10% FCS to raise the Ca2+ levels to 375 μmol/L or with dialyzed 10% FCS, which did not raise the Ca2+ (i.e., 100 ± 5 μmol/L). Like PrEC cells, when 957E/hTERT cells are allowed to reach confluence in low-Ca2+ KSFD complete medium, they up-regulate (i.e., 3.9-fold) expression of p27/kip1 and stop proliferating due to contact inhibition (Fig. 6D). Again, such up-regulation of p27/kip1 protein and resultant contact inhibition are not observed when the prostate cancer cell lines are grown in their standard high-Ca2+ medium. If subconfluent cultures are supplemented to raise the Ca2+ level, 957E/hTERT cells aggregate, stratify, and stop proliferating. This cessation of proliferation is coincident with a 3-fold increase in cellular E-cadherin and a 2.7-fold increase in p27/kip1 content (Fig. 6D). Like the situation with PrEC cells, this aggregation/stratification and proliferation inhibition of 957E/hTERT cells is completely prevented by cotreatment with neutralizing E-cadherin antibody. The cellular aggregation and proliferation inhibition as well as the elevation of E-cadherin and p27/kip1 (Fig. 6D) are negated by using dialyzed FCS, which did not raise medium Ca2+.

The enhancement of growth of PrEC and 957E/hTERT cells when E-cadherin-dependent contact inhibition is neutralized in medium containing >600 μmol/L Ca2+ is not observed for prostatic cancer cells. Morphologically, human prostate cancer cell lines (e.g., CWR22-Rv1, LAPC-4, and LNCaP) form multicell clusters when growing at subconfluence in high (i.e., <600 μmol/L)–Ca2+ serum-containing medium. If E-cadherin neutralizing antibody is added to the high-Ca2+ serum-containing medium, homotypic cell clustering does not occur. Such inhibition of homotypic interaction, however, does not stimulate the overall growth of any of the tested lines as documented by identical numbers of viable cells at the end of a 7-day observation period in the antibody-treated versus control cells.

Previous studies have documented that a critical level of Ca2+ is required both for converting Notch-1 from a constitutively active form allowing cell autonomous signaling to a ligand-dependent process (43) and E-cadherin from an inactive conformation to one mediating homotypic cell-cell interaction (44). Structural biology studies have documented that the mechanism for this Ca2+ dependence is based on cooperative effects involving multiple Ca2+ binding sites whose occupancy regulates the structure of Notch-1 (43) and E-cadherin (44). At the 1 to 2 mmol/L physiologic tissue Ca2+ level, Notch-1 signaling is not cell autonomous but involves reciprocal cell-cell interactions (42). This is because the auto-inhibitory extracellular Notch NH2-terminal fragment (i.e., Nec) must bind with the Jagged-1 ligand of a neighboring cell to allow its dissociation from the Notch-CTF fragment. Once this occurs, the Notch-CTF fragment is proteolyzed to produce NICD, which translocates to the nucleus where it functions as a transcriptional coactivator (38). At low (i.e., <300 μmol/L)–Ca2+ level, however, the Notch-Nec does not associate with the Notch-CTF fragment allowing ligand-independent (i.e., cell autonomous) signaling (43). Like Notch-1, E-cadherin functions are Ca2+ dependent. At 50 to 100 μmol/L Ca2+, E-cadherin is a monomer with a rod-like structure, which does not allow lateral (cis) dimerization at the plasma membrane (44). Cis-dimerization occurs at >500 μmol/L Ca2+ and is required for the trans-interaction between cis-dimerized E-cadherin molecules of adjacent cells responsible for homotypic interaction (44). Use of low (i.e., <300 μmol/L)–Ca2+ medium inhibits such E-cadherin-mediated homotypic interaction in culture. This is significant because such homotypic interaction is the mechanism for contact inhibition of proliferation via its ability to up-regulate expression of p27/kip1 (45).

As documented in the present study, normal human prostatic transit-amplifying cells respond dramatically to different levels of Ca2+. In low (i.e., <260 μmol/L)–Ca2+-SFD medium, Notch-1 signals cell autonomously; thus, the cells survive at low density and E-cadherin is not able to homotypically interact between transit-amplifying cells, and p27/kip1 expression is low. Therefore, these transit-amplifying cells do not aggregate and are not contact inhibited at low cell density in low-Ca2+ medium. This Ca2+-dependent behavior is consistent with why these normal transit-amplifying cells have a low rate of proliferation in vivo (46), although they are chronically supplied by sufficient levels of andromedins produced by the androgen-stimulated supporting stromal cells. Presumably, under in vivo conditions where the Ca2+ level is 1 to 2 mmol/L and cell density is high, E-cadherin-mediated homotypic interaction induces p27/kip1; thus, proliferation of these transit-amplifying cells is low due to contact inhibition. In contrast, when prostate tissue is dissociated and cultured in such low-Ca2+-SFD medium, the transit-amplifying cells (a) lose E-cadherin-mediated homotypic interaction and thus lack high p27/kip1 expression, (b) undergo Notch-1 cell autonomous survival signaling, and (c) proliferate due to the added growth factor in the medium even at initially low cell density. In contrast, this low-Ca2+-SFD medium lacks sufficient levels of attachment/growth factors to allow malignant prostate cancer cells to adhere and grow. This would explain why using such low-Ca2+-SFD medium to establish cultures from either normal prostate or tissue-containing prostate cancer results in outgrowth of normal transit-amplifying cells (e.g., 957E/hTERT) and not prostate cancer cells.

As documented by the present studies, the overgrowth of these normal transit-amplifying cells is prevented by raising the medium Ca2+ to >600 μmol/L. At this Ca2+ level, Notch-1 is not cell autonomous but dependent on cell-cell (i.e., ligand receptor) interaction. Thus, transit-amplifying cells cannot survive at low cell density if the Ca2+ is raised to >600 μmol/L and if the cells do not interact to activate Notch-1 signaling. Such interaction while activating Notch-1, however, allows E-cadherin to form trans-dimers (i.e., homotypic interaction) on the plasma membrane between transit-amplifying cells. This results in the up-regulation of p27/kip1 and thus contact inhibition of growth of these normal prostatic cells. In contrast, prostate cancer cells are not growth inhibited at >600 μmol/L Ca2+. This is because the consequence of Notch-1 signaling and E-cadherin homotypic interaction is characteristically different in malignant versus normal prostatic epithelial cells. During prostatic carcinogenesis, Notch-1 signaling is no longer required for survival but instead is one of multiple signaling pathways stimulating proliferation of prostate cancer cells. Previous studies have documented that E-cadherin expression is decreased and its location is dysregulated in the progression of prostate cancer (4749). In addition, there are changes in the E-cadherin signaling cascade (i.e., α-catenin and β-catenin) such that E-cadherin-mediated homotypic interaction does not induce contact inhibition of prostatic cancer cells (49, 50).

In summary, normal transit-amplifying prostatic epithelial cells activate Notch-1 signaling required for their survival by Ca2+-dependent, E-cadherin-mediated homotypic interaction. Such homotypic interaction, however, restricts proliferation of transit-amplifying cells via up-regulation of p27/kip1 producing contact inhibition. In contrast, Ca2+-dependent, E-cadherin-mediated homotypic interaction does not induce contact inhibition in prostate cancer cells and Notch-1 signaling is not required for survival of these malignant cells. Instead, E-cadherin enhances cancer cell attachment/motility without requiring Notch-1 signaling. These characteristic changes may help explain why prostate cancer cells can dissociate from the primary site, disseminate as single cells, and survive and establish metastases, particularly to the bone, a site of high-Ca2+ levels.

Grant support: NIH grant RO1DK52645 and Specialized Programs of Research Excellence in Prostate Cancer grant P50CA58236.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dorothea K. Stieff for the preparation of the article, Leslie Meszler (Cell Imaging Core, Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins), Anita L. Hawkins and Dr. Constance A. Griffin (Cytogenetics Core, Department of Pathology, Johns Hopkins School of Medicine) for the karyotyping data, Jessica L. Hicks for immunocytochemical staining, and Dr. William B. Isaacs for helpful criticisms and suggestions.

1
Litvinov IV, De Marzo AM, Isaacs JT. Is the Achilles' heel for prostate cancer therapy a gain of function in androgen receptor signaling?
J Clin Endocrinol Metab
2003
;
88
:
2972
–82.
2
Isaacs JT. Control of cell proliferation and cell death in the normal and neoplastic prostate: a stem cell model. In Rogers CH, Coffey DS, Cunha G, Grayhack JT, Hinman F, Jr., Horton R, editors. Benign prostatic hyperplasia. Report no. NIH 87-2881. Bethesda (MD): Department of Health and Human Services, NIH; 1987. p. 85–94.
3
Richardson GD, Robson CN, Lang SH, Neal DE, Maitland NJ, Collins AT. CD133, a novel marker for human prostatic epithelial stem cells.
J Cell Sci
2004
;
117
:
3539
–45.
4
Bonkhoff H, Fixemer T, Remberger K. Relation between Bcl-2, cell proliferation, and the androgen receptor status in prostate tissue and precursors of prostate cancer.
Prostate
1998
;
34
:
251
–8.
5
De Marzo AM, Meeker AK, Epstein JI, Coffey DS. Prostate stem cell compartments: expression of the cell cycle inhibitor p27Kip1 in normal, hyperplastic, and neoplastic cells.
Am J Pathol
1998
;
153
:
911
–9.
6
van Leenders G, Dijkman H, Hulsbergen-van de Kaa C, Ruiter D, Schalken J. Demonstration of intermediate cells during human prostate epithelial differentiation in situ and in vitro using triple-staining confocal scanning microscopy.
Lab Invest
2000
;
80
:
1251
–8.
7
Tran CP, Lin C, Yamashiro J, Reiter RE. Prostate stem cell antigen is a marker of late intermediate prostate epithelial cells.
Mol Cancer Res
2002
;
1
:
113
–21.
8
van Leenders GJ, Gage WR, Hick JL, et al. Intermediate cells in human prostate epithelium are enriched in proliferative inflammatory atrophy.
Am J Pathol
2003
;
162
:
1529
–37.
9
Garraway LA, Lin D, Signoretti S, et al. Intermediate basal cells of the prostate: in vitro and in vivo characterization.
Prostate
2003
;
55
:
206
–18.
10
Uzgare AR, Xu Y, Isaacs JT. In vitro culturing and characteristics of transit amplifying epithelial cells from human prostate tissue.
J Cell Biochem
2004
;
91
:
196
–205.
11
Gao J, Arnold JT, Isaacs JT. Conversion from a paracrine to an autocrine mechanism of androgen-stimulated growth during malignant transformation of prostatic epithelial cells.
Cancer Res
2001
;
61
:
5038
–44.
12
Kurita T, Wang YZ, Donjacour AA, et al. Paracrine regulation of apoptosis by steroid hormones in the male and female reproductive system.
Cell Death Differ
2001
;
8
:
192
–200.
13
Signoretti S, Waltregny D, Dilks J, et al. p63 is a prostate basal cell marker and is required for prostate development.
Am J Pathol
2000
;
157
:
1769
–75.
14
Liu AY, Peehl DM. Characterization of cultured human prostatic epithelial cells by cluster designation antigen expression.
Cell Tissue Res
2001
;
305
:
389
–97.
15
Friedrichs B, Riedmiller H, Goebel HW, Rausch U, Aumuller G. Immunological characterization and activity of transglutaminases in human normal and malignant prostate and in prostate cancer cell lines.
Urol Res
1995
;
23
:
301
–10.
16
Luo J, Zha S, Gage WR, et al. α-Methylacyl-CoA racemase: a new molecular marker for prostate cancer.
Cancer Res
2002
;
62
:
2220
–6.
17
Wang XD, Shou J, Wong P, French DM, Gao WQ. Notch1-expressing cells are indispensable for prostatic branching morphogenesis during development and re-growth following castration and androgen replacement.
J Biol Chem
2004
;
279
:
24733
–44.
18
van Bokhoven A, Varella-Garcia M, Korch C, et al. Molecular characterization of human prostate carcinoma cell lines.
Prostate
2003
;
57
:
205
–25.
19
Uzgare AR, Isaacs JT. Enhanced redundancy in Akt and mitogen-activated protein kinase-induced survival of malignant versus normal prostate epithelial cells.
Cancer Res
2004
;
64
:
6190
–9.
20
Kaighn ME, Kirk D, Szalay M, Lechner JF. Growth control of prostatic carcinoma cells in serum-free media: interrelationship of hormone response, cell density, and nutrient media.
Proc Natl Acad Sci U S A
1981
;
78
:
5673
–6.
21
Peehl DM, Stamey TA. Serial propagation of adult human prostatic epithelial cells with cholera toxin.
In Vitro
1984
;
20
:
981
–6.
22
Peehl DM, Stamey TA. Growth responses of normal, benign hyperplastic, and malignant human prostatic epithelial cells in vitro to cholera toxin, pituitary extract, and hydrocortisone.
Prostate
1986
;
8
:
51
–61.
23
Peehl DM, Stamey TA. Serum-free growth of adult human prostatic epithelial cells.
In Vitro Cell Dev Biol
1986
;
22
:
82
–90.
24
Peehl DM, Wong ST, Bazinet M, Stamey TA. In vitro studies of human prostatic epithelial cells: attempts to identify distinguishing features of malignant cells.
Growth Factors
1989
;
1
:
237
–50.
25
Chopra DP, Grignon DJ, Joiakim A, et al. Differential growth factor responses of epithelial cell cultures derived from normal human prostate, benign prostatic hyperplasia, and primary prostate carcinoma.
J Cell Physiol
1996
;
169
:
269
–80.
26
Yasunaga Y, Nakamura K, Ewing CM, Isaacs WB, Hukku B, Rhim JS. A novel human cell culture model for the study of familial prostate cancer.
Cancer Res
2001
;
61
:
5969
–73.
27
Yasunaga Y, Nakamura K, Ko D, et al. A novel human cancer culture model for the study of prostate cancer.
Oncogene
2001
;
20
:
8036
–41.
28
Brothman AR, Peehl DM, Patel AM, et al. Cytogenetic evaluation of 20 cultured primary prostatic tumors.
Cancer Genet Cytogenet
1991
;
55
:
79
–84.
29
Konig JJ, Teubel W, van Dongen JW, Hagemeijer A, Romijn JC, Schroder FH. Tissue culture loss of aneuploid cells from carcinomas of the prostate.
Genes Chromosomes Cancer
1993
;
8
:
22
–7.
30
Bright RK, Vocke CD, Emmert-Buck MR, et al. Generation and genetic characterization of immortal human prostate epithelial cell lines derived from primary cancer specimens.
Cancer Res
1997
;
57
:
995
–1002.
31
Gindler EM, King JD. Rapid colorimetric determination of calcium in biologic fluids with methylthymol blue.
Am J Clin Pathol
1972
;
58
:
376
–82.
32
Denmeade SR, Sokoll LJ, Dalrymple S, et al. Dissociation between androgen responsiveness for malignant growth vs. expression of prostate specific differentiation markers PSA, hK2, and PSMA in human prostate cancer models.
Prostate
2003
;
54
:
249
–57.
33
Chin RK, Hawkins AL, Isaacs WB, Griffin CA. E1A transformed normal human prostate epithelial cells contain a 16q deletion.
Cancer Genet Cytogenet
1998
;
103
:
155
–63.
34
Shearman MS, Beher D, Clarke EE, et al. L-685,458, an aspartyl protease transition state mimic, is a potent inhibitor of amyloid β-protein precursor γ-secretase activity.
Biochemistry
2000
;
39
:
8698
–704.
35
Chrucher I, Williams S, Kerrad S, et al. Design and synthesis of highly potent benzodiazepine γ-secretase inhibitors: preparation of (2S,3R)-3-(3,4-difluorophenyl)-2-(4-fluorophenyl)-4-hydroxy-N-((3S)-1-methyl-2-oxo-5-phenyl-2,3-dihydro-1H-benzo[e][1,4]-diazepin-3-yl)butyramide by use of an asymmetric Ireland-Claisen rearrangement.
J Med Chem
2003
;
46
:
2275
–8.
36
Oda D, Bigler L, Mao EJ, Disteche CM. Chromosomal abnormalities in HPV-16-immortalized oral epithelial cells.
Carcinogenesis
1996
;
17
:
2003
–8.
37
Shou J, Ross S, Koeppen H, de Sauvage FJ, Gao W-Q. Dynamics of Notch expression during murine prostate development and tumorigenesis.
Cancer Res
2001
;
61
:
7291
–7.
38
Iso T, Kedes L, Hamamori Y. HES and HERP families: multiple effectors of the Notch signaling pathway.
J Cell Physiol
2003
;
194
:
237
–55.
39
Struhl G, Adachi A. Nuclear access and action of Notch in vivo.
Cell
1998
;
15
:
649
–60.
40
Mumm JS, Schroeter EH, Saxena MT, et al. A ligand-induced extracellular cleavage regulates γ-secretase-like proteolytic activation of Notch-1.
Mol Cell
2000
;
5
:
197
–205.
41
LaVoie MJ, Selkoe DJ. The notch ligands, jagged and δ, are sequentially processed by α-secretase and presenilin/γ-secretase and release signaling fragments.
J Biol Chem
2003
;
278
:
34427
–37.
42
Ross DA, Kadesch T. Consequences of Notch-mediated induction of Jagged-1.
Exp Cell Res
2004
;
296
:
173
–82.
43
Rand MD, Grimm LM, Artavanis-Tsakonas S, et al. Calcium depletion dissociates and activates heterodimeric notch receptors.
Mol Cell Biol
2000
;
20
:
1825
–35.
44
Partz O, Bozic D, Koch AW, Fauser C, Brancaccio A, Engel J. A new crystal structure, Ca2+ dependence and mutational analysis reveal molecular details of E-cadherin homoassocation.
EMBO J
1999
;
18
:
1738
–47.
45
Croix BS, Sheehan C, Rak JW, Florenes VA, Slingerland JM, Kerbel RS. E-cadherin-dependent growth suppression is mediated by the cyclin-dependent kinase inhibitor p27KIP1.
J Cell Biol
1998
;
142
:
557
–71.
46
Berges RR, Vukanovic J, Epstein JI, et al. Implication of cell kinetic changes during the progression of human prostatic cancer.
Clin Cancer Res
1995
;
1
:
473
–80.
47
Bussemakers MJ, van Moorselaar RJ, Giroldi LA, et al. Decreased expression of E-cadherin in the progression of rat prostatic cancer.
Cancer Res
1992
;
52
:
2916
–22.
48
Umbas R, Schalken JA, Aalders TW, et al. Expression of the cellular adhesion molecule E-cadherin is reduced or absent in high-grade prostate cancer.
Cancer Res
1992
;
52
:
5104
–9.
49
Morton RA, Ewing CM, Nagafuchi A, Tsukita S, Isaacs WB. Reduction of E-cadherin levels and deletion of the α-catenin gene in human prostate cancer cells.
Cancer Res
1993
;
53
:
3585
–90.
50
Rashid MG, Sanda MG, Vallorosi CJ, Rios-Doria J, Rubin MA, Day ML. Posttranslational truncation and inactivation of human E-cadherin distinguishes prostate cancer from matched normal prostate.
Cancer Res
2001
;
61
:
489
–92.