Stable transfection of protein kinase C alpha cDNA in hormone-dependent breast cancer cell lines

An inverse relationship between protein kinase C (PKC) activity and oestrogen receptor (ER) expression in human breast cell lines and tumours has been firmly established over the past 10 years. To determine whether specific alterations in PKC expression accompany hormone-independence, we examined the expression of PKC isozymes in the hormone-independent human breast cancer cell clones MCF-7 5C and T47D:C42 compared with their hormone-dependent counterparts, MCF-7 A4, MCF-7 WS8 and T47D:A18 respectively. Both hormone-independent cell clones exhibit elevated PKCα expression and increased basal AP-1 activity compared with the hormone-dependent cell clones. To determine whether PKCα overexpression is sufficient to mediate the hormone-independent phenotype, we stably transfected an expression plasmid containing PKCα cDNA to the T47D:A18 and MCF-7 A4 cell lines. This is the first report of PKCα transfection in T47D cells. In contrast to MCF-7 cells, T47D has the propensity to lose the ER and more readily forms tamoxifen-stimulated tumours in athymic mice. We find that in T47D:A18/PKCα clones, there is concomitant up-regulation of PKC βI and δ, whereas in the MCF-7 A4/PKCα transfectants PKC ɛ is up-regulated. In T47D:A18, but not in MCF-7 A4, PKCα stable transfection is accompanied by down-regulation of ER function whilst basal AP-1 activity is elevated. Our results suggest PKCα overexpression may play a role in growth signalling during the shift from hormone dependent to hormone-independent breast cancers. © 2000 Cancer Research Campaign

represent the most aggressive forms of breast cancer and an effective treatment modality is not yet available. We and others have explored the basis of tamoxifen resistance by examining key steps along the ER-mediated signal transduction pathway (Encarnacion et al, 1993;Johnston et al, 1993;Osborne, 1993;Wolf et al, 1993;Karnik et al, 1994;Wolf and Jordan, 1994;Tonetti and Jordan, 1995). However, to date, examination of this pathway has yielded relatively few clues concerning the molecular mechanism of tamoxifen resistance. Identification of a suitable molecular target in these tumour types would facilitate the development of a logical therapeutic approach.
It is well documented that ER and protein kinase C (PKC) activity and abundance is inversely related in breast cancer cell lines  and PKC is elevated in malignant versus normal breast tissue O'Brian et al, 1989;Gordge et al, 1996). Furthermore increased AP-1 (activator protein-1) activity has been reported to occur in hormone-independent breast cancer cell lines and tumours (Dumont et al, 1996;Johnston et al, 1999). PKC is a family of serine-threonine protein kinases which is now comprised of at least 11 isozymes, α, βI, βII, γ, δ, ε, ζ, η, θ, ι/λ and µ (Dekker and Parker, 1994), and is a known upstream activator of the AP-1 signal transduction pathway. These isozymes have been categorized into three groups based mainly on their co-factor requirements and sequence similarity, substrate specificity, and marked tissue and intracellular distribution (Nishizuka, 1992;Kazanietz et al, 1993). Activation of PKCs by the tumour promoter 12-O-tetradecanoylphorbol-13-acetate (TPA) results in the expression of the c-fos and c-jun family of nuclear protooncogenes, and increased transcriptional activity of genes containing a TPA-responsive element (TRE) (Angel and Karin, 1991). The AP-1 complex, comprised of either a Fos-Jun heterodimer or a Jun-Jun homodimer, binds to the TRE and induces gene transcription (Angel et al, 1987b;Lee et al, 1987). PKC isozymes differ in their ability to mediate transcriptional activation through the TRE (Hata et al, 1993;Reifel-Miller et al, 1996). TPA, which specifically binds to and activates PKC, causes down-regulation of the ER by post-transcriptional destabilization of the mRNA (Saceda et al, 1991) and inhibition of ER function (Martin et al, 1995). Furthermore, since ER and PKC expression is inversely related , it is possible that PKC could play a fundamental role in the slip from hormone-dependent to hormone-independent breast cancer.
Therefore, there is sufficient evidence in the literature indicating an involvement of ER and PKC in AP-1 activation. In this study we have examined PKC isozyme expression and AP-1 activation in two hormone-independent human breast cancer cell lines, MCF-7 5C (Jiang et al, 1992) and T47D:C42 (Pink et al, 1996) compared with their respective hormone-dependent counterparts, MCF-7 A4(WS8) and T47D:A18. We find the hormone-independent cell lines overexpress PKCα and have elevated basal AP-1 activity. To determine whether PKCα overexpression is sufficient to create both the hormone-independent phenotype and induce elevated basal AP-1 activity, we stably transfected PKCα into the hormone-dependent T47D:A18 and MCF-7 A4 cell lines. Transfection of PKCα into MCF-7 cells was previously reported by two independent investigators (Ways et al, 1995;Manni et al, 1996). Ways et al report that overexpression of PKCα caused a reduction in ER expression, decreased oestrogen-dependent gene expression and a more aggressive neoplastic phenotype. In contrast, Manni et al. reported a less aggressive phenotype in MCF-7 cells as a result of PKCα overexpression. Since we have both hormone-dependent and -independent pairs of each cell type, we set out to determine the effect of PKCα overexpression in T47D cells compared to MCF-7. This question is especially relevant since T47D cells lose the ER more readily (Pink et al, 1996) and form tamoxifen-resistant tumours in athymic mice at an accelerated rate compared with MCF-7 cells (MacGregor-Schaffer et al, 2000). In consideration of the well-documented inverse relationship between ER and PKC, T47D cells provide a more appropriate cellular context in which to address the role of PKCα overexpression on the hormone-dependent phenotype. We find that in T47D:A18 cells overexpressing PKCα, there is concomitant upregulation of PKC βI and δ, whereas in MCF-7 A4/PKCα transfectants PKCε is up-regulated. In T47D:A18, but not in MCF-7 A4, PKCα stable transfection is accompanied by down-regulation of ER function whilst basal AP-1 activity is elevated. PKCα stable transfection is not sufficient, however, to allow the antioestrogen tamoxifen to switch on the AP-1 pathway in either T47D:A18 or MCF-7:A4 cells. These results suggest that in the breast cancer cell lines T47D:A18 and MCF-7 A4, PKCα, and consequently βI, δ and ε overexpression, is not sufficient to trigger AP-1 activation, and is not likely to singly lead to tamoxifen-stimulated tumour growth. However, PKCα overexpression does lead to decreased ER function, and may play a role in de novo hormone-independent breast cancers.

Cell lines and culture conditions
MCF-7 WS8 and T47D:A18 are hormone-dependent human breast cancer cell clones that have been previously described (Jiang et al, 1992;Pink et al, 1996). MCF-7 A4 is a hormonedependent single cell clone of MCF-7 cells obtained from the American Type Culture Collection (ATCC, Rockville, MD). MCF-7 WS8, A4 and T47D:A18 were maintained in phenol redcontaining RPMI-1640 medium supplemented with 10% FBS as previously described (Jiang et al, 1992;Pink et al, 1996). The hormone-independent MCF-7 5C cell clone was maintained in phenol red-free MEM supplemented with 5% 3 × dextran-coated charcoal (DCC)-treated calf serum (oestrogen-free media) as previously described (Jiang et al, 1992). The hormone-independent T47D:C42 clone was passaged in phenol red-free RPMI-1640 supplemented with 10% 3 × DCC-treated FBS (oestrogen-free media) as previously described (Pink et al, 1996). Prior to transient transfection experiments and RNA isolation, all cell lines were placed in oestrogen-free media for 3 days.

Poly A + RNA isolation and Northern blot
Total RNA was isolated from cell lines using the Trizol reagent (GibcoBRL, Rockville, MD) prior to isolation of poly A + RNA using the PolyATract mRNA isolation kit (Promega, Madison, WI). Equal amounts of polyA + RNA were loaded per lane (4-5 µg), electrophoresed in a 1% agarose-0.66M formaldehyde gel, transferred to a Magnagraph nylon membrane (MSI, Westborough, MA) and the RNA was fixed to the membrane by UV cross-linking. The PKC α, δ, ε and η cDNA clones were obtained from ATCC (Rockville, MD) and the following cDNA fragments were labelled with [α-32 P]dCTP by random priming: a 1.3 Kb Eco RI PKC α fragment from phPKC-α7; a 1.7 Kb Eco RI PKC δ fragment from the I.M.A.G.E. Consortium Clone ID 236079; a 2.2 Kb Nhe I PKC ε fragment from pBluebac/PKC ε; a 1.7 Pvu II PKC fragment from pAcMP3/PKC eta. Equivalent loading of mRNA in each lane was verified by stripping and reprobing with β-actin cDNA.

Growth assays
Cells were grown in oestrogen-free media for 2 days prior to each experiment and were seeded into 24-well plates (15 000 cells/well) (Day 0). The following day (Day 1) medium containing either ethanol (control), E2 (10 -13 -10 -7 M) or 4-OHT (10 -12 -10 -6 M) was added. All compounds were dissolved in 100% ethanol and added to the medium at a 1:1000 dilution. Medium was changed on Days 3 and 5. DNA content was determined on Day 6 using a fluorocolorimeter (Labarca and Paigen, 1980).

ER protein expression
Nuclear extracts were prepared from MCF-7A4 and T47D:A18 cell clones using a freeze-thaw method (Chen et al, 1996). ER expression was assessed by Western blot analysis using the ER mouse monoclonal antibody AER311 (human cross-reactive) (Neomarkers, Fremont, CA) followed by band detection with the ECL Western blotting kit (Amersham).

Statistical analysis
When comparing to 1 as in Figures 4 and 9A, data were analysed using a one sample, one-tailed t-test. When comparing groups, data were analysed using one-way analysis of variance. Significant differences were indicated when P < 0.05.

PKC expression in hormone-dependent and hormoneindependent breast cancer cell lines
Elevated PKC expression (upstream activator of AP-1) is reported to correlate with hormone-independent breast cancer O'Brian et al, 1989;Davidson and Kennedy, 1996;Gordge et al, 1996). We have examined two pairs of hormone-dependent and hormone-independent human breast cancer cell lines to determine whether a particular PKC isozyme(s) is associated with hormone independence. MCF-7 5C and T47D:C42 cell clones were derived from their respective parental cell lines MCF-7 and T47D by long-term culture in oestrogen-free media (Jiang et al, 1992;Pink et al, 1996) and are both E2 and antioestrogen (4-OHT and ICI 164,384) insensitive. The T47D:C42 cell clone has lost the expression of ERα, whereas the MCF7 5C cell clone exhibits reduced ERα expression ( Figure 1) and impaired ER function (Jiang et al, 1992). We have compared PKC isozyme expression in these two hormone-independent breast cancer cell lines with their respective hormone-dependent counterpart cell clones, MCF-7 Figure 2 PKC isozyme mRNA expression in hormone-dependent and hormone-independent human breast cancer cell lines. PolyA + RNA was isolated (as described in 'Materials and Methods') following 24 hours treatment: C, control (ethanol vehicle); E, E2 10 -9 M; O, 4-OHT 10 -7 M. A Northern blot was done by hybridization with PKCα, ε and β-actin cDNAs.
(A) MCF-7 clones WS8 (hormone-dependent) and 5C (hormoneindependent). (B) T47D clones A18 (hormone-dependent) and C42 (hormone-independent). Approximate sizes of specific messages were as follows: PKCα -9.5, 3.7 and 3.2 kb; PKCε -7.4 and 6.8. These results were reproduced at least three times and the Northern blot shown is a representative experiment A4(WS8) and T47D:A18, by both Northern and Western blot analysis. PKC α, δ, ε, and η mRNAs were detected in all of the cell lines by Northern blot. The hormone-independent MCF-7 5C exhibits elevated expression of the PKCα transcripts compared with the hormone-dependent WS8 clone (Figure 2A). The MCF-7 WS8 and 5C clones express similar levels of PKC ε (Figure 2A) δ and η mRNAs (results not shown). Treatment of these cell clones with either E2 or 4-OHT did not alter the expression level of the PKCα isozyme transcripts examined, however there is a slight up-regulation of PKCε in response to E2 treatment (Figure 2A). Similarly, whereas the T47D:A18 (ER-positive, hormone-dependent) and C42 (ER-negative, hormone-independent) clones express comparable levels of PKC δ and η mRNAs in the untreated, E2 or 4-OHT-treated conditions (results not shown), PKC α and ε mRNAs are more abundant in the hormone-independent T47D C42 cell line ( Figure 2B). In particular, expression of PKCα mRNA is undetectable in the T47D:A18 cells in contrast to the abundant expression observed in the T47D:C42 cell clone, whereas only a smaller PKC ε mRNA is more abundant in the C42 cells compared with A18. Therefore, whereas the hormone-dependent cell clones (MCF-7 WS8 and T47D A18) express very low or undetectable levels of PKC α and ε mRNAs, the hormone-independent MCF-7 5C and T47D C42 express comparatively more abundant PKC α and ε transcript expression.
To determine whether the abundance of PKC α and ε mRNA expression is reflected in the amount of protein produced, we examined PKC isozyme expression by Western blot analysis. The relative PKCα protein levels expressed in MCF-7 5C compared with MCF-7 A4 correspond to the PKCα transcript abundance observed; PKCα protein is expressed at higher levels in the hormone-independent MCF-7 5C cells ( Figure 3A). PKC ε protein is more abundant in MCF-7 5C compared with the A4, whereas the PKC ε basal mRNA levels are comparable (Figure 2A). There is no difference in PKC βI expression.
Accordingly, the overexpression of PKCα mRNAs directly correlates with the PKCα protein levels expressed in T47D:A18 and C42; PKCα protein is more abundant in the hormone-independent C42 cells relative to the A18 cells ( Figure 3B). There is no difference in PKC ε or η expression and PKC δ is barely detectable (results not shown) in the T47D clones.

Induction of AP-1 transcriptional activation
Elevated AP-1 activity is associated with tamoxifen-resistant breast cancer (Astruc et al, 1995;Dumont et al, 1996;Smith et al, 1999). Since PKC is a known upstream activator of the AP-1 pathway, we compared the amount of basal and treatment-induced AP-1 activity in the hormone-dependent and hormone-independent breast cancer cell lines. Following transient co-transfection of the TRE-tk-luc and pCMVβ-gal plasmids, the MCF-7 A4, 5C and T47D:A18, C42 cell lines were either untreated (control, ethanol vehicle) or treated with PMA (10 -7 M), E2 (10 -9 M) or 4-OHT (10 -7 M) for 5 and 24 hours prior to determining luciferase and βgal activities. Basal AP-1 activity was elevated in both of the hormone-independent cell clones MCF-7 5C and T47D:C42 compared with their hormone-dependent counterparts MCF-7 A4 and T47D:A18 ( Figure 4A). The basal AP-1 activity of the MCF-7 5C cell clone was 7-fold higher than the MCF-7 A4 clone, although this did not achieve statistical significance at the P < 0.05 level. However, the basal AP-1 activity of T47D:C42 was elevated 2.6-fold compared with T47D:A18 (P < 0.05). Treatment of the cells with 10 -7 M PMA (activator of PKC) for 5 hours caused an induction of AP-1 activity in all cells lines, verifying the functionality of the reporter construct ( Figures 4B and 4C). PMA induced AP-1 activity to a greater extent in MCF-7 A4 cells compared with MCF-7 5C cells (16-fold vs 5-fold) ( Figure 4B), most likely due to the fact that MCF-7 5C cells already exhibit high basal AP-1 activity. PMA-induced AP-1 activation was comparable (2-fold) in both T47D:A18 and C42 cell lines. Neither E2 nor 4-OHT treatment of any of the cell lines resulted in transcriptional activation of AP-1 at 5 h ( Figures 4B and 4C) or 24 h (results not shown). We can conclude that the hormone-independent breast cancer cell lines (MCF-7 5C and T47D C42) have elevated basal AP-1 activity compared with their hormone-dependent counterparts (MCF-7 A4 and T47D A18). However this activity is not regulated by E2 or 4-OHT treatment.

Stable transfection of PKCα in hormone-dependent T47D:A18 and MCF-7 A4 breast cancer cells
Since we have determined that both hormone-independent cell lines overexpress PKCα, we set out to determine the role of this isozyme in hormone-independent breast cancer. Therefore, we stably transfected the PKCα expression plasmid, pSPKCα, into MCF-7 A4 and T47D:A18 cells. After screening several clones for PKCα expression by Western blotting, we chose to further characterise two clones from each cell line, MCF-7 A4/PKCα12 and α29 and T47D:A18/PKCα5 and α20.
To determine whether overexpression of PKCα would affect the expression of other PKC isozymes, we performed Western blots using antibodies recognizing PKC βI, δ, and ε ( Figure 5). We find that in MCF-7, overexpression of PKCα causes concomitant upregulation of PKC ε. This pattern of PKC isozyme expression is similar to the MCF-7 5C cell line ( Figure 3A). In the T47D:A18 cell line however, PKC βI and δ are up-regulated ( Figure 5), whereas this PKC expression pattern is not evident in the T47D:C42 cell line. Cross-regulation between PKCα and β was previously reported in MCF-7 cells (Ways et al, 1995;Manni et al, 1996), however cross-regulation between PKCα and PKCs δ and ε has not been reported. To examine whether the previously characterized inverse relationship of PKC and ER expression exists in our PKCα stable transfectants, we examined cell growth in response to E2, E2induced transcriptional activation of an oestrogen response element (ERE), and ERα protein expression. Growth assays were performed measuring DNA content as the endpoint. We find that the T47D:A18/PKCα transfectants have a reduced oestrogen-induced growth response compared with the T47D:A18/neo control. Whereas the T47D:A18/neo cells demonstrated a 5-fold increase cell growth over the control, the PKCα stable clones exhibited only a 2-fold induction of growth following E2 treatment ( Figure 6A). 4-OHT did not differentially affect the growth of T47D:A18/neo versus PKCα transfectants and was capable of competitively inhibiting the E2-induced growth response (results not shown). No difference in E2-( Figure 6B) nor 4-OHT-(results not shown) induced cell growth was apparent comparing the MCF-7 A4/neo, MCF-7A4/PKCα12 or α29 cell clones.
Transcriptional activation of an ERE was measured by transient transfection of an ERE-luciferase reporter plasmid into each cell line and luciferase activity was measured following 24 hour Western blot analysis of PKC isozyme expression in T47D:A18 and MCF-7 A4 PKCα stable cell clones. PKC protein was partially purified from MCF-7 A4 and T47D:A18 neo and PKCα stable clones by DEAEcellulose chromatography as described in Materials and Methods. Western blot analysis was performed using specific antibodies to PKCα, βI, δ, and ε isozymes recognizing 82, 80, 78 and 90 kDa proteins respectively. Equivalent protein loading was assessed by stripping the membrane and re-probing with a β-actin monoclonal antibody treatment with E2. The ability of the T4D:A18/PKCα clones to activate transcription of an ERE-luciferase reporter plasmid was abrogated by an order of magnitude compared with the T47D:A18/neo clone ( Figure 7A). Luciferase activity was induced 200-fold in the neo transfectants compared with only a 20-fold induction in the PKCα transfectants. The MCF-7 A4 clones however, displayed no difference in the ability to activate the ERE-tk-luc plasmid compared with the neo transfected control ( Figure 7B). These results indicate an inverse relationship of PKCα and ER function in T47D:A18 cells as assessed by E2induced growth response and activation of an ERE. The MCF-7 A4/PKCα clones, however, do not exhibit this inverse relationship  Figure 7 ERE transcriptional activation in response to E2 in T47D:A18/PKCα and MCF-7/PKCα stable clones. ERE-TK-luc and pCMVβ plasmids were transiently co-transfected by electroporation into T47D:A18 (A) and MCF-7 (B) stable transfectants. Twenty-four hours following electroporation, the cells were treated with either ethanol vehicle (Control), E2 (10 -9 M) or ICI 182,780 (10 -7 M). Luciferase and β-gal activity was measured 20 hours following treatment. Data is expressed as the normalized E2-induced fold-increase over ICI 182,780 treatment (mean ± SE) of 3 independent experiments performed in triplicate. Data were analysed by one-way ANOVA. *Indicates significant difference among groups at P = 0.0263 as was previously reported (Ways et al, 1995). Our results in MCF-7 cells may be due to either an insufficient level of PKCα to produce this effect, or may be attributed to differences in the particular MCF-7 cell clones used in the two laboratories. Alternatively, it may be the fact that PKCα overexpression in T47D:A18 leads to increased PKC βI and δ levels, whereas the MCF-7 A4/PKCα transfectants overexpress PKC ε.
To determine whether the impaired ER function in the T47D:A18/PKCα clones as exhibited by the reduced E2-induced growth response and transcriptional activation of an ERE was due to loss of ER expression, ERα protein levels were assessed by Western blotting. The T47D:A18 neo and PKCα5 clones appear to express comparable ERα, whereas the PKCα20 shows a slight decrease in ERα protein (Figure 8). This result suggests that although ERα protein is expressed, functionality is impaired by PKCα, and perhaps βI and δ over-expression. No difference in ERα levels were observed in the MCF-7 A4 transfectants (results not shown).

Basal AP-1 activity in T47D:A18/PKCα stable transfectants
Since basal AP-1 activity is elevated in the hormone-independent T47D:C42 clone compared to T47D:A18, and T47D:C42 overexpresses PKCα, we wanted to determine whether the basal AP-1 activity was also elevated in the PKCα stable transfectants. We transiently transfected a TRE-luciferase reporter construct that contained four tandem repeats of the collagenase TRE sequence (Angel et al, 1987a). The PKCα clones exhibited 6-7-fold higher basal AP-1 activity compared with the neo transfectant ( Figure  9A). These results suggest that overexpression of PKCα, and perhaps β and δ, is sufficient to raise the basal AP-1 activity to levels comparable to that of the hormone-independent T47D:C42 cell line. It is interesting to note that T47D:C42 cells do not exhibit PKC βI nor δ overexpression (results not shown). The MCF-7 A4 transfectants showed no increase in AP-1 basal activity (results not shown). TRE-tk-luc and pCMVβ plasmids were transiently co-transfected by electroporation into T47D:A18/neo and PKCα stable transfectants. Luciferase and β-gal activity was measured 40 hours following transfection. (A) Basal AP-1 transcriptional activity shown as the normalized luciferase activity (luciferase/β-gal) of the PKCα clones expressed relative to the T47D:A18/neo clone (set = 1). *Indicates statistical significance at P < 0.05 using one-sample t-test. (B) AP-1 transcriptional activation in response to PMA and 4-OHT in T47D:A18/PKCα stable clones. Twenty-four hours following electroporation the cells were treated with either ethanol vehicle (control); PMA (10 -7 M) or 4-OHT (10 -7 M). Luciferase and β-gal activity was measured 20 hours following treatment. Data is expressed as the normalized luciferase activity (mean ± SE) of 3 independent experiments performed in triplicate. Data were analysed by one-way ANOVA. *Indicates statistical difference among groups treated with PMA at P = 0.044

4-OHT is incapable of activating AP-1 in T47D:A18/PKCα transfectants
We examined the ability of PMA, which is a potent PKC and AP-1 activator, to induce transcription of the TRE-tk-Luc reporter in the T47D:A18/PKCα transfectants. As we expected, the T47D:A18/PKCα transfectants exhibited a 6-8-fold higher AP-1 activation in response to PMA compared with the neo transfected T47D:A18 cells ( Figure 9B). It was previously reported that tamoxifen is capable of activating AP-1 following long-term antioestrogen treatment of breast cancer cells in a PKC-dependent manner (Astruc et al, 1995). We hypothesized that PKC up-regulation may be a prerequisite step to tamoxifen-stimulated growth. Therefore we examined whether PKCα overexpression in T47D:A18 is sufficient to allow 4-OHT to activate AP-1. We found that 4-OHT did not activate AP-1 in these T47D:A18/PKCα transfectants ( Figure 9B). This result is not surprising since the T47D:C42 (Pink et al, 1995) and T47D:A18/PKCα cell clones exhibit a hormone-independent phenotype and not a tamoxifenstimulated phenotype.

T47D:A18/PKCα proliferative rate
It was previously reported that overexpression of PKCα in MCF-7 cells leads to an increased proliferative rate compared to neo transfected MCF-7 cells (Ways et al, 1995). The T47D:A18/PKCα stable transfectants appear to have a similar growth rate as the T47D:A18/neo up until day 5. However after 5 days, T47D:A18/PKCα5 and α20 can grow to a higher cell density than the neo transfectant ( Figure 10A). After 10 days, the T47D:A18/PKCα5 and α20 cell clones have not yet reached saturation density at 1.9 × 10 7 and 2.4 × 10 7 total cells respectively, and appear to be in good condition and remain attached to the plate. In contrast, the T47D:A18/neo cells at day 10 (1.4 × 10 7 total cells) are beginning to detach from the plate and the remaining cells are in poor condition.
To determine whether PKCα overexpression is sufficient to confer hormone-independent growth we examined the rate of proliferation in oestrogen-free medium. We find that both T47D:A18/PKCα clones are able to grow at a statistically significant faster rate than the T47D:A18/neo clone in the absence of oestrogen between days 5-10 ( Figure 10B).

DISCUSSION
Approximately one half of all breast cancers are hormoneindependent, and will not respond to hormonal therapy. Although tamoxifen is the endocrine treatment of choice for ER/PR positive breast cancers of all stages, eventually most tumours will become refractory to this therapy and the disease will recur. Based on reports in the literature, activation of the AP-1 signal transduction pathway may be involved in hormone-independent and tamoxifenresistant breast cancers (Astruc et al, 1995;Webb et al, 1995;Dumont et al, 1996;Johnston et al, 1999;Smith et al, 1999). To explore this possible mechanism, we examined the PKC isozyme expression profile and AP-1 activity of MCF-7 and T47D breast cancer cell clone pairs that are hormone-dependent and -independent. We report a particular PKC isozyme, PKCα, is commonly overexpressed in both of the hormone-independent cell clones and is associated with elevated AP-1 activity. We established stable PKCα transfectants in the hormone-dependent T47D:A18 and MCF-7 A4 cell lines to determine whether PKCα overexpression is sufficient to confer elevated AP-1 activity and hormoneindependence. We find that the T47D:A18/PKCα clones exhibited decreased ER function, increased hormone-independent growth rate, decreased contact inhibition and elevated basal AP-1 activity whereas the MCF-7/PKCα clones showed no effect on ER function or AP-1 activity.
The observation that ER and PKC expression are inversely related has been previously reported . In particular, inverse expression of PKCα and ER is reported in several breast cancer cell lines, and is associated with increased cell invasion (Morse-Gaudio et al, 1998;Platet et al, 1998;Johnson et al, 1999;Ng et al, 1999;Sun and Rotenberg, 1999). PKCα and ERα are inversely related in the hormone-independent cell lines T47D:C42 and MCF-7 5C. The T47D:C42 cells have lost the expression of ERα and express elevated PKCα protein, whereas the MCF-7 5C cells have reduced ERα expression, concomitant with elevated PKCα expression (Figures 1-3). ER function is likely impaired in the 5C cells since oestrogen-stimulated PR expression and activation of an ERE-reporter construct is dramatically reduced in 5C cells (Jiang et al, 1992).
Other laboratories have previously reported stable transfection and overexpression of PKCα in the MCF-7 breast cancer cell line. Ways et al (1995) reported that MCF-7/PKCα stable transfectants exhibited reduced expression of the ER and of oestrogen-responsive genes, and resulted in a more aggressive neoplastic phenotype. Interestingly Manni et al (1996) reported that transfection of MCF-7 with PKCα produced no concomitant reduction in ER expression, and induced a less aggressive phenotype. These contrasting reports were attributed in part to differences in the expression of other PKC isozymes; Ways et al report upregulation of PKCβ, and downregulation of PKCs δ and η, whereas the transfectants described by Manni et al overexpressed only α and β isozymes (Manni et al, 1996). We find our MCF-7 5C cells overexpress PKCα, and unlike either study, PKC ε ( Figure 3A). Similar to 5C cells, the MCF-7 A4/PKCα transfectants overexpress both PKCα and ε. Despite the fact that MCF-7 5C cells exhibit non-functional ER (reduced ERE activation, absence of E2-induced cell growth), the MCF-7/PKCα transfectants exhibited intact ER function. These results are in contrast to those of Ways et al (1995). Interestingly the T47D:A18/PKCα transfectants exhibited cross-up-regulation of PKC βI and δ, loss of ER function, increased AP-1 activity, hormone-independent growth and increased saturation density. This is the first report of PKCα transfection in T47D cells and is in agreement with the results in MCF-7 cells reported by Ways et al (1995). Our inability to observe these effects in MCF-7 may be due to either differences in MCF-7 clones, or perhaps insufficient levels of PKCα expression were achieved.
It has been reported that the AP-1 pathway may be activated by long-term tamoxifen treatment (Astruc et al, 1995) and tamoxifenresistant breast cancer is associated with elevated AP-1 activity (Dumont et al, 1996;Johnston et al, 1999). Since PKCα is known to play a role in AP-1 signalling (Hata et al, 1993;Reifel-Miller et al, 1996), we investigated this pathway. It is important to point out, however, that the cell lines utilized in this study are hormone-independent. They were derived by long-term culture in oestrogen-free media and do not represent a model for long-term tamoxifen treatment. Although it was previously reported that both E2 and 4-OHT are capable of activating the AP-1 pathway (Astruc et al, 1995;Webb et al, 1995), E2 and 4-OHT did not induce AP-1 activity in either the MCF-7 nor T47D PKCα cell clones ( Figure  4B and C). This is in contrast to the previous reports showing that E2 induces a 2-fold induction of an AP-1 reporter plasmid in MCF-7 cells (Philips et al, 1993;Webb et al, 1995). In addition, 4-OHT has been reported to activate AP-1 in a cell-type specific manner, it acts as an agonist in endometrial cell lines, but not in breast cell lines (Webb et al, 1995). In addition, prolonged treatment of MCF-7 cells with 4-OHT (4-12 days) was reported to activate an AP-1 reporter plasmid (Astruc et al, 1995). Clearly our hormone-independent cell clones have not been treated with 4-OHT long-term nor represent a model of tamoxifen-stimulated growth, and therefore such a response may not be expected.
Based on our results and others (Astruc et al, 1995;Webb et al, 1995;Dumont et al, 1996), activation of the AP-1 signal transduction pathway is a likely mechanism of hormone-independent breast cancer. We hypothesise that in the transition from tamoxifen-responsive to tamoxifen-resistant breast cancer, or in the case of de novo hormone-independent breast cancers, perhaps the equilibrium shifts from predominantly ER-mediated signalling to AP-1 signalling. The report of Webb et al (1995) provided evidence that the antioestrogen/ER complex can interact with the AP-1 components through a protein:protein interaction. However, in those cases of hormone-independence due to loss of ER or a non-functional ER, PKCα up-regulation may be an alternate route to AP-1 activation. At present the mechanism of PKCα up-regulation as well as the specific intermediate steps leading towards AP-1 activation are unknown. We suggest that perhaps the overexpression of PKCα may be a prerequisite step in initiating this switch in signalling. To directly test this hypothesis, we are using antisense technology to downregulate PKCα expression in tamoxifen-stimulated breast tumours in athymic mice in an attempt to abrogate tamoxifen-stimulated tumour growth. We have preliminary evidence that the T47D:A18/PKCα clones form tumours in ovariectomised athymic mice both in response to tamoxifen and in the absence of estrogen supplementation (Tonetti, unpublished observations). We are in the process of characterising these tumours. Identification of a specific PKC isozyme as a potential therapeutic target would be an important advance in the management of hormone-independent breast cancer.