Background
Breast cancer is a heterogeneous disease which encompasses several subgroups with different morphology, genetic changes, and response to therapies [
1,
2]. It is therefore important to gain more insight into relevant therapeutic targets for each subgroup to optimize tailored treatment protocols for individual patients. Numerous intracellular signaling proteins have been suggested to be promising targets for blocking the malignancy of breast cancer cells. The protein kinase C (PKC) isoforms are examples of such potential therapeutic targets.
PKC is a family of serine/threonine kinases involved in several processes including proliferation, differentiation, apoptosis, and migration. The PKC isoforms are divided into three subgroups depending on the structure of the regulatory domain: classical (PKCα, βI, βII, and γ), novel (PKCδ, ε, and θ), and atypical (PKCζ and ι/λ) isoforms. Classical and novel PKCs contain a diacylglycerol (DAG)-binding C1 domain and are therefore regulated by activation of pathways that lead to DAG generation. Atypical PKCs are DAG-insensitive and regulated in a different manner [
3].
Several studies have implicated the DAG-sensitive classical and novel PKC isoforms in promoting malignant features of breast cancer cells. PKCα has been coupled to estrogen receptor (ER) negativity [
4] and estrogen-independent growth of cultured cells [
5,
6] and patients with PKCα-negative tumors had better response to endocrine treatment compared to patients with PKCα-positive tumors [
7,
8]. Moreover, increased PKCα expression leads to a more aggressive phenotype [
4] and is associated with resistance to cytostatic drugs in MCF-7 cells [
9,
10]. PKCα is also evaluated as a therapeutic target for breast cancer [
11]. However, PKCα levels are reduced in breast cancer compared to normal breast tissue [
12,
13]. Thus, there is evidence for both a promoting and a suppressing role for PKCα in breast cancer.
The role of PKCδ in breast cancer is ambiguous. Patients with PKCδ-positive tumors show better endocrine response compared to patients with PKCδ-negative tumors [
8] and PKCδ has been shown to be crucial for UV light-induced apoptosis of cultured breast cancer cells [
14]. However, several studies point to a pro-tumorigenic role of PKCδ in breast cancer. PKCδ can induce resistance to tamoxifen and irradiation in cultured breast cancer cells [
15,
16] and has been shown to promote both metastasis [
17‐
19] and proliferation [
20] of murine mammary cancer and epithelial cells. We have recently shown that depletion of PKCδ is sufficient to drive breast cancer cells into apoptosis [
21].
PKCε has frequently been assigned oncogenic effects in breast cancer. Expression levels of PKCε have been shown to correlate with tumor grade, HER2 expression, ER negativity, and poor survival in breast cancer patients. Moreover, in MDA-MB-231 breast cancer cells, downregulation of PKCε reduced the tumor growth and metastatic capacity in mice [
22]. There is also evidence that PKCε protects cells against apoptotic insults [
23‐
25].
Taken together, the available in vitro and in vivo data highlight PKCα, PKCδ, and PKCε as future candidates for targets in breast cancer therapy and as markers for disease prognosis. However, so far there is limited knowledge on the potential of the different isoforms as diagnostic and prognostic markers in breast cancer. This study sheds light on this issue by analyzing the expression levels of these PKC isoforms in primary breast cancer tissue and our results indicate that PKCα is a potential marker of breast cancer aggressiveness.
Methods
Cell culture
All cell lines were obtained from ATCC. MCF-7, MDA-MB-231, and MDA-MB-468 breast cancer cells were maintained in RPMI 1640 medium (Sigma) supplemented with 10% fetal bovine serum (FBS; Invitrogen), 1 mM sodium pyruvate (PAA laboratories Gmbh), 100 IU/ml penicillin, and 100 μg/ml streptomycin (both Gibco). T47D cells were grown in DMEM supplemented with 10% FBS, 10 mM HEPES (PAA laboratories Gmbh), 100 IU/ml penicillin, and 100 μg/ml streptomycin. The media for MCF-7 and T47D cells were additionally supplemented 0.01 mg/ml insulin (Novo Nordisk A/S).
Transfections
For siRNA transfections, cells were seeded at 35-50% confluency and grown in complete medium without antibiotics for 24 hours. Cells were transfected for 48 hours using 4 μl/ml Lipofectamine 2000 (Invitrogen) and 40 nM siRNA (Invitrogen, table
1) in Optimem (Gibco) according to supplier's protocol.
Table 1
siRNA oligonucleotides
Control | GACAGUUGAACGUCGAUUUGCAUUG |
PKCα #1 | CCGAGUGAAACUCACGGACUUCAAU |
PKCα #2 | CCAUCGGAUUGUUCUUUCUUCAUAA |
PKCδ | CCAAGGUGUUGAUGUCUGUUCAGUA |
PKCε | CACAAGUUCGGUAUCCACAACUACA |
Plasmid transfections were carried out for five hours replacing normal medium with Optimem containing 2 μl/mL Lipofectamine 2000 and 2 μg/mL DNA according to supplier's protocol. Plasmids encoding PKC constructs fused to enhanced green fluorescent protein (EGFP) have been described previously [
26].
Tumor material
Cohort I originally consisted of tumors from 114 patients diagnosed with breast cancer at Umeå University Hospital and treated according to regional guidelines. The cohort is described in table
2. Due to lack of tumor material in the tissue microarray, 42-60 tumors could be analyzed, depending on the parameter investigated. Ki-67 had been classified in two groups <20% and >20% positive cells.
Table 2
Characteristics of the cohorts investigated
Number of patients | 114 | 512 |
Age at diagnosis, median (range) | 60 (30-80) | 65 (27-96) |
Tumor size (mm), median (range) | 22 (8-100) | 16 (1-100) |
Nodal status | | |
Positive | 53 | 168 |
Negative | 48 | 298 |
Missing | 13 | 58 |
ER-status | | |
Positive | 82 | 417 |
Negative | 31 | 72 |
Missing | 1 | 35 |
Cohort II included 512 consecutive breast cancer cases diagnosed at the department of Pathology, Malmö University Hospital, between 1988 and 1992. The cohort is described in table
2. Due to lack of tumor material in the tissue microarray, 223-263 tumors could be analyzed, depending on the parameter investigated. Ki-67 had been classified in three groups, 0-10%, 11-25% and 26-100% positive cells.
The cohorts represented a mix of all histological subtypes in proportions corresponding to the common incidence. The construction of tissue microarrays and clinicopathological properties of the cohorts have been described in detail elsewhere [
27‐
32]. Ethical permissions were obtained from the Lund and Umeå Ethical Boards. The number of tumors that could not be evaluated for PKC expression due to lack of material is indicated as not evaluated in the tables.
Cell pellet arrays
Cells were washed in phosphate-buffered saline (PBS) and fixed for 25 minutes in 4% paraformaldehyde in PBS with Mayer's hematoxylin (5 μl/ml) present during the last 5 minutes. The cells were pelleted, paraformaldehyde was removed and they were thereafter incubated over night in 70% ethanol followed by dehydration using increasing concentrations of ethanol and finally xylen. After dehydration, cell pellets were embedded in paraffin and arranged in a cell line array.
Immunohistochemistry
Sections (4 μm) of the paraffin blocks were dried, deparaffinized, rehydrated and microwave-treated in 1× target retrieval solution with high pH (DAKO). All sections were stained in a DAKO Techmate™ machine and visualized using DAB. The antibodies used were PKCα (1:2000), PKCδ (1:1000), PKCε (1:400; all Santa Cruz Biotechnology, product numbers sc-208, sc-937 and sc-214), and Ki-67 (1:200; DAKO). All tissue microarray slides of a cohort were stained simultaneously with the same staining solutions ensuring identical conditions for each tumor. PKC stainings were scored according to cytoplasmic staining intensity were 0 representing lack of staining, 1 low staining, 2 moderate staining, and 3 strong staining. All cohorts were examined independently by two investigators and disconcordant results were re-evaluated. For Ki-67 analyses of breast cancer cell lines, Ki-67 staining intensity was scored as negative-low or moderate-strong staining.
Sample preparation and Western blot
Cells were washed twice in ice-cold PBS and lyzed with RIPA buffer (10 mM Tris-HCl pH 7.2, 160 mM NaCl, 1% Triton-X 100, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 1 mM EDTA, 1 mM EGTA) supplemented with 40 μl/ml complete protease inhibitor (Roche Applied Science) for 30 minutes on ice. Lysates were cleared by centrifugation at 14,000 × g for 10 minutes at 4°C.
Proteins were separated with SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore). Membranes were blocked with PBS containing 0.05% tween and 5% non-fat milk, and probed with antibodies towards PKCδ (1:500), PKCε (1:500), PKCα (1:3000), and actin (1:2000; MP Biomedicals, clone C4). Proteins were visualized with horseradish peroxidase-labeled secondary antibody (Amersham Biosciences) using the SuperSignal system (Pierce Chemical) as substrate. The chemiluminescence was detected with a CCD camera (Fuji Film).
Analysis of cell growth (WST-1 assay)
Cell were seeded at a density of 2000 cells per well in 96-well culture plates, and incubated for 24 hours. For cell line comparison, viable cell number was measured 24 and 48 hours after seeding. For experiments with inhibition or activation of PKC, 2 μM Gö6976 (Calbiochem) or equal volume DMSO, or 16 nM 12-O-tetradecanoylphorbol-13-acetate (TPA; Sigma) was added in complete medium (CM) or serum-free medium (SFM) 24 hours after seeding, and cells were incubated for 72 hours prior to estimation of viable cell number. The amount of viable cells was assessed by a WST-1 cell viability assay (Roche Applied Science). Absorbance was measured in an ELISA plate reader Antos 2020 (Antos Labtech Instruments).
Immunofluorescence and confocal microscopy
Immunofluorescence of PKCα was done as described [
33] using Alexa Fluor 488-conjugated secondary antibodies. Cells were examined with a Zeiss LSM 710 confocal system using standard settings for Alexa Fluor 488.
Cell cycle analysis
MDA-MB-231 cells were seeded at a density of 100,000 cells per 35 mm cell culture dish and transfected with siRNA. After transfection, cells were incubated in SFM or CM for 24 hours. Cells were trypsinized and fixed in 70% ethanol for 20 minutes at -20°C, washed in PBS, and incubated with a solution containing 3.5 μM Tris-HCl pH 7.6, 10 mM NaCl, 50 μg/ml propidium iodide (PI), 20 μg/ml RNase, and 0.1% igepal CA-630 for 20 minutes on ice in order to label DNA. 10,000 events were acquired on the FL-2 channel for the PI signal. Sample acquisition and analyses were performed with CellQuest software (Becton Dickinson).
Wound healing assay
MDA-MB-231 or MCF-7 cells were seeded at a density of 150,000 cells per 35 mm cell culture dish and transfected with siRNA. After transfection, a scratch was made with a 200 μl pipette tip in a confluent area of the cell culture dish. Photographs of a selected area of each scratch were taken at indicated time points. For experiments with PKC inhibitors, MDA-MB-231 cells were seeded at a density of 350,000 cells per 35 mm cell culture dish and incubated for 24 hours before a scratch was made and PKC inhibitors were added. Photographs of a selected area of each scratch were taken 0 and 16 hours after scratching. The remaining wound area was measured using ImageJ software.
Statistics
For TMA analysis, correlations between variables were calculated using Pearson's two-tailed significance test. Differences in distribution of various clinicopathological parameters in regard to PKC expression were also calculated using the χ2-test. The Kaplan-Meier analysis and the log rank test were used to illustrate differences between recurrence-free survival (RFS) and breast cancer-specific survival (BCSS) according to PKC expression. Cox regression proportional hazards models were used to estimate the impact of PKCα expression on breast cancer-specific survival in both univariate and multivariate analysis, adjusted for Nottingham histological grade (NHG), age, lymph node status, and tumor size of cohort II. For in vitro experiments, the significance of difference was assessed by analysis of variance (ANOVA) followed by Duncan's multiple range test. The difference was considered significant if the p-value was < 0.05. All statistical calculations were performed using SPSS V.11.0.
Discussion
Several studies have suggested that DAG-sensitive PKC isoforms contribute to the progression of breast cancer and the malignant characteristics of breast cancer cells. In particular the PKCα, PKCδ, and PKCε isoforms have been highlighted as potential targets for therapy of breast cancer or of specific subsets of the disease. This led us to design this study in which the expression of these PKC isoforms in primary breast cancer tumors has been examined to assess their utility as markers of tumor aggressiveness.
To substantiate the analysis, two different cohorts of primary breast cancer tumors were used, and the results from the cohorts were similar. We found significant correlations between PKCα levels and several markers of tumor aggressiveness including ER negativity. A correlation between PKCα levels and ER negativity has also been observed in a recent study [
8] using 70 tumors from patients that had received systemic endocrine therapy. They also showed that high PKCα levels predicted a worse outcome in response to endocrine therapy, which is also supported in an earlier study with a smaller number of patients [
7]. Our data, as demonstrated in two separate cohorts, firmly establish the relationship between PKCα and ER negativity. In addition, we identify a clear correlation between PKCα and PR negativity, and a positive correlation with tumor grade and high proliferation rate, further supporting the notion that PKCα expression is associated with parameters related to tumor aggressiveness. Finally, PKCα expression predicts a worse disease outcome with a significantly poorer 10-year breast cancer-specific survival for patients with primary tumors that were PKCα-positive. The results from the multivariate analysis further indicate that PKCα is an independent prognostic factor in breast cancer.
However, other studies have indicated that PKCα levels actually are decreased in breast cancer compared to normal breast tissue [
12,
13]. This may not necessarily contradict our findings since a vast majority of the cancer samples in our material were essentially PKCα-negative, which is in line with the notion that PKCα downregulation is a common event during breast cancer progression. Therefore, the published data, together with our results, suggest that most breast cancers are PKCα-negative, but that there are smaller subgroups with higher PKCα levels, displaying more aggressive clinicopathological features. If PKCα is to be used as a target for breast cancer therapy these data highlight the need to evaluate PKCα levels prior to such intervention.
In the largest cohort (II) there was also a significant association of PKCα levels with histological subtypes. Cancers with medullary histology were over-represented in the group with high PKCα levels. In addition, the only medullary cancer in cohort I was PKCα-positive.
In line with the tumor data demonstrating that PKCα levels correlate with features of aggressiveness, an association of PKCα with malignant features has also been seen in breast cancer cell lines. Increased levels of PKCα correlate to and can induce tamoxifen [
38,
39] and multidrug [
9,
10] resistance of ER-positive cell lines. In addition, overexpression of PKCα in MCF-7 cells leads to increased proliferation which is in line with our tumor data [
4], but also make them more susceptible to apoptotic insults [
40,
41]. We found that the cell line with no detectable PKCα expression, T47D, had a lower percentage of Ki-67-positive cells and slower growth rate than the other cell lines, similar to the tumor data. However, the cell line with the highest PKCα expression, MDA-MB-231, only had marginally higher Ki-67 positivity compared with cell lines with detectable but much lower expression. This may be related to the fact that the fraction of Ki-67-positive cells was high (97%) and could not be further elevated.
Neither inhibition of PKCα in the presence of serum nor PKC activation under serum-starvation influenced the growth of the cell lines in a manner that would support an essential role for PKCα for breast cancer cell growth. A recent paper has shown that the PKCα protein, but not activity, is essential for glioma cell proliferation [
35]. We found that downregulation of PKCα with siRNA caused a modest effect on cell cycle distribution of MDA-MB-231 cells grown in complete medium. However, under serum-free conditions, PKCα silencing clearly reduced the amount of proliferating cells suggesting that, as in glioma cells, the PKCα protein, but maybe not its catalytic activity, supports proliferation under sub-optimal conditions.
Studies
in vitro have demonstrated that increasing the PKCα expression in MCF-7 cells makes them more migratory in response to PKC activation [
42]. Our experiments with cell lines indicate that PKC activity supports migration of MDA-MB-231 cells. However, downregulation of PKCα with siRNA in this cell line did not affect cell migration. MDA-MB-231 cells have high basal expression levels of PKCα and incomplete downregulation of PKCα may explain the lack of effect on cell migration. This assumption is supported by the more efficiently suppressed wound healing upon downregulation of PKCα in MCF-7 cells. Our data therefore indicate that PKCα activity is important for migration of breast cancer cells, in line with previous findings using PKCα overexpression. A migratory propensity might facilitate the metastasation of a tumor. However, our tumor data did not support a role for PKCα in dissemination. There was no correlation between PKCα expression and nodal or distant metastases. Thus, the effects of PKCα on migration may primarily be of importance
in vitro.
For PKCδ there is less information regarding expression levels in primary tumors. One study has shown that low levels of PKCδ, particularly in combination with high PKCα, predict resistance to endocrine therapy [
8]. In this study, we could not observe any significant associations between PKCδ expression and relevant clinicopathological parameters. Thus, altered PKCδ expression does not seem to be a prerequisite for breast cancer progression.
PKCε has been proposed to be a marker of aggressive breast cancer since its expression was reported to be elevated in hormone receptor-negative breast cancers with high tumor grade and
HER2 amplification [
22]. However, a relationship between PKCε expression and tumor grade could not be confirmed in this study, suggesting that it might not be apparent in cohorts representing all histological subtypes of breast cancer. It is possible that the level of PKCε is a marker of aggressiveness in more defined subgroups of breast cancer.
Acknowledgements
We are grateful for excellent technical assistance from Elise Nilsson. This work was supported by grants from The Swedish Cancer Society, The Swedish Research Council, The Children's Cancer Foundation of Sweden, Malmö University Hospital Research Funds, and the Kock, Crafoord, Ollie and Elof Ericsson and Gunnar Nilsson Foundations.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
GKL evaluated clinical samples, performed statistical analyses, participated in the design of and did most of the experimental work, and assembled the drafts of the manuscript.
LC did some experimental work.
IOZ took part in the evaluation of the antibodies and clinical samples.
GL supervised the analysis of cohort I and participated in interpretative discussions.
KJ supervised the analysis of cohort II and statistical analyses of patient data, and helped draft the manuscript.
CL conceived of the study, participated in the design of the experimental work and helped draft the manuscript.
All authors read and approved the final manuscript.