|Home | About | Journals | Submit | Contact Us | Français|
The runt-related transcription factor, Runx2 may have an oncogenic role in mediating metastatic events in breast cancer, but whether Runx2 has a role in the early phases of breast cancer development is not clear. We examined the expression of Runx2 and its relationship with estrogen receptor (ER) and progesterone receptor (PR) in breast cancer cell lines and tissues.
Two human breast cancer cell lines, MCF-7 and MDA-MB-231 were transiently transfected with vectors expressing either Runx2 or ER and the levels of both proteins and mRNA were examined by Western blot analysis and quantitative real-time PCR respectively. Runx2 expression was also examined in tissue microarray sections of 123 breast cancer patients by immunohistochemistry and results were correlated with clinico-pathological characteristics.
Expression of Runx2 and ER was reciprocal in the breast cell culture models and Runx2 suppressed ERβ but ERα not mRNA levels. In contrast, functional expression of Runx2 was evident in the nucleus in 28% of the breast cancer tissues and in both early and late stages of tumor growth. Importantly, Runx2 expression was significantly more frequent in Grade 2 compared to Grade 1 and Grade 3 tumors (48% vs 39% vs 13%) and the expression was significantly associated with ER (p=0.005), PR (p=0.008) expression in Grade 2 & 3 tumors than the Grade 1 tumors.
We propose that Runx2, ER and PR triple positivity in Grade 2 & 3 defines a biological subtype in breast cancer.
Breast cancer occurs due to the action of the two main female steroid hormones, estrogen and progesterone and their cognate nuclear receptors (respectively, ER and PR) that are known to be potent mitogenic factors. The pathological roles of both hormone receptors in the etiology of breast cancer have been demonstrated by in vitro and in vivo studies1. Both ER and PR have been implicated in the initiation or promotion of breast cancer by stimulating breast cell proliferation while bypassing cell cycle checkpoints that normally prevent damage DNA in response to genotoxic insult and permitting the incorporation of random genetic mutations2.
The estrogen receptor (ER) exists in two isoforms, ERα and ERβ that are encoded by two distinct genes, ESR1 and ESR2, located in chromosomal loci at 6q25 and 14q23, respectively3. The ratio of ERα and ERβ varie s between tissues and results in different physiological responses to estrogen. Breast cancer shows mostly higher levels of ERαand its expression is associated with more differentiated tumors and a more favorable prognosis, although the potential role of ERβ a s a favorable prognostic indicator has also been suggested4. Progesterone binds to the progesterone receptor and also has two isoforms, PR-A and PR-B that differ in protein sequence and function but are encoded by the same gene. PR-A lacks the N-terminal 164 amino acids of PR-B and has only two of the three activation function domains that are present in PR-B. Thus, PR-B shows a higher transcriptional activity and acts as a more potent transcriptional activator of target genes than PR-A5, 6. Imbalance in the ratio of these two forms of PR may be associated with the development and progression of breast cancer6-8.
The majority of breast cancers are either positive for ER alone or positive for both ER and PR. Patients that are both ER and PR positive have a- good prognosis9 and are most likely to benefit from endocrine therapy10. Other forms of breast cancer such as ER negative/PR negative or triple negative breast cancers that lack ER, PR and human epidermal growth factor receptor 2 (HER-2) also occur, though less frequently. HER-2 is encoded by the CerbB2 gene (abbreviation for ‘v-erb-b2 erythroblastic leukemia viral oncogene homolog 2’) and is also known as the neuro/glioblastoma (c-neu) proto-oncogene that is located on chromosome 17 and encodes a 185kDa transmembrane tyrosine kinase receptor. Mutations in the CerbB2/HER-2 gene predict poor outcome in breast cancer and the encoded protein is over expressed in at least 30% of breast cancers diagnosed11.
Steroid hormones including estrogen action can enhance bone formation by blocking bone resorption by osteoclasts and/or promoting the bone anabolic functions of osteoblasts. Breast cancer that metastasizes to bone is predominantly osteolytic12-16. The Runt-related transcription factor Runx217 is known to mediate the expression of proteins that support breast cancer metastasis18-22. Runx2 is essential for osteogenesis and differentiation of osteoblasts 23-24. Oncogenic properties of Runx2 have also been demonstrated in T-cell lymphomas 25 and prostate cancer 26. Apart from oncogenic functions upon ectopic expression, endogenous expression of Runx2 in osteoblasts is linked to control of cell growth 27-30 and its expression in mammary epithelial cells may support normal breast development 31-32. Other reports have suggested links between steroids and Runx2 25, 33, 34. Steroids such as glucocorticoids alter the expression of genes associated with osteoblast function at low doses 35, while high doses of this adrenal hormone cause bone loss and fractures 36, 37. Glucocorticoids have inhibitory effects on Runx2 function by depleting its nuclear accumulation 38. Runx2 has also been linked with ERα and both were shown to interact through the ERα DNA binding domain. Estrogen enhances Runx2 activity in dose and estrogen receptor-dependent ways. Notably, deletion of the DNA binding domain of ER, eliminates the stimulatory effects of estrogen on Runx2 activity 39. ER and PR serve as the basis for many therapeutic interventions in breast cancer and Runx2 has been implicated as a regulator of early metastatic events in this cancer. However, links between steroid hormone signaling and pathological functions of Runx2 in breast tumorigenesis remain to be explored.
In this study, we investigate the relationship between Runx2 and the two main female hormone receptors, ER and PR, as well as the oncogene CerbB2 (HER-2/c-neu), in a panel of breast cancer clinical samples. Our results show that nuclear presence of Runx2 is significantly and positively correlated with ER-PR positive breast cancer tissues in Grade 2 & 3 tumors. Our results provide direct clinical evidence indicating a biological role of Runx2 in the pathology of steroid hormone related breast cancers.
Two breast cancer cell lines, MDA-MB-231 and MCF-7 cells and an osteosarcoma cell line, Saos2 (ATCC, Manassas,VA,USA) were cultured in DMEM media supplemented with 10% fetal bovine serum. MDA-MB-231 and MCF-7 cells were transfected either with ER or Runx2 pcDNA3 expression vector using Fugene 6 (Roche, Basel, Switzerland) and Lipofectamine 2000 transfection reagent (Invitrogen Corporation,CA, USA) respectively. The empty vector (pcDNA3) was used as a control in both cases. Three days post transfection, protein and RNA were extracted for Western blot analysis and quantitative real-time PCR respectively.
All the three cell lines were first checked for the endogenous levels of ER, PR and Runx2. Osteosarcoma cells were used as Runx2 positive controls. Cells were lysed with lysis buffer (2% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.002% bromophenol blue and 62.6mM Tris HCl, pH 6.8), briefly sonicated and pre-cleared with high speed centrifugation. Supernatants were loaded and proteins separated by 10% SDS-PAGE and transferred to a nitrocellulose membrane. Proteins were probed with anti-Runx2 (MBL International, Woburn, MA, USA), anti-ER, anti-PR and anti-β-actin (Santa Cruz Biotechnology Inc.,CA,USA) primary antibodies followed by appropriate horse radish peroxidase conjugated to anti-rabbit or anti-mouse secondary antibodies and detected with luminol on exposure to X-ray film.
For quantitative PCR measurements, total RNA from the transfected cells (MCF-7 and MDA-MB-231) was extracted and reverse transcribed using High capacity cDNA reverse transcription kit (Applied Biosystems, USA). The cDNA (15ng) from these cells was amplified in an ABI 7300 Prism system using primers of target genes and Fermentas Maxima SYBR Green qPCR Master Mix. Cycling conditions were 10 mins at 95°C followed by 40 cycles of 15 sec at 94°C, 30 sec at 60°C and 20 sec at 72°C. GAPDH was used as an internal control. The sequences of the forward and reverse primers of the genes were as follows: GAPDH_Fwd:5′ GAGTCCACTGGCGTCTTCA 3′ GAPDH_Rev : 5′ GTTCACACCCATGACGAACA 3′ ERα_Fwd : 5′ CCTATCTCAGGGAGGGAAGG 3′ ERα_Rev : 5′ TCTCCAAGTCCCACTCTGCT 3′ ERβ_Fwd : 5′ CACCAACGAGTGCGAGATCA 3′ ERβ_Rev : 5′ TCCCCACTTTGAGGCATTTC 3′ PR_Fwd : 5′ GATTCAGAAGCCAGCCAGAG 3′ PR_Rev : 5′ GACCTTACAGCTCCCACAGG 3′ Runx2_Fwd: 5′ GGAGTGGACGAGGCAAGAGTTT 3′ Runx2_Rev: 5′ AGCTTCTGTCTGTGCCTTCTGG 3′
Tissue microarray (TMA) sections from 123 breast cancer patients (invasive ductal carcinomas representing all grade and stage) and 2 normal breast tissues (controls) were obtained from the department of pathology, National University Hospital, Singapore upon ethics approval (B06/006) by the institutional review board. These were constructed according to previously reported methods 40, representing all grades and stages of the commonest breast cancer (the invasive ductal carcinoma type). The median age of patients was 52 years (range 29-86). The distribution of patients according to the three most common ethnic groups in Singapore showed that they were of Chinese (81%), Malay (15%) or Indian (4%) descent. Histopathological staging of BC was based on the TNM staging system 41 and grading 42 of tumors. Patients that were alive and being followed up in the hospital after initial diagnosis and surgery and at the time of our data analyses were 91/123 (74%), those that did not follow up were 14/123 (11%) and those that died of the disease were 18/123 (15%).
Immunohistochemistry was performed on breast cancer TMAs constructed and stained according to our protocol described previously with minor modifications 40. Briefly, sections were microwaved at 98 °C for 20 mins in citrate buffer (pH6.0) (DAKO, Denmark) followed by peroxidase blocking for 1 hr at room temperature. Anti-Runx2 mouse mAb 43 at 1: 500 dilution, ER1D5 antibody at 1:1000, for ER detection, NCL-PGR1A6 at 1:1000 for PR and Hercep Test K5204 antibody at 1:500 for CerbB2, (DAKO, Denmark) 44 were used as primary antibodies, incubated for 2hrs at room temperature followed by diaminobenzidine (DAKO, Denmark) staining and detection.
Runx2, ER, PR and CerbB2 immunostained TMA sections were visually scored by two independent observers (AG and MST) that were blinded to clinical outcome of the patients. The staining was semi-quantitatively assessed on a scale ranging from 0 to 3 (0-no staining; 1- weak staining, 2- moderate staining and 3- strong staining) following our previously published protocols for Runx family of proteins43 and also ER, PR and CerbB2 proteins44. The status CerbB2 positive and CerbB2 negative in this study is scored by immunohistochemistry following standard practice, with +0 and +1 representing negative staining and +2 and +3 representing positive. In our hands, this scoring method in TMA is equivalent to the scoring in full sections (see reference 44) and also to HER2 status by FISH.
For statistical analyses negative and weak nuclear or cytoplasmic staining were grouped together and termed as ‘negative expression’ of the proteins and moderate and strong nuclear or cytoplasmic staining was termed as ‘positive expression’ of Runx2. Fisher’s exact test was used to test for the association between Runx2 expression and the clinico-pathological parameters of breast cancer. Survival analysis was performed according to Kaplan and Meier method. All statistical analyses were performed using SPSS version 16.0 for Windows (SPSS Inc., Chicago IL). The p values were from two-sided tests and p< 0.05 was considered significant.
We analyzed the relationship between Runx2, ER and PR genes in vitro by selecting two breast cancer cell lines, MCF-7 and the highly metastatic MDA-MB-231. The cell lines were examined for endogenous levels of Runx2 and ER and it was observed that MCF-7 was ER positive and negative for Runx2 while MDA-MB-231 was ER negative and positive for Runx2 expression level (Figure 1A). This finding is consistent with a requirement for ER in early stages of breast cancer2, 4 and the metastatic potential of Runx2 in later stages14,18,19.
To assess whether the relationship between ER and Runx2 reflects regulatory coupling, the cells were then transiently transfected with vectors expressing either Runx2 or ER and the levels of both proteins and mRNA were examined by Western blot analysis and quantitative real-time PCR respectively. In the Runx2 negative MCF-7 cell line, transfection of Runx2 suppressed expression of ER (Figure 1B). In contrast, exogenous expression of ER in negative MDA-MB-231 cells appreciably altered the levels of Runx2 protein (Figure 1C). The protein data for Runx2 and ER were confirmed by examination of transcripts level. The endogenous Runx2 mRNA level was low or diminished in ER positive cell lines (MCF-7). In contrast, the ER-negative cell lines (MDA-MB-231) showed high levels of Runx2 mRNA (Figure 2A). Interestingly, exogenous expression of Runx2 in MCF-7 cells exhibited low levels of ERα , ERβ and PR (Figure 2B) indicating that Runx2 effects the expression of the two ER isoforms. Similarly, when MDA-MB-231 cells were over expressed with ERα, the Runx2 mRNA level was suppressed (Figure 2C). Thus, ER and Runx2 expression is inversely correlated in these cell lines, consistent with their respective roles in early and late stages of breast cancer, and Runx2 is capable of attenuating ER levels.
To begin understanding the clinical relationships between steroid hormone responsive breast tumors and Runx2, we performed Runx2 immunostaining and analyzed the localization of the protein in a spectrum of breast tumor clinical samples. Of the 123 tumors , Runx2 nuclear expression was detected in 14 (11%) cases (Figure 3a), cytoplasmic expression in 24 (19%) cases (Figure 3b) and concurrent nuclear and cytoplasmic expression was observed in 20 (16%) cases (Figure 3c). There was no expression of Runx2 in 65 (53%) cases. Two normal tissues that were considered as controls showed weak nuclear expression only, consistent with previous data suggesting a role for Runx2 in specific stages of normal breast epithelial development 31.
Based on the known role of Runx2 as a nuclear transcription factor, we propose that nuclear expression of Runx2 reflects a functionally active form of the protein, whereas its cytoplasmic localization indicates loss of function in gene regulation at least in a basal state. We do not exclude the possibilities that cytoplasmic Runx2 may provide a reservoir of sequestered Runx2 that might translocate to the nucleus only upon stimulation, or simply represents an inactive bystander molecule. We interpreted concurrent expression of Runx2 in both nucleus and cytoplasm as indicating that the protein has retained its transcriptional function, and thus is at least in part active in gene regulation.
Table 1 shows the correlation of Runx2 nuclear expression with the pathological parameters in the breast tumors. There was no significant correlation between ER, PR and Runx2 nuclear expression in the breast tumors although Runx2 nuclear protein was more evident in CerbB2 negative tumors than the CerbB2 positive tumors (70% vs 30%) that was nearly statistically significant (p=0.092),..
The ER and PR status in the tumors, when compared with pathological grade showed that ER was positive for (95%) grade 1 tumors (G1), positivity was 81% in grade 2 (G2) tumors and 47% in grade 3 (G3) tumors that was statistically significant (p<0.0005) (Figure 4a). Similar relation was also observed between PR and tumor grades (91% vs 76% vs 37%; p<0.0005) (Figure 4b). This inverse relation is consistent with previous studies and consistent with loss of estrogen-responsiveness in late stage breast tumors (44). Expression of CerbB2 (HER-2/c-neu) is elevated in tumors of higher grades (15% vs 40% vs 68%; p<0.0005).
We next examined the expression of Runx2 in different grades of breast cancer. The nuclear expression of Runx2 varied significantly across the three grades of breast cancer with the highest expression in G2 tumors (Figure 5). Of the 31 tumors that showed Runx2 nuclear expression, 48% were G2, 13% were G3 and the remaining 39% represented G1 cancer (Table1). The Runx2 expression was then correlated with ER and PR positivity in different tumor grade and a statistically significant association was observed between them in G2 & G3 tumors compared with G1 tumours (Table 2). However, CerbB2 showed a negative correlation with Runx2 expression in G2 & G3 tumors which was also significant (Table2). In other words, Runx2 was more frequently expressed in CerbB2 negative tumors than the positive ones (39% vs 17%). There were no significant differences between Runx2 expression and different breast cancer stages or with axillary lymphnode metastasis. Runx2 cytoplasmic expression also did not show any significant difference when compared with the pathological parameters. Our statistical analyses did not show any correlation with patient survival. This is fully consistent with the predominant intermediate grade of Runx2 expression in breast cancer.
There are breast cancer conditions as reflected by cell lines in our study that are ER positive and Runx2 negative as observed in MCF-7 and those that are ER negative and Runx2 positive in MDA-MB-231 cell lines. Reciprocal relationship between Runx2 and ER expression was observed at both the mRNA and protein level suggesting that high levels of Runx2 may be may be effecting the expression of ER in MDA-MB-231 cell lines. Because Runx2 and ER can synergize in biological processes (e.g., they are both bone anabolic), one key question is whether there is an in vivo transition in breast cancer in which both factors can co-exist. We tested this by co-staining breast cancer clinical samples for both ER and Runx2 and assessed the biological role of Runx2 in all grades of this disease together with the hormone receptors.
Runx2 and ER expression results were similar to those of the cell lines in G1 tumors and G3 tumors while in G2 tumors, Runx2, ER and PR expression were increased significantly and there was a significant positive association between Runx2 nuclear protein and ER and PR expression. These data suggest that Runx2 is expressed early on during tumor growth and development.
A link between estrogen and Runx2 has been reported through an increase in the number of Runx2 expressing cells 45. Selective ER modulators, raloxifene having different modes of action from estrogen enhanced Runx2 promoter activity 46.To bind to the target gene promoters, nuclear import of transcription factors is an essential element of the molecular pathways and when these proteins are present in the cytoplasm, they are functionally inactive and may act as bystanders or as a reservoir of unstimulated forms. Runx activity is regulated by tissue—specific expression and also those elements that control its subcellular localization. Cytoplasmic and nuclear localization has been shown for Runx3 in gastric cancer 47 and colon cancer 48 and Runx2 in osteosarcoma cells49. Besides nuclear and cytoplasmic expression of Runx2, our study also showed concurrent expression of this protein in the two subcellular components in 16% of the tumors and more frequently in the G2 tumors. This reflects ectopic induction of Runx2 expression during breast cancer progression. Although Runx2 expression was prevalent during tumorigenesis of mammary epithelial cells as evidenced by our breast cancer tissues, only a subset of tumors showed nuclear localization that would enable it to regulate its target genes.
Strikingly, the nuclear staining correlates with “ER positivity” suggesting that both factors synergize to stimulate tumor growth and development at one or more specific stages. Our observation that 32 % of ER+ breast cancer clinical samples express Runx2 (n=88) is consistent with a patho-physiological role of ectopic Runx2 expression. However, because only 28% of breast cancer tissue exhibits nuclear localization of Runx2 (n=123), it appears Runx2 is not actively participating in gene regulatory events, although we cannot exclude the possibility that Runx2 may perturb gene regulatory events by its cytoplasmic localization (e.g., sequestering of key co-factors like Smad and Yap that transduces TGF-beta and Src/Yes related signals respectively).
ER and Runx2 are both known to contribute to carcinogenesis independently. However, both proteins can also synergize at the molecular level in a manner that may reinforce their biological functions in the pathology of breast cancer. The co-expression of ER and Runx2 that we have established in this study suggests that both proteins may have biological synergism that may be further supported by molecular synergy. In vitro studies have shown that the expressions of ER versus Runx2 are inversely correlated50, parallel to what we have observed. But the novelty of our study is that we also found a subset of breast cancer that showed co-existence of ER and Runx2 suggesting the transition from positive to negative or vice versa as a gradual process.
Our results show that the in vivo disease is much more heterogeneous than the perceptions based on cell culture studies, highlighting the importance of analyzing clinical samples to understand the pathological linkages between different classes of transcription factors like ER and Runx2. Indeed, our full analysis showed cases with this reciprocal relation but, in addition, there were three important subclasses of breast cancer: ER+/PR+/Runx2-, ER+/PR+/Runx2+ and ER-/PR-/Runx2+ (in order of disease progression). We propose that ER+/PR+/Runx2+, triple positivity defines a biological subtype in breast cancer.
In summary, our results stress the idea that solid tumors including breast cancer, are not biologically homogenous disease and defining the biological nature of potential subgroups allows a better understanding of the disease. Finally, it is interesting to note that a defining biomarker of metastasis such as Runx2 is predominantly expressed in G2 breast cancers. Further investigation of its relationship with ER& PR may lead to a better understanding of the molecular processes related to breast cancer metastasis.
We would like to thank Dr Doris Mayer, German Cancer Research Center, Germany for the kind gift of ER expression plasmid. Our study was supported by SCS grants MN05 and MN077 (to MST), NIH grants P01CA82834 (to G.S.S.) and AR049069 (to AJvW.) and Lee Kuan Yew Postdoctoral Fellowship awarded to DTL.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.