Article Text

Download PDFPDF

Expression of the forkhead box transcription factor FOXP1 is associated with oestrogen receptor alpha, oestrogen receptor beta and improved survival in familial breast cancers
  1. M Rayoo1,
  2. M Yan1,
  3. E A Takano1,
  4. G J Bates2,
  5. P J Brown2,
  6. A H Banham2,
  7. S B Fox1
  1. 1
    Department of Pathology, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia
  2. 2
    Nuffield Department of Clinical Laboratory Sciences, John Radcliffe Hospital, University of Oxford, Oxford, UK
  1. Correspondence to Professor S B Fox, Department of Pathology, Peter MacCallum Cancer Centre, St Andrews Place, East Melbourne, Victoria 3002, Australia; stephen.fox{at}petermac.org

Footnotes

  • AHB and SBF contributed equally to this work.

  • kConFab investigators also took part in the study (kConFab, Peter MacCallum Cancer Centre, St Andrews Place, East Melbourne, Victoria, 3002, Australia).

  • Funding This study was partly funded by the Victorian Breast Research Consortium and the Victorian Cancer Biobank, Australia.

  • Competing interests None.

  • Ethics approval Obtained from Oxford and Peter MacCallum Cancer Centre.

  • Provenance and peer review Not commissioned; externally peer reviewed.

View Full Text

Statistics from Altmetric.com

It is estimated that 5–10% of breast cancers are attributable to mutations in several inherited high penetrance susceptibility genes, the two most important being BRCA1 and BRCA2.1 The cumulative risk of breast carcinoma in carriers of BRCA1 or BRCA2 mutations ranges from 45% to 84% by the age of 70 years.2 Morphological and immunohistochemical differences exist in tumours of spontaneous versus germline mutation carriers, with BRCA1 breast cancers being predominantly high grade basal-like, and showing a so called triple negative phenotype for oestrogen receptor alpha (ERα), progesterone receptor (PgR) and human epidermal growth factor receptor 2 (HER2).3 4 5

We and others have been looking for other potential targets in basal-like BRCA1 tumours, since unlike other subtypes of breast cancer, BRCA1 associated tumours may be resistant to conventional hormonal and chemotherapeutic regimens.5 This may be in part due to their enhanced hypoxic drive.6 However, during our investigations in familial breast cancer phenotypes, we noted that ERβ is expressed in ∼20% of BRCA1 associated tumours and that cytoplasmic ERβ expression is associated with a shorter survival (M Rayoo, personal communication, 25 May 2009). This suggests that a significant proportion of BRCA1 associated tumours may benefit from hormonal agents such as aromatase inhibitors. Furthermore, we had noted in sporadic breast cancers a significant correlation between both ERα and ERβ, with FOXP1.7 8 FOXP1 is a member of the winged helix or forkhead group of transcriptional factors that have roles in cell proliferation, differentiation, chromatin remodelling, mitotic programme and neoplastic transformation.9 Our studies have demonstrated that FOXP1 expression is most closely related to nuclear ERβ, rather than ERα expression.8 Since recent studies have suggested that FOXP1 is a tumour suppressor gene, we have investigated the role of FOXP1 in hereditary breast tumours. Our aims were to: (1) determine the frequency and pattern of expression of FOXP1 in familial breast cancers; (2) assess the relationship between FOXP1 and different genotypes; and (3) explore for associations between FOXP1 expression, clinicopathological parameters and survival in familial breast cancer.

Methods

Patients and tumour tissue microarrays

Familial breast carcinomas from female patients were retrospectively collected from kConFab between 1980 and 2005. The classification of BRCA1 and BRCA2 mutations was performed according to designations listed on the kConFab website (www.kconfab.org). The BRCAX breast cancers were defined by familial breast cancer in families without a known BRCA1 and BRCA2 pathogenic mutation, meeting kConFab category 1 and 1B eligibility criteria.

Table 1 lists the flow of patients through the study according to the REMARK criteria.10 Of the 147 cases, 8 were excluded due to lack of tissue for array construction and 13 were excluded due to the absence of tumour on the array. The final cohort was composed of 126 cases (35BRCA1, 34BRCA2 and 57 BRCAX cases). Of the 126 samples, 119 cases had survival data available and 99 had survival data plus the full complement of ER, PgR, HER2, CK5/6 and epidermal growth factor receptor (EGFR) staining.

Table 1

Flow of familial breast cancer patients through the study, according to REMARK criteria10

These were compared with a cohort of 99 sporadic cancers, matched for age and intrinsic subtype, collected from the John Radcliffe Hospital, Oxford, UK. The median ages for the familial and sporadic cohorts were 47.1 and 47.0 years, respectively. This study has ethics committee approvals (01/38 and JR C02.216).

All patients had operable breast carcinomas without metastatic disease at the time of presentation. Information including age, tumour size, grade, histology, nodal status, ER and HER2 status were collected from the clinical and pathological records (table 2). Patients less than 50 years of age with node positive, ER negative tumours or tumours larger than 3 cm received adjuvant chemotherapy. Patients with hormone responsive tumours who were more than 50 years of age received 5 years of endocrine therapy. Patients were followed up for a median period of 64.0 months. During this time 38 patients relapsed and 31 died in the familial group (the recorded deaths were breast cancer related, otherwise were censored).

Table 2

Clinical and tumour characteristics (familial, n = 126; matched sporadic, n = 99)

Immunohistochemistry

For each tumour, four 1 mm cores were investigated using tissue microarrays (TMA), cut at intervals of 4 μm. Tissues were dewaxed and antigen retrieval was performed using a low pH buffer (Envision FLEX, DAKO, Denmark), for 20 minutes at 100°C. FOXP1 staining was performed using JC12 mouse anti-human monoclonal antibody at a dilution of 1/20, using the Envision detection kit (DAKO). Positive myoepithelial cell staining and negative stromal cell staining in normal breast tissue were used as internal positive and negative controls respectively.7 FOXP1 nuclear expression was scored using the following system: negative = 0; weak/focal = 1; strong focal/widespread moderate staining = 2; or strong/widespread staining = 3. Score 2 and 3 tumours were considered positive for FOXP1 in statistical analysis, using the previously defined cut-off value that was based on the median.7 11 12 13

ERβ staining was performed using mouse ERβ monoclonal antibody 14C8 (Abcam, Cambridge, UK) at a 1/200 dilution, overnight at 4°C. Antigen–antibody complex was detected using the same system as for FOXP1 staining. This antibody is specific for ERβ and stain ERβ1 and ERβ2 isoforms.

ERβ was scored for nuclear staining, using the following system: negative = 0; weak = 1; moderate staining = 2; or strong staining = 3. The percentage of tumour cells stained in the given core was scored as: 0% = 0; 1–10% = 1; 11–50% = 2; 51–80% = 3; 81–100% = 4. The ERβ scores for both staining intensity and the percentage of positive tumour cells were added together to give a maximum score of 7. A cut-off of 7 for nuclear expression was used to define two approximately equal size groups of positive and negative tumours for subsequent statistical analyses as previously defined that also was based on the median.8 The highest score from the four cores of the tissue array was used where any discordance between cores was noted.

HER2 competitive in situ hybridisation (CISH) and immunoperoxidase staining for ERα, PgR, HER2, CK5/6 and EGFR was performed for all tumours. HER2 CISH was performed using the Invitrogen Spotlight system (Invitrogen, San Francisco, California, USA). Tumour cells were regarded as positive for amplification if there were more than 6 signals per nucleus.14 Tumours in the equivocal group (4–6 signals) were further probed with chromosome 17 and considered amplified with a ratio of >2.2. Using stratification of intrinsic phenotypes based on Nielsen et al,15 tumours were placed into luminal (ERα positive, HER2 negative, cytokeratin (CK) 5/6 and EGFR negative or positive), basal (HER2 and ERα negative; CK5/6 or EGFR positive), HER2 (HER2 positive, ERα, CK5/6 and EGFR negative or positive) and null/negative (HER2, ERα, CK5/6 and EGFR negative) subtypes. For HER2, EGFR and CK5/6, the same cut-offs were derived from Neilsen et al. An Allred score of 2/8 was considered as positive for ERα.16

Statistical analysis

Associations of threshold data with clinicopathological parameters were evaluated using the χ2 test for independence. Correlation analyses were performed using Spearman’s rank order correlation coefficient. Kaplan–Meier survival curves were plotted using tumour recurrence (at any site following treatment) and cancer-related death as the endpoints and compared using a log rank test. Binary logistic regression was used for multivariate analyses. The Cox proportional hazard regression model was used to identify independent prognostic factors for disease-free and overall survival, with variables being entered in a single step. Analyses were performed with SPSS V.16.0 (SPSS, Illinois, USA). A two-tailed p value test was used in all analyses; a p value of less than 0.05 was considered statistically significant.

Results

FOXP1 expression in familial and sporadic breast tumours

FOXP1 was expressed in the nuclei of normal luminal epithelial and myoepithelial cells. For neoplastic cells staining was seen predominantly in the nuclei as previously described, with cytoplasmic staining being found in a minority of cases (11 of 126 tumours).7 Expression ranged from focal weak positivity to widespread strong positivity for both nucleus and cytoplasm (fig 1). Since FOXP1 is a nuclear transcription factor, and staining predominantly occurred in the nucleus, nuclear staining was used for statistical analyses. Using a cut-off of 2, positive FOXP1 expression was seen in 77% of BRCA1, 74% of BRCA2 and 74% of BRCAX, with no significant difference in FOXP1 expression between the different familial breast cancer groups (p = 0.921) (table 3).

Figure 1

Immunohistochemistry of FOXP1 staining in familial breast tumours: (A) weak focal FOXP1 staining in the tumour nuclei from a patient with BRCA2; (B) moderate homogenous FOXP1 staining in the nuclei of neoplastic cells from a patient with BRCAX; (C) strong FOXP1 nuclear and cytoplasmic staining in the tumour cells from a patient with BRCAX; (D) normal breast tissue with positive FOXP1 staining of ductal epithelium and negative staining for stromal cells; (E) diffuse weak ERβ staining in the nuclei of neoplastic cells from a patient with BRCA1; (F) strong ERβ staining in the nuclei and cytoplasm of neoplastic cells in BRCA2.

Table 3

Immunohistochemical scores for nuclear FOXP1 in familial and sporadic breast tumours (p = 0.004)

When familial cancers were stratified into intrinsic phenotypes using the protocol of Neilson et al,15 using a cut-off value of 2, FOXP1 was positively expressed in 76% (n = 38/50), 69% (n = 25/36), 67% (n = 4/6) and 71% (n = 5/7) of luminal, basal, HER2 and null type tumours, respectively. This compared with matched sporadic tumours where FOXP1 expression was present in 60% (n = 30/50), 33% (n = 12/36), 50% (n = 3/6) and 57% (n = 4/7) of luminal, basal, HER2 and null types, respectively. There was significantly higher expression in familial compared with sporadic cancers for luminal (p = 0.021) and basal (p<0.001) but not HER2 (p = 0.414) or null (p = 0.445) phenotypes. Familial cancers, as a combined group, were more likely to be positive for FOXP1 (54/99, 54%), when compared to matched sporadic tumours (46/99, 46%, p<0.001).

For familial cancers, there was no significant difference in FOXP1 expression between the molecular subtypes (p = 0.498). In contrast, for sporadic breast cancers, basal phenotype was significantly associated with lower FOXP1 expression compared with the luminal phenotype, in a multivariate analysis using binary logistic regression that was independent of tumour grade and size (p = 0.044, hazard ratio (HR) = 0.378, 95% CI 0.147 to 0.973) (table 4).

Table 4

Binary logistic regression of factors associated with positive nuclear FOXP1 in sporadic cancers

Comparison of clinicopathological parameters between FOXP1 negative and positive familial breast cancers

FOXP1 expression varied widely within both familial and sporadic groups. No significant differences were seen between FOXP1 negative and positive tumours, with respect to tumour grade, lymph node status, tumour size and HER2 status (p>0.05) (table 5).

Table 5

Contingency table of correlation between FOXP1 expression and clinicopathological variables in familial tumours

Correlation analysis for FOXP1, ERα and ERβ in familial breast cancers

There was a significant correlation between FOXP1 and ERα (r = 0.196, n = 113, p = 0.038), and ERβ (r = 0.243, n = 122, p = 0.007) in familial breast cancers. FOXP1 correlated with ERα (r = 0.436, n = 51, p = 0.001) and not ERβ (p>0.05) in BRCAX tumours. There was no significant correlation between FOXP1 with either ERα or ERβ in BRCA1 and BRCA2 cancers (all p>0.05).

FOXP1 and survival in familial breast cancers

There was a significant shorter relapse-free (p = 0.025) and overall survival (p = 0.009) for patients with tumours negative for FOXP1 expression in the familial group (fig 2). When the sporadic and familial tumours were combined, this shorter overall survival was confirmed on multivariate Cox regression analysis including ERα, HER2, age, history of familial cancer, grade, lymph node status and size (p = 0.0497, HR = 0.576, 95% CI 0.33 to 0.99) (table 6).

Figure 2

(a) Kaplan–Meier curves of relapse-free survival stratified by FOXP1 expression in all familial patients (n = 121, p = 0.025). (b) Kaplan–Meier curves of overall survival stratified by FOXP1 expression in all familial patients (n = 121, n = 0.009).

Table 6

Multivariate analysis of overall survival using Cox regression in familial tumours as a combined group, with variables entered as a single step

Although negativity for FOXP1 was associated with a significantly worse overall survival in BRCA2 (p = 0.021) (fig 3a,b), there was a non-significant separation of the survival curves for BRCA1 (p = 0.183) (fig 3c,d). No similar separation was seen for BRCAX cancers (p = 0.762) (fig 3e,f).

Figure 3

(a) Kaplan–Meier curves of relapse-free survival stratified by FOXP1 expression in BRCA2 patients (n = 33, p = 0.022). (b) Kaplan–Meier curves of overall survival stratified by FOXP1 expression in BRCA2 patients (n = 33, p = 0.021). (c) Kaplan–Meier curves of relapse free survival stratified by FOXP1 expression in BRCA1 patients (n = 34, p = 0.440). (d) Kaplan–Meier curves of overall survival stratified by FOXP1 expression in BRCA1 patients (n = 34, p = 0.183). (e) Kaplan–Meier curves of relapse free survival stratified by FOXP1 expression in BRCAX patients (n = 54, p = 0.865). (f) Kaplan–Meier curves of overall survival stratified by FOXP1 expression in BRCAX patients (n = 54, p = 0.762).

Discussion

We have examined for the first time the expression of the transcription factor FOXP1 in familial breast cancers. We used a validated antibody JC12 that has been demonstrated to be a sensitive and specific marker for FOXP1 that does not recognise the closely related FOXP2, FOXP3 and FOXP4 proteins.7 However, at least 10 isoforms of FOXP1 have been reported and the monoclonal antibody JC12 recognises eight of these that contain the epitope located within exons 18–21.13 17 As with lymphomas, preferential expression of some of the smaller isoforms may occur17 that may effect tumour behaviour and it would be of interest to investigate these in familial breast cancer subtypes. Our rationale for performing this particular analysis was not only the absence of data on this marker in familial breast cancer, but also the fact that the FOXP1 gene that maps to chromosome 3p14.1 is a region that shows widespread loss of heterozygosity in breast tumours,18 particularly in hereditary tumours with BRCA2 mutations.19

In accordance with sporadic breast tumours (and endometrial carcinomas) we observed a significant correlation between FOXP1 and both ERα and ERβ in familial breast tumours as a combined group. However, the relationship held for ERα in BRCAX only, with no association being retained for BRCA1 or BRCA2 cancers. Furthermore, in familial cancers there was no significant difference in FOXP1 expression between the molecular subtypes, in contrast to sporadic breast cancers where basal phenotype was significantly associated with lower FOXP1 expression compared with the luminal phenotype. Since both familial and sporadic tumours were stratified using the same immunohistochemical methodology, this is unlikely to be due to technical considerations, suggesting there is a true biological difference. Thus, although familial and sporadic tumour subtypes such as basal-like or luminal carcinomas share some features, there appear to be significant biological differences that are likely to extend to the manner in which they respond to therapies, emphasising the necessity for careful evaluation of both familial and sporadic tumours.

The association between FOXP1 and ER in breast cancer is unclear. In prostate cell lines there is some evidence of direct interaction of FOXP1 with the androgen receptor, resulting in down-regulation of transcription.20 However, we have been unable to show any effects of FOXP1 on androgen signalling in our studies11 and similarly have been unable to show any regulatory activity of FOXP1 on ERα in MCF7 breast cell lines.7 However, there is a reported interaction for other members of the forkhead box transcription factor family such as that described for FOXO3a and its up-regulation of ERα promoter activity and ERα protein expression in NF639 cell line derived from a mouse mammary tumour virus (MMTV)-Her-2/neu transgenic mouse.21 Although a direct regulatory effect of FOXP1 on ERα regulation cannot be fully discounted to occur in particular biological contexts, there may be an indirect effect through FOXP1 being a co-regulator of ERα. Ligand-dependent activation of gene transcription by nuclear receptors such as ER are dependent on the recruitment of coactivators. The helical LXXLL motif found in some coactivators is sufficient for ligand-dependent interaction with nuclear receptors; this signature motif, the nuclear receptor box, is present in the amino terminus of the FOXP1 protein, raising the possibility that FOXP1 might physically associate with the ER. This notion is supported by the findings that in developmental systems SMRT and FOXP1 physically interact and are recruited to the p21 promoter, and that they are required to control critical features of organogenesis.22 Indeed, the link between HOX genes that are dysregulated in breast cancers and the role of FOXP1 in regulating HOX genes in development warrant further investigation.23

The present study shows that the loss of FOXP1 expression in familial breast tumours was associated with a shorter relapse free and overall survival, the latter in a multivariate analysis. This is in keeping with the previously suggested tumour suppressor role of FOXP1 and the observations in sporadic breast carcinomas.7 24 This association with survival is in contrast to the findings in B-cell non-Hodgkin lymphomas where FOXP1 is often the target of recurring chromosome translocations frequently but not solely with the immunoglobulin heavy chain locus, where the abnormalities are associated with a poor prognosis.17 It is unknown whether the lack of FOXP1 expression in non-Hodgkin lymphomas is due to their origin from B-cells that lack FOXP1 expression or due to the silencing of FOXP1 gene expression. The correlation between loss of FOXP1 expression and poor prognosis may be in keeping with FOXP1 acting as a tumour suppressor gene.25

In summary, although FOXP1 in familial breast cancers share some similarities with sporadic cancers, there appear to be true biological differences, with an elevated expression particularly in basal-like/BRCA1 associated groups. Thus it is possible that FOXP1 has an important role in this specific subtype unrelated to its association with ERα and ERβ. It is interesting to note that sporadic basal-like tumours do not share the same FOXP1 phenotype as BRCA associated tumours, suggesting that they have different biological properties.

Take-home messages

  • FOXP1 is a tumour suppressor gene involved in the regulation of cell proliferation, differentiation, chromatin remodelling, mitosis and neoplastic transformation.

  • FOXP1 correlates with oestrogen receptor (ER)α and ERβ expression in familial cancers.

  • Different expression patterns are seen for FOXP1 in familial and sporadic breast cancers, even in cancers showing similar phenotypes.

  • Data suggest that FOXP1 expression in familial cancers is associated with improved survival, which is independent of ER expression, and may impact therapeutic options.

Acknowledgments

We wish to thank Heather Thorne, Eveline Niedermayr, kConFab, the Family Cancer Clinics, and the Clinical Follow up Study (funded by NHMRC grants 145684, 288704 and 454508), and the many families who contribute to kConFab. kConFab is supported by grants from the National Breast Cancer Foundation, the National Health and Medical Research Council (NHMRC), and by the Queensland Cancer Fund, the Cancer Councils of New South Wales, Victoria, Tasmania and South Australia, and the Cancer Foundation of Western Australia.

REFERENCES

View Abstract

Footnotes

  • AHB and SBF contributed equally to this work.

  • kConFab investigators also took part in the study (kConFab, Peter MacCallum Cancer Centre, St Andrews Place, East Melbourne, Victoria, 3002, Australia).

  • Funding This study was partly funded by the Victorian Breast Research Consortium and the Victorian Cancer Biobank, Australia.

  • Competing interests None.

  • Ethics approval Obtained from Oxford and Peter MacCallum Cancer Centre.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.