Article Text


Forkhead box A1 expression in breast cancer is associated with luminal subtype and good prognosis
  1. M A Thorat1,
  2. C Marchio2,3,
  3. A Morimiya1,
  4. K Savage2,
  5. H Nakshatri4,5,
  6. J S Reis-Filho2,
  7. S Badve1,6
  1. 1
    Department of Pathology and Laboratory Medicine, IU School of Medicine, Indianapolis, IN 46202, USA
  2. 2
    Molecular Pathology Laboratory, The Breakthrough Breast Cancer Research Centre, Institute of Cancer Research, Fulham Road, London SW3 6JB, UK
  3. 3
    Department of Biomedical Science and Human Oncology, University of Turin, Turin, Italy
  4. 4
    Department of Surgery, IU School of Medicine, Indianapolis, IN 46202, USA
  5. 5
    Department of Biochemistry and Molecular Biology, IU School of Medicine, Indianapolis, IN 46202, USA
  6. 6
    Department of Internal Medicine, IU School of Medicine, Indianapolis, IN 46202, USA
  1. Sunil Badve, Department of Pathology, Indiana University School of Medicine, 635 Barnhill Drive, MS-A128, Indianapolis, IN 46202, USA; sbadve{at}


Aims: Forkhead box A1 (FOXA1) is a forkhead family transcription factor expressed in breast cancer cells. It is essential for optimal expression of ∼50% of oestrogen receptor (ER)-related genes. This study explored the FOXA1 relationship with luminal and basal breast cancer subtypes, proliferation markers, and survival in breast cancer patients who had received similar treatment.

Methods: A tissue microarray comprising tumours from 245 invasive breast cancer patients with 67 months of median follow-up was analysed for FOXA1 expression by immunohistochemistry. Interpretable FOXA1 expression, obtained in 184 patients, was analysed along with other variables such as tumour grade, size, nodal status, ER, progesterone receptor, HER2/neu, proliferation and basal markers.

Results: FOXA1 expression (score >3) was seen in 139 of 184 breast cancers. It correlated positively with ERα (p<0.0001), progesterone receptor (p<0.0001), and luminal subtype (p<0.0001); negatively with basal subtype (p<0.0001), proliferation markers and high histological grade (p = 0.0327). Univariate analysis showed nodal status, tumour grade, ER, progesterone receptor, FOXA1, basal markers and p53 as significant predictors of overall survival. Multivariate analysis showed that only nodal status (p = 0.0006) and ER (p = 0.0017) were significant predictors of OS. In luminal subtype patient subgroup, FOXA1 expression was associated with better survival (p = 0.0284) on univariate analysis.

Conclusion: Based on this study in patients treated with surgery followed by adjuvant anthracycline-based chemotherapy, FOXA1 expression is associated with good prognosis. It correlates with luminal subtype breast cancer, and could possibly serve as a clinical marker for luminal subtype A. Prognostic ability of FOXA1 in these low-risk breast cancers may prove to be useful in treatment decision making.

View Full Text

Statistics from

Oestrogen receptor (ER) expression is an important prognostic factor in breast cancer and it is an excellent predictive factor for response to endocrine therapy.1 ER-positive breast cancers are generally less aggressive; however, not all ER-positive breast cancers behave alike. ER-positive breast cancers do seem to have different subgroups, which are yet to be well characterised clinically. Such clinically reproducible characterisation of ER-positive breast cancer subgroups will require better understanding of the ER pathway and allied factors, followed by clinical validation of markers identified through this mechanism.

Oestrogen plays an important role in the growth, proliferation and differentiation of mammary epithelium. Oestrogen action is mediated through the ERs, which function as oestrogen-activated transcription factors.1 2 Although the actual role of ERβ in breast cancer is still a matter of controversy,3 ERα has been shown to play a pivotal role in breast cancer. This receptor is expressed in 10–15% of luminal epithelial cells of normal breast.4 ERα plays a distinct role in the growth of normal, immortalised, and transformed mammary epithelial cells, since more than 50% of breast cancers express ERα at the time of initial diagnosis.1 Identifying factors controlling these distinct functions of ERα is important for understanding transformation of ERα-positive normal luminal epithelial cells from a slowly proliferating and well-differentiated phenotype to a rapidly proliferating aggressive malignant phenotype. These factors could also be explored therapeutically in future.

Breast cancer is now classified into six distinct subtypes based on unique molecular characteristics (gene-expression) and prognostic significance.5 6 These subtypes are: luminal subtypes A, B and C, human epidermal growth factor receptor 2 (HER2)+/ER−, basal-like, and normal-like subtypes. Basal-like and HER2+/ER− breast cancer subtypes are now well characterised clinically, in congruence with the molecular subtype classification.7 Furthermore, it has been shown that patients with HER2+/ER− cancers benefit from anti-HER2 tailored therapies, including anti-HER2 humanised monoclonal antibodies and HER2 tyrosine kinase inhibitors. It has not yet been possible to characterise luminal A and B subtypes and normal-like subtype in a clinically reproducible manner. According to gene-expression-based classification, Luminal subtype A and -B are ERα-positive breast cancers. Luminal subtype A has better prognosis than subtype-B and expresses higher levels of ERα.6 It is possible that ERα functions differently in luminal A versus luminal B cancers; this may be due to the influence of additional factors, including transcription factors, co-activators and co-repressors modulating ERα activity.

Forkhead box A1 (FOXA1), also known as hepatocyte nuclear factor 3 alpha (HNF3α) is one such transcription factor co-expressed with ERα in luminal subtype A.5 8 FOXA1 promotes gene expression by altering chromatin structure. FOXA1 binds to specific DNA sequences on the nucleosome core and de-compacts chromatin to facilitate binding of other transcription factors to gene promoters.9 10 FOXA1 helps the recruitment of ERα on a loosened chromatin environment.11 Additionally, it is a key factor in controlling glucose availability and it is speculated that it could specifically favour the regulation of metabolism-associated genes by ERα in breast cancer.12 Indeed, recent studies have shown that FOXA1 is required for optimum expression of ∼50% of ERα-regulated genes and oestrogen-induced proliferation.10 13 14 Co-expression of FOXA1, GATA3, XBP-1 and ERα in cell lines and breast cancer samples at the mRNA level has been reported earlier.8 12 15 Furthermore, recent studies have described GATA-3, FOXA1 and ERα as components of the master cell type-specific transcriptional network that dictates the phenotype of hormone-dependent breast cancer.16 Interestingly, FOXA1 has been shown to transcriptionally activate the cell cycle inhibitor p27 and forced expression of this transcription factor in the luminal (MCF7 and SKBR3) and basal (MDA-MB-231) breast cancer cell lines led to a reduction in number and size of surviving colonies of all cell lines compared with empty vector transfected controls.17 In the first detailed immunohistochemistry (IHC) study18 examining FOXA1 and its relationship with ERα and other disease markers, we reported that FOXA1 expression is associated with ERα-positivity, the luminal subtype, and better breast cancer-specific survival. That study18 was performed on an old patient cohort stretching over 20 years; survival data in that cohort was used to decide FOXA1-positive and -negative cut-offs. Patients in that cohort had received different treatments and treatment details were not available for majority of patients.18 Due to these limitations of previous study, it was necessary to validate the findings in an entirely different set of patients who have been treated similarly to rule out confounding due to treatment interactions. In present study, we validate our earlier findings in a cohort of patients uniformly treated with therapeutic surgery followed by adjuvant anthracycline-based chemotherapy and, in particular, the cut-off for FOXA1 expression in breast cancer. Additionally, we also explore FOXA1 relationship with basal and proliferation markers.


Patient information and tissue microarray

A tissue microarray (TMA) was constructed with primary breast cancer samples with replicate 0.6 mm cores of 245 invasive breast carcinomas (186 invasive ductal carcinomas, 27 invasive lobular carcinomas, 24 invasive mixed carcinomas and eight invasive breast carcinomas of other special types). These samples were obtained from patients primarily treated with therapeutic surgery (69 mastectomies and 155 wide local excisions) at the The Royal Marsden Hospital, London, UK. Patients in this series were selected on the basis of having received standard anthracycline-based adjuvant chemotherapy. Adjuvant endocrine therapy was prescribed for all patients with ER-positive tumours (tamoxifen alone in 96.4% of the patients for the available follow-up period). Follow-up data were available for 244 patients, ranging from 0.5 to 125 months (median 67 months, mean 67 months). Patient information, pathological characteristics of the tumours, detailed TMA preparation and the expression of a number of biomarkers have been reported previously19 20 and they are summarised in table 1. This study has been approved by The Royal Marsden Hospital Research Ethics Committee.

Table 1 Description of the patients, and characteristics of their tumours

This TMA has previously been subjected to IHC analysis to evaluate expression of ER, progesterone receptor (PR), HER2/neu, CK5/6, CK14, CK17, p53, epidermal growth factor receptor (EGFR), MIB1, topoisomerase IIα and cyclin D1. Details of antibodies used, antibody concentrations and antigen retrieval methods have been reported earlier.2022

Immunohistochemical staining for FOXA1

Expression of FOXA1 was analysed using goat anti-human FOXA1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) by IHC, as previously described.18 After de-waxing and hydration, 4 μm sections from formalin-fixed paraffin-embedded tissue were treated with 0.05% citraconic anhydride, pH 7.40, in a decloaking chamber (BioCare, Walnut Creek, CA, USA), with chamber settings of SP1: 98°C for 45 min. The slides were then cooled for 20 min at room temperature. Endogenous peroxidase activity was blocked by Peroxo-block (Invitrogen, Carlsbad, CA, USA) for 2 min. The slides were then incubated with polyclonal goat FOXA1 antibody (1:200; Santa Cruz Biotechnology) for 1 h at room temperature. The sections were incubated with anti-goat horseradish peroxidase polymer conjugate (Biocompare, San Francisco, CA, USA) according to the manufacturer’s instructions. The stain was developed using DAB plus (Dako, Glostrup, Denmark) and haematoxylin QS (Vector Laboratories, Burlingame, CA, USA) counterstain. To verify the specificity of staining, non-immune goat serum and PBS negative controls were used. A previously described semi-quantitative scoring system was used;18 briefly, percentage of staining was categorised as 0 if there was no nuclear expression, 1 for up to 10% positive tumour nuclei, 2 for 11–20%, and so on to a maximum score of 10. Intensity was scored as 1+, 2+ and 3+ for weak, moderate and strong staining respectively. Percentage (P) and intensity (I) of nuclear expression were multiplied to generate numerical score (S = P × I). Based on our previous work, scores of 0–3 were defined as a low level of FOXA1 expression (FOXA1neg), and 4–30 were defined as a high level of FOXA1 expression (FOXA1pos).

Statistical methods

FOXA1 expression correlation with tumour size, histological type, tumour grade, nodal status, lymphovascular invasion (LVI), expression of ER, PR, HER2, proliferation markers such as cyclin D1, MIB1, topoisomerase IIα, basal subtype markers like CK5/6, CK14, EGFR, CK17, p53, and breast cancer subtypes such as luminal7 (defined as ER positive and HER2 negative), basal, and HER2 was tested. Information regarding each of these parameters was not available for all patients; the number of patients for whom information was available is given parameter-wise in table 2. Kaplan–Meier analysis was performed using log-rank test for comparison of linear trends with overall survival (OS) as primary endpoint. Cox proportional hazard ratio model was used for the multivariate survival analysis. We used univariate Cox regression to calculate hazard ratios (HR) for the subgroups. All tests were two-sided. We used the 5% alpha-level to determine significance. Statview 5.0 (SAS Institute, Cary, NC, USA) statistical package was used for all statistical analyses.

Table 2 Correlations between forkhead box A1 (FOXA1) expression, clinicopathological parameters and immunohistochemical markers in 245 invasive breast carcinomas


FOXA1 expression in breast cancer

Representative FOXA1 immunostaining of breast cancers is shown in fig 1. FOXA1 expression was restricted to the nucleus with negligible cytoplasmic staining. Tumours showed variable FOXA1 expression: none (24.5%), weak (10.3%), moderate (32.6%) and strong (32.6%). Staining was nuclear in normal and cancerous tissues. A high level of FOXA1 expression (FOXA1pos; score greater than 3) was seen in 139 of 184 (75.5%) interpretable breast cancers.

Figure 1 Immunohistochemical analysis of forkhead box A1 (FOXA1) in breast cancer. Invasive carcinomas show three distinct patterns of expression: no expression (A), weak expression (B) and strong expression (C).

Correlation of FOXA1 expression with other disease markers

FOXA1 expression was compared with histological grade, nodal status, LVI, expression of ER, PR, HER2, p53, cyclin D1, MIB1, topoisomerase IIα, basal subtype markers like CK5/6, CK14, EGFR, CK17, and luminal subtype (table 2). FOXA1 expression correlated positively with ERα (p<0.0001), PR (p<0.0001), and luminal subtype (p<0.0001) and negatively with basal subtype (p<0.0001). FOXA1 expression showed an inverse correlation with high histological grade (p = 0.0327) and proliferation as defined by MIB1 (Ki67, p<0.0001). In addition, a direct correlation between the expression of FOXA1 and cyclin D1 (p = 0.0012), a gene whose transcription is activated by ER,23 was found. No correlation between FOXA1 expression and tumour size, histological type, nodal status, LVI, topoisomerase IIα or HER2/neu expression was found.

Survival analysis

Univariate analysis

Univariate analysis showed (table 3, fig 2) nodal status, tumour grade, ER, PR, FOXA1 (HR 0.3593, 95% confidence interval (CI) 0.1723 to 0.7494), basal markers including p53, MIB1, cyclin D1, and molecular subtypes as significant predictors of overall survival. FOXA1 was significant predictor of metastasis-free survival (p = 0.0185) and did not reach statistical significance (p = 0.0543) for disease-free-survival endpoint (fig 3).

Figure 2 Kaplan–Meier estimates of overall survival: all patients (p = 0.0044). FOXA1pos, high level of FOXA1 expression (score greater than 3); FOXA1neg, low level of FOXA1 expression (score 3 or less).
Figure 3 Kaplan–Meier estimates of disease-free survival: all patients (p = 0.0543). FOXA1pos, high level of FOXA1 expression (score greater than 3); FOXA1neg, low level of FOXA1 expression (score 3 or less).
Table 3 Univariate analysis of various parameters (overall survival)

Multivariate analysis

Multivariate analysis was performed using a Cox regression model, which included nodal status, grade, ER, FOXA1 and HER2/neu as variables (n  = 174, patients with complete data available on all included parameters). It showed nodal status (HR = 6.38, 95% CI 2.13 to 19.15, p = 0.0009) and ER (HR = 0.21, 95% CI 0.08 to 0.57) to be the significant predictors of OS, while FOXA1 (HR = 0.82, 95% CI 0.30 to 02.19; p = 0.6857 not significant), grade, and HER2 did not significantly affect survival.

Subgroup analysis

We performed a luminal subtype subgroup analysis (n = 130). In this subgroup, FOXA1 failed to show any correlation with tumour size, histological grade, LVI, and lymph node metastasis. In addition, there was no correlation with proliferation as defined by MIB1 and topoisomerase IIα, or with cyclin D1 expression (table 4). Luminal subtype patients expressing high FOXA1 had significantly better survival with relative risk of death being 0.28 (p = 0.0284, fig 4). FOXA1neg and FOXA1pos patients in this subgroup had 5-year OS of 83.3% and 93.7% respectively.

Figure 4 Kaplan–Meier Analysis of overall survival: luminal subtype patients (p = 0.0284). FOXA1pos, high level of FOXA1 expression (score greater than 3); FOXA1neg, low level of FOXA1 expression (score 3 or less).
Table 4 Correlations between forkhead box A1 (FOXA1) expression, clinicopathological parameters and immunohistochemical markers in luminal* breast carcinomas carcinomas

Take-home messages

  • Forkhead box A1 (FOXA1) is a transcription factor expressed in ∼75% of breast cancers that is essential for optimal expression of ∼50% of oestrogen-receptor-related genes.

  • FOXA1 expression is directly correlated with luminal subtype and inversely associated with basal-like subtype.

  • FOXA1 is a predictor of good outcome in univariate analysis and a good prognosis indicator for luminal subtype cancers.

  • FOXA1 may be an immunohistochemical marker for the luminal A subgroup of breast cancers.


The pattern of expression of FOXA1 in normal breast and tumours is strikingly similar to that of ERα.18 FOXA1 is a transcription factor with “forkhead” or “winged helix” as its DNA-binding domain, which causes recruitment of ERα to the genome.10 This winged helix region of FOXA1 is very similar to histones H1 and H5 except for the basic amino acids required for chromatin compaction and hence binding of FOXA1 to chromatin results in de-compaction.24 25 This opening of heterochromatin permits efficient interaction of ERα with promoters of various ERα−regulated genes. Depletion of FOXA1 protein in MCF-7 breast cancer cells causes reduced oestrogen-dependent gene expression and proliferation.13 14 Consistent with this possibility, about half of oestrogen-regulated genes on chromosomes 21 and 22 contain binding sites for FOXA1.13 FOXA1, GATA-3 and ERα form a transcriptional circuitry required for growth, differentiation and hormone dependency of the lineage of luminal cells co-expressing these factors.16 26

FOXA1−/− mice are viable for only 2 weeks, and prostate development studies27 in these animals show a severely altered ductal pattern lacking differentiated or mature luminal epithelial cells. It has been suggested that FOXA1, in concert with the androgen receptor, is involved in differentiation of prostate epithelium.27 Although the functional role of FOXA1 in normal breast is yet to be determined, by analogy, FOXA1 in concert with ERα may be involved in mammary ductal morphogenesis and differentiation. It has been demonstrated that ERα-expressing cells control proliferation in stromal and ERα-negative luminal cells through paracrine mechanisms.2830 Hence, it is plausible that ERα/FOXA1-expressing cells, after acquiring tumorigenicity, may promote selective clonal expansion of these cells resulting in a specific subtype of breast cancer, luminal subtype A.

Our previous study18 had some limitations as described earlier; therefore, one of the aims of the present study was to validate the cut-off for FOXA1 positivity in a completely independent set of patients who were treated using single standard protocol. Furthermore, as a secondary aim, we sought to determine the correlations between FOXA1 expression with that of basal and proliferation markers in breast cancer. Consistent with our earlier findings, in this study, we observed FOXA1 to be a positive prognostic factor in this set of breast cancer patients (HR = 0.36, univariate analysis). FOXA1 expression positively correlated with ER, PR, cyclin D1 and luminal subtype breast cancers (p<0.0001). This positive correlation with oestrogen responsive genes such as PR and cyclin D1 further supports the in-concert action of ER and FOXA1. FOXA1 correlated negatively with high tumour grade, EGFR, p53, MIB1 and basal markers such as CK5/6, CK14 and CK17. It did not correlate with tumour size, nodal status or HER2 expression. We have previously studied FOXA1 correlations with tumour size, nodal status, ER, PR, HER2 and luminal subtype; these findings are consistently reproduced in the present study. This validation of FOXA1 expression by a clinically reproducible method (IHC) adds robustness to the possibility that such expression analysis can provide a clinically useful prognostic factor especially in low-risk breast cancer patients.

Luminal subtype breast cancers are good prognosis tumours.5 6 FOXA1 expression in luminal subtype breast cancers patients further sub-classifies this group in high and low-risk groups. FOXA1neg and FOXA1pos patients in this subgroup had 5-year survival of 83.3% and 93.7% respectively; it should be noted here that almost half of the patients in our study had T2 or larger tumours. In agreement with expression profiling studies,5 8 FOXA1 expression subdivided tumours of luminal profile into two prognostically significant groups. In fact, our results suggest that FOXA1 IHC may be used as a marker for tumours pertaining to the luminal subtype A breast cancer, which has an exceptionally good prognosis. Studies comparing in parallel FOXA1 IHC evaluation and gene-expression profiling to confirm this hypothesis are warranted.

In the present study, we also explored whether good prognosis associated with FOXA1 expression in luminal subgroup is merely due to low proliferation than rather having a mechanistic role. We observed that FOXA1 expression in luminal subgroup did not correlate with proliferation markers such as MIB1 and cyclin D1, pointing to a mechanistic role. It should however be noted here that this subgroup analysis lacked sufficient statistical power.

Adjuvant treatment decisions, especially decisions such as whether to give chemotherapy or not in the ER-positive lymph-node-negative patient subgroup are difficult to make. A prognostic marker that can further identify patients with very low risk of recurrence will be very useful to avoid toxic therapies in patients who are not likely to derive any additional benefits. FOXA1 expression showed a trend towards better survival in this subgroup.18 In the present study, this analysis was not possible due to the limited sample size of this subgroup. Based on the similar findings pertaining to low-risk subgroups in these two studies, it may be possible in future to use FOXA1 as a prognostic marker in low-risk patients.

To conclude, the present study demonstrates the prognostic role of FOXA1 expression by IHC in patients treated with current regimens of adjuvant chemotherapy. FOXA1 expression strongly correlates with ER expression and luminal subtype of breast cancer, and divides luminal subtype tumours into two distinct prognostic groups: tumours expressing FOXA1 have significantly better survival and are most likely to be luminal subtype A tumours. FOXA1 is a promising candidate for identification of luminal subtype A tumours in the clinic.


View Abstract


  • Funding: This study was funded in part by Breakthrough Breast Cancer.

  • Competing interests: None.

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.