Steroid hormone receptor expression in breast cancer stroma

Abstract

Around two-thirds of all breast carcinomas respond to adjuvant endocrine therapy with the presence of ERa in tumour epithelial cells used to guide therapy. Other members of the steroid receptor superfamily have also been detected in breast cancer epithelial cells, notably PR, AR, whose presence and potential prognostic role in breast cancer is gaining momentum [1, 2], and there is ongoing interest in the potential role of ERb in breast carcinogenesis [3]. However, scientists are now looking beyond the epithelial landscape with the recognition that interaction of tumour epithelial cells with its surrounding microenvironment, particularly the stroma plays a role in promoting tumour growth and progression. While recognising that the stroma is multicellular, much of this attention has focused on fibroblasts residing in tumour stroma, termed carcinoma-associated fibroblasts (CAFs) as there is growing appreciation that CAFs may provide a favourable environment to augment tumour growth [4]. A recent report in Breast Cancer Research and Treatment by Knower et al. [5] used low density Taqman arrays followed by qPCR and immunohistochemical validation to examine the presence of a panel of 48 nuclear receptors in CAFs, derived from ERa-positive breast carcinomas and in normal breast adipose fibroblasts (NAFs). Receptor expression was grouped as absent, low, moderate or high. Intriguingly, of the four nuclear hormone receptors previously associated with breast carcinogenesis listed above, none were identified as being highly expressed in breast stroma; ERb was absent, ERa and PR showed low expression, while moderate expression of AR was seen. As a group with a major interest in the role of ERs in breast carcinogenesis, we have routinely monitored nonepithelial expression of ERs in clinical (ERa) and research settings (ERb) over the years. In contrast to Knower et al. [5], we readily detect ERa in cultured breast fibroblasts from adjacent normal and cancer-associated breast tissue by qPCR. This is carried forward in observations by our group during the routine histopathological reporting of breast carcinoma, where the presence of ERa in stroma is seen in 60 % of all cases, although this can be very focal. Aware of reported issues with some ERa antibodies [6], we evaluated this further using three different antibodies directed against ERa, 1D5 (Dako), 6F11 (Novocastra) and SP1 (Ventana). Using all three antibodies four fifths of tumours expressed ERa in fibroblasts in concordant areas suggesting the signal is genuine (Fig. 1). The lack of ERa expression observed by Knower et al. [5] is thus surprising. We do agree that ERb gene expression is low, although it is usually detectable by qPCR. However, it is commonly found in breast fibroblasts by immunohistochemistry both total ERb [7] and specific isoforms (ERb1, -2 and -5) [8]. ERb is also subject to complex regulation [9], hence nonconcordance of gene and protein is anticipated. The nuclear receptors whose expression was modified in CAFs included RAR-related orphan receptor-a (ROR-a); thyroid hormone receptor-b (TR-b); vitamin D receptor (VDR); and peroxisome proliferator-activated receptor-c (PPAR-c) [5], which one may not immediately associate This letter to the editor refers to the article at 10.1007/s10549-0132716-6 and a rebuttal letter to this letter to the editor is available at 10. 1007/s10549-014-2837-6