Expression of the glucocorticoid receptor in breast cancer-associated fibroblasts

Abstract

Cancer- associated fibroblasts (CAFs) are actively involved in breast carcinoma. Our previous study demonstrated that the majority of these CAFs were smooth muscle actin (SMA) positive and were therefore termed peritumoral myofibroblast (PMY). Glucocorticoid, linked or not with its receptor (GR), has been postulated to serve a major role in normal breast and breast carcinoma; however, their role in CAFs remains poorly understood. The aim of the present study was to assess the presence of GR in breast CAFs and particularly in PMY in 56 cases of invasive breast carcinoma in correlation with clinicopathological parameters, by immunohistochemistry. GR was observed in CAFs in 51 cases (91%) and were more frequent in luminal A subtype (19/19 cases; 100%). The stromal expression was statistically correlated with the tumor grade (P=0.03), the Ki-67 index (P=0.003) and the presence of GR in the epithelial component (P=0.01). The demonstration of a frequent expression of GR in breast CAFs may serve as an interesting target for future therapeutics for the regulation of the tumoral breast microenvironment.

Introduction

Glucocorticoids (GCs) are essential for survival and serve a major role in embryonic development, tissue homeostasis and in the regulation of the inflammatory response (1,2). In breast, the functions of GCs are complex and depend, in part, if they are linked to their receptor (GR) and consist of the control of milk secretion, differentiation and apoptosis (3). Morphologically, previous studies have demonstrated that GR nuclear expression is observed both in normal breast, in situ carcinoma and less frequently in invasive carcinoma (4–7). In invasive tumors, their expression is limited in tumors with a small size, low grade, good prognosis and expressing estrogen receptor (ER) (4,7). In addition, our previous study clearly demonstrated that dexamethasone has an antiproliferative effect in the MCF-7 breast cells that express GR (8).

Over previous years, the tumor-associated stroma and, in particular, the cancer associated fibroblasts (CAFs) have been demonstrated to serve a crucial role in cancer pathogenesis (9,10).

Our previous study clearly demonstrated that the majority of these CAFs were smooth muscle actin (SMA)-positive with a myofibroblastic-like phenotype and that the presence of these peritumoral myofibroblasts (PMY) is important both in situ and in invasive breast carcinoma of no special type (11). This is also important in metastatic disease by promoting tumor invasion, growth and angiogenesis through paracrine factors and/or direct cell-cell crosstalk (11–13).

Our previous study demonstrated the presence of ER or progesterone receptors (PR); however, the presence and potential role of GR is poorly understood in breast carcinoma PMYs (4).

Therefore, the present study aimed to assess, by immunohistochemistry, the presence or absence of GR in breast CAFs and in CAFs smooth muscle positive/PMY in correlation with clinicopathological variables. Investigating this may assist with elucidating the role of GR in breast carcinoma.

Materials and methods

Patient selection

Breast tissue samples were retrieved from the Departments of Pathology at the Erasme Hospital and IRIS South Hospital (Brussels, Belgium), and consisted of 56 cases of invasive carcinoma. The present study was approved by the Ethics Committee from Erasme University Hospital (no. P2014/418).

Immunohistochemistry

The immunohistochemical assessment of ER, PR, Ki-67 and human epidermal growth factor receptor (HER)2 was routinely performed using an antigen retrieval method using the Leica BOND-III fully automated system (Leica Microsystems, Ltd., Newcastle, UK), as previously described (4). According to these parameters, carcinoma were divided into five groups, as previously described (14): Luminal A (n=19), Luminal B (n=12), HER2+/ER+ (n=7), HER2+/ER- (n=9) and triple negative (n=9). In addition, the following parameters were also included for each patient: Age, stage, tumoral size and lymph node status. All parameters are shown in the Table I.

Table I.

Association of clinicopathological characteristics with immunohistochemical levels of GR in the peritumoral stroma.

		GT expression in the stroma
Characteristic	No. cases (%)	Strong	Weak	Negative	P-value
Age, years					0.56
≤50	24 (43)	18	5	1
>50	32 (57)	23	5	4
Tumor size, mm					1
<20	28 (50)	21	5	2
≥20	28 (50)	20	5	3
Stage					0.81
T1	28 (50)	21	5	2
T2	21 (37.5)	15	3	3
T3	7 (12.5)	5	2	0
Tumor grade					0.03
Grade 1	10 (17.9)	10	0	0
Grade 2	22 (39.2)	12	8	2
Grade 3	24 (42.9)	19	2	3
Lymph node status					0.19
Negative	29 (51.8)	24	4	1
Positive	27 (48.2)	17	6	4
ER status					0.64
Negative	18 (32.1)	14	2	2
Positive	38 (67.9)	27	8	3
PR status					0.39
Negative	23 (41.1)	15	6	2
Positive	33 (58.9)	26	4	3
Ki-67 index, %					0.003
≤14	19 (33.9)	19	0	0
>14	37 (66.1)	22	10	5
HER 2 status					0.2
Negative	40 (71.4)	31	7	2
Positive	16 (28.6)	10	3	3
Intrinsic subtype					0.189
Luminal A	19 (33.9)	19	0	0
Luminal B/HER2−	12 (21.4)	5	5	2
Luminal B/HER2+	7 (12.5)	4	2	1
HER2+	9 (16.1)	6	1	2
Triple negative	9 (16.1)	7	2	0
GR status in glands					0.01
Negative	26 (46.4)	15	6	5
Positive	30 (53.6)	26	4	0

P-values in bold are statistically significant (P<0.05). ER, estrogen receptor; PR, progesterone receptor; HER2, human epidermal growth factor receptor; GR, glucocorticoid receptor.

For the demonstration of GR, a manual technique was applied. Tissue sections (4 µm) were cut sequentially and mounted onto superfrost-treated slides (Menzel-Gläser, Braunschweig, Germany). The slides were dried overnight at 37°C prior to deparaffinization in xylene and rehydration through graded ethanols. For epitope retrieval, the slides were immersed in a waterbath at 95–99°C for 90 min with an ethylenediamine tetraacetic acid buffer (pH 9.0; S236; Dako Corp., Glostrup, Denmark). The slides were subsequently cooled in the buffer for 20 min at room temperature. H₂O₂ (0.3%) was subsequently added to the slides and incubated for 30 min. The tissues were then incubated for 1 h with a monoclonal antibody against the N-terminus of the GR (clone 4H2; cat. no. NCL-GCR; 1:25; Novocastra Laboratories, Newcastle, UK) (4).

Double immunostaining

In addition, for the specific visualization of the expression of ER/PR and GR in SMA-positive CAFs, a double stain was also performed by using the EnVision G/2 double stain system (Dako Corp.), as previously described (15). The same monoclonal antibodies (ER, PR and GR) described above were applied to rehydrated paraffin tissue sections and allowed to incubate for 1 h at room temperature. Endogenous peroxidase was inhibited and 3,3′-diaminobenzidine was used to visualize the binding of these primary antibodies. The sections were subsequently incubated for 1 h with a secondary antibody against SMA (clone αSM-1; 1:50; Novocastra Laboratories, Newcastle, UK). Alkaline phosphatase-conjugated secondary antibody and fuchsin as substrate chromogen system were used to complete the secondary immunostain. Negative controls used the replacement of the different primary antibodies with the corresponding isotypes. In addition, to ensure the absence of PMY in the normal breast, 10 cases of normal breast tissue from patients who underwent plastic surgery were also included.

Immunohistochemical evaluation

All the slides were examined by two independent observers (Xavier Catteau and Jean-Christophe Noël) and the evaluation of ER, PR and GR was made independently by the two pathologists using the Allred score (16), and estimated the proportion of positive CAFs (0, no positive cells; 1, ≤1; 2, 1–10; 3, 11–33; 4, 34–66; 5, 67–100% positive cells) and the average staining intensity (0, negative; 1, weak; 2, Intermediate; 3, strong). The proportion score and the intensity score were added to obtain a total score ranging from 0–8. Subsequently, three grades of immunoreactivity were established: Score 0–2, negative; score 3–4, weak positivity; score 5–8, strong positivity.

Statistical analysis

The correlation analysis was performed. χ²-test and Fisher's exact test were used. All statistical analyses were performed using XLSTAT software (Addinsoft, Paris, France). P<0.05 was considered to indicate a statistically significant difference.

Results

Weak or strong expression of GR in CAFs was observed in 10 cases (18%) and 41 cases (73%), respectively. A total of 5 cases were negative (5%; Table I). The stromal expression was frequent in luminal A tumor (100% of cases; Fig. 1A) and is statistically correlated with the tumor grade (P=0.03), the Ki-67 index (P=0.003) and GR status in glandular/carcinomatous component (P=0.01); however, was not correlated with age, tumor size, lymph node status and the expression of ER or PR, at least with a positive status for the latest as ≥1% in accordance with the World Health Organization recommendations (Table I). The double stain immunohistochemistry confirming unequivocally that among these CAFs, SMA-positive PMY clearly showed a nuclear staining of GR (Fig. 1B).

Figure 1.

Immunohistochemical staining of CAFs. (A) A typical example of glucocorticoid receptor in CAFs (arrows). Note the strong immunoreactivity score of these cells by comparison with the epithelial carcinomatous component, which is negative (magnification, ×10). (B) The double staining immunohistochemistry confirmed the positivity of these nuclear receptors (brown) in α-smooth muscle actin positive CAFs/peritumoral myofibroblats stained in red (arrow) (magnification, ×40). The immunoreactivity for glucocorticoid receptor is moderate. CAF, cancer associated fibroblast.

Discussion

For numerous years, the majority of studies in breast carcinoma have been focused predominantly on the epithelial component; however, recently CAFs and in particular CAFs SMA-positive PMY have been demonstrated to serve an important role in cancer pathogenesis as a result of paracrine cross-interaction between these and epithelial cancer cells. Indeed CAFs/PMY are able to secrete various factors implicated in invasion, matrix remodeling, cell proliferation, differentiation and apoptosis. In breast carcinoma, the hormonal regulation of epithelial cells is well documented and is the result of interaction between estrogen and progesterone, and their respective receptors, at least in hormone-dependent tumor types (16,17). The role of GCs in normal breast is more controversial and likely depends on the balance between targets of linked and non-linked GR with opposing functions: linked GR being involved in maintaining functional differentiation and non-linked GR appearing to be proapoptotic (3,6). In breast carcinoma, the GCs acting through their nuclear receptors are considered as a potential tumor suppressor promoting accurate chromosome segregation during mitosis occurring tumoral cell division (6,18,19). Indeed, according to these data, our previous study demonstrated an antiproliferative effect of dexamethosone in the MCF-7 breast cancer cells line that contains nuclear GR (8). The underlying mechanisms of hormonal regulation of CAFs/PMY remain to be determined; however, our previous study and other previous studies have clearly demonstrated that ER and PR were not present in these cells.

By contrast, the present study clearly demonstrated the presence of a marked GC nuclear immunoreactivity of CAFs in 73% of cases. In addition, this strong immunoreactivity was demonstrated by double labeling in CAFs SMA-positive PMY (Fig. 1) for the first time, to the best of our knowledge. This strong CAFs immunoreactivity was typically more frequent in luminal A (100%) compared with in other subtypes. In addition, it appeared to be correlated with different conventional clinicopathological parameters, including the grade (P=0.03) and Ki-67 index (P=0.003). The CR expression in CAFs was also more frequent when these receptors were present in the carcinomatous counterpart (P=0.01). These data suggested that, as previously shown for matrix metalloproteinase-2, the characteristics and properties of CAFs present in breast carcinoma microenvironment are probably complex and different from one subtype to another (10,20–22).

GCs are considered as agents capable of regulating the proliferation of myofibroblasts in different pathologies where they serve a major role as wound healing or asthma (23,24). Previously, in a myofibroblast cell line associated with colonic carcinoma, dexamethasone inhibited the expression of different classical procarcinogenic factors, including tenascin C, hepatocyte growth factor and transforming growth factor-β, in a receptor-dependent manner (25).

The present data are interesting since over the last few years it appears that in addition to the development concerning the classical therapies (hormone therapies, chemotherapies and immunotherapies), the peritumoral stroma served as potential target therapy in various carcinomas (9,24,25). Finally, even if it remains hypothetical, it has been postulated that the stress through GCs can be a promoting agent in breast cancer (26,27). From this point of view, demonstrating the presence of GR in the CAFs may be important.

The demonstration of a frequent expression of GR in breast CAFs may serve as an interesting target for future therapy in the regulation of the tumoral breast microenvironment. Naturally, future research is required, firstly to establish with larger cohorts the assocaition between the presence of GR in CAFs and the overall survival, and to understand how a therapy may influence the CAF associated with breast carcinoma. Such investigations are in progress.