Expression of CAF-Related Proteins Is Associated with Histologic Grade of Breast Phyllodes Tumor

Purpose. The purpose of this study was to investigate the expression of cancer-associated fibroblast- (CAF-) related proteins and the implications in breast phyllodes tumor (PT). Methods. Tissue microarrays of 194 PT cases (151 benign PT, 27 borderline PT, and 16 malignant PT) were constructed. We performed immunohistochemical staining for CAF-related proteins (podoplanin, prolyl 4-hydroxylase, FAPα, S100A4, PDGFR α/β, and NG2) and analyzed the results according to clinicopathologic parameters. Results. Expression of PDGFRα and PDGFRβ in the stromal component increased with increasing histologic grade of PT (p = 0.003 and p = 0.034, resp.). Among clinicopathologic parameters, only expression of FAPα in stroma was associated with distant metastasis (p = 0.002). In univariate analysis, stromal expression of PDGFRα was associated with shorter overall survival (p = 0.002). In Cox multivariate analysis, stromal overgrowth and PDGFRα stromal positivity were associated with shorter overall survival (p = 0.006 and p = 0.050, resp.). Furthermore, expression of PDGFRβ in stroma was associated with shorter overall survival in patients with malignant PT (p = 0.041). Conclusion. Stromal expression of PDGFRα and PDGFRβ increased with increasing histologic grade of PT. In addition, PDGFR stromal positivity was associated with shorter overall survival. These results suggest that CAFs are associated with breast PT progression.


Introduction
Progress in cancer research has increasingly revealed the clinical significance of the tumor microenvironment. Among various components of the tumor microenvironment, cancerassociated fibroblasts (CAFs), one of the most important elements, have been widely studied [1]. CAFs are located near the cancer cells and have been reported to be involved in tumor initiation, tumor-stimulatory inflammation, metabolism, metastasis, drug response, and immune surveillance [2]. Despite their significant effect on cancer cells, the exact cell origin of CAFs is not completely understood and there is even controversy concerning the definition of CAF [1,2]. Various proteins have been suggested as markers for CAFs, including -smooth muscle actin (SMA) [3], tenascin-C [4], chondroitin sulfate proteoglycan (NG2) [5], platelet-derived growth factor receptor (PDGFR) / [6], fibroblast activation protein (FAP) [7], podoplanin [8], prolyl 4-hydroxylase [9], and fibroblast-specific protein-(FSP-) 1 [5]. CAFs have been suggested to show various functional subtypes that exhibit different characteristics [10], supporting the hypothesis that CAFs have various phenotypes.
Phyllodes tumor (PT) is a relatively rare biphasic breast tumor that accounts for only 0.3-1.5% of all breast tumors [11]. It is composed of atypical spindle cell stroma and usually benign epithelium [12]. Pathologically, differential diagnosis is difficult because several histologic findings of PT overlap with those of fibroadenoma, which is also a fibroepithelial tumor, and PT shows heterogeneous histologic features within the tumor [11,13]. In addition, some cases of PT show several clinical features of malignancy such as relapse or distant metastasis [14]. The histologic classification of PT varies according to authors, but PT has been classified as benign, borderline, and malignant according to the World Health Organization classification of tumors of the breast [11]. A higher histologic grade of PT is associated with increased risk of tumor recurrence or distant metastasis, indications of tumor aggressiveness. Furthermore, among malignant PT, 9-36% of cases experience local recurrence, 9-40% of cases progress with metastasis to the lungs, brain, or liver [15],  and some cases suffer death due to recurrence and distant metastasis [16,17]. The fibroblast is an important component of PT [18,19]; therefore, variations in tumor biology of PT could be accounted for by the phenotype of associated fibroblasts. However, the role of CAF-related proteins in breast PT is poorly understood. The purpose of this study was to investigate the expression of CAF-related proteins and its implications in breast PT.

Patient Selection.
Tissue samples from patients who were diagnosed with breast PT from 2000 to 2010 in Severance hospital were selected. All tissues were fixed in 10% buffered formalin and embedded in paraffin. Archival hematoxylin and eosin-(H&E-) stained slides for each case were reviewed by a single pathologist (Ja Seung Koo). The PT histologic grade was based on the World Health Organization classification of tumors of the breast [11]. The histologic grade was evaluated with H&E-stained slides and clinical data including patient age, recurrence, distant metastasis, and patient survival were obtained by reviewing the patients' medical records. The study was approved by the Institutional Review Board of Yonsei University Severance Hospital.

Tissue Microarray.
Construction of tissue microarray was performed as previously described. Briefly, a representative area was selected on the H&E-stained slide of the tumor and a corresponding spot was marked on the surface of the paraffin block. Using a biopsy needle, the selected area was punched out and the 5 mm tissue core was placed in a 5 × 6 recipient block. To minimize extraction bias, we extracted two tissue cores for every PT case. Finally, each separate tissue core was assigned a unique tissue microarray location number linked to the database including other clinicopathologic data and was used for immunohistochemical staining.
2.3. Immunohistochemistry. The antibodies used for immunohistochemical staining in this study are shown in Table 1. All immunostaining was performed using the constructed tissue microarray. For immunohistochemistry, 5 m sections were obtained with a microtome, transferred onto adhesive slides, and dried at 62 ∘ C for 30 min. After incubation with primary antibodies, immunodetection was performed with biotinylated antimouse immunoglobulin followed by peroxidaselabeled streptavidin using a labeled streptavidin biotin kit with 3,3 -diaminobenzidine chromogen as substrate.
For the negative control, the primary antibody incubation step was omitted. Harris hematoxylin was used for tissue counterstaining. All immunohistochemical markers were assessed using light microscopy and evaluation of staining was performed by calculating the proportion of stained cells and immunostaining intensity. The proportion of stained cells was defined as follows: 0, negative; 1, less than 30% positive; and 2, more than 30% positive. The immunostaining intensity was defined as follows: 0, negative; 1, weak; 2, moderate; and 3, strong. The scores for proportion of stained cells and immunostaining intensity were multiplied, and staining was defined as positive when the final score was >1 [20].

Statistical
Analysis. Data were analyzed using SPSS for Windows, Version 21.0 (SPSS Inc., Chicago, IL, USA). Continuous variables were compared using two-tailed Students'test and categorical data were compared using the Chi square test. To evaluate the time to tumor recurrence and compare the survivals between groups, Kaplan-Meier survival curves and the log-rank test were used. Multivariate survival analyses using a Cox's proportional hazard model were performed to characterize the prognostic factors in PT. A two-tailed value < 0.05 was considered statistically significant.

Basal Characteristics of Phyllodes Tumors.
This study included 194 cases of breast PT, including 151 cases of benign PT, 27 cases of borderline PT, and 16 cases of malignant PT. The basal characteristics of the patients are shown in Table 2.
Greater patient age and larger tumor size were associated with higher PT histologic grade ( = 0.017 and = 0.001, resp.). Higher rate of tumor recurrence and distant metastasis were also associated with higher PT histologic grade ( < 0.001).
Eight cases with distant metastasis showed lung metastasis.

Expression of CAF-Related Proteins according to Phyllodes
Tumor Histologic Grade. There were no significant differences in immunohistochemical staining in the epithelial component according to PT histologic grade for any of the proteins analyzed (Table 3). For the stromal component, the expression of PDGFR and PDGFR increased with increasing PT histologic grade ( = 0.003 and = 0.034, resp.) ( Table 4 and Figure 1).

Correlation between the Expression of CAF-Related Proteins in Phyllodes Tumor and Clinicopathologic Parameters.
We investigated the correlation between the expression of CAF-related proteins in PT and clinicopathologic parameters. Only expression of FAP in the stromal cells was associated with distant metastasis ( = 0.002) (Figure 2). Other clinicopathologic parameters, including age, tumor size, stromal cellularity, stromal atypia, stromal mitosis, stromal overgrowth, tumor margin, and tumor recurrence, were not associated with the expression of CAF-related proteins in PT.

Impact of Expression of CAF-Related Proteins on Patient
Prognosis. Univariate analysis showed that stromal expression of PDGFR was associated with shorter overall survival ( = 0.002) ( Table 5 and Figure 3(a) (Table 6). Furthermore, stromal expression of PDGFR ( = 0.052) (Figure 3(b)) and PDGFR ( = 0.041) (Figure 3(c)) was associated with shorter overall survival in malignant PT, although statistical significance was not reached regarding PDGFR expression.

Discussion
In this study, we investigated the expression of CAF-related proteins in breast PT according to histologic grade with the aim of identifying a new therapeutic target for PT.
Recently, there have been many studies on the tumor microenvironment as a novel therapeutic target. The tumor microenvironment includes nontumor cells with nontransformed elements, including immune system elements (such as macrophages and lymphocytes), blood vessel cells, fibroblasts, myofibroblasts, mesenchymal stem cells, adipocytes, and extracellular matrix, in close proximity to tumor cells. PT is a representative fibroepithelial tumor characterized by fibrous stroma composed of fibroblasts surrounding the epithelium. Previous studies reported that CD34-expressing fibroblasts exist in mammary stroma. Because CD34expressing fibroblasts were observed in both fibroadenoma 4 Disease Markers  and PT [19], efficient treatment for fibroepithelial tumors might be achieved by targeting the fibrous stroma. However, there are no studies on the expression of CAF-related proteins in PT, and as PT shows heterogeneous tumor stromal features, the expression of CAF-related proteins is predicted to vary in each type. In this study, we performed immunohistochemical staining for CAF-related proteins focusing on the stromal component. Our results showed that the expression of PDGFR and PDGFR increased with increasing PT histologic grade. Platelet-derived growth factor is a major mitogen for several cell types, including connective tissue cells, that is activated by binding to two protein tyrosine kinase receptors (PDGFR and PDGFR ). PDGF signaling in tumor cells induces point mutations, amplification, and translocations, which stimulate autocrine stimulatory loops [6,21]. For example, PDGFR and are two well-known receptors that participate in breast cancer progression [22]. In addition, in breast cancer, the desmoplastic response appears to be mediated by PDGF-AA signaling in PDGFR type CAFs [23]. Desmoplasia refers to the growth of fibrous or connective tissue near the tumor and is increased in tumors with aggressive properties. The results from our study revealed that higher grade PT with more active fibroblasts showed higher expression of PDGFR and , suggesting that fibroblasts with a CAF phenotype are associated with PT progression.
Stromal PDGFR expression was related to the prognosis of patients with PT. Furthermore, subgroup analysis in malignant PT showed that stromal PDGFR expression was associated with shorter overall survival and, although statistical significance was not noted, there was also a trend for increased stromal PDGFR expression. In previous studies of glioma [24], squamous cell carcinoma of the head and neck [25], colorectal cancer [26], pancreatic cancer [27], and T cell lymphoma [28], PDGFR activation induced the intracellular signaling pathway and promoted cell migration, invasion, survival, and proliferation [29]. Similarly, PDGFR activation is related to lymphatic metastasis in pancreatic cancer [30] and gastric cancer [31], suggesting that PDGFR might be used as a prognostic marker in malignant PT. This should be validated in future studies.
We found that expression of FAP in stroma was associated with distant metastasis. Fibroblast activation protein is expressed by reactive CAFs in tumor stroma or granulation tissue and is known to be involved in wound healing. The expression of FAP in CAFs is reported in various carcinomas and is used as an important marker of CAF [32,33]. In a meta-analysis that analyzed the clinical implication of FAP overexpression in solid tumors of colorectal cancer, pancreatic adenocarcinoma, non-small cell lung cancer, breast cancer, medullary thyroid carcinoma, and oral squamous cell carcinoma, high FAP expression was related to the risk of   distant metastasis (OR: 2.56) [34], showing similar results to our study. FAP regulates proteolysis of the extracellular cell matrix in tumor stroma, causing stromal cell proliferation and invasiveness. Therefore, it could be speculated that tumors with high expression of FAP are prone to metastasis to distant organs [35,36].
The clinical significance of this study is the role of CAFs as a potential therapeutic target. Sunitinib, a US Food and Drug Administration-(FDA-) approved PDGFR inhibitor, is currently used in the treatment of advanced renal cell carcinoma [37], advanced progressive pancreatic neuroendocrine tumor [38], and advanced radioiodine refractory thyroid carcinoma [39]. Furthermore, ongoing preclinical trials are testing the therapeutic effect of another PDGFR inhibitor, imatinib mesylate, in gastrointestinal stromal tumor [40,41] and in vitro and in vivo studies of malignant peripheral nerve sheath tumors have shown promising results [42]. Therefore, inhibition of PDGFR in PT might be a potential therapeutic strategy. However, further investigation regarding the effect of these inhibitors in PT is required.
Taken together, our data show that expression of PDGFR and PDGFR in the stromal component increased with increased PT histologic grade. Also, stromal expression of PDGFR was associated with shorter overall survival. Overall, the expression of CAF-related proteins is related to the histologic grade of breast PT and PDGFR inhibitors in particular have potential as novel treatments for PT.