Discoidin Domain Receptor 2 Expression as Worse Prognostic Marker in Invasive Breast Cancer

Discoidin domain receptor 2 (DDR2) is arising as a promising therapeutic target in breast carcinoma (BC). The ability of DDR2 to bind to collagen promotes protumoral responses in cancer cells that influence the tumor microenvironment (TME). Nonetheless, the interrelation between DDR2 expression and TME modulation during BC progression remains poorly known. For this reason, we aim to evaluate the correlation between intratumoral expression of DDR2 and the infiltration of the main TME cell populations, cancer-associated fibroblasts (CAFs), and tumor-associated macrophages (TAMs). First, collagen and DDR2 expression levels were analyzed in human invasive BC samples. Then, DDR2 status correlation with tumor aggressiveness and patient survival were retrieved from different databases. Subsequently, the main pathways, cell types, and tissues correlated with DDR2 expression in BC were obtained through bioinformatics approach. Finally, we studied the association of DDR2 expression with the recruitment of CAFs and TAMs. Our findings showed that, together with the expected overexpression of TME markers, DDR2 was upregulated in tumor samples. Besides, we uncovered that altered TME markers were linked to DDR2 expression in invasive BC patients. Consequently, DDR2 modulates the stromal reaction through CAFs and TAMs infiltration and could be used as a potential worse prognostic factor in the treatment response of invasive BC.


Introduction
Breast carcinoma (BC) is the most prevalent malignancy in women and the leading cause of death among female population [1,2]. Approximately 80% of all BC cases diagnosed correspond to invasive ductal carcinoma (IDC). Invasive lobular carcinoma (ILC) represents the second most common subclass of BC, with an incidence of around 15% [3]. Both subtypes differ in clinical, pathological, and biological aspects, with the expression of E-cadherin being the main hallmark of IDC [4]. Apart from its histology, BC tumor is classified according to its molecular phenotype regarding the presence of three different biomarkers: estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2). us, luminal-like BC is characterized by ER/PR positive and HER2 negative, HER2-enriched is defined by ER/PR negative or positive and HER2 positive, and triple-negative tumor is described by the absence of these markers [5]. Current adjuvant treatments are based on targeting these proteins before or after surgery [6,7], excepting triple-negative BC, for which chemotherapy in combination with immunotherapy regimens are commonly used [8]. However, despite medical advances, more than 20% of women with early-stage BC will suffer metastatic disease [9], which reflects the need for new targeted therapies.
One of the risk factors involved in BC malignancy is the dysregulation of extracellular matrix (ECM) composition and architecture, which is associated with the menopausal status [10] and predisposes women to develop invasive BC later on [11]. Aberrant ECM remodeling is characterized by a progressive collagen enrichment or desmoplasia, which increases mammary gland stiffness and alters the tissue mechanosignaling [12]. Collagen deposition and crosslink provide biochemical and biophysical signals to enhance diverse protumoral characteristics, regarding proliferation, migration, and invasion [13]. Several proteins, such as integrins and specific tyrosine kinase receptors (RTKs), mediate the crosstalk between cancer cells and ECM components [14]. One of the receptors involved in the tumor-ECM interplay is the discoidin domain receptor 2 (DDR2), an atypical RTK, due to its capacity to bind collagens [15].
DDR2 overexpression has been linked to progression of different cancer types [16], including BC [17][18][19]. e role of DDR2 in BC development has been uncovered recently, with special emphasis on its ability to stimulate the secretion of collagen-dependent proteases by tumor cells in postpartumrelated carcinomas, activating their migratory and invasive potentials, as well as triggering metastasis [20]. Interestingly, apart from its implication on cancer cells response, DDR2 receptor also mediates key BC cells' functional features that influence the phenotype and reaction of stromal cells, particularly cancer-associated fibroblasts (CAFs) [21,22]. CAFs conform to the major component of the cellular fraction of the tumor microenvironment (TME) and actively participate in the desmoplastic process as major producers of extracellular collagen [23]. CAFs increase cancer cells growth through various mechanisms, including proliferation, migration, cytokine and chemokine secretion, and resistance to anticancer drugs [24,25]. As well as CAFs, tumor-associated macrophages (TAMs) are a very relevant stromal population during ECM reorganization and carcinogenesis, due to the secretion of several cytokines and chemokines that create a favorable microenvironment for tumor growth, promoting cancer cells proliferation, migration, and chemoresistance [26]. Although CAFs and TAMs communication with tumor cells is essential for BC progression, the role of DDR2 during this interaction remains unclear.
Given the importance of DDR2 in invasive BC process, we evaluated collagen and DDR2 status in tumor samples compared to healthy adjacent tissue, together with DDR2 implication on metastasis progression. Besides, we studied the association of DDR2 expression level and menopausal status with patient survival. Moreover, based on the capacity of DDR2 to modulate the TME, we aimed to analyze the relation between DDR2 expression and the recruitment of CAFs and TAMs into the breast tumors. Our findings revealed that collagen and DDR2 expression levels were significantly higher in BC compared to normal adjacent tissue. In addition, DDR2 upregulation was detected in metastatic tumors and represented a worse prognostic factor in combination with postmenopausal age. Moreover, DDR2 overexpression correlated with a higher CAFs and TAMs infiltration into the mammary tumor tissue. Taking these results together, DDR2 could play an important role in the modulation of the TME and may represent a potential target for antitumor early and metastatic BC treatments.

Patients and Tissue Samples.
Breast IDC samples and the corresponding adjacent nontumoral or healthy tissues were obtained by mammary resections from 25 patients (n � 25). Adjacent nontumoral samples were collected from regions at least 5 mm away from tumor boundaries and were classified as normal breast tissue by a histopathologist. Although histological normalcy was detected, we must consider that carcinogenic alterations could affect the phenotypic and genetic profiles of the tissue regions up to 1 cm from the tumor margins [27]. Additionally, clinical and pathological features including menopausal status, tumor size, histological grade, tumor stage, proliferative index, e-cadherin expression, hormonal receptors-HER2/neu status, and lymph node invasion were recorded from the medical history (Table 1). Tumor staging was established according to the International Union against Cancer Tumour-Node-Metastasis (UICC-TNM) classification. All samples were provided by Basurto University Hospital Biobank (Basque Country, Spain) under the guidelines of the Institutional Ethics Committee with the corresponding informed consent.

Masson's Trichrome Staining.
e detection of collagen type I fibers was performed through Masson's trichrome staining. In brief, 5 μm tissue slides were incubated with Weigert's iron hematoxylin solution for 10 minutes and Biebrich scarlet-acid fuchsin Solution for 15 minutes. Finally, samples were treated with phosphomolybdic acid solution for 15 minutes and aniline blue solution for 5 minutes. All staining solutions were purchased from Sigma-Aldrich, United States (cat. numbers HT1079, HT151, HT153, and B8563, respectively). Collagen fibers were observed in blue, and expression level was quantified using FIJI-ImageJ (Colour Deconvolution plugin).

Immunohistochemistry and Scoring.
e expressions of DDR2, α-SMA, and CD68 proteins were analyzed in consecutive sections of breast IDC and nontumoral samples. 5 μm tissue slides were pretreated with Antigen Retrieval Solution (R&D Systems, United States, cat. number CTS013) for 10 minutes at 90°C. Endogenous peroxidase activity was blocked using 3% hydrogen peroxide (PanReac AppliChem, United States, cat. number 131077) in phosphate-buffered saline (PBS) 1X for 30 minutes at room temperature. en, slides were incubated with 0.4% Triton X-100 ( ermo Fisher Scientific, United States, cat. number 85112) and 5% fetal bovine serum (FBS, ermo Fisher Scientific, cat. number 10091148) in PBS 1X for 1 hour at room temperature and overnight at 4°C with one of the following primary antibodies: anti-DDR2 1 : 500 (GeneTex, United States, cat. number GTX25520), anti-α-SMA 1 : 200 (Dako, United States, cat. number GA61161-2), and anti-CD68 1 : 250 ( ermo Fisher Scientific, cat. number MA5-13324). Antibodies were detected with either biotinylated anti-2 e Breast Journal rabbit 1 :1500 or anti-mouse 1 : 2000 for 1 hour at room temperature, followed by streptavidin-HRP 1 : 500 for 30 minutes at room temperature and DAB Substrate Kit for 2-3 minutes (all purchased from ermo Fisher Scientific, cat. numbers 31820, 31800, SA10001, and 34002, respectively). Finally, slides were counterstained in hematoxylin and mounted with Sub-X Mounting Medium (Leica Biosystems, Germany, product ID SUB-X-MOUNTING-MEDIUM). Protein expression was quantified using FIJI-ImageJ (Colour Deconvolution plugin). e expressions of DDR2, α-SMA, and CD68 proteins were evaluated in the whole tissue for intensity of staining comparing breast IDC samples to nontumor ones. Staining with relative values up to 3-fold increase was classified as low status, while that with values superior to 3-fold increase was classified as high status. is evaluation was made following the criteria described by Toy et al. [28]. Samples 4, 12, and 23 (α-SMA staining), as well as 23 and 24 (CD68 staining), were excluded from the correlation analysis due to the high percentage of fatty tissue detected, which made it complicated to classify them as low or high status.

RT-qPCR.
Total RNA was extracted and purificated from breast IDC and nontumoral samples (including the stromal compartment) using the Total RNA Purification Kit (Norgen Biotek Corp., Canada, cat. number 17250). RNA concentration and quality were assessed by NanoDrop spectrophotometer (ND-1000; ermo Fisher Scientific) and 0.5-1 μg of RNA was reverse-transcribed into cDNA with iScript ™ Reverse Transcription Supermix (BIO-RAD, United States, cat. number 1708841). Quantification of cDNA template was performed with real-time PCR using iTaq ™ Universal SYBR ® Green Supermix (BIO-RAD, cat. number 1725121) in ABI 7900HT (Life Technologies). PCR primers (Life Technologies) were as follows: DDR2 F, GGAGGTCATGGCATCGAGTT, and R, GAGTGC-CATCCCGACTGTAATT; GAPDH (housekeeping) F, GTATGACTCCACTCACGGCAA, and R, CTTCCCATTCTCGGCCTTG; RPS15 (housekeeping) F, AGACGAGTTTCAGTGTTGCC, and R, AGACCA-CAGCCTCAGACAAG. Relative expression of target genes was normalized to the internal control genes GAPDH and RPS15 by the ΔΔCt method. Data were generated by the use of specific software (ABI Prism, SDS2.0, Life Technologies) after normalization. e experiments were performed in triplicate.
2.6. Statistical Analysis. Data were analyzed using GraphPad Prism software (version 6) and are presented as the mean ± standard deviation (SD). Differences in gene and protein expression levels were analyzed by two-tailed Student's t-test. e association between DDR2 and stromal markers expression was analyzed by Chi-square, Fisher's exact test, and Pearson's correlation coefficient. e criterion for significance was p ≤ 0.05 for all comparisons.

Collagen and DDR2 Expression Are Upregulated in Human Invasive BC Compared to Nontumoral Tissue.
Collagen accumulation and DDR2 high expression have been reported in invasive BC both in vitro and in vivo [17-19, 38, 39]. We analyzed the intratumoral levels of collagen type I and DDR2 in 25 mammary tissue samples from IDC (e-cadherin positive) and their matched adjacent normal tissues. As observed in Figure 1, the vast majority of the tumors showed collagen deposition (p � 0.0209) and upregulated DDR2 protein (p < 0.0001) compared to that in nontumoral tissue. Besides, DDR2 quantification in the RNA samples obtained from IDC patients supported this finding, as demonstrated by RT-qPCR (p � 0.0004). Further analysis of DDR2 status using METABRIC Database confirmed that DDR2 mRNA levels were significantly higher in ductal and lobular BC subtypes, with respect to those in the normal breast tissue (p � 0.0381 and p � 0.0023, respectively).

DDR2 Upregulation Is Associated with BC Metastatic
Progression and Low Postmenopausal Patient Survival. DDR2 overexpression has been related to poor prognosis in several cancer types, including BC [18,28]. As observed in Figure 2, metadata analysis showed that DDR2 gene level was higher in those invasive BC tumors with a more aggressive phenotype (p � 0.0318), defined by the CCND1 amplification as metastatic marker [33]. Moreover, regarding patient age, high DDR2 expression notably increased postmenopausal BC patients' mortality, compared to the postmenopausal women who exhibited low DDR2 expression levels (p � 0.03). On the contrary, DDR2 expression was not relevant in pre-and perimenopausal women. us, DDR2 status was only significant when we analyzed postmenopausal BC patients.  Figure 2).

Enhanced Expression of α-SMA and CD68 in Human
Breast IDC. CAFs and TAMs are active players in BC progression and chemoresistance [25,26]. We evaluated the infiltration of CAFs and TAMs in breast IDC samples by α-SMA and CD68 staining quantification, respectively. In the healthy breast tissue, these markers correspond to fibroblast-like cells and macrophages, which maintain mammary gland homeostasis. us, the comparison of α-SMA and CD68 expression levels between normal and tumor samples represents the mobilization of these stromal cell populations during the tumorigenic process. As observed in Figure 3, we found increased expressions of α-SMA and CD68 markers in the peritumoral and intratumoral areas compared to normal adjacent tissue.
ese results revealed a pronounced recruitment of CAFs and TAMs in the developing tumoral tissue from breast IDC patients.

Overexpression of DDR2 Correlates with Increased CAFs Infiltration in Human Invasive BC.
e TME is a dynamic compartment that drives tumor growth. CAFs are a major player in this ecosystem, supporting BC growth through several pathways [25]. We aimed to elucidate whether the intratumoral DDR2 expression may be related to the recruitment of CAFs and, therefore, could be used as a marker of stromal reorganization during BC. Using Chisquare and Fisher's exact test, we found a strong correlation between intratumoral DDR2 protein levels and CAFs infiltration (characterized by α-SMA expression) in breast IDC samples. As observed in Figure 4, low DDR2 expression was directly linked with reduced CAFs recruitment (p � 0.0008). In depth, 75% of the low DDR2 expressing tumors exhibited low α-SMA staining, while CAFs infiltration was high in only 25% of the low DDR2 levels breast tumors. On the other hand, 64% of the high DDR2 expressing tumors showed elevated α-SMA protein levels, while barely 36% of these tumors presented low α-SMA expression. Additional analysis of DDR2 and α-SMA correlation in IDC and ILC using METABRIC Database supported this finding (p < 2.2e − 16 and p < 1.9e − 08, respectively), uncovering a direct relation between tumoral DDR2 expression and CAFs recruitment.

TAMs Recruitment Correlates with DDR2 Upregulation in
Human Invasive BC. TAMs, key immunomodulatory cells, are a significant source of cytokines and chemokines that promote disease progression [26]. Here we analyzed the correlation of infiltrating CD68 positive macrophages and the expression of intratumoral DDR2 in breast IDC samples. Chi-square and Fisher's exact test revealed a deep correlation between high intratumoral DDR2 expression and elevated infiltration of CD68 positive TAMs (p � 0.0001). Strikingly, low macrophages counts were found only in patients with low levels of intratumoral DDR2. However, 67% of high DDR2 expressing tumors showed elevated macrophages recruitment, and only 33% of these tumors presented low CD68 positive cells counts ( Figure 5). Further analysis of DDR2 and CD68 correlation in IDC and ILC using METABRIC Database confirmed our results (p < 2.2e − 16 and p < 3.3e − 07, respectively), suggesting a feasible link between intratumoral DDR2 expression and TAMs infiltration.

Discussion
e ECM is one of the main components of the TME. e importance of these extracellular elements resides in their ability to promote significant changes in the behavior of cancer cells through receptor-mediated ECM-tumor cell interactions.
ese alterations exert a positive effect on cancer cells, by means of chemoresistance, as well as increased proliferation and migration, among others [40]. Regarding BC, the first most prevalent malignant disease among women, collagen accumulation has been postulated as a risk factor [38,41]. Despite the limited number of patients, and taking into account the fact that the heterogeneity of the BC population represents a limiting factor, the analyzed invasive BC tissue samples exhibit a marked collagen deposition, pointing out the crosstalk between ECM and BC cells as a feasible starting point for tumor development. Even though collagen constitutes an integral and functional key player of cancer tissue regulation, current therapies against the variety of histological and molecular BC subtypes do not target cellular components directly linked with this ECM protein.
e collagen-binding RTK DDR2 is known to play important roles in tumor progression [16]. In the mammary carcinogenic tissue, DDR2 expression has been postulated as an independent prognostic value [18], with a critical role during postpartum-associated BC [20,42]. In this regard, increased DDR2 promotes tumor aggressiveness in the breast, by means of enhanced proliferation, migration, colony formation, and metastasis [17,19]. In this work, we show that DDR2 expression is upregulated in a small sample size composed by different invasive BC molecular profiling cohorts, including luminal-like, HER2-enriched, and triplenegative subclasses. Besides, the later bioinformatics analysis including a large number of patients confirms these results. As for the last, Toy et al. [28] described that high DDR2 protein levels were significantly associated with poor triplenegative BC patients' outcome, evidencing the involvement of DDR2 in tumor malignancy. In this line, we detect DDR2 overexpression in metastatic BC patients with CCND1 gene amplification, a worse survival predictive biomarker in breast tumors [33,34]. Furthermore, we uncover that mortality increase in those women with a high DDR2 expression status along with postmenopausal age. us, elevated expression of DDR2 could reflect a worse prognosis in those postmenopausal BC patients. Taken together, these results suggest that DDR2 may participate in tumor cells growth and permanence. In fact, bioinformatics analyses corroborate that DDR2 signaling in BC is widely connected to cancer-associated pathways, apart from those related to ECM interaction. Interestingly, the examination of the DDR2 expression pattern indicates that this receptor represents an important part of the stromal compartment, mostly detected in fibroblasts and macrophages. is goes in line with the immunohistochemical analysis, in which we observe DDR2 expression in the tumor stroma of BC tissue slides. e TME is known to support cancer progression, influencing therapeutic response and clinical outcome. us, infiltrating CAFs and TAMs, the main TME cell populations, maintain cancer development [43]. Regarding stromal cells markers detection in 25 BC patients, α-SMA and CD68 upregulation reveals a high recruitment of these cells into the mammary tumor tissue. Concerning α-SMA detection, it is also found in vascular muscular cells and pericytes, which account in high densities in tumor tissues [44]. Even though the heterogeneity of CAFs in the TME cannot be described using a single marker, α-SMA expression has been extensively analyzed as one of the main myofibroblast markers of the tumor stroma [45,46]. Intriguingly, adding our small number of samples with the extensive datasets available, we observe a correlation between DDR2 overexpression and increased levels of CAFs and TAMs in the malignant tissue. is finding suggests that the stromal fraction of BC may be modulated by tumor DDR2. Aside from cancer cell intrinsic effects, DDR2 is also involved in several processes that alter the breast TME, facilitating disease progression. is could be controlled by the wide spectrum of DDR2 mediated changes in the tumor stroma. In this line, Corsa et al. proposed that DDR2 expressed by BC cells determines CAFs activation to enhance tumor invasion and metastasis [22]. Similarly, tumor DDR2 regulates collagen signaling in CAFs during BC progression, as recently reported by Bayer et al. [21]. Comparably, we uncover a plausible positive interrelation between the levels of DDR2 and the mobilization of CAFs and TAMs into the tumor foci, which points out to DDR2 as a possible regulator of attracting signals for these cell populations. Concerning the latter, it is well established that DDR2 is implicated in metalloproteinases (MMPs) secretion by tumor cells [47], resulting in ECM degradation and facilitating stromal cells infiltration and cancer cells migration and invasion [48]. us, MMPs production by BC cells through DDR2 modulation could promote CAFs and TAMs recruitment. Eventually, although CAFs and TAMs are defined as cancer progression promotors, these stromal cells can also delay tumor growth, which depends on the subtype of CAFs and TAMs populations [49][50][51][52]. Considering its  Other DDR2 stimulated proteins are TGF-β and parathyroid hormone-related protein (PTHrP), known to play a role in bone metastasis [53]. On the one hand, TGF-β contributes to CAFs infiltration and proliferation, which produce excessive ECM deposition, providing a scaffold for the entrance of immune cells, such as TAMs, and a substrate for cell migration [54,55]. On the other hand, PTHrP induces the epithelial-to-mesenchymal transition (EMT), by which epithelial cells transform in fibroblasts in several tissues [56,57,58].
is process may account for breast epithelium, therefore consisting in a fibroblast source related to DDR2. Interestingly, in bone metastasis, PTHrP seems to stimulate the recruitment of macrophages into the tumor promoting the secretion of CCL2 by osteoblasts [56]. It is tempting to hypothesize that a similar process may occur in breast tissue, driving to TAMs accumulation. is work represents one of the few preliminary studies correlating DDR2 expression with CAFs and TAMs recruitment during invasive BC progression. However, further ongoing studies will expand the relation between these markers during BC development. Our prior conclusions reflect that DDR2 may constitute an alternative therapeutic target, especially for those individuals with invasive BC or  8 IDC DDR2 expression 9 R = 0.5 p < 1.9e-08 (d) Figure 4: Interplay between the expressions of DDR2 and α-SMA in human invasive BC. Images represent characteristic immunohistochemical samples of DDR2 and α-SMA low and high expression (arrows). Scale bar: 100 μm. Insets: 2x magnification (a). Graph shows that DDR2 expression is related to α-SMA levels (p � 0.0008) (b). DDR2 and α-SMA status positive correlation in breast IDC (R � 0.5; p < 2.2e − 16) and ILC (R � 0.5; p < 1.9e − 08) samples obtained from METABRIC BC Datasets (c, d).   Figure 5: Interplay between the expressions of DDR2 and CD68 in human invasive BC. Images represent characteristic immunohistochemical samples of DDR2 and CD68 low and high expression (arrows). Scale bar: 100 μm. Insets: 2x magnification (a). Graph shows that DDR2 expression is related to CD68 levels (p � 0.0001) (b). DDR2 and CD68 status positive correlation in breast IDC (R � 0.32; p < 2.2e − 16) and ILC (R � 0.46; p < 3.3e − 07) samples obtained from METABRIC BC Datasets (c, d). 12 e Breast Journal with a high risk of relapse, due to the ineffectiveness of hormone and HER-2 therapies, like in triple-negative subtype, as well as the development of chemoresistance. e observed potential connection between DDR2 and the TME markers could put some light on the treatment options for these patients, focusing on the combination of DDR2 inhibitors and stromal-directed drug regimens with the present clinical approaches.

Data Availability
e data used to support the findings of this study are available upon request from Dr. Beatriz Arteta through the email address beatriz.arteta@ehu.eus.

Conflicts of Interest
e authors declare that they have no conflicts of interest.