Prostate Tissue Specimens
The Grade/Stage tissue microarray (TMA) slides containing de-identified human PCa specimens were obtained from the Prostate Cancer Biorepository Network (PCBN), a public bioresource for PCa investigators funded by the Department of Defense resulting from a collaboration between Johns Hopkins University and New York University School of Medicine (Provider Investigator: Dr. Bruce Trock, Johns Hopkins University, Baltimore, MD). Five TMAs referred to as TMA 18, 19, 22, 24, and 34 containing 1,600 cores obtained from radical prostatectomies performed in PCa patients were analyzed in our study. The TMAs comprise a total of 200 cases of PCa with various GS and pathology stages represented in quadruplicate cores, and matching non-cancerous prostate tissues referred to as “normal” (also 4 cores per case). All subjects were annotated based on age, race, treatments received, biopsy GS, serum PSA levels at diagnosis, clinical stage (T) as defined by the American Joint Committee on Cancer (Table 1), and occurrence of BCR (increase of postoperative serum PSA level to 0.2 ng/ml) (Additional file 1: Table S1).
Immunohistochemistry
TMA sections were deparaffinized with xylene and rehydrated with decreasing percentages of ethanol (100% to 70%). Unmasking of antigenic epitopes was performed with Antigen Retrieval Citrus Plus Solution (Cat. # HK0805K, BioGenex, Fremont, CA) in a pressure cooker placed in a microwave for two cycles of 15 min at heat levels 5 and 2. After 20 min at room temperature, the sections were washed with phosphate-buffered saline (PBS) and endogenous peroxidase quenched with 3% hydrogen peroxide in PBS for 30 min. After washing with PBS and blocking of non-specific binding sites with 2.5% normal horse serum for 20 min at room temperature, the tissues were incubated with an in-house raised rabbit monoclonal antibody (mAb) against the extracellular domain of human DDR1 [42] referred to as Rab‑819 antibody, at a 1:100 dilution (from a stock solution of 1 mg/ml) overnight at 4oC. Immunostaining was performed by incubation with a peroxidase micropolymers attached to anti-rabbit IgG made in horse (Cat. # MP-7401, ImmPRESSTM, Vector Laboratories, Burlingame, CA) for 30 min at room temperature, followed by detection with ImmPACTTM DAB peroxidase substrate (Cat. # SK-4105, Vector Laboratories) and light nuclear counterstaining with Mayer’s hematoxylin (Cat. # HMM500, ScyTek Laboratories, Logan, UT). DDR1 immunoreactivity was evaluated and reported by two independent pathologists (SV and DS) as positive or negative staining in different subcellular localizations (i.e., plasma membrane, nucleus, and cytoplasm) in cancerous and adjacent normal (no evidence of neoplastic changes) tissues. For membrane immunostaining, only full membranous staining was considered positive, whereas no staining or staining at basal or basolateral locations were defined as negative.
Tissue Culture
The human PCa PC-3 and C4-2B cell lines were purchased from the American Type Culture Collection (ATCC). Normal human prostate epithelial cell lines, CF-91, ML-8891, CLR-2221, RWPE-1, RWPE-2, and benign prostatic hyperplasia (BPH) epithelial cell lines were kindly provided by Dr. S. Sheng (Wayne State University [WSU]), whereas human PCa LNCaP and DU145 cell lines were supplied by Dr. H-R.C. Kim (WSU). Normal prostate epithelial cell lines, CF-91, ML-8891, CLR-2221, RWPE-1, RWPE-2 were maintained in Keratinocyte Serum Free Medium (Cat. # 17005042, Thermo Fisher Scientific, Waltham, MA.) supplemented with 0.05 mg/ml bovine pituitary extract and 5 ng/ml epidermal growth factor. PC3, LNCaP and C4-2B cell lines were maintained in RPMI-1640 Medium, HEPES (Cat. # 22400121 Thermo Fisher Scientific), supplemented with 10% Fetal Bovine Serum (FBS) (Cat. # 16000044, Thermo Fisher Scientific). The DU145 cell line was cultured in DMEM Medium (Cat. # 10313039, Thermo Fisher Scientific) supplemented with 10% FBS and 1% Penicillin-Streptomycin solution. All the human cell lines used were periodically confirmed negative for mycoplasma contamination and authenticated through short tandem repeat profiling by the Research Technology Support Facility of Michigan State University.
Immunoblot Analyses, Collagen Stimulation, and Cell Fractionation.
For analyses of DDR1 expression in whole cell lysates, cultured prostate epithelial cells were lysed in RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% NP-40, 0.25% sodium deoxycholate and 1 mM EDTA) supplemented with protease inhibitors Cocktail, EDTA-free (Cat. # 539134, Sigma-Aldrich, St. Louis, MO) and 10 mM sodium fluoride and 1 mM sodium orthovanadate. The cell lysates were cleared by centrifugation at 13,000 g at 4°C for 15 min, and the protein concentration was determined using the BCA kit from Pierce (Cat. # 23227, Waltham, MA). Equal amounts of protein from each lysate were resolved by reducing 7.5% SDS-PAGE. Proteins were then transferred to a nitrocellulose membrane using conventional methods. The blots were probed with anti-DDR1 polyclonal antibody Sc-532 (Santa Cruz Biotechnology, Inc. Dallas, Texas), which recognizes a DDR1 epitope at the C-terminal end of the receptor. For loading control, the same blot was reprobed with anti-b-actin antibody.
For analyses of DDR1 subcellular localization in malignant (PC-3) and non-malignant (RWPE-1) cells as a function of collagen stimulation, cells were washed twice with PBS and incubated in serum-free media, overnight. The cells were then treated with 20 µg/ml rat tail collagen type I (Cat. # 354236, Discovery Labware Inc., Corning™, Bedford, MA) for 90 min at 37°C, washed twice with cold PBS and then gently dissociated from the plates using Cell Dissociation Buffer (Cat. # 13151014, Thermo Scientific). The cytoplasmic/membrane and nuclear fractions were isolated using the NE-PER Nuclear Cytoplasmic Extraction Reagent kit (Cat. # 78833, Thermo Scientific, Grand Island, NY, USA), according to the manufacturer's instruction. Cytoplasmic and nuclear extraction buffers were supplemented with protease inhibitors (Roche, complete, Mini, EDTA-free) and 10 mM sodium fluoride and 1 mM sodium orthovanadate. Protein concentrations were determined using the BCA kit. For immunoblot analyses, the nuclear and cytoplasmic/membrane fractions were resolved by SDS-PAGE in two gels: 7.5% polyacrylamide for DDR1 and 4-20% polyacrylamide for glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and Histone. After transfer, the first membrane was probed with an antibody recognizing phosphorylated DDR1 at Tyr513, namely DDR1 rabbit mAb E1N8F (Cat. # 14531) from Cell Signalling Technology (CST), Danvers, MA. After stripping, the membranes were re-probed with total anti-DDR1 (D1G6) rabbit mAb (Cat. # 5583, CST), which recognizes the C-terminal region of DDR1. The membrane was also reprobed for presence of transferrin receptor as a marker of membrane-anchored protein, using an anti-transferrin mouse mAb (Cat. # 612124) from BD Transduction Laboratories, San Jose, CA. The second membrane was probed with anti-Histone H1 mouse mAb (Cat. # 05-457, Sigma-Aldrich) and re-probed with anti-GAPDH mouse mAb (Cat. # MA5-15738, Thermo Fisher Scientific), as nuclear and cytoplasmic protein markers, respectively. Antigen/antibody complexes were visualized using the SuperSignal West Pico Plus and/or the SuperSignal West Femto Maximum Sensitivity Substrate from Thermo Fisher Scientific (Rockford, IL; Cat. # 34580 and 34095, respectively).
Statistical Methods
The primary objective was to evaluate the association between DDR1 IHC expression and GS, comparing between “low grade” (GS 3+4 or lower) and “high grade” (GS 4+3 or higher) tumors. For each PCa case, we summarized the quadruple core-level DDR1 data to patient-level data with overall staining percentage (OSP), a semi quantitative score defined as the percentage of positive stained cores among all quadruple cores, per tissue type (cancerous and adjacent normal), and per subcellular location (membrane, nuclear, or cytoplasm). TMA cores that had stroma, no glands, or no tissue after the staining process were considered as missing at random rather than negative staining.
Association between high/low grade GS and dichotomized DDR1 staining OSP (OSP = 0% vs OSP > 0%) was evaluated with Fisher’s exact test, per tissue type and sub-cellular location. To take into account the paired-tissue design of this TMA, we further evaluated DDR1 relative expression, which was categorized in three expression patterns: Equal expression (cancerous OSP = adjacent benign OSP), higher expression (cancerous OSP > adjacent benign OSP), and lower expression (cancerous OSP < adjacent benign OSP). The association between high/low grade GS and DDR1 relative expression patterns was then evaluated using Fisher’s exact test, per subcellular location.
The association between DDR1 and GS was further evaluated with multivariate logistic regression adjusted for tumor/node/metastasis (TNM) staging, classified as local (T0N0MX, T2N0MX, T2NXMX, and T2XN0MX) and advanced (T2XN1MX, T3AN0MX, T3AN1MX, T3ANXMX, T3BN0MX, and T3BN1MX) for each subcellular location. These evaluations were performed with DDR1’s OSP from cancer, OSP from normal, and relative DDR1 relative expression pattern (higher, equal, or lower expression), respectively.
For this cohort of TMA, we obtained de-identified baseline characteristics and clinical outcomes overall survival (OS), biochemical recurrence free survival (BCRFS), and cause-specific death. Associations between DDR1 expression and clinical outcomes such as OS and BCRFS were also performed with the Cox model. Competing risk analysis of cause-specific death was not performed as there were a very low event of death due to PCa in this cohort. All p values are 2-sided with a significance level of 0.05. The results should be regarded only as descriptive findings and multiple testing were not adjusted. All calculations were performed with statistical software R version 3.6.1.