This study was performed in accordance with the ethical standards, and approved by the Bioethics Committee of the University of Warmia and Mazury in Olsztyn (decision no. 3/2010 and 34/2010), and informed written consent regarding the use of tissue was obtained from each patient included in the study.

The specimens were collected at the Hospital of the Ministry of Internal Affairs and Administration in Olsztyn from 2010 to 2013. The study included 121 patients with CRC (demographic and clinicopathological data are presented in Table I). None of the CRC patients had a second neoplastic disease or had previously undergone chemo- or radiotherapy. The control group consisted of 72 healthy individuals (24 males and 48 females, average age 57.2±6.65 years, range 36–82 years; mean ± SD) who underwent colonoscopy as a part of a routine screening for CRC (within the National screening program for early detection of colorectal cancer). Control subjects had no family history of CRC. None of the CRC patients or control subjects suffered from inflammatory bowel disease. Clinical and demographic data were obtained at the time of enrollment. Data on the overall survival were collected for all patients. Median follow-up time was 36.1 months.

CRC samples were obtained during the partial surgical resection of the large intestine, and control group specimens were collected during colonoscopy. In CRC patient group, two types of matched samples were taken within 20 min after tumor resection: i) tumor tissue and ii) macroscopically unchanged mucosa from a distant part of resected large intestine. Specimens were immediately cut in two pieces for qPCR and western blot analyses, frozen in liquid nitrogen, and stored at −80°C, whereas for routine histological evaluation and immunohistochemistry, the samples were fixed in 10% neutral buffered formalin and further processed into paraffin blocks. In the control group of healthy patients, one biopsy was fixed in 10% neutral buffered formalin for routine histological examination, and two specimens from the adjacent location to the biopsy site were collected for qPCR or western blot assays.

Human CRC cell lines with different characteristics (HT-29, SW-480, LoVo) (14) and a control line, the CCD 841 CoN cell line (epithelial-like, established from normal colonic tissue) were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA). Cells were cultured under conditions recommended by the manufacturer, and were harvested at ~80% confluence.

Total RNA was extracted from cell lines as well as paired samples of cancer tissue and unchanged mucosa derived from 121 CRC patients and 40 colonoscopic biopsies of healthy subjects using a Total RNA Prep Plus kit (A&A Biotechnology, Gdynia, Poland), following the manufacturer's protocol. Isolated RNA was quantified with spectrophotometry (NanoDrop 1000, NanoDrop products, Wilmington, DE, USA). Reverse transcription was carried out in a vial containing 20 μl reaction mixture of 2 μg of total RNA, 0.5 μg of oligo dT primers (Sigma-Aldrich, St. Louis, MO, USA), 200 U of RevertAid™ Reverse Transcriptase, 20 U of RiboLock™ RNase inhibitor, and 1 mM of each dNTP (all Thermo Scientific, Waltham, MA, USA). Reactions were performed according to the manufacturer's instructions, and resulting complementary DNAs (cDNAs) were stored at −20°C after 10-fold dilution with nuclease-free water to be used as the template in qPCR analysis.

Quantification of PLAGL1 gene expression was carried out using ABI 7500/7500 Fast Real-Time PCR system (Life Technologies, Applied Biosystems, Foster City, CA, USA). Hypoxanthine phosphoribosyltransferase 1 (HPRT1) gene was used as an internal control to normalize the transcript levels of PLAGL1. The levels of PLAGL1 and HPRT1 cDNAs in collected isolates were determined using TaqMan^®Fast Advanced Master Mix and a respective TaqMan^® Gene Expression assay (all from Life Technologies, Applied Biosystems) according to the manufacturer's instructions. The primers and TaqMan probes used were PLAGL1: Hs00414677_m1, and HPRT1: Hs02800695_m1. All samples were amplified in duplicates using the following conditions: polymerase activation for 20 sec at 95°C, followed by 40 cycles of denaturation at 95°C for 3 sec and annealing/extension at 60°C for 30 sec. No template control reactions were performed for each qPCR run. Standard curves consisting of serial dilutions of the appropriate cDNA were used to control the efficiency of qPCR reactions. Relative quantification of PLAGL1 expression was evaluated using the ΔΔCt method (15). The fold change in the relative PLAGL1 gene expression was determined by calculating the 2^−ΔΔCt value. Fold increase above 1 (2^−ΔΔCt >1) indicated PLAGL1 overexpression in CRC tissue, and fold decrease under 1 (2^−ΔΔCt <1) indicated PLAGL1 downregulation.

Paired samples of tumor tissue and unchanged mucosa derived from 95 CRC patients and 32 colonoscopic biopsies were homogenized in RIPA lysis buffer supplied with 1:100 protease inhibitor cocktail, 1:100 phosphatase inhibitor cocktail 2 and 5 mM EDTA (all from Sigma-Aldrich). Homogenates were briefly centrifuged to remove tissue debris. Then, samples were centrifuged twice at 9,000 × g for 10 min at 4°C. After centrifugation, supernatants were collected and the total protein content was determined by the Bradford method (16). Samples were aliquoted and stored at −80°C until further analyses.

To determine the level of PLAGL1 protein in tissue lysates, the SDS-PAGE followed by western blotting assays were performed. Isolated protein samples were denatured for 5 min at 95°C and loaded on polyacrylamide gel (40 μg/lane). Gels were run at the 10 mA/gel during migration in the stacking gel and 15 mA/gel in the separating gel (10%). Proteins were transferred onto PVDF membrane (western blotting membrane, Roche, Mannheim, Germany). Blots were blocked in 5% non-fat dry milk dissolved in Tris-buffered saline pH 7.5 with 0.1% Tween-20 (TBS-T) followed by overnight incubation at 4°C with respective primary antibody. Rabbit anti-human monoclonal antibodies against PLAGL1/ZAC (diluted 1:1,000 in TBS-T; #ab129063, Abcam, Cambridge, UK), and polyclonal antibodies anti-actin (ACTB; 1:100; #A2066, Sigma-Aldrich) were used. ACTB level was used as the internal protein load control. After the incubation, primary antibodies were washed out with TBS-T. Then, the membranes were treated with the specific HPR-conjugated goat anti-rabbit IgG secondary antibodies (1:40,000; #A0545, Sigma-Aldrich) for 90 min at room temperature (RT), developed with an enhanced chemiluminescence (SuperSignal West Pico Chemiluminescent Substrate, Thermo Scientific), and visualized with G:BOX iChemi XR imaging system (Syngene, Cambridge, UK). For the negative control, a primary antibody was omitted and substituted with phosphate-buffered saline (PBS). Protein extracts from 293T cells or HeLa cells (both from Abcam) were used as the positive controls for PLAGL1 immunoblotting. Molecular weight standard (Spectra Multicolor Broad Range Protein Ladder, Thermo Scientific) was included into each blotting experiment to confirm the molecular weight of detected bands. Band intensity was quantified using ImageJ software (NIH, Bethesda, MD, USA) (17).

PLAGL1 protein optical density (OD) was normalized on the basis of ACTB protein OD. OD ratios between tumor and the corresponding unchanged tissue of CRC patients were calculated. The ratios >1 indicated that the expression of PLAGL1 protein was upregulated in CRC tissue while those <1 were regarded as downregulated.

PLAGL1 immunoreactivity was analyzed in matched tumor and unchanged colorectal tissues of 60 CRC patients. Immunohistochemistry was performed on 4-μm-thick paraffin sections. Sections were subjected to antigen retrieval procedure by microwaving for 20 min in retrieval solution buffer, pH 6.0 (Leica, Wetzlar, Germany), followed by the incubation in 3% H₂O₂ in methanol for 10 min, and next in 2.5% normal horse serum (Vector Laboratories, Burlingame, CA, USA) for 30 min. The sections were incubated overnight at 4°C with rabbit anti-human monoclonal antibody against PLAGL1/ZAC (diluted 1:2,000 in PBS; #ab129063, Abcam). After washing with PBS, the sections were treated with HRP-conjugated secondary antibody (ready-to-use dilution; ImmPRESS Universal reagent Anti-Mouse/Rabbit Ig, Vector Laboratories) for 30 min at RT. Then, the sections were immersed in 3′,3′-diaminobenzidine (DAB; Dako, Glostrup, Denmark), counterstained with Harris' haematoxylin (Sigma-Aldrich), dehydrated in ethanol, cleared in xylene, and mounted with DPX (Sigma-Aldrich). For each set of staining, the negative controls were performed by omitting the primary antibody.

The PLAGL1 immunostained sections were evaluated using Olympus BX41 light microscope (Olympus, Tokyo, Japan) by an independent pathologist in a blinded manner regarding the clinical data of the patients. Immunoreactivity of PLAGL1 was assessed in cancer cells of CRC sections and enterocytes of the unchanged colon mucosa using the scale based on the reaction intensity (0, no reaction; 10, ≤10%; 30, 11–30%; 60, 31–60%; 80, 61–80% and 100, >80%). Ratios in PLAGL1 score between tumor cells and cells of the matched unchanged tissue of CRC patients were calculated. The ratios >1 indicated that the immunoreactivity of PLAGL1 in CRC tissue was upregulated while those <1 were regarded as down-regulated.

Statistical analyses were performed using Prism 6 (GraphPad, La Jolla, CA, USA) and Statistica v.10 (StatSoft, Tulsa, OK, USA) software. The differences in mRNA and protein levels between matched tumor and unchanged samples of CRC patients were examined by the Wilcoxon matched-pairs test, whereas differences between colon mucosa biopsies of healthy subjects and tissues of CRC patients as well as between cell lines were assessed by the Mann-Whitney U test. The correlations between the demographic, clinicopathological, and molecular parameters were analyzed by the Fisher's exact and Chi-square tests and confirmed using the Mann-Whitney U test and Kruskal-Wallis test. Survival curves were plotted using Kaplan-Meier method. The statistical significance of differences in survival between groups of patients was evaluated using the log-rank test and confirmed by Cox regression method. In all the analyses, results were considered statistically significant when p<0.05.

To determine the expression of PLAGL1 at the mRNA level, matched tumor and unchanged tissues derived from CRC patients and colonic biopsies of healthy group were subjected to qPCR analysis. PLAGL1 mRNA was found in all studied tissue samples of CRC patients and colonic biopsies of healthy individuals. Among the 121 tumor specimens tested, the relative PLAGL1 mRNA level (tumor tissue vs. matching unchanged mucosa of CRC patients) was decreased in 87 (71.9%) tumors while it was increased in 34 (28.1%) cases (Table II). The expression of PLAGL1 mRNA was significantly decreased in the tumor tissues when compared to unchanged tissue of CRC patients and the colon mucosa of healthy individuals (0.44±0.02 vs. 0.83±0.02 and 1.00±0.04, respectively; p<0.0001; Fig. 1). The levels of PLAGL1 mRNA in unchanged tissues of CRC patients did not differ significantly from those in colonic biopsies of the healthy group (0.83±0.02 vs. 1.00±0.04; p>0.05; Fig. 1).

QPCR analysis was also used to determine the expression of PLAGL1 gene in CRC cell lines and the control line established from normal colonic tissue. PLAGL1 mRNA levels in HT-29, LoVo and SW-480 cells were significantly lower than those in the CCD 841 CoN cell line (0.0004±0.0003, 0.024±0.0006 and 0.047±0.0087, respectively, vs. 1.00±0.0087; p<0.05; Fig. 2). Among the CRC cell lines the lowest level of PLAGL1 expression was observed in HT-29 cells (p<0.05; Fig. 2). The level of PLAGL1 mRNA in the LoVo cell line was significantly lower than that in SW480 cells (p<0.05; Fig. 2).

To assess the impact of PLAGL1 expression at the mRNA level on CRC pathogenesis, the relationships between PLAGL1 mRNA content and selected demographic and clinicopathological parameters were tested (Table II). The relative PLAGL1 mRNA level did not correlate with the parameters gender, age, tumor localization, TNM disease stage, depth of invasion, lymph node involvement, or the presence of metastases (p>0.05; Table II).

To determine the expression of the PLAGL1 gene at the protein level, matched tumor and unchanged tissues derived from CRC patients and colonic biopsies of healthy subjects were analyzed by western blotting. PLAGL1 protein was found in all studied tissues of CRC patients and in 30/32 (93.75%) colonic biopsies of healthy subjects. The average content of PLAGL1 protein did not differ significantly between tumor and unchanged tissues of CRC patients or the colon mucosa of healthy individuals (3.47±0.27 vs. 3.12±0.24 and 2.73±0.26, respectively; p>0.05; Fig. 3). Among 95 tumor tissue specimens tested, the relative content of PLAGL1 protein (tumor tissue vs. matching unchanged mucosa of CRC patients) was downregulated in 47 (49.5%) tumors while it was upregulated in 48 (50.5%) cases (Table III).

PLAGL1 immunoreactivity was observed in enterocytes (Fig. 4A) as well as cancer cells of the analyzed tissues (Fig. 4B, D, E, G, H and J–M). Immunoreactivity of the PLAGL1 protein was noted in cancer cells in 49/60 (81.7%) of the analyzed CRC cases, whereas in enterocytes of unchanged large intestine tissue its expression was observed in 57/60 (95%) cases. The average intensity of PLAGL1 immunostaining did not differ significantly between tumor and unchanged tissues of CRC patients (18.5±2.07 vs. 24.83±2.85, respectively; p>0.05). Among 60 tumor tissue specimens tested, the relative intensity of PLAGL1 staining was decreased in 27 (45%) tumors and elevated in 15 (25%) cases, while 18 (30%) tumor samples demonstrated similar level of PLAGL1 immunoreactivity as cells of the corresponding unchanged tissues (Table IV).

Possible correlations of PLAGL1 expression at the protein level with selected demographic and clinicopathological parameters were analyzed based on the results obtained by western blot and immunohistochemical analyses. The relative PLAGL1 protein levels in tumor specimens did not correlate with the patient gender, age, tumor localization, and depth of invasion. However, PLAGL1 protein content was lower in tumor specimens derived from patients diagnosed with: i) lymph node involvement [p=0.0074, the Fisher's exact test, Table III; confirmed by the Mann-Whitney U test: 1.66±0.18 vs. 1.21±0.17 (N0 vs. N1+N2), p=0.0122, Fig. 5A], ii) the presence of metastases [p=0.0037, the Fisher's exact test, Table III; confirmed by the Mann-Whitney U test: 1.57±0.14 vs. 0.80±0.16 (M0 vs. M1), p=0.0309, Fig. 5B], and iii) a higher TNM disease stage [p=0.001, the Fisher's exact test, Table III; confirmed by the Mann-Whitney U test: 1.73±0.19 vs. 1.17±0.16 (I+II vs. III+IV), p=0.0014, Fig. 5C].

Similarly, decreased intensity of PLAGL1 immunohistochemical staining in tumor specimens was associated with lymph node involvement [p=0.0081, the Fisher's exact test, Table IV; confirmed by the Mann-Whitney U test: 1.10±0.14 vs. 0.45±0.11 (N0 vs. N1+N2), p=0.0020, Fig. 4D–F] and a higher TNM disease stage [p=0.0031, the Fisher's exact test, Table IV; confirmed by the Mann-Whitney U test: 1.12±0.14 vs. 0.46±0.10 (I+II vs. III+IV), p=0.0014, Fig. 4J–N].

To estimate the significance of the PLAGL1 expression level as a prognostic factor, all patients were followed up for 36.1 months. During this observation period, 43 (35.5%) patients died.

The levels of PLAGL1 mRNA and protein expression, or the intensity of PLAGL1 immunostaining did not correlate significantly with the patient overall survival (Table V; Fig. 6A–C); however, the hazard ratio (HR) for patients whose tumor tissues showed reduced immunostaining of the PLAGL1 protein was twice higher than in patients whose tumor tissues revealed increased PLAGL1 immunoreactivity (Table V). Of the analyzed demographic and clinicopathological parameters, advanced age at diagnosis (>67 years; p=0.0014; Table V and Fig. 6D), depth of invasion (p=0.0477; Table V and Fig. 6E), lymph node involvement (p<0.0001; Table V and Fig. 6F), the presence of metastases (p<0.0001; Table V and Fig. 6G), and a high TNM disease stage (p<0.0001; Table V and Fig. 6H) were associated with poor patient outcome.