Background
The Polycomb-group (PcG) proteins function as epigenetic transcriptional regulators through multiple mechanisms. PcG proteins are mainly categorized into two groups, polycomb repressive complex 1 (PRC1) and 2 (PRC2) [
1]. CBX7 is a Polycomb protein that has shown tumor suppressive function and is a component of PRC1. Recent studies have utilized
cbx7-knockout mice to validate the tumor suppressor role of
cbx7 in liver and lung carcinogenesis [
2]. However, analysis of many types of cancers revealed that the
Cbx7 expression levels were significantly altered, but the changes were often considerably contradictive. While some studies showed that
Cbx7 behaved as an oncogene in lymphoma, prostate cancer, ovarian cancer, and gastric cancer [
3-
6], other studies indicated
Cbx7 was acting as a tumor suppressor gene in the bladder, colon, pancreas, and thyroid cancers [
7-
11]. Currently, the reasons leading to these conflicting results have not been investigated.
Colon carcinoma (CC) is the third most common cancer and the fourth leading cause of cancer death worldwide [
12]. One large-scale patient study indicated that the loss of
Cbx7 expression in CC correlates with poor prognosis and short survival using results analyzed through immunohistochemistry in tissue microarray [
11]. Using quantitative RT-PCR (qRT-PCR), similar results were obtained among 22 Chinese CC patients [
13]. The qRT-PCR assay offers significant advantages over the hemi-quantitative immunohistochemistry assays for detecting
Cbx7 expression changes in cancers. In the present study, a qRT-PCR assay capable of correlating
Cbx7 mRNA level with protein expression was developed, and the
Cbx7 transcription levels from 97 CC patients were quantified and statistically analyzed for clinicopathological correlations.
Methods
Clinical samples
CC and surgical margin (SM) tissue specimens were obtained from surgical resection patients without neo-adjuvant therapy (n = 97) in Peking University Cancer Hospital (Beijing, China) from 2004 to 2011. The patient population contained 49 males and 48 females with a mean age of 62 years (range 34–89), who were not included in the previous study [
13]. Tumors were staged using the tumor-node-metastasis (pTNM) staging of the International Union Against Cancer (2003) [
14]. The number of patients staged from I to IV were 1, 33, 36 and 27, respectively (Table
1). The SM tissues were more than 5 cm from the tumor and were validated by an experienced pathologist. None of the patients received preoperative chemotherapy. All patients had follow-up data for at least 36 months (
median 61). 49 CC patients were post-operatively treated with adjuvant chemotherapy (Folfox for 30 patient). 45 patients (46.4%) suffered recurrent CC and 52 patients (53.6%) died during follow-up. Paraffin blocks were selected from suitable formalin-fixed paraffin-embedded tissue with an average age of 62.2 years (range 34 ~ 89). Normal colon biopsies were obtained from non-cancerous patients (n = 51). The Institutional Review Boards of Peking University Cancer Hospital and Institute approved the study. All samples were obtained with the patients’ informed written consent.
Table 1
Association of
Cbx7
mRNA level and clinicopathological features of colon cancer patients
Age (years) | ≤62 | 42 |
2.69 (1.07-4.61)
b
| 2.42 (1.31-4.88) |
| >62 | 55 |
1.38 (0.44-3.37)
| 1.72 (0.98-4.43) |
Sex | Male | 49 | 2.04 (0.74-4.54) | 1.77 (0.97-4.00) |
| Female | 48 | 1.60 (0.90-3.40) | 2.22 (1.31-5.29) |
Location | Ascending/Transverse | 42 | 1.94 (1.05-4.52) | 1.70 (1.27-3.27) |
| Descending/Sigmoid | 37 | 1.90 (0.72-4.61) | 1.94 (1.04-6.26) |
| Undefined | 18 | 1.29 (0.61-2.75) | 2.42 (1.20-7.74) |
Differentiation | Poor | 24 | 1.60 (0.89-3.59) | 1.60 (1.31-3.54) |
| Moderate/Well | 70 | 2.02 (0.76-4.59) | 2.22 (0.949-6.05) |
Vascular invasion | Absent | 69 | 1.95 (0.79-4.48) | 1.94 (1.27-5.25) |
| Present | 26 | 1.74 (0.69-4.13) | 1.90 (1.13-4.71) |
pTNM | I&II | 39 | 2.06 (1.01-4.54) | 1.94 (0.66-5.41) |
| III&IV | 58 | 1.63 (0.69-3.92) | 1.90 (1.30-4.66) |
Invasion | T1 ~ 2 | 5 | 5.77 (4.48-8.25) | 1.31 (0.30-3.09) |
| T3 | 51 | 1.40 (0.38-3.73) | 2.39 (0.94-5.83) |
| T4 | 41 | 1.83 (0.96-3.81) | 1.82 (1.31-3.39) |
Lymph metastasis | Negative | 50 |
2.54 (1.07-4.57
)
b
| 1.77 (1.18-4.63) |
| Positive | 47 |
1.29 (0.24-2.94)
| 2.22 (1.12-4.85) |
Distant metastasis | Negative | 69 | 2.04 (0.96-4.27) | 1.94 (1.08-4.95) |
| Positive | 29 | 1.31 (0.42-4.56) | 1.97 (1.22-5.15) |
Post-operative therapy | No | 48 | 1.93 (0.75-3.77) | 1.75 (1.03-5.03) |
| Yes | 49 | 1.64 (0.94-4.48) | 2.22 (1.27-4.00) |
qRT-PCR
Total RNA was extracted from 30–50 mg of tumor tissue using a commercial RNA isolation kit according to the manufacturer’s protocol (Ultrapure RNA Kit, CWBIO, Beijing). Subsequently, the RNA concentration was checked using 1.0% agarose gel electrophoresis stained with 0.5 μg/ml ethidium bromide and quantified with a NanoVue spectrophotometer (GE Healthcare). For reverse transcription, 1 μg RNA, 20 units reverse transcriptase, 1× reaction buffer, 1 mM deoxynucleotides, 3 mM MgCl2, and 4.0 mg of random hexamers were used. The reaction mixtures were incubated at 25°C for 10 min, 42°C for 1 h, and 95°C for 5 min according to the manufacturer’s protocol (Improm-II Reverse Transcription System A3800, Promega, USA). The cDNA was stored at −20°C.
qRT-PCR was performed using an ABI 7500 Fast Realtime System (Applied Biosystems, Foster City, CA, USA). Primers and a TaqMan probe for
Cbx7 were designed and synthesized according to the Taqman Gene Expression Assay (Roche Diagnostics, Mannheim, German). The primer sequences follow: human
Cbx7 gene 5′-cgtcatggcctacgagga-3′ (sense), 5′-tgggtttcggacctctctt-3′ (antisense); TaqMan probe 5′-FAM-aggaggag-TEMER-3′ [
8,
15];
GAPDH 5′-gaaggtgaaggtcggagt-3′ (sense) and 5′-gaagatggtgatgggatttc-3′ (antisense); the
Alu elements 5′-gaggctgaggcaggagaatcg-3′ (sense), 5′- gtcgcccaggctggagtg-3′ (antisense) [
16]. PCR reactions were carried out in a final volume of 10 μL containing 5 μL Maxima Probe/ROX qPCR Master Mix (2×) (K0233, Thermo Scientific), 0.5 μM of each primer and DNase-free water. The PCR conditions were 5 min at 95°C, followed by 40 cycles of 95°C for 15 s, 54°C for 30s, and 72°C for 35 s and finished with a melting curve analysis. The relative copy number [2
-ΔCT] of
Cbx7 mRNA was determined from the difference in cycle threshold (CT) values between the target and reference genes.
Immunohistochemistry (IHC)
The paraffin was removed from the embedded CC and SM tissue samples using xylene. The samples were then rehydrated in a graded series of ethanol solutions. Antigen retrieval was performed in Tris/EDTA (pH 9.0) for 3 min at 120°C. The sections were then incubated for 20 min in 3% H
2O
2 and washed with 0.025% Triton X-100/TBS (TBST). Blockage was performed with 10% goat serum for 2 h at room temperature. The slides were then incubated overnight at 4°C with anti-CBX7 monoclonal antibody (ab21873, Abcam, Cambridge, UK). Subsequently, the sections were incubated with an HRP-conjugated anti-mouse EnVision system (DAKO, Glostrup, Denmark) for 20 min at 37°C followed by staining with diaminobenzidine hydrochloride (DAB, DAKO). Normal mouse IgG was applied as negative control (Additional file
1: Figure S1). The sections were counterstained with hematoxylin. The intensity of nuclear CBX7-staining in the epithelial and stromal cells was grouped as negative (−), weak (visible at high magnification = +1), moderate (visible at low magnification = +2), or strong (strikingly positive at low magnification = +3).
Western blot
Whole-cell protein extracts were prepared from primary tumors. The samples underwent electrophoresis in a 12% SDS-PAGE gel followed by blotting onto a Polyvinylidene-Fluoride membrane (Bio-Rad). The membranes were incubated in PBS containing 5% skim milk and 0.05% Tween-20 for 2 h at room temperature, then probed overnight with anti-CBX7 antibody (1:1000) (Abcam) in blocking solution at 4°C. An HRP-labeled goat anti-rabbit secondary antibody was then used (DAKO K5007, Glostrup, Denmark). The results were visualized on a Fluor chem system (Cell Biosciences). Integrated Option Density (IOD) was used to quantify the amounts of CBX7 and β-Actin. β-Actin was used as a reference control. In order to normalize the levels of CBX7 between samples, the IODCBX7/IODReference ratio was calculated.
Cbx7 plasmid construction and transfection
The coding region of
Cbx7 was inserted into the pEGFP-C1 vector and used to transfect cultured cells as described previously [
15].
Transwell migration and matrigel invasion assays
The migration and invasion capacity of colon cancer cell lines SW480 (5 × 10
4 cells/well) and HCT116 (4 × 10
4 cells/well) (kindly provided by Dr. Yuanjia Chen at Peking Union Medical College Hospital) were tested using the Transwell migration and invasion assays 48 hrs after transiently transfection with the
Cbx7 plasmid or empty vectors for 48 hrs. [
15].
SW480 cells (2 × 106 cells in 0.2 ml) were injected into SCID mice via the tail vein 48 hrs after transient transfection with the Cbx7 plasmid or empty vectors. The mice (8/Group) were harvested during the 6th experimental week. Chest wall, number of pulmonary metastasis tumor nodules, and the lung weight were then measured for each mouse. The lung organs were fixed with Bouin solution, paraffin-embedded and cut into 5 μm slides along the maximum area, and examined microscopically following H.E. staining.
Statistical analysis
All statistical analyses were performed using SPSS software (SPSS version 17.0). The correlation between CBX7 protein level and mRNA level was analyzed using the Spearman’s rank correlation coefficient or the Pearson product–moment correlation coefficient. The nonparametric Wilcoxon test, Kruskal-Wallis test, and Mann–Whitney test were applied to evaluate the association between Cbx7 transcription and clinicopathological tumor features. Kaplan–Meier survival curves were generated and compared using the log-rank test. A multivariable Cox regression model was applied to determine if a particular factor was an independent predictor of survival in multivariate analysis. All statistical tests were two-sided, and P-values < 0.05 were considered statistically significant.
Discussion
The
Cbx7 expression levels in human cancers are highly variable and often contradictive. The organ/tissue-specificity, assay used for analysis, and sample differences may account for these conflicting results. In the present study, it was determined that by using
Alu as the normalization control,
Cbx7 mRNA levels in the CC and SM tissues were positively and significantly correlated with the amount of CBX7 protein detected by both IHC and Western blot. Such a correlation was not observed when
GAPDH was used as the normalization control. By using the
Alu RNA reference, it was found that
Cbx7 transcription was significantly downregulated in CC and
Cbx7 expression-low tumors contributed to a high risk of lymph metastasis and poor overall survival in Chinese CC patients. These results are consistent with a previous report, which showed that loss of CBX7 expression in CC correlates with poor outcomes based on large-scale patient analysis by IHC in tissue microarray [
11].
GAPDH is the traditional reference used in qRT-PCR to normalize mRNA levels in cell/ tissue samples containing different cell numbers. It is especially effective for samples of homogeneous cellularity/constitution; however, its reliability for normalization in samples that are cellularly heterogenic is problematic. This shortcoming is especially highlighted when attempting to compare gene mRNA levels in cancer tissues with corresponding normal tissues. A number of strategies have been proposed for proper data normalization in qRT-PCR [
16-
18]. There are about 800,000 copies of
Alu elements in each haploid of the human genome. More than one million copies of the
Alu element (about 300 bp) are located in untranslated regions (UTRs) of 1,500 genes. Therefore, transcriptional dysregulation of a few
Alu elements has no significant affect on altering the total amount of
Alu transcripts in the whole human transcriptome. This unique characteristic makes
Alu transcripts a good reference gene for qRT-PCR. It is known that
Alu elements are globally hypomethylated in cancers [
19]; however, it is not known if the hypomethylation leads to upregulation of
Alu expression in the genome and subsequently decreases its reliability. The high consistency between CBX7 protein level and
Alu-normalized
Cbx7 mRNA level in colon tissues found in the present study is in strong agreement with recent reports suggesting that
Alu RNA is a reliable reference in qRT-PCR analysis [
16], and could be particularly beneficial for the evaluation of gene expression changes between cancer and normal tissues.
It has been reported that
Cbx7 transcription was downregulated among 22 Chinese CC patients using the typical
GAPDH-normalized qRT-PCR analysis [
13]. Although a similar result was also observed among the 10 pairs of CC and SM samples in the present study, the
GAPDH-normalized
Cbx7 mRNA levels did not correlate with the amount of CBX7 protein detected through IHC and Western blot analysis. This implies that the observed difference in
GAPDH-normalized
Cbx7 mRNA levels between CC and SM samples may not be an accurate indicator of protein level. Although other factors, such as stability and degradation differences, could have played a role in the inconsistency observed between mRNA and protein levels of a gene in tissues of interest, the high consistency and repeatability seen using the
Alu-normalized
Cbx7 mRNA and CBX7 protein levels shows that selecting a suitable reference is very critical in drawing a reliable conclusion. Use of an unreliable reference gene for qRT-PCR analysis may contribute more to the noted inconsistency than has previously been estimated. It is increasingly necessary to determine if the correct normalization control is being used when qRT-PCR assays are utilized in studies on tumor biology [
20,
21].
Interestingly, although the average Cbx7 mRNA level in SMs was significantly higher than CCs, it was also significantly higher than in normal colon control tissues from non-cancer patients. Age also seems to be associated with the levels of CBX7. The average level of Cbx7 mRNA in CC and SM samples from patients ≤62 yrs was significantly higher than was observed in older patients. Therefore, because the CC patients were much older than the controls, the Cbx7 mRNA level in the normal colon tissues distant from the site of malignancy (not available in the present study) should be lower than that from the non-cancer controls. In other words, the expected Cbx7 expression difference between the SMs and distant normal colon from the CC patients might be higher than the observed difference between the SMs and normal colon biopsies from non-cancer patients. This suggests that alteration of Cbx7 expression in colon carcinogenesis may be far more complex than expected.
According to the map of human proteome [
22], CBX7 is highly expressed in B cells (Additional file
3: Figure S3). In the present study, strong nuclear CBX7 staining was frequently observed in the lymphoid cells in the SMs and the normal colon biopsies, but not in CC samples. It is likely that both the decreased number of CBX7-positive lymphoid cells in the CC tissues, and the increased number of lymphocytes infiltrating into stromal tissues around cancer cells contributes to the downregulation of
Cbx7 expression in CCs. It cannot be excluded that
Cbx7 may be upregulated in SMs as a B-cell infiltration-related host response to the presence of cancer cells. More studies are necessary to further elucidate the role of
Cbx7 in tumor development and modulation.
Contribution of CBX7 to cancer development may be organ-dependent [
3-
11]. CBX7 is suggested to help suppress the progression of human colon cancers [
10,
13]. The results of our study were in close agreement with this conclusion. However, the results of the Transwell tests and mouse experimental pulmonary metastasis assay showed that enforced
Cbx7 overexpression might slightly increase the migration/invasion capacity of colon cancer cells. This suggests that the role of the exogenous
Cbx7 overexpression in epithelial cancer cells might be different from that of the endogenous
Cbx7 in colon cancer. Whether downregulation of
Cbx7 in stromal lymphoid cells involves in the progression of colon cancers should be studied further.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interest
The authors declare that they have no competing interests.
Authors’ contributions
XZ carried out the IHC and WB analysis, participated in the design of the study, performed the statistical analysis, and drafted the manuscript. JZ extracted protein and RNA, and carried out the qRT-PCR assay. BZ carried out Transwell and experimental pulmonary metastasis tests. JZ participated in the collection of clinical samples and their clincopathological information and statistical analysis. JW performed English language editing. LG participated in the collection of clinical samples and their clincopathological information. BZ collected the biopsy samples. JG and JJ treated the patients and collected the surgical samples. DD conceived the study, participated in its design and coordination, and revised the manuscript. All authors have read and approved the final manuscript.