Abstract
Background/Aim: CD133 is an important, but not exclusive, biomarker of colorectal cancer (CRC) stem cells. Materials and Methods: In order to identify other CRC stem cell-specific genes, we performed a comparative expression profiling of CD133+ and CD133- cell populations in primary and metastatic tumors from four patients with CRC. CD133+ and CD133- CRC cells were isolated using MagSweeper and used for whole-transcriptome analysis with RNA-Seq. Results: We found that in CD133+ cells, 17 genes (RNASE2, PRB2, IL4, MGC27382, CLEC4C, SALL3, GIMAP1, ISG15, LOC728875, ZIK1, ICAM2, CCDC7, CDYL2, LRRC2, ZEB1, OSTF1 and CCDC144B) were significantly up-regulated compared to CD133- CRC cells. Among them, IL4 has been known as an inducer of survivin implicated in the survival and proliferation of cancer cells. However, the prognostic value of survivin in CRC is controversial. We evaluated survivin expression in formalin-fixed paraffin-embedded tumor samples of 188 patients with CRC by immunohistochemistry. Survivin over-expression was detected in 85 patients (45.2%) and was significantly associated with primary tumor sites (p=0.028), lymph node metastasis (p=0.029) and advanced III/IV CRC stages (AJCC 7; p=0.001). Furthermore, survivin up-regulation correlated with reduced disease-free survival (DFS; p=0.021) and overall survival (OS; p<0.000) and was proved to be an independent prognostic factor for both DFS and OS in multivariate analysis. Conclusion: Our data suggest that CD133+ CRC stem cells have a distinct expression pattern and that survivin, up-regulated by differentially expressed IL-4, is a candidate biomarker for the prediction of recurrence and survival in CRC.
Colorectal cancer (CRC) is the fourth most common cancer in men and the third most common cancer in women worldwide (1-3). Although progress has been made in CRC treatment over the last several years (4-6), the mortality rate remains very high. Thus, in South Korea, the overall five-year survival for patients with CRC is approximately 70%; for those with advanced CRC it is less than 10%. There is increasing evidence that a small population of tumor cells with stem cell-like properties is responsible for tumor initiation, progression and recurrence. These cells are capable of proliferation, self-renewal and differentiation into other cell types, while the dysregulation of their dynamics is implicated in cancer progression and metastasis (7-10).
CD133, a pentaspan transmembrane protein, is one of the biomarkers used to identify and isolate stem cells from cancer tissues, including those of CRC (7, 8, 11, 12). A previous study showed that a CD133+ cell sub-population in CRC can initiate xenograft tumors in immunodeficient mice; however, CD133-deficient cells are also capable of tumorigenesis, suggesting that CD133 is not the only molecule responsible for the tumor-initiating stem cell phenotype in CRC (13). Therefore, expression profiling of CD133+ and CD133- intestinal stem cells should help in the accurate definition of carcinogenic stem cell population in CRC. Given that cancer stem cells are resistant to conventional chemoradiotherapy, the characterization of CRC-initiating stem cells could suggest a novel therapeutic approach that specifically targets this cell group.
CRC stem cells are known to acquire resistance to death stimuli through the autocrine production of interleukin-4 (IL-4) (14, 15). IL-4 induction results in the activation of the signal transducer and activator of transcription 6 (STAT6) (16) shown to regulate the expression of genes involved in the suppression of apoptosis and promotion of metastasis (17, 18). Survivin is a member of the inhibitor of apoptosis (IAP) family implicated in the regulation of cell division and expressed in fetal tissues and human cancers (19-21). IL-4/STAT6 activation causes the up-regulation of survivin expression in CRC stem cells (15, 22), suggesting that survivin contributes to apoptosis resistance of CRC stem cells. This notion is supported by several reports showing that survivin expression correlates with poor survival of patients with several types of malignancies including non-small cell lung, breast and gastric cancers (23-26). However, the data on the role of survivin as a stem cell marker in CRC prognosis are controversial.
Molecular profiling of rare cells is important in biological and clinical studies. MagSweeper technology, based on immunomagnetic separation, is used for the purification of rare cells, including stem cells present in a mixed population; it is an efficient technique to isolate highly enriched populations of rare cells (27, 28) without compromising gene expression (29). The advanced technologies of targeted isolation and comparative genomic analysis of small cell populations provide an opportunity for the development of more personalized approaches to target specific genes and signaling pathways involved in cancer progression.
In the present study, we used the MagSweeper technique and high-throughput RNA expression analysis to assess the difference in expression profiles between CD133+ and CD133- CRC stem cells. Furthermore, we evaluated the impact of IL-4-regulated survivin on cancer recurrence and survival in 188 patients with CRC.
Materials and Methods
Ethics statement. The Institutional review board of the Samsung Medical Center approved the process of acquiring tissue samples from patients with CRC. All study participants provided a written informed consent also approved by the Institutional review board.
Patients and sample collection. The patients participated in this study had advanced stages of CRC and all specimens were collected before chemotherapy or radiotherapy. For the purification of CD133+ and CD133- CRC cells by the MagSweeper method, the specimens were obtained from four patients with CRC, including a primary tumor sample from one patient, metastatic tissue sample from another patient and both primary and metastatic tumor samples from the other two patients. Survivin expression was analyzed by immunohistochemistry (IHC) in 188 CRC tissue samples. All samples were characterized by the patient’s age at diagnosis, gender, primary site, operation, cancer stage, pathological differentiation, lymphovascular and perineural invasion and adjuvant therapy. All hematoxylin and eosin-stained slides were examined and representative paraffin blocks were selected for further studies.
Cell isolation using MagSweeper. Streptavidin-coated MyOne magnetic beads (Life technologies, Carlsbad, CA, USA) were functionalized by biotinylated CD133 antibodies. Tumor cell suspensions were incubated with the functionalized beads for 30 min at 4°C and the magnetically-labeled CD133+ cells were purified using the MagSweeper technology; CD133+ and CD133- cells were counted under an inverted light microscope (Olympus, Center Valley, USA). To confirm the purity of the isolated cells, an aliquot of CD133+ population was stained by a CD133-specific antibody and analyzed by fluorescence microscopy (Olympus, Center Valley, PA, USA).
RNA-Seq and data-analysis. Total RNA was isolated from CD133+ and CD133- CRC cells of each specimen and reverse transcribed using the SMARTer Ultra Low Input RNA for Illumina Sequencing kit (Clontech, Mountain View, CA, USA). cDNA was amplified using the Advantage 2 PCR kit (Clontech) for 18-25 cycles prior to the conversion into an Illumina-compatible DNA sequencing library using the Nextera DNA Sample Prep Kit (Illumina, Hayward, CA), which was amplified by 12 cycles of PCR. Libraries were quantified by using a BioAnalyzer (Agilent, Waldbronn, Germany) and by qPCR using a KAPA SYBR Green PCR kit (Kapa Biosciences, Foster City, CA) and an Illumina ECO qPCR system. Paired end flow cells were prepared using 8 pM of Nextera library per lane on a cBot (Illumina) and sequenced using single 50-bp reads on an Illumina GAIIx instrument. The Bowtie and NEUMA programs were used for mapping and quantification, respectively, of RNA-Seq (30, 31). After replacing 0 values with 0.001, the read values were log-transformed and 21,225 genes and six CD133-paired samples were included for further statistical analysis.
Statistic tests for clustering. For each patient, CD133+ and CD133- sites were indexed as k=1 and k=2, respectively. For each gene, χkij denoted the expression level from the sample of site j(=1,...mj) of patient i(=1,...n) with CD133 outcome, k(=1,2), where mj=1 or 2 depending on whether patient had only primary tumor or both primary and metastatic sites.
mi denoted the total number of sites from n patients; in our dataset, n=4 and M=6.
Let μk=E(χkij) Under H0: μ1=μ2, 
is approximately N(0,1) where
refer to Liang and Zeger. For each gene, a two-sided p-value was calculated from the N(0,1) distribution.
. Isolation of CD133+ stem cells and CD133- cells in tumor specimens of four patients with colon cancer.
Survivin detection using immunohistochemistry. Immunohistochemistry (IHC) was performed on formalin-fixed, paraffin-embedded, 4-μm-thick tissue sections. The sections were deparaffinized three times in xylene for a total of 15 min, rehydrated and antigen retrieval was carried out in ER2 buffer at 97°C for 20 min. After blocking the endogenous peroxidase activity with 3% hydrogen peroxidase for 10 min, the slides were incubated with a rabbit monoclonal antibody against survivin (1:5,000; Abcam, Cambridge, UK) for 15 min at room temperature. Immunostaining was analyzed by a Bond-max autoimmunostainer (Leica Biosystems, Melbourne, Australia) with the Bond™ Polymer refine detection system (Vision Biosystems, Melbourne, Australia). Staining for survivin was considered positive when more than 90% of tumor cells showed strong nuclear reactivity.
Statistical analysis. Correlations between survivin expression and clinicopathological parameters were analyzed using the Kruskal-Wallis test. Disease-free survival (DFS) and overall survival (OS) were analyzed by the Kaplan-Meier estimation and log-rank testing. The Cox proportional hazards model was used for the multivariate analysis to assess the independent effects of survivin and to obtain hazard ratio (HR) estimates. p<0.05 values were considered significant.
Results
Isolation of CD133+ and CD133- CRC cells. CD133+ and CD133- CRC cells were isolated in six tumor specimens of four patients with CRC by the MagSweeper technique. Table I shows the total numbers of CD133+ and CD133- cells isolated in each specimen. There was a significant variation in the percentage of CD133+ cells among patients and within a patient depending on the site of tissue collection. Thus, in CRC tissues, patient 2 had 27.3% of CD133+ cells, whereas in the liver it was only 6.0%. For each specimen, RNA extracted from the purified CD133+ and CD133- CRC cell populations was used for gene expression profiling.
Gene-expression profiles in CD133+ and CD133- CRC cells. To determine the specific difference between CD133+ and CD133- stem cell phenotypes in CRC, the expression profiles were compared in these two cell populations. The expression of 21,225 genes was analyzed for paired CD133+/CD133- cells in six CRC specimens and an unsupervised clustering analysis was performed (Figure 1). We identified 17 genes that were significantly over-expressed in CD133+ compared to CD133- CRC cells (Table II): RNASE2, PRB2, IL4, MGC27382, CLEC4C, SALL3, GIMAP1, ISG15, LOC728875, ZIK1, ICAM2, CCDC7, CDYL2, LRRC2, ZEB1, OSTF1 and CCDC144B. Among them, the expression of IL4 in CD133+ cells exceeded that in CD133- CRC cells by 43-fold. Given that the IL-4-survivin pathway has been associated with “stemness” in CRC in a previous study (REF), we next investigated survivin as a potential biomarker of CD133+ CRC stem cells.
Survivin expression in CRC. Survivin expression was evaluated in 188 patients with CRC (114 male and 74 female patients; median age=59 years, range=28-84 years). Among these patients, 23 (12.2%), 79 (42.0%), 83 (44.1%) and 3 (1.6%) had stages I, II, III, and VI CRC, respectively, of well-to-moderate (171; 91.0%) to poor (11; 9.0%) differentiation. Three stage IV patients had CRC with resectable liver metastasis and were subjected to curative surgery for primary and metastatic disease. The clinicopathological characteristics of all 188 patients are summarized in Table III. Survivin over-expression (+) was observed in 85 of 188 patients (45.2%). The analysis of survivin association with clinicopathological features (Table IV) indicates that survivin expression significantly correlated with primary tumor location (p=0.028), lymph node metastasis (p=0.029) and CRC stage (AJCC 7; p=0.001).
Unsupervised clustering of over-expressed genes of CD133+ CRC stem cells and CD133- CRC cells isolated from tumor specimens. Red indicates high gene expression, green indicates low expression and black represents undetectable expression.
Survivin as a biomarker for CRC survival. The Kaplan-Meier survival analysis showed a strong correlation of survivin expression with a shorter disease-free (DFS) and overall (OS) survival. Univariate analysis indicated that survivin expression, poor differentiation, perineural invasion, stage III/IV, lymph node metastasis and absence of adjuvant chemotherapy were associated with shorter DFS and OS. Multivariate analysis showed that survivin expression (p=0.039), perineural invasion (p<0.001) and lymph node metastasis (p=0.032) were independent factors associated with shorter DFS (Table V), whereas survivin positivity (p=0.001), perineural invasion (p=0.024) and lymph node metastasis (p=0.019) were independent factors associated with shorter OS (Table VI).
Kaplan-Meier survival curves, (a) disease-free survival (DFS) and (b) overall survival (OS) of CRC patients grouped by survivin IHC (+) vs. (-) in tumors.
Over-expressed genes in CD133+ CRC stem cells compared CD133- CRC cells.
Discussion
In the present study, we generated expression profiles and identified some functional roles of stem cells in CRC. We found that a number of genes such as RNASE2, PRB2, IL4, MGC27382, CLEC4C, SALL3, GIMAP1, ISG15, LOC728875, ZIK1, ICAM2, CCDC7, CDYL2, LRRC2, ZEB1, OSTF1 and CCDC144B had higher levels in CD133+ CRC stem cells than in the CD133- CRC population. Among these genes, IL4 has been known as a positive regulator of survivin, a biomarker of an anti-apoptotic stem cell-like phenotype, which, in this study, was found to be strongly associated with more aggressive stage (III/IV CRC (p=0.001)) characterized by lymph node metastasis (p=0.029). Univariate and multivariate analyses revealed that survivin expression was an independent prognostic factor for both DFS and OS.
Recent studies have shown that a small population of stem-like cells, capable of proliferation and self-renewal, is responsible for the initiation and progression of various cancer types (7, 8), as well as tumor metastasis and resistance to chemotherapy and radiotherapy (9, 10). CD133 has been recently proposed as an important biomarker of colon cancer stem cells (11); however, CD133 expression does not identify the entire population of tumor-initiating cells in CRC. In order to comprehensively characterize the CRC stem cell phenotype, we generated and compared gene expression profiles of CD133+ and CD133- cells by using whole-transcriptome RNA-Seq technology. The isolation of CD133+ CRC stem cells from surgical specimens is a challenge because the CD133+ population in CRC is very small. To circumvent this obstacle, we used MagSweeper, an advanced technology based on targeted immunomagnetic separation of rare cells with high specificity (27). Previous studies have shown that MagSweeper can be reliably used to extract circulatory tumor cells from the blood of cancer patients (32) and does not significantly alter gene expression (29). By using MagSweeper, herein we could effectively separate CD133+ and CD133- populations of CRC cell in each surgical specimen and generate transcriptome profiles by RNA-Seq to identify differentially expressed genes.
. Baseline characteristics of CRC patients (N=188).
The survivin gene, located in chromosome 17q25, is a member of the IAP family expressed predominantly in fetal tissue but is also found in many common tumors, including colon cancer (33). Survivin is implicated in multiple signaling networks essential for tumor progression and metastasis, including cell division and stress response pathways (34). In CRC stem cells, survivin has been shown to be up-regulated by IL-4 (15) and suggested as one of CRC stem cell biomarkers. Our comprehensive transcriptome analysis by RNA-Seq showed that CD133+ CRC stem cells over-expressed IL-4, which could induce survivin up-regulation previously demonstrated in CRC stem cell. These observations suggest a potential prognostic value of survivin expression in patients with CRC. Indeed, our results indicate that survivin over-expression appeared to be a significant independent predictive factor for tumor recurrence and overall survival of patients with CRC and could be used as a molecular target for anti-cancer therapies in CRC.
Correlation between survivin expression and clinicopathological parameters in CRC patients (N=188).
By using RNA-Seq transcriptomic analysis, we also identified other potential biomarkers of CRC stemness. However, the comprehensive characterization of a tumor-initiating cell population in CRC requires additional analysis of all genetic aberrations, including mutations and translocations, as well as gene interactions, which could significantly contribute to stem cell-like CRC phenotyping but cannot be assessed by RNA-Seq analysis. In our future studies, we plan to analyze CD133+ and CD133- cells by whole-genome sequencing, which will help to further elucidate functional mechanisms involved in cancer promotion by stem cells. Another limitation of this study is the small sample size: only six specimens from four patients with CRC were used in the RNA-Seq analysis, which is likely to be a confounding factor in data interpretation.
In conclusion, we identified the genes, including IL4, that were up-regulated in the CD133+ CRC stem cell population compared to CD133- CRC population. Moreover, survivin, known to be positively regulated by IL-4, was found to have an independent prognostic value for tumor recurrence and survival of patients with CRC. Complete genetic characterization of CRC stem cells may be fundamentally important for the development of individualized approaches to the treatment of patients with CRC.
Conflicts of Interest
The Authors declare that they have no conflict of interest.
Univariate and multivariate analysis for disease-free survival (DFS).
Univariate and multivariate analysis for overall survival (OS).
Acknowledgments
This study was supported by a grant from the Samsung Biomedical Research Institute (#SBRI-CB11092, CB0220 and #MX1132701).
Footnotes
-
Seung Tae kim, Insuk Sohn, In-Gu Do and Jeeyun Lee contributed equally to this work.
- Received August 12, 2014.
- Revision received September 12, 2014.
- Accepted September 15, 2014.
- Copyright © 2014 The Author(s). Published by the International Institute of Anticancer Research.
References
In this issue
Jump to section
Related Articles
Cited By...
- DPHL v2: An updated and comprehensive DIA pan-human assay library for quantifying more than 14,000 proteins
- Prognostic Significance of Survivin Expression and Combined Analysis with Cancer Stem Cell and Epithelial-Mesenchymal Transition-related Markers in Patients with Rectal Cancer Undergoing Preoperative Chemoradiotherapy