SEARCH

SEARCH BY CITATION

Keywords:

  • colon cancer;
  • gene profile;
  • rectal cancer;
  • signal pathway

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFFERENCES

OBJECTIVE:  Colon cancer is more common in the USA and Europe than that in China, for reasons that are unclear. The aim of this study was to investigate the differences in gene expression profiles and carcinogenesis pathways between colon and rectal cancer.

METHODS:  Expression profiling of primary tumor tissues from 12 colon and 12 rectal cancers was performed using oligonucleotide microarray analysis. All samples were strictly matched by clinical features. Bioinformatic analyses such as the Kyoto Encyclopedia of Genes and Genomes were used to identify genes and pathways specifically associated with colon or rectal cancers.

RESULTS:  A total of 824 genes were differentially expressed in colon and rectal cancers. All differential gene interactions in the Signal-Net were analyzed. More genes were differentially expressed and included in the Signal-Net for rectal than colon cancer. Of the genes differentially expressed between colon and rectal cancer, S100P, the Reg family, ACTN1, CAMK2G and ACAT1 were the most significantly altered. Genes involved in the cell cycle pathway were present in rectal and colon cancers, but were more important in rectal cancer. The p53 and metabolic signaling pathways were significantly different in colon and rectal cancers. Gene expression profiles differed between colon and rectal cancer, with metabolic pathways being more important in rectal cancer.

CONCLUSION:  The oncogenesis of rectal cancer may be more complex than that of colon cancer. Some genes could be new biomarkers for distinguishing between these two cancers.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFFERENCES

Colorectal cancer is a common malignant tumor of the gastrointestinal tract. Morbidity due to colorectal cancer is increasing rapidly in China. Epidemiological evidence suggests that colon and rectal cancers differ in terms of their morbidities and etiology. Rectal cancer is more common in China, where it accounts for over 50% of colorectal cancer, compared with less than 30% in Western countries. However, this percentage decreased from 72.6% in the 1980s to 66.9% in the 1990s. The morbidity due to colon cancer increased by 78% in the last two decades, while that of rectal cancer increased by only 6%.1,2 Data from Peking Union Medical College Hospital indicated that colon and rectal cancers accounted for 55.6% and 44.4% of colorectal cancer, respectively, during 1989 through 2008.3 In contrast, colon cancer was shown to account for over 60% of colorectal cancer in the USA and Europe.4,5 These results suggest differences in the carcinogenesis of colon and rectal cancers.

A previous study has provided evidence for the molecular oncogenesis of colorectal cancer from adenomatous polyps, indicating the abnormal expression of many genes including APC and k-ras.6 However, these mechanisms mainly refer to colon cancer, and whether the same mechanisms apply to rectal cancer remains unknown. Most of the studies investigated the oncogenetic mechanisms of colorectal cancer with the combination of colon and rectal cancers. However, several studies7–9 have indicated differences in the epidemiology, etiology, clinical manifestations, pathological features and genetic abnormalities between colon and rectal cancer. Gene hybridization techniques have also shown the amplification of 20q in colon cancer, compared with the amplification of 12p in rectal cancer.10 Preliminary studies have supported the theory that the oncogenesis of left-sided and right-sided colorectal cancers may involve, at least partially, different mechanisms, and epidemiological evidence also suggests that the oncogenetic mechanisms of colon and rectal cancers should be studied separately.

The advent of genome-wide technologies, such as gene expression microarray, has made it possible to achieve a comprehensive view of the alterations involved in cancer. Although several results have been published based on human colon tumors,10–12 colon and rectal cancers have been mixed together in these studies and no reported studies have yet applied genome-wide technology to investigate the differences between colon and rectal cancer. In this study, we therefore investigated the differences in gene expression between colon and rectal cancer using gene expression microarray analysis.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFFERENCES

Patient selection

Surgical cancer tissue samples from 24 patients (12 right colon and 12 rectal cancers) were obtained at Peking Union Medical College Hospital in 2009, and kept in liquid nitrogen. The patients in each group were matched by age and gender, as shown in Table 1. Fresh tissue of tumor specimens and adjacent normal mucosa were identified under microscopy by an experienced pathologist. All tumors were adenocarcinomas at stages I–II, with no adjacent lymph node metastasis according to TNM Classification. Informed consent was obtained from all participants. The protocol for this study was approved by the Ethics Committee of Peking Union Medical College Hospital.

Table 1.  Clinical features of patients in rectal and colon cancer groups
 Rectal cancer (n)Colon cancer (n)
  1. SD, standard deviation.

No. of patients1212
Gender  
 Male66
 Female66
Age (years, mean ± SD)65.2 ± 4.565.7 ± 7.2
Cancer site  
 Rectal120
 Right colon012
Pathology  
 Well-moderate adenocarcinoma1212
Lymph node metastasis00
TNM stage (I-II)1212

RNA preparation, amplification and labeling

Total RNA was extracted from each frozen sample using TRIzol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions and purified using NucleoSpin RNA Clean-Up Kit (Macherey-Nagel, Duren, Germany). RNA amplification, cDNA labeling, hybridization and scanning were performed according to the Illumina TotalPrep RNA Amplification Kit (Illumina, San Diego, CA, USA) instructions. Briefly, single-stranded cDNA containing a T7 promoter sequence was synthesized using a T7 Oligo(dT) Primer (ABI, New York, NY, USA), and then converted into a double-stranded DNA (dsDNA) template for transcription from 5 µg total RNA. Multiple copies of biotinylated cRNA were then synthesized from the double-stranded cDNA templates.

Gene expression detection

Total genome gene expression was detected using Illumina HT-12 V4.0 Expression Beadchip oligonucleotide microarray (Illumina, San Diego, CA, USA). Labeled samples and the reference pool were hybridized with the microarray overnight at 42°C and unbound probe was removed by stringent washing. Arrays were scanned using a LuxScan 10 K/A confocal scanner (CapitalBio Corporation, San Diego, CA, USA), and the images were transformed into digital data using LuxScan 3.0 image analysis software (CapitalBio Corporation, San Diego, CA, USA). Microarray data were normalized by LOWESS normalization of each array using the R package.

Microarray data analysis of genes differentially expressed in colon and rectal cancers

Genes differentially expressed in normal, colon and rectal cancer tissues were identified using the TwoClassDif method. The differentially expressed genes for the normal tissue and cancer groups were then filtered by applying the Random Variance Model (RVM) t-test, which might increase the degrees of freedom in cases with small sample size. Significance and false discovery rate analyses were performed, and differentially expressed genes between colon and rectal cancer were then selected if P < 0.05.12,13

Differential pathway analysis based on microarray data

Pathway analysis was used to identify significantly different regulatory pathways according to the Kyoto Encyclopedia of Genes and Genomes (KEGG), Biocarta and Reatome. Fisher's exact test was used and the threshold for statistical significance was set at P < 0.05. The pathway distributions could then be constructed based on the data.14–16

Signal-Net analysis

Based on the observation of many protein-protein and protein-DNA interactions, which were experimentally validated, we first constructed an interactions repository derived from KEGG by MySQL 5.1.3 (Oracle, CA, USA). The repository is distinguishable from pathway maps on the KEGG website because the repository breaks down the protein interactions by pathway class and supplies the integrated protein interactions involved in multiply pathway categories. We were thus able to search for differential expression genes. Two nodes were connected when their corresponding encoded gene products were either directly connected or indirectly connected by a linker gene in the interaction network. The network for each gene was measured by counting the number of upstream and downstream genes or binding genes, which were expressed as in-degree and out-degree or degree, respectively. The higher degree of a gene indicates that it is regulating or being regulated by a greater number of genes, implying it has a more important role in the signaling network. A P value < 0.05 was considered statistically significant. Gene interactions could then be drawn based on the data.17,18

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFFERENCES

Overall differential gene expression profiles and carcinogenesis pathways between colon and rectal cancer

The DNA microarray data are shown in Table 2. Compared with normal tissue, a total of 676 genes related to 11 signal pathways were differentially expressed in colon cancer development, which was much less than the 1 789 genes and 30 signal pathways altered in rectal cancer. A total of 824 genes were identified as significantly differentially expressed, being either upregulated or downregulated, in colon and rectal cancers, in which more genes were upregulated than downregulated. Compared with normal tissue, there were more differential pathways and more cross-linked genes in the Signal-Net analysis for rectal cancer tissues than those for colon cancer tissues.

Table 2.  Overall differential gene expression profiles and carcinogenesis pathways between colon and rectal cancer
GroupNumber of significant differential genesNumber of significant differential pathwaysNumber of genes crosslink in Signal-Net
upperdowntotalupperdowntotal
  1. upper, upregulated; down, downregulated.

Rectal/colon611213824
Rectal/normal99679317891812302824
Colon/normal3952816765611214

Differential gene expression profiles between colon and rectal cancer

The changes in expression levels of the 824 differentially expressed genes differed. Tables 3 and 4 show the 10 genes that were the most obviously upregulated or downregulated according to their P values. A diagram of the gene interaction network was drawn up based on the genes differentially expressed in colon and rectal cancers (Fig. 1). There were many more genes involved in the Signal-Net analysis in rectal cancer, of which YUHAIB (upregulated), ACTN1 (downregulated) and CAMK2G (downregulated) were the three main genes, while the main regulatory genes included in the Signal-Net analysis in colon cancer were ACAT1 (downregulated), TUBA4A (upregulated) and TPM1 (downregulated).

Table 3.  The top 10 differential upregulated genes expression between colon and rectal cancer
Gene symbolP valueMean of intensities: rectalMean of intensities: colonFold change
RGS43.4094 × 10−571.167517.69404.0221
SGIP14.4737 × 10−547.460112.48203.8023
ANKRD380.00012581117.270320.01615.8588
OMD0.0001430536.492111.63433.1366
RRAD0.0001612464.638311.09095.8281
AQP10.0001718358.210611.00485.2896
AP1S20.00019011378.7380135.86822.7875
MEG30.00020284283.576979.74623.5560
MAP6D10.0003067484.561731.27302.7040
PLD60.00034789172.966960.38712.8643
Table 4.  The top 10 differential downregulated genes expression between colon and rectal cancer
Gene symbolP valueMean of intensities: rectalMean of intensities: colonFold change
S100P6.2745 × 10−51844.97366904.21220.2672
C4BPB6.4368 × 10−541.0429275.56130.1489
REG1A0.0001433279.55492010.71200.0396
TFF10.00025749131.03432930.57860.0447
NPDC10.00050186118.7494418.56330.2837
SYTL10.0005166917.3281105.02900.1650
REG1B0.0007832336.0111503.77560.0715
ALG1L0.0007840344.5817181.72100.2453
TLR40.0011040727.991484.29460.3321
LOC4007680.0013258415.800356.23340.2810
image

Figure 1. Network of differentially expressed genes in (a) colon cancer and (b) rectal cancer. Circles represent genes (red shows upregulated, blue shows downregulated). The size of the circle represents the number of other genes that interact with this gene. Lines indicate interactions between genes.

Download figure to PowerPoint

Differential carcinogenesis pathways between colon and rectal cancer

The oncogenetic pathways of colon and rectal cancers were analyzed according to the functions and interactions of the differential genes. Figure 2 shows the signal pathways in colon and rectal cancers. As for differential gene expression, the pathways were divided into upregulated and downregulated pathways. Some signal pathways were involved in both colon and rectal cancers, but their roles were different. As shown in Figure 2, the cell cycle pathway was involved in both rectal and colon cancers, but was more important in rectal cancer than that in colon cancer. The p53 signal pathway was the most strongly upregulated pathway in colon cancer. Metabolic pathways were more affected in rectal cancer, both upregulated and downregulated, than those in colon cancer.

image

Figure 2. Histogram of signal pathways that were significantly different in colon and rectal cancers showing (a) upregulated pathways in colon cancer; (b) downregulated pathways in colon cancer; (c) upregulated pathways in rectal cancer; (d) downregulated pathways in rectal cancer. X axis, negative logarithm of the P value (-LgP); Y axis, the name of the pathway. The larger the value, the smaller the P value.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFFERENCES

No data of the carcinogenesis of right-sided vs. left-sided colon cancers have yet been published. This study investigated variations in gene expression profiles and signal pathways between colon and rectal cancer using microarray techniques. The clinical data in the Peking Union Medical College Hospital showed that there was no significant difference in the gender of patients with colon cancer and rectal cancer.3 In order to study the difference of gene expression using microarray in the limited tissue sample, all affected factors were removed by strictly matching all clinical features. Furthermore, tumors and normal tissues were carefully identified by pathological findings.

The mechanisms responsible for colorectal cancer development have been well studied since 1998,6 including the progression from adenoma to cancer, and the involvement of various gene mutations, such as APC, k-ras and p53 during the different stages of carcinogenesis. However, no studies have demonstrated that the same mechanisms are involved in both colon and rectal cancers. The proximal colon, distal colon and rectum have different embryological origins. In addition, colon cancer is related to fatty foods, less exercise and a Caucasian ethnic origin, while rectal cancer is more prevalent in younger, Asian individuals.19,20 Some studies21,22 have found that tumor suppressor genes, point mutations and genetic instability differ according to the subsite of the colorectum. Colon cancer has also been reported to be more likely to have CpG island methylator phenotype and k-ras mutations, whereas rectal and distal colon tumors are more likely to have p53 and APC mutations.7,23 Further studies clarifying the differences in gene expression will help to identify potentially useful markers for differential diagnosis and screening in colon and rectal cancers.

More than 800 genes were differentially upregulated or downregulated in colon and rectal cancers in this study. The first upregulated differentially expressed gene in colon and rectal cancer, the RGS4 gene, encodes for a protein that acts as a GTPase-activating protein, and has recently been studied in breast and ovarian tumors,24,25 while no research data exists for it in colorectal cancer. The second most significantly upregulated gene, SGIP1, encodes for an endocytic protein that affects receptor signaling in neuronal systems involved in energy homeostasis.26 The function of the third most upregulated gene, ANKRD38, is still unclear, and its roles in gastrointestinal cancers and in colorectal cancer development are poorly understood. Most of the genes that were strongly downregulated showed close relationships with gastrointestinal cancers, including the gene for S100P, which is a member of the large family of S100 calcium-binding proteins involved in the regulation of a number of vital cellular processes. The results of many studies have strongly supported a significant role for S100P gene in the development and progression of a range of cancers, such as those of the colon, pancreas, lung and prostate.27,28 Birkenkamp-Demtroder et al. reported that S100P gene was significantly differentially expressed exclusively in right-sided colon adenocarcinomas.8 Our data confirmed that the S100P gene was downregulated in rectal cancer. S100P can be secreted and act via the RAGE gene, which was also differentially downregulated in the current study, resulting in increased cell proliferation and survival.29REG1A is a type I member of the multigene Reg family, which is involved in cellular proliferation and differentiation. Reg I mRNA has been detected at various levels in gastric and colorectal cancer and the Reg gene family has been found to be upregulated in human colorectal cancer.30 However, there have been no reports of a differential expression of Reg family members in colon and rectal cancers.

Signal-Net analysis included more genes that are involved in rectal cancer than in colon cancer, because of the relative numbers of differentially expressed genes. This suggests that the carcinogenesis of rectal cancer is complex. ACTN1 and CAMK2G were identified as key genes in rectal cancer and play crucial roles in cell adhesion,31 which may be related to the higher incidence of metastasis in rectal cancer. The Signal-Net analysis of colon cancer included fewer key genes. Of these, ACAT1 plays a major role in ketone body metabolism and cholesterol esterification, and its encoding protein acts as a modulator of estrogen receptors in colon cancer, partially explaining the higher associations between fatty foods and estrogen in colon cancer, compared with rectal cancer.32 Based on these data, further studies of these genes' expression and the protein functions of S100P, the Reg family, ACTN1, CAMK2G and ACAT1 need to be performed in more samples using reverse transcriptase-polymerase chain reaction and western blotting; moreover, the regulation of identified genes and protein functions also should be studied. These further studies will help to improve the clinical diagnosis and treatment of patients with colon and rectal cancers.

As for differentially expressed genes, more signal pathways were altered in rectal cancer than in colon cancer, further supporting the relative complexity of the mechanisms involved in rectal carcinogenesis. Some signal pathways involved in colon and rectal cancers were the same, such as cell cycle and metabolic and vascular muscle contraction pathways. This shows that the carcinogenesis of colon and rectal cancer are complex, and the main pathways involved are the same. However their roles in different types of carcinogenesis differed. For example, the cell cycle pathway was the most significantly altered pathway in rectal cancer but only the second most altered in colon cancer. A recent study33 indicated that there were significant differences between rectal and colon cancer in terms of amplification of the genes for cyclin A2, cyclin B1, cyclin D1 and cyclin E, with gene amplification and protein expression in rectal cancer significantly correlated for cyclin B1, D3 and E. It has thus been suggested that even though the signal pathways are the same, the components and regulatory mechanisms differ in colon and rectal cancers.

Our results also suggested that metabolic pathways play a more important role in rectal than in colon cancer. Metabolic pathways were the third most upregulated and second most downregulated pathways in rectal cancer, compared with only the fifth most downregulated pathways in colon cancer. The role of metabolic pathways in colorectal cancer development has been a recent focus of attention because of the association between metabolic disorders and cancer.34,35 Slattery et al.36 reported no significant difference in the expression of genes related to metabolic pathways between colon and rectal cancer. However, smoking has been shown to be associated with an increased risk of rectal tumors through changes in the expression of genes related to metabolic pathways.37 The most upregulated pathway in colon cancer was the p53 signal pathway. p53 mutations have been found in 34% of proximal colon tumors and 45% of distal colon and rectal tumors.38 Further studies of the expression of genes related to metabolic and p53 pathways will help in the diagnosis and target treatment of patients with rectal and colon cancers.

The above results all suggest that differences in gene expression exist between colon and rectal cancer. These genes encode proteins involved in different signal pathways, the disruption of which can promote cancer development. Colon cancer and rectal cancer thus represent two distinct kinds of tumors, and several genes, such as S100P, the Reg family, ACTN1, CAMK2G, and ACAT1 provide potential candidates for distinguishing between colon and rectal cancer in the future. This distinction will aid in the diagnosis and prevention of colon and rectal cancers, based on their different features.

ACKNOWLEDGMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFFERENCES

This work was supported by the Beijing Municipal Science and Technology Commission (No. Z080507030808012). Microarray data were analyzed by Genminix Informatics (Shanghai, China).

REFFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. REFFERENCES