This article is mentioned in:

Introduction

The size and shape of a lesion in colorectal cancer (CRC) are believed to be connected directly to clinical phenotype and to serve as predictors of malignant behavior (1,2). Early superficial CRC lesions can be classified as elevated (polypoid and non-polypoid) or depressed (3,4). Polypoid lesions grow above the surface of the mucosa and the volume of the polypoid component appears to be correlated with malignant behavior. Non-polypoid lesions may grow flat or slightly elevated, eventually progressing, into polypoid lesions or lateral spreading tumors. Finally, depressed lesions (0-IIc, 0-IIc+IIa, 0-IIa+IIc), which comprise only 2.3% of all superficial lesions (5), warrant particular attention due to the difficulty of detection and removal by fiberscope (6,7). Since depressed-type lesions are frequently located in the right colon, they can be difficult to detect. Moreover, depressed lesions, independent of size, have been associated with an increased risk of rapid progression to cancer, as shown in endoscopy and pathology units in Japan (4,5,8,9). Depressed-type lesions and tumors, such as earlier phase lesions (0-IIc, 0-IIc+IIa, 0-IIa+IIc) and/or advanced tumors (types 2 and 3), invade and metastasize to lymph nodes and other distant organs and indicate a poorer prognosis than do type 1 tumors.

Microarray analysis of RNA from the cancer cells of 144 CRC cases revealed that the fibronectin 1 (FN1) gene is significantly associated with tumor shape in CRC. FN1 is a glycoprotein that is present in a soluble dimeric form in the plasma, and in a dimeric or multimeric form at the cell surface and in the extracellular matrix (10). It is involved in cell adhesion and migration processes including embryogenesis, wound healing, blood coagulation, host defense and metastasis. The gene has 3 regions that are subject to alternative splicing, with the potential to produce 20 transcript variants. However, the full-length nature of some of these variants has not been determined.

In the present study, we performed a comprehensive analysis to identify genes that determine tumor shape (11). Evaluation of changes in the expression patterns of these genes may allow physicians to make a precise, non-invasive diagnosis of depressed-type lesions in early CRC. The correlation between FN1 expression and tumor shape was validated by quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) before single nucleotide polymorphism (SNP) analysis was performed to identify polymorphisms in the FN1 coding region that could be used as predictors of tumor shape (1) in CRC.

Materials and methods

CRC patients

The study group comprised 146 patients with primary CRC. The patients ranged in age from 32 to 96 years, with an average age of 66 years. They underwent operations at major hospitals in Japan: Kyushu University, Kitazato University, Tokyo Medical and Dental University, National Defense University, Mie University, Takano Hospital, National Cancer Center and Osaka University from 2004 to 2009. None of the patients received preoperative treatments such as radiation or chemotherapy. Immediately after surgical resection, tumor samples (T) were carefully removed from primary cancerous lesions for cell isolation using laser microdissection (LMD). Clinicopathological patient data were obtained from clinical records. Histopathological assessments were made using the Japanese Classification of Colorectal Carcinoma, 7th edition.

Collection of CRC cells

Tissues from the 146 CRC patients were collected for LMD using the Leica Laser Microdissection System (Leica Microsystems, Wetzlar, Germany). In brief, 5-μm frozen sections were fixed in 70% ethanol for 30 sec, stained with hematoxylin and eosin and dehydrated as follows: 5 sec each in 70, 95 and 100% ethanol and a final 5 min in xylene. The sections were air-dried, then microdissected with the LMD system. Target cells, at least 100 cells/section, were excised and bound to transfer film for total DNA extraction.

Total RNA extraction and first-strand cDNA synthesis

CRC tissue specimens or cultured cell lines at subconfluency were homogenized, and total RNA was extracted using the modified acid-guanidine-phenol-chloroform method. Total RNA (8.0 μg) was reverse transcribed to cDNA using M-MLV RT (Invitrogen Corporation, Carlsbad, CA, USA).

qRT-PCR

The sequences for FN1 mRNA were: FN1, sense primer, 5′-GAACTATGATGCCGACCAGAA-3′ and antisense primer, 5′-GGTTGTGCAGATTTCCTCGT-3′. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control and GAPDH primers were: sense primer 5′-TTGGTATCGTGGAAGGACTCTA-3′ and antisense primer, 5′-TGTCATATTTGGCAGGTT-3′. Real-time monitoring of PCR reactions was performed using the LightCycler system (Roche Applied Science, Indianapolis, IN, USA) and SYBR-Green I dye (Roche Diagnostics, Tokyo, Japan). Monitoring was performed according to the manufacturer’s instructions. In brief, a master mixture containing 1 μl of cDNA, 2 μl of DNA Master SYBR-Green I mix, 50 ng of primers and 24 μl of 25 mM MgCl2 was prepared on ice, and the final volume was adjusted to 20 μl with water. After the reaction mixture was loaded into glass capillary tubes, qRT-PCR was performed with the following cycling conditions: initial denaturation at 95°C for 10 min, followed by 45 cycles of 95°C for 10 sec, annealing at 60°C for 10 sec and extension at 72°C for 10 sec. After amplification, products were subjected to a temperature gradient from 65° to 95°C at 0.2°C/sec, under continuous fluorescence monitoring, to produce a melting curve for analysis of primer specificity.

Expression array analysis

For microarray expression analysis, we used a commercially available Whole Human Genome Oligo DNA Microarray kit (Agilent Technologies, Santa Clara, CA, USA). A list of genes on this cDNA microarray is available from http://www.chem.agilent.com/scripts/generic.asp?lpage=5175&indcol=Y&prodcol=Y&prodcol=N&indcol=Y&prodcol=N. Cyanine (Cy)-labeled cRNA was prepared using T7 linear amplification as described in the Agilent Low RNA Input Fluorescent Linear Amplification Kit Manual (Agilent Technologies). Labeled cRNA was fragmented and hybridized to an oligonucleotide microarray (Whole Human Genome 4×44K Agilent G4112F). Fluorescence intensities were determined with an Agilent DNA Microarray Scanner and were analyzed using G2567AA Feature Extraction Software version A.7.5.1 (Agilent Technologies), which uses locally weighted linear regression curve fit (LOWESS) normalization. This microarray study followed the MIAME guidelines issued by the Microarray Gene Expression Data group. Further analyses were performed using GeneSpring version 7.3 (Silicon Genetics, San Carlos, CA, USA).

Evaluation of representative SNPs in the FN1 coding region

Genomic DNA was extracted from the peripheral blood of 64 patients with primary CRC using a QIAamp DNA Mini kit according to the manufacturer’s protocol (Qiagen, Valencia, CA, USA). SNPs in the exonic and intronic regions of FN1 were evaluated simultaneously using an Affymetrix genome-wild SNP array to determine the association between FN1 expression and the genotype of each SNP in the FN-1 gene: rs6707530, rs11651, rs7594168, rs10498037, rs33996776, rs1250214, rs7568287, rs10201850, rs41347752, rs1250204, rs7588661, rs2577302, rs1968510, rs34255697, rs10172425, rs7572169, rs2372545, rs1250264, rs12105173, rs10199059, rs1437799, rs1250270, rs11693652, rs7567647, rs1898536, rs10202483, rs1250247, rs1250233, rs6753702 and rs1250252. However, rs2372545 and rs1250264 were excluded due to low-quality data. The Ethics Committee of each institute approved this project.

Results

Differential expression of FN-1 in cells isolated from CRC primary tumors

Microarray analysis revealed a significant difference in the expression of the FN1 gene between depressed-type tumors and lesions (0-IIb, 0-IIc, 0-IIa+IIc, 0-IIc+IIa, type 2 and type 3 tumors) and elevated-type tumors (0-Ip, 0-Isp, 0-Is, 0-IIa and type 1 tumors). The microarray contained three probes for FN1, all of which were significantly upregulated in depressed-type lesions (n=129) compared to elevated-type tumors (n=17). The 3 probes were also used to validate these findings in representative samples of depressed (n=19) and elevated (n=9) tumors and normal tissues (n=9). The 19 depressed tumors showed significantly higher expression than the other groups (raw p-value, 1.03985×10−6; q-value, 5.17319×10−3; fold-change, 32.4169; Fig. 1). In comparison to that of normal samples, the average expression of FN1 in depressed type lesions was upregulated, while expression in elevated-type lesions was downregulated.

Figure 1 Differences in the growth type in Tis-T2 CRC tumor. Three probes covered the coding region of FN1 (NM-212482) on the Agilent expression array. FN1 expression detected by each probe is represented by the black lines in each column. Left column, expression in normal colorectal epithelial cells (n=9); middle column, expression in depressed-type tumors (n=19); right column, expression in elevated-type tumors (n=9). The upregulated (red lines) and downregulated (green lines) expression levels of whole genes in the microarray analysis are also denoted.

qRT-PCR supported the finding that the expression of FN1 in the subgroup of depressed tumors (19 tumors; 1.47±0.36) was significantly higher (p<0.05) than that in the subgroup of elevated tumors (n=18; 0.39±0.37; Fig. 2).

Figure 2 Validation of the association between tumor shape and FN1 expression. The expression of FN1 in the subgroup of depressed tumors (n=19; 1.47±0.36) vs. the subgroup of elevated tumors (n=18; 0.39±0.37; p<0.05). FN1, fibronectin 1.

Clinicopathological significance of FN1 in CRC cases

In addition to tumor shape, 2 other clinicopathological factors were associated with FN1 expression (Table I). Higher expression was observed in larger tumors (>5 cm; n=71) than in smaller tumors (n=75; p=0.0024). There was also a significant difference in FN1 expression between lymphatic permeation negative (n=65; 4.38±0.24) and positive (n=61; 3.70±0.22) tumors (p=0.036).

Table I Clinicopathologic significance of the expression of fibronectin-1/GAPDH in CRC cases.

Table I

Clinicopathologic significance of the expression of fibronectin-1/GAPDH in CRC cases.

Clinicopathologic analysis	N	Fibronectin-1/GAPDH log	P-value
Gender
Male	82	4.27+0.28	NS
Female	64	3.67+0.25
Location of tumor
Rectum	52	4.40+0.27	NS
Proximal colon	43	3.57+0.30
Distal colon	51	3.97+0.28
Size of tumor (cm)
Large (>5)	71	4.51+0.23	0.0024
Small (<5)	75	3.53+0.22
Type of tumora
Depressed	129	4.16+0.17	0.0098
Elevated	17	2.84+0.47
Histologic differentiationb
Well	82	4.02+1.92	NS
Mod	61	4.03+2.11
Poor	2	2.76+0.13
Muc	1	3.62
Depth of tumor invasion
pTis-pT1(pM-pSM)	8	2.92+0.70	NS
pT2 (pMP)	19	4.25+0.46
pT3 (pSS/pA)	88	4.12+0.21
pT4(pSE-PSO/pAI)	31	3.82+0.36
Lymphatic permeation
Negative	65	4.38+0.24	0.036
Positive	81	3.70+0.22
Venous permeation
Negative	54	4.15+0.27	NS
Positive	91	3.93+0.21
Lymph node metastasis
Negative	75	4.01+0.23	NS
Positive	70	3.98+0.24
Peritoneal dissemination
Negative	140	4.00+0.17	NS
Positive	6	3.933+0.81
Liver metastasis
Negative	126	4.01+0.18	NS
Positive	20	3.98+0.45
Dukes
A	21	3.86+0.44	NS
B	49	4.09+0.29
C	53	4.09+0.28
D	23	3.77+0.42

a Shape of tumors: depressed-type tumor; 0-IIc, type 2; type 2. Elevated type tumor: 0-Isp, 0-Ip; type 1.

b Histologic differentiation: well, well differentiated adenocarcinoma; mod, moderately differentiated adenocarcinoma; poor, poorly differentiated adenocarcinoma; muc, mucinous adenocarcinoma.

FN1 expression is associated with 1 SNP

Of the 30 SNPs in the FN1 region, rs6707530 was associated with tumor shape in CRC (Fig. 3). The expression of FN1/GAPDH was higher in CRC samples with a GG genotype (n=44; 4.09±0.32) at this locus rather than a GT or TT genotype (n=23; 3.32±0.45).

Figure 3 FN1 expression is associated with a single nucleotide polymorphism. The GG and GT/TT genotypes of rs6707530 were identified in 44 and 23 CRC cases, respectively. The expression of FN1 was significantly higher in GG cases of CRC than in GT/TT cases. FN1, fibronectin 1.

Discussion

The tumorigenesis of depressed tumors progresses along a different pathway from the conventional path for elevated tumors that was advocated by Vogelstein et al (4). It is believed that malignant cells in serrated and adenomatous lesions proliferate and grow laterally and top-down from the surface of the lesion (1–3). However, little is known about how the molecular biology of certain cancer cells is determined. In the present study, we focused on clarifying the mechanism responsible for the differentiation of the shapes of malignant cells. In order to do this, we extracted cancer cells from primary tumor tissues, avoiding contamination with interstitial cells or non-malignant cells with the use of LMD. Microarray analysis following the extraction of total RNA and purified mRNA identified 1 gene, FN1 (10). FN1 was more highly expressed in the interstitial tissues than in the superficial glands in the section (Fig. 4). We propose that the abundant expression of FN1 may allow for the generation of traction forces through its surface receptors. However, further in vitro and in vivo studies are required to answer this question.

Figure 4 FN1 expression in representative CRC tissues. FN1 expression was much stronger in the interstitial tissues of tumors than in the corresponding normal tissues. FN1, fibronectin 1; CRC, colorectal cancer.

CRC characterized by depressed tumors has a higher incidence of recurrence than CRC cases characterized by elevated tumors (Fig. 5). Considering practical clinical applications, reliable markers that can be used to predict tumor shape may facilitate the early diagnosis of malignancy prior to a colonoscopy. In the present study, we identified 1 SNP in FN1 that was significantly associated with tumor shape.

Figure 5 Comparison of disease-free survival rates based on Dukes classification and tumor shape. A Dukes D classification was indicative of the highest recurrence rate among this collection of cases. CRC cases with depressed tumor (n=129) had a much higher incidence of recurrence than those with elevated tumors (n=17; log-rank test; p<0.05). CRC, colorectal cancer.

In conclusion, we found that the majority of CRC cases with depressed tumors had a higher frequency of elevated FN1 expression. In addition, we could predict the presence of depressed tumors by evaluation of 1 SNP (rs6707530) in the FN1 region in germline DNA from peripheral blood. This discovery will beneficial in the clinical setting, providing a method for the early diagnosis of depressed-type tumors by colon fiberscope.

References