Abstract
Background: In a previous study, we reported a critical role of epidermal growth factor-like domain 7 (EGFL7) in the metastasis of hepatocellular carcinoma (HCC) and documented it to be a prognostic biomarker as well as a potential therapeutic target for HCC. However, the role of EGFL8, the only known paralog of EGFL7, in human malignancies is currently unclear. Patients and Methods: EGFL8 expression in 101 cases of colorectal cancer (CRC) patients was determined by quantitative reverse transcription-polymerase chain reaction and the clinicopathological features of the CRC patients were correlated with the EGFL8 down-regulation scores. In addition, the survival curve and Cox regression model were also employed to assess the prognostic value of EGFL8 down-regulation. Results: EGFL8 was significantly decreased in CRC tissues (p<0.0001) and the down-regulation of EGFL8 was evidenced in 74.3% (75/101) of the CRC patients. EGFL8 down-regulation correlated significantly to distant metastasis (p=0.038) and high TNM stage (p=0.012) of CRC. The CRC patients with high EGFL8 down-regulation showed either poorer disease-free survival (p=0.0167) or poorer overall survival (p=0.0310) than those with low EGFL8 down-regulation. Multivariable analysis identified EGFL8 down-regulation as an independent prognostic factor for CRC patients (hazard ratio, 12.974; p=0.037). Conclusion: The reduced expression of EGFL8 is closely related to metastastic potential and poor prognosis of CRC, suggesting the down-regulation of EGFL8 as a novel prognostic biomarker for CRC patients.
Colorectal carcinoma (CRC) ranks the third most common cancer in the world and the incidence of CRC is increasing rather rapidly in the Asia-Pacific region, especially in Japan (1-3). Abundant evidence has shown that a variety of genetic and epigenetic alterations in both oncogenes and tumor suppressors are involved in the pathogenesis of CRC. Activation of oncogenes such as rat sarcoma (RAS) gene and the inactivation of tumor suppressors such as adenomatous polyposis coli (APC) and p53 genes have been well documented in CRC (4-6). In addition, we also have identified some genetic as well as epigenetic changes related to this disease (7-12). However, further investigations are still necessary to clarify the tumorigenic pathway of CRC (13).
In a previous study, we demonstrated the over-expression of epidermal growth factor-like domain 7 (EGFL7), an essential gene in vascular development during embryogenesis (14), in hepatocellular carcinoma (HCC) tissues and revealed an important role of the EGFL7/focal adhesion kinase (FAK)/epidermal growth factor receptor (EGFR) signaling pathway in metastasis of HCC, which provided the first evidence for the expression pattern and role of EGFL7 in human malignancy (15). Egfl8 is the only known paralog of Egfl7 and the proteins they encode share the same overall domain structure, including an EGF-like domain, a Ca2+ binding EGF-like domain and an N-terminal signal peptide (14, 16). Moreover, the expression profiles of Egfl8 and Egfl7 are similar in adult mouse organs, with the highest levels of expression in kidney, brain, thymus and lung, but absolutely different in embryonic tissues (16). However, the expression pattern of EGFL8 in human malignancies is currently unknown. Whereas Fitch et al. thought that Egfl7 and Egfl8 proteins might not overlap in their function during embryonic development (16), the similar structure shared by these two proteins led us to hypothesize that EGFL8 may play an important role in human malignancies just like EGFL7. To test this hypothesis, the present study was carried out to determine the EGFL8 expression pattern in human CRC tissues and to explore the correlations between EGFL8 expression and clinicopathological characteristics, as well as prognosis of this malignancy.
Patients and Methods
Patients and specimens. Matched cancer and normal tissue specimens were obtained from 101 CRC patients who underwent surgery at Showa University Fujikaoka Hospital from February 2007 to August 2009. Informed written consent was obtained from all the patients and the study was approved by the Institutional Review Board of Showa University. All the specimens were collected and frozen in liquid nitrogen immediately after surgery and then stored at −80°C until analysis. The diagnoses of CRC were all confirmed by histopathological examination. The clinicopathological data such as patient age, patient gender, tumor histology, lymph node status and metastasis situation were also collected and the characteristics of the colorectal tumors are summarized in Table I.
RNA preparation and reverse transcription. The total RNA was extracted from the CRC tissues and the corresponding normal tissues by using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instruction as described previously (17). The quality of the total RNA was measured by absorbance at 260nm with a U-2001 spectrophotometer (Hitachi Ltd., Chiyoda, Tokyo, Japan). First-strand cDNA was generated from RNA as described previously (18).
Quantitative polymerase chain reaction (QPCR). QPCR was performed by the Thermal Cycler Dice Real-time System TP800 (TaKaRa Bio Inc., Otsu, Shiga, Japan) using a SYBR Premix Ex Taq II kit (TaKaRa Bio Inc.). Thermocycling was carried out in a final volume of 25 μl containing 1.0 μl of cDNA sample, 0.5 μl of each primer (forward and reverse, 100 nM), 12.5 μl of the SYBR Premix Ex Taq II (including Taq DNA polymerase, reacting buffer, and deoxynucleotide triphosphate mixture). The PCR amplification consisted of 40 cycles (95°C for 5 s, 53°C for 30 s after an initial denaturation step [95°C for 10 s]). The PCR primers for EGFL8 were designed by Primer Premier 5.0 software (Premier Biosoft International, Palo Alto, CA, USA): forward, 5’-AGCCCTA CCTGACCTTGTG-3’; reverse, 5’-GTGCGAGCAGA GGGT-GAT-3’. To correct for differences in both quality and quantity of cDNA samples, β-actin gene was measured in the same samples as an internal control. All the QPCR analyses were performed in duplicate.
Score of EGFL8 down-regulation and grouping of CRC patients. The relative expressions of Egfl8 in the tissue samples were normalized to the internal control β-actin and calculated by the 2−ΔCt method. To show the degree of Egfl8 down-regulation in CRC patients better, the down-regulation of EGFL8 was scored using the logarithmic scale as described previously (19):
Egfl8 down-regulation score=log e (the relative expression of Egfl8 in normal tissue/the relative expression of Egfl8 in cancer tissue).
By the mean of all the CRC patients' Egfl8 down-regulation scores, the CRC patients could be divided into two groups: high Egfl8 down-regulation group (Egfl8 down-regulation score ≥the mean) and low Egfl8 down-regulation group (Egfl8 down-regulation score <the mean).
Follow-up and candidate prognostic factors of CRC. Follow-up data were obtained from all the patients with CRC. The follow-up period was defined as the interval between the date of operation and the patient's death or the end of follow-up (February 23, 2010). The median follow-up time was 384 days. Deaths from other causes were treated as censored cases. Recurrence was diagnosed by clinical examination, colonoscopy, B-ultrasound, computed tomography (CT) or magnetic resonance imaging (MRI). To determine factors influencing survival after surgery, 10 conventional variables along with the EGFL8 down-regulation score were tested in all the CRC patients: age (year, continuous data), gender (male versus female), maximal tumor size (mm, continuous data), depth of tumor invasion (≤muscular tunic invasion versus>muscular tunic invasion), pathological type (well/moderately differentiated CRC versus poorly-differentiated CRC), lymph node metastasis (presence versus absence), hepatic metastasis (presence versus absence), peritoneal dissemination (presence versus absence), distant metastasis (presence versus absence), TNM stage (I, II, III stage versus IV stage), and the EGFL8 down-regulation score (continuous data).
Statistical analysis. The Mann-Whitney U-test was applied to analyze the EGFL8 expression levels in the CRC tissues and the corresponding normal tissues. Student's t-test was used to analyze the correlations between the EGFL8 down-regulation scores and the clinicopathological characteristics of CRC. Survival curves were constructed by using the Kaplan-Meier method and the differences in disease-free survival and overall survival were evaluated by log-rank test. A Cox proportional hazards regression model was constructed to identify the factors that were independently associated with overall survival of CRC. In this model, a step-wise selection was used for variable selection with entry and removal limits of p≤0.05 and p>0.10, respectively. The continuous data were expressed as mean±SE. All the statistical analyses were two-sided and performed by SPSS 13.0 software package (SPSS, Chicago, IL, USA). A p-value <0.05 was considered statistically significant.
Results
EGFL8 expression in CRC tissues. EGFL8 was detectable in all the tissue specimens, but its relative expression level in the CRC tissues was significantly decreased compared with the corresponding normal tissues (0.141±0.082 versus 0.523±0.233, p<0.0001) (Figure 1). In addition, down-regulation of EGFL8 was found in 74.3% (75/101) of the CRC patients. The EGFL8 down-regulation score in the CRC tissues was 1.63±0.27. According to the mean of EGFL8 down-regulation scores, 51 cases of CRC patients fell into the low EGFL8 down-regulation group and the other 50 cases fell into the high EGFL8 down-regulation group.
Correlations between EGFL8 down-regulation and clinicopathological characteristics. The high EGFL8 down-regulation scores were significantly correlated to distant metastasis (p=0.038) and high TNM stage (p=0.012) of CRC (Table II). However, no significant correlation was found between EGFL8 down-regulation scores and the other clinicopathological features of CRC.
Correlation of down-regulation of EGFL8 and poor prognosis. The CRC patients in the high EGFL8 down-regulation group had both poorer disease-free survival and poorer overall survival than those in the low EGFL8 down-regulation group (2-year disease-free survival, 49% versus 96%, p=0.0167; 2-year overall survival, 80% versus 97%, p=0.0310) (Figure 2). By univariable Cox regression analysis, high TNM stage (hazard ratio [HR], 0.127; p=0.005), peritoneal dissemination (HR, 0.150; p=0.010), and high EGFL8 down-regulation score (HR, 7.19; p=0.045) were correlated to overall survival of the CRC patients. Four variables including patient gender, maximal tumor size, tumor differentiation and distant metastasis did not enter the multivariable Cox regression model. In this multivariable model, interestingly, only high TNM stage (HR, 0.012; p=0.034) and high EGFL8 down-regulation score (HR, 12.924; p=0.037) were found to be independent prognostic factors for survival of CRC (Table III).
Discussion
The present study showed for the first time that EGFL8 expression significantly decreased in human CRC compared to the normal tissues and the down-regulation of EGFL8 found in was most (74.3%) of the patients. These data suggested that EGFL8 may have a distinct expression pattern and mechanism of regulation from those of EGFL7 in human malignancies (15, 20-22). The EGFL8 gene is located on human chromosome 6 (6p21.32) and synthenic to the major histocompatibility complex regions (16). In light of the fact that loss of heterozygosity (LOH) at 6p21.32 is common in human malignancies and able to cause the down-regulated expression of human leukocyte antigen class I in CRC (23-25), the down-regulation of EGFL8 may also be induced by LOH at its loci.
When correlating the EGFL8 down-regulation scores with the clinicopathological features of CRC, the down-regulation of EGFL8 was significantly associated with distant metastasis of CRC. In addition, a trend for association with EGFL8 down-regulation and hepatic metastasis of CRC was also found (p=0.071). These criteria are well-accepted features of metastasis of CRC (26, 27), so these data suggested that EGFL8 may be involved in metastasis of CRC. Since the EGFL8 protein shares a similar protein structure with EGFL7 protein, which is also called vascular endothelial statin and inhibits the maturation of vessels through repressing the motility of smooth muscle cells (SMC) (28, 29), we propose that EGFL8 protein may also restrain the migration of SMCs and the down-regulation of EGFL8 in CRC might improve the maturation and validity of tumor vessels and subsequently prompt the metastasis. Further studies, of course, will be necessary to test this possibility.
Down-regulation of EGFL8 was also significantly correlated to high TNM stage (stage IV) of CRC, suggesting the down-regulation of EGFL8 as a late event during the progression of gastroenterological malignancies. Because TNM stage is the most important prognostic factor for CRC (30), the Kaplan-Meier method was employed to assess the prediction power of EGFL8 expression in the prognosis of CRC patients. The CRC patients with high EGFL8 down-regulation scores had both poorer disease-free survival and poorer overall survival than those with low EGFL8 down-regulation scores, further supporting the involvement of EGFL8 in metastasis of CRC and implying the prognostic value of EGFL8. The multivariable Cox regression analysis identified the down-regulation of EGFL8 as an independent risk factor for overall survival in CRC, suggesting EGFL8 down-regulation as a novel prognostic marker for CRC patients.
Collectively, the current study has provided the first evidence indicating the significant decrease of EGFL8 in CRC and its close association with metastasis and poor prognosis of CRC. The study suggests that the down-regulation of EGFL8 may serve as a novel prognostic biomarker for CRC. However, further studies are required to elucidate the exact mechanisms underlying the down-regulation of EGFL8.
Acknowledgements
We would like to thank M. Ogada for her technical assistance in RNA preparation and reverse transcription.
- Received February 22, 2011.
- Revision received May 16, 2011.
- Accepted May 17, 2011.
- Copyright© 2011 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved