Mady, Hussam H. M.D.*‡; Hasso, Sean B.S.†; Melhem, Mona F. M.D.*‡
Colorectal cancer accounts for more than 90% of the malignant tumors of the large bowel. After lung and breast cancer, it is most common cause of death from malignant disease in Western countries (1). Colorectal tumorigenesis is one of the best-characterized examples of the multistep nature of the cancer. Dominant oncogenes and tumor suppressor genes are consistently involved in colorectal tumorigenesis (2).
The E2F family of transcription factors plays a key role in the control of cell-cycle progression, and some of the members have been implicated in the regulation of apoptotic cell death (3–4). Some family members may act as oncogenes, whereas others act as tumor-suppressor genes (5). The E2F family consists of five members (E2F-1 to E2F-5) that can be divided into two subfamilies based on their structures (6). The E2F proteins have a similar structure, although E2F-1, E2F-2, and E2F-3 are more closely related to each other than are E2F-4 and E2F-5. The E2Fs are both activators and repressors of transcription, and several results suggest that the E2Fs have distinct functions in regulating cell-cycle progression. E2F-4 and E2F-5 act mainly as repressors of E2F-dependent transcription in early parts of the cell cycle; E2F-1, E2F-2, and E2F-3 act as repressors and activators of transcription in later parts of the cell cycle (7). Some in vitro studies found that E2F-1 has both tumor-suppressing and oncogenic activity, and they suggest that E2F-1 overexpression can cause carcinogenesis (7,8).
Studies performed on transgenic E2F-4(−/−) knockout mice models showed delayed intestinal epithelial development with abnormally fewer crypts and proliferating cells than their healthy counterparts (9). E2F-4 is involved in the transition from the resting state (G0) to G1 phase and in facilitating the transactivation of genes necessary for cellular proliferation (10). The tumor-suppressor protein p53 appears to function at the G1 phase of the cell cycle as a checkpoint in response to DNA damage. Mutations in the p53 gene lead to an increased rate of genomic instability and tumorigenesis. E2F-1 has been implicated in increased apoptosis in the absence of p53 (11). No specific correlation between p53 mutation and the E2F-family function is known.
The Rb family of proteins must interact with the E2F family to effect their growth-suppressive functions. E2F-4 has been associated preferentially with p107 and p130 members of the Rb-family (7). This is thought to contribute to the antitumorigenic effect of the Rb family, and disruption of this interaction may lead to the loss of cell growth control.
Some experimental evidence implicating genetic anomalies in the E2F-4 gene in gastrointestinal tumors have been described: the E2F-4 gene contains a polymorphic trinucleotide repeat, coding for a polyserine array, and microsatellite instability affecting this repeat has been documented in gastrointestinal tumors (12,13).
In this study, we examine the expression of two E2F proteins (E2F-1 and E2F-4) and their possible correlation with apoptosis in colorectal cancer by evaluating immunohistochemistry through a computerized image analysis technique that has been previously proven to accurately measure the expression of protein within the cells of interest (14). This was correlated with results from an immunoblot technique by using the same tumors and their corresponding histologically healthy covering epithelium.
MATERIALS AND METHODS
Patients and Tissues
Tissue samples, including both tumors and their histologically healthy corresponding mucosa, were collected at the Veterans Affairs Hospital from 20 patients with colorectal cancer who were surgically treated and histologically diagnosed. The patients were admitted to the Veterans Affairs Hospital between 1991 and 1997. Sections that were 5 μg thick from formalinfixed, paraffin-embedded tissues were dewaxed and stained with hematoxylin and eosin stain for histologic assessment.
Immunohistochemical staining of tissue samples were performed by using the mouse monoclonal antibody for E2F-4 transcription factor (Ab-4 clone 4E2F04; NeoMarkers, Fremont, CA) and by using the streptavidin biotin peroxidase complex technique. For E2F-1 immunohistochemical staining, we used the Ab-6 mouse (clone KH95; NeoMarkers) monoclonal antibody, which recognizes a protein of 46kDa, identified as E2F-1.
In Situ Labeling of Apoptotic Cells
To detect apoptotic bodies and cells, we used the Apoptag Peroxidase Kit (Intergen, Purchase, NY) as well as a working strength peroxidase substrate developing system for color developing (following the manufacturer's protocol).
Quantitative Analysis of Positively Stained Slides by Using the Image Analyzer
Objective analysis of the amount of and distribution of the immunostained tissues was achieved by using a computerized image analysis system for slides stained with E2F-4, E2F-1, and Apoptag. The slides were evaluated under 40× magnification by using Samba 4000 Image Analysis System with Immuno 4.00 (Image Products International, Chantilly, VA) quantitative program in the Microsoft Windows (Microsoft, Redmond, WA) environment. Eight representative high-power fields (40×) of the covering mucosa and another eight high-power fields of the tumor from the same sample stained with E2F-4 were analyzed in each slide to detect 1) the labeling index (LI) of the staining (equal to the ratio between area of positively labeled tumor and total area of the tumor); 2) the mean optical density (the concentration of the stain as measured per positive pixels); and 3) the quick score (the numerical product of the mean optical density and the LI). The mean LI of the Apoptag stain per object in the tumors was examined by the image analysis system in eight high power fields per slide, which included approximately 640 cells.
Protein Isolation and Immunoblot Analysis
Protein extracts from whole tissues were obtained by tissue homogenization in lysis buffer (sodium chloride, 150 mmol/L; Tris HCl, 50 mmol/L; EDTA, 5 mmol/L; 0.5% NP40, pH 8.0) containing both protease (complete protease inhibitors [Boehringer, Ingelheim, Germany]: leupeptin, 5 μg/mL; pepstatin, 0.7 μg/mL), and phosphatase inhibitors (sodium fluoride, 50 mmol/L; sodium orthovanadate, 1 mmol/L). Cellular membranes were removed by centrifugation, and the protein concentration of the soluble protein fraction was determined by a protein assay reagent (Bradford, 1976, Bio-Rad, Hercules, CA). Thirty micrograms of each protein lysate were resolved in polyacrylamide-SDS gels and transferred to polyvinyl diflouride membranes (Immobilon-P; Millipore, Bedford, MA). The membranes were treated with specific primary antibodies and secondary agents (protein A or goat antimouse immunoglobulin, linked to horseradish peroxidase) according to standard procedures.
We examined the distribution of LI and the expression of E2F-4 in the tumor cells and in the covering epithelium, as well as the LI of the Apoptag stain. We compared the LI of E2F-4 in tumor cell with the corresponding covering epithelium and the expression of E2F-4 in tumor cell with that of E2F-4 in the covering epithelium (both comparisons performed by using the paired t test). Bivariate sample correlation was assessed by using Spearman's rank correlation coefficients. All analyzes were performed by using Statistical Analysis System (version 6.12) software (SAS, Cary, NC).
Histology, Immunohistochemistry, and Image Analysis
Histologic subtypes of the 20 cases included 4 well-differentiated adenocarcinomas, 13 moderately differentiated adenocarcinomas, 1 poorly differentiated adenocarcinoma, and 2 mucin-producing carcinomas. The tumors were classified according to the predominant histologic component.
Expression of E2F-4 and E2F-1 Proteins in Human Colorectal Cancer Versus the Covering Epithelium
Immunohistochemical results showed that the tumor cells express E2F-4 in a heterogeneous manner with positive cells varying in number, the protein expression being strictly nuclear and varying in intensity. Brownish staining of E2F-4 protein was detected mainly in the nuclei of the cells and, in a few cases, in the cytoplasm. The histologically healthy covering epithelium showed E2F-4 expression varying from negative to weakly positive staining, whereas the whole tumor was strongly stained (Fig. 1 A and B). Image analysis of immunohistochemical staining detected expression of E2F-4 in a much higher manner in the tumors than that in the covering epithelium. The LI of E2F-4 was high in 12 cases, moderate in 5 cases, and very low in 3 cases of the 20 cases included in this study. The mean LI of E2F-4 expression in the tumors was 49%, whereas that in their covering epithelium was 15% (Fig. 2). Our data show that, as the LI of E2F-4 in tumor cell increases, the LI of E2F-4 in covering epithelium increases (Spearman correlation coefficient, 0.62) (Fig. 3A).
Image analysis of E2F-1 immunohistochemical staining found weak positive staining in the tumors of five cases and in the covering mucosa of three cases. The mean LI of E2F-1 staining in the tumors was 8.1% and that of the covering mucosa was 3.9%. Figure 4 shows the difference of expression between E2F1 (Figure 4A) and E2F-4 (Figure 4B) within the same tumor.
Correlation of E2F-4 Protein Expression and Apoptosis in Human Colorectal Cancer
There is a negative correlation between the E2F-4 immunostain and the Apoptag stain, whereas the LI of E2F-4 in the tumors was inversely proportional to the LI of Apoptag immunostain (Spearman correlation, −0.33) (Fig. 3B). Twelve cases showed a very high E2F-4 LI and four cases with moderate expression showed a corresponding low apoptosis LI. Three cases showed relatively lower levels of E2F-4 LI, and two of them were characterized with high apoptotic rate.
Protein Analysis of Tumors and the Corresponding Mucosa by Immunoblotting
Immunodetection of human E2F-4 by immunoblot analysis showed proteins migrating with an apparent molecular weight between 46 and 48 Kd, as expected for human E2F-4 (theoretical molecular weight, 44 Kd). By both immunohistochemistry and immunoblot analysis,
77% of tumor samples showed a relative increase of E2F-4 in comparison with the corresponding mucosa. In clear contrast, we were not able to detect E2F-1, a member of the E2F family and part of the subgroup comprising E2F-1, E2F-2 and E2F-3, although the control E2F-1 protein present in murine extracts was easily apparent (Fig. 5). We have not found any obvious correlation among the type of tumor, the site of surgery, or the age of the patients and the relative levels of E2F-4 protein.
In this study, we have evaluated E2F-4 protein expression in colorectal tumors and their covering epithelium by using both immunohistochemical and immunoblot techniques. Some reports suggest that, within the E2F family, E2F-4 is the main factor responsible for maintaining the quiescent and/or differentiated status in adult tissues and it is not required for active induction of cell proliferation (15). Nonetheless, some reports contradict this hypothesis, showing that E2F-4 transcripts are more highly expressed in proliferating cells through embryonic development (16) and that E2F-4 is able to induce cell proliferation and transformation of certain cell types both in vivo and in vitro (17,18).
In agreement with this possibility, we show that the E2F-4 protein is specifically present in the proliferative compartment of the benign colonic epithelium and that the protein levels increase selectively and significantly in cells from colonic tumors, whereas they remain within healthy levels in nonaltered adjacent and distant mucosa. This may reflect a role this protein plays in the development of colorectal tumors. In view of this, Yoshitaka et al. (12) suggest that E2F-4 may be a cancer-related gene, and the polymorphisms in E2F-4 gene may be one of the risk factors for colorectal carcinogenesis. This is supported by the recent report that found the initiation and the maintenance of neuronal differentiation induced by nerve growth factor in PC12 cells is promoted by E2F4 (19).
Suzuki et al. (5), using Southern, Northern, and Western blot analysis, observed increases in the expression of E2F-1 in 60% (three of five) of colorectal carcinomas in comparison with the corresponding nonneoplastic mucosa. These data seem to be in disagreement with the observed expression pattern of E2F-1. In this study, we could not detect E2F-1 in either the colorectal tumors or in the covering mucosa by the immunoblot technique; however, we did detect it in 5 of 20 cases (20%) by using the immunohistochemical technique, but the mean LI of its image analysis is very low (8.1% in tumors and 3.9% in covering mucosa) when compared with that of E2F-4 (48.3% in tumors and 14.8% in the covering epithelium). This can be attributed to a dilution effect of the protein because of whole tissue homogenate used in immunoblots that contain a large amount of interstitial stroma, whereas microscopic section examination and image analysis accurately assess the stain within epithelial cells in both benign and malignant tumors.
Paramio et al. (20) found that E2F-1 and E2F-4 are differentially expressed during HaCaT epidermal keratinocyte cell line differentiation. Their data suggest that E2F-1 and E2F-4 have opposite functions in keratinocyte differentiation. Although E2F-1 inhibits differentiation, E2F4 facilitates the process. The presence of free E2F-4 can trigger the expression of genes positively involved in this process (19,21).
Differential E2F protein expression has also been found during differentiation in hematopoietic, neuronal, and muscle cells (18,22,23). In the current study, we detected that apoptosis levels observed in colonic adenocarcinoma samples from different patients inversely correlate with the amount of E2F-4 expressed in the same tumors.
Several reports show the involvement of E2F-1 in the induction of apoptosis, independent of p53 function (11). Therefore, E2F-1 can induce events leading to S phase, but this process is not normal and results in the activation of dell death (apoptosis) pathway (3,4,24). Because we do not find any E2F-1 expression in colon tumors, then the rate of apoptosis detected in these tumors is most likely E2F-1 independent.
In this study, we have shown that immunohistochemical staining coupled with computerized image analysis techniques accurately assesses the level of E2F-4 expression, as confirmed by immunoblot analysis in normal and neoplastic colonic epithelium. Furthermore, colon tumors showed heterogeneity in E2F-4 expression and overexpressed E2F-4 relative to healthy mucosa. We also found an inverse correlation between E2F-4 levels and apoptosis in tumors.
In conclusion, our data suggest that E2F-4 gene overexpression plays a role, either directly or indirectly, in the development of colorectal tumors and in the suppression of apoptosis. On the other hand, E2F-1 does not appear to contribute significantly in most cases of colorectal carcinoma, or its absence may further confirm its role as a tumor suppressor. These findings clearly warrant further investigation.
The authors would like to thank Dr. James Pipas and Dr. M.T. Saenz for stimulating discussion; Ms. Diane George, HT, ASCP, for her excellent technical support in performing the immunohistochemical staining; and Mary Kelley, MS, for her help with the statistical analysis.
1. Houlston RS, Tomlinson IPM. Genetic prognostic markers in colorectal cancer. Mol Pathol 1997; 50:281–8.
2. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990; 61:759–67.
3. Wu X, Levine AJ. P53 and E2F-1 cooperate to mediate apoptosis. Proc Natl Acad Sci U S A 1994; 91:3602–6.
4. Kowalik TF, DeGregori J, Schwarz JK, et al. E2F-1 overexpression in quiescent fibroblasts leads to induction of cellular DNA synthesis and apoptosis. J Virol 1995; 69:2491–500.
5. Suzuki T, Yasui W, Yoko-Zaki H, et al. Expression of E2F family in human gastrointestinal carcinomas. Cancer 1999; 81:535–8.
6. Helin K. Regluation of cell proliferation by the E2F transcription factors. Curr Opin Genet Dev 1998; 8:28–35.
7. Johonson DG, Schneider-Broussard R. Role of E2F in cell cycle control and cancer. Front Biosci 1998; 3:d447–58.
8. Field SJ, Tsai F, Kuo F, et al. E2F-1 functions in mice to promote apoptosis and suppress proliferation. Cell 1996; 85:546–61.
9. Rempel RE, Saenz-Robles MT, Storms R, et al. Loss of E2F4 activity leads to abnormal development of multiple cellular lineages. Mol Cell 2000; 6:293–306.
10. Sardet C, Vidal M, Cobrinik D, et al. E2F-4 and E2F-5, two members of the E2F family are expressed in the early phases of the cell cycle. Proc Natl Acad Sci U S A 1995; 92:2403–7.
11. Agah R, Krishenbaum LA, Abdellatif M, et al. Adenoviral delivary of E2F-1 directs cell cycle reentry and p53-independent apoptosis in postmitotic adult myocardium in vivo. J Clin Invest 1997; 100:2722–8.
12. Yoshitaka T, Matsubara N, Ikeda M, et al. Mutations of E2F-4 trinucleotide in colorectal cancer with microsatellite instability. Biochem Biophys Res Commun 1996; 227:553–7.
13. Souza RF, Yin J, Smolinski KN, et al. Frequent mutation of the E2F-4 cell cycle gene in the primary human gastrointestinal tumors. Cancer Res 1997; 57:2350–3.
14. Grandis JR, Melhem MF, Barnes EL, et al. Quantitative immunohistochemical analysis of transforming growth factor-alpha and epidermal growth factor receptor in patients with squamous cell carcinoma of the head and neck. Cancer 1996; 78:1284–92.
15. Dyson N. The regulation of E2F by pRB-family proteins. Genes Dev 1998; 12:2245–62.
16. Dagnino L, Fry CJ, Bartley SM, et al. Expression patterns of the E2F family of transcription factors during murine epithelial development. Cell Growth Differ 1997; 8:553–63.
17. Beijersbergen RL, Kerkhoven RM, Zhu L, et al. E2F-4, a new member of the E2F gene family, has oncogenic activity and associates with p107 in vivo. Genes Dev 1994; 8:2680–90.
18. Wang J, Helin K, Jin P, et al. Inhibition of in vitro myogenic differentiation by cellular transcription factor E2F1. Cell Growth Differ 1995; 6:1299–1306.
19. Persengiev SP, Kondova II, Kilpatrick DL. E2F4 actively promotes the initiation and maintenance of nerve growth factor-induced cell differentiation. Mol Cell Biol 1999; 19:6048–56.
20. Paramio JM, Segrelles C, Casanova ML, et al. Opposite functions for E2F-1and E2F-4 in human epidermal keratinocyte differentiation. J Biol Chem 2000; 275:41219–26.
21. Paramio JM, Laín S, Segrelles C, et al. Differential expression and functionally co-operative roles for the retinoblastoma family of proteins in epidermal differentiation. Oncogene 1998; 17:949–57.
22. Richon VM, Venta-Perez G. Changes in E2F DNA-binding activity during induced erythroid differentiation. Cell Growth Differ 1996; 7:31–42.
23. Strom DK, Cleveland JL, Chellappan S, et al. E2F-1 and E2F-3 are functionally distinct in their ability to promote myeloid cell cycle progression and block granulocyte differentiation. Cell Growth Differ 1998; 9:59–69.
24. Itoshima T, Fujiwara T, Waku T, et al. Induction of apoptosis in human esophageal cancer cells by sequential transfer of the wild-type p53 and E2F-1genes: involvement of p53 accumulation via ARF-mediated MDM2 downregulation. Clin Cancer Res 2000; 6:2851–9.
© 2002 Lippincott Williams & Wilkins, Inc.