|Home | About | Journals | Submit | Contact Us | Français|
Flap endonuclease1 (FEN1) is a structure-specific nuclease best known for its critical roles in Okazaki fragment maturation, DNA repair and apoptosis-induced DNA fragmentation. Functional deficiencies in FEN1, in the forms of somatic mutations and polymorphisms, have recently been shown to lead to autoimmunity, chronic inflammation, and predisposition to and progression of cancer. In order to explore how FEN1 contributes to cancer progression, we examined FEN1 expression using 241 matched pairs of cancer and corresponding normal tissues on a gene expression profiling array and validated differential expression by quantitative real-time PCR, and immunohistochemistry. Furthermore, we defined the minimum promoter of human FEN1 and examined the methylation statuses of the 5' region of the gene in paired breast cancer tissues. We demonstrate that FEN1 is significantly up-regulated in multiple cancers and the aberrant expression of FEN1 is associated with hypomethylation of the CpG island within the FEN1 promoter in tumor cells. The over-expression and promoter hypomethylation of FEN1 may serve as biomarkers for monitoring the progression of cancers.
The development of cancer involves altered expression of many genes during tumor formation and progression as a result of both genetic and epigenetic changes in the genome (1, 2). The identification of genes that contribute to cancer outcome and progression is critical for the development of appropriate therapy. Accumulating evidence shows that flap endonuclease 1 (FEN1) plays a pivotal role in the development of cancer and up-regulation of FEN1 may be correlated with disease progression (3-5).
FEN1 is a structure-specific nuclease that is best known for its involvement in the maturation of Okazaki fragments during DNA replication and its efficient removal of 5′-flaps during long-patch base excision repair (LP-BER)(6-11). In addition to its 5′-flap endonuclease activity, FEN1 is also characterized as a 5′ exonuclease (EXO activity) (6, 12) and a gap-dependent endonuclease (GEN activity) (13, 14). In response to apoptotic stimuli, the GEN and EXO activities of FEN1 can be stimulated to promote apoptosis-induced DNA fragmentation. Considering its critical role in multiple DNA metabolic pathways, FEN1 is a key enzyme for maintaining genomic stability (15) and its deficiency results in predisposition to cancer (4) and rapid development of tumors (3).
FEN1 is widely expressed in mammalian cells and shows a high level of expression in testes, thymus, bone marrow and other highly proliferative tissues, in keeping with its role in DNA replication(16). In mouse embryonic fibroblasts, FEN1 was shown to be up-regulated in a p53-dependent manner upon UV-C exposure (17). In addition, FEN1 protein is associated with proliferative cell populations (18-20). Its expression is up-regulated in metastatic prostate cancer cells(21), gastric cancer cells (22), neuroblastomas (23), pancreatic cancer (24) and lung cancer cell lines (25). FEN1 is also highly expressed in comprehensive genome-wide tumor micro-array datasets for cell-cycle regulated genes (26). Recently, Lam et al. showed that FEN1 is up-regulated in prostate cancer as compared to matched normal prostate and its expression increases with tumor progression(5), suggesting that FEN1 is a possible biomarker for patients at high risk for prostate cancer and a potential target for therapy.
To re-evaluate the expression of FEN1 in various cancers and to understand the mechanism underlying such transcriptional activation, we performed a comprehensive analysis of FEN1 expression in multiple cancers, and particularly in breast tissues, using a cancer profiling array that included paired normal/tumor specimens and in vivo immunohistochemistry in a breast progression cancer tissue array. We found that the nuclease gene was significantly over-expressed in cancer cells. We also tested whether DNA methylation plays a role in the regulation of FEN1 expression, and identified a region of the FEN1 promoter that was hypomethylated in the same cancer cells in which we observed gene over-expression.
A cancer profiling array was used to examine FEN1 expression in matched tumor tissues versus normal tissues. A general trend of FEN1 over-expression was observed in all cancers examined (Supplementary Table 3, Figure 1A and 1B), except for in prostate cancer, where the difference was not statistically significant. An approximately 2.5-fold consistent increase in expression was found in breast and uterine cancer samples, while a 1–2-fold increase was found in other tumors, including colon, stomach, lung and kidney. In cases where metastatic samples were included, FEN1 expression was significantly greater in the metastatic tissues as compared to the paired tumor tissues. This trend was observed in samples of breast, uterine, rectal, ovarian and colon cancers.
Of the 50 cases of breast tissues examined, 47 cases (94%) demonstrated greater FEN1 expression than the matched normal tissues (Table 1, Figure 1C). In three breast cancer cases with matched metastatic samples, FEN1 expression was higher in one metastatic tissue compared to the tumor tissues. The observed over-expression of FEN1 in tumor samples from the cancer profiling array supports the hypothesis that increased expression of FEN1 may be associated with tumorigenesis. This is consistent with several other studies demonstrating FEN1 over-expression in various cancer tissues(5, 21, 22). The integrity of the samples on the array was confirmed by re-probing the array with a radio-labeled ubiquitin cDNA (data not shown). The FEN1 cDNA did not hybridize with yeast total RNA, yeast tRNA, E. coli RNA, E coli DNA, poly r(A) and Cot-1DNA.
Immunohistochemical assessment of a breast cancer progression array showed that FEN1 was expressed in nuclei, with occasional cytoplasmic staining (Figure 2). We did not detect FEN1 in three normal breast tissues (Table 2); however, its expression was variable in tissues from different kinds of benign breast disease (BBD). In the 21 samples, including adenosis, fibrofatty tissue, fibroadenosis, adenosis with hyperplasia of epithelium and cystic hyperplasia, there was very low or no FEN1 immunohistochemical staining. The 11 samples with blunt duct adenosis, fibroadenosis with hyperplasia of epithelium, sclerosing adenosis, papillomatosis or papillomatosis with hyperplasia of epithelium tissues, were positive for FEN1 expression. Moreover, in infiltrating ductal carcinoma, FEN1 expression increased with corresponding disease stages and was highest in nonspecific infiltrating duct carcinoma III tissues. Among other three kinds of breast carcinomas, there was greater FEN1 expression in infiltrating lobular carcinoma than in medullary or mucous carcinoma. These findings suggest that the amount of FEN1 expression is directly correlated with the higher stages and grades of breast tumors.
To determine the location of the FEN1 promoter, we studied the transcriptional activity of chimeric constructs of progressively 5-’deleted DNA fragments (-1821/+632). The various deletion fragments were cloned into a firefly luciferase-based reporter vector, pGL4.10, and subsequently co-transfected with a renilla-based reporter vector into HeLa cells to quantify the strength of the reporter activities driven by different deletion fragments (Figure 3). The empty pGL4.10 vector directed 0.46 ± 0.23 RLU of luciferase activity in HeLa cells. No significant transcriptional activity was detected with the -1821/+632 and +352/+2262 fragments (3.31 ± 0.32 RLU and 0.12 ± 0.02 RLU, respectively). 5’ deletions of -1821/+632 fragment (1.5 ± 0.35RLU) ie., -1191/+632, -701/+632, -421/+632 and -281/+632 resulted in 5.7 ± 0.75, 5.8 ± 0.28, 3.5 ± 0.14 and 10.7 ± 1.05 RLU of luciferase acitivity, respectively. However, the maximum activity was shown by the –458/+278 fragment (32.54 ± 3.81 RLU). The –458/+278 fragment has been shown to be the minimum promoter to drive expression of the mouse FEN1(27), which suggests that the homologous fragment is also the minimum promoter to drive human FEN1 expression.
The mechanism responsible for the regulation of FEN1 expression is largely unknown. In silico analysis indicated that there are two CpG islands in the FEN1 promoter (Figure 4A). To elucidate the methylation status of the CpGs in these islands, we performed bisulfite sequencing on samples from six paired breast cancer and normal tissues. We detected methylated DNA only in CpG island 2 in normal breast tissues (Figure 4B). In contrast, breast tumors showed no methylation of either CpG island. Furthermore, we analyzed endogenous FEN1 expression by real-time RT-PCR in the same six paired breast cancer and normal tissues (Figure 4C). FEN1 mRNA was expressed at different levels in the specimens tested, and there was a good correlation with the observation that FEN1 is significantly up-regulated in tumors. These results suggest that hypomethylation of the FEN1 promoter in tumors is associated with increased expression of FEN1.
As a multifunctional nuclease, FEN1 is involved in DNA replication, repair and apoptosis. Although it is up-regulated in multiple types of cancer(21-25), it is unclear whether FEN1 expression is also increased in breast and other major cancers. Therefore, we investigated its mRNA abundance in matched tumor and normal samples using a cDNA array. Consistent with previous observations, FEN1 was over-expressed in 178 tumor cases as compared to normal tissues in all 236 primary tumor samples (75.42%) examined. However, we did not detect up-regulation of FEN1 in prostate cancer, which may be due to the small sample size (only 4 cases). FEN1 expression has been reported to be induced for DNA replication during cell proliferation (20). Thus, increased expression of FEN1 may reflect the increased proliferation rate of cancer cells. Moreover, when cancer cells are exposed to DNA alkylating agents, some cells can adapt to these exposures by increasing the expression of DNA polymerase β, a core enzyme in the BER pathway (28). Because FEN1 is involved in the BER pathway, increased expression of FEN1 may also be a response to increased DNA damage in cancer cells. Interestingly, we found that, in female cancers, not only the ratio of samples with FEN1 over-expression was the highest (86.92%), but also the amount of FEN1 over-expression. It has been reported that FEN1 could interact directly with estrogen receptor-alpha (ERα), enhance the interaction of ERα with ERE-containing DNA and influence estrogen-responsive gene expression (29, 30). Conversely, FEN1 expression can also be regulated by estrogen in the uterus (31). These findings suggest that FEN1 over-expression may be precisely regulated by hormones in female cancers.
In this study, we found greater FEN1 RNA expression in breast cancer than in other cancer types. Based on this observation, we analyzed endogenous FEN1 protein expression by immunohistochemical staining in a breast cancer progression tissue array. Consistent with the RNA expression results, FEN1 was up-regulated in tissues from BBDs with atypical hyperplasia and several kinds of breast cancer tissues. It has been shown that atypical hyperplasia has a relative risk of 4.24, proliferative disease without atypia has a relative risk of 1.88, and nonproliferative lesions have a relative risk of 1.27 (32). Interestingly, our results also show that FEN1 expression increases with an increase in relative cancer risk in BBD tissues. Moreover, in agreement with previous studies that examined the role of FEN1 expression in prostate cancer (5), we found that increased FEN1 expression was associated with tumor dedifferentiation in infiltrating ductal carcinoma. These results suggest that FEN1 might be a potential tumor marker for selecting patients at high risk of progression.
To determine the underlying mechanism of the differential transcriptional regulation of FEN1 between tumor and normal tissues, we investigated the possibility of epigenetic regulation of FEN1 expression via methylation. We found that over-expression of FEN1 was associated with hypomethylation within CpG island 2 of the FEN1 promoter in breast cancer. Genome-wide hypomethylation and regional hypermethylation of certain genes have been observed in several human cancers, including breast cancer (33, 34). The former may lead to activation of genes such as oncogenes and expression of provirus sequences; the latter may result in gene silencing of tumor suppressor genes (35). Therefore, it has been proposed that hypomethylation and hypermethylation in cancer are independent processes, which target different programs at different stages in tumorigenesis (36). Accumulating data have led to the hypothesis that hypomethylation plays a role in activating certain genes required for cancer progression and metastasis of breast cancer (37, 38). In agreement with these studies, we also examined the role of aberrant DNA hypomethylation in FEN1 gene expression in breast cancer progression.
In conclusion, our results suggest that FEN1 promoter hypomethylation may contribute to its over-expression in poorly differentiated carcinomas. To the best of our knowledge, this is the first study to show how methylation of FEN1 promoter region affects gene expression in breast cancer. Furthermore, the elevated expression of FEN1 may serve as a useful molecular marker for predicting prognosis and as a target for therapy.
To compare the expression of FEN1 in human tumors and corresponding normal tissues, we used the Cancer Profiling Array I (BD Biosciences Clontech Inc., Palo Alto, CA). This array includes normalized cDNAs from tumors of 241 individual patients and corresponding normal tissues, including breast, uterus, ovary, cervix, lung, kidney, stomach, colon, rectum, small intestine, pancreas and prostate as well as 12 cDNAs from metastases corresponding to 12 of the tumor/normal pairs. 32P-labelled cDNA probes were synthesized from human FEN1 or ubiquitin control cDNA using a random primer-labeling kit (New England Biolabs, Beverly, MA) followed by probe purification on CHROMA SPIN+STE-100 columns (Clontech). The FEN1 fragment corresponding to 463 bp-734 bp, was used to probe the filter. Hybridization of the Cancer Profiling Array I with FEN1 probes and washing of the array were done according to the manufacturer's recommendations (Clontech). The hybridized arrays were then exposed to phosphorimaging screen, scanned with a Typhoon Phosphorimager and analyzed using ImageQuant 1.2 software (Amersham Biosciences, Sunnyvale, CA). The array was then stripped and hybridized with the human ubiquitin cDNA probe to confirm the integrity of the samples on the array. Statistical analysis of the cancer-profiling array was done using GraphPad Prism 2.01 software and two-tailed paired t-test (P ≤ 0.05).
For in situ and in vivo FEN1 expression analysis, a breast carcinoma progression array (Cybrdi, Frederick, MD) was used. Immunohistochemical staining was performed using a monoclonal anti-FEN1 antibody (NCL-Fen-1; Novocastra, Newcastle upon-Tyne, UK) according to the manufacturer's instructions. Staining was semi-quantitatively assessed by the pathologist at the City of Hope Pathology Core Facility, and two criteria of FEN1 expression were scored: the intensity on a 0–3 scale (0, negative; 1, weakly positive; 2, moderately positive; 3, strongly positive), and the percentage of positively-stained target cells (0, no cells; 1, less than 10% of the cells; 2, 11–50% of the cells; 3, 51–80% of the cells; 4, more than 80% of the cells). Finally, a composite score was obtained by multiplying the values of the mean staining intensity and the percentage of FEN1-positive cells (0–1, negative; 1–2, weakly positive; 2–3, moderately positive; ≥3, strongly positive).
Specific primer pairs (Supplementary Table 1) with SacI and NheI restriction sites were used to amplify multiple deletion fragments spanning 5′ flanking region of FEN1 (from –1821 bp to +2662 bp, relative to the transcription start site) from human genomic DNA (Clontech) using Hot Start Taq (Qiagen, Valencia, CA). The PCR products were then digested with SacI and NheI (New England Biolabs) and ligated into an appropriately digested pGL4.10 vector (Promega, Madison, WI) containing the firefly luciferase gene as a reporter. The constructs were designated as pGL4-FP1 to pGL4-FP9 (Supplementary Table 1). Restriction analysis and complete DNA sequencing confirmed the orientation and integrity of the inserts.
Transient transfection of HeLa cells was performed using LipofectAmine (Invitrogen, Carlsbad, CA). Cultured cells were cotransfected with 1μg of one promoter construct and 0.1 μg of pGL4.74 [hRluc/TK] vector, a renilla luciferase control reporter vector (Promega, Madison, WI) that was used as an internal control to normalize the activities of the experimental reporters. After transfection in serum-free medium, the cells were allowed to recover in serum-containing medium. The cells were lysed 48 hrs post-transfection and assayed for promoter activity using the Dual–Luciferase Reporter Assay System (Promega, Madison, WI). The luciferase activity was measured using a Fluoroskan Ascent FL luminometer (Thermo Electron Corp., Waltham, MA). Luciferase values (relative light units, RLU) were calculated by dividing the firefly luciferase activity by the renilla luciferase activity. The assay was conducted three times in duplicates.
DNA methylation analysis and RNA preparation were performed on six paired specimens of breast tumors and corresponding normal tissues selected from the archives of Cancer Hospital, China. Histopathological diagnosis of the tumors was performed according to the World Health Organization classification (Supplementary Table 2). Patients with metastasized cancer from other organs were excluded. This study was approved by the institutional review board of the Cancer Institute and Hospital, Chinese Academy of Medical Sciences (Beijing, China).
The 5′ flanking region, exon1 and the intron of human FEN1 were analyzed by Methprimer software (http://www.urogene.org/methprimer/index1.html). Two CpG islands, CpG1-297 bp and 24 CpG dinucleotides and CpG2-399 bp and 37 CpG dinucleotides, were detected. Paired normal and breast caner genomic DNA samples were modified by bisulfite reaction using the EZ DNA Methylation Kit (Zymo Research, Orange, CA). Up to 2 μg of genomic DNA was used for conversion with the bisulfite reagent. Approximately 80 ng of bisulfite-converted DNA was used as a template for each PCR analysis. Primers for bisulfite sequencing were designed using the Methprimer software. The primer pairs CpG-1F 5’-AGTTGAGAAATTTAAGGAGT-3’, CpG-1R 5’-CTCCAAAAAAAACAAAATCT-3’ and CpG-2F 5’-GAGGGATTGGTTGTTATGAGAGTAG-3’, C p G-2R 5’-ACCCCATAAAATAAAACTTATTACC-3’ were used to amplify the target CpG islands with Hot Start Taq polymerase (Qiagen). The amplified fragments were cloned into the pSC-A vector (Stratagene). Individual clones were sequenced and compared with the original sequence to calculate the percentage of methylation.
Total RNA was isolated from paired tissues and converted to cDNA using an oligo (dT)15 primer and Superscript II (Invitrogen). Quantitation of relative gene expression for FEN1 and β-actin as an internal reference gene was carried out using the ABI Prism 7300 sequence detection system (Applied Biosystems, Foster City, CA) in triplicates, based on the SYBR-Green method . The primers used for FEN1 were 5′-CTGTGGACCTCATCCAGAAGCA-3′ and 5′-CCAGCACCTCAGGTTCCAAGA-3′; and for β-actin were 5′-GGCGGCACCACCATGTACCCT-3′ and 5′-AGGGGCCGGACTCGTCATACT-3′. The PCR specificity was confirmed by dissociation curve analysis and gel electrophoresis. The expression of individual FEN1 measurements was calculated relative to expression of β-actin using a modification of the method described by Lehmann and Kreipe (39).
Supplementary Table 1: Primer pairs used to amplify the Fen1 5′ flanking region
Supplementary Table 2: Characteristics of the subjects used in methylation analysis
Supplementary Table 3: FEN1 over-expression in matched normal/cancer tissues on cancer profiling array I
This work was supported by NIH grant CA073764 to B.H.S. and State Key Basic Research Program grant 2004CB518701 to D.L.
Conflict of interest statement: We declare that we have no conflict of interest.