The Wnt Antagonist sFRP1 in Colorectal Tumorigenesis (original) (raw)

Regions of the short arm of chromosome 8 are deleted frequently in a range of solid tumors, indicating that tumor suppressor genes reside at these loci. In this study, we have examined the properties of the Wnt signaling antagonist secreted frizzled-related protein (sFRP) 1 as a candidate for this role at c8p11.2. An initial survey of 10 colorectal tumors, selected by the presence of isolated short deletions of the 8p11.2 region, identified three chain-terminating mutations, all within the first exon, which encodes the cysteine-rich domain. None of these tumors exhibited microsatellite instability, indicating intact mismatch repair gene function. The preserved sFRP1 alleles in the remaining seven tumors each contained a polymorphic three-base insertion in the signal sequence, but in a broader study, no association was found between this and the development of colorectal cancer. Epigenetic inhibition of sFRP1 transcription was investigated, and increased methylation of the promotor region was demonstrated in an additional cohort of 51 locally advanced colorectal cancers. Hypermethylation was identified in 40 of 49 (82%) cancers and in only 11 of 36 (30%) matched normal mucosal samples (P < 0.001). Semiquantitative analysis, by real-time PCR, of mRNA expression in 37 of the same cohort of 51 cancers revealed that sFRP1 mRNA expression was down-regulated in 28 (76%) cases compared with matched normal large bowel mucosa. The 3′ end of the sFRP1 mRNA also was found to be alternatively spliced, compared with the prototype liver and lung forms, in the colon and a number of other tissues, yielding an extended COOH terminus, which may influence its activity in a tissue-specific manner. The inactivation and down-regulation of sFRP1 observed are consistent with it acting as a tumor suppressor gene in colorectal carcinogenesis. Because β-catenin is constitutively active in the majority of colorectal tumors, it is unlikely that sFRP1 can act in the canonical Wnt response pathway. Therefore, we propose that the reduced activity or absence of sFRP1 allows the transduction of noncanonical Wnt signals, which contribute to tumor progression.

INTRODUCTION

Interstitial deletions of chromosome 8p occur at a high frequency in a range of different cancers (1, 2, 3, 4, 5, 6, 7), and linkage studies indicate the breast cancer gene BRCA3 may reside on this arm (8, 9). Our studies of the region 8p11.2 in a series of prostate cancers (10), squamous cell head and neck cancers (11), and colorectal carcinomas (12) have shown frequent interstitial deletions in these different tumor types and an association between 8p11.2 deletion and local invasion (10, 12). We also have identified a minimum region of deletion between the Ankyrin-1 gene and the STS marker D8S515. The localization of sFRP1 to 8p11.2 and its function as an antagonist of Wnt signaling led us to investigate this gene as a candidate tumor suppressor gene (13).

The first mammalian Wnt was identified by its ability to promote mouse mammary tumorigenesis (14). The canonical Wnt response pathway operates by stabilizing β-catenin, enabling it to accumulate in the nucleus where it directs transcription of a range of genes in association with Lef/TCF factors (15). In colorectal cancer, inactivating mutations of APC or stabilizing mutations of β-catenin lead to constitutive activation of this pathway, and this has led to the assumption that Wnt signaling can make no additional contribution to tumor progression (16).

However, two β-catenin-independent Wnt response pathways have been described more recently. The planar cell polarity pathway, first identified in Drosophila (17), diverges from the β-catenin response upstream of APC and leads to activation of RhoA and JNK, whereas the Wnt/Ca pathway operates through PKC and CamKII (18). It is possible that at least one of these contributes to tumor progression because Wnt3 has been shown to direct cyclooxygenase 2 transcription through a β-catenin-independent route (19).

β-catenin regulates the expression of a number of Wnt pathway factors, including Axin 2 and hNkd1 (20), the latter acting to direct the response toward the planar cell polarity pathway (21). Thus, dysregulation of β-catenin activity is likely to lead to a redirection of subsequent Wnt responses toward the noncanonical pathways, and this may play a role in progression of the tumor.

Wnt inhibitory factor 1 (22), Dickkopfs (23), and the secreted frizzled-related proteins (sFRPs; Refs. 13, 24) are distinct classes of extracellular Wnt antagonists and may be expected to counter persistent or excessive stimulation by Wnts. Dickkopfs genes do not interact directly with Wnts but block the action of the lrp 5/6 coreceptor in canonical signaling (25), whereas Wnt inhibitory factor 1 may be specific for Wnt8 (22).

The sFRPs comprise an N-terminal domain homologous to the cysteine-rich domain (CRD) of the frizzled family of Wnt receptors and a COOH-terminal domain with some homology to netrin (13). The CRD domain of the sFRPs competes with the frizzled receptors for Wnt binding, modulating the signal (26).

The sFRP1 gene was proposed to lie at 8p11.2, within the region found to be deleted in our earlier tumor studies. Loss of expression has been shown recently to correlate with lymph node metastases and increased mortality in breast tumors (27).

We present data on the localization, gene structure, a frequent polymorphism in the NH2 terminus of sFRP1, and tissue-specific alternative splicing of the COOH terminus. Mutational analysis of the coding region demonstrated homozygous inactivating mutations in 30% of a selected series of colorectal tumors, indicating that sFRP1 might act as a tumor suppressor gene in colorectal carcinogenesis. We proceeded to investigate sFRP1 in a larger cohort of colorectal tumors, looking at the influence of epigenetic phenomena on expression levels, and genomic alterations.

MATERIALS AND METHODS

Microsatellite Instability Analysis.

Microsatellite instability was studied at five loci: BAT26 and BAT40, which are mononucleotide repeat markers, and D2S123, D8S255, and D13S175, which are dinucleotide repeat markers. Oligonucleotides used were as follows: BAT26 forward, 5′-TGACTACTTTTGACTTCAGCC; BAT26 reverse, 5′-AACCATTCAACATTTTTAACCC; BAT40 forward, 5′-ATTAACTTCCTACACCACACC; BAT40 reverse, 5′-GTAGAGCAAGACCACCTTG-3′; D2S123 forward, 5′-ACATTGCTGGAAGTTCTGGC; D2S123 reverse, 5′-CCTTTCTGACTTGGATACCA; D8S255 forward, 5′-TTTTGGAATTTCTAGCCTCC; D8S255 reverse, 5′-TGAAACCCACAGATATTGGG; D13S175 forward, 5′-TATTGGATACTTGAATCTGCTG; and D13S175 reverse, 5′-TGCATCACCTCACATAGGTTA. BAT26 and BAT40 were labeled fluorescently with tetrachloro-6-carboxyfluorescein, D2S123 and D13S175 were labeled with 6-carboxyfluorescein, and D8S255 was labeled with hexachloro-6-carboxyfluorescein.

PCR conditions for all of the reactions were as follows: 95°C for 5 min, then 24 cycles of 95°C for 1 min, 55°C for 1 min, and 72°C for 1 min, with a final extension at 72°C for 5 min. PCR products (1 μl) were combined with 1.5 μl loading buffer, size standard 0.5 μl 6-carboxytetramethylrhodamine, and 1 μl formamide (all supplied by Applied Biosystems, Foster City, CA). Products were denatured at 95°C for 4 min and resolved on an ABI 377 6% denaturing polyacrylamide gel (19:1; Applied Biosystems). Results were analyzed using ABI Genescan software (Applied Biosystems).

Immunohistochemistry.

The streptavidin-biotin indirect immunoperoxidase method was used as we have published previously (28). Briefly, 5-μm sections were dewaxed, rehydrated, and blocked by 10% H2O2 in methanol for 10 min. Microwave antigen retrieval was undertaken for 1 h. Sections were incubated overnight with primary antibody. Primary antibodies hMLH-1 and hMSH-2 were used at a dilution of 1:20 and 1:100, respectively (PharMingen, San Diego, CA). After washing with PBS, sections were incubated with biotinylated goat antimouse (Dako, Carpinteria, CA) according to the manufacturer’s instructions for 30 min. Serial PBS washing and incubation with streptavidin-peroxidase conjugate (Dako) was undertaken before incubation with diaminobenzidine tetrahydrochloride (Sigma, Poole, United Kingdom). Sections were counterstained with hemalum, dehydrated, and then analyzed by light microscopy.

Loss of Heterozygosity Analysis.

The colorectal tumors were examined for loss of heterozygosity using four microsatellite markers (D8S505 forward, 5′-AGCCTGCTATTTGTAGATAATGTTT; reverse, 5′-AGTGCTAAGTCCCAGACCA; D8S1722 forward, 5′-CCTTGCTGGGAATTGTG; reverse, 5′-AGCTGCCTGGCTAAGAG; D8S532 forward, 5′-GCTCAAAGCCTCCAATGAC; reverse, 5′-GACTTCGTGATCCACCTGC; D8S519 forward, 5′-CTGTCACCCCAGCGTC; and reverse, 5′-AGTGGCCTTTCTGCTCC) found on chromosome 8p11.2 close to the sFRP1 gene. Oligonucleotide primers were obtained from Alta Biosciences (University of Birmingham, Birmingham, United Kingdom).

All of the PCR reactions were prepared to a final volume of 25 μl, containing 2.5 μl of 10× BIOTAQ NH4-based reaction buffer (Bioline, Randolph, MA), 1 mm MgCl2, 0.2 mm deoxynucleoside triphosphate stock (Amersham Pharmacia Biotech, Piscataway, NJ), 5 pmol 33P γATP-labeled forward primer, 5 pmol of unlabeled reverse primer, 0.5 units of BIOTAQ DNA polymerase (Bioline), and 100 ng of genomic DNA. Amplification involved an initial step of 2 min at 94°C, followed by 35 cycles of 94°C for 10 s, 55°C or 59°C for 20 s, and 72°C for 30 s, with a final extension of 5 min at 72°C. The DNA was denatured and separated on a 6% acrylamide gel followed by autoradiography.

Analysis of Promoter Methylation Status.

The method used for DNA modification was essentially that of Grunau et al. (29). In short, 10 μg tRNA (Sigma) were added to 1 μg of genomic DNA and made up to 100 μl. Freshly prepared NaOH (Sigma) was added to a final concentration of 0.3 m, and the sample was incubated at 42°C for 20 min. A total of 1.2 ml 5.2–5.69 m sodium bisulfite (Sigma) and 10 mm hydroquinone (Sigma; pH 5) was added, and the solution was overlaid with mineral oil and incubated at 55°C for 4 h. DNA was desalted and redissolved into 100 μl of Tris-Cl (pH 8; Sigma). NaOH was added to a final concentration of 0.3 m, and the solution then was incubated at 37°C for 20 min. After incubation, the solution was neutralized; 10 μg of tRNA were added; and nucleic acids were precipitated with ethanol at −20°C overnight. Precipitated DNA was washed with 70% ethanol, dried, and resuspended in 50 μl of 1 mm Tris-HCl (pH 8).

Modified DNA was amplified with oligonucleotides specific for unmethylated DNA (forward, 5′-GAGTTAGTGTTGTGTGTTTGTTGTTTTGT; reverse, 5′-CCCAACATTACCCAACTCCACAACCA) with cycle conditions of 95°C for 5 min, 35 cycles of 95°C for 1 min, 59°C for 1 min, 72°C for 1 min, and 72°C for 7 min. Oligonucleotides specific for methylated DNA used were forward 5′-GTGTCGCGCGTTCGTCGTTTCGC and reverse 5′-AACGTTACCCGACTCCGCGACCG. Cycle conditions were the same as with unmethylated oligonucleotides with an annealing temperature of 63°C. Amplified products were analyzed on a 2% agarose gel.

Methylation status also was analyzed using the combined bisulfite restriction analysis (COBRA) method (30). The sFRP1 CpG island region was predicted using CpG plot (www.ebi.ac.uk). A strong CpG island (island size > 100 bp; GC percent > 50.0; and Obs/Exp > 0.6) was detected within the region −180 bp to +530 bp relative to the transcription start site of the sFRP1 gene. To analyze this region of the gene for methylation, PCR primers were designed that were specific for the bisulfite modified sequence. The sFRP1 CpG island was analyzed using nested PCR with the following primers: sFRP1 COBRA forward, 5′-GGTTAGTAGTTGGGTGTTTTTGTTTA and sFRP1 COBRA reverse, 5′-CCTTACCTTAAAACTTAAAAACTTC. One two hundred fiftieth of this then was used in a subsequent nested PCR reaction using the primers sFRP1 COBRA forward-nested 5′-TTGGGTGTTTTTGTTTAATAAGAATT and sFRP1 COBRA reverse-nested 5′-AAAACTTATCACACTTAAACATCTC. The PCR conditions used in both reactions were 94°C for 3 min; 30 cycles of 94°C for 20 s, 54°C for 30 s, 72°C for 40 s, and 72°C for 7 min; and 35 cycles were used in the nested PCR reaction. PCR products were incubated with restriction enzymes _Taq_I and _Bst_UI for 2 h at 65°C and 60°C, respectively, to assay for methylation and were visualized on a 2% agarose gel.

Semiquantitative Real-Time PCR.

Total RNA was extracted using TRI reagent (Helena Biosciences, Sunderland, United Kingdom). First-strand cDNA was synthesized from 2 μg of DNase-treated total RNA using Ready-To-Go You-Prime First-Strand Beads (Amersham Pharmacia Biotech) and random hexamers (Promega, Madison, WI).

Oligonucleotide primers and TaqMan probes were designed using Primer Express, version 1.5 (Applied Biosystems). The primers for sFRP1 (NM 003012) gene amplification were 5′-CCAATCCCACCGAAGCCT and 5′-ATGATGGCCTCAGATTTCAACTC. The sequence for the TaqMan fluorogenic probe for sFRP1 was 5′-CAAGCCCCAAGGCACAACGGTG. Data for the sFRP1 gene were normalized to the epithelial cell-specific gene keratin 8 (KRT8; NM 002273). For KRT8, the primers and probe were 5′-GATCGCCACCTACAGGAAGCT, 5′-ACTCATGTTCTGCATCCCAGACT, and 5′-CCGGCTCTCCTCGCCCTCCA, respectively. The TaqMan Universal PCR Master Mix and the TaqMan probes were purchased from Applied Biosystems. Primers were obtained from Alta Biosciences (University of Birmingham). Multiplex PCR amplifications were performed using an ABI PRISM 7700 sequence detector in a final volume of 25 μl (Applied Biosystems). Each reaction contained 12.5 μl of 2× TaqMan Universal PCR Master Mix (Applied Biosystems), 90 nm KRT8 and sFRP1 primers, 150 nm sFRP1 TaqMan probe, 175 nm KRT8 TaqMan probe, 1 μl of cDNA sample, and water. The thermal cycling conditions comprised an initial step at 50°C for 2 min and 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min.

Informed consent was obtained from each patient for molecular analysis of the resected specimen. To analyze differences between the tumor and normal specimens, χ2 test was used.

RESULTS

The Genomic Map of sFRP1.

A pool of PAC clones, already localized to 8p11.2, was screened for sFRP1. This identified two clones, 100H11 and 152M3, which contained the entire coding sequence of sFRP1, as judged by their ability to support amplification of 5′- and 3′-specific amplicons. Fluorescence in situ hybridization analysis of selected colorectal tumors, using these PAC clones as probes, confirmed them to lie within the deletion region at 8p11.2. These clones contained the STS marker D8S268, and deletion was confirmed in 30% of squamous cell head and neck carcinomas (data not shown).

The PAC clones were used to confirm the genomic map of sFRP1. A combination of PCR amplification and cycle sequencing revealed the presence of two introns in the coding sequence at nucleotides 540/541 and 618/619, as shown in Fig. 1.

This map corresponds to the expected domain structure of the sFRP1 protein. The first coding exon contains the whole of the frizzled-related CRD, whereas the third exon contains the netrin-related domain (31). The short middle exon probably represents a spacer between these two.

Identification of Protein-Truncating Mutations in Tumor DNA.

Flanking primers were designed to amplify each exon from genomic DNA for analysis by direct cycle sequencing. Genomic DNA was analyzed from 10 advanced colorectal tumors shown previously to carry heterozygous interstitial deletions at 8p11.2.

Mutations leading to premature termination of the translation product were identified in 3 of the 10 samples. These were two single-base deletions (26delG and 67delG) and a single-base change (G450A), generating an in-frame stop codon (Fig. 2 B). Each of these mutations was found within the first exon, shown previously to be sufficient for Wnt antagonist activity (32, 26). Of the 10 tumors analyzed, none contained truncating mutations in the second or third exons of sFRP1.

The possibility that the mutations had been generated by a mutator genotype was investigated. MMR protein expression (hMLH-1 and hMSH-2) was investigated in each tumor by immunohistochemistry. The presence of microsatellite instability was investigated using five different mono/dinucleotide repeat markers (BAT26, BAT40, D2S123, D8S255, and D13S175). No loss of protein expression was detected, and no instability at a DNA level was observed in any of these 10 tumors, indicating the mutations were unlikely to have arisen as a consequence of failure of MMR gene function.

An additional 51 tumors were analyzed by direct sequence analysis, of which 49 provided clearly interpretable results. Only the first exon was sequenced for stop codon mutations, but none were found. The remaining exons were not sequenced because no mutations had been identified in those regions in the preliminary analysis. This finding indicated that point mutation is not a frequent method of inactivation of the sFRP1 gene in colorectal cancer. The same bank of 51 locally advanced (T3/4) tumors then was analyzed for evidence of other methods of DNA disruption.

Exon 1 of sFRP1 Contains a Common Polymorphism.

In each of the seven tumors from the preliminary investigation without an identified truncating mutation, the retained sFRP1 allele contained an in-frame three-base insertion after nucleotide 37 (Fig. 1). This is predicted to lead to an extra alanine in the protein, after codon 13, and is represented in the expressed sequence tag (EST) database.

The presence of this variant was investigated by analysis of blood-derived DNA from 51 colorectal cancer patients (Table 1) and a separate cohort of 102 patients without a history of colorectal neoplasia. Thirty-two of 102 (31%) colorectal cancer-free individuals carried the variant compared with 18 of 51 (35%) patients with colorectal cancer. No significant association between the development of colorectal cancer and the presence of the 3-bp insertion was identified in this cohort of patients.

Increased Methylation of the sFRP1 Promotor Region in Tumor DNA.

The possibility that sFRP1 expression was modified by epigenetic factors was investigated in the same cohort of 51 cancers (Table 1), of which 49 gave interpretable results. Investigation of the methylation status of the promotor region of sFRP1 by methylation-specific PCR and COBRA (shown in Fig. 3) revealed hypermethylation in 40 (82%) cancers. Although these two approaches agreed broadly, COBRA was more sensitive and informative. We proceeded to analyze a selection of the matched normal colonic mucosal samples (n = 36; Table 1). Gels were scored by eye for the presence or absence of a methylated band, and images were quantified using the GeneTools analysis package (Syngene, Cambridge, United Kingdom) to provide the percentages given in Table 1. There were no cases in which sFRP1 was unmethylated in the tumor but methylated in the matched normal mucosa. Only 11 methylated matched normal mucosa samples (P < 0.001) were found. The mean (median) level of DNA modification differed between the groups, at 35% (33%) for the cancers and 10% (9%) for normal mucosae.

These data demonstrate that hypermethylation of the sFRP1 promoter region is a frequent event in colorectal cancer and is increased significantly compared with normal mucosa from the same patient.

The 3′ End of the sFRP1 Coding Region Is Alternatively Spliced.

Attempts to amplify sFRP1 from colonic mucosa cDNA with 3′ primers immediately downstream of the stop codon failed, whereas more distal primers yielded amplicons smaller than the predicted size (Fig. 4). Sequencing revealed that nucleotides 913-1005 were absent from this transcript, removing seven amino acids and the stop codon predicted in the original sequence and extending the reading frame by an additional 40 amino acids, terminating at nucleotide 1125.

A review of the splicing patterns in cDNAs from a range of tissues (Fig. 4) revealed that the extended protein is the predominant species. The unspliced form is the major species in only the lung and liver, whereas the prostate expresses another variant, which maintains the stop codon used in liver but lacks sequences from nucleotides 942-1092, downstream of the stop codon. Heart cDNA contains the colon and the prostate forms at low, but approximately equivalent, levels.

The extended sequence contains a hydrophobic region, which may act as a transmembrane anchor, modifying the localization of the protein. This may influence the function of sFRP1 in different tissues because an untethered protein may be more effective in antagonizing Wnt signaling to tumor cells in trans than would a membrane-bound form.

sFRP1 Transcription Is Down-Regulated in Colorectal Tumors.

We compared the sFRP1 transcript expression level in the same cohort of tumors. Thirty-seven colorectal tumor samples with matched normal mucosa gave analyzable RNA (Table 1; Fig. 5). The sFRP1 TaqMan probe was designed to span the exon1/intron1 boundary of the gene. sFRP1 expression in each tumor and normal colon sample was standardized to cytokeratin 8 (KRT8) gene expression. KRT8 was used as an epithelial cell-specific marker because stromal and inflammatory cell components may vary in tumor and matched normal epithelium. The expression of KRT8 was consistent between normal and tumor samples. Expression of sFRP1 in tumors was normalized to the mean of sFRP1 expression in the matched normal mucosa.

As shown in Fig. 5, sFRP1 mRNA was down-regulated by >10-fold in 28 of the 37 (76%) tumors compared with normal mucosa. In 6 of 37 (16%) tumors, there was a <10-fold change in expression level, and sFRP1 expression was up-regulated in 3 of 37 (7%) tumors.

The clinical and pathologic data from these 37 tumors were investigated, but no correlation was found between sFRP1 expression level and patient age, sex, tumor site, serosal spread, or presence of lymph node metastases.

DISCUSSION

Our studies indicate that sFRP1 acts as a tumor suppressor gene in colorectal carcinogenesis, as demonstrated by homozygous inactivation through interstitial deletion and truncating mutations. We also have identified that hypermethylation of the promotor region is a frequent event in a series of 51 locally advanced colorectal cancers. Reduced transcript levels were seen in >75% of these cases.

Our previous studies (10, 11, 12) showed localized deletion of the s_FRP1_ locus in advanced colorectal, prostate, and squamous head and neck tumors, indicating that it plays a role in the progression of many solid tumors. Other studies have demonstrated reduced transcription of sFRP1 in breast cancer (27, 33, 34) and gastric cancer (35).

s_FRP1_ levels were reduced in breast carcinomas and were maintained in benign breast tumors (34). One study of 70 breast tumors found reduced levels of sFRP1 mRNA to be associated with lymph node metastases and increased mortality (27). In our series of locally advanced colorectal tumors, we found no association between sFRP1 expression levels and the development of lymph node metastases or any other clinical or pathologic characteristics, although this had been suggested by our previous loss of heterozygosity studies (12). This issue may be clarified by analysis of less advanced colorectal cancers and premalignant adenomas for levels of sFRP1 expression. One possibility is that sFRP1 influences tumor progression at an earlier stage of tumor development in colorectal cancer than in breast cancer.

We identified a high frequency (82%) of methylation, comparable with the 95% found by Suzuki et al. (37), but there was not an absolute inverse correlation between methylation and transcription. There could be a number of reasons for this, including heterogeneity in the cancer samples, but the sFRP1 promotor has yet to be characterized fully; therefore, the contributions of each of its three CpG islands to transcriptional control are not known. Thus, our analysis provides an estimate of the frequency of methylation at the sFRP1 locus rather than a direct measurement of modification of a functional site. However, the observation that demethylation leads to expression of sFRP1 in RKO cells demonstrates that there can be a relationship between these processes, and our results show that methylation and transcriptional repression are common events in colorectal cancer (37).

The primary translation product of sFRP1 contains an atypical signal sequence, in which the hydrophobic domain is preceded by a stretch of 15 hydrophilic amino acids. We have identified a common polymorphism, which results in an extra amino acid after codon 13. This insertion was over-represented in the retained alleles of our primary cohort of colorectal tumors with interstitial loss, but secondary analysis of a larger series of tumors failed to show a statistically significant correlation between its presence and the risk of developing colorectal cancer. Additional studies are underway to determine whether this polymorphism has any direct effect on protein activity.

The COOH-terminal domain of sFRP1 is related to netrin 1 (31), a regulator of apoptosis via its interaction with DCC (38), and this netrin-related motif also is found in a range of other proteins where it is thought to mediate protein-protein interactions (39). The identification of sFRP1 as SARP2 (24) demonstrated its potential to promote apoptosis. If this is a response to Wnt occupancy of the N-terminal CRD, it may be modified by differences at the COOH terminus. We have identified alternative splicing, which leads to an extended COOH terminus in a range of tissues, including colon and prostate, and this may influence the role of sFRP1 in tumorigenesis.

According to the model of colorectal cancer progression that has emerged in recent years, APC and β-catenin mutations lead to constitutive stimulation of β-catenin transcription as an early or initiating event, and mutations in the canonical Wnt response pathway upstream of β-catenin would have little effect (16). However, Wnt signaling can act through at least two β-catenin-independent, noncanonical pathways (40, 41). This suggests a model whereby chronic β-catenin signaling leads to a shift in the Wnt response toward these alternative pathways and loss of the antagonist function of sFRP1 hypersensitizes the tumor to Wnt. This loss of stringency in growth factor responses could be an important step in tumor progression.

Grant support: Supported by Cancer Research United Kingdom.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Notes: K. Gensberg is currently at the School of Biosciences, The University of Birmingham, Birmingham, United Kingdom; S. Jan is currently at the Birmingham Clinical Trials Unit, Birmingham, United Kingdom; R. G. Hardy is currently at the Department of Clinical and Surgical Sciences (Surgery), Royal Infirmary of Edinburgh, Edinburgh, United Kingdom; S. Chughtai is currently at the Division of Medical Genetics, The University of Birmingham, Birmingham, United Kingdom.

Requests for reprints: Glenn Matthews, Department of Surgery, Queen Elizabeth Holspital, Birmingham, B15 2TH, United Kingdom.

Fig. 1.

Genomic map of sFRP1. PACs containing the sFRP1 locus (100H11 and 152M3) were analyzed by cycle sequencing using primers directed toward the expected location of the intron(s). The sequence obtained diverged from the cDNA sequence at nucleotides 540/541 and 618/619. The sequence positions relate to the cDNA sequence given by Finch et al. (13), GenBank accession no. AF001900, numbering modified to start from the A of the initiator codon. A, the positions of the introns relative to the coding regions of the exons. The figure also shows the location of the three mutations (26delG, 67delG, and G450A) and the polymorphism (37insGCA) identified within the gene. The unshaded region indicates the extent of the frizzled-related cysteine-rich domain. B, the sequence of the first 50 bases of each intron. These data were used to design primers to amplify the coding regions for direct sequence analysis. Additional sequence data are available (GenBank accession no. AF229189, AF229190, AF229191, and AF229192).

Fig. 1.

Close modal

Fig. 2.

sFRP1 mutations in tumors. The coding region of sFRP1 from 10 advanced colorectal tumors showing allelic imbalance at c8p11.2 was analyzed by direct sequencing of exon-derived amplicons. A, one of the inactivating mutations identified. This and the other two mutations found are indicated against the normal sequence in B. In the first two cases, single-base deletions (at nucleotides 26 and 67) lead to frameshifts, which would be predicted to cause truncation of the expression product at nucleotide 106. The third mutation (G450A) introduces an in-frame stop codon.

Fig. 2.

Close modal

Fig. 3.

Fig. 3. Combined bisulfite restriction analysis of DNA methylation. Normal (N) and tumor (T) genomic DNA was treated with bisulphate, amplified, and analyzed before (uncut) or after (cut) digestion with TaqI. Methylated DNA remains capable of cleavage, whereas unmethylated DNA is resistant. Samples A, B, C, and D correspond to samples 19, 21, 46, and 35 in Table 1<$REFLINK>, respectively.

Combined bisulfite restriction analysis of DNA methylation. Normal (N) and tumor (T) genomic DNA was treated with bisulphate, amplified, and analyzed before (uncut) or after (cut) digestion with _Taq_I. Methylated DNA remains capable of cleavage, whereas unmethylated DNA is resistant. Samples A, B, C, and D correspond to samples 19, 21, 46, and 35 in Table 1, respectively.

Fig. 3.

Close modal

Fig. 4.

Alternative splicing at the 3′ end of the sFRP1 mRNA. Amplification of the 3′ end of the sFRP1 mRNA by reverse transcription-PCR yielded a product from colonic mucosa (Co) mRNA that was 87 bases shorter than the product from liver mRNA (Li), as seen in A. B, sequencing revealed that this difference results from the loss of sequences between nucleotides 917 and 1005 (lowercase letter indicates the sequence removed by splicing), extending the reading frame to the stop codon at 1125. C, reverse transcription-PCR analysis of a range of tissues demonstrated that the shorter mRNA (longer protein) also is the major species expressed in small intestine (SI), ovary (Ov), spleen (Sp), and muscle (M), whereas the longer mRNA is restricted to liver (Li) and lung (Lu) in this series. The prostate (P) was found to express a third mRNA isoform, which lacks the sequences from nucleotide 942 to nucleotide 1093 (underlined in B). Because this region is downstream of the stop codon, the predicted protein from prostate is unaffected. Heart (H) contained low, but approximately equal, levels of the forms found in the prostate and the colon.

Fig. 4.

Close modal

Fig. 5.

Real-time PCR quantitation. Real-time PCR quantitation of relative sFRP1 mRNA expression in a series of 37 matched colorectal tumors according to the comparative CT method. The first step in the calculations is the normalization of the sFRP1 gene to the KRT8 gene to normalize quantity and quality of the cDNA samples. The level of sFRP1 mRNA expression in each tumor sample then was normalized to the mean of the results obtained for the adjacent matched normal tissue. sFRP1 normalized expression is given by _sFRP1_N = 2−ΔΔCT, where −ΔΔCT = ΔCT cancer − ΔCT normal. The _sFRP1_N values are shown using a logarithmic scale. sFRP1 mRNA levels were reduced substantially (>10-fold) in 28 of the tumors, and 6 of the tumors showed changes of <10-fold. Three of the tumors contained more sFRP1 mRNA than normal mucosa.

Fig. 5.

Close modal

Table 1

Summary of analysis of FRP1 gene in series of matched colorectal cancers

Tumor	Polymorphism	Methylation, %a	FRP1 expression (vs. normal)	Loss of heterozygosity
Normal	Insertion	Cancer	Normal
1	++		M31	M 11	7 × 10	+
2		++	M 49		1 × 10	NI
3	+	+	M 69	U 8	1 × 10	−
4	+	+	M 30	U 5	2 × 10	−
5		++	M 63		2 × 10	−
6	+	+	M 54	U 7	3 × 10	−
7	++		M 61	M 14	3 × 10	NI
8	+	+	M 33	U 11	3 × 10	−
9	+	+	M 51	M 17	3 × 10	−
10		++	M 48		3 × 10	−
11	++		M 58		4 × 10	−
12		++	U 8	U 9	4 × 10	−
13	+	+	M 41	M 14	5 × 10	−
14	++		M 31	U 6	9 × 10	−
15	+	+	U 14		9 × 10	−
16	+	+			1 × 10	+
17	+	+	M 28	M 13	1 × 10	NI
18	+	+	M 66	U 7	2 × 10	NI
19		++	U 9	U 7	2 × 10	NI
20	+	+	U 7	U 6	3 × 10	−
21	+	+	M 27	U 6	7 × 10	−
22		++	M 68	U 6	1 × 10	−
23	+	+	M 30	U 9	1 × 10	+
24	+	+	M 66	M 22	1 × 10	−
25		++	M 19		2 × 10	+
26	+	+	U 14	U 10	2 × 10	−
27		++	M 48	M 18	6 × 10	+
28		++	M 15	U 10	9 × 10	−
29		++	M 26		3 × 10	NI
30		++	M 33	M 12	3 × 10	NI
31	+	+	M 41		4 × 10	−
32	++		U 16	U 9	1.4	+
33	+	+			1.8	−
34		++	M 34	M 15	5	+
35	+	+	M 69	M 23	11	−
36	+	+	M 40		11	−
37		++	M 36		20	+
38	++		M 28	U 3		−
39		++	M 16	U 5		−
40	+	+	M 42	U 6		−
41	++		M 16	M 23		NI
42	++		U 8			−
43	++		M 38	U 11		+
44	+	+	M 16	U 11		−
45		++	M 40	U 5		−
46	+	+	M 56	U 6		−
47		++	U 9			−
48		++	M 79	U 8		−
49	+	+	M 20	U 7		−
50		++	M 21	U 7		−
51		++	U 8			−

Tumor	Polymorphism	Methylation, %a	FRP1 expression (vs. normal)	Loss of heterozygosity
Normal	Insertion	Cancer	Normal
1	++		M31	M 11	7 × 10	+
2		++	M 49		1 × 10	NI
3	+	+	M 69	U 8	1 × 10	−
4	+	+	M 30	U 5	2 × 10	−
5		++	M 63		2 × 10	−
6	+	+	M 54	U 7	3 × 10	−
7	++		M 61	M 14	3 × 10	NI
8	+	+	M 33	U 11	3 × 10	−
9	+	+	M 51	M 17	3 × 10	−
10		++	M 48		3 × 10	−
11	++		M 58		4 × 10	−
12		++	U 8	U 9	4 × 10	−
13	+	+	M 41	M 14	5 × 10	−
14	++		M 31	U 6	9 × 10	−
15	+	+	U 14		9 × 10	−
16	+	+			1 × 10	+
17	+	+	M 28	M 13	1 × 10	NI
18	+	+	M 66	U 7	2 × 10	NI
19		++	U 9	U 7	2 × 10	NI
20	+	+	U 7	U 6	3 × 10	−
21	+	+	M 27	U 6	7 × 10	−
22		++	M 68	U 6	1 × 10	−
23	+	+	M 30	U 9	1 × 10	+
24	+	+	M 66	M 22	1 × 10	−
25		++	M 19		2 × 10	+
26	+	+	U 14	U 10	2 × 10	−
27		++	M 48	M 18	6 × 10	+
28		++	M 15	U 10	9 × 10	−
29		++	M 26		3 × 10	NI
30		++	M 33	M 12	3 × 10	NI
31	+	+	M 41		4 × 10	−
32	++		U 16	U 9	1.4	+
33	+	+			1.8	−
34		++	M 34	M 15	5	+
35	+	+	M 69	M 23	11	−
36	+	+	M 40		11	−
37		++	M 36		20	+
38	++		M 28	U 3		−
39		++	M 16	U 5		−
40	+	+	M 42	U 6		−
41	++		M 16	M 23		NI
42	++		U 8			−
43	++		M 38	U 11		+
44	+	+	M 16	U 11		−
45		++	M 40	U 5		−
46	+	+	M 56	U 6		−
47		++	U 9			−
48		++	M 79	U 8		−
49	+	+	M 20	U 7		−
50		++	M 21	U 7		−
51		++	U 8			−

U, unmethylated; M, methylated; NI, non-informative.

Acknowledgments

We thank John Gregory, who assisted with the immunohistochemistry work on MMR gene expression.

References

Takanishi D. M., Jr., Kim S. Y., Kelemen P. R., Yaremko M. L., Kim A. H., Ramesar J. E., Horrigan S. K., Montag A., Michelassi F., Westbrook C. A. Chromosome 8 losses in colorectal carcinoma: localization and correlation with invasive disease.

Mol. Diagn.

-10,

1997

Adelaide J., Chaffanet M., Imbert A., Allione F., Geneix J., Popovici C., van Alewijk D., Trapman J., Zeillinger R., Borresen-Dale A. L., Lidereau R., Birnbaum D., Pebusque M. J. Chromosome region 8p11–p21: refined mapping and molecular alterations in breast cancer.

Genes Chromosomes Cancer

186

-199,

1998

El-Naggar A. K., Coombes M. M., Batsakis J. G., Hong W. K., Goepfert H., Kagan J. Localization of chromosome 8p regions involved in early tumorigenesis of oral and laryngeal squamous carcinoma.

Oncogene

2983

-2987,

1998

Wright K., Wilson P. J., Kerr J., Do K., Hurst T., Khoo S. K., Ward B., Chenevix-Trench G. Frequent loss of heterozygosity and three critical regions on the short arm of chromosome 8 in ovarian adenocarcinomas.

Oncogene

1185

-1188,

1998

Qin L. X., Tang Z. Y., Sham J. S., Ma Z. C., Ye S. L., Zhou X. D., Wu Z. Q., Trent J. M., Guan X. Y. The association of chromosome 8p deletion and tumor metastasis in human hepatocellular carcinoma.

Cancer Res.

5662

-5665,

1999

Piao Z., Kim N. G., Kim H., Park C. Deletion mapping on the short arm of chromosome 8 in hepatocellular carcinoma.

Cancer Lett.

138

227

-232,

1999

Lerebours F., Olschwang S., Thuille B., Schmitz A., Fouchet P., Buecher B., Martinet N., Galateau F., Thomas G. Fine deletion mapping of chromosome 8p in non-small-cell lung carcinoma.

Int. J. Cancer

854

-858,

1999

Seitz S., Rohde K., Bender E., Nothnagel A., Kolble K., Schlag P. M., Scherneck S. Strong indication for a breast cancer susceptibility gene on chromosome 8p12–p22: linkage analysis in German breast cancer families.

Oncogene

741

-743,

1997

Charafe-Jauffret E., Moulin J. F., Ginestier C., Bechlian D., Conte N., Geneix J., Adelaide J., Noguchi T., Hassoun J., Jacquemier J., Birnbaum D. Loss of heterozygosity at microsatellite markers from region p11–21 of chromosome 8 in microdissected breast tumor but not in peritumoral cells.

Int. J. Oncol.

989

-996,

2002

Crundwell M., Chughtai S., James N., Lee S., Wallace D., Neoptolemos J., Morton D., Phillips S. Allelic loss of DCC, chromosome 22 and chromosome 8 in human prostate cancer.

Int. J. Cancer

295

-300,

1996

Kuo M. J., Crundwell M. C., Knowles M., Oates J., Watkinson J. C., Warfield A. T., Morton D. G., Neoptolemos J. P. Allelic deletion at 8p11.2 is a frequent event in head and neck cancer and identifiable using FISH on tumour imprints.

Br. J. Surg.

1563

1998

Chughtai S. A., Crundwell M. C., Cruickshank N. R. J., Affie E., Armstrong S., Knowles M., Takle L., Kuo M., Khan N., Phillips S. M. A., Neoptolemos J. P., Morton D. G. Two novel regions of interstitial deletion on chromosome 8p in colorectal cancer.

Oncogene

657

-665,

1999

Finch P. W., He X., Kelley M. J., Üren A., Schaudies R. S., Popescu N. C., Rudikoff S., Aaronson S. A., Varmus H. E., Rubin J. S. Purification and molecular cloning of a secreted, frizzled-related antagonist of Wnt action.

Proc. Natl. Acad. Sci. USA

6770

-6775,

1997

Nusse R., van Ooyen A., Cox D., Fung Y. K., Varmus H. Mode of proviral activation of a putative mammary oncogene (int-1) on mouse chromosome15.

Nature (Lond.)

307

131

-136,

1984

Wodarz A., Nusse R. Mechanisms of Wnt signaling in development.

Annu. Rev. Cell. Dev. Biol.

-88,

1998

Morin P. J., Sparks A. B., Korinek V., Barker N., Clevers H., Vogelstein B., Kinzler K. W. Activation of β-catenin-Tcf signaling in colon cancer by mutations in β-catenin or APC.

Science (Wash. DC)

275

1787

-1790,

1997

Boutros M., Paricio N., Strutt D. I., Mlodzik M. Disheveled activates JNK and discriminates between JNK pathways in planar polarity and wingless signaling.

Cell

109

-118,

1998

Kuhl M., Sheldahl L. C., Park M., Miller J. R., Moon R. T. The Wnt/Ca2+ pathway: a new vertebrate Wnt signaling pathway takes shape.

Trends Genet.

279

-283,

2000

Haertel-Wiesmann M., Liang Y., Fantl W. J., Williams L. T. Regulation of cyclooxygenase-2 and periostin by Wnt-3 in mouse mammary epithelial cells.

J. Biol. Chem.

275

32046

-32051,

2000

Yan D., Wiesmann M., Rohan M., Chan V., Jefferson A. B., Guo L., Sakamoto D., Caothien R. H., Fuller J. H., Reinhard C., Garcia P. D., Randazzo F. M., Escobedo J., Fantl W. J., Williams L. T. Elevated expression of axin2 and hnkd mRNA provides evidence that Wnt/β-catenin signaling is activated in human colon tumors.

Proc. Natl. Acad. Sci. USA

14973

-14978,

2001

Yan D., Wallingford J. B., Sun T. Q., Nelson A. M., Sakanaka C., Reinhard C., Harland R. M., Fantl W. J., Williams L. T. Cell autonomous regulation of multiple disheveled-dependent pathways by mammalian Nkd.

Proc. Natl. Acad. Sci. USA

3802

-3807,

2001

Hsieh J-C., Rattner A., Smallwood P. M., Nathans J. Biochemical characterization of Wnt-frizzled interactions using a soluble, biologically active vertebrate Wnt protein.

Proc. Natl. Acad. Sci. USA

3546

-3551,

1999

Krupnik V. E., Sharp J. D., Jiang C., Robison K., Chickering T. W., Amaravadi L., Brown D. E., Guyot D., Mays G., Leiby K., Chang B., Duong T., Goodearl A. D., Gearing D. P., Sokol S. Y., McCarthy S. A. Functional and structural diversity of the human Dickkopf gene family.

Gene

238

301

-313,

1999

Melkonyan H. S., Chang W. C., Shapiro J. P., Mahadevappa M., Fitzpatrick P. A., Kiefer M. C., Tomei L. D., Umansky S. R. SARPs: a family of secreted apoptosis-related proteins.

Proc. Natl. Acad. Sci. USA

13636

-13641,

1997

Mao B., Wu W., Li Y., Hoppe D., Stannek P., Glinka A., Niehrs C. LDL-receptor-related protein 6 is a receptor for Dickkopf proteins.

Nature (Lond.)

411

321

-325,

2001

Bafico A., Gazit A., Pramila T., Finch P. W., Yaniv A., Aaronson S. A. Interaction of frizzled related protein (FRP) with Wnt ligands and the frizzled receptor suggests alternative mechanisms for FRP inhibition of Wnt signaling.

J. Biol. Chem.

274

16180

-16187,

1999

Wong S. C., Lo S. F., Lee K. C., Yam J. W., Chan J. K., Hsiao W. L. Expression of frizzled-related protein and Wnt-signalling molecules in invasive human breast tumours.

J. Pathol.

196

145

-153,

2002

Hardy R. G., Tselepsis C., Hoylands J., Wallis Y., Pretlow T. P., Talbot I., Sanders S. A., Matthews G., Morton D., Jankowski J. A. Z. Aberrant P-cadherin expression is an early event in hyperplastic and dysplastic transformation in the colon.

Gut

513

-519,

2002

Grunau C., Clark S. J., Rosenthal A. Bisulfite genomic sequencing: systematic investigation of critical experimental parameters.

Nucleic Acids Res.

E65

2001

Xiong Z., Laird P. W. COBRA: a sensitive and quantitative DNA methylation assay.

Nucleic Acids Res.

2532

-2534,

1997

Leyns L., Bouwmeester T., Kim S. H., Piccolo S., De Robertis E. M. Frzb-1 is a secreted antagonist of Wnt signaling expressed in the Spemann organizer.

Cell

747

-756,

1997

Lin K., Wang S., Julius M. A., Kitajewski J., Moos M., Jr., Luyten F. P. The cysteine-rich frizzled domain of Frzb-1 is required and sufficient for modulation of Wnt signaling.

Proc. Natl. Acad. Sci. USA

11196

-11200,

1997

Zhou Z., Wang J., Han X., Zhou J., Linder S. Up-regulation of human secreted frizzled homolog in apoptosis and its down-regulation in breast tumors.

Int. J. Cancer

-99,

1998

Ugolini F., Charafe-Jauffret E., Bardou V. J., Geneix J., Adelaide J., Labat-Moleur F., Penault-Llorca F., Longy M., Jacquemier J., Birnbaum D., Pebusque M. J. WNT pathway and mammary carcinogenesis: loss of expression of candidate tumor suppressor gene SFRP1 in most invasive carcinomas except of the medullary type.

Oncogene

5810

-5817,

2001

To K. F., Chan M. W., Leung W. K., Yu J., Tong J. H., Lee T. L., Chan F. K., Sung J. J. Alterations of frizzled (FzE3) and secreted frizzled related protein (hsFRP) expression in gastric cancer.

Life Sci.

483

-489,

2001

Uren A., Reichsman F., Anest V., Taylor W. G., Muraiso K., Bottaro D. P., Cumberledge S., Rubin J. S. Secreted frizzled-related protein-1 binds directly to Wingless and is a biphasic modulator of Wnt signaling.

J. Biol. Chem.

275

4374

-4382,

2000

Suzuki H., Gabrielson E., Chen W., Anbazhagan R., van Engeland M., Weijenberg M. P., Herman J. G., Baylin S. B. A genomic screen for genes upregulated by demethylation and histone deacetylase inhibition in human colorectal cancer.

Nat. Genet.

141

-149,

2002

Mehlen P., Rabizadeh S., Snipas S. J., Assa-Munt N., Salvesen G. S., Bredesen D. E. The DCC gene product induces apoptosis by a mechanism requiring receptor proteolysis.

Nature (Lond.)

395

801

-804,

1998

Banyai L., Patthy L. The NTR module: domains of netrins, secreted frizzled related proteins, and type I procollagen C-proteinase enhancer protein are homologous with tissue inhibitors of metalloproteases.

Protein Sci.

1636

-1642,

1999

Li L., Yuan H., Xie W., Mao J., Caruso A. M., McMahon A., Sussman D. J., Wu D. Dishevelled proteins lead to two signaling pathways. Regulation of lef-1 and c-jun N-terminal kinase in mammalian cells.

J. Biol. Chem.

274

129

-134,

1999

Sheldahl L. C., Park M., Malbon C. C., Moon R. T. Protein kinase C is differentially stimulated by Wnt and frizzled homologs in a G-protein-dependent manner.

Curr. Biol.

695

-698,

1999

2004

1,029 Views

249 Web of Science

The Wnt Antagonist sFRP1 in Colorectal Tumorigenesis (original) (raw)

Abstract

INTRODUCTION

MATERIALS AND METHODS

Microsatellite Instability Analysis.

Immunohistochemistry.

Loss of Heterozygosity Analysis.

Analysis of Promoter Methylation Status.

Semiquantitative Real-Time PCR.

RESULTS

The Genomic Map of sFRP1.

Identification of Protein-Truncating Mutations in Tumor DNA.

Exon 1 of sFRP1 Contains a Common Polymorphism.

Increased Methylation of the sFRP1 Promotor Region in Tumor DNA.

The 3′ End of the sFRP1 Coding Region Is Alternatively Spliced.

sFRP1 Transcription Is Down-Regulated in Colorectal Tumors.

DISCUSSION

Acknowledgments

References

Citing articles via

Email alerts