Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Frequent loss of SFRP1 expression in multiple human solid tumours: association with aberrant promoter methylation in renal cell carcinoma

Abstract

Oncogenic wingless-related mouse mammary tumour virus (Wnt) signalling, caused by epigenetic inactivation of specific pathway regulators like the putative tumour suppressor secreted frizzled-related protein 1 (SFRP1), may be causally involved in the carcinogenesis of many human solid tumours including breast, colon and kidney cancer. To evaluate the incidence of SFRP1 deficiency in human tumours, we performed a large-scale SFRP1 expression analysis using immunohistochemistry on a comprehensive tissue microarray (TMA) comprising 3448 tumours from 36 organs. This TMA contained 132 different tumour subtypes as well as 26 different normal tissues. Although tumour precursor stages of, for example kidney, colon, endometrium or adrenal gland still exhibited moderate to abundant SFRP1 expression, this expression was frequently lost in the corresponding genuine tumours. We defined nine novel tumour entities with apparent loss of SFRP1 expression, i.e., cancers of the kidney, stomach, small intestine, pancreas, parathyroid, adrenal gland, gall bladder, endometrium and testis. Renal cell carcinoma (RCC) exhibited the highest frequency of SFRP1 loss (89% on mRNA level; 75% on protein level) and was selected for further analysis to investigate the cause of SFRP1 loss in human tumours. We performed expression, mutation and methylation analysis in RCC and their matching normal kidney tissues. SFRP1 promoter methylation was frequently found in RCC (68%, n=38) and was correlated with loss of SFRP1 mRNA expression (p<0.05). Although loss of heterozygosity was found in 16% of RCC, structural mutations in the coding or promoter region of the SFRP1 gene were not observed. Our results indicate that loss of SFRP1 expression is a very common event in human cancer, arguing for a fundamental role of aberrant Wnt signalling in the development of solid tumours. In RCC, promoter hypermethylation seems to be the predominant mechanism of SFRP1 gene silencing and may contribute to initiation and progression of this disease.

Introduction

Human cancer cells show distinct patterns of deregulated gene expression. This deregulation is expected to affect 1–3% of the transcriptome (Zhang et al., 1997) including a variety of tumour suppressor genes whose expression is predominantly downregulated in cancer cells. DNA methylation of CpG-rich promoter regions is now becoming recognized as one fundamental mechanism for the inactivation of tumour suppressor genes in cancer cells (Esteller, 2005). For a number of tumour entities, it has already been reported that the gene encoding the soluble wingless-related mouse mammary tumour virus (Wnt) antagonist and putative tumour suppressor secreted frizzled-related protein 1 (SFRP1) is frequently inactivated by promoter methylation in solid tumours, including colon cancer (Caldwell et al., 2004; Suzuki et al., 2004), ovarian cancer (Takada et al., 2004), bladder cancer (Stoehr et al., 2004; Marsit et al., 2005), mesothelioma (Lee et al., 2004), prostate cancer (Lodygin et al., 2005), lung cancer (Fukui et al., 2005) and breast cancer (Veeck et al., 2006).

The innovative research of Suzuki et al. (2004) was able to uncover the direct link between epigenetic inactivation of the SFRP1 tumour suppressor gene and constitutive Wnt signalling in colorectal cancer. The binding of Wnt ligands to its corresponding Frizzled transmembrane receptors initiates a complex signalling cascade, which leads to a stabilization of cytosolic β-catenin. After translocation into the nucleus, β-catenin interacts with T-cell factor (TCF)/lymphoid-enhancing factor (LEF) transcription factors inducing an activation of downstream target genes as an immediate consequence (Behrens et al., 1996). The fact that several oncogenes like c-myc, cyclin D1 and c-jun (He et al., 1998; Mann et al., 1999; Tetsu and McCormick, 1999) are Wnt responsive genes argues for an involvement of constitutive Wnt signalling in the development of cancer after loss of SFRP1 expression. There is now growing evidence that oncogenic Wnt signalling may be involved in renal cell carcinoma (RCC) development as well (Togashi et al., 2005).

RCCs arise from the renal epithelium and account for approximately 85% of all renal tumours. According to the American Cancer Society, 38 890 new cases of renal cancer are predicted to occur in the United States in 2006 (Statistics for 2006; http://www.cancer.org). Despite recent improvements in RCC diagnostics, a quarter of patients are diagnosed with advanced disease with median survival times of approximately one year. Therefore, 12 840 individuals are estimated to have died due to RCC in the US in 2006, emphasizing the need for better diagnostic and therapeutic strategies.

Recently, it was reported that the hypoxia inducible gene 2 (HIG2) is abundantly overexpressed in RCC (Togashi et al., 2005). These authors found that an ectopic expression of HIG2 confers a growth advantage to RCC cell lines. HIG2 was supposed to exert its effect via binding to the Wnt receptor Frizzled-10 and subsequently initiating the translocation of β-catenin into the nucleus. Since the HIG2 promoter itself contains several putative TCF-binding motifs, it was suggested that oncogenic Wnt signalling in RCC cells is stimulated via a positive feedback system (Togashi et al., 2005). Additional studies have postulated a possible link between RCC and oncogenic Wnt signalling: Janssens et al. (2004) detected abundant Frizzled-5 expression in RCC and a close correlation to the accumulation of cyclin D1 in the nucleus. Thus, it is apparently important to analyse the expression of further Wnt genes and their negative regulators including SFRP1 in RCC.

In the present study, we addressed two major questions to further clarify the role of SFRP1 in carcinogenesis. The first goal was a comprehensive expression analysis of SFRP1 by immunohistochemistry (IHC) using a multitumour tissue microarray (TMA) containing 3448 human tumours derived from 132 tumour subtypes and 26 different normal tissues. This analysis defined the spectrum of human tumours showing loss of SFRP1 expression. Second, we selected RCC as the tumour entity with the highest frequency of SFRP1 loss for further molecular analysis. We investigated whether the putative tumour suppressor gene SFRP1 is inactivated by mutation or by methylation-dependent silencing in RCC. Promoter methylation has been shown to be the major mechanism of SFRP1 inactivation in breast and colon cancer (Suzuki et al. 2004; Veeck et al. 2006). To this end, we have analysed mutation, expression and promoter methylation of the SFRP1 gene in RCC and correlated the obtained data from the DNA, RNA and protein levels.

Results

Abundant loss of SFRP1 protein expression in numerous human tumour types

Of 3703 human tumour and normal tissue specimens on the TMA, 2616 (70.6%) were successfully analysed. Non-informative results were due to loss of tissue cores during the experimental process or due to missing tumour cells in specific TMA cores. Altogether, SFRP1 expression was analysed in 132 tumour subtypes derived from 36 organs and in 26 normal tissues (Supplementary Table 1). For statistical analysis, loss of SFRP1 expression was only analysed in those tumour subtypes with five or more analysed cases and where the corresponding normal or benign tissues exhibited moderate to abundant SFRP1 expression in at least 75% of analysed cases. This criterion was fulfilled by 17 cancer subtypes (marked with an asterisk in Supplementary Table 1) that included major tumour entities already known to exhibit loss of SFRP1 expression like adenocarcinoma of the colon (37% loss) and ductal carcinoma of the breast (44% loss). Altogether, 14 (82%) of these 17 cancer subtypes exhibited clear loss of SFRP1 expression compared to their benign or normal tissues (cutoff: at least 10% of cases shifted from moderated/strong expression to no or weak expression). Regarding the organ level, we defined nine novel tumour entities where loss of SFRP1 expression has not been described before (Table 1). These entities include tumours of the kidney, testis, pancreas, adrenal gland, parathyroid, endometrium, small intestine, stomach and gall bladder. Representative examples of SFRP1 IHC stainings in these entities are shown in Figure 1 (for pancreas, small intestine, stomach and testis) and Figure 2 (for renal cell carcinoma). Seminomas of the testis and adenocarcinomas of the pancreas showed negative or weak SFRP1 expression in 59 and 49% of analysed tumours, respectively (Table 1). In stomach cancer, there was considerable difference in SFRP1 loss between intestinal adenocarcinomas (48% negative or weak staining) and diffuse adenocarcinomas (27% negative or weak staining) (see Supplementary Table 1). In RCC we found most abundant loss of SFRP1 expression in clear cell carcinoma (51% weak or negative staining), whereas papillary RCC showed a smaller fraction of SFRP1 loss (32% weak or negative staining). In contrast, benign oncocytoma of the kidney retained SFRP1 expression in every specimen analysed (Table 1). Similar tendencies could be observed for cancer precursor stages of the colon and adrenal gland. Although colon adenomas in variable severities of dysplasia as well as adenomas of the adrenal gland retained SFRP1 protein expression, this expression was frequently lost in the corresponding genuine tumours.

Table 1 Nine novel tumour entities showing loss of SFRP1 expression as defined in this study
Figure 1
figure1

Representative examples of SFRP1 IHC in different tumour entities and corresponding normal tissues. (a and b) Pancreas tissue. Normal pancreatic tissue (a) exhibits strong SFRP1 expression, whereas adenocarcinoma of the pancreas (b) is mostly negative. (c and d) Small intestine tissue. Normal ileum tissue (c) exhibits strong SFRP1 expression. Small intestinal adenocarcinoma (d) shows loss or weak expression of SFRP1 protein. (e and f) Stomach tissue. Normal tissue derived from the corpus region of the stomach shows strong expression of SFPR1 (e) whereas SFRP1 expression is almost completely lost in intestinal adenocarcinomas of the stomach (f). (g and h) Testis tissue. Normal testis exhibits strong SFRP1 expression in Leydig cells (g). Seminoma of the testis shows clear loss of SFRP1 expression (h). Magnification in (a–h): × 200.

Figure 2
figure2

Representative examples of SFRP1 IHC in normal kidney tissue and different subtypes of kidney cancer. (a) Kidney, normal tissue strong positivity in proximal and distal tubules. (b) Kidney, clear cell carcinoma, no expression of SFRP1 is detectable. (c) Kidney, clear cell carcinoma, weak SFRP1 positivity. (d) Kidney, papillary carcinoma, focal SFRP1 positivity. (e) Kidney, chromophobe carcinoma, moderate SFRP1 expression. (f) Kidney, oncocytoma, abundant SFRP1 expression. Magnification in (a–f): × 200.

Verification of loss of SFRP1 expression by analysis of a cancer profiling array

To verify the loss of SFRP1 protein expression detected by IHC, we performed an expression analysis on the SFRP1 mRNA level. A complementary DNA (cDNA) dot blot array (Clontech cancer profiling array) containing 494 cDNAs synthesized from 241 primary tumours and 241 matched normal tissues and 12 cDNAs from metastases, altogether representing 13 different organs, was hybridized to a SFRP1-specific gene probe (Figure 3). The obtained hybridization signals were quantified by densitometry (Table 2). This analysis confirmed the expression loss obtained with our SFRP1-specific antibody by IHC in most tumour types. Loss of SFRP1 protein expression in the newly described entities stomach cancer and RCC could be confirmed on the RNA level, showing at least fivefold expression loss in 46% of stomach cancers and 89% of RCC. Absence of SFRP1-mRNA could also be confirmed in endometrium and small groups of tumours derived from small intestine, prostate and pancreas (see Figure 3). Interestingly, there was no evidence of SFRP1 loss in precursor stages of uterus/endometrium and kidney cancer. Of all tissues analysed on the mRNA level, SFRP1 was most clearly downregulated in RCC (Figure 3). These tumour samples included 17 primary RCCs, one carcinoid tumour (Figure 3, lane D), one transitional cell carcinoma (TCC) (Figure 3, lane F) and one oncocytoma (Figure 3, lane I). In summary, this analysis showed downregulation of SFRP1 in 15 of 17 RCC tumours and downregulation in one TCC and carcinoid tumour, but no loss in one benign oncocytoma analysed, compared to matching normal kidney tissue. Owing to this high frequency of SFRP1 loss in RCC, we subsequently concentrated our efforts on evaluating the mechanisms of SFRP1 loss in RCC.

Figure 3
figure3

Loss of SFRP1 mRNA expression in human solid tumours. Expression profiles were determined using the Clontech cancer profiling array containing cDNA pairs derived from cancer tissues (T), non-tumourous tissues (N) and metastatic tissues (marked with arrows and asterisks). No loss of SFRP1 expression was observed in benign tumours versus normal tissues (outlined pairs).

Table 2 Loss of SFRP1 mRNA expression as determined by the Cancer Profiling Array

Methylation of the SFRP1 promoter in primary renal cell carcinoma

Promoter methylation is implicated to be an effective mechanism of SFRP1 inactivation in breast, colon and prostate tumours. A key question was whether promoter methylation is also the driving mechanism in RCC that directs SFRP1 gene silencing. Since it has been reported that normal human tissue can also exhibit some degree of age-related gene methylation (Waki et al., 2003), we analysed matching samples of tumour and normal kidney tissue by methylation-specific polymerase chain reaction (MSP) analysis. Forty RCC, all of which had matching normal kidney tissue, were analysed. Representative results are shown in Figure 4. A clear PCR product could be amplified in 38 RCC samples using methylation-specific primers that can distinguish between unmethylated (TU, tumour unmethylated; NU, normal unmethylated) and methylated (TM, tumour methylated; NM, normal methylated) sequences in tumour and normal tissue of the kidney. Note that tumour tissue usually displayed a PCR product in the U-reaction as well, owing to contaminating normal tissue (stromal cells, endothelial cells) present in the tumour specimens. This effect has already been described by Suzuki et al. (2004) in colon carcinoma. However, only a few RCCs (e.g., #17, #26 in Figure 4) showed no evidence of aberrant SFRP1 promoter methylation in MSP. Overall, clear SFRP1 promoter methylation was found in 26/38 (68.4%) of RCCs. The majority of tumours investigated were clear cell carcinomas, although a few papillary and chromophobes were also analysed (Table 3).

Figure 4
figure4

SFRP1 promoter methylation in RCC. MSP was performed on bisulphite-treated DNA from RCC (T) and matching normal kidney tissue (N). MSP results from eight representative patients are shown. DNA bands in lanes labelled U indicate PCR products amplified with primers recognizing the unmethylated promoter sequence. DNA bands in lanes labelled M represent amplified products with methylation-specific primers. Human placenta (PLC) and BT20, a breast cancer cell line with methylated promoter region, served as positive controls for the U and M reactions, respectively. Water was used instead of template in the negative control (NTC).

Table 3 Clinicopathological parameters of 38 renal cell carcinomas analysed in this study

Correlation analysis of SFRP1 expression and methylation in RCC

To examine a potential correlation between SFRP1 transcription level and methylation, SFRP1 mRNA expression was determined by real-time PCR in the same cohort of RCC used previously for methylation analysis. Figure 5a compares SFRP1 mRNA expression in RCC with that found in matching normal kidney tissues, which demonstrates a statistically significant downregulation (median: 247-fold; p<0.001) in the malignant tissue samples. The fold change of SFRP1 downregulation with respect to the matching normal tissue is illustrated in Figure 5b. A distinct decrease of SFRP1 mRNA expression of more than fold change 5 could be observed in 34 of the 38 (90%) RCC tumour/normal pairs. The evaluation of statistical correlation data revealed significant differences between SFRP1 promoter methylation and SFRP1 mRNA downregulation (Mann–Whitney U test: p=0.048, Figure 5c). No correlation was found between SFRP1 methylation and available clinicopathologic data (age at diagnosis, tumour stage, histological grade and type) in this group of RCCs.

Figure 5
figure5

SFRP1 mRNA expression as quantified by real-time PCR and correlation with MSP data. (a) The level of SFRP1 mRNA expression in each RCC and normal kidney sample was normalized against the SFRP1 expression found in a commercially available normal kidney tissue sample (Clontech). (b) SFRP1 mRNA expression in RCC normalized to its expression in the matching normal kidney sample and presented as fold change (black bars, downregulation >5-fold versus normal; grey bars, non-differential expression). (c) Box plot showing the correlation between SFRP1 promoter methylation and loss of SFRP1 expression. Factor of SFRP1 mRNA downregulation relative to the matching normal kidney is shown on the y axis (fold change N/T). Methylated RCC exhibited a very distinct SFRP1 downregulation (median: 146-fold), whereas unmethylated RCC exhibited a SFRP1 downregulation that was less pronounced (median: 16-fold). Horizontal lines: group medians; boxes: 25–75% quartiles, range, peak and minimum.

To analyse whether the correlation between SFRP1 promoter methylation and loss of SFRP1 mRNA is also reflected on the protein level, we performed a small parallel analysis of SFRP1 DNA, RNA and protein in four tumour and normal tissue pairs of the kidney and found concordance in three of four samples. All tumour samples showed at least weak methylation of the SFRP1 promoter region. The tumour sample with the weakest signal and the tumour sample with the strongest methylation signal were selected for the representative data set shown in Figure 6. A RCC specimen with a weak SFRP1 methylation signal (Figure 6a) exhibited a 10-fold downregulation in its SFRP1 mRNA (Figure 6b) and the tumour still retained some SFRP1 protein expression (Figure 6d). In RCC specimens with clear SFRP1 methylation signals (Figure 6e), the downregulation of SFRP1 mRNA was much more prominent (160-fold) (Figure 6f) and correlated with almost complete loss of SFRP1 on the protein level (Figure 6h).

Figure 6
figure6

Comparison of SFRP1 promoter methylation, SFRP1 RNA expression and SFRP1 protein expression in representative samples of RCC and corresponding normal tissue. (a and e) Promoter methylation analysis. (b and f) Real-time PCR analysis. (c and d) and (g and h) SFRP1 immunohistochemistry of normal kidney tissues (c and g) as well as corresponding RCC (d and h). See text for details.

Loss of heterozygosity and mutation analysis of the SFRP1 gene in RCC

Eight RCC samples exhibited a clear reduction (fold change >5) of SFRP1 expression, despite the absence of SFRP1 promoter methylation (data not shown), indicating further inactivating mechanisms in RCC, for example gene loss or somatic mutations. Therefore, we screened our complete series of RCC samples with good-quality DNA (n=38) for allelic loss on chromosome 8p. Tumour DNA and matched normal DNA were analysed using the microsatellite markers D8S255, D8S1817, D8S532, D8S268 and D8S311 in the close vicinity of the SFRP1 gene, the markers D8S1469 and D8S1145 upstream at 8p23.1 and 8p22, respectively, and the marker D8S587 downstream (8q11.21) of the SFRP1 gene locus. Allelic loss on chromosome 8p in at least one marker was found in 16% (six of 38) of kidney tumours (representative examples are shown in Figure 7). Interestingly, all of them showed loss of heterozygosity (LOH) at chromosome 8p12, in close proximity to the SFRP1 gene. Three of the tumours showed smaller deletions with a minimal region of deletion at 8p12–11.1 (D8S311) or 8p12–11 (D8S255), whereas the remaining three tumours displayed LOH in at least two markers on chromosome 8p. The six tumours with LOH were further analysed for SFRP1 mutations by direct genomic sequencing of all SFRP1 exons and splice sites as well as the SFRP1 promoter region. No inactivating mutations within the SFRP1 gene were detected and no alterations of the SFRP1 promoter region were observed, indicating that mutational inactivation of the SFRP1 gene might not play a major role in kidney cancer.

Figure 7
figure7

LOH analysis at the SFRP1 gene locus. N, normal tissue; T, tumour tissue. Representative examples of (a) D8S311 and (b) D8S255. The shorter and the longer allele are lost in the tumours, respectively (arrows).

Discussion

The majority of human malignancies, i.e. cancers of the colon, breast and prostate, are characterized by a significant expression loss of the putative tumour suppressor gene SFRP1. SFRP1 is a negative modulator of the Wnt signalling pathway because its loss in colon cancer has been associated with aberrant activation of the Wnt signalling cascade (Suzuki et al., 2004). SFRP1 is thought to bind Wnt ligands and to prevent or modulate Wnt binding to its cognate receptors of the Frizzled family (Finch et al. 1997; Melkonyan et al., 1997).

In this report, we have comprehensively analysed SFRP1 expression by IHC in 26 different human normal tissues and 132 different subtypes of human cancers. Our expression data indicate that loss of SFRP1 expression is a very prevalent event in human tumours. In summary, 82% of analysed cancer subtypes exhibited clear loss of SFRP1 expression compared to their benign or normal tissues. For example, in breast cancer we found reduced SFRP1 expression in 44% of ductal and 56% of lobular carcinomas, thus confirming the data from Klopocki et al. (2004), who found loss of SFRP1 protein expression in 46% of invasive breast tumours. SFRP1 promoter hypermethylation and loss of SFRP1 mRNA expression in primary colorectal carcinomas have been found in 82 and 76% of cases, respectively (Suzuki et al., 2002, 2004). This fits in our observed frequency of SFRP1 downregulation on the mRNA level in cancer profiling array analyses (50–77%).

In addition to the previously characterized tumour entities, we defined nine novel tumour entities with apparent loss of SFRP1 expression, that is, kidney cancer, gastric cancer and pancreatic cancer, as well as cancers of the small intestine, parathyroid, adrenal gland, gall bladder, endometrium and testis. SFRP1 promoter methylation has already been eported to be relevant in gastric cancer cell lines (Suzuki et al., 2002) and matches well with our findings of reduced protein expression in diffuse and intestinal adenocarcinoma of the stomach. Aberrant SFRP1 promoter methylation in 85% (n=33) of genomic DNAs derived from pancreatic juices of patients with pancreatic cancer has been demonstrated by Watanabe et al. (2006). Our data showing loss of SFRP1 expression in 49% of pancreatic adenocarcinomas support the hypothesis that SFRP1 methylation in pancreatic cancer is a major factor for SFRP1 gene silencing in this tumour entity.

Recently, tumours of the testis have been studied for SFRP1 mRNA expression in a small group of ten cases and seven controls (Hoei-Hansen et al., 2004). These testis cancers represented one embryonal carcinoma, one teratoma, three seminomas and five intratubular germ cell neoplasias carcinoma in situ (CIS). Abundant expression of SFRP1 was found in all CIS samples and the teratoma sample, whereas no or weak SFRP1 mRNA expression was found in seminoma and normal tissue (Hoei-Hansen et al., 2004). Similar to this study, we found loss of SFRP1 expression in 59% of seminomas (n=46); however, we detected moderate or strong protein expression in the corresponding normal testis tissue.

Recently, a distinct role of SFRP1 in wound healing of the skin was reported by Li and Amar (2006). Loss of SFRP1 in a variety of skin tumours, especially histiocytoma, capillary hemangioma and Kaposi's sarcoma, might also indicate a possible involvement of aberrant Wnt signalling in these skin tumours. The relevance of loss of SFRP1 expression in the carcinogenesis of the stomach, the endometrium, the small intestine and the parathyroid has not been reported so far and is a matter of ongoing studies in our laboratory.

The Wnt signalling pathway also plays an important role in embryonic kidney development (Vainio et al., 1999; Dressler, 2002) and its reactivation by either mutations or altered expression of individual components has been postulated to be a determining factor in the development of RCCs. Several Wnt components have already been examined: despite lack of adenomatosis polyposis coli (APC) and only rare β-catenin mutations (Bohm et al., 1997; Suzuki et al., 1997; Kim et al., 2000; Ueda et al., 2001), cytoplasmic accumulation of β-catenin has repeatedly been described (Kim et al., 2000; Janssens et al., 2004). Moreover, aberrant activity of the Wnt signalling pathway was reported in renal-cancer-derived cell lines associated with higher expression levels of Wnt5a and Frizzled 5 mRNA and/or increased β-catenin expression (Zang et al., 2000). Furthermore, mRNA as well as protein levels of the Wnt target cyclin D1 were reported to correlate largely with increased Frizzled 5 expression in clear cell RCC (Janssens et al., 2004). As already mentioned above, SFRP1 functions as a Wnt antagonist and its loss has been reported to be associated with activation of the Wnt signalling pathway. However, the role of SFRP1 in RCC development has not been investigated yet. Thus, after verifying the immunohistochemistry results on the RNA level by means of a cancer profiling array (CPA) hybridization, we focussed on the underlying mechanism of loss of SFRP1 expression in kidney cancer. Hypermethylation of CpG islands in gene promoter regions has been associated with transcriptional silencing and represents, in addition to genetic alterations, an important mechanism of gene inactivation in tumorigenesis. We found SFRP1 promoter hypermethylation in 68% of RCCs and a correlating downregulation of SFRP1 mRNA expression by a fold change >5 in 90% of these cases, resulting in attenuated protein expression. Additionally, we performed 8p LOH analysis on the same, predominantly clear cell RCC specimens and corresponding normal tissue. LOH at 8p in clear cell RCCs was previously reported to occur in 15–32% of cases, depending on the detection method applied (Reutzel et al., 2001; Presti et al., 2002). We focussed mainly on markers located near the SFRP1 locus 8p12.1 and detected LOH in only 16% of cases, being in concordance with the data described above. Moreover, we could not find any mutations within the SFRP1 gene, like they were infrequently found in colorectal carcinomas (Caldwell et al., 2004), and no alterations in the SFRP1 promoter region. Therefore, we conclude that promoter methylation rather than mutational inactivation is the main mechanism for SFRP1 gene silencing in RCC. This is in line with previous findings such as in bladder (Stoehr et al., 2004), breast (Veeck et al., 2006) and prostate cancer (Lodygin et al., 2005). It is to note that these data were obtained by analysing mainly clear cell RCC (33/40). This entity displayed the highest fraction of loss of SFRP1 protein among kidney tumours analysed on the multitumour TMA and is also the most frequent and clinically aggressive variant. Like papillary RCC, this entity is thought to develop from the tubular epithelium (Mancilla-Jimenez et al., 1976; Hughson et al., 1993). Contrarily, chromophobe RCCs, which retained SFRP1 protein expression in 54% of cases, are supposed to descend from the intercalating cells of the collecting duct epithelium and are therefore of different origin (Storkel et al., 1989; Ortmann et al., 1991).

In summary, our study provides additional insights into the prevalence, incidence and cause of SFRP1 gene inactivation and its potential contribution to neoplastic progression of human tumours. In clear cell RCCs, aberrant cytosine methylation in relevant SFRP1 promoter regions seems to be a fundamental mechanism that might lead to ectopic activation of the Wnt signalling pathway. This work is a prerequisite for further functional studies, clarifying the importance of the Wnt signalling pathway in the pathogenesis of RCC and other tumour entities.

Materials and methods

Clinical material

Matched tumour/normal samples of renal cancer specimens (n=40) were obtained from patients who had undergone primary surgery for RCC at the Department of Urology at the University Hospital of Regensburg. All patients gave informed consent to participate in the study. Part of the tumour material and macroscopically normal kidney was snap-frozen in liquid nitrogen or embedded in paraffin shortly after surgery. Hematoxylin and eosin-stained frozen sections were prepared for assessment of the percentage of tumour cells. Only samples with more than 50% tumour cells were selected for analysis and prepared for manual microdissection. In all cases two board-certified pathologists (AH and MW) agreed on the diagnosis of RCC. Patient characteristics and histopathological data are shown in Table 3.

TMA construction

Formalin-fixed and paraffin-embedded tissue samples were collected from the archives of the Institute of Pathology of the University Hospital, Basel, Switzerland. The assessment of tissue and clinical data was performed according to the regulations of the local institutional review board. All relevant patient data were anonymized. A comprehensive TMA was created by transferring representative tissue cylinders with a diameter of 0.6 mm to seven new paraffin blocks as described previously (Bubendorf et al., 2001). Representative areas of different subtypes for the most frequent tumour entities and their corresponding normal tissues were selected for analysis. Sections of 4 μm from the resulting TMA block were cut and mounted to an adhesive-coated slide system (Instrumedics Inc., Hackensack, NJ, USA). The multitumour TMA consisted of 3448 primary tumours from 132 different tumour subtypes and 26 different normal tissues and allowed us to determine the prevalence of SFRP1 expression in normal tissues and corresponding malignant tumours. The major tumour tissues included on this TMA comprise skin (n=330), brain (n=228), breast (n=218), lung (n=217), colon (n=204), salivary gland (n=152), soft tissue (n=150), kidney (n=144), ovary (n=140) and testis (n=126), Detailed tumour characteristics can be found in Supplementary Table 1.

SFRP1 immunohistochemistry

Section of 3 μm were deparaffinized in xylene, rehydrated in a decreasing ethanol series and subsequently boiled for 30 min in citrate buffer (pH 6.0) for antigen retrieval. IHC was performed using an NEXES Immuno Stainer (Ventana, Tucson, AZ, USA) according to the manufacturer's specifications. Polyclonal anti-SFRP1 antibodies (Eurogentec, Seraing, Belgium) were verified for specificity as described previously (Klopocki et al., 2004) and used in a 1:200 dilution using the ChemMate Envision kit (Dako, Hamburg, Germany). Normal breast tissue sections in which SFRP1 expression has already been described by Klopocki et al. (2004) served as positive controls. Sections were counterstained with Mayer's hematoxylin and embedded in Entellan mounting medium (EMS Diatome, Fort Washington, PA, USA). The application of primary antibodies to tissue sections was omitted in negative controls. Immunohistochemical analysis was performed without the knowledge of histopathological data. Cytoplasmic staining was scored by a four-step scale from 0 (no expression) to 3+ (strong expression) for semiquantitative evaluation of the multiple tumour TMA by two pathologists (MW and AH).

Expression analysis using the CPA

The CPA (Clontech, Heidelberg, Germany) has been previously described in detail (Zafrakas et al., 2006). Histological type, patient age and a complete list of tissues can be found on the provider's website (http://www.clontech.com/clontech/techinfo/manuals/PDF/pt7841-1.pdf). Hybridization using 25 ng of a gene-specific 32P-labelled cDNA probe digested from Unigene cDNA clone (acc. no. W21306) was performed according to the manufacturer's recommendations. The tumour/normal intensity ratio was calculated using a STORM-860 phosphorimager (Molecular Dynamics, Sunnyvale, CA, USA) and normalized against the background.

DNA/RNA isolation from RCC

Frozen tissue samples were prepared for manual microdissection and dissolved in lysis buffer for subsequent DNA isolation using the blood and cell culture DNA kit (Qiagen, Hilden, Germany) or by RNA isolation using Trizol (Gibco-BRL, Glasgow, UK) according to the protocol supplied by the manufacturer.

Reverse transcription PCR

Of the total RNA, 1 μg was reverse transcribed using the Reverse Transcription System (Promega, Madison, WI, USA). To improve transcription rate we mixed oligo-dT and pdN(6)-primers 1:2. For PCR, 1 μl cDNA was amplified using SFRP1 (product size 599 bp) and glycerinaldehyde-3-phosphate dehydrogenase (GAPDH) (510 bp) primers given in Supplementary Table 2. Reactions were initiated as ‘Hot Start’ PCR at 95°C for 5 min and held at 80°C before the addition of 1 unit of Taq DNA polymerase (Roche, Mannheim, Germany). Cycle conditions applied for both genes were as follows: 95°C for 5 min, 35 cycles of 95°C for 1 min, 60°C for 1 min, 72°C for 1 min and a final extension at 72°C for 10 min. PCR analyses were carried out in a PTC-200 cycler (Bio-Rad, formerly MJ Research, Hercules, CA, USA). The amplification products were analysed on a 1% agarose gel containing ethidium bromide under UV light.

Semiquantitative real-time PCR

Semiquantitative PCR was performed using the LightCycler system together with the LightCycler DNA Master SYBR Green I Kit (Roche, Mannheim, Germany) as described previously (Veeck et al., 2006). Gene expression was quantified by the comparative CT method, normalizing CT-values to the housekeeping gene GAPDH and calculating relative expression values (Fink et al., 1998). Postamplification melting curve analyses were performed to assure product specificity.

Primer sequences for SFRP1 (product size 168 bp) and GAPDH (150 bp) are listed in Supplementary Table 2. The relative SFRP1 expression levels were standardized to the expression level of a normal kidney tissue sample that contained approximately 40% of epithelial (renal tubular) cells (tumours generally contained >50% of tumour cells). To ensure experiment accuracy, all reactions were performed in triplicate.

Bisulphite modification and MSP

Bisulphite modification and MSP were performed as described previously (Veeck et al., 2006). Of the genomic DNA, 1 μg was bisulphite-treated using the DNA modification Kit (Chemicon, Ternecula, CA, USA) according to the manufacturer's specifications. For MSP, 1 μl of modified DNA was amplified using MSP primers (see Supplementary Table 2) that specifically recognize the unmethylated (product size 135 bp) or methylated (126 bp) SFRP1 promoter sequence after bisulphite conversion. DNA derived from human placenta was bisulphite-treated to serve as a control for the unmethylated promoter sequence. DNA derived from human mamma carcinoma cell line BT20, recently described as hypermethylated in the SFRP1 promoter (Veeck et al., 2006), was used to generate a positive control for methylated alleles. Amplification products were visualized by UV illumination on a 3% low-range ultra agarose gel (Bio-Rad Laboratories, Hercules, CA, USA) containing ethidium bromide.

LOH analysis

The microsatellite markers used in this study and their localization (see Supplementary Table 2) were taken from the Genome Database (http://www.gdb.org). All primers were obtained from Proligo (Hamburg, Germany). Matched normal/tumour DNA samples were amplified by PCR in a 25-μl volume containing 0.2 mM dNTP, 0.18 μ M primers, 1.5 mM MgCl2 and 0.25 units Taq-polymerase (Promega, Mannheim, Germany) using 3 μl isolated DNA as template. The reaction mixture was subjected to 3 min of denaturing at 95°C and 35 cycles of 95°C for 1 min and specific annealing temperature for 1 min and 72°C for 1 min, followed by a final extension step at 72°C for 10 min. PCR conditions were optimized by gradient PCR and were carried out in a PTC-100 Thermocycler (Bio-Rad, formerly MJ Research, Hercules, CA, USA). Primer sequences and annealing temperatures are provided in Supplementary Table 2. The amplification products were visualized by polyacrylamide gel electrophoresis and silver staining as described (Schlegel et al., 1995). The silver-stained gels were assessed visually by two independent observers (CH and RSt) and informative cases were scored as allelic loss when intensity of the signal for the tumour allele was decreased to at least 50% relative to the matching normal allele. To avoid errors due to preferential amplification of one allele during the PCR, all LOH analyses were run in duplicates following independent PCRs.

SFRP1 genomic sequencing

Genomic DNA obtained from six LOH-positive clinical specimens, which contained >50% of tumour cells, were used as template for amplification of 500–1000 bp fragments containing the regions of interest. Oligonucleotide sequences are provided in Supplementary Table 2. Gel-purified PCR products were applied to the sequencing reaction containing Big Dye Terminator Mix and Big Dye Sequencing Buffer according to the BigDye Terminator v.1.1 Cycle Sequencing Kit (ABI, Warrington, UK) and were sequenced in forward and reverse directions. The sequencing mixture was subjected to 1 min of denaturing at 96°C and 25 cycles of 96°C for 10 s and specific annealing temperature for 5 s and elongation at 60°C for 4 min, followed by a final extension step at 72°C for 10 min. Sequence detection was performed using an ABI Prism 310 Genetic Analyzer (ABI, Weiterstadt, Germany) and analysed using BioEdit Sequencing Alignment Editor v.7.0.5.3 (see: http://www.mbio.ncsu.edu/BioEdit). The coding region of all three exons and the promoter region 1000 bp upstream from the transcription start site were analysed for base substitutions.

Statistical analysis of clinicopathologic patient data

Statistical analyses were completed using SPSS version 10.0 (SPSS, Chicago, IL, USA). Differences were considered statistically significant when p values were below 0.05. The non-parametric Mann–Whitney U test was used to compare the delta CT values of the real-time reverse transcriptase–PCR results between the RCC group and the normal kidney group and also to compare the different methylation groups.

References

  1. Behrens J, von Kries JP, Kuhl M, Bruhn L, Wedlich D, Grosschedl R et al. (1996). Functional interaction of beta-catenin with the transcription factor LEF-1. Nature 382: 638–642.

    CAS  Article  Google Scholar 

  2. Bohm M, Wieland I, Sinhofer C, Otto T, Rubben H . (1997). Detection of loss of heterozygosity in the APC tumor suppressor gene in nonpapillary renal cell carcinoma by microdissection and polymerase chain reaction. Urol Res 25: 161–165.

    CAS  Article  Google Scholar 

  3. Bubendorf L, Nocito A, Moch H, Sauter G . (2001). Tissue microarray (TMA) technology: miniaturized pathology archives for high-throughput in situ studies. J Pathol I 95: 72–79.

    Article  Google Scholar 

  4. Caldwell GM, Jones C, Gensberg K, Jan S, Hardy RG, Byrd P et al. (2004). The Wnt antagonist sFRP1 in colorectal tumorigenesis. Cancer Res 64: 883–888.

    CAS  Article  Google Scholar 

  5. Dressler GR . (2002). Tubulogenesins in the developing mammalian kidney. Trends Cell Biol 12: 390–395.

    CAS  Article  Google Scholar 

  6. Esteller M . (2005). Aberrant DNA-methylation as a cancer inducing mechanism. Annu Rev Pharmacol Toxicol 45: 629–665.

    CAS  Article  Google Scholar 

  7. Finch PW, He X, Kelley MJ, Uren A, Schaudies RP, Pupesco NC et al. (1997). Purification and molecular cloning of a secreted, frizzled-related antagonist of Wnt action. Proc Natl Acad Sci USA 94: 6770–6775.

    CAS  Article  Google Scholar 

  8. Fink L, Seeger W, Ermert L, Hanze J, Stahl U, Grimminger F et al. (1998). Real-time quantitative RT–PCR after laser-assisted cell picking. Nat Med 4: 1329–1333.

    CAS  Article  Google Scholar 

  9. Fukui T, Kondo M, Ito G, Maeda O, Sato N, Yoshioka H et al. (2005). Transcriptional silencing of secreted frizzled related protein 1 (SFRP 1) by promoter hypermethylation in non-small-cell lung cancer. Oncogene 24: 6323–6327.

    CAS  Article  Google Scholar 

  10. He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT et al. (1998). Identification of c-MYC as a target of the APC pathway. Science 281: 1509–1512.

    CAS  Article  Google Scholar 

  11. Hoei-Hansen CE, Nielsen JE, Almstrup K, Hansen MA, Skakkebaek NE, Rajpert-DeMeyts E et al. (2004). Identification of genes differentially expressed in testes containing carcinoma in situ. Mol Hum Reprod 10: 423–431.

    CAS  Article  Google Scholar 

  12. Hughson MD, Johnson LD, Silva FG, Kovacs G . (1993). Nonpapillary and papillary renal cell carcinoma: a cytogenetic and phenotypic study. Mod Pathol 6: 449–456.

    CAS  PubMed  Google Scholar 

  13. Janssens N, Andries L, Janicot M, Perera T, Bakker A . (2004). Alteration of frizzled expression in renal cell carcinoma. Tumor Biol 25: 161–171.

    CAS  Article  Google Scholar 

  14. Kim AS, Kang YK, Kim JB, Kondo Y, Hirohashi S . (2000). Beta-catenin expression and mutational analysis in renal cell carcinomas. Pathol Int 50: 725–730.

    CAS  Article  Google Scholar 

  15. Klopocki E, Kristiansen G, Wild PJ, Klaman I, Castanos-Velez E, Singer G et al. (2004). Loss of SFRP1 is associated with breast cancer progression and poor prognosis in early stage tumors. Int J Oncol 25: 641–649.

    CAS  PubMed  Google Scholar 

  16. Lee AY, He B, You L, Dadfarmay S, Xu Z, Mazieres J et al. (2004). Expression of the secreted frizzled-related protein gene family is downregulated in human mesothelioma. Oncogene 23: 6672–6676.

    CAS  Article  Google Scholar 

  17. Li CH, Amar S . (2006). Role of secreted frizzled-related protein 1 (SFRP1) in wound healing. J Dent Res 85: 364.

    CAS  Article  Google Scholar 

  18. Lodygin D, Epanchintsev A, Menssen A, Diebold J, Hermeking H . (2005). Functional epigenomics identifies genes frequently silenced in prostate cancer. Cancer Res 65: 4218–4227.

    CAS  Article  Google Scholar 

  19. Mancilla-Jimenez R, Stanley RJ, Blath RA . (1976). Papillary renal cell carcinoma: a clinical, radiologic, and pathologic study of 24 cases. Cancer 38: 2469–2480.

    CAS  Article  Google Scholar 

  20. Mann B, Gelos M, Siedow A, Hanski ML, Gratchev A, Ilyas M et al. (1999). Target genes of beta-catenin-T cell-factor/lymphoid-enhancer-factor signaling in human colorectal carcinomas. Proc Natl Acad Sci USA 96: 1603–1608.

    CAS  Article  Google Scholar 

  21. Marsit CJ, Karagas MR, Andrew A, Liu M, Darnaee H, Schned AR et al. (2005). Epigenetic inactivation of SFRP genes and TP53 alteration act jointly as markers of invasive bladder cancer. Cancer Res 65: 7081–7085.

    CAS  Article  Google Scholar 

  22. Melkonyan HS, Chang WC, Shapiro JP, Mahadevappa M, Fitzpatrick PA, Kiefer MC et al. (1997). SARPs. A family of secreted apoptosis-related proteins. Proc Natl Acad Sci USA 94: 13636–13641.

    CAS  Article  Google Scholar 

  23. Ortmann M, Vierbuchen M, Fischer R . (1991). Sialyated glycoconjugates in chromophobe cell renal carcinoma compared with other renal cell tumors. Indication of its development from the collecting duct epithelium. Virchows Arch B Cell Pathol Mol Pathol 61: 123–132.

    CAS  Article  Google Scholar 

  24. Presti Jr JC, Wilhelm M, Reuter V, Russo P, Motzer R, Waldman F . (2002). Allelic loss on chromosomes 8 and 9 correlates with clinical outcome in locally advanced clear cell carcinoma of the kidney. J Urol 167: 1464–1468.

    CAS  Article  Google Scholar 

  25. Reutzel D, Mende M, Naumann S, Schorkel S, Brenner W, Zabel B et al. (2001). Genomic imbalances in 61 renal cancer from the proximal tubulus detected by comparative genomic hybridization. Cytogenet Cell Genet 93: 221–227.

    CAS  Article  Google Scholar 

  26. Schlegel J, Bocker T, Zirngibl H, Hofstadter F, Ruschoff J . (1995). Detection of microsatellite instability in human colorectal carcinomas using a non-radioactive PCR-based screening technique. Virchows Arch 426: 223–227.

    CAS  Article  Google Scholar 

  27. Stoehr R, Wissmann C, Suzuki H, Knuechel R, Krieg RC, Klopocki E et al. (2004). Deletions of chromosome 8p and loss of sFRP1 expression are progression markers of papillary bladder cancer. Lab Invest 84: 465–478.

    CAS  Article  Google Scholar 

  28. Storkel S, Steart PV, Drenckhahn D, Thoenes W . (1989). The human chromophobe cell renal carcinoma: its probable relation to intercalated cells of the collecting duct. Virchows Arch B Cell Pathol Mol Pathol 56: 237–245.

    CAS  Article  Google Scholar 

  29. Suzuki H, Ueda T, Komiya A, Okano T, Isaka S, Shimazaki J et al. (1997). Mutational state of von Hippel-Lindau and adenomatous polyposis coli genes in renal tumors. Oncology 54: 252–257.

    CAS  Article  Google Scholar 

  30. Suzuki H, Geabreilson E, Chen W, Anbazhagan R, van Engeland M, Weijenberg MP et al. (2002). A genomic screen for genes upregulated by demethylation and histone deacetylase inhibition in human colorectal cancer. Nat Genet 31: 141–149.

    CAS  Article  Google Scholar 

  31. Suzuki H, Watkins DN, Jair KW, Schuebel KE, Markowitz SD, Chen WD et al. (2004). Epigenetic inactivation of SFRP genes allows constitutive WNT signaling in colorectal cancer. Nat Genet 36: 417–422.

    CAS  Article  Google Scholar 

  32. Takada T, Yagi Y, Maekita T, Imura M, Nakagawa S, Tsao SW et al. (2004). Methylation-associated silencing of the Wnt antagonist SFRP1 gene in human ovarian cancers. Cancer Sci 95: 741–744.

    CAS  Article  Google Scholar 

  33. Tetsu O, McCormick F . (1999). Beta-catenin regulates expression of cyclin D1 in colon carcinoma cells. Nature 398: 422–426.

    CAS  Article  Google Scholar 

  34. Togashi A, Katagiri T, Ashida S, Fujioka T, Maruyama O, Wakumoto Y et al. (2005). Hypoxia-inducible protein 2 (HIG2), a novel diagnostic marker for renal cell carcinoma and potential target for molecular therapy. Cancer Res 65: 4817–4826.

    CAS  Article  Google Scholar 

  35. Ueda M, Gemmill RM, West J, Winn R, Sugita M, Tanaka N et al. (2001). Mutations of the β- and γ-catenin genes are uncommon in human lung, breast, kidney, cervical and ovarian cancers. Br J Cancer 85: 64–68.

    CAS  Article  Google Scholar 

  36. Vainio SJ, Itaranta PV, Pcrasaari JP, Uusitalo MS . (1999). Wnts as kidney tubule inducing factors. Int J Dev Biol 43: 419–423.

    CAS  PubMed  Google Scholar 

  37. Veeck J, Niederacher D, An H, Klopocki E, Wiesmann F, Betz B et al. (2006). Aberrant methylation of the Wnt antagonist SFRP1 in breast cancer is associated with unfavourable prognosis. Oncogene 25: 3479–3488.

    CAS  Article  Google Scholar 

  38. Waki T, Tamura G, Sato M, Motoyama T . (2003). Age-related methylation of tumor suppressor and tumor-related genes: an analysis of autopsy samples. Oncogene 22: 4128–4133.

    CAS  Article  Google Scholar 

  39. Watanabe H, Okada G, Ohtsubo K, Yao F, Jiang PH, Mouri H et al. (2006). Aberrant methylation of secreted apoptosis-related protein 2 (SARP2) in pure pancreatic juice in diagnosis of pancreatic neoplasms. Pancreas 32: 382–389.

    Article  Google Scholar 

  40. Zafrakas M, Petschke B, Donner A, Fritzsche F, Kristiansen G, Knuchel R et al. (2006). Expression analysis of mammaglobin A (SCGB2A2) and lipophilin B (SCGB1D2) in more than 300 human tumors and matching normal tissues reveals their co-expression in gynecologic malignancies. BMC Cancer 6: 88–100.

    Article  Google Scholar 

  41. Zang T, Zhuang L, Zhang Z, Xin D, Guo Y . (2000). Aberrant activity of WNT/frizzled signaling pathway in renal cancer cell lines. Chin Sci Bull 45: 1703–1707.

    Article  Google Scholar 

  42. Zhang L, Zhou W, Velculescu B, Kern SE, Hruban RH, Hamilton SR et al. (1997). Gene expression profiles in normal and cancer cells. Science 276: 1268.

    CAS  Article  Google Scholar 

Download references

Acknowledgements

The expert technical assistance of Sonja von Serényi (Institute of Pathology, RWTH Aachen), Rudolf Jung (Institute of Pathology, University Regensburg), Kerstin Reher and Nina Niessl (Department of Urology, University Regensburg) is greatly appreciated. The study was supported by the German Ministry for Education and Research (BMBF grant 01KW0404 to Edgar Dahl) as part of the German Human Genome Project (DHGP), by a grant from the RWTH Aachen (START program) to Edgar Dahl and by the University of Regensburg to Christine Hammerschmied (Regensburger Forschungsfoerderung in der Medizin: ReForM A).

Author information

Affiliations

Authors

Corresponding author

Correspondence to E Dahl.

Additional information

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Dahl, E., Wiesmann, F., Woenckhaus, M. et al. Frequent loss of SFRP1 expression in multiple human solid tumours: association with aberrant promoter methylation in renal cell carcinoma. Oncogene 26, 5680–5691 (2007). https://doi.org/10.1038/sj.onc.1210345

Download citation

Keywords

  • loss of SFRP1 in human tumours
  • renal cell carcinoma
  • tumour suppressor gene
  • Wnt pathway
  • SFRP1 methylation
  • tissue microarray

Further reading

Search

Quick links