Hongzhi Zou

Next-Generation Stool DNA Test Accurately Detects Colorectal Cancer and Large Adenomas

BACKGROUND & AIMS Technical advances have led to stool DNA (sDNA) tests that might accurately detect neoplasms on both sides of the colorectum. We assessed colorectal neoplasm detection by a next-generation sDNA test and effects of covariates on test performance. METHODS We performed a blinded, multicenter, case-control study using archived stool samples collected in preservative buffer from 252 patients with colorectal cancer (CRC), 133 with adenomas ≥ 1 cm, and 293 individuals with normal colonoscopy results (controls); two-thirds were randomly assigned to a training set and one-third to a test set. The sDNA test detects 4 methylated genes, a mutant form of KRAS, and the α-actin gene (as a reference value) using quantitative, allele-specific, real-time target and signal amplification; it also quantifies hemoglobin. We used a logistical model to analyze data. RESULTS The sDNA test identified 85% of patients with CRC and 54% of patients with adenomas ≥1 cm with 90% specificity. The test had a high rate of detection for all nonmetastatic stages of CRC (aggregate 87% detection rate for CRC stages I-III). Detection rates increased with adenoma size: 54% ≥ 1 cm, 63% >1 cm, 77% >2 cm, 86% >3 cm, and 92% >4 cm (P < .0001). Based on receiver operating characteristic analysis, the rate of CRC detection was slightly greater for the training than the test set (P = .04), whereas the rate of adenoma detection was comparable between sets. Sensitivities for detection of CRC and adenoma did not differ with lesion site. CONCLUSIONS Early-stage CRC and large adenomas can be detected throughout the colorectum and with high levels of accuracy by the sDNA test. Neoplasm size, but not anatomical site, affected detection rates. Further studies are needed to validate the findings in a larger population and optimize the sDNA test.

Secreted Frizzled-Related Protein 1 Loss Contributes to Tumor Phenotype of Clear Cell Renal Cell Carcinoma

Purpose: Incidence and mortality rates for renal cell carcinoma (RCC) have been rising for decades. Unfortunately, the molecular events that support RCC carcinogenesis remain poorly understood. In an effort to gain a better understanding of signaling events in clear cell RCC (cRCC), we investigated the antitumor activity of secreted frizzled-related protein 1 (sFRP1), a negative regulator of Wnt signaling. Experimental Design: Genomic profiling of cRCC tumors and patient-matched normal tissues was done and confirmed using quantitative PCR and immunohistochemistry. Methylation-specific PCR was done on patient samples to evaluate the mechanism responsible for sFRP1 loss. sFRP1 expression was restored in cRCC cells and the effects on tumor phenotype were characterized. Results: Genomic profiling, quantitative PCR, and immunohistochemistry indicated that loss of sFRP1 occurred in cRCC and papillary RCC patient tissues. Twelve Wnt-regulated genes were up-regulated in cRCC tissues, including c-myc and cyclin D1, potentiators of cell proliferation and survival. Methylation of the sFRP1 gene was one mechanism identified for attenuation of sFRP1 mRNA. Stable reexpression of sFRP1 in cRCC cells resulted in decreased expression of Wnt target genes, decreased growth in cell culture, inhibition of anchorage-independent growth, and decreased tumor growth in athymic nude mice. Conclusions: To our knowledge, this is the first report to show that stable restoration of sFRP1 expression in cRCC cells attenuates the cRCC tumor phenotype. Our data support a role for sFRP1 as a tumor suppressor in cRCC and that perhaps loss of sFRP1 is an early, aberrant molecular event in renal cell carcinogenesis.

Sulfatase 2 up-regulates glypican 3, promotes fibroblast growth factor signaling, and decreases survival in hepatocellular carcinoma.

It has been shown that the heparin‐degrading endosulfatase, sulfatase 1 (SULF1), functions as a liver tumor suppressor, but the role of the related sulfatase, sulfatase 2 (SULF2), in liver carcinogenesis remains to be elucidated. We investigated the effect of SULF2 on liver tumorigenesis. Expression of SULF2 was increased in 79 (57%) of 139 hepatocellular carcinomas (HCCs) and 8 (73%) of 11 HCC cell lines. Forced expression of SULF2 increased HCC cell growth and migration, whereas knockdown of SULF2 using short hairpin RNA targeting SULF2 abrogated HCC cell proliferation and migration in vitro. Because SULF1 and SULF2 desulfate heparan sulfate proteoglycans (HSPGs) and the HSPG glypican 3 (GPC3) is up‐regulated in HCC, we investigated the effects of SULF2 on GPC3 expression and the association of SULF2 with GPC3. SULF2‐mediated cell growth was associated with increased binding of fibroblast growth factor 2 (FGF2), phosphorylation of extracellular signal‐regulated kinase and AKT, and expression of GPC3. Knockdown of GPC3 attenuated FGF2 binding in SULF2‐expressing HCC cells. The effects of SULF2 on up‐regulation of GPC3 and tumor growth were confirmed in nude mouse xenografts. Moreover, HCC patients with increased SULF2 expression in resected HCC tissues had a worse prognosis and a higher rate of recurrence after surgery. Conclusion: In contrast to the tumor suppressor effect of SULF1, SULF2 has an oncogenic effect in HCC mediated in part through up‐regulation of FGF signaling and GPC3 expression. (HEPATOLOGY 2008.)

Aberrant methylation of secreted frizzled-related protein genes in esophageal adenocarcinoma and Barrett's esophagus.

Hypermethylation of secreted frizzled‐related proteins (SFRP) genes frequently occurs with several cancers but has not been studied in esophageal adenocarcinoma or its precursor—Barretts esophagus. To explore the role of SFRP methylation in the neoplastic progression of Barretts esophagus and to evaluate methylated SFRP genes as biomarkers for Barretts esophagus and cancer, methylation of SFRP genes was determined in esophageal adenocarcinomas, Barretts esophagus and normal epithelia using methylation‐specific PCR. Protein expression of SFRP genes was then assessed in these tissues by immunohistochemistry. The mRNA expression of SFRP genes was quantified by real‐time reverse‐transcription PCR in esophageal adenocarcinoma cell lines with and without demethylation by 5‐aza‐2′deoxycytidine and inhibition of deacetylation by trichostatin A treatment. Hypermethylation of SFRP1, 2, 4 and 5 was detected in 93%, 83%, 73% and 85% of 40 cancers; 81%, 89%, 78% and 73% of 37 Barretts epithelia; 25%, 64%, 32% and 21% of 28 adjacent normal epithelia from Barretts patients; and 10%, 67%, 0% and 13% of 30 normal esophagogastric epithelia from healthy individuals, respectively (p < 0.001 for SFRP1, 4 and 5; p < 0.05 for SFRP2). Protein expression of SFRP1, 2 and 4 was downregulated in 87%, 67% and 90% of cancers, and expression correlated inversely with grade and stage of cancers and with grade of dysplasia. Expression of SFRP2 and SFRP4 proteins was lower in cancers with corresponding gene methylation (p < 0.05). Demethylation treatment effectively re‐expressed SFRP mRNA in cancer cell lines. Thus, hypermethylation of SFRP genes is a common early event in the evolution of esophageal adenocarcinoma, and methylation of SFRP1, 4 and 5 might serve as biomarkers for Barretts neoplasia. Aberrant promoter methylation appears to functionally silence SFRP gene expression in esophageal adenocarcinoma.

Highly Methylated Genes in Colorectal Neoplasia: Implications for Screening

Discriminant markers are required for accurate cancer screening. We evaluated genes frequently methylated in colorectal neoplasia to identify the most discriminant ones. Four genes specifically methylated in colorectal cancer [bone morphogenetic protein 3 (BMP3), EYA2, aristaless-like homeobox-4 (ALX4), and vimentin] were selected from 41 candidate genes and evaluated on 74 cancers, 62 adenomas, and 70 normal epithelia. Methylation status was analyzed qualitatively and quantitatively and confirmed by bisulfite genomic sequencing. Effect of methylation on gene expression was evaluated in five colon cancer cell lines. K-ras and BRAF mutations were detected by sequencing. Methylation of BMP3, EYA2, ALX4, or vimentin was detected respectively in 66%, 66%, 68%, and 72% of cancers; 74%, 48%, 89%, and 84% of adenomas; and 7%, 5%, 11%, and 11% of normal epithelia (P < 0.01, cancer or adenoma versus normal). Based on area under the curve analyses, discrimination was not significantly improved by combining markers. Comethylation was frequent (two genes or more in 72% of cancers and 84% of adenomas), associated with proximal neoplasm site (P < 0.001), and linked with both BRAF and K-ras mutations (P < 0.01). Cell line experiments supported silencing of expression by methylation in all four study genes. This study shows BMP3, EYA2, ALX4, and vimentin genes are methylated in most colorectal neoplasms but rarely in normal epithelia. Comethylation of these genes is common, and pursuit of complementary markers for methylation-negative neoplasms is a rational strategy to optimize screening sensitivity. (Cancer Epidemiol Biomarkers Prev 2007;16(12):2686–96)

The Stool DNA Test Is More Accurate Than the Plasma Septin 9 Test in Detecting Colorectal Neoplasia

BACKGROUND & AIMS Several noninvasive tests have been developed for colorectal cancer (CRC) screening. We compared the sensitivities of a multimarker test for stool DNA (sDNA) and a plasma test for methylated septin 9 (SEPT9) in identifying patients with large adenomas or CRC. METHODS We analyzed paired stool and plasma samples from 30 patients with CRC and 22 with large adenomas from Mayo Clinic archives. Stool (n = 46) and plasma (n = 49) samples from age- and sex-matched patients with normal colonoscopy results were used as controls. The sDNA test is an assay for methylated BMP3, NDRG4, vimentin, and TFPI2; mutant KRAS; the β-actin gene, and quantity of hemoglobin (by the porphyrin method). It was performed blindly at Exact Sciences (Madison, Wisconsin); the test for SEPT9 was performed at ARUP Laboratories (Salt Lake City, Utah). Results were considered positive based on the manufacturers specificity cutoff values of 90% and 89%, respectively. RESULTS The sDNA test detected adenomas (median, 2 cm; range, 1-5 cm) with 82% sensitivity (95% confidence interval [CI], 60%-95%); SEPT9 had 14% sensitivity (95% CI, 3%-35%; P = .0001). The sDNA test identified patients with CRC with 87% sensitivity (95% CI, 69%-96%); SEPT9 had 60% sensitivity (95% CI, 41%-77%; P = .046). The sDNA test identified patients with stage I-III CRC with 91% sensitivity (95% CI, 71%-99%); SEPT9 had 50% sensitivity (95% CI, 28%-72%; P = .013); for stage IV CRC, sensitivity values were 75% (95% CI, 35%-97%) and 88% (95% CI, 47%-100%), respectively (P = .56). False positive rates were 7% for the sDNA test and 27% for SEPT9. CONCLUSIONS Based on analyses of paired samples, the sDNA test detects nonmetastatic CRC and large adenomas with significantly greater levels of sensitivity than the SEPT9 test. These findings might be used to modify approaches for CRC prevention and early detection.

High Detection Rates of Colorectal Neoplasia by Stool DNA Testing With a Novel Digital Melt Curve Assay

BACKGROUND & AIMS Current stool DNA tests identify about half of individuals with colorectal cancers and miss most individuals with advanced adenomas. We developed a digital melt curve (DMC) assay to quantify low-abundance mutations in stool samples for detection of colorectal neoplasms and compared this test with other approaches. METHODS We combined a melt curve assay with digital polymerase chain reaction and validated the quantitative range. We then evaluated its ability to detect neoplasms in 2 clinical studies. In study I, stool samples from patients with colorectal tumors with known mutations (KRAS, APC, BRAF, TP53) were assayed. In study II, archived stool samples from patients with advanced adenomas containing known KRAS mutations were assayed, along with controls. Results were compared with those from the stool DNA test PreGenPlus (Exact Sciences, Marlborough, MA), Hemoccult, and HemoccultSensa (both Beckman-Coulter, Fullerton, CA). RESULTS The DMC assay detected samples in which only 0.1% of target genes were mutated. In study I, the DMC assay detected known mutations in 28 (90%) of 31 tumor samples and 6 (75%) of 8 advanced adenoma samples. In study II, the DMC assay detected 16 (59%) of 27 advanced adenoma samples that contained KRAS mutations, compared with 7% with the Hemoccult, 15% with the HemoccultSensa, and 26% with the PreGenPlus assays (P < .05 for each, compared with the DMC assay); specificities did not differ significantly. CONCLUSIONS The DMC assay has a high level of sensitivity in detecting individuals with colon neoplasms and is better than current stool screening methods in detecting those with advanced adenomas. Further studies are indicated.

A Sensitive Method to Quantify Human Long DNA in Stool: Relevance to Colorectal Cancer Screening

Human long DNA in stool may reflect nonapoptotic exfoliation and has been used as a colorectal cancer (CRC) marker. Targeting human-specific Alu repeats represents a logical but untested approach. A real-time Alu PCR assay was developed for quantifying long human DNA in stool and evaluated in this study. The accuracy and reproducibility of this assay and the stability of long DNA during room temperature fecal storage were assessed using selected patient stools and stools added to human DNA. Thereafter, long DNA levels were determined in blinded fashion from 18 CRC patients and 20 colonoscopically normal controls. Reproducibility of real-time Alu PCR for quantifying fecal long DNA was high (r2 = 0.99; P < 0.01). Long DNA levels in nonbuffered stools stored at room temperature fell a median of 75% by 1 day and 81% by 3 days. An EDTA buffer preserved DNA integrity during such storage. Human long DNA was quantifiable in all stools but was significantly higher in stools from CRC patients than from normal controls (P < 0.05). At a specificity of 100%, the sensitivity of long DNA for CRC was 44%. Results indicate that real-time Alu PCR is a simple method to sensitively quantify long human DNA in stool. This study shows that not all CRCs are associated with increased fecal levels of long DNA. Long DNA degrades with fecal storage, and measures to stabilize this analyte must be considered for optimal use of this marker. (Cancer Epidemiol Biomarkers Prev 2006;15(6):1115–9)

Frequent Methylation of Eyes Absent 4 Gene in Barrett's Esophagus and Esophageal Adenocarcinoma

Most esophageal adenocarcinomas arise within Barretts esophagus but the cause of this increasingly prevalent condition remains unknown. Early detection improves survival and discriminant screening markers for Barretts esophagus and cancer are needed. This study was designed to explore the natural history of eyes absent 4 (EYA4) gene methylation in the neoplastic progression of Barretts esophagus and to evaluate methylated EYA4 as a candidate marker. Aberrant promoter methylation of EYA4 was studied by methylation-specific PCR using bisulfite-treated DNA from esophageal adenocarcinomas, Barretts esophagus, and normal epithelia, and then confirmed by sequencing. Eight cancer cell lines were treated with the demethylation agent 5-aza-2′-deoxycytidine, and EYA4 mRNA expression with and without treatment was quantified by real-time reverse-transcription PCR. EYA4 hypermethylation was detected in 83% (33 of 40) of esophageal adenocarcinomas and 77% (27 of 35) of Barretts tissues, but only in 3% (2 of 58) of normal esophageal and gastric mucosa samples (P < 0.001). The unmethylated cancer cell lines had much higher EYA4 mRNA expression than the methylated cancer cell lines. Demethylation caused by 5-aza-2′-deoxycytidine increased the mRNA expression level by a median of 3.2-fold in methylated cells, but its effect on unmethylated cells was negligible. Results indicate that aberrant promoter methylation of EYA4 is very common during tumorigenesis in Barretts esophagus, occurs in early metaplasia, seems to be an important mechanism of down-regulating EYA4 expression, and represents an intriguing candidate marker for Barretts metaplasia and esophageal cancer.

Stool DNA testing for the detection of colorectal neoplasia in patients with inflammatory bowel disease

Current approaches to the detection of colorectal neoplasia associated with inflammatory bowel disease (IBD‐CRN) are suboptimal.