Patent Application
20130041047
The present invention relates to methods of detecting bladder cancer in an individual using urinary methylation markers.
Cancer of the urinary bladder is the fifth most common neoplasm among the populations of industrialized countries.
Epigenetics is the study of mitotically and/or meiotically heritable changes in gene function that cannot be explained by changes in DNA sequence. Several forms of epigenetic regulation are known, including histone modifications and DNA methylation. DNA methylation occurs during critical normal processes like development, genomic imprinting, and X-chromosome inactivation. Alterations in epigenetic control have been associated with several human pathologic conditions including cancer. See, Egger G, et al., Epigenetics in human disease and prospects for epigenetic therapy, Nature 2004; 429:457-63. CpG sites are sparsely distributed throughout the genome except for in CpG islands. See, Takai D, et al., Comprehensive analysis of CpG islands in human chromosomes 21 and 22, Proc. Natl. Acad. Sci. 2002; 99:3740-45; Gardiner-Garden M, et al., CpG islands in vertebrate genomes, J. Mol. Biol. 1987; 196:261-82. CpG dinucleotides outside CpG islands are generally hypermethylated in normal cells and undergo a substantial loss of DNA methylation in cancers.
CpG sites within CpG islands are usually in an unmethylated state permissive to transcription in normal cells, but become hypermethylated at certain promoters in cancers. Transcriptional inactivation by CpG island promoter hypermethylation is a well-established mechanism for gene silencing in cancer, including bladder cancer, and aberrant methylation is associated with stage, and grade of the tumors as well as recurrence rate and progression.
In 75% of all cases of urinary bladder cancer, the primary tumor will present as a non-muscle invasive tumor stage Ta or T1 (NMIBC). The remaining 25% of the cases will present with invasion of the bladder muscle, stage T2-4 (MIBC). Stage Ta bladder cancer is characterized by frequent recurrences after resection, in as many as 60% of patients. See Millan-Rodriguez F, et al., Primary superficial bladder cancer risk groups according to progression, mortality and recurrence. J Urol 2000; 164:680-4. Often one or more tumors will appear each year over an 8-10 years period without any progression, however, up to 25% will eventually develop an aggressive invasive phenotype. See Wolf H, Kakizoe T, et al. Bladder tumors, Prog Clin Biol Res 1986; 221:223-55.
Patients diagnosed with superficial bladder cancer are generally monitored over an extended time period with a cystoscope, which is extended into the bladder through the urethra. Such monitoring causes patient discomfort and is costly. Markers for bladder cancer which can be detected in patient urine would decrease the cost of monitoring and lessen the discomfort of patients. Markers which indicate the likelihood of cancer progression would have additional value in determining courses of treatment.
Hypermethylation of one or more of the urinary markers HOXA9, ZNF154, POU4F2, and EOMES indicates existence of urinary bladder cancer. Hypermethylation of TBX4 was not found to be a urinary marker for detecting the presence of bladder cancer, but was newly discovered to be associated with a likelihood of bladder cancer progression from stage Ta to stage T1 or T2, or another more advanced stage and is disclosed herein as a marker for bladder cancer progression. Disclosed is assaying a subject's urine sample whether one or more of the markers HOXA9, ZNF154, POU4F2, and EOMES is hypermethylated, for example within the promoter region, relative to the level of methylation in said markers in a control, or, relative to the methylation level of other or all genomic material, such as genomic DNA material or genomic DNA in the assay. Following the determination, the information can be used to initiate a monitoring program for subjects with hypermethylation of the relevant markers such as increasing frequency of cystoscopies if hypermethylation is observed or reducing monitoring for patients with no observed hypermethylation of the relevant markers. The determination can also be used to alter the course of treatment—for example, treating more aggressively (e.g. with cystectomy, chemotherapy or immunotherapy) because of the increased progression risk. The determination can be done using any assay technique, including those described herein, i.e., Infinium Array, bisulfate sequencing or Methylation-Sensitive High Resolution Melting. Monitoring programs and treatment methods are known to those of skill in the art.
According to certain aspects of the present disclosure, a method is provided for identifying bladder cancer in a subject or predicting a likelihood of a subject developing bladder cancer. According to one aspect, the method includes collecting urine from a subject, assaying genomic material in the urine for one or more of the markers HOXA9, ZNF154, POU4F2, and EOMES being hypermethylated relative to the level of methylation in said markers in a control representative of a subject. i.e. a control sample from a subject, who is negative for bladder cancer or relative to the level of methylation of the total genomic material, such as genomic DNA material or genomic DNA, in the assay, and wherein hypermethylation indicates bladder cancer in the subject.
According to a certain aspect, a determination is made by hybridizing the genomic material to an array of probes where the array is capable of determining the average percentage of methylation of the markers. According to an additional aspect, bisulfite sequencing is used in the determination of the average percentage of methylation of the markers. According to one aspect, following bisulfite sequencing, a high resolution melting analysis is performed.
According to certain aspects, methods described herein include determining whether any markers other than HOXA9, ZNF154, POU4F2, and EOMES are hypermethylated or hypomethylated in a tissue sample from the subject. According to one aspect, the tissue sample is obtained by performing a cystoscopy on the patient. According to one aspect, the tissue sample is obtained by performing a transurethral resection of a bladder tumor on the patient. According to an additional aspect, the markers are one or more of PTGDR; ZNF135; TBX4; ACOT11; PCDHGA12; CA3; CHRNB1; BRF1; SOCS3; PTGDR; or SCARF2. According to a still additional aspect, the one or more of the markers PTGDR; ZNF135; TBX4; ACOT11; PCDHGA12; or CA3 is hypermethylated relative to the level of methylation in the markers in the control; and/or one or more of the markers CHRNB1; BRF1; SOCS3; PTGDR; or SCARF2 is hypomethylated relative to the level of methylation in said markers in the control.
According to an additional aspect, hypermethylation of the markers is observed, and a monitoring program for the subject for bladder cancer development or progression or recurrence is undertaken. According to one aspect, hypermethylation of the markers is observed, and initiation of treatment, or a change in existing treatment regimens, is undertaken. According to one aspect, the array described herein with respect to the methods described herein is analyzed by establishing a threshold which reflects a significant level of methylation. According to one aspect, the assaying described herein includes amplification of portions of the markers HOXA9, ZNF154, POU4F2, and EOMES. According to a certain aspect, the amplification step includes use of primers targeting the methylated or unmethylated portions of the markers.
According to one aspect of the present disclosure, a method is provided for identifying, detecting, confirming, determining, diagnosing, or prognosing bladder cancer in a subject including assaying genomic material in urine from the subject for one or more of the markers HOXA9, ZNF154, POU4F2, or EOMES being hypermethylated relative to the level of methylation in respective HOXA9, ZNF154, POU4F2, or EOMES non-bladder cancer control markers or relative to the level of methylation of total genomic material in the assay; and wherein hypermethylation of one or more of the markers HOXA9, ZNF154, POU4F2, or EOMES indicates bladder cancer in the subject.
According to one aspect, the step of assaying includes hybridizing the genomic material to an array of probes where the array indicates the average percentage of methylation of the markers. According to an additional aspect, the step of assaying includes bisulfite sequencing to determine the average percentage of methylation of the markers. According to an additional aspect, a high resolution melting analysis is performed following bisulfite sequencing.
According to one aspect, methods described herein including determining whether markers other than HOXA9, ZNF154, POU4F2, and EOMES are hypermethylated or hypomethylated in a tissue sample from the subject. According to one aspect, the tissue sample is obtained by performing a cystoscopy on the patient. According to one aspect, the tissue sample is obtained by performing a transurethral resection of a bladder tumor on the patient. According to one aspect, the markers from the tissue sample are one or more of PTGDR; ZNF135; TBX4; ACOT11; PCDHGA12; CA3; CHRNB1; BRF1; SOCS3; PTGDR; or SCARF2.
According to a certain aspect, the one or more of the markers PTGDR; ZNF135; TBX4; ACOT11; PCDHGA12; or CA3 is hypermethylated relative to the level of methylation in respective control markers; or one or more of the markers CHRNB1; BRF1; SOCS3; PTGDR; or SCARF2 is hypomethylated relative to the level of methylation in respective control markers.
According to one aspect, hypermethylation of the markers is observed and the subject is monitored for bladder cancer development, recurrence or progression. According to one aspect, hypermethylation of the markers is observed and the subject is treated for bladder cancer. According to one aspect, the array described herein with respect to the methods described herein is analyzed by establishing a threshold which reflects a significant level of methylation. According to one aspect, the assaying described herein includes amplification of portions of the markers HOXA9, ZNF154, POU4F2, or EOMES. According to a certain aspect, the amplification step includes use of primers targeting the methylated or unmethylated portions of the markers.
According to certain aspects, methods are provided where hypermethylation of two or more of the markers HOXA9, ZNF154, POU4F2, or EOMES indicates bladder cancer in the subject. According to certain aspects, methods are provided where hypermethylation of three or more of the markers HOXA9, ZNF154, POU4F2, or EOMES indicates bladder cancer in the subject. According to certain aspects, methods are provided where hypermethylation of the markers HOXA9, ZNF154, POU4F2, or EOMES indicates bladder cancer in the subject.
According to certain aspects, the step of assaying in the methods described herein includes assaying for markers TWIST1 or VIM being hypermethylated relative to the level of methylation in respective TWIST1 or VIM non-bladder cancer control markers or relative to the level of methylation of total genomic material in the assay, and wherein hypermethylation of one or more of the markers TWIST1 or VIM indicates bladder cancer in the subject.
According to certain aspects, a method is provided for identifying bladder cancer in a subject including assaying genomic material in urine from the subject for the marker HOXA9 being hypermethylated relative to the level of methylation of HOXA9 in a non-bladder cancer control sample or relative to the level of methylation of total genomic material in the assay, and wherein hypermethylation of HOXA9 indicates bladder cancer in the subject. According to certain aspects, a method is provided for identifying bladder cancer in a subject including assaying genomic material in urine from the subject for the marker ZNF154 being hypermethylated relative to the level of methylation of ZNF154 in a non-bladder cancer control sample or relative to the level of methylation of total genomic material in the assay, and wherein hypermethylation of ZNF154 indicates bladder cancer in the subject. According to a certain aspect, a method is provided for identifying bladder cancer in a subject including assaying genomic material in urine from the subject for the marker POU4F2 being hypermethylated relative to the level of methylation of POU4F2 in a non-bladder cancer control sample or relative to the level of methylation of total genomic material in the assay, and wherein hypermethylation of POU4F2 indicates bladder cancer in the subject. According to a certain aspect, a method is provided for identifying bladder cancer in a subject including assaying genomic material in urine from the subject for the marker EOMES being hypermethylated relative to the level of methylation of EOMES in a non-bladder cancer control sample or relative to the level of methylation of total genomic material in the assay, and wherein hypermethylation of EOMES indicates bladder cancer in the subject.
The novel urinary markers of methylation were found by analysis of a large number of samples by using microarray analysis for an initial identification followed by confirmation using Methylation-Sensitive High Resolution Melting (MS-HRM). Bisulfite sequencing was also performed on the urinary markers as an additional assay for marker methylation. Statistical correlations were determined as described below.
Patient material: A total of 119 tissue samples from patients with bladder cancer but without other malignant disease, were analyzed by Infinium Array or Methylation-Sensitive High Resolution Melting (MS-HRM). Most patients provided metachronous tumors. The samples were obtained fresh from transurethral resection of bladder tumors from patients, embedded in Tissue-Tek (O.C.T) Compound (Sakura Finetek), and immediately snap frozen in liquid nitrogen. Normal bladder urothelium (for controls) was obtained from individuals who had benign prostate hyperplasia or bladder stones.
Samples were macro dissected (for tumor samples) or laser dissected (normal samples) to obtain a urothelial cell percentage of at least 75%. Sample composition was confirmed by H&E evaluation of sections cut before and after those used for extraction. Voided urine was collected from 115 bladder cancer patients (for evaluating urinary markers) and 59 individuals with benign prostate hyperplasia or bladder stones (for controls). Nineteen of the controls were stix positive for nitrite, indicating bacterial infection. Urine specimens were collected immediately before urinary cytology or cystoscopy, pelleted by centrifugation, and frozen at −80° C.
DNA Extraction and Bisulfite Modification:
Tissue DNA was extracted using the Puregene DNA purification kit (Gentra Systems, Minneapolis Minn.). One microgram (μg) of DNA extracted from fresh frozen tissue was bisulfite modified using the EZ-96 DNA methylation D5004 (Zymo Research, Irvine Calif.) for the Infinium array, or EpiTect (Qiagen, Inc., MA) for the MS-HRM, respectively. Urinary DNA was extracted using the Puregene DNA purification kit (Gentra Systems) according to the manufacturer's recommendations. Tissue and urine DNA purity was assessed using the OD 260/280 ratio.
Infinium Array:
One μg of DNA from each sample was whole genome amplified and hybridized overnight to Infinium Arrays, scanned by a BeadXpress Reader instrument (Illumina, Inc.) and data were analyzed by the Bead Studio Methylation Module Software (Illumina) and exported to Excel for further analysis. The CpG island status was obtained from Bead Studio. For each of the 27,578 probes the Infinium assay returns a beta value (β), which approximately corresponds to the average percentage of methylation in the sample analyzed. Illumina reports that the Infinium array is accurate with Δβ-values above 0.2. The Δβ cutoff value for differential methylation was conservatively set to ±0.25.
Cloning and Bisulfite Sequencing:
Primers for bisulfite sequencing of CpG island regions were designed using Methprimer and primer sequences are shown in Supplementary Table 1. PCR for cloning was carried out with the Accuprime™ Taq DNA Polymerase System (Invitrogen) according to the manufacturer's instructions, in a final volume of 25 microliters (μl) using 4 μl of bisulfite modified DNA as template. Amplification cycling temperatures can be seen in Supplementary Table 1, for each primer pair. PCR amplicons were gel purified using the QIAQUICK Gel Extraction Kit (Qiagen) and TOPO TA cloned for sequencing (Invitrogen) according to the manufacturer's instructions. Twelve random colonies from each gene were used for colony PCR in a final volume of 25 μl using the TEMPase Kit (Ampliqon) according to the manufacturer's instructions. Primers were M13 forward and M13 reverse from the TOPO TA Cloning Kit (Invitrogen). The sequencing reactions were analyzed in a 3130× Genetic Analyzer (Applied Biosystems).
Methylation Sensitive High Resolution Melting (MS-HRM):
Methylation-Sensitive High Resolution Melting (MS-HRM) was carried out in triplicate with 15 sets of primers (Supplementary Table 1) using 1.5 μl (15 nanograms (ng)) of bisulfite modified DNA as template in a final volume of 10 μl using the LightCycler™ 480 High Resolution Melting Master (Roche). Each plate included a no template control (NTC) and a standard curve (100%, 75%, 50%, 25%, 5%, and 0% methylated samples, CpGenome™ Universal Methylated DNA (from Millipore) diluted with unmethylated peripheral blood DNA. Melting curves were analyzed on a LightScanner (Idaho Technology Inc.).
RNA Purification and Gene Expression Microarray:
RNA was purified using the RNeasy Kit (Qiagen). The RNA integrity and RNA Integrity Number (RIN) was assessed with the 2100 Bioanalyzer (Agilent). 500 ng of RNA from each sample were loaded on a Human Exon 1.0 ST Arrays (Affymetrix). Microarray analysis and data handling was performed as is conventional, and described for example, in Dyrskjot L, et al., Identifying distinct classes of bladder carcinoma using microarrays, Nat Genet 2003; 33:90-96).
Data Analysis:
Genespring GX 10 software (Agilent) was used for Exon array analysis. Data was quantile normalized using ExonRMA16 software with transcript level core (17881 transcripts) and by using antigenomic background probes. The statistical analysis was performed with independent samples only, except for the two analyses of metachronous tumors. The independent tumor analysis included analysis of methylation in the Ta(stable) tumor group compared with the Ta(stable2) tumor group and the analysis of the methylation level in the Ta(prog) tumor group compared with the methylation level in the subsequent progressed tumor (T1 or T2-4). Ta(stable) and Ta(stable2) respectively consist of the first and second tumor from patients with a stable Ta disease. The second tumor is a recurrent tumor. Ta(prog) consists of Ta tumors from patients with subsequent progression to T1 or T2-4. When patients had several Ta tumors before the disease progressed to stage T1 or higher, the Ta tumor closest to the stage progression, i.e., the Ta tumor with the shortest timespan to the progressed tumor, was used in the analysis.
Gene Ontology (GO) and Ingenuity Pathway Analysis (IPA):
ene symbols of genes showing hypo or hyper methylation were used as input in GO analysis. The undivided list was submitted to IPA (2000-2008 Ingenuity Systems) and the data were analyzed to identify (adjusted for multiple testing by the Benjamini-Hochberg method) top network associated functions and Canonical pathways.
Statistical Analysis:
Stata 10 (Statacorp, Tex., USA) was used for analyzing methylation data from MS-HRM using the nonparametric Wilcoxon-Mann-Whitney test. The inter observer agreement coefficient (κ) was calculated for MS-HRM. The Infinium Array data was analyzed using nonparametric Wilcoxon-Mann-Whitney or Wilcoxon signed-rank test in R (see r-project.org) to evaluate differential methylation between independent groups (based on stage of progression) or related samples, respectively. As synchronous lesions were very similar in methylation only one from each patient was included for statistical calculation. There was no adjustment for multiple testing due to limited group sizes. The most interesting CpG sites were instead validated on an independent sample set. The Chi-squared test was used for evaluation of chromosomal distribution. Excel (Microsoft) was used for a two tailed student's t-test to evaluate different mRNA expression between groups and Pearson correlations to progression.
Genome Wide Methylation in Urinary Bladder Cancer:
The genome wide DNA methylation status of six normal urothelium samples and 50 urothelial carcinomas (UC) of the bladder, were first profiled using microarrays interrogating 27,000 known CpG sites. To study the methylation over time in single individuals, we analyzed metachronous tumors (two-three tumors from 18 patients). We subdivided patients with stage Ta bladder cancer into stable disease (named Ta(stable), or Ta(stable2)) when taken from the same patient) if no progression to higher stages was observed and progressing disease (named Ta(prog)), indicating the bladder cancer progressed from stage Ta to stage T1 or higher. The average CpG site methylation within CpG islands was increased (p=0.013; student's t-test) in the aggressive Ta(prog), T1 and T2-4 tumors, compared to normals and Ta(stable) tumors. Sites outside CpG islands measured a decrease (p=0.0095; student's t-test) in average CpG site methylation reaching 18.5% in the Ta(stable) group and 10.6% in the T2-4 tumor group compared to normal tissue. Using the Ta(stable) tumors as a reference group it was evident that the majority of changes in methylation occurred in the transition from normal to cancer. These findings are in concordance with other findings in cancer tissues compared to normal tissues.
Gene Specific Methylation Differences:
Table 1 below is a list of the 19 most highly differentiated methylated genes between controls and tumors, as well as genes selected as indicated in the flow chart depicted in
<0.0001
ZNF154*
0.0018
−0.68
0.0001
0.0009
HOXA9*
0.0003
−0.46
0.0009
ZNF154
0.0049
−0.75
0.0004
14
POU4F2*
0.0004
HIST1H4F
#
0.0005
ACOT11*
0.0004
EOMES*
0.0004
PCDHGA12*
0.0001
CA3*
0.0001
PTGDR*
0.0218
GRM4
#
<0.0001
SLC22A12
#
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
=0.0001
<0.0001
<0.0001
#Validated by bisulfite sequencing.
$Not determined.
Nine other genes showed a high sensitivity and specificity when comparing normal and cancer (see flow chart for gene selection at
Validation of Microarray Data:
In order to confirm the microarray findings, the MS-HRM technique was also used on an independent sample set consisting of 8 normals and 55 cancers as indicated in Table 2 above showing the demographic and clinical characteristics of the bladder cancer patients and control individuals.
Technical Validation of the MS-HRM Technique:
To test PCR based MS-HRM, a technical validation was performed prior to independent validation. MS-HRM primers for eight bladder cancer marker genes (selected as per
All eight tumor markers ZNF154, HOXA9, POU4F2, EOMES, ACOT11, PCDHGA12, CA3 and PTGDR) were validated by the independent validation set (p<0.011) (see
The inter observer agreement (Kappa-value) of the MS-HRM validation assay indicated its effectiveness (0.58 to 1.00, see Supplementary Tables 2 and 3). None of the markers identified were independent of each other (see Supplementary Table 4). This indicates that one single methylation mechanism may account for the majority of the methylation alterations discovered.
Bisulfite Sequencing of DNA Surrounding Infinium Probes:
Eleven tumor marker genes and one stage marker were selected for analytical validation by bisulfite sequencing to obtain detailed information on the sequence surrounding the Infinium Array probe source sequence, and the sequence analyzed by MS-HRM. Bisulfite sequencing corresponded well with the array and MS-HRM based findings (See
Association Between Methylation Status and Clinicopathological Variables in the Validation Set:
The possible association of methylation status with two clinicopathological parameters, stage and grade, were investigated. Table 3 below indicates association between methylation markers and stage, grade and age in the validation set. Methylation values were dichotomized as positive or negative based on Receiver Operating Characteristic (ROC) analysis. The frequency of methylation is shown, as well as the number of methylation positive tumors and the total number of tumors.
Only methylation of ACOT11 was associated with stage (Fisher's exact test, p=0.049). ACOT11 was more frequently methylated in the T1 and T2-4 stage tumors than in the superficial Ta tumors. Another marker CA3 was less frequently methylated in grade I tumors compared to grade II and III tumors (Fisher's exact test, p=0.011). When the tumor patients were divided into two groups by mean age (72) of the patients, no significant association with age was found. However, higher stage was associated with increasing age (Fisher's exact test, p=0.041).
Identification of Methylated Biomarkers in Urinary Specimens from Bladder Cancer Patients:
To test the potential of the validated tumor specific methylation of the genes ZNF154, POU4F2, HOXA9, and EOMES as urinary markers for early detection of bladder cancer, urine from 115 patients with cancer and 59 control urine samples was analyzed using MS-HRM. The results are set forth in Table 4 below.
#Mann-Whitney U test
$Not applicable
The methylation difference between urine from healthy individuals and patients was highly significant for ZNF154 (p<0.0001), POU4F2 (p<0.0001), HOXA9 (p<0.0044), and EOMES (p<0.0001). The sensitivity observed for the individual markers was 62%-74%, while the specificity was 100% for ZNF154, POU4F2, and EOMES, and 96% for HOXA9 using cut-off values decided by receiver operating characteristic (ROC) analysis. Combining all four markers increased the sensitivity to 82% and the specificity to 97%; with positive predictive value (PPV) of 98% and negative predictive value (NPV) of 73%.
Given that cytology has less sensitivity in low stage lesions, the combined markers were analyzed on urine from 59 patients with Ta tumors. The sensitivity was 80% and specificity 97%, the AUC (95% CI) 0.88 (0.82-0.94), the PPV 96% and the NPV 83% (see Supplementary Table 5). The sensitivity in urine from patients with T1 and T2-4 tumors was slightly higher than for Ta tumors. Cytology also has less sensitivity in low grade tumors. The performance of the combined markers on urine from patients with grade one tumors was: sensitivity 71%, specificity 97%, AUC (95% CI) 0.84 (0.72-0.95), PPV 86%, and NPV 92% (see Supplementary Table 5). The sensitivity on urine specimens with tumor cells detected by the pathologist was 95%, while it was 87% in urine in which the pathologist did not detect cells. Methylation markers were able to detect cancer in 13 of 15 patients where a pathologist did not detect tumor cells in urine samples. Based on this, the urinary methylation assay is much more sensitive than urine cytology for the detection of bladder tumors. Methylation data from urine specimens and tumor samples were matched for 33 patients. The analytical sensitivity of the methylation data on these patient samples ranged from 81%-97%, with a combination of the four methylation markers ZNF154, HOXA9, POU4F2, and EOMES achieving 94% analytical sensitivity (see Supplementary Table 6).
Association Between Methylation Status and Clinicopathological Variables on Urine Specimens:
The association of the four urinary markers of bladder cancer ZNF154, HOXA9, POU4F2, and EOMES with stage, grade, age, cytology, and nitrite status was analyzed (see Supplementary Table 7). Methylation of ZNF154 was associated with higher stage (Fisher's exact test, p=0.019) and grade (Fisher's exact test, p=0.002), whereas methylation of EOMES was associated with high grade (Fisher's exact test, p=0.036). The frequency of methylation of HOXA9 and EOMES was independent of cytology being positive or negative for tumor cells (Fisher's exact test, p>0.05). No association was observed between the frequency of methylation and age for any of the markers (Fisher's exact test, p>0.05). Nitrite positivity did not influence the methylation assay in tumor urine samples nor in normal control urine samples.
Correlation Between DNA Methylation and Transcription:
Considering the genes in Table 1, only HOXA9 and ZNF154 had an absolute Pearson correlation between methylation and expression equal to or larger than 0.4, and only HOXA9 was differentially expressed between normal and tumor samples (p=0.0022, student's t-test). As expected, the level of HOXA9 transcript was lower in tumor compared to normal samples. The bisulfite sequencing did not provide additional information—as the array probes seemed to reflect the methylation event well in the sequenced areas (see
Intrapatient Variation in Methylation:
The intrapatient stability of methylation was high for both Ta(stable) and Ta(prog) tumors, as 92% and 89% of changes, respectively, found in early tumors were present later on. The number of changes was independent of time between tumors (R2=0.0029) and mRNA transcript level of DNA-methyltransferases. However, to study if this was based on a systematic change in methylation of certain genes over time; a group comparison was made across the metachronous samples. This analysis revealed that no single genes were differentially methylated between the first and second tumor within the stable or progressing groups (p>0.05; Wilcoxon signed-rank test).
Pathway Analysis of Differentially Methylated Genes:
Using Gene Ontology (GO), the 149 differentially methylated genes between Ta stable and Ta progressing tumors belonged mainly to 22 overrepresented pathways, having up to 7 methylation changes. Hypermethylated pathways were related to cellular development, in particular, epidermal development (p<0.037). Hypomethylated pathways were related to cell-cell signaling, in particular negative regulators of cell death (p<0.038). Using Ingenuity pathway analysis (IPA), the main network associated functions altered by methylation were cell movement of eukaryotic cells (p=1.65E-010), tumorigenesis (p=3.37E-08), and growth of cancer cells (p=4.46E-07) (see Supplementary Table 8) as well as apoptosis (p<1.24E-06) and proliferation of cells (p<3.91E-06). The top canonical pathway was G-protein coupled receptor signaling (p=9.96E-06 to p=1.56E-02, see Supplementary Table 8).
Pathway analysis on superficial papillomas of low histological grade versus high grade superficial and invasive tumors, showed that many of the top networks identified between Ta stable and Ta progressing tumors were also present in this analysis (see Supplementary Table 8). These results suggest that methylation may hit selected networks and pathways at multiple levels, thereby impacting the malignant process.
Epigenetic Regulation of Keratin (KRT), Keratin Associated Proteins (KRTAP), and Small Proline Rich Proteins (SPRR):
Chromosome 21 was found to encompass more differentially methylated genes outside CpG islands, than any other chromosome after correction for number of CpG sites (p<0.0001) (see
Patient Material:
A total of 652 voided urine samples were collected at the Department of Urology at Aarhus University Hospital from 390 bladder cancer patients, and 47 individuals with benign prostatic hyperplasia or bladder stones, but no history of bladder cancer (control individuals). From these, 227 samples were excluded, as the DNA amount was below a set threshold. See Table 5 below.
aBergkvist
bNot available
The remaining 425 samples (390 samples from 184 BC patients and 35 from control individuals) are indicated in Table 6 below and
aN/A Not available.
bBergkvist.
cOf the 184 patients, 26 were lost for follow-up.
Ten to fifty milliliters (mL) of urine was collected at regular follow-up visits. Urine specimens were collected immediately before cystoscopy; cells were sedimented by centrifugation, and frozen at −80° C. The tumors were staged according to the TNM system described in Sobin, TNM Classification of malignant Tumours, International Union Against Cancer, 2002, 6th Edition (New York, N.Y.: John Wiley & Sons and graded according to Bergkvist et al., Classification of bladder tumours based on the cellular pattern. Preliminary report of a clinical-pathological study of 300 cases with a minimum follow-up of eight years, Acta Chir Scand, 1965, 130(4): p. 371-78. Fifteen of the control individuals were stix positive for nitrite in the urine indicating bacterial infection. Informed written consent was obtained from all patients, and research protocols were approved by The Central Denmark Region Committees on Biomedical Research Ethics. Patient treatment and follow-up were performed in accordance with the guidelines of European Association of Urology as set forth in Babjuk et al., EAU Guidelines on non-muscle-invasive urothelial carcinoma of the bladder, Eur Urol, 2008, 54(2):p. 303-14.
DNA Extraction and Bisulfite Modification:
DNA was extracted with the QIAsymphony Virus/Bacteria Midi kit (96) (Qiagen) using the QIAsymphony® SP instrument and employing the Complex800_V5_DSP protocol. Five hundred nanograms (500 ng) of DNA was bisulfite modified using the EZ-96 DNA methylation D5004 kit (Zymo Research) according to the manufacturers recommendations and eluted in 60 microliters (μl) of elution buffer and stored at −20° C. until use.
Real-Time Quantitative Methylation-Specific Polymerase Chain Reaction (MethyLight):
Methylation analysis was performed using MethyLight as described in Campan et al., MethyLight, Methods Mol Biol, 2009, 507: p. 325-37. Primers and probes for the six genes of interest were designed to include eight to ten CpG dinucleotides as indicated in Table 7 below. All probes contain a 6-FAM fluorophore at the 5′ end and a black hole quencher-1 (BHQ-1) at the 3′ end.
, (95° C. 15
,
) × 45
, (95° C. 15
,
) × 45
, (95° C. 15
,
) × 45
, (95° C. 15
,
) × 45
, (95° C. 15
,
) × 45
, (95° C. 15
,
) × 45
, (95° C. 15
,
) × 45
indicates data missing or illegible when filed
For normalization of DNA input material, the ALU-C4 repeat element sequence was used as described in Weisenberger et al., Analysis of repetitive element DNA methylation by MethyLight, Nucleic Acids Res, 2005, 33(21): p. 6823-36. Quantitative PCR amplifications were carried out with the TaqMan Universal PCR Master Mix No AmpErase (Applied Biosystems) according to the manufacturer's instructions in duplicates using 2 μl (5 ng) of bisulfite modified DNA in a final volume of 5 μl in 384-well plates on a ABI 7900 HT Fast Real Time PCR System (Applied Biosystems). When inconsistency between duplicates occurred, the analysis was repeated. Amplification protocols for real-time quantitative methylation-specific polymerase chain reactions were used. Amplification data was analyzed by the sequence detector system (SDS 2.4, Applied Biosystems). Each plate included a serial dilution (25-0.04 ng) of fully methylated DNA: CpGenome™ Universal Methylated DNA) (Millipore) with the gene of interest and ALU-C4, several no template control (NTC) wells, five nanograms (5 ng) of a methylated control sample: CpGenome™ Universal Methylated DNA (Millipore), and five nanograms (5 ng) unmethylated sample consisting of whole genome amplified DNA from peripheral blood DNA. The percentage of methylated reference (PMR) was calculated for each sample according to the equation: 100×[(gene-x copy value)sample/(ALU-C4 copy value)sample][(gene-x copy value)Universal Methylated DNA (ALU-C4copy value)Universal Methylated DNA]. To classify each sample as methylated or unmethylated, a cutoff value was defined on the basis of mean+2× standard deviation of the methylation levels in urine samples from control individuals (including only those samples having methylation values above zero). PMR values used to define hypermethylation for each marker were: PMR (ZNF154)≧1.51, PMR (EOMES)≧0.348, PMR (HOXA9)≧0.077, PMR (POU4F2)≧0.371, PMR (TWIST1)≧0.405, and PMR (VIM)≧0.368.
Statistical Analysis:
Stata 11 (Statacorp, Tex., USA) was used for all statistical calculations. Two tailed tests were considered statistically significant if P<0.05. Methylation differences were evaluated by nonparametric Wilcoxon-Mann-Whitney test. Fisher's exact test was used for analyzing dichotomous variables. The exact χ2test was used for analyzing associations between clinico-pathological parameters with two or more categories. Correlations of the methylation levels of the markers were calculated with Spearman correlation coefficients. A ROC curve was prepared for each marker and combinations of markers by plotting sensitivity against (1-specificity) and the area under the curve (AUC) was calculated. Log-Rank tests were applied to evaluate equality of survival and Kaplan-Meier survival plots were used for visualization. Multivariate Cox regression analysis was used to analyze covariation between methylation markers, stage, grade, tumor multiplicity, and CIS.
Results:
The analysis of urine data was separated into two parts. The first available urine from each patient was analyzed and compared to non-malignant control urine samples. Then, the methylation markers in urine samples taken during follow-up of each patient were analyzed. The first available urine was from the incident tumor visit in 44 out of 184 cases, and from later recurrences in 140 cases. The level of the six markers: ZNF154, EOMES, HOXA9, POU4F2 (see Reinert et al., Clin Cancer Res, 2011), TWIST1 (see Renard et al., Eur Urol, 2009), and VIM (see Costa et al., Clin Cancer Res, 2010, 16(23): p. 5842-51) in the first available urine was compared to urine from 35 controls as indicated in Table 6. All six markers were highly significantly hyper-methylated in the urine from bladder tumor patients compared to controls, when analyzing both incident and recurrent tumors (Mann-Whitney, P<0.0001) as indicated in Table 8 and Table 9 below.
aSome urine samples provided inconclusive results for some markers
bPositive predictive value
cNegative predictive value
dMann-Whitney U test
eNot available
aSome urine samples provided inconclusive results for some markers
bPositive predictive value
cNegative predictive value
dMann-Whitney U test
eNot available
Better sensitivities and specificities of the markers were observed when analyzing urine from incident tumor visits compared to urine from recurrent tumor visits as indicated in Table 9. No association was observed between the individual markers and stage, but ZNF154, EOMES, POU4F2, and VIM were more methylated in grade III lesions compared to grade I lesions (Fisher's exact test, P≦0.048). See Table 10 below.
ZNF154, EOMES, and POU4F2 were less methylated in tumors with a size below 3 cm. (Fisher's exact test, P≦0.047). The methylation level of EOMES was associated with age (Fisher's exact test, P≦0.003). There was no association between any of the markers and nitrite status that indicated bacterial infection. None of the markers identified were independent of each other (Spearman's p test, P<0.0001) (data not shown).
Detection of Recurrences by Methylation Markers:
To test the clinical usefulness of the markers, 206 urine samples from the follow-up of 158 patients, 139 samples from patients with a recurrent bladder tumor, and 67 samples from patients with no tumor recurrence were analyzed. See Table 6. Employing the cut-points determined initially, using the control individuals, and only analyzing samples where the first sample was positive for methylation, sensitivity in the range from 87% to 94% was obtained, and specificity in the range from 28%-47%, AUC (95% CI) ranged from 0.70 (0.61-0.80) to 0.78 (0.71-0.86), PPV ranged from 72%-78%, and NPV from 55%-78% as indicated in Table 11 below.
aSome urine samples provided inconclusive results for some markers
bMann-Whitney U test
In comparison, the sensitivity of cytology was 77% and the specificity was 60%, the AUC (95% CI) was 0.68 (0.61-0.76), the PPV was 79%, and the NPV was 56%. Attempts to combine the markers resulted in lower specificity without much gain in sensitivity when combining two or more markers (results not shown). Of notice, urine samples from patients with recurrent tumors showed no significant associations between methylation and clinicopathologic variables as indicated in Table 12 below.
The above data was obtained from cystoscopy results from the urine sampling visit. Using cystoscopy results from the following 12 months of follow-up, many of the samples formerly classified as false positives were determined to be true positives. The adjusted methylation marker values were: sensitivity 88% to 94%, specificity 37%-66%, AUC (95% CI) 0.78 (0.68-0.89) to 0.84 (0.77-0.91), PPV 81%-90%, and NPV 55%-78%. See Table 13 below.
aSome urine samples provided inconclusive results for some markers
bMann-Whitney U test
Adjusted cytology values were: sensitivity 79%, specificity 74%, AUC 0.77 (0.69-0.84), PPV 89% and NPV 56%.
Prognostic Value of Methylation Markers for Predicting Recurrences:
The prognostic value of the methylation markers at non-recurrent visits was assessed. For all markers, it was determined that a positive marker at a tumor negative visit was significantly associated with later tumor recurrence in a 24 or 60 month follow-up period (Log-Rank test, P≦0.0397). See
Discovery of an Epigenetic Field Defect in Bladder Cancer Patients:
If the methylation of the biomarkers was confined to malignant cells forming tumors, the markers in urine should only be detected when a tumor was present, or occurring within a foreseeable future depending on the growth rate of the tumor. However, results indicate that even high urinary levels of methylation could be present at visits without recurrences. See
Primer and DNA Sequences for Methylation Markers:
This application claims the benefit of provisional application Ser. No. 61/491,912 filed Jun. 1, 2011 which is hereby incorporated by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| 61491912 | Jun 2011 | US |
