CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2005;3:142–149
`
`Analysis of Promoter Methylation in Stool: A Novel Method for
`the Detection of Colorectal Cancer
`
`KONSTANZE LENHARD,* GUIDO T. BOMMER,* SILKE ASUTAY,* ROLF SCHAUER,‡
`THOMAS BRABLETZ,§ BURKHARD GÖKE,* ROLF LAMERZ,* and FRANK T. KOLLIGS*
`*Department of Medicine II, and ‡Department of Surgery, Klinikum Grosshadern, University of Munich, Munich, Germany; and the §Institute
`of Pathology, University of Erlangen, Erlangen, Germany
`
`Background & Aims: Detection of tumor-derived DNA
`alterations in stool is an intriguing new approach with
`high potential for the noninvasive detection of colorectal
`cancer (CRC). Because of heterogeneity of tumors, usu-
`ally multiple markers distributed throughout the human
`genome need to be analyzed. This is labor intensive and
`does not allow for high through-put screening. Therefore,
`markers with high sensitivity and good specificity are
`needed. We explored the potential of a single epigenetic
`marker in comparison with fecal occult blood testing
`(FOBT) for the discrimination of patients with CRCs and
`adenomas from those without. Methods: Methylation-
`specific polymerase chain reaction (PCR) was per-
`formed to analyze hypermethylated in cancer 1 (HIC1)
`promoter methylation status in a blinded fashion in
`stool samples from 26 patients with CRC, 13 with ade-
`noma >1 cm, 9 with hyperplastic polyps, 9 with chronic
`inflammatory bowel disease, and 32 with endoscopi-
`cally normal colon. Results: Ninety-seven percent of the
`stool samples contained amplifiable DNA. Forty-two per-
`cent of the samples from patients with CRC and 31% of
`the samples from patients with colorectal adenoma >1
`cm were positive for HIC1 promoter methylation. No
`methylated HIC1 promoter DNA was detected in the
`fecal DNA from patients with endoscopically normal
`colon or hyperplastic polyps. Conclusions: The epige-
`netic marker HIC1 promoter methylation carries high
`potential for the remote detection of CRCs. We postu-
`late that a panel of merely a few genetic and epigenetic
`markers will be required for the highly sensitive and
`specific detection of CRCs and adenomas in fecal sam-
`ples from affected patients.
`
`Colorectal cancer (CRC) is one of the leading causes of
`
`cancer-related morbidity and mortality.1 About
`40% of patients die within 5 years of being diagnosed.
`This mainly is attributable to late presentation with
`locally advanced or metastatic disease in one third of
`cases, precluding curative surgery.2 The majority of
`CRCs follow the adenoma-carcinoma sequence, requiring
`a time period of usually more than 10 years.3 Detection
`
`of early disease and precancerous adenomatous lesions
`leads to a decrease of CRC-related mortality.4 –7 There-
`fore, there is a strong rationale for screening programs.
`Sigmoidoscopy, colonoscopy, and fecal occult blood test
`(FOBT) alone or in combination with endoscopy, are
`recommended for screening of the average-risk popula-
`tion.8 However, FOBT, when applied regularly, can
`decrease CRC mortality by only 15%–33%.5–7 Compli-
`ance with endoscopic screening is not satisfactory. Over-
`all, in 1998, only 37% of eligible adults in the United
`States had been screened for CRC in the previous 3
`years.9
`Detection of tumor-derived genetic changes in stool
`is a promising new approach for CRC screening. Stud-
`ies published so far focus on the detection of muta-
`oncogenes,10 –16
`tions
`in
`and
`tumor-suppressor
`genes,17–19 as well as changes in microsatellite mark-
`ers.20 Two recent studies required the analysis of up to
`15 different genetic markers21 or 144 aliquots from
`each stool18 for the detection of 91% and 61% of
`cancers and 82% and 50% of adenomas ⱖ1 cm, re-
`spectively. This precludes their application for routine
`high throughput screening owing to high labor inten-
`sity and costs. The identification of new markers with
`high sensitivities and specificities is needed. So far,
`little attention has been paid to DNA hypermethyl-
`ation as a potential stool marker for CRC.
`Methylation of CpG islands of promoters leads to
`silencing of transcription of the affected gene.22,23
`Methylation-specific polymerase chain reaction (PCR)
`(MSP)24 has been used successfully to detect DNA
`methylation in primary tumors and in various body
`fluids.25 The feasibility of amplification of methylated
`
`Abbreviations used in this paper: CI, confidence interval; CRC, colo-
`rectal cancer; FOBT, fecal occult blood test; HIC1, hypermethylated in
`cancer 1; MSP, methylation-specific PCR; PCR, polymerase chain re-
`action.
`© 2005 by the American Gastroenterological Association
`1542-3565/05/$30.00
`PII: 10.1053/S1542-3565(04)00624-X
`
`Geneoscopy Exhibit 1004, Page 1
`
`

`

`February 2005
`
`HIC1 METHYLATION IN STOOL 143
`
`Table 1. Clinicopathologic Data and HIC1 Methylation Status of Patients With CRC
`
`No.
`
`Sex
`
`Age (y)
`
`UICC stage
`
`Localization
`
`Symptomsa
`
`HIC1 methylation
`
`FOBT
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`M
`F
`M
`M
`M
`M
`M
`F
`M
`M
`M
`M
`M
`M
`F
`M
`M
`M
`F
`F
`M
`M
`F
`M
`M
`F
`
`71
`75
`66
`69
`65
`56
`63
`72
`74
`68
`66
`75
`71
`67
`72
`59
`59
`53
`55
`61
`61
`80
`53
`63
`76
`60
`
`IV
`III
`IV
`II
`III
`IV
`IV
`III
`III
`IV
`IV
`III
`IV
`II
`IV
`II
`II
`III
`IV
`IV
`IV
`II
`IV
`I
`III
`IV
`
`Sigmoid colon/rectum
`Sigmoid colon
`Descending/sigmoid colon
`Ascending colon
`Transverse colon
`Rectum
`Ascending colon
`Cecum
`Sigmoid colon/rectum
`Rectum
`Rectum
`Cecum
`Sigmoid colon
`Hepatic flexure
`Sigmoid colon
`Rectum
`Rectum
`Descending colon
`Rectum
`Sigmoid colon
`Rectum
`Transverse colon
`Ascending colon
`Cecum
`Transverse colon
`Ascending colon
`
`1
`1,3
`0
`0
`0
`1
`2
`2
`0
`3
`1
`0
`0
`3
`1,2
`0
`0
`0
`2
`2,3
`0
`2
`1,3
`4
`2
`1
`
`UICC, International Union Against Cancer; M, male; F, female; NA, no amplification.
`a0, no symptoms; 1, change in bowel habits; 2, blood admixed with stool; 3, abdominal pain; 4, anemia.
`
`⫺
`⫺
`⫺
`⫹
`⫹
`⫹
`⫺
`⫹
`⫺
`⫹
`⫺
`⫺
`⫹
`⫹
`⫹
`⫹
`⫺
`⫺
`⫺
`⫺
`⫺
`⫹
`⫺
`NA
`⫺
`⫹
`
`⫺
`⫺
`⫺
`⫺
`⫹
`⫺
`⫺
`⫹
`⫹
`⫹
`⫹
`⫺
`⫺
`⫺
`⫺
`⫺
`⫹
`⫺
`⫹
`⫺
`⫺
`⫺
`⫺
`⫹
`⫹
`⫺
`
`DNA from stool samples of patients with CRC has
`been reported.26,27 In one study evaluating several
`potentially methylated genes, SFRP2, SFRP5, and
`PGR were found to be methylated differentially in the
`stool of patients with CRC.27 Of these markers,
`SFRP2 was found to be the most sensitive fecal meth-
`ylation marker, detecting 77%–90% of CRCs. How-
`ever, specificity of SFRP2 methylation was quite poor,
`at 77%. Therefore, identification of both sensitive and
`highly specific methylation markers is required before
`MSP can be evaluated as a stool-based screening pro-
`cedure for CRC in prospective studies.
`We studied the potential of hypermethylated in
`cancer 1 (HIC1) promoter methylation as a stool-based
`DNA marker. The promoter of HIC1, a candidate
`tumor-suppressor gene localized on 17p13.3 and the
`first gene cloned based on the finding of CpG island
`hypermethylation in cancer,28,29 frequently is meth-
`ylated in CRC,30 but not in normal or aging colonic
`tissue.31 We show that HIC1 promoter methylation
`can be detected frequently and with high specificity in
`stool samples from patients with CRCs. The combi-
`nation of HIC1 methylation analysis with FOBT al-
`lowed for the detection of two thirds of CRCs.
`
`Materials and Methods
`Patients and Stool Samples
`
`The investigation was approved by the ethical com-
`mittee of the Medical Faculty of the University of Munich.
`Stool samples were collected preoperatively from patients
`with verified CRCs (Table 1), and before endoscopy from
`patients with adenomas ⱖ1 cm (Table 2), endoscopically
`normal colons, hyperplastic polyps, and chronic inflamma-
`tory bowel disease (Table 3). Most patients with CRCs had
`been referred with a known diagnosis, all persons with
`adenomas and normal colons were asymptomatic and un-
`derwent colonoscopies for surveillance reasons.
`Stool samples were collected before cathartic preparation for
`scheduled surgery or colonoscopy. A period of at least 12 days
`was respected between previous biopsy examinations of colo-
`rectal tumors before stool sampling. Samples were received
`within 10 hours after defecation at the laboratory, aliquoted,
`and stored at ⫺80°C.
`Isolation of DNA
`
`Stool samples were thawed on ice and DNA was
`isolated using the QiaAmp DNA Stool Mini-Kit (Qiagen,
`Hilden, Germany) according to the manufacturer’s instruc-
`tions. Briefly, DNA was isolated from stool samples of
`approximately 200 mg, and purified DNA was eluted from
`the column in 50 ␮L of deionized water. The total yield of
`
`Geneoscopy Exhibit 1004, Page 2
`
`

`

`144 LENHARD ET AL
`
`CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 3, No. 2
`
`Table 2. Clinicopathologic Data and HIC1 Methylation Status of Patients With Adenomatous Polyps
`
`No.
`
`Sex
`
`Age (y)
`
`Diagnosis
`
`HIC1 methylation
`
`FOBT
`
`27
`28
`29
`30
`31
`32
`33
`34
`35
`36
`37
`38
`39
`
`M
`F
`M
`M
`M
`F
`M
`F
`M
`M
`M
`M
`F
`
`72
`63
`63
`76
`83
`65
`72
`60
`69
`70
`64
`72
`63
`
`Tubular adenoma, ascending colon, 3 cm
`Tubulovillous adenoma, ascending colon, hepatic flexure, sigmoid colon, 1–1.5 cm
`Tubulovillous adenoma, ascending colon, 2 cm
`Tubulovillous adenoma, ascending and transverse colon, 1, 3, 4 cm
`Tubulovillous adenoma, transverse colon, 3 cm
`Tubulovillous adenoma, sigmoid colon, 1.5 cm
`Tubulovillous adenoma, ascending colon, 1 cm
`Tubular adenoma, sigmoid colon, 1.5 cm
`Tubular adenoma, 1.5 cm, ascending colon
`Tubulovillous adenoma, sigmoid colon, 1 cm
`Tubulovillous adenoma, rectum, 1 cm
`Adenoma, cecum, 2 cm
`Tubulovillous adenoma, sigmoid colon, 2 cm
`
`⫹
`⫺
`⫺
`⫹
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫹
`⫹
`
`⫺
`⫺
`⫺
`⫹
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`
`M, male; F, female.
`
`DNA was determined by ultraviolet absorbance at 260 nm;
`the range of DNA concentrations was 120 –300 ng/␮L.
`Leukocyte DNA was extracted with the Qia-Amp Blood
`Mini Kit (Qiagen) according to the manufacturer’s recom-
`mendations. Microdissected paraffin-embedded sections of
`normal colon and CRCs were treated with lysis buffer
`containing .5% Tween 20 (Sigma, Taufkirchen, Germany),
`20 ␮g proteinase K (Sigma), 50 mmol/L trizma base at pH
`8.9, and 2 mmol/L ethylenediaminetetraacetic acid. After
`overnight incubation at 56°C, proteinase K was heat inac-
`tivated. Samples were subjected to phenol:chloroform/chlo-
`roform DNA extraction twice, chloroform extraction, and
`ethanol precipitation. Precipitated DNA was dissolved in
`Tris-ethylenediaminetetraacetic acid buffer and stored at
`4°C.
`
`Methylation-Specific Polymerase Chain
`Reaction
`MSP is based on a 2-step process.24 First, DNA is
`treated with sodium bisulfite, which efficiently converts un-
`methylated cytosine to uracil but leaves methylated cytosines
`unchanged. Consequently, after treatment, methylated and
`unmethylated alleles have different sequences. Second, allele-
`specific primers allow for the specific amplification of methyl-
`ated and unmethylated alleles.
`A total of 2 ␮g DNA from peripheral blood leukocytes,
`microdissected colonic tissues, and stool was diluted to 50
`␮L of distilled water. Then 2 ␮g of salmon sperm DNA was
`added as a carrier to stool and tissue DNA. DNA was
`denaturated by .2 mol/L NaOH for 15 minutes at 37°C. A
`total of 30 ␮L of 10 mmol/L hydrochinone (Sigma) and 520
`␮L of 3 mol/L sodium bisulfite (Sigma) at pH 5 were added,
`and the samples were incubated for 16 hours at 50°C. After
`bisulfite treatment, DNA was purified using the Wizard
`DNA Clean-up System (Promega, Mannheim, Germany),
`following the manufacturer’s protocol, and incubated for 5
`minutes at room temperature with .3 mol/L NaOH. DNA
`was ethanol-precipitated and resuspended in 50 ␮L of
`deionized water and stored at ⫺80°C for up to 8 weeks.
`
`Primers were designed to discriminate between methylated
`and unmethylated alleles. Primers were designed to hybridize
`to the CpG-rich region of the HIC1 promoter 1b (Figure 1A),
`which is affected by methylation in carcinomas.31 All PCR
`reactions were performed in a blinded fashion. The person
`performing the reactions was not aware of the diagnosis of the
`patients. A 2-step nested PCR was established to improve
`sensitivity. The methylated and unmethylated primer pair
`sequences for the first-step PCR were as follows: 5=-TTAGG-
`TATTGAAGTTGTGAAGTTGT-3= (unmethylated forward),
`5=-ACCATAACAACCAATAAAATACACC-3=
`(unmethyl-
`ated reverse), 5=-GTTAGGTATCGAAGTCGTGAAGTC-3=
`(methylated forward), and 5=-TTACCATAACAACCGATA-
`AAATACG-3= (methylated reverse). The second-step PCR
`was performed with the following primers: 5=-GAAGTTGT-
`GAAGTTGTTTGTGTAG -3= (unmethylated forward), 5=-
`ACTTAACAACCACACTCACCA -3= (unmethylated reverse),
`5=-GAAGTCGTGAAGTCGTTTGCGTAG -3= (methylated
`forward), and 5=-ACTTAACAACCGCGCTCGCCG -3= (meth-
`ylated reverse). The sizes of the first- and second-round PCR
`products were 231 and 181 bp, respectively.
`Both PCRs were performed in 20 ␮L reaction volumes
`containing 1 ⫻ PCR buffer (Qiagen), 4 mmol/L MgCl2, 250
`␮mol/L deoxynucleotide triphosphate mixture (Invitrogen),
`1 ␮mol/L of each primer, 1 ⫻ Q-Solution (Qiagen), 2 ␮L
`bisulfite-treated DNA, and 1.7 units taq DNA polymerase
`(Hot Star Taq; Qiagen). Thermocycler conditions for the
`first-step PCR were as follows: 95°C for 15 minutes (95°C
`for 30 s; 60°C for unmethylated and 56°C for methylated
`primers for 30 s; 72°C for 30 s) for 20 cycles followed by a
`final extension at 72°C for 10 minutes. PCR products were
`purified using the QIAquick PCR Purification Kit (Qiagen)
`according to the manufacturer’s instructions and one tenth
`of the elution volume was subjected to the second-step
`PCR. Here the cycling conditions were as follows: 95°C for
`15 minutes (95°C for 30 s; 54°C for unmethylated and
`64°C for methylated primers for 30 s; 72°C for 30 s) for 40
`cycles followed by a final extension at 72°C for 10 minutes.
`PCR products were analyzed by 3% agarose gel electro-
`
`Geneoscopy Exhibit 1004, Page 3
`
`

`

`February 2005
`
`HIC1 METHYLATION IN STOOL 145
`
`Table 3. HIC1 Methylation Status of Healthy Controls and
`Patients With Nonneoplastic Disease
`
`No.
`
`Sex
`
`Age
`
`Diagnosis
`
`HIC1
`methylation
`
`FOBT
`
`SssI (New England Biolabs, Frankfurt, Germany). Indicated
`genome equivalents32 corresponding to the DNA content of
`1 to 1000 human cells were spiked into normal stool DNA
`and the nested PCR was performed.
`
`Fecal Occult Blood Testing
`
`All stool samples were subjected to FOBT (Beckman
`Coulter, Fullerton, CA) immediately on receipt and before
`freezing. One drop of the peroxide catalyst was added to the
`reverse side of each window of the test cards 1 day after
`applying stool samples. A blue color reaction within 60
`seconds was considered a positive result.
`
`Statistical Analysis
`
`Sensitivity and specificity were estimated relative to
`the results of colonoscopy in the usual manner. Exact bi-
`nominal estimate was used to calculate 95% confidence
`intervals (CIs) for the estimated parameters.33 P values were
`calculated with the Fisher exact test. Calculations were
`performed using SISA online statistics package (http://
`home.clara.net/sisa/binomial.htm; Daan G. Uitenbroek,
`Hilversum, The Netherlands).
`
`100 bp
`100 bp
`
`PROMOTER 1b
`PROMOTER 1b
`
`CpG
`CpG
`
`H2O
`C 1
`C 2
`C 3
`H2O
`C 1
`C 2
`C 3
`C 5
`C 4
`C 6
`C 5
`C 4
`C 6
`U M U M U M U M U M U M U M
`U M U M U M U M U M U M U M
`
`N 1 N 3
`N 4
`H2O
`
`N 1 N 3 N 1 N 3
`
`N 4N 4
`
`H2OH2O
`N 5
`N 8
`N 9
`
`N 5N 5
`
`N 8N 8
`
`N 9N 9
`U M U M U M U M U M U M U M
`U M U M U M U M U M U M U M
`U M U M U M U M U M U M U M
`
`H2O
`L 1
`L 2
`L 3
`
`L 1 L 1
`
`L 2 L 2
`
`L 3L 3
`
`H2OH2O
`L 5
`L 4
`L 6
`
`L 4L 4
`
`L 6L 6
`
`L 5L 5
`U M U M U M U M U M U M U M
`
`U M U M U M U M U M U M U MU M U M U M U M U M U M U M
`
`100copies
`100copies
`
`1000copies
`1000copies
`
`DNA
`DNA
`
`ol
`ol
`sto
`sto
`U M
`U M
`
`size marker
`size marker
`
`1copy
`1copy
`
`10copies
`10copies
`
`O
`O
`H2
`H2
`U M U M U M U M U M
`U M U M U M U M U M
`
`A
`A
`
`B
`B
`
`C
`C
`
`D
`
`DD
`
`E
`E
`
`200 bp-
`200 bp-
`
`40
`41
`42
`43
`44
`45
`46
`47
`48
`49
`50
`51
`52
`53
`54
`55
`56
`57
`58
`59
`60
`61
`62
`63
`64
`65
`66
`67
`68
`69
`70
`71
`72
`73
`74
`75
`76
`77
`78
`79
`80
`81
`82
`83
`84
`85
`86
`87
`88
`89
`
`F
`F
`F
`M
`M
`F
`M
`F
`F
`M
`M
`F
`F
`F
`M
`M
`F
`F
`M
`M
`F
`F
`M
`F
`F
`F
`M
`F
`M
`M
`M
`F
`M
`F
`F
`F
`F
`F
`F
`F
`F
`M
`M
`M
`M
`M
`M
`F
`M
`M
`
`61
`61
`73
`63
`47
`41
`53
`59
`73
`36
`71
`65
`60
`66
`80
`70
`27
`57
`55
`23
`73
`79
`68
`78
`38
`62
`76
`61
`23
`74
`60
`49
`59
`66
`67
`65
`58
`76
`54
`76
`60
`47
`64
`27
`67
`53
`51
`44
`30
`37
`
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Normal
`Hyperplastic polyp
`Hyperplastic polyp
`Hyperplastic polyp
`Hyperplastic polyp
`Hyperplastic polyp
`Hyperplastic polyp
`Hyperplastic polyp
`Hyperplastic polyp
`Hyperplastic polyp
`Ulcerative colitis
`Ulcerative colitis
`Ulcerative colitis
`Ulcerative colitis
`Ulcerative colitis
`Ulcerative colitis
`Crohn’s disease
`Crohn’s disease
`Crohn’s disease
`
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`NA
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`NA
`⫺
`⫺
`⫺
`⫺
`⫹
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`
`F, female; M, male; ND, not determined; NA, no amplification.
`
`⫹
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫹
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`ND
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫺
`⫹
`⫹
`⫺
`⫹
`⫺
`⫺
`⫺
`⫺
`⫺
`
`phoresis and ethidium bromide staining followed by visu-
`alization with ultraviolet illumination using a digital im-
`aging system (Intas, Göttingen, Germany). The first-step
`MSP followed by the second-step MSP independently was
`performed twice for each sample. For the sensitivity assay,
`human leukocyte DNA was treated with the CpG methylase
`
`Figure 1. (A) Location of the amplified region (underlined) of the HIC1
`promotor by MSP. (B) Detection of unmethylated (U) and methylated
`(M) HIC1 promoter DNA in primary CRCs, (C) normal colonic epithe-
`lium, and (D) human peripheral blood leukocytes. (E) Analysis of
`sensitivity of the HIC1 MSP assay. Indicated genome equivalents of
`fully methylated DNA were spiked into normal stool DNA.
`
`Geneoscopy Exhibit 1004, Page 4
`
`

`

`146 LENHARD ET AL
`
`CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 3, No. 2
`
`
`
`AA
`
`
`
`
`
`# 6 # 6
`# 7 # 7
`# 8# 8
`# 11 H2O# 11 H2O
`
`# 9 # 10# 9 # 10
`
`U M U M U M U M U M U M U MU M U M U M U M U M U M U M
`
`
`
`BB
`
`
`
`# 27 # 28 # 29# 27 # 28 # 29
`# 30 # 31 # 32 H2O# 30 # 31 # 32 H2O
`
`U M U M U M U M U M U M U MU M U M U M U M U M U M U M
`
`
`
`CC
`
`
`
`# 63 # 64 # 65# 63 # 64 # 65
`# 66 # 67 # 81 H2O# 66 # 67 # 81 H2O
`
`U M U M U M U M U M U M U MU M U M U M U M U M U M U M
`
`Figure 2. Detection of unmethylated (U) and methylated (M) HIC1
`promoter DNA in stool samples from patients with (A) CRC, (B) ade-
`noma, and (C) normal controls.
`
`Results
`Establishment of Hypermethylated in
`Cancer 1 Promoter Methylation Analysis
`and Determination of the Sensitivity of
`the Assay
`
`DNA was extracted from microdissected tissues
`from primary CRCs, microdissected samples from nor-
`mal colonic epithelium from individual patients, and
`peripheral blood leukocytes from healthy individuals. A
`nested MSP amplifying a CpG island within the HIC1
`promoter 1b was established (Figure 1A). We found 7 of
`9 cancers (78%) to contain both methylated and unmeth-
`ylated and 2 cancers to contain unmethylated HIC1
`DNA only (Figure 1B and data not shown). Furthermore,
`none of 8 samples from normal colonic epithelium con-
`taining analyzable DNA after bisulfite conversion re-
`vealed any detectable methylated HIC1 promoter DNA
`(Figure 1C and data not shown). In accordance with
`previously reported data,30 we found that HIC1 pro-
`
`moter methylation is frequent in primary CRCs and
`absent in healthy colorectal epithelium. Because stool of
`a significant fraction of patients with CRC and nonma-
`lignant disease contains overt or occult blood we ana-
`lyzed the HIC1 methylation status of peripheral blood
`leukocytes from 10 healthy individuals. MSP revealed
`that none of the 10 leukocyte samples contained any
`detectable methylated HIC1 DNA (Figure 1D and data
`not shown). Based on these data we were prompted to
`analyze the potential of HIC1 promoter methylation
`analysis as a stool-based marker for CRC detection.
`To determine the sensitivity of the assay to detect
`tumor-derived methylated DNA in the presence of a
`surplus of normal stool DNA, 1–1000 genome equiva-
`lents of SssI-treated fully methylated leukocyte DNA
`were diluted into stool DNA from a healthy individual.
`MSP revealed that the equivalent of methylated HIC1
`DNA from 10 cells could be detected with the assay
`(Figure 1E). This shows the high sensitivity of the assay
`because the vast majority of DNA is degraded during
`bisulfite conversion.34 Tumors shedding less methylated
`DNA into the lumen of the colon would not be detected
`by this assay.
`
`Analysis of Hypermethylated in Cancer 1
`Promoter Methylation in Stool DNA
`
`DNA was extracted and purified from small
`amounts of stool from 26 patients diagnosed with CRC
`before surgery (Table 1) and of 63 patients before
`colonoscopy (32 with confirmed healthy colons, 13 with
`adenomas ⬎1 cm, 9 with hyperplastic polyps, and 9 with
`chronic inflammatory bowel disease; Tables 2 and 3).
`Amplification of the unmethylated allele, which was
`expected for all samples, was positive in 97% of cases and
`was used as a control for amplifiable DNA after bisulfite
`conversion. Five of 13 patients with cancers of the rec-
`tum or sigmoid colon tested positive for HIC1 methyl-
`ation (sensitivity 39%; 95% CI, 14%– 68%) (Figure 2A
`
`Table 4. Positivity Rates of the Fecal DNA Assay, FOBT, and Combination of the Tests
`
`HIC1
`
`FOBT
`
`Combined
`
`n (%)
`
`95% CI
`
`n (%)
`
`95% CI
`
`n (%)
`
`95% CI
`
`All cancers
`Localized cancers
`Advanced cancers
`Rectum and sigmoid colon
`From descending colon to cecum
`Asymptomatic cases
`Symptomatic cases
`Adenomas
`
`11/26 (42)
`4/6 (67)
`5/20 (25)
`5/13 (39)
`6/13 (46)
`4/10 (40)
`7/16 (44)
`4/13 (31)
`
`23%–63%
`22%–96%
`9%–49%
`14%–68%
`19%–75%
`12%–74%
`20%–70%
`9%–61%
`
`9/26 (35)
`2/6 (33)
`7/20 (35)
`5/13 (39)
`4/13 (31)
`3/10 (30)
`6/16 (38)
`1/13 (8)
`
`17%–56%
`4%–77%
`15%–59%
`14%–68%
`14%–68%
`7%–65%
`15%–65%
`.2%–36%
`
`17/26 (65)
`6/6 (100)
`11/20 (55)
`9/13 (69)
`8/13 (62)
`6/10 (60)
`11/16 (69)
`4/13 (31)
`
`44%–83%
`12%–188%a
`32%–77%
`39%–91%
`39%–91%
`26%–88%
`41%–89%
`9%–61%
`
`aEstimated by normal approximation of Poisson distribution.
`
`Geneoscopy Exhibit 1004, Page 5
`
`

`

`February 2005
`
`HIC1 METHYLATION IN STOOL 147
`
`and Table 4). Sensitivity for the 13 cancers of the cecum,
`and ascending, transverse, and descending colon was
`46% (95% CI, 19%–75%) (Figure 2A and Table 4).
`Altogether the assay detected 42% (95% CI, 23%– 63%)
`of all CRCs. Interestingly, the sensitivity was marginally
`higher for localized (International Union Against Cancer
`stages I and II) than advanced disease (67% vs. 25%, P
`⫽ .084). Next to detecting CRCs, a stool-based screen-
`ing assay also should detect relevant premalignant lesions
`(ie, adenoma). Thirteen stool samples from patients with
`lesions ⱖ1 cm were collected before
`adenomatous
`polypectomy. Four of the 13 patients (31%) with ade-
`noma tested positive for methylated HIC1 DNA (Figure
`2B, and Table 4).
`To test for specificity, stool samples from 32 patients
`with normal colonoscopy results, 9 patients with hyper-
`plastic polyps, and 9 patients with a known history of
`chronic inflammatory bowel disease were analyzed by
`MSP. One stool sample from a patient with ulcerative
`colitis who did not reveal any malignant process of the
`colon on colonoscopy contained methylated HIC1 DNA.
`All other samples from normal controls, and patients
`with hyperplastic polyps or chronic inflammatory bowel
`disease, tested negative for methylated HIC1 DNA
`(Figure 2C, and Table 3). Altogether only 1 of 50
`samples (2%) from patients lacking neoplastic disease of
`the colon was positive, resulting in a specificity of 98%
`(95% CI, 92%–100%).
`HIC1 MSP was compared with the best evaluated
`stool-based screening test to date: FOBT. Sensitivity and
`specificity of FOBT found in this study is comparable
`with published data.35 The HIC1 assay was only slightly
`more sensitive for all cancers (42% vs. 35%; P ⫽ .19;
`Table 4). However, although there was no difference
`between both tests in detecting rectal and sigmoid can-
`cers, symptomatic or asymptomatic cases (P ⫽ .66, P ⫽
`.26, and P ⫽ .33, respectively), the HIC1 assay detected
`more localized cancers (67% vs. 25%, P ⫽ .085) and
`more adenomas (31% vs. 8%, P ⫽ .14). In addition,
`specificity of the HIC1 assay was higher (97.9% vs.
`89.9%, P ⫽ .102). The combination of both assays
`resulted in increased detection rates for CRCs (65%;
`95% CI, 44%– 83%). Although the combined test de-
`tected all localized cancers, no increase in the sensitivity
`for adenoma was seen.
`
`Discussion
`Highly sensitive, specific, and easily analyzable
`markers are required for noninvasive stool-based CRC
`screening. We show here that DNA methylation can be
`detected in stool of CRC patients with high specificity.
`
`The use of the DNA methylation marker, HIC1 pro-
`moter methylation, allowed for the highly specific detec-
`tion of CRCs in 42% and of adenomas ⱖ1cm in 31% of
`cases. Interestingly, sensitivity was quite high for early
`cancers and higher for adenoma than FOBT. The com-
`bination of HIC1 MSP with FOBT resulted in the de-
`tection of almost two thirds of CRCs.
`DNA methylation was first described in primary tu-
`mors. More recently, tumor-derived methylated DNA
`has been detected in various remote media, including
`serum, urine, and sputum.25 In 30%–70% of cases with
`methylation of a specific marker in the primary colorectal
`tumor, the same methylated marker also could be de-
`tected in serum.36 –38 However, to date, methylation
`markers have not been assessed thoroughly in stool. Our
`data suggest that we detected at least 50% of cases
`expected to be methylated at the HIC1 locus, because
`methylation of HIC1 has been reported to occur in
`60%– 80% of CRCs.28 Nonetheless, in at least 25% of
`expected cases with HIC1 methylation no methylation
`was detected by our assay. Because only small amounts of
`stool were used for DNA extraction and 84%–96% of
`DNA is lost during bisulfite conversion owing to deg-
`radation,37 it is conceivable that tumors shedding only
`small amounts of DNA into the lumen of the colon will
`not be detected. The use of larger stool samples might
`increase sensitivity because only a fraction of the stool
`DNA is of human origin. It has been reported that the
`feces of healthy individuals contains 200 –500 ng and the
`feces of CRC patients contains 1500 –2000 ng human
`DNA per mg of stool.39 The amount and quality of
`extracted and converted DNA is most likely also the
`reason for the failure to amplify an unmethylated product
`in 3 cases. Although our nested PCR allowed for the
`amplification of stool DNA in 97% of cases, another
`study using a single-step fluorescent PCR approach failed
`to amplify DNA after bisulfite conversion in more than
`40% of cases.27
`DNA methylation of promoters has been shown to be
`an early event in colorectal carcinogenesis already occur-
`ring in aberrant crypt foci.40 Promoters of some genes are
`affected more frequently by methylation than oth-
`ers.40 – 42 The reason for this currently is not very well
`understood. Promoters of several genes, including the
`estrogen receptor and insulin-like growth factor II, have
`been reported to be methylated in both CRCs and in
`aging colon,31,43 rendering these promoters unattractive
`markers for screening owing to low specificity for CRC.
`So far, no HIC1 promoter methylation has been found in
`the colon mucosa of healthy controls31 and our analysis of
`samples from normal colon did not reveal any HIC1
`
`Geneoscopy Exhibit 1004, Page 6
`
`

`

`148 LENHARD ET AL
`
`CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 3, No. 2
`
`study proposed
`promoter methylation. One recent
`SFRP2 hypermethylation as a marker for stool-based
`detection of CRC.27 However, this marker has low spec-
`ificity for CRC. Although SFRP2 methylation was de-
`tected in 77%–90% of CRCs with amplifiable DNA,
`23% of colonoscopically healthy controls revealed meth-
`ylation of this locus as well. This low specificity is most
`likely owing to the SFRP2 methylation that occurs fre-
`quently in premalignant aberrant crypt foci.44 In one
`study, 13 of 15 of these lesions revealed hypermethyl-
`ation of SFRP2. Because aberrant crypt foci are not
`detected by routine colonoscopy and their progression
`rate to adenoma and CRC is not known, their detection
`by a screening assay does not seem to be favorable.
`Moreover, K-ras mutation, the best evaluated molecular
`stool marker to date, has a sensitivity for CRC detection
`in stool samples of 40%– 60% of cases.10 –16 But K-ras
`mutations also have been found in the stool of controls
`with normal colons. In one study, K-ras mutations were
`detected in 7% of healthy individuals.20 Self-limiting
`hyperplastic polyps, nondysplastic aberrant crypt foci,
`and morphologically normal colonic mucosa,45– 47 which
`do not progress to cancer, frequently contain K-ras mu-
`tations. This shows that methylation markers as well as
`genetic markers need to be selected carefully to achieve
`both high sensitivity and specificity. Choosing sensitive
`markers that already are positive in a premalignant sit-
`uation will compromise specificity severely. In addition,
`low specificity of a test results in a low positive predictive
`value and consecutively in high rates of unnecessarily
`performed colonoscopies.
`One of the stool samples from a patient with ulcerative
`colitis who did not reveal any detectable colonic malig-
`nancy contained methylated HIC1 DNA. The reason for
`this is not entirely clear. However, aberrant methylation
`of several promoters has been reported in normal and
`dysplastic colonic epithelium of patients with long-
`standing and extensive ulcerative colitis48,49 as well as in
`ulcerative colitis–associated CRCs.50,51 Alternatively,
`the patient could have an as yet unknown disease or
`malignant process somewhere else throughout the gas-
`trointestinal tract, shedding methylated HIC1 DNA
`into the lumen, or a pathologic finding was overseen by
`the colonoscopy.
`Although methylated DNA of several markers has
`been found in hyperplastic polyps from patients with
`multiple hyperplastic polyps as well as in patients with
`hyperplastic polyposis, no DNA methylation has been
`described in sporadic hyperplastic polyps.52,53 In accor-
`dance, none of the stools analyzed from patients with
`single hyperplastic polyps revealed methylated HIC1
`
`DNA. Nonetheless, detection of hyperplastic polyps in
`patients with hyperplastic polyposis would be favorable
`because these lesions have been found to be associated
`with CRC.53 Altogether, of the 50 stool samples from
`controls with either endoscopically normal colons, hy-
`perplastic polyps, or inflammatory bowel disease, only
`one was MSP positive. This specificity of 98% empha-
`sizes the value of HIC1 promoter methylation as a sen-
`sitive and specific marker for stool-based CRC screening.
`In summary, this study shows HIC1 as a sensitive and
`specific marker well suited for stool-based CRC screen-
`ing. Combination of HIC1 with a few additional sensi-
`tive and specific methylation markers may allow for the
`highly sensitive and specific stool-based detection of
`CRCs and adenomas. Evaluation of other genes known to
`be methylated in CRC such as hMLH154 and HPP1,55
`identification of other useful methylation markers, pro-
`spective screening studies, as well as further simplifica-
`tions and automatization of the assays used, clearly are
`indicated.
`
`References
`1. Jemal A, Thomas A, Murray T, et al. Cancer statistics. CA Cancer
`J Clin 2002;52:23–47.
`2. McArdle C. ABC of colorectal cancer. Primary treatment-does the
`surgeon matter? BMJ 2000;321:1121–1123.
`3. Kinzler KW, Vogelstein B. Lessons from hereditary colorectal
`cancer. Cell 1996;87:159 –170.
`4. Winawer SJ, Zauber AG, Ho MN, et al, the National Polyp Study
`Workgroup. Prevention of colorectal cancer by colonoscopic
`polypectomy. N Engl J Med 1993;329:1977–1983.
`5. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from
`colorectal cancer by screening for fecal occult blood. Minnesota
`Colon Cancer Control Study. N Engl J Med 1993;328:1365–1371.
`6. Hardcastle JD, Chamberlain JO, Robinson MH, et al. Randomised
`controlled trial of faecal-occult-blood screening for colorectal can-
`cer. Lancet 1996;348:1472–1477.
`7. Kronborg O, Fenger C, Olsen J, et al. Randomised study of
`screening for colorectal cancer with faecal-occult-blood test. Lan-
`cet 1996;348:1467–1471.
`8. Burt W. Colon cancer screening. Gastroenterology 2000;119:
`837– 853.
`9. Breen N, Wagener DK, Brown ML, et al. Progress in cancer
`screening over a decade: results of cancer screening from the
`1987, 1992, and 1998 National Health Interview Surveys. J Natl
`Cancer Inst 2001;93:1704 –1713.
`10. Sidransky D, Tokino T, Hamilton SR, et al. Identification of ras
`oncogene mutations in the stool of patients with curable colorec-
`tal tumors. Science 1992;