Data Software

Location:

Seattle, WA

Posted:

November 23, 2012

Contact this candidate

Resume:

February ****

Application Note

Barrett s Esophagus and Esophageal

Adenocarcinoma Epigenetic Biomarker

Discovery Using Infinium Methylation

Contributed by Patricia C. Galipeau, Dennis L. Chao, Xiaohong Li, Jessica D. Arnaudo, Heather D. Kissel,

Carissa A. Sanchez, and Brian J. Reid, Fred Hutchinson Cancer Research Center, Seattle, Washington

from individuals diagnosed with Barrett s

I ntroductIon

There is a growing body of evidence that esophagus (BE).

epigenetic alterations are not only common, BE is a premalignant condition in

but may precede major genetic changes which the normal squamous epithelium

that lead to cancer1 3. Selected gene-specific is replaced by a specialized intestinal

hypo- and hyper-methylation of CpG metaplasia in response to acid reflux. It

islands have been shown to cause overex- is the only known precursor to esophageal

pression of oncogenes and transcriptional adenocarcinoma (EA). Methylation changes

silencing of tumor suppressor genes, specific to BE/EA samples have been

respectively, possibly leading to clonal previously identified6,7. These observations

evolution4,5. Therefore, we pursued an lead us to hypothesize that BE tissue has

epigenetic evaluation of DNA samples a unique, tissue-specific methylation

From top left to bottom right, Carissa Sanchez, Heather Kissel, Patricia Galipeau, Dennis Chao

(Bottom) Xiaohong Li, Brian Reid, Jessica Arnaudo.

S ampleS and data acquISItIon

F igure 1. Comparison oF methylation We obtained 60 fresh, frozen biopsies from

levels between repliCates

24 patients diagnosed with BE. The patients

represented a cross-section of all stages

of BE, from early BE (no genomic instability

and no progression to cancer over mean

0.80

11.59 years) to advanced EA. Samples from

Replicate 1

0.60

three different tissue types, including

The Infinium

Barrett s epithelium and adjacent normal

0.40

Methylation Assay

proximal squamous and distal gastric

tissue, were collected from 18 of these

was highly repro-

0.20

patients. EA samples from visible tumors

ducible, showing

were collected from surgical resections

0 0.20 0.40 0.60 0.80 1

of six patients.

excellent agree-

Replicate 2

The methylation fraction of >27,500

ment of CpG

the beta methylation values at 28,245 cpG

individual CpG sites was measured in all

sites in two replicate biological samples

the samples using Illumina s Infinium

site methylation

were compared. this beadStudio plot shows

excellent agreement between the same dna Methylation Assay. As input to the assay,

fractions between

sample that was bisulfite-treated and run in

500 ng of DNA was bisulfite-converted

replicate.

replicates

using the EZ DNA Methylation Gold kit

(Zymo Research). After bisulfite conversion,

signature that distinguishes it from the methylation level at each CpG site was

adjacent squamous and gastric tissue, determined by measuring the methylation

regardless of the stage of neoplastic pro- fraction (beta), defined as the fraction of

gression. In addition, methylation events methylated signal over the total signal

are selected during progression to EA. (unmethylated + methylated). The Infinium

Previous evaluations of CpG methylation Assay includes redundant, built-in, bisulfite

in BE/EA studies have been limited by the conversion quality controls that measure

inability to characterize the methylation the conversion rate of non-CpG cytosines

status of large numbers of genes at once, and background signal. Using controls eval-

allowing only 1 20 genes per sample to be uated with Illumina BeadStudio software,

evaluated in a single study. The Infinium we determined that non-methylated

HumanMethylation27 BeadChip from cytosines in all 72 biological samples were

Illumina represents a significant technologi- efficiently bisulfite converted, providing a

cal advance in epigenetic studies. This within-assay assessment of background

platform characterizes the methylation signal and bisulfite conversion rate.

status of over 27,500 CpG sites across more

than 14,000 genes. It provides semi-quanti- r eproducIbIlIty and valIdatIon of the

methylatIon reSultS

tative methylation data without the use

of a standard curve or reference locus and

Reproducibility

is technically compatible with small,

The Infinium Methylation Assay was highly

endoscopic biopsy and surgical resection

reproducible, showing excellent agreement

specimens.

of CpG site methylation fractions between

In this application note, we describe the

replicates (average r2=0.98 for 12 pairs

use of the Infinium HumanMethylation27

of replicates, with a range of 0.97 to 0.99).

BeadChip to measure CpG methylation in

Figure 1 shows a typical high level of

samples obtained from individuals with

concordance between two replicates of

Barrett s esophagus. We validate these data

a single DNA sample processed in indepen-

by Pyrosequencing and demonstrate that

dent bisulfite treatments and amplified in

achieved results are similar to those for

two different Infinium Assay reactions.

an earlier experiment performed using

Illumina s GoldenGate Methylation Assay.

F igure 2. Consistent goldengate and inFinium methylation results

CpG Site 1 CpG Site 2 CpG Site 3

1.0

A B C

B BSTS BBB

G G

S G

B B

Infinium Beta Methylation

B SB G B

T=tumor B

0.8

S B

B=Barrett s

S=squamous

0.6

S S

G=gastric SS S

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

GoldenGate Beta Methylation

GoldenGate Beta Methylation GoldenGate Beta Methylation

methylation fractions of three cpG sites measured by Infinium and GoldenGate assays. the plots

show that the results from the Infinium assay were generally consistent with those from the

GoldenGate assay.

...highly

methylation values between the two assays

Comparison of Infinium

HumanMethylation27 BeadChip with the (Figure 2). We found 60 of 117 sites to

quantitative

GoldenGate Methylation Cancer Panel I

have significantly different methylation

We found that the measurements of

Pyrosequencing

fractions between the normal squamous

methylation levels from the new Infinium

and the BE samples on the GoldenGate

data shows strong

Methylation Assay agreed with our earlier

platform (Wilcoxon test p

study conducted on the GoldenGate plat-

correlation with

of the 60 sites, or 88%, also had significantly

form using Illumina s Methylation Cancer

different fractions on the Infinium platform.

data obtained

Panel I, which included 1,536 CpG sites

Thus, we find that the Infinium platform

using the high-

across 808 genes. The two experiments

produced results consistent with our

used independent biopsies from the same

density Infinium

GoldenGate study.

patients. We compared the beta methyla-

Assay

tion values between the GoldenGate and Technical validation of Infinium

HumanMethylation27 BeadChip with

Infinium Assays at the 117 CpG sites that

Pyrosequencing

are targeted on both platforms. Although

The Infinium Methylation Array was

the assays were performed in different

validated using Pyrosequencing. Assays

biopsy samples, used different probes,

were designed and optimized to sequence

and were based on different technologies,

the same CpG sites for nine different genes

we found general agreement of beta

Figure 3. high Correlation between inFinium humanmethylation27 beadChip and

quantitative pyrosequenCing data

SFRP2 EYA4

r = 0.928

r = 0.984

Infinium Beta Methylation

0.8

0.6

0.4

0.2

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

Pyrosequencing Methylation Pyrosequencing Methylation

comparison of methylation fractions in eya4 and Sfrp2 assessed using pyrosequencing (x-axis)

and the Infinium methylation platform (y-axis beta methylation value) in the same bisulfite-treated

biological samples and methylation controls (0).

F igure 4. apC methylation levels at multiple Cpg sites

S q u am o u s BE

Sample Sample

Average Beta Methylation Value

Results from

this pilot study

agree with those

from a similar

study conducted

using Illumina s

GoldenGate

platform. The

* *

Infinium Assay has

the advantage

of covering nearly

1-234*-*-*-*-**** 6 7

CpG Site CpG Site

20 times more

genes, with greatly

average beta methylation values (y-axis) for seven cpG sites associated with apc in squamous tissue

(left panel) and be tissue (right panel). each sample is represented as a different color, as customized

increased CpG site

within the beadStudio software.

*methylated control

coverage

assayed on the Infinium platform (two of 36 genes (Wilcoxon rank sum test p

llumina.com.

for reSearch uSe only

Contact this candidate