Current and emerging technologies for DNA methyla7on analysis. What a pathologist needs to know. RCPA Short Course in Medical Gene6cs and Gene6c Pathology 20 June 2011 Alex Dobrovic Molecular Pathology Research and Development Laboratory Peter MacCallum Cancer Centre Epigene6c Diagnos6cs •  DNA methyla6on as a marker of response to therapy •  diagnosis of a condi6on or predisposi6on to a condi6on •  diagnos6c use requires appropriately robust sensi6ve and specific methodology What are we measuring?
PCR based methods • capable of detec6ng methyla6on at all sites from limi6ng amounts of material • some may be used with degraded material eg formalin fixed paraffin sec6ons • PCR amplifica6on loses methyla6on informa6on! Bisulfite modifica6on • converts non-­‐methylated cytosine residues of single stranded DNA to uracil • uracil converted to thymine by PCR •  any Cs correspond to methylated Cs GACAT CG
!-> GATATTG!
m
GACAT CG
!-> GATATCG!
Bisulfite sequencing Methyla6on specific PCR (MSP) •  rapid screening •  specifically amplifies methylated sequences by using 3’ mismatches at Cs of CpG sites •  -­‐only assays a few sites •  sensi6ve to approximately 0.1% •  not quan6ta6ve trol
+ con
1
MB23
MB45
3
35
MB 4
BT20
8t
HS57
MB 4
68
7
MCF
ZR75
T42D
T24
er
Mark
MSP detection of APC methylation in
breast cancer cell lines
When do we want to sensi6vely measure methyla6on? •  cancer biomarker for early detec6on •  cancer biomarker for monitoring disease aVer therapy (mrd) –  plasma DNA •  marker of disease predisposi6on in normal 6ssues (cons6tu6onal soma6c methyla6on) MethyLight FAM
FAM
TAMRA
MethyLight •  Quan6ta6ve adapta6on of MSP based on TaqMan •  Real 6me analysis of amplifica6on lends itself to high throughput analysis –  Poten6ally sensi6ve to 1/10,000 –  Allows discrimina6on of signal from background •  Suitable for monitoring treatment using methyla6on as a tumour marker –  Eads et al. (2000) Methodologies based on high
resolution melting (HRM)
•  fluorescent dyes
•  dyes fluoresce only when
intercalated into ds DNA
•  heat PCR product
–  strands separate (melt)
–  monitor fluorescence
–  derive melting profile
•  melt temperature dependent on sequence
•  closed tube analysis of the PCR product
Methyla6on-­‐sensi6ve HRM(MS-­‐HRM) MGMT
BRCA1 methyla6on from FFPE A
B
no BRCA1 methylation in NSCLC
•  No methylation
detected in 56 N1
NSCLC samples
•  18% - Wang 2007
•  30% - Lee 2007
SMART-­‐MSP uses dye instead of probe
MGMT methyla6on -­‐ a predic6ve marker Heterogeneous MGMT methyla6on Patient
100%
MDA-MB-231
10%
WGA
MGMT methyla6on -­‐ Tm curves WGA
100%
Patient
10%
MGMT methyla6on associated with the T allele of a promoter SNP in normal 6ssues Methylated
Not
methylated
CC
(3)
74
77
CT and TT
7
5
12
10
79
89
p < 0.0001
MGMT •  are we measuring tumour specific methyla6on –  tumour MGMT methyla6on frequency es6mates by MSP are exaggerated •  Is the tumour heterozygously or homozygously methylated? What does quan6fica6on of methyla6on mean? •  Are we quan6fying CpG sites? •  Are we quan6fying methylated alleles (“epialleles”) Overall readout between primers
HRM
SSCA
DGGE
C
D
Average reading at each position between primers
Direct bisulfite sequencing
Bisulfite pyrosequencing
MassARRAY
->Quantitation at individual CpGs
Sequenom MassArray Pyrosequencing DAPK1 MS-­‐HRM Pyrosequencing reveals heterogeneity Sample
S13
S14
S15
S18
S23
1
8
7
6
37
20
2
8
7
6
43
27
3
10
5
6
36
17
4
11
7
8
38
17
5
8
4
6
12
9
6
7
4
4
7
10
7
30
16
18
40
51
8
69
11
6
7
21
9
48
13
10
8
46
10
17
28
6
8
17
11
24
22
11
13
46
12
36
44
14
57
36
0 10 20 30 40 50 60 70 80 90
13
23
26
14
46
22
14
19
37
10
31
27
15
13
18
8
42
15
16
13
34
9
15
15
Reading of epialleles
Cloned from undiluted templates:
Bisulfite genomic Sequencing
Limiting dilution
Digital MS-HRM & sequencing
Massively parallel sequencing
->epiallele quantification
Digital MS-­‐HRM •  eliminates heteroduplexes •  eliminates bias •  will dis6nguish PCR amplifica6on posi6ve wells from primer-­‐dimers •  counts the methylated alleles •  assesses extent of methyla6on of each allele •  only methylated alleles need to sequenced DAPK1 dMS-­‐HRM 13,14,23,18
15
Future prospects •
•
•
•
•
deep massively parallel sequencing methylBEAMing cheap digital nanotechnology approaches bisulfite modifica6on free methodologies •  Array based methodologies. Infinium: CDH1 methyla6on Take home messages •  appropriate choice of methodology •  relates to ques6on asked •  diagnos6c – –  needs to be FFPE friendly –  needs to be quan6ta6ve/ related to tumour purity •  most methods perform well for homogeneous methyla6on –  heterogeneous methyla6on is difficult to quan6fy without digital methologies •  much of the current literature is unreliable