cDNA analyses of the MMR genes

MLH1, MSH2, MSH6, and PMS2

investigate the effect of VUS upon splicing,

detect unexpected splicing defects,

and find allelic losses indicating a germline defect

E. Holinski-FederLMU MünchenMGZ München

Pathogenicity assessment by RNA-analysis

Prerequists for RNA based pathogenicity assessment

• Allelic balance• Normal splicing pattern• Efficiency of NMD• Abnormal splicing pattern

Cut off definition by analysing more than 900 cDNAs Full lenght cDNA analysis protocol - EMMR working group

• missense variants

• splice site variants

• intronic

• synonymous

• 5`/ 3`UTR

• exon duplication

• ins/del of one amino acid

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

MLH1 MSH2 MSH6 PMS2

Isolation of patient’s total RNA

RNA derived from patient blood samples

short-term

2x heparine lymphocyte NMD-inhibition RNA isolation cDNA synthesis

cultures +puromycin dT(20)

-puromycin

72-96 h, 37°C Nonsense-Mediated-mRNA-Decay (NMD) 48°C, 75-90 min

early truncating > 55 nt prior to last exon-junction

T G MLH1 ~ 2.5 kb

MSH2 ~ 3 kb

MSH6 ~ 4.2 kb

PMS2 ~ 2.8 kb

Alternative splicing in controls

-> alternative splicing in cDNA-P and cDNA+P usually 0-10%

-> gene- and exon-specific exceptions with higher values

of up to 15% for MLH1 exon 1q, PMS2 exon 6p and exon 11,

of up to 25% for MSH6 exon 4

Cut off definition• alle

Allelic balalnce 50% +/- 10%Allelic loss < 10%Normal splicing up to 10%, some exceptionsAbberrant splicing above 30%

Universal protocol used within the european mismathch repair consortium

Testing 5 Class 5 variants

MLH1 class 4-5 variants • c.793C>T p.(Arg265Cys) • c.986A>C p.(His329Pro)• c.1984A>C p.(Thr662Pro)• c.1852_1854del p.(Lys618del)• c.350C>T p.(Thr117Met)

bioinformatical prediction: experimental analysis:

3 predicted deficiency 5 MMR deficiencyproteinfunction

5 variants predicted to affect one amino acid (putative missense/indel), class 4-5:

RNA analysis 5 not predicted as splice defects splice site defect: 2 complete, 1 partial

Effect on mRNA splicing• c.793C>T p.(Arg265Cys) • c.986A>C p.(His329Pro)• c.1984A>C p.(Thr662Pro)• c.1852_1854del p.(Lys618del)• c.350C>T p.(Thr117Met)

Testing 25 class 3 variants

variant cDNA result re-classification

3 potential SSD all splicing-defects 2x class 5 (complete SSD)

13 missense variants all splice-neutral, biallelic 0, remain class 3 due to unclear

1 exon duplication duplication verified 1x class 4 (funct. protein domain)

4 synonymous variants all splice-neutral, biallelic 4x class 2

4 VUS + pathogenic variant all splice-neutral, biallelic 2x re-classification to class 2

-> re-classification of 36% of VUS

-> splicing defects 12%

Testing MMR deficient unsolved patients

26 „mutation-negative“ patients with unsolved IHC loss in CRC:

MLH1/PMS2 in 16 patients

PMS2 in 5 patients

MSH2/MSH6 in 2 patients

MSH6 in 3 patients

➢ 5/26 abnormal cDNA results (19%)

➢ 21/26 normal FLT = no germline deficiency in MMR gene analyzed

cDNA result genomic cause

MLH1 NMD-P, PTC+P MLH1 truncating variant (overseen)

MLH1 NMD-P, SSD oof+P MLH1 splice site variant (overseen)

PMS2 NMD-P, SSD ins oof+P PMS2 partial intron inclusion

MLH1 allelic loss –P/+P ?

MLH1 allelic loss –P/+P ?

• protocol for full-length cDNA analysis for the four MMR transcripts

• definition of cut-off values for allelic representation, normal and aberrant splicing

• testing the system on different variants

missense variants (class 4-5) -> 60% splicing defects

25 different types of VUS class 3 -> 12% splicing defects, 36% re-classified

Summary

• testing the system on 26 MMR deficient, unsolved patients

19% with abnormal cDNA result = MMR-defect

81% with normal cDNA results = no MMR-defect in germline

Transfer to NGS-bases RNA-analysis by using these data to calibrate the system

and educate bio-IT systems

Creating a meaningfull algorithm for pathogenicity assessment

Thanks to all contributors

Monika Morak

Verena Steinke-Lange

Kerstin Schaefer, Trisari Maßdorf,

Brigitte Mauracher, Susanne Keinath

MDs & German HNPCC Consortium for patient contribution

Medizinische Klinik und Poliklinik IV, Campus Innenstadt, Klinikum der Universität München

MGZ – Medizinisch Genetisches Zentrum,Munich, Germany

Andreas Laner, Melanie Locher,

Udo Koehler, Tanja Haeussler

Nils Rahner

Medical Faculty, Institute of Human Genetics, Heinrich-Heine University, Düsseldorf

Jessica Bailey

Clinical Genetics, St. George's University Hospital NHS Foundation Trust, London

Christiane Kling

MVZ Labor Dr. Fenner und Kollegen, Hamburg

cDNA analysis using NGS technologies

Next step:

Lab-project

Use the defined cut-offs to validate

automated, NGS-based cDNA analyses

IT-project

Improve skills of bio-IT

InSiGHT guidelines

Classification of MMR gene variants

probably non-pathogenic

non-pathogenic

probablypathogenic pathogenic

VUS uncertain significance

1 2 3 4 5

Thompson et al. Nat Genet. 2014

VariantInterpretation

Committe

• mRNA splicing/expression

• functional protein assays

class

• splice site variants

• intronic

• synonymous

• 5`/ 3`UTR

• exon duplication

• predicted missense variants

• ins/del of one amino acid

=> RNA based analysis for pathogenicity assessment of sequence variants

• LR-PCR from cDNA: primers in first and last exon

• agarose gel electrophoresis

• Sanger sequencing

complete sequence analysis in overlapping reads of one orientation informative variant shows allelic representation

Full-length transcript amplification and sequencing

MLH1 ~ 2.5 kb

MSH2 ~ 3 kb

MSH6 ~ 4.2 kb

PMS2 ~ 2.8 kb

successful FLT amplification 98% PBL cDNA

T G

Allelic representation in controls

peak heights in sequence electropherogram -> relative intensities in %

results compared between cDNA-P and cDNA+P

Allelic representation MMR genes

0

10

20

30

40

50

60

0 1 2 3 4 5 6 7 8 9 10 11

MLH1-P MLH1+P MSH2-P MSH2+P MSH6 -P MSH6+P PMS2-P PMS2+P

Alternative splicing in controls

10 MMR-proficient controls for each MMR full-length transcript in cDNA –P and +P

cDNA-P cDNA+P

NMD-inhibition in controls

10 samples with early PTC variants :

-> NMD = allelic loss = AR of 0-10%

-> NMD-block usually works, but informative variant is needed for calculation of AR

-> splicing in PTC transcripts may be enhanced (15% in 1/10 cDNA+P samples)

0

10

20

30

40

50

0 2 4 6 8 10 12

cDNA-P cDNA+P

Definition of aberrant splicing

6 variants of class 4-5 predicted as splice site defect/exon skipping:

-

0

10

20

30

40

50

60

MLH

1 E

9-1

2 if

/E1

2 o

of

MLH

1 I1

3 o

of

+ A

L

MLH

1 E

14

oo

f/E1

4_

15

oo

f

MSH

2 E

5 if

MSH

2 E

16

p la

te t

r

PM

S2 E

8 o

of

cDNA-P cDNA+P-> aberrant splicing ≥30%

-> Splicing above alternative pattern/10% <30% = unclear, needs further investigation or AR

0

5

10

15

20

25

30

35

40

45

MLH

1 N

MD

1

MSH

2 N

MD

1

MSH

2 N

MD

2

MSH

2 N

MD

3

MSH

2 N

MD

4

MSH

2 N

MD

5

MSH

6 N

MD

1

PM

S2 N

MD

1

PM

S2 N

MD

2

PM

S2 N

MD

3

allelic representation with active/blocked NMD

Allelic representation of benign variants in controls

10 MMR-proficient controls, informative class 1-2 variant for each MMR gene

-> biallelic= allelic representation of 50% +/-10%

0

10

20

30

40

50

60

0 1 2 3 4 5 6 7 8 9 10 11

allelic representation of MMR class 1-2 variants

MLH1-P MLH1+P MSH2-P MSH2+P MSH6 -P MSH6+P PMS2-P PMS2+P

biallelic{

Definition of aberrant splicing

6 variants of class 4-5 predicted as splice site defect/exon skipping:

in-frame SSD

• 50±10% intensity in cDNA-P/+P

out-of-frame SSD -> NMD

• 0-20% intensity in cDNA-P (allelic loss)

• 30-40% intensity in cDNA+P

• exception: 15% SSD + allelic reduction 20%

-> aberrant splicing ≥30%

-> Splicing above alternative pattern/10% <30% = unclear, needs further investigation or AR

InSiGHT: demonstrate allele-specific splicing defect in absence of wildtype transcript

partial splicing defects = unclear clinical relevance, class 3

0

10

20

30

40

50

60

MLH

1 E

9-1

2 if

/E1

2 o

of

MLH

1 I1

3 o

of

+ A

L

MLH

1 E

14

oo

f/E1

4_

15

oo

f

MSH

2 E

5 if

MSH

2 E

16

p la

te t

r

PM

S2 E

8 o

of

} splice sitedefect

cDNA-P cDNA+P

*

AR 20%

{

{0

10

20

30

40

50

60

MLH

1 IV

S17

c.1

98

9+4

_1

98

9+5

insC

wit

h…

MSH

2 e

7 c

.12

75

A>G

p.(

Glu

42

5G

lu)

wit

h…

MSH

6 IV

S6 c

.35

56

+3_3

55

6+

13

de

l wit

h 6

2-…

MLH

1 e

10

c.7

99

_8

00

del

GTi

nsA

G…

MLH

1 e

11

c.1

01

3A

>G p

.(A

sn3

38

Ser)

MLH

1 e

13

c.1

41

8A

>G p

.(H

is4

73

Arg

)

MLH

1 e

17

c.1

96

4T>

C (

p.ll

e65

5Th

r)

MLH

1 e

19

c.2

13

5G

>C p

.(Tr

p7

12

Ser)

MSH

2 e

2 c

.31

9G

>C p

.(A

la1

07

Pro

)

MSH

2 e

8 c

.13

16

_1

31

8d

el p

.(P

ro4

39

de

l)

MSH

2 e

14

c.2

39

9T>

C p

.(Le

u8

00

Pro

)

MSH

6 e

5 c

.32

64

_3

26

6d

el p

.(P

he1

08

8d

el)

MSH

6 e

5 c

.32

86

C>

T p

.(H

is1

09

6Ty

r)

MSH

6 e

8 c

.37

58

T>A

p.(

Val

12

53

Glu

)

PM

S2 e

11

c.1

43

7C

>G

p.(

His

47

9G

ln)

PM

S2 e

13

c.2

24

0G

>C p

.(A

rg7

45

Thr)

PM

S2 e

12

c.2

00

7-?

_21

74

+?d

up

PM

S2 e

12

c.2

00

7-?

_21

74

+?d

up

MLH

1 e

12

c.1

40

1C

>T

p.(

Ser4

67

Ser)

MSH

2 e

10

c.1

56

0A

>G p

.(G

ly5

20

Gly

)

MSH

6 e

1 c

.13

5C

>T

p.(

Gly

45

Gly

)

MSH

6 e

4 c

.19

14

T>G

p.(

Leu

63

8Le

u)

splicing

exon splicing with VUS -P exon splicing with VUS +P AS tolerance

0

10

20

30

40

50

60

0 5 10 15 20 25cDNA-P cDNA+P cDNA+P controls

Testing VUs

splicing in exon of VUS:

splice site defect

alternative splicing

Patient samples with 25 different types of class 3 MMR variants:

• universal protocol for coding MMR variants to investigate their effect on splicing

allows re-classification of VUS

determines the allelic status in cis/trans of VUS to a pathogenic variant

Pros and cons of full-length transcript analysis

• discrimination against PMS2 pseudogenes

• needs informative variant for significant cDNA result (not in all samples, especially MSH2)

• semi-quantitative assesment of allelic representation and splicing

• requires good RNA quality for FLT amplification

• labour-intensive analysis with hands-on-time

• selection for wildtype FLT

allows detection of allelic loss by informative variant (transcript integrity test)

excludes irrelevant transcripts using alternative first exons/poly-adenylation