Acute myeloid leukemia leads the way in molecular cancer genetics: - PowerPoint Presentation

Slide1

Acute myeloid leukemia leads the way in molecular cancer genetics: precision medicine at reach?

Lars BullingerUniversity of Ulm / MDC Berlin-Buch

MODERN

PROBLEMS OF GENETICSdedicated to the 115th anniversary of the birth of N. W. Timofeeff-RessovskyCancer genetics & Cancer therapy4th June 2015, St. Petersburg

Slide2

Introduction

Acute Myeloid Leukemia (AML)

Slide3

Acute Myeloid Leukemia (AML)Epidemiology:- 3% of all cancers-

Incidence increases with age Clinical course:

CR

Survivalpts < 65 years70-80% ~35%

pts

> 65 years

40-60%

~5%

Definition:

- Clonal expansion of myeloid blasts in bone marrow, blood or tissue.

NCI, SEER data base

Slide4

AML CytogeneticsLow-riskt(15;17) PML-RARA

t(8;21) AML1-ETOinv(16) CBFB-MYH11Intermediate-risk

normal karyotypet(9;11)

MLL-AF9High-riskinv(3) EVI1 complex karyotypeOverall survival

0

1

2

3

4

5

6

7

8

9

10

0

20

40

60

80

100

time (years)

low-risk; n=261

intermediate-risk; n=698

high-risk; n=171

Slide5

Omics and NGS in AML

Slide6

Approved

treatment options for AML

vs. other hematologic malignancies

1. NCI Drug Info. http://www.cancer.gov/cancertopics/druginfo. 2. EMA Drug Approvals. http://www.ema.europa.eu/ema/index.jsp?curl=pages/includes/ medicines/medicines_landing_page.jsp&mid=. 3. FDA Drug Approvals. http://www.accessdata.fda.gov/ scripts/cder/drugsatfda/index.cfm. 3. NCCN clinical practice guidelines in oncology: acute myeloid leukemia. National Comprehensive Cancer Network website. V.2.2014. http://www.nccn.org/professionals/physician_gls/PDF/aml.pdf.

US approvals

EU approvals

Subsequently withdrawn

Slide7

18-45 yrs, n=1,57445-60 yrs, n=2,156

60-70 yrs, n=963>70 yrs, n=437

T

ime (years)Survival (%)

0

1

2

3

4

5

6

7

8

9

10

0

25

50

75

100

Improving

outcome

in AML

remains

a

major challenge

No. of pts: n=5,130; only pts considered eligible for intensive induction therapy

Slide8

Genomic landscape of de novo AML

TCGA Research Network. N Engl J Med 2013

Slide9

Genomic landscape of de novo AML

TCGA Research Network. N Engl J

Med 2013

59%18%22%Signaling genes FLT3, KIT, NRASMyeloid transcription-factor genesRUNX1, CEBPATranscription-factor fusion genes RUNX1-RUNX1T1, MYH11-CBFB

Slide10

Genomic landscape of de novo AML

TCGA Research Network. N Engl J Med 2013

NPM1

Tumor-suppressor genesTP53, WT1, PHF6DNA-methylation-related genes DNMT3A, TET2, IDH1/2Signaling genes FLT3, KIT, NRASChromatin-modifying genesMLL-X, ASXL1, EZH2Myeloid transcription-

factor genes

RUNX1

,

CEBPA

Cohesin

-

complex genes SMC1A, SMC3, RAD2127%22%59%

44%

30%

16%13%14%Spliceosome-complex genesSF3B1, U2AF118%Transcription-factor fusion genes RUNX1-RUNX1T1, MYH11-CBFB

Slide11

Genomic heterogeneity and evolution

=> improved understanding of clonal heterogeneity at diagnosis might provide means to prevent relapse caused by evolution of persisting subclonesWelch et

al. Cell

2012

Slide12

Clonal architecture and genetic heterogeneity

Lindsley and Ebert. Blood 2013

Slide13

Genomic heterogeneity and evolution

Krönke

et

al. Blood 2013

Slide14

? DNMT3A

DNMT3A

DNMT3A

DNMT3ANPM1

Loss of

NPM1

BHMT2, CDH26, FSD1,

KAT6B, PDLIM1,

PSTPIP2,

TOP2B, ZBTB47

ARID4B, BRAF, CACNA1E, CHUK,

COL5A2, DDX58,

FAT2,

LINGO4,

MDN1, NF1, NPAT,

PARS2, PTPRO

Diagnosis

Relapse

2000

2010

pre

-leukemic HSC ?

time

(

years

)

SureSelect

Human All Exon 50Mb

Average coverage: diagnosis sample, 73.6 fold;

r

elapse

sample,

88.9 fold

pre-leukemic

HSC

persistence

?

Remission

Exome

sequencing

of

NPM1

mut

loss cases

Slide15

DNMT3A

/

ABL1

x104NPM1/ABL1x104diagnosisinduction IICons. III

f-

up

(2,75

yrs

)

induction

IRelapseLoss of

NPM1mut

NPM1mut (BM)DNMT3Amut

AMLSG HD98A

trial

,

pt

.

died

, in

relapse

BM

PB

BM

BM

BM

MRD of

DNMT3A

mut

-R882H and

NPM1

mut

Slide16

Identification of pre-leukemic HSCs in AML

Shlush

et

al. Nature 2014

DNMT3A

mutation precedes

NPM1

mutation in human

AML

DNMT3A

mutations present in stem/ progenitor cells at diagnosis and remission

Slide17

Therapy related AML (t-AML)

Wong et al. Nature 2015The mutational burden in t-AML is similar to de novo AMLHSC clones harboring somatic TP53 mutations are detected

in patients before cytotoxic therapy exposure

Slide18

Loss of TP53 confers a clonal advantage

Wong et al. Nature 2015

Slide19

Clonal evolution model in t-AML/t-MDS

Bullinger. Hematotopics 2015

Slide20

Targeted re-sequencing in AML

Basis for precision medicine?

Slide21

Clinical diagnostics

: which biomarkers should be tested for?

Prognosis: which biomarkers provide prognostic information independent from others?

Prediction: which biomarkers are able to predict response to a specific therapy (novel agents)?Molecular therapy: which molecular lesions can be targeted therapeutically?Genetics guided therapeutic approaches:clinical practical challenges

Slide22

Systematic characterization of

myeloid neoplasms

1540 adult patients with AML

Enrolled in 3 trials of the German-Austrian AML Study GroupTargeted re-sequencing of 111 genes involved in pathogenesis of myeloid neoplasms (SureSelect target enrichment)Objectives:Identify genetic lesions that contribute to disease pathogenesis and classificationIdentify secondary and tertiary gene-gene interactionsEvaluate prognostic and predictive

impact

E. Papaemmanuil, M.

Gerstung

, P. Campbell

R. Schlenk, K. Döhner, L. Bullinger, H. Döhner

Slide23

Genomic landscape of AML

6 genes in >10% of pts; 13 genes 5-10%; 24 genes 2-5%; 37 genes <2%

CN-AML => more gene mutations than in AML with

chrom. abnormalitiesDriver mutations significantly increased with age (p<0.001)

Slide24

Timing of driver mutation acquisition

Genes

involved with epigenome modeling (

DNMT3A, ASXL1, TET2) were typically acquired earliestGenes involved in receptor tyrosine kinase (RTK) / RAS signaling occurred as late eventsvirtual time axisSubclonal heterogeneity and informative timings could be inferred for 690 (64%) of 1076 pts. with two or more mutations

Slide25

Implication of genomics for classification

Formal statistical analysis (Bayesian latent class models)

11 non-overlapping molecular classes can be identified

NPM1 mutation, with significant contribution from DNA methylation / hydroxymethylation genes DNMT3A, TET2, IDH1, IDH2Biallelic CEBPA mutationTP53 mutation and / or chromosomal aneuploidies

Splicing factor genes or regulators of chromatin and transcription

DNMT3A

/

IDH2

(in the absence of

NPM1

)6 balanced rearrangements:

inv(16), t(15;17), t(8;21), t(11q23),

inv(3), t(6;9)

Slide26

Implication of genomics for classification

11 “non-overlapping” molecular AML classes

1332/1540 (86%) of AML classified, with minimal overlap across categories

Slide27

Reference map of gene-gene interactions

>200 significant interactions

CA,

chrom. aneuploidies I SF, splicing factor I RTK, receptor tyrosine kinases

Slide28

Risk contributions in each patient

RED: short survival

BLUE: long survival

Slide29

Unique constellations of risk

factors

Slide30

Risk groups

Quartiles of predicted risk

121 intermediate-1 risk patients will be reclassified as very high risk case

very low low high very highFavorable 308 131 33 3inter-1 19 104 174 121inter-2 36 65 85 84Adverse 4 23 44 185

Slide31

Personally

tailored cancer management based

on

genomic and clinical dataComprehensive genomic profiles of 111 cancer genes and cytogenetic data from 1,540 AML patient can be used in conjunction with clinical data sets to accurately predict outcome for each patientData can be used to compute absolute probabilities of competing events such as relapse, non-relapse mortality or salvage rates on an individual level for each patient and under different treatment options -> basis for rationalized clinical decision support

Incorporating all genomic driver mutations into prognostic models outperforms models using conventional prognostication schemes

Genomic data account for approx. 2/3 of the predicted risk of overall survival

Slide32

AML prediction tool

29-year old female patient with t(8;21)-positive AML

Intensive Chemotherapy

Allogeneic HCT in 1

st

CR

Slide33

Targeted resequencing in the clinic?

using e.g. Illumina sequencing by synthesis technology (MiSeq

)

=> AML panel of 31 genes (560 target regions, 222kb)Library preparation e.g. HaloPlex (~500kb) ~6h

Cluster generation and sequencing

(2x100bp -> 1Gb)

~14h

Data analysis

(supported workflow)

<

2

h

ASXL1, CBL, CEBPA, CTCF, DNMT1, DNMT3A, ETV6, EZH2, FLT3, GATA2, HIPK2, IDH1,

IDH2, JAK2, KIT, KRAS, MLL3, MLL5, NF1, NPM1, NRAS, NSD1, PHF6, RAD21, RUNX1, SF3B1, SFPQ, TET1, TET2, TP53, WT1

Slide34

NGS based targeted personalized therapy

Slide35

Clinical implementation?

Current status

Slide36

Translation into the clinic

Newly diagnosed AML

IC molecular screening

Registration AMLSG-BiO AMLSG BiO-ID

0-8 hours

BM&PB samples

 Reference Lab

Overnight

Molecular screening

-

PML-RARA

-RUNX1-RUNX1T1

-CBF



-MYH11

-MLL-AF9

-

FLT3

-ITD

-

FLT3

-TKD

-NPM1

-CEBPA

overnight

24-48 hours

Genotype

adapted

strategy

APL

CBF

NPM1

mut

FLT3

-ITD

Other

Slide37

Genetics guided AML therapy

Trial

Midostaurin

AMLSG 16-10ATRA +/- GO AMLSG 09-09NAPOLEON GIMEMA/AMLSG/SALAPOLLO +/- ATO-ATRA-Ida

+/-

Dasatinib

AMLSG 21-13

+/-

Panobinostat

AMLSG 22-14+/- Volasertib AMLSG 20-13

+/-

Crenolanib

AMLSG 19-13EPZ 5676 (DOT1L) Palbociclib (CDK6) AMLSG 23-14 Genotype

AML

FLT3

mut

CBF-AML [

KIT

]

Molecular

Screening

24-48

hrs

AML

NPM1

mut

Other

subtypes

,

mainly

high-

risk

APL

[

PML-RARA

]

AML

MLL

rearr

Slide38

PML-RARA ATRA, ATOKIT mutations dasatinib, midostaurin

FLT3 mutations midostaurin, sorafenib; quizartinib, crenolanibIDH mutations AG-221MLL-rearranged DOT1L, CDK6 inhibitor

Epigenetic mutations /

azacitidine, decitabine, SGI-110alterations (?) OTX015, I-BET-762Selected targets for molecular therapy

Slide39

Precision medicine in AML:

fact

or fiction?

We have entered a new era in leukemia genomics=> however, large gene panel testing and whole exome/genome sequencing remain research toolsCurrently, cytogenetics and NPM1, CEBPA, FLT3-ITD mutational screening are standard of care (WHO / ELN update in 2016)The explosion of knowledge has yet to be translated into therapeutic benefit=> however, a number of novel compounds are at the horizon that hold promise to enter the clinicMajor challenge: identification of predictive biomarkers that help selecting the appropriate therapy for an individual patient=> integrate biosampling, companion studiesEnter your patients, younger or older, on a clinical trial!

Slide40

P. Campbell

E

. Papaemmanuil

CambridgeM. HeuserG. GöhringF. TholB. SchlegelbergerA. GanserMHH, Hannover

S.

Cocciardi

A.

Dolnik

V. Gaidzik

S. Kapp-Schwörer

J.

KrönkeK. Lang

F. KuchenbauerP. Paschka

F. RückerF. StegelmannD. WeberK. HolzmannK. DöhnerR. F. SchlenkS. StilgenbauerH. DöhnerUlm University

A.

Krivtsov

S. Armstrong

New York

G. Martinelli

I.

Iaccobucci

Bologna

P. Valk

B. Löwenberg

Rotterdam

SFB 1074

S. Fröhling

C.

Plass

P. Lichter

C. Scholl

Heidelberg

K.

Rajewsky

S. Sander

Berlin