ph-like (bcr/abl1-like): un apporoccio baato sul target ...€¦ · characterization of...

Ph-like (BCR/ABL1-like): un apporoccio baato sul target molecolare

Sabina Chiaretti, MD PhD

LE SFIDE DELLA MEDICINA DI PRECISIONE IN EMATOLOGIABologna 28 Giugno, 2018

Topics: BCR/ABL1-like and other subgroups

• Molecular background

• Incidence

• Diagnosis

• Outcome and MRD

• Treatment

• Molecular background

• Incidence

• Diagnosis

• Outcome

• Treatment

Topics: BCR/ABL1-like

Haferlach et al, Blood 2005Chiaretti et al, CCR 2005

First report in adult ALLs

2005: first identification, by GEP, of a subset of adult B-lineage ALL clustering together with BCR/ABL1+ ALL cases

Characterization of BCR-ABL1-like in children

GEP: Identification, within children (n=297), of a subset with a transcriptional profile resembling that of BCR/ABL1+cases (≈15-20%)

Clinical features: Hyperleukocytosis, poor response to VCR, ASP and DNM, poor prognosis (reduced DFS at 5 years andincreased resistance to induction)

Array-CGH: IKZF1, PAX5, TCF3 and VPREB1 deletions , CRLF2 deregulationDen Boer et al, Lancet 2009; Mullighan et al, NEJM 2009

Deletions are visualised in red, whereas amplifications are shown in blue.

Impact of CRLF2 expression on B-ALL survival

CRLF2-high

p=0.004

Surv

ival

pro

bab

ility

Chiaretti et al, Leuk Res 2016

CRLF2-low

del(X)(p22.33p22.33)del(Y)(p11.32;p11.32)

t(X;14)(p22;q32)t(Y;14)(p11;q32)

Point mutations(F232C)

• 5-7% of pediatric BCR/ABL1-negative cases • >50% of cases associated with Down's syndrome• 5-15% of adult BCR/ABL1-negative cases • Associated with mutant JAK and IKZF1

d-CRLF2

P2RY8/CRLF2

IGH@/CRLF2

CRLF2-low CRLF2-highCRLF2 deregulationDCt>8 DCt<8

100

75

50

25

00 10 20 30 40 50 60 70

Months

Associated with the BCR/ABL1-like profile, but not useful as a diagnostic marker

In high risk ALL, RNA-seq has identified novel mutations that involve TKs in the majority of cases. They appear to havetransforming capability and to respond to TKIs.

Roberts KG. Cancer Cell 2012; 22:153-66 Roberts KG, et al. N Engl J Med 2014;371:1005–1015

BCR/ABL1-like ALL in adults. Genetics

Roberts KG et al, JCO 2016

Higher frequency of other kinase involvement.Some cases do no have any lesions.

Kinase fusions identified in Ph-like ALLKinase gene

Fusion partners, n

Patients, n 5’ genes

ABL1 6 14 ETV6, NUP214, RCSD1, RANBP2, SNX2, ZMIZ1

ABL2 3 7 PAG1,* RCSD1,* ZC3HAV1*

CSF1R 1 4 SSBP2*

PDGFRB 4 11 EBF1, SSBP2,* TNIP1,* ZEB2*

CRLF2 2 30 IGH, P2RY8

JAK2 10 19 ATF7IP,* BCR, EBF1,* ETV6, PAX5, PPFIBP1,* SSBP2, STRN3, TERF2,* TPR*

EPOR 2 9 IGH, IGK*

DGKH 1 1 ZFAND3*

IL2RB 1 1 MYH9*

NTRK3 1 1 ETV6†

PTK2B 2 1 KDM6A,* STAG2*

TSLP 1 1 IQGAP2*

TYK2 1 1 MYB*

Roberts KG, et al. N Engl J Med 2014;371:1005–1015

• Molecular background

• Incidence

• Diagnosis

• Outcome

• Treatment

Topics: BCR/ABL1-like

BCR-ABL1-like. Incidence

Incidence is higher in AYA (10% in children; 27% in AYA). NEVER detected in cases positive for known fusion transcripts (BCR/ABL1, KMT2A-r, TCF3/PBX1)

However:- It highly depends on the denominator (all B-lineage ALL or B-neg ALL) and

- on the assay used for BCR/ABL1-like identification

- More adult cases are being evaluated →incidence in adults is almost equal to AYA ≥ 25%

Roberts KG, NEJM 2014 371:1005-15; Heroldt T, NEJM 2014 371:2235; Chiaretti S et al, in press

• Molecular background

• Incidence

• Diagnosis

• Outcome

• Treatment

Topics: BCR/ABL1-like

BCR/ABL1-like ALL diagnosis: a difficult issue

• Well identified subgroup by GEP

• Poor prognosis documented: therapy?

Difficult to identify these cases by techniques otherthan GEP

Furthermore, non univocal gene signature

Still a difficult issue…

LDA

Q-RT-PCR

Quantification of expression of 15 transcripts(Kang BW et al, ASH 2013)

Quantification of expression of 10 transcripts(Chiaretti S et al, BJH 2018)

LDA, FISH, RT-PCR, NGS (RNA-seq, WGS, WES) (Roberts KG et al, NEJM 2014)

FISH, RT-PCR, Q-RT-PCR, NGS (Fasan A et al, ASH 2015)

Known fusion transcripts, JAK2 mutations , CRLF2-r(Herold T et al, Haematologica 2016)

CRLF2 expression (FC), JAK2 mutations, FISH for TK-rearrangements and CRLF2-r (Jain N et al, ASH 2017)

As far as possible, diagnostic assays should be available in most centers (or in centralized laboratories)

Integrated algorithms

BCR/ABL1-like ALL diagnosis: a difficult issue

• Well identified subgroup by GEP

• Poor prognosis documented: therapy?

Difficult to identify these cases by techniques other than GEP

Furthermore, non univocal gene signature

To set up a diagnostic assay for a straightforward identification of BCR/ABL1-like ALL cases

To analyze the clinico-biologic and molecular features of BCR/ABL1-like adult ALL cases

To verify BCR/ABL1-like cases response to TKIs (ponatinib)

BCR/ABL1-like ALL cases: methods (I)

342 Probsets

337 Probsets

Harvey R.C. et al.

9

CRLF2

• By meta-analysis of published and in house GEP data of B-NEG ALL, selection of 9 BCR/ABL1-like specifictranscripts (FC 1.5, p<0.001) + CRLF2

• Confirm the differential expression of the 10 transcripts by a Q-PCR approach

• Build a “BCR/ABL1-like predictor” on the basis of Q-PCR results (easy, fast and economic!)

BCR/ABL1-like ALL cases: methods (II)

26 negative B-ALL

52 adult ALL samples previously evaluated by GEP

• Q-RT expression values shrinked into 3 principal components.

• A logistic regression model was used to examine the association among the 3 principalcomponents and BCR/ABL1-like cases.

• Generation of a score on principal components, used to classify the remaining samples(screening panel).

26 BCR/ABL1-like ALL cases

Discovery panel

Development of the BCR/ABL1-like predictor and screening

Selection of BCR/ABL1-like specificgenes and validation by Q-PCR

Development ofBCR/ABL1-like predictor*

Screening

Factor loadings

PC1 PC2 PC3Gene1 2^(-ΔCt) 0.87921 -0.08451 0.19315

Gene2 2^(-ΔCt) 0.87162 0.41262 0.05974

Gene3 2^(-ΔCt) 0.80540 0.30143 0.22085

Gene4 2^(-ΔCt) 0.66775 0.40903 0.41369

Gene5 2^(-ΔCt) 0.57903 0.48726 0.48781

Gene6 2^(-ΔCt) 0.17357 0.90884 0.08792

Gene7 2^(-ΔCt) 0.61788 0.68881 0.14372

Gene8 2^(-ΔCt) 0.13521 0.65104 0.55717

Gene9 2^(-ΔCt) 0.17012 0.01622 0.90066

Gene10 2^(-ΔCt) 0.22407 0.26363 0.89565

54/194 newly identified BCR/ABL1-like patients (28%)- 9.5% children, 29% AYA, 30% adults -

1. Q-PCR of predictor genes in 129 B-NEG ALL

2. BCR/ABL1-like predictor

1. Selection of 10 predictor genes

2. Validation by Q-PCR in 52 B-NEG ALL

1. Identification of principal components (PCs)

2. Definition of a score

Chiaretti S et al, BJH in press*https://redcap.gimema.it/redcap

BCR/ABL1-like ALL cases: clinical features

BCR/ABL1-like (n=54)

non-BCR/ABL1-like (n=140)

P-value

Gender (M/F) 36/18 78/62 p=ns

Age 32 (6-72) 28 (0-78) p=p=ns

Wbc x 109/l 22.6 (1.89-239) 12.4 (0.6-425) p=0.023

Plts x 109/l 47 (0.15-283) 47 (1-308) p=ns

Hb g/dl 9.7 (4.1-15.3) 8.9 (3.7-15.8) p=ns

CR rate 77.8% 89.2% p=0.06

No significant differences in age and gender: significantly higher WBC and lower remission rate

Molecular features of the cases identified by the BCR/ABL1-like predictor

BCR/ABL1-like (N=28) non-BCR/ABL1-like (N=26)

• 96% of cases had a lesion typical of the BCR/ABL1-like subset• JAK/STAT mutations often concurrent with CRLF2 overexpression• CRLF2 overexpression detected in 69% of BCR/ABL1-like cases, though not exclusively• ABL-class rearrangements detected in 18% of BCR/ABL1-like ALL• TK-r cases often do not express high levels of CRLF2 Chiaretti S et al, BJH 2018

• Molecular background

• Incidence

• Diagnosis

• Outcome and MRD

• Treatment

Topics: BCR/ABL1-like

Outcome in childrenAssociated with male gender, hyperleucocytosis and increased MRD levels, also when patients are considered as standard-risk

Roberts KG et al, JCO 2014

Intensive and MRD-driven treatment seems to abolish the negative prognosticimpact of the BCR/ABL1-like signature in childhood ALL

Response to induction therapy in adults

• Contrasting results on the CR rate

- Lower than in other B-neg ALL cases (den Boer J et al, Haematologica 2015; Chiaretti S et al, BJH in press)

-No differences with other ALL cases (Herold T et al, Haematologica 2016; Jain N et al, Blood 2017)

-MRD persistence more frequent in BCR/ABL1-like cases (Roberts KG et al, JCO 2016; Herold T et al Haematologica 2017; Jain N etal Blood 2017; personal data). Not yet clear if in adults MRD negativity overcomes BCR/ABL1-like signature

• Significantly inferior survival (EFS, DFS, OS) in all reported studies

Survival in adults

Roberts KG et al, JCO 2016

• Molecular background

• Incidence

• Diagnosis

• Outcome and MRD

• Treatment: two options

Topics: BCR/ABL1-like

Treating the target

Pui CH et al, Clin Lymp Myel Leuk, 2017

Requires a deep knowledge of the genomic background of each case. Time and cost consuming. Feasible only in a few centers.

9 R/R pts have been treated. Median age 24 yrs (range 18-62).8 pts treated on the ruxolitinib arm (7 pts CRLF2-high, 1 with a JAK2 fusion (HMBOX1-JAK2).1 pt on the dasatinib arm (NUP214/ABL1).

No DLT, but no reponse on ruxo or dasa.

Jain N et al, ASH 2017

In vitro use of ponatinib on primary cells: effect on

proliferation and apoptoticresponse similar in

BCR/ABL1+ and BCR/ABL1-like cases (2 EBF1/PDGFRB-

positive, 1 JAK2-mutated and P2RY8/CRLF2-positive, 1

RCSD1/ABL1, 3 WT for JAK/STAT and RAS

mutations)

Wide-spectrum appraoch. Ponatinib

Chiaretti S et al, BJH 2018

p=0.0007

p=0.023

Conclusions

BCR-ABL1-like ALL represents a novel ALL subtype.

- Detected only in B-neg ALL cases and more frequently from adolescence onwards.

- The genomic background is highly heterogeneous and variable from case to case. It can be summarized in lesions involving ABL class genes, JAK/STAT genes and other kinases. Some cases are devoid of lesions.

- Diagnosis is still not standardized and represents an unmeet need.

- Therapy should include a TKI, possibly in combination with steroids and chemotherapy.

Upfront? At MRD persistence (different strategies in children and adults). In combination with immunotherapy?

Monica Messina Alessia LaurettiAlfonso PiciocchiNadia PeragineGiorgio InghiramiOliver Elemento Antonella VitaleAnna GuariniRobin Foà

Acknowledgments