ZAP-70, LPL, and TCF7 Best Markers for Mutational Status

[Guest guest]

QUANTITATIVE GENE EXPRESSION ANALYSIS OF SURROGATE MARKERS FOR

GENETIC RISK GROUPS AND SURVIVAL IN CLL

0909

Author Kienle, Dirk , Ulm, University of Ulm Germany

Co-author(s) Benner, A. , Heidelberg, Central Unit Biostatistics,

DKFZ

Schneider, Christof , Ulm, University of Ulm

Winkler, Dirk , Ulm, University of Ulm

Habermann, Annett , Ulm, University of Ulm

Hensel, Manfred , Heidelberg, University of Heidelberg

Lichter, P. , Heidelberg, Department of Molecular Genetics, DKFZ

Dalla-Favera, Riccardo , New York, Columbia University

Döhner, H. , Ulm, University of Ulm

Stilgenbauer, S. , Ulm, University of Ulm

Background: The genetic factors VH mutation status, V3-21 gene

usage, and genomic deletions at 11q22-q23 and 17p13 have been shown

to be important prognostic markers in CLL. Given the high complexity

of these analyses in the recent years several molecular surrogate

markers were developed aiming at the facilitation of routine

prognosis assessment.

Aims: To assess the value of potential surrogate markers for the

prediction of genetic risk groups and survival.

Methods: Real-time RT-PCR (RQ-PCR) of candidate genes was performed

in a CD19-purified and a non-purified CLL cohort each comprising the

relevant genetic subgroups (VH mutated, VH unmutated, V3-21 usage,

11q-, 17p-). 17 markers (ADAM29, ATM, CLLU1, DMD, GLO1, HS1,

KIAA0977, LPL, MGC9913, PCDH9, PEG10, SEPT10, TCF7, TP53, Vimentin,

ZAP-70, ZNF2) identified in previous studies were investigated in

the non-purified cohort of 102 CLL patients. Of these, 10 markers,

either with an overexpression in non-CLL cells or an impact on

survival or risk group prediction, were analyzed in the purified

cohort of 112 cases. VH sequencing and FISH screening for genomic

aberrations were carried out for all cases. Survival information was

available for 80 (purified) and 88 cases (non-purified). Logistic

regression was performed to test the predictive value of gene

expression for genetic risk groups, proportional hazards

statistics for survival analysis.

Results: The genetic risk groups in both cohorts showed the expected

correlation with survival with significantly shorter survival of VH

unmutated, 17p-, and 11q- cases indicating a representative

composition of the cohorts under study.

In non-purified cases, the best predictive marker for VH status was

LPL (p=0.001). While no reliable predictive markers were identified

for V3-21 usage or17p-,lower ATM expression was predictive for 11q-.

In survival analysis including all candidate genes TCF7 (p=0.001)

and KIA0977 (p=0.016) were of prognostic value. In multivariate

survival analysis including candidate gene expression and the

genetic risk factors as variables, only 17p- remained as a

significant parameter.

In the purified cohort, significant markers (p<0.05) for genetic

risk groups were: ZAP70, LPL, and TCF7 for VH mutation status (TCF7

expression associated with mutated VH); SEPT10, ZAP70, and ADAM29

for V3-21 usage; ATM and TCF7 for 11q- (both with a negative

association); ZNF2 for 17p- (negative association). In multivariate

analysis, the parameters 17p-, 11q-, V3-21 usage, TCF7, and ZAP70

expression were identified as independent prognostic factors. In

contrast to ZAP70, TCF7 expression was positively correlated with

survival times.

Conclusions: Several results obtained in CD19-purified cases could

not be reproduced in unpurified cases strongly arguing for a tumor

cell selection prior to expression analysis. In purified cases,

ZAP70, LPL, and TCF7 were the best predictors for VH mutation

status.

Additional markers such as ATM and ZNF2 may help to identify genomic

risk groups such as 11q- and 17p-. Multivariate survival analysis

suggests TCF7 as a strong survival predictor and points to a

pathogenic role for this gene in CLL. The patient series have been

extended to 115 non-purified and 150 purified cases, statistical

evaluation is currently ongoing.