LPL is the Strongest Prognostic Factor in a Comparative Study of RNA-based

Markers in Chronic Lymphocytic Leukemia

Mohd Arifin Kaderi1a, Meena Kanduri1a, Mahmoud Mansouri1a, Anne Mette Buhl2, Marie
Sevov1, Nicola Cahill1, Rebeqa Gunnarsson3, Mattias Jansson1, Karin Ekström Smedby4,
Henrik Hjalgrim5, Gunnar Juliusson3, Jesper Jurlander2, Richard Rosenquist1
1
Department of Genetics and Pathology, Uppsala University, Uppsala, Sweden,
2
Department of Hematology, Leukemia Laboratory, Rigshospitalet, Copenhagen,
Denmark, 3Department of Laboratory Medicine, Stem Cell Center, Hematology and
Transplantation, Lund University, Lund, Sweden, 4Department of Medicine, Clinical
Epidemiology Unit, Karolinska Institutet, Stockholm, Sweden, 5Department of
Epidemiology Research, Statens Serum Institut, Copenhagen, Denmark.
a
These authors contributed equally as first authors

Introduction: Chronic lymphocytic leukemia (CLL) is a heterogeneous disease with

varying clinical outcome, where many patients have an indolent course for many years,
whereas others show a more aggressive disease despite treatment. This has prompted the
search for biomarkers that can predict outcome in this disease. Recent studies have
proposed the RNA expression levels of certain genes, i.e. LPL, CLLU1, TCL1, MCL1 and
ZAP70 to be novel predictors of clinical outcome in CLL. However, a comprehensive
assessment of these RNA-based markers is still lacking. The current study aimed to
investigate the potential of these markers in CLL prognostication, either as single
markers or in combination with established markers. Patients and Methods: By applying
real-time quantitative PCR, we measured the RNA expression levels of LPL, CLLU1,
TCL1, MCL1 and ZAP70 in 256 newly diagnosed CLL samples from a Scandinavian
population-based cohort collected from 1999 to 2002 (median follow-up, 89 months) and
correlated with clinical outcome. The expression cut-offs for each RNA marker was
determined by constructing ROC curves. Additionally, Binet stage, IGHV mutation
status, CD38 expression (cut-off 7%) and the presence of recurrent genomic aberrations
(i.e. 11q-, 17p-, 13q- and +12) were evaluated for all cases. Results: High expression of
all RNA-based markers except MCL1 predicted significantly shorter overall survival
(OS) and time to treatment (TTT), with LPL being the most significant prognostic marker
in both log-rank and Cox univariate regression analyses. In multivariate analysis
including the RNA markers, LPL expression was the only independent prognostic factor
for OS and TTT (Table 1). When including all established markers, LPL lost its
significance in the model, due to its close association to the IGHV mutation status. Once
the mutation status was excluded from the analysis, LPL regained its prognostic power in
addition to genomic aberrations and CD38. Interestingly, all of the RNA-based markers
added further prognostic information to established markers in subgroups of patients,
with LPL expression status giving the most significant results. Notably, high LPL
expression predicted a worse outcome in favorable prognostic subgroups such as patients
with Binet stage A, CD38 negativity or favorable genomic aberrations (Table 2).
Conclusions: Altogether, we conclude that LPL expression appear to be the strongest
among the RNA-based markers for prediction of clinical outcome in CLL and thus could
potentially be applied in the clinical laboratory to predict outcome, particularly in
combination with established markers.

Table 1. Multivariate Cox-regression analysis of RNA-based markers.

Variable Overall survival (N=248) Time to Treatment (N=216)

HR 95% CI p HR 95% CI P
TCL1 1.3 0.86 – 1.1 0.79 –
0.180 0.415
8 2.19 8 1.75
LPL 4.6 2.65 – 3.6 2.32 –
<0.00001 <0.00001
3 8.09 4 5.71
ZAP70 1.3 0.82 – 1.2 0.84 –
0.253 0.282
2 2.11 5 1.86
CLLU1 0.8 0.51 – 1.1 0.77 –
0.443 0.473
3 1.33 6 1.74
MCL1 1.2 0.82 – 0.9 0.62 –
0.276 0.652
9 2.04 1 1.34

The threshold values used in the analysis were determined based on ROC curve analysis.
HR: Hazard ratio, CI: Confidence interval.

Table 2. The prognostic information of RNA-based markers in subgroups of established

markers

Log-rank p values
Variable LPL ZAP70 CLLU1 TCL1 MCL1
OS TTT OS TTT OS TTT OS TTT OS TTT

IGHV mutation status

Mutated NS NS NS NS NS NS NS 0.007 NS NS
Unmutated 0.01 NSa NS NS 0.03* NS NS NS NS NS

Binet Stage
A 0.00000 0.00000 0.005 NS NS 0.02 0.007 0.0001 NS NS
B/C NS NA NS NA NS NA 0.05 NA 0.03 NA

Genomic aberrations
del(13q)/No aberration 0.00000 0.00002 NS NS NS NS 0.04 0.007 NS NS
Trisomy 12/ del(11q) NS NS 0.04 NS NS NS NS NS NS NS
del(17p) NA NA NA NA NA NA NA NA NA NA

CD38
< 7% 0.00009 0.00004 NS NS NS NS NS 0.02 NS NS
≥ 7% 0.0009 0.005 0.02 NS NS NS NS NS NS NS

a
only 3 cases with low LPL expression in this subgroup
* high CLLU1 had longer OS accoding to Kaplan-Meier curve
NS not significant
NA not available; reliable log-rank test could not be performed due to very low number of cases