GENE EXPRESSION PROFILES AS PROGNOSTIC
FACTORS FOR HIGH-DOSE THERAPY AND BORTEZOMIB
IN PATIENTS WITH MULTIPLE MYELOMA
MULTIPLE MYELOMA
Standard therapy: high dose melphalan supported with
autologous peripheral blood stem cells (PBSC) transplantation
Harousseau JL et al., Blood Rev.
Rev 2002
Attal M et al, N Engl J Med. 2003
Lee CK et al., Bone Marrow Transplant. 2002
Barlogie B et al., Blood. 2004
Treatment choice depends on the extent of the disease,
t
symp oms, age, comorbidit
bidity
d
an t l
o erance
Significant inter-subject variations in outcome
Segeren CM et al., Br J Haematol. 1999
Lokhorst HM et al., J Clin Oncol. 2003
Segeren CM et al., Blood. 2003
van Agthoven M et al., Eur J Cancer. 2004
Gt
Gene itic factors d t
e
i
erm i
n ng response are unknown
GOAL
To identify gene expression profiles,
relevant for treatment outcome and toxicity
To investigate if there is heterogeneity
between patient subgroups with respect to
gene profiles
Myeloma plasma
plasma cells
cells
To study differences between Bortezomib
based therapy and standard treatment
Hovon 65/GMMG HD4 study
VAD
PAD
vs.
Randomization
Iv push
Bortezomib d 1,4,8,11
CAD
Stem cell collection
Cyclo/adria/dex
1-2 HDM
Auto PBSCT
200mg/m2
200mg/m
HLA-id Sib
Nl
Nonmyel S
o T
ST
Thalidomide
Bt
Bortezomib
ib
vs.
Maintenance/
200cGy
50 mg/daily
vs.
2x /month
Consolidation
MATERIALS & METHODS
152 PC
l
samp es
RNA isolation
FACS
F
analys
analy is
s
2 cycle biotin labeling,
·CD138
·Annexine
hybridization to U133 2.0 arrays
·7AAD
Quality controls (SF, % genes present)
137 samples for correlation in Omniviz
Visualization in MADex*
Clustering, Significance Analysis of
of
U 133 2.0 array
Microarrays (SAM)
*Valk et al. NEJM, 2004
Omniviz Correlation View
Pearson's correlation
correlation coefficient
coef
A
B
C
A
A
A
D
E
F
A
A
A
Subgroup formation based on correlated gene-
expression profiles
MADex,
e, visua
su li
a sa
s tio
aon too
ool for co
correlatio
aon in Omniviz*
1
2
3
4
5
6
7
*Valk et al. NEJM, 2004
7 clusters and known genes upregulated by IgH
translocations, plotted in MADex
MMSET
FGFR3
CCND2
MAF
MAFB
CCND1
CCND3
1
2
3
4
5
6
7
Gene Clustering in Spotfire in cluster division as
determined in MADex
F
FB
MMSET
MA
MA
FGFR3
CCND2
CCND1
CCND3
1
2
3
4
5
6
7
Down
Up
Up: genes involved in:
protein
protein biosynthesis (40%) ribosome biogenesis, odd
1
chromosomes*
Down: genes involved in transcription and DNA repair
2
Up: Genes mapping to 1p, INHB, ASS1, KLHL4
(CD 1 group)**
3
Up: CCND1, VPREB1, MS4A1 (CD20), (CD 2 group)**
Down: Il6R
Up: protein biosynthesis
biosynthesis, transport and transcription,
4
genes mapping to odd chromosomes.
Down: protein binding, proteolysis, genes mapping to 1q
5
Up: HRK, NOTCH2, DUSP16 (inactivity of MAPK)
6
Up: MAF downstream targets: ARK5, NTRK2, ARID5A,
bone resorption (P2RX7) and osteoclastogenesis (TLR4)
FGFR3, MMSET cluster, FZD2, FZD8, MITF involved in
WNT signaling. DSG2( cadherin family), PBX1 (B-cell
7
oncogene)
* Chng et al. Cancer Res, 2007
** Zhan et al. Blood, 2006
Upregulated genes in Bortezomib responders
Bortezomib responders:
At
Apoptosis
Angiogenesis
NF
NF B pathway, osteoclastogenesis
Protein biosynthesis, transport
False Discovery Rate 13%
Response measured ft
a er 3 cycles PAD
PA (
43)
n=
CONCLUSIONS
A division of myeloma patients into subgroups with correlated
behavior based on their gene-expression profiles
SAM analysis could confirm 3 subgroups (MMSET/FGFR3,
MAF/MAFB, and a group with high CCND1 expression) as
described
f
be ore
A number of
of genes
genes was associated with Bortezomib response.
Response data of more patients and at more time points will
follow soon to validate genes associated with response to
Bt
Bortezomib
ib
FUTURE DIRECTIONS
Cytogenetic/clinical data to confirm gene signature specific for the
different clusters
Validate the gene expression profiles in Bortezomib (non)responders
Confirm cluster specific genes by RT-PCR
Build classifiers of cluster specific genes to be used as class/cluster
predictor using
pg Prediction analysis for micro-arrays (
y(PAM)
Expand the analysis by correlating gene expression profiles with
clinical data:
Response data
data at
at different
different stages of treatment, TTP,
TTP OS
OS
Myeloma related complications: bone lesions
Adverse events: neuropathy, thrombocytopenia
Acknowledgements
Erasmus Medical Center
University of Heidelberg, Germany
Yvonne de Knegt
Dirk Hose
Martijn Schoester
Uta Mazitchek
Sophie Corthals
Hartmut Goldschmidt
Goldschmidt
Bronno van der Holt
Laila El Jahiri
Berna Beverloo
Skyline Diagnostics BV
Roel Ve
V rhaak
e
Henk Vietor
Matthijs Sanders
Peter Valk
Bas Wouters
Supported by:
Bas Wouters
Mark van Duin
EMCR Translational Research Grant
Pieter Sonneveld
EHA Clinical Research Grant 2006
Jan
Ja ssen Orthob
obiotech
Utrecht University
University Medical Center
Henk Lokhorst