Transcript Diagnostic and Prognostic Significance of Gene Expression

Diagnostically and Prognostically
Significant Genetic Alterations in
Diffuse Large B-Cell Lymphoma
Friederike Kreisel, MD
Department of Pathology and Immunology
Washington University in St. Louis
Frequency of B-and T-Cell
Lymphomas
Diffuse Large B-cell
Lymphoma
Follicular Lymphoma
8.6%
11.7%
2.5%
33%
6.0%
Mantle Cell
Lymphoma
Burkitt Lymphoma
6.7%
9.4%
Marginal Zone B-cell
Lymphoma
CLL/SLL
22.1%
T-cell Lymphomas
Other Types
Incidence of non-Hodgkin lymphoma (NHL) has increased more than 73% between
1973 and 1991
Estimated rate for DLBCL is ~4.68 cases per 100,000 persons/year
Normal B-Cell Differentiation
NEJM 2000, 343;108-117
The Germinal Center
Lymph node
Germinal Center
CD10+/BCL-6+/BCL-2-
NEJM 2000, 343;108-117
Diffuse Large B-Cell Lymphoma
• Usually de novo
(primary)
• Transformation from:
– CLL/SLL
– Follicular lymphoma
• Immunodeficiency is
strongest known risk
factor
WHO 2008, Tumors of
Hematopoietic and
Lymphoid Tissues
Diffuse Large B-Cell Lymphoma
• Clinically,
morphologically, and
genetically
heterogenous group
with only a subset
falling into the WHO
subcategories
• 20-30% of DLBCL
continue to be defined
by their nuclear size
only
DLBCL Subcategories (WHO 2008)
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Diffuse large B-cell lymphoma, NOS
T-cell/histiocyte rich DLBCL
Primary DLBCL of CNS
Primary cutaneous DLBCL, leg type
EBV positive DLBCL of the elderly
Primary mediastinal (thymic) LBCL
Intravascular LBCL
DLBCL associated with chronic inflammation
Lymphomatoid granulomatosis
ALK-positive LBCL
Plasmablastic lymphoma
LBCL arising in HHV8+ multicentric Castleman’s disease
Primary effusion lymphoma
B-cell lymphoma, unclassifiable
International Prognostic Index (IPI) most
commonly used to predict outcome in DLBCL
• Unfavorable variables
– Age >60 years
– Poor performance status (ECOG2)
– Advanced Ann Arbor stage (III-IV)
– Extranodal involvement 2 sites
– High serum LDH (>normal)
Risk group
Low
Low/intermediate
High/intermediate
High
Unfavorable variables
All patients
Patients60
0 or 1
0
2
3
4 or 5
1
2
3
Major Recurring Genetic Events in DLBCL
Genetic Defect
Frequency
Location
BCL6
35-40%
3q27
BCL2
18q21
cMYC
t(14;18) 13%,
amplification 24%
15%
FAS(CD95)
20%
10q24
Aberrant SHM
45%
Genes aberrantly
targeted: BCL6,
cMYC, PAX5
TP53
16%
17p
8q24
BCL6
• Zinc-finger transcription
repressor normally
expressed within germinal
center (GC) B-cells 
BCL6 null animals fail to
generate GCs in response
to antigen
• Constitutive expression of
BCL6 might  the p53mediated apoptosis,
promoting persistence of
malignant clones
Wikipedia
BCL2
• Proto-oncogene located at
18q21 that promotes B-cell
survival via inhibition of
apoptosis and confers
chemotherapy resistance
• BCL-2 expression is
normally down-regulated
in the GC where apoptosis
plays a critical role in
negative B-cell selection
• t(14;18)  fusion gene
leading to transcription of
 levels of BCL2
Wikipedia
Gene Expression Profiling
• Measurement of the activity (expression)
of thousands of genes at once  relative
amount of mRNA expressed  provides a
global picture of cellular function
Gene Expression Profiling
• Highlight similarities between subsets of
DLBCL and normal B cells
• Define robust and highly reproducible
DLBCL subtypes with comprehensive
transcriptional signatures
• Identify features associated with
unfavorable responses to empiric
combination chemotherapy
DNA Microarray - Methodology
DLBCL1
–
DLBCL2
–
–
+
+
–
+
+
+
LYMPHOCHIP
Slides are coated with polyLysine, which is positively charged.
DNA is negatively charged, so it
“sticks” to the slide through ionic interaction
LYMPHOCHIP
cDNA is made from different DLBCL
tumors
DNA Microarray - Methodology
DLBCL1
DLBCL1
DLBCL2
LYMPHOCHIP
Both sets of DLBCL cDNA are labelled
with different fluorescent tags, in this
case red and green
DLBCL2
LYMPHOCHIP
Gene array lymphochip is incubated
with the tagged DLBCL cDNAs, which
bind to the matching genes printed
on the array
DNA Microarray - Methodology
DLBCL1
DLBCL2
LYMPHOCHIP
Since positions of the genes on the DNA array are known,
levels of gene expression can be figured out based on color signal.
If the gene is only expressed in DLBCL1, the square is red; if the gene
is only expressed in DLBCL2, the square is green; if the gene is
equally expressed in both DLBCL, the square is yellow.
Gene Expression Profiling in DLBCL
(Alizadeh et al., Nature 2000)
Ratio of
hybridization of fluorescent
experimental mRNA to
fluorescent pooled
reference mRNA
Gene Expression Profiling in DLBCL
(Alizadeh et al., Nature 2000)
a.
b.
c.
Hierarchical clustering of DLBCL
(orange and blue) and germinal center
B cells (black) based on the genes of
the germinal center B-cell gene
expression signature
Discovery of genes that are selectively
expressed in GC B-like DLBCL and
activated B-like DLBCL based on
genes of pan B-cells, germinal center
B-cells, and activated B-cells
Hierarchical clustering of the genes
selectively expressed in GC B-like
DLBCL and activated B-like DLBCL,
which was determined from Fig 3b
Differences Between GC-like and
Activated B-like DLBCL
GC B-like DLBCL
Activated B-like
DLBCL
Cell of origin
Germinal center B-cell
Postgerminal center Bcell
Chromosomal
alterations
t(14;18); gains of 12q
Gains of 3q and
18q21-22; deletion of
6q21-q22, BCL6
translocations
Oncogenic
mechanisms
BCL2 translocation;
amplification of cREL locus
Constitutive NFkB
activation
Ongoing Ig
mutation
Yes
No
IHC
Expression of CD10, BCL-6
Expression of MUM1
Clinical outcome
Favorable (60% 5-year
survival)
Poor (35% 5-year
survival)
Overall survival of DLBCL patients grouped
on the basis of gene profiling shows worse
outcome for patients with “activated B-like”
subtype
A.A. Alizadeh et al. Nature 2000; 403:503-511
Immunohistochemistry (IHC) can be used to
determine the GC B-like and Activated B-like
subtypes of DLBCL
Normal
Lymph Node
CD10
BCL-6
GC B-like DLBCL
Activated B-like/non-GC
B-like DLBCL
C.P. Hans et al. Blood 2004, 103:275-282
Overall survival curves comparing immunohistochemical
and genetic classification of GC B-like and activated B-like
DLBCL are worse for the activated B-like subtype
C.P. Hans et al. Blood 2004, 103:275-282
Decision Tree for IHC Classification
of DLBCL
C.P. Hans et al. Blood 2004, 103:275-282
• Sensitivity of IHC 71% for
the GC B-like group and
88% for the activated B-like
/ non GC B-like group
• PPV of IHC was 87% for
the GC B-like group and
73% for the activated B-like
/ non GC B-like group
• 30/142 (21%) of cases had
discordant IHC and cDNA
microarray results
Does the type of therapy for DLBCL
affect prognosis in patients?
• Drugs in CHOP:
–
–
–
–
C = Cyclophosphamide
H = Doxorubicin (Hydroxydaunomycin)
O = Vincristine Sulfate (Oncovin)
P = Predisone
• Drugs in R-CHOP (since 2000):
–
–
–
–
–
R = Rituximab
C = Cyclophosphamide
H = Doxorubicin (Hydroxydaunomycin)
O = Vincristine Sulfate (Oncovin)
P = Predisone
Rituximab
• CD20 antigen
Human IgG1
constant region
B-cell CD20
Human k constant
region
Murine antigen binding domain
– Expressed only on B-cells
– Present in >90% of B-cell
lymphomas
– Does not shed off cell
surface
– Important for cell cycle
initiation and
differentiation
Effects of Rituximab
• Mediates antibody-dependent cellular
cytotoxicity and complement-dependent
cytotoxicity
• Elicits shedding of CD23
• Down-regulates the B-cell receptor
• Induces apoptosis of CD20+ cells
• Half-life of 30-400 hours (varies by dose
and length of treatment)
OS Curves of Patients with GC B-Like
and Activated B-like DLBCL by IHC
After CHOP or R-CHOP Treatment
R. Seki et al. Cancer Sci 2009, 100:1842-47
The overall survival rates for control and
immunochemotherapy-treated DLBCL patients
according to immunohistochemical and clinical
factors
CHOP
R-CHOP
R-CHOP
H. Nyman et al. Blood 2007;109:4930-4935
The Hans Classifier – Controversies in
the CHOP and R-CHOP Era
Hans et al. Blood 2004
Berglund et al. Mod Path 2005
Haarer et al. Arch Pathol 2006
Muris et al. J Pathol 2006
Survival association present
Sjo et al. Eur J Hematol 2007
Van Imhoff et al. J Clin Oncol 2007
Nyman et al. Blood 2007
Amara et al. Mod Pathol 2008
Fu et al. JCO 2008 (R-CHOP)
Colomo et al. Blood 2003
Nyman et al. Blood 2007 (R-CHOP)
De Paepe et al. JCO 2005
Dupuis et al. Haematologica 2007
Survival association absent
Natkunam et al. JCO 2008
Veelken et al. Ann Oncol 2007
Amen et al. Histopathology 2007
Wilson et al. JCO 2008 (DE-EPOCH-R)
Ott et al. Leukemia Research 2012 (R-CHOP)
Transcriptional Profile of DLBCL
Subsequently Treated with R-CHOP
J.P. Jais et al. Leukemia 2008;22:1917-24
The recognition of GCB and ABC DLBCL is highly
reproducible by gene expression profiling between
different centers
Center A
Center B
GCB
Unclass
ABC
GCB
53
3
0
Unclass
3
16
2
ABC
0
3
48
Concordant results in 117/128 (91%)
Unpublished data from Elias Campo, MD
Prognostic Impact of Gene Alterations
in DLBCL in the Era of Rituximab
• MYC gene rearrangements
• CDKN2A (p16INK4/p14ARF) and
CDKN2B (p15INK4) deletions
• p53 aberrations
cMYC
• Located on 8q24
• Regulates gene expression
through binding on Enhancer
Box sequences and recruiting
histone acetyl-transferases
• Burkitt lymphoma t(8;14) cMYC gene next to the
immunoglobulin heavy chain
locus  fusion gene causing
overexpression of MYC protooncogene in lymphoma cells
Wikipedia
Outcomes of Patients with MYC+ or
MYC- DLBCL
• 135 cases of DLBCL
analyzed
• 12/134 cases (8.8%)
were positive for MYC
rearrangement
• All patients were
treated with R-CHOP
K.J. Savage et al. Blood 2009; 114:3533-37
CDKN2A/B (p16,p14,p15)
• Located on 9p21.3
• Tumor suppressor gene that
regulates cell cycle via the
retinoblastoma and p53
apoptosis pathways
• Defective p53/CDKN2A/B
signaling can lead to
apoptosis resistance in
DLBCL
Kreisel et al. Cancer Genetics 2010;204:129-37
TP53
• Located on 17p13.1
• Tumor suppressor gene
• “The guardian of the
genome”
• p53 is in reference to its
molecular weight (53-kDa
protein)
• 50% of human tumors
contain a mutation or
deletion of the TP53 gene
Wikipedia
OS according to the TP53 and CDKN2A allelic status in DLBCL. (A) Overall survival (OS)
probability in the entire lymphoma population (n = 114). (B) OS probability in the R-CHOP
treated subgroup (n=78)
Jardin F et al. Blood 2010;116:1092-1104
Schematic representation of the 9p21 locus and its genomic alterations detected by a
dedicated quantitative multiplex PCR of short fragments (QMPSF) assay in DLBCL.
CDKN2A/B
p16 p14 p15
Jardin F et al. Blood 2010;116:1092-1104
Gray: heterozygous deletion
Black: homozygous deletion
Summary
• DLBCL represents a morphologically,
biologically, and clinically heterogenous group of
tumors
• ~20-30% of DLBCL cannot be subcategorized
into a current WHO subclassification
• Despite DLBCL subcategories clinical outcome
within a specific category is heterogenous
• Genetic alterations of BCL-6, BCL-2, and c-MYC
are among the most common recurrent genetic
abnormalities
Summary
• Gene expression profiling of DLBCL
offered
– Comparison of gene signatures between
subsets of DLBCL and normal B cells
– Delineation of robust and highly reproducible
DLBCL subtypes with comprehensive
transcriptional signatures (GC-like vs.
activated-like DLBCL)
– Identification of gene expression signatures
associated with unfavorable response to
empiric combination chemotherapy
Summary
• Immunohistochemical sub-categorization
into GC-like vs. non GC B-type in
Rituximab era remains controversial,
although robust with cDNA microarray
• Additionally, c-MYC, TP53, and
CDKN2A/B aberrations appear to have
prognostic significance in Rituximab era
Journal Club
4/10/12
Aim of study
• Show that the variability in natural history of
DLBCL reflects unrecognized molecular
heterogeneity in the tumors
• Measure the activity (expression) of
thousands of genes at once (DNA
microarrays)
• Identify molecularly distinct subgroups of
DLBCL
• Correlate different subgroups with clinical
outcome
Construction of DNA microarray
• “Lymphochip”: selection of genes that are
preferentially expressed in lymphoid cells and
genes with known or suspected roles in
immunology or cancer
– 12,069 out of 17,856 cDNA from a germinal center Bcell library
– 2,338 cDNA from libraries derived from DLBCL,
follicular lymphoma, mantle cell lymphoma, chronic
lymphocytic leukemia
– cDNA from genes that are induced or repressed
during B- and T-cell-lymphocyte activation by
mitogens or cytokines
– 3,186 genes of importance to lymphocyte and/or
cancer biology
cDNA Microarray - Methodology
Experimental
sample
Experimental
sample
Reference pool
of 9 different
lymphoma cell
lines
17,856-GENE ARRAY
LYMPHOCHIP
Reference pool
of 9 different
lymphoma cell
lines
17,856-GENE ARRAY
LYMPHOCHIP
The fluorescence ratio was quantified
for each gene and reflected the relative
abundance of the gene in each
experimental mRNA sample compared
to the reference pool
Each row: separate cDNA
clone on the microarray
Each column: separate
experimental mRNA
sample
Results: ratio of
hybridization of fluorescent
experimental mRNA to
fluorescent pooled
reference mRNA
Figure 1
dendrogram
Figure 2
“Proliferation signature”, a
quantitative measurement of tumor
cell proliferation, confers an inferior
survival
“Germinal center B-cell signature”
with BCL-6 and CD10 expression
“Lymph node survival
signature”  host response to
lymphoma, is associated with 
survival times
Figure 3
Figure 4
Figure 5