Proteomic Analysis of Waldenstrom Macroglobulinemia
Steven P. Treon,
Thomas E. Witzig,
1Alexey A. Leontovich,
2Morie A. Gertz,
1Anne Sophie Moreau,
2Kenneth C. Anderson,
2and Irene M. Ghobrial
1Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts;2Division of Hematology,
Department of Internal Medicine, Mayo Clinic College of Medicine, Rochester, Minnesota;
Healthcare System; and
Institute, University of Pittsburgh, Pittsburgh, Pennsylvania
3Department of Pathology, VA Pittsburgh
4Division of Hematology and Oncology, Department of Internal Medicine, University of Pittsburgh Cancer
To better understand the molecular changes that occur in
Waldenstrom macroglobulinemia (WM), we employed anti-
body-based protein microarrays to compare patterns of
protein expression between untreated WM and normal bone
marrow controls. Protein expression was defined as a >2-fold
or 1.3-fold change in at least 67% of the tumor samples.
Proteins up-regulated by >2-fold included Ras family proteins,
such as Rab-4 and p62DOK, and Rho family proteins, such
as CDC42GAP and ROKA. Other proteins up-regulated by
>1.3-fold included cyclin-dependent kinases, apoptosis regu-
lators, and histone deacetylases (HDAC). We then compared
the samples of patients with symptomatic and asymptomatic
WM and showed similar protein expression signatures,
indicating that the dysregulation of signaling pathways occurs
early in the disease course. Three proteins were different by
>2-fold in symptomatic versus asymptomatic, including the
heat shock protein HSP90. Elevated protein expression was
confirmed by immunohistochemistry and immunoblotting.
Functional significance was validated by the induction of
apoptosis and inhibition of proliferation using specific HDAC
and HSP90 inhibitors. This study, therefore, identifies, for the
first time, multiple novel proteins that are dysregulated in
WM, which both enhance our understanding of disease
pathogenesis and represent targets of novel therapeutics.
[Cancer Res 2007;67(8):3777–84]
Waldenstrom macroglobulinemia (WM) is a low-grade lympho-
proliferative disorder characterized by the presence of a lympho-
cytic, plasma cell and lymphoplasmacytic infiltrate in the bone
marrow and the presence of a serum monoclonal protein
immunoglobulin M (IgM; refs. 1–3). To date, WM remains
incurable, with a median survival of 5 years (4–6). The current
therapeutic modalities available include alkylator agents, purine
nucleoside analogues, and rituximab. The main risk factor for the
development of WM is preexisting IgM-monoclonal gammopathy
of undetermined significance (MGUS; refs. 7, 8).
Comprehensive investigation of the underlying molecular
alterations in WM has been made possible with the introduction
of gene expression profiling. A recent study of gene expression
profiling of 23 cases of WM compared with cases of multiple
myeloma and chronic lymphocytic leukemia (CLL) revealed that
WM had a homogenous expression profile on unsupervised
clustering analysis and had a similar expression profile to CLL
(9). A small set of genes had a unique expression profile in WM,
which included the mitogen-activated protein kinase (MAPK)
pathway and interleukin-6 (IL-6). However, gene expression
profiling determines the level of mRNA in the samples and
controls, and changes in mRNA levels do not always translate into
changes at the protein level (10).
The recent development of protein array techniques allows for a
comprehensive analysis of molecular changes at a functional
protein level. Conventional methods of proteins quantification,
such as immunoblotting, ELISA, and two-dimensional gel electro-
pheresis are not amenable to high-throughput applications.
The antibody-based protein microarray allows the screening of
multiple samples simultaneously and generation of signature
profiles (11–14). In this study, we defined the differential expression
of proteins in tumor samples from patients with WM versus normal
bone marrow lymphocytes and plasma cells and then validated in
functional assays their biological importance as potential targets of
novel therapeutics in WM.
Materials and Methods
Patient samples. This study was approved by the Mayo Foundation and
the Dana-Farber Cancer Center Institutional Review Boards and conducted
in accordance with the Declaration of Helsinki. All patients gave written
informed consent for their tissue to be used for research and for review of
their clinical records. For the protein array analysis, 10 frozen bone marrow
samples were obtained from patients with newly diagnosed WM between
August 1999 and February 2003. Another four samples were obtained from
patients with symptomatic WM, between September 2005 and April 2006,
for confirmation studies by immunoblotting. The diagnosis of WM was
based on the characteristic tumor cell morphology and immunophenotype,
as well as the demonstration of an IgM monoclonal protein by
immunoelectropheresis and immunofixation. Cells were obtained from
bone marrow aspirates using immunomagnetic bead selection (Miltenyi
Biotech, Auburn, CA) for concomitant isolation of CD19+and CD138+cells
as previously described (9). This technique led to the selection of malignant
cells with more than 85% homogeneity as confirmed by n and E
immunostaining of purified cells.
Normal CD19+lymphocytes and CD138+plasma cells were isolated in a
similar fashion from four bone marrows that were pooled and used as
nondiseased controls for all microarray experiments to provide a constant
reference point for arrays. These cells were obtained from bone marrows
because they represent the normal counterpart of the WM cells in the bone
marrow microenvironment context. Another three normal controls were
Requests for reprints: Irene M. Ghobrial, Dana-Farber Cancer Institute, 44 Binney
Street, Mayer 547, Boston, MA 02115. Phone: 617-632-4198; Fax: 617-632-4862; E-mail:
I2007 American Association for Cancer Research.
Cancer Res 2007; 67: (8). April 15, 2007
used for confirmation studies with immunoblotting and immunohisto-
chemistry. A total of 11 separate experiments with antibody-based protein
microarrays were done. All samples were compared with a pooled control
sample of CD19+and CD138+cells from four normal bone marrow controls.
One experiment was done with a normal control extract versus itself
to detect any nonspecific binding to the antibodies imprinted on the
microarray slides and to provide a normalized baseline ratio for comparison
of all WM samples to control.
To determine differences between symptomatic and asymptomatic WM,
we selected 10 cases of WM, 5 with symptomatic WM and 5 asymptomatic.
The asymptomatic WM/MGUS samples were selected by the presence of an
IgM monoclonal protein with an M-spike <1.0 g/dL, and the presence of
lymphoplasmacytic cells of <10% in the bone marrow and the absence of
symptoms related to WM. In addition, follow-up of these patients showed
no progression to symptomatic WM to the date these studies were done.
Protein microarray procedure. The Ab Microarray (BD Clontech, Palo
Alto, CA) detects a wide variety of proteins (both cytosolic and membrane
bound) representing a broad range of biological functions, including signal
transduction, cell-cycle regulation, gene transcription, and apoptosis. The
microarray contains 512 highly specific and sensitive monoclonal anti-
bodies (mAb) against human polypeptides. Because each antibody is
printed twice on each microarray slide for internal control of each sample,
each slide contains 1,024 antibodies.5
After CD19 and CD138 enrichment, cells (1–6 ? 106) were sedimented.
Total protein was extracted by a single freeze-thaw cycle of the cell pellet in
liquid nitrogen followed by homogenization in the manufacturer’s
extraction/labeling buffer. After measurement of the total protein
concentration by the bicinchoninic acid method (Pierce, Rockford, IL,
USA), each sample was diluted to a final total protein concentration of
1.1 mg/mL. For each sample and control, 90 AL of total protein was labeled
with 10 AL of labeling dyes, either Cy3 or Cy5 (Amersham Biosciences,
Piscataway, NJ). The dual-fluorescence detection method is designed so
that inherent variations in dye labeling do not affect the outcome of the
experiment. After 90 min of incubation and 30 min of blocking, the
unbound dye was removed using PD-desalting columns (Amersham
Biosciences). Protein concentration was again determined using the
bicinchoninic acid method with subtraction of the dye’s contribution to
the overall absorbance at 562 nm. The average number of dye molecules
covalently coupled to each protein was measured as per the user manual
and usually ranged from two to four molecules of Cy3 or Cy5 per molecule
Two slides, each imprinted with 512 mAbs, were provided for reverse
color labeling to allow normalization of the samples. The arrays are printed
on standard-size (75 ? 25 ? 1-mm) glass slides with 512 antibodies printed
in duplicate on each slide. Protein samples from normal B cells labeled with
Cy5 were mixed with protein samples from WM B cells labeled with Cy3
protein and added to slide 1. For slide 2, the protein samples from normal B
cells were labeled with Cy3, and the proteins from WM B cells labeled with
Cy5. Labeling the proteins from both the normal and malignant B cells with
Cy3 and Cy5 allows the microarray to detects differences in specific protein
abundance between the WM sample and the control sample with each
experiment. Total protein (20 Ag) was added to each slide and incubated at
room temperature for 30 min before a series of washes. The slides were
dried and scanned according to the instructions of the supplier using the
Axon GenePix 4000B scanner set to 635 nm (Cy5 channel), PMT 670 V,
power 33%, and 532 nm (Cy3 channel), PMT 550 V, Power 33% to produce a
text file with signal intensities. Two ratios were generated from the spot
images, WM-Cy5/normal-Cy3 (slide 1) and normal-Cy5/MCL-Cy3 (slide 2),
for each protein target.
Data analysis and clustering. The mean of the ratios of Cy5/Cy3 of
both slides were analyzed using Clontech Excel software developed
specifically for each microarray lot by the manufacturers. The two ratios
were used to calculate an internally normalized ratio (INR), or ratio of
ratios, for each spot on the array. This calculation normalizes for differences
due to labeling efficiency and antibody-antigen binding affinity, greatly
enhancing the precision and accuracy of the assay. The replicate values
within each slide were then averaged, and an INR was calculated where
INR =pratio 1/ratio 2 and ratios 1 and 2 correspond to slides 1 and 2. ratio
1 = normal-Cy5 relative fluorescent units/WM-Cy3 relative fluorescent units,
and ratio 2 = WM-Cy5 relative fluorescent units/normal-Cy3 relative
fluorescent units. The average INR was calculated for each antibody; and
duplicate INR values that varied by >30% were discarded. The Genespring
software was used for analysis of all 10 experiments and normalized to the
control versus control experiment. An unsupervised clustering analysis was
done; and changes that were 1.3- or 2-fold or higher in at least 6 of 10 (67%)
WM samples compared with the control were identified. These differences
were statistically significant at P = 0.05.
Cell lines and reagents. The WM cell lines (BCWM.1 and WSU-WM)
and IgM-secreting low-grade lymphoma cell lines (MEK1, RL) were used.
The BCWM.1 is a recently described WM cell line that has developed from a
patient with untreated IgMn WM (15). The cells express the typical
lymphoplasmacytic phenotype (15). WM-WSU was a kind gift from Dr.
AlKhatib (Wayne State University, Detroit, MI). MEK1 was a kind gift from
Dr. Neil Kay (Mayo Clinic Rochester, MN). In addition, we used a multiple
myeloma cell line, MM.1S, a kind gift from T. Hideshima (Dana Farber
Cancer Institute, Boston, MA) to determine the expression of proteins in
plasma cells. All cell lines were cultured in RPMI 1640 containing 10% fetal
bovine serum (Sigma Chemical, St. Louis, MO), 2 Amol/L L-glutamine, 100
units/mL penicillin, and 100 Ag/mL streptomycin (Life Technologies, Grand
Island, NY). The histone deacetylase (HDAC) inhibitor trichostatin A
(Upstate, New York, NY) and HSP90 inhibitor 17-AAG (A.G. Scientific Inc.,
San Diego, CA) were used to determine the functional effect of inhibition of
these proteins on survival of WM cell lines.
Immunohistochemistry. Paraffin-embedded bone marrow biopsies
available from the same patients with the same dates of sample acquisition
were analyzed by immunohistochemistry using monoclonal antibodies to
human proteins: cyclin-dependent kinase 2 (CDK2; LabVision Neomarker),
Rab4, HDAC3, RCC-1, fatty acid synthase (FAS), and HSP90 (BD
PharMingen, San Diego, CA). Anti-CD20 antibodies (Ventana Stainer,
Ventana Medical System, Tucson, AZ) were used to localize the areas of
lymphocyte aggregates. Another four normal controls were used to confirm
the expression level of proteins in the control samples. Images were
visualized under a Leica DM IL microscope (Wetzlar, Germany) using an
objective lens of 40/0.6 ? 2 (magnification, 80?) equipped with a camera
Leica dfc300 FX and software Leica IM50 version 6.2.1.
Immunoblotting. WM and IgM-secreting lymphoma cell lines, along
with primary CD19+or CD138+cells or concomitant CD19+/CD138+cells
from patients with WM (n = 10) and normal donors (n = 5), were harvested
and lysed. To determine that the changes in protein expression were not
due to different ratios of plasma cells and lymphocytes in the samples, we
did immunoblotting using primary WM cells (n = 3) sorted separately for
CD19 and for CD138 and compared them to normal controls (n = 2) that
were selected in a similar fashion. Similarly, we confirmed the expression of
these proteins using another three samples of WM cells selected for
concomitant CD19+/CD138+. Immunoblotting was done using lysis buffer
(Cell Signaling Technology, Beverly, MA) reconstituted with 5 mmol/L NaF,
2 mmol/L Na3VO4, 1 mmol/L phenylmethylsulfonyl fluoride, 5 Ag/mL
leupeptine, and 5 Ag/mL aprotinin. Whole-cell lysates were subjected to
SDS-PAGE and transferred to polyvinylidene diflouride membrane (Bio-Rad
Laboratories, Hercules, CA). The antibodies used for immunoblotting
included Rab4, p62DOK, ROK, CDC42, and HSP90 (Cell Signaling
Technology). Anti-actin antibody (Santa Cruz Biotechnology, Santa Cruz,
CA) was used for loading control.
Apoptosis assay. Apoptosis was detected by using annexin V/propidium
iodide (PI) staining. In brief, cells (1 ? 106) from 24- and 48-h cultures with
trichostatin or 17-AAG were washed with ice-cold PBS and resuspended in
binding buffer [10 mmol/L HEPES (pH, 7.4), 140 mmol/L NaCl, 2.5 mmol/L
CaCl2]. WM cells were incubated with annexin V-FITC (5 AL/mL; Caltag
Laboratories, Burlington, CA) for 15 min at 4jC. Analysis of the data was
5The complete list of the arrayed antibodies, including Swiss-Prot ID numbers of
the target antigens, is available at http://bioinfo2.clontech.com/abinfo/ab-list-
Cancer Res 2007; 67: (8). April 15, 2007
done using flow cytometry (Beckman Coulter Inc., Fullerton, CA) as
previously described (18). Similarly, WM cells from two patients were
cultured with serial concentrations of trichostatin or 17-AAG for 48 h and
analyzed for apoptosis using annexin/PI staining.
Baseline patient characteristics. The median age of the 10
patients of whom samples were used for protein microarray was
69.4 years (range, 55.8–78). There were six men and four women in
this cohort. At the time of sample collection, all patients had no
previous therapy. The median M-spike for the patients with
asymptomatic WM/MGUS was 0.7 mg/dL (range, 0–1.3) and 3.3
mg/dL (range, 1.6–5.9) for those with symptomatic WM (P = 0.009).
The percent bone marrow involvement with lymphoplasmacytic
malignant cells was 3% (range, <5–5%) for patients with
asymptomatic WM/MGUS and 30% (range, 10–60) for patients
with symptomatic WM.
Protein expression. Unsupervised clustering of the WM
samples showed a homogenous pattern of expression in all the
samples (Fig. 1A). We analyzed polypeptides that were up- or
down-regulated by >1.3-fold or >2-fold as compared with normal
control. The >1.3-fold expression cutoff was recommended by the
manufacturer as a significant value for up-regulated proteins. We
also used a cutoff of >2-fold because it has been the standard cutoff
used in the analysis of cDNA array data. The expression level of
each of the polypeptides dysregulated by >1.3-fold is shown in the
unsupervised clustering analysis in Fig. 1A, and the fold change in
the level of each of these proteins in all the samples as compared
with the control is shown in Fig. 1B. Using the >2-fold cutoff, the
microarrays identified six dysregulated polypeptides in at least
60% samples of WM (Table 1 and Fig. 1C). All polypeptides were
overexpressed in the WM cells as compared with control cells.
These polypeptides were signal transduction regulators, such as
Ras-related proteins, including Rab4 and p62DOK; Rho-related
proteins, including CDC42GAP, and ROKa; and other proteins, such
as SNX-1, Roaz, and FAS. Using the >1.3 cutoff, 105 polypeptides
were up-regulated, and 74 were down-regulated in at least 60%
samples of WM (Tables 2 and 3). These included polypeptides
involved in cell cycle regulation, such as CDK2 and RCC-1, histone
deacetylases such as HDAC3, and modulators of apoptosis, such as
the proteins in the phosphoinositide-3-kinase pathway and
Figure 1. A, heat map of unsupervised clustering analysis of polypeptides up-regulated by 1.3-fold in samples of 10 patients with WM as compared with normal control
(CTRL). All samples were normalized to control versus control samples (first sample, left). Yellow, protein level similar to the control; red/orange, overexpressed;
blue/green, underexpressed. B, expression levels of polypeptides up-regulated by 1.3-fold or more in samples of patients with WM. Rab4 and p62DOK were
up-regulated by 8-fold as compared with the normal control. C, heat map of polypeptides up-regulated by 2-fold or more in the 10 samples as compared with control.
These included seven proteins, including Rab4, p62DOK, FAS, CDC42, Roaz, SNX1, and ROKa.
Proteomics Analysis of Waldenstrom Macroglobulinemia
Cancer Res 2007; 67: (8). April 15, 2007
We then determined whether there was a difference in protein
expression in patients with asymptomatic disease/MGUS as
compared with those with symptomatic WM who required therapy.
As shown in Fig. 2, unsupervised clustering showed no difference in
protein expression between samples of patients with symptomatic
versus asymptomatic disease. However, there were three proteins
identified as up-regulated in symptomatic WM as compared with
asymptomatic WM//MGUS by >2-fold expression level. These
included the heat shock protein HSP90, the Ras family protein
CDC25C, and the chemotaxis protein p43/EMAPII.
Validation of the results. To assess reproducibility, a protein
array was done of protein from normal CD19+and CD138+cells
against the protein from the same cells, and no difference was
found between the two samples, indicating equal labeling of the
proteins by the two fluorescent dyes (data not shown). To validate
the results of the protein microarray, immunohistochemistry on
paraffin-embedded tissue from the same biopsies used for the
protein array analysis was done. Figure 4 shows the expressions of
Rab4, HDAC3, and RCC1 that were overexpressed in symptomatic
WM samples as compared with control. In addition, to further
confirm the overexpresssion of HSP90 in symptomatic WM, we
determined the level of HSP90 on samples of symptomatic WM as
compared with control and showed that HSP90 is overexpressed
in these samples consistent with the protein array data (Fig. 3).
Although some of the predictions of the arrays were confirmed,
immunohistochemistry also showed that the antibody arrays had
false-positive results. FAS was identified as an overexpressed
protein in the WM samples compared with the normal controls. On
validation with immunohistochemistry, we determined that FAS
identifies only adipocytes present in the bone marrow, which may
have been present in a higher quantity in the WM samples as
compared with the control, leading to a false-positive result. To
Table 1. Proteins up-regulated by 2-fold or more in WM samples compared with control
Ras-related protein (Rab4)Small GTPase in the Ras family that regulates recycling of proteins from the early
endosomes to the cell surface. Overexpression of Rab4 causes a redistribution of
receptors on plasma membrane versus endocytic compartments.
p62dok is a major substrate for many tyrosine kinases, including c-kit, v-abl, vFPS,
epidermal growth factor, and platelet-derived growth factor. Upon phosphorylation
by kinases, p62dok forms a complex with Ras GTPase-activating protein.
Fatty acid synthesis. Up-regulated in most human carcinomas. False positive in WM.
Along with other members of the GTP binding proteins (Rho and Rac), CDC42GAP is
implicated in regulating a variety of cellular functions, including actin cytoskeleton
organization, cell growth control and development, transcriptional activation,
membrane trafficking, and cell transformation.
Involved in intracellular trafficking. This endosomal protein regulates the cell surface
expression of epidermal growth factor receptor. This protein also has a role in sorting
protease-activated receptor-1 from early endosomes to lysosomes.
Alters the expression of target genes during cell lineage determination and differentiation.
Downstream of Ras/Raf-1. Binds RhoA, B, and C. Involved in reorganization of
RasGAP-associated docking protein p62DOK(P62DOK)
Sorting nexin 1 (SNX1)
Rat O/E-1–associated zinc finger protein (Roaz)
Table 2. Proteins up-regulated by >1.3-fold (n = 105) in WM as compared with control
Apoptosis/proliferation14-3-3e; caspase-7/Mch3, iNOS/type II, FADD/Mort-1, Smac/DIABLO,
caspase-8/FLICE, PCNA, perforin, Inhibitor 2, PTP1B, PI-4 kinase b,
Janus-activated kinase 1, Stat 3, HSF4, PKC I
Cdk2, cdk1, FBP, Rb2, c-Myc, HDAC3, E2F-2, GCIP, RCC1.
FYB/SLAP-130, STI1, NM23-H1 (nucleoside Di-P kinase), NEK3, Clk1 (Sty),
guanylate kinase, CaM kinase kinase, MKP2, TNIK, Rac1
Integrin b3 (CD61), MDC9 (ADAM), JAM-1, Mena, Maspin, LAR, annexin II,
p62 lck ligand, pp120 src substrate, CLA-1 (CD36), RPTPb, nexilin,
CBP, DP-1 Ku70, Max, p54nrb, TLS, Pax-5
AP-180 (AP-3) adenovirus 5E1A, GFAP, SIII p15, IGFBP-3, M33, DBP2, SV40
large T antigen, La Protein, Brm, Neurogenin3, Rch-1, annexin IV, CD3z,
ECA39, SRP54, ZFP-37, PECI, DCC, Lamp-1, AKAP 79, cGB-PDE, IL-12 (p70),
GAGE, SH2-B, FEN-1, rSec8, PYK2, Striatin, L1, GRIP, Rag-1, PDI, DLP1, JIP-1,
GIT1, G3BP, Rad 50, RanBP3
Adhesion, invasion, and cell motility
Cancer Res 2007; 67: (8). April 15, 2007
further confirm that the level of expression in normal controls
was reproducible in multiple samples, we analyzed three different
normal bone marrow samples and showed that the levels of Rab4,
HDAC3, RCC1, and HSP90 were similar in all samples (data not
In addition, to further validate our results, we did immunoblot-
ting on 10 samples of newly diagnosed symptomatic WM and 5
different controls that were not included in the protein array
analysis. As shown in Fig. 4A, p62DOK, Rab4, and HSP90 were
overexpressed in WM samples (n = 3) compared with normal
control cells. Similarly, the WM cells lines and IgM-secreting
lymphoma cells lines (BCWM1, RL, and WM-WSU) had a high
expression of all three proteins. In this experiment, we used CD19+
cells and CD138+cells selected separately to detect a true
differential expression of these proteins between malignant cells
and normal control and exclude changes that may occur due to the
ratio of lymphocytes and plasma cells in each sample. As shown in
Fig. 4A and B, both the CD19+cells and the CD138+cells of the
same samples expressed a higher protein concentration of HSP90,
p62DOK, and Rab4 compared with control. We then confirmed
those changes in a different set of WM samples (n = 4) and control
normal BM CD19+cells (n = 2) as shown in Fig. 4C. Finally, we used
a third group of primary WM samples (n = 3) that were selected for
concomitant CD19+/CD138+cells and compared with normal
controls selected in a similar fashion. As shown in Fig. 4D, the
expression of HSP90, p62DOK, and Rab4 was higher in the WM
samples compared with normal control. These data confirm the
overexpression of p62DOK, Rab4, and HSP90 in 10 symptomatic
WM samples that were not used in the protein array analysis.
The HDAC inhibitor trichostatin and HSP90 inhibitor 17-
AAG showed inhibition of proliferation and induction of
apoptosis in WM cell lines and patient samples. To confirm the
functional significance of protein elevation, we used the HDAC
inhibitor trichostatin and HSP90 inhibitor 17-AAG. BCWM1 and
WM-WSU cells were cultured for 24 and 48 h in the presence of
trichostatin (1–100 ng/mL) or 17-AAG (300–900 nmol/L). As shown
in Fig. 5, trichostatin inhibited WM cell survival at 24 and 48 h in a
dose-dependent fashion. Similarly, 17-AAG (900 nmol/L) induced
apoptosis at 24 and 48 h (Fig. 5B). Finally, we confirmed the
apoptotic effects of 17-AAG and trichostatin on primary WM cells
obtained from patients and showed significant induction of
apoptosis at 48 h in response to both 17-AAG and trichostatin.
There is an urgent need to elucidate the molecular pathways that
mediate proliferation and resistance to apoptosis in WM to provide
targets for novel therapies. Transcriptional profiling in WM has
identified some pathways that are up-regulated in WM (9).
Proteomic analysis represents a technique that yields more
information at the functional protein level. The antibody array
technology represents a high-throughput new technology to
identify novel proteins and rapidly screen multiple samples yielding
molecular signatures and profiles. However, this technique has its
limitations with false-positive and negative results (11, 13, 16–20),
and confirmation studies with conventional techniques are
In this study, we identified for the first time novel proteins that
are dysregulated in WM compared with normal controls. Using >2
cutoff, the microarrays identified seven up-regulated polypeptides
involved in signaling pathways regulating Ras-related proteins,
including Rab4 and p62DOK; Rho-related proteins, including
Figure 2. Heat map of unsupervised clustering analysis of samples of patients
with symptomatic WM (n = 5, left), control normal samples, and asymptomatic
WM/MGUS (n = 5, right). All samples were normalized to control versus control
samples (center of the clustering analysis). The clustering analysis showed
similar expression patterns in the symptomatic and asymptomatic samples.
Three proteins were differentially expressed in the symptomatic compared with
the asymptomatic samples by >2-fold expression difference. These included
HSP90, CDC25C, and p43/EMAPII.
Table 3. Proteins down-regulated by >1.3-fold in WM compared with control (n = 74)
Signaling pathways Proteins
Survival and cell cycleJNKK1, NPAT, hRAD9, RalA, GOK/Stim1, PP2A Catalytic a, PKC b, PKC a,
PKCE, IKKg/3/NEMO, NF-nB, IKKa/1, IkBe, PARP, Btf, p55Cdc
Integrin a3 (CD49c/VLA-3 a), Fibronectin, Phospholipase Cg 1,
p140mDia, MCP-3, Melusin
eIF-4g, p300, NTF2, TFII-I/BAP-135
Cyclooxygenase-2/PGHS, CDC27, heme oxygenase 1, acetylcholine receptor (b),
IL-3, IL-13, Mint1, Gelsolin, cathepsin L, VHL, Attractin, PMCA2, annexin VI,
ORC5, Tim23, CTBP1, chromogranin A, caveolin 1, LEDGF syntaxin 11, TRF2,
ZAP70 kinase nNOS, WT1 (Wilm’s tumor protein), Mint3, MyoD, L-Caldesmon,
K+ Channel a, XIN, COMT, SATB1, CAF-1 p150, SRPK2, FMS (CD115), annexin XI,
plakophilin 2a, Mxi-1, AMACR, Rab24, amphiphysin, MUPP1, KSR-1,
Psme3/PA28-g, CNTFRa, PMF-1, TFIIB, SIP1, MONA
Adhesion, cell motility
Transcription- and translation-related proteins
Proteomics Analysis of Waldenstrom Macroglobulinemia
Cancer Res 2007; 67: (8). April 15, 2007
CDC42GAP, and ROKa; and other proteins, such as SNX-1, Roaz,
and FAS. We then confirmed the expression of these polypeptides
using immunohistochemistry and immunoblotting and delineated
some of the biological networks that these proteins signal through.
Although the number of samples analyzed in this study was small,
we confirmed the results using another 10 WM samples using
immunoblotting. Therefore, protein array techniques can be used
as an exploratory tool, and other traditional techniques, such as
immunoblotting or immunohistochemistry, can be used for
confirmation of the identified proteins in a larger number of
Some of the polypeptides identified in this analysis might
contribute to the pathogenesis of WM, including those in the Ras
and Rho families of kinases. Ras proteins included Rab4 and
p62DOK. Oncogenic Ras expression occurs in up to 40% of multiple
myeloma cases and correlates with aggressive disease (21, 22). This
study, therefore, identifies a role of Ras signaling pathway in WM.
Rab4 is a Ras-like small GTPase that coordinates protein transport
Figure 4. A, immunoblotting for Rab4, p62DOK, and HSP90 in CD19+symptomatic WM samples (n = 3) and CD19+control cells (n = 2) and WM and IgM-secreting
cell lines (BCWM.1, RL, and WM-WSU). Expression of these proteins was higher in the WM samples and in the cell line as compared with the control samples.
p45DOK was used as loading control. B, immunoblotting for Rab4, p62DOK, and HSP90 in the same WM cell used in (A), but using the CD138+fraction of the
cells and control cells. Expression of these proteins was higher in the WM samples compared with the control samples. p45DOK was used as loading control.
C, immunoblotting for Rab4, p62DOK, and HSP90 in another group of symptomatic WM patient samples (n = 4) and CD19+control cells (n = 2) and BWM.1 cell line.
Expression of these proteins was higher in the WM samples compared with the control samples, confirming the results obtained in (A). Actin was used as loading
control. D, immunoblotting for Rab4, p62DOK, and HSP90 in another set of WM samples (n = 3) and control samples (n = 2) that were selected using concomitant
CD19+/CD138+selection. Expression of these proteins was higher in the WM samples compared with the control samples. AKT was used as loading control.
Figure 3. Immunohistochemistry for
HDAC3, HSP90, CDK2, Rab4, and FAS
antibodies in three samples of patients with
symptomatic WM compared with the
expression levels of normal bone marrow
control. The expression of Rab4, HDAC3,
CDK2, and HSP90 was elevated in WM as
compared with control. FAS expression
was only present in the fat globules. The
presence of lymphocytes was confirmed
by CD20 staining (data not shown).
Confirmation with three other normal
controls was done (data not shown).
Images were visualized under a Leica DM
IL microscope using an objective lens of
40/0.6 ? 2 (magnification, 80?) equipped
with a camera Leica dfc300 FX and
software Leica IM50 version 6.2.1.
Cancer Res 2007; 67: (8). April 15, 2007
from the endosome to the plasma membrane (23). It is associated
with prolonged activation of MAP kinase in some malignancies
(24). P62DOK or RasGAP-associated docking protein was originally
defined as a tyrosine-phosphorylated 62-kDa protein that coim-
munoprecipitated with p21Ras GTPase-activating protein (RasGAP;
ref. 25). RasGAP is an essential component of Ras-activated
signaling pathways (26, 27). RasGAP down-regulates Ras activity
and plays a role in cell growth and differentiation (26, 27).
Similarly, proteins in the Rho pathway were up-regulated in WM
as compared with normal controls. The GTPase RhoA has been
implicated in various cellular activities, including the formation of
stress fibers, motility, and cytokinesis (28, 29). Cdc42 belongs to the
Rho family of small GTP binding proteins along with Rac and Rho
(30). It is involved in regulating a variety of cellular functions,
including actin cytoskeleton organization, cell growth control and
development, transcriptional activation, membrane trafficking, and
cell transformation (30). ROKa is a p150 serine/threonine kinase
binding RhoA only in its active GTP-bound state, promoting the
formation of stress fibers and focal adhesion complexes (31).
Other polypeptides that were up-regulated by 1.3-fold include
HDAC3. Histone acetyltransferases can stimulate gene transcrip-
tion by acetylating histones, facilitating an open chromatin state
(32). Alteration in the chromatin structure allows access of
transcription factors to the promoter regions and results in the
activation of gene transcription (32). HDACs play a critical role on
the pathogenesis of B cell malignancies, such as in large B cell
lymphoma and multiple myeloma (32). In addition, we showed that
the HDAC inhibitor trichostatin inhibited growth and survival of
primary WM cells and WM cells lines, confirming that HDACs are
important regulators of survival in WM.
We further showed that the molecular changes occurred early in
the disease in cases with asymptomatic WM/MGUS analogous to
results in patients with multiple myeloma where the molecular
abnormalities identified in MGUS are similar to those identified in
symptomatic multiple myeloma (33). HSP90 was up-regulated in
symptomatic WM as compared with asymptomatic WM/MGUS,
indicating that this protein is up-regulated with the progression of
disease. HSP90 has been implicated in the pathogenesis and
resistance of many malignancies including multiple myeloma,
another plasma cell dyscrasia (34). We further confirmed the
functional significance of this protein in the survival of WM cells by
demonstrating that the HSP90 inhibitor 17-AAG induced signifi-
cant apoptosis and inhibition of growth in WM cell lines and
primary patient samples.
Previous studies of gene expression profiling in 23 patients
diagnosed with WM identified a homogenous expression profile of
WM cells that was similar to that of CLL. The most significantly up-
regulated gene was IL-6, and the most significantly associated
pathway for this set of genes was MAPK signaling. Although
changes in mRNA levels do not always translate into changes at
the protein level, we have identified multiple members of the Ras/
MAPK pathway up-regulated in this protein array analysis
reflecting consistency between gene and protein expression
In summary, our studies have identified for the first time novel
proteins that are differentially dysregulated in WM, which both
enhances our understanding of disease pathogenesis and represent
targets for novel specific inhibitors.
Received 8/21/2006; revised 1/16/2007; accepted 2/12/2007.
Grant support: Research Fund for Waldenstrom, ASH Scholar, and Leukemia and
Lymphoma Society. I.M. Ghobrial is a Lymphoma Research Scholar.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
Figure 5. A, apoptosis assay using annexin V/PI staining showed induction of
apoptosis with 17-AAG at 24 and 48 h incubation. B, apoptosis assay using
annexin V/PI staining showed induction of apoptosis with trichostatin at 24 and
48 h incubation. C, apoptosis assay using annexin V/PI staining showed
induction of apoptosis with 17-AAG at 48 h incubation in patient samples.
D, apoptosis assay using annexin V/PI staining showed induction of apoptosis
with trichostatin at 48 h incubation in patient samples.
Proteomics Analysis of Waldenstrom Macroglobulinemia
Cancer Res 2007; 67: (8). April 15, 2007
Cancer Research Download full-text
Cancer Res 2007; 67: (8). April 15, 2007
1. Dimopoulos MA, Kyle RA, Anagnostopoulos A, Treon
SP. Diagnosis and management of Waldenstrom’s
macroglobulinemia. J Clin Oncol 2005;23:1564–77.
2. Ghobrial IM, Gertz MA, Fonseca R. Waldenstrom
macroglobulinaemia. Lancet Oncol 2003;4:679–85.
3. Owen RG, Treon SP, Al-Katib A, et al. Clinicopatho-
logical definition of Waldenstrom’s macroglobulinemia:
consensus panel recommendations from the Second
International Workshop on Waldenstrom’s Macroglob-
ulinemia. Semin Oncol 2003;30:110–5.
4. Ghobrial IM, Fonseca R, Gertz MA, et al. Prognostic
model for disease-specific and overall mortality in newly
diagnosed symptomatic patients with Waldenstrom
macroglobulinaemia. Br J Haematol 2006;133:158–64.
5. Dhodapkar MV, Jacobson JL, Gertz MA, et al.
Prognostic factors and response to fludarabine therapy
in patients with Waldenstrom macroglobulinemia:
results of United States intergroup trial (Southwest
Oncology Group S9003). Blood 2001;98:41–8.
6. Gertz MA, Fonseca R, Rajkumar SV. Waldenstrom’s
macroglobulinemia. Oncologist 2000;5:63–7.
7. Kyle RA, Therneau TM, Rajkumar SV, et al. Long-term
follow-up of IgM monoclonal gammopathy of undeter-
mined significance. Blood 2003;102:3759–64.
8. Kyle RA, Rajkumar SV, Therneau TM, Larson DR,
Plevak MF, Melton LJ. Prognostic factors and predictors
of outcome of immunoglobulin M monoclonal gamm-
opathy of undetermined significance. Clin Lymphoma
profiling of Waldenstrom’s macroglobulinemia reveals a
phenotype more similar to chronic lymphocytic leukemia
than multiple myeloma. Blood 2006;108:2755–63.
10. Chen G, Gharib TG, Huang CC, et al. Discordant
protein and mRNA expression in lung adenocarcinomas.
Mol Cell Proteomics 2002;1:304–13.
11. Ghobrial IM, McCormick DJ, Kaufmann SH, et al.
Proteomic analysis of mantle cell lymphoma by protein
microarray. Blood 2005;105:3722–30.
12. Mitsiades CS, Mitsiades N, Treon SP, Anderson KC.
Proteomic analyses in Waldenstrom’s macroglobuline-
mia and other plasma cell dyscrasias. Semin Oncol 2003;
13. Petricoin EF, Bichsel VE, Calvert VS, et al. Mapping
molecular networks using proteomics: a vision for
patient-tailored combination therapy. J Clin Oncol
14. Zhu H, Snyder M. Protein chip technology. Curr Opin
Chem Biol 2003;7:55–63.
15. Santos D, Ho A, Tournilhac O, et al. Establishment of
a Waldenstrom’s macroglobulinemia cell line (BCWM.1)
with productive in vivo engraftment in SCID-hu mice
[abstract]. Blood 2005;106:106.
16. Skynner HA, Amos DP, Murray F, et al. Proteomic
analysis identifies alterations in cellular morphology
and cell death pathways in mouse brain after chronic
corticosterone treatment. Brain Res 2006;1102:12–26.
17. Hideshima T, Catley L, Yasui H, et al. Perifosine,
an oral bioactive novel alkylphospholipid, inhibits
Akt and induces in vitro and in vivo cytotoxicity in
18. Raje N, Kumar S, Hideshima T, et al. Combination of
the mTOR inhibitor rapamycin and CC-5013 has
synergistic activity in multiple myeloma. Blood 2004;
19. Carr KM, Rosenblatt K, Petricoin EF, Liotta LA.
Genomic and proteomic approaches for studying
human cancer: prospects for true patient-tailored
therapy. Hum Genomics 2004;1:134–40.
20. Eckel-Passow JE, Hoering A, Therneau TM, Ghobrial
I. Experimental design and analysis of antibody micro-
arrays: applying methods from cDNA arrays. Cancer Res
21. Hideshima T, Bergsagel PL, Kuehl WM, Anderson KC.
Advances in biology of multiple myeloma: clinical
applications. Blood 2004;104:607–18.
22. Fonseca R, Barlogie B, Bataille R, et al. Genetics and
cytogenetics of multiple myeloma: a workshop report.
Cancer Res 2004;64:1546–58.
23. Murray JW, Wolkoff AW. Assay of Rab4-dependent
trafficking on microtubules. Methods Enzymol 2005;403:
24. Kostenko O, Tsacoumangos A, Crooks D, Kil SJ,
Carlin C. Gab1 signaling is regulated by EGF receptor
sorting in early endosomes. Oncogene 2006;25:6604–17.
25. Yamanashi Y, Baltimore D. Identification of the Abl-
and rasGAP-associated 62 kDa protein as a docking
protein, Dok. Cell 1997;88:205–11.
26. Niki M, Di Cristofano A, Zhao M, et al. Role of Dok-1
and Dok-2 in leukemia suppression. J Exp Med 2004;200:
27. Dube N, Cheng A, Tremblay ML. The role of protein
tyrosine phosphatase 1B in Ras signaling. Proc Natl
Acad Sci U S A 2004;101:1834–9.
28. Ghiaur G, Lee A, Bailey J, Cancelas J, Zheng Y,
Williams DA. Inhibition of RhoA GTPase activity
enhances hematopoietic stem and progenitor cell
proliferation and engraftment in vivo. Blood 2006;108:
29. Fritz G, Kaina B. Rho GTPases: promising cellular
targets for novel anticancer drugs. Curr Cancer Drug
30. Wang L, Yang L, Filippi MD, Williams DA, Zheng Y.
Genetic deletion of Cdc42GAP reveals a role of Cdc42
in erythropoiesis and hematopoietic stem/progenitor
cell survival, adhesion, and engraftment. Blood 2006;
31. Rattan R, Giri S, Singh AK, Singh I. Rho/ROCK
pathway as a target of tumor therapy. J Neurosci Res
32. Mitsiades N, Mitsiades CS, Richardson PG, et al.
Molecular sequelae of histone deacetylase inhibition in
human malignant B cells. Blood 2003;101:4055–62.
33. Davies FE, Dring AM, Li C, et al. Insights into
the multistep transformation of MGUS to myeloma
using microarray expression analysis. Blood 2003;102:
34. Mitsiades CS, Mitsiades NS, McMullan CJ, et al.
Antimyeloma activity of heat shock protein-90 inhibi-
tion. Blood 2006;107:1092–100.