The AP1-dependent secretion of galectin-1 by
Reed–Sternberg cells fosters immune privilege
in classical Hodgkin lymphoma
Przemyslaw Juszczynski*, Jing Ouyang*, Stefano Monti†, Scott J. Rodig‡, Kunihiko Takeyama*, Jeremy Abramson*,
Wen Chen*, Jeffery L. Kutok‡, Gabriel A. Rabinovich§, and Margaret A. Shipp*¶
*Department of Medical Oncology, Dana–Farber Cancer Institute, 44 Binney Street, Boston, MA 02115;†Broad Institute, Cambridge Center,
Cambridge, MA 02142;‡Department of Pathology, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA 02115; and§Institute of
Biology and Experimental Medicine, Consejo Nacional de Investigaciones Cientı ´ficas y Te ´cnicas de Argentina, Vuelta de Obligado 2490 and
Department of Biological Chemistry, Faculty of Exact and Natural Sciences, University of Buenos Aires, C1428ADN, Buenos Aires, Argentina
Communicated by Klaus Rajewsky, Harvard Medical School, Boston, MA, June 26, 2007 (received for review April 23, 2007)
Classical Hodgkin lymphomas (cHLs) contain small numbers of neo-
plastic Reed–Sternberg (RS) cells within an extensive inflammatory
cells. The skewed nature of the T cell infiltrate and the lack of an
effective host antitumor immune response suggest that RS cells use
potent mechanisms to evade immune attack. In a screen for T
cell-inhibitory molecules in cHL, we found that RS cells selectively
overexpressed the immunoregulatory glycan-binding protein, galec-
tin-1 (Gal1), through an AP1-dependent enhancer. In cocultures of
cell Gal1 increased T cell viability and restored the Th1/Th2 balance.
In contrast, Gal1 treatment of activated T cells favored the secretion
of Th2 cytokines and the expansion of CD4?CD25highFOXP3?Treg
maintenance of an immunosuppressive Th2/Treg-skewed microenvi-
ronment in cHL and provide the molecular basis for selective Gal1
expression in RS cells. Thus, Gal1 represents a potential therapeutic
target for restoring immune surveillance in cHL.
immunomodulation ? microenvironment ? Th2 ? Treg? c-Jun
Europe each year; ?90% of these patients are young adults.
Classical HLs include small numbers of malignant Reed–Sternberg
(RS) cells within an extensive inflammatory infiltrate (1). The
tumor cells derive from preapoptotic germinal center B cells that
2). Classical HL RS cells lack B cell receptor-mediated signals and
rely on alternative survival and proliferation pathways activated by
transcription factors such as NF-?B and activator protein 1 (AP1)
(3–5). In cHL, the tumor cells exhibit constitutive AP1 activation,
express high levels of the AP1 components, c-Jun and JunB, and
depend on AP1-mediated proliferation signals (3).
Although primary cHLs have a brisk inflammatory infiltrate,
there is little evidence of an effective host antitumor immune
response. The reactive T cell population includes predominantly T
(Tregcells) that directly suppress immune responses and protect
cHL RS cells from immune attack (1, 6–8); Th1, natural killer, and
cytotoxic T cells are markedly underrepresented. In addition,
primary cHLs are characterized by a unique cytokine and chemo-
kine profile, including IL-4, IL-5, IL-10, and IL-13 (1, 9). In fact,
IL-13 is a critical growth factor for cHL RS cells (1, 9). However,
the molecular signals and endogenous factors responsible for
creating and maintaining the Th2-skewed immunosuppressive mi-
croenvironment in cHL remain to be defined.
In a screen for T cell-inhibitory molecules in cHL, we found that
cHL RS cells overexpressed the carbohydrate-binding lectin, ga-
of immune cell homeostasis and tumor-immune escape (10–13).
lassical Hodgkin lymphoma (cHL) is a B cell malignancy
diagnosed in ?20,000 new patients in North America and
Gal1, an evolutionarily conserved member of this family (14),
preferentially recognizes multiple Gal ?1,4 GlcNAc (N-
acetyllactosamine) units that may be presented on the branches of
N- or O-linked glycans on cell surface glycoproteins such as CD45,
CD43, and CD7 (15). Through binding and cross-linking of specific
glycoconjugates, Gal1 has the potential to inhibit T cell effector
functions and regulate the inflammatory response (16–20). In
several murine models of chronic inflammatory diseases, recom-
binant Gal1 suppressed Th1-dependent responses and increased T
cell susceptibility to activation-induced cell death (17–20).
In a recently described solid tumor (murine melanoma) model,
T cell-dependent immunity and conferring immune privilege to
tumor cells (12). In this model, Gal1 blockade markedly enhanced
syngeneic tumor rejection and tumor-specific T cell-mediated im-
mune responses (12). In another recently described solid tumor
(head and neck squamous cell carcinomas), Gal1 overexpression
was inversely correlated with the number of infiltrating T cells and
was an independent prognostic factor for shorter overall survival
(13). Because cHL is a unique lymphoid malignancy characterized
by an extensive but ineffective host inflammatory/immune re-
sponse, we evaluated the causes and consequences of Gal1 over-
expression in this disease.
Gal1 Is Overexpressed in cHL RS Cells. To identify novel cHL-specific
T cell-inhibitory molecules, we compared the gene expression
profiles of a series of cHL and diffuse large B cell lymphoma
(DLBCL) and mediastinal large B cell lymphoma (MLBCL) cell
lines than in the LBCL lines (P ? 0.002, FDR ? 0.014, Fig. 1 A and
and low or undetectable in DLBCL and MLBCL lines by Western
sections revealed abundant Gal1 expression in cHL RS cells,
Author contributions: P.J. and J.O. contributed equally to this work; P.J., J.O., S.M., J.A.,
J.L.K., G.A.R., and M.A.S. designed research; P.J., J.O., S.M., S.J.R., K.T., W.C., and J.L.K.
performed research; G.A.R. contributed new reagents/analytic tools; P.J., J.O., S.M., K.T.,
and J.L.K., G.A.R., and M.A.S. analyzed data; and P.J., J.O., G.A.R., and M.A.S. wrote the
The authors declare no conflict of interest.
Freely available online through the PNAS open access option.
Abbreviations: AP1, activator protein 1; cHL, classical Hodgkin lymphoma; DLBCL, diffuse
large B cell lymphoma; DN, dominant-negative; Gal1, galectin-1; MLBCL, mediastinal large
B cell lymphoma; PE, phycoerythrin; QPCR, quantitative PCR; rGal1, recombinant Gal1; RS
cell, Reed–Sternberg cell; SCR, scrambled control shRNA; shRNA, short hairpin RNA; Treg,
regulatory T cell; TDG, thiodigalactoside; Th cell, T helper cell.
¶To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2007 by The National Academy of Sciences of the USA
August 7, 2007 ?
vol. 104 ?
no. 32 www.pnas.org?cgi?doi?10.1073?pnas.0706017104
whereas LBCLs were uniformly negative (Fig. 1D). In a series of
primary lymphoid tumors, 10 of 10 cHLs were Gal1?, whereas 10
of 10 primary DLBCLs and 5 of 5 primary MLBCLs lacked Gal1
expression (data not shown). Taken together, these data indicate
that Gal1 is selectively up-regulated in the RS cells of cHLs.
RS Cell Gal1 Expression Is Regulated by an AP1-Dependent Enhancer.
To elucidate the mechanisms responsible for Gal1 overexpression
in cHL RS cells, we first analyzed the Gal1 locus on chromosome
22. We identified a candidate GC-rich regulatory element with a
Gal1 transcription start site. Because the AP1 components, c-Jun
and JunB, are overexpressed in cHL and are critical for the
pathogenesis of the disease (3), we asked whether AP1 mediates
by the previously described Gal1 promoter (?403 to ?67) (21) and
the putative Gal1 enhancer element (or mutated controls), and we
assessed associated luciferase activity in a cHL cell line (HD-
MY-Z) known to have constitutive activation of AP1 (Fig. 2A) (3).
or 1346–1746) up-regulated luciferase expression ?8- to 10-fold,
whereas constructs lacking the candidate sequence (459–777) or
containing a deletion in the AP1-binding site (1346–1746del) ex-
hibited significantly lower luciferase activity (Fig. 2A). Similar
results were obtained with a set of constructs in which the regula-
tory element was cloned upstream from the Gal1 promoter, dem-
onstrating that the identified sequence (bp 1346–1746) is a bona
fide Gal1 enhancer (data not shown).
Given the AP1 dependence of the Gal1 enhancer and the
constitutive AP1 activity in cHL (3), we next asked whether the
Gal1 enhancer was selectively active in this disease. For these
experiments, we transfected cHL, DLBCL, and fibroblast cell lines
or the promoter–enhancer construct (pGL3-Gal1403–67-Luc-
e1346–1746) and compared the respective luciferase activities (Fig.
2B). The Gal1 promoter–enhancer construct specifically up-
fibroblasts (Fig. 2B).
After demonstrating the specificity and activity of the Gal1 AP1
enhancer element in a cHL cell line (Fig. 2 A and B), we directly
evaluated the requirement for AP1 transcription factors in elec-
trophoretic mobility shift assays (Fig. 2C). Nuclear extracts from
three cHL and three DLBCL cell lines were incubated with
radiolabeled wild-type (WT) or mutant probes corresponding to
AP1 element in the Gal1 enhancer. Gal1 WT, but not mutant
probe, directly bound to nuclear proteins extracted from cHL, but
not DLBCL, cell lines (Fig. 2C). The complexes formed with Gal1
WT probe were displaced by unlabeled WT competitor, further
Gal1–AP1 complex was retarded by c-Jun antibody (Fig. 2C).
Furthermore, the simultaneous overexpression of a dominant-
negative c-Jun construct (c-Jun-DN) significantly reduced Gal1-
driven luciferase activity in cHL cells (Fig. 2D). In addition, when
AP1 was at least partially inhibited by the overexpression of
c-Jun-DN, there was a significant decrease in Gal1 transcript
abundance in cHL cells (Fig. 2E). Taken together, these studies
indicate that cHL RS cells selectively overexpress Gal1, at least in
part, by an AP1-driven enhancer.
Endogenous cHL Gal1 Expression Contributes to the Immunosuppres-
sive and Th2-Skewed Microenvironment in cHL. After delineating the
mechanism for cHL-specific overexpression of Gal1, we assessed
the functional consequences of cHL RS cell Gal1 expression on the
associated inflammatory/immune infiltrate. We first generated
stable HD-MY-Z transfectants expressing Gal1-specific short hair-
HD-MY-Z cells grown as adherent monolayers, and the cHL line
and T cells were cocultured. Thereafter, total T cell and Th cell
viabilities were assessed by triple-color annexin-V, CD3, and CD4
flow cytometry. There were significantly fewer viable total (CD3?)
in cocultures of HD-MY-Z cells with Gal1 knockdown (Fig. 3B).
These studies directly demonstrate that endogenous cHL RS cell
Gal1 decreases the viability of infiltrating activated T cells.
Given the skewed nature of inflammatory infiltrate in cHL, we
next asked whether endogenous cHL RS cell Gal1 may contribute
to this Th1/Th2 imbalance. To address this question, we isolated
CD4?Th cells from the cHL/T cell cocultures and analyzed the
relative expression of the Th1- and Th2-specific transcription
factors, T-bet and GATA-3, by real-time quantitative PCR
cells exhibited significantly lower expression of T-bet and higher
expression of GATA-3 than CD4?T cells from Gal1 shRNA
HD-MY-Z cocultures (Fig. 3C). Taken together, these results
cHL cell lines. Color scale at the bottom indicates relative expression ? SEM.
Red connotes high-level expression; blue indicates low-level expression. (B)
Median expression of Gal1 (boxes) in LBCL vs. cHL cell lines ? 25–75 percentile
(bars) and ? range (whiskers). (C) Respective cHL cell lines (KMHZ, HDLM2,
SupHD1, L1236, L540, L428, and HD-MY-Z), the MLBCL cell line (Karpas 1106),
and DLBCL cell lines (all others). (D) Gal1 immunohistochemistry (IHC). Gal1
IHC of a representative primary cHL (Upper) and DLBCL (Lower) are shown.
[Original magnifications: ?40 (Upper) and ?400 (Lower).]
Gal1 is overexpressed in cHL cell lines and primary tumors. Relative
Juszczynski et al. PNAS ?
August 7, 2007 ?
vol. 104 ?
no. 32 ?
viability of associated Th1 cells, resulting in a skewed Th2-type
Gal1 Promotes Immune Privilege by Favoring the Secretion of Th2
Cytokines and the Expansion of CD4?CD25highFOXP3?Treg Cells.
Given the profound immunosuppressive activity of Gal1 in
skewed Th2 cytokine profile associated with primary cHL. For
this purpose, we treated activated T cells with recombinant Gal1
(rGal1) in the presence or absence of the Gal1 inhibitor,
thiodigalactoside (TDG) (17). As expected, rGal1 induced apo-
ptosis of total activated T cells [supporting information (SI) Fig.
4]. Concurrent treatment with TDG completely blocked rGal1-
induced apoptosis, confirming the specificity of the Gal1 effect
(SI Fig. 4). We next quantified Th2 cytokines in supernatants
from activated T cells that were untreated or treated with rGal1
in the presence or absence of TDG. Supernatants from rGal1-
treated T cells contained significantly higher amounts of the
Hodgkin-associated Th2 cytokines, IL-4, IL-5, IL-10, and IL-13,
and TDG specifically blocked this effect (Fig. 3D). These data
further support the hypothesis that RS cell Gal1 expression
promotes Th2-type cytokine production in primary cHLs.
In addition to Th2 cells, the inflammatory infiltrate in
(CD4?CD25highFOXP3?) that directly blunt the host antitumor
assay. Activated T cells were treated with rGal1 in the presence or
absence of TDG and analyzed thereafter for Treg cells
(CD4?CD25highFOXP3?) by triple-color immunofluorescence as
described (23). The CD4?CD25highFOXP3?population was sig-
nificantly increased in rGal1-treated cells, and TDG completely
blocked this effect (Fig. 3E). Taken together, these results demon-
strate that Gal1 fosters the skewed and immunosuppressive micro-
environment in cHL by enhancing the production of Th2 cytokines
identify the mechanism as a phylogenetically conserved AP1-
responsive enhancer. In functional in vitro assays, we directly show
that cHL RS cell Gal1 decreased the viability of activated T cells
cytokines, including IL-4, IL-5, IL-10, and IL-13. In addition, Gal1
in the development and maintenance of the unique Th2/Treg-
skewed immunosuppressive microenvironment in primary cHL.
Although cHL RS cells exhibit near uniform Gal1 expression,
DLBCL and MLBCL are largely Gal1-negative, prompting spec-
ulation that Gal1 may distinguish cHL from certain ‘‘gray-zone’’
lymphomas that share characteristics of DLBCL and cHL (24). A
inflammatory response, highlighting the interaction between the
tumor cells and their host microenvironment (24). Gal1 overex-
pression is a defining feature of cHL that is not shared with its
closely related counterpart, primary MLBCL (25), providing in-
sights into the relative efficacy of host immune responses in these
The differential expression of Gal1 in these lymphomas is likely
factor components, c-Jun and JunB, and the constitutive activation
of the AP1 pathway (3). Gal1 expression is regulated, at least in
part, by a cHL-specific, AP1-driven enhancer. AP1 also functions
in synergy with NF-?B to control the proliferation and limit the
apoptosis of cHL RS cells (3). Therefore, in addition to its prosur-
vival functions in cHL RS cells, AP1 also regulates the interplay
between RS cells and the tumor microenvironment through a
AP1-dependent Gal1 enhancer. The previ-
ously described Gal1 promoter (21) and pu-
the predicted AP1-binding site (represented
by a black bar) were cloned into a luciferase
reporter vector, transiently transfected into
activities. Representative luciferase activities
from three independent experiments are
normalized to Renilla luciferase activity and
are presented as bars ? SD. (B) Selective ac-
tivity of the Gal1 enhancer. Classical HL, DL-
BCL, and fibroblast lines were transfected
with either the Gal1 promoter-only vector
(pGL3-Gal1?403 ?67-Luc) or the promoter–
enhancer construct (pGL3-Gal1403 ?67-Luc-
e1346–1746) and assessed as in A for their
respective luciferase activities. (C) AP-1 de-
pendence of the Gal1 enhancer in electro-
phoretic mobility shift assays. Nuclear
extracts from DLBCL cell lines (DHL4, DHL7,
and SupHD1) were incubated with WT or
mutant32P-labeled, double-stranded DNA
probe corresponding to AP1-binding site in
Gal1 enhancer. Specific, unlabeled competi-
tor and antibodies against c-Jun or ?-actin (control) were included in certain assays as indicated. The gel-shift band corresponding to the probe–protein complex is
indicated with an arrow, and supershift bands corresponding to the probe–protein–antibody complex are noted with asterisks. (D) c-Jun dependence of the Gal1
enhancer. HD-MY-Z cells were cotransfected with the Gal1 promotor-only vector or the Gal1 promotor–enhancer construct with either the dominant-negative c-Jun
were transfected with either c-Jun-DN or empty vector. Thereafter, relative Gal1 mRNA abundance was assessed by real-time QPCR.
www.pnas.org?cgi?doi?10.1073?pnas.0706017104Juszczynski et al.
RS Gal1 is likely to be a critical factor shaping the immuno- and
specifically induced the apoptosis of activated T lymphocytes,
suggesting that a similar mechanism operates in primary cHLs and
that these tumors represent sites of immune privilege. Of note,
Our short-term in vitro assays likely underestimate the long-term
in vivo effects of Gal1 in cHL because the lectin is also deposited
in the extracellular matrix and stroma where it kills susceptible T
cells (27). In our in vitro assays, Gal1-expressing RS cells selectively
decreased the viability of infiltrating Th1 cells, resulting in a
in Th1 cells and enrichment in Th2 cells, Gal1 significantly in-
The Gal1-associated cytokine profile suggests that this glycan-
binding protein may have additional functions beyond modulating
T cell responses in primary cHLs. Because IL-13 is a critical RS cell
growth factor (9), Gal1 may indirectly stimulate tumor growth by
fostering the production of this Th2 cytokine. Through its effect on
another Th2 cytokine, IL-5, Gal1 may also promote the character-
istic eosinophilic infiltrate in primary cHL (28).
Our observations regarding Gal1 function in cHL are consistent
with reports regarding the role of the lectin in murine models of
Th1-driven chronic inflammatory and autoimmune disorders, in-
cluding collagen-induced arthritis, inflammatory bowel disease,
graft vs. host disease, and autoimmune uveitis (17–20). In these
studies, the administration of Gal1 dramatically suppressed Th1-
dependent responses and skewed toward Th2 cytokine profiles
(17–20, 29). A careful examination of the mechanisms involved in
the repertoire of cell surface glycans required for Gal1 binding and
subsequent cell death, whereas Th2 cells are protected from Gal1
by differential sialylation of their cell surface glycoproteins (30).
In addition to the Th2 shift, we provide evidence showing that
Gal1 treatment increases
immune response. It is possible that the specific glycosylation
pattern of Gal1 receptors on Tregcells renders them resistant to
Gal1-induced apoptosis. In fact, Tregcells have been reported to
exhibit increased ?2,6 sialylation (compared with effector T cells)
cell death (32). This hypothesis is further supported by recent
studies demonstrating that Tregcells overexpress Gal1 and remain
resistant to Gal1-mediated apoptosis (33).
cells and identifies a key AP1-dependent mechanism regulating
cHL-specific immune privilege. Because Gal1 blockade dramati-
(12), it is possible that Gal1 inhibition may augment host antitumor
responses in primary cHL. Furthermore, this lectin is likely to have
Gal1 also promotes tumor cell motility and enhances tumor angio-
genesis (11, 34–36), processes critical for tumors such as cHL that
spread by contiguous involvement of adjacent nodes and organs.
For all of these reasons, Gal1 represents a promising rational
therapeutic target in cHL.
Materials and Methods
Cell Lines. Twenty-one DLBCL cell lines (Ly19, Ly18, Ly10, Ly8,
DHL4, DB, HT, WSU, Karpas 422, Toledo, and Farage), one
MLBCL cell line (Karpas 1106), and seven cHL cell lines (KMH2,
HDLM2, SupHD1, L1236, L540, L428, and HD-MY-Z) were
maintained as described previously (3, 37). Of note, the cHL cell
lines were previously demonstrated to have constitutive AP1 ac-
tivity and increased expression of c-Jun and JunB (Fig. 2 and SI
Identification of Genes Overexpressed in cHL Cell Lines by Gene
Expression Profiling. Total RNAs from a panel of 21 DLBCL and
7 cHL cell lines were hybridized to U133A and B oligonucleotide
and the data were analyzed as described previously (38). The top
9,586 genes that met threshold and variation index criteria were
analyzed with the GenePattern program (www.broad.mit.edu/
cancer/software/genepattern/CEF) to identify differentially ex-
pressed genes in cHL and DLBCL. Genes correlated with the class
template (HL vs. DLBCL) were identified by ranking them ac-
cording to their signal-to-noise ratio. For each gene, a specific P
false discovery rate (39, 40).
Analysis of Gal1 Expression in Cell Lines by Immunoblot. DLBCLand
(Invitrogen, Carlsbad, CA), and transferred to PVDF membranes
(Millipore Corp., Bedford, MA). Membranes were immunostained
with purified Gal1 rabbit IgG (12) and horseradish peroxidase-
conjugated donkey anti-rabbit antibody (GE Healthcare, Piscat-
of Th2 cells and Tregcells. (A) RNAi-mediated blockade of Gal1 expression in the
encoding Gal1-specific shRNA (Gal1 shRNA, G) or scrambled control shRNA (SCR
shRNA, S) and analyzed thereafter for Gal1 protein expression. (B) Viability of
total (CD3?) and CD4?T cells cocultured with Gal1 shRNA cHL or control SCR
Th2-specific transcription factors, T-bet and GATA-3, in CD4?cells from the Gal1
by Gal1-treated T cells. Activated T cells were either untreated or treated with
production was then assessed by cytometric bead arrays. (E) Tregcell abundance
rGal1?TDG or left untreated. The percentage of CD4?CD25?FOXP3?T cells was
tative of three separate experiments averaged to obtain the percent of
CD4?CD25highFOXP3?cells in the bar graph (Right).
Juszczynski et al.PNAS ?
August 7, 2007 ?
vol. 104 ?
no. 32 ?
Immunohistochemistry. Immunohistochemistry was performed as
described previously (42) with 5-?m-thick formalin-fixed, paraffin-
embedded tissue sections of newly diagnosed primary cHL and
DLBCL and purified Gal1 rabbit IgG.
Analysis of Regulatory Elements in the Gal1 Locus and Generation of
Gal1 Enhancer Constructs. Computational analysis of the Gal1 locus
(chr22:36,400,510–36,406,802, alignment with Human NCBI Ge-
nome assembly v36, March 2006) was performed with the publicly
available version of Genomatix suite (www.genomatix.de) (43) and
rVISTA (http://rvista.dcode.org) (44), and a putative downstream
regulatory element (enhancer) containing a conserved AP1-
promoter–enhancer reporter constructs, we first PCR-amplified
the Gal1 promoter region (?403 to ?67) and ligated this sequence
into the pGL3 promoterless reporter vector (Promega, Madison,
WI), generating pGL3-Gal1?403?67-Luc. Thereafter, fragments
spanning nucleotides 459-1746, 459–777, and 1346–1746 from the
Gal1 transcription start site were PCR-amplified and cloned into
pGL3-Gal1?403?67-Luc 3? of the luciferase gene. Deletions in
AP1 site (TGACTCA to TGxxxCA) were generated by using the
pGL3-Gal1?403?67-Luc-e1346?1746 construct and the Gene
Tailor site-directed mutagenesis system (Invitrogen) as recom-
mended by the manufacturer. An additional set of constructs was
generated with the candidate enhancer elements cloned upstream
from the Gal1 promoter.
Generation of DN c-Jun Constructs. DN c-Jun constructs were gen-
erated as described previously (45) with minor modifications. In
brief, a c-Jun fragment that lacked the transactivation domain
(amino acids 123–223) was PCR-amplified from intronless c-JUN
genomic DNA and ligated in the pFLAG-CMV2 vector (pFLAG-
CMV2-c-Jun-DN) (Sigma–Aldrich, St Louis, MO). All primer
sequences are available upon request.
Analysis of Gal1 Promoter–Enhancer Constructs with Luciferase As-
says. The HD-MY-Z cHL and SU-DHL7 DLBCL cell lines were
grown to 60–80% confluence on 24-well plates and cotransfected
(WT or mutant Gal1) per well and 100 ng of the control reporter
plasmid, pRL-TK (Promega) per well by using FuGENE 6 trans-
fection reagent (Roche Applied Science, Indianapolis, IN) accord-
ing to the manufacturer’s protocol. For cotransfection experiments
with c-Jun-DN-FLAG, HD-MY-Z cells were transfected with 150
ng of pGL3-Gal1?403 ?67-Luc-e1346 ?1746, 250 ng of pFLAG-
cells were lysed, and luciferase activities were determined by
chemiluminescence assay with the dual luciferase assay kit (Pro-
mega) and Luminoskan Ascent luminometer (Thermo Lab Sys-
tems, Franklin, MA) as described (42).
Electrophoretic Mobility Shift Analyses of the AP1-Binding Site in the
Gal1 Enhancer. Nuclear extracts from three cHL cell lines (HD-
MY-Z, L428, and SupHD1) and three DLBCL cell lines (SU-
DHL7, SU-DHL4, and Toledo) were obtained as described pre-
viously (42). Double-stranded WT and mutant probes
corresponding to the AP1-binding region in Gal1 enhancer (WT,
5?-TTTTCTGGGTGACTCACTTCCCCCG-3?; and mutant, 5?-
TTTTCTGGGTtcagtACTTCCCCCG-3?) (mutant bases in are in
lowercase letters) were end-labeled with [?-32P]ATP, purified, and
used in binding reactions as described (42). DNA binding was
carried out by using 5 ?g of nuclear extracts and ?10,000 cpm of
radiolabeled probe in 20 ?l of binding buffer (42). After a 30-min
incubation, reactions were loaded onto a 5% polyacrylamide gel
and electrophoresed. Gels were vacuum dried and exposed to x-ray
films overnight at ?80°C. For competitor studies, a 100? molar
excess of unlabeled WT or mutant probe was included in the
binding reactions. For supershift studies, 1 ?l of c-Jun antibody or
?-actin [Santa Cruz Biotechnology (Santa Cruz, CA) and Sigma–
Aldrich, respectively] was added to the reaction 15 min before the
QPCR Analysis of Gal1 Transcript Abundance After AP1 Inhibition.The
HD-MY-Z cHL cell line was grown to 60–80% confluence on
100-mm plates and transiently transfected with 15 ?g of pFLAG-
CMV2 (empty vector) or pFLAG-CMV2-c-Jun-DN plasmids and
FuGENE 6 transfection reagent (Roche Applied Science) accord-
ing to the manufacturer’s protocol. After 72 h in culture, RNA was
extracted with TRIzol reagent (Invitrogen), and cDNA was syn-
thesized from total RNA (3 ?g) by using SuperScript II reverse
transcriptase (Invitrogen) and random hexamer primers. Gal1 and
GAPDH (housekeeping control) transcript abundance was evalu-
Biosystems, Foster City, CA) and the following primers: GAPDH
forward, GATTCCACCCATGGCAAATTC; GAPDH reverse,
TGATTTTGGAGGGATCTCGCTC; Gal1 forward, TCGC-
CAGCAACCTGAATCTC; Gal1 reverse, GCACGAAGCTCT-
TAGCGTCA. PCR was performed using an ABI 7700 thermal
cycler (Applied Biosystems), and threshold cycle (CT) values were
generated with the Sequence Detection Software, version 1.2
(Applied Biosystems). Gal1 transcript abundance was calculated
relative to the housekeeping control GAPDH by using the
2?(?CTGal1??CTGAPDH)method according to the manufacturer’s
instructions. Standard deviations were calculated from triplicate
RNAi-Mediated Gal1 Knockdown. Gal1-specific siRNA was designed
by the siRNA Selection Program (46) (http://jura.wi. mit.edu/bioc/
siRNAext), synthesized as single-stranded DNA oligonucleotides
by Integrated DNA Technologies (IDT, Inc., Coralville, IA) and
annealed. Gal1-specific oligonucleotide (Gal1 RNAi, GATC-
CATCTGGCAGCTTTTTTG) or scrambled oligonucleotide
GAGACGCGCGAGATATGGAGGTTTTTTG) was ligated
into the linearized pSIREN-RetroQ retroviral vector (BD Clon-
tech, Mountain View, CA). Generation of recombinant retrovirus
and infection of HD-MY-Z cells were performed as described
previously (42). After infection, cells were subjected to puromycin
selection (0.5 ?g/ml) and subcloning by limiting dilution. Thereaf-
ter, whole-cell extracts of obtained subclones were prepared and
screened for Gal1 expression by immunoblotting as described
or viability of transduced HD-MY-Z cells.
Cocultures and Analyses of T Cell Responses in cHL Microenvironment.
Coculture. Peripheral blood mononuclear cells were obtained from
centrifugation. Thereafter, T cells were purified by using a Pan T
cell isolation kit II (Miltenyi Biotec, Auburn, CA) and were
activated with 1 ?g/ml phytohemagglutinin (Sigma–Aldrich) for
64 h. Activated T cells (1 ? 106) were then cocultured with
monolayers (3 ? 106cells) of HD-MY-Z cells expressing either
scrambled shRNA or Gal1-specific shRNA for 6 h at 37°C.
Analyses of viable T cells. After coculture, all cells were harvested
and sequentially stained with CD3-phycoerythrin (PE) and
CD4-PE-Cy5 (Beckman Coulter, Fullerton, CA) followed by an-
nexin V-FITC and analyzed with a Beckman Cytomics
FC500 flow cytometer. The numbers of viable (annexin V and
propidium iodide-negative) CD3?and CD4?T cells in Gal1
shRNA or scrambled shRNA HD-MY-Z/T cell cocultures were
Analyses of T-bet and GATA-3 expression in cocultured CD4?T cells.After
phytohemagglutinin-activated T cells were cocultured with HD-
MY-Z cells expressing either scrambled shRNA or Gal1-specific
shRNA for 24–48 h, cells were collected, and nonviable cells were
www.pnas.org?cgi?doi?10.1073?pnas.0706017104Juszczynski et al.
depleted by Ficoll–Hypaque gradient centrifugation. Remaining
viable cells were washed, incubated with CD4 MACS microbeads,
and purified by using a MACS LS column according to the
manufacturer’s protocol (Miltenyi Biotec). Thereafter, RNA was
obtained from the isolated CD4?cells, and cDNA was synthesized
from total RNA (3 ?g) as described above. GATA-3, T-bet, and
GAPDH (housekeeping control) transcript abundance was evalu-
ated by QPCR using Power SYBR Green PCR Master Mix
(Applied Biosystems), the GAPDH primers listed above, and the
following GATA-3 and T-bet primers: GATA-3 forward, TAA-
CATCGACGGTCAAGGCA; GATA-3 reverse, ACACCTG-
GCTCCCGTGGT; T-bet forward, TGGACGTGGTCTTGGTG-
GACC; T-bet reverse, TGGACGTACAGGCGGTTTCC. PCR
and transcript abundance analysis were performed as described
above. GATA-3 and T-bet expression in purified CD4?cells from
SCR shRNA and Gal1 shRNA HD-MY-Z/T cell cocultures were
Cytokine production in T cell subpopulations treated with recombinant
Gal1 (rGal1). Recombinant Gal1 was obtained and purified essen-
simultaneously activated with anti-CD3- (0.1 ?g/ml) and anti-
rGal1 (20 ?M) in RPMI medium 1640 alone or rGal1 and the Gal1
inhibitor TDG (100 mM) (17) or left untreated for 16 h. Superna-
tants were then analyzed for IL-4, IL-5, IL-10, and IL-13 by using
cytometric bead array Flex Set beads according to the manufac-
turer’s protocol (BD Biosciences, San Jose, CA). In brief, multi-
plexed antibody-conjugated beads were incubated with culture
supernatants or serial dilutions of cytokine standards for 1 h.
Thereafter, the PE detection reagent was added, and samples were
incubated for additional 2 h, washed, and analyzed by a FACS Aria
flow cytometer (BD Biosciences). Results were captured with
Analyses of Tregcells. T cells were purified, activated with anti-CD3-
and TDG or left untreated for 24 h. Thereafter, cells were stained
control as described previously (23) (human regulatory T cell
Tregcells were identified and quantified by a Beckman Cytomics
FC500 flow cytometer as described previously (23).
Statistical Analysis. All statistical analyses were done by using
Statistica 6.0 software (Statistica, Tulsa, OK). Student’s t test was
used for comparisons between two groups; ANOVA was used for
1. Re D, Kuppers R, Diehl V (2005) J Clin Oncol 23:6379–6386.
2. Kanzler H, Kuppers R, Hansmann ML, Rajewsky K (1996) J Exp Med
3. Mathas S, Hinz M, Anagnostopoulos I, Krappmann D, Lietz A, Jundt F,
Bomment K, Mechta-Grigoriou F, Stein H, Dorken B, et al. (2002) EMBO J
4. Kuppers R, Schwering I, Brauinger A, Rajewsky K, Hansmann ML (2002) Ann
5. Schwering I, Brauninger A, Klein U, Jungnickel B, Tinguely M, Diehl V,
Hansmann ML, Dalla-Favera R, Rajewsky K, Kuppers R (2003) Blood
6. Marshall NA, Christie LE, Munro LR, Culligan DJ, Johnston PW, Barker RN,
Vickers MA (2004) Blood 103:1755–1762.
7. Gandhi MK, Lambley E, Duraiswamy J, Dua U, Smith C, Elliott S, Gill D,
Marlton P, Seymour JF, Khanna R (2006) Blood 108:2280–2289.
8. Ishida T, Ishii T, Inagaki A, Yano H, Komatsu H, Iida S, Inagaki H, Ueda R
(2006) Cancer Res 66:5716–5722.
9. Skinnider BE, Kapp U, Mak TW (2002) Leuk Lymphoma 43:1203–1210.
10. Rabinovich GA, Baum LG, Tinari N, Paganelli R, Natoli C, Liu FT, Iacobelli
S (2002) Trends Immunol 23:313–320.
11. Liu F-T, Rabinovich GA (2005) Nat Rev Cancer 5:29–41.
12. Rubinstein N, Alvarez M, Zwirner NW, Toscano MA, Llarregui JM, Bravo A,
Mordoh J, Fainboim L, Poshajcer OL, Rabinovich GA (2004) Cancer Cell
13. Le Q-T, Shi G, Cao H, Nelson DW, Wang Y, Chen EY, Zhao S, Kong C,
Richardson D, O’Byrne KJ, et al. (2005) J Clin Oncol 23:8932–8941.
14. Vasta GR, Ahmed H, Odom EW (2004) Curr Opin Struct Biol 14:617–630.
15. Stillman BN, Hsu DK, Pang M, Brewer CF, Johnson PJ, Liu F-T, Baum LG
(2006) J Immunol 176:778–789.
16. Perillo NL, Pace KE, Seihamer JJ, Baum LG (1995) Nature 378:736–739.
17. Rabinovich GA, Daly G, Dreja H, Tailor H, Riera CM, Hirabayashi J,
Chernajovsky Y (1999) J Exp Med 190:385–397.
18. Toscano MA, Commodaro AG, Ilarregui JM, Bianco GA, Liberman A, Serra
HM, Hirabayashi J, Rizzo LV, Rabinovich GA (2006) J Immunol 176:6323–
19. Santucci L, Fiorucci S, Rubinstein N, Mencarelli A, Palazzetti B, Federici B,
Rabinovich GA, Morelli A (2003) Gastroenterology 124:1381–1394.
20. Baum LG, Blackall DP, Arias-Magallano S, Nanigian D, Uh SY, Browne JM,
Hoffmann D, Emmanouilides CE, Territo MC, Baldwin GC (2003) Clin
21. Salvatore P, Benvenuto G, Caporaso M, Bruni CB, Chiariotti L (1998) FEBS
22. Gabrilovich DI (2007) Curr Cancer Drug Targets 7:1.
23. Zorn E, Nelson EA, Mohseni M, Porcheray F, Kim H, Litsa D, Bellucci R,
Raderschall E, Canning C, Soiffer RJ, et al. (2006) Blood 108:1571–1579.
24. Abramson J, Shipp M (2005) Blood 106:1164–1174.
25. Savage K, Monti S, Kutok J, Cattoretti G, Neuberg D, de Leval L, Kurtin P,
Dal Cin P, Ladd C, Feuerhake F, et al. (2003) Blood 102:3871–3879.
26. Gandhi MK, Moll G, Smith C, Dua J, Lambley E, Ramuz O, Gill D, Marlton
P, Seymour JF, Khanna R (April 16, 2007) Blood, 10.1182/blood-2007-01-
27. He J, Baum LG (2004) J Biol Chem 279:4705–4712.
28. von Wasielewski R, Seth S, Franklin J, Fischer R, Hubner K, Hansmann ML,
Diehl V, Georgii A (2000) Blood 95:1207–1213.
29. van der Leij J, van den Berg A, Harms G, Eschbach H, Vos H, Zwiers P, van
Weeghel R, Groen H, Poppema S, Visser L (2006) Mol Immunol 10:1–8.
30. Toscano MA, Bianco GA, Ilarregui JM, Croci DO, Correale J, Hernandez JD,
Zwirner NW, Poirier F, Riley EM, Baum LG, et al. (June 24, 2007) Nat
Immunol, 10.1038/ni 1482.
31. Jenner J, Kerst G, Handgretinger R, Muller I (2006) Exp Hematol 34:1211–
32. Amano M, Galvan M, He J, Baum LG (2003) J Biol Chem 278:7469–7475.
33. Garin MI, Chu C-C, Golshayan D, Cernuda-Morollon E, Wait R, Lechler RI
(2007) Blood 109:2058–2065.
34. Rabinovich GA (2005) Br J Cancer 92:1188–1192.
35. Thijssen VL, Postel R, Brandwijk RJ, Dings RP, Nesmelova I, Satijn S,
Verhofstad N, Nakabeppu Y, Baum LG, Bakkers J, et al. (2006) Proc Natl Acad
Sci USA 103:15975–15980.
36. Camby I, Belot N, Lefranc F, Sadeghi N, de Launoit Y, Kaltner H, Musette S,
Darro F, Danguy A, Salmon I, et al. (2002) J Neuropathol Exp Neurol
37. Smith PG, Wang F, Wilkinson KN, Savage KJ, Klein U, Neuberg DS, Bollag
G, Shipp MA, Aguiar RCT (2005) Blood 105:308–316.
38. Monti S, Savage KJ, Kutok JL, Feuerhake F, Kurtin P, Mihm M, Wu B,
Pasqualucci L, Neuberg D, Aguiar RC, et al. (2005) Blood 105:1851–1861.
39. Benjamini Y, Drai D, Elmer G, Kafkafi N, Golani I (2001) Behav Brain Res
40. Reiner A, Yekutieli D, Benjamini Y (2003) Bioinformatics 19:368–375.
41. Polo JM, Juszczynski P, Monti S, Cerchietti L, Ye K, Greally JM, Shipp M,
Melnick A (2007) Proc Natl Acad Sci USA 104:3207–3212.
42. Juszczynski P, Kutok JL, Li C, Mitra J, Aguiar RCT, Shipp MA (2006) Mol Cell
43. Scherf M, Klingenhoff A, Werner T (2000) J Mol Biol 297:599–606.
44. Loots GG, Ovcharenko I (2004) Nucleic Acids Res 32:W217–W221.
45. Ludes-Meyers JH, Liu Y, Munoz-Medellin D, Hilsenbeck SG, Brown PH
(2001) Oncogene 20:2771–2780.
46. Yuan B, Latek R, Hossbach M, Tuschl T, Lewitter F (2004) Nucleic Acids Res
Juszczynski et al. PNAS ?
August 7, 2007 ?
vol. 104 ?
no. 32 ?