Oncogene (1998) 16, 2131 ± 2139
 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00
http://www.stockton-press.co.uk/onc
Defective IkBa in Hodgkin cell lines with constitutively active NF-kB
Katrina M Wood1, Marilynn Ro€2,3 and Ronald T Hay2
1
Department of Pathology, Royal Victoria In®rmary, Newcastle upon Tyne NE1 4LP, England; 2School of Biological and Medical
Science, Irvine Building, University of St. Andrews, St. Andrews KY169AL Scotland
The molecular mechanisms underlying Hodgkin's disease
remain obscure, but it has been recognized that the
neoplastic cells display high levels of constitutively active
nuclear NF-kB. Here we demonstrate that although
nuclear NF-kB is transcriptionally active, the Hodgkin
cells fail to activate NF-kB dependent transcription in
response to CD40 ligand. In three Hodgkin cell lines
examined each had abnormalities in expression of IkBa
which could account for the deregulated NF-kB.
Although all three cell lines had greater than normal
levels of IkBa mRNA no IkBa protein could be detected
in the KM-H2 cells, while the L428 cell line contains a
C-terminally truncated IkBa species that fails to
associate with NF-kB. The HDLM-2 cell line contains
a more slowly migrating form of IkBa that can associate
with NF-kB, but increasing the level of this protein
within the cell fails to inhibit nuclear NF-kB. Addition of
recombinant IkBa to nuclear extracts from all three cell
lines resulted in complete inhibition of NF-kB DNA
binding activity and introduction of a plasmid expressing
IkBa into the cells inhibited the transcriptional activity
of an NF-kB dependent reporter plasmid. Thus the
constitutive expression of NF-kB in Hodgkin cells is a
direct consequence of the abnormal expression of IkBa
rather than changes in NF-kB that render it refractory
to inhibition by IkB proteins. These changes could, at
least in part, account for the characteristic activated
phenotype of Hodgkin cells and their pattern of cytokine
secretion, which determine the pathological appearance
and clinical manifestations of Hodgkin's disease.
Keywords: Hodgkin's lymphoma; NF-kB; IkBa
Introduction
Hodgkin's disease is a group of lymphomas characterised by malignant Reed-Sternberg and Hodgkin
cells in a reactive background of lymphocytes, plasma
cells, histiocytes and eosinophils (Kadin, 1994). The
nodular lymphocyte predominance subtype of Hodgkin's disease is now recognised to be a distinct
clinicopathological entity, a form of B cell lymphoma
(Mason et al., 1994). The origin of the malignant cell in
the other histological subtypes of Hodgkin's disease
remains obscure, although recent evidence has tended
Correspondence: RT Hay
3
Present address: Protein Fractionation Centre, Scottish National
Blood Transfusion Service, Edinburgh EH17 7QT, Scotland
Received 14 August 1997; revised 19 November 1997; accepted 19
November 1997
to favour a lymphoid origin, with both B and T cell
types (Stein and Hummel, 1993).
Reed-Sternberg and Hodgkin cells characteristically
express CD30, which is also found on activated B and
T cells, and a range of other cytokine receptors such as
the interleukin-2 receptors p55 (CD25) and p75, CD71
and the interleukin-6 receptor. They produce a large
number of cytokines including interleukin-1a (IL-1),
tumour necrosis factor (TNFa) and lymphotoxin-a
(LTa) which are associated with the constitutional or
`B' symptoms of Hodgkin's disease ± fever, night
sweats, itching and weight loss (Gorschluter et al.,
1995; Gruss et al., 1997).
Several authors have reported the presence of CD40
on Hodgkin and Reed-Sternberg cells (Carbone et al.,
1995a;b; Gruss et al., 1994; O'Grady et al., 1994). The
number of cases expressing CD40 varied between 70%
and 100% in the di€erent studies, with much of this
variation attributable to di€erences in ®xation and the
use of di€erent monoclonal antibodies to CD40. A
variety of functional e€ects have been demonstrated in
the Hodgkin cell lines in response to stimulation with
soluble trimeric CD40-ligand (Carbone et al., 1995a;
Gruss et al., 1994, 1995), suggesting that the CD40 on
the cell surface is competent to transduce signals
following interaction with its ligand.
A typical response to engagement of the IL-1,
TNFa, LTa, CD30 and CD40 receptors by their
cognate ligands is activation of the transcription
factor NF-kB. A common feature of Reed-Sternberg
and Hodgkin cells is that they have high levels of
constitutively active NF-kB (Bargou et al., 1996; Gruss
et al., 1997). The NF-kB family of transcription factors
plays a crucial role in the activation of many genes, the
products of which have important roles in cell
proliferation, the immune response and in¯ammation.
In unstimulated cells NF-kB is held in the cytoplasm in
an inactive form by IkB proteins. In most cases
activation of NF-kB involves signal induced degradation of IkBa, thus releasing the transcription factor
which translocates to the nucleus (Baeuerle and
Baltimore, 1996; Baldwin, 1996; Matthews and Hay,
1995; Roulston et al., 1995). The IkB family of proteins
(a, b, g, e and Bcl-3) are characterised by the presence
of multiple ankyrin repeats and their ability to
physically associate with NF-kB proteins (Baeuerle
and Baltimore, 1996; Baldwin, 1996; Matthews and
Hay, 1995; Roulston et al., 1995). Signal induced
activation of NF-kB is preceded by phosphorylation
and rapid degradation of IkBa (Beg and Baldwin,
1993; Brown et al., 1993; Henkel et al., 1993; Mellits et
al., 1993) mediated by the ubiquitin proteasome
pathway (Palombella et al., 1994). The target residues
for signal induced modi®cation are located in the
exposed N-terminus of the IkBa protein (Ja€ray et al.,
1995) with S32 and S36 being the sites of phosphoryla-
Defective IkBa in Hodgkin cells
KM Wood et al
2132
tion (Brockman et al., 1995; Brown et al., 1995;
DiDonato et al., 1996; Ro€ et al., 1996; Traenckner
et al., 1995) and K21 and K22 being the principal sites
of ubiquitination (Baldi et al., 1996; Chen et al., 1995;
Rodriguez et al., 1996; Scherer et al., 1995). Signal
induced modi®cation of the IkBa N-terminus, although
necessary, is not sucient to promote degradation of
the protein. Additional signals located within the Cterminus of the protein are also required (Brown et al.,
1995; Rodriguez et al., 1995; Sun et al., 1996;
Whiteside et al., 1995). One of the results of NF-kB
activation is induction of the IkBa gene and
consequent resynthesis of the IkBa protein (Chiao et
al., 1994; LeBail et al., 1993; Sun et al., 1993).
Resynthesized IkBa accumulates transiently in the
nucleus where it is responsible for post-induction
repression of NF-kB dependent gene expression
(Arenzana-Seisdedos et al., 1995) and nuclear export
of the transcription factor (Arenzana-Seisdedos et al.,
1997).
The objective of this work was to investigate the
mechanism which leads to NF-kB activation in
Hodgkin cells. In all three cell lines examined the
nuclear NF-kB was shown to be transcriptionally
active with both p50 and p65 subunits present.
Analysis of IkBa protein levels indicated that in
KM-H2 cells IkBa could not be detected under
normal circumstances, but a short fragment of the
protein accumulated when inhibitors of proteasomal
degradation were present. In L428 cells a C-terminally
truncated version of IkBa was detected which was
unable to associate with NF-kB and which failed to
undergo signal induced degradation. HDLM-2 cells
produced a more slowly migrating form of IkBa
which could associate with NF-kB, but increasing the
levels of this protein failed to inhibit NF-kB DNA
binding activity. In each cell line the NF-kB could be
inhibited by recombinant IkBa in vitro and by IkBa
expressed from a transfected plasmid in vivo. Thus
constitutive activation of NF-kB in Hodgkin cells
appears to be a direct consequence of defective IkBa
expression, even although IkBa mRNA levels appear
to be normal.
show a substantial increase in kB dependent transcription following CD40L stimulation. Hodgkin cells also
display constitutive NF-kB dependent reporter activity
but in contrast to the Daudi cells they fail to respond
substantially to CD40L (Figure 1). All cells do respond
weakly to PMA (Figure 1).
High levels of NF-kB DNA binding activity in the nuclei
of Hodgkin cell lines
Nuclear extracts prepared following stimulation of
Hodgkin cell lines and Daudi cells with CD40L were
analysed for NF-kB DNA binding activity. The three
Hodgkin cell lines display a high level of constitutive
NF-kB DNA binding activity which is not increased in
response to CD40L. Although the Daudi cells also
have constitutively active NF-kB DNA binding activity
it is less than that observed in the Hodgkin cells and it
is increased substantially in response to CD40L (Figure
2). Analysis of the NF-kB DNA complex by supershifting with antibodies to Rel family members reveals
that heterodimers of p50 and p65 are present in all
cases. While supershifting with the p65 antibody was
incomplete no supershifts were detected with antibodies
to p52, c-rel and rel B although there were indications
that DNA binding activity was reduced after addition
of c-rel antibody. DNA binding activity was abolished
by incubation with recombinant IkBa (Figure 2). It is
therefore likely that the constitutive DNA binding
activity present in the nuclei of Hodgkin cell lines is
composed of p50 and p65 heterodimers which are not
abnormal in their ability to be inhibited by IkBa.
Defective IkBa in Hodgkin cell lines
In cells of the haematopoietic lineage NF-kB activity is
regulated by IkBa (Beg et al., 1995; Haskill et al.,
1991). The metabolism of IkBa was therefore examined
in Hodgkin cells and Daudi cells after treatment with
CD40L. Western blot analysis with a monoclonal
antibody directed against the N-terminus of IkBa
Results
Hodgkin cell lines fail to activate NF-kB dependent
transcription in response to CD40 ligand
As engagement of CD40 by the CD40 ligand (CD40L)
has been reported to activate the transcription factor
NF-kB and CD40 is present on the surface of many
Hodgkin cell lines (Carbone et al., 1995a) it was of
interest to determine if the signal transduction pathway
from CD40 to NF-kB was functional in these cells.
Three well characterized Hodgkin cell lines (L428,
KM-H2 and HDLM-2) and the Daudi B-cell line were
transfected with either a kB dependent luciferase
reporter plasmid or a similar luciferase reporter
lacking NF-kB binding sites. Twenty-four hours after
transfection the cells were exposed to either soluble
CD40L, control supernatant or phorbol ester myristate
acetate (PMA) and the reporter activity determined.
Although the Daudi cells display a high level of
constitutive NF-kB dependent transcription, they also
Figure 1 Hodgkin cell lines fail to activate NF-kB dependent
transcription in response to CD40 ligand. 3enhconAluc or
conAluc luciferase reporter plasmids were transfected into the
indicated cell lines by the DEAE dextran procedure. Twenty-four
hours after transfection cells were exposed to either CD40 ligand
(CD40L), a control supernatant or PMA. Six hours later cells
were harvested and luciferase activity determined in a luminometer. Activity is expressed as relative light units (r.l.u.) per mg
of protein
Defective IkBa in Hodgkin cells
KM Wood et al
electrophoretic mobility than the form apparent in
Daudi cells. IkBa was not detected in KM-H2 cells
(Figure 3). Identical results were obtained when a
polyclonal antibody raised against recombinant IkBa
(Mellits et al., 1993) was used in Western blot analysis
–
+
0
30
60
120
+
0
30
60
120
–
+
–
0
30
60
120
+
KM-H2
0
30
60
120
–
0
30
60
120
HDLM–2
0
30
60
120
L428
0
30
60
120
Daudi
CD40L
0
30
60
120
revealed apparently normal IkBa of 38 kD in Daudi
cells (Figure 3). In L428 cells the 38 kD form of IkBa
was not detected, but was replaced by a faster
migrating species of 30 kD (Figure 3). HDLM-2 cells
contain two species of IkBa that have a slightly slower
time (min)
B
relB
rIκBα
p65
c-rel
p52
p50
PI
relB
rIκBα
p65
c-rel
p50
p52
PI
rIκBα
p65
p50
p52
PI
p65
p50
PI
F
Figure 2 NF-kB DNA binding activity in the nuclei of Hodgkin cells. Cell lines were exposed to either control supernatant (7) or
supernatant from cells expressing CD40 ligand (+) for the indicated period of time prior to nuclear extract preparation. NF-kB
DNA binding activity in 5 mg of nuclear extract was determined in a gel electrophoresis DNA binding assay using a 32P-labelled
DNA representing a kB motif present in the HIV-1 LTR. The position of free DNA (F) and protein bound DNA (B) is indicated.
Subunit composition of DNA protein complexes was determined by incubation with either the indicated antibodies or by addition
of 10 ng recombinant IkBa (rIkBa). In each case the nuclear extract was from cells incubated for 30 min with CD40 ligand. Images
were obtained from a Fuji Phosphorimager
Daudi
–
+
–
KM-H2
+
–
+
0
30
60
120
0
30
60
120
0
30
60
120
0
30
60
120
+
HDLM–2
0
30
60
120
0
30
60
120
0
30
60
120
0
30
60
120
–
L428
CD40L
time (min)
45 kD
IκBα
29 kD
IκBβ
45 kD
29 kD
Figure 3 Defective IkBa in Hodgkin cell lines. Cell lines were exposed to either control supernatant (7) or supernatant from cells
expressing CD40 ligand (+) for the indicated periods of time prior to preparation of cytoplasmic extracts. 25 mg of cytoplasmic
extract was analysed by Western blotting using the 10B monoclonal antibody speci®c for IkBa (upper panels). Antibodies were
stripped from the blots, which were then exposed to peptide antibodies speci®c for IkBb (lower panels). The positions of prestained
molecular weight markers are indicated
2133
Defective IkBa in Hodgkin cells
KM Wood et al
To determine if the truncated IkBa observed in L428
cells and the more slowly migrating form of IkBa in
HDLM-2 cells could associate with NF-kB, proteins
were immunoprecipitated from cell extracts with
a
Daudi
KM-H2
L428
HDLM–2
– –+ + – – + + – – + + – – + +
Dex
– + – + – +– +
–
–+
–+ – +
+ ZLLLH
B
F
b
Daudi
KM-H2
L428
HDLM–2
Dex
– –+ + – – + + – – + + – – + +
– + – + – +– +
–
–+
–+ – +
+ ZLLLH
68 kD
45 kD
IκBα
29 kD
Glio.
60
120
0
ZLLLH
30
TPCK
120
c
30
60
Dexamethasone treatment of cells has been reported to
increase IkBa levels as a consequence of increasing the
level of IkBa mRNA (Auphan et al., 1995; Scheinman
et al., 1995). IkBa levels can also be increased by
inhibiting proteasome mediated degradation with the
peptide aldehyde inhibitor ZLLLH (Ro€ et al., 1996).
Daudi and Hodgkin cell lines were therefore exposed
to either dexamethasone, ZLLLH or a combination of
both and NF-kB DNA binding activity was determined
in a gel electrophoresis DNA binding assay (Figure
4a). IkBa levels were determined by Western blotting
(Figure 4b). In Daudi cells treatment with ZLLLH
either in the absence or presence of dexamethasone
results in a small decrease (30 ± 40%) in NF-kB DNA
binding activity (Figure 4a) and a corresponding
increase in the levels of IkBa (Figure 4b). Treatment
with ZLLLH, dexamethasone or the combination does
not alter the level of NF-kB DNA binding activity
detected in KM-H2, L428 or HDLM-2 cells. No IkBa
is detected in the KM-H2 cells and this is not altered
by treatment with dexamethasone or ZLLLH. Treatment of L428 cells with ZLLLH causes accumulation of
the truncated protein, along with a more slowly
migrating forms which may represent phosphorylated
derivatives. There is a slight increase in the level of the
truncated protein following treatment with dexamethasone, but the combination of dexamethasone and the
inhibitor causes a more striking increase in IkBa
protein level (Figure 4b). HDLM-2 cells show a
striking shift to more slowly migrating forms of IkBa
when the cells are treated with ZLLLH. Although
dexamethasone alone has only a minor e€ect on IkBa
levels, the combination treatment results in a signi®cant
increase in the total amount of IkBa protein (Figure
4b).
An identical result was obtained when the blots were
probed with a previously described (Mellits et al., 1993)
polyclonal antiserum raised against recombinant IkBa
(data not shown). Stripping and reprobing the blot
with an antibody to IkBb indicated that over the time
scale of the experiment levels of this protein were not
a€ected by treatment with dexamethasone or ZLLLH
(data not shown). It is therefore clear that in L428 and
HDLM-2 cells increasing the level of IkBa protein does
not bring about a corresponding decrease in NF-kB
DNA binding activity. In an attempt to detect
fragments of IkBa that may result from rapid
degradation of the protein, KM-H2 cells were treated
with either the protease inhibitor TPCK, the fungal
Association of IkBa with NF-kB in Hodgkin cell lines
60
Agents which allow IkBa to accumulate do not in¯uence
NF-kB activity in Hodgkin cell lines
gliotoxin or the proteasome inhibitor ZLLLH and
proteins separated on a high concentration polyacrylamide gel. Western blot analysis of the transferred
protein with IkBa antibodies revealed the presence of a
small fragment of IkBa that was detected after
treatment with ZLLLH, but not TPCK or gliotoxin
(Figure 4c). This fragment is derived from the Nterminus of the protein as the monoclonal antibody
used for detection recognises an epitope located in the
N-terminus of IkBa (Ja€ray et al., 1995).
120
(data not shown). While stimulation with CD40L
produces the characteristic loss and resynthesis of
IkBa in Daudi cells, this is not the case for the IkBa
species in HDLM-2 and L428 cells (Figure 3). After
CD40L stimulation of HDLM-2 cells the faster
migrating species of IkBa disappears, but there is
little change in the level of the more slowly migrating
form of IkBa. Stripping and reprobing these membranes with the polyclonal antibody to IkBb
(Thompson et al., 1995) shows that all three Hodgkin
cell lines and the Daudi cells contain the IkBb protein.
The level of IkBb does not change over the 2 h period
of stimulation with CD40L (Figure 3).
30
2134
time (min)
29 kD
KM-H2
18.4 kD
14.3 kD
Figure 4 Agents which allow IkBa to accumulate do not
in¯uence NF-kB activity in Hodgkin cell lines. (a,b) The
indicated cell lines were incubated in the absence (7) or presence
(+) of 1 mM Dexamethasone (Dex) or 10 mM proteasome inhibitor
ZLLLH for 2 h. Cells were harvested and fractionated into nucleus
and cytoplasm. NF-kB DNA binding activity in 5 mg of nuclear
extract was determined in a gel electrophoresis DNA binding
assay (a) while IkBa protein in cytoplasmic extracts was detected
by Western blotting with the 10B monoclonal antibody. (c) In a
separate experiment KM-H2 cells were treated with TPCK,
gliotoxin (Glio.) or ZLLLH for the indicated times and
cytoplasmic extracts fractionated in a 15% polyacrylamide gel
containing SDS prior to detection of IkBa by Western blotting as
described above
Defective IkBa in Hodgkin cells
KM Wood et al
antibodies to NF-kB and associated proteins identi®ed
by Western blotting. Proteins from Daudi cells and the
three Hodgkin lines were immunoprecipitated with
antibodies to NF-kB p50, NF-kB p65 or pre-immune
serum and analysed by Western blotting with the IkBa
speci®c monoclonal antibody. IkBa coprecipitates with
p50 and p65 in Daudi and HDLM-2 cells but not in
KM-H2 and L428 cells (Figure 5a). As IkBa could not
be detected in KM-H2 cells (Figures 3 and 4) this was
expected but it demonstrates that the truncated version
of IkBa observed in L428 cells is unable to associate
with NF-kB. The protein present in L428 was readily
immunoprecipitated with polyclonal antisera to recom-
–
p50
p65
L428 HDLM–2
p50
p65
–
p65
–
KM-H2
p50
p65
c.e
Daudi
p50
a
–
IP Ab
68 kD
Ab H-chain
45 kD
IκBα
IkBa mRNA in Hodgkin cells
As IkBa protein was not detected in KM-H2 cells and
a truncated form of the protein was present in L428
cells, RNA analysis was performed. Total RNA from
Jurkat T-cells and the Hodgkin cells was analysed by
Northern blotting using an IkBa cDNA as probe. It is
clear that all of the Hodgkin cell lines have greater
than normal levels of IkBa mRNA that, within the
resolving power of the gel, appears to be full length
(Figure 6).
Introduction of IkBa into Hodgkin's cell inhibits
endogenous NF-kB activity
29 kD
The high level of nuclear NF-kB in the Hodgkin cells
and the presence of apparently defective forms of IkBa
suggested that transfer of the wild type IkBa gene into
these cells could inhibit NF-kB activity. A potential
explanation for the defective IkBa in these cells is that
the signal transduction pathway which results in IkBa
Western blot MAb 10B
MAb 10B
IκBα
–
p50
L428 IPs
L428
HDLM–2
KM-H2
Daudi
cell extract
p65
b
binant IkBa and the 10B monoclonal antibody but not
with antibodies to p50 or p65 (Figure 5b). The form of
IkBa present in L428 cells is likely to be truncated at
the C-terminus as the 10B monoclonal antibody
recognizes the N-terminus of the protein. This was
con®rmed when the blot was stripped and reprobed
with an antibody directed against the extreme Cterminus of IkBa. This antibody recognised the
proteins present in Daudi and HDLM-2 cells but did
not react with the truncated IkBa in L428 cells (Figure
5c).
KM-H2
L428
IκBα
Jurkat
Ab H-chain
45 kD
HDLM–2
68 kD
∆IκBα
29 kD
Western blot MAb 10B
c
68 kD
Ab H-chain
45 kD
IκBα
29 kD
Western blot C-ter peptide Ab
Figure 5 Association of IkBa with NF-kB in Hodgkin cell lines.
Cytoplasmic extracts from the indicated cell lines were
immunoprecipitated with either preimmune serum (7), antibodies to p50 or antibodies to p65. Immunoprecipitates were
analysed by Western blotting using the 10B monoclonal antibody
speci®c for IkBa (a). Cell extracts from the indicated cell lines
were analysed by Western blotting using the 10B monoclonal
antibody, which recognises an epitope in the N-terminus of IkBa
(b) or antibodies directed against the C-terminus of IkBa (c).
Prior to Western blotting extracts from L428 cells were
immunoprecipitated with either preimmune serum (7) or
antibodies to p50, p65 and IkBa. The position of prestained
molecular weight markers and cross reacting IgG Heavy chains
are indicated
IκBα
Figure 6 IkBa m RNA in Hodgkin cells. Total RNA (20 mg)
from the indicated cell lines was analysed by Northern blotting
using a 32P-labelled random primed DNA probe speci®c for IkBa
2135
Defective IkBa in Hodgkin cells
KM Wood et al
IκBα S32A, S36A
IκBα
pcDNA
2136
Figure 7 Introduction of IkBa in Hodgkin cells inhibits constitutive NF-kB activity. An NF-kB dependent luciferase reporter (HIV
LTR luc) and a control plasmid lacking functional NF-kB binding sites (DkB HIV LTR luc) were co-electroporated into KM-H2
(a), L428 (b) and HDLM-2 (c) cells with expression plasmids for either IkBa, S32A S36A IkBa or empty vector pCDNA3. Twentyfour hours after electroporation cell extracts were prepared for determination of luciferase activity (R.L.U. in 100 mg protein) and
expression of IkBa in KM-H2 cells was determined by Western blotting (10 mg protein) with the 10B monoclonal antibody
degradation has been activated. To test this possibility
an IkBa S32A, S36A mutant, which fails to undergo
signal induced degradation (Ro€ et al., 1996) was also
tested. KM-H2 cells were electroporated with an NFkB dependent luciferase reporter (HIV LTR luc) along
with plasmids expressing wild type IkBa, IkBa S32A,
S36A or empty vector (pcDNA). An equivalent
reporter in which the kB motifs were mutated (DkB
HIV LTR luc) was used as control. The high level of
kB dependent reporter activities in the KM-H2, L428
and HDLM-2 cells is reduced by introduction of either
wt IkBa or IkBa S32A, S36A into these cells, although
the extent of inhibition by the ectopic IkBa species
varies between cell lines (Figure 7). Inhibition of NFkB dependent reporter activity by ectopic wild type
ImBa is less pronounced in HDLM-2 cells, but reporter
activity is reduced to basal levels by the S32A, S36A
mutant of IkBa. Levels of expressed IkBa were
determined by Western blotting of the electroporated
cells. No substantial di€erences in the accumulated
levels of IkBa or IkBa S32A, S36A were detected
(Figure 7, insert), suggesting that the lack of IkBa in
the KM-H2 cell is not due to permanent activation of
the normal signal transduction pathway that leads to
IkBa degradation.
Discussion
The conclusion of this study is that constitutive
activation of NF-kB observed in Hodgkin cell lines
(Bargou et al., 1996; Gruss et al., 1997; this study) is a
consequence of defective IkBa. In KM-H2 cells IkBa
cannot be detected under normal circumstances, but an
18 kD fragment of IkBa is detected after treatment of
the cells with an inhibitor of proteasomal degradation
(Figure 5). We were unable to detect any full length
IkBa with a range of polyclonal and monoclonal
antibodies. These data indicate that in KM-H2 cells an
unstable 18 kD fragment of IkBa is made which is
rapidly degraded by the proteasome. It is unlikely that
an unusual, proteasome independent, processing event
is taking place as wild type IkBa introduced into these
cells appears at the expected molecular weight (Figure
7). A similar situation is apparent in L428 cells in
which a truncated version of IkBa is also detected. In
this case the truncated protein has a molecular weight
of about 30 kD and so should contain about 250 of the
317 amino acids in IkBa. The short proteins from KMH2 and L428 cells both appear to be truncated at the
C-terminus as they are detected by the 10B monoclonal
antibody which recognises an epitope located between
K21 and E48 in IkBa (Ja€ray et al., 1995), but are not
detected by an antibody raised against the C-terminal
12 amino acids of IkBa. Consistent with previous
deletion and partial proteolysis experiments (Ernst et
al., 1995; Hatada et al., 1993; Ja€ray et al., 1995;
Rodriguez et al., 1995) the truncated form observed in
L428 cells is unable to interact with NF-kB and would
thus be incapable of retaining NF-kB in the cytoplasm
or inhibiting NF-kB DNA binding activity. Treatment
of L428 cells with the proteasome inhibitor ZLLLH
leads to signi®cant accumulation of the truncated IkBa
protein (Figure 5) suggesting that, as in KM-H2 cells,
the truncated forms of IkBa which fail to bind NF-kB
are being rapidly turned over. This would be expected
as it is known that free and NF-kB bound IkBa have a
di€erent susceptibility to proteolytic degradation (Sun
et al., 1993). The situation in HDLM-2 cells is
somewhat di€erent as they contain constitutively
active NF-kB even although full length IkBa, which
can associate with NF-kB, is detected. However, this
protein appears to have a slightly slower electrophoretic mobility, in polyacrylamide gels containing SDS,
than IkBa from Daudi cells (Figure 3 ± 5). Treatment
of these cells with a combination of ZLLLH and
Defective IkBa in Hodgkin cells
KM Wood et al
dexamethasone results in the accumulation of substantial quantities of what appears to be phosphorylated IkBa (Figure 5), again suggesting that IkBa is
being rapidly turned over. Although IkBa accumulates
under the above conditions it does not reduce the NFkB DNA binding activity present in the cell nucleus,
thus suggesting that the IkBa present in HDLM-2 cells
is functionally defective. An alternative explanation for
these ®ndings is that the nuclear NF-kB is refractile to
inhibition by IkBa as would be expected if the nuclear
NF-kB was being shielded bound to IkBb (Suyang et
al., 1996). This is unlikely as the nuclear DNA binding
activity is eciently inhibited by recombinant IkBa in
vitro (Figure 2) and NF-kB dependent transcription is
inhibited after electroporation of a plasmid expressing
IkBa (Figure 7).
There are interesting parallels between this work on
Hodgkin cell lines and observations made on IkBa
`knockout' mice (Beg et al., 1995). The IkBa de®cient
mice die within a few days of birth and show high
levels of constitutive NF-kB activation, particularly in
haematopoietic cells. As in the Hodgkin cell lines, the
mice also show increased transcription of some, but
not all, NF-kB responsive genes (such as G-CSF).
The extent to which the neoplastic phenotype of
Hodgkin cells is dependent on their defective IkBa and
deregulated NF-kB remains to be established. In a
previously published study on Hodgkin cells all cells
contained constitutively active NF-kB and while IkBa
was not analysed in L428 cells, Western blotting of
extracts from KM-H2 and HDLM-2 cells indicated
that immunoreactive material which comigrated with
wild type IkBa was detected at a very low level
(Bargou et al., 1996). Although not mentioned, these
authors also detected a more slowly migrating form of
IkBa which could correspond to that observed in this
report. A reason for the apparently contradictory
results observed in KM-H2 cells is not clear, but
di€erent antibodies to IkBa were employed. The IkBa
speci®c monoclonal antibody used in this study is very
sensitive allowing excellent discrimination between the
signal for IkBa and non-speci®c signals. Deregulated
expression of NF-kB in the neoplastic Hodgkin and
Reed-Sternberg cells could account for the characteristic phenotype of these cells with expression of
activation markers and production of cytokines.
Involvement of NF-kB in cell proliferation is
suggested by activation of NF-kB during the cell
cycle (Baldwin et al., 1991), by virtue of a p300
mediated interaction between p65 and cyclin E-Cdk3
(Perkins et al., 1997). A further important role of
constitutively active NF-kB in the Hodgkin cells may
be to prevent apoptosis. It has been demonstrated that
NF-kB activation is required to protect cells from TNF
mediated apoptosis (Beg and Baltimore, 1996; van
Antwerp et al., 1996; Wang et al., 1996), while in Bcells TGFb promotes apoptosis by inducing synthesis
of IkBa which blocks the activity of NF-kB (Arsura et
al., 1996). The role played by constitutively active
NF-kB in B cells appears to be in transcription of the
c-myc proto-oncogene, the promoter of which is
responsive to NF-kB (Duyao et al., 1990). Direct
evidence for the oncogenic potential of NF-kB is
provided by the known variants of Rel family members
which are associated with tumours. v-Rel is a mutated
form of c-Rel isolated from the acutely transforming
REV-T retrovirus. It is a potent oncogene in avian cells
and infection of young chickens with the REV-T
retrovirus causes multiple haematopoietic tumours of
the spleen and liver, typically resulting in death after
7 ± 10 days (Bose, 1992; Gilmore, 1991).
The human IkBa gene on chromsome 14 (Glavac et
al., 1994) does not appear to be commonly associated
with chromosome breaks or deletions in cases of
Hodgkin's disease (Koduru et al., 1993). Although
this is consistent with our data indicating the IkBa
mRNA is of apparently normal electrophoretic
mobility on a Northern blot (Figure 6) cytogenetic
studies are dicult to perform in these tumours and
the results do not exclude the possibility of smaller
scale alterations in the IkBa gene.
The results presented here suggest that loss of
functional IkBa in Hodgkin cell lines could account
for the characteristic activated phenotype of these cells,
their deregulated cytokine secretion and consequently
the clinical and pathological manifestations of Hodgkin's disease.
Materials and methods
Cell lines
Three well-characterised Hodgkin cell lines were used for
these experiments. The HDLM-2 and KM-H2 lines were
obtained from the DSM cell line collection and the L428
cell line was provided by Professor Harald Stein. All three
cell lines were independently derived from patients with
disseminated Hodgkin's disease. The KM-H2 line originated from a patient with mixed cellularity Hodgkin's
disease, and the other lines from patients with nodular
sclerosing Hodgkin's disease. The HDLM-2 cell line was
cultured in RPMI supplemented with 20% heat inactivated
foetal calf serum. The other cell lines were grown in RPMI
1640 with 10% foetal calf serum.
Antibodies
The Mad 10B monoclonal antibody was generated against
recombinant full-length IkBa and recognises an epitope in
the N-terminus (Ja€ray et al., 1995). Rabbit antibodies
against the IkBa C terminus (residues 297 ± 317), IkBb C
terminus (residues 339 ± 358), and p65/RelA (residues 531 ±
550) were obtained from Santa Cruz Biotechnology, Inc.
Polyclonal rabbit antisera to recombinant p50, p65 and
IkBa were as described (Arenzana-Seisdedos et al., 1995).
Rabbit antibodies to p49 were generated by immunisation
with bacterially produced recombinant protein.
Cell stimulations and preparation of extracts
Cells were collected by centrifugation and resuspended in 8
ml of medium at a concentration of 36106 cells/ml. 24-well
Costar plates were prepared with stimulating reagents at
the appropriate concentration.
CD40-ligand CHO supernatant and CHO control supernatant were obtained from Immunex, Seattle and were used
at a dilution of 1/250. The proteasome inhibitor ZLLLH was
synthesised as described (Ro€ et al., 1996) and used at 10 mM.
Dexamethasone (Sigma) was used at a concentration of 1 mM.
Cell suspensions were added to the plate at 1 ml/well and
cells which were not removed immediately for lysis were
incubated at 378C. At each time point, cells from a single well
were washed in ice-cold PBS and lysed in 300 ml bu€er
containing 20 mM sodium phosphate pH 7.5, 50 mM sodium
2137
Defective IkBa in Hodgkin cells
KM Wood et al
2138
¯uoride, 5 mM tetra sodium pyrophosphate, 10 mM bglycerophosphate, 0.5% NP-40 and protease inhibitors (1 mM
leupeptin, 1 mM pepstatin, 1 mM pefablock and 20 mM
TPCK). After centrifugation to collect nuclei, the cytoplasmic fraction was removed and the nuclei washed twice in lysis
bu€er before resuspension in lysis bu€er containing 425 mM
NaCl. After 30 min on ice, the extracts were clari®ed by
centrifugation at 13 000 r.p.m. for 20 min at 48C in an
Eppendorf microfuge. Extracts were used immediately or
stored at 7708C.
Western blotting
Cytoplasmic extracts containing 25 mg protein (Bradford,
1976) were resolved in a 10% polyacrylamide gel containing SDS, and transferred to Immobilon PVDF membranes
(Millipore) by semi-dry electroblotting. Membranes were
incubated in blocking bu€er (TBS containing 5% skimmed
milk and 0.1% Tween 20) for 1 h, followed by primary
antibody diluted in the same bu€er. Monoclonal antibody
Mad 10B hybridoma supernatant was diluted 1/100 and
commercial polyclonal antibodies to IkBb or the Cterminus of IkBa (Santa Cruz Biotechnology, Inc.) were
used at a concentration of 2 mg/ml. Membranes were
washed in TBS containing 0.1% Tween 20. Speci®callybound antibody was detected using peroxidase-conjugated
goat anti-mouse (DAKO, 1/750) or goat anti-rabbit
(DAKO, 1/500) and the Amersham ECL system.
When indicated, membranes were stripped by incubation
in 60 mM Tris HCl pH 6.8, 2% SDS and 100 mM bmercaptoethanol at 708C for 30 min. After extensive washing
in TBS-Tween, membranes were blocked again and probed
with antibody as before.
Gel electrophoresis DNA binding assays
NF-kB binding assays were performed as described
(Arenzana-Seisdedos et al., 1995) with 5 mg nuclear
protein and a 32P-labelled double stranded DNA oligonucleotide representing the upstream kB motif present in the
HIV-1 enhancer. Free DNA was resolved from the DNAprotein complexes on a native 6% polyacrylamide gel and
the positions of radioactive species determined by
phosphorimaging of the dried gel. The subunit composition of DNA-protein complexes containing NF-kB was
determined by pre-incubation of nuclear extracts with
polyclonal antibodies to p50, RelA/p65, c-Rel, p49 or
RelB. Additional samples were pre-incubated with 10 ng
recombinant IkBa (Ja€ray et al., 1995).
NF-kB dependent reporter assays
NF-kB dependent luciferase reporters were HIVLTRluc
and 3EnhConAluc, while control reporters lacking functional NF-kB binding sites were DkBHIVLTRluc and
ConAluc (Alcami et al., 1995; Rodriguez et al., 1996).
Expression plasmids for IkBa and S32A, S36A IkBa have
been described previously (Ro€ et al., 1996). Plasmids were
introduced into cells either by DEAE dextran mediated
transfection (Pazzagli et al., 1992) or electroporation (Alcami
et al., 1995). Where indicated cell stimulations were carried
out 24 h after transfection for a period of 6 h with either
CD40 ligand supernatant (diluted 1 in 250), control supernatant or 100 ng/ml phorbol myristate acetate (PMA). Cells
were collected by centrifugation and the luciferase activity in
cell lysates determined as described previously (Rodriguez et
al., 1996).
Northern blotting
RNA was isolated by acid guanidinium thiocyanate phenol
chloroform extraction (Chomczynski, 1992), fractionated
in agarose gels containing formaldehyde and transferred to
nylon membranes as described (Chomczynski and Sacchi,
1987). The RNA was hybridised to 32P-labelled DNA
representing the coding region of the human IkBa gene
prepared by random priming (Pharmacia). Radioactive
species were detected by phosphorimaging.
Immunoprecipitations
Cells (26107) were collected by centrifugation, washed in
PBS and lysed in 1 ml of bu€er containing 50 mM Tris.
HCl pH 7.4, 150 mM sodium chloride, 0.5% Nonidet-P40,
1 mM sodium orthovanadate and protease inhibitors as
above. Following centrifugation at 6000 r.p.m. for 3 min in
an Eppendorf microfuge, the supernatant was pre-cleared
three times with 100 ml of Protein G-Sepharose
(Pharmacia). Tubes containing 25 ml Protein G (Pharmacia) were pre-loaded with: 200 ml PBS, 2 mg polyclonal antip65 (Santa Cruz Biotechnology, Inc.), 1 ml anti-p50 rabbit
serum, 1 ml anti-IkBa rabbit serum, or 200 ml anti-IkBa
(Mad 10B) tissue culture supernatant. All volumes were
made up to 200 ml with PBS and the tubes allowed to
rotate for 1 h at 48C. Pre-loaded Protein G was washed
twice with lysis bu€er and the pre-cleared lysate divided
between the antibody tubes. These were then allowed to
tumble again for 1 h at 48C. Immunoprecipitates were
washed three times in lysis bu€er, boiled in reducing
sample bu€er diluted 1:1 in distilled water and analysed by
SDS ± PAGE and Western blotting.
Acknowledgements
We would like to thank Richard Armitage, Immunex,
Seattle for the CD40-ligand. Advice on electroporation and
transfection provided by Dr Fernando Arenzana-Seisdedos, Institut Pasteur, Paris was greatly appreciated, as were
Eric Cabannes' comments on the manuscript. KMW was
supported by the Wellcome Trust and MR was supported
by the Medical Research Council. This work bene®ted
from the support of EC project ROCIO.
References
Alcami J, Delera TL, Folgueira L, Pedraza MA, Jacque JM,
Bachelerie F, Noriega AR, Hay RT, Harrich D, Gaynor
RB, Virelizier JL and Arenzana-Seisdedos F. (1995).
EMBO J., 14, 1552 ± 1560.
Arenzana-Seisdedos F, Thompson J, Rodriguez MS,
Bachelerie F, Thomas D and Hay RT. (1995). Mol. Cell.
Biol., 15, 2689 ± 2696.
Arenzana-Seisdedos F, Turpin P, Rodriguez M, Thomas D,
Hay RT, Virelizier JL and Dargemont C. (1997). J. Cell
Sci., 110, 369 ± 378.
Arsura M, Wu M and Sonenshein GE. (1996). Immunity, 5,
31 ± 40.
Auphan N, Di DJ, Rosette C, Helmberg A and Karin M.
(1995). Science, 270, 286 ± 290.
Baeuerle PA and Baltimore D. (1996). Cell, 87, 13 ± 20.
Baldi L, Brown K, Franzoso G and Siebenlist U. (1996). J.
Biol. Chem., 271, 376 ± 379.
Baldwin AJ, Azizhan JC, Jensen DE, Beg AA and Coodly
LR. (1991). Mol. Cell. Biol., 11, 4943 ± 4951.
Baldwin AS. (1996). Ann. Rev. Immunol. 14, 649 ± 683.
Defective IkBa in Hodgkin cells
KM Wood et al
Bargou RC, Leng C, Krappmann D, Emmerich F, Mapara
MY, Bommert K, Royer HD, Scheidereit C and Dorken
B. (1996). Blood, 87, 4340 ± 4347.
Beg AA and Baldwin AJ. (1993). Genes Dev., 7, 2064 ± 2070.
Beg AA and Baltimore D. (1996). Science, 274, 782 ± 789.
Beg AA, Sha WC, Bronson RT and Baltimore D. (1995).
Genes Dev., 9, 2736 ± 2746.
Bose HJ. (1992). Biochim. Biophys. Acta, 1114, 1 ± 17.
Bradford MM. (1976). Anal. Biochem., 72, 248 ± 254.
Brockman JA, Scherer DC, Mckinsey TA, Hall SM, Qi XX,
Lee WY and Ballard DW. (1995). Mol. Cell. Biol., 15,
2809 ± 2818.
Brown K, Gerstberger S, Carlson L, Franzoso G and
Siebenlist U. (1995). Science, 267, 1485 ± 1488.
Brown K, Park S, Kanno T, Franzoso G and Siebenlist U.
(1993). Proc. Natl. Acad. Sci. USA, 90, 2532 ± 2536.
Carbone A, Gloghini A, Gattei V, Aldinucci D, Degan M,
De PP, Zagonel V and Pinto A. (1995a). Blood, 85, 780 ±
789.
Carbone A, Gloghini A, Gruss HJ and Pinto A. (1995b). Am.
J. Pathol., 146, 780 ± 781.
Chen ZJ, Hagler J, Palombella VJ, Melandri F, Scherer D,
Ballard D and Maniatis T. (1995). Genes Dev., 9, 1586 ±
1597.
Chiao PJ, Miyamoto S and Verma IM. (1994). Proc. Natl.
Acad. Sci. USA, 91, 28 ± 32.
Chomczynski P. (1992). Anal. Biochem., 201, 134 ± 139.
Chomczynski P and Sacchi N. (1987). Anal. Biochem., 162,
156 ± 159.
DiDonato J, Mercurio F, Rosette C, Wu-Li J, Suyang H,
Ghosh S and Karin M. (1996). Mol. Cell. Biol., 16, 1295 ±
1304.
Duyao MP, Buckler AJ and Sonenshein GE. (1990). Proc.
Natl. Acad. Sci. USA, 87, 4727 ± 4731.
Ernst MK, Dunn LL and Rice NR. (1995). Mol. Cell. Biol.,
15, 872 ± 882.
Gilmore TD. (1991). Trends Genet., 7, 318 ± 322.
Glavac D, Ravnik GM, O'Brien SJ and Dean M. (1994).
Hum. Genet., 93, 694 ± 696.
Gorschluter M, Bohlen H, Hasenclever D, Diehl V and
Tesch H. (1995). Ann. Oncol., 6, 477 ± 482.
Gruss HJ, Pinto A, Duyster J, Poppema S and Herrmann F.
(1997). Immunology Today, 18, 156 ± 163.
Gruss HJ, Hirschstein D, Wright B, Ulrich D, Caligiuri MA,
Barcos M, Strockbine L, Armitage RJ and Dower SK.
(1994). Blood, 84, 2305 ± 2314.
Gruss HJ, Ulrich D, Braddy S, Armitage RJ and Dower SK.
(1995). Eur. J. Immunol., 25, 2083 ± 2089.
Haskill S, Beg AA, Tompkins SM, Morris JS, Yurochko AD,
Sampson JA, Mondal K, Ralph P and Baldwin AJ. (1991).
Cell, 65, 1281 ± 1289.
Hatada EN, Naumann M and Scheidereit C. (1993). EMBO
J. 12, 2781 ± 2788.
Henkel T, Machleidt T, Alkalay I, Kronke M, Ben NY and
Baeuerle PA. (1993). Nature, 365, 182 ± 185.
Ja€ray E, Wood KM and Hay RT. (1995). Mol. Cell. Biol.,
15, 2166 ± 2172.
Kadin ME. (1994). Curr. Opin. Oncol., 6, 456 ± 463.
Koduru PR, Susin M, Schulman P, Catell D, Goh JC, Karp
L and Broome JD. (1993). Am. J. Hematol., 44, 117 ± 124.
LeBail O, Schmidt-Ullrich R and Israel A. (1993). EMBO J.,
12, 5043 ± 5049.
Mason DY, Banks PM, Chan J, Cleary ML, Delsol G, Peters
CD, Falini B, Gatter K, Grogan TM, Harris NL, Isaacson
PG, Ja€e ES, Knowles DM, Mullerhermelink HK, Pileri
S, Ralfkiaer H, Stein H and Warnke R. (1994). Am. J.
Surg. Pathol., 18, 526 ± 530.
Matthews JR and Hay RT. (1995). Inter. J. Biochem. Cell
Biol., 27, 865 ± 879.
Mellits KH, Hay RT and Goodbourn S. (1993). Nucl. Acids
Res., 21, 5059 ± 5066.
O'Grady JT, Stewart S, Lowrey J, Howie SE and Krajewski
AS. (1994). Am. J. Pathol., 144, 21 ± 26.
Palombella VJ, Rando OJ, Goldberg AL and Maniatis T.
(1994). Cell, 78, 773 ± 785.
Pazzagli M, Devine JH, Peterson DO and Baldwin TO.
(1992). Anal. Biochem., 204, 315 ± 323.
Perkins ND, Felzien LK, Betts JC, Leung K, Beach DH and
Nabel GJ. (1997). Science, 275, 523 ± 527.
Rodriguez MS, Michalopoulos I, Arenzana-Seisdedos F and
Hay RT. (1995). Mol. Cell. Biol., 15, 2413 ± 2419.
Rodriguez MS, Wright J, Thomson J, Thomas D, Baleux F,
Virelizier JL, Hay RT and Arenzana-Seisdedos F. (1996).
Oncogene, 12, 2425 ± 2435.
Ro€ M, Thomson J, Rodriguez MS, Jacque JM, Baleux F,
Arenzana-Seisdedos F and Hay RT. (1996). J. Biol.
Chem., 271, 7844 ± 7850.
Roulston A, Lin RT, Beauparlant P, Wainberg MA and
Hiscott J. (1995). Microbiol. Rev., 59, 481 ± 505.
Scheinman RI, Cogswell PC, Lofquist AK and Baldwin AJ.
(1995). Science, 270, 283 ± 286.
Scherer DC, Brockman JA, Chen Z, Maniatis T and Ballard
AJ. (1995). Proc. Natl. Acad. Sci. USA, 92, 11259 ± 11263.
Stein H and Hummel M. (1993). Leuk. Lymphoma, 10, 99 ±
102.
Sun S-C, Elwood J and Greene WC. (1996). Mol. Cell. Biol.,
16, 1058 ± 1065.
Sun SC, Ganchi PA, Ballard DW and Greene WC. (1993).
Science, 259, 1912 ± 1915.
Suyang H, Phillips R, Douglas I and Ghosh S. (1996). Mol.
Cell. Biol., 16, 5444 ± 5449.
Thompson JE, Phillips RJ, Erdjument BH, Tempst P and
Ghosh S. (1995). Cell, 80, 573 ± 582.
Traenckner EBM, Pahl HL, Henkel T, Schmidt KN, Wilk S
and Baeuerle PA. (1995). EMBO J., 14, 2876 ± 2883.
van Antwerp DJ, Martin SJ, Kafri T, Green DR and Verma
IM. (1996). Science, 274, 787 ± 789.
Wang CY, Mayo MW and Baldwin AS. (1996). Science, 274,
784 ± 787.
Whiteside ST, Ernst MK, Lebail O, Laurent-Winter C, Rice
N and Israel A. (1995). Mol. Cell. Biol., 15, 5339 ± 5345.
2139