Mutated cylindromatosis gene affects the functionalstate of dendritic cells

Matthias Bros1, Nadine Dexheimer1, Verena Besche1, Joumana Masri2,

Stefanie Trojandt1, Nadine Hovelmeyer2, Sonja Reissig2,

Ramin Massoumi3, Stephan Grabbe1, Ari Waisman2 and

Angelika B. Reske-Kunz1

1 Clinical Research Unit Allergology, Department of Dermatology, University Medical Center of

the Johannes Gutenberg-University, Mainz, Germany2 Institute for Molecular Medicine, University Medical Center of the Johannes Gutenberg-

University, Mainz, Germany3 Department of Laboratory Medicine, Lund University, Malmo University Hospital, Malmo,

Sweden

Cylindromatosis gene (CYLD) is a ubiquitously expressed deubiquitinating enzyme, which

interacts with members of the NF-jB signaling pathway and attenuates NF-jB and JNK

signaling. Here, we report that DC derived from transgenic mice, which solely express a

naturally occurring CYLD isoform (CYLDex7/8), display a higher content of nuclear RelB and

express elevated levels of NF-jB family members as well as of known NF-jB-target genes

comprising costimulatory molecules and pro-inflammatory cytokines, as compared with

WT DC. Accordingly, unstimulated CYLDex7/8 DC exhibited a significantly higher primary

allogenic T-cell stimulatory capacity than WT DC and exerted no tolerogenic activity.

Transduction of unstimulated CYLDex7/8 DC with relB-specific shRNA reduced their T-cell

stimulatory capacity. Treatment with the synthetic glucocorticoid dexamethasone known

to inhibit NF-jB and AP-1 activity reverted the pro-immunogenic phenotype and function

of CYLDex7/8 DC and re-established their tolerogenic function. DC derived from CYLD

knockout mice showed no functional alterations compared with WT DC. Therefore,

although complete loss of CYLD may be compensated for by other endogenous NF-jB

inhibitors, CYLDex7/8 acts in a dominant negative manner. Our findings raise the question

of whether genetic defects associated with increased NF-jB activity may result in distur-

bed maintenance of peripheral tolerance.

Key words: Autoimmunity . Costimulation . DC . Tolerance . Transgenic/knockout mice

Introduction

Under steady-state conditions, DC reside as sentinel cells in most

tissues and sample their micro-environment for antigens. In an

immature or semi-mature state, DC exert tolerogenic function

and interact with thymus-derived natural as well as induced

adaptive Treg to maintain peripheral tolerance [1]. Contact with

pathogen-derived components and pro-inflammatory cytokines

released by pathogen-sensing cells results in stimulation of DC

mediated by activation of NF-kB transcription factors [2]. These

transcription factors induce a transition of the DC phenotype

characterized by strong upregulation of costimulatory receptors

and pro-inflammatory cytokines [3]. In their acquired mature

state, DC constitute the most potent APC of the immune system

and may transiently override the function of Treg [4, 5] to ensure

the activation of T-effector cells.Correspondence: Dr. Matthias Brose-mail: mbros@uni-mainz.de

& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji-journal.eu

DOI 10.1002/eji.200939285 Eur. J. Immunol. 2010. 40: 2848–2857Matthias Bros et al.2848

NF-kB activity is under tight control of endogenous inhibitors

[6] and mutation-associated deregulation of NF-kB activity

results in functional defects of the immune system, immunode-

ficiencies in case of impaired NF-kB activity, and autoimmune

diseases characterized by chronic inflammation due to persistent

NF-kB activity [7]. In addition, hyperactivity of NF-kB is

frequently apparent in cancer cells and may be associated with

their increased resistance toward apoptotic signals and their

enhanced proliferation.

The tumor suppressor gene cylindromatosis gene (CYLD) was

identified as being mutated in many cases of familial cylin-

dromatosis [8]. Patients suffer from benign skin appendages,

which originate from hair follicle cells on the scalp and facial skin

[9]. However, CYLD is expressed rather ubiquitously [8] and in

different types of tumors CYLD was frequently found to be

downregulated [10], which indicates a general importance of

CYLD as a tumor suppressor.

CYLD is a deubiquitinating enzyme, which interacts with the

signaling components TRAF2, TRAF6, TRAF7 (TRAF, TNF recep-

tor-associated factor) and NF-kB essential modulator (NEMO)

[11–14]. Ubiquitination of these proteins triggers assembly of the

IkB kinase complex, which results in NF-kB activation, and

accordingly CYLD was shown to inhibit canonical NF-kB activation

via IkB kinase complex. Moreover, CYLD controls nuclear entry of

the proto-oncogene Bcl3 [15], which forms heterodimers with

NF-kB family members to exert transcriptional control [16].

Recently, a transgenic mouse strain (CYLDex7/8) engineered to

solely express a naturally occurring CYLD isoform devoid of NEMO

and TRAF2 interaction domains but retaining the ability to interact

with Bcl3 was shown by us to harbor an expanded population of

mature B cells displaying higher activity of NF-kB and enhanced

nuclear accumulation of Bcl3 [17].

Here, we show that unstimulated BM-derived DC (BMDC)

isolated from CYLDex7/8 mice showed an increased content of

nuclear RelB which was associated with higher expression of DC-

activation markers, and induced stronger allogenic T-cell prolif-

eration than WT BMDC. Accordingly, CYLDex7/8 BMDC were

unable to induce the differentiation of naıve T cells to Treg in

vitro. Transduction of CYLDex7/8 BMDC with relB-specific shRNA

reduced their T-cell stimulatory capacity. Treatment of CYLDex7/8

BMDC with the synthetic glucocorticoid dexamethasone (DEX)

previously shown to inhibit NF-kB activity, reverted the hyper-

activity of CYLDex7/8 BMDC and re-established their tolerogenic

function.

Results

Murine BMDC express two isoforms of CYLD

It was recently shown that murine B cells expressed a CYLD

isoform lacking exons 7 and 8 (termed CYLDex7/8) [17] that

encode TRAF/NEMO interaction domains (Fig. 1A). Transgenic

mice (C57BL/6 background) engineered to express only the short

isoform CYLDex7/8 showed a strong expansion of the mature

B-cell population due to constitutive NF-kB activation. As shown

in Fig. 1B, BMDC of C57BL/6 origin expressed mRNA of the two

CYLD isoforms both at immature state and following stimulation

with LPS. As expected, BMDC derived from transgenic CYLDex7/8

mice expressed only the shorter CYLD isoform.

CYLDex7=8 BMDC display altered function

Compared with WT BMDC, unstimulated CYLDko BMDC displayed

lower MHCII expression, but similar surface levels of the

costimulatory molecules CD40, CD80 and CD86 (Fig. 1C and D).

On the other hand, CYLDex7/8 BMDC expressed comparable levels

of MHCII and CD80, but moderately higher levels of CD40 and

CD86 as shown by the MFI. Upon stimulation with LPS, the

fraction of cells expressing MHCII at high level was enhanced in all

three groups (Fig. 1C), and the average MHCII expression level

was elevated (Fig. 1D). The expression of the costimulatory

molecules monitored was enhanced in response to stimulation as

well. Interestingly, stimulated CYLDex7/8 BMDC displayed lower

CD40 expression than WT BMDC, but higher CD80 expression.

In MLR, both WT and CYLDko BMDC showed comparable

allogeneic T-cell stimulatory capacity under basal conditions,

which was significantly increased upon LPS stimulation (Fig. 1E).

This finding suggests that the absence of NF-kB inhibitory func-

tion of CYLD (CYLDko) may be compensated for by other endo-

genous NF-kB inhibitors [6]. On the contrary, unstimulated

CYLDex7/8 BMDC were almost as potent as stimulated WT BMDC

in inducing T-cell proliferation (Fig. 1F), suggesting that sole

expression of truncated CYLD lacking TRAF and NEMO interac-

tion domains, but retaining catalytic activity, counteracts

compensatory mechanisms. However, the presence of a WT CYLD

allele (CYLDwt/ex7/8) was sufficient to counteract the excessive

stimulatory potency observed in unstimulated CYLDex7/8 BMDC

(Fig. 1G).

CYLDex7=8 BMDC display a pro-immunogenic geneexpression signature

To analyze more thoroughly the molecular base for the stronger

APC activity of CYLDex7/8 BMDC as compared with WT BMDC at

unstimulated state, we monitored the expression of a panel of

well-known genes whose expression state compared with

unstimulated WT BMDC, unstimulated CYLDex7/8 BMDC showed

significantly higher mRNA expression of NF-kB family member

c-Rel as well as enhanced expression of relB (Fig. 2A), both of

which mediate upregulation of costimulatory molecules and pro-

inflammatory cytokines [18, 19]. Concomitantly, unstimulated

CYLDex7/8 BMDC displayed higher mRNA expression of CD40,

OX40L and SLAM, (Fig. 2B), but significantly lower expression of

mRNA encoding for LIGHT and the coinhibitory molecules B7-H3

and B7-DC (Fig. 2C). All the pro-inflammatory cytokine mRNA

species analyzed showed higher expression in CYLDex7/8 BMDC

than in WT BMDC at basal state (Fig. 2D). Interestingly, mRNA

Eur. J. Immunol. 2010. 40: 2848–2857 Immunomodulation 2849

& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji-journal.eu

Figure 1. CYLD gene mutations affect the phenotype and APC activity of DC. (A) Schematic representation of the murine CYLD gene. MurineB cells were reported to express a CYLD isoform lacking exons 7 and 8 (CYLDex7/8) [17], which encodes binding sites for TRAF2 and NEMO.(B) Expression of CYLD isoforms was assessed by conventional PCR using cDNA derived from WT and CYLDex7/8 BMDC. The additional bandapparent in each reaction was previously shown to constitute an unspecific PCR product [17]. (C) Phenotype of BMDC derived from WT, CYLDex7/8

and CYLDko mice. BMDC were stained with Ab specific for the indicated surface markers. Histograms represent specific Ab binding tounstimulated (dashed line) and LPS stimulated (solid line) DC. Binding of isotype-matched control Ab is indicated (dotted line). Graphs arerepresentative of two to three independent experiments. (D) MFI corrected for isotype background. Data show mean7SEM of two to fourindependent experiments and were normalized to the MFI of unstimulated WT BMDC arbitrarily set to one. (E–G) Titrated numbers of irradiatedBMDC were cocultured with 3� 105 nylon wool-enriched T cells of BALB/c mice for 4 days in 0.2 mL culture medium. T-cell proliferation wasassessed by uptake of [3H] thymidine for the final 16 h of culture. Data show mean7SEM of triplicate cultures and are representative of two to fourindependent experiments. (D–G) Statistically significant differences: 1(unstimulated CYLD mutated DC and stimulated WT DC versusunstimulated WT DC), ] (CYLD mutated DC at stimulated versus unstimulated state), $(stimulated CYLD mutated DC versus stimulated WT DC).(�] $po0.05; ��]] $$po0.01; ���]]] $$$po0.001).

Eur. J. Immunol. 2010. 40: 2848–2857Matthias Bros et al.2850

& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji-journal.eu

encoding the DC maturation marker Fascin (Fscn) was signifi-

cantly increased in CYLDex7/8 BMDC as well (Fig. 2E). Upon

stimulation with LPS, CYLDex7/8 BMDC exhibited higher mRNA

expression of the NF-kB family members p52 and relB compared

with stimulated WT BMDC. Further, the mRNA levels of IL-1b

and TGF-b1 and of the DC maturation marker Fascin were

increased, but decreased mRNA expression in case of LIGHT and

the coinhibitory receptors B7-H3 and B7-DC was seen.

NF-kB transcription factors are complexed in the cytoplasm

and exert regulatory activity upon translocation into the nucleus.

Figure 2. CYLDex7/8 DC exhibit a pro-immunogenic gene expression signature and enhanced accumulation of RelB in the nucleus. BMDC werederived from WT and CYLDex7/8 mice and aliquots were stimulated with LPS. The relative mRNA expression level of genes encoding (A) NF-kBfamily members, (B, C) costimulatory/coinhibitory receptors, (D) cytokines and (E) other molecules was assessed by quantitative real-time PCR.Data indicate relative differences in mRNA expression of LPS-stimulated WT and of CYLDex7/8 BMDC at unstimulated and stimulated statecompared with unstimulated WT BMDC (dashed line). Data show mean7SEM of three to five independent experiments performed in duplicate.�po0.05, statistically significant differences versus unstimulated WT DC. (F) Protein extracts were probed for RelB and c-Rel protein content in thecytoplasmic and nuclear compartments. b-Actin detection served as loading control. Data are representative of two (c-Rel) and three (RelB)experiments. (G–I) BMDC of WT and CYLDex7/8 genotype were transduced on days 5 and 6 of culture with lentiviral particles harboring either emptyvector as control (Ctrl) or relB shRNA (sh relB). On day 8, BMDC were harvested and subjected to experiments. (G) Protein extracts of transducedBMDC populations were probed for RelB protein content. b-Tubulin detection served as loading control. Data presented are representative of twoindependent experiments. (H) Titrated numbers of irradiated BMDC of either population were cocultured with 3� 105 nylon wool-enriched T cellsof BALB/c mice for 4 days as described (see legend of Fig. 1E–G). Data show mean7SEM of triplicate cultures and are representative of threeindependent experiments. (I) Quantification of relative allogeneic T-cell proliferation induced by 1.9� 104 differentially transduced BMDC derivedfrom WT and CYLDex7/8 mice cocultured with 3� 105 BALB/c T cells. T-cell proliferation induced by unstimulated WT BMDC transduced withcontrol vector (Ctrl) was arbitrarily set to one. Data show mean7SEM of three independent experiments performed in triplicate. Statisticallysignificant differences: �(versus WT DC transduced with Ctrl vector), 1(CYLDex7/8 DC transduced with relB shRNA versus Ctrl vector), ](CYLDex7/8 DCversus WT DC, each transduced with relB shRNA). �]1po0.05; ��11po0.01; ���po0.001).

Eur. J. Immunol. 2010. 40: 2848–2857 Immunomodulation 2851

& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji-journal.eu

The activity of the NF-kB transcription factor family members

relB and c-Rel correlates with the activation state of DC [20]. Due

to our finding of higher relB and c-Rel mRNA levels in CYLDex7/8

BMDC than in WT BMDC, we assessed their expression on the

protein level by Western blot analysis. WT and CYLDex7/8 BMDC

harbored comparable amounts of RelB in the cytoplasm,

whereas the amounts of cytoplasmic c-Rel were very low

in both cell types (Fig. 2F). Stimulation with LPS induced

enhanced expression of both NF-kB proteins. Most interestingly,

in the nuclear compartment unstimulated CYLDex7/8 BMDC

displayed a higher RelB level than WT BMDC, whereas c-Rel was

apparent at a low level in either BMDC population. In LPS-

stimulated BMDC, a general increase in nuclear RelB and c-Rel

was noted.

Li et al. [21] have shown that inhibition of RelB activity by

RNA interference in BMDC diminished their T-cell stimulatory

capacity. Based on our finding of a higher expression of relB both

on the mRNA and on the protein level in unstimulated CYLDex7/8

BMDC, we asked for the relevance of this NF-kB member for the

observed T-cell hyperstimulatory capacity of these cells. Inhibi-

tion of relB by transduction of unstimulated BMDC with specific

shRNA attenuated RelB expression in BMDC of both genotypes

(Fig. 2G), and resulted in impaired T-cell stimulatory capacity of

CYLDex7/8 BMDC (Fig. 2H and I).

CYLDex7=8 BMDC treated with DEX exert tolerogenicfunction

CYLD was shown to limit the activity of various AP-1 and NF-kB

transcription factors besides relB [6, 22]. Therefore, we tested,

whether glucocorticoids (GC), which were shown to broadly

inhibit NF-kB and AP-1 activity and are frequently prescribed in

the treatment of autoimmune diseases and severe allergic

diseases [23], were able to diminish the T-cell hyperstimulatory

capacity of CYLDex7/8 BMDC more efficiently than relB shRNA.

GC activates the normally retained GC receptor, which exerts its

anti-inflammatory effects by direct inhibition of NF-kB and AP-1

transcription factors through protein–protein interaction. In

addition, activated GC receptor binds to GC-response elements

and diminishes the expression of genes encoding pro-inflamma-

tory products.

We treated DC progenitors from day 3 of culture onward with

the synthetic GC DEX. After treatment with DEX, with or without

LPS stimulation, BMDC of WT or CYLDex7/8 origin expressed

MHCII as well as the costimulatory molecules CD40, CD80 and

CD86 at significantly reduced levels as compared with BMDC

cultivated in the absence of DEX (Fig. 3A and B).

Compared with unstimulated WT BMDC, in many cases DEX-

treated unstimulated BMDC of either genotype displayed lowered

expression of mRNA encoding for different NF-kB family

members (Fig. 3C), costimulatory (Fig. 3D) and coinhibitory

(Fig. 3E) molecules, pro-inflammatory cytokines (Fig. 3F) as well

as Fscn and CCR7 (Fig. 3G). However, DEX treatment induced

mRNA expression of LIGHT in BMDC of either genotype and of

B7-H2 in CYLDex7/8 BMDC (Fig. 3E). DEX treatment, in general,

prevented stimulation-dependent increases of NF-kB family

members as well as of costimulatory molecules and diminished

LPS-mediated upregulation of pro-inflammatory cytokine mRNA

species, most evident in case of IL-1b and IL12p40 mRNA

(compare with the data in Fig. 2). In addition, expression of Fscn

and CCR7 was strongly inhibited.

In accordance with their pro-tolerogenic phenotype, both WT

and CYLDex7/8 DEX-DC did not mediate efficient allogeneic T-cell

stimulation (Fig. 4A and B). Their T-cell stimulatory capacity was

lower than the moderate stimulatory potential exerted by unsti-

mulated WT BMDC (Fig. 4C; n 5 3). The same effect was noted

for CYLDko BMDC (data not shown).

Unstimulated WT BMDC were shown to exert tolerogenic

activity [1, 4]. Accordingly, T cells primed by unstimulated

WT BMDC inhibited proliferation of naıve T cells. On the

contrary, unstimulated CYLDex7/8 BMDC, which displayed a

strongly enhanced T-cell stimulatory capacity, did not exert

tolerogenic function (Fig. 4D). In accordance with our finding of

a drastically reduced allogeneic T-cell stimulatory potential of

DEX-DC, T cells primed by CYLDex7/8 DEX-DC exerted suppres-

sive activity on cocultured naıve T cells, although at lower

potency than T cells primed by unstimulated WT BMDC or WT

DEX-DC.

Discussion

Although the tumor suppressor gene CYLD was initially identified

as a candidate gene mutated in patients suffering from familial

cylindromatosis [8], the finding that its expression is strongly

diminished in tumors of distinct origin like colon and hepatocel-

lular carcinomas [10] suggested a more general role of CYLD.

Since the aa sequences of human and murine CYLD are highly

conserved (around 95% homology), several groups have gener-

ated CYLD knockout mouse strains in order to elucidate its role

in vivo.

Despite several in vitro studies which congruently showed that

CYLD inhibits NF-kB [11–13] and JNK [22] activity by interfering

with signaling molecules TRAF2, TRAF6 and NEMO as well as

Bcl3 [15], the various CYLDko mouse strains showed marked

phenotypical differences [24–26]. Recently, we identified a

naturally occurring CYLD isoform, which lacks exons 7 and 8

(CYLDex7/8) that encode aa 394–557 which encompass the

binding sites for TRAF2 and NEMO. Notably, transgenic mice

engineered to solely express CYLDex7/8 showed very similar B-cell

defects as noted for the CYLDko mouse strain described by Jin

et al. [27].

Taken together, CYLD seems to be critically involved in the

differentiation of T and B lymphocytes and controls their state of

activation. In light of this concept, we asked whether CYLD

function may influence APC function as well. We focused on DC

because they are known to exert tolerogenic function in imma-

ture or semi-mature states [1], and to acquire potent T-cell

stimulatory function upon activation by pathogen-derived signals

Eur. J. Immunol. 2010. 40: 2848–2857Matthias Bros et al.2852

& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji-journal.eu

Figure 3. DEX induces a pro-tolerogenic phenotype in CYLDex7/8 DC. BMDC were derived from WT and CYLDex7/8 mice in the absence or presenceof 10�6 M DEX and aliquots were stimulated with LPS. (A) DC were stained with Ab specific for the indicated surface markers. Histograms representspecific Ab binding to unstimulated (black dashed line) and stimulated DC (black solid line), and to DEX-treated unstimulated (gray dashed line)and stimulated DC (gray solid line). Binding of isotype-matched control Ab is indicated (black stippled line). Graphs are representative of two tothree independent experiments. (B) MFI corrected for isotype background. Data show mean7SEM of two to three independent experiments andwere normalized to the MFI of unstimulated WT BMDC arbitrarily set to one. Statistically significant differences: �(differentially treated DCpopulation versus untreated DC of the same genotype), 1(CYLDex7/8 DC versus WT DC at corresponding state), ](DEX-treated DC at stimulated versusunstimulated state) (�]1po0.05; ��11po0.01; ���po0.001). (C–G) The gene expression signature of differentially treated BMDC was assessed by real-time PCR as described in the legend of Fig. 2. Data represent the mean7SEM of three to five independent experiments performed in duplicate.�po0.05 DEX-DC (WT and CYLDex7/8) versus unstimulated WT DC (dashed line) and between groups as indicated.

Eur. J. Immunol. 2010. 40: 2848–2857 Immunomodulation 2853

& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji-journal.eu

which activate NF-kB [2] and JNK [28], potentially controlled by

CYLD.

The detection of WT CYLD and CYLDex7/8 mRNA in WT BMDC

both at immature state and after stimulation with LPS suggested

that both isoforms indeed participate in the control of DC func-

tion. To examine their functional properties in DC, we examined

BMDC generated in parallel from WT, CYLDko and CYLDex7/8 BM

cells. We noted elevated surface expression of CD40 and CD86 in

unstimulated CYLDex7/8 BMDC in comparison with WT BMDC.

In addition, we observed higher mRNA expression levels of

various costimulatory molecules and pro-inflammatory

cytokines in CYLDex7/8 BMDC than in WT BMDC at basal state,

which were enhanced in CYLDex7/8 BMDC in response to stimu-

lation. Moreover, the expression of coinhibitor-encoding mRNA

species was significantly lower in CYLDex7/8 BMDC at either state

of activation.

Accordingly, unstimulated CYLDex7/8 BMDC mediated strong

T-cell proliferation at similar extent as fully stimulated WT

BMDC. This hyperactive state was not apparent in BMDC derived

from CYLDko or heterozygous CYLDwt/ex7/8 mice, which implies

that the stimulatory activity of CYLDex7/8 is under tight control of

WT CYLD, probably exerted at least in part by direct competition

of both isoforms for the same substrate(s).

Our findings suggest that loss of WT CYLD is largely

compensated by other endogenous NF-kB inhibitors [6], which

on the other hand were obviously unable to compensate the

biological function of CYLDex7/8 which still binds at least to Bcl3

and exerts catalytic activity [17].

Figure 4. DEX-treatment counteracts T-cell hyperstimulatory capacity of CYLDex7/8 BMDC and re-establishes their tolerogenic function. BMDCwere differentiated in the absence or presence of 10�6 M DEX and aliquots were stimulated with LPS. (A and B) The primary allogeneic T-cellstimulatory capacity of BMDC was assessed as described in the legend of Fig. 1. Data show mean7SEM of triplicate cultures and are representativeof three independent experiments. (C) Quantification of relative allogeneic T-cell proliferation induced by 5� 104 differentially treated BMDCderived from WT and CYLDex7/8 mice cocultured with 3� 105 BALB/c T cells. T-cell proliferation induced by unstimulated WT BMDC was arbitrarilyset to one. Data show mean7SEM of three independent experiments. Statistically significant differences: ](DEX-treated versus untreated DC),�(stimulated versus unstimulated DC), 1(DEX-treated DC at stimulated versus unstimulated state), $(stimulated DEX-treated DC versus stimulateduntreated DC) and y(unstimulated CYLDex7/8 DC versus unstimulated WT DC) (�]$1po0.05; ��]]$$11yypo0.01; ���]]] $$$ 111po0.001). (D) BMDC (106)derived from WT and CYLDex7/8 mice were irradiated and cocultured with nylon wool-enriched T cells (6� 106) of BALB/c mice in 4 mL volume inwells of 6-well cluster plates. After 7 days of coculture, primed T cells (pTC) were harvested. Graded numbers of pTC were cocultured with 3� 105

naıve BALB/c T cells (nTC) and 3� 105 C57BL/6 spleen cells in 0.2 mL volume for 6 days. Data show mean7SEM of triplicate cultures and arerepresentative of two independent experiments. The dashed line indicates the proliferation of 3� 105 nTC stimulated in the absence of pTC.Briefly, 6� 105 nTC stimulated in the absence of pTC yielded 29.25073.511 cpm. Statistically significant differences: �(nTC1pTC versus nTC), ](nTC1pTC (primed by CYLDex7/8 DC left untreated or DEX treated) versus nTC1pTC (primed by WT DC at corresponding state of treatment)), 1(nTC1

pTC (primed by DEX-treated DC of either origin) versus nTC1pTC (primed by untreated DC of either origin)) (]1po0.05; ��]]11po0.01; ���]]]po0.001).

Eur. J. Immunol. 2010. 40: 2848–2857Matthias Bros et al.2854

& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji-journal.eu

Interestingly, unstimulated CYLDex7/8 BMDC showed higher

transcript expression levels of NF-kB family members relB and

c-Rel than WT BMDC, and displayed a higher nuclear RelB

content. Similarly, CYLDex7/8 B cells were shown by us to express

higher amounts of RelB protein than WT B cells, but the relB

mRNA level remained unaltered in these B cells, suggesting post-

transcriptional regulation [17]. In line with our finding of higher

nuclear RelB levels in CYLDex7/8 than in WT BMDC at basal state,

we noted elevated expression of costimulatory molecules and

pro-inflammatory cytokines in unstimulated CYLDex7/8 BMDC on

the transcriptional and protein level as compared with WT

BMDC. Accordingly, in contrast to unstimulated WT BMDC,

unstimulated CYLDex7/8 BMDC induced hyperproliferation of

interacting allogeneic T cells and lacked suppressive activity. In

accordance with the prominent nuclear localization of RelB in

CYLDex7/8 BMDC, but not WT BMDC at unstimulated state, we

achieved partial inhibition of the T-cell hyperstimulatory activity

of CYLDex7/8 BMDC by their transduction with lentivirus encod-

ing relB-specific shRNA. Our results suggest that in CYLDex7/8

BMDC other factors besides RelB may contribute to the pro-

immunogenic phenotype and T-cell hyperstimulatory activity of

CYLDex7/8 BMDC. This notion is in line with the reports showing

that CYLD limits the activity of various NF-kB and AP-1 tran-

scription factors [6, 22]. A more extensive inhibition of NF-kB

and AP-1 family members is achieved by GC [23]. Therefore, in

an attempt to re-establish tolerogenic function of CYLDex7/8

BMDC, we have employed the synthetic GC DEX. We have

recently shown that DEX-treated DC acquired a tolerogenic

phenotype and induced Treg [29]. Cotreatment of BMDC of

either genotype with DEX diminished expression of NF-kB family

members, costimulatory molecules and pro-inflammatory cyto-

kines. In accordance with their DEX-mediated pro-tolerogenic

phenotype, DEX-treated BMDC of either origin displayed low APC

activity, and CYLDex7/8 BMDC regained the ability to induce Treg.

In an analogous manner, Brummelkamp et al. [12] have shown

that hyperactivation of NF-kB induced by CYLD gene silencing

can be compensated by applying the NF-kB inhibitor aspirin.

Our study indicates that germline mutations affecting NF-kB

activity may determine the functional state of DC, if not compen-

sated by other endogenous control mechanisms. In clinical settings,

it may be possible to overcome these defects by generating DC

ex vivo, educated by applying suitable drugs to prevent DC hyper-

activity [30] (as shown here) or to induce strong DC activation,

followed by reinfusion of antigen-loaded DC into the patient.

Materials and methods

Mice

Female C57BL/6 and BALB/c mice as well as female CYLD

mutant mice (knockout and ex7/8) on C57BL/6 background [15,

17] were bred and maintained in the Central Animal Facilities of

the University of Mainz under specific pathogen-free conditions

on a standard diet. The ‘‘Principles of Laboratory Animal Care’’

(NIH publication no. 85-23, revised 1985) were followed.

Cells

BMDC were generated from BM progenitors of mice as first

described by Scheicher et al. [31] with some modifications [32]

using BMDC medium (IMDM with 10% FBS; (PAA, Colbe,

Germany), 2 mM L-glutamine (Biochrom AG, Berlin, Germany),

100 U/mL penicillin, 100mg/mL streptomycin (Gibco, Paisly, UK)),

supplemented with 5% of GM-CSF containing cell culture super-

natant [33] (a kind gift from Dr. B. Stockinger, MRC National

Institute for Medical Research, Mill Hill, London). Where indicated

DEX (Sigma-Aldrich, Deisenhofen, Germany) was added on day 3

of culture (10�6 M). On days 7 or 8 of culture, part of the

nonadherent cells was harvested (485% CD11c1 cells), reseeded

on 100-mm2 Cellstar tissue-culture dishes (Greiner) and stimu-

lated with LPS (1mg/mL; Calbiochem, Schwalbach, Germany) for

24 h. GM-CSF and DEX were replenished in the course of change of

culture media on days 3 and 6, and were added to cell cultures

stimulated with LPS. HEK 293T cells used for production of

lentiviral particles were kept as described previously [34].

Flow cytometry

BMDC (5� 105) were washed in staining buffer (PBS/2% FBS).

To block Fc receptor-mediated staining, cells were incubated with

rat anti-mouse CD16/CD32 (2.4.G2) (Dianova, Hamburg,

Germany) for 15 min on ice. Afterward, cells were incubated

with FITC-conjugated rat mAb recognizing MHC class II I-A/I-E

(2G9), CD80 (1G10), CD86 (GL1) (BD Pharmingen, San Diego,

CA, USA) and CD40 (1C10) (Southern Biotechnology, Birming-

ham, AL, USA) for 20 min on ice. Appropriate isotype controls

were used. Flow cytometric analysis was performed using a

FACScan flow cytometer (BD Biosciences) equipped with

CellQuest Software.

Western blot analysis

Western blot analysis was carried out as described previously

[17] using the following Ab: anti-RelB (sc-226) and anti-c-Rel

(sc-71) (Santa Cruz, Santa Cruz, CA, USA), as well as anti-b-Actin

(AC-15; Sigma-Aldrich) and b-tubulin (ab21058; Abcam,

Cambridge, UK). Preparation of nuclear extracts was performed

as described previously [35]. Nuclear and cytoplasmic fractions

were prepared according to the standard procedures [36].

Lentiviral transduction

HEK 293T cells were cotransfected with HIV-1-derived packaging

plasmid pCMVDR8.91 and the vesicular stomatitis virus

Eur. J. Immunol. 2010. 40: 2848–2857 Immunomodulation 2855

& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji-journal.eu

glycoprotein G-encoding plasmid pMD2.G, and either relB shRNA

encoding or empty TRIPZ vector (Open Biosystems, Huntsville,

AL, USA) to generate lentiviral particles as described previously

[34]. Unstimulated BMDC were transduced on days 5 and 6 as

described and were used for experiments on day 8.

MLR

BALB/c splenocytes were depleted of erythrocytes by incubation

in lysis buffer (155 mM NH4Cl, 10 mM KHCO3, 100 mM EDTA-

sodium, pH 7.4) for 1 min and were passed over nylon wool

columns to enrich T cells. T cells (3� 105) were cocultured for 4

days on flat-bottom 96-well tissue culture plates with graded

numbers of irradiated (30 Gy) BMDC in 200 mL of BMDC

medium. Cell proliferation was assessed by the uptake of [3H]

thymidine (0.25 mCi/well) for the last 16 h of culture. Cells were

harvested onto glass fiber filters and retained radioactivity was

measured in a liquid scintillation counter (1205 Betaplate, LKB

Wallac, Turku, Finland).

T-cell restimulation and suppression assays

BMDC (106/well) were cocultured on 6-well tissue culture plates

with nylon wool-enriched BALB/c T cells (6� 106/well) in a

volume of 4 mL for 7 days, and prestimulated T cells were

harvested. C57BL/6 splenocytes, depleted of erythrocytes and

g-irradiated (30 Gy), were cocultured with freshly isolated

BALB/c T cells (3�105 each) and graded numbers of prestimu-

lated T cells for 6 days. T-cell proliferation was assessed as

described above.

PCR and real-time PCR analysis

Total RNA was isolated from at least 5� 105 BMDC by using the

RNeasy Plus Mini kit (Qiagen, Hilden, Germany) as recom-

mended. RNA was reverse-transcribed applying a 1:1 mix of

Oligo-dT and random hexamer primers by using iScript (Bio-Rad,

Munich, Germany) as recommended. Expression of CYLDex7/8

and the corresponding WT sequence was detected by PCR using

primers binding to sequences located in neighboring exons as

described previously [17]. Primer sequences have been described

[29] and primers were purchased from Operon (Koln, Germany).

The house-keeping gene ubiquitin C served as internal control.

Real-time PCR reactions included 200 ng of cDNA and SYBR-

Green mastermix (ABgene, Hamburg, Germany) and were

performed and analyzed as described previously [29].

Statistical analysis

Data were analyzed for statistically significant differences by

applying Student’s t-test.

Acknowledgements: This work was supported by the

Deutsche Forschungsgemeinschaft, Sonderforschungsbereich

548 (to M. B., A. W. and A. B. R. -K.) and 432 (to A. W.), and

the Stiftung Rheinland-Pfalz fur Innovation (15212-386261/761)

(to A. B. R. -K.).

Conflict of interest: The authors declare no financial or

commercial conflict of interest.

References

1 Tang, Q. and Bluestone, J. A., The Foxp31 regulatory T cell: a jack of all

trades, master of regulation. Nat. Immunol. 2008. 9: 239–244.

2 Hofer, S., Rescigno, M., Granucci, F., Citterio, S., Francolini, M. and

Ricciardi-Castagnoli, P., Differential activation of NF-kappa B subunits in

dendritic cells in response to Gram-negative bacteria and to lipopoly-

saccharide. Microbes Infect. 2001. 3: 259–265.

3 Adams, S., O’Neill, D. W. and Bhardwaj, N., Recent advances in dendritic

cell biology. J. Clin. Immunol. 2005. 25: 87–98.

4 Fehervari, Z. and Sakaguchi, S., Control of Foxp31 CD251CD41 regulatory

cell activation and function by dendritic cells. Int. Immunol. 2004. 16:

1769–1780.

5 Ahn, J. S., Krishnadas, D. K. and Agrawal, B., Dendritic cells partially

abrogate the regulatory activity of CD41CD251 T cells present in the

human peripheral blood. Int. Immunol. 2007. 19: 227–237.

6 Chen, F., Endogenous inhibitors of nuclear factor-kappaB, an opportunity

for cancer control. Cancer Res. 2004. 64: 8135–8138.

7 Courtois, G. and Gilmore, T. D., Mutations in the NF-kappaB signaling

pathway: implications for human disease. Oncogene 2006. 25: 6831–6843.

8 Bignell, G. R., Warren, W., Seal, S., Takahashi, M., Rapley, E., Barfoot, R.,

Green, H. et al., Identification of the familial cylindromatosis tumour-

suppressor gene. Nat. Genet. 2000. 25: 160–165.

9 Massoumi, R. and Paus, R., Cylindromatosis and the CYLD gene: new

lessons on the molecular principles of epithelial growth control. Bioessays

2007. 29: 1203–1214.

10 Hellerbrand, C., Bumes, E., Bataille, F., Dietmaier, W., Massoumi, R. and

Bosserhoff, A. K., Reduced expression of CYLD in human colon and

hepatocellular carcinomas. Carcinogenesis 2007. 28: 21–27.

11 Trompouki, E., Hatzivassiliou, E., Tsichritzis, T., Farmer, H., Ashworth, A.

and Mosialos, G., CYLD is a deubiquitinating enzyme that negatively

regulates NF-kappaB activation by TNFR family members. Nature 2003.

424: 793–796.

12 Brummelkamp, T. R., Nijman, S. M., Dirac, A. M. and Bernards, R., Loss of

the cylindromatosis tumour suppressor inhibits apoptosis by activating

NF-kappaB. Nature 2003. 424: 797–801.

13 Kovalenko, A., Chable-Bessia, C., Cantarella, G., Israel, A., Wallach, D.

and Courtois, G., The tumour suppressor CYLD negatively regulates NF-

kappaB signalling by deubiquitination. Nature 2003. 424: 801–805.

14 Yoshida, H., Jono, H., Kai, H. and Li, J. D., The tumor suppressor

cylindromatosis (CYLD) acts as a negative regulator for toll-like receptor 2

signaling via negative cross-talk with TRAF6 and TRAF7. J. Biol. Chem.

2005. 280: 41111–41121.

Eur. J. Immunol. 2010. 40: 2848–2857Matthias Bros et al.2856

& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji-journal.eu

15 Massoumi, R., Chmielarska, K., Hennecke, A., Pfeifer, A. and Fassler, R.,

Cyld inhibits tumor cell proliferation by blocking Bcl-3-dependent NF-

kappaB signaling. Cell 2006. 125: 665–677.

16 Hayden, M. S., West, A. P. and Ghosh, S., NF-kappaB and the immune

response. Oncogene 2006. 25: 6758–6780.

17 Hovelmeyer, N., Wunderlich, F. T., Massoumi, R., Jakobsen, C. G., Song, J.,

Worns, M. A., Merkwirth, C. et al., Regulation of B cell homeostasis and

activation by the tumor suppressor gene CYLD. J. Exp. Med. 2007. 204:

2615–2627.

18 Boffa, D. J., Feng, B., Sharma, V., Dematteo, R., Miller, G., Suthanthiran,

M., Nunez, R. and Liou, H. C., Selective loss of c-Rel compromises

dendritic cell activation of T lymphocytes. Cell. Immunol. 2003. 222:

105–115.

19 Li, M., Zhang, X., Zheng, X., Lian, D., Zhang, Z. X., Ge, W., Yang, J. et al.,

Immune modulation and tolerance induction by RelB-silenced dendritic

cells through RNA interference. J. Immunol. 2007. 178: 5480–5487.

20 Neumann, M., Fries, H., Scheicher, C., Keikavoussi, P., Kolb-Maurer, A.,

Brocker, E., Serfling, E. and Kampgen, E., Differential expression of Rel/

NF-kappaB and octamer factors is a hallmark of the generation and

maturation of dendritic cells. Blood 2000. 95: 277–285.

21 Li, M., Zhang, X., Zheng, X., Lian, D., Zhang, Z. X., Ge, W., Yang, J. et al.,

Immune modulation and tolerance induction by RelB-silenced dendritic

cells through RNA interference. J. Immunol. 2007. 178: 5480–5487.

22 Reiley, W., Zhang, M. and Sun, S. C., Negative regulation of JNK signaling

by the tumor suppressor CYLD. J. Biol. Chem. 2004. 279: 55161–55167.

23 Smoak, K. A. and Cidlowski, J. A., Mechanisms of glucocorticoid receptor

signaling during inflammation. Mech. Ageing Dev. 2004. 125: 697–706.

24 Reiley, W. W., Zhang, M., Jin, W., Losiewicz, M., Donohue, K. B., Norbury,

C. C. and Sun, S. C., Regulation of T cell development by the

deubiquitinating enzyme CYLD. Nat. Immunol. 2006. 7: 411–417.

25 Reiley, W. W., Jin, W., Lee, A. J., Wright, A., Wu, X., Tewalt, E. F., Leonard,

T. O. et al., Deubiquitinating enzyme CYLD negatively regulates the

ubiquitin-dependent kinase Tak1 and prevents abnormal T cell

responses. J. Exp. Med. 2007. 204: 1475–1485.

26 Zhang, J., Stirling, B., Temmerman, S. T., Ma, C. A., Fuss, I. J., Derry, J. M.

and Jain, A., Impaired regulation of NF-kappaB and increased suscept-

ibility to colitis-associated tumorigenesis in CYLD-deficient mice. J. Clin.

Invest. 2006. 116: 3042–3049.

27 Jin, W., Reiley, W. R., Lee, A. J., Wright, A., Wu, X., Zhang, M. and Sun, S.

C., Deubiquitinating enzyme CYLD regulates the peripheral development

and naive phenotype maintenance of B cells. J. Biol. Chem. 2007. 282:

15884–15893.

28 Nakahara, T., Uchi, H., Urabe, K., Chen, Q., Furue, M. and Moroi, Y., Role

of c-Jun N-terminal kinase on lipolpolysaccharide induced maturation of

human monocyte-derived dendritic cells. Int. Immunol. 2004. 16:

1701–1709.

29 Bros, M., Jahrling, F., Renzing, A., Wiechmann, N., Dang, N. A., Sutter, A.,

Ross, R. et al., A newly established murine immature dendritic cell line

can be differentiated into a mature state, but exerts tolerogenic function

upon maturation in the presence of glucocorticoid. Blood 2007. 109:

3820–3829.

30 Morelli, A. E. and Thomson, A. W., Tolerogenic dendritic cells and the

quest for transplant tolerance. Nat. Rev. Immunol. 2007. 7: 610–621.

31 Scheicher, C., Mehlig, M., Zecher, R. and Reske, K., Dendritic cells from

mouse bone marrow: in vitro differentiation using low doses of

recombinant granulocyte-macrophage colony-stimulating factor. J. Immu-

nol. Methods 1992. 154: 253–264.

32 Lutz, M. B., Kukutsch, N., Ogilvie, A. L., Rossner, S., Koch, F., Romani, N.

and Schuler, G., An advanced culture method for generating large

quantities of highly pure dendritic cells from mouse bone marrow.

J. Immunol. Methods 1999. 223: 77–92.

33 Zal, T., Volkmann, A. and Stockinger, B., Mechanisms of tolerance

induction in major histocompatibility complex class II-restricted T cells

specific for a blood-borne self-antigen. J. Exp. Med. 1994. 180: 2089–2099.

34 Besche, V., Wiechmann, N., Castor, T., Trojandt, S., Hohn, Y., Kunkel, H.,

Grez, M. et al., Dendritic cells lentivirally engineered to overexpress IL-10

inhibit contact hypersensitivity responses, despite their partial activation

induced by transduction-associated physical stress. J. Gene Med. 2010. 12:

231–243.

35 Weih, F., Carrasco, D. and Bravo, R., Constitutive and inducible Rel/NF-

kappa B activities in mouse thymus and spleen. Oncogene 1994. 9:

3289–3297.

36 Schreiber, E., Matthias, P., Muller, M. M. and Schaffner, W., Rapid

detection of octamer binding proteins with ‘mini-extracts’, prepared

from a small number of cells. Nucleic Acids Res. 1989. 17: 6419.

Abbreviations: BMDC: BM-derived DC � CYLD: cylindromatosis gene �DEX: dexamethasone � GC: glucocorticoid � NEMO: NF-kB essential

modulator � nTC: naıve BALB/c T cells � pTC: primed T cells � TRAF:

TNF receptor-associated factor

Full correspondence: Dr. Matthias Bros, Clinical Research Unit

Allergology, Department of Dermatology, University Medical Center of

the Johannes Gutenberg-University, Ob. Zahlbacher Street 63, D-55131

Mainz, Germany

Fax: 14961313933360

e-mail: mbros@uni-mainz.de

Received: 27/1/2009

Revised: 15/6/2010

Accepted: 19/7/2010

Article accepted online: 28/7/2010

Eur. J. Immunol. 2010. 40: 2848–2857 Immunomodulation 2857

& 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.eji-journal.eu