CD40 regulation of death domains containing receptorsand their ligands on lymphoma B cells

PATRICIA RIBE IRO, NATHALIE RE NA RD, KRZYSZTOF WARZOCHA,† CAROLE CHARLOT, LORE NA JEANDENANT,EVELYNE CALLE T-BAUCHU,* BERTRAND COIFF IER AND GILL ES SAL LES Service d’Hematologie,Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon et UPRES-JE 1879 ‘Hemopathies Lymphoıdes Malignes’,and *Laboratoire Central d’Hematologie, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, France

Received 22 May 1998; accepted for publication 28 August 1998

Summary. Within the tumour necrosis factor (TNF) familythe induction of apoptosis is restricted to some ligand–receptors pairs, including TNF–TNF receptor type I (TNFRI/p55), FasL-Fas, TNF-related apoptosis-inducing ligand(TRAIL) and its death-receptors (DR)-4 and -5. The pairCD40L–CD40 belongs to the same family but rescues B cellsfrom apoptosis. To investigate how these opposing actionsare cross-linked, purified follicular lymphoma (FL) cells were

activated upon a human CD40L-transfected murine fibro-blastic layer, then RNA messengers for the above moleculeswere analysed using RT-PCR. The observed down-modula-tion of TRAIL and up-regulation of TNF and Fas transcriptsmight account for CD40-CD40L-mediated FL cell survival.

Keywords: CD40 activation, apoptosis, lymphoma, B cells.

Within the TNF family, TNF-TNFRI, FasL-Fas (APO-1) andthe recently reported TRAIL (APO2L)-DR4/DR5 share theability to induce apoptosis due to a death domain sequencethat recruits common adapter molecules of the apoptoticpathway (Golstein, 1997). Through the TNF receptorassociated factors (TRAFs), nuclear factor-kappa B (NF-kB)may also be activated, which however rescues cells fromapoptosis (Rothe et al, 1995). CD40 is a TNF family B-cellreceptor (Gruss et al, 1997) that leads to activation andsurvival of normal and malignant B cells, such as non-Hodgkin follicular lymphoma (FL) (Johnson et al, 1993).CD40 signalling through TRAF2 induces NF-kB activation(Rothe et al, 1995) but the mechanisms which are relevantfor FL cells survival remain unclear. One proposed mecha-nism is the anti-apoptotic bcl-2 protein overexpression in FLcells which bear the t(14;18) chromosomal translocation.However, without CD40 stimulation, FL cells die in vitrodespite bcl-2 overexpression.

CD40 triggering modulated Fas and TNF expression on FLand normal B cells (Plumas et al, 1998; Worm & Geha,1994), suggesting that additional signals may enable B-celltumours to escape apoptosis. TRAIL has been shown toinduce apoptosis in several cell lines and fresh lymphoid

malignancies including FL (Snell et al, 1997). DR4 and DR5seem to induce death through the same adapter molecules asFas and TNFRI (Golstein, 1997). The present study addressesthe role of CD40L on the regulation of the expression of FasL-Fas, TNF-TNFRI and TRAIL-DR4/DR5, on FL cells.

MATERIALS AND METHODS

Patients. Lymph nodes were obtained from biopsies of sixFL patients, according to institutional guidelines. Thediagnosis was assessed by morphological and immuno-phenotyping analysis. All samples had the t(14;18) trans-location and bcl-2 gene rearrangement.

Cells. The mononuclear cell fraction of lymph nodes cellsuspension was isolated by Ficoll-Hypaque (Gibco BRL,Grand Island, N.Y.) and E rosetting (BioMerieux, Marcyl’Etoile, France). The nonrosetting cells were incubated withanti-CD3 (OKT3, American Type Culture Collection, Rocke-ville, Md., ascite fluid) anti-CD14 and anti-CD16 monoclonalantibodies (mAbs) (Immunotech, Marseille, France), thenwith magnetic beads (Dynabeads, Dynal, Oslo, Norway)coated with anti-mouse IgG. The isolated cells were analysedby flow-cytometry (FACScan, Becton Dickinson, MountainView, Calif.) after staining with fluorescein isothiocyanateconjugated CD3, CD14 (Immunotech) and CD19 (Dako,Glostrup Denmark) and phycoerythrin-goat anti-mouseimunoglobulin (Dako) conjugated sIg and light chain mAbs.All samples showed 98% cell purity and were monoclonal.

British Journal of Haematology, 1998, 103, 684–689

684 q 1998 Blackwell Science Ltd

† Present address: Department of Haematology, Medical Universityof Lodz, Poland.

Correspondence: Dr Gilles Salles, Service d’Hematologie, CentreHospitalier Lyon-Sud, 69495 Pierre-Benite cedex, France.

685CD40L Up-regulates TNF and fas but Down-modulates TRAIL

q 1998 Blackwell Science Ltd, British Journal of Haematology 103: 684–689

Normal tonsil B cells (CD10–, CD19þ, CD40þ, sIgþ) obtainedafter immunopurification were provided by Dr S. Saeland(Schering Plough, Dardilly, France).

Culture conditions. Murine fibroblast cells transfected withthe human CD40L and untransfected controls (provided byDr J. Banchereau, Schering Plough) were treated with 25 mg/mlof mitomycine C (Sigma Chemical Co., St Louis, Mo.) for 1 hat 378C, washed and seeded in 48- and 96-well plates (Costar,Dutscher, Brumath) in culture medium at a final concentra-tion of 5 × 105 cells/well. Purified FL were distributed on thelayers at 1 × 106 cells/well in culture medium containingRPMI 1640 (Gibco BRL), 10% fetal calf serum (Dutscher),Penicillin G, Streptomycine and L-glutamine (all from Sarbach,Surennes, France). At the onset of culture, 100 U/ml ofinterleukin-4 (IL-4), 10 mg/ml of anti-huCD40L mAb (LL48)(produced by Dr Cees van Kooten) or 10 mg/ml of isotype-matched control IgG mAb with unrelated specificities (allfrom Schering Plough) were added. After 24, 48, 72, 96 hand 7 d of culture, B cells were removed for analysis. To verifythat transcripts regulation was not exclusive of FL cells,normal purified tonsil B cells were submitted to the sameexperimental procedures.

Cell cycle and apoptosis analysis. After removal, cells werefixed in 70% ethanol at 48C for 30 min, treated with 1 mg/mlpropidium iodide and 10 mg/ml RNAse (Sigma) andanalysed on FACScan (Becton Dickinson). Apoptosis wasassessed by incubation with 2·5 mg/ml of fluorochromeHoechst 33342 (Sigma) for 20 min, stopped and fixed with10% formaldehyde (Sigma). Cells spread on glass slides werestudied under an UV microscope.

Polymerase chain reaction (PCR) studies. Total mRNA frompurified FL and normal B cells was extracted using TRIZOL(Gibco BRL) with RNAse-free glycogen (Boeringher

Mannhein, Germany) or a single-step method of acidguanidium thiocyanate–phenol–chloroform (Warzocha etal, 1997).

Primers used (Table I) were designed using GenBanksequences, except for Fas-L (Xerri et al, 1997). After 10 minof heating at 958C, PCR reactions were carried out for 36cycles, each consisting of denaturation for 60 s at 958C,annealing for 60 s at 588C (for TRAIL and Fas-L), 628C (forDR5) or 648C (for Fas, DR4, TNF and TNFRI) and extensionfor 60 s at 728C. As a PCR control, the same procedure wasapplied to a reaction mixture without cDNA. After 11 PCRcycles, glyceraldehyde-3-phosphate dehydrogenase (G3PD)primers were added for the remaining 25 cycles, to normalizethe amplifiable cDNA content. PCR products were resolvedon ethidium bromide-stained 1·5–2% agarose gel, and UV-photographed.

To ensure PCR products specificity, each was purified anddigested by an appropriate restriction enzyme (Table I),except TRAIL, DR4 and DR5 which were isolated andsequenced after cloning.

RESULTS

Apoptosis and cell cycle analysisCell counts of purified FL and B cells plated in medium aloneshowed a progressive decrease in cell numbers, whereas inthose with CD40L a sustained cell number increase wasobserved (not shown). Hoechst 33342 staining was per-formed to verify that the dying cells showed the character-istic dye uptake with peripheral chromatin condensation. Inmedium alone the proportion of apoptotic cells was higherthan in the CD40L culture condition (Table II). In cell cycleanalysis the hypodiploid DNA cell content increased in the

Table I. Oligonucleotide sequences chosen as PCR primers for RT-PCR analysis andrespective restriction enzymes used to digest the amplification products. TRAIL, DR4 andDR5 amplified cDNAs were sequenced.

RestrictionMolecules Oligonucleotide sequence enzyme

TRAIL Sense 50-CTGCGTGCTGATCGTGATC-30

Antisense 50-GCCAACTAAAAAGGCCCCG-30 –

DR4 Sense 50-GAGATGTGCCGGAAGTGCAGC-30

Antisense 50-GTGGATCGAGGCGTTCCGTCC-30 –

DR5 Sense 50-CTCCGGGTCCCCAAGACCC-30

Antisense 50-CTGTGGCCCGCTGCCTCAG-30 –

Fas-L Sense 50-CCAGATGCACACAGCATCATC-30

Antisense 50-CAACATTCTCGGTGCCTGTAAC-30 Sac I

Fas Sense 50-CACCACTATTGCTGGAGTCATG-30

Antisense 50-GGGTACTTAGCATGCCACTGC-30 Hind III

TNF Sense 50-ATGAGCACTGAAAGCATGATCCGG-30

Antisense 50-GATAGATGGGCTCATACCAGGGC-30 Pvu II

TNFRI Sense 50-GTATCGCTACCAACGGTGGAAG-30

Antisense 50-CTCGATGTCCTCCAGGCAGCC-30 Pst I

G3PD Sense 50-CCTCCTGCACCACCAACTGC-30

Antisense 50-GGATACATGACAAGGTGCGGC-30 Xba I

medium alone condition, whereas with CD40L, cells in S andG2 phases progressively increased (not shown). Anti-CD40L mAb addition blocked the CD40L anti-apoptoticeffect (Table II) whereas control IgG induced no change.

Immunophenotypic analysis of the cultured cells neverdisclosed growth of T cells, NK and monocytes (not shown).

TNF and TNFRI expression and regulationTNF transcript was detected in 5/6 isolated FL fresh samplesand three samples were cultured. In medium alone or in thepresence of the untransfected layer, the relative amount ofTNF transcript compared to G3PD remained unchanged, butupon CD40 activation TNF expression increased more thanG3PD (Fig 1).

To rule out the possibility that residual cells played a rolein the CD40-mediated effect on TNF transcription, althoughFL cell purity was up to 98%, anti-CD40L mAb and controlIgG were added at the onset of culture. In the presence ofcontrol IgG the TNF signal remained unchanged, whereasanti-CD40L mAb addition decreased the TNF transcript’s

expression to levels that matched those for the untransfectedcondition. This result, consistent for all samples and inseveral repeated experiments, demonstrated that theobserved effect was B-cell specific, since anti-CD40L mAbblocked CD40-CD40L interaction (Fig 1).

In contrast, TNFRI transcript was absent in all fresh andcultured FL samples although present on normal B cells. NoCD40L-induced regulation was observed (not shown).

Fas and FasL expression and regulationFas transcript was undetectable in unstimulated freshlypurified FL cells but it was induced after culture in everycondition. With CD40L, Fas transcript expression was up-regulated (Fig 2, top panel). Anti-CD40L mAb additionspecifically blocked Fas up-regulation. Fresh and unstimu-lated normal B cells expressed Fas transcript which wasalso up-regulated by CD40L (not shown). FasL transcriptwas undetectable in all but one fresh FL sample, andremained unchanged upon CD40L stimulation (Fig 2, bottompanel). Although FasL transcript was faintly expressed by

q 1998 Blackwell Science Ltd, British Journal of Haematology 103: 684–689

686 Patricia Ribeiro et alTable II. Percentage of FL cells in apoptosis assessed by Hoechst staining.

Cells in apoptosis

Sample Culture conditions Day 2 Day 3 Day 4

1 þ medium alone 40 52 ndþ CD40L þ IL-4 13 15 ndþ Orient þ IL-4 33 14 nd

2 þ medium alone 12 21 28þ CD40L þ IL-4 6 10 7þ Orient þ IL-4 8 15 18þ CD40L þ IL-4 þ cIg 11 23 9þ CD40L þ IL-4 þ anti-CD40L 10 19 21

3 þ medium alone 27 31 59þ CD40L þ IL-4 5 7 6þ Orient þ IL-4 20 20 6

4 þ medium alone 38 42 ndþ CD40L þ IL-4 10 17 ndþ Orient þ IL-4 26 26 nd

5 þ medium alone 27 34 46þ CD40L þ IL-4 8 5 6þ Orient þ IL-4 13 12 15þ CD40L þ IL-4 þ cIg 4 6 6þ CD40L þ IL-4 þ anti-CD40L 14 13 13

6 þ medium alone 30 32 51þ CD40L þ IL-4 21 5 9þ Orient þ IL-4 16 13 15þ CD40L þ IL-4 þ cIg 16 5 7þ CD40L þ IL-4 þ anti-CD40L nd 19 21

All slides for each condition and time were read twice and the number of cells in apoptosis were counted ina total of 200 cells. The mean value was then calculated and expressed as a percentage. A marked decrease ofthe number of cells in apoptosis were observed when FL cells were plated on CD40L transfected layer in thepresence of IL4 compared to medium alone or the untransfected layer (Orient) culture condition whereintermediate counts were observed. The addition of the anti-CD40L mAb specifically blocked the protectiveeffect of the transfected CD40L. Cells were also counted at days 1 and 7 when similar effects were observed.nd: not done.

687CD40L Up-regulates TNF and fas but Down-modulates TRAIL

q 1998 Blackwell Science Ltd, British Journal of Haematology 103: 684–689

unstimulated and stimulated normal B cells, no CD40Lregulation was observed (Fig 2, bottom panel).

TRAIL, DR4 and DR5 expression and regulationTRAIL transcript was detected in 5/6 freshly purified FLsamples (Fig 3a). In the presence of CD40L a marked decreaseof TRAIL expression was observed with a maximum decreaseat day 3 (Fig 3b), contrasting with the sustained level of G3PD

(Figs 3b and 3c). Anti-CD40L mAb specifically blockedTRAIL transcript decrease so its level of expression remainedhigh, matching the one obtained for the untransfectedcondition in all three patients’ samples tested (Fig 3b).

In normal purified B cells, TRAIL transcript was alsodetected at day 0 and decreased by CD40L (Fig 3d) but with adifferent kinetics analysis when compared to FL cells (Figs 3band 3d).

Fig 1. Analysis of TNF transcript. Purified FL cells from patients (three samples) and normal B cells were seeded in culture with medium alone(M), control untransfected fibroblastic layer (C) and human CD40L transfected fibroblastic layer (CD40L). At the onset of culture with transfectedhuman CD40L fibroblastic layer, additional wells were done where anti-CD40L mAb (Ab) and control IgG (cIg) were added separately. RT-PCRanalysis with specific oligonucleotides for TNF and G3PD were done as described in the Material and Methods section. A representative patientsample is shown here. Results at days 1, 2 and 3 show that the relative amount of TNF cDNA was increased upon CD40L culture conditioncompared to G3PD transcript for both FL and normal B cells. The addition of anti-CD40L mAb blocked this effect whereas the addition of controlIgG did not modify CD40L-induced TNF up-regulation (day 2).

Fig 2. Analysis of Fas (top panel) and FasL (bottom panel) transcripts. Purified FL cells and normal B cells were seeded in culture with mediumalone (M), control untransfected fibroblastic layer (C) and human CD40L transfected fibroblastic layer (CD40L), in the presence of anti-CD40L mAb (Ab) or control IgG (cIg). RT-PCR analysis with specific oligonucleotides for Fas, FasL and G3PD were done as described.A representative sample is shown where upon CD40L an increased relative amount of Fas transcript, compared also to a higher G3PD signal, canbe observed at days 2 and 4. On the bottom panel, FasL transcript on normal B cells and on the only positive FL sample where it was faintlyexpressed, are represented. There was no CD40L-induced modulation of FasL transcript.

DR4 and DR5 transcripts were detected in all fresh andcultured FL and normal B cells, but in both cell types nomodulation upon CD40L-induced activation was observed(not shown).

DISCUSSION

CD40–CD40L interaction rescued malignant and normalB cells from apoptosis and regulated the expression of TNF,

Fas and TRAIL specific transcripts, thus illustrating theimportance of this interaction upon the death–survivalbalance.

TNF transcript was found in FL and normal B cells andwas up-regulated after CD40 stimulation (Worm & Geha,1994). Several observations indicate that this up-regulationmay not lead necessarily to apoptosis. In fact, TNF has beenshown to induce B-cell proliferation in some models (Digel etal, 1989) and soluble human recombinant TNF did not

q 1998 Blackwell Science Ltd, British Journal of Haematology 103: 684–689

688 Patricia Ribeiro et al

Fig 3. Analysis of TRAIL transcript. (a) RT-PCR analysis with specific oligonucleotides for TRAIL and G3PD on six fresh FL samples at day 0.TRAIL transcript is present in five FL samples. (b) Then three samples of purified FL cells were seeded in culture with medium alone (M), controluntransfected fibroblastic layer (C) and human CD40L transfected fibroblastic layer (CD40L) with anti-CD40L mAb (Ab) or control IgG (cIg).RT-PCR analysis with specific oligonucleotides for TRAIL and G3PD were done as described. In all FL samples there was a decrease in the level ofTRAIL expression after day 1, being undetectable by day 2/3 for patients 2 and 4. The addition of anti-CD40L mAb blocked this effect, whereasupon control IgG no change on TRAIL level of expression was observed. (c) The kinetic assay shows a representative sample where TRAILdecrease can still be observed by day 7. Note that the major decrease was at day 3, for the same patient in different experiments, as shown inFig 3(b). (d) Normal B cells were seeded in the same culture conditions as mentioned above and RT-PCR analysis for TRAIL and G3PD transcriptswere done at the same time points. TRAIL transcript expression was also decreased, although this was noted later, at day 2, and the differencebetween the relative amount of TRAIL and G3PD cDNA was not so marked. The addition of anti-CD40L mAb blocked this effect whereas thecontrol IgG did not change CD40L-induced TRAIL down-regulation.

689CD40L Up-regulates TNF and fas but Down-modulates TRAIL

q 1998 Blackwell Science Ltd, British Journal of Haematology 103: 684–689

modify CD40L-induced death protection in the presentmodel (unpublished data). Growth-inducing effects of TNFcan be mediated through TNF receptor type II (p75), thetranscripts of which were detected in resting and activated Bcells (not shown), in contrast to TNFRI that was practicallyundetectable in the present study.

Fas transcript was absent on FL cells but induced afterculture and up-regulated upon activation (Plumas et al,1998). Although Fas seems to be highly expressed on bothFL and normal B cells upon activation, untransformed cellswere rendered sensitive to Fas-mediated apoptosis whereasFL cells were found to be intrinsically resistant to this deathsignal (Plumas et al, 1998; Xerri et al, 1997). Althoughfunctional consequences of the activation of CD40 and Faspathways may depend on the cell type, CD40 has beenshown to be overexpressed in low-grade non-Hodgkinlymphomas (Gruss et al, 1997), which might account forthe different susceptibility to Fas-mediated apoptosis. Inter-estingly, it has been recently shown that Fas may alsotransduce stimulatory growth signals (Newton et al, 1998).Finally, the present report confirms that FasL transcript waspractically undetectable from FL cells, although faintlypresent in their normal counterpart (Xerri et al, 1997).

Fresh FL cells and their normal counterpart co-expressedTRAIL, DR4 and DR5. CD40 activation had no influence onthe expression levels of DR4 and DR5 transcripts but TRAILwas markedly and specifically down-regulated. TRAIL hasfour related receptors so far described, including DR4 andDR5 which contain a functional death-domain motif butmay also activate NF-kB and TRAIL-R3 and TRAIL-R4.These act as decoys receptors since they compete for TRAILbut lack a death-domain, although TRAIL-R4 retains anintracytoplasmatic functional motif which may also induceNF-kB activation (Degli-Esposeti et al, 1997; Golstein, 1997).In this complex setting, the present study showed that theCD40-mediated down-regulation of TRAIL was more rapidand deep in FL cells than in normal B cells. In regard to thiskinetic difference, it is noteworthy that normal B cells havebeen found to express decoy receptors which are seldompresent in malignant cell lines (Degli-Esposeti et al, 1997;Golstein, 1997). Therefore TRAIL down-regulation in FLcells may be crucial for their survival, although other anti-apoptotic pathways can not be excluded namely thoseinvolving the bcl-2/bax family (Plumas et al, 1998). Futurestudies are needed to evaluate the expression and function ofthe common transducer signalling molecules upon the CD40system.

In conclusion, CD40–CD40L interaction regulates theexpression of death-domain containing receptors and theirrespective ligands. Since apoptosis can be inhibited ordelayed by the modification of expression of specifictranscripts, we may postulate that TRAIL transcript’sdown-regulation contributes to CD40L-induced FL cellsurvival.

ACKNOWLEDGMENTS

We are grateful to Dr P. Felman for her support and to R.Rouviere (Centre Hospitalier Lyon-Sud, Lyon, France) fortechnical help. This study was supported by Hospices Civilsde Lyon-PHRC (96·044) and by INSERM (Paris). P.R. andN.R were supported with grants funded respectively by theEuropean Society for Medical Oncology (ESMO, Lugano) andthe Hospices Civils de Lyon.

REFERENCES

Degli-Esposeti, M., Dougall, W., Smolak, P., Waugh, J., Smith, C. &Goodwin, R. (1997) The novel receptor TRAIL-R4 induces NF-kappaB and protects against TRAIL-mediated apoptosis, yetretains an incomplete death domain. Immunity, 7, 813–820.

Digel, W., Stefanic, M., Schoniger, W., Buck, C., Raghavachar, A.,Frickhofen, N., Haimel, H. & Porzslt, F. (1989) Tumor necrosisfactor induces proliferation of neoplastic B cells from chroniclymphocytic leukemia. Blood, 73, 1242–1246.

Golstein, P. (1997) Cell death: TRAIL and its receptors. CurrentBiology, 7, R750–R753.

Gruss, H.-J., Herrmann, F., Gattei, V., Gloghini, A., Pinto, A. &Carbone, A. (1997) CD40/CD40 ligand interactions in normal,reactive and malignant lympho-hematopoietic tissues. Leukemiaand Lymphoma, 24, 393–422.

Johnson, P., Watt, S., Betts, D., Davies, D., Jordan, S., Norton, A. &Lister, T. (1993) Isolated follicular lymphoma cells are resistant toapoptosis and can be grown in vitro in the CD40/stromal cellsystem. Blood, 82, 1848–1857.

Newton, K., Harris, A. W., Bath, M.L., Smith, K.G.C. & Strasser, A.(1998) A dominant interfering mutant of FADD/MORT1 enhancesdeletion of autoreactive thymocytes and inhibits proliferation ofmature T lymphocytes. EMBO Journal, 17, 706–718.

Plumas, J., Jacob, M.-C., Chaperot, L., Molens, J.-P., Sotto, J.-J. &Bensa, J.-C. (1998) Tumor B cells from non-Hodgkin’s lymphomaare resistant to CD95 (Fas/Apo-1)-mediated apoptosis. Blood, 91,2875–2885.

Rothe, M., Sarma, V., Dixit, V. & Goeddel, D. (1995) TRAF2-mediatedactivation of NF-kB by TNF receptor 2 and CD40. Science, 269,1424–1427.

Snell, V., Clodi, K., Zhao, S., Goodwin, R., Thomas, E., Morris, S.,Kadin, M., Cabanillas, F., Andreeff, M. & Younes, A. (1997)Activity of TNF-related, apoptosis-inducing ligand (TRAIL) inhaematological malignancies. British Journal of Haematology, 99,618–624.

Warzocha, K., Renard, N., Charlot, C., Bienvenu, J., Coiffier, B. &Salles, G. (1997) Identification of two lymphotoxin beta isoformsexpressed in human lymphoid cell lines and non-Hodgkin’slymphomas. Biochemical and Biophysical Research Communications,238, 273–276.

Worm, M. & Geha, R. (1994) CD40 ligation induces lymphotoxinbeta gene expression in human B cells. International Immunology,6, 1883–1890.

Xerri, L., Devilard, E., Hassoun, J., Haddad, P. & Birg, F. (1997)Malignant and reactive cells from human lymphomas frequentlyexpress Fas ligand but display a different sensitivity to Fas-mediated apoptosis. Leukemia, 11, 1868–1877.