ORIGINAL ARTICLE

Inhibition of CD40–CD154 costimulatory pathwayby a cyclic peptide targeting CD154

Ilaria Deambrosis & Sara Lamorte & Fulvia Giaretta &

Lorenzo Tei & Luigi Biancone & Benedetta Bussolati &Giovanni Camussi

Received: 18 April 2008 /Revised: 11 September 2008 /Accepted: 20 October 2008 / Published online: 5 November 2008# Springer-Verlag 2008

Abstract Disruption of the CD40–CD154 interaction wasfound to be effective in the prevention and treatment ofseveral immune-mediated diseases. The antibody-basedstrategy of inhibition was in humans limited by plateletactivation leading to thrombotic effects. Other strategiesdifferent from antibody technology may be useful to createtools to interfere with CD40–CD154 pathway. In thepresent study, we selected and characterized from a phagedisplay library, cyclic hepta-peptides specific for humanCD154 through biopanning against plate-immobilizedrecombinant hCD154-muCD8. Nine phage clones wereselected for the ability to bind CD154 expressed on thesurface of J558L cells transfected with human CD154.From the nine selected phage clones, we obtained sevendifferent amino acidic sequences, and the correspondinghepta-peptides rendered cyclic by two cysteines weresynthesized. All the peptides specifically bound CD154expressed on J558L. However, only the peptide 4.10(CLPTRHMAC) was found to recognize the active binding

site of CD154, as it competed with the blocking anti-CD154 antibody. When changes in the amino acidcomposition were introduced in the sequence of 4.10peptide, the binding to CD154 was abrogated, suggestingthat the amino acid sequence was critical for its specificity.This peptide was found to inhibit the CD40–CD154interaction, preventing CD40-dependent activation of Blymphocytes in vitro as it was able, as the blocking anti-human CD154 mAb, to prevent the expression of CD80and CD86 costimulatory molecules and switching of Igisotype induced by CD154. Moreover, the peptide 4.10inhibited the in vitro endothelial cell motility and organi-zation into capillary-like structures, and the in vivoangiogenesis of human umbilical cord-derived endothelialcells implanted in Matrigel in severe combined immunode-ficiency mice. In vitro studies on platelet activationdemonstrated that the 4.10 peptide, at variance of the anti-CD154 mAb, was unable to prime human platelet activa-tion and aggregation. In conclusion, we identify a cyclichepta-peptide able to displace the binding of human CD154to CD40 expressed on cell surface and to abrogate somebiological effects related to the CD40 stimulation, such asB cell activation and endothelial triggered angiogenesis.

Keywords Phage display . Angiogenesis .

Blocking peptide . CD40 . B cell activation

Introduction

CD154 (CD40L or gp39), the ligand of CD40, is a type IIintegral membrane protein, related to tumor necrosis factor(TNF) and Fas ligand [1]. After activation, it is transientlyexpressed on the surface of T cells, macrophages, platelets,monocytes, NK cells, and endothelial cells [2]. Its receptor,

J Mol Med (2009) 87:181–197DOI 10.1007/s00109-008-0416-1

I. Deambrosis : S. Lamorte : F. Giaretta : L. Biancone :B. Bussolati :G. CamussiCattedra di Nefrologia, Dipartimento di Medicina Internaand Centro Ricerca Medicina Sperimentale (CeRMS),Università di Torino,Turin, Italy

L. TeiDipartimento di Scienze dell’Ambiente e della Vita,Università del Piemonte Orientale, A. Avogadro,Alessandria, Italy

G. Camussi (*)Cattedra di Nefrologia, Dipartimento di Medicina Interna,Ospedale Maggiore S. Giovanni Battista,Corso Dogliotti 14,10126, Turin, Italye-mail: giovanni.camussi@unito.it

CD40, is a 50-kDa transmembrane glycoprotein of the TNFreceptor superfamily, originally identified on B cells andessential for B cell proliferation, differentiation, Ig produc-tion, isotype switching, and maturation into memory B cells[3]. It has been subsequently found that a multitude of othercell types express CD40, including monocytes, dendriticcells, endothelial cells, smooth muscle cells, fibroblasts,and epithelial cells, suggesting multiple functions [2, 4]. Itis now recognized that the CD40–CD154 is a T cellcostimulatory pathway involved in antigen presentation aswell as in the regulation at different levels of the immuneresponse and of inflammation [2–4].

In the last decade, many studies have demonstrated theinvolvement of CD40–CD154 interaction in a number ofinflammatory processes, ranging from atherosclerosis toautoimmune diseases [5–8] and in the development ofcancer [9–11].

All these findings primed an impressive number ofstudies aimed to functionally characterize this receptorialsystem in view of therapeutically exploiting its properties.Indeed, various approaches (e.g., blocking antibodies, Fabfragments, or fusion proteins), aimed to disrupt the CD40–CD154 interaction, were found to be effective in theprevention and treatment of several experimental modelsof autoimmune disease, atherosclerosis, and transplantrejection [12]. In particular, selective blockade of CD40–CD154 interaction with anti-CD154 monoclonal antibody(mAb) remarkably prolonged survival of islet, skin, bonemarrow, heart, and kidney allografts in murine and largeanimal models, including non-human primates [13–16].Such blockade has also been demonstrated to inhibitautoantibody production in several animal models ofautoimmune disease like systemic lupus erythematosus,rheumatoid arthritis, psoriasis, and Crohn's disease [6, 7].These results have led to clinical trials in both transplan-tation and autoimmune diseases [17] that were halted whenan unusually high incidence of thromboembolic complica-tions was noted to be associated with anti-CD154 mAbtreatment [18]. This effect was related to the expression ofCD154 by activated human platelets and subsequentstimulation of their aggregation by the anti-CD154 mAb[19, 20]. Other strategies different from antibody technol-ogy may be useful to create tools to interfere with CD40–CD154 interaction. Peptide mimics may provide analternative method to block the receptor–ligand interaction,based on the assumption that proteins only exert theirbiological effects through small regions on their surfaces.These short sequences can be reproduced in a small,conformational correct form that will bind to the receptorand provide steric hindrance between the receptor andnative ligand. Peptide mimics have the advantage of beingwater soluble and non-immunogenic, which allows them tobe administered for long periods of time [21].

The aims of the present study were to select andcharacterize from a phage display library, cyclic peptidesspecific for human CD154, and to evaluate whether specificpeptides may bind and interfere with the activation ofCD40–CD154-dependent pathways. In particular, we eval-uated whether specific peptides inhibited the activation of Bcells and the angiogenesis induced by CD40 ligation.

Materials and methods

Reagents

Flourescein isothiocyanate (FITC)-conjugated anti-humanCD14, CD16, CD56, and CD27 mAbs were from CaltagLaboratories, Burlingame, CA, USA. PE-conjugated anti-human CD19 mAb was from Becton Dickinson Biosciences,San Jose, CA, USA. FITC-conjugated anti-human CD3 mAband anti-FITC microbeads were from Miltenyi Biotec,Auburn, CA, USA. FITC-conjugated anti-human CD154mAb was from Serotec, Oxford, UK. PE-conjugated anti-human CD40 mAb was from Immunotech, Coulter Company,Marseille, France. Recombinant fusion proteins constituted bymurine CD8 fused to human CD153 (hCD153-muCD8) orhuman CD154 (hCD154-muCD8), anti-human CD154 block-ing mAb (MK13A4), recombinant human-soluble CD154 kit(rhsCD154 plus enhancer), recombinant mouse-solubleCD154 (rmsCD154), and recombinant human-soluble FAS-L (rhsFAS-L) were from Alexis (Vinci Biochem, FI, Italy).Trypsin, non-enzymatic cell dissociation solution, Tween-20,BSA, basic-Fibroblast Growth Factor (b-FGF), heparin,thrombin, and anti-M13 mAb were from Sigma, St. Louis,CA, USA. PE-conjugated anti-mouse Ab was from DakoCy-tomation, Glostrup, Denmark. Ficoll separating solution wasfrom Biochrom AG, Berlin. Fetal bovine serum (FBS) wasfrom Gibco, Invitrogen Corporation, Paisley, UK.

Cell preparations and cell lines

Human B lymphocytes were collected from peripheralblood of healthy donors with informed consent obtainedin accordance with the Declaration of Helsinki. Mononu-clear cells from whole blood were isolated by Ficoll densitygradient centrifugation, plated in Dulbecco’s modifiedEagle’s medium (DMEM; Sigma) + 10% FBS andincubated overnight (O.N.) at 37°C in the presence of 5%of CO2. After an O.N. adhesion, non-adherent cells wereremoved by gentle rinsing and depleted of T cells, NK, andresidual macrophages by incubation with a mixture ofFITC-conjugated mAbs (anti-human CD3, CD16, CD56,CD14) followed by magnetic depletion using anti-FITCmicrobeads and MACS system (Miltenyi Biotec). Thepurity of eluted CD19-positive cells was always >90%.

182 J Mol Med (2009) 87:181–197

Human umbilical vein endothelial cells (HUVEC) wereisolated by treatment of human umbilical cord veins with0.5% trypsin (1 h at 37°C) and maintained in culture withendothelial basal medium (EBM), supplemented withhuman epidermal growth factor (10 ng/ml), hydrocortisone(1 mg/ml), bovine brain extract (all from CambrexBioscience, Walkersville, MD, USA), and 10% FBS. Theywere used at early passages (second or third).

The murine myeloma cell line J558L transfected withhuman CD154 cDNA [22] was kindly provided by Prof.Mantovani (Mario Negri Institute, Milan, Italy) and culturedin HL-1 serum-free hybridoma medium (Cambrex Bioscien-ces), added with 2.14 mg/ml of L-histidinol dihydrochloride(Sigma); while J558 cell line, transfected with the emptyvector, was cultured in HL-1 serum-free medium.

In vitro biopanning and sequencing of specific phagepeptides

The disulfide-constrained (seven amino acids with twoflanking cysteines at both ends of the peptide) cyclic M13phage display library containing 1.2×109 possible seven-residue sequences (Ph.D.C7C system; New England Biolabs,Hitchin, UK) was used throughout this study. The 96-wellplates were coated with hCD154-muCD8 or hCD153-muCD8 fusion proteins (100 μg/ml in NaHCO3 0.1 M,pH 8.6) and incubated O.N. at 4°C, gently shaken. Excess ofprotein was washed away with Tris-buffered saline (TBS)–0.1% Tween-20 (TBST). Then plates were blocked 2 h atroom temperature (R.T.) with NaHCO3 0.1 M, pH 8.6+ BSA 5 mg/ml + 0.02% NaN3 and washed again withTBST. Ten microliters of the phage library containing 1×1011 plaque-forming units (PFU) was diluted 1:10 with 90 μlof TBST and sequentially added to uncoated and hCD153-muCD8 plates for preabsorption. In each case, the librarywas gently shaken at R.T. for 1 h. Finally the preabsorbedlibrary was applied to hCD154-muCD8-coated plate forspecific screening 1 h at R.T. After washing ten times withTBST, plate-bound phage clones were eluted at R.T. with100 μl of elution buffer (0.2 M glycine–HCl, pH 2.2) for10 min and neutralized with 15 μl of Tris–HCl 1 M, pH 9.1.The eluted phage clones were titrated and amplified, asdescribed [23]. Briefly, to determine the number of phages inthe eluate, in each round of selection, titered triplicatesamples of the eluate were added with the Escherichia colihost Tet-resistant ER2738 (New England Biolabs) intomelted Luria–Bertani (LB) agar tops (7 mg/l agarose, 1 mgMgCl2·6H2O; Sigma), which were then plated onto Tet/IPTG/X-Gal LB agar plates (Kramel Biotech, Cramlington,UK). After O.N. incubation at 37°C, the peptide phages,appearing as blue plaques, were counted and the yield ofrecovered phages determined. The residual eluate wasamplified by culturing the phages as individual plaques on

Tet/IPTG/X-Gal LB agar plates as described above fortitering. The amplified phages in the plaques were recoveredfrom the agar by homogenizing the agar top layer in LBmedium (Sigma), centrifuging, and then precipitating thesupernatant with 3.3% polyethylene glycol 8,000/0.4 MNaCl (Sigma). The resultant pool of phages was resuspendedin TBS and titered, as described above, for a subsequentrandom selection. Three other rounds of selection andamplification were performed with reduced incubation withimmobilized hCD154-muCD8 fusion protein (15 min) andincreased amount of Tween-20 (0.5%) for washes after thebinding process. After the fourth round, individual plaqueswere picked up randomly, amplified in Tet-resistant ER2538E. coli and subjected to FACS analysis (Becton DickinsonBiosciences) and DNA sequencing (MWG Biotech, UK).The primer used for sequencing was 5′-HOCCC TCA TAGTTA GCG TAA CG-3′ (-96 gIII sequencing primer, providedin C7C kit, New England Biolabs) [23]. Alignment bymanual comparison of the sequences was used to identifyconsensus motifs.

Phage binding to CD154-positive cells

For binding experiments, 2.5×105 J558L, or, as control,J558, were incubated with 1×1011 PFU of each clone in200 μl of PBS–0.25% BSA for 1 h at R.T., then washedtwice and incubated with anti-M13 mAb (diluted 1:50 inPBS–BSA) for 1 h at 4°C, washed, and finally incubatedwith PE-conjugated anti-mouse Ab (diluted 1:20 in PBS–BSA) for 30 min in the dark at 4°C. After washing, cellswere analyzed on a FACScan. As negative control, cellswere stained with anti-M13 mAb and PE-conjugated anti-mouse Ab.

Synthesis of anti-CD154 peptides

The synthesis of the peptides was accomplished by solidphase synthesis using the standard Fmoc strategy with aRink Amide resin (Advanced Biotech, Italy). After cleav-age from the resin and precipitation with diethyl ether, thedisulfide bridge between the two cysteines was formed byair oxidation in water (10 ml of H2O for each milligram ofpeptide) with 1% of DMSO. The final products wereobtained in good yields after purification with semi-preparative HPLC (Amersham AKTA Purifier 10/100 usinga Waters X Terra RPC18 19/50 column, final purity >90%)and characterized by Maldi mass spectra (Bruker Daltonics,Bremen, Germany).

Peptide binding to CD154-positive cells

For binding experiments, we used peptides labeled at theN-terminal end with a tag of six histidines (his6-peptides).

J Mol Med (2009) 87:181–197 183

Briefly, 2.5×105 J558L, or J558 cells, were incubated withhis6-peptides (30 μM) in 200 μl of PBS–BSA for 30 minat R.T., then washed twice with PBS–BSA and incubatedwith PE-conjugated anti-poly-histidine mAb (R&D Sys-tems, Minneapolis, MN, USA) for 30 min at 4°C in thedark, washed, and analyzed on a FACScan. As negativecontrol, cells were stained with PE-conjugated anti-poly-histidine mAb.

Competitive assay for human CD154

For competition experiments, 2.5×105 J558L cells wereincubated with his6-peptides (30 μM) for 30 min at R.T.,then cells were further incubated with FITC-conjugatedanti-human CD154 mAb for 30 min in the dark at R.T.Cells were then washed twice and analyzed on a FACScan.As control, cells were stained only with FITC-conjugatedanti-human CD154 mAb. Binding of peptides was evalu-ated as reduction of FITC fluorescence intensity versuscontrol. To verify the effective binding of peptides to thecell surface, we further incubated the cells with PE-conjugated anti-poly-histidine mAb for 30 min at 4°C andverify cell positivity for PE fluorescence on a FACScan.Each peptide was tested at least three times. In selectedexperiments, J558L cells were preincubated with increasingconcentrations of unlabeled 4.10 peptide followed byincubation with 30 μM of his6-labeled 4.10 peptide.

4.10 Peptide binding to human CD154 detectedby enzyme-linked immunosorbent assay

To confirm peptide 4.10 specificity for human CD154, wecoated a 96-well Immuno-Plate (Nunc Maxisorp plates,EuroClone SpA, Life Sciences Division, MI, Italy) with4.10 peptide (500 μg/ml), or, as control, 4.10-ala peptide(500 μg/ml) in NaHPO4 100 mM–HCO3 50 mM, pH 5.0coating buffer, for 6 h at 4°C. The wells were washed withPBS-T (0.2% Tween-20) and saturated O.N. with PBS–10% BSA at 4°C. Then, FLAG-tagged rhsCD154 (1 μg/ml;Alexis) was added to the wells and incubated for 1 h at R.T.After washing, binding of FLAG-tagged rhsCD154 topeptides was detected by incubation with an HRP-conjugated anti-FLAG mAb (Sigma) diluted 1:2,000 inPBS–BSA for 1 h at R.T. The plates were washed andthe enzymatic activity was determined using 3,3′,5,5′-tetramethyl-benzidine substrate in tablets (Sigma). Thereaction was stopped by the addition of 2 M H2SO4, andoptical density was measured at 450 nm, with thecorrection wavelength set at 620 nm. In some experi-ments, to test the possible cross-reaction of 4.10 peptidewith human FAS-L, or murine CD154, plates precoatedwith 4.10 peptide were incubated with FLAG-taggedrhsFAS-L or mrsCD154 (1 μg/ml; Alexis).

B cell activation assay

The ability of the selected anti-CD154 peptides to blockCD40-mediated B cell activation has been evaluated asinhibition of CD40-mediated CD80 (B7-1) and CD86 (B7-2)upregulation. Briefly, 2×105 enriched B cells per well (48-well plate) were plated in 500 μl of complete medium(DMEM, 10% FBS, 2 mM L-glutamine, 100 U penicillin,100 μg/ml streptomycin, 2 mM non-essential amino acid)and stimulated for 48 h with rhsCD154 (100 ng/ml; plus1 μg/ml enhancer) alone or preincubated with variousdoses of peptides. As control, B cells were stimulated withrhsCD154 preincubated with blocking anti-human CD154mAb (5 μg/ml). Then, cells were washed; resuspended inPBS–BSA; incubated at 4°C for 30 min with PE-conjugated anti-human CD80, or PE-conjugated anti-human CD86, and FITC-conjugated anti-human CD20mAbs; and analyzed on a FACScan. As negative control,cells were stained with isotype-matched control Ab.

Induction of B cell isotype switching

Human B cells were isolated by positive selection using CD19MultiSort Kit (Miltenyi Biotec) and more than 98% of theresulting cell population was positive for CD19. Then, naiveCD27-negative B cells were purified through magneticseparation by labeling total B cells with FITC-conjugatedanti-human CD27 mAb and anti-FITC microbeads (MiltenyiBiotec). The obtained CD27-negative naive B cells were>98%, as evaluated by FACS analysis. Naive B cells (1×105/250 μl) were cultured in 96-well plate in DMEM + 10%FBS, and Ig isotype switching was induced by stimulationfor 96 h with rhsCD154 (100 ng/ml; plus 1 μg/ml enhancer)and human IL-4 (0.4 ng/ml). For inhibition studies,rhsCD154 was preincubated at 37°C for 15 min withblocking anti-human CD154 mAb (5 μg/ml) or with anti-CD154 4.10 peptide (30 μM). Then, cells were harvested,incubated with PE-conjugated anti-human CD20 mAb andFITC-conjugated isotype control, anti-human IgM, IgD, orIgG mAbs at 4°C for 30 min, and analyzed on a FACScan.

Platelet aggregation assay

Aggregation assays were performed in an aggregometer(ELVI 840, Logos Milano, Italy) equipped with a stirringdevice, a heated cuvette holder, and time-driven recordingof transmission values. Whole blood was collected fromhuman healthy donors into heparinized tubes at a finalconcentration of 5 U/ml of heparin. Platelet-rich plasmawas prepared by centrifuging whole blood at 900 rpm for20 min. After the upper layer was collected, platelet-poorplasma was obtained by centrifugation of the remainingblood at 3,000 rpm for 10 min. Adenosine diphosphate

184 J Mol Med (2009) 87:181–197

(ADP) 0.5 μM was used as agonist to induce plateletaggregation. To investigate a possible prothrombotic effectof anti-CD154 4.10 peptide, we tested platelet aggregationafter priming platelet-rich plasma for 10 min at 37°C withrhsCD154 (200 ng/ml), or immune complexes (formed by200 ng/ml rhsCD154 and 5 μg/ml anti-human CD154mAb), or anti-CD154 4.10 peptide (60 μM), and thenstimulating with ADP 0.5 μM. In selected experiments,stimulation was performed with 0.3 U/ml thrombin. Aspositive control, to verify the effective responsiveness ofplatelets and to obtain an irreversible aggregation, we usedADP 1 μM or 0.6 U/ml thrombin. Platelet-poor plasma wasused as blanch. Platelet aggregation was measured as ODvariation versus blanch.

Enzyme-linked immunosorbent assay for P-selectin

To evaluate P-selectin release from platelets, platelet-richplasma obtained from healthy donors was primed for10 min at 37°C with rhsCD154 (200 ng/ml), or 4.10peptide (60 μM), or immune complexes (formed byrhsCD154 and anti-human CD154 mAb), or anti-CD1544.10 peptide (60 μM), and then stimulated for further10 min at 37°C under gentle shaking with ADP 0.5 μM. Aspositive control, we used ADP 1 μM. After stimulation,samples were centrifuged at 3,000 rpm for 5 min at 4°C toeliminate platelets, and supernatants were transferred tonew tubes and stored at −20°C until P-selectin detection.After thawing, samples were analyzed for the presence ofsoluble P-selectin using a specific enzyme-linked immuno-sorbent assay (ELISA) kit (R&D Systems, Abingdon, UK),according to the manufacturer’s instructions.

In vitro endothelial cell migration

Before starting the experiment, every culture of HUVEChas been stained for CD40 positivity. HUVEC (1×105)were plated on gelatin-coated T25 flask in EBM with 10%FBS and allowed to adhere O.N. Then, cells were rested for5 h in DMEM containing 5% FBS, washed with PBS, andincubated in DMEM containing 2.5% FBS and theagonists. To obtain CD40 activation, cells were stimulatedwith rhsCD154 (100 ng/ml; plus 1 μg/ml enhancer). Forinhibition studies, rhsCD154 was preincubated at 37°C for15 min with blocking anti-human CD154 mAb (5 μg/ml) orwith anti-CD154 4.10 peptide (30 μM), and then incubatedwith cells. The 4.10-ala peptide was used as negativecontrol peptide. Cell division did not start to any significantdegree during the experiments. Cell migration was studiedover a 4-h period under a Nikon (Melville, NY, USA)Diaphot inverted microscope with a ×10 phase-contrastobjective in an attached, hermetically sealed PlexiglasNikon NP-2 incubator at 37°C [24]. Cell migration was

recorded using a JVC (Tokyo, Japan) 1-CCD video camera.Image analysis was performed with a MicroImage analysissystem (Cast Imaging, Venice, Italy) and an IBM-compat-ible system equipped with a video card (Targa 2000; TrueVision, Santa Clara, CA, USA). Image analysis wasperformed by digital saving of images at 15-min intervals.Migration tracks were generated by marking the position ofthe nucleus of individual cells on each image. The netmigratory speed (straight-line velocity) was calculated bythe MicroImage software based on the straight-line distancebetween the starting and ending points divided by the timeof observation. Migration of at least 35 cells was analyzedfor each experimental condition.

In vitro angiogenesis assay

The assay was performed as previously described [24].Briefly, 24-well plates were coated with growth factor-reduced Matrigel (Becton Dickinson Biosciences) at 4°Cand incubated for 30 min at 37°C, 5% CO2, in a humidifiedatmosphere. To evaluate the endothelial organization intocapillary-like structures, HUVEC were washed twice withPBS, detached with non-enzymatic cell dissociation solu-tion, and seeded onto Matrigel-coated wells in RPMI plus5% FBS at the density of 3.5×104 cells/well under differentexperimental conditions. Cells were stimulated withrhsCD154 (100 ng/ml; plus 1 μg/ml enhancer) alone, orpreincubated at 37°C for 15 min with blocking anti-humanCD154 mAb (5 μg/ml) or with anti-CD154 4.10 peptide(30 μM). The 4.10-ala peptide was used as negative controlpeptide. Cells were periodically observed with a Nikoninverted microscope (Nikon Corporation, Tokyo, Japan),and experimental results were recorded after 4 h ofstimulation. The extension of capillary-like structures wasmeasured with the MicroImage analysis system (CastImaging Srl, Venice, Italy) and expressed in arbitrary units.

In vivo angiogenesis assay

Severe combined immunodeficiency (SCID) mice (CharlesRiver, Jackson Laboratories, Bar Harbor, ME, USA) werekept under standardized conditions at 12 h light/12 h darkcycle with free access to food and water. Animal housing,care, and applications of experimental procedures compliedwith the Guide for the Care and Use of Laboratory Animalsof the Government of Italy and are in accordance to therecommendations of the Society for Laboratory AnimalScience and the Federation of European Laboratory AnimalScience Associations. In vivo Matrigel angiogenesis assaywas performed as previously described [24]. Briefly,HUVEC were harvested using cell dissociation non-enzymatic solution, washed with PBS, resuspended in200 μl of Hanks’ balanced salt solution (Sigma), and

J Mol Med (2009) 87:181–197 185

added to 500 μl of growth factor-reduced Matrigel, inliquid form at 4°C, mixed with 200 ng/ml rhsCD154 plus2 μg/ml enhancer and heparin (64 U/ml). Cells (2.5×106)were injected subcutaneously into the mid-abdominal regionof SCID mice via a 26-gauge needle and a 1-ml syringe. Forinhibition studies, rhsCD154 was preincubated for 15 min at37°Cwith 60μM4.10 peptide and then added to theMatrigel.As control, HUVEC were added to 500 μl of growth factor-reduced Matrigel mixed with heparin (64 U/ml). At day 7,mice were killed and the plugs were recovered and processed

for histology. Typically, gels were cut out by retainingperitoneal lining for support, fixed in 10% buffered formalin,and embedded in paraffin. Sections (3 μm) were cut, stainedwith hematoxylin and eosin, and examined under a lightmicroscopy system. Morphometric analysis was performed tocount vessels that were expressed as percent area per field.Vessel structures were counted only if showing a patentedlumen with red globuli and/or leukocytes. The vessels areawas planimetrically assessed using the MicroImage analysissystem (Cast Imaging Srl). The human nature of endothelial

Fig. 1 Binding of selected phage clones to CD154-transfected J558Lcells. a Histogram representing cytofluorimetric expression ofhCD154 on CD154-transfected J558L surface. Black area shows theanti-CD154 mAb, white area shows the isotypic control. b Histogramsrepresenting binding of selected phage clones to CD154-transfectedJ558L cells (dark lines) by FACS analysis. J558L (2.5×105) were

incubated with 1×1011 PFU and the binding was revealed with anti-M13 mAb and PE-conjugated anti-mouse Ab (dark lines). All theselected phage clones bound to J558L. Staining with anti-M13 mAband PE-conjugated anti-mouse Ab in the absence of phages (dashedlines) was used as control. Data are representative of three differentexperiments

186 J Mol Med (2009) 87:181–197

cells was assessed by double staining in immunofluorescencewith an anti-HLA class I polyclonal Ab (Santa CruzBiotechnology, Santa Cruz, CA, USA) and a rat anti-humanCD31 Ab (Abcam, Cambridge, UK).

Cell viability

Cell viability was assessed by trypan blue exclusion.Human PBMC were plated in 96-well plates in DMEM+ 10% FBS and incubated at 37°C, 5% CO2 for 24 to 96 hin the presence/or not of 4.10 peptide (60 μM). Then, cellswere harvested and stained with trypan blue (Sigma).Viable cells (which exclude dye) were counted, andviability was expressed as percent of viable cells.

Results

Selection and sequencing of peptides that bindto human CD154

The recombinant fusion protein hCD154-muCD8 was usedto develop peptides that bind to human CD154. For thispurpose, a random cyclic M13 phage display peptidelibrary (Ph.D.-C7C) containing 1.2×109 possible seven-residue sequences was screened through biopanning againstplate-immobilized recombinant hCD154-muCD8. To elim-inate potential non-specific binding of phage clones withaffinity for plastic or muCD8, we previously performed twonegative selections versus plastic and a non-related recom-binant fusion protein hCD153-muCD8. The human CD154-binding phage population was enriched over the course offour cycles of biopanning and subsequent amplifications.After the fourth round of biopanning, nine phage cloneswere randomly chosen, expanded, and tested for the abilityto bind human CD154 expressed on cell membrane. FACS

analysis showed that all the phage clones resulted were ableto recognize CD154 expressed in its native configuration onthe surface of J558L cells transfected with human CD154(Fig. 1), whereas they did not bind to CD154-negative J558cells transfected with the empty vector [22], as control (datanot shown). The sequences of the peptides that wereencoded by the selected phage clones were determinedthrough sequencing of the phage DNA. From the nineselected phage clones, we obtained seven different aminoacidic sequences, as the DNA sequences of 4.2 and 4.3, aswell as 4.7 and 4.12 peptides, were identical (Table 1). Noconsensus motifs, which might be responsible for bindingto CD154, could be identified. This was probably due to therecognition of different regions of the CD154 extracellulardomain by each peptide.

Peptide synthesis and binding to CD154-positive cells

We synthesized for each amino acidic sequence thecorresponding cyclic peptide and peptides tagged at theN-terminal with six histidines (his6-tagged peptides), for aneasier detection of the peptide binding. Experiments werethen performed to verify that the binding to human CD154observed in phage clones was still present in the syntheticpeptides. FACS analysis confirmed that all the sevenpeptides maintained the ability to bind to human CD154-positive J558L cells, whereas they did not bind to J558control cells (Fig. 2 and Table 1).

The anti-CD154 4.10 peptide competed for CD154active site with an anti-human CD154 mAb

We evaluated the capacity of the cyclic peptides to preventbinding to human CD154-positive J558L cells of FITC-conjugated anti-human CD154 mAb (clone MK13A4),specific for the CD154 active site [25]. For this purpose,J558L cells were preincubated for 30 min with 30 μM ofeach peptide before the addition of the mAb. Preincubationof J558L cells with 4.10 peptide, as well as with 4.10-his6-peptide, induced a reduction of 73% of the FITC-fluorescent signal due to mAb binding to CD154 on theJ558L cell surface. All the other peptides resulted wereunable to significantly block this interaction (Fig. 3a,b),suggesting that only the 4.10 peptide recognized the CD154active site, whereas the other peptides bound differentCD154 extracellular domains. To confirm the specificity ofthe binding of the 4.10 peptide to human CD154, wedemonstrated that the binding of 4.10-his6-peptide could bedisplaced by increasing concentrations of unlabeled 4.10peptide (Fig. 3c). In order to identify critical amino acidicresidues in 4.10 sequence, we synthesize several variants of4.10 peptide with selective changes, according to twodifferent criteria. Three variants were obtained by substitu-

Table 1 Peptide inserts of the selected phage clones obtained fromthe final in vitro selection were sequenced and analyzed for theirability of binding to CD154 expressed on J558L cells transfected withhuman CD154

Phage clone Sequence Percent binding to J558L

4.2=4.3 CPSGHTKAC 70.9±44.4 CGTHSSRIC 98.2±24.6 CLGTQNKEC 40.8±34.7=4.12 CTPGKPHSC 97.5±14.8 CKAASANIC 51.1±54.10 CLPTRHMAC 85.6±104.11 CLSAVHNMC 43.2±6

Inserts of clones 4.2 and 4.3, as well as of clones 4.7 and 4.12, wereidentical. Binding was expressed as percentage ± SD of positive cellsby FACS analysis in five different experiments.

J Mol Med (2009) 87:181–197 187

Fig. 2 Binding of syntheticanti-CD154 peptides toCD154-transfected J558L cellsor to J558 control cells. J558L(2.5×105) or, as control, J558,were incubated with the sevenselected peptides (30 μM),tagged with a six histidine-tagand then with PE-conjugatedanti-poly-histidine mAb. Bind-ing of his6-peptides (dark lines)to the J558L or J558 surface wasevaluated by FACS analysis inrespect to cells stained with PE-conjugated anti-poly-histidinemAb alone (dashed lines). Allthe seven peptides bound toJ558L (a), while no peptidebound to J558 control cells (b).Data are representative of threedifferent experiments

188 J Mol Med (2009) 87:181–197

Fig. 3 Competition of anti-CD154 synthetic peptides for binding ofanti-human CD154 mAb to CD154 active site. a Histograms showinginhibition of anti-human CD154 mAb binding to the CD154 active siteby anti-CD154 selected peptides. J558L cells 2.5×105 were incubatedwith each of the seven selected his6-peptides (30 μM) and subsequentlyincubated with FITC-conjugated anti-human CD154 mAb. The abilityof single peptides to inhibit binding of anti-human CD154 mAb toCD154 expressed on J558L surface was evaluated by FACS analysis asreduction of fluorescence intensity versus cells stained with FITC-conjugated anti-human CD154 mAb alone (gray column). Results arethe mean of three different experiments ± SD. ANOVA with Dunnett’smulticomparison test: peptides plus anti-human CD154 mAb versusanti-human CD154 mAb alone (***p<0.001). b Representativehistogram showing that the binding of anti-human CD154 mAb toJ558L (dark lines) was reduced by preincubation of cells with anti-CD154 4.10 peptide (thin line; dotted line isotypic control Ab). cEvaluation of specificity of 4.10-his6 binding to J558L by displacementafter incubation with increasing doses of unlabeled 4.10 peptide (peptide

4.10-no tag). The percentage of binding was reduced by the addition ofunlabeled peptide. Results are the mean of four different experiments± SD. ANOVA with Dunnett’s multicomparison test: his6-peptide plusunlabeled peptide versus his6-peptide alone (**p<0.01; ***p<0.001). dComparison between 4.10 and 4.10-ala peptides in competing with anti-human CD154 mAb for binding to J558L. Substitution of the residue ofarginine in position 5 with a residue of alanine, abrogated the ability of4.10 peptide to block binding of anti-human CD154 mAb to CD154 onJ558L surface. Results are the mean of four different experiments ± SD.ANOVAwith Dunnett’s multicomparison test: peptides plus anti-humanCD154 mAb versus anti-human CD154 mAb alone (***p<0.001). eThe binding of rhsCD154, rhsFas-L, or rmsCD154 to 4.10 peptideimmobilized on plastic was evaluated by ELISA, as described in“Materials and methods”. A significant binding in respect to vehiclealone (*p<0.001) was observed only for the human CD154. Results arethe mean of four different experiments ± SD. ANOVA with Dunnett’smulticomparison test was performed versus vehicle

J Mol Med (2009) 87:181–197 189

tion of amino acids in positions 4, 5, and 6, respectively,with a residue of alanine; the other four peptidic variantswere obtained by substitution of one or two residues withamino acid with similar properties. Table 2 summarizes thepeptidic variants that we synthesized. All the mutationsresulted in the abrogation of the capacity of the 4.10peptide to bind to human CD154 expressed on J558L(Table 2). One of these peptides, characterized by selectivechange of the arginine in position 5 with a residue ofalanine, has been chosen as negative control peptide forsubsequent experiments and named 4.10-ala. Incubation ofJ558L with 4.10-ala peptide, even at the concentration250 μM, did not inhibit the binding of FITC-conjugatedanti-human CD154 mAb to J558L (Fig. 3d). The specificbinding of 4.10 peptide, but not of 4.10-ala peptide (datanot shown), to CD154 was confirmed by ELISA onrhsCD154-coated plates (Fig. 3e). Moreover, no significantcross-binding of 4.10 peptide with rhsFas-L or rmsCD154was observed (Fig. 3e).

Anti-CD154 4.10 peptide blocked CD40-mediated B cellactivation and Ig switching

We utilized rhsCD154 to stimulate CD40 on quiescenthuman B cells and to induce enhanced expression of CD80(B7-1) and CD86 (B7-2) on their surface. The stimulationof peripheral B cells with rhsCD154 (100 ng/ml; plus 1 μg/ml enhancer) determined a significant increase of CD80and CD86 expression within 48 h. Preincubation ofrhsCD154 with 30 μM anti-CD154 4.10 peptide for15 min at 37°C significantly prevented B cell activationreducing the expression of CD80 and CD86 (Fig. 4).Preincubation of rhsCD154 with other peptides, at the sameconcentration, did not significantly affect the expression of

CD80 and CD86 induced by CD40 stimulation. Peptides4.6 and 4.11 inhibited B cell activation only at higherconcentration (250 μM; data not shown). As shown inFig. 4c, the inhibitory effect of anti-CD154 4.10 peptide onCD80 expression induced by CD154 was dose dependentand reached its maximal effect at 30 μM. Preincubation ofrhsCD154 with the 4.10-ala peptide, the negative controlpeptide, even at a concentration of 250 μM, did not affect Bcell activation (Fig. 4a). Preincubation with the anti-humanCD154 mAb (clone MK13A4), used as positive control forblocking CD40-CD154 interaction, reduced CD80 andCD86 expression to basal levels (Fig. 4a,b). Moreover, weevaluated the ability of the 4.10 peptide to inhibit CD40-mediated B cell Ig isotype switching. For this purpose,CD27-negative human naive B cells were cultured in thepresence of rhsCD154 (100 ng/ml; plus 1 μg/ml enhancer)and IL-4 (0.4 ng/ml). After 4 days of culture, a significantexpression of IgG by B cells was observed. When B cellswere stimulated with IL-4 and rhsCD154 in the presence ofthe anti-CD154 4.10 peptide (30 μM), a significantreduction of IgG expression was found (Fig. 5). Acomparable inhibitory effect on Ig switching was observedwith the blocking of anti-human CD154 mAb (Fig. 5), butnot with the inactive 4.10-ala peptide (not shown). Nosignificant difference in viability of cells treated with 4.10peptide for 24–96 (96±1.3%: mean ± SD of five experi-ments) or with vehicle (94±0.6%: mean ± SD of fiveexperiments) was observed.

Anti-CD154 4.10 peptide blocked CD40-mediated in vitroendothelial cell migration and angiogenesis

CD40 stimulation is known to induce endothelial migrationand angiogenesis [9, 10, 24, 26]. Therefore, we evaluated

Table 2 The 4.10 peptide and several variants were synthesized andtested for the binding to CD154 expressed on J558L cells transfectedwith human CD154

Peptides Sequence Percent binding to J558L

4.10 CLPTRHMAC 85.6±104.10-I CLPARHMAC 0.5±0.24.10-II (or 4.10-ala) CLPTAHMAC 0.7±0.14.10-III CLPTRAMAC 1.0±0.64.10-IV CIPTRHMAC 0.4±0.24.10-V CLPSRHMAC 0.6±0.14.10-VI CIPTRHMVC 0.5±0.34.10-VII CLPTRWMAC 0.8±0.4

Three peptide variants (4.10 I–III) were obtained by substitution ofamino acids in positions 4, 5, and 6, respectively, with a residue ofalanine. The other four variants (4.10 IV–VII) were obtained bysubstitution of one or two residues with amino acids with similarproperties. Binding was expressed as percentage ± SD of positive cellsby FACS analysis of three different experiments.

Fig. 4 Effect of selected anti-CD154 peptides on B cell CD80 andCD86 upregulation by rhsCD154. a CD80 (white bars) and CD86(black bars) expression on B cell membrane, after stimulation of Bcells for 48 h with rhsCD154 (100 ng/ml; plus 1 μg/ml enhancer)alone or preincubated for 15 min at 37°C with peptides (30 μM).Stimulation with rhsCD154 significantly enhanced CD80 and CD86expression, in respect to basal level (basal). The preincubation ofrhsCD154 with 4.10 peptide, but not with the other peptides,significantly reduced CD80 and CD86 expression. The control 4.10-ala peptide did not inhibit the effect of rhsCD154. As control,rhsCD154 was preincubated with blocking anti-human CD154 mAb(5 μg/ml) that completely prevented B cell activation. Results aremean of three different experiments ± SD. ANOVA with Newman–Keuls multicomparison test: rhsCD154 versus control (§p<0.001),rhsCD154 plus peptides or anti-human CD154 mAb versus rhsCD154(**p<0.01). b FACS analysis representative of CD80 and CD86expression by B cells in basal condition (basal) or stimulated withrhsCD154 alone or preincubated with anti-CD154 4.10 peptide(30 μM) or blocking anti-human CD154 mAb (5 μg/ml). Controlisotypic control mAb. c The dose response of 4.10 peptide on CD154-induced expression of CD80 was evaluated. Results are the mean± SD of four different experiments

�

190 J Mol Med (2009) 87:181–197

whether the 4.10 peptide inhibited CD40-induced endothe-lial activation. Cell motility of HUVEC was studied bytime-lapse recording migration assay over a 4-h period.Unstimulated cells were found to remain steady for thewhole period of observation never exceeding the averagespeed of 12 μm/h, whereas stimulation of endothelial CD40

with rhsCD154 significantly enhanced HUVEC migration,which remained sustained until the end of the observationperiod (Fig. 6). The preincubation of rhsCD154 with the4.10 peptide inhibited CD40-dependent endothelial migra-tion reducing endothelial speed average to basal levelswithin 30 min (Fig. 6) and the effect persisted for the whole

J Mol Med (2009) 87:181–197 191

period of observation (Fig. 6a). Similar results were obtainedby preincubation of rhsCD154 with the blocking anti-humanCD154 mAb (Fig. 6). In contrast, preincubation of rhsCD154with the negative control 4.10-ala peptide did not affectCD40-mediated increase in cell migration (Fig. 6). Moreover,the 4.10 peptide inhibited CD40-triggered endothelial angio-genesis in vitro. As shown in Fig. 7a,c, rhsCD154 stimulatedthe organization of HUVEC plated on growth factor-reducedMatrigel increasing the formation of vessel-like structures.Preincubation of rhsCD154 with the anti-CD154 4.10 peptide(30 μM) completely inhibited the formation of vessel-likestructures (Fig. 7a,d). In contrast, preincubation of rhsCD154with the 4.10-ala peptide did not affect endothelial organi-zation (Fig. 7a,e). Preincubation of rhsCD154 with theblocking anti-human CD154 mAb, as positive control,completely abrogated the proangiogenic effect of rhsCD154(Fig. 7a, f). The peptides, 4.6 and 4.11, did not reduce invitro angiogenesis (data not shown).

Anti-CD154 4.10 peptide blocked CD40-mediated in vivoangiogenesis

To evaluate the effect of anti-CD154 4.10 peptide onangiogenesis in vivo, HUVEC were injected subcutaneous-ly in SCID mice within growth factor-reduced Matrigelcontaining heparin (64 U/ml) and rhsCD154 preincubated/or not with the 4.10 peptide. After 7 days, plugs wererecovered and processed for histological analysis. Asshown in Fig. 8, the 4.10 peptide, but not 4.10-ala peptide,significantly blocked the angiogenic effect of rhsCD154 onHUVEC. Unstimulated HUVEC did not form vesselsin vivo.

Anti-CD154 4.10 peptide did not affect platelets activation

To test the effect on platelets aggregation of the 4.10peptide, we stimulated platelets with thrombin at low dose

Fig. 5 Effect of anti-CD154 4.10 peptide on CD40-mediated isotypeswitching on B cells. Expression of membrane IgG by CD27− naive Bcells stimulated for 96 h with rhsCD154 (100 ng/ml; plus 1 μg/mlenhancer) and IL-4 (0.4 ng/ml) alone, or preincubated for 15 min at 37°C with anti-CD154 4.10 peptide (30 μM) or with blocking anti-humanCD154 mAb (5 μg/ml). a Representative FACS dot plots and bhistogram showing the Ig isotype switching induced by stimulation with

rhsCD154 and IL-4 and the inhibition of Ig isotype switching bypreincubation of rhsCD154 with anti-CD154 4.10 peptide (30 μM) orwith blocking anti-human CD154 mAb (5 μg/ml), used as control.Histograms are mean of three independent experiments ± SD. ANOVAwith Newman–Keuls multicomparison test: rhsCD154 and IL-4 versusbasal control (§p<0.05); rhsCD154 and IL-4 plus 4.10 peptide or plusanti-human CD154 mAb versus rhsCD154 and IL-4 alone (*p<0.05)

192 J Mol Med (2009) 87:181–197

and the 4.10 peptide or, as positive control, the anti-CD154mAb. Anti-CD154 mAb, but not 4.10 peptide, induced asignificant increase in thrombin-induced platelet aggrega-tion (Fig. 9a). Moreover, we also evaluated the effect ofcomplexes formed by rhsCD154 and anti-human CD154mAb or 4.10 peptide on 0.5 μM ADP-induced plateletaggregation, following the procedure described by Langeret al. [20]. Immunocomplexes formed by rhsCD154 andanti-human CD154 mAb primed platelet aggregationinduced by ADP (Fig. 9b). In contrast, complexes formed

by rhsCD154 and the 4.10 peptide (60 μM) did not(Fig. 9b). Moreover, we investigated a possible stimulatoryeffect of the 4.10 peptide on platelets by evaluating therelease of P-selectin after platelet stimulation. As shown inFig. 9b, stimulation of platelets with immunocomplexesformed by rhsCD154 and anti-human CD154 mAb and0.5 μM ADP induced a significant increase in P-selectinrelease, whereas no significant release was observed withthe 4.10 peptide (60 μM), or with complexes formed byrhsCD154 and 4.10 peptide.

Discussion

In the present study, we identify a peptide able toselectively bind the human costimulatory moleculeCD154. This peptide was shown to recognize the activesite of CD154 and to inhibit the CD40–CD154 interaction,preventing CD40-dependent activation of B lymphocytes invitro and CD40-induced angiogenesis both in vitro and invivo.

CD154 expressed on activated T lymphocytes, platelets,monocytes/macrophages, vascular smooth muscle, andendothelial cells is the ligand of the CD40, a costimulatorymolecule crucial for the induction of effective adaptiveimmune and inflammatory response [2–4]. The binding ofCD154 in its trimeric surface expressed form, as well as inits biologically active soluble form released after cellactivation, may lead to a broad range of biological activitiesin target cells expressing CD40 [2, 4]. Therapeuticallytargeting the CD40–CD154 pathway in the hope ofinducing antigen-specific tolerance or inhibition of a widearray of chronic inflammatory and autoimmune diseases [7,8, 12, 17] has been an attractive idea in the scientificcommunity for the last decade. Improvement of allograftsurvival and induction of tolerance have been observed inrodent models using anti-CD154 mAb [15, 16]. CD154blockade has also been shown to prevent acute rejectionand to promote long-term allograft acceptance in non-human primates [13, 14]. Moreover, the interruption of theCD40–CD154 signaling was shown to reduce the severityof arthritis, allergic encephalomyelitis, and atherosclerosis[17] in murine models. Unfortunately, the clinical trialstesting anti-CD154 mAb in autoimmune diseases andtransplantation were terminated due to an unexpectedincidence of thromboembolic complications [18]. Theseadverse effects have been ascribed to surface expression ofCD154 by human platelets [19, 20]. Indeed thrombin-activated human platelets were found to rapidly expressCD154 that can interact with CD40 on endothelial cellsmediating chemotaxis and upregulation of adhesion mole-cules [19]. Moreover, it has been suggested that theactivation of platelets by the complex soluble CD154–

Fig. 6 Effect of anti-CD154 4.10 peptide on CD154-induced HUVECmotility. Time course (a) and speed average (b) of HUVEC motility inresponse to stimulation with rhsCD154 (100 ng/ml; plus 1 μg/mlenhancer), or rhsCD154 associated with 4.10 peptide (30 μM), 4.10-ala peptide (30 μM), or anti-human CD154 blocking mAb (5 μg/ml).Stimulation with rhsCD154 strongly enhanced cell motility; preincu-bation of rhsCD154 with anti-CD154 4.10 peptide reduced the effectof CD154, whereas preincubation of rhsCD154 with 4.10-ala peptidedid not affect CD40 stimulation. Anti-human CD154 blocking mAbwas used as control to inhibit rhsCD154 stimulation. Results are mean± SD of three different experiments evaluating at least 35 cells foreach experimental condition. ANOVA with Newman–Keuls multi-comparison test: rhsCD154 versus vehicle (§p<0.05); rhsCD154 plus4.10 peptide or plus 4.10-ala peptide or plus anti-human CD154 mAbversus rhsCD154 alone (**p<0.01)

J Mol Med (2009) 87:181–197 193

anti-CD154 mAb could be mediated by FcγRII similar towhat was observed in other immune thrombophilias such asthe heparin-induced thrombocytopenia syndrome [25].Recently, it has been shown that CD154 expressed onactivated platelet can sustain a proaggregatory effect ofCD154 mAb by a mechanism involving the Fc domain[27]. These results raised the necessity to find alternativeapproaches for the inhibition of CD40–CD154-dependentpathways.

In the present study, we attempted to develop peptidesable to bind the CD154 with the aim to inhibit itsinteraction with the CD40 expressed on the cell membrane.Being the FcR implicated in platelet activation, we chose touse a peptide phage display library. Peptide screening bothin vitro and in vivo using the phage display technology hasbeen previously used to identify selected peptides able tobind a number of cell surface molecules [23, 28–31].Through the use of phage display, it is now possible to find

peptides that bind protein targets with high affinity andspecificity, in some cases comparable with that of anti-bodies [32].

In the present study, using this technology, we identifiedseven different cyclic hepta-peptides able to bind thehuman CD154. Only one of these peptides was shown tobe able to displace the binding of an anti-CD154 mAb,described as specific for the active site of the molecule [25].The absence of a consensus motif among the identifiedpeptides able to bind CD154, as well as the unique abilityof the 4.10 peptide to recognize the active site of CD154,suggests that different peptides recognized different regionsof the CD154 extracellular domain. When changes in theamino acid composition were introduced in the sequence ofthe 4.10 peptide, the binding to CD154 was abrogated,suggesting that the amino acid sequence is critical for itsspecificity. To test whether the binding of the 4.10 peptideto CD154 would interfere with the biological activities

Fig. 7 Effect of anti-CD1544.10 peptide on CD154-inducedHUVEC in vitro angiogenesis. aOrganization in capillary-likestructures of HUVEC (3.5×104)plated on growth factor-reducedMatrigel was evaluated after 4 hof stimulation with vehicle alone(control), rhsCD154 (100 ng/ml;plus 1 μg/ml enhancer), orrhsCD154 in association withanti-CD154 4.10 peptide(30 μM) or 4.10-ala peptide(30 μM), or anti-human CD154blocking mAb (5 μg/ml). Dataare expressed as the mean ± SDof the length of capillary-likestructures evaluated by the com-puter analysis system in arbitraryunits in at least four differentfields at ×20 magnification offour different experiments.ANOVA with Newman–Keulsmulticomparison test: rhsCD154versus control (§p<0.05);rhsCD154 plus 4.10 peptide orplus anti-human CD154 mAb orplus 4.10-ala peptide versusrhsCD154 alone (**p<0.01). b-fMicrographs representative of invitro formation of vessel-likestructures by HUVEC stimulatedwith vehicle alone (b), rhsCD154(c), or rhsCD154 in associationwith anti-CD154 4.10 peptide (d)or 4.10-ala peptide (e), or anti-human CD154 blocking mAb (f).Original magnification ×200

194 J Mol Med (2009) 87:181–197

triggered by CD154 on cells bearing CD40, we studied theactivation of B lymphocytes and the angiogenic effect onendothelial cells triggered by the CD40–CD154 pathway.CD40 is constitutively expressed on B lymphocytes, and itsligation by CD154-bearing cells triggers B cell differenti-ation, proliferation, and switching of Ig isotype, withgeneration of memory B cells [2]. These events areassociated with the expression of other costimulatory

molecules, such as CD80 and CD86 by activated Blymphocytes [2]. In the present study, we found that the4.10 peptide, but not the modified 4.10-ala control peptide,was able to prevent in vitro expression of costimulatorymolecules, and switching of Ig isotype induced by CD154

Fig. 8 Effect of anti-CD154 4.10 peptide on CD154-induced in vivoangiogenesis. a Formation of vessels after 7 days by HUVEC (2.5×106)implanted in Matrigel subcutaneously in SCID mice after stimulationwith rhsCD154 (200 ng/ml; plus 1 μg/ml enhancer), rhsCD154 plusanti-CD154 4.10 peptide (60 μM), or 4.10-ala peptide (60 μM), or inthe absence of stimulation (control). Angiogenesis was evaluated as thepercentage of vessels area in five different fields at ×20 magnification

and data are expressed as mean ± SD of seven different experiments.ANOVA with Newman–Keuls multicomparison test: rhsCD154 versusvehicle alone, as control (§p<0.01); rhsCD154 plus 4.10 peptide or plus4.10-ala peptide versus rhsCD154 alone (***p<0.001). b, c Represen-tative light microscopy micrographs of the HUVEC-formed vessels inthe Matrigel plug after stimulation with rhsCD154 (b) or rhsCD154 plusanti-CD154 4.10 peptide (c). Original magnification ×200

�Fig. 9 Effect of anti-CD154 4.10 peptide on platelet aggregation andP-selectin release. a Histograms representing platelet aggregation,evaluated as variation in light transmission, were recorded by anaggregometer. Platelets were stimulated with thrombin at low dose(Thr, 0.3 U/ml) and the 4.10 peptide (60 μM) or the anti-CD154 mAb(5 μg/ml). Thr (0.6 U/ml) was used as positive control. Anti-CD154mAb, but not 4.10 peptide, induced a significant increase in thrombin-induced platelet aggregation. b Effect of complexes formed byrhsCD154 (200 ng/ml) and anti-human CD154 mAb (5 μg/ml) or4.10 peptide (60 μM) was evaluated on 0.5 μM ADP-induced plateletaggregation (dark column) and on P-selectin release (white column).ADP (1 μM) was used as positive control. Complexes formed byrhsCD154 and anti-human CD154 mAb primed platelet aggregationand P-selectin release induced by 0.5 μM ADP, whereas the 4.10peptide (60 μM) did not. Histograms are mean of three independentexperiment ± SD. ANOVA with Dunnett’s multicomparison test wasperformed: 0.3 U/ml Thr plus anti-human CD154 mAb or 4.10peptide versus 0.3 U/ml Thr alone (§p<0.05); complex rhsCD154-anti-CD154 mAb or complex rhsCD154–4.10 peptide plus 0.5 μMADP versus 0.5 μM ADP alone (§p<0.01)

J Mol Med (2009) 87:181–197 195

in B lymphocytes, in a dose-dependent manner. Theefficacy of the inhibition was comparable to that of ablocking anti-human CD154 mAb, suggesting that the 4.10peptide may compete with CD40 for the binding to theactive site of CD154.

Moreover, it has been shown that the CD40–CD154-dependent pathway may trigger an angiogenic response thatmay be relevant in the pathogenesis of different inflamma-tory states, atherosclerosis and cancer [6, 10, 33, 34]. CD40is expressed by endothelial cells and its interaction witheither soluble or surface expressed CD154 may triggermotility and coordinated cell to cell interaction leading toneo-angiogenesis [24, 26]. We found that blockade ofCD40–CD154 interaction with the 4.10 peptide resulted inthe inhibition of endothelial cell motility and of in vitro andin vivo angiogenesis, comparable to that obtained with theblocking anti-human CD154 mAb. In vitro studies onplatelet activation demonstrated that the 4.10 peptide, atvariance of anti-CD154 mAb, was unable to prime humanplatelet aggregation and P-selectin release.

In conclusion, the results of the present study demon-strate that a cyclic hepta-peptide was able to displace thebinding of human CD154 to CD40 expressed on cellsurface and to abrogate some biological effects related tothe CD40 stimulation, such as B cell activation andendothelial triggered angiogenesis.

Acknowledgments This work was supported by the ItalianMinistry ofUniversity and Research (MIUR): FIRB project (RBNE01HRS5-001)and COFIN; the Associazione Italiana per la Ricerca sul Cancro (AIRC);Regione Piemonte-Ricerca Scientifica Applicata; ONCOPROT; and byProgetto S. Paolo Oncologia.

References

1. Noelle RJ, Roy M, Shepherd DM, Stamenkovic I, Ledbetter JA,Aruffo A (1992) A 39-kDa protein on activated helper T cellsbinds CD40 and transduces the signal for cognate activation of Bcells. Proc Natl Acad Sci U S A 89:6550–6554

2. Schönbeck U, Libby P (2001) The CD40/CD154 receptor/dyad.Cell Mol Life Sci 58:4–43

3. Grewal IS, Flavell RA (1996) The role of CD40 ligand incostimulation and T-cell activation. Immunol Rev 153:85–106

4. Quezada SA, Jarvinen LZ, Lind EF, Noelle RJ (2004) CD40/CD154 interactions at the interface of tolerance and immunity.Annu Rev Immunol 22:307–328

5. Schönbeck U, Libby P (2001) CD40 signaling and plaqueinstability. Circ Res 89:1092–1103

6. Biancone L, Cantaluppi V, Camussi G (1999) CD40–CD154interaction in experimental and human disease. Int J Mol Med3:343–353

7. Liossis SN, Sfikakis PP (2004) Costimulation blockade in thetreatment of rheumatic diseases. BioDrugs 18:95–102

8. Caproni M, Torchia D, Antiga E, Giomi B, Mercuri M, Volpi W,Fabbri P (2007) The CD40/CD40 ligand system in the skin ofpatients with subacute cutaneous lupus erythematosus. J Rheu-matol 34:2412–2416

9. Bussolati B, Russo S, Deambrosis I, Cantaluppi V, Volpe A,Ferrando U, Camussi G (2002) Expression of CD154 on renal cellcarcinomas and effect on cell proliferation, motility and platelet-activating factor synthesis. Int J Cancer 100:654–661

10. Biancone L, Cantaluppi V, Boccellino M, Del Sorbo L, Russo S,Albini A, Stamenkovic I, Camussi G (1999) Activation of CD40favours the growth and vascularization of Kaposi's sarcoma. JImmunol 163:6201–6208

11. Bereznaya NM, Chekhun VF (2007) Expression of CD40 andCD40L on tumor cells: the role of their interaction and newapproach to immunotherapy. Exp Oncol 29:2–12

12. Howard LM, Miller SD (2004) Immunotherapy targeting theCD40/CD154 costimulatory pathway for treatment of autoim-mune disease. Autoimmunity 37:411–418

13. Kirk AD, Burkly LC, Batty DS, Baumgartner RE, Berning JD,Buchanan K, Fechner JH Jr, Germond RL, Kampen RL,Patterson NB, Swanson SJ, Tadaki DK, TenHoor CN, White L,Knechtle SJ, Harlan DM (1999) Treatment with humanizedmonoclonal antibody against CD154 prevents acute renalallograft rejection in non-human primates. Nat Med 5:686–693

14. Kenyon NS, Chatzipetrou M, Masetti M, Ranuncoli A, Oliveira M,Wagner JL, Kirk AD, Harlan DM, Burkly LC, Ricordi C (1999)Long-term survival and function of intrahepatic islet allografts inrhesus monkeys treated with humanized anti-CD154. Proc NatlAcad Sci U S A 96:8132–8137

15. Molano RD, Pileggi A, Berney T, Poggioli R, Zahr E, Oliver R,Ricordi C, Rothstein DM, Basadonna GP, Inverardi L (2001)Prolonged islet graft survival in NOD mice by blockade ofthe CD40–CD154 pathway of T-cell costimulation. Diabetes50:270–276

16. Nanji SA, Hancock WW, Luo B, Schur CD, Pawlick RL, Zhu LF,Anderson CC, Shapiro AM (2006) Costimulation blockade ofboth inducible costimulator and CD40 ligand induces dominanttolerance to islet allografts and prevents spontaneous autoimmunediabetes in the NOD mouse. Diabetes 55:27–33

17. Daoussis D, Andonopoulos AP, Liossis SN (2004) TargetingCD40L: a promising therapeutic approach. Clin Diagn LabImmunol 11:635–641

18. Kawai T, Andrews D, Colvin RB, Sachs DH, Cosimi AB (2000)Thromboembolic complications after treatment with monoclonalantibody against CD40 ligand. Nat Med 6:114

19. Henn V, Slupsky JR, Gräfe M, Anagnostopoulos I, Förster R,Müller-Berghaus G, Kroczek RA (1998) CD40 ligand onactivated platelets triggers an inflammatory reaction of endothelialcells. Nature 391:591–594

20. Langer F, Ingersoll SB, Amirkhosravi A, Meyer T, Siddiqui FA,Ahmad S, Walker JM, Amaya M, Desai H, Francis JL (2005) Therole of CD40 in CD40L- and antibody-mediated platelet activa-tion. Thromb Haemost 93:1137–1146

21. Allen SD, Rawale SV, Whitacre CC, Kaumaya PT (2005) Thera-peutic peptidomimetic strategies for autoimmune diseases: costimu-lation blockade. J Pept Res 65:591–604

22. Sozzani S, Allavena P, D'Amico G, Luini W, Bianchi G, Kataura M,Imai T, Yoshie O, Bonecchi R, Mantovani A (1998) Differentialregulation of chemokine receptors during dendritic cell maturation:a model for their trafficking properties. J Immunol 161:1083–1086

23. Bussolati B, Grange C, Tei L, Deregibus MC, Ercolani M,Aime S, Camussi G (2007) Targeting of human renal tumor-derived endothelial cells with peptides obtained by phagedisplay. J Mol Med 85:897–906

24. Russo S, Bussolati B, Deambrosis I, Mariano F, Camussi G(2003) Platelet-activating factor mediates CD40-dependent angio-genesis and endothelial–smooth muscle cell interaction. J Immu-nol 171:5489–5497

25. Solanilla A, Pasquet JM, Viallard JF, Contin C, Grosset C,Déchanet-Merville J, Dupouy M, Landry M, Belloc F, Nurden P,

196 J Mol Med (2009) 87:181–197

Blanco P, Moreau JF, Pellegrin JL, Nurden AT, Ripoche J (2005)Platelet-associated CD154 in immune thrombocytopenic purpura.Blood 105:215–218

26. Flaxenburg JA, Melter M, Lapchak PH, Briscoe DM, Pal S (2004)The CD40-induced signaling pathway in endothelial cells result-ing in the overexpression of vascular endothelial growth factorinvolves Ras and phosphatidylinositol 3-kinase. J Immunol172:7503–7509

27. Mirabet M, Barrabés JA, Quiroga A, Garcia-Dorado D (2008)Platelet pro-aggregatory effects of CD40L monoclonal antibody.Mol Immunol 45:937–944

28. Anderson ME, Yakovleva T, Hu Y, Siahaan TJ (2004) Inhibitionof ICAM-1/LFA-1-mediated heterotypic T-cell adhesion to epi-thelial cells: design of ICAM-1 cyclic peptides. Bioorg Med ChemLett 14:1399–1402

29. Kim MY, Byeon CW, Hong KH, Jeong S (2005) Inhibition of theangiogenesis by MCP-1 (monocyte chemoattractant protein-1)binding peptide. FEBS Lett 579:1597–1601

30. Liang S, Lin T, Ding J, Pan Y, Dang D, Guo C, Zhi M, Zhao P,Sun L, Hong L, Shi Y, Yao L, Liu J, Wu K, Fan D (2006)Screening and identification of vascular–endothelial-cell-specificbinding peptide in gastric cancer. J Mol Med 84:764–773

31. Lee L, Buckley C, Blades MC, Panayi G, George AJ, Pitzalis C(2002) Identification of synovium-specific homing peptidesby in vivo phage display selection. Arthritis Rheum 46:2109–2120

32. Hajitou A, Pasqualini R, Arap W (2006) Vascular targeting: recentadvances and therapeutic perspectives. Trends Cardiovasc Med16:80–88

33. Bergmann S, Pandolfi PP (2006) Giving blood: a new role forCD40 in tumorigenesis. J Exp Med 203:2409–2412

34. Danese S, Scaldaferri F, Vetrano S, Stefanelli T, Graziani C,Repici A, Ricci R, Straface G, Sgambato A, Malesci A, Fiocchi C,Rutella S (2007) Critical role of the CD40 CD40–ligand pathwayin regulating mucosal inflammation-driven angiogenesis in in-flammatory bowel disease. Gut 56:1248–1256

J Mol Med (2009) 87:181–197 197