Humanfusionproteins between interleukin 2 …BstEII BsEII GGATCC I..Xbal IL2-Um FIG. 1. Structures of IL2-Mu and IL2-Um cDNAs and the corresponding proteins. IL2 sequences are identified.

Proc. Nail. Acad. Sci. USAVol. 89, pp. 11337-11341, December 1992Immunology

Human fusion proteins between interleukin 2 and IgM heavy chainare cytotoxic for cells expressing the interleukin 2 receptorHENRI VIE*, THIERRY GAUTHIERt, RICHARD BREATHNACH*, MARC BONNEVILLE*, ANNE GODARD*,JACQUES DIETRICHt§, GEORGES KARAM**, MARIE-CLAUDE GESNEL*, MARIE-ALIX PEYRAT*,YANNICK JACQUES*, AND J. P. SOULILLOU*tII*Institut National de la Santd et de la Recherche M6dicale (U.211), UnitE de Recherche sur les Effecteurs Lymphocytaires T, tLynatech S. A., §SangstatAtlantique, and tCentre Hospitalo-Universitaire, Plateau Technique C. H. R., Quai Moncousu, Nantes 44035 Cedex, France

Communicated by Jean Dausset, July 23, 1992

ABSTRACT We have constructed a hybrid cDNA codingfor a fusion protein between human interleukin 2 and atruncated heavy chain from human immunoglobulin M. Theprotein encoded by this cDNA contains the entire interleukin 2sequence including its signal peptide, fused at its C terminus todomains 2 to 4 of the immunoglobulin heavy-chn constantregion. Cells transfected with the hybrid cDNA secrete multi-meric forms of the fusion protein, which bind specifically tocells bearing high-affinity interleukin 2 receptors. This bindingleads either to T-cell proliferation or, if complement is added,to T-cell death. Multimeric forms of the fusion protein with amolecular mass above 500 kDa mediate complement-dependentlysis but trigger proliferation inefficiently when compared withforms with a low molecular mass (<500 kDa). In contrast, thelatter efficiently mediate T-cell proliferation without inducingcomplement-dependent lysis. The high molecular mass fusionproteins could thus constitute valuable tools for specific im-munosuppression in humans.

Organ transplants are being carried out with, increasingfrequency and represent a powerful approach to the treat-ment of various organ end-stage diseases. One of the prob-lems associated with this technique is graft rejection. Immu-nological events leading to rejection are initiated by theactivation of genetically precommitted recipient T cellswhich recognize donor antigens. Reagents that would permitthe selective destruction of these T lymphocytes, whichrepresent <1% of the T-cell population (1, 2), without inter-fering with the resting T-cell immune repertoire would thusfind many uses. Activated, but not resting, T lymphocytescarry on their surface a high-affinity receptor for interleukin2 (IL2) comprising at least two different chains, the a chain,or p55 (3), and the P3 chain, or p75 (4). Interaction of thisreceptor with IL2 is of crucial importance for T-lymphocyteproliferation. Various strategies for blocking graft rejectionhave been successfully developed which target this high-affinity receptor that is specific for the activated T-lympho-cyte population (see refs. 5 and 6 for reviews).Mouse monoclonal antibodies directed against IL2-binding

epitopes of the p55 chain inhibit IL2 binding to the high-affinity receptor and block the IL2-driven growth ofactivatedT lymphocytes (7). These antibodies decrease alloimmune ordelayed-type reaction in both animals (8-10) and humans (5,6). However, drawbacks of the antibodies are their relativelylow affinity (as compared with IL2) for the receptor and theirimmunogenicity. Furthermore, binding of the IL2 receptorby the monoclonal antibodies does not lead to efficientcomplement-induced cytotoxicity. Hybrid human/mousemonoclonal antibodies against the IL2 receptor have beendeveloped (11, 12). In principle, these humanized antibodies

should show reduced immunogenicity (production of feweranti-isotypes), though anti-idiotypes may still be producedand act as blocking antibodies. Another approach is the useof IL2 molecules fused to bacterial toxins such as thediphtheria (13) or Pseudomonas (14) toxins. These moleculeshave a higher binding affinity for the IL2 receptor than themonoclonal antibodies (12). Their internalization leads totoxin-induced cell death. They remain immunogenic, how-ever (15). This may be an important drawback with IL2-diphtheria toxin fusion proteins, as many people alreadypossess anti-diphtheria antibodies.The ideal reagent would be a fully humanized, poorly

immunogenic structure with a high affinity for the IL2receptor and good cytotoxic potential for activated T lym-phocytes. Along these lines, an IL2-immunoglobulin fusionprotein may have considerable advantages. Indeed, the util-ity of immunoglobulin (lg) fusion proteins has been demon-strated in studies developing reagents targeted to the AIDSvirus (16, 17). Also, a fusion protein between IL2 and theIgGI heavy-chain constant region (Cy) has been describedwith some complement-dependent cytotoxicity against IL2receptor-positive cells (18). However, a fusion protein be-tween IL2 and the IgM heavy-chain constant region (C.M)would theoretically be more advantageous, resulting in morepotent complement-mediated killing and in a higher avidityfor the IL2 receptor due to multivalency. Furthermore,inhibition of the IL2-mediated signal after interaction of amultivalent fusion protein with the IL2 receptor is a realpossibility (19-21).

In this paper, we describe a human IL2-truncated IgMheavy-chain fusion protein. Multimeric forms of this fusionprotein are toxic for activated T lymphocytes in the presenceof complement but weakly trigger their proliferation in itsabsence. These molecules constitute a potential immunosup-pressive drug for use in transplantation.

MATERIALS AND METHODSHybrid cDNA Construction. cDNA reverse-transcribed

from Jurkat cell RNA was used as a template for thepolymerase chain reaction with oligonucleotide primers de-rived from the human IL2 cDNA sequence: primer 1, 5'-CGTCGACTCCTGCCACAATGTACAGG-3'; primer 2, 5'-CGGATCCAGTCAGTGTTGAGATGATGC-3'. Primer 1contains part of the IL2 cDNA's 5' untranslated region andthe codons specifying the first 3 amino acids of the IL2 signalsequence (ATG-TAC-AGG) linked to a Sal I site. Primer 2 iscomplementary to the sequences coding for the very C-ter-minal amino acids of IL2 and contains aBamHI cleavage site

Abbreviations: IL2, interleukin 2; C;,, IgM heavy-chain constantregion; BLCL, B-lymphoblastoid cell line.'lTo whom reprint requests should be addressed.

11337

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

8, 2

020

Proc. Natl. Acad. Sci. USA 89 (1992)

in place of the IL2 stop codon. Use of these primers gener-ated a 0.48-kilobase-pair (kb) Sal I-BamHI fragment.A partial C, cDNA clone (21) was used as a template for

the polymerase chain reaction with the following primers(22): primer 3, 5'-CGGATCCGTGATTGCTGAGCT-GCCTCCC-3'; primer 4, 5'-CTCTAGAGGGTCAGTAG-CAGGTGCCAG-3'. Primer 3 contains the codons forthe first7 amino acids of the CA2 domain (GTG-ATT-GCT-GAG-CTG-CCT-CCC) linked to a BamHI cleavage site. Primer 4contains the sequences complementary to the codons for theC-terminal amino acids of the C,4 domain, including the stopcodon, and also contains an Xba I cleavage site. Use oftheseprimers generated a 1.06-kb BamHI-Xba I fragment, whichwas introduced between the corresponding sites ofthe vectorpKCSRa (a derivative of pKCR3; ref. 23 and R.B., unpub-lished work). The IL2 Sal I-BamHI fragment was then placedupstream of the C,. sequences in the resulting plasmid togenerate pIL2-Mu.The orientation of a BstEII fiagment within the C., coding

sequences of pIL2-Mu was inverted to generate pIL2-Um. A1.5-kb IL2-IA cDNA Kpn I fragment excised from pIL2-Muwas inserted in the unique Kpn I site ofpKCR6 (23) to obtainpIL2-Mu6.

Preparation of the Anti-p-Chain Column. A goat anti-human ,u chain antibody (Biosys, Compiegne, France) wascoupled to Affi-Gel 10 (Bio-Rad, Ivry sur Seine, France).COS supernatant concentrate (141 ml) was loaded on theimmunoaffinity column by continuous recirculation during 18hr at a flow rate of 0.2 ml/min. After the column was washedsuccessively with phosphate-buffered saline (PBS: 140 mMNaCl/20 phosphate, pH 7.4/0.1% Tween 20) and with 0.1 xPBS, elution was performed with 100mM glycine/HCl buffer(pH 2.5) and the pH of the fractions was immediatelyneutralized by adding 0.2 volume of 1 M K2HPO4.

Immunoprecipitation. Transfected cells were metabolicallylabeled with [35S]methionine and [35S]cysteine. Fifty micro-liters of labeled supernatant from transfected cells was addedto 150 ,ul of 10 mM Tris HCl, pH 7.5/150 mM NaCl/0.2%Nonidet P-40/2 mM EDTA and subjected to immunoprecip-itation for 1 hr at 4°C with 25 jig of the goat anti-human 1chain polyclonal antibody or with 2.5 &g of rabbit anti-humanIL2 polyclonal antibody (Genzyme). After addition of excessprotein A-Sepharose (Bio-Rad), samples were incubatedovernight at 4°C with gentle agitation. After extensive wash-ings, bound proteins were subjected to SDS/PAGE analysis.

Eectrophoresis. SDS/PAGE was carried out (24) in 4.5-16% and 3-7% polyacrylamide gradient slab gels. Gels werethen subjected to autoradiography.Cation-Exchange and Gel Filtration Chromatography. Two

liters of supernatant from transfected CHO cells was diluted2-fold in 20 mM phosphate buffer (pH 7) (buffer A), and thepH was adjusted to 7 with HCL. The diluted supernatant waspassed at 4°C and at 3 ml/min through a cation-exchangecolumn (Hiload S-HP 16/10 from Pharmacia France) previ-ously equilibrated in buffer A. The column was rinsed withbuffer A and two elution steps were carried out with bufferA containing first 0.5 M and then 1 M NaCl. The fractioneluted at 0.5 M NaCl was then loaded at 4°C onto a SephacrylS-300 gel filtration column (Hiload 26/60 from PharmaciaFrance) previously equilibrated with buffer A containing 1.5M NaCl. Fractions were collected every 4 min at 0.5 nl/min.

Cell Lines. CTLL-2 is an IL2-dependent murine T-cell line;4AS is a human alloreactive T-cell clone (25); the B-lympho-blastoid cell line (BLCL) is a human B-cell line obtained bytransformation with Epstein-Barr virus; DA-la is an IL3-dependent murine cell line (26); K562 is a human erythro-myeloblastoid cell line; JAR is a human neuroblastoma line(ATCC; HTB 144); and Jurkat is a human leukemic T-cellline.

IL2-Mu Assay. The IL2-l4 fusion protein IL2-Mu wasdetected by an enzyme-linked immunosorbent assay(ELISA) with polyclonal rabbit anti-human IL2 antibody(Biosys) coated on the plates and goat anti-human At chainpolyclonal antibody coupled to peroxidase (Clinisciences,Paris). Immunopurified IL2-Mu was used as a standard in theassay.Flow Cytometry. IL2-Mu binding to target cells was de-

tected with a standard staining procedure using fluoresceinisothiocyanate-conjugated goat anti-human 1L chain antibody(Kallestedj Austin, TX), and was analyzed on a FACScan(Becton Dickinson).Complement-Dependent Cytotoxicity. Target cells were in-

cubated with ice-cold culture supernatant from pIL2-Mu-,pIL2-Um-, or pKCSRa-transfected COS-1 cells for 30 minand washed twice before addition of rabbit complement(Cedarlane, Toronto, Canada) at 10%6 (vol/vol) in RPMI 1640(GIBCO/BRL). After a 45-min incubation at 37C, cell via-bility was assayed either by counting cells under the micro-scope in the presence of eosin or by measuring radioactivityreleased by cells that had been prelabeled with Na251Cr2O4(2209 MBq/ml; Amersham France) (27).

Estimation of IL2-Mu Binding Affinity. The affinity ofpurified high molecular mass (>500-kDa) IL2-Mu species forthe IL2 receptor on CTLL-2 cells was estimated with flowcytometry by titrating the mean fluorescence obtained with0-20 nM IL2-Mu and a biotinylated anti-human IgM antise-rum (Kallested).

RESULTSHybrid cDNA Construction. We constructed a hybrid

cDNA (see Materials and Methods and Fig. 1) coding for asecreted fusion protein, IL2-Mu, in which the IL2 and C,sequences are joined together by a two amino acid linker,Gly-Ser. The fusion protein should be able to use the IL2sequence to bind to the IL2 receptor and use the C, sequenceto bind complement. In addition, because the fusion proteincontains the cysteine residues implicated in multimerizationof IgM heavy chains (Fig. 1), it should itself form multimers.A variant form of the hybrid cDNA was also produced by

reversing the orientation of a BstEII fragment (see IL2-UmcDNA, Fig. 1) contained within the heavy-chain codingsequence. This cDNA codes for a truncated fusion protein,IL2-Um, which lacks part of the C,,2 domain and all of theIgC,.3 and IgC,,4 domains and thus should be unable to formmultimers or bind complement but retain the ability to bindto the IL2 receptor. IL2-Mu and IL2-Um cDNA were intro-duced into the vectors pKCSRa or pKCR6 under the controlof a simian virus 40 early gene promoter and transfected intoeither COS (pKCSRa constructs) or CHO cells (pKCR6

IL2-Mu 1 ,2 IIL2-Mu Sall..ATG IiL2l7cDNA

112-Urn Sa~llATG L II2 I]cDNA

cys cys cys

GLY SER | 2 | 3 | 4 1BsEII BdEII

GGATCC | 2J I3J|jTGA..XbalBstEII BsEII

GGA TCC I ..Xbal

IL2-Um

FIG. 1. Structures of IL2-Mu and IL2-Um cDNAs and thecorresponding proteins. IL2 sequences are identified. The g&-chainsequences are represented by boxes 2-4 corresponding to domains2-4 of the constant region. The three cysteine residues important formultimerization are marked. IL2 and ,u sequences are linked in thecDNA by a BamHI recognition site (GGA-TCC), and in the proteinsby Gly-Ser. The IL2-Mu and IL2-Um cDNAs differ in the orientationof a BstEII fragment (restriction sites marked).

11338 Immunology.- Vid et al.

IL2 I GLY SER

Dow

nloa

ded

by g

uest

on

Mar

ch 2

8, 2

020

Proc. Nat!. Acad. Sci. USA 89 (1992) 11339

constructs). Gene amplification was carried out for trans-fectedCHO cells in the presence ofincreasing concentrationsof methotrexate up to 4 uM, a concentration at which aconsistent production of 10-20 ,ug IL2-Mu per ml was ob-tained.Both IL2-Mu and EL2-Um Stimulate Cel Growth. Fusion

proteins were tested for their ability to promote the growth ofthe IL2-dependent murine T-cell line CTLL-2 or lectin-activated human T lymphocytes. Supernatant from COS-1cells transfected with pIL2-Mu or pIL2-Um, unlike that fromcells transfected with the "empty" expression vectorpKCSRa, specifically elicited proliferation of both murine(Fig. 2) and human (data not shown) activated T cells.Although this suggests that IL2-Mu and IL2-Um can bind tothe IL2 receptor and stimulate cell growth, proliferationmeasured might have been due to the presence of IL2 itselfin supernatant from transfected cells. Such IL2 could intheory be derived from the fusion proteins by proteolyticcleavage. This possibility was ruled out by the finding that thegrowth-stimulatory activity in the culture supernatant couldbe retained on an anti-IgM column, whereas the same columndid not bind IL2.

Specifc Bing of EI2 Mu to Cels Bearing IL2 Reeptors.Various cells types were incubated with COS-1 cell super-natant containing IL2-Mu. Binding of IgM sequences to cellswas assayed by staining the cells with fluorescently labeledgoat anti-human ju chain antibody. Staining of IL2 receptor-positive cells (CTLL-2 and 4AS) was observed, whereas IL2receptor-negative cells (Jurkat and JAR) were not stained(Fig. 3). When COS-1 cell supernatant containing IL2-Umwas used, none of the cell types was stained. A kinetic studyfollowing antigenic stimulation ofan alloreactive T-cell clonedemonstrated parallel increases in the amount of cell surfaceIL2 receptor, as assayed by using a monoclonal antibodydirected against the a chain (7), and the amount ofcell-boundp-chain sequences (data not shown). These data demonstratethat IL2-Mu binds specifically to cells expressing the IL2receptor.BindlngofIL2-Mu, but Not IL2-Um, Leads toCme t-

d Cttoxicit. 51Cr-labeled CTLL-2 lymphocytes orhuman alloreactive T-cells (clone 4AS) were incubated withrabbit complement and supernatants from COS-1 cells trans-fected with pIL2-Mu, pIL2-Um, or pKCSRa. Lysis ofCTLL-2 cells was observed after incubation with IL2-Mu-containing supernatant and rabbit complement, but no killingwas seen when IL2-Um-containing supernatant and rabbit

CTLL-2 4AS Jurkat JAR

E8 L~iJLLJLLFluorescence intensity

FIG. 3. Specific biding of IL2-Mu to IL2 receptor-positive cells.Receptor-positive (CTLL-2 and 4AS), or receptor-negative (Jurkatand JAR) cells were incubated with supernatant from COS-1 cellstransfected with pIL2-Mu, or with PBS alone, and binding ofIL2-Muwas analyzed by fluorescence with a fluorescein-labeled goat anti-human ju chain antibody fragment. Fluorescence was detected witha FACScan.

complement were used (Fig. 4). The phenomenon was spe-cific, as IL2 receptor-negative cell lines (such as K562 orBLCL) were not killed under the same assay conditions.Purificao and Anyis of the Varos 112-Mu Molecular

Forms. Supernatant from pIL2-Mu6-transfected CHO cellswas subjected to immunoprecipitation, and bound proteinswere eluted and analyzed by SDS/PAGE in 4.5-16% (Fig. 5)and 3-7% (data not shown) polyacrylamide gradient slab gelsunder reducing and nonreducing conditions. Identical immu-noprecipitation patterns were obtained with either anti-IL2 oranti-gM antibodies (Fig. 5), demonstrating that the recom-binant material secreted by transfected CHO cells comprisedboth the IL2 and the it-chain moieties. Under nonreducingconditions, various molecular forms could be distinguishedranging from 68 to 650 kDa (Fig. 6), whereas under reducingconditions, a single band was observed with the expectedmolecular mass of the monomeric form (68 kDa). We de-duced that IL2-Mu was obtained as a mixture of formsraning from monomers to decamers. A two-step chromato-graphic protocol allowed purification of IL2-Mu from crudeCHO supernatant. A cation exchanger bound all the IL2-Muwhile eliminating 95% of the total protein input in theflow-through (data not shown). After elution with 0.5 MNaCl, the EL2-Mu fraction obtained was enriched 60-foldcompared with the crude supernatant. Further purificationwas obtained by gel filtration. Total IL2-Mu recovery astested by ELISA was 82%. Each fraction was tested in anELISA to quantitate the amount of IL-2Mu, and in bothproliferation and cytotoxicity assays to compare the respec-tive functional activity of the various IL2-Mu forms. Twomain IL2-Mu-containing fractions could be distinguished,

K562

CTLL-2

BLCL

_,-

1.1.n 10 2C 3C

,- 'Cr release

40C

IL2 or COS supematant dilution

FIG. 2. lL2-Mu and IL2-Um induce proliferation of the 1L2-dependent murine T-cell line CTLL-2, as estimated by [3Hlthynuidineincorporation. Comparative effects on CTLL-2 proliferation of var-ious dilutions of recombinant IL2 (150 units/ml) (0) or COS-1supernatant dilutions containing IL2-Mu (i, *) or IL2-Um (*).IL2-Mul and IL2-Mu2 are supernatants from two different trans-fections.

FIG. 4. IL2-Mu, but not IL2-Um, induces complement-dependent cytotoxicity of IL2 receptor-positive cells. Receptor-positive (CTLL-2) or receptor-negative (K562 and BLCL) cellsprelabeled with 51Cr were incubated first with ice-cold supernatantfrom COS-1 cells transfected with pIL2-Mu (black bars), pIL2-Um(open bars), or pKCSRa (gray bars), for 30 min and then with a 1:10dilution of rabbit complement. Results are expressed as percent 51Crrelease.

7---- i0---

Immunology: Vid et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

8, 2

020

Proc. Nadl. Acad. Sci. USA 89 (1992)

M E . :3 rV1 E 60

11^ r ~ -i

._ t-2

40

20

0 111d iij ,,ll.d 111in

0.0001 0.001 0.01 0.1 1 10 100IL2-Mu, nM

FIG. 7. Dose-dependent binding of the high molecular mass1L2-Mu fraction to CTLL-2 cells. CTLL-2 cells were labeled withincreasing concentrations of 1L2-Mu and analyzed on a FACScancytometer. The mean fluorescence intensity (MFI) derivedfrom eachanalysis is plotted against 1L2-Mu concentration.

FIG. 5. Detection ofdifferent 1L2-Mu forms. Culture supernatantfrom CHO cells transfected with pIL2-Mu6 or untransfected wassubjected to immunoprecipitation with protein A alone (protein A),protein A plus anti-,u (anti-Mu), or protein A plus anti-IL2 (anti-IL2)and analyzed by SDS/PAGE (4.5-16%) in the absence or presenceof 2-mercaptoethanol (BME).

one of high molecular mass (>500 kDa) and one of lowmolecular mass (<500 kDa). The former fraction was foundto contain most of the cytotoxic activity and showed a lowproliferative activity. Conversely, the latter fraction was notcytotoxic but contained most ofthe proliferative activity. Forexample, in the experiment shown in Fig. 6, the IC50 cyto-toxicity/IC50 proliferation ratio was 256/7.9 for fractionnumber 2 and 0/40 for fraction number 36.

Estination ofHih Molecular Mass L2-Mu Bindng Affnityfor the IL2 Receptor. Attempts to radioiodinate IL2-Mufractions led to a dramatic loss of their biological activity.Therefore, flow cytometry was used as an alternative toevaluate the binding affinity of IL2-Mu for the IL2 receptor.The results from such an experiment on CTLL-2 cells areshown in Fig. 7. The mean fluorescence intensity of thelabeled cells was found to increase in a dose-dependent way,

100 1900kDa 6692kDa 40kDa 240 0a

0

0.

~60

E

~20-

0 10 20 30 40

Fraction

FIG. 6. IL2-Mu purification for gel filtration analysis of IL2-Muspecies. Five milliliters of the fraction eluted at 0.5 M NaCl from thecation exchanger was loaded onto a Sephacryl S-300 column. Frac-tions were collected every 4 min at a flow rate of 0.5 ml/min. Theconcentration ofIL2-Mu in each fraction was determined by ELISA(- - -). Also shown are the specific proliferative (o) and cytotoxic (o)activities of each fraction as estimated from the ELISA data. Toshow both proliferation and cytotoxicity data on the same figure,each value within each test was normalized to the maximal response.

eventually reaching a plateau. The apparent affinity of thehigh molecular mass IL2-Mu fraction (>500 kDa) couldtherefore be derived by fitting the data to a simple hyperbolicsaturation function. The apparent dissociation constant (Kd)derived from five independent experiments was 450 + 130pM.

DISCUSSION

Our results allow some observations to be made about thestructural behavior of IL2 fusion proteins. First, it is clearthat IL2 can be fused to other sequences either via its Cterminus (our results) or via its N terminus [IL2-diphtheriatoxins (13)] without losing its capacity to bind to the IL2receptor. Second, the ability ofbound IL2-Mu to induce lysisof IL2 receptor-positive cells in the presence of complementdemonstrates that its Clq binding site is exposed under thesecircumstances. In a natural IgM, an IgM-ligand interactiontriggers a conformational change required to expose the Clqbinding site. It remains to be determined whether this site isalways exposed or whether it becomes exposed only afterbinding of IL2-Mu to its receptor. Finally, the truncated CAchain induces multimerization of IL2-Mu. Gel filtration anal-ysis demonstrated that although the fusion protein was se-creted as a mixture of multimers of different molecular mass,only the high molecular mass forms possessed cytotoxicpotential.The low mitotic induction by large polymeric IL2-Mu

forms (>500 kDa) is likely to be related to modifications inIL2-receptor internalization (28). Indeed, this is reminiscentof the inhibitory effect on IL2-driven growth of activatedcells by IgM monoclonal antibodies directed against anIL2-noncompeting epitope on the p55 chain of the IL2receptor, whereas corresponding monovalent fragmentsfrom the same IgM were not inhibitory (19). This has beenconfirmed recently in a human system (20). Whether thisinhibition of IL2-driven proliferation by high molecular massforms of IL2-Mu would play a role in vivo, where comple-ment-mediated cytotoxicity, opsonization, or antibody-dependent cell-mediated cytotoxicity can operate, remainsunknown.

Similarly, only in vivo experiments will tell whether thelarge multimeric forms of IL2-Mu have access to the site ofthe immune response. However, it is noteworthy that IL2receptor-positive cells are not restricted to graft tissues butmigrate in the blood. It has also been reported that cytotoxicIgM monoclonal antibodies have a high efficiency in vivo(29).The affinity of large forms of IL2-Mu (>500 kDa) for the

IL2 receptor has been repeatedly evaluated by flow cyto-

1 IC C Cl

CC-.

u...

1;C -* 41'

14 ->

11340 Immunology: Vid et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

8, 2

020

Proc. Natl. Acad. Sci. USA 89 (1992) 11341

metric analysis on the murine CTLL-2 line, which is a widelyused standard for such experiments. The apparent dissocia-tion constant obtained on this murine line (Kd, 450 ± 130 pM)compares with that of human IL2 for the murine high-affinityreceptors (100 pM). Since polymeric IL2-Mu is expected tocross-link a large number of IL2 receptors on CTLL-2 cells,one might have expected a higher apparent affinity as a resultof multiple binding units. The apparent dissociation constantwe report here may thus be only an average of differentbinding strength interactions that the IL2 molecules can formwith different IL2 receptor conformations on CTLL-2 cells(high-affinity p55/p75 complexes and low-affinity p55 chainsin excess). Use ofhuman and/or murine cells expressing onlyhigh-affinity IL2 receptors or isolated p55 or p75 chainswould help in further understanding the binding ofIL2-Mu toits receptor.

Besides its ability to effect complement-mediated killing, amajor potential advantage of the presented fusion moleculelies in its human nature. Indeed, so far, all bioreagents usedin immunosuppressive treatments elicit a rapid immune re-sponse in vivo. This is well known for unmodified monoclonalantibodies against the IL2 receptor (5) or CD3 determinants(30). Chimeric or complementary determinant-region recom-binant molecules derived from them (11) will probably de-crease the incidence of host immunization but cannot intheory avoid production of blocking anti-idiotypic antibod-ies. IL2-toxin fusion proteins cannot trigger IL2-blockingantibodies, but the bacterial toxin component of these fusionproteins may elicit antibodies (15). In contrast, the fusionprotein we describe contains only human sequences and willprobably not elicit antibody production when administered tohumans, which would allow long-term treatment with thisreagent, although we cannot rule out the possibility thatconformational antibodies can be elicited against this fusionprotein.The protein we describe may thus prove to be particularly

advantageous in the treatment of allograft recipients and ofpatients with autoimmune diseases or bearing tumors ex-pressing the IL2 receptor. In addition, the strategy wedescribe here can clearly be adapted for the construction ofother fusion proteins containing different ligands linked to theIgM heavy chain.

We are grateful to M. Fougereau and C. Schiff (Marseille, France)for providing the partial C, cDNA and to Ms. Aline Bertho andNathalie Lopes for excellent secretarial assistance. This work wassupported in part by the Ministere de la Recherche et de la Tech-nologie, Paris.

1. Zang, L., Van Bree, S., Van Rood, J. F. & Li, F. H. J. (1990)Transplantation 50, 1008-1010.

2. Ryser, J. E. & MacDonald, H. R. (1979) J. Immunol. 122,1691-1696.

3. Uchiyama, T., Broder, S. & Waldman, T. A. (1981) J. Immu-nol. 126, 1393-1397.

4. Sharon, M., Klausner, R. D., Cullen, B. R., Chizzonite, R. &Leonard, W. J. (1986) Science 234, 859-863.

5. Soulillou, J. P., Cantarovich, D., Le Mauff, B., Giral, M.,

Robillard, N., Hourmant, M., Him, M. & Jacques, Y. (1990) N.Engl. J. Med. 322, 1175-1182.

6. Le Mauff, B., Cantarovich, D., Jacques, Y. & Soulillou, J. P.(1990) Transplant. Rev. 4, 79-91.

7. Jacques, Y., Le Mauff, B., Godard, A., Olive, D., Moreau,J. F. & Soulillou, J. P. (1986) J. Immunol. 136, 1693-1699.

8. Kirkman, R. L., Barrett, L. V., Gaulton, G. N., Kelley, V. E.,Ythier, A. & Strom, T. B. (1985) J. Exp. Med. 162, 358-362.

9. Kelley, V. E., Bacha, P., Pankewycz, O., Nichols, J. C.,Murphy, J. R. & Strom, T. B. (1988) Proc. Natl. Acad. Sci.USA 85, 3980-3984.

10. Dantal, J., Jacques, Y. & Souliliou, J. P. (1991) Transplanta-tion 52, 110-115.

11. Queen, C., Schneider, W. P., Selick, H. E., Payne, P. W.,Landolfi, N. F., Duncan, J. F., Avdalovic, N. M., Levitt, M.,Junghans, R. P. & Waldmann, T. A. (1989) Proc. Natl. Acad.Sci. USA 86, 10029-10033.

12. Brown, P. S., Parenteau, G. L., Dirbas, F. M., Garsia, R. J.,Goldman, C. K., Bukowski, M. A., Junghans, R. P., Queen,C., Hakimi, J., Benjamin, W. R., Clark, E. & Waldmann, T. A.(1991) Proc. Natl. Acad. Sci. USA 88, 2663-2667.

13. Bacha, P., Williams, D. P., Waters, C., Williams, J. M., Mur-phy, J. R. & Strom, T. B. (1988) J. Exp. Med. 167, 612-622.

14. Lozberboum-Golski, H., Fitzgerald, D., Chandhary, V., Ad-hya, S. & Pastan, I. (1988) Proc. Natl. Acad. Sci. USA 85,1922-1926.

15. Kirkman, R. L., Bacha, P., Barrett, L. V., Forte, S., Murphy,J. R. & Strom, T. B. (1989) Transplantation 47, 327-330.

16. Traunecker, A., Schneider, J., Kiefer, H. & Karjalainen, K.(1989) Nature (London) 339, 68-70.

17. Capon, D. J., Chamow, S. M., Mordenti, J., Marsters, S. A.,Gregory, T., Mitsuya, H., Byrn, R. A., Lucas, C., Wurm,F. M., Groopman, J. E., Broder, S. & Smith, D. H. (1989)Nature (London) 337, 525-531.

18. Landolfi, N. F. (1991) J. Immunol. 146, 915-919.19. Kumar, A., Moreau, J. F., Gibert, M. & Theze, J. (1987) J.

Immunol. 139, 3680-3684.20. Duprez, V., Ferrer, M., Comet, V., Olive, D. & Dautry-Varsat,

A. (1991) J. Biol. Chem. 266, 1497-1501.21. Milili, M., Fougereau, M., Guglielmi, P. & Schiff, C. (1991)

Mol. Immunol. 28, 753-761.22. Word, C. J., White, M. B., Kuziel, W. A., Shen, A. L., Blatt-

ner, F. R. & Tucker, P. W. (1989) Int. Immunol. 1, 298-309.23. Matrisian, L. M., Bowden, G. T., Krieg, P., Furstenberger,

G., Briand, J. P., Leroy, P. & Breathnach, R. (1986) Proc.Natl. Acad. Sci. USA 83, 9413-9417.

24. Laemmli, U. K. (1970) Nature (London) 227, 680-685.25. Chabannes, D., Le Mauff, B., Hallet, M. M., Jacques, Y. &

Soulillou, J. P. (1988) Transplantation 46, 97S-100S.26. Moreau, J. F., Bonneville, M., Godard, A., Gascan, H., Gru-

art, V., Moore, M. A. S. & Soulillou, J. P. (1987) J. Immunol.138, 3844-3849.

27. Vie, H., Chevalier, S., Garand, R., Moisan, J. P., Praloran, V.,Devilder, M. C., Moreau, J. F. & Soulillou, J. P. (1989) Blood74, 285-290.

28. Robb, R. J. & Green, W. C. (1987) J. Exp. Med. 165, 1201-1209.

29. Friend, P. J., Hale, G., Waldmann, H., Gore, S., Thiru, S.,Joysey, V., Evans, D. B. & Calne, R. Y. (1989) Transplanta-tion 48, 248-253.

30. Chatenou, L., Baudrihaye, M. F. & Chkoff, N. (1986) J.Immunol. 137, 830-838.

Immunology: Vie' et aL

Dow

nloa

ded

by g

uest

on

Mar

ch 2

8, 2

020