AU714902B2 - Means for detecting and preventing HIV infection involving use of receptors or binding sites capable of interacting with GP120 - Google Patents

Description

WO 96/41884 PCT/IB96/00571 1 MEANS FOR DETECTING AND PREVENTING HIV INFECTION INVOLVING USE OF RECEPTORS OR BINDING SITES CAPABLE OF INTERACTING WITH gp120 The invention relates to means for detecting and preventing from an infection by the human immunodeficiency virus, HIV, AID causative agent.

It more particularly relates to new receptors or binding sites capable of interacting with gpl20 and participating in virus infection. The invention also relates to the biological applications of said receptors or binding sites.

Since 1984, the CD4 molecule was described to be the HIV-1 receptor by which infection occurs and (numbers correspond to the bibliographic references which are given at the end of the description). The HIV envelope protein gpl20 binds with high affinity to the CDR2 domain of CD4 and to purified recombinant CD4 and CD4-negative cells could also be infected, which led to the description of other receptors, such as complement receptors and glycolipids in glia cells receptors on fetal astrocytes or on brain capillary endothelial cells A secondary receptor differing from CD4 has also been described on CD4+ Langerhans cells in the skin By carrying out investigations on the effect of serum on the binding of recombinant gpl20 to CD4+ whole cells, in ELISA and FACS, the inventors have identified new binding sites. Such results are in contrast to those on the interaction of rgpl20 with isolated recombinant, soluble CD4 reported in (11) and (12).

The inventors have also found accessory molecules for virus in erythrocyte ghosts and have then if CONFIRMATION COPY WO 96/41884 PCT/IB96/00571 2 explored gpl20 binding to erythrocytes and its membrane proteins.

It is then an object of the invention to provide new gpl20 receptors or binding sites.

It is also an object of the invention to provide binding sites to said new receptors on the molecule.

It is another object of the invention to provide serum factors capable to compete with gpl20 for the binding to binding sites on CD4+ cells.

The new gpl20 receptors or binding sites of the invention are characterized by the fact they are found on CD4+ cells, but differ from the CD4 binding receptor since they are involved when gpl20 binds to CD4+ cells which are blocked by a monoclonal antibody directed against the gpl20 binding site of CD4, they give interactions with gpl20 such as those obtained when reacting gpl20 with CD4+ cells in the absence of serum.

As demonstrated by the examples hereinafter given, the binding of gpl20 to CD4+ cells is indeed cited to the so-called CD4 receptor, but involves other interactions which can be identified in the absence of serum and at certain gpl20/cell ratios.

Said interactions are considerably inhibited in the presence of serum.

The new gpl20 receptors or binding sites of the invention are such as involved in a process comprising incubation of gpl20 with CD4+ cells under duration and temperature conditions allowing binding of gpl20 to cells with a CD4 blocked receptor.

In such a process, the CD4+ cells are for example pre-incubated for about 1 hour at about 37*C with WO 96/41884 PCT/IB96/00571 3 a monoclonal antibody specific for the gpl20 binding domain of CD4 and gpl20 is pre-incubated with bovine serum albumin at about physiological concentration.

The CD4+ cells have intact CD4 as shown by the binding of a labelled monoclonal antibody directed against the gpl20 binding site of CD4. At least 90 of this binding is inhibited by pre-incubation with The CD4+ cells are living or dried cells. Such cells are for example, peripheral blood monocyte cells (PBMC) and living or microtiter plate adherent, dried, CD4+ CEM cells.

The gpl20 secondary receptors or binding sites of the invention are further characterized by the fact that, in the presence of about 10 of serum, their interactions with gpl20 are inhibited by about 80 More particularly, said secondary receptors or binding sites are capable of interacting in the presence of about 10 of serum, with 25 to 30 of gpl20, for concentration of 0.5 and 1 pg/ml and with 40 of gpl20 for 2 pg/ml depending on the gpl20/cell ratio.

The new gpl20 receptors or binding sites of the invention are still further characterized by the fact that, in the absence of serum, they are capable of interacting with more than 20% of gpl20 as shown by FACS analysis.

More particularly, said new receptors or binding sites are capable of interacting with about 80% of gpl20 at gpl20 doses of 0,5 or 1 pg/ml, and 50 60 at 2 pg/ml, in ELISA with BSA at about physiological concentration and in the absence of fetal calf serum or human serum.

WO 96/41884 PCT/IB96/00571 4 Such results demonstrate the secondary interactions of gpl20 with whole living cells.

The interaction is more pronounced on dried CEM cells with accessible intracellular receptors.

Said new gpl20 receptors or binding sites are localized on the outer cell membrane of living cells.

They appear to possess lower affinities than the CD4 receptor, since they appear to be present on all cells, as shown by double labelling, but bind about at least of total bound Binding to said receptors is considerably increased when intracellular sites become available, as found with dried cells by ELISA or with different kinds of permeabilized cells by FAGS analysis.

The new gp 120 receptors or binding sites are further characterized by the fact that their binding activity for gp 120 are not destroyed when CD4 is deleted on CD4+ cells, for example by trypsin, formaldehyde or deoxycholate.

Accordingly, they are more resistant than said CD4 receptor to proteolytic enzymes and chemical alteration.

Said gpl20 receptors or binding sites will also be hereinafter designated as serum inhibited receptors

(SIR)

Investigation experiments carried out by the inventors have shown a strong interaction between and the proteins belonging to the families of the multipassage transmembrane proteins, such as present at the surface of human erythrocytes, or lymphocytar cells such as CEM cells.

Accordingly, the invention relates to new receptors or binding sites such as above defined, further characterized by the fact they comprise at least a part WO 96/41884 PCT/IB96/00571 of the proteins of multi-passage transmembrane protein families, or lymphocyte cells such as CEM cells.

Transmembrane proteins comprise proteins whose MW is in a range of about 40 kDa to 150 kDa.

For example, multi-passage transmembrane proteins in erythrocyte comprise band 3 protein (100 kD), glucose transporter (54 kD), Rh50 (44 kD), calcium transporting ATPase (134 kD), sodium transporting ATPase (112 kD).

In order to examplify the invention, it will hereinafter be referred to proteins of band 3-like protein family, although the invention encompasses any multi-passage transmembrane protein capable of interacting with gpl20 under the conditions disclosed in the examples.

Said band 3-like proteins are known to be ubiquitous and may be found in the plasmic membrane, the Golgi apparatus, as well as in mitochondria in various cells.

It comprises proteins having homology, i.e. the so-called band 3 protein which, when originating from erythrocytes, possesses a MW of 102 kD, band 3 AE 2 (AE anion exchange) with a MW of 112 kD, and band 3 AE3 having a MW of 137 kD (the MW were evaluated in SDS-

PAGE).

According to the cDNA deduced amino acid sequence (13) band 3 protein from erythrocytes (EXPASY, Swiss-prot P 02 730) possesses 911 amino acids, several hydrophobic regions allowing to predict 12-14 transmembrane passages, a 40 kD C-terminal and a short Nterminal cytoplasmic tail The 4 major extracellular loops contain approximately 14 positively and 6 WO 96/41884 PCT/IB96/00571 6 negatively charged amino acids with a ratio of 2.3 in favour of the positive charges. The domination of the positive charges could have a physiological role for anion exchange as they seem to constitute a positively charged funnel contributing to repulsion of cations and selection for anions Band 3 AE 2 protein with a total of 1240 AA and a 303 AA larger cytoplasmic tail (EXPASY, Swiss-protP 04 920) is an anion exchanger of wide distribution, first described for the leucocyte derived cell line K562 Band 3 protein AE 3 is very similar with 1232 AA containing also a 303 larger cytoplasmic tail (EXPASY, Swiss-prot P 48 751) and is localized in heart and brain. Band 3-like protein A2 present in lymphocyte derived cells (16 EXPASY Swissprot.) were shown to have high crossreactivity with rabbit antibodies to erythrocyte band 3 protein.

According to the invention, the receptors or binding sites of HIV gpl20 are thus characterized in that they comprise at least a part of proteins of the band 3like protein family.

The experiments carried out by the inventors, which are given hereinafter in the examples, have shown absorption of gpl20 by band 3 from erythrocytes or competition of band 3 protein with CD4 blocked CEM cells for gpl20 binding or virus infection. The high crossreactivity between both band 3-like molecules indicates the involvment of this kind of transmembrane molecules as accessory receptors for virus infection.

The invention particularly relates to the receptors or binding sites comprising at least a part of band 3 protein such as present on human erythrocytes, or WO 96/41884 PCT/IB96/00571 7 lymphocyte cells such as CEM cells, particularly on the outer membrane surface.

Said receptors or binding sites are also characterized in that they are capable of interacting with gpl20, in the absence of human serum, at 370C.

The interaction between gpl20 and band 3 protein is indeed inhibited by serum.

Said receptors are such as present on HIVinfectable cells, such as CD4+ CEM cells.

The invention also relates to transmembrane proteins, with several passages from inside to outside of the cell, under a purified form, and the fragments thereof, particularly the extra-membrane regions, especially those positively charged, inasmuch as said fragments and regions have a specific binding activity with respect to the gpl20 present on HIV.

The term "protein", such as used in the description and the claims, designates a protein of the transmembrane protein families, under a purified form, or in vesicles or in reconstituted liposomes, as well as the fragments or regions of the protein, such as above identified.

Said term also encompasses the recombinant forms of the protein, for example the fusion proteins, or the forms which are totally or partially devoid of their intramembane sequence.

The invention relates to the purified band 3 like proteins separated from at least partially of the membrane associated proteins. Particularly, such purified proteins, especially band 3 proteins, are practically totally devoid of spectrin, ankyrin, only once crossing WO 96/41884 PCT/IB96/00571 8 transmembrane proteins such as glycophorin and complement receptor CR1.

As already above-mentioned, it is specifically referred to band 3 protein to examplify the invention, but said expression designates any other multi-passage transmembrane protein and correspond to the accessible transmembrane proteins on the outer membrane surface, or fragment thereof, inasmuch as they are able to react with as above defined.

According to another aspect, said new receptors or binding sites are such as involved when reacting gpl20 with erythrocytes adsorbed to plates or pretreated by high proteolytic enzyme concentrations of potent enzymes, such as Pronase R The invention also relates to the polyclonal or monoclonal antibodies directed against said protein. The fragments of said antibodies, particularly the Fab region are also included in the invention.

The invention also relates to anti-Ig antibodies, i.e. the anti-idiotype antibodies.

Said anti-idiotype antibodies are capable of reacting, according an antigen-antibody reaction, with the above defined antibodies. The invention also covers the fragments of said antibodies.

The polyclonal antibodies are obtained according to usual methods by immunizing animals with band 3 protein, or fragments or regions thereof, recovering the antibodies obtained from splenic cells, and purifying the same.

Monoclonal antibodies are obtained by cultivating hybridoma, under the usual conditions which comprise fusioning splenic cells secreting antibodies WO 96/41884 PCT/IB96/00571 9 with myeloma cells, selecting those clones secreting hybridomas, and injecting said clones to mice for producing tumoral ascites from which the desired monoclonal antibodies are recovered and purified.

The above defined proteins and antibodies are useful for elaborating detecting means and therapeutic or prophylactic compositions.

The invention thus relates to detection or diagnostic means, compositions and kits, comprising said proteins or said antibodies in a sufficient amount to characterize an HIV infection and, optionally, to quantify the HIV level in a sample to be analyzed or the immune response against HIV.

In detection or diagnostic compositions, the proteins or the antibodies advantageously comprise a marker revealing the immunological reaction.

According to the invention, the method for detecting the presence of HIV or the immune response comprises contacting the sample to be analyzed with the secondary receptor such as above defiied, said receptor being capable to bind to the HIV-RIS reaction product optionally contained by the sample, or to the SIR site, or alternatively, with the above defined protein, capable to bind to HIV, where the inhibition of the fixation of the viral proteins on SIR will indicate the presence of antibodies capable to hinder the virus fixation on said receptor, revealing the antigen-antibody reaction when occurred.

The contacting step is carried out under appropriate conditions, particularly with respect to the WO 96/41884 PCT/1B96/00571 duration of the reaction, the temperature, and the buffer, so as obtain the HIV fixation when HIV is present.

The sample to be analyzed is more particularly a body fluid such as blood, plasma, urine, salive, cerebro-spinal fluid, seminal liquid.

Antibodies or proteins are used in solution or fixed on a support.

Labelling means are used to reveal the reaction, such as fluorescent agents for example fluorescein, enzymes like peroxidase, or coloring agents.

The detection of the presence of HIV or anti- HIV antibodies in view of studies or diagnostic, will be advantageously carried out by using a kit comprising said protein or said antibody, optionally fixed on a microplate, in a sufficient amount for performing the test, reagents and buffers for the detection as well as instructions for use.

According to another aspect, the invention relates to compositions useful for therapy, comprising, in association with pharmaceutically acceptable carriers, an efficient amount of at least one antibody, or one protein, such as above-defined.

The antibodies are thus used as inhibitor agents with respect to the interaction with HIV.

Band 3 protein, or its fragments, is used to compete with the transmembranar protein the SIR site cells.

Advantageously, the pharmaceutical compositions of the invention comprise agents capable of protecting the antibodies or proteins administered to a patient against the effect of cellular proteases.

WO 96/41884 PCT/IB96/00571 11 The composition of the invention may also be used in association with antiviral agents.

For such therapeutical applications, the proteins or the antibodies are administered under galenic forms for parenteral or intravenous administration.

They are also administrable as liposomes.

Gp 120 used in the investigation experiments is a recombinant glycosylated gpl20 from HIV-1 from baculovirus. But the intact virus with oligomeric may be used instead of the recombinant gpl20 Inhibition of virus infection in vitro by band 3 protein vesicles, as shown by the examples hereinafter given, demonstrates indeed that virus also possesses very similar structures to recombinant gpl20 concerning the interactions with said new receptors or binding sites.

The term "gpl20" as used in the specification and the claims encompasses all the forms of gpl20, i.e.

the recombinant form or the natural viral form. The invention also relates to the part of gpl20 such as bound on living cells, distinct from the binding to CDR2 of CD4, when incubating CD4 cells with monoclonal antibodies specific for the gpl20 binding domain of CD4.

The presence of gpl20 on the cells is revealed with an anti-gpl20 rabbit anti-serum followed by an antirabbit Ig coupled to peroxidase for OPD coloured reactiver as usually done in ELISA. The corresponding binding site on the gpl20 molecule is characterized by the fact that it is localized, with monoclonal antibodies, near the V3 and the CD4 binding region. The localization of the binding region near the V3 region is WO 96/41884 PCT/IB96/00571 12 highly interesting, since it is known to interact with neutralizing antibodies.

Taking into account the low efficacy of CD4 blockade observed on dried cells in ELISA and in FACS analysis, it is believed that the gpl20 molecule possesses at least two binding sites, with different specificities, one for the CD4 receptor and an other for said new receptors or binding sites. Gp120 might then bind to both cellular sites simultaneously, thereby inducing the formation of a bridge between both cellular receptors.

The invention also relates to the part of gp120 whose binding to said outer cell membrane sites of living cells can be partially inhibited by pre-incubation of gpl20 with fetal calf serum or human serum.

It will be appreciated that said gp binding site is of great interest for determining the amino acid and the corresponding nucleotidic sequences involved in the binding to the new receptors and for studying their anti-HIV effect.

Polyclonal and monoclonal antibodies directed against the binding sites situated on the gp120 molecule are also part of the invention as well as their diagnostic on therapeutic applications.

The invention also relates to the serum factors capable of at least partially inhibiting gpl20 binding to said new gpl20 receptors or binding sites.

Said serum factors are then characterized by the fact they compete with gpl20 binding on said receptors or binding sites and/or modify the WO 96/41884 PCT/IB96/00571 13 molecule itself by direct binding or by partial degradation.

The serum factors and the fractions are useful to study the inhibitory effect of the VIH infection of target cells. The results obtained concerning said receptors or binding sites give useful tools to test the validity of vaccinating substances.

The invention will be further illustrated by the following examples taken in conjunction with the accompanying drawings, wherein Figures 1A and 1B represent the binding of or of mAB F101.69 to CEM cells and inhibition of mAB binding by gpl20 (FACS), Figures 2A and 2B, the gpl20 binding on CD4 CEM cells in presence of blocking mAB F101.69 or of FCS or both (FACS), Figures 3A to 3C, and 4A to 4C, the binding to treated or dried cells under different conditions (FACS and ELISA respectively), Figure 5A to 5D, the double labelling of blocked CD4 receptor and cell bound rgpl20 (FACS), Figure 6, the double labelling of gpl20 and blocked CD4 on CEM cells in the cytofluorometric microscope.

Figure 7, the localisation on the molecule of the binding site to the new receptors, figures 8A and 8B, the total binding to dried CEM cells or specific binding to CD4 after prior iodination of gpl20 by idogen, figures 9A and 9B, the inhibition of binding by mAB F101.69 or serum, or both together, figures 10A and 10B, the gpl20 binding to human erythrocytes and the inhibition thereof by serum, WO 96/41884 PCT/IB96/00571 14 figure 11, the SDS electrophoresis of highly enriched band 3 protein, figure 12, the gpl20 binding to purified transmembrane proteins from erythrocytes, figure 13A and 13B, the inhibition of binding to CD4 CEM cells by human serum and band 3 protein in erythrocyte vesicles, respectively, figure 14, the absorption of gpl20 by enzyme treated erythrocytes or band 3 protein vesicles, and figure 15, the inhibition of RT after adsorption of HIV to band 3 vesicles.

MATERIAL AND METHODS

MATERIALS

Microtiter plates were purchased from NUNC (Roskilde, Denmark) bovine serum albumin (BSA) glycophorin, octyl-glycopyranoside from Sigma (St. Louis, Pronase inhibitor and trypsin inhibitor from Boehringer (Mannheim, Germany) trypsin from Serobal, FCS from Gibco 4-15 PAGE SDS gels from Pharmacia (Upsala) biobeads SM-2 from Bio-Rad (Richmond, Ca.) erythrocytes and normal human serum from the Regional Blood Distribution Center (Montpellier) Antigens and antibodies Monoclonal mouse antibodies (mABs) known to block gpl20 binding -to the CD4 receptor, F101.69 and F92.3A11 (Sanofi Research, Montpellier, France) and OKT4a (purchased from Ortho Diagnostics, France) were used. The mAB F101.69 had been classified as an anti-CD4 antibody during the Third International Workshop on Human Leukocyte Differentiation Antigens It binds to the CDR2 domain of.CD4, also known to bind gpl20 from HIV-1, and is inhibited by mutations of amino acids 42-43 WO 96/41884 PCT/IB96/00571 Recombinant, glycosylated gpl20 from HIV-1/IIIB produced in baculovirus was used (ABT, now Intracell purchased from NEOSYSTEM, Strasbourg, France, or AGMED, Bedford, MA, USA).

Gp 120 was labelled by antigen-antibody complex formation with anti-HIV mAB 110.4 (2 molar excess) which was directed against V3 of gpl20 (GENETIC SYSTEMS). This antibody had been conjugated before (16 and 17) with periodate (PROLABO, Paris, France) to peroxidase (BOEHRINGER, Mannheim, Germany). The peroxidase labeled antibody in the absence of gpl20 served as a specific control for every experimental point and was subtracted from the experimental values. Unlabelled gpl20 was revealed with an anti-gpl20 hyperimmune antiserum from rabbits, followed by peroxidase conjugated to anti-rabbit Ig from sheep (Amersham, England). The rabbit antiserum was obtained after a first immunization with in complete Freund's adjuvant, followed by monthly hyperimmunizations without adjuvants in doses of 1 pg/kg.

The rabbit anti-gpl20 antiserum was obtained after a first immunization with gpl20 in complete Freunds adjuvant, followed by monthly hyperimmunizations without adjuvants in doses of 1 pg/kg. Anti-rabbit Ig and antimouse Ig conjugated to peroxidase came from Amersham, England, and anti-rabbit Ig or anti-mouse Ig conjugated to FITC or phycoerythrin from Sigma, France.

sCD4 was obtained from the NIH AIDS Research and Reference Reagent Program (cat. n* 1813, lot 394150 and cat. n* 1246, lot 1CD51009) and also from ABT, now Intracel, Cambridge, MA (cat. n* 13001, lot 55-90-1A).

Cells Peripheral blood mononuclear cells (PBMC) were freshly isolated from healthy donors with Ficoll-Paque (PHARMACIA, Paris, France). CD4+CEM cells were obtained WO 96/41884 PCT/IB96/00571 16 from the ATCC, USA, and cultured in RPMI plus 10% FCS.

Microtiter plate adherent dried cells, were obtained after three washes of CEM cells in PBS without addition of proteins and distribution of 10 s

CEM

cells/well in non saturated maxisorb U-microtiter plates pl per well). Cells were dried overnight at 37C in closed boxes with silica gel (Prolabo, Paris, France) distributed around the microtiter plates. Before use, the wells were saturated with 3% BSA. Non-adherent dried cells for FACS analysis were obtained in the same way, but in microtiter plates after prior saturation with BSA 3% to avoid adherence of cells.

ELISA

After saturation of U.maxisorb microtiter plates (Nunc, Roskilde, Denmark) with 3% BSA for 30 min at 37*C (RIA grade, Sigma) 10 5 living cells, human PBMC or CEM (ATCC, USA) were distributed per well and washed twice with PBS+BSA 0.3% (200 pl/well). Gpl20 was preincubated with 10% FCS (decomplemented 30 min. at 56'C) or 3% BSA for 1 h at 37'C and then incubated in quadruplicates with the cells for 30 min. at 37*C in the indicated incubation buffer (PBS plus BSA 3% or 10% FCS).

Cells were then washed twice in PBS with 0.3% BSA and further incubated (30 min. at 37*C) with a 1/1000 dilution of a rabbit anti-gpl20 antiserum. After 2 washes cells were incubated with a peroxidase labeled antirabbit Ig from sheep Amersham, UK) for 30 min.

at 37'C. After another 3 washes peroxidase was revealed with OPD and stopped with H 2

SO

4 4N. Optical densities were read at 492 nm in a Multiskan (Flow Laboratories, France).

In blocking experiments, cells were incubated min. at 37*C) with CD4 blocking antibodies F101.69 or OKT4a at 10 pg/ml concentrations in PBS-BSA 3% followed WO 96/41884 PCT/IB96/00571 17 by two washes before reaction with rgpl20. In some experiments (as indicated), the blocking mAB was not washed away before incubation with Every ELISA plate included two internal standards of 0.5 pg/ml gpl20 in BSA which served as references for comparison between different microtiter plates. Experimental values were expressed as ratios of the corresponding reference values. Optical densities of references varied between 1.0 to 2.0. For each experimental quadruplicate, controls consisted of quadruplicates containing peroxidase labeled antibody without gpl20, the values of which were subtracted from the experimental values. The control for peroxidase labeled F101.69 consisted of a similarly labeled mAB at the same concentration. Wells without cells, but saturated with BSA indicated low nonspecific binding to the wells, which never exceed Inhibition studies of the binding of gpl20 to cells were effected with anti gpl20 mABs after prior incubation of gpl20 with the mABs at 10 pg/ml for lh at 37'C. The mixture was then incubated and further processed on dried cells, as described.

In experiments with erythrocytes anti-species Ig antibodies linked to alkaline phosphatase were used and revealed with nPPD.

Analysis by a fluorescence activated cell sorter (FACS) Experiments with CEM cells for FACS analysis were carried out in microtiter plates under the same conditions as for ELISA. Only during the last step was anti-Ig coupled to peroxidase replaced by anti-Ig antibodies linked to fluorochromes at determined optimal dilutions. For dual fluorescence analysis mouse mAB F101.69 directed against CD4 was revealed with anti-mouse WO 96/41884 PCT/IB96/00571 18 Ig-FITC and rgpl20 incubated with rabbit antiserum with anti-rabbit coupled to phycoerythrin.

During FACS analysis (FACSort, Becton Dickinson) homogeneous cell populations (5x103 cells) were selected according to scatter parameters and then analyzed for mean fluorescence intensity (MFI) for each fluorochrome.

An internal reference consisted of 2 pg/ml rgpl20, which was run in all experiments. In order to compare different experiments with each other by mean values and standard deviations, the reference MFI was defined as 1 and all experimental values were transformed into relative MFI values by calculating the corresponding ratios (exp./ref.).

In some experiments, cells were incubated with propidium iodide (PI) from Sigma, France for 10 min. at 4*C before FACS analysis, to determine and exclude dead cells. In such experiments rgpl20 was revealed with rabbit anti-rgpl20 antiserum followed by anti-rabbit Ig- FITC (Sigma).

The Kruskal-Wallis test, a Chi-Square approximation,was used for the statistical analysis.

Cvto-fluorimetric microscopy CEM cells (10 cells per ml) were prepared and stained as for FACS analysis. Cells were then concentrated by centrifugation, layered on a slide and air dried at room temperature. Sections were analyzed on the scanning stage of the ACAS 570 Interactive laser Cytometer (Moridian, Okemos, MI, USA). Dual fluorescence expression was derived from pseudo-color two-dimensional image scans generated by a laser beam and an X-Y scanning stage (0.800 pM) and analyzed with two detectors each generating an image. Pseudo color linear scale range from 0 to 4096, 570 and 930 nM filters were used for PE and TRI-COLOR analysis respectively.

WO 96/41884 PCT/IB96/00571 19 CD4 deletion Living or microtiter adherent dried cells were after saturation with BSA 3 incubated in microtiter plates (30 min, 37*C) with increasing concentrations of trypsin, paraformaldehyde or deoxycholate.

After treatment, cells were immediately washed three times in PBS-BSA 0.3 and resaturated with BSA 3 The presence of CD4 was detected in ELISA by incubation of treated cells with the F101.69 mAB linked to peroxidase (30 min, 22'C) and revealed with OPD. For FACS analysis cells were incubated with F101.69 (30 min, 37*C), followed by an anti-mouse IgG-30 min, 22'C), similar to the above described method. Binding of after incubation with cells (lh, 37C) was studied by further incubation with rabbit anti-gpl20 antiserum min, 22'C) revealed with an anti rabbit Ig POD (30 min, 22*C) and OPD for ELISA or anti-rabbit Ig (30 min, 22*C) for FACS.

Iodination of gp 120 lodogen, 70 nmol (30 pg in 1 ml chloroform), was precoated on glass tubes (14 x 100 nm) by evaporation under nitrogen. After one wash in citrate phosphate buffer pH 7.4 and complete elimination of liquid, 50 pg/ml of rgpl20 were incubated in a final volume of 180 ul (citrate phosphate buffer pH 7.4 with 0.5 BSA) containing 0.3 pg/ml cold NaIO 4 with the iodogen coated tube under continuous stirring with a vortex. After different time periods the reaction was stopped by transfer of gpl20 to an other tube and immediate 1:20 dilution in PBS containing 3 BSA and 0.2 mg/ml tyrosine.

Modified gpl20 was examined for its binding activity to CD4 by inhibition of the mAB F101.69 to bind WO 96/41884 PCT/IB96/00571 to CD4. After saturation of microtiter adherent CEM cells with PBS-BSA 3 cells were incubated with the modified (lh, 37*C) and after two washes with PBS-BSA 0.3 further incubated with F101.69 linked to peroxidase min, 22"C) and revealed with OPD.

Binding of modified gpl20 to said receptor or binding site was tested by its binding to dried CEM cells, which predominantly show binding to said receptor or binding site. After saturation with 3 BSA cells were incubated with modified gpl20 (lh, 37*C) and, after 2 washes, further incubated with an rabbit antiserum (30 min, 22*C) followed by an antirabbit Ig coupled to peroxidase (30 min, 22*C) and revealed with OPD.

Modification of the gpl20 structure was tested by adsorption of unmodified or modified gpl20 at the concentrations of 25 pg/ml and two 1:4 dilution on maxisorb microtiter plates (lh, 37'C). After saturation with BSA 3 wells were incubated with a rabbit antigpl20 antiserum (30 min, 22'C), followed by anti-rabbit Ig-POD (30 min, 22*C) and OPD. At 50 inhibition the concentrations of modified and unmodified gpl20 were compared to each other.

Enzyme treatment Freshly taken erythrocytes from healthy donors were washed 3x in PBS, incubated with Pronase trypsine in PBS for lh at 37*C under agitation, washed 3x in PBS BSA 0.3 inhibitor and then incubated with 2 pg/ml for Ih at 37*C under agitation. After centrifugation the supernatant was incubated with microtiter adsorbed CEM cells and gpl20 binding revealed by the inhibition.

WO 96/41884 PCT/IB96/00571 21

RESULTS

1- INHIBITION OF GP120 BINDING TO WHOLE CELLS BY CD4 BLOCKADE WITH mABs (ELISA) from HIV-1/IIIB was preincubated with 10%FCS and controls with BSA 3% for 1 h at 37*C, PBMC, living CEM or microtiter plate adherent, dried CEM cells were saturated with 0.3% BSA and incubated with the anti- CD4 mABs: F101.69 or OKT4a and control with BSA or anti- CD2 mAB F92.3A11 for 30 min. at 37*C. Cells were further incubated with the pre-incubated gpl20 for 1 h at 37*C in the indicated pre-incubation medium. Gp120 complexed to mAB 110.4 which was linked to peroxidase, could be directly revealed with OPD. Unlabelled gp120 was further incubated with rabbit anti-gpl20 antiserum followed by peroxidase labelled sheep anti-rabbit Ig and then revealed with OPD.

The results are given in Table 1 (experiments A to K) Incubation mediulm control cell gpl12O binding nn CflA hln-L-rl ,'all C! I% medium Illualla 0 OKT4a FI11.69 (anti CD4 mAB) PBMC (living) FCS 10%

BSA

B3SA BSA or aCD2

BSA

aCD2 23 23 5 17 77 83 74 72 82 67 77 86 19 77 79 8.8 4.4 7.7 AvsB A vs C B vs C 0.01 0.*01 n. s.

CEM cells (living)

FCS

BSA

BSA

10% BSA or aCD2

BSA

aCD2 3 2 16 17 26 77 77 72 85 82 79 81 98 9.9 5.1 10.3 "Dvs E 0.01 "Dvs F 0.01 E vsF n.s.

CEM cells (dried) FCS 10% FCS 100%

BSA

BSA or aCD2 BSA or aCD2

BSA

83 76 25 22 90 98 97 98 99 98 Different gpl.2O lot on CEM cells (living) 2.7 0.9 (G-IH) vs J 0.02 K BSA BSA 51 45 62 34 48 11.7 Compared to anti-CD2 mAB F92.3A1 1 instead of BSA SD standard deviation n.s. not significant WO 96/41884 PCT/IB96/00571 23 binding to PBMC in ELISA In order to favour serum interactions, recombinant gpl20 from baculovirus was pre-incubated with FCS (lh, 37*C) and further incubated with PBMC in FCS (30 min. at 37*C). It bound to the cells with optical densities 1, which were considered as 100% binding.

Preincubation of PBMC with mAB F101.69 which recognises the gpl20 binding region in CD4 and inhibits binding, reduced bound gpl20 to 19%, as compared to controls (Table 1A, 5 experiments). Therefore gpl20 in presence of FCS bound essentially to the CD4 receptor, as expected from the literature.

The influence of serum on gpl20 binding was now tested by omitting serum completely. Gpl20 was preincubated in BSA an almost physiological concentration, and further incubated with PBMC also in BSA Gpl20 bound to the cells and its OD was considered as 100% gpl20 binding to control cells. Other target cells were first blocked at the CD4 receptor with mAB F101.69 and then incubated with gpl20. Now, the blocking antibody reduced gpl20 binding only to 77% with respect to controls (Table 1B, 5 experiments). The results in absence and presence of FCS were significantly different (Table 1, A vs. B p 0.01). This means that, in the absence of FCS, about 80% of cell bound gpl20 was not inhibited by the anti-CD4 antibody and appear to be directed to an other cellular binding site, different from the known CD4 epitope.

It was verified that the blocking mAB F101.69 was in sufficient excess to inhibit gpl20 binding to CD4 completely. In all experiments, the antibody was used at pg/ml, which is more than 10 x higher than saturation concentrations in ELISA. It was also tested, if the blocking antibody was directed against the binding site WO 96/41884 PCT/IB96/00571 24 of gpl20 on CD4. Preincubation of the cells with 2 pg/ml inhibited subsequent binding of mAB F101.69 by more than Since it is possible that the blocking mAB could have altered the surface characteristics of the target cells, while the control in BSA without mABs were unchanged, complementary experiments were carried out, in which the controls were exposed to mAB F92.3All, directed to the unrelated CD2 receptor. Such experiments were first effected in the presence of FCS and gpl20 binding could be reduced to 20% by mAB F101.69 (table lA). When the experiments were effected in BSA 3% i.e. without FCS the blocking mAB reduced gpl20 binding only to 79% with respect to the contraol cells (Table 1C). The results in absence and presence of FCS were highly significant (A vs. C p<0.01). When target cells were blocked with the reference mAB OKTa, instead of mAB F10169, identical results of 79% were obtained (Table IC). This confirms that the residual gpl20 binding should be directed to a receptor different from the classical CD4 receptor, which is only diminished in the absence of serum.

GD120 binding to CEM cells in ELISA PBMCs are a heterogeneous population containing different cell types, i.e. CD4+, CD8+lymphocytes and monocytes. In order to show that secondary gpl20 binding can be found on purified CD4+ cells, CEM cells were examined in the same way as PBMCs. When experiments were effected in the presence of 10% FCS, blockade of CEM cells with the anti CD4 mAB F101.69, reduced binding to 13% (Table ID), indicating predominant binding to the CD4 receptor. However, in experiments performed in the absence of FCS, the same blocking antibody reduced gpl20 binding only to 79% (Table lE).

The differences were highly significatn (D vs.E: p<0.1), WO 96/41884 PCT/IB96/00571 whether control cells were maintained in BSA or incubated with the control mAB anti CD2 (Table 1F).

Replacement of mAB F101.69 by the reference mAB OKT4a also reduced gpl20 binding to 88% only (Table 1F). This confirms the existence of gpl20 binding to secondary receptors on CEM cells in the absence of serum. The secondary binding sites and CD4 are simultaneously present on the same CEM cell.

When similar experiments were effected with another gpl20 lot, slighty different values were obtained. In the absence of serum mAB F101.69 reduced binding on CEM cells to about 50% instead of Therefore gpl20 binding to secondary receptors may be influenced by the gpl20 lot used and could be generally estimated by ELISA in the range of 50% 80% of total bound rgpl20 (Table 1K).

GD20 binding to dried CEM cells in ELISA In order to exclude modulation of DC4 during incubation at 37*C as a responsible factor, which could have diminished the total amount of accessible CD4 receptors, microtiter adsorbed dried CEM cells were studied, which take up trypan blue, do not undergo endocytosis but present accessible membrane and intracellular receptors. It was verified with the corresponding mABs that CD4 binding sites for gpl20 were fully expressed on dried cells, as were CD5 and CD7, although CD2 had disappeared. When experiments were performed in the presence of 10% FCS, blockade of CD4 reduced gpl20 binding to 80% (Table 1G), which is much less than on living cells. An increased FCS concentration of 100% reduced gpl20 binding to 24% (Table 1H), showing a dose dependent effect of FCS. When experiments were effected in BSA i.e. in the absence of FCS, there was no significant reduction of gpl20 binding (Table lJ) WO 96/41884 PCT/IB96/00571 26 and this was also observed with the reference mAB OKT4a (Table 1J). Binding of an other glycoprotein to dried CEM cells, such as peroxidase, was not significantly increased. In these experiments, peroxidase conjugated to mAB 110.4 was used in the absence of gpl20 and its binding to dried cells did not exceed 10 20%. Therefore secondary binding sites on dried cells are inhibited by FCS in a dose dependent manner, requiring high concentrations for total inhibition. Modulation of CD4 could be excluded as a responsible factor for decreased binding to CD4 receptors.

2- GP120 BINDING TO CEM CELLS BY FACS ANALYSIS The existence of secondary binding sites was verified by a second method, FACS analysis, on living CEM cells.

SBinding of Dpl20 or of mAB F101.69 to CEM cells and inhibition of mAB binding by pl20 the results are illustrated on figures 1A and lB.

A) CEM cells were first incubated with (lh, 37'C) and then with anti-gpl20 rabbit antiserum, followed by anti-rabbit Ig-phycoerythrin (black squares) or, after incubation with rgpl20, CEM cells were incubated with anti CD4 mAB F101.69 which was detected with anti mouse Ig-phycoerythrin (open squares). Results were expressed as relative mean fluorescence intensities (MFI), which are the ratio of experimental versus internal standard arbitrary units.

B) CEM cells or PBMC were incubated with increasing concentration of mAB F101.69 (30 min. 37*C) and then revealed with an anti mouse Ig-FITC by FACS.

(mean values of 4 experiments with different donors for PBMC, compared by the intern reference for 10 pg/ml F101.69).

WO 96/41884 PCT/IB96/00571 27 Binding of gpl20 to CEM cells was saturable.

It increased in a dose dependent manner up to plateau levels at about 30 pg/ml (Fig 1A). There is competition between gpl20 and mAB F101.69 on the same binding site, since after incubation with 10 pg/ml of gpl20 subsequent binding of the mAB F101.69 is inhibited (Fig. 1A).

Saturation curves of mAB F101.69 on PBMC (4 experiments with different donors) or CEM cells indicated the minimum saturation concentration at 0.3 pg/ml.

Therefore a concentration of 10 pg/ml, which was used for CD4 blockade, was more than 1 log above the minimal saturation concentration in FACS analysis (Fig. 1B).

Gp120 binding on CD4+ CEM cells in presence of blocking mAB F101.69 or of 10% FCS or both the results are illustrated on figures 2A and 2B.

binding to CEM cells was studied at three concentrations 0.5, 1 and 2 pg/ml.

A) CEM cells were incubated with 10 pg/ml anti-CD4 mAB F101.69 (30 min. at 37*C), a concentration more than 1 log over saturation concentrations, in order to block all CD4 receptors and then further incubated with rgpl20 (black squares). (ii) Controls were in BSA 3% only (open squares). (iii) rgpl20 was pre-incubated with FCS (for 1 h at 37C) and further incubated with the cells in presence of 10% FCS (open triangles) or (iv) both treatments, on the cells and on rgpl20, were effected simultaneously :cells were incubated with blocking mAB F101.69 and gp20 was pre-incubated with FCS (black triangles). rgpl20 was detected with a rabbit anti-gpl20 antiserum and revealed with an anti-rabbit Igphycoerythrin antiserum for analysis in a cell sorter.

(mean values of 6 independent experiments, as compared by the intern reference of 2 pg/ml WO 96/41884 PCT/IB96/00571 28 B) The same results are represented as percent binding or percent inhibition, as compared to the controls for each rgpl20 concentration in BSA 3%.

With increasing gpl20 concentrations binding increased about 3 fold (Fig. 2A). The preincubation of with 10% FCS diminished its binding to cells to about 60% (Fig. 2B). In the presence of FCS, the CD4 receptors were blocked by mAB F101.69 and gpl20 binding was reduced to indicating binding essentially to CD4 (Fig. 2B).

In experiments with BSA 3% the blockade of the CD4 receptor by mAB F101.69 reduced gpl20 binding to about 20% at all three concentrations, (Fig. 2B) indicating that 20% of total gpl20 bound to secondary binding sites, different from the known CD4 binding domain (6 independent experiments). The percentage found by FACS is inferior to the one by ELISA, but confirms the existence of secondary binding sites by an independent method.

Binding of rgDl20 to CEM cells after their treatment under different conditions (FACS) CD4 deletion followed up by FACS In order to further demonstrate the existence of two different cellular receptors for experiments were carried out to destroy one of them by enzymatic or chemical methods without affecting the other. CEM cells were exposed to trypsin or paraformaldehyde at different concentrations and then analyzed by FACS for the presence of the CD4 receptor and for continuing gpl20 binding. Dead cells, labeled by PI uptake, were excluded. CD4 and gpl20 were stained with different fluorochromes, the anti-CD4 mouse mAB F101.69 was revealed with an anti-mouse IgG and the WO 96/41884 PCT/IB96/00571 29 rabbit antiserum by an anti-rabbit Ig-PE (phycoerythrine).

A) Living CEM cells were treated with trypsin and, after washing, resaturated with BSA 3 Treated cells were incubated with the anti-CD4 mAB F101.69 revealed by an anti-mouse Ig FITC for analysis in a FACs.

The results are given on figure 3A to visualize total CD4 (open triangles) or treated cells were incubated with rgpl20 (HIV-1 IIIB) from baculovirus and further revealed with rabbit anti-gpl20 serum, followed by anti-Ig phycoerythrin (PE) to show gpl20 binding (open squares). Finally trypsin treated cells were blocked at the CD4 with mAB F101.09 (at 1 log over saturation concentration) before incubation with gpl20, which was then revealed with rabbit anti-gp antiserum and antirabbit Ig-PE, as above (black squares) B) and C) The target cells were treated with paraformaldehyde (PFA) up to 10 mg/ml for 30 min. at 37'C, in order to get pore-formation and access to intracellular binding sites. These experiments were performed in BSA only. After resaturation, the cells were blocked by mAB F101.69 at the CD4 receptor and then double labeled for propidium iodide uptake and for the presence of The results are illustrated on figures 3B and 3C.

CEM cells were treated with paraformaldehyde (PFA) at different concentrations for 30 min. at 37*C.

After resaturation with BSA the CD4 of target cells was blocked with mAB F101.69 (30 min. at 37*C) and thereafter incubated with gpl20 for 30 min. at 37'C.

was detected with an anti gpl20 rabbit antiserum min. at 4'C) followed by an anti-rabbit Ig-FITC WO 96/41884 PCT/IB96/00571 min. at Propidium iodine uptake was simultaneously revealed before analysis in a cell sorter. (means values of 3 independent experiments, compared by the intern reference of 2 pg/ml Dead cells contained in the experiment or dried cells or PFA treated dead cells showing PI uptake were blocked or no blocked with mAB F101.69, and further incubated with rgpl20 as in Fig.3B for FACS analysis.

At low PFA concentrations, up to 0.1 mg/ml, PI uptake was low and increased rapidly between 1 and mg/ml PFA. Binding of gpl20 increased in a similar way between PFA concentrations of 1 10 mg/ml (Fig. 3B).

Since CD4 was blocked by mAB F101.69, only gpl20 bound to secondary receptors was measured. Therefore secondary receptors are more abundant in permeabilized cells, probably intracellularly, or they possess higher affinities.

Specific gpl20 binding to secondary sites was found to be increased in different kinds of permeabilized cells. In dried cells, PFA treated cells, or permeabilized cells, occurring during the experiment, 60%-70% of total bound gpl20 were fixed to secondary binding sites by FACS analysis (Fig. 3C).

CD4 deletion followed up by ELISA Deletion of CD4 was studied in parallel by the ELISA method, on microtiter plate adsorbed dried cells, which allow more drastic treatment of cells with detergents. It was first verified with specific mABs, that a variety of CD antigens resisted to the drying procedure of CEM cells (18h, 37'C). CD4, CD5 and CD7 were preserved after drying, so that dried CEM cells presenting intact CD4 binding sites, could be treated under the same conditions as living cells with enzymes, WO 96/41884 PCT/IB96/00571 31 PFA and even detergents to destroy CD4 receptor integrity.

treatment with trypsin (figure 4A) Trypsin was used at different concentrations, from 0.03 to 0.5 mg/ml.

Microtiter adherent, dried CEM cells were saturated with BSA 3 and treated with trypsin and after washing and resaturation with BSA 3 incubated with the anti-CD4 mAB F101.69 linked to peroxidase and then revealed with OPD to show presence of CD4 (open triangles), or dried cells were incubated with rgpl20 and revealed with a rabbit anti-gpl20 antiserum, followed by anti-rabbit Ig linked to peroxidase and OPD to demonstrate gpl20 binding (black squares).

With increasing trypsin concentrations, CD4 detection diminished progressively and disappeared almost completely at 0.25 mg/ml trypsin, very similar to FACS analysis.

The binding capacity of gpl20 in the absence of serum to trypsin treated cells did not diminish in ELISA, even after exposure to the highest trypsin concentration (4 experiments). This confirms a higher resistance of the receptors of the invention than CD4 to enzymatic degradation, as already found by FACS nalysis.

treatment with paraformaldehyde (figure 4B wherein Pi means propidium iodide incorporation).

When dried CEM cells were chemically treated with paraformaldehyde from 0.01 to 100 mg/ml, the integrity of CD4 domain 1 was reduced to only 25 as revealed with mAB F101.69 Nevertheless, gpl20 binding without serum increased slightly (3 experiments). Dried cells in ELISA simultaneously expose extra and intra cellular sites. Thus the increased gpl20 binding to the receptors of the invention, found with living CEM cells WO 96/41884 PCT/IB96/00571 32 by FACS analysis potentially after the appearance of holes in the membrane, could not be seen in ELISA. PFA treatment of dried cells also confirmed a higher resistance of said receptors to chemical modification.

treatment with deoxycholate (figure 4C) To further characterize both receptors, cells were exposed to more drastic conditions such as detergents, which could no longer be followed up by FACS, due to cell disruption. Dried CEM cells were incubated with deoxycholate in increasing concentrations from 6 to 100 mM and lost CD4 in a concentration dependent manner.

CD4 was no longer detectable after treatment with 50 mM deoxycholate. Nervertheless in the absence of serum, binding of rgpl20 to these cells was not decreased (4 experiments), showing that the binding site according to the invention resisted even a high concentration of an anionic detergent.

Double labelling of blocked CD4 receptor and CEM cell bound Secondary binding sites capture about 20% of bound gpl20 to living cells. To check that the receptor sites of the invention were not restricted to a subpopulation of 20% target cells, or that secondary binding sites were not present on all cells, but with reduced affinity for gpl20, CEM cells were double labeled for CD4 and for gpl20 simultaneously.

The results are given on figures 5A to A) CEM cells were incubated with rgpl20, which was revealed with an anti rgpl20 rabbit antiserum, followed by anti rabbit Ig-phycoerythrin.

B) CEM cells were incubated with 10 pg/ml mAB F101.69 (more than 1 log over saturation concentration) to block the CD4 receptor and revealed with an anti mouse Ig-FITC antiserum.

WO 96/41884 PCT/IB96/00571 33 D) Cells were double labeled by a first incubation with the mAB F101.69 to block the CD4 receptor and a second with rgpl20. Both, gpl20 and the mAB F101.69, were then revealed as in Fig.5A) and B) before analyses for double fluorescence in a FACS cell sorter.

C) unlabelled cells served as controls. (a representative experiment out of Gpl20 was revealed with rabbit anti gpl20 antiserum, followed by anti rabbit Ig-PE. The presence of CD4 was identified with an anti mouse Ig-FITC. More than 96% of all cells were double labeled by FACS analysis (5 experiments), so that almost all cells possessed CD4 and simultaneously bound Similar results were obtained, whether blocking mABs were washed away or not before contact of the target cells with gpl20. All detectable gpl20 was bound to secondary binding sites, since the CD4 receptor was blocked.

Therefore secondary receptors were present on almost all living cells, but with lower affinities than those of CD4 (Fig. 3 CYTO-FLUORIMETRIC MICROSCOPY Living CEM cells were first blocked with the anti-CD4 mAB F 101.69 (at a concentration of 10 pg/ml).

Without washing the mAb away, in order to maintain an antibody excess for total blockade, cells were further incubated with gpl20. After washing, gp 120 was then revealed with an anti-gp 120 rabbit antiserum followed by an anti-rabbit Ig-tricolor. The CD4 blocking monoclonal mouse antibody represented the CD4 receptor and was labeled in parallel with an anti mouse Ig coupled to phycoerythrin. The results are given on figure 6.

In the cyto-fluorimetric microscope, most cells showed the simultaneous presence of CD4 (red color) and WO 96/41884 PCT/IB96/00571 34 (green color) on the same cell, as verified by their double labeling (yellow color). This confirms that under saturation conditions and in the absence of serum a complete blockade of CD4 was unable to prevent binding to the same cell. The presence of said receptor or binding site was further confirmed on some cells which, despite staining very weakly for CD4, nevertheless, presented considerable gpl20 binding.

Moreover, gpl20 and CD4 showed a different cellular distribution. CD4 appeared more homogeneoulsy distributed over the whole cell, while gpl20 was found in patches, suggesting a different modulation of both receptors.

Double labeling and cyto-fluorimetric microscopic clearly distinguished said gpl20 receptors or binding sites, which were different from the classical CD4 receptor.

4 INHIBITION OF GP120 BINDING BY ANTI-GP120 MABs (figure 7) In order to localize the expected binding sites for secondary interactions on the gpl20 molecule itself, gpl20 was pre-incubated with mABs known to be directed to defined regions of gpl20, before the mixture of plus mABs was further incubated with dried cells. Out of 12 mABs, 8 directed to the V2, C3, C4 and the C-term.

regions had no or no significant binding inhibition with respect to gpl20 without mABs. (TABLE 2).

WO 96/41884 WO 9641884PCTIIB96/00571 TABLE 2 Region AB or sCD4 Epitope gp12O binding controls) concentrations (pgg/m1) 10 10 934 1 4476 -51 -470 -43 -410 2 I110-C 1 10-D 282 FTC 284 38 1-394 CD4-BS sCD4 -240 -220 -9.80 -380 conformationel CD41 C-term D7324 7E2/4 7E/4-9.9 -3 WO 96/41884 PCT/IB96/00571 36 Only 4 mABs, 2 of them directed to V3 (5042, 110.4), and the 2 others to the CD4 binding site (F91, 110-K), inhibited gpl20 binding to the secondary binding sites on cells to about 50% in two experiments. The antibody concentration of 10pg/ml corresponded to plateau levels of binding inhibition. The partial inhibition of binding suggests that the ligand recognition regions for secondary binding were localized near the V3 and the CD4 binding regions of gpl20 Since the anti V3 mAB (4476) did not inhibit suggests that the V3 region is only partially implecated.

MODIFICATION OF THE BINDING SITE OF GP120 FOR CD4 AND THE RECEPTORS BINDING SITES OF THE INVENTION In order to demonstrate the existence of at least two different binding sites on the gpl20 molecule, experiments were carried out to destroy one of them by chemical modification and the results are given on figures 8A and 8B. Iodination with iodogen and cold Nal was tried to alter tyrosines.

Recombinant gp 120 (50 pg/ml) in 0.5 BSA was iodinated by insolubilized iodogen (30 pg/ml) in presence of cold Nal for different time periods. Dried CEM cells were then incubated with the modified gpl20 to block the CD4 receptors followed by the anti-CD4 mAB F101.69 linked to peroxidase and revealed with OPD (open triangles).

Dried CEM cells were also incubated with the modified and then with rabbit anti-gp 120 antiserum, followed by anti-rabbit Ig-peroxidase and OPD, to show gpl20 binding to secondary receptors, which can only be seen on dried cells (black squares).

The CD4 binding site of gp 120 was completely destroyed after 1 h of iodination, while the binding site to the other receptor was only partially lost (about 30 WO 96/41884 PCT/IB96/00571 37 in 3 experiments). This makes a clear distinction between two different binding sites on rgpl20 with different specificities (Fig 8A). When similar experiments were carried out with a lower BSA concentration of 0.25 a similar complete loss of the CD4-binding site and only partial reduction of the other receptor-binding site activity was found, but now kinetics were considerably shorter, in the order of 15 minutes (3 experiments, not shown).

In order to verify if the antigenic structure of gpl20 was modified by iodination, treated and untreated gpl20 and two dilution of each were adsorbed on microtiter plates, starting with a 25 pg/ml concentration.

Gpl20 was then revealed with a polyclonal rabbit anti-gpl20 antiserum, followed by an anti-rabbit Ig-peroxidase. Both dilution curves were similar, but with a slight difference at the 50 inhibitory concentration. This suggest that the described iodination procedure did not substantially alter the antigenic structure of gpl20 (Fig. 8B).

6- BRIDGE FORMATION BY GP120 BETWEEN TWO

RECEPTORS

Since gpl20 appears to possess two different binding sites for CD4 and the other receptor respectively, inhibition of one receptor would not be sufficient to avoid binding of gpl20 to the other.

Complete inhibition of gpl20 binding would need the simultaneous presence of inhibitors for both receptors.

This was shown by inhibiting rgpl20 binding by blockade of CD4 with mAB F101.69 or by preincubation of gpl20 with 100 FCS and finally with both inhibitors together.

WO 96/41884 PCT/IB96/00571 38 A) ELISA (figure 9A) Living CEM cells were incubated with rgpl20 in BSA 3 and, after washing with rabbit anti-gpl20 serum, and then with anti-rabbit Ig-linked to peroxidase, revealed with OPD (open circles) or the cells were first blocked with mAB F101.69 before incubation with gpl20 in BSA 3 and then further treated as before (crosses) or was preincubated with 10 FCS before incubation with cells and further treated as before (open triangles) or gpl20 was preincubated with FCS and CEM cells were simultaneously blocked with mAB F101.69. Gp 120 was further revealed as before (black squares).

a:mAB 0.1 pg/ml, FCS: 25 b: mAB 1 pg/ml, c:mAB 10 pg/ml, FCS:100 B) FACS (figure 9B) Living CEM cells were incubated with rgpl20 in BSA 3 and after washing further incubated with rabbit anti-gpl20 serum and then with anti-rabbit Ig-linked to phycoerythrin before analysis with a FACsort (open circles) or cells were first blocked with mAB F 101.69 (10 pg/ml, 30 min, 37*C,) then incubated with gpl20 in BSA 3 and revealed as before (crosses), or gpl20 was preincubated with 10 FCS before incubation with CEM cells and revealed as before (open triangles) or cells were first blocked with mAB F101.69 and gpl20 preincubated with 10 FCS before it was incubated with the blocked CEM cells and further revealed as before (black squares).

d:mAB 0.1 pg/ml, FCS e:mAB 1 pg/ml, FCS f:mAB 10 pg/ml, FCS 100 In ELISA with living CEM cells maximal CD4 blockage alone reduced gpl20 binding to about 60 and preincubation of gpl20 with 100 FCS alone to about WO 96/41884 PCT/IB96/00571 39 When both inhibitions were effected simultaneously, gpl20 binding was almost completely inhibited (4 experiments). This suggested that a certain percentage of gpl20 might bind to both receptors, thereby forming a bridge (figure 9A). The same phenomenon was confirmed by FACS on living CEM cells. Now CD4 blockage with mAB F101.69 reduced gpl20 binding to 20 FCS to about 80 and both inhibitions together to about 7 of maximum gpl20 binding (figure 9B). Results with ELISA and FACS therefore indicate the same phenomenon.

Inhibition on the CD4 and on the other receptor are complementary, suggesting that part of gpl20 may form a bridge between both receptors. Additionally it was found that even total blockage of CD4 and presence of 100 serum did not avoid residual binding of gpl20 to living cells completely. This means that serum inhibition at physiological serum concentrations does not completely inhibit all interactions with said receptor.

7- GD 120 BINDING TO HUMAN ERYTHROCYTES Gp 120 binding to human erythrocytes was studied with an ELISA method A) Relation between gpl20 concentration and its binding.

Red blood cells (French Blood Transfusion Center) were washed in PBS (3x) and incubated in Maxisorb R microtiter plates (NUNC) at 4 x 105 cells well for adhesion. The wells were then saturated with BSA 3% and after washing, incubated with different doses of in PBS-BSA 3% for lh at 37'C. After 3 washings in PBS-BSA 0.3 gpl20 was detected with a rabbit anti-gpl20 antiserum, followed by anti-rabbit Ig coupled to alkaline phosphatase and revealed with pNPP.

The wells were evaluated at 414 nm.

WO 96/41884 PCT/IB96/00571 Figure 10A shows the OD variations as a function of increasing doses of Gp 120 bound to erythrocytes (optical densities 1.0) in a dose dependent manner and could be detected at concentrations superior to 0.25 pg/ml (6 experiments).

Binding was linear up to 4 pg/ml concentrations and then reached saturation levels.

Specificity was tested by omitting gpl20, or when it was present by replacing the anti-gpl20 antiserum by normal rabbit serum, resulting in marginal signals in ELISA, which were subtracted from the experimental values. Thus specific binding of gpl20 to erythrocytes was confirmed. Similar experiments were performed with freshly taken erythrocytes from different human healthy donors. The dose dependent gpl20 binding was similar, although at lower doses binding was slightly decreased with respect to conserved erythrocytes from the blood transfusion center (Montpellier).

In order to test gpl20 interactions with erythrocytes under more physiological conditions, serum was added during incubation of gpl20 with erythrocytes.

Erythrocytes were first adsorbed on microter plates and saturated as above-described in BSA 3 In parallel, gpl20 was pre-incubated with different dilutions of human serum or fetal calf serum for 1 h at 370C. The mixture was then incubated with the adsorbed erythrocytes. The washing and the revealing steps were identical to the above-disclosed phosphatase alkaline method.

Figure 10B gives the OD variation as a function of the concentration in human serum curve) and in fetal calf serum curve).

WO 96/41884 PCT/IB96/00571 41 In presence of 100 normal serum, binding to erythrocytes was considerably reduced to about This reduction was dependent on serum concentration and was effective between 1 and 100 serum (3 experiments). At lower concentrations inhibition was no longer significant. Human serum or FCS did not show significant differences. The presence of serum therefore presents an efficient in vivo protection of erythrocytes for gpl20 binding.

8- BINDING OF GP 120 TO THE TRANSMEMBRANE PROTEIN GLYCOPHORIN Since, in absence of serum, gpl20 could only bind to molecules accessible at the outer surface of erythrocytes, the major transmembrane proteins were tested. Several transmembrane proteins are known to be present in erythrocytes, but only two of them are quantitatively dominant, glycophorin and the anion exchanger band 3 protein. These proteins were individually examined for gpl20 binding.

Commercially obtained, purified glycophorin was bound to an anti-glycophorin monoclonal mouse antibody, which had been previously adsorbed at an optimal concentration of 20 pg/ml to MaxisorpR microtiter plates (NUNC, Denmark) and saturated with BSA 3 Recombinant gpl20 was incubated with the antibody adsorbed glycophorin and revealed with a rabbit antiserum followed by an anti-rabbit Ig linked to peroxidase and OPD.

Only low optical densities were obtained.

Although a slight binding of gpl20 to insolubilized glycophorin could be detected, it was too weak to explain all gpl20 binding to erythrocytes (figure 11).

WO 96/41884 PCT/IB96/00571 42 9- PURIFICATION OF THE TRANSMEMBRANE PROTEIN BAND 3 OF ERYTHROCYTES Band 3 protein, commercially not available, was purified from human erythrocytes according to the method This method allows an enrichment of band 3 protein to high purity in its original erythrocyte membrane by eliminating the other transmembrane or membrane associated proteins. Erythrocyte ghosts were incubated in low ionic strength buffers at alkaline pH, in order to extract membrane adherent proteins. The transmembrane protein glycophorin was eliminated by the weak neutral detergent, octyl-glycopyranoside at about 2 CMS. Other membrane associated proteins were finally extracted by a rapid step of pH 12 at low ionic forces. The resulting small erythrocyte vesicles were revealed with PBS and used for the studies with gpl20 and HIV-1. They contained about 3 mg/ml protein, as evaluated with the BCC method in presence of 1 SDS (2 preparations) In SDS-PAGE electrophoresis only one major band was found at 100 kd showing identical migration to a reference Some small contaminations of higher molecular weight about 120 kD, 170 kD and >220 could represent membrane associated proteins, such as ankyrin, spectrin or others. The contamination by glycophorin was assessed by ELISA and compared to a glycophorin standard curve. Microtiterplate adsorbed band 3 vesicles were incubated with the antiglycophorin mAb and revealed with an anti-mouse Ig coupled to peroxidase. The contamination of the band 3 vesicles by glycophorin was very low, in the order of 0.3%.

BINDING OF GP 120 BINDING TO BAND 3 PROTEIN VESICLES WO 96/41884 PCT/IB96/00571 43 Band 3 vesicles were adsorbed on microtiter plates at experimentally determined optimal concentrations of 0.3 mg/ml. Gpl20 was incubated in increasing concentrations with these insolubilized vesicles for lh at 37°C and binding detected with antirabbit antibodies followed by anti-rabbit IG coupled to peroxidase in the described way. Gpl20 bound in a dose dependent manner to band 3 vesicles. From pg/ml to about 4 pg/ml concentrations gpl20 binding was linear and the reached plateau levels. As shown by figure 12, binding to the band 3 vesicles curve) was much stronger than to glycophorin curve). With band 3, a gpl20 binding was achieved at about 16 times lower concentrations than with glycophorin, suggesting that gpl20 has higher affinity to band 3 protein.

11- ACCESSORY GP 120 BINDING SITES ON CD4 CEM

CELLS

Since band 3 like proteins are ubiquitous transmembrane proteins found in all cells and tissues tested, they could represent an accessory receptors for HIV, if their presence was also demonstrated on CD4+ cells.

Therefore it was first studied, if accessory binding receptors for gpl20 could be found on CD4+ CEM cells.

Intact CEM cells (105/ well) were insolubilized on anti-CD5 and anti-CD100 antibodies, which had been adsorbed to MaxisorpR microtiter plates and saturated with BSA 3 before. Their CD4 receptor were then blocked or not blocked at the gpl20 binding site by incubation with an oversaturating dose of mAb F101.69 known for its gpl20 binding inhibition. The blocking mAb WO 96/41884 PCT/IB96/00571 44 was maintained during incubation (lh at 370C, in BSA 3 and gpl20 binding was revealed, as described above, with a rabbit anti-gpl20 antibody and an anti-rabbit Ig coupled to peroxidase. Although the CD4 binding site was completely blocked, as tested by inhibition of mAb F101.69 binding (3 experiments), about 50 of bound to these cells, when no serum was present as shown by figure 13A.

In order to find out if these binding sites were inhibited by human serum in a similar way as the erythrocyte receptors, gpl20 was pre-incubated with different serum concentrations and then incubated with CD4 mAb on blocked or not blocked CEM cells.

Binding was followed with anti-gpl20 rabbit antibodies in the described way. On CD4 blocked cells possessing only accessory binding receptors, increasing serum concentrations decreased gpl20 to about 5 residual binding, thus demonstrating considerable serum inhibition. On cells with no CD4 blockade, containing simultaneously CD4 and accessory binding sites, serum inhibition was much less pronounced and reduced gpl20 to about 70 residual binding only. When serum inhibition was tested on the CD4 receptor alone by inhibition of mAb F101.69, gpl20 binding was only slightly reduced to 80 Therefore serum inhibited the CD4 receptor only slightly, but the accessory gpl20 binding sites almost completely, which clearly demonstrated a different nature of both binding sites (Fig. 13A). As above-mentioned said secondary sites are designated in the specification "serum inhibited receptors" or SIR.

12- COMPETITION OF GP 120 BINDING WITH BAND 3 PROTEINS TO ACCESSORY CEM BINDING SITES WO 96/41884 PCT/IB96/00571 Since band 3-like proteins are ubiquitous in almost all cell types and tissues, it was studied, if isolated band 3 vesicles from erythrocytes were able to inhibit the interaction of gpl20 with the SIRs of CEM cells.

Band 3 vesicles were pre-incubated in increasing concentrations from 2 1500 pg/ml with for lh at 370C and the mixture with CD4 blocked CEM cells, presenting only SIRs.

Binding of gpl20 was revealed by rabbit antibodies as described.

Prior contact of gpl20 with band 3 vesicles inhibited its binding to CEM cells in a dose dependent manner, starting with band 3 concentrations of about g/ml.

With high concentrations of about 1 mg/ml was reduced to less than 20 residual binding. The efficient inhibition of gpl20 binding to CEM cells by prior contact to band 3 protein suggests the presence of a crossreacting band 3 like protein on CEM cells (Fig.

14). Nevertheless pre-incubation with normal erythrocytes in concentrations 1000 times higher than the target CEM cells, did not inhibit gpl20 binding to SIRs of CEM cells, although they contained band 3 protein in considerable amounts.

13- ABSORPTION OF GP 120 AFTER PROTEOLYTIC TREATMENT OF ERYTHROCYTES In order to find out, if band 3 protein was not directly accessible to gpl20, it was tried to pretreat erythrocytes with potent proteolytic enzymes.

Erythrocytes were treated with increasing Pronase R concentrations from 10 1000 pg/ml for 1 h at WO 96/41884 PCT/IB96/00571 46 370C and, after extensive washing and addition of enzyme inhibitors, incubated with gpl20. After proteolytic digestion, erythrocytes were able to absorb depending on the PronaseR concentration used. With high Pronase R concentrations, up to 1 mg/ml, about 85 of could be absorbed by the pretreated erythrocytes.

This suggests that the access of gpl20 to the transmembrane protein band 3 on intact erythrocytes is difficult and requires first an intervention on the glycocalix of erythrocyte membrane (Fig. 13B).

INHIBITION OF VIRUS INFECTION IN VITRO BY BAND 3 VESICLES The influence of prior contact of intact virus with band 3 vesicles was tested.

The diameter of band 3 vesicles was similar to the one of virus particles about 0.1 pm., so that virus could be adsorbed to these particles and eliminated by centrifugation.

Virus was first adsorbed to band 3 vesicles and elimination by centrifugation before testing the supernatant for residual virus activity.

HIV-1 LAI virus was pre-incubated in BSA 3 for lh at 37°C in doses of 100TCID 50 with increasing concentrations of band 3 vesicles, centrifuged (15 min at 15000 g) and then incubated in BSA 3 with uninfected CEM cells for another hour at 370C. Control virus was pre-incubated in BSA 3 under the same conditions. After washing cells were cultured for 14 days in normal culture medium and the presence of RT in the supernatant was measured at days 4 and 8 after infection. Adsorption of whole virus to band 3 vesicles diminished the amounts of virus within the supernatant and subsequently the virus WO 96/41884 PCT/IB96/00571 replication by about 80 at the day 4 after infection of CEM cells (Fig. WO 96/41884 PCT/IB96/00571 48

REFERENCES

1) KLATZMANN, CHAMPAGNE, CHAMARET,

GRUEST,

GUETARD, HERGENT, T, GLUCKMAN, J.C. MONTAGNIER, (1984). T-lymphocyte T4 molecule behaves as the receptor for human retrovirus LAV. Nature 312, 767-768, 2) DALGLEISH, BEVERLEY, P.C.L. CLAPHAM, P.R., CRAWFORD, GREAVES, M.F. WEISS, R.A. (1984). The CD4 (T4) antigen is an essential component of the receptor for the AIDS retrovirus. Nature 312, 763-767.

3) LASKY, NAKAMURA, SMITH FENNIE, C., SHIMASAKI, PATZER, BERMAN, GREGORY T. CAPON, D. (1987). Delineation of a region of the human immunodeficiency virus type lgpl20 glycoprotein critical for interaction with the CD4 receptor. Cell, 50, 975-985.

4) MOORE WALLACE FOLLET E.A.C. McKEATING J.A. (1989). An enzyme-linked immunosorbent assay for antibodies to the envelope glycoproteins of divergent strains of HIV-1. AIDS, 3, 155-163.

5) GILBERT BRIGIDO MULLER HANSEN J.E., EZEKOWITZ R.A. MILLS J. (1993). Screening for inhibitors of HIV gpl20-CD4 binding using an enzymelinked immunoabsorbant assay. Journal of Virological Methods. 42, 1-12.

WO 96/41884 PTI9/07 PCT/IB96/00571 49 6) MONTEFIORI, STEWART, K, AHEARN, ZHOU, J. ZHOU J. Complement-mediated binding of naturally glycosylated and glycosylation-modified human immunodeficiency virus type 1 to human CR2 (CD21).

(1993). Journal of Virology, 2699-2706.

7) HAROUSE, BHAT, SPITALNIK, LAUGHLIN, M.

STEFANO, SILBERBERG D.H. GONZALE Z-SCARANO, F.

(1991). Inhibition of entry of HIC-1 in neural cell lines by antibodies against galactosyl ceramide. Science, 253:320-323.

8) MA, GEIGER, J.G. MATH, A (1994).

Characterization of a novel binding site for the human imnmunodeficiency virus type 1 envelope protein gpl2O on human fetal astrocytes. Journal of Virology, 69, 6824- 6828.

9) MOSES, BLOOM, PAUZA, D.D. NELSON, J.A.

(1993). Human imnmunodeficiency virus infection of human brain capillary endothelial cells occurs via a CD4 /galactosylceramide independent mechanism. Proceedings of the National Academy of Science 90, 10474-10478.

10) DEZUTTER-DANBUYANT, SCHMITT, DUSSERRE, N., HANAU, KOLBE KIENY CAZENAVE J.P., SCHMITT, PASQUALI, OLIVIER GAZZOLO, THIVOLET, (1991). Interactions of human epidermnal Langemans cells with HIV-1 viral envelope proteins (GP12O and GP1-160s) involves a receptor mediated endocytosis WO 96/41884 PCT/IB96/00571 independent of CD4 T4 epitope. The Journal of Dermatology 18, 377-392.

11) MOORE, J.P. (1990). Simple methods for monitoring HIV-1 and HIV-2 gpl20 binding to soluble CD4 by enzymelinked immunoabsorbent assay: HIV-2 has a 25 fold lower affinity than HIV-1 for soluble CD4, AIDS, 4, 297-305.

12) GILBERT BRIGIDO MOLLER HANSEN J.E., EZEKOWITZ R.A. MILLS J. (1993). Screening for inhibitors of HIV gpl20-CD4 binding using an enzymelinked immunoabsorbant assay. Journal of Virological Methods. 42, 1-12.

13) Tanner M.J.A. Martin P.G. High S. 1988. The complete amino acid sequence of the human erythrocyte membrane anion-transport protein deduced from the cDNA sequence, Biochemical Journal 256: 703-712.

14) Wang D.N. Sarabia V.E. Reithmeier R.A.F. Kuhlbrandt, W. 1990. Three-dimensional map of the dimeric membrane domain of the human erythrocyte anion exchanger, Band 3.

EMBO J. 13:3230-3235.

15) Jay et al., 1986, Amm. Rev. Biochem. 55:SM-538 16) Demuth, D.R. Showe Ballantine, Palumbo, Fraser, P.J. Cioe, Rovera, Curtis P.J. 1986, Cloning and structural characterization of a human non erythroid band 3 like protein. EMBO 5:1205-1214.

WO 96/41884 PCT/IB96/00571 51 17) Wolosin, J.M. Ginsburg Cabantchik Z.I. 1977.

Functional characterization of anion transport system isolated from human erythrocyte membranes. J. Biol.

Chem., 252, 2419-2427.

18) McMICHAEL BEVERLY P.C.L, COBOLD CRUMPTON GILKS GOTCH F.M. eds. (1987). Leucocyte Typing III. White cell differentiation antigens. Oxford, Oxford University Press, pp 991-992.

19) SATTENTAU, ARTHOS, DEEN, HANNA, N., HEALEY, BEVERLEY, SWEET, R. TRUNEH, A., (1989). Structural analysis of the human immunodeficiency virus binding domain of CD4. Epitope mapping with sitedirected matants and anti-idiotypes. Journal of Experimental Medicine 170, 1319-1334.

NAKANE, P.K. KAWAOI; A. (1974). Peroxidase labeled antibody. A new method of conjugation. Journal of Histochemistry and Cytochemistry, 22: 1084-1091.

21) CARRIERE, FONTAINE, BERTHIER, A.M., ROUQUETTE, CARAYON, LAPRADE, JUILLARD, R., JANSEN, PAOLI, PAOLUCCI, GROS, P. PAU, B.

(1994). Two-site immunoassay of CD4 and CD8 molecules on the surface of T lymphocytes from healthy subjects and HIV-1 infected patients. Clinical Chemistry 40, 30-37.

WO 96/41884 PCT/IB96/00571 52 22) WALKER, LLULL, BALKIAN, KO, FLORES, RAMSAMMOJ, BLACK, HEWITT, C.W. MARTIN, D.C. (1992). A rapid and sensitive cellular enzyme-linked immunoabsorbant assay (CELISA) for the detection and quantitation of antibodies against cell surface determinants. I.A. comparison of cell fixation and storage techniques. Journal of Immunological Methods 153, 121-130.

Claims (13)

1. Use of a protein of the multi-passage transmembrane protein families as gp120 receptor, wherein said protein is a band 3-like protein.

2. Use according to claim 1, wherein said protein is band 3 protein such as present on human erythrocytes.

3. Use according to claim 1, wherein said protein is band 3 protein such as present on lymphocyte cells.

4. Use according to claim 3, wherein said lymphocyte cells are CEM cells.

Use according to anyone of claims 1 to 4 wherein said protein is capable of interacting with gp120, at 37°C, in the absence of human serum.

6. Use according to anyone of claims 1 to 4 wherein said protein is such as present on the surface of HIV-infectable cells, such a CD4 CEM cells.

7. Polyclonal or monoclonal antibodies directed against the gp120 receptor region of the protein used in anyone of claims 1 to 6, and the anti-idiotype antibodies directed against said antibodies. i

8. Means for detecting the presence of HIV or the immune response to HIV, including using the gp120 receptor or binding site as used according to anyone of claims 1 to 6 or the antibodies according to claim 7.

9. A method for detecting the presence of HIV or the immune response, including: contacting the sample to be analyzed with a protein used as a gp120 receptor according to anyone of claims 1 to 6, capable to bind to HIV, where the inhibition of the fixation of the viral proteins on said receptor will indicate the presence of antibodies capable to hinder the virus fixation on said receptor.

A kit for detecting the presence of HIV, or the immune response to HIV, including at least one protein as used according to anyone of claims 1 to 6, or at least one antibody according to claim 7, in a sufficient amount for performing the test, as well as appropriate reagents and buffers and instructions for use.

11. A kit according to claim 10 wherein the protein or antibody is fixed on a microplate.

12. Pharmaceutical compositions containing an efficient amount of the protein as used according to anyone of claims 1 to 6 or the antibodies according to claim 7 in association with a suitable carrier.

13. Prophylactic compositions, characterized by the fact that they are elaborated from the protein as used according to anyone of claims 1 to 6 or the antibodies according to claim 7. 9* DATED this 4th day of June 1999 INSTITUT FRANCAIS DE RECHERCHE POUR LE DEVELOPPEMENT EN COOPERATION CORSTOM WATERMARK PATENT TRADEMARK ATTORNEYS 290 BURWOOD ROAD HAWTHORN VICTORIA 3122 AUSTRALIA LCG/KMH/MEH DOC 27 AU5776896.WPC