HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Jacobin, M.-J. Articles by Clofent-Sanchez, G. Search for Related Content PubMed PubMed Citation Articles by Jacobin, M.-J. Articles by Clofent-Sanchez, G.

Human IgG Monoclonal Anti-_IIb3-Binding Fragments Derived from Immunized Donors Using Phage Display¹

Marie-Josée Jacobin^*, Jeanny Laroche-Traineau^*, Melvyn Little^2,, Armin Keller^3,, Karlheinz Peter, Martin Welschof^3,, Alan Nurden^* and Gisèle Clofent-Sanchez^4,*

^* Centre National de la Recherche Scientifique, Unité Mixte de Recherche 5533, Hopital Cardiologique, Pessac, France; German Cancer Research Center, Recombinant Ab Group, Heidelberg, Germany; and Department of Cardiology, University of Freiburg, Freiburg, Germany

	Abstract

Top Abstract Introduction MATERIALS AND METHODS Results Discussion References

 
Previous studies of the immune response in polytransfused Glanzmann thrombasthenia (GT) patients and in autoimmune thrombocytopenic purpura (AITP) have relied on serum analysis and have shown the frequent development of Abs directed against the _IIb3 integrin. However, little is known about the molecular diversity of the humoral immune response to _IIb3 due to the paucity of mAbs issuing from these pathologies. We have isolated human IgG anti-_IIb3 binding fragments using combinatorial libraries of single-chain IgG created from the B cells of a GT and an AITP patient, both with serum Abs. Ab screening was performed using activated platelets or activated _IIb3-expressing Chinese hamster ovary cells. Sequencing of selected phage Abs showed that a broad selection of genes from virtually all V gene families had been used, indicating the diversity of the immune response. About one-half of the V_H and V_L segments of our IgG anti-_IIb3 fragments displayed extensive hypermutations in the complementarity-determining region, supporting the idea that an Ag-driven immune response was occurring in both patients. The H chain complementarity-determining region 3 analysis of phage Abs revealed motifs other than the well-known RGD and KQAGDV integrin-binding sequences. To our knowledge, our study is the first to illustrate multiple human IgG anti-_IIb3 reactivities and structural variations linked to the anti-platelet human immune response. Human _IIb3 Abs preferentially directed against the activated form of the integrin were further characterized because platelet _IIb3 inhibitors are potential therapeutic reagents for treating acute coronary syndromes. Currently available _IIb3 antagonists do not specifically recognize the activated form of the integrin.

	Introduction

Top Abstract Introduction MATERIALS AND METHODS Results Discussion References

 
The human platelet membrane glycoprotein IIb-IIIa complex is a member of the ₃ subclass of integrin receptors (_IIb3). It mediates platelet aggregation by binding multivalent adhesive proteins, which then form bridges linking adjacent platelets (1). Anti-_IIb3 serum Abs have been detected in a majority of patients with either the chronic or acute forms of autoimmune thrombocytopenic purpura (AITP)⁵ and in patients with Glanzmann thrombasthenia (GT), the inherited disorder of _IIb3, after receiving blood transfusions to stop bleeding (2, 3, 4). Although _IIb3 acts as an Ag in these pathologies, relatively little is known about the structural and genetic features, V gene usage, and degree of somatic mutation of the anti-_IIb3 Abs. Efforts to study the genetic structure of the disease-relevant mAbs have been impeded in part by the difficulty of obtaining monoclonal IgG human anti-_IIb3 Abs from patients by conventional cell immortalization methods. Thus, in a rare report, Olee et al. (5) described a monoclonal IgG anti-_IIb3 generated after EBV transformation of secreting B cells from an AITP patient. The production of human mAbs has been more recently addressed by genetic engineering (6). Ishida et al. (7) isolated, by phage display, an Ab specific for _IIb3 from the B cells of a polytransfused GT patient. Recombinant Abs against the human platelet alloantigen-1a determinant carried by ₃ have been produced from patients with post-transfusion purpura or neonatal alloimmune thrombocytopenia (8, 9, 10).

We have adapted current technology to produce recombinant Abs using, as starting material, combinatorial libraries from B cells of a polytransfused patient with GT and another with AITP, both patients possessing serum IgG Abs against _IIb3 (4, 11). Combinatorial libraries from immune donors have been often used to determine the nature of the humoral immune response, for example, in patients with autoimmune or neoplastic diseases or with viral infection (12, 13, 14, 15, 16, 17, 18, 19). Nucleotide sequences of variable regions of such Abs may help us to understand the origin of the secreting B cells. Natural autoantibodies are present in the sera of all individuals, and early studies gave support to the idea that they are mostly encoded by germline genes (20, 21, 22). In contrast, autoantibodies encountered in pathology display, in most instances, somatic mutations, as they will have been affinity matured by the immune system (23, 24). Knowledge of the genetic status of anti-_IIb3 Abs in terms of the number of accumulated mutations could help advance our understanding of the pathogenesis of Ab development. Moreover, questions have been raised concerning the oligo or polyclonal nature of the immunological response developed in GT patients after transfusion or during AITP. Up to now, rarely more than one mAb has been obtained from the B cells of immunized anti-_IIb3 developing donors, and Ishida et al. (7) demonstrated a restricted usage of the V_H4 gene family. The variable gene family usage and structure of the H chain complementarity-determining region 3 (HCDR3) motifs of the anti-_IIb3 Abs obtained from our patients is now reported. Our results confirm the molecular diversity of the immunological response in hematological disorders.

The _IIb3 integrin is creating interest because of the role of platelets in acute coronary syndromes (25, 26). In this context, _IIb3 antagonists represent a new class of antithrombotic drugs that block the common end point of the platelet activation pathways that lead to aggregation. Agents that have been approved for clinical use, including a chimeric anti-_IIb3 Fab (c7E3) and synthetic peptide and nonpeptide antagonists modeled on the RGD motif (27), bind to the nonactivated form of the integrin and some may stimulate a transition of _IIb3 from a resting to a ligand competent state (28). These properties have been implicated in the thrombocytopenia occasionally reported after administration of such antagonists in man (29, 30). Thus, activation-dependent human mAbs would be interesting to develop for therapeutic use. Therefore, our study was expanded to screen for phages recognizing active sites on _IIb3. These were selected using activated platelets and Chinese hamster ovary (CHO) cells expressing _IIb3 locked in a high affinity state (28). Our findings show that patients who have developed an immune response to this integrin may, in fact, have the potential to generate Abs against a whole range of epitopes on _IIb3.

	MATERIALS AND METHODS

Top Abstract Introduction MATERIALS AND METHODS Results Discussion References

 
Patients

Patient 1 was an elderly man whose life-long history of bleeding and abnormal platelet function tests were consistent with a clinical diagnosis of GT (4). His platelets were severely deficient in _IIb3 and fibrinogen, findings typical of type I disease (31). He was transfused on several occasions to arrest bleeding episodes. Preliminary characterization showed that his serum contained a high-titer Ab that reacted with ₃ in Western blotting performed against normal platelet proteins. Studies were performed after informed consent and according to the declaration of Helsinki, Finland. Blood cells were used in this study; the resulting phage display library was named EB.

Patient 2 was a 54-year-old man who was diagnosed with autoimmune Hashimodo thyroiditis some 20 years ago. The symptoms abated with medication (levothyrox, 200 µg/day). In 1993, he was found to have excessive bruising and mucosal bleeding in his mouth. Testing for serum anti-platelet Abs (IgM and IgG) by flow cytometry was positive (11). mAb-specific immobilization of platelet Ag testing (3) for anti-_IIb3 was also positive, and he was diagnosed as having chronic AITP. He was treated with i.v. Igs (30 g/day). His platelet count rose to 55 x 10³/µl. He underwent splenectomy in 1996, but his platelet count remained low. Cells used in the present study were obtained from the spleen. Studies were performed after informed consent and according to the declaration of Helsinki, Finland. The resulting phage display library was named TE.

RNA isolation and creation of combinatorial scFv libraries

PBL from patient EB and spleen tissue from patient TE were obtained as a source of mRNA for the generation of IgG libraries constructed largely according to published protocols (32, 33). Briefly, mRNA was extracted using Dynabeads oligo(dT)₂₅ (Dynal Biotech, Compiegne, France) from total RNA purified with Trizol LS Reagent (Life Technologies, Cergy Pontoise, France) from 5 x 10⁷ cells. To generate first-strand cDNA, three C-region primers, kindly provided by Dr. D. Nugent (Orange County, CA), specific for anti-sense sequences of the C1, C, and C genes were used to amplify Ig-specific mRNA (7). The V domains of the -, -, and -chains were amplified by RT-PCR using primers previously described by us (33) and Klentaq polymerase (Sigma-Aldrich, St. Quentin, France). Reactions were conducted using a Gene-Amp 9600 PCR system (PerkinElmer, Foster City, CA) and a denaturing cycle of 3 min at 95°C, 30 cycles of 30 s at 95°C, and 3 min at 68°C, completed with a final cycle of 5 min at 68°C. For cloning into pSEX81, the V_H and V_L genes were reamplified with homologous primers containing restriction endonuclease sites NcoI (5') and HindIII (3') for V_H and MluI (5') and NotI (3') for V_L using, after a denaturing cycle, 15 cycles of 30 s at 95°C, 1 min at 57°C, and 1 min at 75°C, completed with 5 min at 75°C. The amplified products were purified on a 1.5% agarose gel. Bands of 400 bp were excised from the gel, and DNA was extracted and purified with the QIAquick gel extraction kit (Qiagen, Courtaboeuf, France).

Purified PCR products were double digested with MluI/NotI and NcoI/HindIII (Roche, Meylan, France) for the cloning of the L and H chains into pSEX81. After purification with Qiaquick PCR purification kit (Qiagen), the L chain genes were ligated in the pSEX81 phage display vector. L chain repertoires were then electroporated into Escherichia coli electrocompetent XL-1 blue cells (Stratagene, La Jolla, CA) using a Gene Pulser (Bio-Rad, Ivry sur Seine, France). The - and L chain sublibraries were prepared by Qiaquick plasmid extraction (Qiagen). The combinatorial phagemids were constructed by ligating H chain genes with - and L chain sublibraries. The final constructs were electroporated in E. coli XL-1 blue. The transformed bacteria were stored at -80°C. Phage libraries displaying recombinant scFv were then rescued by infection with M13KO7 helper phage (Amersham, Saclay, France) in 2TY_AK (15 g/L bacto-agar, 17 g/L bacto-tryptone, 10 g/L bacto-yeast extract, 0.1 M NaCl, 50 µg/ml ampicillin, 50 µg/ml kanamycin) medium as described (34).

Library affinity selection on activated coated platelets

Washed platelets were prepared from acid-citrate-dextrose-anticoagulated blood according to described procedures (3). Samples (10⁸ platelets/ml) were activated with 0.5 U/ml human -thrombin (Fibrindex; Ortho Diagnostics, Raritan, NJ). After light fixation with 0.025% paraformaldehyde, activated platelets were coated overnight at 4°C in 100 µl of 0.1 M of sodium bicarbonate (pH 9) at a concentration of 10⁷ platelets/well on Costar 96-well plates (Polylabo, Strasbourg, France). The wells were blocked with 200 µl of 2% dry fat milk in PBS. After a 2-h incubation, 100 µl of phage library (typically 10¹² CFU) were added to each well and were incubated for 1 h at room temperature. A total of 16 individual wells were used for the screening of each library. Unbound phages were removed through repeated washing first with 0.1% (v/v) Tween 20 in PBS and then by PBS. Bound phages were eluted with 0.1 M of triethylamine for 5 min at room temperature, neutralized with 2 M of Tris-base, and used to infect 20 ml of exponentially growing E. coli XL-1 blue (A₆₀₀ of 0.4). The next day, bacteria were scraped with 2TY medium supplemented with 50 µg/ml ampicillin and 100 mM of glucose to rescue the amplified eluted phages for a new round of panning. Two rounds of panning were conducted. Small-scale phage rescues were effected from the isolated colonies and transferred to a sterile 96-deep-well master plate in separated wells. Phages from single-infected colonies were stored at 4°C until required for phage-ELISA. The positive clones were analyzed by restriction site digestions and by DNA sequencing.

Library affinity selection on activated platelets and CHO cells in suspension

A volume of 5 x 10⁷ thrombin-activated platelets was mixed with 2.5 x 10¹¹ CFU and the volume was made up to 1 ml with 2% milk-PBS. Binding was allowed to proceed for 1 h at room temperature with end-over-end rotation (35). The unbound phages were removed by five cycles of centrifugation (5000 x g) and washings with PBS. The resulting platelet/phage pellet was incubated in 0.5 ml of 6 M urea (pH 3) for 15 min at room temperature followed by neutralization with 10 µl of 2 M Tris base. The eluted phages were titered and amplified using E. coli XL-1 blue before being used for subsequent rounds of selection. For the second round, bound phages were competitively eluted using purified _IIb3 (Enzyme Research Laboratories, Swansea, U.K.) at a 10-fold higher concentration than present on platelets. A third affinity selection was performed on _IIb3-expressing CHO cells (modified protocol of Goodson et al.; Ref. 36). To preferentially select phagemid clones recognizing epitopes exclusive to the activated form of _IIb3, 10¹² CFU from round two were preadsorbed for 1 h with 5 x 10⁶ CHO cells expressing the nonactivated (low-affinity) wild-type (WT) _IIb3 (WT-CHO). The cells were sedimented at 4°C, and the supernatant was incubated with 10⁶ CHO cells displaying activated (high-affinity), GFFKR-deleted _IIb3 (Del-CHO). The cells were washed five times with IMDM. The eluted phages (see above) were amplified and titered on E. coli XL-1 blue, and phages from single-infected colonies were again recovered on a master plate.

Expression and purification of soluble scFv

For cloning into the vector pHOG21 (37), the scFv genes of our selected anti-_IIb3 phage Abs were digested with NcoI and NotI, separated on an agarose gel, and purified. These products were then ligated to NcoI/NotI-digested pHOG21. Transformed XL1-blue E. coli bacteria were grown overnight in 2TY medium supplemented with 50 µg/ml ampicillin and 100 mM glucose at 37°C. Dilutions (1/60) of the overnight cultures were grown as 50-ml cultures in shake-flasks at 37°C with shaking at 280 rpm. When the culture reached an OD₆₀₀ of 0.7, bacteria were pelleted at 1500 x g for 10 min and were resuspended in the same volume of fresh 2TY medium containing 50 µg/ml ampicillin and 0.4 M sucrose. ScFv expression was induced by the addition of 1 mM isopropyl--D-thiogalactopyranoside. The culture was allowed to grow for 7–8 h at 25°C. Bacteria were harvested by centrifugation at 5000 x g for 10 min, and the pellet was stored at -80°C. In the pHOG21 vector, the scFv fragment is immediately followed by a c-myc tag recognized by the mAb 9E10 and by six C-terminal histidine residues. The 6-His-tagged scFv were purified by immobilized metal affinity chromatography on Ni-NTA spin columns (Qiagen). To isolate soluble scFv fragments in native conditions, bacteria were resuspended in lysis buffer (50 mM NaH₂PO₄ (pH 8), 300 mM NaCl, and 10 mM imidazole) to which was added 1 mg/ml lysozyme (Sigma-Aldrich). After 30 min on ice and sonication, the lysate was centrifuged at 10,000 x g for 30 min at 4°C and the supernatant was loaded onto a Ni-NTA spin column. ScFv fragments were isolated under the conditions recommended by the supplier. Eluted material was dialyzed against PBS and was used for functional analysis by ELISA.

Analysis of Ab reactivity by ELISA

The wells of 96-well microtiter plates were coated with 5 µg/ml purified _IIb3. Control wells were coated with 3.5 µg/ml BSA, 5 µg/ml myosin, or 5 µg/ml actin (all from Sigma-Aldrich). Thrombin-activated washed platelets were coated overnight at 4°C at 10⁷ platelets/well. Del-CHO and WT-CHO cells were seeded and grown as monolayers in 96-well microtiter plates. Each well was blocked with 200 µl of 2% milk-PBS for 2 h at 37°C. Rescued phages (10⁸–10⁹ CFU) were then incubated overnight at 4°C. Nonspecific binding of helper phages was evaluated by adding to control wells the same quantity of M13KO7. An anti-Phox phage Ab that did not recognize _IIb3 was also included in the test as a negative control. The wells were then treated with 100 µl of 10 µg/ml murine anti-M13 mAb (Amersham). Two murine mAbs were used as positive controls for _IIb3 binding: Y2/51, an anti-CD61 mAb (DAKO, Trappes, France) and PAC-1, an IgM mAb that recognizes _IIb3 and mimics the binding characteristics of fibrinogen (38). After incubation for 2 h at room temperature, 100 µl of a 1/1000 dilution of HRP-conjugated anti-mouse IgG (Immunotech, Marseille, France) was added and incubated for 90 min at room temperature. Color was developed with 100 µl of ABTS (Sigma-Aldrich), and the absorbance was read at 405 nm using an Emax precision microplate reader (Molecular Devices, Sunnyvale, CA).

Soluble scFv fragments isolated by immobilized metal affinity chromatography on Ni-NTA spin columns were tested on purified _IIb3, human platelets, and BSA using the anti-c-myc mAb 9E10 and HRP-conjugated anti-mouse IgG.

SDS-PAGE and immunoblot analysis

SDS-PAGE was conducted under reducing conditions, and immunoblot analysis using anti-c-myc mAb 9E10 was performed as previously described (37, 39).

Determination of specificity and affinity of TEG4 and EBB3 scFv

The functional affinity of the scFv fragments was determined using a soluble Ag competition ELISA with each scFv used at a dilution that gave 50% maximal binding in PBS-dry fat milk 2% (w/v). Binding of scFv to solid-phase _IIb3 was evaluated colorimetrically using the mAb 9E10 (anti-c-myc; see above). Diluted scFv fragments were first preincubated for 18 h at 4°C with between 2.5 and 3000 nM of soluble purified _IIb3 (a0) before they were added to wells precoated with 5 µg/ml _IIb3. Thus, the amount of unbound scFv in the mixture was evaluated. Residual anti-_IIb3 activity (A) was measured relative to a control sample that was incubated in an equal volume of buffer in the absence of soluble _IIb3 (A0).

Capacity of fibrinogen to inhibit the interaction _IIb3-soluble scFv

The ability of fibrinogen to inhibit the interaction between _IIb3 and EBB3 or TEG4 scFv was evaluated at a dilution that gave 50% maximal binding to solid-phase _IIb3 in 2% milk-PBS (1/2 dilution) in a competition ELISA. The scFv were mixed with increasing concentrations of fibrinogen (9.3 x 10^-10–1.4 x 10^-5 M) at 37°C. The mixture was then added to ELISA wells precoated with 5 µg/ml _IIb3, and binding of the scFv fragments on _IIb3 was measured using anti-c-myc Ab. Equivalent dilutions of the scFv fragments without inhibitor (fibrinogen) served as the 100% binding control.

The same incubations were performed using 5 x 10^-9–3 x 10^-4 M actin (Sigma-Aldrich) as a substrate control.

Fingerprinting

To fingerprint isolated clones, two restriction enzymes, BstNI and RsaI (Roche), were used. Two micrograms of each phagemid was digested with 10 U/µg BstNI for 2 h at 60°C or with 15 U/µg RsaI for 1 h at 37°C. Digestion products were then analyzed by electrophoresis in 3% (w/v) Nusieve agarose-gels (Life Technologies).

Restriction analysis of rearranged H and L chains

One microgram of the ELISA-positive clones were digested with restriction enzymes NcoI/NotI, MluI/NotI, and NcoI/HindIII (10 U/µg) for 1 h at 37°C. Digestion products were then analyzed by electrophoresis in 1.5% agarose gels.

Determination and analysis of nucleotide and amino acid sequences of selected clones

Selected clones containing both functionally rearranged H and L chains were purified with Midiprep (Qiagen). The sequence reaction was performed with the sequenase sequencing kit (Amersham) and infrared fluorophore (infrared dye 800)-labeled primers using a DNA sequencer (LI-COR DNA sequencer 4000; Sciencetec, Les Ulis, France). H and L chains were sequenced on both strands a minimum of three times using two different primers. The infared dye-labeled sequencing primers used to determine the H chain sequence were HOG5 (5'-ATTAAAGAGGAGAAATTAACCA-3') hybridizing to the pSEX81-positive strand, and YOL3 (5'-CGCGTGCTTCTG-3') hybridizing to the negative strand of the pSEX81 linker. The primers used for the L chain were YOL2 (5'-GCTTGAAGAAGG-3') and SEQIV (5'-GA(AG)TTTTCTGTATGGGAT-3'), hybridizing to the positive strand of the pSEX81 linker and the pSEX81-negative strand, respectively. Automatic sequence analysis was done by computer program Base ImagIR Image Analysis Version 4. The programs DNAsis V2.1 and AlignIR V1.1 were used to further analyze all sequence data. The Ig V region sequences of the H and L genes were submitted to the DNA-Plot program for IMGT (40) and to FASTA, BLAST-n, and WU-BLASTX+BEAUTY programs for GenBank/EMBL.

The probability that the excess of replacement (R) mutations in the complementarity-determining region (CDR) or framework region (FR) (with respect to the closest germline genes) arose by chance was calculated according to the binomial distribution model (41, 42): p = [n!/k! (n - k)!] x q^k x (1 - q)^{n- k}. In this equation, n is the total number of observed mutations, q is the probability that an R mutation will localize to CDR or FR (q = Rf x CDRf or FRf where Rf is the expected proportion of R mutations (0.75), and CDRf or FRf is the relative size of the CDR or FR), and k is the number of observed R mutations in the CDR or FR. Amino acid residues occurring as a result of primer sequence in the FR1 region were excluded from the analysis. A p value of <0.05 indicated that the R mutations had occurred in a nonrandom fashion.

	Results

Top Abstract Introduction MATERIALS AND METHODS Results Discussion References

 
Construction of the pSEX81 libraries

Separate scFv libraries were prepared from the lymphocytes (EB) or splenocytes (TE) of donors previously shown to possess serum Abs to integrin _IIb3. After cloning into the phagemid display vector pSEX81 and electroporation into E. coli, sublibraries for the - and -chains were obtained for each patient. The V_H DNA repertoire was then cloned in a second step into these recombinant and phagemid vectors, followed by electroporation in E. coli. The size of the individual sublibraries and libraries were determined after serial dilutions followed by plating (Table I). Restriction digestion of the phagemid vector extracted from the resultant EB and TE libraries with NcoI/HindIII (V_H digestion), MluI/NotI (V_L digestion), and NcoI/NotI (scFv digestion) indicated that the libraries contained the expected 400-bp inserts for V_H and V_L and 800-bp inserts for the scFv. For each library, 20 clones were selected at random and were digested with the restriction enzymes BstNI and RsaI (i.e., fingerprinting). All the clones checked for diversity were different (data not shown).

View this table: [in this window] [in a new window]   Table I. Number of independent transformants in each and sublibrary and library

Isolation of _IIb3-specific clones

Two types of selection procedures were used for both IgG-derived scFv libraries displayed on phagemid particles. The first consisted of incubating the phages with activated platelets immobilized within the wells of a microtiter plate. After the second round of panning, an 10-fold increase (from 3.1 x 10³ to 2.8 x 10⁴ CFU) in the titer of phage binding was seen for EB. A similar amplification was not seen for TE. A total of 96 individual E. coli clones isolated from the second round of panning of the EB library was rescued for phage expression in a deep-well master plate. Phages derived from these 96 single-infected colonies were checked for specificity for _IIb3 by performing phage-ELISAs. Purified integrin was used as Ag and BSA was used as the control. A large number of colonies displayed specificity for _IIb3, as illustrated in Fig. 1.

View larger version (46K): [in this window] [in a new window]   FIGURE 1. ELISA in which phage Ab supernatants from the EB library after the second round of panning on coated platelets were tested against purified IIb3. Clones issuing from the master plate are numbered from A1 to H12. The same ELISA experiment using plates coated with BSA gave OD values lower than 0.1.

Clones producing scFv fragments giving high OD values in the ELISA assay were digested with the restriction enzymes NcoI and NotI. In this way, we ensured that functional restriction sites were still present in the construction for recloning into the pHOG21 vector for the production of soluble scFv. Further restriction analyses of their isolated DNA with NcoI, NotI, MluI, and HindIII revealed that some contained truncated H and/or L chain inserts. Sequencing revealed that they lacked part of the H or L chain V domains. These clones were not studied further.

Specificities against _IIb3 in an activated and a resting form

Ten selected phage Abs were tested by ELISA for their capacity to bind to a panel of unrelated Ags, including BSA, actin, and myosin. They all failed to react with these proteins (OD < 0.05). Fig. 2 shows the binding characteristics of the phage Abs against _IIb3 expressed in different conditions. A large majority of the phage Abs recognized the integrin independently of the source of _IIb3 used. However, differences in the binding intensity sometimes occurred, as shown by TEG4 and EBE12 phage Abs, which showed greater reactivity with coated activated platelets than to _IIb3-expressing CHO cells. For TEH11 and TEF2, the contrary is observed with a better recognition of the Del-CHO cell line. To further characterize the binding of our phage Abs, they were compared with two murine mAbs that bind to the _IIb3 complex. One of them, Y2/51, equally recognizes ₃ in both the activated and unactivated conformations of the integrin. The other, PAC-1, is an IgM mAb that recognizes the activated conformation of the complex. Fig. 2 demonstrates the binding of PAC-1 to thrombin-activated platelets and to the Del-CHO cell line with little recognition of the WT-CHO cell line. These results clearly underline a preferential recognition of the activated form of _IIb3 expressed on the CHO cell line for clones like TEH11, TEF2, EBB3, and EBB10.

View larger version (46K): [in this window] [in a new window]   FIGURE 2. Binding of individual phage clones after two rounds of panning against coated platelets (CP) and three rounds of panning against CHO cells in suspension (SC)-expressing IIb3. Individual pSEX81 clones were also designated according to their library/donor source (EB/TE) and their position in the master plate. Microtiter plates were coated with: 1) purified IIb3, 2) thrombin-activated platelets (AP), 3) WT-CHO cells expressing IIb3 in a low affinity state, and 4) Del-CHO cells expressing recombinant mutated IIb3 (GFFKR-deleted) in a high-affinity conformation. The binding of individual clones was detected using anti-M13 MAb followed by HRP-labeled anti-mouse Ab. The murine MAbs PAC-1 and Y2/51 were used as positive controls. Y2/51 is an IgG mAb that recognizes the 3 subunit and PAC-1 is an IgM mAb that recognizes the activated state of the IIb3 integrin. The binding of IgG and IgM MAbs were detected with peroxidase-labeled anti-IgG and anti-IgM Abs, respectively.

The 10 selected phage Abs were recloned into the pHOG21 vector for the production of soluble scFv. Western blot experiments were realized on purified bacterial lysates to show properly expressed soluble scFv Ab fragments of the expected size (Fig. 3). We obtained full-length scFv fragments for all scFv fragments except TEA9, TEB7, and EBE12, which could not be expressed as soluble fragments using our bacterial system despite several attempts and despite the fact that the pHOG21 expression vector contains a DNA insert of the expected size. The purified bacterial lysates showed specific binding to purified _IIb3 or _IIb3 on platelets by ELISA analysis (Fig. 4).

View larger version (48K): [in this window] [in a new window]   FIGURE 3. Immunoblot analyses of bacterial extracts from E. coli clones expressing soluble scFv fragments issuing from anti-IIb3 phage-Abs. The scFv were detected using mAb 9E10 recognizing the C-terminal c-myc epitope. Each sample corresponded to the scFv in 1 ml of culture. As a control, the scFv fragment derived from an hybridoma secreting an anti-myosine mAb (B7) was used, isolated from periplasmic extracts (60 ). The positions of the molecular mass markers are shown on the right.

View larger version (40K): [in this window] [in a new window]   FIGURE 4. ELISA analysis of the binding of soluble scFv to purified IIb3 or IIb3 on platelets. BSA was included in the test as an unrelated Ag. The scFv were detected using mAb 9E10 recognizing the C-terminal c-myc epitope. The binding of mAb 9E10 mouse IgG was detected with peroxidase-labeled goat anti-IgG Abs.

The scFv-binding affinities of TEG4 and EBB3 were assessed according to a previously described method that uses the Klotz equation (43). As shown in Fig. 5, TEG4 was found to have a K_d value of 2.6 x 10^-6 M and EBB3 has a K_d value of 1.8 x 10^-7 M. Binding specificity was confirmed by the demonstration that soluble _IIb3 inhibited the scFv interaction with immobilized _IIb3.

View larger version (11K): [in this window] [in a new window]   FIGURE 5. Functional affinity of TEG4 and EBB3 soluble scFv. Affinities were deduced from the Klotz equation 1/v = 1 + Kd/a0 where v is the fraction of bound Ab [(A0 - A)/A0] and a0 is the total Ag concentration. A0 is the absorbance measured in ELISA for the soluble scFv in the absence of Ag and A is the absorbance measured for the soluble scFv with a given concentration of Ag (a0).

Fig. 6 shows the inhibition of the binding of EBB3- and TEG4-soluble scFv samples to _IIb3 by fibrinogen in the micromolar range (TEG4, 3 x 10^-6 M; EBB3, 7 x 10^-6 M), clearly indicating specific binding and biological potency. Thus, these scFv fragments were competing with fibrinogen, the physiologic ligand of _IIb3. Concomitantly, the same experiments realized with actin (5 x 10^-9–3 x 10^-4 M) showed no relevant inhibition of scFv binding (shown for EBB3).

View larger version (21K): [in this window] [in a new window]   FIGURE 6. Capacity of fibrinogen to inhibit the interaction of soluble scFv with IIb3. The ability of fibrinogen to inhibit the interaction between IIb3 with the scFv EBB3 or TEG4 was evaluated in a competition ELISA performed in the presence of increasing concentrations of fibrinogen (9.3 x 1010–1.4 x 10-5 M). Results were expressed as a percentage of inhibition. The same competition ELISA performed for EBB3 with actin (5 x 10-9–3 x 10-4 M) showed no relevant inhibition of scFv binding on IIb3.

Sequence analyses of selected _IIb3-specific clones

The nucleotide sequences of the V_H and V_L genes of the selected _IIb3-specific clones were determined and aligned to the most homologous germline genes in the IMGT/DNA PLOT directory (http://imgt.cines.fr:8104/). For example, V_H domains of clones EBB10, TEH11, and TEA9 used V_H germline genes from the V_H1 family (see Table II). It should be noted that V_H domains from the V_H1 and V_H3 families all belonged to different germline genes. The number of somatic mutations in the anti-_IIb3 V_H domains ranged from 0 to 27. All have distinct CDR3 sequences with the use of different D segment genes. Only the JH4*02 gene, encoding the carboxy terminal part of CDR3, appeared to be preferentially rearranged, although it was truncated differently in the eight clones that used this segment. The 10 distinct V_H domains identified paired with a variety of V_L domains (Table II). For example, the V_L domains of clones EBA11, EBG10, EBB3, and EBE12 used V genes from the V3 family. However, only two V genes were used; TEA9 clones were encoded by a V1 germline gene and the TEH11 clone was encoded by a V2 gene. All V genes used the J3 segment (eight clones), whereas the J2 segment was used for remaining clones. The number of somatic mutations in the anti-_IIb3 V_L domains ranged from 0 to 19.

View this table: [in this window] [in a new window]   Table II. Nucleotide sequences of the VH and VL genes of IIb3-specific clones compared with the closest germline gene sequences in the IMGT/DNA PLOT directory

R mutations and Ag selection of the anti-_IIb3 phage Abs

An Ig gene displays a higher number of R mutations in the CDRs than given by chance alone when a positive pressure is exerted by Ag to enhance the fit of the Abs (42). The distribution and nature of the point mutations in the V segments were studied to determine whether our phage Abs display mutations that are consistent with selection by Ag. The results are given in Table III. In the case of a positive antigenic pressure in vivo that results in the mutation of the Ig V CDR structure, the likelihood that the excess of R mutations in the CDR arose by chance is low (taking the phage Ab EBA11 as an example, p = 0.017 for the V_H CDR portions). Concomitantly, a lower proportion of R mutations than was expected is observed in the FR portion (11 expected in the HFR of EBA11 for 8 observed). The calculated ratios between R and silent (S) mutations gave a value of 6 for the CDR and 2.6 for the FR. This strongly supports our hypothesis, because for an Ag-driven process, the R/S value was determined to be above 2.9 in the CDRs (44). Similarly, the V_H segment of clones EBE12, TEH11, TEF2, and TEB7, and the V_L segment of clones EBB10, EBB3, EBG10, TEF2, and TEA9 showed a higher concentration of R mutations and a higher R/S mutation ratio in the CDR regions than in the FR regions, findings that were highly consistent with selection by Ag. No R mutations occurred in EBB10, EBG10, TEG4, and TEA9 V_H segment CDRs, the V_H of EBG10 displaying an entirely germline configuration. Concerning the V_L segment CDR, no R mutations occurred in EBE12, TEG4, and TEB7. The V_L segment of TEH11 had two R mutations in the CDR without evidence of a positive antigenic pressure (R/S = one in CDR with R/S = two in FR). Overall, about one-half of our anti-_IIb3 phage Abs contained a statistically significant number of mutations in their V_H or V_L domains. V_H and V_L nucleotide sequences of each phage Ab were made available from EMBL data library, and accession numbers (from AJ308447 to AJ308466) were included in the database.

View this table: [in this window] [in a new window]   Table III. Distribution and nature of mutations in the variable genes of the anti-IIb3 phage-Absa

V_H CDR3 analysis of anti-_IIb3 Abs

The HCDR3 regions of our phage Abs were compared with those of other murine and human mAbs, particularly with respect to the presence of RGD and AGDV motifs found in fibrinogen domains reactive with the integrin (Table IV). Some of the sequences determined for our phage Abs contained variations of the RGD motif where the glycine was replaced (RWD and RVD) or lacking (RD). These type of sequences have previously been observed in phage libraries expressing cyclic CX6–7C peptides selected after panning with _IIb3 integrin (52). One phage Ab, EBA11, was found to contain another variation in its CDR3 region, an RNGD sequence. Variations were also found within the AGDV motif, with, for example ARDV in TEA9. The PAC-1 IgM murine mAb, which binds selectively to activated platelets, contained a variation of RGD: the RYD motif (38). Interestingly, a portion of the HCDR3 region of PAC-1, YYRYD containing the RYD motif, is similar to a portion of the HCDR3 of EBA11, YFSFD, according to multiple alignment analysis using the Genebee service (http://www.genebee.msu.su/services/malign_reduced.html). TEF2 also shares with PAC-1 a YDSSGRY motif and EBB3 shares a YG-SGSQ motif. However, the RYD sequence by itself is not able to induce activation dependence because two other mAbs, OP-G2 and LJ-CP3, also contain the RYD motif in the HCDR3 region and bind to both activated and nonactivated forms of _IIb3 (53). As many CDR3 sequences were found to harbor neither the RGD nor the AGDV sequences, multiple alignment analyses were performed to reveal other repeated motifs shared by reported anti-_IIb3 Abs and our phage Abs. These motifs are shown in bold type in Table IV and are aligned according to the strongest homologies using the Genebee service. The GXYY(F)XXD motif is encountered in EBA11, PAC-1, OP-G2, and LJ-CP3 Abs. The YFDY motif is encountered in four Abs, XIIF9, TEG4, U38663, and TEB7. The WDSRWDAF-DL sequence of EBE12 is similar to the WGSYRDPYFDY sequence of U38663 and to the WGG-WDHYMDV sequence of PDG31. The GNFGYFDY portion of TEB7 is highly homologous to the GNYGWFAY portion of U05216 and to the GRYSYNDH motif of TEF2. Finally, the RDVTLVR sequence carried by TEA9 was also determined by Genebee analysis as being structurally similar to the RDITVLP region of XIIF9, a murine mAb that was found to bind particularly well to activated platelets as shown by Scatchard analysis (54).

View this table: [in this window] [in a new window]   Table IV. Amino acid sequences of the CDR3H regions of anti-IIb3 mAbs from murine or human origin reported from GenBank/EMBL/IMGT databases and from the literaturea

	Discussion

Top Abstract Introduction MATERIALS AND METHODS Results Discussion References

We analyzed the somatic mutations present in the V genes of the Abs obtained from our patients to gain information on the secreting B cells. In general, V regions from Ag-activated B cells have nonrandom patterns of somatic mutations. These patterns result from the positive selection of cells with R mutations in CDRs, presumably because some of these mutations improve Ab affinity for Ag (20, 21, 22, 23, 24). Mutation analysis performed with our phage Abs revealed that 40–50% of the V_H and V_L genes coding for the Abs in both patients showed evidence of somatic hypermutation, supporting the idea that a significant part of the anti-_IIb3 response developed in AITP or GT patients is driven by Ag. Our results confirm recent experiments where phage display was successfully used to isolate Abs from individuals with demonstrable serum Ab responses to a variety of Ags, including infectious agents such as HIV-1 (17), mutated protein in malignancy (18), and self Ags in autoimmune diseases (12, 13, 14, 15, 16, 19). Isolated Abs have been shown to reflect the in vivo Ab response, and, even if theoretically, the combinatorial approach implies the possibility of pairings of L and H chains not present in the original source tissue. The existence of a nonrandom mutational process for one-half of our phage Abs provides strong support for the contention that they reflect an in vivo Ag-driven immune response in these patients. Moreover, previous studies reported that the recombination freedom of a V_H domain shaped by somatic mutation is limited, probably based on structural restriction (58).

Interestingly, both Ag-driven and germline clones specifically recognized the _IIb3 integrin, thus confirming previous results that germline chain genes can code highly specific Abs (59, 60). Studies are in progress to determine whether the highly specific germline IgG clones really secrete pathogenic Abs that retain a germline sequence or whether they represent anergized autoreactive B cells also found in nonimmunized subjects that possess the genetic potential to produce anti-_IIb3 Abs (61). However, despite several attempts, we have been unable to select anti-_IIb3 phage Abs from a complex naive library made from mRNA isolated from pooled blood of 50 healthy donors (33). Autoreactive B cells, even if present in naive blood samples, are certainly not proliferating, and, as a result, the RNA may not be in sufficient amounts to be amplified. Significantly, IgG anti-ds DNA were recovered from the library of a clinically active systemic lupus erythematosus patient, but not from the library of his healthy identical twin (12). Therefore, the minimally mutated Abs in our study may be either pathogenic by themselves or may be the template for the expansion of pathogenic Abs.

The events triggering the clonal proliferation process are as yet undefined, and a lot of questions remain open. Natural autoreactivity is highly regulated by active cellular and humoral control mechanisms. The selection of human natural self-reactive IgG Ab repertoires requires normal T/B cell interactions (62). More particularly, _IIb3-autoreactive T cells that share characteristics with anergic T cells have been demonstrated in healthy individuals under physiological conditions (63). In autoimmune diseases, altered T/B cell interactions may have an impact on the selection of uncontrolled self-reactive Ab repertoires. A defective control of IgG autoreactivity by anti-idiotypic Abs may also be an underlying mechanism for emergence of autoimmunity (64). Moreover, many cases of autoimmune pathologies such as AITP are a direct result of autoantibody binding. In AITP, highly specific anti-_IIb3 Abs are frequently developed. The differences between natural Abs directed against self-components and autoantibodies developed in autoimmune diseases often rely on polyspecificity and idiotopes on their V region targeted by anti-idiotypic Abs (22). Defects in the idiotypic network that ensures homeostasis of autoreactivity could have thus contributed, in AITP disease, to the uncontrolled emergence and expansion of highly specific anti-_IIb3 "pathogenic" autoantibodies presenting with a favorable VDJ rearrangement. Similar results in terms of percentage of mutated Abs and highly specific clones apply to polytransfused GT patients, in which case the stimulus would be the _IIb3 integrin present on donor platelets.

Our report describes the first recombinant phage Abs derived from splenocytes of an AITP patient. One other study reported the isolation of an anti-_IIb3 mAb from lymphocytes of a GT patient (7). When further tested under various conditions against activated and unactivated forms of _IIb3, some bound preferentially to the activated form. To our knowledge, this is the first time that human anti-_IIb3 Abs have been isolated from donors immunized against this integrin and shown to recognize its different activation states. The isolation of anti-_IIb3 Abs with different activation-dependent requirements, especially those that preferentially bind to an activated conformation, is of potential interest both from a fundamental and a therapeutic point of view. Such Abs may aid the further study of the ligand recognition specificity of the integrin. There are still many unknowns relating to the mechanisms leading to the activation of _IIb3 and to the structure of the ligand-binding pocket (65, 66, 67, 68). At least two ligand binding sites exist on _IIb3, and the activation requirement for ligand binding appears to depend on the type of ligand being examined. Binding of soluble adhesive proteins such as fibrinogen (AGDV-type ligand) or fibronectin (RGD-type ligand) require the ligand-binding pocket to be expressed, whereas low molecular mass-RGD-mimetic compounds can bind to _IIb3 in an activation-independent manner. The RGD sequence has often been used as a starting structure for designing anti-thrombotic agents. However, the CDR3 regions of previously described murine and human anti-_IIb3 mAbs are highly variable and very few of them contain the well-known RGD or AGDV sequences (see Table IV). In the current study, our phage Abs contained HCDR3 portions with various motifs whose binding specificities can be investigated in future work.

Studies from Bednar et al. (69) clearly indicate that agents recognizing the platelet _IIb3 receptor with high affinity and selectivity for activated rather than resting platelets possess a higher therapeutic potential for thromboembolic events. Development of activation-dependent human Abs that could mimic fibrinogen in their binding capacity may alleviate some adverse effects such as the increased bleeding risk encountered with the use of the present generation of _IIb3 antagonists in which abciximab, chimeric Fab₂ of the mAb 7E3 is the most widely used (29). Fully human mAbs would be more suitable for repeated therapy because of their lower immunogenicity (30). Furthermore, Abs preferentially recognizing the activated form may be active at lower therapeutic levels. Fully recombinant human scFv fragments as reported in Table IV able to preferentially bind to the activated form of _IIb3 will be investigated in further studies as potential antagonists of this integrin.

	Acknowledgments

 
We thank J. L. Chagnaud for technical assistance, T. Kunicki and D. Nugent for their gift of oligonucleotides, and P. Nurden for providing us with the blood samples from the patients. We are very grateful to B. Guillemain for providing space in his laboratory. Special thanks to Dr. J.-H. Bézian for providing space for the GISIR.

	Footnotes

 
¹ This study was funded by the Centre National de la Recherche Scientifique, by Université de Bordeaux 2, by Conseil Régional d’Aquitaine, by Ministère de l’Education Nationale de l’Enseignement et de la Recherché (Actions Concertés-Sciences du Vivant No. 12), by Ligue Departementale contre le Cancer de Bayonne, by Institut Electricité Santé, and by the Groupement d’Intérêt Scientifique sur les Immunoglobulines Recombinantes. M.-J.J. was financed by European funding to the Groupement d’Intérêt Scientifique sur les Immunoglobulines Recombinantes. Her training courses in Heidelberg were sponsored by a bourse of the Fondation de la Recherche Médicale (France).

² Current address: Affimed Therapeutics, Dr. A. Reimann Strasse 2, 68526 Lademburg, Germany.

³ Current address: BASF-LYNX Bioscience AG, Im Neuenheimer Feld 515, 69120 Heidelberg, Germany.

⁴ Address correspondence and reprint requests to Dr. Gisèle Clofent-Sanchez, Centre National de la Recherche Scientifique, Unité Mixte de Recherche 5533, Hopital Cardiologique, Avenue de Magellan, 33604, Pessac, France. E-mail address: Gisele.Clofent{at}umr5533.u-bordeaux2.fr

⁵ Abbreviations used in this paper: AITP, autoimmune thrombocytopenic purpura; GT, Glanzmann thrombasthenia; CDR, complementarity-determining region; CHO, Chinese hamster ovary; WT, wild type; WT-CHO, WT _IIb3; Del-CHO, high- affinity, GFFKR-deleted _IIb3; FR, framework region; R, replacement; S, silent; HCDR3, H chain CDR3.

Received for publication February 9, 2001. Accepted for publication December 3, 2001.

	References

Top Abstract Introduction MATERIALS AND METHODS Results Discussion References

 

1. Nurden, A. T., C. Poujol, C. Durrieu-Jais, P. Nurden. 1999. Platelet glycoprotein IIb/IIIa inhibitors: basic and clinical aspects. Arterioscler. Thromb. Vasc. Biol. 19:2835.[Free Full Text]
2. McMillan, R.. 2000. Autoantibodies and autoantigens in chronic immune thrombocytopenic purpura. Semin. Hematol. 37:239.[Medline]
3. Clofent-Sanchez, G., S. Lucas, J. Laroche-Traineau, P. Rispal, J. L. Pellegrin, P. Nurden, A. Nurden. 1996. Autoantibodies and anti-mouse antibodies in thrombocytopenic patients as assessed by different MAIPA assays. Br. J. Haematol. 95:153.[Medline]
4. Jallu, V., M. Diaz-Ricart, A. Ordinas, M. Pico, G. Vezon, A. T. Nurden. 1994. Two human antibodies reacting with different epitopes on integrin ₃ of platelets and endothelial cells. Eur. J. Biochem. 222:743.[Medline]
5. Olee, T., J. En, C.-J. Lai, L. Mo, C. S. Cho, X. Wei, X.-F. Wang, V. L. Woods, P. Chen. 1997. Generation and analysis of an IgG anti-platelet autoantibody reveals unusual molecular features. Br. J. Immunol. 96:836.
6. Kipriyanov, S. M., M. Little. 1999. Generation of recombinant antibodies. Mol. Biotechnol. 12:173.[Medline]
7. Ishida, F., Y. Gruel, E. Brojer, D. J. Nugent, T. J. Kunicki. 1995. Repertoire cloning of a human IgG inhibitor of _IIb3 function: OG idiotype. Mol. Immunol. 32:613.[Medline]
8. Griffin, H. M., W. H. Ouwehand. 1995. A human monoclonal antibody specific for the leucine-33 (P1A1, HPA-1a) form of platelet glycoprotein IIIa from a V gene phage display library. Blood 86:4430.[Abstract/Free Full Text]
9. Proulx, C., P. Chartrand, V. Roy, M. Goldman, F. Decary, A. Rinfret. 1997. Human monoclonal Fab fragments recovered from a combinatorial library bind specifically to the platelet HPA-1a alloantigen on glycoprotein IIb-IIIa. Vox Sang. 72:52.[Medline]
10. Okamoto, N., S. D. Kennedy, E. A. Barron-Casella, J. F. Casella, H. Inoko, T. S. Kickler. 1998. Identification of a human heavy chain antibody fragment directed against human platelet alloantigen 1a by phage display library. Tissue Antigens 51:156.[Medline]
11. Macchi, L., G. Clofent-Sanchez, G. Marit, C. Bihour, C. Durrieu-Jais, P. Besse, P. Nurden, A. T. Nurden. 1996. PAICA: a method for characterizing platelet-associated antibodies: its application to the study of idiopathic thrombocytopenic purpura and to the detection of platelet-bound c7E3. Thromb. Haemost. 76:1020.[Medline]
12. Roben, P., S. M. Barbas, L. Sandoval, J. M. Lecerf, B. D. Stollar, A. Solomon, G. J. Silverman. 1996. Repertoire cloning of lupus anti-DNA autoantibodies. J. Clin. Invest. 98:2827.[Medline]
13. Chukwuocha, R. U., E. T. Hsiao, P. Shaw, J. L. Witztum, P. P. Chen. 1999. Isolation, characterization and sequence analysis of five IgG monoclonal anti-₂-glycoprotein-1 and anti-prothrombin antigen-binding fragments generated by phage display. J. Immunol. 163:4604.[Abstract/Free Full Text]
14. Van den Brink, E. N., E. A. M. Turenhout, J. Davies, N. Bovenschen, K. Fijnvandraat, W. H. Ouwehand, M. Peters, J. Voorberg. 2000. Human antibodies with specificity for the C2 domain of factor VIII are derived from VH1 germline genes. Blood 95:558.[Abstract/Free Full Text]
15. McIntosh, R. S., M. S. Asghar, P. F. Watson, E. H. Kemp, A. P. Weetman. 1996. Cloning and analysis of IgG and IgG anti-thyroglobulin autoantibodies from a patient with Hashimoto’s thyroiditis: evidence for in vivo antigen-driven repertoire selection. J. Immunol. 157:927.[Abstract]
16. Graus, Y. F., M. H. de Baets, P. W. Parren, S. Berrih-Aknin, J. Wokke, P. J. van Breda Vriesman, D. R. Burton. 1997. Human anti-nicotinic acetylcholine receptor recombinant Fab isolated from thymus-derived phage display libraries from myasthenia gravis patients reflect predominant specificities in serum and block the action of pathogenic serum antibodies. J. Immunol. 158:1919.[Abstract]
17. Toran, J. L., L. Kremer, L. Sanchez-Pulido, I. M. de Alboran, G. del Real, M. Llorente, A. Valencia, M. A. de Mon, A. C. Martinez. 1999. Molecular analysis of HIV-1 gp120 antibody response using isotype IgM and IgG phage display libraries from a long-term non-progressor HIV-1-infected individual. Eur. J. Immunol. 29:2666.[Medline]
18. Coomber, D. W., N. J. Hawkins, Clark M. A., R. L. Ward. 1999. Generation of anti-p53 Fab from individuals with colorectal cancer using phage display. J. Immunol. 163:2276.[Abstract/Free Full Text]
19. Welschof, M., P. Terness, S. M. Kipriyanov, D. Stanescu, F. Breitling, H. Dorsam, S. Dubel, M. Little, G. Opelz. 1997. The antigen-binding domain of a human IgG anti-F(ab')₂ autoantibody. Proc. Natl. Acad. Sci. USA 94:1902.[Abstract/Free Full Text]
20. Huetz, F., F. Jacquemart, C. P. Rossi, F. Varela, A. Coutinho. 1988. Autoimmunity: the moving boundaries between physiology and pathology. J. Autoimmun. 1:507.
21. Pascual, V., Y. J. Li, A. Magalski, O. de Bouteiller, J. Banchereau, J. D. Capra. 1994. Analysis of somatic mutation in five B cell subsets of human tonsil. J. Exp. Med. 180:329.[Abstract/Free Full Text]
22. Lacroix-Desmazes, S., S. V. Kaveri, L. Mouthon, A. Ayouba, E. Malanchere, A. Coutinho, M. D. Kazatchkine. 1998. Self-reactive antibodies (natural autoantibodies) in healthy individuals. J. Immunol. Methods 216:117.[Medline]
23. P. Kasaian, M. T., H. Ikematsu, J. E. Balow, P. Casali. 1994. Structure of the V_H and V_L segments of monoreactive and polyreactive IgA autoantibodies to DNA in patients with systemic lupus erythematosus. J. Immunol. 152:3137.[Abstract]