help button home button Endocrine Society Endocrinology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Chapal, N.
Right arrow Articles by Péraldi-Roux, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chapal, N.
Right arrow Articles by Péraldi-Roux, S.
Endocrinology Vol. 142, No. 11 4740-4750
Copyright © 2001 by The Endocrine Society

ARTICLES

Thyroid Peroxidase Autoantibodies Obtained from Random Single Chain Fv Libraries Contain the Same Heavy/Light Chain Combinations as Occur in Vivo

Nicolas Chapal, Thierry Chardès, Damien Bresson, Martine Pugnière, Jean-Claude Mani, Bernard Pau, Majida Bouanani and Sylvie Péraldi-Roux

Centre National de la Recherche Scientifique-Unité Mixte de Recherche 5094, Faculté de Pharmacie (N.C., D.B., M.P., J.-C.M., B.P., M.B., S.P.-R.), 34060 Montpellier, France; Institut National de la Recherche Agronomique-Centre National de la Recherche Scientifique-Unité Mixte de Recherche 5087 (T.C.), Saint Christol 30380 Lez Alès, France

Address all correspondence and requests for reprints to: Dr. Sylvie Peraldi-Roux, Faculté de Pharmacie, Centre National de la Recherche Scientifique-UMR 5094, Institut de Biotechnologie et Pharmacologie, 15 avenue Charles Flahault, 34060 Montpellier Cedex 2, France. E-mail address: sylvie.roux{at}ibph.pharma.univ-montp1.fr


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Three combinatorial libraries were constructed from unpurified, CD19+, and antithyroid peroxidase (anti-TPO) B cells extracted from thyroid tissue of Graves’ disease patients. Fifteen of the 41 randomly derived anti-TPO single chain variable region fragments (scFvs), showed VH1–3/V{lambda}1–51 or VH1–69/V{lambda}1–40 heavy/light chain pairing similar to that obtained with TPO-specific scFv derived from an in-cell library. One VH1–3/V{lambda}1–51 scFv, A16, showed exactly the same nucleotide sequence as in-cell scFv ICB7, demonstrating that in vivo rearrangement can be obtained from a random combinatorial library. The majority of the scFvs used a heavy chain gene derived from the VH1–3 gene segment, whereas the light chain gene segments used were more heterogeneous, with dominance of the V{kappa}1–39 and V{lambda}1–51 gene segments. The anti-TPO scFvs showed high affinities to TPO, with values between 0.77 and 12.3 nM, and defined seven antigenic regions on the TPO molecule. The anti-TPO fragments, particularly VH1–3/V{lambda}1–51 randomly associated scFv B4, which mimic natural H/L pairing, and VH1–3/V{lambda}1–40 in-cell-derived scFv ICA5, efficiently displaced the TPO binding of serum autoantibodies from 20 Graves’ disease patients. Our study directly demonstrates that antibodies derived from combinatorial libraries are likely to represent in vivo pairing, leading to high affinity antibody fragments mimicking the binding of serum autoantibodies to TPO.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
AUTOIMMUNE THYROID DISEASE is one of the most common organ-specific autoimmune syndromes. In this disease the humoral response is directed against one or more of the thyroid antigens: Tg, thyroid peroxidase (TPO), or the TSH receptor (TSHr). The use of combinatorial libraries for producing recombinant antibodies allowed several laboratories to obtain and characterize numerous recombinant {kappa} and some {lambda} anti-TPO antibodies (Abs) (1, 2, 3, 4, 5, 6, 7). One major question when using this technology is the possibility of finding the in vivo H/L pairing in recombinant antibodies obtained by a random combinatorial library. Furthermore, it is not clear whether the generation of artificial heavy (VH) and light chain (VL) combinations from random combinatorial library reflects the in vivo situation. Although two reports suggest that in vivo pairing of the H and L chains is never found in random combinatorial libraries (8, 9), one group hypothesized that Abs from a combinatorial library are closely related to the immune response of the donor (10). Furthermore, the likelihood that chain pairing in a TPO-selected random library contains in vivo H/L combinations has been emphasized (11, 12). We recently obtained three {lambda} anti-TPO single chain variable region fragments (scFvs) by an in-cell library (13) that reflect in vivo rearrangements. These results point out the need to develop an extended antibody repertoire from combinatorial libraries to definitively discriminate between natural H/L pairing and artificial association.

With these perspectives in mind, we constructed three different random combinatorial libraries using unpurified, CD19+ or TPO+ cells from thyroid-infiltrating B lymphocytes (TIBL) and compared the H/L pairing of the single chain variable region fragments (scFvs) obtained by random combinatorial libraries with those previously produced by in-cell PCR. We provide herein the first direct demonstration that in vivo VH1–3/V{lambda}1–51 or VH1–69/V{lambda}1–40 H/L pairing can be found in random TPO-specific combinatorial libraries. Secondly, we confirmed that the TPO-specific Ab repertoire shows, independently of the starting B cells, large VH gene restriction to the VH1–3 germline usage, whereas a more heterogeneous VL gene use was observed with dominance of V{kappa}1–39 and V{lambda}1–51 genes. Finally, the anti-TPO fragments, particularly VH1–3/V{lambda}1–51 randomly associated scFv B4 and VH1–3/V{lambda}1–40 in-cell-derived scFv ICA5, specifically inhibited the binding of autoantibodies (aAbs) to TPO. Our study demonstrates that a large Ab repertoire derived from combinatorial libraries can reflect the in vivo situation, and that the effective TPO-specific aAb response in Graves’ disease is more diverse than that previously described (1, 2, 3, 4, 5, 6, 7).


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Isolation and purification of TIBL
The TIBL used for library construction were isolated from biopsies from seven patients (bdw11, bdw16, bdw41, bdw42, bdw43, bdw44, and bdw45) suffering from Graves’ disease. These patients were between 23–50 yr of age. The anti-TPO aAbs of these patients’ sera were analyzed by direct ELISA on human purified TPO (purity, >85%; HyTest, Turku, Finland). They had high anti-TPO [titers, 1:2,500 to 1:13,000 (titer defined as the dilution giving an A450nm of 1)]. The sera of 20 Graves’ disease patients with high anti-TPO titers (A450nm = 1 for a serum dilution of 1:300 to 1:13,000) were used for ELISA inhibition studies. Most of the patients’ sera were further characterized for anti-Tg Abs and for anti-TSHr Abs by ELISA (Bio-Rad Laboratories, Inc., Marnes-La-Coquette, France; Table 1Go).


View this table:
[in this window]
[in a new window]
  		Table 1. Titer of thyroid autoantibodies in serum of Graves’ disease patients

 
TIBL were isolated from the thyroid gland of patients bdw11 and bdw16 and were purified with anti-CD19 magnetic beads as previously described (14). The B cells from patient bdw43 were purified on human TPO with anti-TPO magnetic beads. Briefly, the cells were resuspended in 1 ml PBS and 1% BSA containing 5 µg/ml human TPO and incubated for 30 min at room temperature with gentle shaking. After three washings with PBS/1% BSA, the cell-TPO complexes were resuspended in 1 ml PBS/1% BSA containing anti-TPO-coupled magnetic beads (Estapor, Rhone-Poulenc, France) and incubated for 30 min at room temperature with gentle shaking. The bound cells were washed extensively with PBS using a magnet before cell recovery. Finally, dissociated-thyroid tissues from patients bdw41, bdw42, bdw43, and bdw45 were used as an unpurified B cell suspension (unpurified B cells).

Combinatorial library construction, selection, and expression of scFvs
Different scFv combinatorial libraries were obtained depending on the starting cells, i.e. unpurified cells (unpurified library), CD19+ cells (CD19+ library), or TPO+ B cells (TPO+ library). After total RNA extraction from cells by using TRIzol reagent (Life Technologies, Inc., Gaithersburg, MD), RT was performed on the total RNA using SuperScript II reverse transcriptase (Life Technologies, Inc.). For construction of the CD19+ library, the PCR amplification protocol was the same as the one we used for construction of the in-cell library (14). For construction of the two other libraries, two PCR steps were performed, one for the amplification of the VH and VL Ig genes and one for the amplification of scFv products after loxP-Cre recombination and the introduction of restriction sites. The amplified and associated scFv genes were cloned and transformed in Escherichia coli XL1-Blue-competent cells by electroporation. The loxP site, which allows recombination between VH and VL genes, was removed by NheI (New England Biolabs, Inc., Beverly, MA) restriction enzyme digestion.

Phage stock solutions were prepared using a standard procedure involving M13K07 helper phages (Bio-Rad Laboratories, Inc., Hercules, CA) routinely yielding phage titers between 1011 and 1012 titration units/ml. One or two pannings were performed on human TPO as previously described (13). Briefly, immunotubes (Nunc, Roskilde, Denmark) were coated overnight at 4 C with 2 ml 5 µg/ml human TPO in 0.1 M carbonate/bicarbonate buffer, pH 9.6. After washing and a saturation step, 2 x 109 titration unit phage-displayed scFv fragments were incubated in PBS/2% nonfat milk at room temperature with gentle shaking. The bound phages were eluted by adding 0.1 M glycine-HCl (pH 2.2), neutralized, and used to infect the XL1-Blue cells. The bacteria were plated and then used to make a new phage preparation for the next round of panning.

Soluble expression of scFvs was performed in HB2151 E. coli cells after induction with 1 mM isopropyl-ß-thiogalactopyranoside for 3 h at 25 C. The cells were then pelleted, resuspended at a 1:40 dilution of the culture volume in lysis buffer [20 mM HEPES (pH 8) and 1 mg/ml polymixin B supplemented with protease inhibitors], and incubated for 30 min on ice (15). The suspension was centrifuged for 10 min at 13,000 rpm, and the supernatant containing the scFv fragments was stored or used directly for the immunoreactions.

Sequencing of recombinant clones
Sequences were determined by cycle sequencing using the ABI Prism Rhodamine Terminator Cycle Sequencing Ready Reaction Kit (PE Applied Biosystems, Foster City, CA). The primers used were the same as those previously described (14). Sequences were run on an ABI Prism 377 electrophoresis system (PE Applied Biosystems). Sequence alignments with germline genes were performed using the IMGT sequence directory (http://imgt.cnusc.fr) (16).

TPO binding activities of scFv fragments and determination of TPO antigenic regions
The binding of soluble scFv fragments to TPO was assessed by ELISA as previously described (13). The microtiter plates were coated overnight at 4 C with human TPO (1 µg/ml in 0.1 M carbonate/bicarbonate buffer, pH 9.6). After washing and saturation, the scFv fragments were added to the plates and incubated for 2 h at room temperature. Bound scFv fragments were detected by the anti-Myc monoclonal antibody 9E10 (17) incubated for 1.5 h at room temperature. Next, an alkaline phosphatase-conjugated Fc-specific antimouse IgG (Sigma, St. Louis, MO) was added and incubated for 1.5 h at room temperature. Enzyme activity was assayed by addition of 4-nitrophenyl phosphate as substrate.

The affinity and the epitope mapping of the anti-TPO scFv fragments were determined using BIACORE 2000 (Biacore AB, Uppsala, Sweden). For the affinity experiments, the anti-Myc monoclonal antibody 9E10 was covalently immobilized on the flow cell of a CM5 sensor chip surface activated with 100 mM N-ethyl-N'-(3-dimethylaminopropyl) carbodiimide hydrochloride and 400 mM hydroxysuccinimide (EDC/NHS). The anti-TPO scFv periplasmic extracts were diluted 1:4 in HBS-EP buffer, pH 7.4, buffer (Biacore AB) and injected at a flow rate of 20 µl/min over the mAb 9E10. The scFv fragments were retained by the c-Myc peptide present at the C-terminal end of each scFv. Known concentrations of TPO were then injected to determine the affinity of the scFv fragments for TPO. The CM5 sensor chip with immobilized 9E10 was regenerated after each kinetic experiment by a 20-sec pulse with 100 mM HCl. The same sensor chip was used for all reported kinetic studies. The kinetic variables were calculated using the BIAevaluation 3.0 software (Biacore AB) (18). For epitope mapping, TPO was covalently immobilized on the flow cell of a CM5 sensor chip surface activated as described above. The epitopes recognized by the scFv fragments were determined by assaying all scFv pairs for their capacity to bind simultaneously to the TPO. The scFv periplasmic extracts diluted 1:2 in HBS, pH 7.4, were successively passed over the immobilized TPO as either first or second Ab. Three injections were required to saturate the epitope of the first Ab so that the binding of the second Ab could be interpreted as being due to the existence of a different epitope. An irrelevant antidigoxin scFv fragment was used as a negative control.

Competition studies between scFv fragments and serum anti-TPO aAbs
scFv fragment binding inhibition by serum anti-TPO autoantibodies of 20 patients suffering from Grave’s disease was performed by ELISA (13) with some modifications. Native human TPO (1 µg/ml in carbonate/bicarbonate buffer) was coated overnight at 4 C. After saturation, patients’ sera diluted 1:10 in PBS, 0.1% Tween, and 1% nonfat milk were added, and the microtiter plates were incubated for 1.5 h at room temperature. The microtiter plates were washed, and the scFv fragments, diluted 1:10, were added and incubated for 2 h at room temperature. Bound scFv fragments were detected as described above for the binding of soluble scFv to TPO by ELISA.

For inhibition of serum anti-TPO autoantibody binding by scFv fragments, microtiter plates were coated with TPO as described above. After saturation, scFv fragments diluted 1:2 in PBS, 0,1% Tween, and 1% nonfat milk were added and incubated for 2 h at room temperature. Then, the patients’ sera, at a dilution giving an absorbance of 1, were added to the wells, and the microtiter plates were incubated for 1.5 h at room temperature. After three washings with PBS/0.1% Tween, bound anti-TPO aAbs were detected by an alkaline phosphatase-conjugated Fc-specific antihuman IgG (Sigma) diluted 1:2000 in PBS, 0.1% Tween, and 1% nonfat milk. Enzyme activity was assayed by addition of 4nitrophenyl phosphate as substrate.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Library construction and anti-TPO scFv fragment selection
Three distinct combinatorial libraries were constructed from the unpurified, CD19-purified, and TPO-purified B cells. The sizes of these libraries were 2 x 105, 106, and 4 x 106 clones, respectively. To obtain at least 80% of colonies specific for human TPO, the CD19+ and the unpurified B cell phage-displayed libraries were subjected to two rounds of panning on highly purified TPO (>95%). In contrast, one round of panning was sufficient to select 100% colonies showing TPO-specific scFv fragments from the TPO+ library. After sequencing, we obtained 17 different anti-TPO scFv fragments (A1 to A17) of 22 clones sequenced from the CD19+ library, 11 different anti-TPO scFv fragments (B1 to B11) of 23 clones sequenced from the unpurified B cell library, and finally 13 different anti-TPO scFv fragments (T1 to T13) of 15 clones sequenced from the TPO+ library (Table 2Go).


View this table:
[in this window]
[in a new window]
  		Table 2. Anti-TPO scFv fragments isolated from three different combinatorial libraries

 
H/L pairing of TPO-specific scFv fragments
As shown in Table 2Go, dominant genes in H/L pairing (VH1–3, V{lambda}1–51, and V{lambda}1–40) were used by the anti-TPO scFv whatever their library of origin. Analysis of H/L pairing at the VH/VL family level showed that the VH1–3-derived scFvs were found associated with almost all (8 VL chains of 10) of the light chains (Table 2Go). Among them, 14 anti-TPO scFvs showed the same VH1–3/V{lambda}1–51 pairing as the in-cell scFvs ICA1 and ICB7 previously described (13). Similarly, identical VH1–69/V{lambda}1–40 pairing was found in scFv A10 derived from the CD19+ combinatorial library and in ICA5 scFv obtained from the in-cell library (13).

Analysis of H/L pairing at the VDJ-H/VJ-L level indicates that dominant associations were found in scFvs derived from the unpurified and CD19+ libraries. More precisely, 8 scFvs of 17 from the CD19+ library used the VH1–3 gene associated with the D4–17/JH4 gene segment and paired with the V{lambda}1–40/J{lambda}1 or V{lambda}1–51/J{lambda}1 rearrangement. This latter pairing corresponds to that described for in-cell scFvs ICA1 and ICB7 (13). The anti-TPO scFv A16, derived from the CD19+ combinatorial library, showed VDJ-H/VJ-L pairing similar to that observed for in-cell ICB7 (13) with an identical nucleotide sequence. Five of 10 scFvs derived from the unpurified library used the VH1–3/D5–24/JH4 gene association paired with 4 different VJ-L genes from either the {kappa} or {lambda} locus. On the other hand, scFvs produced from the TPO+ library used various VDJ-H rearrangements paired with diverse VJ-{kappa} or -{lambda} genes. These results are indicative of a greater diversity of gene segment usage and H/L pairing for scFvs derived from the TPO+ library.

V gene analysis of anti-TPO scFv fragments
The V genes encoding the 41 TPO-specific scFv fragments, selected on the basis of their nucleotide sequence and derived from the three combinatorial libraries, were analyzed and compared with the closest putative germline genes (Tables 2Go, 3Go, and 4Go). The scFvs used 6 different families of VH genes, involving 3 VH1, 2 VH3, and 1 VH5 germline genes. The majority of the H chain V regions were encoded by genes derived from the germline gene VH1–3 (33 clones). These VH1–3 scFvs mainly used D4–17 and D5–24 putative D regions, and all but 4 had the same JH4 segment. However a greater diversity of D regions used by the VH1–3 scFvs obtained from the TPO+ library was observed by comparison with those derived from the 2 other libraries. The 5 other scFv groups, each represented by 1–3 clones, were derived from VH1–8, VH1–69, VH3–30, VH3–64, and VH5–51 germline genes. No apparent D restriction was observed in these cases. The 2 scFvs using the VH1–69 germline gene were associated with a JH4 segment, and the 3 scFvs using the VH3–64 germline gene were associated with a JH6 segment (Table 2Go).


View this table:
[in this window]
[in a new window]
  		Table 3. Amino acid sequences of TPO-specific scFv heavy chains aligned with the putative closest germline genes1

 

View this table:
[in this window]
[in a new window]
  		Table 4. Amino acid sequences of TPO-specific scFv light chains aligned with the putative closest germline genes1

 
The L chain V regions were mainly encoded by genes closest to V{lambda}1–51 (16 clones), V{lambda}1–40 (9 clones), and V{kappa}1–39 (6 clones) germline genes, with dominance of V{lambda}1–51 and V{kappa}1–39 germline genes inside each {lambda} and {kappa} locus. The scFvs used V{lambda}1–51 and V{lambda}1–40 germline genes in association with either J{lambda}1 or J{lambda}3 segments, whereas V{kappa}1–39 scFvs used J{kappa}2, J{kappa}4, and J{kappa}5 segments. The 10 other scFv fragments were derived from V{lambda}1–44, V{lambda}2–14, V{lambda}2–8, V{kappa}1–12, V{kappa}1–5, and V{kappa}3–11 germline genes and used different J segments (Table 2Go). Comparison of the unpurified B cell and TPO+ libraries constructed from cells at different stages of differentiation shows a different VL germline gene usage. Concerning the 6 VL germline genes (V{lambda}1–51, V{lambda}2–8, V{lambda}1–40, V{kappa}1–5, V{kappa}3–11, and V{kappa}1–39) in the TPO+ library, a majority of V{kappa} germline genes (8 of 13) was used by the 13 scFvs (T1-T13), whereas among the 6 VL germline genes (V{lambda}1–40, V{lambda}2–14, V{lambda}7–43, V{lambda}1–51, V{kappa}1–39, and V{kappa}1–12) obtained from the unpurified B cell library, only 2 of the 11 scFvs used a {kappa} gene.

All of the anti-TPO scFv sequences showed evidence of somatic hypermutation, with a replacement/silent ratio typically higher in the VH chain than in the VL chain. The replacement/silent ratio was greater in the complementarity-determining region (CDR) regions than in the framework region (FR) regions (Table 2Go). Although some scFv VL chain sequences showed high somatic hypermutation (A9, B3, B7, T5, and T9), a large majority showed little, if any, mutation. The number of mutations was not dependent on the germline gene used by the anti-TPO scFvs, as the scFvs derived from the same germline genes, V{lambda}1–51, V{lambda}1–40, and V{kappa}1–39, were either highly mutated or completely unmutated. The fact that 80% of the characterized scFvs showed the same VH1–3 germline gene, whereas a wider VL gene usage was observed, suggests that the sequences encoding anti-TPO heavy chains are more restricted than those encoding light chains.

Amino acid sequence analysis of anti-TPO scFv fragments
Hypermutations were analyzed in detail for the 33 different VH1–3-derived scFv sequences, including those showing H/L pairing similar to that which occurred in vivo, and revealed 3 distinct patterns (Table 3Go). First, somatic gene mutations, mainly located in the CDRs, were common to most of these anti-TPO scFvs, whatever the patient and the library of origin. In CDR1, the threonine in positions 29 and 31 was mainly replaced by a serine, and the alanine in position 34 was replaced by an asparagine or a glycine; in FR2, the methionine in position 39 was replaced by an isoleucine. In CDR2, the alanine in position 58 was replaced by a glutamine, and the asparagine 60 was replaced by a threonine in most of the anti-TPO scFvs (Table 3Go). Second, some mutations were restricted to anti-TPO scFvs derived from only 1 random library. For example, the arginine in position 49 in FR2 was replaced by a glycine, and the threonine in position 63 in CDR2 was replaced by an arginine in most of the scFvs from the CD19+ library. Mutation of valine 12 to methionine was also restricted to scFvs obtained from the CD19+ library (Table 3Go). Third, similar hypermutations were observed in scFvs derived from TPO+ and unpurified B cell combinatorial libraries, i.e. residue mutations in positions 6, 13, 14, 42, 50, 53, 62, 89, and 92. On the other hand, replacement of asparagine 57 by histidine was mainly represented in scFvs from CD19+ and TPO+ libraries. Taken together, the pattern of gene somatic hypermutation in the VH region could be correlated with the starting B cells used for the construction of the libraries. In contrast, no specific assignment of a particular mutation was observed for the VL region (Table 4Go).

TPO binding activity of scFv fragments
Among the 41 scFv fragments selected on the basis of their gene sequences, 17 Ab fragments presenting a TPO Ab titer, by ELISA, greater than 100 were further characterized (Table 5Go). The affinity of these anti-TPO scFvs was determined on BIACORE 2000 by real-time interaction analysis, as exemplified by clone T2 (Fig. 1Go). The scFvs had affinities for TPO ranging from 0.77–12.3 nM (Table 5Go). As exemplified by clones B2 (VH1–3/V{kappa}1–39) and B4 (VH1–3/V{lambda}1–51), the affinity was similar independently of natural vs. artificial H/L pairing and of V{lambda}- vs. V{kappa}-chain usage (Table 5Go). The scFv-binding domains on TPO were analyzed by competition studies on BIACORE 2000. All scFv pairs were assayed by sequential injections on TPO as previously described (13). Binding of the second scFv to TPO was interpreted as recognition by this scFv of a different antigenic region than that recognized by the first scFv, as exemplified in Fig. 2Go for the pair of scFvs B4 and B10. Antigenic regions I and II were previously described as Ab-binding regions on TPO for in-cell scFvs ICA1 and ICA5, respectively (13). Seven other antigenic regions (III–IX) were defined by scFvs obtained from the 3 random combinatorial libraries (Table 5Go). Antigenic domains VI and VIII, found in the 3 libraries, are closely related and are dominant in the TPO+ and unpurified B cell libraries.


View this table:
[in this window]
[in a new window]
  		Table 5. Characterization of the anti-TPO scFv fragments

 


View larger version (8K):
[in this window]
[in a new window]
  		Figure 1. Affinity measurement of anti-TPO scFvs by BIACORE 2000. Kinetics of TPO binding to scFv T2 immobilized onto a CM5 sensor chip through the anti-Myc mAb 9E10. The experimental curve and the theoretical curve, obtained by the global method using BIAevaluation 3.0 software, are superimposed.

 


View larger version (12K):
[in this window]
[in a new window]
  		Figure 2. Epitope mapping of anti-TPO scFvs by BIACORE 2000. After three injections of the scFv B4 to saturate its binding site on TPO, the scFv B10 still bound TPO (a), whereas the scFv B6 did not (b). The scFvs B4 and B6, but not B10, recognized the same domain on TPO. RU, Resonance units.

 
To study more precisely the relationship between antigen specificity and H/L pairing or V chain gene usage, we focused our attention on the scFvs derived from the same VH gene. The same VH1–3-derived gene was found associated with two closely related V{lambda}1–51-derived genes in scFvs A2 and A11 (Tables 3Go and 4Go); these scFvs recognized two different domains, domains III and V, respectively (Table 5Go), on TPO. In contrast, scFvs B2 and B4, which used H chain genes close to the VH1–3 germline gene and showed different L chain genes derived from V{kappa}1–39 and V{lambda}1–51, respectively, recognized the antigenic region VI on TPO (Table 5Go). Similarly, scFvs T8 and T13, which used the same H chain genes close to the VH1–3 germline gene (Table 3Go) and different L chain genes derived from V{kappa}1–39 and V{lambda}1–40 (Table 4Go), respectively, recognized the same domain (namely VIII) on TPO (Table 5Go). Taken together, these results indicate no relationship between antigenic specificity and variable chain usage or H/L pairing.

Competition studies between scFv fragments and serum anti-TPO autoantibodies
For each epitope defined by the BIACORE study, we chose 1 scFv for competition studies with the serum autoantibodies. Twenty sera from patients suffering from Graves’ disease and 20 sera from healthy subjects were used to test their ability to inhibit the binding of the scFvs to TPO. Strong inhibitions, ranging between 60–100%, were obtained with each patients’ serum, but not with normal sera (data not shown), indicating that the serum aAb from patients with Graves’ disease recognized regions on TPO similar to those recognized by the scFvs tested. We also tested the ability of the scFv to inhibit patient serum binding to TPO. As shown in Table 6Go, the anti-TPO scFv defining the antigenic domains I, III, IV, V, VII, and IX showed moderate inhibition, whereas the scFvs VH1–3/V{lambda}1–51 B4 (domain VI), VH1–3/V{lambda}1–40 T13 (domain VIII), and VH1–69/V{lambda}1–40 ICA5 (domain II) were able to strongly inhibit the binding of most patients’ sera to TPO, suggesting that these scFvs recognized immunodominant epitopes.


View this table:
[in this window]
[in a new window]
  		Table 6. Serum anti-TPO autoantibody binding inhibition by scFv fragments

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
The phage display system has the potential for greatly facilitating the generation and characterization of human Abs. Gene families and epitope recognition of recombinant anti-TPO aAbs (2, 3, 4, 5, 6, 7, 19, 20, 21) have been characterized using combinatorial libraries. Although two studies stressed the likelihood that H/L pairing in TPO-specific combinatorial libraries reflects the in vivo situation (11, 12), the direct demonstration that original H/L pairing can be obtained from a random combinatorial library remained to be elucidated. To clarify this point, we constructed three random combinatorial libraries from different B cell subsets extracted from thyroid biopsies from patients suffering from Graves’ disease. The H/L pairing of the anti-TPO aAbs obtained as well as the nucleotide sequence of the scFv fragments were compared with the recombinant aAbs, previously produced from an in-cell library (13), which reflects the in vivo situation. The three in-cell anti-TPO scFvs, ICA1, ICA5, and ICB7, used VH1–3 and VH1–69 germline genes associated with V{lambda}1–51 and V{lambda}1–40 genes, respectively. ICA1 and ICB7 show the same VDJ-H rearrangement, but differ in VL CDR3. Regarding the H/L pairing, 15 of the 41 scFvs obtained by random pairing used the same association as that found for in-cell scFvs. Particularly, one VH1–3/V{lambda}1–51 anti-TPO scFv derived from the CD19+ random combinatorial library shows exactly the same nucleotide sequence as in-cell ICB7 (13) and consequently the same H/L pairing. These results clearly indicate that scFv from random combinatorial libraries can be formed by H/L pairs similar as those found in vivo, as previously suggested by others (10, 11, 12). On the other hand, the VH1–69/V{lambda}1–40 family pairing, originally found with scFv ICA5, over represented in our in-cell library (13), was obtained with only one scFv derived from the TPO+ random library, but the original VDJ-H/VJ-L pairing was never found in our random combinatorial libraries. It seems that the possibility to obtain recombinant antibodies similar to those produced in vivo is strongly dependent on the development of a large random library.

Among the 41 different human anti-TPO scFv fragments selected, 31 possessed a {lambda} light chain, including the 15 anti-TPO scFvs that showed natural H/L pairing. In contrast with the findings of Portolano et al. (7), who reported lower affinity for the majority of the {lambda} light chain anti-TPO Fabs than for the {kappa} light chain ones, our anti-TPO scFv fragments showed comparable affinities to TPO, independently of the chain pairing and the light chain type used. In accordance with studies describing the TPO antibody repertoire (6, 7, 19, 20, 21), most of the specific anti-TPO scFvs we produced were also derived from the VH1–3 germline gene in association with the V{lambda}1–40, V{lambda}1–51, or V{kappa}1–39 germline genes, indicating that a restriction in gene usage is found in thyroid-infiltrating B lymphocytes from patients with Graves’ disease. In addition, we derived scFvs encoded by germline genes VH3–30, VH3–64, VH5–51, V{lambda}1–44, V{lambda}2–8, V{lambda}7–43, V{kappa}1–5, and V{kappa}3–11 that have never been described as being implicated in the TPO antibody response. These results extend the repertoire of anti-TPO aAbs and enlarge our knowledge of genes encoding such antibodies. Autoantibodies with {lambda} light chains have been described in various autoimmune diseases (21, 22, 23, 24, 25), in particular {lambda} anti-TSHr autoantibodies are involved in thyroid stimulation in Graves’ disease (26, 27, 28). The role of {lambda} autoantibodies in thyroid diseases has recently been emphasized, because 5 anti-Tg antibodies close to the V{lambda}1–40 germline gene and 1 close to the V{lambda}1–51 germline gene have been isolated from a combinatorial library constructed from a patient with Hashimoto’s thyroiditis (29). Few {lambda}-TPO recombinant antibodies have been reported (7, 13, 21), probably because of the lack of information concerning V{lambda} germline genes leading to an inadequate choice of amplification primers and the low proportion of TPO autoantibodies present in the sera of patients with Graves’ disease (30, 31, 32, 33). Our data, obtained from an in-cell VH/VL assembled library (13) or with random libraries, indicate that anti-TPO scFvs derived from the V{lambda}1 gene family and particularly from the V{lambda}1–51 and V{lambda}1–40 germline genes represent one part of the in vivo TPO Ab repertoire of patients with autoimmune thyroid diseases.

We define nine antigenic regions on the TPO molecule by using the anti-TPO scFvs. With regard to the four antigenic domains described by using Fabs (2), it seems that the TPO molecule shows a more diverse set of Ab-binding regions. Furthermore, results obtained for epitope mapping of scFvs on TPO by using BIACORE methodology are difficult to compare with other methods used (2) for defining antigenic domains. Unlike Chazenbalk et al. (2), we did not observe any association between gene usage and epitope recognition by the anti-TPO scFvs. More precisely, scFvs with VL genes closest to the same germline gene, notably V{kappa}1–39, were found to recognize different regions on TPO, and, inversely, scFvs with VL genes closest to different germline genes, including V{kappa}1–39, were found to recognize the same epitope on TPO. However, other anti-TPO scFv fragments we obtained were encoded by newly described VL genes, which are not represented in the study by Chazenbalk et al. (2).

Among the three random combinatorial libraries we prepared, the library constructed from TPO-selected B cells (TPO+ library) showed the greatest clonal diversity. This is in accordance with Hawkins and Winter (34), who suggested that the construction of Ab fragments from antigen-selected cells gives rise to a more diverse repertoire. As surface Igs are not expressed on plasmocytes, our TPO+ library is probably composed of antibody fragments from other B cells and, notably, memory cells. On the other hand, the CD19 marker is present at all stages of B cell development, including plasma cells (35), suggesting that the CD19+ and unpurified B cell libraries are derived from the same B cell population. By comparing the CD19+ library with the previously reported in-cell library (13), the clonal diversity of the CD19+ random library is 5 times greater than that of the in-cell library. This difference can be explained by the fact that one single H (or L) chain can be associated with different L (or H) chains and still bind to TPO in a random combinatorial library. This cannot occur in the case of scFvs selected from an in-cell library, where the H/L pairing occurs within the cells.

Amino acid analysis of scFvs derived from the dominant germline gene VH1–3 clearly indicates that certain residue mutations are systematically found in the majority of anti-TPO recombinant antibodies, whatever the starting cells used for the library construction, but other amino acid replacements are closely related to a given library, suggesting that residue mutations could be a signature of the anti-TPO antibodies. Some of them have been described in TPOspecific Fabs (5, 6, 7, 29, 36), but others have never been demonstrated before. Epitope recognition does not appear to be linked to the starting cells, because the three scFvs showing strong inhibition of the autoantibody response to TPO belong to three different random libraries. These observations suggest that particular amino acid patterns in the VH region can be assigned to anti-TPO Abs, with a partial dependence on the library construction.

By competition studies between the anti-TPO scFvs and the serum aAbs from patients with Graves’ disease, we have shown that serum aAbs and anti-TPO scFvs recognize the same or closely related domains on TPO. Furthermore, some of our anti-TPO scFvs were able to inhibit more than 80% of the binding of some of the serum aAbs to TPO, independently of serum titer, demonstrating that these recombinant Ab represent the major part of the TPO Ab repertoire. This is in correlation with previous studies described for the anti-TPO Fabs (37) and suggests that our scFvs recognize epitopes similar to those recognized by the Fabs. It would be instructive to compare TPO antigenic regions probed by our scFvs and by recombinant antibodies (1, 2, 3, 4, 5, 6, 7) or monoclonal Abs whose epitope has been determined (38). Interestingly, a strong inhibition of the binding of patients’ sera to TPO was obtained by the random scFvs B4, T13 and in-cell scFv ICA5, which recognized three different antigenic domains on TPO, suggesting that these scFvs define immunodominant regions on TPO that might be identical to domains A and B (3). In addition, these scFvs use V{lambda}1–40 or V{lambda}1–51 genes in association with a VH1–3 gene, reflecting the in vivo pairing.

In conclusion, random H/L assortment libraries can give rise to scFvs that reflect the in vivo H/L situation and are similar to those obtained with the in-cell scFv library (13), which give rise to high affinity Ab fragments mimicking the binding of serum autoantibodies to TPO.


   Acknowledgments
 
We thank Dr. Sharon L. Salhi for critical comments and help in preparing the manuscript, and Drs. Isabelle Navarro-Teulon and Jean-Michel Portefaix for technical assistance with DNA sequencing. We also thank Drs. Line Baldet and Bernard Guerrier for providing the sera and the thyroid biopsies.


   Footnotes
 
This work was supported by in part by Elf Aquitaine, Bio-Rad Laboratories, Inc., and a fellowship from the Ministère de la Recherche et de l’Enseignement Superieur (to N.C.). The sequence data of the anti-TPO scFvs are available from the EMBL/GenBank/DNA database in Japan under accession numbers AJ399801 to AJ399839 for the heavy chain genes and AJ399840 to AJ399879 for the light chain genes.

Abbreviations: aAbs, Autoantibodies; Ab, antibodies; CDR, complementarity-determining region; FR, framework region; H, heavy chain; L, light chain; scFvs, single chain variable fragments; TIBL, thyroid-infiltrating B lymphocytes; TPO, thyroid peroxidase; TSHr, TSH receptor.

Received May 14, 2001.

Accepted for publication July 9, 2001.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Weetman AP, MacGregor AM 1994 Autoimmune thyroid disease: further developments in our understanding. Endocr Rev 15:788–830[Abstract]
  2. Chazenbalk GD, Portolano S, Russo D, Hutchison JS, Rapoport B, McLachan S 1993 Human organ-specific autoimmune disease. Molecular cloning and expression of an autoantibody gene repertoire for a major autoantigen reveals an antigenic immunodominant region and restricted immunoglobulin gene usage in the target organ. J Clin Invest 92:62–74
  3. McLachlan SM, Rapoport B 1995 Genetic and epitopic analysis of thyroid peroxidase (TPO) autoantibodies: markers of the human thyroid autoimmune response. Clin Exp Immunol 101:200–206[Medline]
  4. McIntosh RS, Asghar MS, Watson PF, Kemp EH, Weetman AP 1996 Cloning and analysis of IgG {kappa} and IgG {lambda} anti-thyroglobulin autoantibodies from a patient with Hashimoto’s thyroiditis: evidence for in vivo antigen-driven repertoire selection. J Immunol 157:927–935[Abstract]
  5. Portolano S, McLachlan SM, Rapoport B 1993 High affinity, thyroid-specific human autoantibodies displayed on the surface of filamentous phage use V genes similar to other autoantibodies. J Immunol 151:2839–2851[Abstract]
  6. Hexham JM, Partridge LJ, Furmaniak J, Peterson VB, Colls JC, Pegg C, Rees Smith B, Burton DR 1994 Cloning and characterisation of TPO autoantibodies using combinatorial phage display libraries. Autoimmunity 17:167–179[Medline]
  7. Portolano S, Prummel MF, Rapoport B, McLachlan SM 1995 Molecular cloning and characterization of human thyroid peroxidase autoantibodies of lambda light chain type. Mol Immunol 32:1157–1169[CrossRef][Medline]
  8. Gherardi E, Milstein C 1992 Original and artificial antibodies. Nature 357:201–202[CrossRef][Medline]
  9. Burton DR, Barbas III CF 1992 Antibodies from libraries. Nature 359:782–783
  10. Caton AJ, Koprowski H 1990 Influenza virus hemagglutinin-specific antibodies isolated from a combinatorial expression library are closely related to the immune response of the donor. Proc Natl Acad Sci USA 87:6450–6454[Abstract/Free Full Text]
  11. Portolano S, Chazenbalk GD, Hutchison JS, McLachan SM, Rapoport B 1993 Lack of promiscuity in autoantigen-specific H and L chain combinations as revealed by human H and L chain. J Immunol 150:880–887[Abstract]
  12. Costante G, Portolano S, Nishikawa T, Jaune JC, Chazenbalk GD, Rapoport B 1994 Recombinant thyroid peroxidase-specific autoantibodies. II. Role of individual heavy and light chains in determining epitope recognition. Endocrinology 135:25–30[Abstract]
  13. Chapal N, Peraldi-Roux S, Bresson D, Pugnière M, Mani JC, Granier C, Baldet L, Guerrier B, Pau B, Bouanani M 2000 Human anti-thyroid peroxidase scFv isolated from a combinatorial library assembled in-cell: insights into the in vivo situation. J Immunol 114:4162–4169
  14. Chapal N, Bouanani M, Embleton MJ, Navarro-Teulon I, Biard-Piechaczyk M, Pau B, Peraldi-Roux S 1997 In-cell assembly of scFv from human thyroid-infiltrating B cells. BioTechniques 23:518–524[Medline]
  15. Lehrer RI, Barton A, Ganz T 1988 Concurrent assessment of inner and outer membrane permeabilization and bacteriolysis in E. coli by multiple-wavelength spectrophotometry. J Immunol Methods 108:153–158[CrossRef][Medline]
  16. Lefranc MP, Giudicelli V, Ginestoux C, Bodmer J Muller W, Bontrop R, Lemaitre M, Malik A, Barbie V, Chaume D 1999 IMGT, the international ImMunoGeneTics database. Nucleic Acids Res 27:209–212[Abstract/Free Full Text]
  17. Munro S, Pelham HR 1986 An Hsp70-like protein in the ER: identity with the 78 kd glucose-regulated protein and the immunoglobulin heavy chain binding protein. Cell 46:291–300[CrossRef][Medline]
  18. Karlsson R, Falt A 1997 Experimental design for kinetic analysis of protein-protein interactions with surface plasmon resonance biosensors. J Immunol Methods 200:121–133[CrossRef][Medline]
  19. Rapoport B, Portolano S, McLachlan SM 1995 Combinatorial libraries: new insights into human organ-specific autoantibodies. Immunol Today. 16:43–49
  20. Chazenbalk GD, Costante G, Portolano S, McLachlan SM, Rapoport B 1993 The immunodominant region on human thyroid peroxidase recognized by autoantibodies does not contain the monoclonal antibody 47/c21 linear epitope. J Clin Endocrinol Metab 77:1715–1718[Abstract]
  21. Prummel MF, Portolano S, Costante G, Rapoport B, McLachlan SM 1994 Isolation and characterization of a monoclonal human thyroid peroxidase autoantibody of {lambda} light chain type. Mol Cell Endocrinol 102:161–166[CrossRef][Medline]
  22. Serrano MP, Cardona A, Vernet der Garabedian B, Bach JF, Pleaus JM 1994 Nucleotide sequences of variable regions of an human anti-acetylcholine receptor autoantibody derived from a myasthenic patient. Mol Immunol 31:413–417[CrossRef][Medline]
  23. Cairns E, Kwong PC, Misener V, Ip P, Bell DA, Siminovitch KA 1989 Analysis of variable region genes encoding a human anti-DNA antibody of normal origin. Implications for the molecular basis of human autoimmune responses. J Immunol 143:685–691[Abstract]
  24. Song YW, Kim HA, Lee EB, Baek HJ, Kim JM, Chung JH, Chung ES, Hong KM 1998 Molecular and genetic characterization of two anti-DNA antibodies derived from patients with systemic lupus erythematosus. Rheumatol Int 17:223–228[CrossRef][Medline]
  25. Ravirajan CT, Rahman MA, Papadaki L, Griffiths MH, Kalsi J, Martin AC, Ehrenstein MR, Latchman DS, Isenberg DA 1998 Genetic, structural and functional properties of an IgG DNA-binding monoclonal antibody from a lupus patient with nephritis. Eur J Immunol 28:339–350[CrossRef][Medline]
  26. Zakarija M, McKenzie JM 1983 Thyroid-stimulating antibody (TSAb) of Graves’ disease. Life Sci 32:31–44[CrossRef][Medline]
  27. Knight J, Laing P, Knight A, Adams D, Ling N 1986 Thyroid-stimulating autoantibodies usually contain only lambda-light chains: evidence for the "forbidden clone" theory. J Clin Endocrinol Metab 62:342–347[Abstract]
  28. Williams RC, Marshall NJ, Kilpatrick K, Montano J, Brickell PM, Goodall M, Ealey PA, Shine B, Weetman AP, Craig RK 1988 {kappa}/{lambda} Immunoglobulin distribution in Graves’ thyroid-stimulating antibodies. Simultaneous analysis of C lambda gene polymorphisms. J Clin Invest 82:1306–1312
  29. McIntosh R, Watson P, Weetman A 1998 Somatic hypermutation in autoimmune thyroid disease. Immunol Rev 162:219–231[CrossRef][Medline]
  30. Parkes AB, McLachlan SM, Bird P, Rees Smith B 1984 The distribution of microsomal and thyroglobulin antibody activity among the IgG subclasses. Clin Exp Immunol 57:239–243[Medline]
  31. Kotani T, Umeki K, Matsunaga S, Kato E, Ohtaki S 1986 Detection of autoantibodies to thyroid peroxidase in autoimmune thyroid diseases by micro-ELISA and immunoblotting. J Clin Endocrinol Metab 62:928–933[Abstract]
  32. Weetman AP, Black CM, Cohen SB, Tomlinson R, Banga JP, Reimer CB 1989 Affinity purification of IgG subclasses and the distribution of thyroid auto-antibody reactivity in Hashimoto’s thyroiditis. Scand J Immunol 30:73–82[CrossRef][Medline]
  33. Williams SC, Winter G 1993 Cloning and sequencing of human immunoglobulin V {lambda} gene segments. Eur J Immunol 23:1456–1461[Medline]
  34. Hawkins RE, Winter G 1992 Cell selection strategies for making antibodies from variable gene libraries: trapping the memory pool. Eur J Immunol 22:867–870[Medline]
  35. Weng WK, Shah N, O’Brien D, Van Ness B, LeBien TW 1997 Differential induction of DNA-binding activities following CD19 cross-linking in human B lineage cells. J Immunol 159:5502–5508[Abstract]
  36. McIntosh RS, Asghar MS, Kemp EH, Watson PF, Gardas A, Banga JP, Weetman AP 1997 Analysis of immunoglobulin G {kappa} antithyroid peroxidase antibodies from different tissues in Hashimoto’s thyroiditis. J Clin Endocrinol Metab 82:3818–3825[Abstract/Free Full Text]
  37. Nishikawa T, Costante G, Prummel MF, McLachlan SM, Rapoport B 1994 Recombinant thyroid peroxidase autoantibodies can be used for epitopic "fingerprinting" of thyroid peroxidase autoantibodies in the sera of individual patients. J Clin Endocrinol Metab 78:944–949[Abstract]
  38. Finke R, Seto P, Ruf J, Carayon P, Rapoport B 1991 Determination at the molecular level of a B-cell epitope on thyroid peroxidase likaly to be associated with autoimmune thyroid disease. J Clin Endocrinol Metab 73:919–921[Abstract]



This article has been cited by other articles:


Home page
 Int ImmunolHome page
S. A. Rebuffat, D. Bresson, B. Nguyen, and S. Peraldi-Roux
The key residues in the immunodominant region 353-363 of human thyroid peroxidase were identified
Int. Immunol., July 1, 2006; 18(7): 1091 - 1099.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Endocrinol. Metab.Home page
F. Latrofa, M. Phillips, B. Rapoport, and S. M. McLachlan
Human Monoclonal Thyroglobulin Autoantibodies: Epitopes and Immunoglobulin Genes
J. Clin. Endocrinol. Metab., October 1, 2004; 89(10): 5116 - 5123.
[Abstract] [Full Text] [PDF]

Home page
 J. Biol. Chem.Home page
D. Bresson, M. Cerutti, G. Devauchelle, M. Pugniere, F. Roquet, C. Bes, C. Bossard, T. Chardes, and S. Peraldi-Roux
Localization of the Discontinuous Immunodominant Region Recognized by Human Anti-thyroperoxidase Autoantibodies in Autoimmune Thyroid Diseases
J. Biol. Chem., March 7, 2003; 278(11): 9560 - 9569.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Endocrinol. Metab.Home page
F. Latrofa, P. Pichurin, J. Guo, B. Rapoport, and S. M. McLachlan
Thyroglobulin-Thyroperoxidase Autoantibodies Are Polyreactive, Not Bispecific: Analysis Using Human Monoclonal Autoantibodies
J. Clin. Endocrinol. Metab., January 1, 2003; 88(1): 371 - 378.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Chapal, N.
Right arrow Articles by Péraldi-Roux, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chapal, N.
Right arrow Articles by Péraldi-Roux, S.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Endocrinology Endocrine Reviews J. Clin. End. & Metab.
Molecular Endocrinology Recent Prog. Horm. Res. All Endocrine Journals