| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,§
*
Department of Medicine, University of Texas Health Science Center, San Antonio, TX 78284;
Departments of Medicine, Microbiology, and Immunology and Cancer Center, University of Rochester Medical Center, Rochester, NY 14642;
Department of Medicine, Arthritis and Immunology Program, Oklahoma Medical Research Foundation, University of Oklahoma, Oklahoma City, OK 73104; and
§
Department of Veterans Affair Medical Center, Oklahoma City, OK 73104
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Patient CS was a 52-year-old African-American female with long-standing SLE clinically characterized by World Health Organization type IV glomerulonephritis, photosensitive rash, and symmetrical nonerosive polyarthritis. Serologically the patient had high titer antinuclear Ab with strongly positive anti-dsDNA and anti-Sm Abs. At the time that serum and cell samples were obtained for this study, the patient was being treated only with hydroxychloroquine and nonsteroidal anti-inflammatory drugs. Previously, she had received six courses of monthly i.v. cyclophosphamide followed by 1 additional year of i.v. cyclophosphamide administered every 3 mo. Altogether, cyclophosphamide therapy had been discontinued 10 mo before samples were obtained. At that time, the patient’s blood cell counts were normal except for moderate absolute lymphopenia (950 lymphocytes/µl).
Construction of an IgG1/
combinatorial library
A combinatorial library of randomly assorted heavy chains of the
IgG1 isotype and light chains was constructed as previously
described according to established protocols (10).
Briefly, after informed consent was obtained, 50 ml of peripheral blood
was collected from our patient and mononuclear cells were isolated by
Ficoll-Hypaque gradient (Sigma, St. Louis, MO). Total RNA was extracted
from 2 x 107 cells and 5 µg of RNA was
reverse transcribed using RNaseH- reverse
transcriptase and random hexamer primers (Pharmacia, Piscataway, NJ).
The resulting ds-cDNA was used as template for the PCR amplification of
the Fd fragment of IgG1 heavy chains and full-length light chains.
Sense primers were designed to amplify all of the known human
VH and V families
(10) and contained appropriate restriction sites
(SacI for light chains and XhoI for heavy chains)
required for directional cloning in the phagemid pComb3 (kindly
provided by D. Burton and C. Barbas, The Scripps Research Foundation)
(11, 12). The restriction sites generated by the antisense
primers were SpeI for the heavy chain and XbaI
for the light chain.
Library screening and generation of soluble Fab fragments
The recombinant library was panned against an affinity-purified preparation of the Sm Ag containing both the B/B' and D polypeptides (ImmunoVision, Springdale, AR). Briefly, individual microtiter wells (Costar, Cambridge, MA) were coated overnight at 4°C with the Sm preparation at 20 µg/ml in 0.1 M sodium bicarbonate (pH 8.6), washed with PBS (pH 7.4), and then blocked with 3% BSA/PBS for 2 h at 37°C. Approximately 1011 PFU of the recombinant library were applied to each of the blocked wells, unbound phage was removed by vigorously washing 10 times with PBS/0.05% Tween 20, and the remaining phage was eluted with 50 µl of 0.1 M glycine/HCl (pH 2.2) and neutralized immediately with 2 M Tris. The eluate was used to reinfect XL-1Blue Escherichia coli cells followed by superinfection with VCS-M13 helper phage. This enrichment process was monitored by comparing the number of phage eluted in each round with the phage eluted in the previous round. After four rounds of panning, phagemid DNA was isolated from the rounds with higher enrichment and the sequence encoding the minor coat protein (cpIII) was removed by digesting with SpeI and NheI to produce soluble Fab fragments. After religation, phagemid DNA was used to transform XL-1 Blue cells. Randomly selected individual colonies were grown in Super Broth with 50 µg/ml carbenicillin and protein production was induced with 1 mM IPTG (Boehringer Mannheim, Indianapolis, IN). The supernatant from these cultures was used without further purification to determine anti-Sm reactivity by conventional ELISA as described below
Purification of recombinant Fab proteins
Ab Fab fragments were obtained from individual bacterial cultures induced with 1 mM IPTG. The Fab protein was purified from the bacterial pellets by affinity chromatography as previously described (10), and the concentration of purified Fab was estimated by sandwich ELISA using as standard a commercially available IgG Fab preparation of known concentration (Pierce, Rockford, IL).
Anti-Sm ELISAs
Dose-dependent reactivity of the recombinant Fabs with either the Sm-B/B'-D Ag or a panel of irrelevant control Ags was determined by conventional titration ELISA using 1 µg of the corresponding Ag/well of polysterene 96-well plates (Costar) as previously described (4). Both serum samples and mAbs were tested in duplicate. The ability of either Sm or various control Ags in fluid phase to block the binding of recombinant Abs to immobilized Sm was also tested by ELISA in competitive inhibition assays. Purified Fabs were used at 75% of the concentration previously determined to provide optimum binding in standard titration assays. The relative affinity constant of each Fab was calculated according to Friguet et al. (13) as the concentration of soluble competing Ag providing 50% inhibition of maximum binding to Sm. Competing Ags, including Sm itself, were tested at concentrations ranging from 10-6 to 10-10 M.
Epitope mapping
Overlapping octapeptides spanning the coding regions of Sm-B' and Sm-D1 were constructed on a solid-phase support system as previously described in detail (4). These overlapping octapeptides were simultaneously synthesized on radiation-derivatized polyethylene pins (Chiron Technologies, Clayton, Victoria, Australia). Positive and negative control pins were synthesized from known antigenic and nonantigenic regions of Sm B'.
Octapeptides were tested for reactivity with whole patient sera or
purified Fab fragments of various mAbs, including Fati-1, F-17, F-14,
and F-4. First, octapeptides were blocked with 3% low fat milk in PBS
for 1 h at room temperature. Second, primary Ab, either patient
sera (1:100 dilution) or purified Fab mAbs (at 
1 ng/well) in 3%
milk/PBS with 0.05% Tween 20 (PBST) were allowed to incubate overnight
at 4°C. Pins were then washed four times with PBST and then reacted
with a 1:1000 dilution of antihuman IgG (Fab')2
raised in a goat, affinity-purified and conjugated to alkaline
phosphatase (Pierce) for 2 h at room temperature.
p-Nitrophenyl phosphate disodium was used as a substrate for
alkaline phosphatase, and plates were read at 405 nm with a MicroELISA
Reader (Dynatech, Alexandria, VA). Results for each plate were then
standardized by comparison to positive control pins.
After completion of an assay, pins were sonicated for 2 h in sonication buffer (4), washed in hot water, and boiled in methanol for 2 min. Pins were then allowed to air dry and were either stored with desiccant or used for another assay (4).
DNA sequencing and analysis
Heavy and light chain nucleotide sequences were determined by
automated sequencing of both DNA strands using a Taq
fluorescent dideoxy terminator cycle sequencing kit
(Perkin-Elmer/Cetus, Norwalk, CT ). The sequencing was performed in an
Applied Biosystems PRISM 377 automatic sequencer (Perkin-Elmer/Cetus)
by the Oligonucleotide and DNA Sequencing Core Laboratory of the
University of Rochester. The germline counterparts of the rearranged
VH and V sequences were
determined on-line using the V-BASE search program
(http://www.mrc-cpe.cam.ac.uk; Medical Research Council Center for
Protein Engineering, Cambridge, U.K.), and the degree of similarity
between the corresponding sequences was established using the MegAlign
program of the DNAstar sequence analysis software (DNAstar, Madison,
WI). Complementarity-determining regions were assigned according to the
definition of Kabat (14). Ag-contact loops were defined
according to Chotia et al. (14).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As estimated by the efficiency of transformation of XL1-Blue
E. coli competent cells, the IgG1/ library made by in
vitro pairing of the Ab heavy and light chains expressed in the PBMCs
of SLE patient CS contained 2 x 107
primary clones. The representation of all VH
families in the library was ensured by the successful PCR amplification
obtained with all of the primer combinations described above and
confirmed by a combination of dot blot hybridization experiments using
VH family-specific oligonucleotide probes and
random sequencing of bacterial colonies isolated from consecutive
rounds of panning against the Sm Ag (results not shown).
Isolation and characterization of recombinant anti-Sm mAbs
A 20-fold increase in the titer of eluted phage was obtained after
three rounds of panning against Sm and an additional 10-fold increase
was observed after the fourth round, strongly suggesting a significant
enrichment in phage particles bearing specific anti-Sm Abs. To
further evaluate these clones, phagemid DNA purified from the last
round of panning was modified as described and used to generate
E. coli colonies that secreted soluble Fab fragments. Four
of 10 randomly picked colonies produced significant binding to Sm and
were selected for detailed analysis. The characteristics of the
recombinant Abs produced by these bacterial clones constitute the focus
of this paper. Fab fragments were purified from large-scale cultures
induced with IPTG and used for conventional titration ELISA to
determine binding against Sm and a panel of irrelevant control Ags
(Fig. 1
). All four Fabs demonstrated very
high binding against Sm in a dose-dependent fashion, with low binding
against other Ags and low background binding. The binding specificity
for the Sm Ag was also confirmed by competitive ELISA in which a set of
soluble Ags, including Sm itself, were used to compete against
immobilized Sm (Fig. 2). These
experiments show that only Sm produced significant inhibition as
compared with control Ags. Using the method of Friguet et al.
(13), the relative affinity constants of the recombinant
anti-Sm Abs was estimated in the range of 2 x
10-8 to 1 x
10-8 M.
|
|
The fine specificity of the recombinant Fabs was defined by ELISA
in which the Abs were tested against all of the possible overlapping
octapeptides derived from the amino acid sequence of the Sm-B/B' and
-D1 polypeptides (233 and 112 octapeptides, respectively). These assays
were also performed with serum samples obtained from patient CS
concomitantly with the cells used for the construction of the
combinatorial library. The results of these experiments are presented
in Figs. 3 and 4. Overall, the recombinant Fabs showed
very restricted specificity and reacted mainly with only 1 or 2 of the
345 octapeptides tested (the only exception, mAb F-14, will be
discussed below). As a rule, they recognized the same epitopes also
recognized by the patient’s serum Abs. Thus, mAb FATI-1 only
recognized the two major and earliest epitopes of the anti-Sm
response, the recurrent PPPGMRPP and the similar and overlapping
PPPGMRGP sequence with a reactivity of 5 and 14 SDs, respectively, over
background levels (Fig. 5). Despite its
monovalent nature, the reactivity of this Fab fragment with Sm epitopes
was equivalent to that of a whole anti-Sm mAb derived from
autoimmune mice, KSm5 (15). In turn, mAb F-17 exquisitely
recognized another octapeptide sequence (MAPPPGMR) but not the
overlapping PPPGMRPP octapeptide. Neither one of these mAbs showed any
reactivity with Sm-D octapeptides. The reactivity of the third and
fourth mAbs was concentrated on the Sm-D polypeptide. Interestingly,
both mAbs F-4 and F-14 recognized the same octapeptides 12–15 derived
from the amino terminus of the Sm-D molecule (spanning residues 12–22,
12HETVTIELKNG22) with a
reactivity >10 SD above the normal mean. A similar, partially
overlapping sequence (octapeptide 17,
17IELKNGTQ24) was also
recognized by the patient’s serum. In addition, F4 gave minor
reactivity (2–3 SD above the mean) with octapeptides 76–78
(76LDTILVDVEP85). This
sequence has partial overlap with the main peak obtained with the
patient’s serum (octapeptide 79,
79ILVDVEPK86). F-14
also recognized a carboxyl-terminal epitope which contains the GRG
repeat characteristic of this part of the Sm-D molecule (octapeptide
105, 105GRGRGRGR112)
(16). Interestingly, this mAb displayed some significant
degree of cross-reactivity with the region encompassed by overlapping
octapeptides 144–151 and 150–157
(144PQGRGTVAAAAAAA157) as
well as with multiple epitopes of Sm-B/B'.
|
|
|
DNA sequencing analysis of the genes encoding the variable regions
of both heavy and light chains revealed a remarkable restriction in the
use of VH genes. Thus, all four mAbs were encoded
by different members of the VH4 family
(VH4-30.1/31,VH4-34,
VH4-61, and VH4-59). In
contrast, the light chains were encoded by genes derived from the
V1 and V3 families.
Both heavy and light chain genes displayed characteristics of
Ag-selected somatic hypermutation with a degree of similarity with the
corresponding germline sequences ranging from 92 to 98%. A detailed
analysis of these sequences will be published
elsewhere.4
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
It should be emphasized that the epitope reactivity of the recombinant
mAbs closely resembled the one obtained with the patient’s polyclonal
serum (see Figs. 3
and 4). This fact strongly suggests that the
individual heavy and light chains used by these Abs had been selected
in vivo by the Sm Ag and most likely reflect the gene repertoire
expressed by the patient’s Sm-specific memory B cells. This conclusion
is consistent with the results of other investigators. Most
significantly, Barbas and colleagues (19) have used the
same technology to isolate anti-DNA mAbs from patients with SLE.
Yet, in the same study, the investigators failed to obtain anti-DNA
Abs from patients infected with the HIV despite the presence of some
anti-DNA reactivity in the patients’ sera. These results suggest
that phage-display libraries do not contain specific IgG autoantibodies
unless they are significantly represented in the in vivo repertoire.
This is important since combinatorial technology can create in vitro
artificial pairings of heavy and light chains potentially
misrepresenting the antigenic specificity of the actual repertoire. It
is entirely possible however that the specific H + L chain combinations
found in our Abs may not correspond to the ones originally expressed in
vivo, and indeed several studies have demonstrated that multiple light
chains can be paired with a heavy chain while retaining the original
specificity of the Ab (20). This issue could be partly
elucidated by H-L chain gene recombination experiments. However, a
final answer will only be provided by the isolation and analysis of
single anti-Sm Ab-producing B cells or plasma cells.
Only in one case did a mAb react with epitopes not recognized by the patient’s Abs. Thus, F-14 bound the sequence GRGRGRGR of Sm-D, which did not react with CS serum. This serum is exceptional in this regard since the great majority of lupus sera bind to the carboxyl-terminal GR repeat of Sm-D, whether or not they have anti-Sm autoantibodies (16, 21, 22, 23). The classic murine anti-Sm monoclonal autoantibody Y12 is another example of binding to this GR repeat peptide (23, 24). Interestingly, it has been shown that Y12 also recognizes Sm-B/B' (15, 23, 24). It is possible that the presence of the GRG motif within the B/B' sequence recognized by F-14 (144PQGRGTVAAAAAAA157) could contribute to this cross-reactivity. Interestingly, another murine mAb, ANA125, recognizes a similar stretch of the B/B' sequence, 146GRGTVAAAAAAAT158 (22). These results could be explained by postulating that F-14 might be encoded in vivo by a B cell clone that had not yet broken tolerance in the patient under study and, thus, was part of the "dormant" autoimmune repertoire but was not yet expressed in the actual repertoire of activated B cells at the time of sampling. In this scenario, in vitro cloning and expression would have overcome in vivo anergy. Alternatively, F-14 reactivity might have gone undetected in total serum if it represented a minor fraction of the patient’s secreted autoantibody repertoire, perhaps by having been absent from the plasma cell compartment at the time of the study.
The genetic restriction encountered in the heavy chain used by the anti-Sm Abs represents another remarkable feature of our study. Thus, all of the heavy chains were encoded by genes derived from the VH4 family including one (F-17) encoded by the VH4-34 gene which is otherwise found in 100% of pathogenic cold agglutinins and in a large variety of other autoantibodies but not in conventional, protective Abs (25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35). This restriction suggests that either a triggering event or an amplification mechanism appears to preferentially recruit VH4-encoded anti-Sm autoantibodies into the pathogenic lupus repertoire. Possible mechanisms that could explain this observation include a superantigen effect of the Sm Ag or related Ags and/or an enhanced pathogenic potential of VH4-encoded autoantibodies. In summary, we report the detailed characterization of the first panel of human anti-Sm IgG monoclonal autoantibodies ever generated from an SLE patient. Remarkable aspects of this work include the fact that the Abs were generated by phage-display technology from PBLs and that the epitope recognition profile of the recombinant autoantibodies overlapped significantly with the in vivo activity of the patients’ Abs. This suggests that this technology can indeed recapitulate in vitro the Ab repertoire expressed by recirculating memory B cells and validates the use of this approach while indicating that the availability of tissues enriched in memory B cells or plasma cells is not an absolute requirement for this type of study. The ability to generate relevant monoclonal autoantibodies with this approach should allow investigators to define the earliest antigenic epitopes targeted by autoimmune responses as well as to understand the genetic and structural basis of pathogenic autoantibody responses and of epitope spreading.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Iñaki Sanz, Clinical Immunology and Rheumatology Unit, University of Rochester Medical Center, 601 Elmwood Avenue, Rochester, NY 14642.
3 Abbreviations used in this paper; Sm, Smith; SLE, systemic lupus erythematosus.
4 I. del Rincon, M. Zeidel, E. Rey, M. Fischbach, and I. Sanz. Genetic analysis of SLE anti-Smith monoclonal antibodies reveals restriction to the VH4 family. Submitted for publication.
Received for publication July 3, 2000. Accepted for publication September 21, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Seagal, E. Edry, H. Naftali, and D. Melamed Generation and selection of an IgG-driven autoimmune repertoire during B-lymphopoiesis in Ig{micro}-deficient/lpr mice Int. Immunol., July 1, 2004; 16(7): 905 - 913. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 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 | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
