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 Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Freed, J. H. Articles by Eisenberg, R. A. Search for Related Content PubMed PubMed Citation Articles by Freed, J. H. Articles by Eisenberg, R. A.

MHC Class II-Bound Self Peptides from Autoimmune MRL/lpr Mice Reveal Potential T Cell Epitopes for Autoantibody Production in Murine Systemic Lupus Erythematosus¹

John H. Freed^2,*,, Amy Marrs^*, Jennifer VanderWall^*, Philip L. Cohen and Robert A. Eisenberg

^* Division of Basic Immunology, Department of Medicine, National Jewish Medical and Research Center, Denver, CO 80206; Department of Immunology, University of Colorado Health Sciences Center, Denver, CO 80262; and Division of Rheumatology, University of Pennsylvania School of Medicine, Philadelphia, PA 19104

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
The systemic lupus erythematosus-like syndrome in MRL/lpr mice involves high-titered IgG autoantibodies, particularly antinuclear Abs that target histones, DNA, and RNA particles. Although T cell help is required for the generation of antinuclear Abs, the epitopes recognized by such helper T cells are unknown. To address this question, we isolated and sequenced self peptides bound by MHC class II molecules from MRL/lpr mice. We identified a number of peptides that are not seen in similar preparations from nonautoimmune C3H animals. The "abnormal" peptide donors include histone, a protein component of a small nuclear ribonucleoprotein, ribosomal proteins, and RNA processing enzymes. We postulate that the peptides from these donors are T cell epitopes required for the generation of the most frequent antinuclear Abs specificities seen in MRL/lpr mice.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Systemic lupus erythematosus (SLE)³ is an autoimmune disease characterized by the production of autoantibodies of multiple specificities, some of which are known to cause immune-mediated damage. Although the disease is heterogeneous in nature and the clinical manifestations include the involvement of multiple organ systems, the hallmark of the disease remains the production of antinuclear Abs that are seen in 95–99% of all patients. Although it is clear that some of the pathology of SLE may be attributed to these autoantibodies, the underlying etiology of the disease remains obscure. One popular model (1) postulates that the disease arises in genetically susceptible individuals who are exposed to as-yet-unknown environmental trigger(s), although in the MRL/lpr mouse model exogenous Ags are not necessary (2). The central feature of this model is an autoantigen-specific CD4 T cell driving force of unknown specificity that, most likely in combination with other cellular defects, is required for the production of the pathogenic IgG molecules. The production of IgG requires cognate recognition of Ag, which raises the question of the natures of the B cell and T cell epitopes involved in the generation of each of the autoantibodies. It has been possible to determine the antigenic specificity of the autoantibodies produced and thereby understand the B cell epitopes. The T cell epitopes, however, have been more elusive, because there is no direct readout of their specific recognition. TCRs do not recognize intact Ag directly, as do the surface Ig receptors of B cells; rather, the TCR of the helper T cells bind to peptide fragments of the Ag that are bound by self MHC class II molecules of APC.

Among the commonly used mouse models for lupus, the MRL/MpJ-Fas^lpr (MRL/lpr) strain is particularly relevant because it produces a number of SLE-specific autoantibodies (3). These mice are homozygous for the H-2^k haplotype, a haplotype shared by several other strains, such as C3H, that do not develop SLE. Most notably, MRL/lpr animals have a mutation in their Fas (CD95 gene) that produces striking, but nonmalignant, lymphoproliferation. This Fas defect dramatically accelerates the autoimmune state inherent to the MRL strain (1, 3, 4). MRL/lpr mice develop pathogenic IgG autoantibodies, strikingly similar in specificity to the autoantibodies seen in human lupus (3), and die of glomerulonephritis at an early age, with 50% mortality by 21 wk of age (4).

The use of bone marrow chimeras allowed the demonstration of intrinsic defects in both T cells (5) and B cells (6) of lpr mice. Subsequent studies that employed a similar approach revealed that the production of autoantibodies in lpr animals required the cognate, MHC-restricted interaction between lpr T and B cells (7). Central to such a process is the recognition of complex Ag (and Ag complexes) by both B and T cells (7). This concept has been applied to the generation of anti-DNA autoantibodies (8, 9, 10). In these models, nucleosomes are taken into the B cells by virtue of reactivity of the DNA-specific Ig receptor of the autoreactive B cell; the entire nucleosomal complex is processed; and peptides, presumably derived from the histone component of the nucleosome, are bound and presented by the MHC class II molecules of the B cell (9). The autoreactive CD4⁺ helper T cell then interacts with the peptide-class II complex and delivers the appropriate helper signals that allow the B cell to proliferate and differentiate into an autoantibody-secreting plasma cell.

Two recent reports provide indirect support for the role of B cells as APC in the generation of the autoimmune state. The first demonstrates that when MHC class II-deficient MRL/lpr mice were created, they failed to produce significant levels of autoantibodies or to develop glomerulonephritis, although they still exhibited lymphadenopathy (11). This would suggest that, in the absence of the class II MHC structure required to present peptides and therefore initiate T cell help, the B cells cannot make the pathogenic IgG autoantibodies despite whatever other defects might be inherent in the autoimmune B cells. The second study demonstrated that MRL/lpr mice that are genetically engineered to lack all B cells not only have the obvious lack of autoantibodies, but also lack all T cell-mediated manifestations of the disease (12). This, likewise, implies that cognate recognition of Ag and its presentation by B cells to T cells lies at the center of the autoimmune process.

Despite the presumed importance of cognate recognition of autoantigen by T cells and thus the presentation of peptides as T cell epitopes in SLE-prone mice, the nature and specificity of such autoreactive T cells has been difficult to ascertain. Some progress has been made by using autoimmune T cells specific for nuclear Ags to begin to define T cell epitopes expressed on histones (13), as well as on topoisomerase I (14) and Sm (15). Although each of these studies has provided information about T cell epitopes in autoimmune disease, the approach as a whole has permitted analysis of only a small number of epitopes due, in part, to the difficulty of isolating Ag-reactive T cell clones and to the labor-intensive nature of the experiments required to define the exact peptide that represents the T cell epitope. We reasoned that direct biochemical inspection of the MHC class II-bound self peptides expressed by sick SLE animals might reveal a broader spectrum of potential T cell epitopes and, thus, provide initial insight into the specificities of the CD4⁺ T cells generally considered to be the driving force for lupus.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Cell sources, Abs, and chemicals

Mice of the following strains were used: MRL-/lpr, MRL-lpr-Thy1.1, and MRL-lpr-Igh^b. All animals were bred and maintained in the animal colony at the University of North Carolina (Chapel Hill). Animals (96 total) were sacrificed, and single cell suspensions of pooled lymph nodes (from all three strains) were pelleted by centrifugation and frozen at -80°C for transport to Denver. The anti-I-A^k mAb 17/227 (16), the anti-I-E^k mAb 14-4-4S (17), and the anti-I-A^d mAb MK-D6 (18) were produced at National Jewish Medical and Research Center from hybridoma culture supernatants and purified by passage over protein A-agarose (Sigma, St. Louis, MO). All chemicals, unless noted otherwise were obtained from Sigma.

Affinity purification of class II molecules, and extraction and fractionation of bound peptides

Frozen lymph node pellets (2.7 x 10¹¹ cells) were lysed in 1% Nonidet P-40 (NP-40) and 0.01 M iodoacetamide in Tris-buffered saline (TBS: 0.01 M Tris and 0.14 M NaCl (pH 7.4)) at 1 x 10⁸ cells/ml for 1.25 h at 0°C. The lysate was centrifuged at 100,000 x g for 1 h, and the clarified supernatant was combined with a similar supernatant prepared from 1 x 10⁸ fresh MRL/lpr lymph node cells that had been surfaced labeled with 1 mCi NA¹²⁵I-using Iodogen (Pierce, Rockford, IL). The supernatants were run over three mAb affinity columns, each prepared by coupling 30 mg mAb to 10 ml of cyanogen bromide-activated agarose. The supernatant was divided into three equal portions, and each portion was passed over the three columns in a different sequence to equalize nonspecific sticking of molecules to each column. The columns were washed separately with 0.25% NP-40 in TBS followed by 1% octylglucoside (n-octyl-ß-D-glucopyranoside) in TBS. The bound class II molecules were eluted from each column separately with 0.05 M diethylamine and 1% octylglucoside (pH 11.0), and the fractions were immediately neutralized with Tris (pH 6.5) and 1% octylglucoside. Fractions were pooled based on the ¹²⁵I profiles.

Pooled fractions were concentrated using Centricon 30 concentrators (Amicon, Beverly, MA) that had been preincubated three times with 2.5 M acetic acid for 30 min at 37°C and then neutralized. The concentrates were diluted to 2 ml with 0.01 M Tris (pH 7.5) and were reconcentrated. Peptides were extracted from the class II molecules by incubating the concentrates in 2.5 M acetic acid for 30 min at 37°C. Each entire mixture was transferred to a prewashed Centricon 10 concentrator, and the peptides were collected in the filtrate and taken to dryness in a Speed-Vac (Savant/EC, Holbrook, NY). The resultant pellets were stored at -80°C until needed.

Each thawed pellet was dissolved in 100 µl of 0.1% trifluoroacetic acid (TFA) in water and applied to a narrow bore Vydac C₈ reverse-phase HPLC column (The Separations Group, Hesperia, CA) at 250 µl/min. Peptides were separated with a 5 min isocratic elution (0.1% TFA in H₂O), followed by a 60-min linear gradient (0.1% TFA in H₂O to 100% of 0.08% TFA in acetonitrile). Peptides were monitored using absorbance at 214 nm. Fractions (130) were collected by time (0.5 min) starting at the injection of the mixture. Individual fractions were stored at -80°C until needed for sequence analysis.

Edman sequencing and matrix-assisted laser desorption ionization time of flight (MALDI-TOF) mass spectrometry

N-terminal sequences of individual fractions were analyzed by diluting 85 µl of the sample with two volumes of water and applying the diluted sample to a polyvinylidene difluoride (PVDF) disk (Pro-Sorb, PE-Applied Biosystems, Foster City, CA). The sample was sequenced using the "Pulsed Liquid PVDF Protein" program on an Applied Biosystems Procise 492 sequencer (PE-Applied Biosystems). Sequence assignments were made manually by overlaying chromatograms using Applied Biosystems model 610A software. Most fractions contained multiple peptides producing multiple amino acids at each cycle. The estimated relative yields of each amino acid were used to assign an amino acid to the "major" or "minor" sequence for that fraction. The most probable major sequence was compared with sequences in the nonredundant database at National Center for Biotechnology Information (NCBI) using the BLAST search engine with an ungapped BLOSUM62 homology matrix. When a match was found, the known sequence was used to refine the assignment of residues to the major sequence. The process was repeated using the remaining amino acids at each step forming the minor sequence. For a peptide to be assigned as being derived from a specific protein donor, no more than a single mismatched amino acid was permitted when the relevant mouse protein was in the NCBI database. (Because Edman sequencing cannot detect underivatized cysteine or tryptophan in low yield, mismatches at these positions were not counted as significant.) When the mouse sequence was not available and the alignment had to be made to a rat or human sequence, then up to two to three mismatches were tolerated as long as the substitutions were mostly conservative. Once a protein donor was assigned, the absolute yield of the peptide derived from that donor was determined by using the yields of three to four residues (established by comparison with a 10 pmol standard run as part of the sequence) to calculate an initial yield for the peptide.

If a protein donor for a peptide could be identified, then the remaining sample was submitted to mass determination using a PerSeptive Biosystems/Perkin-Elmer Voyager-DE/RP MALDI-TOF instrument with -cyano-cinnamic acid as the matrix. Mass peaks were compared with calculated masses for the known peptide donor protein (as determined by Edman sequencing) and those matching were used to establish the C termini, and possible presence of nested sets, of peptides derived from that donor. NP-40 carryover from the affinity chromatography purification prevented mass spectrometric analysis on a number of fractions in all three runs.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Isolation of MHC class II molecules from MRL/lpr lymph nodes and sequence determination of bound peptides

MRL/lpr animals with marked lymphadenopathy (12–16 wk of age) were used as the source of material for analysis. Cells from three different congenic strains (see Materials and Methods), all on the MRL/lpr background, were combined; the cells will be designated as MRL/lpr. Frozen pellets containing an estimated 2.7 x 10¹¹ lymphocytes, derived from a pool of 96 animals, were lysed with 1% NP-40. Because lymph nodes from diseased MRL/lpr mice are estimated to contain 95% T cells, only 1.3 x 10¹⁰ cells were estimated to be B cells and these were presumably the primary source of the class II molecules analyzed. The lysate was passed over mAb affinity columns to isolate I-A^k and I-E^k molecules. Class II-containing fractions were combined and concentrated; peptides were eluted from the class II molecules with acetic acid and collected by passage through a 10-kDa cut-off membrane. Each set of peptides was fractionated by reverse-phase HPLC (Fig. 1). The profiles of the I-A^k-bound peptides and the I-E^k-bound peptides were separately compared with the control profile of peptides eluted from material passed over an irrelevant (anti-I-A^d) affinity column. Peaks that were found in the specific profiles and absent from the control profile were sequenced. Peaks eluting before 25 min and after 47 min either were also found in the control run or did not yield sequence information (data not shown). Not all peaks contained material that yielded sequence (see, for example, the large peak at 35 min in the I-A^k profile), and the yield of peptide did not always correlate with the size of the peak.

View larger version (24K): [in this window] [in a new window]   FIGURE 1. Separation of class II-bound self peptides from MRL/lpr mice by reverse-phase HPLC. A detergent lysate of lymph node cells from sick MRL/lpr animals was passed over anti-I-Ak, anti-I-Ek, and control (anti-I-Ad) affinity chromatography columns. The eluted class II molecules were treated with acetic acid to remove bound peptides, which were collected by passage through a 10-kDa cut-off membrane. Peptides from each column were separated by reverse-phase HPLC on a C18 column eluted by 5 min of isocratic conditions (0.1% TFA in H2O), followed by a linear gradient (0.1% TFA in H2O to 100% of 0.08% TFA in acetonitrile) over 60 min. Fractions (0.5 min) were collected beginning at the time of injection. The small numbers indicate the peaks that gave identifiable, allele-specific sequences, and the "X" indicates a peak that contained a peptide bound nonspecifically to all three affinity columns.

Analysis of peptides bound by the I-A^k molecules isolated from MRL/lpr lymph nodes (Table I) revealed that peptide yields ranged from 1.6 pmol for the 26S proteasome p112 peptide to 73.5 pmol for the transferrin peptides beginning with residue 119 and found in peak 1 (see Fig. 1). When all forms of peptide from a single protein donor were combined, transferrin-derived peptides accounted for 107.2 pmol, A_ß^k-derived peptides for 153.9 pmol, and lysozyme-derived peptides for 22.1 pmol. Most of the protein donors of the peptides listed in Table I are molecules expected to have access to the class II processing pathway (19). These include the extracellular proteins transferrin and the 1-chain of IgG1 (although the IgG1 peptide could be derived from the membrane form as well); the plasma membrane protein and class II component A_ß^k; and the lysosomal proteins lysozyme and saposin D. However, in addition, several protein donors were identified that are found in compartments that normally do not have access to the class II processing pathway: nucleoporin NUP155, 14-3-3 protein, and the 26S proteasome p112 protein. It is also apparent from Table I that the isolated peptides have "ragged" N and C termini, presumably reflecting variable degrees of processing by endosomal exopeptideases (20, 21).

Each fraction in the control run that corresponded to a fraction containing a sequence shown in Tables I and II was also sequenced by Edman sequencing. The vast majority of these gave no sequence or produced a weak sequence that could not be identified, either by comparison with the corresponding I-A^k-binding and I-E^k-binding sequences or by searching the NCBI database. Three fractions, while not yielding a coherent sequence, did have a number (5 each) of widely scattered residues that matched the corresponding positions in the I-E^k-bound peptides from ferritin L chain and the 60S ribosomal proteins L36 and L14. Two alternative explanations are possible for this: first, because multiple amino acids occur in the control run at almost every cycle, including the cycles where the match to the I-E^k-binding peptides occurred, the match could be random and not due to a specific peptide in the control sample. Second, the matches could indicate that the peptides are indeed found in the control runs, but in amounts too low to yield a bona fide sequence. Although we favor the first possibility, accepting the second possibility does not undermine significantly the specific binding of the peptides by the I-E^k molecules, especially for the ferritin L chain peptide and the 60S ribosomal protein L14 peptide, both of which fit the I-E^k-binding motif (see below).

The two Ig-derived peptides (1 for I-A^k and 2b for I-E^k) were not found in the control run; nevertheless, because the samples were purified by affinity chromatography, it is possible that the peptides were derived from the mAbs used in the affinity chromatography step. We can rule this out, however, because all three mAbs used in the affinity chromatography step were IgG2a and the corresponding 2a sequence differs from both the 1 and 2b peptides found in the HPLC runs.

Finally, we obtained a sequence for the fraction marked with an "X" in Fig. 1. The same sequence was found both in the I-A^k and I-E^k runs and in the control run, all in roughly comparable yields. This peptide clearly represents a nonspecific peptide and accordingly lacked the binding motifs for both the I-A^k and I-E^k molecules. (The sequence determined was MRAKWRKKRMRRLKRKRRKM; this highly basic sequence is derived from the N-terminal 80% of the 25-aa 60S ribosomal protein L41.) It was not included in either Table I or II.

Peptides that are bound by I-A^k and I-E^k fit the established motifs for these molecules

The binding motif for the I-A^k molecule has recently been described by Fremont et al. (22). In this motif, the index position (position 1, P1) must be aspartic acid (D), although asparagine (N) has been described in a few cases. P4 is often hydrophobic (V, I, L) or asparagine (N). P6 is usually E or Q, although T and G have been found as well. Finally, P9 is polar (S, T, Q) or small (G or A). As indicated, the side chain at P1 is critical and residues acceptable at P6 appear to be more restricted than for P4 or P9. Comparison of Tables I and III reveals that 11 of the 14 peptides possess the critical D or N at the P1 position for I-A^k motif. Of the 11 peptides with D or N at P1, 7 (including all four variants of the transferrin peptide) fit the motif exactly; 2 fit at 3 of 4 anchor positions; and 2 have structurally conservative substitutions at the non-P1 anchor positions. Of the three peptides that do not possess the motif, A_ß^k (143-) has a D that can be assigned to P1 but lacks the other anchor residues, whereas the 1 and proteasome peptides lack the motif-specified P1 residue but have motif-compatible residues at P4, P6, and P9. It should be pointed out that other studies have found peptides bound by I-A^k that lack the motif (23).

The binding motif for the I-E^k molecule has been known for some time based both on the analysis of peptides bound by I-E^k (see Ref. 23 for a summary) and on the solved crystal structure (24). P1 is an amino acid with a hydrophic side chain (usually I, L, or V, but also F, Y, or W); P4 is also hydrophobic (I, L, V, F) or S; P6 is usually polar (Q, N, E) or A; and P9 (or occasionally P10) is R or usually K. The P1 and P9 positions appear to exert a greater role in controlling binding specificity than do P4 and P6 (24). As may be seen by comparing Tables II and IV, 7 of 17 of the peptides bound by the I-E^k molecules isolated from MRL/lpr animals exhibit this exact motif; the peptide derived from histone H2A2.2 differs only at P4; the peptides derived from axin and the 60S ribosomal protein L15 differ only at P9; the ferritin L-chain-derived peptide differs at both P4 and P6; and the peptide derived from the 2b-chain differs at P1 and P6. The peptide derived from RNA editase does not possess the motif.

Comparison of peptides bound by class II molecules in MRL/lpr mice with those bound in C3H mice

One of us (J.H.F.) has previously analyzed peptides bound by the class II molecules of nonautoimmune C3H/HeJ mice that share the H-2^k haplotype with MRL/lpr animals (25). We compared peptides bound by class II molecules in the two strains. As revealed in Table V, significant differences in relative abundance of peptides were found in the two strains. In the C3H animals, the transferrin (119–) peptide accounted for only 7.8% of the recovered peptides, whereas in the MRL/lpr animals, this peptide was the major peptide (23.4%). In addition, new N-terminal variants (those beginning at positions 120 and 121) of this peptide were found in MRL/lpr compared with C3H animals. These bring the overall yield of all forms of transferrin peptides in MRL/lpr to 34.1% compared with only 11.1% in C3H. Although the overall yield of A_ß^k-derived peptides was roughly comparable in the two strains (42.8% in C3H vs 49.0% in MRL/lpr), the specific peptides (and specific yields) found in the two strains differed widely. Thus, the A_ß^k (37–) peptide accounted for 28.0% of the peptides in C3H and only 4.6% of the peptides in MRL/lpr. Furthermore, C3H animals formed the A_ß^k (3–) and A_ß^k (10–) peptides in modest yield, while these peptides were absent from MRL/lpr animals, having been replaced by the high yield peptides A_ß^k (110–) and A_ß^k (143–).

Comparison of the peptides bound by the I-E^k molecules from the two strains revealed the same general patterns (Table VI). A clusterin-derived peptide, which is relatively minor in C3H (13.9%), was the major peptide (46.0%) in MRL/lpr, whereas the peptide derived from ß₂-microglobulin was reduced from 63.5% in C3H to 8.5% in MRL/lpr, and the HSC70 peptide was decreased in yield from 14.6% to 3.8% when comparing C3H to MRL/lpr. A serum albumin-derived peptide was found only in C3H. Of particular significance was the large number of peptides found only in I-E^k isolated from MRL/lpr animals. Not only were the majority of these derived from proteins that would not normally have access to the class II pathway, but many of them are candidates for T cell determinants for IgG isotype autoantibody formation (see below).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The abundance of transferrin-derived peptides eluted from I-A^k molecules is of considerable interest. Although not readily detectable in serum because of the high level of circulating transferrin, autoantibodies to this protein can be detected easily at the single-cell level in spleen cell suspensions from MRL/lpr mice by using a hemolytic plaque assay with transferrin-coated erythrocytes (28). It is possible that transferrin-specific T cells recognize the eluted peptide and provide help for anti-transferrin autoantibodies.

The most striking aspect of the self peptides found in the MRL/lpr animals, but not in the C3H animals, is that a large proportion of the peptides bound by the I-E^k molecules could plausibly function as T cell epitopes for the cognate interactions that are thought to lead to the production of the pathogenic IgG autoantibodies. In these interactions, autoantibody on the autoreactive B cell would facilitate uptake of complexes of nucleic acid and protein (e.g., nucleosomes or ribosomes) by reacting with the nucleic acid portion of the complex (i.e., DNA or RNA, respectively). Normal processing of the complex in the class II Ag processing pathway would result in display of peptides derived from the protein component(s) of the complex (i.e., histone or ribosomal proteins). These peptides would be recognized by CD4⁺ autoreactive T cells that provide help to the autoreactive B cells and drive the production of pathogenic IgG. In a simpler variation of this model, proteins may contain both B cell and T cell epitopes, so that B cell receptor-mediated uptake of such proteins would produce the T cell epitopes for the autoreactive T cells.

Table VII lists the major autoantibody specificities found in MRL/lpr mice and correlates these with proteins that serve as donors of peptides that represent potential epitopes for autoreactive T cells. Histone proteins, as components of the nucleosome, can provide epitopes that serve as cognate T cell determinants for autoantibodies to DNA and to the histones themselves. In addition, the autoantibody specificity Su is a determinant on a macromolecular complex of non-histone nucleosomal proteins (29). Thus, it is reasonable to propose that the peptide derived from the histone H2A.2 molecule can serve as the T cell epitope for anti-histone, anti-DNA, and anti-Su autoantibodies seen in MRL/lpr mice. This model is supported by the recent demonstration in the (SWR x NZB)F₁ model of SLE that i.v. injection of soluble synthetic histone peptides can suppress the formation of autoantibodies to dsDNA, ssDNA, histones, and non-histone (HMG) proteins (30). Finally, RNA polymerase I is the nucleolar polymerase responsible for transcription of ribosomal RNA. The RNA polymerase I transcription-initiation complex consists of the core histones (H2A, H2B, H3, and H4) and an upstream transactivator complex known as UBF (31). Thus, peptides derived from histone H2A.2 can serve as T cell epitopes for autoantibodies directed against RNA polymerase I.

Our results are the first demonstration of a potential basis for T cell help in the anti-RNA and anti-ribosome autoantibody responses. The NOD mouse model for autoimmune insulin-dependent diabetes mellitus (IDDM) provides an interesting comparison, because self peptides eluted from the I-A^g7 molecules obtained from spleens of NOD animals contained heterogeneous nuclear RNP (hnRNP)-derived peptides (35). The hnRNP proteins, like several proteins in the current study (especially SRp20), are RNA-binding proteins. This observations may suggest an unanticipated correlation between class II-bound peptides derived from RNA-binding proteins and autoimmune disease.

The autoantibody specificity Sm is an epitope on several of the core proteins that make up the protein components of the U1 (and U2, U4/U6, and U5) small nuclear RNPs (snRNPs) (36). Because SRp20 and U1 snRNPs have been demonstrated to be components of a macromolecular complex that regulates the splicing and generation of alternative polyadenylation sites in pre-mRNAs (37), it is possible that SRp20 can serve as a donor of peptides for the T cell epitope for the generation of anti-Sm autoantibodies. The observations that SRp20, and not the U1 proteins, contains sequence information necessary and sufficient for localization to nuclear speckles (38) and that anti-Sm Abs stain nuclear speckles (39, 40) support our model that SRp20 and the Sm-containing U1 components exist as a complex capable of providing determinants for both autoreactive T cell and B cell recognition.

Finally, the presence of IgG H-chain-derived peptides (bound by both I-A^k and I-E^k) provides a demonstration of potential T cell determinants for the generation of rheumatoid factors in MRL/lpr animals. Alternatively, recognition of Ig sequences by autoreactive T cells could provide noncognate help for any autoreactive B cell that might present its own Ig.

We did not find peptides to serve as potential T cell epitopes for two other autoantibody specificities, i.e., gp70 and collagen. However, it is possible that peptides from the proteins were bound by I-E^k and/or I-A^k molecules in MRL/lpr animals, but at levels below our limits of detection. Analysis of self peptides with currently available technology allows identification of only a handful of the most abundant of the estimated 2000 peptides bound by a single MHC allele (41).

It should be stressed that, in the absence of functional data, the peptides we describe must be considered potential T cell autoantigenic peptides, despite our ability to rationalize a role for most of the "abnormal" peptides that are bound by the I-E^k molecule. In future studies we will attempt to demonstrate the presence of T cells reactive with these peptides in MRL/lpr mice and will further attempt to use these peptides to modulate disease, as has recently been described for histone peptides in the (SWR x NZB)F₁ model for SLE (30).

It is interesting that many of the self peptides bound by both I-A^k and I-E^k molecules in the MRL/lpr animals are derived from donor proteins from the cytosol and nucleus that normally should not have access to the class II processing pathway. It is theoretically possible that this "cross-talk" between the class I and class II Ag processing pathways could result from alterations in the class II pathway caused somehow by the "hyperresponsive" nature of the B cells in the MRL/lpr animals. Similarly, the hyperresponsive state of the B cells in MRL/lpr animals could explain differences in peptides from the same donor protein in the two strains. Although this is a plausible explanation, no data currently available demonstrate activation-state dependent alterations in Ag processing in B cells of the type required to explain the results in the current study.

A second possibility is that the differences in peptides reflect differences between the organs used as the source of B cells in the two strains, i.e., spleen in C3H and lymph nodes in MRL/lpr animals. However, our previous studies (25) demonstrated essentially identical peptides in preparations isolated from thymus and spleen in nonautoimmune animals and also revealed no peptides that were abnormal for the class II processing pathway. Therefore, we feel it is unlikely that splenic and lymph node B cells should express different peptides. Although organ-specific differences in the expression of cathepsins have been reported (42), this seems to be more related to cell type than to the organ per se and does not appear to be a sufficient difference to account for the large numbers of cytosolic and nuclear proteins that provide class II-bound peptides in the MRL/lpr animals. It could possibly explain how different peptides are derived from the same donor protein in the two strains, however.

A third possible mechanism for explaining the existence of peptides derived from cytosolic and nuclear proteins in the MRL/lpr class II molecules is the possibility that cytosolic macromolecular complexes (e.g., the proteasome) or organelles could be turned over in the lysosomes (or perhaps in the late endosomes). This occurs in some systems, such as the autophagy of cytosol, including mitochondria and endoplasmic reticulum, by hepatocytes (43). However, APC, and specifically B cells and B cell lymphomas, do not appear to carry out such processes, as witnessed by the rare occurrence of cytosol-derived peptides in class II preparations from nonautoimmune sources (23, 44). Generally, only a single cytosol-derived peptide was found for a given class II molecule, and then it was obtained in very low yield (44). In contrast, the I-E^k molecules isolated from MRL/lpr animals bound multiple peptides from cytosolic and nuclear proteins in yields that approached or often exceeded the yield of peptides from proteins that are normally considered to have access to the class II Ag processing pathway.

An additional explanation for the different array of peptides found in the MHC class II molecules of MRL/lpr mice, in comparison with C3H, is that the Fas/Fas ligand system regulates the life span of APC. Fas is expressed on B cells, dendritic cells, and macrophages and serves to induce their apoptosis when engaged by a cell expressing Fas ligand. In the absence of Fas, senescent APC may persist and undergo a prolonged cell death via pathway(s) independent of Fas, possibly leading directly to presentation of self peptides or to the inappropriate entry of peptide donors into the class II processing pathway from which they are normally excluded in healthy cells. It is also possible that debris generated by the death of Fas-deficient cells contains a different array of autoantigens and, consequently, leads to the appearance of unusual class II-associated peptides.

A fifth possible mechanism that has particular relevance for autoimmunity is that cytosolic and nuclear proteins are taken in from the extracellular milieu by Ag-specific receptor-mediated uptake. Prerequisites for this possibility include the expression of autoreactive surface Ig (B cell receptor) on the B cells and a significant degree of cellular destruction that will result in an abundance of Ag in the extracellular milieu. Both of these conditions could be present in SLE mice. This possibility is consistent with the model presented above for cognate recognition of nucleic acid/protein complexes or of complex proteins.

Our findings bear on several suggestions that autoantigenic peptides may possess unique properties. It has been observed for human serologically defined (B cell) epitopes that most of the protein donors of the common epitopes can be cleaved by the protease granzyme B (45). We therefore examined the protein donors of the potential T cell epitopes for autoantibody formation for the existence of sequences recognized by granzyme B. We found that only two of eight potential T cell autoepitopes possessed granzyme B cleavage sites. As noted by Casciola-Rosen et al. (45), not all theoretical cleavage sites are cleaved by granzyme B in practice, so our finding of two proteins with cleavage sites must be considered an upper limit to the number of epitope donors that can actually serve as substrates for the protease. Thus, our data are not consistent with the postulate that autoantigenic peptides most frequently are derived from substrates that are cleaved by granzyme B.

A second model of autoantigenic determinants posits that charged residues (46), usually basic amino acids (47, 48), are critical for serologically defined epitopes. We calculated isoelectric point (pI) values for the putative autoantigenic peptides in our study and found that the pI values ranged from weakly acidic (pI = 5.95) to basic (pI = 10.74), but that this range completely overlapped the range of pI values found in the normal peptides that we isolated. Likewise, incidence of charged residues in normal and potentially autoantigenic peptides did not differ significantly. This is perhaps not surprising when one considers that the majority of peptides described in this study fit the motifs for binding to I-A^k or I-E^k, regardless of their autoantigenic potential. However, we did observe that all of the protein donors of potential autoantigenic T cell epitopes (except the Ig H chains) had calculated pI values that were quite basic, ranging from pI = 9.03 for RNA editase-1 to pI = 11.64 for SRp20. This most likely reflects the requirement for the proteins to bind to nucleic acids, but could also be an inherent feature of their autoantigenicity and, thus, may have some bearing on the work with B cell determinants cited above.

We excluded a highly basic peptide (MRAKWRKKRMRRLKRKRRKM) derived from the 60S ribosomal protein L41 from our analysis because it binds equally well to the I-A^k, I-A^d and I-E^k molecules, but did not possess the motif for binding to any of these MHC molecules. However, this peptide is reminiscent of some of the basic, histone-derived peptides that stimulate autoreactive (SWR x NZB)F₁ T cells in an MHC-nonrestricted manner (49). It is possible that such peptides, which bind MHC molecules promiscuously, are binding outside the peptide-binding groove and act more as a superantigen than a traditional peptide epitope. Future studies, directed at assessing the T cell stimulatory capacity of the peptides described in the present study, will include the L41-derived peptide.

Finally, we wish to address the choice of C3H mice as control animals. The MRL/lpr strain does have a congenic partner MRL/MpJ a/a Tyr^c/Tyr^c-+^Fas-lpr/+^Fas-lpr (MRL/+). These animals have a completely normal Fas gene and do not develop marked lymphoid hyperplasia early in life. However, they do develop a lupus-like syndrome in their second year of life with 50% mortality at 73 wk of age (4). This observation has led some to suggest that the Fas defect is merely an accelerator of lupus-like autoimmune disease that is inherent to the MRL strain (1, 4). Similarly, the potential control strain, C3H/lpr, while not exhibiting the autoimmune pathology present in the MRL/lpr strain, does make autoantibodies (50, 51, 52). Thus, neither MRL/+ nor C3H/lpr animals can be used as "normal" controls. Instead, MRL/+ animals will be used in future studies, along with young MRL/lpr animals, to study the kinetics of expression of the "abnormal" self peptides observed in older MRL/lpr mice.

	Acknowledgments

	Footnotes

 
¹ This study was supported in part by a Lupus Foundation of Colorado grant (to J.H.F.) and by U.S. Public Health Service Grants AI37523 (to J.H.F.), AR33887 (to P.L.C.), and AR26574, AR40620, and AR34156 (to R.A.E.).

² Address correspondence and reprint requests to Dr. John H. Freed, National Jewish Medical and Research Center, 1400 Jackson Street, Denver, CO 80206.

³ Abbreviations used in this paper: SLE, systemic lupus erythematosus; NP-40, Nonidet P-40; TFA, trifluoroacetic acid; MALDI-TOF, matrix-assisted laser desorption ionization time of flight.

Received for publication November 23, 1999. Accepted for publication February 15, 2000.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Drake, C. G., B. L. Kotzin. 1992. Genetic and immunological mechanisms in the pathogenesis of systemic lupus erythematosus. Curr. Opin. Immunol. 4:733.[Medline]
2. Maldonado, M. A., V. Kakkanaiah, G. C. MacDonald, F. Chen, E. A. Reap, E. Balish, W. R. Farkas, J. C. Jennette, M. P. Madaio, B. L. Kotzin, et al 1999. The role of environmental antigens in the spontaneous development of autoimmunity in MRL-lpr mice. J. Immunol. 162:6322.[Abstract/Free Full Text]
3. Cohen, P. L., R. A. Eisenberg. 1991. lpr and gld: single gene models of systemic autoimmunity and lymphoproliferative disease. Annu. Rev. Immunol. 9:243.[Medline]
4. Theofilopoulos, A. N., F. J. Dixon. 1985. Murine models of systemic lupus erythematosus. Adv. Immunol. 37:269.[Medline]
5. Katagiri, T., P. L. Cohen, R. A. Eisenberg. 1988. The lpr gene causes an intrinsic T cell abnormality that is required for hyperproliferation. J. Exp. Med. 167:741.[Abstract/Free Full Text]
6. Sobel, E. S., T. Katagiri, K. Katagiri, S. C. Morris, P. L. Cohen, R. A. Eisenberg. 1991. An intrinsic B cell defect is required for the production of autoantibodies in the lpr model of murine systemic autoimmunity. J. Exp. Med. 173:1441.[Abstract/Free Full Text]
7. Sobel, E. S., V. N. Kakkanaiah, M. Kakkanaiah, R. L. Cheek, P. L. Cohen, R. A. Eisenberg. 1994. T-B collaboration for autoantibody production in lpr mice is cognate and MHC-restricted. J. Immunol. 152:6011.[Abstract]
8. Mohan, C., S. K. Datta. 1995. Lupus: key pathogenic mechanisms and contributing factors. Clin. Immunol. Immunopathol. 77:209.[Medline]
9. Mohan, C., S. Adams, v. Stanik, S. Datta. 1993. Nucleosome: a major immunogen for pathogenic autoantibody-inducing T cells of lupus. J. Exp. Med. 177:1367.[Abstract/Free Full Text]
10. Theofilopoulos, A. N.. 1995. The basis of autoimmunity. I. Mechanisms of aberrant self-recognition. Immunol. Today 16:90.[Medline]
11. Jevnikar, A. M., M. J. Grusby, L. H. Glimcher. 1994. Prevention of nephritis in major histocompatibility complex class II-deficient MRL-lpr mice. J. Exp. Med. 179:1137.[Abstract/Free Full Text]
12. Chan, O., M. J. Shlomchik. 1998. A new role for B cells in systemic autoimmunity: B cells promote spontaneous T cell activation in MRL-lpr/lpr mice. J. Immunol. 160:51.[Abstract/Free Full Text]
13. Kaliyaperumal, A., C. Mohan, W. Wu, S. K. Datta. 1996. Nucleosomal peptide epitopes for nephritis-inducing T helper cells of murine lupus. J. Exp. Med. 183:2459.[Abstract/Free Full Text]
14. Kuwana, M., Jr T. A. Medsger, T. M. Wright. 1997. Highly restricted TCR-ß usage by autoreactive human T cell clones specific for DNA topoisomerase I. J. Immunol. 158:485.[Abstract]
15. Talken, B. L., M. M. Holyst, D. R. Lee, R. W. Hoffman. 1999. T cell receptor ß-chain third complementarity-determining region gene usage is highly restricted among Sm-B autoantigen-specific human T cell clones derived from patients with connective tissue disease. Arthritis Rheum. 42:703.[Medline]
16. Hammerling, G. J., U. Hammerling, H. Lemke. 1979. Isolation of twelve monoclonal antibodies against Ia and H-2 antigens. Immunogenetics 8:433.
17. Ozato, K., N. Mayer, D. H. Sachs. 1980. Hybridoma cell lines secreting monoclonal antibodies to mouse H-2 and Ia antigens. J. Immunol. 124:533.[Abstract]
18. Kappler, J. W., B. Skidmore, J. White, P. Marrack. 1981. Antigen-inducible, H-2-restricted, interleukin-2-producing T cell hybridomas: lack of independent antigen and H-2 recognition. J. Exp. Med. 153:1198.[Abstract/Free Full Text]
19. Germain, R.. 1994. MHC-dependent antigen processing and peptide presentation: providing ligands for T lymphocyte activation. Cell 76:287.[Medline]
20. Rudensky, A. Y., P. Preston-Hurlburt, S.-C. Hong, A. Barlow, Jr C. A. Janeway. 1991. Sequence analysis of peptides bound to MHC class II molecules. Nature 353:622.[Medline]
21. Nelson, C. A., I. Vidavsky, N. J. Viner, M. L. Gross, E. R. Unanue. 1997. Amino-terminal trimming of peptide for presentation on major histocompatibility complex class II molecules. Proc. Natl. Acad. Sci. USA 94:628.[Abstract/Free Full Text]
22. Fremont, D. H., D. Monnaie, C. A. Nelson, W. A. Hendrickson, E. R. Unanue. 1998. Crystal structure of I-A^k in complex with a dominant epitope of lysozyme. Immunity 8:305.[Medline]
23. Rammensee, H.-G., T. Friede, S. Stevanovíc. 1995. MHC ligands and peptide motifs: first listing. Immunogenetics 41:178.[Medline]
24. Fremont, D. H., W. A. Hendrickson, P. Marrack, J. Kappler. 1996. Structures of an MHC class II molecule with covalently bound single peptides. Science 272:1001.[Abstract]
25. Marrack, P., L. Ignatowicz, J. W. Kappler, J. Boymel, J. H. Freed. 1993. Comparison of peptides bound to spleen and thymus class II. J. Exp. Med. 178:2173.[Abstract/Free Full Text]
26. Shores, E. W., R. A. Eisenberg, P. L. Cohen. 1986. Role of the Sm antigen in the generation of anti-Sm autoantibodies in the SLE-prone MRL mouse. J. Immunol. 136:3662.[Abstract]
27. Reeves, W. H., Z. M. Sthoeger, R. G. Lahita. 1989. Role of antigen selectivity in autoimmune responses to the Ku (p70/p80) antigen. J. Clin. Invest. 84:562.
28. Cohen, P. L., R. G. Rapoport, R. A. Eisenberg. 1985. Hidden autoantibodies against common serum proteins in murine systemic lupus erythematosus: detection by in vitro plaque-forming cell assay. J. Exp. Med. 161:1587.[Abstract/Free Full Text]
29. Satoh, M., J. J. Langdon, C.-H. Chou, D. P. McCauliffe, E. L. Treadwell, T. Ogasawara, M. Hirakata, A. Suwa, P. L. Cohen, R. A. Eisenberg, W. H. Reeves. 1994. Characterization of the Su antigen, a macromolecular complex of 100/102 and 200-kDa proteins recognized by autoantibodies in systemic rheumatic diseases. Clin. Immunol. Immunopathol. 73:132.[Medline]
30. Kaliyaperumal, A., M. A. Michaels, S. K. Datta. 1999. Antigen-specific therapy of murine lupus nephritis using nucleosomal peptides: tolerance spreading impairs pathogenic function of autoimmune T and B cells. J. Immunol. 162:5775.[Abstract/Free Full Text]
31. Kermekchiev, M., J. L. Workman, C. S. Pikkard. 1997. Nucleosome binding by the polymerase I transactivator upstream binding factor displaces linker histone H1. Mol. Cell. Biol. 17:5833.[Abstract]
32. Melcher, T., S. Maas, A. Herb, R. Sprengel, P. H. Seeburg, M. Higuchi. 1996. A mammalian RNA editing enzyme. Nature 379:460.[Medline]
33. Scott, J.. 1995. A place in the world for RNA editing. Cell 81:833.[Medline]
34. O’Connell, M. A., A. Gerber, W. Keller. 1997. Purification of human double-stranded RNA-specific editase 1 (hRED1) involved in editing of brain glutamate receptor B pre-mRNA. J. Biol. Chem. 272:473.[Abstract/Free Full Text]
35. Reich, E.-P., H. von Grafenstein, A. Barlow, K. E. Swenson, K. Williams, Jr C. A. Janeway. 1994. Self peptides isolated from MHC glycoproteins of non-obese diabetic mice. J. Immunol. 152:2279.[Abstract]
36. Kambach, C., S. Walke, R. Young, J. M. Avis, E. de la Fortelle, V. A. Raker, R. Luhrmann, J. Li, K. Nagai. 1999. Crystal structures of two Sm protein complexes and their implications for the assembly of the spliceosomal snRNPs. Cell 96:375.[Medline]
37. Lou, H., K. M. Neugebauer, R. F. Gagel, S. M. Berget. 1998. Regulation of alternative polyadenylation by U1 snRNPs and SRp20. Mol. Cell. Biol. 18:4977.[Abstract/Free Full Text]
38. Caceres, J. F., T. Misteli, G. R. Screaton, D. L. Spector, A. R. Krainer. 1997. Role of the modular domains of SR proteins in subnuclear localization and alternative splicing specificity. J. Cell Biol. 138:225.[Abstract/Free Full Text]
39. Satoh, M., K. J. Hamilton, A. K. Ajmani, X. Dong, J. Wand, Y. S. Kanwar, W. H. Reeves. 1996. Autoantibodies to ribosomal T antigens with immune complex glomerulonephritis in SJL mice treated with pristane. J. Immunol. 157:3200.[Abstract]
40. Nyman, U., H. Hallman, G. Hadlaczky, I. Petersson, G. Sharp, N. R. Ringertz. 1986. Intranuclear localization of snRNP antigens. J. Cell Biol. 102:137.[Abstract/Free Full Text]
41. Hunt, D. F., H. Michel, T. A. Dickinson, J. Shabanowitz, A. L. Cox, K. Sakaguchi, E. Appella, H. M. Grey, A. Sette. 1992. Peptides presented to the immune system by the murine class II major histocompatibility complex molecule I-A^d. Science 256:1817.[Abstract/Free Full Text]
42. Nakagawa, T. Y., W. H. Brissette, P. D. Lira, R. J. Griffiths, N. Petrushova, J. Stock, J. D. McNeish, S. E. Eastman, E. D. Howard, S. R. M. Clarke, et al 1999. Impaired invariant chain degradation and antigen presentation and diminished collagen-induced arthritis in cathepsin S null mice. Immunity 10:207.[Medline]
43. Stromhaug, P. E., T. O. Berg, M. Fengsrud, P. O. Seglen. 1998. Purification and characterization of autophagosomes from rat hepatocytes. Biochem. J. 335:217.
44. Chicz, R. M., R. G. Urban, J. C. Gorga, D. A. A. Vignali, W. S. Lane, J. L. Strominger. 1993. Specificity and promiscuity among naturally processed peptides bound to HLA-DR alleles. J. Exp. Med. 178:27.[Abstract/Free Full Text]
45. Casciola-Rosen, L., F. Andrade, D. Ulanet, W. B. Wong, A. Rosen. 1999. Cleavage by granzyme B is strongly predictive of autoantigen status: implications for initiation of autoimmunity. J. Exp. Med. 190:815.[Abstract/Free Full Text]
46. Hahn, B. H., R. R. Singh, F. M. Ebling. 1998. Self Ig peptides that help anti-DNA antibody production: importance of charged residues. Lupus 7:293.[Free Full Text]
47. Pelsue, S., K. D. Jung, P. F. Agris. 1993. Immunochemical analysis of an arginine-rich systemic lupus erythematosus autoepitope. Autoimmunity 15:231.[Medline]
48. James, J. A., R. H. Scofield, J. B. Harley. 1994. Basic amino acids predominate in the sequential autoantigenic determinants of the small nuclear 70K ribonucleoprotein. Scand. J. Immunol. 39:557.[Medline]
49. Shi, Y., A. Kaliyaperumal, L. Lu, S. Southwood, A. Sette, M. A. Michaels, S. K. Datta. 1998. Promiscuous presentation and recognition of nucleosomal autoepitopes in lupus: role of autoimmune T cell receptor chain. J. Exp. Med. 187:367.[Abstract/Free Full Text]
50. Warren, R. W., S. A. Caster, J. B. Roths, E. D. Murphy, D. S. Pisetsky. 1984. The influence of the lpr gene on B cell activation: differential antibody expression in lpr congenic mouse strains. Clin. Immunol. Immunopathol. 31:65.[Medline]
51. Ishigatsubo, Y., A. D. Steinberg, D. M. Klinmnan. 1988. Autoantibody production is associated with polyclonal B cell activation in autoimmune mice which express the lpr or gld genes. Eur. J. Immunol. 18:1089.[Medline]
52. Giese, T., W. F. Davidson. 1994. Chronic treatment of C3H-lpr/lpr and C3H-gld/gld mice with anti-CD8 monoclonal antibody prevents the accumulation of double negative cells but not autoantibody production. J. Immunol. 152:2000.[Abstract]

This article has been cited by other articles:

				B. C. Beaudette-Zlatanova, T. Ling, M. J. Shlomchik, A. Marshak-Rothstein, and I. R. Rifkin B Cells and Dendritic Cells from V{kappa}8 Light Chain Transgenic Mice Activate MRL-lpr/gld CD4+ T Cells J. Immunol., July 1, 2006; 177(1): 45 - 52. [Abstract] [Full Text] [PDF]

				L. Bennett, A. K. Palucka, E. Arce, V. Cantrell, J. Borvak, J. Banchereau, and V. Pascual Interferon and Granulopoiesis Signatures in Systemic Lupus Erythematosus Blood J. Exp. Med., March 17, 2003; 197(6): 711 - 723. [Abstract] [Full Text] [PDF]

				C.-H. Suh, J. H. Freed, and P. L. Cohen T Cell Reactivity to MHC Class II-Bound Self Peptides in Systemic Lupus Erythematosus-Prone MRL/lpr Mice J. Immunol., February 15, 2003; 170(4): 2229 - 2235. [Abstract] [Full Text] [PDF]

				A. Kaliyaperumal, M. A. Michaels, and S. K. Datta Naturally Processed Chromatin Peptides Reveal a Major Autoepitope That Primes Pathogenic T and B Cells of Lupus J. Immunol., March 1, 2002; 168(5): 2530 - 2537. [Abstract] [Full Text] [PDF]

				S. C. Henderson, M. M. Kamdar, and A. Bamezai Ly-6A.2 Expression Regulates Antigen-Specific CD4+ T Cell Proliferation and Cytokine Production J. Immunol., January 1, 2002; 168(1): 118 - 126. [Abstract] [Full Text] [PDF]

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 Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Freed, J. H. Articles by Eisenberg, R. A. Search for Related Content PubMed PubMed Citation Articles by Freed, J. H. Articles by Eisenberg, R. A.