The specificity of peptides bound to human histocompatibility leukocyte antigen (HLA)-B27 influences the prevalence of arthritis in HLA-B27 transgenic rats

Abstract

Human histocompatibility leukocyte antigen B27 is highly associated with the rheumatic diseases termed spondyloarthropathies, but the mechanism is not known. B27 transgenic rats develop a spontaneous disease resembling the human spondyloarthropathies that includes arthritis and colitis. To investigate whether this disease requires the binding of specific peptides to B27, we made a minigene construct in which a peptide from influenza nucleoprotein, NP383-391 (SRYWAIRTR), which binds B27 with high affinity, is targeted directly to the ER by the signal peptide of the adenovirus E3/gp19 protein. Rats transgenic for this minigene, NP1, were made and bred with B27 rats. The production of the NP383-391 peptide in B27(+)NP1(+) rats was confirmed immunologically and by mass spectrometry. The NP1 product displaced approximately 90% of the 3H-Arg-labeled endogenous peptide fraction in B27(+)NP1(+) spleen cells. Male B27(+)NP1(+) rats had a significantly reduced prevalence of arthritis, compared with B27(+)NP- males or B27(+) males with a control construct, NP2, whereas colitis was not significantly affected by the NP1 transgene. These findings support the hypothesis that B27-related arthritis requires binding of a specific peptide or set of peptides to B27, and they demonstrate a method for efficient transgenic targeting of peptides to the ER.

Figure 1

Construction of the minigenes encoding the adenovirus E3/19K signal peptide and influenza NP383-391 peptide. (A) Procedure for generation of the double-stranded DNA insert. The indicated overlapping oligonucleotides were synthesized and annealed. Double stranded synthesis was completed with Klenow enzyme, and the resulting construct was cleaved with SalI and BamHI and ligated into the expression vector pHSE3′ (21). For the NP2 minigene, the residue at P2 in the B27-binding epitope was changed from Arg to Leu by replacing the codon AGG with the codon TTG. This substitution abrogates binding of the peptide to B27 (22). (B) Map of the complete construct for NP1 and NP2.

Figure 2

The NP1 transgene construct directs expression of antigenically authentic NP1 peptide. (A) Con A–stimulated LN lymphoblast targets from the indicated transgenic rats were tested for lysis by the human CTL line Q124, which is specific for HLA-B27 and the NP383-391 peptide. Only cells from rats carrying both the B27/hβ2m and NP1 transgene loci were lysed significantly. (B) Con A blast targets from a B27⁺ NP1⁻ rat were pulsed with synthetic NP peptide (10 nM), then washed and tested for lysis by the Q124 CTL line at an E:T of 5:1, in the presence of varying concentrations of the indicated unlabeled (cold) targets. Lysis was efficiently inhibited by double transgenic B27⁺NP1⁺ targets and by B27⁺ targets pulsed with synthetic NP peptide. (C) B27⁺NP1⁻ rats were immunized in vivo with cells from RT1-matched B27⁺NP1⁺ rats, and the immune LN cells were restimulated in an MLR. The resulting CTL effectors lysed B27/hβ2m-transfected mouse EL-4 tumor cells to which authentic NP1 peptide, but not NP2 peptide, had been added.

Figure 3

Isolation and immunologic identification of NP peptide from transgenic rats. B27 molecules were immunoprecipitated from 10 ⁸ B27⁺NP1⁺ (A) and B27⁺NP⁻ (B) Con A–stimulated spleen cells and the bound peptides were extracted with 0.1% TFA and separated by RP-HPLC. One-tenth of each fraction was added to 5,000 ⁵¹Cr-labeled B27⁺ C1R cells, which were then tested for lysis by 2.5 × 10⁴ line Q124 CTL effectors. Lysis above background was seen only with the peptides from the B27⁺NP1⁺ cells, with all of the immunologic activity in fraction 83.

Figure 4

Mass spectrometric analysis of the immunologically active NP fraction, as shown in Fig. 3. (Top) Electrospray MS spectrum. The [M+3H]³⁺ ions of components A–E are indicated, with the charge state evident from the m/z ratio separation of the isotopic components. Corresponding [M+2H]²⁺ and [M+4H]⁴⁺ ions were also observed outside the m/z range shown. (Bottom) Product ion spectrum recorded during tandem MS analysis. The selected precursor ion was m/z 403.6 ([M+3H]³⁺). Fragment ions are designated using the nomenclature of Biemann (41). Confirmatory evidence was obtained from the product ion spectrum of m/z 604.8 ([M+2H]²⁺); b₇ ²⁺ was observed only in the latter spectrum.

Figure 4

Mass spectrometric analysis of the immunologically active NP fraction, as shown in Fig. 3. (Top) Electrospray MS spectrum. The [M+3H]³⁺ ions of components A–E are indicated, with the charge state evident from the m/z ratio separation of the isotopic components. Corresponding [M+2H]²⁺ and [M+4H]⁴⁺ ions were also observed outside the m/z range shown. (Bottom) Product ion spectrum recorded during tandem MS analysis. The selected precursor ion was m/z 403.6 ([M+3H]³⁺). Fragment ions are designated using the nomenclature of Biemann (41). Confirmatory evidence was obtained from the product ion spectrum of m/z 604.8 ([M+2H]²⁺); b₇ ²⁺ was observed only in the latter spectrum.

Figure 5

The NP1 transgene construct dramatically alters the profile of peptides bound to B27. Spleen cells from NP1⁺ (A), NP⁻ (B), and NP2⁺ (C) LEW.33-3 rats were metabolically labeled with ³H-Arg, and the B27 molecules were immunoprecipitated from detergent lysates of the labeled cells, as previously described (24). The bound peptides were eluted and separated by RP-HLPC, as in Fig. 3, A and B, and the resulting fractions counted for ³H. Total ³H cpm in the B27 immunoprecipitates were (×10⁻⁷): NP1⁺ 2.75, NP2⁺ 3.00, NP⁻ 2.75. Total ³H cpm in the peptide pools eluted from B27 were (×10⁻⁴): NP1⁺ 6.5, NP2⁺ 8.4, NP⁻ 6.8.

Figure 5

The NP1 transgene construct dramatically alters the profile of peptides bound to B27. Spleen cells from NP1⁺ (A), NP⁻ (B), and NP2⁺ (C) LEW.33-3 rats were metabolically labeled with ³H-Arg, and the B27 molecules were immunoprecipitated from detergent lysates of the labeled cells, as previously described (24). The bound peptides were eluted and separated by RP-HLPC, as in Fig. 3, A and B, and the resulting fractions counted for ³H. Total ³H cpm in the B27 immunoprecipitates were (×10⁻⁷): NP1⁺ 2.75, NP2⁺ 3.00, NP⁻ 2.75. Total ³H cpm in the peptide pools eluted from B27 were (×10⁻⁴): NP1⁺ 6.5, NP2⁺ 8.4, NP⁻ 6.8.

Figure 5

The NP1 transgene construct dramatically alters the profile of peptides bound to B27. Spleen cells from NP1⁺ (A), NP⁻ (B), and NP2⁺ (C) LEW.33-3 rats were metabolically labeled with ³H-Arg, and the B27 molecules were immunoprecipitated from detergent lysates of the labeled cells, as previously described (24). The bound peptides were eluted and separated by RP-HLPC, as in Fig. 3, A and B, and the resulting fractions counted for ³H. Total ³H cpm in the B27 immunoprecipitates were (×10⁻⁷): NP1⁺ 2.75, NP2⁺ 3.00, NP⁻ 2.75. Total ³H cpm in the peptide pools eluted from B27 were (×10⁻⁴): NP1⁺ 6.5, NP2⁺ 8.4, NP⁻ 6.8.

Figure 6

Expression of the NP1 transgene construct reduces recognition of other B27-presented peptides. (A) CTL against the male-specific HY minor histocompatibility antigen presented by HLA-B27 were generated in LEW transgenic rats as previously described (10, 16, 17, 24). These were tested for lysis of Con A lymphoblasts from male rats carrying the indicated transgenes. The rats were from a backcross of the SD 300-5 line to the DA.33-3 line, and therefore did not share any RT1 alleles with the effectors. (B) Allospecific CTL were generated by in vivo immunization of nontransgenic DA rats with cells from sex-matched B27/hβ2m LEW 21-4L rats congenic for RT1 ^av1, and the immune LN cells were similarly restimulated in an MLR. The CTL were tested for lysis of the Con A blast target cells described in A.