Apoptotically modified forms of autoantigens have been hypothesized to participate in lupus immunopathogenesis. This study identifies a major B cell epitope present on the apoptotic but not the intact form of the U1-70-kDa ribonucleoprotein lupus autoantigen (70k). Human autoimmune sera with strong recognition of apoptotic 70k and minimal recognition of intact 70k were identified and tested for reactivity to truncated forms of 70k by immunoblot and ELISA. Patient sera that preferentially recognized apoptotic 70k were specific for an epitope dependent on residues 180–205 of the protein. This epitope was also recognized by 19 of 28 (68%) intact anti-70k-positive autoimmune human sera with Abs also recognizing apoptotic but not the intact form 70k, but only 1 of 9 (11%) intact 70k-positive sera without such Abs (Fisher’s exact, p = 0.0055). Immunization of HLA-DR4-transgenic C57BL/6 mice with a peptide containing this epitope induced anti-70k immunity in 13 of 15 mice, including Abs recognizing apoptotic but not intact forms of autoantigens in 12 of 15 mice. Anti-70k responder mice also developed spreading of immunity to epitopes on the endogenous form of 70k, and proliferative lung lesions consistent with those described in patients with anti-70k autoimmunity. Thus, a major epitope in the B cell response to U1-70 kDa localizes to the RNA binding domain of the molecule, overlaps with the most common T cell epitope in the anti-70k response, and is not present on the intact form of the 70k molecule. Immunization of mice against this epitope induces an immune response with features seen in human anti-70k autoimmune disease.

The observation that the U1-70-kDa ribonucleoprotein (70k)3 is cleaved in apoptosis was one of the early events that led to the proposal that apoptotically modified forms of autoantigens may drive autoimmunity in lupus (1), and thus, that the potential to undergo structural modification in apoptosis may be a feature that favors lupus antigenicity. In support of this hypothesis, the U1-70-kDa Ag is often an early target of IgG responses in the development of human anti-ribonucleoprotein (RNP) autoimmunity (2). Abs to the apoptotic form of 70k exist that are distinct from Abs to intact U1-70 kDa (3), and patients with high recognition of the apoptotic form of 70k have increased prevalence of lupus skin disease (4).

Although Abs to U1-70 kDa appear relatively frequently in patients with lupus (5), the prevalence of immunity to other apoptotically modified Ags such as NuMa and Keratin-18 in lupus is much lower (6, 7, 8). Thus, some investigators have argued that other features of the 70k protein are more likely to account for its antigenicity. One such alternate view emphasizes the importance of RNA binding domains in lupus antigenicity. A number of lupus autoantigen proteins are remarkable for their binding to RNA (9, 10). Regarding 70k immunity, recent T cell epitope mapping in humans with anti-70k immunity supports a role for the RNA binding regions in immunogenicity; the only anti-70k T cell epitopes identified targeted the U1-70-kDa RNA binding domain, which extends from aa 97–202 in a 437-aa protein (11). This area is distant from the apoptotic caspase cleavage site of the 70k molecule at aa 341 (12). Animal studies have similarly observed overlapping B and T cell epitopes arising early in the anti-70k autoimmune process that preferentially target an area within the RNA binding domain of 70k (13). Thus, the hypothesis has been proposed that immunity to the RNA binding domain of 70k drives anti-70k autoimmunity (11, 13).

By mapping B cell epitopes in humans with anti-70k immunity that are present on apoptotic but not intact forms of 70k, this study has linked the apoptosis hypothesis for 70k autoimmunity to the RNA binding domain hypothesis. We examined the Ab reactivity of human sera with Abs to apoptotic but not intact U1-70 kDa; confirmatory studies were performed with other 70k-reactive antisera and in a mouse model of 70k immunity. Using these approaches, a major epitope of 70k present on the apoptotic form of the molecule but not on the intact form was identified that overlaps with the RNA binding domain of the 70k protein. Thus, apoptotic modification and the presence of an RNA binding domain may both contribute to the autoantigenicity of the U1-70-kDa protein. Moreover, immunization with this epitope led to spreading of immunity to the endogenous form of 70k, and to manifestations of autoimmune tissue injury.

Sera from all 60 patients referred from the University of Missouri Medical Center or the Harry S. Truman Memorial Veterans Affairs Medical Center between 1997 and 2001 for extractable nuclear Ag (ENA) testing who had ENA titers of 1:2000 or greater by hemagglutination were evaluated. This patient sample was selected independently from previous reports on apoptosis-specific Abs. All samples and data for this study were obtained through protocols approved by the University of Missouri Institutional Review Board.

Jurkat cells grown in standard culture conditions were washed and lysed in SDS sample buffer. To generate apoptotic samples, cells were exposed to 1650 J/m2 UV-B light and incubated for 8 h at 37°C. Truncated forms of 70k were expressed as maltose-binding protein (MBP) fusion proteins in Escherichia coli and affinity purified over amylose columns, as previously described (3). Truncation mutants were generated from the pMALc2g-61-70-kDa expression plasmid using QuikChange mutagenesis (Stratagene, La Jolla, CA), as previously described (11). The modified plasmids were confirmed to have the proper sequences before expression. The identity and purity of all products was confirmed by immunoblot against standard 70k-recognizing sera. In some experiments, the MBP fusion protein partner was removed by preincubation with 1 μl of Genenase I (New England Biolabs, Beverly, MA) per milligram of MBP-conjugated protein for 8 h at 25°C in a 20 mM Tris (pH 8.0) and 200 mM NaCl buffer, followed by gel purification of the 70k peptide as previously described (14). Purified, cell culture-grade MBP was from New England Biolabs.

The 180–205 peptide form of U1-70 kDa and the collagen II peptide 14 were synthesized using Fmoc chemistry on an Applied Biosystems (Foster City, CA) model 432A peptide synthesizer (11, 15), and analyzed for purity and sequence fidelity using HPLC and mass spectrometry.

Control and apoptotic Jurkat lysates in 50-μg aliquots, as well as purified protein aliquots as indicated, were subjected to SDS-PAGE and transferred to nitrocellulose. Blots were blocked with 3% BSA, exposed to 1/10,000 dilutions of primary Ab (patient sera) in the presence or absence of blocking soluble Ag, incubated with HRP-linked goat anti-human IgG (Fc region-specific) secondary Ab (Sigma-Aldrich, St. Louis, MO), and visualized with chemiluminescence. For experiments on mouse sera, 1/1000 dilutions of primary Ab were used, and HRP-linked goat anti-mouse IgG (Sigma-Aldrich) was used as secondary Ab. Experiments with competitive blocking Ags used in the primary incubations were performed as described (3), using purified 70k peptide at 2 μg/ml dilutions. Control antisera, test antisera, and competitively blocked specimens were prepared on the same gel, processed in parallel, and analyzed at the same exposure.

Ninety-six-well flat-bottom microtiter plates were incubated overnight at 4°C with 100 ng of purified Ag in 25 μl/well PBS. After washing in PBS/0.05% Tween 20 (PBS/Tween) buffer, plates were incubated with 1% BSA in PBS/Tween for 4 h at room temperature. Sera were tested at final dilutions between 1/100 and 1/2000 in PBS/Tween with 1% BSA with similar results; final dilutions of 1/1000 for human sera and 1/250 for mouse sera are presented. Plates were incubated with primary sera for 4 h at room temperature, with HRP-linked Fc region-specific goat anti-human IgG or goat anti-mouse IgG secondary Abs (Sigma-Aldrich) overnight at 4°C, and then with 0.55 mg/ml orthophenylenediamine in 100 μl of sodium citrate (pH 5.0) buffer containing 0.03% hydrogen peroxide for 60 min at room temperature. Washes were performed between each incubation with PBS/Tween. Absorbance at 450 nm was measured in a Bio-Tek (Winooski, VT) microtiter plate reader. All assays were performed in duplicate wells. Standard positive and negative control sera and test antisera were assayed on each plate, as previously described (3). Data are reported as mean ± SD.

Assays for fluorescent antinuclear Abs (FANA) on mouse sera were performed using a Hep-2 cell substrate (DiaSorin, Minneapolis, MN) with a fluorescence-tagged goat anti-mouse IgG secondary Ab (Jackson ImmunoResearch Laboratories, West Grove, PA), as previously described (14). Assays for anti-dsDNA Abs (dsDNA) on mouse sera were performed using a Crithidia lucillae substrate (Scimedex, Denville, NJ) with the same second Ab, as previously described (14). FANA assays were performed at serum titers of 1:40 and dsDNA assays were performed at serum titers of 1:20; all assays were read by trained technicians, who were blinded to the test conditions of the mice studied. Established negative and positive murine control sera were used in each assay.

Because the human anti-70k response in autoimmune disease patients is strongly associated with HLA-DR4 (16), a DR4 transgenic strain of C57BL/6 with concurrent knockout of endogenous mouse class II molecules (C57BL/6Ntac-[KO]Abb-[Tg]DR-4) was obtained from R. Schwartz (National Institute for Allergy and Infectious Diseases, Bethesda, MD). This strain uses a hybrid MHC class II molecule composed of the peptide binding domains of human DR4 and the membrane proximal domains of mouse I-E. As the α2 and β2 domains of mouse MHC class II are preserved, interactions with CD4 coreceptors on murine T cells are maintained (17).

Mice were immunized s.c. with 50 μg of the bacterially produced and purified 70k peptides emulsified in CFA (Difco, Detroit, MI). Control mice were immunized with equal quantities of human collagen II (Elastin Products, Owensville, MO) or MBP (Sigma-Aldrich) following the same protocol. Sera were drawn 1 mo after initial immunization. All mouse studies were approved by the Institutional Animal Care and Use Committee at our center, and all animals were housed in American Association for Accreditation of Laboratory Animal Care-approved facilities.

Mice were euthanized by cervical dislocation after induction of isoflurane anesthesia 1 mo after immunization, and organs were immediately harvested. Formalin-fixed, paraffin-embedded sections were stained with H&E, and imaged with a Nikon (Melville, NY) CoolPix 990 camera mounted on an American Optical/Spencer (Buffalo, NY) model 1036A microscope. Images were converted to grayscale using Microsoft Photo Editor; no additional image manipulation was performed.

Statistics were calculated used Prism 3.0 (GraphPad Software, San Diego, CA). Differences in group prevalences were tested using Fisher’s exact test or χ2 analysis.

To identify sera that would react preferentially with the apoptotic form of U1-70 kDa vs the intact form, we tested sera from all patients seen at the University of Missouri Medical Center or the Harry S. Truman Memorial Veterans Affairs Hospital between 1997 and 2001 in whom serum Abs to ENA were measured at the University of Missouri Antinuclear Antibody Testing Laboratory by hemagglutination at a titer of 1:2000 or greater. Serum samples from 60 patients were evaluated for Abs to the intact and apoptotic forms of U1-70 kDa using an immunoblot assay at serum dilutions of 1/10,000, as previously described (3). IgG Abs clearly recognizing the intact form of U1-70 kDa were found in 37 subjects, 28 of whom also had apoptosis-specific anti-70k Abs identified by blocking immunoblot (3). In seven additional subjects in this group tested concurrently, IgG Abs recognizing the apoptotic form of 70k were noted in association with very weak or absent recognition of the intact form of the Ag (Fig. 1).

FIGURE 1.

Identification of antisera that preferentially recognize the apoptotic vs the intact form of U1-70 kDa. A series of human antisera with positive tests for ENA were tested for Abs to intact and apoptotic U1-70 kDa by immunoblot at serum dilutions of 1/10,000 as previously described (3 ). Representative results are shown from at least three separate experiments per serum. The antisera A1, A2, A3, A4, A5, A6, and A7 showed recognition of a 40-kDa band in apoptotic Jurkat lysates that comigrated with the apoptotic form of U1-70 kDa identified by anti-70k control sera such as C1 (shown). However, these sera had little (sera A4 and A6) or no recognition of the intact form of 70k in control Jurkat lysates, whereas anti-70k control sera run on the same gels showed similar recognition of intact 70k from control lysates compared with recognition of apoptotic 70k from apoptotic lysates.

FIGURE 1.

Identification of antisera that preferentially recognize the apoptotic vs the intact form of U1-70 kDa. A series of human antisera with positive tests for ENA were tested for Abs to intact and apoptotic U1-70 kDa by immunoblot at serum dilutions of 1/10,000 as previously described (3 ). Representative results are shown from at least three separate experiments per serum. The antisera A1, A2, A3, A4, A5, A6, and A7 showed recognition of a 40-kDa band in apoptotic Jurkat lysates that comigrated with the apoptotic form of U1-70 kDa identified by anti-70k control sera such as C1 (shown). However, these sera had little (sera A4 and A6) or no recognition of the intact form of 70k in control Jurkat lysates, whereas anti-70k control sera run on the same gels showed similar recognition of intact 70k from control lysates compared with recognition of apoptotic 70k from apoptotic lysates.

Close modal

To map B cell epitopes of sera reactive with apoptotic but not intact 70k, we examined the reactivity of the seven antisera that preferentially recognized the apoptotic form of U1-70 kDa using C-terminal truncation mutants of 70k. All seven antisera recognized a recombinant, bacterially produced, purified preparation of the 1- to 341-aa apoptotic form of U1-70 kDa, as well as C-terminally truncated forms of 70k down to aa 205 by immunoblot (not shown). However, although all sera recognized the 1–205 form of 70k by immunoblot, five of seven sera did not recognize a further C-terminally truncated 1–180 form of 70k as strongly by immunoblot (Fig. 2,A). Additional truncation to a 1–150 form of 70k was not recognized by any of the tested sera (not shown). The patient sera with preferential recognition of apoptotic 70k by immunoblot also had substantially higher recognition of the 1–205 form than of the 1–180 form by ELISA (Fig. 2 B). This suggested that the 180–205 region was critical to recognition of apoptotic U1-70 kDa by these sera.

FIGURE 2.

Recognition of C-terminally truncated U1-70-kDa peptides by human antisera that preferentially recognize the apoptotic form of U1-70 kDa. A, Immunoblot recognition of purified, bacterially produced 1–205 and 1–180 peptide forms of U1-70 kDa by seven sera that preferentially recognize apoptotic 70k (A4, A2, A3, A1, A7, A5, and A6) by immunoblot, and a representative serum with similar immunoblot recognition of both intact and apoptotic forms of U1-70 kDa (C2) (3 ). Immunoblots were performed with sera diluted 1/10,000 as in Fig. 1, with 1 μg of peptide loaded per lane. Sera A4, A2, and A3 have substantially stronger recognition of the 1–205 form than the 1–180 form of the protein; sera A1 and A7 have modestly increased recognition of the 1–205 vs the 1–180 form; no difference in recognition of 1–205 vs 1–180 was seen with sera A5 and A6. B, ELISA testing for recognition of purified, bacterially produced 1–205 and 1–180 peptides of U1-70 kDa by seven sera that preferentially recognize apoptotic 70k (A4, A2, A3, A1, A7, A5, and A6), two representative sera with similar recognition of both intact and apoptotic forms of U1-70 kDa (C1 and C2), and one normal control serum (N1). Serum dilutions of 1/1000 are shown. Values are expressed as mean OD ± SD. All the sera that preferentially recognize apoptotic 70k show substantially higher recognition of the 1–205 form compared with the 1–180 form.

FIGURE 2.

Recognition of C-terminally truncated U1-70-kDa peptides by human antisera that preferentially recognize the apoptotic form of U1-70 kDa. A, Immunoblot recognition of purified, bacterially produced 1–205 and 1–180 peptide forms of U1-70 kDa by seven sera that preferentially recognize apoptotic 70k (A4, A2, A3, A1, A7, A5, and A6) by immunoblot, and a representative serum with similar immunoblot recognition of both intact and apoptotic forms of U1-70 kDa (C2) (3 ). Immunoblots were performed with sera diluted 1/10,000 as in Fig. 1, with 1 μg of peptide loaded per lane. Sera A4, A2, and A3 have substantially stronger recognition of the 1–205 form than the 1–180 form of the protein; sera A1 and A7 have modestly increased recognition of the 1–205 vs the 1–180 form; no difference in recognition of 1–205 vs 1–180 was seen with sera A5 and A6. B, ELISA testing for recognition of purified, bacterially produced 1–205 and 1–180 peptides of U1-70 kDa by seven sera that preferentially recognize apoptotic 70k (A4, A2, A3, A1, A7, A5, and A6), two representative sera with similar recognition of both intact and apoptotic forms of U1-70 kDa (C1 and C2), and one normal control serum (N1). Serum dilutions of 1/1000 are shown. Values are expressed as mean OD ± SD. All the sera that preferentially recognize apoptotic 70k show substantially higher recognition of the 1–205 form compared with the 1–180 form.

Close modal

The 25-mer 180–205 peptide itself was synthesized chemically as previously described (11), and was evaluated for serum binding characteristics. It was not recognized by any of the sera that reacted with the apoptotic but not the intact form of 70k when tested by ELISA, it did not competitively inhibit recognition by these sera of the 1–205 form by ELISA, and it was unable to block recognition of apoptotic 70k in Jurkat lysate with any of these sera (data not shown). Thus, although the epitope recognized on apoptotic but not intact 70k is dependent on the presence of residues 180–205, it is also dependent on the presence of additional components of the U1-70-kDa sequence. An N-terminal truncation of the first 62 residues of 70k had no effect on recognition of the 180- to 205-dependent epitope in any of the sera tested (data not shown). Denaturation in 6 M urea reduced recognition of the 180- to 205-dependent epitope (data not shown), suggesting that this epitope is conformational.

To examine the prevalence of Abs to the 180- to 205-dependent epitope of U1-70 kDa, we assessed for ELISA recognition of forms of 70k C-terminally truncated at residue 180 and 205 in our unselected group of 37 intact U1-70-kDa-positive patient sera described above (Fig. 3). A priori, we stipulated that, to be positive for the 180- to 205-dependent epitope, sera must have recognition of the 63–205 peptide from 70k (expressed as mean OD) at least 5 SDs above normal control sera, and that the recognition of the 63–205 form must be at least 50% greater than the recognition of the 63–180 form. Using these criteria, patient sera that had been found to recognize areas on apoptotic 70k that were not recognized on the intact form of 70k by immunoblot frequently showed recognition of the 180- to 205-dependent epitope (19 of 28; 68%), whereas other 70k-positive patient sera showed recognition of the 180- to 205-dependent epitope significantly less frequently (1 of 9; 11%; Fisher’s exact, p = 0.0055). Our analysis was not dependent on the precise criteria that we set to define positivity for the 180- to 205-dependent epitope; for example, using criteria as permissive as 2 SDs and strict as 7 SDs above normal control sera did not significantly affect our results (data not shown). The high prevalence of Abs to the 180- to 205-dependent epitope in an unselected population of anti-70k antisera that recognize areas on apoptotic 70k that are not recognized on intact 70k suggests that the 180- to 205-dependent epitope is a frequent target of such responses.

FIGURE 3.

Recognition of the 180- to 205-dependent epitope in intact U1-70-kDa-positive patient sera. A, Twenty-eight antisera with substantial levels of recognition of the intact form of 70k and also persistent recognition of apoptotic 70k after competitive blockade of recognition of intact 70k on immunoblot from an unselected cohort of ENA-positive patients were assayed for recognition of the p205 and p180 forms of 70k as described in Fig. 2 B. Sera 1–19 met criteria for specific recognition of the 180- to 205-dependent epitope (see Results), whereas sera 20–28 did not show specific recognition. B, Nine antisera from an unselected cohort of ENA-positive patients with substantial levels of recognition of the intact form of 70k but without recognition of apoptotic 70k after competitive blockade of recognition of intact 70k on immunoblot were assayed for recognition of the p205 and p180 forms of 70k. Serum 1 met criteria for specific recognition of the 180- to 205-dependent epitope (see Results), whereas sera 2–9 did not show specific recognition.

FIGURE 3.

Recognition of the 180- to 205-dependent epitope in intact U1-70-kDa-positive patient sera. A, Twenty-eight antisera with substantial levels of recognition of the intact form of 70k and also persistent recognition of apoptotic 70k after competitive blockade of recognition of intact 70k on immunoblot from an unselected cohort of ENA-positive patients were assayed for recognition of the p205 and p180 forms of 70k as described in Fig. 2 B. Sera 1–19 met criteria for specific recognition of the 180- to 205-dependent epitope (see Results), whereas sera 20–28 did not show specific recognition. B, Nine antisera from an unselected cohort of ENA-positive patients with substantial levels of recognition of the intact form of 70k but without recognition of apoptotic 70k after competitive blockade of recognition of intact 70k on immunoblot were assayed for recognition of the p205 and p180 forms of 70k. Serum 1 met criteria for specific recognition of the 180- to 205-dependent epitope (see Results), whereas sera 2–9 did not show specific recognition.

Close modal

To further investigate the immune response to the 1–205 form of U1-70 kDa, we immunized a series of C57BL/6 HLA-DR4-transgenic mice with this form of 70k or with a series of controls, and studied serum samples taken 1 mo after immunization. Immunizations were performed with 50 μg of peptide administered s.c. in CFA. Using Jurkat lysates to assess for Abs to intact and apoptotic forms of 70k by immunoblot, we found that 13 of 15 mice immunized with the 1–205 form of 70k had Abs recognizing the apoptotic form of U1-70 kDa, with 10 of these mice also recognizing the intact form of 70k (Fig. 4,A and data not shown). No Abs to 70k were observed in sera collected from 12 mice 1 mo after immunization with MBP (χ2, p < 0.0001), or 3 mice immunized with collagen II following the same protocol (χ2, p = 0.002) (data not shown). Additionally, 0 of 6 unimmunized littermate controls developed anti-70k Abs (χ2, p = 0.0002). Five mice were also immunized with the 1–180 form of 70k in CFA following the same protocol; only 2 of these animals mounted measurable anti-70k responses (χ2, p = 0.037, vs 1–205 immunization) (Fig. 4 B).

FIGURE 4.

A, Response to murine immunization with the 1–205 form of U1-70 kDa. Control and apoptotic Jurkat lysates were immunoblotted with 1/1000 dilutions of sera from mice taken 1 mo after immunization with the 1–205 form of 70k. Results shown are representative of other mouse samples. The identity of the intact and apoptotic 70k bands was confirmed by comigration with 70k bands identified from intact and apoptotic Jurkat lysates using human control sera on the same gels (not shown). Mice A, B, C, and D show recognition of the intact and apoptotic forms of 70k, plus recognition of a 32-kDa activity in apoptotic but not intact lysates; mice E and F show preferential recognition of the apoptotic form of 70k. B, Immunization with the 1–180 form of 70k does not induce immunity to epitopes on apoptotic but not intact Ag forms. Sera taken 1 mo after immunization with the 1–180 70k peptide from three mice were screened by immunoblot in the presence and absence of soluble U1-70-kDa blocking Ag for recognition of the intact and apoptotic forms of 70k in Jurkat lysate as previously described (3 ). Competitive inhibition of recognition of the intact and apoptotic forms of 70k were comparable in mice 1 and 2; mouse 3 had no measurable response to 70k. Although inclusion of the peptide for competitive binding led to increased background signal (particularly with mouse 2), no additional bands were observed in apoptotic but not intact lysates.

FIGURE 4.

A, Response to murine immunization with the 1–205 form of U1-70 kDa. Control and apoptotic Jurkat lysates were immunoblotted with 1/1000 dilutions of sera from mice taken 1 mo after immunization with the 1–205 form of 70k. Results shown are representative of other mouse samples. The identity of the intact and apoptotic 70k bands was confirmed by comigration with 70k bands identified from intact and apoptotic Jurkat lysates using human control sera on the same gels (not shown). Mice A, B, C, and D show recognition of the intact and apoptotic forms of 70k, plus recognition of a 32-kDa activity in apoptotic but not intact lysates; mice E and F show preferential recognition of the apoptotic form of 70k. B, Immunization with the 1–180 form of 70k does not induce immunity to epitopes on apoptotic but not intact Ag forms. Sera taken 1 mo after immunization with the 1–180 70k peptide from three mice were screened by immunoblot in the presence and absence of soluble U1-70-kDa blocking Ag for recognition of the intact and apoptotic forms of 70k in Jurkat lysate as previously described (3 ). Competitive inhibition of recognition of the intact and apoptotic forms of 70k were comparable in mice 1 and 2; mouse 3 had no measurable response to 70k. Although inclusion of the peptide for competitive binding led to increased background signal (particularly with mouse 2), no additional bands were observed in apoptotic but not intact lysates.

Close modal

In addition to having Abs that recognized the apoptotic form of 70k, 12 of 15 mice immunized with the 1–205 form of 70k showed evidence of an immune response targeting apoptotically modified Ags but not intact forms of these Ags. In 3 mice, this consisted of strongly preferential recognition of the apoptotic form of 70k. In 9 additional mice, strongly preferential recognition of a 32-kDa band in apoptotic lysates without a corresponding band in control lysates was noted (Fig. 4,A). In contrast, none of the 1- to 180-immunized mice had strongly preferential recognition of the apoptotic form of 70k or recognition of the apoptotic 32-kDa band (χ2, p = 0.002) (Fig. 4,B). Likewise, none of the mice immunized with MBP or collagen II, or the unimmunized mice developed these Abs (not shown). To assess whether the apoptosis-specific 32-kDa band developing in mice was seen in humans with autoimmune disease, we re-examined our original ENA-positive cohort of 60 patients. Of these, we found that 16 patients’ sera also recognized a 32-kDa band in apoptotic but not control lanes that comigrated with the band seen in mouse sera (Fig. 5). Of this group of patients with apoptosis-specific anti-32-kDa immunity, 14 of 16 subjects also had anti-70k immunity.

FIGURE 5.

Mouse and human antisera recognize a 32-kDa activity in apoptotic but not intact Jurkat lysates. Control and apoptotic Jurkat lysates were immunoblotted on the same gel with either 1/10,000 dilutions of three autoimmune human sera (H1–H3) or 1/1,000 dilutions of three mice immunized with the 1–205 form of 70k (M1–M3). All sera shown recognize a 32-kDa band in apoptotic but not intact lysates. The human sera and the mouse sera also share recognition of additional bands, including the U1-A (H1, M1, and M3) and Sm B/B′ (H3, M1) components of the U1-small nuclear RNP.

FIGURE 5.

Mouse and human antisera recognize a 32-kDa activity in apoptotic but not intact Jurkat lysates. Control and apoptotic Jurkat lysates were immunoblotted on the same gel with either 1/10,000 dilutions of three autoimmune human sera (H1–H3) or 1/1,000 dilutions of three mice immunized with the 1–205 form of 70k (M1–M3). All sera shown recognize a 32-kDa band in apoptotic but not intact lysates. The human sera and the mouse sera also share recognition of additional bands, including the U1-A (H1, M1, and M3) and Sm B/B′ (H3, M1) components of the U1-small nuclear RNP.

Close modal

Because some of the 1- to 205-immunized mice developed Abs to both the intact and apoptotic forms of 70k, and shared recognition of other bands with human anti-RNP antisera (Fig. 5), we considered the possibility that the immune response against 70k in our model could spread to other epitopes on the 70k endogenous protein. In 2 mice followed with serial blood draws for up to 11 wk after immunization, Abs to a Genenase-cleaved, gel-purified form of the p205 peptide persisted, and IgG Abs to a Genenase-cleaved, gel-purified 342–437 C-terminal fragment of U1-70 kDa appeared later in the course of follow-up, consistent with intramolecular epitope spreading to the endogenous form of the 70k protein (Fig. 6). Overall, in sera taken 1 mo after immunization, 6 of 15 mice immunized with the 1–205 form of 70k that developed Abs to the 1–205 form also had recognition of the C-terminal 70k peptide by ELISA that was over 5 SDs greater than in unimmunized controls (data not shown). This was in contrast to mice that were immunized with the 1–180 form of 70k, in which 0 of 5 mice developed Abs reactive with the endogenous C-terminal peptide of 70k.

FIGURE 6.

Immunization with the 1–205 form of 70k leads to the development of Abs to the C-terminal region of the endogenous 70k Ag. Mice were immunized as described with the p205 peptide of 70k or with collagen II. Unimmunized control mice and responder mice (identified by ELISA recognition of the 1–205 peptide of 70k or the peptide 14 region of collagen II (15 )) were tested by ELISA for recognition of Genenase-cleaved, gel-purified p205 peptide and the Genenase-cleaved, gel-purified 342–437 C-terminal 70k peptide; results with 1/250 serum dilutions are shown. Recognition of p205 is present in immunized mice at 3 wk and persists to 11 wk, whereas recognition of the 70k C-terminal fragment develops between 4 and 11 wk.

FIGURE 6.

Immunization with the 1–205 form of 70k leads to the development of Abs to the C-terminal region of the endogenous 70k Ag. Mice were immunized as described with the p205 peptide of 70k or with collagen II. Unimmunized control mice and responder mice (identified by ELISA recognition of the 1–205 peptide of 70k or the peptide 14 region of collagen II (15 )) were tested by ELISA for recognition of Genenase-cleaved, gel-purified p205 peptide and the Genenase-cleaved, gel-purified 342–437 C-terminal 70k peptide; results with 1/250 serum dilutions are shown. Recognition of p205 is present in immunized mice at 3 wk and persists to 11 wk, whereas recognition of the 70k C-terminal fragment develops between 4 and 11 wk.

Close modal

We also evaluated study mice for additional autoantibodies. We found that 5 of 20 mice immunized with the 1–205 form of 70k developed dsDNA Abs using the Crithidia assay at titers of 1:20. This rate of dsDNA positivity was not different from that seen in controls tested to date, where 1 of 13 mice were anti-dsDNA positive (the sole positive animal was an aged unimmunized mouse that had no histologic evidence of autoimmunity at necropsy; additional dsDNA negative control mice included 1 unimmunized mouse, 3 mice immunized with the 1–180 form of 70k, and 8 mice immunized with MBP). Using a Hep-2 immunofluorescence assay for antinuclear Abs (FANA) at 1:40, we found that 4 of 19 p205-immunized mice were FANA positive (of which 4 of 4 were also anti-32-kDa positive, but 0 of 4 were anti-dsDNA positive), and 0 of 13 controls were FANA positive. The same mice were assayed for FANA as for dsDNA with the exception of 1 of the 1- to 205-immunized mice, for which sufficient serum to do the FANA was not present.

Three mice immunized with the 1–205 form of 70k that developed spreading anti-70k immune responses were sacrificed 1 mo after immunization for histological analysis. Two control mice of the same age from the same colony were sacrificed in parallel. One mouse that responded to immunization with the 1–180 form of 70k was also sacrificed 1 mo after immunization. None of the mice demonstrated clinical features of illness at the time of sacrifice. To our surprise, all three of the 1- to 205-immunized mice but not the control or the 1- to 180-immunized mice showed interstitial and perivascular hypercellular lung lesions, consistent with the proliferative pulmonary lesions reported in human mixed connective tissue disease (Fig. 7), an anti-70k-associated syndrome (4, 18). No renal, cardiac, or hepatic lesions were noted in the mice studied. Further studies are underway to confirm that immunization with the 1–205 form of 70k is associated with features of an anti-70k-associated human systemic rheumatic disease.

FIGURE 7.

Pulmonary lesions in mice immunized with the 1–205 form of 70k (p205). Two control (A and B) and three p205-immunized mice (C–F) were sacrificed 1 mo after immunization to assess for evidence of tissue injury associated with immunization. Formalin-fixed, paraffin-embedded sections were stained with H&E. Representative microscopic fields from the mice are shown. A and C, Shown are comparable fields with the ×10 objective from control and p205-immunized mice. The p205-immunized lung has thickened, hypercellular septae (arrows). B, D, and F, Shown are comparable fields with the ×45 objective from control (B) and p205-immunized mice (D and F), illustrating that the hypercellular interstitial infiltrate in p205-immunized mice includes plexiform vascular lesions (F). E, Shown is a field from a p205-immunized mouse with the ×4 objective, demonstrating the patchy nature of the interstitial thickening (arrows).

FIGURE 7.

Pulmonary lesions in mice immunized with the 1–205 form of 70k (p205). Two control (A and B) and three p205-immunized mice (C–F) were sacrificed 1 mo after immunization to assess for evidence of tissue injury associated with immunization. Formalin-fixed, paraffin-embedded sections were stained with H&E. Representative microscopic fields from the mice are shown. A and C, Shown are comparable fields with the ×10 objective from control and p205-immunized mice. The p205-immunized lung has thickened, hypercellular septae (arrows). B, D, and F, Shown are comparable fields with the ×45 objective from control (B) and p205-immunized mice (D and F), illustrating that the hypercellular interstitial infiltrate in p205-immunized mice includes plexiform vascular lesions (F). E, Shown is a field from a p205-immunized mouse with the ×4 objective, demonstrating the patchy nature of the interstitial thickening (arrows).

Close modal

Apoptotic U1-70-kDa Ag has been proposed to be an immunogenic form of the molecule that may drive autoimmune responses. Among patients with anti-70k immunity that we have analyzed, over half have had evidence of immunity to epitopes present on the apoptotic but not the intact form of the protein (3). In the current study, a group of patient sera was identified that had preferential reactivity with the apoptotic form of 70k. Epitope mapping results with these sera lead to the identification of an epitope present on apoptotic but not intact U1-70 kDa that depends on residues 180–205, within the RNA binding domain of the protein. Abs targeting the 180- to 205-dependent epitope were also common when we looked at sera with Abs to the intact form of 70k that continued to recognize the apoptotic form of 70k after competitive blockade of recognition of the intact form. When we immunized DR4-transgenic C57BL/6 mice with the 1–205 truncated form of 70k, we observed frequent anti-70k immunity, diversifying immune responses that included recognition of proteins in apoptotic but not intact cell lysates, and histologic features of target organ injury. Thus, immunity to this form of 70k appears to be associated with apoptotic recognition, RNA binding domain recognition, and clinical features of human autoimmune disease.

The 180–205 region of 70k on which this B cell epitope depends overlaps with the most frequent T cell epitope of 70k (residues 173–187) that we identified by cloning of human autoimmune anti-70k T cells (11). Because diversification of the B cell response to 70k has been well documented, but the T cell response appears to stay limited to epitopes within the RNA binding domain, it is tempting to consider that the 180–205 region may represent an initial immunodominant epitope for both B and T cell anti-U1-70-kDa responses. We theorize that caspase cleavage of the C terminus of 70k may make the RNA binding domain of 70k, including the 180–205 region, more accessible to immune recognition on the U1-small nuclear RNP molecule (19). If this is true, then other forms of Ag modification in addition to caspase cleavage may also expose this epitope (4, 20).

Although we have emphasized that the 180- to 205-dependent epitope was present on the apoptotic form but not the intact form of 70k, our results do not prove that the apoptotic form of 70k is in fact involved in the development of this immune response. Other modified forms of 70k or molecular mimics of 70k that display the same epitope could also account for this response.

Given the significant rate of diversification of the anti-70k response even to the C-terminal area, we speculate that we could not identify a higher prevalence of 1- to 205-immunized mice exclusively with Abs to the 180- to 205-dependent epitope, because rapid spreading to other areas on the 70k molecule occurs after immunization. In support of the validity of this animal system as a model of human disease, the percentage of 70k-reactive individuals with strongly preferential Abs to the apoptotic form of 70k was similar in our human cohort (7 of 44; 16%) as in our mouse cohort (4 of 13; 30%).

The diversification of anti-70k immunity to a 32-kDa band in apoptotic but not intact lysates was unexpected. The 70k molecule itself does not form a 32-kDa fragment in apoptosis, and anti-70k immunity had not previously been correlated with frequent immune responses to another protein that is preferentially recognized in apoptotic lysates. The identity and clinical significance of this 32-kDa band is under further investigation. Using radiolabeled U1-RNA to hybridize to intact and apoptotic cell lysates, we do not observe a 32-kDa apoptosis-specific activity that binds the U1-RNA (data not shown). It is unclear whether frequent development of the anti-32-kDa response in 1- to 205-immunized animals (and in humans with anti-70k immunity) is due to structural homology between this protein and 70k, or potentially, whether the 32-kDa activity closely associates with modified forms of 70k, facilitating a spreading response.

In our immunoblot experiments, mouse sera were assayed at titers of 1:1,000, whereas human sera were assayed at titers of 1:10,000. This reflects the fact that the human sera come from patients with established clinical disease who have had the potential to develop extremely high Ab titers in response to potentially longstanding and persistent antigenic stimulation. The mice, by contrast, have received only a single immunization with Ag and a brief period of follow-up. We posit that mice at this stage may model what human subjects’ immune responses might have been at the preclinical stage where they initially were breaking tolerance to self Ags, a process that often occurs years before the development of health complaints that lead to clinical autoantibody testing (21). This may also explain why the prevalence of humans with apoptotic-only Abs was somewhat lower than in the mice.

Although no previous studies have performed epitope mapping of B cell responses that recognize apoptotic but not intact 70k, there are several reports in which other B cell epitopes of 70k have been mapped. In some of these studies, peptide forms of 70k extending to residue 195 but not to 205 were used, and 70k epitopes encompassing residues within the 180–205 region were identified (22, 23, 24, 25). We found that a 1–195 form of 70k was not strongly recognized by our apoptosis-only anti-70k sera, an intermediate result between the strong reactivity with the 1–205 peptide and the often absent reactivity of the 1–180 peptide (data not shown). Other investigators (26) have mapped the reactivity of anti-70k sera using octamer peptides. Because even the 180–205 25-mer was insufficient to induce recognition of the epitope we identify, the use of smaller peptides would similarly not be anticipated to recognize this epitope. The lack of recognition of our 180- to 205-dependent 70k epitope in these small peptide studies may therefore be explained by differences in the forms of the Ag studied.

The 1–205 peptide sequence with which we immunized mice is completely homologous between mice and humans (GenBank accession nos. AK011564.1 (mouse) and XM_027872.3 (human)). The immunization that induced apoptosis-specific Abs in the mice can thus be legitimately viewed as an immunization with a modified form of a self protein, rather than as an alloantigen stimulus. Diversification of the immune response from the 1–205 peptide immunogen to a response against the distant 342–437 peptide in immunized mice demonstrates that an immune response that initially targets the epitope we identify can lead to a broader autoimmune response. Preliminary identification of pulmonary lesions in immunized mice similar to those seen in mixed connective tissue disease, albeit in a limited number of mice, provides further evidence that the process induced in this experimental model may be relevant to anti-70k-associated human disease. The response to the 1–205 form is notable in its immunogenicity, because in prior studies in MRL/lpr mice, we found that immunization with either an intact 70k-MBP construct or with MBP alone in CFA led to a delayed anti-RNP response, even though strong anti-MBP immunity was induced (27). Further studies are underway to examine mice immunized with the 1–205 form of 70k for additional autoantibody diversification and development of clinical disease.

In summary, this report has identified an epitope present on the apoptotic but not the intact form of a human autoantigen. The epitope is in the RNA binding domain and distant from the caspase cleavage site in the Ag, suggesting that epitope mapping strategies for other autoantigens need not focus on the region of direct apoptotic modification. Rather, autoimmunity may be influenced by other immunogenic properties of the autoantigen. The B cell epitope identified overlaps with a major T cell epitope of U1-70 kDa, emphasizing the potential for linked B and T cell responses to this Ag. Immunization of a murine model with this epitope induces manifestations suggestive of human anti-70k disease. If further evidence supports the view that this epitope is important for driving human anti-70k autoimmune responses, it may be a candidate to target for therapeutic interventions.

1

This study was supported by Department of Veterans Affairs, Lupus Foundation of America (Kansas City Chapter), Amgen Rheumatology fellowship program, Eastern Missouri Chapter of the Arthritis Foundation, and Grants AI-01842 (to E.L.G.) and AR-43309 (to R.W.H.) from the National Institutes of Health.

3

Abbreviations used in this paper: 70k, U1-70-kDa ribonucleoprotein; RNP, ribonucleoprotein; ENA, extractable nuclear Ag; MBP, maltose-binding protein; FANA, fluorescent antinuclear Ab.

1
Casciola-Rosen, L. A., G. Anhalt, A. Rosen.
1994
. Autoantigens targeted in systemic lupus erythematosus are clustered in two populations of surface structures on apoptotic keratinocytes.
J. Exp. Med.
179
:
1317
.
2
Greidinger, E. L., R. W. Hoffman.
2001
. The appearance of U1 RNP antibody specificities in sequential autoimmune human antisera follows a characteristic order that implicates the U1-70 kd and B′/B proteins as predominant U1 RNP immunogens.
Arthritis Rheum.
44
:
368
.
3
Greidinger, E. L., M. F. Foecking, R. W. Hoffman.
2002
. Apoptotic U1-70 kd is antigenically distinct from the intact form of the U1-70 kd molecule.
Arthritis Rheum.
46
:
1264
.
4
Greidinger, E. L., L. Casciola-Rosen, S. M. Morris, R. W. Hoffman, A. Rosen.
2000
. Autoantibody recognition of distinctly modified forms of the UI-70-kd antigen is associated with different clinical disease manifestations.
Arthritis Rheum.
43
:
881
.
5
St. Clair, E. W., C. C. Query, R. Bentley, J. D. Keene, R. P. Polisson, N. B. Allen, D. S. Caldwell, J. R. Rice, C. Cox, D. S. Pisetsky.
1990
. Expression of autoantibodies to recombinant (U1) RNP-associated 70K antigen in systemic lupus erythematosus.
Clin. Immunol. Immunopathol.
54
:
266
.
6
Earnshaw, W. C., L. M. Martins, S. H. Kaufmann.
1999
. Mammalian caspases: structure, activation, substrates, and functions during apoptosis.
Annu. Rev. Biochem.
68
:
383
.
7
Auer-Grumbach, P., B. Achleitner.
1994
. Epidemiology and clinical associations of NuMa (nuclear mitotic apparatus protein) autoantibodies.
J. Rheumatol.
21
:
1779
.
8
Youinou, P., P. Le Goff, C. B. Colaco, J. Thivolet, D. Tater, J. Viac, M. Shipley.
1985
. Antikeratin antibodies in serum and synovial fluid show specificity for rheumatoid arthritis in a study of connective tissue diseases.
Ann. Rheum. Dis.
44
:
450
.
9
Hardin, J. A., D. R. Rahn, C. Shen, M. R. Lerner, S. L. Wolin, M. D. Rosa, J. A. Steitz.
1982
. Antibodies from patients with connective tissue diseases bind specific subsets of cellular RNA-protein particles.
J. Clin. Invest.
70
:
141
.
10
Satoh, M., V. M. Shaheen, P. N. Kao, T. Okano, M. Shaw, H. Yoshida, H. B. Richards, W. H. Reeves.
1999
. Autoantibodies define a family of proteins with conserved double-stranded RNA-binding domains as well as DNA binding activity.
J. Biol. Chem.
274
:
34598
.
11
Greidinger, E. L., M. F. Foecking, K. R. Schafermeyer, C. W. Bailey, S. L. Primm, D. R. Lee, R. W. Hoffman.
2002
. T cell immunity in connective tissue disease patients targets the RNA binding domain of the U1-70kD small nuclear ribonucleoprotein.
J. Immunol.
169
:
3429
.
12
Greidinger, E. L., D. K. Miller, T. T. Yamin, L. Casciola-Rosen, A. Rosen.
1996
. Sequential activation of three distinct ICE-like activities in Fas-ligated Jurkat cells.
FEBS Lett.
390
:
299
.
13
Monneaux, F., J.-P. Briand, S. Muller.
2000
. B and T cell immune response to small nuclear ribonucleoprotein particles in lupus mice: autoreactive CD4+ T cells recognize a T cell epitope located within the RNP80 motif of the 70K protein.
Eur. J. Immunol.
30
:
2191
.
14
Takeda, Y., G. S. Wang, R. J. Wang, S. K. Anderson, I. Pettersson, S. Amaki, G. C. Sharp.
1989
. Enzyme-linked immunosorbent assay using isolated (U) small nuclear ribonucleoprotein polypeptides as antigens to investigate the clinical significance of autoantibodies to these polypeptides.
Clin. Immunol. Immunopathol.
50
:
213
.
15
Krco, C. J., J. Pawelski, J. Harders, D. McCormick, M. Griffiths, H. S. Luthra, C. S. David.
1996
. Characterization of the antigenic structure of human type II collagen.
J. Immunol.
156
:
2761
.
16
Hoffman, R. W., L. J. Rettenmaier, Y. Takeda, J. E. Hewett, I. Pettersson, U. Nyman, A. M. Lugar, G. C. Sharp.
1990
. Human autoantibodies against the 70-kd polypeptide of U1 small nuclear RNP are associated with HLA-DR4 among connective tissue disease patients.
Arthritis Rheum.
33
:
666
.
17
Ito, K., H. J. Bian, M. Molina, J. Han, J. Magram, E. Saar, C. Belunis, D. R. Bolin, R. Arceo, R. Campbell, et al
1996
. HLA-DR4-IE chimeric class II transgenic, murine class II-deficient mice are susceptible to experimental allergic encephalomyelitis.
J. Exp. Med.
183
:
2635
.
18
Wiener-Kronish, J. P., A. M. Solinger, M. L. Warnock, A. Churg, N. Ordonez, J. A. Golden.
1981
. Severe pulmonary involvement in mixed connective tissue disease.
Am. Rev. Respir. Dis.
124
:
499
.
19
Stark, H., P. Dube, R. Luhrmann, B. Kastner.
2001
. Arrangement of RNA and proteins in the spliceosomal U1 small nuclear ribonucleoprotein particle.
Nature
409
:
539
.
20
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
.
21
Arbuckle, M. R., J. A. James, K. F. Kohlhase, M. V. Rubertone, G. J. Dennis, J. B. Harley.
2001
. Development of anti-dsDNA autoantibodies prior to clinical diagnosis of systemic lupus erythematosus.
Scand. J. Immunol.
54
:
211
.
22
Netter, H. J., H. H. Guldner, C. Szostecki, H. Will.
1990
. Major autoantigenic sites of the (U1) small nuclear ribonucleoprotein-specific 68-kDa protein.
Scand. J. Immunol.
32
:
163
.
23
Cram, D. S., N. Fisicaro, R. L. Coppel, S. Whittingham, L. C. Harrison.
1990
. Mapping of multiple B cell epitopes on the 70-kilodalton autoantigen of the U1 ribonuclear complex.
J. Immunol.
145
:
630
.
24
Nyman, U., I. Lundberg, E. Hedfors, I. Pettersson.
1990
. Recombinant 70-kD protein used for determination of autoantigenic epitopes recognized by anti-RNP sera.
Clin. Exp. Immunol.
81
:
52
.
25
Takeda, Y., U. Nyman, A. Winkler, K. S. Wise, S. O. Hoch, I. Pettersson, S. K. Anderson, R. J. Wang, G. S. Wang, G. C. Sharp.
1991
. Antigenic domains on the U1 small nuclear ribonucleoprotein-associated 70K polypeptide: a comparison of regions selectively recognized by human and mouse autoantibodies and by monoclonal antibodies.
Clin. Immunol. Immunopathol.
61
:
55
.
26
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
.
27
Chang, M., S. E. Walker, R. W. Hoffman.
2000
. Immunization with a bacterial ATP-binding cassette transporter fragment suppresses autoimmunity and prolongs survival in MRL/lpr lupus-prone mice.
Lupus
9
:
1
.