Since apoptotic cell Ags are thought to be a source of self-Ag in systemic lupus erythematosus, we have examined the role of apoptotic cells in the regulation and activation of B cells specific for Sm, a ribonucleoprotein targeted in human and murine lupus. Using Ig-transgenic mice that have a high frequency of anti-Sm B cells, we find that apoptotic cell injection induces a transient splenic B cell response, while simultaneously causing extensive splenic and peritoneal anti-Sm B cell death. In contrast, mice deficient in the clearance of apoptotic cells develop a chronic anti-Sm response beginning at 1–2 mo of age. These mice have expanded marginal zone and B-1 B cell populations and anti-Sm B cells of both types are activated to form Ab-secreting cells. This activation appears to be Ag-specific, suggesting that activation is due to increased availability of apoptotic cell Ags. Since marginal zone and B-1 cells are positively selected, these data suggest a loss of ignorance rather than a loss of tolerance.
Systemic lupus erythematosus is an autoimmune disorder characterized by the production of autoantibodies specific for a variety of autoantigens of nuclear and cytoplasmic origin. Autoantibodies characteristic of this disease include anti-dsDNA and Abs directed against the ribonucleoprotein Sm (1). Although anti-Sm Abs are present in only 15–30% of lupus patients (1), it is a highly specific disease marker and is associated with late onset renal disease and a poor prognosis (2). To better understand the regulation of anti-Sm autoreactive B cells, we have generated Sm-specific H chain transgenic (Tg) 3 mice using an unmutated, rearranged VHJ558 gene from the MRL/lpr anti-Sm hybridoma 2-12 (3, 4, 5). These Tg mice (2-12H) produce substantial numbers of B cells specific for Sm, allowing the tracking of anti-Sm B cells in both nonautoimmune and autoimmune mice. Tolerance to Sm is maintained in nonautoimmune 2-12H Tg mice, since the level of anti-Sm in circulation is not elevated. Many splenic Sm-specific B cells are transitional, suggesting that many do not become mature B-2 cells (3). Low-affinity anti-Sm B cells are able to become B-2, but are anergic (6). In contrast to this regulation in the spleen, the peritoneum harbors a significant population of positively selected, functional anti-Sm B-1 cells (4). Thus, both functional and nonfunctional anti-Sm B cells are present in nonautoimmune mice.
Apoptotic cells are a suspected source of Ag in the activation of autoreactive B cells in human and mouse lupus. Apoptosis results from a series of biochemical and morphological changes. These include shrinkage in cell volume, condensation of the nucleus, and blebbing of the plasma membrane. Small surface blebs become highly enriched for nuclear Ags, including Ro, La, small nuclear ribonucleoproteins, Ku, and DNA (7, 8, 9) that are targets of the immune system in systemic lupus erythematosus. Autoantigens associated with apoptotic blebs could efficiently stimulate B cells possibly by inducing multivalent cross-linking of B cell Ag receptors. Indeed, injection of apoptotic cells into nonautoimmune mice induces the production of autoantibodies characteristic of those seen in human and mouse lupus (10, 11, 12).
Deficiencies in the clearance of apoptotic cells can also lead to autoantibody production and disease. Impaired uptake of apoptotic cells has been implicated in human and mouse lupus (13, 14, 15). Recent murine studies indicate a role for the receptor tyrosine kinase Mer in the clearance of apoptotic cells (15). Mice that have been engineered to have a disrupted tyrosine kinase domain in mer (merkd) develop autoantibodies to nuclear Ags, such as ssDNA, dsDNA, and chromatin (16) and develop a lupus-like disease (14, 15). The disease is more severe if all three Tyro-3 family members, Tyro-3, Axl, and Mer, are lacking (14). The absence of polyclonal B cell activation in merkd mice (16) suggests that only B cells specific to apoptotic Ags are activated. Because B cells do not express these receptors and B cell function is normal in Tyro-3-deficient mice (14), the activation of B cells is likely the result of stimulation by apoptotic cell Ags (15). C1q-deficient mice have a marked accumulation of apoptotic cells in the kidney (17) and C1q deficiency also leads to the development of a severe lupus-like disease in both humans and mice (18). Thus, defects in apoptotic cell clearance can contribute to the development of lupus in mice and humans. To further examine the effect of apoptotic cells on the regulation and activation of autoreactive B cells, we have followed the regulation and activation of anti-Sm B cells upon acute and chronic exposure to apoptotic cell Ags. We demonstrate that apoptotic cells activate the positively selected Sm-specific MZ B cells in the spleen and B-1 cells in the peritoneum.
Materials and Methods
2-12H Tg and 6-1 Tg mice have been described previously (3, 4, 5, 19) and are maintained by backcrossing to C.B-17 mice. Tg offspring are identified by PCR analysis of tail genomic DNA as previously described (3, 19). Merkd mice were obtained from Dr. G. Matsushima (University of North Carolina at Chapel Hill) and bred with 2-12H Tg and 6-1 Tg mice. Merkd mice were identified by Southern blot analysis of tail genomic DNA as described by Camenisch et al. (20).
Flow cytometry was done and data were analyzed as previously described (4). The Abs specific for IgMa (DS-1), IgM (II/41), B220 (RA3-6B2), CD11b (M1/70), CD21(7G6), CD23 (B3B4), CD43 (S7), CD5 (53-7.3), CD40 (HM40-3), CD80 (16-10A1), and CD86 (GL1) were obtained from BD PharMingen (San Diego, CA) and were labeled with FITC, PE, allophycocyanin, or biotin.
Immunization with apoptotic cells
Apoptotic cells were prepared as described by Mevorach et al. (10) with slight modification. Briefly, thymocytes were prepared from young mice (5–6 wk) and irradiated at 600 rad. The irradiated thymocytes were cultured overnight and ∼5 × 107 cells were injected i.v. into the recipient mice.
Both apoptotic and fresh thymocytes were stained with the anti-Sm Vκ31T hybridoma Ab for 15 min and washed with PBS twice. The cells were then stained with annexin V in binding buffer (BD PharMingen) for 15 min and washed twice with RPMI 1640 medium (HyClone Laboratories, Logan, UT) containing 0.1% sodium azide and 3.0% FBS (HyClone Laboratories). After staining with anti-mouse IgM-Cy3 (Jackson ImmunoResearch Laboratories, West Grove, PA) and Alexa Fluor 488 conjugate (Molecular Probes, Eugene, OR), the cells were washed twice with RPMI 1640/FBS/NaN3 and transferred to poly-d-lysine-coated coverslips. The cells on the coverslips were fixed with 3% sucrose and 3% paraformaldehyde in PBS. The slides were then mounted and analyzed using a LeicaTCSS-NT confocal microscope (Leica Microsystems, Bannockburn, IL).
2-12H Tg mouse peritoneal cell transfer
Approximately 5 × 106 peritoneal cells isolated from 2-12H Tg mice were transferred i.p. into non-Tg littermate or non-Tg merkd mice. Sera were collected at 1 and 2 wk after transfer. Spleen and peritoneal cells were analyzed by flow cytometry 2 wk after transfer.
Quantitation of anti-Sm Abs and total IgM in mouse serum was done by ELISA as previously described (3).
ImmunoSpot Mid Plates (Cellular Technology, Cleveland, OH) were coated with 10 U/well Sm Ag (Immunovision, Springdale, AR) in PBS overnight at 4°C. The plates were washed five times in PBS and then blocked with 1% BSA in PBS. Lymphocytes isolated from different tissues were resuspended in HL-1 medium (BioWhittaker, Walkerville, MD) supplemented with 1% l-glutamine and 1% penicillin/streptomycin. One million or 2 × 105 cells were added to each well and incubated for 24 h at 37°C. Plates were washed with PBS and then 0.05% Tween 20 in PBS (PBST). Biotin-labeled anti-IgMa or IgMb Ab (BD PharMingen) diluted in 1% BSA in PBST was added to each well and incubated overnight at 4°C. Plates were then washed with PBST and streptavidin-HRP (BD PharMingen) diluted in 1% BSA in PBST was added to each well and the plates were incubated for 2 h at room temperature. After washing with PBST and PBS, the plates were developed with 3-amino-9-ethylcarbazole (Sigma-Aldrich, St. Louis, MO) in 3-amino-9-ethylcarbazole buffer. ELISPOT on sorted splenic marginal zone (MZ) and follicular (FO) B cells was performed according to the method described by Grimaldi et al. (21) with slight modification. The plates were scanned and analyzed using an ELISPOT Analyzer (Cellular Technology, Cleveland, OH).
Statistical analysis was performed using the one-tailed Student’s t test. A value of p < 0.05 was considered to be significant.
Ribonucleoprotein Sm is associated with apoptotic cell blebs
To determine whether Sm is associated with apoptotic cells, apoptotic thymocytes were stained with the anti-Sm mAb Vκ31T (22). Four hours postirradiation thymocytes stain with annexin V, indicating apoptosis (23, 24, 25, 26), but not with Vκ31T (data not shown). However, by 24 h about half of the annexin V-staining thymocytes also stain with Vκ31T, but not an isotype control IgM mAb (Fig. 1,A). Similar results were obtained at 48 and 72 h postirradiation (data not shown). Thus, anti-Sm staining occurs only during the late stages of apoptosis similar to that seen with anti-DNA binding. To assess the localization of Sm on apoptotic cells, we examined Vκ31T binding to apoptotic cells by fluorescence microscopy. As shown in Fig. 1,B, Sm appears to be associated with apoptotic blebs, and like anti-DNA binding (9) does not colocalize with annexin V staining. Neither annexin V nor Vκ31T stain nonirradiated live thymocytes (Fig. 1 B).
Apoptotic cells induce both activation and loss of anti-Sm B cells
To determine whether the Sm associated with apoptotic cells can activate anti-Sm B cells, 2-12H Tg and non-Tg mice were given a single i.v. injection of 3–5 × 107 apoptotic thymocytes. Serum anti-Sm levels increase in both 2-12H Tg and non-Tg mice, peaking at day 6 and returning to normal by day 15 (Fig. 2,A, left). The response to Sm by non-Tg mice is expected since they have anti-Sm B cells in both the spleen and peritoneum (3, 4). Anti-Sm B cells in 2-12H Tg mice are not activated by immunization with the cell lysate from an equivalent number of thymocytes (Fig. 2,A, left) or by immunization with an equivalent number of live cells (data not shown). 2-12H Tg mice were also given a series of four weekly injections (days 0, 7, 14, and 21) followed by two more injections on days 60 and 67, and serum anti-Sm levels were determined at multiple times as indicated (Fig. 2 A, right). Serum anti-Sm levels in mice receiving multiple injections never exceed those seen after only a single injection. Thus, there is no evidence of an enhanced secondary response.
To assess the impact of apoptotic cell injection at a cellular level, we examined the splenic and peritoneal B cell populations in 2-12H Tg and non-Tg mice after injection. By 3 days postinjection, nearly half of the anti-Sm B cells in the spleen and peritoneum are depleted (Fig. 2,B, lower left and right). These populations return to normal by day 15. Total B cell numbers are reduced in the spleen (Fig. 2,B, upper left), although not significantly (p = 0.0578), but are significantly decreased in the peritoneum (p < 0.01; Fig. 2,B, upper right). Apoptotic cell injection appears to affect anti-Sm B cells preferentially, since there is no detectable effect on the total number of splenic or peritoneal B cells in non-Tg mice injected with apoptotic cells (Fig. 2 B, upper left and right), suggesting that anti-Sm B cell depletion is B cell receptor driven and involves cells specific for apoptotic cell Ags.
Depletion of anti-Sm B cells following apoptotic cell injection could be due to differentiation of anti-Sm B cells to Ab-secreting cells (ASCs) or to Ag-induced cell death. To address this issue, the number of ASCs generated by apoptotic cell injection was determined. There is a 3-fold increase in anti-Sm ASCs in the spleen 6 days after apoptotic cell injection (Fig. 2,C) when Ab levels have reached their peak, and no change in anti-Sm ASCs in the bone marrow (BM), mesenteric lymph node (MLN), or lamina propria (LP) (Fig. 2,C). Although we have been unable to detect anti-Sm B cells undergoing cell death, the number of anti-Sm B cells lost from the spleen (∼7.5 × 106) is three orders of magnitude larger than the number of anti-Sm plasma cells gained in the spleen (∼8 × 103), suggesting that the majority of anti-Sm B cells lost have undergone cell death. The absence of anti-Sm ASCs in MLN and LP (Fig. 2 C) suggests that anti-Sm B-1 cells are also deleted rather than activated upon apoptotic cell injection, since the MLN and LP are the primary sites for plasma cells of B-1 origin (Refs.27, 28, 29, 30 , and see below). Taken together, these data indicate that the majority of splenic and peritoneal anti-Sm B cells are induced to undergo cell death upon apoptotic cell injection, while only a small number of splenic B cells are induced to differentiate to ASCs.
A deficiency in apoptotic cell clearance results in anti-Sm B cell activation
To determine whether a deficiency in apoptotic cell clearance leads to an anti-Sm response, 2-12H Tg and non-Tg mice homozygous for the merkd mutation were generated (2-12H/merkd and non-Tg/merkd mice, respectively). Anti-Sm Ab is first detected in about half the mice at 6 wk, and all mice are anti-Sm positive by 8 wk (Fig. 3 A). The level of anti-Sm remains relatively constant after 2 mo. 2-12H Tg and non-Tg mice have very low levels of anti-Sm. However, 4 of 22 non-Tg/merkd mice have serum anti-Sm at levels similar to those of 2-12H/merkd mice, indicating that anti-Sm responses occur in non-Tg merkd mice.
Anti-Sm production by merkd mice was also assessed by ELISPOT to identify the location of anti-Sm ASCs. 2-12H/merkd mice have significantly more anti-Sm ASCs in their BM, spleen, MLN, and LP than 2-12H Tg and non-Tg control mice (Fig. 3,B). The presence of anti-Sm ASCs in BM and spleen suggests a splenic B cell response (31, 32, 33, 34, 35, 36), while the presence of anti-Sm ASCs in the MLN and LP suggests a peritoneal B-1 cell response (27, 28, 29, 30). To determine whether FO (transitional and mature) or MZ B cells are involved in this response, cells of each population were sorted and analyzed by ELISPOT. As shown in Fig. 3 C, approximately four times as many 2-12H/merkd MZ B cells as 2-12H MZ B cells secrete anti-Sm, and negligible numbers of FO B cells from mice of either strain secrete anti-Sm. Thus, anti-Sm MZ B cells are activated and form ASCs in 2-12H/merkd mice.
To determine the effect of the merkd mutation on B cell differentiation, we examined B cells from Tg and non-Tg merkd mice by flow cytometry. The merkd mutation has little effect on the total splenic B cell numbers, as seen by the comparison of B cell numbers in non-Tg and non-Tg/merkd mice (Fig. 4,A, left). However, the merkd mutation in 2-12H Tg mice causes a significant reduction in the total B cell number and in the number of splenic anti-Sm B cells (cf 2-12H and 2-12H/merkd mice in Fig. 4,A, left and right). Not just anti-Sm B cells are affected, as anti-Sm B cells account for only ∼3 × 106 of the ∼15 × 106 cells lost from the spleen (Fig. 4 A). The additional B cells affected by the merkd mutation could be low-affinity anti-Sm B cells that do not stain well with Sm (6) or cells of other apoptotic cell specificities, such as the anti-ssDNA B cells present in these mice (3).
To more carefully examine how the merkd mutation affects the differentiation of anti-Sm B cells, we examined the distribution of splenic anti-Sm B cells into splenic B cell subsets. Using IgM, CD21, and CD23 to distinguish between FO, transitional 1 (T1), and MZ B cells, we find that a significantly (p < 0.05) higher percentage of anti-Sm B cells are MZ in 2-12H/merkd mice (34.6 ± 2.0%, n = 6) than in 2-12H Tg mice (18.8 ± 1.8%, n = 6; Fig. 4,B). In addition, 2-12H/merkd mice have a greater number of anti-Sm MZ B cells than 2-12H Tg mice (Fig. 4,C). The merkd mutation also increases the frequency (7.4 ± 0.67% in non-Tg merkd mice vs 4.3 ± 0.41% in wild-type mice; p < 0.05, n = 5) and number (Fig. 4,C) of MZ B cells in non-Tg mice. This increase cannot be attributed solely to an increase in anti-Sm B cell numbers, although this increase is significant (Fig. 4,C). Splenic anti-Sm B cells of 2-12H/merkd mice are generally larger (forward scatter), more granular (side scatter), and express higher levels of CD40, CD80, and CD86 than non-Tg mice (Fig. 4,D), suggesting that they have encountered Ag. In comparison to 2-12H Tg mice, it appears that a greater number of anti-Sm B cells in 2-12H/merkd mice express high levels of these markers (Fig. 4 D).
Because anti-Sm B cells are selected into the peritoneal B-1 cell population, we examined how the merkd mutation affected this population. The peritoneal anti-Sm B cells of 2-12H/merkd mice are B-1 cells (Fig. 5,A). Non-Tg merkd and 2-12H/merkd mice have roughly twice the number of B-1 cells as their mer wild-type counterparts (Fig. 5,B). However, 2-12H/merkd mice have fewer anti-Sm B-1 cells and their number declines as mice age from a peak at 2 mo. In contrast, the number of anti-Sm B-1 cells in 2-12H Tg mice remains constant. Despite this decline in anti-Sm B-1 cell numbers in 2-12H/merkd mice, the total number of peritoneal B-1 cells remains constant as mice age (Fig. 5 B), indicating that the loss of anti-Sm B-1 cells is specific for a subset of B-1 cells and is compensated by an increase in the number of B-1 cells of other specificities.
The loss of anti-Sm B-1 cells from the peritoneum is due at least partly to activation
The gradual loss of anti-Sm B-1 cells from the peritoneum of 2-12H/merkd mice could be due to either activation or deletion. Activation is suggested by the presence of anti-Sm ASCs in the LP and MLN. To determine whether anti-Sm B-1 cells in 2-12H/merkd mice are activated, 5 × 106 peritoneal cells were transferred i.p. to wild-type and merkd recipient mice. Two weeks after transfer, ∼1.5 × 106 B cells of donor origin were recovered from wild-type recipients, but only ∼3 × 105 were recovered from merkd recipients. Significantly, donor anti-Sm B-1 cells were present at the same frequency in the recovered cells from wild-type recipients as in donor peritoneum, but donor anti-Sm B-1 cells were nearly absent from recovered cells from merkd recipients (Fig. 6,A). Serum anti-Sm levels in merkd recipients increased at 1 wk after transfer and again in some mice at 2 wk after transfer (Fig. 6,B). In addition, ELISPOT assay indicates that donor anti-Sm ASCs are present in MLNs and the LP, but not in spleen or BM (Fig. 6,C). Neither increased serum anti-Sm nor increased anti-Sm ASCs occur in wild-type recipients. Not all recipient merkd mice developed elevated levels of serum anti-Sm (Fig. 6 B), even though the number of peritoneal anti-Sm B-1 cells decreased in these same mice (data not shown). Thus, at least some peritoneal anti-Sm B-1 cells from 2-12H/merkd differentiate to ASCs found in the MLNs and LP, accounting for at least some of the anti-Sm B-1 cells lost from the peritoneum.
Phosphatidylcholine (PtC)-specific B-1 cells are not affected by the merkd mutation
The decline in the number of anti-Sm B-1 cells, but not the total number of B-1 cells in the peritoneum of 2-12H/merkd mice, suggests that only B-1 cells specific for apoptotic cell Ags are affected by the merkd mutation. To more carefully test this possibility, the effect of the merkd mutation on PtC-specific B-1 cells was examined, since anti-PtC Abs do not stain apoptotic cells (data not shown). 6-1 anti-PtC Tg mice homozygous for the merkd mutation (6-1/merkd) were bred and analyzed. 6-1 Tg mice express a rearranged VH12 H chain gene that encodes PtC-specific Abs, and they have a greatly expanded anti-PtC B-1 cell population (19). 6-1/merkd and 6-1 mice have comparable numbers of PtC-specific B cells in the spleen and peritoneum (Fig. 7,A) and these cells have a B-1 phenotype (IgMhigh, CD23−, CD43+, and CD5+; Fig. 7,B), indicating that anti-PtC B-1 cell development is not grossly altered. The number of anti-PtC B-1 cells in 6-1/merkd mice does not decrease over time and there is no significant increase in serum IgMa levels (data not shown), suggesting that there is no activation-induced depletion of B-1 cells from the peritoneum. This was verified by transfer of 6-1 peritoneal anti-PtC B-1 cells to wild-type and merkd mice. As shown in Fig. 7 C, 2 wk after transfer PtC-specific B-1 cells of donor origin persist in both wild-type and merkd recipients and at nearly equal numbers (1.4 ± 0.7 × 106 and 1.8 ± 0.5 × 106, respectively, n = 3). Thus, anti-PtC B-1 cells are not depleted from merkd recipients, consistent with the idea that the merkd mutation affects only B-1 cells specific for apoptotic cell Ags.
Apoptotic cells are considered a source of Ag in lupus because many of the self-Ags targeted in this disease are associated with apoptotic blebs (8, 26, 37, 38). Apoptotic cell immunizations support this idea (10, 11, 12), as does the evidence that an apoptotic cell Ag is selected in autoimmunity (9). We demonstrate here that the Sm autoantigen targeted uniquely in lupus is also associated with apoptotic cell blebs and that apoptotic cells induce an anti-Sm response in nonautoimmune mice.
The response to apoptotic cell immunization by non-Tg and 2-12H Tg mice may be T independent. Multiple apoptotic cell injections do not increase anti-Sm levels above those generated by a single injection, suggesting that memory cells are not formed, which would be expected to occur in a T-dependent response. This is similar to the anti-ssDNA and anticardiolipin responses induced by apoptotic cell immunization (9). The anti-Sm response to apoptotic cells by non-Tg mice indicates that the anti-Sm response is not an artifact of 2-12H Tg mice.
The most significant effect of apoptotic cell immunization on anti-Sm B cells appears to be the loss of anti-Sm B cells. Nearly half of the splenic anti-Sm B cells (∼7 × 106) are lost by day 3 (Fig. 2,B), while only ∼3.6 × 103 ASCs/spleen are formed (Fig. 2,C). Also, nearly half of the peritoneal anti-Sm B-1 cells (∼5 × 105) are lost (Fig. 2,B). Cell death may be explained by the fact that transitional and B-1 B cells undergo apoptosis upon B cell receptor cross-linking with apoptotic cell Ags (39, 40, 41, 42) or because of differences between anti-Sm B cells in their affinity for Sm. A particularly interesting possibility is that these cells have been activated to become plasmablasts, but are unable to differentiate to plasma cells because there is an insufficient number of dendritic cells to mediate this differentiative step. MacLennan and colleagues (43) have shown that dendritic cells are required for differentiation of plasmablasts to plasma cells and that this requirement is effectively a bottleneck to the response in Ig Tg mice in which large numbers of B cells are activated. Most Ag-specific B cells die as plasmablasts because they are unable to associate with dendritic cells. As a result, the serum Ab levels in Tg and non-Tg mice are not different. This possibility could also account for why non-Tg and 2-12H Tg mice have equivalent Ab responses to apoptotic cell injection (Fig. 2). We exclude the involvement of a nonspecific mechanism since apoptotic cell immunization of non-Tg mice has no effect on splenic B cell and peritoneal B-1 cell numbers (Fig. 2 B). Because B cell numbers, at least in the peritoneum, as measured by B220 expression, are also reduced, we exclude capping of surface IgM as a reason for decrease in anti-Sm B cell numbers.
The merkd mutation induces a chronic anti-Sm response in 2-12H Tg mice that begins between 1 and 2 mo of age (Fig. 3). Although at a lower prevalence, non-Tg/merkd littermates also develop an anti-Sm response; 4 (18%) of 22 non-Tg merkd mice over the age of 2 mo have serum anti-Sm at levels comparable to those in 2-12H/merkd mice. Since B cells do not express mer (15), the occurrence of an anti-Sm response in merkd mice does not require an intrinsic B cell defect. The activation of 2-12H anti-Sm B-1 cells after transfer to non-Tg merkd recipient mice confirms this (Fig. 6). Interestingly, a defect in macrophage clearance of apoptotic cells has been noted in human lupus and in autoimmune MRL/Mp mice (13, 44, 45), suggesting that the initiation of an anti-Sm response in MRL/Mp mice and lupus patients does not require a B cell intrinsic defect. However, a role for a B cell intrinsic defect in murine or human lupus cannot be excluded.
Anti-Sm MZ B cells are activated in 2-12H/merkd mice. The merkd mutation nearly doubles the size and proportion of the MZ population in non-Tg mice and nearly doubles the number of anti-Sm MZ B cells in 2-12H Tg mice (Fig. 4). In addition, some anti-Sm MZ B cells are activated to form ASCs (Fig. 3 C). These ELISPOTs are smaller than those detected by analysis of unsorted spleen and BM and are likely plasmablasts and precursors to the larger ELISPOTs formed by plasma cells detected with unsorted spleen cells. Whether other splenic B cells differentiate to ASCs has yet to be determined. These data add to an increasing body of evidence that MZ B cells contribute to autoantibody production in murine lupus. Anti-DNA MZ B cells are activated in a murine model of estrogen-induced lupus (21) and are activated in anti-DNA Tg MRL/lpr mice (46, 47) and NZB/NZW mice (48, 49, 50). However, BXSB mice are impaired in their ability to generate MZ B cells, indicating that MZ B cells are not involved in all lupus models (51).
The activation of anti-Sm MZ B cells could be due to the increased availability of apoptotic cell Ags or to nonspecific mechanisms caused by the effects of merkd on non-B cells. We favor the idea that an increased availability of apoptotic cell Ags is responsible since the number of anti-Sm B cells is increased in 2-12H Tg mice, but not the total number of MZ B cells (Fig. 4,C), indicating a specificity restricted expansion of the MZ population in these mice. In 2-12H Tg mice, increased apoptotic cell Ags would affect primarily anti-Sm B cells because of the bias conferred by the transgene, but in non-Tg/merkd mice this would presumably include anti-Sm B cells (Fig. 4 C) and B cells specific for other apoptotic cell Ags.
B-1 cells also contribute to the anti-Sm response in merkd mice. The number of B-1 cells in non-Tg/merkd and 2-12H/merkd mice is nearly twice that in their mer wild-type counterparts (Fig. 5,B) and anti-Sm B-1 cells differentiate into ASCs (Fig. 6). Transfer of 2-12H Tg peritoneal cells to merkd mice demonstrates that anti-Sm B-1 cells generate ASCs in the LP and MLNs and contribute to serum anti-Sm. We note that some mice did not show an increase in serum anti-Sm even though they developed increased numbers of anti-Sm ASCs in the MLNs and LP. The reason for this is unknown. The lack of serum anti-Sm could be due to the small number of anti-Sm B-1 cells transferred and to the secreted Ab being bound up by apoptotic cells. As with anti-Sm MZ B cells, the activation of B-1 cells appears to be dependent on specificity for apoptotic cells. Anti-Sm B-1 cells are selectively lost from the peritoneum and anti-PtC B-1 cells are unaffected by the merkd mutation (Fig. 7). B-1 cells are implicated in the production of other serum autoantibodies in murine lupus. Autoimmune NZB mice have an expanded B-1 cell population (52, 53, 54, 55), and the sle2 lupus susceptibility locus results in an expanded B-1 cell population (55). In addition, B-1 cells produce antierythrocyte autoantibody in autoimmune hemolytic anemia of Ig Tg mice homozygous for the lpr mutation (30). Thus, B-1 cells may be a significant source of some autoantibodies in murine lupus.
The decline in B-1 cell numbers in 2-12H/merkd mice as they age (Fig. 5,B) was seen previously in 2-12H MRL/lpr mice (5) and in B6/lpr mice (56). In 2-12H/merkd mice, this decline can be at least partly explained by migration to MLNs and LP, as indicated by the transfer experiments (Fig. 6). Activation may also explain the loss of anti-Sm B-1 cells in 2-12H MRL/lpr mice, as they too have elevated numbers of anti-Sm ASCs in the MLNs and LP (Y. Qian, K. Conway, and S. H. Clarke, unpublished observation). However, both Sm-binding and nonbinding B-1 cells are affected in 2-12H MRL/lpr mice and B6/lpr mice, suggesting that the underlying reason for the decline in MRL/lpr and B6/lpr mice differs from that in 2-12H/merkd mice and involve mechanisms that are independent of B cell specificity.
The activation of anti-Sm MZ and B-1 cells in this model, in which presumably the level of Ag is altered, has implications for the pathogenesis of lupus. The loss of tolerance and an intrinsic B cell defect are not necessarily required. Defects in apoptotic cell clearance, as has been detected in MRL mice (45), may be sufficient to at least initiate an anti-Sm response from functional MZ and B-1 cells. However, intrinsic defects may play a role in activation, e.g., by increasing sensitivity to Ag. We have demonstrated that increased sensitivity can activate anti-Sm B cells that overexpress CD19 or do not express CD22 (4). One critical issue that has yet to be addressed in merkd mice is the status of anti-Sm T cells. In lupus models, loss of T cell tolerance would lead to affinity maturation and H chain class switch. T cells can also activate anergic B cells (57) and thereby expand the repertoire of anti-Sm B cells involved in the response beyond MZ and B-1 cells. It is not clear whether T cell tolerance to Sm is lost in merkd mice, but anti-Sm T cells are present in 2-12H MRL/lpr mice. It will be interesting to determine the status of anti-Sm T cells in merkd mice and the effect of T cells on the anti-Sm B cell response.
In summary, merkd mice have expanded splenic MZ and peritoneal B-1 cell populations and an Ag-driven mechanism activates the anti-Sm B cells of both subsets. MZ and B-1 cells are normally positively selected and contribute to the initial response to pathogenic organisms (58). We suggest that under normal circumstances the level of apoptotic cell Ags to which the anti-Sm MZ and B-1 cells are exposed is insufficient to drive differentiation to ASCs. However, an increase in the availability of apoptotic cell Ags, as through a deficiency in the clearance of apoptotic cells, drives differentiation of anti-Sm MZ and B-1 cells to ASCs. Thus, the activation of anti-Sm MZ and B-1 cells in merkd mice is more consistent with a loss of ignorance than a loss of tolerance.
We gratefully acknowledge the Flow Cytometry Facility and the Physiology Imaging and Neuroscience Confocal Facility at the University of North Carolina for their assistance.
This work was supported by National Institutes of Health Grants AI29576, AI43587, and AI48085. Y.Q. is a recipient of an Arthritis Foundation Postdoctoral Fellowship.
Abbreviations used in this paper: Tg, transgenic; BM, bone marrow; ASC, Ab-secreting cell; MLN, mesenteric lymph node; LP, lamina propria; MZ, marginal zone; FO, follicular; T1, transitional 1; PtC, phosphatidylcholine.