Autoreactive MZ and B-1 B-cell activation by Fas^lpr is coincident with an increased frequency of apoptotic lymphocytes and a defect in macrophage clearance

A goal of B-cell tolerance studies is to understand how tolerance is lost in autoimmune diseases such as systemic lupus erythematosus (SLE). Mice of SLE-susceptible strains carry unique sets of susceptibility genes.¹  MRL/Fas^lpr mice develop a disease that closely resembles human SLE including the production of anti-Sm, a marker SLE antibody.²  The MRL bkg is required for anti-Sm production,³  but the genes predisposing to autoimmunity have not yet been identified. One gene that can contribute to both murine and human autoimmunity is the proapoptosis gene Fas.^4-6  The lpr mutation of Fas (Fas^lpr) causes a loss of function⁴  and is sufficient to induce SLE, but the severity is dependent on the bkg genes. It accelerates and exacerbates the disease induced by the MRL bkg,⁷  but induces only a mild disease in C57BL/6 (B6) mice.⁷  Despite producing a wide spectrum of autoantibodies in B6 mice, Fas^lpr does not induce anti-Sm production in these mice.⁸ 

To examine how autoreactive B cells are regulated, we developed 2-12H transgenic (Tg) mice expressing an anti-Sm H chain transgene.⁹  Both positively and negatively selected anti-Sm B cells coexist in 2-12H mice, yet they are not activated to secrete antibody. Anti-Sm B cells constitute approximately 30% of the follicular (FO) B-cell subset, a majority of MZ B cells, and approximately 30% of peritoneal B-1 cells in 2-12H mice. Regulation of FO anti-Sm B cells occurs by anergy and developmental arrest,^9,10  but how the anti-Sm MZ and B-1 B cells are regulated is unknown. We have recently determined that anti-Sm B cells are activated and begin plasma cell (PC) differentiation, but arrest at a pre-PC stage before becoming antibody secreting cells (ASCs).¹¹  However, in 2-12H MRL/Fas^lpr mice, anti-Sm pre-PCs overcome this block and become ASCs, producing elevated anti-Sm levels with complete penetrance by 2 months of age.¹² 

MZ and B-1 cells are likely sources of autoantibodies in MRL/Fas^lpr mice, as the normal B-1 and MZ B-cell repertoires contain a high frequency of anti-self B cells including anti-DNA.^13,14  MZ B-cell involvement in autoantibody production has been documented in an Ig Tg model where these cells produce anti-ssDNA upon estrogen treatment.¹⁵  Also, excess BAFF expression rescues self-reactive B cells from peripheral deletion and leads to autoimmunity,^16,17  and interestingly induces an expansion of the MZ B-cell subset,¹⁸  implying a role for this population. Antierythrocyte B-1 cells are responsible for the production of the hemolytic autoantibodies in Fas^lpr mice,¹⁹  implicating a role for B-1 cells in autoantibody production. Also, the sle2 locus of NZW origin contributes to autoimmunity and is responsible for, among other effects, an expansion of the B-1a cell population.^20,21 

We have recently demonstrated anti-Sm MZ and B-1 B cells are activated in 2-12H Tg mice carrying a mutation of the receptor tyrosine kinase Mer (Mer^kd).²² Mer^kd macrophages are impaired in their ability to phagocytize apoptotic cells, and Mer^kd mice develop a mild lupuslike disease with the production of autoantibodies.^22,23  The anti-Sm MZ population is expanded in 2-12H/Mer^kd mice, and there is a progressive loss of peritoneal anti-Sm B-1 cells in young adult mice, which is caused at least in part by B-1–cell activation and differentiation to ASCs of the mesenteric lymph nodes (MLNs) and lamina propria (LP). Apoptotic cells are a likely source of Sm antigen in vivo, as Sm is exposed on the surface of apoptotic cell blebs,²²  similar to other nuclear antigens,²⁴  and immunization of nonautoimmune mice results in a transient anti-Sm response.²²  Thus, the increased availability of apoptotic cells caused by Mer^kd mice could contribute to the induction of an anti-Sm response in 2-12H/Mer^kd mice, or Mer^kd could alter the function of dendritic cells or macrophages that indirectly lead to defects in T-cell and B-cell regulation. Since Fas has an opposing function in apoptosis to Mer, we sought to understand how Fas^lpr differs from Mer^kd in the dysregulation of anti-Sm B cells. In this report, we show that similar to Mer^kd, Fas^lpr induces the activation of anti-Sm MZ and B-1 B cells by an antigen-specific mechanism. This unexpected similarity can be explained by the observations that Fas^lpr also induces an increase in apoptotic cell frequency in lymphoid tissues and a defect in macrophage phagocytosis of apoptotic cells.

2-12H, 2-12H MRL/Fas^lpr, 2-12H/Mer^kd, and 6-1 mice have been described.^{9,12,22,25,26}  2-12H/MRL mice were generated by backcrossing 2-12H MRL/Fas^lpr mice with Fas intact MRL/Mp^+/+ (MRL) mice (Jackson Laboratory, Bar Harbor, ME) through 2 generations to eliminate the Fas^lpr mutation. 2-12H and 6-1 mice were crossed with C57BL/6 Fas^lpr mice (Jackson Laboratory) to generate 2-12H/Fas^lpr and 6-1/Fas^lpr mice. Ig H chain transgenes of offspring were identified by polymerase chain reaction (PCR) as previously described.^9,25 Fas and Fas^lpr were identified by analysis of tail genomic DNA using the forward primer CAA GCC GTG CCC TAG GAA ACA CAG, and reverse primers GCA GAG ATG CTA AGC AGC AGC CGG and GTG GAG CTC CAA TGC AGC GTT CCT. All animal experiments were carried out with institutional IACUC approval.

The antibodies specific for IgM^a (DS-1), IgM^b (AF6-78), B220 (RA3-6B2), CD11b (M1/70), CD21 (7G6), CD23 (B3B4), CD43 (S7), and CD5 (53-7.3) were obtained from PharMingen (San Diego, CA), and were labeled with FITC, PE, APC, or biotin. Biotinylated Sm (Immunovision, Scottsdale, AZ) was used to stain for Sm-specific B cells. Biotinylated reagents were revealed with streptavidin-PerCP. Cells were stained as described before.^22,26  Liposomes encapsulating FITC and composed of membranes containing PtC were used as described to identify PtC-specific B cells.^25,27  CaspACE FITC-VAD-FMK was purchased from Promega (Madison, WI), and cells were stained according to the manufacturer's instructions. Data were analyzed using WinMDI (Scripps Institute, La Jolla, CA). All data represent cells that fall within the lymphocyte gate determined by forward and 90-degree light scatter. Per sample, 1 to 5 × 10⁵ cells were analyzed.

For cell sorting of FO, MZ, and CD138⁺ B cells for enzyme-linked immunospot (ELISpot) analysis, spleen cells were stained for CD19, CD21, CD23, and CD138. The CD19⁺, CD21^hi, CD23^–/lo cells were sorted as MZ B cells, the CD19⁺, CD21^low, CD23⁺, CD138^– cells were sorted as FO B cells, and the CD19⁺, CD138⁺ cells were sorted as pre-PCs. For cell sorting of peritoneal B-1 and B-2 cells for transfer, peritoneal cells were stained for CD19 and CD11b and the CD19⁺ CD11b⁺ (B-1) and CD19⁺ CD11b^– (B-2) cells sorted. The cells were sorted using a MoFlo high-speed sorter (DakoCytomation, Fort Collins, CO). Sorted populations were more than 90% pure as determined by reanalysis.

Sorted peritoneal B-1 and B-2 cells (1 × 10⁵) were injected intraperitoneally into non-Tg recipients. In other experiments, approximately 5 × 10⁶ peritoneal cells isolated from 2-12H Tg or 2-12H/Fas^lpr mice were transferred intraperitoneally into non-Tg wild-type or non-Tg Fas^lpr mice. MLN and LP ELISpot assays were conducted at 2 weeks after transfer.

Quantification of anti-Sm antibodies and total IgM in mouse serum was performed by enzyme-linked immunosorbent assay (ELISA) as previously described.^9,22,26  A capture ELISA was used to measure levels of serum IL-10 as described previously.²⁸  Anti-Sm and IgM ELISpot assays were conducted as described previously using Sm- and anti-IgM–coated plates.²²  Small ELISpots were between 1 × 10^–4 and 3.2 × 10^–3 mm², and large ELISpots were more than 3.2 × 10^–3 mm².

Spleen and MLN tissues were fixed in 10% formalin, embedded in paraffin, and sectioned. Deoxynucleotidyl-transferase–mediated UTP nick-end labeling (TUNEL) assays were performed using ApopTag PLUS peroxidase In Situ Apoptosis Detection Kit (Chemicon, Temecula, CA) according to the manufacturer's instructions. Methyl green was used for counterstaining. Images were obtained with a Leica DMRE microscope (Leica, Bannockburn, IL) with a 40×/0.7 NA PH2 objective. Images were captured using an RT Spot CCD camera (Diagnostic Instruments, Sterling Heights, MI) and Spot 4.6 software (Diagnostic Instruments). Image processing was done using Adobe Photoshop 6.0 software (Adobe Systems, San Jose, CA).

In vitro phagocytosis experiments were done as described by Scott et al.²³ 

Statistical analysis was performed using one-tailed Student t test. A value of P below .05 was considered significant.

We previously demonstrated that serum anti-Sm levels were elevated in 2-12H MRL/Fas^lpr mice.¹²  To investigate the contribution of Fas^lpr and the MRL bkg to anti-Sm production, we examined mice of a mixed C57BL/6-BALB/c bkg (2-12H and 2-12H/Fas^lpr) and an autoimmune MRL bkg (2-12H MRL and 2-12H MRL/Fas^lpr). Compared with 2-12H mice, serum anti-Sm levels were significantly elevated in all 2-12H/Fas^lpr, 2-12H MRL, and 2-12H MRL/Fas^lpr mice examined, indicating that both the MRL bkg and Fas^lpr induced anti-Sm production (Figure 1A). The levels were not significantly different between 2-12H MRL and 2-12H MRL/Fas^lpr mice, suggesting that the MRL bkg and Fas^lpr were not additive. Fas^lpr did not induce anti-Sm production in non-Tg mice, as previously demonstrated.⁸  Anti-Sm ASCs of 2-12H/Fas^lpr and 2-12H MRL/Fas^lpr mice were demonstrated in the bone marrow (BM), spleen, MLN, and LP (Figure 1B).

To understand the effects of the MRL bkg and Fas^lpr on anti-Sm B cells, we examined their effects on B-cell differentiation. We observed a higher frequency and number of anti-Sm B cells in 2-12H MRL than 2-12H mice (41% ± 3% vs 30% ± 3%, respectively; P = .013) (Figure 2A-B). This increase was against the backdrop of an overall 2-fold increase in the total number of splenic B cells in 2-12H MRL and non-Tg MRL mice compared with their nonautoimmune counterparts (Figure 2B). Thus, the MRL bkg preferentially expanded the anti-Sm B-cell subset. This expansion was opposed by Fas^lpr, as both the frequency (16% ± 3% in 2-12H MRL/Fas^lpr vs 41% ± 3% in 2-12H MRL mice; P = .002) and number of anti-Sm B cells were reduced by half in 2-12H MRL/Fas^lpr mice (Figure 2A-B). This Fas^lpr-induced reduction of anti-Sm B cells in MRL mice did not occur in nonautoimmune 2-12H mice (30% ± 3% in 2-12H vs 36% ± 4% in 2-12H/Fas^lpr mice; P = .114; Figure 2A-B). Thus, Fas^lpr prevented the expansion of both the splenic B-cell and anti-Sm B-cell populations induced by the MRL bkg, but had no effect on the sizes of these populations in nonautoimmune mice.

The MRL bkg and Fas^lpr preferentially affected the anti-Sm MZ B-cell subset. MRL mice had an approximately 4-fold expansion of the MZ B-cell subset compared with nonautoimmune mice (Figure 2C-D bottom), as has been previously reported,^29,30  but had an approximately 8-fold expansion of the anti-Sm MZ B-cell subset (Figure 2D top). These increases exceeded the approximately 2-fold expansion of total splenic B-cell numbers (Figure 2B), indicating a preferential effect on MZ B cells and in particular anti-Sm MZ B cells. Again, Fas^lpr opposed the expansion of the MZ B-cell subset, as 2-12H MRL/Fas^lpr mice had one third the number of MZ B cells and one seventh the number of anti-Sm MZ B cells as 2-12H MRL mice (Figure 2D). Fas^lpr affected only the MRL-induced expansion of MZ B cells, since non-Tg/Fas^lpr and non-Tg mice did not differ in the number of MZ B cells (Figure 2D). Thus, the MRL bkg preferentially expanded the anti-Sm MZ B-cell subset, and this expansion was opposed by Fas^lpr.

While Fas^lpr opposed an anti-Sm MZ expansion in autoimmune 2-12H MRL mice, this same mutation was responsible for the loss of anti-Sm MZ B cells in 2-12H/Fas^lpr mice. 2-12H and 2-12H/Fas^lpr mice have equal numbers of anti-Sm MZ B cells up to 2 months of age, but while 2-12H anti-Sm MZ numbers remained the same thereafter, 2-12H/Fas^lpr numbers declined to background levels by 3 months of age (Figure 2C-D). Because there was no decline in MZ B-cell numbers in non-Tg Fas^lpr mice (Figure 2D bottom) and the majority of MZ B cells in 2-12H but not non-Tg mice were anti-Sm (Figure 2E), we conclude that Fas^lpr induces an autoantigen-specific loss of MZ B cells.

We previously determined that anti-Sm B cells in 2-12H mice differentiate to an early pre-PC stage. Similar to PCs, these cells express intermediate levels of the PC marker CD138 (CD138^int), have up-regulated CXCR4, and have become larger and more granular; yet, similar to B cells, they still express near-normal levels of CD19, B220, and IgM.¹¹  However, their differentiation is arrested prior to becoming ASCs. In contrast, the anti-Sm B cells of autoimmune 2-12H MRL/Fas^lpr mice can differentiate beyond this checkpoint, become CD138^hi, increase in size and granularity, and express low levels of CD19, B220, and IgM.¹¹  Most significantly, some differentiate to ASCs.¹¹  To assess the role of the MRL bkg and Fas^lpr on pre-PC differentiation, we examined anti-Sm pre-PCs in these autoimmune mice. We found that the frequency of CD138^int anti-Sm B cells was similar in 2-12H and 2-12H/Fas^lpr mice (P = .48) (Figure 3A) and that there were no differences in size or expression of IgM, CXCR4, CD19, and CD80 (Figure 3B and data not shown). Moreover, Fas^lpr did not induce an increase in CD138^hi pre-PCs (P = .38). In contrast, the MRL bkg induced a significant decrease in the frequency of CD138^int anti-Sm B cells in 2-12H MRL mice (P = .008 versus 2-12H) and a significant increase in CD138^hi anti-Sm B cells (P = .049 versus 2-12H). Compared with CD138^int pre-PCs, the CD138^hi pre-PCs were larger, more granular, and had higher levels of CXCR4 and lower levels of CD19, B220, and IgM (Figure 3B and data not shown). This phenotype was similar to that of the anti-Sm CD138^hi pre-PCs of 2-12H MRL/Fas^lpr mice (Figure 3A-B and Culton et al¹¹ ). Thus, only the MRL bkg exhibits evidence of significant differentiation beyond the early pre-PC tolerance checkpoint.

Since 2-12H/Fas^lpr mice generate anti-Sm ASCs (Figure 1B), we sought to determine whether bypass of the early pre-PC checkpoint occurs inefficiently, so as not to alter the frequency of CD138^int and CD138^hi pre-PCs, yet still generate ASCs. Thus, we determined the frequency of anti-Sm ASCs among CD138⁺ pre-PCs from both nonautoimmune and autoimmune 2-12H mice by ELISpot analysis. Since anti-Sm MZ B cells in nonautoimmune mice secrete small amounts of antibody,²²  we also examined the frequency of ASCs among FO and MZ B cells. The sorting criteria were as illustrated in Figure 3C. PCs were excluded from the sorted CD138^int B-cell population since the sorted population excluded all CD19^– cells (Figure 3B). The FO and MZ B cells of 2-12H and 2-12H/Fas^lpr mice formed only small anti-Sm ELISpots, with the highest frequency among MZ B cells (Figure 3D). Of interest, 2-12H/Fas^lpr mice had a significantly lower frequency of anti-Sm ELISpots among MZ B cells than 2-12H mice, and anti-Sm ELISpots were undetectable among MZ and FO B cells of 2-12H MRL/Fas^lpr mice. In contrast, large anti-Sm ELISpots were detectable only among the CD138⁺ pre-PCs from 2-12H/Fas^lpr and 2-12H MRL/Fas^lpr mice, with the highest frequency among the latter (Figure 3D). Thus, although Fas^lpr does not induce an increase in CD138^hi pre-PCs, it induces some pre-PCs to become ASCs, indicating bypass of the early pre-PC tolerance checkpoint. Moreover, the inverse relationship between the numbers of small anti-Sm ELISpots of FO and MZ B cells and large anti-Sm ELISpots of the CD138⁺ pre-PCs suggests that MZ and FO B cells are induced to differentiate to CD138^int pre-PCs in autoimmune mice.

Fas^lpr and the MRL bkg also affected the number of anti-Sm B-1 cells. B-1 cells were identified as CD5⁺, CD43⁺, CD23^–, and CD11b⁺ (Figure 4A and data not shown). As with the MZ B-cell population, the MRL bkg genes induced a 1.5- to 2-fold expansion of both the anti-Sm B-1–cell population and the total B-1–cell population (Figure 4B), and Fas^lpr opposed this increase (Figure 4B). However, it had no effect on the number of B-1 cells in nonautoimmune 2-12H or non-Tg mice (Figure 4B). As with anti-Sm MZ B cells, the anti-Sm B-1–cell population underwent limited expansion in perinatal 2-12H/Fas^lpr and 2-12H MRL/Fas^lpr mice and declined to background numbers by 3 months of age, whereas the anti-Sm B-1–cell numbers were constant in 2-12H and 2-12H MRL mice. Thus, Fas^lpr induces the loss of anti-Sm B-1 cells, but not the total number of B-1 cells, suggesting that like the loss of anti-Sm MZ B cells, anti-Sm B-1–cell loss is autoantigen specific.

We previously showed that peritoneal anti-Sm B-1–cell numbers declined in 2-12H/Mer^kd mice after approximately 2 months of age and that this was due at least in part to differentiation to ASCs of the MLNs and LP.²²  To determine whether a similar explanation accounted for the loss of peritoneal anti-Sm B-1 cells in 2-12H/Fas^lpr mice, sorted B-1 and B-2 cells from 2-12H/Fas^lpr mice were transferred intraperitoneally to Fas^lpr recipient mice. As shown in Figure 5A, only B-1–cell transfers generated anti-Sm ASCs in the MLN and LP. Thus, Fas^lpr induced anti-Sm B-1 cells, but not B-2 cells, to differentiate to MLN and LP ASCs. Since these data indicate that only B-1 cells are activated after transfer, all subsequent experiments used unsorted peritoneal cells to maximize the number of B-1 cells transferred.

To determine whether the Fas^lpr environment is required for activation, 2-12H/Fas^lpr peritoneal cells were transferred to wt and Fas^lpr recipient mice. Two weeks after transfer, Fas^lpr recipients, but not wt recipients, had an elevated level of serum anti-Sm (Figure 5B inset) and an elevated number of anti-Sm ASCs of donor origin in the MLN and LP of Fas^lpr recipients (Figure 5B). Anti-Sm ASCs were not detected in the spleen or BM of either Fas^lpr or wt recipients. B-1–cell activation did not require that they be Fas-deficient, since transfer of 2-12H peritoneal cells to Fas^lpr mice, but not wt recipient mice, generated serum anti-Sm and anti-Sm ASCs of donor origin (Figure 5C). Thus, anti-Sm B-1 cells are activated only in Fas^lpr recipients and they need not be Fas-deficient.

That Fas^lpr had no effect on total B-1–cell numbers in nonautoimmune mice (Figure 4B lower panel) suggested that the loss of anti-Sm B-1 cells caused by Fas^lpr is antigen specific. To further test this possibility, we examined B-1 cells specific for phosphatidyl choline (PtC), a common membrane phospholipid to which as many as 7% of B-1 cells are specific.²⁷  For these experiments, we used mice with an H chain transgene that encodes anti-PtC antibodies (6-1). 6-1 Tg mice have a high frequency of splenic and peritoneal anti-PtC B-1 cells. These cells were identified as CD43⁺, CD5⁺ B cells that also stained with PtC-containing liposomes. Fas^lpr had no effect on the percentage of peritoneal anti-PtC B-1 cells (Figure 6A) or on the expansion of the anti-PtC B-1–cell population in neonatal and perinatal mice (Figure 6B), and it did not induce a loss of PtC-specific B-1 cells (Figure 6B).

To determine whether Fas^lpr increased the activation of anti-PtC B-1 cells to ASCs, we compared the activation of anti-PtC B-1 cells in 6-1 and 6-1/Fas^lpr mice by cell transfer. 6-1 peritoneal cells were transferred to wt and Fas^lpr recipient mice. Anti-PtC B-1 cells are normally activated in nonautoimmune mice,²⁷  and, thus, 2 weeks after transfer ASCs of donor origin were present in wt recipients (Figure 6C). However, the numbers of ASCs of donor origin in Fas^lpr mice were not significantly different from those in wt recipients (Figure 6C). Thus, the activation and depletion of B-1 cells in Fas^lpr mice is limited to cells of certain specificities, consistent with antigen dependence.

The activation and loss of anti-Sm B-1 cells of 2-12H/Fas^lpr mice were remarkably similar to our findings with anti-Sm B-1 cells of 2-12H/Mer^kd mice.²²  We attributed this loss in Mer^kd mice to an increase in availability of apoptotic cell antigens such as Sm and a defect in macrophage phagocytosis of apoptotic cells. Since Fas has nonapoptotic functions,^31,32  we sought to determine whether Fas^lpr mice also increased the availability of apoptotic cell antigens. The frequency of apoptotic cells in wt, Fas^lpr, and MRL/Fas^lpr mice was measured by flow cytometry using the fluorescence-labeled caspase inhibitor VAD-FMK. Both 2- and 3-month-old mice were analyzed to span the time frame for anti-Sm MZ and B-1 B-cell loss (Figures 2,4). No difference in the frequency of splenic or MLN apoptotic cells was evident at 2 months, whereas a higher frequency of apoptotic cells was noted in Fas^lpr spleens and MLNs at 3 months (Figure 7A). This was corroborated by the detection of cells with fragmented DNA by TUNEL assay of frozen tissue sections of 3-month-old Fas^lpr mice (Figure 7B). In the spleen, TUNEL⁺ cells were located predominantly in the extrafollicular space, whereas in the MLN they were located predominantly in B-cell–rich follicles and in germinal centers.

The increase in apoptotic cell frequency seen in Fas^lpr mice could be due to a defect in their clearance by phagocytic cells. Since macrophages play a key role in apoptotic cell clearance,²³  we compared the ability of macrophages from 2- and 3-month-old Fas^lpr and wt mice to phagocytize apoptotic cells in an in vitro assay. As shown in Figure 7C, both 2- and 3-month-old Fas^lpr macrophages were significantly less able to phagocytize apoptotic cells than their wt counterparts, although their defect at both ages was less severe than that of Mer^kd macrophages. Nevertheless, these data implicate Fas^lpr macrophages in the high frequency of apoptotic lymphocytes in Fas^lpr mice and in the activation of autoreactive B cells.

We show here that MRL, MRL/Fas^lpr, and Fas^lpr mice with the 2-12H transgene develop a spontaneous anti-Sm response with complete penetrance. Our findings point to a central involvement of MZ and B-1 B cells in this response and suggest a link to defective apoptotic cell clearance in their activation. Our comparison of anti-Sm B cells from 2-12H mice and B cells of diverse specificities from non-Tg mice indicates bias at 2 levels: B-cell subset and antigen specificity. Fas^lpr induces a loss of anti-Sm MZ and B-1 B cells in young adult 2-12H/Fas^lpr and 2-12H MRL/Fas^lpr mice between 2 and 3 months of age. Since Fas^lpr has no effect on non-Tg MZ and B-1 B cells (Figures 2D,4) or on anti-PtC B-1 cells, we conclude that the loss of anti-Sm B cells is antigen specific (Figure 6) and not due to a general effect on the generation or maintenance of the MZ or B-1 B-cell populations.

The most likely explanation for the Fas^lpr-induced loss of anti-Sm MZ and B-1 B cells is that they are activated to differentiate to ASCs. Autoantigen activation would account for the antigen-specific depletion of anti-Sm MZ and B-1 B cells and the presence of ASCs in multiple lymphoid tissues (Figure 1). It would also account for the inverse relationship between FO and MZ ASCs and CD138^int ASCs (Figure 3B) since activated FO and MZ B cells would presumably pass through the CD138^int pre-PC stage before differentiating to PCs. Additionally, the activation of anti-Sm MZ and B-1 B cells would explain the appearance of serum anti-Sm in 2-12H MRL/Fas^lpr mice between 1 and 2 months of age.¹²  Anti-Sm B-1–cell activation is confirmed by the transfer experiments showing that they generate MLN and LP ASCs and produce serum antibody in Fas^lpr recipients, but not wt recipients (Figure 5A). Antigen-independent mechanisms may have a role in this activation, but the number of anti-PtC B-1 cells is the same in 6-1 and 6-1/Fas^lpr mice, indicating that any such effect is likely to be small. Thus, we conclude that loss of anti-Sm B-1 cells in Fas^lpr mice is due to autoantigen-specific activation and differentiation to ASCs. The parallel loss of anti-Sm MZ B cells by Fas^lpr mice suggests that anti-Sm MZ B cells are also activated to become ASCs, although this remains to be confirmed.

The inefficient bypass of the early pre-PC tolerance checkpoint induced by Fas^lpr (Figure 3) indicates that Fas-FasL–induced cell death is not a major mechanism of regulation at this checkpoint. Whether Fas-FasL interactions have a regulatory role at an earlier checkpoint cannot be ruled out. The interaction of Fas on autoreactive B cells with FasL on helper T cells can induce cell death,^33,34  but this is not always the case.³⁵  Although helper T-cell regulation of these B cells may occur through Fas-FasL interactions, the finding that Fas⁺ B-1 cells are activated in Fas^lpr hosts (Figure 5) suggests that the primary defect leading to anti-Sm MZ and B-1 B-cell activation in Fas^lpr mice lies elsewhere.

A possible explanation for anti-Sm and B-1–cell activation in Fas^lpr mice is the defect in macrophage phagocytosis of apoptotic cells and the increase in the presence of apoptotic lymphocytes in lymphoid tissues (Figure 7). Apoptotic cells expose Sm and can induce an anti-Sm response in nonautoimmune mice.²²  A Fas^lpr-induced increase in apoptotic cells has been previously noted in the brains of MRL/Fas^lpr mice.³⁶  This would account for the antigen-specific effects of Fas^lpr and for the observation that both Fas-sufficient and -deficient B-1 cells are activated (Figure 5). Additionally, since anti-PtC antibodies do not bind apoptotic cells (Y.Q. and S.H.C., unpublished data, March 2001), it would explain why Fas^lpr has no effect on anti-PtC B-1 cells (Figure 6). These findings parallel our previous observation that anti-Sm MZ and B-1 B cells are activated in 2-12H/Mer^kd mice, which also have a defect in apoptotic cell clearance.²²  In addition, the appearance of increased numbers of apoptotic cells in lymphoid tissues corresponds with when anti-Sm production begins and when anti-Sm MZ and B-1 cells are lost, further implicating apoptotic cells as the driving force in the activation of Fas^lpr anti-Sm MZ and B-1 cells. Of interest, Fas^lpr mice that lack p53, a proapoptotic gene,³⁷  generate lower levels of serum autoantibodies compared with Fas^lpr mice.³⁸  Thus, the absence of p53 may result in a lower frequency of apoptotic cells in Fas^lpr mice, accounting for the lower autoantibody levels. Defective clearance of apoptotic cells is a recurring feature in autoimmunity, and targeted impairment of apoptotic cell clearance invariably results in autoimmunity.^23,39-41  Defective phagocytosis of apoptotic cells is also a feature of human lupus.^42,43  MRL macrophages are defective in apoptotic cell clearance,⁴⁴  and thus the combination of the MRL bkg- and the Fas^lpr-induced impairment of apoptotic cell clearance may contribute to the more severe autoimmunity of MRL/Fas^lpr mice.^42,43  Of interest, while the defect in macrophage phagocytosis of apoptotic cells is evident in Fas^lpr and MRL/Fas^lpr mice at 2 months (Figure 7A), there is no apparent effect on the frequency of apoptotic lymphocytes relative to wt mice at 2 months (Figure 7C). Thus, either this macrophage defect is not responsible for the increase in apoptotic lymphocytes or that there are compensatory mechanisms for phagocytosis in younger mice.

How Fas^lpr induces this macrophage defect remains to be investigated. Serum from lupus patients significantly inhibits efficient phagocytosis by normal human neutrophils,⁴⁵  and thus there may be a secreted factor in Fas^lpr mice that diminishes macrophage phagocytosis of apoptotic cells. Although incompletely characterized, Fas has nonapoptotic functions, such as cytokine production, induction of DC maturation and survival, and chemokine production^46,47  involving multiple cell types.^31,32,48,49  Thus, Fas-FasL interactions could be necessary for the induction of efficient macrophage phagocytosis of apoptotic cells. IL-10 enhances macrophage phagocytosis of apoptotic cells,⁵⁰  although there is disagreement,⁵¹  but no difference in systemic IL-10 levels was noted (22.3 ± 0.81 pg/mL for wt mice vs 22.9 ± 2.37 pg/mL for Fas^lpr mice, P = .73).

Our finding that B-1 cells are involved in the production of anti-Sm appears to contradict an earlier report.⁵²  This earlier study relied on reconstitution of irradiated recipient mice, and we find that anti-Sm B-1 cells undergo apoptosis upon encounter with apoptotic cells and poorly reconstitute irradiated mice after bone marrow transfer (Y.Q. and S.H.C., unpublished observation, September 2000). Also, B-1 cells may contribute minimally to serum autoantibody since many anti-Sm ASCs of B-1–cell origin are located in the LP and are likely to secrete into the intestinal lumen. Thus, previous studies could reasonably lead to different conclusions.

The activation of anti-Sm MZ and B-1 B cells in MRL and Fas-deficient mice adds to a growing body of evidence that MZ and B-1 B cells are involved in autoimmunity.^{13,20,22,23,30,53-56}  Both MZ and B-1 B cells are antigen-experienced cells that can rapidly differentiate to ASCs in response to antigen and provide protection against blood-borne particulate antigens and enteric organisms.⁵⁷  These features, as well as their location, may make them uniquely positioned to encounter and respond to apoptotic cells in the blood and lymph. MZ B cells are potent activators of CD4⁺ T cells,⁵⁸  and thus activation of MZ B cells may also contribute to disease through the activation of autoreactive T cells. Understanding how autoreactive MZ and B-1 B cells are activated and the nature of the link to increased apoptotic cell frequency and defective macrophage apoptotic cell phagocytosis will be important directions for further investigation.

The authors thank the UNC Flow Cytometry Facility and Larry Arnold for their assistance and advice with the flow cytometry and cell sorting.