APRIL modulates B and T cell immunity (original) (raw)
APRIL Tg mice. The expression pattern of APRIL on lymphoid cells remains ill-defined. Injection of recombinant APRIL nonetheless results in expansion of the B cell population and T cell activation (18), suggesting that it plays a role in the immune system. To analyze whether APRIL, like BLyS (18), is expressed on T cells, we purified TCR Tg CD4+ T cells from DO11.10 mice. These T cells can be activated in vitro using an appropriate OVA peptide. Moreover, these T cells can be directed toward a Th1 or Th2 phenotype depending on the cytokines present during activation (27). Using this model system, we analyzed APRIL and BLyS expression by RT-PCR. As reported previously, BLyS expression was upregulated during T cell activation (Figure 1a) (4). Similarly, APRIL expression was detected in activated Th1 or Th2 cells, but not in naive cells (Figure 1a), pointing to a role for this molecule in T cell–dependent immunity.
(a) APRIL expression in activated, but not in naive, T cells. Activated T cells have increased APRIL mRNA expression levels. DO11.10 spleen CD4+ T cells were isolated and activated in vitro toward a Th1 or Th2 phenotype. RNA was prepared from naive, Th1, and Th2 cells and used to analyze the expression of APRIL, BLyS, and GAPDH by RT-PCR. Generation and characterization of APRIL Tg mice. (b) Schematic diagram of constructs used for peripheral T cell–specific expression of the human APRIL transgene. (c) mRNA expression of APRIL in T cells of four different founders. RNA was prepared from purified T cells derived from spleen and lymph nodes of four Tg mice and analyzed by RT-PCR for APRIL RNA expression levels. To ensure that the amplified human APRIL signal was not due to contaminating genomic DNA, cDNA probes were amplified for Thy1 using primers that amplify a 300-bp fragment of cDNA and a 700-bp fragment of genomic DNA (genomic mouse DNA was used as control). (d) Protein expression of APRIL in Tg mouse T cells. Cell lysates were prepared from purified T cells derived from spleen and lymph nodes of four Tg mice, a control littermate, and a C57BL/6 mouse and analyzed in Western blot using anti-human APRIL antibodies (8). (e) Secreted APRIL circulates in serum of APRIL Tg mice. Sera (1 μl) of control and Tg mice were resolved under nonreducing conditions and immunoblots developed with anti-human APRIL antibodies. Recombinant (Rec.) APRIL protein (0.1–5 ng) added to control serum was used as standard.
To study this finding in further detail, we generated a Tg mouse strain expressing human APRIL under the control of the lck distal promoter, which directs transgene expression to mature thymocytes and peripheral T lymphocytes (Figure 1b) (26). We confirmed in ELISA assays that binding of recombinant human APRIL to mouse TACI and BCMA is comparable to its binding to the corresponding human receptors (not shown), which concurs with a recent report that human APRIL has the same binding capacities for human and mouse BCMA (31). In four independent Tg mouse lines, we detected RNA expression of the transgene (Figure 1c). Only two of these transgenes displayed detectable levels of the transgene protein in T cells (Figure 1d), but in no other types of cells tested (not shown). Moreover, sera from 8-month-old mice of the Tg 3919 line showed detectable circulating APRIL levels at concentrations of 1–5 μg/ml (Figure 1e). Tg mice were born at the expected mendelian ratio, and histological analysis at 3–12 weeks of age revealed no gross abnormalities in either of the Tg strains. In addition, semiquantitative RT-PCR analysis of RNA derived from APRIL Tg mouse spleens revealed no alterations in BLyS, TACI, or BCMA RNA expression levels (not shown). Line 3919 was used for subsequent experiments, but similar observations were made in line 3923.
T and B cell composition of lymphoid organs. As recombinant APRIL results in increased spleen weight due to B cell accumulation (18), we compared the spleens of 12 Tg mice and their corresponding non-Tg littermates at 6–12 weeks of age. No significant difference was observed in spleen weight (Figure 2a) or total number of spleen cells (Figure 2b), and immunohistological analysis of APRIL Tg mouse spleen architecture showed no apparent abnormalities (not shown). Furthermore, no alteration was observed in spleen weight or total cell number in older Tg mice tested (age 4–9 months) (not shown). We next analyzed the B and T cell composition in Tg mouse secondary lymphoid organs by FACS analysis. No variation was observed in the percentage of B220+ B cells in spleen, mesenteric lymph nodes, or Peyer’s patches (Figure 2c). In contrast, a small increase in the percentage of B220+ B cells was observed in peripheral lymph nodes in all Tg mice tested (n = 10) (Figure 2c). The percentages of both CD4+ and CD8+ T cells were proportionally decreased (Figure 2d), resulting in a significant decrease in absolute T cell numbers in Tg mouse peripheral lymph nodes (Figure 2e). Further analysis of T cells revealed a striking increase in CD62L– CD4+ and CD8+ T cells in secondary lymphoid organs, of which only a small percentage showed CD44 upregulation (Figure 3). Whereas a fivefold increase has been observed in CD62L– CD44high effector T cells in BLyS Tg mice (19), APRIL Tg mice revealed an accumulation of CD62L– CD44low T cells, which do not correspond to classically activated T cells. In agreement, there was no reproducible upregulation of CD25 or downregulation of CD45RB in the T cells in the mice analyzed (not shown).
B and T cell homeostasis in APRIL Tg mice. (a) Unaltered spleen weight in APRIL Tg mice. Spleen weight of mice was measured at the age of 6–12 weeks (n = 12). (b) Unaltered splenocyte and peripheral lymph node cell number in APRIL Tg mice. Splenocyte number was determined in 6- to 12-week-old mice (n = 11). Data represent the mean ± SD. (c) Evaluation of B cell percentage (B220+ cells) in spleen, mesenteric lymph nodes (MLN), Peyer’s patches (PP), and peripheral lymph nodes (PLN) (n = 10). (d) Percentage of CD4+ and CD8+ T cell populations in peripheral lymph nodes of APRIL Tg mice (n = 10; 6–12 weeks old). (e) Comparison of absolute B and T cell numbers in peripheral lymph nodes (n = 10). Data are the mean ± SM. The statistical significance of the data was determined using ANOVA. P > 0.05 is considered insignificant, P < 0.05 significant, P < 0.01 very significant, and P < 0.0001 extremely significant. Co, control.
Decreased CD62L (L-selectin) expression in splenic T cells is not accompanied by CD44 upregulation. CD62L versus CD44 expression is shown gated on CD4 and CD8 populations. Downregulation of CD62L was seen in T cells of blood and secondary lymphoid organs; it was already apparent in Tg mouse spleen at 2 weeks of age and was maintained throughout their lives (not shown).
Ectopic APRIL expression prolongs T cell survival in vivo and in vitro, correlating with increased Bcl-2 levels. Recombinant APRIL stimulates B and T cell proliferation in vitro (18). In agreement, purified T cells from APRIL Tg mice showed increased thymidine incorporation following a 2-day stimulation with anti-CD3 in combination with anti-CD28, as well as after stimulation with anti-CD3 alone (Figure 4a). The increased in vitro proliferation of Tg T cells correlated with increased IL-2 production in CD8+ T cells after in vitro stimulation (Table 1).
Increased proliferation of T cells and survival of T cells from APRIL Tg mice in vitro. (a) Increased proliferation of Tg T cells after 2 days’ culturing with anti-CD3 alone, or anti-CD3 in combination with anti-CD28. (b) T cell response in APRIL Tg mice in vivo: Delayed deletion of superantigen-responsive Vβ8+ CD4+ cells in APRIL Tg mice. PBLs were stained for CD4, CD8, and Vβ8 on the days indicated (n = 4). This is one representative experiment of four performed. ANOVA confirmed that the differences for CD4+ cells at days 7, 10, and 13 are statistically significant (Table 2). *P < 0.05. The Vβ6+ subpopulation of CD4+ and CD8+ cells was unresponsive at all time points analyzed (not shown). (c and d) Survival of purified T cells cultured for 3 days alone, with anti-CD3, or with anti-CD3 plus anti-CD28. Cells were stained for CD4 (c) or CD8 (d) and PI as described in Methods. (e) Purified T cells of C57BL/6 mice were cultured for 2 days without stimulation, alone or in the presence of 5 μg/ml recombinant APRIL, and PI-stained. (f) Purified CD4+ T cells were cultured for 3 days without stimulation, alone or in the presence of 50 μg/ml of TACI-Fc, BMCA-Fc, or control Ig. Similar results were obtained for CD8+ T cells. All data in a and c–f are the mean ± SD of triplicate determinations of a representative experiment of at least three performed. The statistical significance of the data was determined using ANOVA.
Cytokine production of in vitro activated Tg T cells
To determine whether the increased in vitro T cell proliferation observed can also be detected in vivo, we analyzed T cell activation using injection of SEB, which interacts predominantly with Vβ8+ T cells (32). Treatment of mice with SEB leads to systemic activation of all Vβ8+ T cells. SEB injection leads to an initial increase in the percentage of both Vβ8+ CD4+ and Vβ8+ CD8+ T cells. This initial proliferation is followed by a deletion phase, with clear decreases in Vβ8+ T cell numbers, which eventually reach levels lower than those found before SEB treatment. The decrease in Vβ8+ CD4+ cells is due to apoptosis of SEB-responsive cells (32, 33); Figure 4b shows a representative experiment tracking the percentage of Vβ8+ T cells in PBLs from APRIL Tg mice and littermates following SEB administration. Both Vβ8+ CD4+ and Vβ8+ CD8+ T cells of APRIL Tg mice show expansion similar to that of corresponding non-Tg littermates at day 3, indicating that APRIL does not affect the expansion phase of the T cell response in vivo. In contrast, a significantly delayed deletion of Vβ8+ CD4+ T cells was observed in comparison with SEB-reactive cells of littermates; this was especially evident when the percentage of Vβ8+ CD4+ T cells was compared 7 days after SEB injection. Despite the clear delay in deletion of CD4+ T cells, no significant difference was observed in the CD8 response (Table 2). This lack of effect on the CD8+ T cell response was confirmed by analyzing the CD8 response induced after infection with adenovirus. Using this approach, we observed no effect of ectopic APRIL expression in the CD8+ T cell response to adenovirus (not shown).
P values for data in Figure 4b
To analyze this delayed T cell deletion in further detail, we tested whether ectopic APRIL-expressing T cells have an increased survival capacity in vitro. To this end, purified T cells were cultured for 3 days under various conditions, and survival was determined by measuring PI uptake. Both CD4+ and CD8+ T cells isolated from Tg mice showed significantly increased survival compared with those of littermates (Figure 4, c and d). This difference became evident only when T cells received suboptimal or no stimulus. Increased survival of purified T cells also became evident by quantification of live and dead cell gates by forward/side scatter (Table 3). Increased survival of APRIL Tg T cells was not blocked by BCMA-Fc or TACI-Fc (Figure 4f). In addition, the increased survival of nonactivated T cells could not be mimicked by the addition of recombinant APRIL to T cell cultures (Figure 4e). Taken together, this suggests that the presence of APRIL in in vitro cultures is not sufficient to increase survival of unstimulated T cells, but that Tg T cells receive this survival signal in vivo.
Survival of cultured T cells determined by forward scatter and side scatter
The augmented survival of B cells observed in BLyS Tg mice has been associated with increased Bcl-2 expression levels (19). Bcl-2 expression is especially effective when cells are deprived of survival factors (34, 35). Moreover, Bcl-2 has been shown to delay the deletion of SEB-primed CD4+ T cells (33). We therefore determined Bcl-2 expression in the APRIL Tg mice and control littermates by FACS analysis. Ex vivo CD4+ and CD8+ T cells, but not B cells, displayed a twofold increase in Bcl-2 levels in Tg mice compared with non-Tg littermates, suggesting that elevated Bcl-2 levels may be responsible for the prolonged survival of these cells (Figure 5a).
T cells of Tg mice have increased Bcl-2 levels. (a) Increased Bcl-2 expression in CD4+ and CD8+ T cells, but not B cells of APRIL Tg mice. Ex vivo CD4+ and CD8+ T cells and B220+ B cells from peripheral lymph nodes were analyzed for Bcl-2 expression as described in Methods. Staining in cells of control littermates is shown in black (gray shadow for isotype control, no shadow for Bcl-2 staining); blue shows isotype control, and red shows Bcl-2 in cells of Tg’s. One representative staining of four is shown. The average increase in Bcl-2 mean fluorescence intensity was 1.5 ± 0.19–fold in Tg CD4 cells (mean ± SD, n = 4), 1.8 ± 0.4–fold in CD8 cells, and 1.1 ± 0.05–fold in B220+ cells. (b) Increased Bcl-2 expression in sorted CD62L+ T cells of Tg mice. Staining in cells of control littermates is shown in black (gray shadow for isotype control, no shadow for Bcl-2 staining); blue shows isotype control, and red shows Bcl-2 in cells of Tg’s. One representative staining of three is shown. (c) T cell proliferation of sorted CD62L+ T cells after 2 days’ culture with anti-CD3 in combination with anti-CD28. (d) Survival of sorted CD62L+ T cells cultured for 3 days without stimulation. Data in c and d are the mean ± SD of triplicate determinations of a representative experiment of at least three performed. The statistical significance of the data was determined using ANOVA.
The increased proliferation, survival, and Bcl-2 expression may be due to the presence of the relatively large CD62L– T cell fraction in the Tg animals, which might represent activated T cells. To exclude this possibility, we analyzed CD62L+ T cells, which represent the naive T cell fraction. Like unsorted T cell populations, Tg mouse CD62L+ T cells showed increased T cell proliferation following anti-CD3/anti-CD28 activation, as well as prolonged survival in 3-day cultures (Figure 5, c and d). In addition, an increase in Bcl-2 expression was detected on sorted Tg mouse CD62L+ T cells compared with those of littermates (Figure 5b). This indicates that the survival and proliferation effects described are not due to the increase in CD62L– T cells in the Tg mice but are a direct result of APRIL expression on naive T cells.
APRIL Tg mice display normal B cell development in spleen. Only mature B cells can recirculate and enter lymphoid follicles of spleen and lymph nodes. Two spleen B cell populations, termed type 1 (T1) and type 2 (T2) transitional B cells, were identified as precursors for mature B cells (36). BLyS preferentially supports the survival of T2 B cells (37). To analyze whether Tg expression of APRIL had a similar effect on the B cell compartment, we determined the composition in the spleen. T1 and T2 transitional B cells and marginal zone B cells can be distinguished on the basis of IgM, CD21, and CD23 expression (36). The T1 and T2 transitional B cell subpopulations are both IgMbright, whereas only T2 cells are CD21bright CD23+, which also distinguishes them from CD21bright CD23– marginal zone B cells (36). Triple staining of spleen B cells from mice Tg for APRIL showed no significant difference in the percentage of T1, T2, mature, or marginal B cells (Figure 6). T1 B cells are the precursors of T2 and mature B cells (36), and T1 B cells are already detectable in 1-week-old mice, whereas T2 and mature B cells only become clearly detectable in mice at about 3 weeks of age (36). To exclude that APRIL influences the kinetics of B cell development, we followed T1 and T2 B cell formation in mice at various ages; we found that T1 and T2 B cell development is comparable between APRIL Tg mice and littermates (Figure 6). Notably, B cells in peripheral lymph nodes of APRIL Tg mice correspond to mature (IgM+, CD23+, CD21bright, and HSA+) B cells (not shown). The increased constitutive IgM levels in sera of APRIL Tg mice suggested an alteration in the B1 cell population (38). B1 cells are characterized by B220 and IgM expression, but by a lack of CD23 expression; CD5 expression analysis allows a further subdivision into B1a (CD5+) and B1b (CD5–) subpopulations. We thus determined the percentages of these cells in the peritoneum. As shown in Table 4, no significant differences were found upon analysis of APRIL Tg T cells compared with those of littermates.
Ectopic APRIL expression does not influence B cell development in spleen. Splenocytes were analyzed for their content of T1 (IgMbright CD21– CD23–), T2 (IgMbright CD21bright CD23+), mature (IgM+ CD21bright CD23+), and marginal zone (IgMbright CD21bright CD23-) B cells (n = 20).
No altered percentages of peritoneal B cell populations in APRIL Tg’s
Ectopic APRIL expression moderately alters the T cell–dependent humoral immune response. APRIL stimulates IgM production by peripheral blood B cells in vitro (25). We therefore compared serum IgM concentrations between APRIL Tg mice and littermates. IgM serum concentrations of Tg mice showed an approximately twofold increase, whereas IgG levels were comparable between Tg mice and littermates (Figure 7, a and b). The twofold difference in IgM serum levels between Tg mice and littermates persisted in older Tg mice (at age 4–9 months; not shown). No increase was found in anti-DNA antibodies in Tg mouse serum (Figure 7c). As ectopic APRIL expression resulted in increased serum IgM levels, we tested whether the humoral response is also modulated in APRIL Tg mice. Mice were challenged with a modified version of Ankara-strain vaccinia virus (MVA) expressing β-galactosidase (gal), a T-dependent antigen (30, 39). Antibody (IgM and IgG) levels against virus and antigen were measured 14 days after virus inoculation. Serum samples of APRIL Tg mice showed increased levels (approximately twofold) of IgM anti-virus antibodies compared with serum samples of control mice, while IgG induction was similar in these mice (Figure 7d). The IgG2a/IgG1 ratio revealed a Th1 response in both Tg and control mice (not shown). A second immunization (boost) with MVAgal virus resulted in a moderate but nonsignificant increase in IgG in the Tg compared with that of non-Tg littermates (Figure 7d). Similar data were obtained for the antibody response to the recombinant antigen (not shown).
Moderate alteration of serum Ig levels and T cell–dependent humoral responses in APRIL Tg mice. Comparison of IgM (a) and IgG (b) levels in APRIL Tg mice (n = 10 in a; n = 6 in b). (c) Anti-DNA autoantibodies were not elevated in APRIL Tg mice. For comparison, pooled sera of five MRL-lpr mice were used at the same dilution (1:100) as those of APRIL Tg mice and littermates, and at a 1:1,600 dilution to obtain a comparable OD value. (d) T cell–dependent humoral response to a modified version of attenuated Ankara-strain vaccinia virus (rMVA) in APRIL Tg mice (n = 4). Serum IgM and IgG levels from 8-week-old mice before immunization (preimmune), 15 days after immunization (priming), and 15 days after second immunization (boost). Statistical significance was determined using ANOVA.
We also analyzed the humoral response in APRIL Tg mice to a wild-type strain of vaccinia virus (a recombinant form of the Western Reserve strain), which showed a pattern similar to that observed for MVA, i.e., a twofold increase in antigen-specific IgM production within 15 days of immunization (not shown).
Elevated TI-2 humoral response in APRIL Tg mice. Polyvalent antigens such as pneumococcal polysaccharides and Ficoll-conjugated antigens are so-called T cell–independent type 2 (TI-2) antigens, as they were originally described to induce T cell–independent immune responses (40). TACI-deficient mice revealed depressed responsiveness to TI-2 antigens, manifesting the importance of that receptor for developing TI-2 responses (22). APRIL is a known TACI ligand; we therefore evaluated the role of APRIL in the generation of the TI-2 humoral response. Immunization of APRIL Tg mice with a model TI-2 antigen, NP-Ficoll, resulted in an increase in IgG response compared with that of non-Tg littermates (Figure 8a). In addition, the IgM response was significantly elevated on day 8 (Figure 8a). The increased IgG response is particularly evident for IgG1, IgG2a, and IgG2b, as they remained significantly elevated even into day 14 after TNP-Ficoll treatment (Figure 8b). IgG3 levels were also significantly increased at day 8, but we observed a leveling at later time points (Figure 8c).
TI-2 humoral response in APRIL Tg mice. Groups (n = 6) of APRIL Tg mice and littermates were immunized with 30 μg NP-Ficoll, a TI-2 antigen, and serum Ig levels were determined by ELISA. (a) IgM and total IgG response to NP-Ficoll. (b) Distribution of IgG1, IgG2a, and IgG2b isotypes. For comparison, IgG2a levels of preimmune sera are shown; similar values were obtained for IgG1 and IgG2b levels in preimmune sera. (c) IgG3 response to NP-Ficoll. The statistical significance of the data was determined using ANOVA.