Dendritic Cell Apoptosis in the Maintenance of Immune Tolerance

See allHide authors and affiliations

Science  24 Feb 2006:
Vol. 311, Issue 5764, pp. 1160-1164
DOI: 10.1126/science.1122545


Apoptosis in the immune system is critical for maintaining self-tolerance and preventing autoimmunity. Nevertheless, inhibiting apoptosis in lymphocytes is not alone sufficient to break self-tolerance, suggesting the involvement of other cell types. We investigated whether apoptosis in dendritic cells (DCs) helps regulate self-tolerance by generating transgenic mice expressing the baculoviral caspase inhibitor, p35, in DCs (DC-p35). DC-p35 mice displayed defective DC apoptosis, resulting in their accumulation and, in turn, chronic lymphocyte activation and systemic autoimmune manifestations. The observation that a defect in DC apoptosis can independently lead to autoimmunity is consistent with a central role for these cells in maintaining immune self-tolerance.

The critical role of apoptosis in maintaining peripheral tolerance is clearly demonstrated by systemic autoimmune diseases that result from mutations in the pro-apoptotic Fas receptor or Fas ligand genes, both in humans and mice (13). Although lymphocytes play a central role in these conditions, the extent to which apoptosis defects in various immune cell types might also be involved has yet to be fully characterized. Transgenic mice expressing apoptosis inhibitors in T cells do not display autoimmune symptoms (410), and conditional deletion of Fas in T or B cells fails to induce the typical autoimmune diseases observed with global mutations in Fas (11). Such results make it likely that defective apoptosis in other cell types plays a prominent role in the onset of autoimmune diseases.

DCs are potent antigen-presenting cells that initiate lymphocyte activation and may also be critical for maintaining immune tolerance (1216). Although immunization with excessive activated DCs induces tissue-specific and systemic autoimmune symptoms (17, 18), the role of DC turnover in the development of autoimmunity remains untested. Previously, we observed DC accumulation in autoimmune patients harboring a deficiency in apoptosis (19), and significant expansion of DCs has also been reported in Fas-deficient lpr mice (20). It is plausible, therefore, that defects in DC apoptosis lead to DC accumulation, and through chronic lymphocyte activation, the development of autoimmunity (19).

To directly test this hypothesis, we generated independent DC-specific transgenic mouse lines (DC-p35; fig. S1) that express the baculovirus p35 protein under the control of the CD11c promoter (21, 22). p35 is capable of inhibiting caspase-8 and several downstream caspases through covalent binding to active sites of these proteases (23). We also generated different lines of transgenic mice selectively expressing p35 in T cells (T-p35) or B cells (B-p35) under the control of the CD2 and CD19 promoters (24, 25), respectively (fig. S1).

DCs, but not T or B cells, from DC-p35 mice were deficient in Fas-mediated apoptosis as compared to those cells from control mice (figs. S1 to S4). In young DC-p35 mice (1 to 3 months), no obvious changes in lymphocytes and DCs were observed (26). However, significant expansion of CD11c+ (and CD40+ or I-Ab+) DCs was observed in the spleens of DC-p35 mice but not in their nontransgenic littermates or age-matched T-p35 and B-p35 mice at 12 months of age (Fig. 1A). Expansion of CD11clowCD11bB220+Gr-1+ plasmacytoid DCs (pDCs) (27) was also observed (Fig. 1B). No significant changes were detected in T cells, B cells, natural killer (NK) cells, or macrophages from DC-p35 mice (fig. S4). These data suggest that inhibition of apoptosis in DCs leads to an accumulation of DCs in DC-p35 mice with increasing age.

Fig. 1.

DC accumulation in DC-p35 transgenic mice. (A) Splenocytes from 12-month-old nontransgenic control, DC-p35, T-p35 and B-p35 mice were stained for different markers and analyzed by flow cytometry. The percentage of CD11c+I-Ab+ (upper panels) or CD11c+CD40+ DCs (lower panels) is shown. (B) The CD11chighCD11b+ mDCs (upper panels) and the Gr-1+B220+ pDCs among the CD11clowCD11b cells (lower panels) were stained and analyzed by flow cytometry. (C) CD69 expression on CD3+ cells (upper panels) or CD19+ cells (middle panels) was analyzed by flow cytometry. Splenocytes were also stained with fluorescein isothiocyanate (FITC)–anti-CD4 (antibody to CD4), followed by intracellular staining with phycoerythrin (PE)–anti-FoxP3 (lower panels).

Marked increases in the activation marker, CD69, were detected in T and B cells in the spleens of DC-p35 mice (Fig. 1C). In contrast, only modest increases in CD69+ T and B cells from T-p35 or B-p35 mice were observed (Fig. 1C), and DC-p35 mice displayed no significant change in the numbers of CD4+FoxP3+ natural T regulatory (Treg) cells relative to nontransgenic controls (Fig. 1C). Moreover, purified CD4+CD25+ cells composing the Treg population from DC-p35 and control mice inhibited the proliferation of CD4+CD25 T cells to a similar extent (fig. S5). Negative selection and Treg development also appeared to be normal in the thymus of DC-p35 mice (fig. S6).

Because caspase-8 deficiency is known to impair the development and activation of T cells (28, 29), it is possible that p35 disrupts the development of DCs in DC-p35 mice by inhibiting caspase-8. However, we found normal development of different subsets of CD11c+ DCs in 1- to 3-month-old DC-p35 mice (26). In addition, continuous labeling of DCs with bromodeoxyuridine (BrdU) in vivo (22) indicated that endogenous DCs were replaced by newly generated BrdU+ DCs at normal rates in DC-p35 mice (Fig. 2A), suggesting that transgenic expression of p35 was not impairing the development of DCs.

Fig. 2.

Enhanced immune responses in DC-p35 mice. (A) Percentage of BrdU+ DCs in the spleens of DC-p35 or nontransgenic mice after continuous BrdU labeling (22). (B) DC-p35 or nontransgenic mice were pulsed with BrdU from day –10 to day –1 and immunized with OVA on day 0. The percentage of BrdU+ cells among mDCs or pDCs in draining lymph nodes (LN) was quantitated at different times after immunization. (C) CFSE-labeled OVA-specific TCR transgenic CD8+ OT1 (upper panels) or CD4+ OT2 T cells (lower panels) and DCs from DC-p35 or control mice pulsed with the corresponding OVA peptides were transferred into recipient mice (22). CFSE+ cells in the draining nodes were gated to show CFSE dilution of dividing cells. Expansion of CD8+ OVA-specific T cells was also probed with H-2kb/OVA-tetramer (middle panels). Phosphate-buffered saline (PBS) solution was used instead of DCs for negative controls. (D) CD4+ Tcells (105 per well) from OT2 mice were mixed with the indicated numbers of DCs of DC-p35 mice and controls in the presence of OVA (10 μg/ml) for 3 days. Cell proliferation was measured by thymidine uptake. (E) DCs from DC-p35, lpr, or control (C) mice (left) or DCs from C57BL/6 mice transduced with a lentiviral vector (Lv), Lv-p35, or Lv–bcl-2 (right) were transferred into recipients. After 24 hours, recipients were injected with lipopolysaccharide, and sera were collected 2 weeks later to quantitate ANAs (22). Data (mean ± SD) were analyzed by Student's t test using GraphPad Prism version 4 for Macintosh. A P value of <0.05 was considered statistically significant.

In light of these results, we examined whether the in vivo survival of DCs during an active immune response might be affected in DC-p35 mice. Mice were injected with BrdU, followed by ovalbumin (OVA) immunization (22). The percentage of BrdU+ myeloid DCs (mDCs) in the draining lymph nodes rapidly decreased from 65% on day 1 to 20% on day 3 after immunization in wild-type mice, whereas the decrease in BrdU+ mDCs was slower in DC-p35 mice (Fig. 2B). In contrast to these effects on mDCs, the number of BrdU+ pDCs increased rapidly after immunization in wild-type mice (Fig. 2B). It is possible that antigen stimulation triggered proliferation of recently labeled pDC precursors, leading to a rapid expansion of pDCs during the initial phase of an immune response when the draining lymph nodes sharply increase in size. The percentage of BrdU+ pDCs then decreased rapidly from 62% on day 2 to 27% on day 4 and more slowly thereafter in wild-type mice (Fig. 2B). Although the initial expansion of BrdU+ pDCs was similar in DC-p35 and control mice, the decline in BrdU+ pDCs was significantly slower in DC-p35 mice (Fig. 2B). These data are consistent with an increase in the survival of recently expanded mDCs and pDCs in DC-p35 mice in an immune response.

To determine whether increased survival of DCs in DC-p35 mice would lead to an increase in T cell priming in vivo, CD8+ T cells from OVA-specific T cell receptor (TCR) transgenic mice (OT1) labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE) were transferred into recipient mice, followed by immunization with DCs from DC-p35 or control mice pulsed with the OVA antigen (22). DCs from DC-p35 mice induced stronger T cell proliferation, assayed by CFSE dilution and expansion of OT1 T cells stained with an H-2Kb/OVA tetramer (Fig. 2C, upper and middle panels) (22). Similarly, DCs from DC-p35 mice pulsed with OVA antigen induced increased proliferation of OVA-specific CD4+ OT2 T cells in vivo (Fig. 2C, lower panels) and in vitro (Fig. 2D). Enhanced immune responses were also observed in DC-p35 mice by immunization with protein antigens (fig. S7). Thus, DCs from DC-p35 mice had an enhanced capacity to induce antigen-specific immune responses.

In adoptive-transfer experiments, DCs from DC-p35 or lpr mice were found to be potent inducers of antinuclear antibodies (ANAs) in recipient mice (Fig. 2E). Bcl-2 have also been shown to prolong the survival and enhance the immunogenicity of DCs (30). However, compared with p35-transduced DCs, bcl-2–transduced DCs were less efficient at inducing ANAs (Fig. 2E; fig. S7F). Thus, in addition to inducing enhanced antigen-specific immune responses, p35-DCs have the potential to induce autoantibody production.

To further investigate the potential for increased DC survival in inducing autoimmunity, we first compared spontaneous production of autoantibodies in both young and old DC-p35 mice. We detected no increases of ANAs in DC-p35 mice at 3 and 6 months of age (fig. S8). However, ANAs could be detected in 9-month-old DC-p35 mice (fig. S8). By 12 months of age in DC-p35 mice, ANAs could be found in most DC-p35 mice, but not in nontransgenic littermates or age-matched T-p35 and B-p35 mice (Fig. 3A). Significant ANA production was also detected in the independent DC-p35 transgenic line 2 (Fig. 3A), suggesting that autoantibody production was not due to a positional effect of the transgene. Unexpectedly, crossing DC-p35 to T-p35 mice (DC/T-p35) or B-p35 mice (DC/B-p35) did not significantly affect the titers of ANAs (Fig. 3A), suggesting that additional inhibition of apoptosis by p35 in T or B cells was not an important factor in promoting autoantibody production in DC-p35 mice. Substantial DC accumulation and lymphocyte infiltration in the lung were evident in all 12-month-old DC-p35 mice with detectable ANAs, but not in T-p35 or B-p35 mice (Fig. 3B). Most 12-month-old DC-p35 mice (>80%) also displayed immunogloblin G (IgG) deposition in the glomeruli of kidneys (Fig. 3B), whereas lymphocyte infiltration in the livers and kidneys was detectable in only ∼20% of 12-month-old DC-p35 mice (26). Our data indicate that DC-p35 mice on the C57BL/6 background developed autoimmune manifestations at old ages.

Fig. 3.

Autoimmunity in DC-p35 mice on the C57BL/6 background. (A) Enzyme-linked immunosorbent assay (ELISA) for ANAs in the sera of 12-month-old control (–), DC-p35, T-p35, B-p35, and DC/T-p35 (left) and DC/B-p35 transgenic mice on the C57BL/6 background (right). (B) Sections of spleens from 12-month-old control, DC-p35, T-p35, and B-p35 transgenic mice were stained with anti-CD11c, anti-CD4, anti-CD8, and anti-B220. Areas of DC accumulation in the spleens of DC-p35 mice are highlighted by arrows. Sections of lungs were stained with hematoxylin and eosin (H&E). Lymphocyte infiltration is highlighted by an arrow. Sections of kidneys were stained with FITC–anti-IgG.

A four-generation backcross of DC-p35 mice to the autoimmune-prone MRL background induced significant DC accumulation at 6 months (Fig. 4A), and greater numbers of CD11c+ DCs were found outside the B cell areas and in T cell areas in the spleens of MRL/DC-p35 mice (Fig. 4F). An increase in spontaneous activation of T cells was also observed in 6-month-old MRL/DC-p35 mice (Fig. 4B) (26), although again, normal numbers of CD4+FoxP3+ Treg cells were present (Fig. 4B). DC accumulation and chronic lymphocyte activation were similar in DC-p35 and lpr mice, but MRL/DC-p35 mice did not show expansion of the unusual TCRαβ+B220+CD4CD8 double-negative T (DNT) cells that are abundant in MRL/lpr mice (Fig. 4C), suggesting that the accumulation of DNT cells is not caused by apoptosis deficiency in DCs. MRL/DC-p35 developed significant levels of ANAs at 4 months of age (Fig. 4D), and by 6 months of age, most MRL/DC-p35 mice had developed ANAs (Fig. 4D). Sera from these DC-p35 mice showed speckled nuclear staining of Hep2 cells similar to that of the lpr sera (Fig. 4E). Lymphocyte infiltrations near the bronchi in the lungs and in the liver and kidney surrounding the blood vessels, and IgG deposition in the glomeruli of kidneys, were observed in all 6-month-old MRL/DC-p35 and MRL-lpr mice with detectable ANAs (Fig. 4G). Therefore, DC-p35 mice have an earlier onset of autoimmunity on the MRL background.

Fig. 4.

DC accumulation, lymphocyte activation, and autoimmunity in DC-p35 mice on the MRL background. (A to C) Total splenocytes from 6-month-old DC-p35 mice, nontransgenic control, and lpr mice on the MRL background were stained for CD11c+CD11b+ mDCs (A) (upper panels) or plasmacytoid dendritic cell antigen–1 (PDCA-1)–positive and CD11clow pDCs (A) (lower panels), CD69 on T cells (B) (upper panels) or with FITC–anti-CD4 followed by intracellular staining with PE–anti-FoxP3 (B) (lower panels), or with FITC–anti-TCRαβ, PE–anti-CD4, PE–anti-CD8, and cychrome–anti-B220 (C). TCRαβ+ cells were gated and B220+CD4CD8 DNT cells were quantitated (C). (D) ANA titers in the sera of control (C), DC-p35, and lpr mice on the MRL background were determined by ELISA. (E) ANAs were also detected by incubating sera from 6-month-old MRL (1:40 dilution), MRL/DC-p35 (1:640), and MRL/lpr (1:640) mice with Hep2 cell slides followed by probing with FITC-conjugated anti-mouse IgG. (F) The spleen sections of 6-month-old control, DC-p35, and lpr mice on the MRL background were stained for T cells (anti-CD3, red) and B cells (anti-B220, blue) (upper panels) or DCs (anti-CD11c, blue) (lower panels) (22). (G) H&E staining of lungs, livers, and kidneys of 6-month-old mice on the MRL background (22). The kidney sections were also stained with FITC–anti-mouse IgG.

Although lymphocytes are likely to be essential in inflicting autoimmune damage, our finding that apoptosis in DCs can independently lead to autoimmune manifestations suggests that these cells may be key initiators of autoimmune responses in individuals harboring apoptosis deficiencies. Negative selection and the development of Treg cells in the thymus were apparently normal in DC-p35 mice (fig. S6), suggesting that only peripheral tolerance is affected. DC accumulation due to apoptosis deficiency in DC-p35 mice may selectively induce overactivation of responder lymphocytes, resulting in the onset of systemic autoimmunity. On the autoimmune-prone MRL background, DC-p35 mice developed accelerated autoimmune responses (Fig. 4), indicating that synergies between apoptosis deficiency in DCs and other genetic and environmental factors likely will induce more severe autoimmune symptoms. The critical role of DC apoptosis in regulating peripheral tolerance suggests that targeting DCs may represent an effective therapeutic approach to limiting the onset of autoimmune diseases.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S9

References and Notes

References and Notes

View Abstract

Navigate This Article