Annexin1 regulates DC efferocytosis and cross-presentation during Mycobacterium tuberculosis infection

The phagocytosis of apoptotic cells and associated vesicles (efferocytosis) by DCs is an important mechanism for both self tolerance and host defense. Although some of the engulfment ligands involved in efferocytosis have been identified and studied in vitro, the contributions of these ligands in vivo remain ill defined. Here, we determined that during Mycobacterium tuberculosis (Mtb) infection, the engulfment ligand annexin1 is an important mediator in DC cross-presentation that increases efferocytosis in DCs and intrinsically enhances the capacity of the DC antigen–presenting machinery. Annexin1-deficient mice were highly susceptible to Mtb infection and showed an impaired Mtb antigen–specific CD8⁺ T cell response. Importantly, annexin1 expression was greatly downregulated in Mtb-infected human blood monocyte–derived DCs, indicating that reduction of annexin1 is a critical mechanism for immune evasion by Mtb. Collectively, these data indicate that annexin1 is essential in immunity to Mtb infection and mediates the power of DC efferocytosis and cross-presentation.

CD8⁺ T cells are essential in resistance to Mycobacterium tuberculosis (Mtb) (1–3). Despite the fact that Mtb is a phagosomal pathogen and initiates antigen presentation via the MHC class II (MHC-II) endocytic pathway, Mtb antigens also access the MHC-I–restricted pathway by a mechanism termed cross-presentation (4, 5). This mechanism allows antigen-presenting cells (APCs), mainly DCs, to efficiently present antigens of exogenous origins to MHC-I–restricted CD8⁺ T cells (6). Several studies have demonstrated that phagocytosis of apoptotic cells, a process called efferocytosis, is an important source of antigens for cross-presentation by DCs (7–10). Kaufmann’s group further showed that engulfment of apoptotic vesicles released from Mtb- or bacillus Calmette-Guerin–infected (BCG-infected) macrophages by DCs leads to CD8⁺ T cell activation (11, 12). In line with these findings, we have demonstrated that Mtb evades host immunity by inhibiting apoptosis and promoting necrosis in infected macrophages (13–15). Importantly, increased apoptosis in Mtb-infected macrophages initiates an early protective immunity to pulmonary tuberculosis (16). Although the role of apoptosis in Mtb infection has been widely studied, our understanding of the mechanisms involved in selective recognition and uptake of apoptotic cells/vesicles (efferocytosis) by DCs for cross-presentation is still very limited.

A recent study by Behar’s group demonstrated that apoptosis per se is not intrinsically bactericidal, but is dependent on efferocytosis by macrophages to control Mtb growth (17). Whereas macrophages are highly efficient in efferocytosis and play an important role in innate immunity to Mtb (17), little is known about the mechanism or mechanisms of efferocytosis by DCs as well as its contribution to immunity against Mtb. Similar to macrophages, efferocytosis by DCs depends on the expression of engulfment ligands, also known as “eat-me” signals, on the surface of dying cells and multiple receptors on the surface of phagocytic cells. These receptors can directly interact with apoptotic cells, but most interactions occur indirectly through bridging proteins (18). The best characterized eat-me signal is the exposure of phosphatidylserine (PS) at the outer leaflet of the plasma membrane (19). Although necessary, PS alone is not efficient in triggering engulfment. Thus, other eat-me signals are required for an optimum uptake of apoptotic cells (20). Though in vitro studies have identified several classes of receptors implicated in efferocytosis (21–23), knowledge of the engulfment ligands and molecular mechanisms involved in DC efferocytosis as well as cross-presentation in vivo continues to represent a major gap in our understanding of this important phenomenon.

Annexin1 (encoded by ANXA1) is a member of the annexin protein superfamily and binds to negatively charged phospholipid membranes in a calcium-dependent manner (24). The N-terminal portion of annexin1 confers its unique biological characteristics, with a wide spectrum of biological functions. In addition to its very well-described antiinflammatory role (25, 26), colocalization of PS with annexin1 on the surface of apoptotic cells enhances recognition, tethering. and internalization of apoptotic cells by endothelial cells (27). Additionally, annexin1 has recently been shown to be a key modulator of adaptive immunity through its ability to control T cell activation (28, 29). Annexin1 expression has been associated with a variety of inflammatory diseases (cystic fibrosis and ulcerative colitis), autoimmune diseases (rheumatoid arthritis and lupus erythematous), and cancer (30), but its role in infectious diseases is poorly understood.

A recent study by Gan and colleagues (31) indicated that in macrophages infected with the virulent strain of Mtb, the amino terminal of annexin1 is removed, resulting in proteolytic truncation of annexin1 and loss of its biological activities in vitro. In the current study, we sought to determine the role of annexin1 during the course of Mtb infection in vivo. We found that annexin1-deficient mice (Anxa1^–/–) were remarkably more susceptible to pulmonary Mtb infection than WT mice. The high levels of pulmonary bacterial burden and mortality in Anxa1^–/– mice were associated with reduced Mtb antigen–specific CD8⁺ T cell responses in the lung. By generating chimeric mice that selectively lack annexin1 in T cells, we have shown that the reduction of Mtb antigen–specific CD8⁺ T cells is extrinsic to the T cell compartment. Interestingly, both in vitro and in vivo, annexin1-deficient DCs demonstrated a markedly reduced capacity to cross-present antigens to CD8⁺ T cells. The reduced capacity of annexin1-deficient DCs for cross-presentation was due to (a) the critical role of annexin1 in efferocytosis and (b) the intrinsic role of annexin1 in antigen-processing machinery. Importantly, infection of human blood monocyte–derived DCs with Mtb induced a downregulation of annexin1 gene expression, and genome-wide gene expression shows a strong correlation between annexin1 and biological pathways involved in endosome, lysosome, and autophagy. Furthermore, we showed that annexin1 is required for an optimal autophagy, suggesting an important link among annexin1, autophagy, and cross-presentation in DCs. Collectively, these data identify annexin1 as a central player in protective immunity against Mtb infection, primarily by regulating the power of DC cross-presentation.

Anxa1^–/– mice are highly susceptible to Mtb infection. To investigate the role of annexin1 during Mtb infection, we initially evaluated survival of WT and annexin1-deficient mice (Anxa1^–/–) infected i.v. with a high dose (~10⁸ CFU) or low dose (~10⁶ CFU) of virulent Mtb (H37Rv). Anxa1^–/– mice were highly susceptible to Mtb infection, and all succumbed to death (Figure 1, A and B). This increased susceptibility of Anxa1^–/– mice to infection was corroborated with higher numbers of pulmonary Mtb (Figure 1C). We next infected WT and Anxa1^–/– mice with a low dose (50–100 CFU) of H37Rv via an aerosol route to assess both pulmonary bacterial burden and survival. Similar to the i.v. model, Anxa1^–/– mice showed higher bacterial burden at 35 and 90 days after infection (Figure 1D). This impaired control of bacterial growth in Anxa1^–/– mice correlated with a significant decrease in survival (Figure 1E). Analyses of histopathology indicated that lungs of Anxa1^–/– mice were affected with a diffuse chronic active histiocytic pneumonitis (Supplemental Figure 1A; supplemental material available online with this article; doi:10.1172/JCI77014DS1), with large numbers of stainable mycobacterial organisms (Supplemental Figure 1B). Together, these results indicate that annexin1 plays a critical role in protection against Mtb infection.

Figure 1

Anxa1^–/– mice are highly susceptible to Mtb infection. (A and B) Survival of WT and Anxa1^–/– mice (n = 9–10/group) infected i.v. with approximately 10⁸ (A) or approximately 10⁶ (B) virulent Mtb bacilli. (C) Lung tissues from WT (upper panel) and Anxa1^–/– (lower panel) mice stained with Ziehl–Neelsen for detection of Mtb bacilli after 20 days infection (i.v.) with approximately 10⁸ virulent Mtb. Original magnification, ×40. (D) Bacterial burden in the lungs after 14, 35, and 90 days of aerosol infection with Mtb (n = 5/group). (E) Survival of WT and Anxa1^–/– mice (n = 7/group) after aerosol infection with virulent Mtb. *P < 0.05 (1-way ANOVA). Survival data were analyzed by log-rank test.

Anxa1^–/– mice have an impaired antigen-specific CD8⁺T cell response. Both CD4⁺ and CD8⁺ T cells are important effectors in protective immunity against Mtb (1–3, 32). We next assessed the quantity of pulmonary T cells in WT and Anxa1^–/– mice after aerosol infection with virulent Mtb. There was no significant difference in the frequency of CD4⁺ or CD8⁺ T cells (Supplemental Figure 1, C and D). However, the frequency of Mtb32-specific CD8⁺ T cells was significantly reduced in the lungs of Anxa1^–/– mice compared with WT mice at days 35 and 56 after infection (Figure 2, A and B). We next investigated whether infection with an avirulent strain would also result in reduced antigen-specific T cell responses in Anxa1^–/– mice. We adoptively transferred (i.v.) CFSE-labeled OT-I TCR–transgenic CD8⁺ T cells, which are specific to OVA peptide (SIINFEKL), to WT and Anxa1^–/– mice, followed by a footpad infection with the recombinant BCG-expressing OVA (BCG-OVA) 1 day later. Four days after infection, we detected a significantly smaller population of SIINFEKL-specific CD8⁺ T cells in the draining popliteal lymph nodes (popLN) of infected Anxa1^–/– mice compared with WT mice (Figure 2, C and D). Consistent with reduced expansion of SIINFEKL-specific CD8⁺ T cells in Anxa1^–/– mice, the proliferation of SIINFEKL-specific CD8⁺ T cells was also significantly decreased, as measured by the percentage of CFSE^loCD8⁺ T cells (Figure 2, F and G). Collectively, these results indicate that following mycobacterial infection, there was a substantial reduction in antigen-specific CD8⁺ T cells in the absence of annexin1.

Figure 2

Anxa1^–/– mice have reduced Mtb antigen–specific CD8⁺ T cell responses to Mtb infection. (A and B) Frequency of Mtb32 309–18–specific CD8⁺ T cells in the lungs of WT and Anxa1^–/– mice (n = 5/group) after 14, 35, 56, and 90 days of aerosol infection with virulent Mtb (30–50 CFU). PE-conjugated streptavidin was used as a control for the tetramer staining. Numbers above outlined areas (A) indicate the percentage of CD8⁺ T cells stained with H-2D^b-Mtb32 309–18. Results of 2 independent experiments are pooled. *P < 0.05 (t test). (C–G) CFSE-labeled OT-I TCR–transgenic CD8⁺ T cells were adoptively transferred to WT or Anxa1^–/– mice (i.v.). The following day, mice were infected (footpad) with BCG-OVA. Frequency (C and D) and total cell number (E) of SIINFEKL-specific CD8⁺ T cells and T cell proliferation (F and G) from draining popLNs at day 4 after infection. Numbers above outlined areas (C) indicate the percentage of CD8⁺ T cells stained with H-2K^b–SIINFEKL; numbers above gate (F) indicate the percentage of CFSE^lo cells. Each symbol represents an individual mouse. Results are pooled from 2 independent experiments (n = 4–5/group). *P < 0.05; **P < 0.01 (t test).

Annexin1 has no intrinsic role in T cells during Mtb infection. It has been previously shown that annexin1 can act as a “tuner” to modulate the strength of TCR signaling, leading to increased T cell proliferation (28, 29). As both the quantity and quality of the T cell response are important for the control of Mtb infection, we next determined whether annexin1 directly affects T cell function or expansion in vivo. To this end, we infected mice with BCG-OVA and measured the cytotoxic activity of CD8⁺ T cells in vivo. Two weeks after infection, splenic SIINFEKL-specific CD8⁺ T cells from WT and Anxa1^–/– mice showed comparable ability to eliminate target cells pulsed with OVA peptide (SIINFEKL; Figure 3, A and B). To evaluate the protective capacity of T cells in the presence or absence of annexin1 against pulmonary infection with virulent Mtb, we adoptively transferred (i.v.) equal numbers of purified naive CD3⁺ T cells from WT or Anxa1^–/– mice to Rag-deficient recipient mice. Quantitative PCR confirmed that the expression of the annexin1 gene was disrupted in Anxa1^–/– T cells. Importantly, the absence of annexin1 had no effect on the expression of other annexin family members or genes involved in eicosanoid pathways (Supplemental Figure 2A). One day after T cell transfer, recipient mice were infected with Mtb by aerosol (Figure 3C). Three weeks after infection, we detected no difference in bacterial burden in the lungs or spleens of recipient mice that received WT or Anxa1^–/– T cells (Figure 3, D and E). To further rule out the potential intrinsic role of annexin1 in T cells, we generated chimeric mice in which only T cells lacked annexin1 (T cell^Anxa1–/– chimeric mice). As shown in Supplemental Figure 2B, irradiated α/β T cell receptor–deficient mice (Tcra^–/–) were reconstituted with BM cells from Tcra^–/– and Anxa1^–/– mice mixed in a 4:1 ratio, respectively. Thus, T cells in reconstituted T cell^Anxa1–/– mice are only derived from Anxa1^–/– precursors, whereas the non-T cells are predominantly Anxa1^+/+. Control chimeric mice were generated in an equivalent manner by using a BM mixture from Tcra^–/– mice and WT littermates (T cell^Anxa1+/+ chimeric mice). Twelve weeks after reconstitution, by using real-time PCR, we confirmed the absence of annexin1 in T cells purified from T cell^Anxa1–/– mice (Supplemental Figure 2C). Considering that a small fraction of APCs were annexin1 deficient in T cell^Anxa1–/– mice, we next generated BM–derived DCs (BMDCs) from both groups of chimeric mice to evaluate their functions. Upon LPS stimulation, there were no significant differences in costimulatory molecule expression (Supplemental Figure 2D) or cytokine production (Supplemental Figure 2E), indicating there were no defects in DCs from the chimeric animals. To evaluate T cell responses as well as bacterial growth, we infected (aerosol) chimeric mice with virulent Mtb. The proportions of total CD4⁺ T and CD8⁺ T cells in the lungs of T cell^Anxa1–/– mice were equivalent to those of T cell^Anxa1+/+ chimeric mice at different time points after infection, indicating that both groups were similarly reconstituted (Supplemental Figure 2, F and G). We did not find differences in the frequency of Mtb32-specific CD8⁺ T cells (Figure 3, F and G) or bacterial burden in the lungs (Figure 3H) or spleens (Supplemental Figure 2H) from either group of chimeric mice. Thus, annexin1 expression in T cells is dispensable for protection against pulmonary Mtb infection.

Figure 3

Expression of annexin1 in T cells is not required for resistance to pulmonary Mtb infection. (A and B) In vivo cytotoxic activity against target cells coated with OVA peptide (SIINFEKL) in WT and Anxa1^–/– mice (n = 4/group) at day 14 after i.v. infection with BCG-OVA. Numbers above bracketed lines (A) indicate the percentage of unpulsed CFSE^lo (left) and SIINFEKL-coated CFSE^hi (right) cells. (B) Percentage of elimination of target cells coated with SIINFEKL was calculated as described in Methods. (C) Rag-deficient mice were used as recipients for splenic T cells (CD3⁺) purified from naive Anxa1^–/– or WT mice. Recipient mice (n = 5/group) were infected with Mtb by the aerosol route within 24 hours after transfer of purified T cells. Bacteria burden in the lungs (D) and spleen (E) after 3 weeks of infection. Each symbol represents an individual mouse. Data are representative of 2 independent experiments. P < 0.05, 1-way ANOVA. (F–H) Mixed BM chimeric mice were generated as described in Methods. Twelve weeks after reconstitution, mice were infected via aerosol route with virulent Mtb. Frequency of Mtb32-specific CD8⁺ T cell (F and G) and bacterial burden (H) in the lung of T cell^Anxa1+/+ or T cell^Anxa1–/– chimeric mice after 35 and 90 days of Mtb infection. PE-conjugated streptavidin was used as a control for the tetramer staining. (H) Numbers above outlined areas indicate the percentage of CD8⁺ T cells stained with H-2K^b–SIINFEKL. Ten mice/group were individually analyzed. There was no significant difference between groups (t test).

Expression of annexin1 on apoptotic vesicles enhances DC efferocytosis and the T cell response. Previous studies from our and other groups (11, 16, 33) have shown that increased apoptosis in Mtb-infected macrophages enhances T cell–mediated immunity via cross-presentation. Furthermore, a study by Gan and colleagues revealed that annexin1 is essential for forming the apoptotic envelope (31). Therefore, we investigated whether differences in the macrophage cell death program could be responsible for reducing protection against Mtb in Anxa1^–/– mice. We infected BM–derived macrophages (BMDMs) from WT and Anxa1^–/– mice with different MOI of virulent Mtb in vitro and evaluated the cell death program. Two days after infection, both groups of macrophages presented equivalent levels of apoptosis and necrosis, as measured by a cell death ELISA (Supplemental Figure 3, A and B). Similarly, there was no significant difference in necrosis (7AAD⁺) or apoptosis (annexinV⁺7AAD^–) of BMDMs infected with avirulent strain H37Ra (Supplemental Figure 3, C–E).

Efferocytosis is also a source of antigens for CD8⁺ T cells via cross-presentation (8, 12). Since annexin1 expression on the surface of apoptotic cells can act as a ligand that mediates engulfment in vitro (27), we next determined whether the presence of annexin1 on the surface of apoptotic cells is important for the recognition and uptake by DCs in vivo. Splenocytes from WT or Anxa1^–/– mice were labeled with CFSE, subjected to osmotic shock to induce apoptosis (34), and injected (i.v.) into recipient WT mice. Similar to findings in the previous study by Iyoda and colleagues (34), CFSE⁺ apoptotic cells were found within both CD11c⁺ and CD8⁺ DCs located just outside of the splenic marginal zone, a region delineated by SIGN-R1⁺ marginal zone macrophages (Figure 4A). These results suggested that CD8⁺ DCs, a dominant source of cross-presenting APCs (34), were able to take up and process apoptotic bodies. To quantify this observation, CD11c⁺ MHC-II^hi DCs were enriched from the spleens of recipient WT mice by CD11c-positive selection (Supplemental Figure 5A) and assessed by flow cytometry. As shown in Figure 4B, only the CD8⁺ subset of DCs had taken up CFSE⁺ apoptotic cells. Importantly, the frequency and total number of CD8⁺ DC^WT that engulfed WT apoptotic cells were significantly higher than CD8⁺ DC^WT that engulfed Anxa1^–/– apoptotic cells (Figure 4, C and D). To further support the importance of annexin1 on apoptotic cells for efferocytosis, we next injected CFSE-labeled apoptotic splenocytes from WT mice into WT or Anxa1^–/– mice. As all donor apoptotic cells contained annexin1, no significant differences were observed in the uptake of apoptotic cells between CD8⁺ DC^WT and CD8⁺ DC^Anxa1–/– (Supplemental Figure 6, B–D).

Figure 4

Expression of annexin1 on apoptotic cells regulates efferocytosis. (A–D) CFSE-labeled WT or Anxa1^–/– apoptotic cells were injected (i.v.) into WT mice (n = 3/group). (A) After 1 hour, localization of injected apoptotic cells was examined in the sections of spleen. The sections were stained with anti-CD11c or anti-CD8 (red) and SIGN-R1 (blue marginal zone [MZ] macrophages). Scale bars: 100 μm. The white squares are enlarged views showing the majority of CFSE-labeled apoptotic cells (green) were localized with either CD11c⁺ or CD8⁺ cells in red pulp. Scale bars: 30 μm. B, B cells; T, T cells. (B–D) CFSE-labeled apoptotic cells were selectively taken up by CD8⁺CD11c⁺ DCs. One hour after injection of apoptotic cells, splenic DCs were enriched with positive selection and CD8⁺ DCs were gated from the CD11c⁺MHC-II^hi population. Frequency (B and C) and total number (D) of CD8⁺ DCs that engulfed apoptotic cells. *P < 0.05 (t test).

We next evaluated whether the presence of annexin1 on the surface of apoptotic vesicles released from mycobacterial-infected macrophages is also important for the recognition and uptake by DCs. Apoptotic vesicles were purified from WT or Anxa1^–/– BMDMs infected with BCG-OVA. No difference in bacterial burden was observed between WT and Anxa1^–/– BMDMs (Supplemental Figure 4A), and there were no viable bacteria in apoptotic vesicles, while OVA protein was detectable by Western blot (data not shown). Scanning electron microscopy revealed vesicle formation of WT and Anxa1^–/– BMDMs undergoing infection-induced apoptosis (Supplemental Figure 4B), and the absence of annexin1 on the membrane of Anxa1^–/– vesicles was confirmed by Western blot (Supplemental Figure 4C). Considering exosomes also can transfer antigens for cross-presentation independently of apoptotic vesicles (35), we further characterized the vesicles in our study. Using flow cytometry, we found that the size of WT and Anxa1^–/– vesicles was similar (Supplemental Figure 4, D and E). WT and Anxa1^–/– vesicles presented a heterogeneous population, with the majority of vesicles in the range of 0.22 μm (Supplemental Figure 4E). Similarly, analysis by a particle sizer using a low-angle laser light-scattering device indicated that the size of vesicles was approximately 0.1–0.2 μm in WT and Anxa1^–/– vesicles, as indicated by size and polydispersity(Supplemental Figure 4F). Both vesicles presented a negative surface charge, as expected (zeta potential; Supplemental Figure 4F). Finally, detection of surface phosphatidylserine by annexinV confirmed that both WT and Anxa1^–/– purified vesicles were derived from apoptotic cells (Supplemental Figure 4G) and were not exosomes.

To determine whether efferocytosis by DCs is modulated by vesicular expression of annexin1, WT and Anxa1^–/– vesicles were labeled with PKH26 fluorochrome and incubated separately with WT BMDCs ( DC^WT). After 30 minutes of incubation, DC^WT were stained with Hoechst and Alexa Fluor 488–conjugated phalloidin to detect nuclei and F-actin, respectively, and subjected to confocal microscopy. The uptake of WT vesicles (~80%) was significantly higher than that of Anxa1^–/– vesicles (~20%) by DC^WT (Figure 5, A and B). Colocalization of F-actin with the vesicles indicated that WT vesicles were confined to the cytosol (Figure 5A). Additionally, the fluorescence intensity ratio of cytosolic vesicles relative to cell nuclei, a quantitative analysis of cellular uptake, was also significantly increased using WT vesicles compared with Anxa1^–/– vesicles (Supplemental Figure 4H).

Figure 5

Expression of annexin1 on apoptotic vesicles regulates efferocytosis. (A and B) In vitro efferocytosis assay with WT BMDCs. (A) Confocal microscopy of WT BMDCs (DC^WT) incubated with PKH26-labeled WT (upper panel) or Anxa1^–/– (bottom panel) apoptotic vesicles (red). DC^WT were stained for F-actin (green) and nuclei (blue). The right pictures represent an overlay demonstrating the presence of red vesicles in the cytoplasm of DC^WT. Scale bars: 20 μm. (B) Percentage of DC^WT-containing vesicles in the cytoplasm. Results are representative of 2 independent experiments. **P < 0.01 (t test). (C–E) PKH26-labeled WT or Anxa1^–/– vesicles mixed or not with recombinant annexin1 were injected into the footpad of WT mice (n = 3/group). (C, upper panels). Frequency of DCs (CD11c⁺MHC-II^hi) from popLNs after 24 hours of injection of apoptotic vesicles. Frequency of DCs (C, lower panels, and D) and CD8⁺ DCs (E) positive for PKH26. Numbers above outlined areas (C, upper panels) indicate the percentage of CD11c⁺MHC-II^hi cells. Numbers above gates (C, lower panels) indicate the percentage of CD11c⁺PKH26⁺ cells. *P < 0.05 (t test).

We next assessed the requirement of annexin1 for an optimal DC efferocytosis in vivo. WT or Anxa1^–/– vesicles labeled with PKH26 were injected into the footpads of WT mice. Twenty-four hours later, the uptake of PKH26⁺ apoptotic vesicles by DCs (CD11c⁺MHC-II^hi) was evaluated in the draining popLN by confocal microscopy. At this time point, we found that the majority of the apoptotic vesicles localized to the interfollicular regions of the draining popLN in close association with CD11c⁺ and CD8⁺ DCs (Supplemental Figure 5B). Similar to the apoptotic cells shown in Figure 4, B and C, quantification by flow cytometry revealed that the frequency and total number of vesicle⁺CD8⁺ DCs from mice that received WT vesicles were significantly higher than from animals receiving Anxa1^–/– vesicles (Figure 5, C–E). Interestingly, addition of recombinant annexin1 to Anxa1^–/– vesicles prior to injection into the footpad of WT recipient mice increased the uptake of Anxa1^–/– vesicles to a level similar to that observed with WT vesicles (Figure 5, C–E). Collectively, these in vitro and in vivo data suggest that expression of annexin1 on the surface of apoptotic cells and vesicles enhances efferocytosis by DCs.

To further investigate whether impairment in efferocytosis reduces cross-priming of CD8⁺ T cells in Anxa1^–/– mice, we cocultured BMDC^WT loaded with either WT or Anxa1^–/– vesicles (generated upon BCG-OVA infection) and CFSE-labeled OT-I TCR–transgenic CD8⁺ T cells in vitro. Three days later, we assessed CD8⁺ T cell proliferation. The percentage of divided OT-I cells (CFSE^lo) after cross-priming by DC^WT loaded with WT vesicles was significantly higher than DC^WT loaded with Anxa1^–/– vesicles (Supplemental Figure 5C). To extend these findings to endogenous DCs, WT mice that received CFSE-labeled OT-I cells were administered apoptotic splenocytes loaded with OVA from WT or Anxa1^–/– mice the following day. Three days after injection of apoptotic cells, the proliferation of donor CD8⁺ T cells was substantially decreased in mice that received Anxa1^–/– compared with WT apoptotic cells (data not shown). Consistent with the reduced proliferation, CD8⁺ T cell expansion was significantly decreased in the spleens of WT mice that received Anxa1^–/– apoptotic cells (Supplemental Figure 5D). Taken together, these results indicate that the presence of annexin1 on the surface of apoptotic cells or vesicles is required for an optimal CD8⁺ T cell response through DC-dependent cross-priming.

Annexin1-deficient DCs are impaired in cross-priming. Previous studies have shown that annexin1 positively regulates DC activation and maturation (36). We next investigated whether annexin1 also has an intrinsic role in DC cross-priming of CD8⁺ T cells following mycobacterial infection. We infected BMDC^WT or DC^Anxa1–/– with BCG-OVA (MOI 2), cocultured them with CFSE-labeled OT-I T cells, and measured CD8⁺ T cell proliferation. CD8⁺ T cells primed by DC^Anxa1–/– underwent less proliferation (CFSE^lo) compared with cells primed by DC^WT (Figure 6, A and B). This difference was not observed in BMDCs loaded with SIINFEKL peptide, in which antigen processing was bypassed (Supplemental Figure 6A). To test whether the presence of annexin1 in DCs is required for an optimal T cell response in vivo, we adoptively transferred (i.v.) CFSE-labeled OT-I T cells to WT mice, followed by transfer of BCG-OVA–infected BMDC^WT or BCG-OVA–infected DC^Anxa1–/– (footpad) 24 hours later (Figure 6C). Five days after adoptive transfer of DCs, we found that the proliferation of CD8⁺ T cells was significantly decreased in the draining popLN of WT mice that received BCG-OVA–infected DC^Anxa1–/– compared with those that received infected DC^WT (Figure 6D).

Figure 6

Annexin1 regulates cross-priming in DCs. (A and B) BMDC^WT and DC^Anxa1–/– were infected with BCG-OVA (MOI ~2). After 4 hours of infection, CFSE-labeled purified OT-I CD8⁺ T cells were added to the culture. At day 3, CD8⁺ T cell proliferation was evaluated using CFSE dilution. Controls were CFSE-labeled OT-I CD8⁺ T cells cultured with DC without BCG-OVA infection. Each symbol represents 1 replicate (B). Results are representative of 6 independent experiments. *P < 0.05 (t test). (C and D) CFSE-labeled, purified OT-I CD8⁺ T cells were adoptively transferred to WT mice (i.v.). The following day, BCG-OVA–infected DC^WT and DC^Anxa1–/– were transferred into the footpads of WT mice (n = 3/group). (D) Percentage of proliferation of SIINFEKL-specific CD8⁺ T cells (CFSE^lo). *P < 0.05 (t test, 2 tailed). (E–H) CFSE-labeled OT-I CD8⁺ T cells were adoptively transferred (i.v.) into WT or Anxa1^–/– mice (n = 4/group). The following day, mice were injected (i.v.) with WT apoptotic cells. At day 3, T cell proliferation (E and F), frequency (G), and total number (H) of SIINFEKL-specific CD8⁺ T cells were assessed in spleens. *P < 0.05; **P < 0.01 (t test). (I–K) CFSE-labeled OT-I CD8⁺ T cells were adoptively transferred (i.v.) to WT or Anxa1^–/– mice. The following day, mice were injected (i.v.) with WT vesicles (n = 3/group). At day 4, T cell proliferation (I), frequency, (J) and total number (K) of SIINFEKL-specific CD8⁺ T cells were evaluated in popLN. *P < 0.05 (t test). Numbers above outlined areas (A and I) indicate the percentage of T cell proliferation (CFSE^lo). Numbers above bracketed lines (E) indicate the percentage of cells that had undergone 3 or more divisions (left) or fewer than 3 divisions (right). Numbers above outlined areas (G and J) indicate the percentage of CD8⁺ T cells stained with H-2K^b–SIINFEKL.

To further evaluate the intrinsic role of annexin1 in DC cross-priming in vivo, WT or Anxa1^–/– mice that received CFSE-labeled OT-I T cells were administered OVA-loaded apoptotic cells from WT mice 24 hours later. The frequency of CD8⁺ T cell proliferation was significantly reduced in the spleen of recipient Anxa1^–/– mice compared with WT controls (Figure 6, E and F). The reduced proliferation correlated with decreased frequency (Figure 6G) and total number (Figure 6H) of OT-I T cells observed in Anxa1^–/– mice. The decreased T cell response detected in Anxa1^–/– mice was not due to impaired efferocytosis because the frequency and total number of DCs containing CFSE⁺ apoptotic cells in the spleen of WT and Anxa1^–/– mice were similar (Supplemental Figure 6, B–D). A similar defect in antigen-specific T cell proliferation was observed in the popLN of Anxa1^–/– mice when apoptotic vesicles were administered (footpad) in place of whole apoptotic cells (Figure 6, I–K). Collectively, these results indicate that the presence of annexin1 in DCs is required for an optimal cross-priming for CD8⁺ T cell proliferation.

Annexin1-deficient DCs are impaired in cross-presentation. We next determined the mechanism by which annexin1 affects the DC’s ability to activate T cells. We initially evaluated whether the impaired CD8⁺ T cell response observed in Anxa1^–/– mice upon Mtb infection is due to an impaired migratory capacity of DC^Anxa1–/– to the draining lymph nodes. To address this issue, WT and Anxa1^–/– mice were intranasally injected with LPS or PBS, and DC migration to the lung-draining lymph nodes was evaluated 1 day later. We did not observe a significant difference in the frequency (Supplemental Figure 6E) or total cell number (Supplemental Figure 6G) of DCs (CD11c⁺MHC-II^hi) in the lung-draining lymph nodes of WT and Anxa1^–/– mice injected with LPS. Similarly, the frequency (Supplemental Figure 6F) and total cell number of CD103⁺ DCs and CD11b⁺ DCs were not altered (Supplemental Figure 6, H and I). We next evaluated DC maturation upon BCG-OVA infection. BMDC^WT or DC^Anxa1–/– infected with BCG-OVA similarly upregulated the expression of CD40, CD80, MHC-II, and MHC-I (Supplemental Figures 6, J–N). Flower’s group previously showed (37) that, in some tissues, the disruption of the annexin1 gene affected the expression of other annexins as well as genes involved in the eicosanoid pathways. As lipids from eicosanoid pathways play an important role in immunity to Mtb (15, 16, 38), we next investigated whether the absence of annexin1 in DCs affects the expression of other annexin family members (annexin2, -4, -6, and -10), cPLA2, sPLA2, mPGES1, 5LO, and NOS2 between BCG-infected DC^WT and DC^Anxa1–/–. Annexin1 expression was downregulated upon BGC-OVA infection, but no significant differences in expression of the other genes were observed (Supplemental Figure 7A). Similarly, there was no difference in the production of PGE2 and LXA4, respective products of mPGES1 and 5LO enzyme activity (Supplemental Figure 7B).

To study the role of annexin1 in MHC-I–restricted cross-presentation of cell-associated antigen, purified splenic DC^WT or DC^Anxa1–/– were incubated with either OVA-loaded or unloaded apoptotic cells from β2-microglobulin deficient mice (B2m^–/–) and cocultured with CFSE-labeled OT-I T cells (39). As β2-microglobulin–deficient cells are not able to directly activate CD8⁺ T cells, DC^WT or DC^Anxa1–/– can only present antigens to CD8⁺ T cells via cross-presentation. DC^WT or DC^Anxa1–/– were loaded with SIINFEKL peptide as a positive control. Three days after coculture, we observed reduced proliferation of SIINFEKL-specific CD8⁺ T cells cross-primed by DC^Anxa1–/– compared with DC^WT (Figure 7, A and B). In contrast, there was no difference in CD8⁺ T cell proliferation primed by SIINFEKL-loaded DC^WT and DC^Anxa1–/–. To study the MHC-I–restricted cross-presentation of soluble antigen, purified endogenous DC^WT or DC^Anxa1–/– was incubated with soluble OVA protein and cocultured with CFSE-labeled OT-I T cells. Similar to cross-presentation of cell-associated antigen shown in Figure 7, A and B, OT-I T cells exhibited reduced proliferation when cross-primed by DC^Anxa1–/– compared with DC^WT. However, there was no difference in the proliferation of CD8⁺ T cells primed by SIINFEKL-loaded DC^WT and DC^Anxa1–/– (Figure 7, C and D).

Figure 7

Annexin1 regulates cross-presentation in DCs. (A and B) DCs were purified from spleen and lymph nodes of WT or Anxa1^–/– mice, and cultured with irradiated B2m^–/– splenocytes, as indicated, that were either untreated (+B2m^–/– –OVA) or loaded with 10 mg/ml soluble OVA (+B2m^–/– +OVA). As a positive control, DCs were loaded with 1 μM OVA peptide without B2m^–/– splenocytes (–B2m^–/– +SIINFEKL). CFSE-labeled OT-I CD8⁺ T cells were cultured with these cells, and T cell proliferation was determined by flow cytometry at day 3. Numbers above gate (A) indicate the percentage of CFSE^lo cells. Results are representative of 2 independent experiments. **P < 0.01 (t test). (C and D) DCs were purified from spleen and lymph nodes of WT or Anxa1^–/– mice and loaded with 2 μg/ml of OVA protein for 4 hours at 37°C or SIINFEKL. After washes, CFSE-labeled OT-I CD8⁺ T cells were cultured with these cells for 3 days and T cell proliferation was determined by flow cytometry. Numbers above gate (C) indicate the percentage of CFSE^lo cells. **P < 0.01 (t test).

To further evaluate cross-presentation, we measured the expression of MHC-I/OVA peptide (SIINFEKL) complex on the surface of BMDC^WT and DC^Anxa1–/– after BCG-OVA infection or loading with soluble OVA protein or OVA peptide (SIINFEKL). DC^WT presented higher levels of MHC-I/SIINFEKL complexes upon loading with OVA protein and infection with BCG-OVA than DC^Anxa1–/– (Figure 8, A and B). In contrast, both groups presented similar levels of MHC-I/SIINFEKL complex when loaded with SIINFEKL. Considering the expression of MHC-I was not affected by annexin1 (Supplemental Figure 6, J–N) while the levels of MHC-I/SIINFEKL complex were reduced, this suggests that there is a defect in DC antigen processing. To test this possibility, we used DQ-OVA, a self-quenching conjugate designed specifically for the study of antigen processing. Upon proteolysis, highly fluorescent peptides were released from DQ-OVA and quantified by flow cytometry. After 30 minutes of incubation, DC^WT processed DQ-OVA at higher levels than DC^Anxa1–/– (Figure 8, C and D). To evaluate antigen processing by endogenous DCs, WT and Anxa1^–/– mice were intranasally injected with DQ-OVA. Similar to the in vitro data, we observed a higher degradation of DQ-OVA by DC^WT than by DC^Anxa1–/– (Figure 8, E–G), indicating that annexin1 enhances antigen processing by DCs. Taken together, our results suggest that annexin1 is required for optimal cross-presentation by DC.

Figure 8

Reduced expression of MHC-I/peptide complex and antigen processing by Anxa1^–/– DCs. (A and B) BMDC^WT and DC^Anxa1–/– were incubated with 10 mg/ml of OVA protein or 10^–5 μM SIINFEKL or infected in vitro with BCG-OVA (MOI ~10). After 16 hours, expression of H-2K^b–SIINFEKL complex was measured on the cell surface. (A) Histogram of stimulated cells (SIINFEKL, blue; OVA protein, red; BCG-OVA, purple) versus unstimulated cells (black). (B) MFI of expression of H-2K^b–SIINFEKL complex on the surface of DC^WT and DC^Anxa1–/–. Results are representative of 3 independent experiments. *P < 0.05 (1-way ANOVA). (C and D) 10⁵ BMDCs were loaded with 15 μg/ml of a quenched OVA protein (DQ-OVA). Antigen processing was quantified as increase in MFI after 30 minutes. Controls were cells loaded with DQ-OVA for 15 minutes on ice. Results are representative of 3 to 6 independent experiments. *P < 0.05 (t test, 1 tailed). (E–G) 20 μg of DQ-OVA were injected (intranasal) into WT and Anxa1^–/– mice. Two hours later, lung cells were stained for DCs (CD11c⁺MHC-II^hi) (E), and degradation of DQ-OVA was evaluated on this cell population as increase in MFI (F and G). Numbers above outlined areas (E) indicate the percentage of CD11c⁺MHC-II^hi cells. Results are representative of 2 independent experiments. *P < 0.05 (t test).

Mtb inhibits ANXA1 expression in human blood monocyte–derived DCs. We next investigated whether annexin1 (ANXA1 in humans) was differently expressed in human blood monocyte–derived DCs in response to Mtb infection. To do so, we interrogated transcript expression levels in human blood monocyte–derived DCs from 65 individuals, before and after 18-hour infection with Mtb (40). We found that ANXA1 was significantly downregulated in human blood monocyte–derived DCs in response to Mtb infection (P = 8.3 × 10^–16; Figure 9A). To understand the effects of ANXA1 on broader Mtb-mediated regulatory networks, we investigated the genome-wide relationships between interindividual variation in ANXA1 expression levels after Mtb infection and the expression levels of 12,957 other genes measured from the same individuals. We found that ANXA1 expression levels positively correlated with the expression of 18 other genes after correction for multiple testing (P < 1.6 × 10^–4; FDR < 10%; Figure 9B and Supplemental Table 1). Interestingly, the most enriched gene ontology terms and biological pathways among the ANXA1-associated genes were related to the endosome and lysosome pathways (FDR < 5%; Supplemental Table 2). Indeed, 4 of the 18 genes (22%) associated with ANXA1 (Supplemental Figure 8) are known to be linked to the endosome membrane and/or to be involved in the lysosome pathway, significantly more than expected by chance (P = 8.1 × 10^–4). Of particular note, TANK-binding kinase 1 (TBK1) was one of the most significant differentially expressed genes, with a P value of 7.9 × 10^–6 (Supplemental Figure 8). It has been recently demonstrated that TBK1 is essential for the maturation of autophagic organelles into autolysosomes for the captured material degradation during Mtb infection (41). To establish the potential link between annexin1 and autophagy, we determined how the loss of annexin1 may affect autophagic flux. BMDC^WT or DC^Anxa1–/– were treated with vehicle or the mTOR inhibitor rapamycin, an inducer of autophagy initiation, for 6 hours, and the levels of lipidated LC3B (LC3B-II) were measured in the absence or presence of the lysosomal inhibitor bafilomycin A1. In DC^WT, rapamycin led to an increase in LC3B-II levels in cells exposed to bafilomycin (Figure 9C). In contrast, LC3B-II levels in DC^Anxa1–/– exposed to bafilomycin were significantly reduced (Figure 9C). We next calculated the average bafilomycin-induced change in LC3B-II expression (i.e., autophagic flux). Both basal and rapamycin-induced autophagic flux were significantly reduced in DC^Anxa1–/– compared with DC^WT (Figure 9D). Collectively, our results demonstrate that ANXA1 is downregulated in response to Mtb and that the regulatory network associated with ANXA1 is linked to the autophagy-dependent endosome/lysosome trafficking pathway, which is critical for cross-presentation.

Figure 9

Mtb downregulates ANXA1 in human blood monocyte–derived DCs. (A) Box plots showing log₂ expression levels of ANXA1 (y axis) in noninfected and Mtb-infected DCs. (B) Gene regulatory network of ANXA1-associated genes. We identified 18 genes for which expression levels were significantly associated with the levels of ANXA1 in Mtb-infected DCs. The thickness of the edges connecting the nodes in the network (i.e., genes) reflects the strength of the association of these genes with ANXA1 as well as among themselves. Only connections with an r² value above 0.2 are shown. Genes associated with the endosome membrane and the lysosome pathways are highlighted. (C and D) Annexin1 mediates autophagy in DCs. (C) DC^WT or DC^Anxa1–/– were cultured with vehicle or the mTOR inhibitor rapamycin for 6 hours. Bafilomycin A1 was added 2 hours prior to cell lysis and Western blots for LC3B, phospho-S6 (pS6), and β-tubulin. (D) The means of bafilomycin-induced fold increase in LC3B-II levels ± SEM (ΔLC3BII), each obtained from 6 individual experiments. * P < 0.05, KO vs. WT (t test).

Apoptosis has evolved not only for the maintenance of tissue homeostasis, but also as a mechanism to transfer the essential immunological information that directly influences the outcome of immune responses. Accumulating studies suggest that apoptosis is critical for immune tolerance or activation. Although it has been shown that the uptake of apoptotic cells by phagocytes can be a source of antigens for immune tolerance or activation, our knowledge of the mechanism(s) involved in efferocytosis of apoptotic vesicles and their capacity in instructing immunity is limited. This is an important issue, considering that during the course of apoptosis, cells shed many vesicles that can be internalized by phagocytic cells. Thus, we hypothesized that efferocytosis of apoptotic vesicles is a critical mechanism for cross-presentation by DCs and protective immunity against Mtb infection, both of which depend on annexin1.

Annexin1 was initially discovered over 30 years ago as an antiinflammatory protein mimicking the effects of glucocorticoids (42, 43). Annexin1 inhibits PLA₂, COX-2, and iNOS expression and activates formyl peptide and lipoxin A₄ receptors as well as inducing IL-10 and clearance of apoptotic cells (44). Such antiinflammatory effects of annexin1 have been shown to ameliorate diseases such as endotoxemia and peritonitis by suppressing leukocyte activation and transmigration (45, 46). Furthermore, it has been shown that activation of T cells leads to secretion of annexin1, which binds to its receptor (formyl peptide receptor [FPR]) in an autocrine/paracrine manner, promoting activation and differentiation of T cells. D’Acquisto and colleagues have shown that activation of naive T cells or Th1-polarized T cells in the presence of annexin1 promotes Th1-producing IL-2 and IFN-γ cells and dampens the Th2 phenotype (28, 29). These in vitro observations were further supported using a mouse model of rheumatoid arthritis, as annexin1 treatment exacerbated disease severity (28). Furthermore, the same group has demonstrated that annexin1-deficient mice showed an increased Th2 phenotype in a mouse model of allergic inflammation (29). Here, we have demonstrated that Anxa1^–/– mice were highly susceptible to Mtb infection and that this was associated with decreased Mtb-specific CD8⁺ T cell responses. By using a T cell–adoptive transfer model as well as generating chimeric annexin1-deficient T cells, we found, however, that annexin1 has no intrinsic role in the regulation of T cell immunity to Mtb infection. The disparate results between our group and others might reflect the immunogenic role of annexin1 in infection compared with its classical tolerogenic role in inflammatory diseases.

Physiological apoptosis has an essential role in development, differentiation, and tissue homeostasis (47). Although the tolerogenic role of annexin1 in recognition and engulfment of apoptotic cells, which results in dampening of the immune response, has been well established (48), little is known about the immunogenic role of annexin1 during infectious diseases. It has only recently been accepted that apoptosis is an important host defense mechanism against microbial infections. Albert and colleagues initially demonstrated the importance of cross-priming through apoptotic cells in antiviral immunity (8). Others (11, 12) as well as our group (14, 16) have previously shown that increased apoptosis in Mtb-infected macrophages initiates an early cross-priming event that markedly enhances T cell–mediated immunity to infection. Activation of pattern recognition receptors, such as Toll-like receptors, can substantially differentiate immunogenic versus tolerogenic apoptosis (49), but the mechanism is still unknown. Winau and coworkers (12) initially demonstrated that generation of apoptotic vesicles from mycobacterial infected macrophages has the capability to cross-prime CD8⁺ T cells. Our current study provides the mechanistic foundation for this observation by demonstrating that the presence of annexin1 on apoptotic vesicles was essential for transfer of antigens from apoptotic vesicles to DCs for cross-presentation and activation of CD8⁺ T cells. Furthermore, we have demonstrated that annexin1 plays an intrinsic role in autophagy and the antigen-processing machinery of DCs, which enhances the efficiency of DCs in presenting antigens to CD8⁺ T cells. Accordingly, we also demonstrated that Mtb substantially suppresses annexin1 expression in human blood monocyte–derived DCs, highlighting the importance of annexin1 in host immunity to Mtb infection. Thus, our results collectively indicate that annexin1 plays a critical role in immunity to Mtb infection by regulating efferocytosis as well as antigen cross-presentation by DCs.

Lung CD103⁺ DCs and splenic CD8⁺ DCs have a greater ability to acquire and cross-present apoptotic cell–associated antigens in vivo than CD11b⁺ DCs and CD8^– DCs, respectively (34, 50, 51). Indeed, mice deficient in the transcription factor BATF3, which lack both CD103⁺CD11b^– DCs in the lungs and CD8⁺ DCs in the spleen, are defective in cross-presentation (39, 52). Corroborating our findings that annexin1 plays an intrinsic role for cross-presentation of apoptotic cells and vesicles in DCs, it has been demonstrated that both CD8⁺ DCs and CD103⁺ DCs express more annexin1 at the mRNA level than CD8^– DCs and CD103^–CD11b⁺ DCs, respectively (53).

The mechanisms by which antigens derived from efferocytosis are processed in DCs for cross-presentation are still poorly understood. Antigen cross-presentation is influenced by many factors, including efficiency of antigen capture, the biochemical nature of antigens, expression levels of MHC-I, and antigen processing. While here we have demonstrated that the presence of annexin1 on the surface of apoptotic vesicles is required for optimal efferocytosis by DCs, we also have demonstrated an intrinsic role for annexin1 in antigen processing (Figure 10). There are 2 major intercellular pathways involved in cross-presentation of exogenous antigens: (a) the cytosolic pathway, in which antigens that are processed by the proteasome are loaded onto MHC-I via the ER or redirected to the phagosome containing MHC-I, and (b) the vacuolar pathway, in which exogenous antigens are directly degraded in the MHC-I–containing phagosome followed by peptide loading onto MHC-I (6). However, after efferocytosis, this process may require additional receptors, as the efferosome contains an additional membrane distinct from the apoptotic vesicles. Thus, the presence of annexin1 on apoptotic vesicles within the phagosome may also facilitate the access of antigens from phagosome to the intercellular pathways for cross-presentation. Additionally, we found that DCs either loaded with exogenous antigens (DQ-OVA) or infected with BCG-OVA required annexin1 for antigen processing.

Figure 10

Regulatory role of annexin1 during Mtb infection. Mtb-infected macrophages undergoing apoptosis express annexin1 on the cell surface, which acts as an engulfment ligand. Bystander macrophages recognize this ligand and engulf apoptotic cells/vesicles containing live bacteria via efferocytosis to eliminate Mtb. Similarly, DCs recognize annexin1 and uptake apoptotic vesicles containing mycobacterial antigens. Mtb antigens are processed and loaded on MHC-I via cross-presentation. The presence of annexin1 on efferosomes, as well as in cytosol, may facilitate cross-presentation by increasing the availability of antigens in efferosomes to MHC-I, either via proteasome-ER–dependent pathway or/and vacular-dependent pathway (effero-lysosome). This will result in CD8⁺ T cell activation and expansion to control Mtb infection.

Importantly, our observations that Mtb induced downregulation of annexin1 gene expression in human blood monocyte–derived DCs and that annexin1 regulates autophagy as well as the expression of genes such as TBK1 and FAM125B, which are involved in cytoplasmic cargo trafficking, reinforce the importance of annexin1 in antigen processing. It has been recently shown that both annexin1 (54) and TBK1 (41) are essential in autophagic maturation/degradation. Autophagy is a host-defense mechanism against mycobacterial infection (55, 56) that enhances peptide presentation by DCs (57). In the current study, we have also found that annexin1 regulates autophagy in DCs and that autophagy can serve as an effective system for the delivery of exogenous antigens in cross-presentation (58, 59). Furthermore, it has been shown that following efferocytosis, autophagy promotes apoptotic cell degradation (60). Thus, it is plausible that annexin1 may influence the cargo selection and access of antigens to either cytosolic or vacuolar pathways. Another possible explanation is that, upon antigen uptake, intracellular annexin1 colocalizes with early endosomes and also with phagosomal membrane (61, 62). As annexin1 has the ability to aggregate membranes in a calcium-dependent manner (63–65), annexin1 could facilitate the fusion of phagosome with lysosome. However, as Mtb is known to arrest phagosomal maturation, this pathway may be inhibited following infection. In contrast, if annexin1 also colocalizes with the surface of efferosome, this would lead to fusion with the lysosome and antigen degradation through the vacuolar pathway. Certainly, further experiments are needed to identify the molecular mechanisms involved in this process.

The high susceptibility of Anxa1^–/– mice to Mtb infection cannot only be explained by reduced T cell–mediated immunity. We anticipate that the lack of annexin1 on apoptotic cells/vesicles also markedly reduces efferocytosis by macrophages, further impairing innate immunity to Mtb infection (Figure 7). A recent study by Martin et al. elegantly demonstrated that efferocytosis is an essential antimycobacterial effector mechanism for controlling bacterial growth (17). In addition, deficiency of annexin1 in macrophages also impairs autophagic maturation (54), which is also important for control of mycobacterial growth (55, 56). This may explain why virulent Mtb specifically targets annexin1 by removing its aminoterminal domain, resulting in a truncated nonfunctional protein (31). Although this study showed that annexin1 is required for formation of apoptotic envelopes, our data indicate that annexin1 does not affect the cell death program of macrophages infected with either virulent or avirulent strains of mycobacteria. These differences might be explained by the use of different experimental models and experimental methods used. Thus, further studies are necessary to precisely identify the role of annexin1 in the cell death program.

Despite the worldwide application of BCG vaccination and other anti–Mtb interventions, Mtb remains one of the most successful human pathogens. Approximately 2 million people die of tuberculosis annually, and 8 to 10 million new cases of active tuberculosis occur each year due to the large reservoir of asymptomatic people chronically infected with Mtb (66). The success of this pathogen, which is transmitted person to person by the aerosol route, is closely linked to its ability to subvert both innate and adaptive immune responses. Thus, by targeting annexin1 and paralyzing efferocytosis, Mtb compromises an essential innate host defense mechanism as well as DC-mediated T cell immunity via cross-presentation. Given the potential importance of efferocytosis in effective innate and adaptive immunity to Mtb infection as well as an important role for annexin1 in facilitating antigen delivery to DCs, synthetic nanoscale particles embedded with annexin1 may provide an effective delivery system for vaccination against Mtb.

View Supplemental data

View Supplemental Tables 1-2

This work was supported by the Canadian Institute of Health Research (CIHR) operating grants (MOP-130310) and National Sciences and Engineering Research Council of Canada (NSERC) (G231908) to M. Divangahi. M. Divangahi also holds a CIHR New Investigator Award. M. Tabrizian’s laboratory is supported by CIHR-Nanomedicine/Regenerative Medicine and Genome Quebec. We thank Roderick Flower (William Harvey Research Institute, Barts and The London School of Medicine) for providing Anxa1^–/– mice.

Address correspondence to: Maziar Divangahi, Assistant Professor of Medicine, Department of Microbiology and Immunology, McGill International TB Centre, McGill University, Meakins-Christie Laboratories, 3626 St. Urbain Street, Montreal, Quebec, H2X 2P2, Canada. Phone: 514.398.3864, ext. 089727; E-mail: maziar.divangahi@mcgill.ca.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2015;125(2):752–768. doi:10.1172/JCI77014.

Anthony Michel Jean Sanchez’s present address is: University of Perpignan Via Domitia, Laboratoire Performance Santé Altitude, Font-Romeu, France.

1. Chen CY, et al. A critical role for CD8 T cells in a nonhuman primate model of tuberculosis. PLoS Pathog. 2009;5(4):e1000392.
   View this article via: PubMed CrossRef Google Scholar
2. Flynn JL, Goldstein MM, Triebold KJ, Koller B, Bloom BR. Major histocompatibility complex class I-restricted T cells are required for resistance to Mycobacterium tuberculosis infection. Proc Natl Acad Sci U S A. 1992;89(24):12013–12017.
   View this article via: PubMed CrossRef Google Scholar
3. Woodworth JSM, Behar SM. Mycobacterium tuberculosis-specific CD8⁺ T cells and their role in immunity. Crit Rev Immunol. 2006;26(4):317–352.
   View this article via: PubMed CrossRef Google Scholar
4. Canaday DH, Ziebold C, Noss EH, Chervenak KA, Harding CV, Boom WH. Activation of human CD8⁺ αβ TCR⁺ cells by Mycobacterium tuberculosis via an alternate class I MHC antigen-processing pathway. J Immunol. 1999;162(1):372–379.
   View this article via: PubMed Google Scholar
5. Grotzke JE, Siler AC, Lewinsohn DA, Lewinsohn DM. Secreted immunodominant Mycobacterium tuberculosis antigens are processed by the cytosolic pathway. J Immunol. 2010;185(7):4336–4343.
   View this article via: PubMed CrossRef Google Scholar
6. Mantegazza AR, Magalhaes JG, Amigorena S, Marks MS. Presentation of phagocytosed antigens by MHC class I and II. Traffic Cph Den. 2013;14(2):135–152.
   View this article via: PubMed CrossRef Google Scholar
7. Albert ML. Death-defying immunity: do apoptotic cells influence antigen processing and presentation? Nat Rev Immunol. 2004;4(3):223–231.
   View this article via: PubMed CrossRef Google Scholar
8. Albert ML, Sauter B, Bhardwaj N. Dendritic cells acquire antigen from apoptotic cells and induce class I-restricted CTLs. Nature. 1998;392(6671):86–89.
   View this article via: PubMed CrossRef Google Scholar
9. Bosnjak L, Miranda-Saksena M, Koelle DM, Boadle RA, Jones CA, Cunningham AL. Herpes simplex virus infection of human dendritic cells induces apoptosis and allows crosspresentation via uninfected dendritic cells. J Immunol. 2005;174(4):2220–2227.
   View this article via: PubMed CrossRef Google Scholar
10. Yrlid U, Wick MJ. Salmonella-induced apoptosis of infected macrophages results in presentation of a bacteria-encoded antigen after uptake by bystander dendritic cells. J Exp Med. 2000;191(4):613–624.
    View this article via: PubMed CrossRef Google Scholar
11. Schaible UE, et al. Apoptosis facilitates antigen presentation to T lymphocytes through MHC-I and CD1 in tuberculosis. Nat Med. 2003;9(8):1039–1046.
    View this article via: PubMed CrossRef Google Scholar
12. Winau F, et al. Apoptotic vesicles crossprime CD8 T cells and protect against tuberculosis. Immunity. 2006;24(1):105–117.
    View this article via: PubMed CrossRef Google Scholar
13. Behar SM, Divangahi M, Remold HG. Evasion of innate immunity by Mycobacterium tuberculosis: is death an exit strategy? Nat Rev Microbiol. 2010;8(9):668–674.
    View this article via: PubMed Google Scholar
14. Behar SM, Martin CJ, Nunes-Alves C, Divangahi M, Remold HG. Lipids, apoptosis, and cross-presentation: links in the chain of host defense against Mycobacterium tuberculosis. Microbes Infect Inst Pasteur. 2011;13(8–9):749–756.
    View this article via: PubMed Google Scholar
15. Divangahi M, et al. Mycobacterium tuberculosis evades macrophage defenses by inhibiting plasma membrane repair. Nat Immunol. 2009;10(8):899–906.
    View this article via: PubMed CrossRef Google Scholar
16. Divangahi M, Desjardins D, Nunes-Alves C, Remold HG, Behar SM. Eicosanoid pathways regulate adaptive immunity to Mycobacterium tuberculosis. Nat Immunol. 2010;11(8):751–758.
    View this article via: PubMed CrossRef Google Scholar
17. Martin CJ, et al. Efferocytosis is an innate antibacterial mechanism. Cell Host Microbe. 2012;12(3):289–300.
    View this article via: PubMed CrossRef Google Scholar
18. Vandivier RW, Henson PM, Douglas IS. Burying the dead: the impact of failed apoptotic cell removal (efferocytosis) on chronic inflammatory lung disease. Chest. 2006;129(6):1673–1682.
    View this article via: PubMed CrossRef Google Scholar
19. Li W. Eat-me signals: keys to molecular phagocyte biology and “appetite” control. J Cell Physiol. 2012;227(4):1291–1297.
    View this article via: PubMed CrossRef Google Scholar
20. Paidassi H, Tacnet-Delorme P, Arlaud GJ, Frachet P. How phagocytes track down and respond to apoptotic cells. Crit Rev Immunol. 2009;29(2):111–130.
    View this article via: PubMed CrossRef Google Scholar
21. Albert ML, et al. Immature dendritic cells phagocytose apoptotic cells via αvβ5 and CD36, and cross-present antigens to cytotoxic T lymphocytes. J Exp Med. 1998;188(7):1359–1368.
    View this article via: PubMed CrossRef Google Scholar
22. Nakayama M, et al. Tim-3 mediates phagocytosis of apoptotic cells and cross-presentation. Blood. 2009;113(16):3821–3830.
    View this article via: PubMed CrossRef Google Scholar
23. Schulz O, Pennington DJ, Hodivala-Dilke K, Febbraio M, Reis e Sousa C. CD36 or αvβ3 and αvβ5 integrins are not essential for MHC class I cross-presentation of cell-associated antigen by CD8α⁺ murine dendritic cells. J Immunol. 2002;168(12):6057–6065.
    View this article via: PubMed CrossRef Google Scholar
24. Morand EF, Hutchinson P, Hargreaves A, Goulding NJ, Boyce NW, Holdsworth SR. Detection of intracellular lipocortin 1 in human leukocyte subsets. Clin Immunol Immunopathol. 1995;76(2):195–202.
    View this article via: PubMed CrossRef Google Scholar
25. Perretti M, D’Acquisto F. Annexin A1 and glucocorticoids as effectors of the resolution of inflammation. Nat Rev Immunol. 2009;9(1):62–70.
    View this article via: PubMed CrossRef Google Scholar
26. Perretti M, Dalli J. Exploiting the Annexin A1 pathway for the development of novel anti-inflammatory therapeutics. Br J Pharmacol. 2009;158(4):936–946.
    View this article via: PubMed CrossRef Google Scholar
27. Arur S, et al. Annexin I is an endogenous ligand that mediates apoptotic cell engulfment. Dev Cell. 2003;4(4):587–598.
    View this article via: PubMed CrossRef Google Scholar
28. D’Acquisto F, et al. Annexin-1 modulates T-cell activation and differentiation. Blood. 2007;109(3):1095–1102.
    View this article via: PubMed Google Scholar
29. D’Acquisto F, Paschalidis N, Sampaio ALF, Merghani A, Flower RJ, Perretti M. Impaired T cell activation and increased Th2 lineage commitment in Annexin-1-deficient T cells. Eur J Immunol. 2007;37(11):3131–3142.
    View this article via: PubMed CrossRef Google Scholar
30. D’Acquisto F, Piras G, Rattazzi L. Pro-inflammatory and pathogenic properties of Annexin-A1: the whole is greater than the sum of its parts. Biochem Pharmacol. 2013;85(9):1213–1218.
    View this article via: PubMed CrossRef Google Scholar
31. Gan H, Lee J, Ren F, Chen M, Kornfeld H, Remold HG. Mycobacterium tuberculosis blocks crosslinking of annexin-1 and apoptotic envelope formation on infected macrophages to maintain virulence. Nat Immunol. 2008;9(10):1189–1197.
    View this article via: PubMed CrossRef Google Scholar
32. Caruso AM, Serbina N, Klein E, Triebold K, Bloom BR, Flynn JL. Mice deficient in CD4 T cells have only transiently diminished levels of IFN-γ, yet succumb to tuberculosis. J Immunol. 1999;162(9):5407–5416.
    View this article via: PubMed Google Scholar
33. Winau F, Kaufmann SHE, Schaible UE. Apoptosis paves the detour path for CD8 T cell activation against intracellular bacteria. Cell Microbiol. 2004;6(7):599–607.
    View this article via: PubMed CrossRef Google Scholar
34. Iyoda T, et al. The CD8⁺ dendritic cell subset selectively endocytoses dying cells in culture and in vivo. J Exp Med. 2002;195(10):1289–1302.
    View this article via: PubMed CrossRef Google Scholar
35. Viaud S, et al. Dendritic cell-derived exosomes for cancer immunotherapy: what’s next? Cancer Res. 2010;70(4):1281–1285.
    View this article via: PubMed CrossRef Google Scholar
36. Huggins A, Paschalidis N, Flower RJ, Perretti M, D’Acquisto F. Annexin-1-deficient dendritic cells acquire a mature phenotype during differentiation. FASEB J. 2009;23(4):985–996.
    View this article via: PubMed CrossRef Google Scholar
37. Hannon R, et al. Aberrant inflammation and resistance to glucocorticoids in annexin 1^–/– mouse. FASEB J. 2003;17(2):253–255.
    View this article via: PubMed Google Scholar
38. Chen M, et al. Lipid mediators in innate immunity against tuberculosis: opposing roles of PGE2 and LXA4 in the induction of macrophage death. J Exp Med. 2008;205(12):2791–2801.
    View this article via: PubMed CrossRef Google Scholar
39. Hildner K, et al. Batf3 deficiency reveals a critical role for CD8α⁺ dendritic cells in cytotoxic T cell immunity. Science. 2008;322(5904):1097–1100.
    View this article via: PubMed CrossRef Google Scholar
40. Barreiro LB, Tailleux L, Pai AA, Gicquel B, Marioni JC, Gilad Y. Deciphering the genetic architecture of variation in the immune response to Mycobacterium tuberculosis infection. Proc Natl Acad Sci U S A. 2012;109(4):1204–1209.
    View this article via: PubMed CrossRef Google Scholar
41. Pilli M, et al. TBK-1 promotes autophagy-mediated antimicrobial defense by controlling autophagosome maturation. Immunity. 2012;37(2):223–234.
    View this article via: PubMed CrossRef Google Scholar
42. Blackwell GJ, Carnuccio R, Di Rosa M, Flower RJ, Parente L, Persico P. Macrocortin: a polypeptide causing the anti-phospholipase effect of glucocorticoids. Nature. 1980;287(5778):147–149.
    View this article via: PubMed CrossRef Google Scholar
43. Flower RJ. Eleventh Gaddum memorial lecture. Lipocortin and the mechanism of action of the glucocorticoids. Br J Pharmacol. 1988;94(4):987–1015.
    View this article via: PubMed CrossRef Google Scholar
44. Parente L, Solito E. Annexin 1: more than an anti-phospholipase protein. Inflamm Res. 2004;53(4):125–132.
    View this article via: PubMed CrossRef Google Scholar
45. Da Cunha EE, Oliani SM, Damazo AS. Effect of annexin-A1 peptide treatment during lung inflammation induced by lipopolysaccharide. Pulm Pharmacol Ther. 2012;25(4):303–311.
    View this article via: PubMed CrossRef Google Scholar
46. Getting SJ, Flower RJ, Perretti M. Inhibition of neutrophil and monocyte recruitment by endogenous and exogenous lipocortin 1. Br J Pharmacol. 1997;120(6):1075–1082.
    View this article via: PubMed CrossRef Google Scholar
47. Poon IKH, Lucas CD, Rossi AG, Ravichandran KS. Apoptotic cell clearance: basic biology and therapeutic potential. Nat Rev Immunol. 2014;14(3):166–180.
    View this article via: PubMed CrossRef Google Scholar
48. Weyd H, et al. Annexin A1 on the surface of early apoptotic cells suppresses CD8⁺ T cell immunity. PLoS One. 2013;8(4):e62449.
    View this article via: PubMed CrossRef Google Scholar
49. Green DR, Ferguson T, Zitvogel L, Kroemer G. Immunogenic and tolerogenic cell death. Nat Rev Immunol. 2009;9(5):353–363.
    View this article via: PubMed CrossRef Google Scholar
50. Desch AN, et al. CD103⁺ pulmonary dendritic cells preferentially acquire and present apoptotic cell-associated antigen. J Exp Med. 2011;208(9):1789–1797.
    View this article via: PubMed CrossRef Google Scholar
51. Schnorrer P, et al. The dominant role of CD8⁺ dendritic cells in cross-presentation is not dictated by antigen capture. Proc Natl Acad Sci U S A. 2006;103(28):10729–10734.
    View this article via: PubMed CrossRef Google Scholar
52. Edelson BT, et al. Peripheral CD103⁺ dendritic cells form a unified subset developmentally related to CD8α⁺ conventional dendritic cells. J Exp Med. 2010;207(4):823–836.
    View this article via: PubMed CrossRef Google Scholar
53. Gene Skyline. annexin A1 probe 10466606. Immunological Genome Project Web site. http://www.immgen.org/databrowser/index.html. Accessed November 19, 2014.
54. Kang J-H, Li M, Chen X, Yin X-M. Proteomics analysis of starved cells revealed Annexin A1 as an important regulator of autophagic degradation. Biochem Biophys Res Commun. 2011;407(3):581–586.
    View this article via: PubMed CrossRef Google Scholar
55. Gutierrez MG, Master SS, Singh SB, Taylor GA, Colombo MI, Deretic V. Autophagy is a defense mechanism inhibiting BCG and Mycobacterium tuberculosis survival in infected macrophages. Cell. 2004;119(6):753–766.
    View this article via: PubMed CrossRef Google Scholar
56. Ponpuak M, et al. Delivery of cytosolic components by autophagic adaptor protein p62 endows autophagosomes with unique antimicrobial properties. Immunity. 2010;32(3):329–341.
    View this article via: PubMed CrossRef Google Scholar
57. Jagannath C, Lindsey DR, Dhandayuthapani S, Xu Y, Hunter RL Jr, Eissa NT. Autophagy enhances the efficacy of BCG vaccine by increasing peptide presentation in mouse dendritic cells. Nat Med. 2009;15(3):267–276.
    View this article via: PubMed CrossRef Google Scholar
58. Patterson NL, Mintern JD. Intersection of autophagy with pathways of antigen presentation. Protein Cell. 2012;3(12):911–920.
    View this article via: PubMed CrossRef Google Scholar
59. Li Y, Wang L-X, Yang G, Hao F, Urba WJ, Hu H-M. Efficient cross-presentation depends on autophagy in tumor cells. Cancer Res. 2008;68(17):6889–6895.
    View this article via: PubMed CrossRef Google Scholar
60. Li W, et al. Autophagy genes function sequentially to promote apoptotic cell corpse degradation in the engulfing cell. J Cell Biol. 2012;197(1):27–35.
    View this article via: PubMed CrossRef Google Scholar
61. Diakonova M, Gerke V, Ernst J, Liautard JP, van der Vusse G, Griffiths G. Localization of five annexins in J774 macrophages and on isolated phagosomes. J Cell Sci. 1997;110(pt 10):1199–1213.
    View this article via: PubMed Google Scholar
62. Seemann J, Weber K, Osborn M, Parton RG, Gerke V. The association of annexin I with early endosomes is regulated by Ca²⁺ and requires an intact N-terminal domain. Mol Biol Cell. 1996;7(9):1359–1374.
    View this article via: PubMed CrossRef Google Scholar
63. Bitto E, Li M, Tikhonov AM, Schlossman ML, Cho W. Mechanism of annexin I-mediated membrane aggregation. Biochemistry. 2000;39(44):13469–13477.
    View this article via: PubMed CrossRef Google Scholar
64. De la Fuente M, Parra AV. Vesicle aggregation by annexin I: role of a secondary membrane binding site. Biochemistry. 1995;34(33):10393–10399.
    View this article via: PubMed CrossRef Google Scholar
65. Hu NJ, Bradshaw J, Lauter H, Buckingham J, Solito E, Hofmann A. Membrane-induced folding and structure of membrane-bound annexin A1 N-terminal peptides: implications for annexin-induced membrane aggregation. Biophys J. 2008;94(5):1773–1781.
    View this article via: PubMed CrossRef Google Scholar
66. World Health Organization. Global Tuberculosis Report 2013. Geneva, Switzerland: World Health Organization; 2013.
67. Tzelepis F, et al. Intrinsic role of FoxO3a in the development of CD8⁺ T cell memory. J Immunol. 2013;190(3):1066–1075.
    View this article via: PubMed CrossRef Google Scholar
68. Qiu CH, Miyake Y, Kaise H, Kitamura H, Ohara O, Tanaka M. Novel subset of CD8{alpha}+ dendritic cells localized in the marginal zone is responsible for tolerance to cell-associated antigens. J Immunol. 2009;182(7):4127–4136.
    View this article via: PubMed CrossRef Google Scholar
69. Smyth GK. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol. 2004;3:Article3.
    View this article via: PubMed Google Scholar
70. Storey JD, Tibshirani R. Statistical significance for genomewide studies. Proc Natl Acad Sci U S A. 2003;100(16):9440–9445.
    View this article via: PubMed CrossRef Google Scholar