Antigenic liposomes displaying CD22 ligands induce antigen-specific B cell apoptosis

Abstract

Antibodies confer humoral immunity but can also be harmful when they target an autoantigen, alloantigen, allergen, or biotherapeutic. New strategies are needed for antigen-specific suppression of undesired antibody responses, particularly to T cell-dependent protein antigens, because they elicit T cell help. Here we show that liposomal nanoparticles, displaying both antigen and glycan ligands of the inhibitory coreceptor CD22, induce a tolerogenic program that selectively causes apoptosis in mouse and human B cells. These SIGLEC-engaging tolerance-inducing antigenic liposomes (STALs, where SIGLEC is defined as sialic acid-binding Ig-like lectin) induced robust antigen-specific tolerance to protein antigens in mice, preventing subsequent immune response to challenge with the same antigen. Since development of inhibitory antibodies to FVIII is a serious problem in treatment of hemophilia A patients, we investigated the potential of this approach for inducing tolerance to FVIII in a hemophilia mouse model. STALs prevented formation of inhibitory FVIII antibodies, allowing for effective administration of FVIII to hemophilia mice to prevent bleeding. These findings suggest that STALs could be used to eliminate or prevent harmful B cell-mediated immune responses.

Figure 1. Induction of tolerance with liposomes displaying antigen and CD22 ligands.

(A) Schematic of STALs. (B) Chemical structures of CD22 ligands used for studies in mice. (C and D) CD22-dependent induction of tolerance to a T-independent (NP; C) and a T cell–dependent antigen (HEL; D). WT or Cd22-KO mice were treated on day 0 (white arrow) as shown and challenged with the immunogenic liposomes on days 15 and 30 (black arrow). Data represent mean ± SEM. (n = 8–10). (E) Titration of ^BPANeuGc and NeuGc on STALs. Titers were determined after 2 challenges with immunogenic liposomes on days 15 and 30 (n = 4). (F) Mice were tolerized to HEL at different times relative to the challenge and titers were determined 2 weeks after challenge with immunogenic liposomes and are plotted as percentage relative to immunization of naive mice (n = 4). Data represent mean ± SEM. (n = 4).

Figure 2. STALs strongly inhibit BCR signaling and cause apoptosis.

(A) Calcium flux in IgM^HEL B cells stimulated with the indicated liposomes. (B) CD86 upregulation of IgM^HEL B cells 24 hours after stimulation with the indicated liposomes. (C) In vitro proliferation of CTV-labeled IgM^HEL B cells 3 days after simulation with the indicated liposomes. (D) Annexin V versus PI staining of IgM^HEL B cells treated for 24 hours with the indicated liposomes. For quantification over time, the percentages of annexinV^–PI^– (live) cells are expressed relative to the controls treated with naked liposomes normalized to 100% at each time point and plotted as the mean ± SEM (n = 3). (E) In vivo proliferation of adoptively transferred CFSE-labeled IgM^HEL B cells 4 days after immunization with the indicated liposomes. The same number of total splenocytes was analyzed for each condition (1 × 10⁶) and gated through the IgM^a+Ly5^a+ population. (F) Analysis of the number of adoptively transferred Ly5^a+IgM^HEL B cells remaining in the spleen of recipient mice 12 days after immunization with the indicated liposomes. Quantitation represents mean ± SEM (n = 4). *P < 0.05.

Figure 3. A CD22-dependent tolerogenic program inhibits basal signaling in the Akt survival pathway and drives nuclear import of FoxO1.

(A) Western blot analysis of BCR signaling components in WT and Cd22-KO IgM^HEL B cells 30 minutes after stimulation of cells with the indicated liposomes or PBS as a control. STALs inhibit phosphorylation of signaling components of all major BCR signaling pathways and induce hypophosphorylation of Akt and FoxO1 in WT B cells, but not Cd22-deficient IgM^HEL B cells. Data are a subset of Supplemental Figure 4. (B) Analysis of FoxO1 staining in IgM^HEL B cells by confocal microscopy. Cells were stimulated for 2 hours with the indicated liposomes and stained with anti-FoxO1, phalloidin, and DAPI. Inserts are a representative cell at 3 times the magnification. Original magnification, ×63.

Figure 4. Antigen-specific tolerization of mice to strong T cell–dependent antigens.

(A) Tolerization of OVA in C57BL/6J mice. (B) Tolerization of MOG(residues 1–120) in BALB/c mice. (C) Tolerization of FVIII in BALB/c. (D) Tolerization is antigen-specific. BALB/c mice tolerized to HEL or OVA have normal responses to other antigens. Mice were immunized on day 0 (white arrows) with the indicated conditions, challenged on day 15 with immunogenic liposomes (black arrows), and titers (IgG₁) determined 2 weeks later on day 29. Black arrows show days 15 and 30. All data represent mean ± SEM (n = 4). *P < 0.05.

Figure 5. Immune tolerization to FVIII prevents bleeding in FVIII-deficient mice.

(A) WT or FVIII-deficient mice were dosed on days 0 and 15 with immunogenic liposomes (Immunogen), STALs, or left untreated. On day 30, mice were reconstituted with rhFVIII at 50 U/kg or saline. FVIII-deficient mice treated with STALs had significantly less blood loss (μl/g) over 20 minutes following tail clip than mice initially treated with immunogenic liposomes. Percentage of bleeding protection (dashed line) represents blood loss of less than 9.9 μl/g as defined by mean + 3 SDs in WT BALB/c mice. (B) FVIII titers in the 3 reconstituted groups demonstrate that bleeding prevention is accompanied by a significant reduction in anti-FVIII antibodies. Data represent mean ± SEM. A 2-tailed Student’s t test was used to establish the level of significance. NS is defined by a P value greater than 0.05.

Figure 6. STALS induce apoptosis in naive and memory human B cells.

(A) Structure of the high-affinity human CD22 ligand ^BPCNeuAc. (B–D) Activation of naive and memory human B cells is inhibited by copresentation of ^BPCNeuAc with cognate antigen (anti-IgM or anti-IgG, respectively) on liposomes, as judged by calcium flux (B), Western blot analysis of BCR signaling components (C), and CD86 upregulation (D). (E) Liposomes displaying cognate antigen and hCD22 ligands decrease viability of both naive and memory human B cells. Data represent mean ± SEM (n = 3). A 2-tailed Student’s t test was used to establish the level of significance. (F) Staining of naive (red) and memory (blue) human B cells with anti-CD22 or isotype control (gray) antibodies. Data are representative of 3 healthy donors.