Modulating antigen processing through metal-organic frameworks to bias adaptive immunity

Abstract

Vaccines induce specific immunity through antigen uptake and processing. However, while nanoparticle vaccines have elevated uptake, the impact of intracellular protein release and how this affects processing and downstream responses are not fully understood. Herein, we reveal how tuning unmodified antigen release rate, specifically through modulation of metal-organic framework (MOF) pore size, influences the type and extent of raised adaptive immunity. We use two MOFs in the NU-100x series with 1.4 nm difference in pore diameter, employ facile postsynthesis loading to achieve significant internalization of model protein antigen ovalbumin (ca. 1.4 mg/mg), and observe distinct antigen release and intracellular processing profiles influenced by MOF pore size. We investigate how this difference in release biases downstream CD8⁺, T_H1, and T_H2 T cell responses. Ovalbumin-loaded NU-1003 induced 1.8-fold higher CD8⁺:CD4⁺ T cell proliferation ratio and displayed 2.2-fold greater ratio of CD4⁺ T_H1:T_H2 cytokines compared to ovalbumin-loaded NU-1000. Antigen released from NU-1000 in vivo exhibited stronger antigen-specific IgG responses, which is dependent on CD4⁺ T cells (up to ninefold stronger long-term antibody production and 5.9-fold higher IgG1:IgG2a ratio), compared to NU-1003. When translated to wild-type SARS-CoV-2 receptor-binding domain (RBD) protein, RBD-loaded NU-1000 induced 60.5-fold higher IgG1:IgG2a compared to NU-1003. Wild-type RBD-loaded NU-1000 immunization also induced a greater breadth of epitope recognition compared to NU-1003, as evidenced by increased binding antibodies to the Omicron RBD variant. Overall, this work highlights how antigen release significantly influences immunity induced by vaccines and offers a path to employ unmodified antigen release kinetics to drive personalized protective responses.

Keywords: adaptive immunity; metal–organic frameworks; nanoscale vaccine design.

Fig. 1.

NU-1000 and NU-1003 successfully load high quantities of native ovalbumin protein antigen. (A) NU-1000 and NU-1003 load similar quantities of ovalbumin (n = 7 to 8 per group). Mean and SD shown. (B and C) Background-subtracted PXRD of NU-1000 (B) or NU-1003 (C) unloaded (dashed lines) and loaded (solid lines) MOFs illustrates the drop in crystallinity following antigen loading. In panel A, analysis was performed using an unpaired, two-tailed t test. ns = non-significant.

Fig. 2.

MOF pore size changes the antigen profile and cell uptake, which influences downstream processing kinetics. (A) Ovalbumin is released more rapidly from Ova-AF555@NU-1000 compared to Ova-AF555@NU-1003. (B) Inset of the first 4 h of release in (A). (C) Ova@MOF is taken up by BMDCs to similar extents over 24 h. Differences in the timecourse of the uptake are shown between the two different nanoparticles (n = 3 per group). (D and E) Representative cell images and colocalization of Ova-AF647 (magenta) and early endosome measured through EEA1 (D) or lysosome measured through LAMP1 (E) (yellow) from isolated BMDCs treated with Ova-AF647@NU-1000 and Ova-AF647@NU-1003 for 4 or 24 h (n = 6 to 10 per group). MOF is cyan. Mander’s overlap coefficient representing the fraction of Ova-AF647 signal colocalized with the respective organelle are shown. (F) Ova@MOF delivery to BMDCs enhances antigen presentation of MHC-I-restricted SIINFEKL epitope over 72 h compared to Free Ova (n = 2 per group). All groups were treated with 2.5 µM of ovalbumin antigen and 2.5 µM of ODN1826 adjuvant. Signal intensity is most elevated after 48 h for Ova@NU-1000 compared to Ova@NU-1003. Mean and SD shown in panels C–F. In panels D and E, analysis was performed using an unpaired, two-tailed t test within each timepoint. In panel F, analysis was performed using an ordinary two-way ANOVA, followed by a Tukey’s multiple comparisons test. *P = <0.05; **P < 0.01; ****P < 0.0001; ns = non-significant. Not all comparisons shown for simplicity.

Fig. 3.

Ova@NU-1000 and Ova@NU-1003 elicit distinct T cell proliferation profiles in vitro. (A) Splenocytes from OT-I and OT-II mice were harvested, stained, and pooled before incubating with free ovalbumin or antigen loaded within NU-100x (Ova@NU-100x). Cells were collected 3 to 5 d following incubation and stained for live flow cytometric analysis. Created with BioRender.com. (B and C) Proliferation dose–response curves of CD8⁺ OT-I (B) and CD4⁺ OT-II (C) T cells after 4 d of incubation (n = 3 per group). (D and E) Proliferation fold change of CD8⁺ OT-I (D) and CD4⁺ OT-II (E) T cells treated at 10 nM by Ova following 3 to 5 d of incubation, compared to untreated cells (n = 3 per group). Mean and SD shown in B–E. In panels D and E, analysis used an ordinary two-way ANOVA, followed by a Tukey’s multiple comparisons test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns = nonsignificant.

Fig. 4.

Ova@NU-1000 and Ova@NU-1003 elicit distinct T_H1 and T_H2 cytokine profiles. (A and B) Ova@MOFs induce greater quantities of IFN-γ (A) and IL-5 (B) compared to free ovalbumin in pooled OT-I and OT-II splenocytes following 4 d of incubation with 1000 nM of treatments by Ova (n = 3 per group). (C) Ova@NU-1000 and Ova@NU-1003 induce distinct ratios of IFN-γ:IL-5 cytokines in pooled OT-I and OT-II splenocytes following 4 d of incubation (n = 3 per group). Mean and SEM shown in all graphs. Analysis used an ordinary one-way ANOVA, followed by Tukey’s (A and C) or Šídák’s (B) multiple comparisons test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. ND = not detected. Not all significances shown for clarity.

Fig. 5.

Adj-Ova@NU-1000 and Adj-Ova@NU-1003 elicit strong ovalbumin-specific CD8⁺ T cell responses in vivo. (A) Female C57BL/6 mice (n = 3-4 per group) were treated with 3 immunizations of ovalbumin and ODN1826 as an admix or in Adj-Ova@MOFs and euthanized as per the schedule shown. Created with BioRender.com. (B and C) Comparisons of CD44⁺CD62L^- effector memory CD4⁺ (B) and CD8⁺ (C) T cells between the different treatments (n = 3 per group). (D and E) Comparisons of CD8⁺ T cells specific to the MHC-I-restricted Ova1 epitope (D) and CD4⁺ T cells specific to the MHC-II-restricted Ova2 epitope (E) of ovalbumin (n = 3-4 per group). (F) Comparisons of IFN-γ-producing T cells following ex vivo restimulation with the Ova1 peptide (n = 2-3 per group). (Left) Representative images showing raw counts and (Right) quantified spot forming cells (SFCs) per 2 × 10⁵ splenocytes. Mean and SD shown in panels B–F. In panels B and C, analysis was performed using Brown-Forsyth and Welch’s ANOVA followed by Dunnett's T3 multiple comparisons test. In panels D–F, an ordinary one-way ANOVA, followed by Tukey’s multiple comparisons test. *P < 0.05; **P < 0.01; ns = nonsignificant. Not all significances are shown for clarity.

Fig. 6.

Antigen-loaded NU-1000 and NU-1003 elicit strong and distinct antibody responses over time in vivo in female C57BL/6 mice with both Ova and RBD. (A) Ovalbumin-specific IgG reciprocal titers comparing the effects of dosage and MOF encapsulation on the humoral response on day 21, with immunizations at a 1:1 dose of Ova and ODN1826 as an admix or ODN1826 and Ova@NU-1000 on day 0 and 7 (n = 3 per group). (B) Ovalbumin-specific IgG reciprocal titers over time from ovalbumin and ODN1826 administered as an admix or coencapsulated in Adj-Ova@NU-100x (n = 3 per group). Mice were primed on day 0, then boosted on week 2 and week 4. Blood collections for serum occurred on weeks 3 to 5. (C) Long-term comparison of humoral responses between Ova@NU-1000 and Ova@NU-1003 with free ODN1826 (n = 2-3 per group), based on the injection schedule in SI Appendix, Fig. S17. (D) Difference in the ratio of ovalbumin-specific T_H2-induced IgG1 to T_H1-induced IgG2a reciprocal titers (n = 3-4 per group). (E) Wild-type RBD-specific IgG reciprocal serum titers from wild-type RBD and ODN1826 administered as an admix or RBD@NU-1000 and RBD@NU-1003 with free ODN1826 (n = 2-3 per group) based on injection and blood collection schedule in SI Appendix, Fig. S20. (F) Difference in the ratio of RBD-specific T_H2-induced IgG1 to T_H1-induced IgG2a reciprocal titers. (G) Omi-RBD-specific IgG reciprocal titer in mice following the aforementioned vaccination using wild-type RBD. Mean and SD shown in panels A–C and E. In panel A, analysis was performed using an ordinary one-way ANOVA followed by Tukey’s multiple comparisons test. In panels B, C, and E, analysis was performed using an ordinary two-way ANOVA, followed by Šídák’s (B and D) or Tukey’s (C) multiple comparisons test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns = nonsignificant. Not all significances are shown for clarity.