Nanoparticle dose and antigen loading attenuate antigen-specific T-cell responses

Abstract

Immune-mediated hypersensitivities such as autoimmunity, allergy, and allogeneic graft rejection are treated with therapeutics that suppress the immune system, and the lack of specificity is associated with significant side effects. The delivery of disease-relevant antigens (Ags) by carrier systems such as poly(lactide-co-glycolide) nanoparticles (PLG-Ag) and carbodiimide (ECDI)-fixed splenocytes (SP-Ag) has demonstrated Ag-specific tolerance induction in model systems of these diseases. Despite therapeutic outcomes by both platforms, tolerance is conferred with different efficacy. This investigation evaluated Ag loading and total particle dose of PLG-Ag on Ag presentation in a coculture system of dendritic cells (DCs) and Ag-restricted T cells, with SP-Ag employed as a control. CD25 expression was observed in nearly all T cells even at low concentrations of PLG-Ag, indicating efficient presentation of Ag by dendritic cells. However, the secretion of IL-2, Th1, and Th2 cytokines (IFNγ and IL-4, respectively) varied depending on PLG-Ag concentration and Ag loading. Concentration escalation of soluble Ag resulted in an increase in IL-2 and IFNγ and a decrease in IL-4. Treatment with PLG-Ag followed a similar trend but with lower levels of IL-2 and IFNγ secreted. Transcriptional Activity CEll ARrays (TRACER) were employed to measure the real-time transcription factor (TF) activity in Ag-presenting DCs. The kinetics and magnitude of TF activity was dependent on the Ag delivery method, concentration, and Ag loading. Ag positively regulated IRF1 activity and, as carriers, NPs and ECDI-treated SP negatively regulated this signaling. The effect of Ag loading and dose on tolerance induction were corroborated in vivo using the delayed-type hypersensitivity (DTH) and experimental autoimmune encephalomyelitis (EAE) mouse models where a threshold of 8 μg/mg Ag loading and 0.5 mg PLG-Ag dose were required for tolerance. Together, the effect of Ag loading and dosing on in vitro and in vivo immune regulation provide useful insights for translating Ag-carrier systems for the clinical treatment of immune disorders.

Keywords: autoimmune; nanoparticles; tolerance.

Figure 1

Soluble, PLG NP‐, and ECDI‐SP‐derived antigen cause differential transcription factor activity and T‐cell fates. C57BL/6 DCs were transduced with an IRF1 TF reporter. (a) At time 0, T cells were added with a range of soluble OVA (0– 2000 ng/ml), PLG‐OVA (8 μg OVA/mg NP, 5–100 μg NP/ml), or SP‐OVA (1:6 to 3:1 SP‐OVA:DC ratio). (b) TF activity was tracked over time by bioluminescence using IVIS imaging. (b) On Day 4, T‐cell activation was determined by flow cytometry using CD25 expression (reported as frequency distributions), and (d) the secretion of IL‐2, IL‐4, and IFNγ was measured in the supernatants using ELISA. Error bars represent SEM. N = 3 replicates per condition. Colors are consistent for conditions throughout subfigures (b–d). Asterisk represent statistical significance from no antigen control by ANOVA. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. ANOVA, analysis of variance; DC, dendritic cells; ELISA, enzyme‐linked immunosorbent assay; IFNγ, Interferon‐gamma; IL, interleukin; TF, transcription factor.

Figure 2

PLG NP‐ and SP‐Ag induce differential TF activity in DCs during Ag presentation to T cells. C57BL/6 DCs were transduced with an array of TF reporters. At time 0, T cells were added with PLG‐OVA (8 μg OVA/mg NP, 10 μg NP/ml) or SP‐OVA (1:1 ratio with DCs). TF activity was tracked by bioluminescence using IVIS imaging. (a) The dynamic TF activity of IRF1, STAT1, NF‐κB, and SMAD3 and (b) a heatmap representing the activity of SP‐OVA treatment relative PLG‐OVA in all TFs measured. N = 3 replicates per condition. DC, dendritic cells; IFNγ, Interferon‐gamma; IL, interleukin; TF, transcription factor.

Figure 3

(a) The activity of IRF1 induced during cognate antigen presentation is positively influenced by concentration and antigen loading of PLG‐Ag NPs and SP‐OVA. C57BL/6 DCs were transduced with an IRF1 TF reporter. At time 0, T cells were added with a PLG‐OVA or SP‐OVA. PLG‐OVA particles of various Ag‐loadings (2, 4, 8, or 80 μg OVA/mg NP) were added at a low or high concentration (10 or 100 μg/ml). SP‐OVA were prepared using 1×, 2×, 4×, or 8× OVA (1, 2, 4, or 8 mg/ml per 3.2 × 10⁸ SPs) during ECDI coupling and were added at a 3:1 ratio to DCs. TF activity was tracked over time by bioluminescence using IVIS imaging. (b) On Day 4, IL‐2, IL‐4, and IFNγ were measured in the supernatants. N = 3 replicates per conditions. Statistical differences compared to no antigen control found using ANOVA. a: p < 0.01, b: p < 0.001, c: p < 0.0001. ANOVA, analysis of variance; DC, dendritic cells; ECDI, carbodiimide; PLG, poly(lactide‐co‐glycolide) nanoparticles; SP, splenocytes; TF, transcription factor.

Figure 4

PLG‐Ag NP‐induced tolerance induced in dependent on antigen‐loading in the DTH model. (a) Mice were intravenously injected with PLG‐OVA NPs prophylactically (Day −7 relative to immunization) or therapeutically (Day 10 relative to immunization). Mice were immunized with OVA/CFA and on Day 14 were primed with an intradermal injection of 1 mg/ml OVA or irrelevant PLP in the pinna of the ear. Ear thickness was determined before priming or 24 h following priming. (b) Mice were not treated (NT) or injected with 50 μg OVA or 1 mg of PLG NP containing a range of antigen loading (0–150 μg OVA/mg NP). (c) The amount of Ag delivered by PLG‐Ag was fixed at 8 μg delivered by 1, 2, or 4 mg of PLG‐OVA NPs (loadings of 8, 4, or 2 μg OVA/mg NP). Statistical differences were determined by two‐way ANOVA with Sidak's multiple comparisons test with nonsignificant differences between test ear and control ear indicated (p > 0.05). N = 5 mice per condition. ANOVA, analysis of variance; PLG, poly(lactide‐co‐glycolide); PLP, proteolipid peptide; SP, splenocytes; TF, transcription factor.

Figure 5

PLG‐Ag NP‐induced tolerance is NP dose‐dependent even at high antigen loading in the EAE model. Mice were intravenously injected with PLG‐PLP NPs 7 days before immunization with PLP/CFA. A range of doses were evaluated (0.01–1 mg/mouse) for these high antigen content particles (128 μg PLP/mg NP). Statistical differences from PBS control were determined by Kruskal–Wallis test (one‐way ANOVA nonparametric) with Dunn's multiple comparisons test. N = 5 mice per condition. ANOVA, analysis of variance; PLG, poly(lactide‐co‐glycolide); PLP, proteolipid peptide.