Nanoparticle-Conjugated Toll-Like Receptor 9 Agonists Improve the Potency, Durability, and Breadth of COVID-19 Vaccines

Abstract

Development of effective vaccines for infectious diseases has been one of the most successful global health interventions in history. Though, while ideal subunit vaccines strongly rely on antigen and adjuvant(s) selection, the mode and time scale of exposure to the immune system has often been overlooked. Unfortunately, poor control over the delivery of many adjuvants, which play a key role in enhancing the quality and potency of immune responses, can limit their efficacy and cause off-target toxicities. There is a critical need for improved adjuvant delivery technologies to enhance their efficacy and boost vaccine performance. Nanoparticles have been shown to be ideal carriers for improving antigen delivery due to their shape and size, which mimic viral structures but have been generally less explored for adjuvant delivery. Here, we describe the design of self-assembled poly(ethylene glycol)-b-poly(lactic acid) nanoparticles decorated with CpG, a potent TLR9 agonist, to increase adjuvanticity in COVID-19 vaccines. By controlling the surface density of CpG, we show that intermediate valency is a key factor for TLR9 activation of immune cells. When delivered with the SARS-CoV-2 spike protein, CpG nanoparticle (CpG-NP) adjuvant greatly improves the magnitude and duration of antibody responses when compared to soluble CpG, and results in overall greater breadth of immunity against variants of concern. Moreover, encapsulation of CpG-NP into injectable polymeric-nanoparticle (PNP) hydrogels enhances the spatiotemporal control over codelivery of CpG-NP adjuvant and spike protein antigen such that a single immunization of hydrogel-based vaccines generates humoral responses comparable to those of a typical prime-boost regimen of soluble vaccines. These delivery technologies can potentially reduce the costs and burden of clinical vaccination, both of which are key elements in fighting a pandemic.

Keywords: Drug delivery; Hydrogels; Immunoengineering; SARS-CoV-2; Vaccines.

Figure 1

Schematic representation of a subcutaneous vaccine injection in a mouse model for in vivo release. Delivery of CpG adjuvant can be achieved in different ways: in its molecular form, tethered to PEG-b-PLA NPs, or tethered to NPs and encapsulated in polymer-nanoparticle (PNP) hydrogels. PNP hydrogels are loaded with vaccine cargo, including antigen and adjuvant (CpG-NPs), and allow for sustained vaccine exposure. After subcutaneous injection of the hydrogel vaccine, vaccine components can be transported to the lymph nodes (LNs) either by drainage through antigen presenting cells (APCs) that have previously infiltrated the hydrogel, or by LN drainage of the single vaccine components themselves. Soluble vaccines, on the other hand, do not create an inflammatory niche for cell infiltration. Vaccine components are rapidly cleared from the body and drained to the lymph nodes, potentially decreasing the potency. Nanoparticle vaccine cargo, such as CpG-NPs, however, may improve immune cell activation and LN-targeting ability.

Figure 2

Design of CpG-functionalized NPs. (A) Synthetic scheme for the fabrication of CpG-based NPs. Formation of azide-terminated PEG-b-PLA NPs via nanoprecipitation followed by copper-free click chemistry with DBCO-CpG to yield to CpG-functionalized NPs. 10%, 20%, 30%, and 50% valencies were achieved by mixing different weight ratios of PEG-b-PLA and N₃-PEG-b-PLA polymer solutions before nanoprecipitation. (B) Normalized UV absorbance of 10%, 20%, 30%, and 50% CpG-functionalized NPs. (C) Hydrodynamic diameters of PEG-b-PLA NPs and CpG-NPs in PBS 1X. (D) Surface zeta potential of PEG-b-PLA NPs and CpG-NPs in PBS 1X.

Figure 3

In vitro activity of CpG-functionalized NPs. (A) Incubation of RAW-Blue macrophage cells and THP1 hTLR9 monocyte cells with either soluble CpG or different valencies of CpG-NPs (10%, 20%, 30%, 50%) induces the activation of NF-kB and AP-1. The magnitude of activation is quantified via colorimetric output using QUANTI-Blue solution. (B) Normalized activation curves across a range of CpG concentrations (3.1–29 μg/mL) delivered on CpG-NPs at different densities to 100000 RAW-Blue cells. The absorbance at 655 nm corresponds to TLR activation. (C) Log EC₅₀ values for each activation curve were extrapolated from (B) using a “log(TLR9 agonist) vs response” nonlinear regression curve fit of the dilution curves. (D) Activation curves across a range of CpG concentrations (3.1–29 μg/mL) delivered with different CpG formulations to 100000 THP1-dual hTLR9 cells. (E) Optical density of different CpG formulations at a CpG concentration of 29 μg/mL at 655 nm. (F) Confocal microscopy images of cellular uptake of RAW-Blue cells incubated with different CpG formulations equivalent to 5 μg of CpG. Cell nucleus was stained with DAPI, cell wall was stained with Alexa Fluor 488 Antialpha 1 Sodium Potassium ATPase antibody, and CpG was conjugated with Cy5. Scale bars are 10 μm. (G) Accumulation of Cy5-conjugated CpG in organs of interest 3 h after injection. Images and signal were determined by an in vivo imaging system. p values listed were determined using a 1-way ANOVA with Tukey’s multiple comparisons test. p values for comparisons between the 30% CpG-NPs group and all other groups are shown above the bars.

Figure 4

Fabrication and characterization of CpG-polymer-nanoparticle hydrogels. (A) Vaccine-loaded CpG-NP hydrogels are formed when aqueous solutions of PEG-b-PLA NPs and dodecyl-modified hydroxypropylmethylcellulose (HPMC-C₁₂) are mixed together with aqueous solutions of vaccine cargo comprising CpG-NPs (adjuvant) and spike protein (antigen). (B) Vaccine cargoes are added to the aqueous NPs solution before loading the aqueous and polymer components in two separate syringes (i); mixing the two phases with an elbow mixer (ii) leads to homogeneous hydrogels (iii). Image of a PNP hydrogel flowing through a 21-gauge needle during injection (iv) and formation of solid-like depot after injection (v). (C) Frequency-dependent oscillatory shear rheology and oscillatory amplitude sweeps (D) of CpG-NP and unloaded PNP hydrogels. (E) Stress-controlled flow sweeps of the CpG-NP hydrogel and yield stress value. (F) Shear-dependent viscosities of the two analyzed hydrogels demonstrate shear thinning and yielding properties, decreasing with increased shear rate. (G) Step-shear measurements over 3 cycles model yielding and healing of the hydrogels. Alternating low shear rates (0.1 1/s) and high shear rates (10.0 1/s, gray color) are imposed for 60 and 30 s, respectively.

Figure 5

Diffusivity of the cargo and gel components in the CpG-NP hydrogel. (A) FRAP microscopy images of the selected area to be photobleached (i) before bleaching, (ii) right after the bleaching process, and (iii) after complete fluorescence recovery. (B) Representative fluorescence recovery curve over time of the spike protein at a concentration of 0.27 mg/mL of hydrogel. Time points representing (A) are outlined on the curve. (C) Diffusivities of spike protein in PNP hydrogels (n = 8) measured via FRAP and diffusivity of spike in PBS 1X calculated using the Stokes–Einstein equation (eq 2). (D) PEG-b-PLA NPs and spike protein diffusivities in the hydrogel are measured via FRAP and are represented normalized by D_gel, the polymer matrix diffusivity. Values close to 1 represent diffusivities similar to that of the polymer matrix and support the assumption that NPs and spike antigen are caught in the hydrogel network. The dotted line shows D_cargo/D_gel = 1 (n = 4–8). (E) Representative schematic of the vaccine loaded PNP hydrogel, showing all the components diffuse slowly within the hydrogel network. All the results are given as mean ± sd.

Figure 6

In vivo kinetics of spike and CpG-NP. Mice were immunized with vaccines formulated with Alexa Fluor 790 labeled spike antigen and Cy5-CpG-NP in either PNP hydrogel or PBS 1X bolus formulation. (A) Representative images showing the different duration of release of spike protein given as a bolus or hydrogel subcutaneous immunization over 16 days. (B) Fluorescent signal from Alexa Fluor 790 labeled spike protein shown in (A). (C) Representative images demontrating the different duration of release of CpG-NP given as a bolus or gel subcutaneous immunization over 16 days. (D) Fluorescent signal from Alexa Fluor 790 labeled spike protein shown in (B). Release half-life of (E) spike and (F) CpG-NP in either bolus or PNP hydrogel. (G) The ratio of release half-lives for spike protein to CpG-NP in bolus or PNP hydrogel. Images and signal were determined by an in vivo imaging system, and results are shown as mean ± sd (n = 5). p values listed were determined using unpaired two-tailed t tests.

Figure 7

In vivo humoral response to COVID-19 subunit vaccine. (A) Timeline of mouse immunizations and blood collection for different assays. Soluble vaccine groups were immunized with a prime dose of 10 μg spike antigen and 20 μg CpG NPs or soluble CpG at day 0 and received a booster injection of the same treatment at day 21. CpG-NP hydrogel group was immunized with a single dose of 20 μg of spike antigen and 40 μg of CpG-NP adjuvant at day 0. Serum was collected over time to determine cytokine levels and IgG titers. IgG1, IgG2b, and IgG2c titers were quantified and neutralization assays were conducted on day 21 and day 35 serum. (B) Antispike total IgG ELISA end point titer of soluble vaccines before and after boosting (arrow) and single-immunization CpG-NP hydrogel. (C) Area under the curve (AUC) of antispike titers from (B). (D) Antispike IgG ELISA titers from serum collected on week 6, 3 weeks after boosting the soluble vaccine groups.Titers were determined for wild-type spike as well as Beta (B.1.351), Delta (B.1.617.2), and Omicron (B.1.1.529) variants of the spike protein. Each point represents an individual mouse (n = 5). Data are shown as mean ± sd. p values listed were determined using a 1-way or 2-way ANOVA with Tukey’s multiple comparisons test on the logged titer values for IgG titer comparisons (including total IgG and spike variants). p values for comparisons are shown above the data points.

Figure 8

Antibody subtype response to COVID-19 subunit vaccine. Antispike IgG1 (A) and IgG2c (B) titers from serum collected on week 5, 2 weeks after boosting the soluble vaccine groups. (C) The ratio of antispike IgG2c to IgG1 postboost titers. Lower values (below 1) suggest a Th2 response or humoral response, and higher values (above 1) suggest a Th1 response or cellular response. Each point represents an individual mouse (n = 5). Data are shown as mean ± sd. p values listed were determined using a 1-way or ANOVA with Tukey’s multiple comparisons test on the logged titer values for IgG titer comparisons. p values for comparisons are shown above the data points.

Figure 9

Single immunization of CpG-NP hydrogel elicits neutralizing antibodies in mice. (A) Preboost (Day 21) spike-pseudotyped viral neutralization assays for the CpG-adjuvanted COVID-19 spike vaccines at a serum dilution of 1:50. (B) Postboost of soluble vaccines (day 35) spike-pseudotyped viral neutralization assays for the CpG-adjuvanted COVID-19 spike vaccines at a serum dilution of 1:50. (C–E) Percent infectivity for all treatment groups at a range of Week 5 serum dilutions as determined by a SARS-CoV-2 spike-pseudotyped viral neutralization assay. (F) Comparison of IC₅₀ values determined from neutralization curves on day 35 for soluble vaccine formulations (prime-boosted) and hydrogel vaccine (single immunization) following immunization with CpG-adjuvanted COVID-19 spike vaccines. Each data point represents an individual mouse (n = 5). Data are shown as mean ± sd. p values listed were determined using a 1-way ANOVA with Tukey’s multiple comparisons. p values for comparisons are shown above the data points.