Introduction

In the context of vaccine immunity, the plasma cells derived from B cells produce neutralizing antibodies (nAbs) for protection against virus entry and infection. Concurrently, CD8+ T cells could be activated and differentiated into cellular toxicity lymphocytes (CTLs) to kill the virus-infected cell. Some CD4+ T cells are differentiated into T follicular helper (TFH) cells, which engage in critical interactions with B cells in germinal center, resulting in the production of high-affinity neutralizing antibodies1,2,3,4. There are multiple critical molecules produced from TFH cells to coordinate with B cells, such as CD40L, IL-21, IL-4, ICOS, etc5,6,7. CD40L, a costimulatory molecule expressed on T cells, interacts with CD40 to stimulate B cell activation and proliferation, triggering quiescent B cells into the cell proliferation cycle3,8. TFH cells could also secrete cytokines like IL-4 and IL-21 to incite B cell activity9,10. IL-21 serves as a critical factor in the differentiation of B cells into plasma cells, and also contributes significantly to the development and function of T follicular helper cells, crucial for the formation of functional germinal centers. Importantly, ICOS interacts with its ligand, ICOSL (Inducible T-cell Co-stimulator Ligand), which is expressed on various antigen-presenting cells such as B-cells, macrophages, and dendritic cells (DCs)11. The interaction between ICOS and ICOSL contributes to the enhancement of immune responses by promoting T cell activation, cytokine production, and the generation of effector T cells. Conversely, this interaction also enhances the BCR-dependent B cell responses12,13,14,15,16.

Vaccines often elicit lower immune responses in populations with compromised immunity or immunodeficiencies17,18,19,20. Elderly individuals experience immune senescence, leading to reduced functionality of the interaction between T cells and B cells21,22. Consequently, the weakened humoral immune responses are accompanied by declines in high-affinity antibodies. Furthermore, human immunodeficiency virus type-1 (HIV-1)-infected individuals suffer from immune deficiencies as the virus directly infects and destroys CD4+ T cells, severely hindering adaptive immune defense23,24. Although CD4+ T cell amount in most HIV-1-infected individuals could be well maintained during long-term combined antiretroviral therapy (cART), TFH cells are identified as HIV-1 latent reservoirs and expanded in the lymph nodes (LNs)25,26,27, which cannot effectively help B cell differentiation and antibody response24,28. The defective TFH is also associated with hypergammaglobulinemia in HIV-1 infected individuals29,30. It highlights that HIV-1 infected individuals or patients with low CD4+ T cell counts caused by other reasons exhibit weakened immune responses elicited by vaccines18,31. In addition, individuals with compromised immune function due to conditions like transplantation32,33, obesity34, hypertension35, diabetes36, among others, show varying degrees of reduced immune response to SARS-CoV-2 vaccines. Therefore, personalized vaccine development for immunocompromised individuals becomes quite necessary to enhance vaccine efficacy.

We have developed SARS-CoV-2 RBD nanoparticle vaccines37, which are self-assembled by twenty-four ferritin units and sequentially conjugated with immunogen through GV/SD linker system38, exhibit potent protection against the infections by SARS-Cov-2 such as alpha, Beta, Delta, or Omicron strains37,38,39,40,41. Compared to mRNA vaccines, the protein-based nanoparticle vaccines have a longer observation on the safety, more stability, and more flexibility in the design, suggesting it may serve as one optimal option for the application in the immunocompromised group42,43,44. Taking advantage of the capability to present multiple antigens or molecules simultaneously on one nanoparticle, we assembled the RBD nanoparticle together with various TFH-generating functional proteins, which may be essential for compensating the T-B cooperation under the compromised conditions. Herein, we unexpectedly found that ICOSL activation played an indispensable role in humoral immunity without TFH, suggesting a rational vaccine design strategy for immunocompromised individuals.

Results

ICOS enhanced the efficacy of SARS-CoV-2 RBD nanoparticles in TFH-deficient mice

Considering the epidemic and persistent mutation of SARS-CoV-2, which still threatens human health, especially those who are immunocompromised, the particular requirement remains unmet. To this end, by virtue of the advances of the Helicobacter pylori ferritin, which could present multiple molecules on the self-assembled nanoparticle surface, we chose three important factors generated from the TFH cells, e.g. CD40L, ICOS, or IL-21, to displayed them individually with SARS-CoV-2 RBD on one nanoparticle surface, to mimic the TFH function to enhance B cell responses (Fig. 1A). The ferritin was incubated with SARS-CoV-2 RBD and CD40L, ICOS, or IL-21 under the condition of equal molars, respectively, to form the CD40L-RBD, ICOS-RBD, or IL12-RBD nanoparticles through the covalent bond between GV and SD tags, followed by confirmation with gel electrophoresis and Transmission Electron Microscopy (TEM) (Fig. 1B, C).

Fig. 1: Design, construction, and evaluation of the modified RBD nanoparticle vaccines in TFH deficient mice.
figure 1

A A Schematic diagram of modified RBD nanoparticle vaccines. The CD40L, IL-21, ICOS were respectively incubated with Ferritin in the presence of RBD in an equal molar manner, to present the immune molecules with RBD on the same nanoparticle. B Coomassie blue staining showing the protein purification and nanoparticle assembly. * RBD-HPF monomer, ** CD40L-HPF monomer, *** IL-21-HPF monomer, **** ICOS-HPF monomer. C Transmission electron microscopy (TEM) images and two-dimensional (2D) reconstruction of each nanoparticle. Scale bars represented 200 nm. D Schematic of CD4creBcl6fl/fl mouse vaccination. All the mice were prime/boost vaccinated with 10 ug of nanoparticle vaccines as indicated at week 0 and 2. Serum was collected every two weeks. E, F ELISA showing SARS-CoV-2 RBD-specific IgG (B) and IgA (C) titers in the immunized CD4creBcl6fl/fl mice at week 2 post-boost. G Time-course curve showing the SARS-CoV-2 RBD-specific IgG titers at each collection point. H Determination of IgG avidity against SARS-CoV-2 RBD after a 7 M urea wash in sera from boost-immunized TFH deficient mice. The relative avidity was calculated by the ratio of OD values between in presence and absence of urea treatment. I pseudotyped virus neutralization test (pVNT) test determined the nAbs titers against SARS-CoV-2 pseudovirus in the sera from the CD4creBcl6fl/fl mice vaccinated with two doses of RBD, CD40L-RBD, IL-21-RBD, ICOS-RBD nanoparticles at day 14 post-boost. ALL data were presented means ± SEM. n = 6 from 3 independent experiments. Adjusted p values were calculated by one‐way ANOVA with Tukey’s multiple comparison test. ns indicated not significant.

To investigate the immune response under the immunocompromised conditions, the CD4creBcl6fl/fl mice were employed, which were not able to generate germinal center due to the absence of TFH cells and had impaired antibody production45. To compare the efficacy of the modified RBDs nanoparticle vaccines, we subcutaneously immunized the transgenic mice at weeks 0 and 2, and collected the sera every two weeks (Fig. 1D). Compared to the RBD nanoparticle vaccination, CD40L-RBD, ICOS-RBD, and IL-21-RBD nanoparticles elicited 9.4-fold, 63-fold, and 13-fold more RBD-specific IgG in the CD4creBcl6fl/fl mice at day 14 post-boost, respectively (Fig. 1E). ICOS-RBD nanoparticle vaccination generated 18-fold or 9-fold more RBD-specific IgA titers in the CD4creBcl6fl/fl mice compared to the CD40L-RBD or IL-21-RBD nanoparticle groups, which was essential for mucosal immunity to prevent various infections (Fig. 1F). In addition, the ICOS-RBD nanoparticle also exerted persistently highest RBD-specific IgG antibodies in the CD4creBcl6fl/fl mice in a long-term observation compared to the CD40L-RBD or IL-21-RBD nanoparticle groups (Fig. 1G).

To further evaluate the efficacy of the modified RBD nanoparticle vaccines, a well-established urea-based assay was applied to analyze the antibody avidity, which would elute the low-avidity or low-affinity antibodies after antigen-antibody binding by the 7 M urea. The ICOS-RBD nanoparticle vaccination generated 46% more high-avidity antibodies than the RBD nanoparticle group in the TFH-deficient mice (Fig. 1H). Moreover, we conducted a 50% of Neutralizing antibody Test against Pseudotyped SARS-CoV-2 Virus (pVNT50) to compare the capability of the nAbs to block the virus entry and infection. Consistently, ICOS-RBD nanoparticle elicited 38-fold or 32-fold higher nAb titers against SARS-CoV-2 viruses than the CD40L-RBD or IL-21-RBD nanoparticle groups (Fig. 1I), indicating that ICOS significantly enhanced the efficacy of SARS-CoV-2 RBD nanoparticle vaccine under the immunodeficient situation.

ICOS-conjugated nanoparticle-induced formation of antibody-secreting B cells (ASCs) in TFH-deficient mice

To investigate how ICOS-RBD nanoparticle vaccination enhanced RBD-specific antibody production without TFH help, we compared the immune responses of various immune cells after the subcutaneous immunization with ICOS-RBD nanoparticle in the CD4creBcl6fl/fl mice. The result showed that ICOS-RBD nanoparticle vaccination did not significantly alter the percentages or numbers of the total T cells in the lymph nodes compared to the RBD nanoparticle group at day 14 post the single dose of vaccination (Fig. 2A, B). However, compared to the RBD nanoparticle group, ICOS-RBD nanoparticle immunization increased 37% higher percentages and 48% more amounts in B cells in the lymph nodes (Fig. 2A, C), indicating that the ICOS may specifically enhanced B cell response. Due to the lack of TFH cells in the CD4creBcl6fl/fl mice45 (Supplementary Fig. 1A), there was almost no GL7+Fas+ germinal center B (GC B) cells in the CD19+ subset in the LNs in either ICOS-RBD or RBD nanoparticle groups (Fig. 2D, Supplementary Fig. 1B)12,46. Therefore, the ICOS-RBD nanoparticle enhanced antibody immunity without altering GC B cells. In addition to promoting the response of total B cells (Fig. 2C), ICOS-conjugation also increased the percentage of plasmablasts in the lymph nodes by 43% and the total number by 85%. (Fig. 2E), which secreted antibodies and correlated with long-lived plasma cells in bone marrow for durable antibody protection. Furthermore, we found ICOS-RBD nanoparticle vaccine also induced 82% more amounts of B cells in the spleen (Fig. 2F), as well as 1.4-fold more memory B cells (MBC) (Fig. 2G, Supplementary Figure 2). In the absence of TFH cells, there were almost no ASCs after the RBD nanoparticle immunization. However, ICOS-RBD nanoparticle vaccine significantly increased RBD-specific ASCs (Fig. 2H). To investigate the efficacy of ICOS nanoparticle on the antibody immunity in the presence of TFH cells, the wild-type C57BL/6 mice were treated with ICOS-RBD or RBD nanoparticles in a prime/boost manner. Two weeks post dose 2, the ICOS-RBD nanoparticle vaccination generated 37% more high-avidity antibodies than the RBD nanoparticle group by using urea-base ELISA in C57BL/6 mice (Fig. 2I). Consistently, the neutralizing antibodies in the ICOS-RBD nanoparticle-treated mice 11-fold more than that in RBD nanoparticle group (Fig. 2J). The SARS-CoV-2-specific antibody-secreting cells in the bone marrow from ICOS-RBD nanoparticle-treated mice were 70% more than that in RBD nanoparticle group (Fig. 2K). Taken together, all these data supported that ICOS potently enhanced humoral response in the TFH-deficient mice, confirming the essential role of ICOS/ICOSL signaling in B cell immune response.

Fig. 2: Analysis of the immune responses elicited by the ICOS-RBD nanoparticle vaccination in TFH deficient mice.
figure 2

A The representative flow gating of total T cells and B cells in the draining lymph nodes. The percentages and cell numbers of the CD3+ T cells (B) and B220 + B cells (C) in the lymph nodes (LN). Flow analysis showing the percentages and cell numbers of GL7+ Fas+ GC B cells (D) and B220+CD138+ plasmablasts (E) in the LN from the CD4creBcl6fl/fl mice with one dose of RBD or ICOS-RBD nanoparticle vaccination at day 14. Flow analysis showing the percentages and cell numbers of total B220+ B cells (F) and IgD-CD38+ memory B cells (G) in the spleen from the CD4creBcl6fl/fl mice with two doses of RBD or ICOS-RBD nanoparticle vaccination at day 14 post-dose 2. H ELISpot assays showing the RBD-specific ASCs number per 1 million splenocytes from the CD4creBcl6fl/fl mice with two doses of RBD or ICOS-RBD nanoparticle vaccination at day 14 post-dose 2. I Determination of IgG avidity against SARS-CoV-2 RBD after a 7 M urea wash in sera from boost-immunized C57BL/6 mice. The relative avidity was calculated by the ratio of OD values between in presence and absence of urea treatment. J ELISA showing SARS-CoV-2 RBD-specific IgG titers in the immunized C57BL/6 mice at week 2 post-boost. K ELISpot assays showing the RBD-specific ASCs number per 1 million splenocytes from the C57BL/6 mice with two doses of RBD or ICOS-RBD nanoparticle vaccination at day 14 post-dose 2. All data were presented means ± SEM. n = 5 or 6, as indicated by the dots in the figures, from 3 independent experiments. The p values were calculated by two-tailed Student t-test. ns means not significant.

Co-conjugation of ICOS and RBD nanoparticles preferentially expanded antigen-specific B cells

The treatment with ICOS-conjugated nanoparticle increased the percentage and cell number of B cells (Fig. 2F), indicating that the binding of ICOS to ICOSL could stimulate B cell activation and proliferation. Therefore, we hypothesized that the co-conjugation of ICOS and antigen on same nanoparticle would induce antigen-specific B cell activation. To this end, we first confirmed the binding of purified ICOS protein to the ICOSL on the B cells by flow cytometry (Fig. 3A). Then we used an anti-ICOSL blockade antibody47,48, to pretreat the CD4CreBcl6fl/fl mice for twenty-four hours, followed by immunization with RBD-ICOS nanoparticle. Fourteen days later, we found that ICOSL blockade significantly suppressed the antibody production elicited by ICOS-RBD nanoparticles compared to the non-blockade group (Fig. 3B). Moreover, ICOSL blockade also reduced percentages of plasmablasts induced by ICOS-RBD nanoparticle (Fig. 3C), suggesting that ICOS-RBD nanoparticle improved the immune response through ICOSL on B cells in the absence of TFH cells. To further investigate the influence of ICOS co-conjugation nanoparticle on the B cell proliferation, we immunized the CD4CreBcl6fl/fl mice with RBD nanoparticle, or with ICOS-RBD nanoparticle, or with RBD nanoparticle plus ICOS monomer. ICOS-RBD nanoparticle elicited 17-fold more RBD-specific than the RBD nanoparticle plus ICOS monomer group (Fig. 3D). However, ICOS-RBD nanoparticle induced 35% less B cells proliferation compared to the RBD nanoparticle plus ICOS monomer group at day 14 post vaccination (Fig. 3E), indicating a limited degree of antigen-dependent B cell expansion by ICOS-RBD. In the 7 M urea-based ELISA, ICOS-RBD nanoparticle vaccination generated 71% more high-avidity RBD-specific antibodies components than the RBD nanoparticle plus ICOS monomer group (Fig. 3F). Taken together, co-conjugation of antigen and ICOS on the same nanoparticle highlights the advantages of controllable and localized responses in nanoparticle vaccine applications.

Fig. 3: Co-conjugation of ICOS and RBD nanoparticles preferentially expanded antigen-specific B cells in the absence of TFH cell.
figure 3

A Flow cytometry showing the staining of the mouse B cells with the BV421-labeled ICOS protein. B ELISA showing SARS-CoV-2 RBD-specific IgG titers in the immunized CD4creBcl6fl/fl mice treated with anti-ICOS blockade antibody or not at week 2 post vaccination. n = 5. C Flow analysis showing the percentages of B220+CD138+ plasmablasts in the LN from the CD4creBcl6fl/fl mice with one dose of RBD or ICOS-RBD nanoparticle vaccination at day 14. n = 5. D ELISA showing SARS-CoV-2 RBD-specific IgG titers in the CD4creBcl6fl/fl mice immunized with RBD nanoparticle, or with ICOS-RBD nanoparticle, or with RBD nanoparticle plus ICOS monomer at week 2 post injection. n = 4. E Flow analysis showing the percentages of total B220+ B cells in the spleen from the CD4creBcl6fl/fl mice immunized with RBD nanoparticle, or with ICOS-RBD nanoparticle, or with RBD nanoparticle plus ICOS monomer at week 2 post injection. n = 4. F Determination of IgG avidity against SARS-CoV-2 RBD after a 7 M urea wash in sera from boost-immunized CD4creBcl6fl/fl mice. The relative avidity was calculated by the ratio of OD values between in presence and absence of urea treatment. All data were presented means ± SEM. Adjusted p values were calculated by one‐way ANOVA with Tukey’s multiple comparison test. ns indicated not significant.

ICOS nanoparticles promoted B cell survival and proliferation through ICOSL/PKCβ/AKT/NF-κB pathway

ICOSL is expressed on antigen-presenting cells, such as dendritic cells, macrophages, and B cells11. The interaction of ICOS and ICOSL is necessary for the activation of TFH cells through the stimulation of ICOS on T cells. However, the mechanism and function of ICOSL in B cells remain to be investigated. Considering the observation that ICOS nanoparticles significantly increase B cell amount in both lymph nodes and spleen after the vaccination, we treated the naïve B cell isolated from the spleen of the CD4creBcl6fl/fl mouse with ICOS, CD40L, IL-21 or ferritin nanoparticles in vitro. Carboxyfluorescein succinimidyl ester (CFSE) analysis showed that ICOS nanoparticles induced the most cell division and proliferation of B cells than the other groups (Fig. 4A). Consistently, ICOS nanoparticles promoted the transition from G1 phase to S phase in B cells, resulting in 40% more B cells in the S phase, or six times the percentages of S phase than that in the CD40L or IL-21 nanoparticle groups (Fig. 4B). Moreover, ICOS nanoparticles also maintained 50% higher cell viability during in vitro culture compared to ferritin nanoparticle treatment (Fig. 4C). To further confirm the direct regulation of ICOS on B cells through ICOSL, we treated the primary mouse B cells with anti-ICOSL antibody for 8 hours in vitro, followed by the stimulation of ICOS nanoparticle. The results showed that anti-ICOSL blockade could significantly inhibit the ICOS nanoparticle-induced proliferation of B cells (Fig. 4D). All the results demonstrated that ICOS nanoparticles could promote the survival and proliferation of primary B cells. Loss-of-function ICOSL in DCs was reported to be associated with Crohn’s disease phenotype, which impaired pattern-recognition receptor-induced cytokine secretion due to insufficient activation of pathways, including the Protein Kinase C (PKC), AKT (Protein Kinase B), Nuclear Factor Kappa B (NF-κB), and MAPK signaling pathways16. To investigate the mechanism how the interaction of ICOS and ICOSL promoted survival and proliferation in B cells, we treated the mouse primary B cells with ferritin or ICOS nanoparticles, followed by the analysis of both protein and mRNA through western blot and qRT-PCR. The results showed that, compared to the ferritin group, the treatment with ICOS nanoparticle significantly increased the phosphorylation levels of PKCβ and AKT in the B cells (Fig. 4E). ICOS nanoparticle induced more total NF-κB and phosphorylated NF-κB in B cells than ferritin did (Fig. 4E). Consistently, the genes related to cell cycle, such as Ccnd1, Ccnd2, Ccnd3, and Cdk4, were upregulated in the B cells treated with ICOS nanoparticle, compared to ferritin (Fig. 4F). In addition, the treatment of ICOS nanoparticle significantly upregulated Mcl1 expression (Fig. 4F), a pro-survival and anti-apoptosis protein. The D-type cyclins can promote cell cycle progression from G1 to S phase by binding to Cdk4 and Cdk6. The western blot results also showed that both Cyclin D2 and D3 were significantly upregulated in the mouse primary B cells treated by the ICOS nanoparticle than that in ferritin group (Fig. 4G). Taken together, ICOS nanoparticle treatment promoted B cell survival and proliferation through activating ICOSL/PKCβ/AKT/NF-κB pathways.

Fig. 4: ICOS nanoparticle promoted B cell proliferation via ICOSL/PKCβ/AKT/NF-Κb.
figure 4

A Flow cytometry showing the CFSE staining of the mouse primary B cells treated with 50ug/mL ferritin, ICOS, CD40L, or IL-21 nanoparticles for 40 hours. Median Fluorescence Intensity (MFI), proliferation index (the average number of divisions that exclude the undivided cells) and frequency of divided cells were analyzed by Flowjo. n = 4 from 3 independent experiments. Adjusted p values were calculated by ANOVA with Tukey’s multiple comparison test. Flow cytometry showing the cell cycle (B) staining of the mouse primary B cells treated as Fig. 4A. The data were presented means ± SEM. n = 3 from 3 independent experiments. Adjusted p values were calculated by ANOVA with Tukey’s multiple comparison test. C Flow cytometry showing the apoptosis of the mouse primary B cells treated with 50ug/mL ferritin, ICOS nanoparticles, or PBS for 24 hours. The cells were stained with anti-Annexin V-PB and PI. Live cells were negative for both Annexin V-PB and PI. The data were presented means ± SEM. n = 5 from 3 independent experiments. Adjusted p values were calculated by one‐way ANOVA with Tukey’s multiple comparison test. D Cell counting showing the growing of B cells isolated from mouse spleen, with or without anti-ICOSL blockade for eight hours, followed by the treatment of 50ug/mL ferritin, ICOS nanoparticles for 72 hours. The data were presented means ± SEM. n = 5 from 3 independent experiments. Adjusted p values were calculated by two‐way ANOVA with Tukey’s multiple comparison test. E Immunoblotting showing the expression of p-PKCβ, PKCβ, p-AKT, AKT, p-NF-κB, NF-κB, and Actin in mouse primary B cells treated with 50 ug/mL ferritin or 50 ug/mL ICOS nanoparticles for 36 hours. Representative images of three independent experiments. F qRT-PCR assay showing relative expression of the cell cycle-related genes in mouse primary B cells treated with ferritin or ICOS nanoparticles for 48 hours. Hprt used for reference. The data were presented means ± SEM. n = 3 from 3 independent experiments. P values were calculated by two‐tailed Student t-test. ns indicated not significant. G Immunoblotting showing the expression of Cyclin D2, Cyclin D3, Actin in mouse primary B cells treated with 50ug/mL ferritin or ICOS nanoparticles for 36 hours. Representative images of three independent experiments.

ICOS-RBD nanoparticle vaccine potently suppressed SARS-CoV-2 infection in immunodeficient mice

Other groups and us reported that some strains of SARS-CoV-2, such as the Beta, Delta or Omicron sublineages, could directly infect the wildtype mice due to the improved affinity to mouse Ace2 from N501Y mutation on Spike49,50. To evaluate the protective efficacy of ICOS-harboring nanoparticles against infections in immunodeficient mice, HPF, RBD, or ICOS-RBD nanoparticles were subcutaneously injected into the CD4creBcl6fl/fl mice with two dose of 10 ug at an interval of 14 days, followed by infection with SARS-CoV-2 Omicron EG.5 strains at day 14 post-dose 2 and pathology analysis at day 5 post-viral challenge (Fig. 5A). The ICOS-RBD nanoparticle induced the highest RBD-specific antibody titers in TFH-deficient mice at day 14 post-boost than the RBD or HPF nanoparticle groups (Fig. 5B). While the viral RNAs in the lungs in the HPF nanoparticle group had an average of 9.02 ×106 copies per µg total RNA after Omicron EG.5 challenge, the ICOS-RBD nanoparticle vaccination could reduce viral replication by a factor of approximate 1660 or 10, compared to HPF or RBD nanoparticle vaccination group (Fig. 5C). Moreover, a significant reduction of inflammation in the lung, detected by H&E staining, in ICOS-RBD nanoparticle-immunized mice was also observed (Fig. 5D). The expression of SARS-CoV-2 nucleocapsid (N) antigen was consistently the lowest in ICOS-RBD nanoparticle vaccination group compared to control HPF or RBD nanoparticle groups (Fig. 5E). These results further demonstrated that the ICOS-RBD nanoparticle vaccine induced robust immune responses against SARS-CoV-2 and significantly improved protection against the infection of authentic Omicron EG.5 strain in the immunocompromised mice.

Fig. 5: Co-conjugation of ICOS and RBD nanoparticles enhanced the protection against SARS-CoV-2 EG.5 infection in TFH-deficiency mice.
figure 5

A Immunization schedule. CD4creBcl6fl/fl mice were immunized with ICOS-RBD, RBD nanoparticle vaccine at days 0, 14. The vaccinated mice were challenged with authentic Omicron EG.5 variant at day 28 and euthanized at day 5 post‐challenge for analysis. B ELISA showing SARS-CoV-2 RBD-specific IgG titers in the immunized CD4creBcl6fl/fl mice at week 4. n = 4–5 for each group. Adjusted p values were calculated by one‐way ANOVA with Tukey’s multiple comparison test. C qRT‐PCR analysis showing viral RNA copies per ug total RNA from the lungs in the EG.5-challenged CD4creBcl6fl/fl mouse at day 5 post-infection. Adjusted p values were calculated by one‐way ANOVA with Tukey’s multiple comparison test. H&E staining (D) and immunohistochemistry (IHC) against SARS-CoV-2 N proteins (E) showing the pathology and viral replication in the lungs of mice. Scale bars represented 200 µm.

ICOS enhanced the efficacy of HIV-1 nanoparticle vaccine independent of TFH cell

In addition to aging-induced immunocompromise or primary immunodeficiency, viral infection can also cause severe immunodeficiencies, such as HIV-1 infection. HIV primarily infects CD4+ T cells, resulted in impaired immunity and reduced TFH cells24,28. Therefore, it is important to consider the inadequate support of TFH cells in HIV-1 therapeutic vaccine development. Germline-targeted design was one of the promising strategies in the development of HIV vaccines to induce broad neutralizing antibodies. Here, we chose MD39 as HIV-1 immunogen, a well-designed epitope targeting N332-supersite in the V3 loop of HIV-1 gp120 envelop (Fig. 6A), followed by expression and purification from HEK293F cells51. The MD39 was then conjugated onto the ferritin nanoparticle together with ICOS, and the ICOS-MD39 nanoparticle was confirmed by both gel electrophoresis and western blot analysis (Fig. 6B, C).

Fig. 6: ICOS potently enhanced the efficacy of HIV-1 MD39 nanoparticle vaccines independent of TFH cell.
figure 6

A A Schematic diagram of ICOS-modified HIV MD39 nanoparticle vaccines. Confirmation of protein purification and nanoparticle assembly by Coomassie blue staining (B) and western blot (C). * MD39 and glycosylated MD39, ** MD39-HPF monomer, *** ICOS-HPF monomer. Anti-gp120 antibody was applied. D Schematic of CD4creBcl6fl/fl mouse vaccination. All the mice were prime/boost vaccinated with nanoparticle vaccines at week 0 and 2. Serum was collected every two weeks. E ELISA showing HIV-1 Envelope-specific IgG titers in the CD4creBcl6fl/fl mice with two doses of MD39 or ICOS-MD39 nanoparticle vaccination at day 14 post-dose 2. n = 6 from 3 independent experiments. The p values were calculated by two-tailed Student t-test. F Time-course curve showing the HIV-1 gp120-specific IgG titers at each collection point. n = 6 from 3 independent experiments. G Determination of IgG avidity against SARS-CoV-2 RBD after a 7 M urea wash in sera from the CD4creBcl6fl/fl deficient mice with two doses of vaccination at day 14 post-dose 2. The relative avidity was calculated by the ratio of OD values between in presence and absence of urea treatment. n = 6 from 3 independent experiments. The p values were calculated by two-tailed Student t-test. H ELISpot assays showing the HIV Envelope-specific ASCs number per 1 million splenocytes from the CD4creBcl6fl/fl mice with two doses of MD39 or ICOS-MD39 nanoparticle vaccination at day 14 post-dose 2. n = 6 from 3 independent experiments. The p values were calculated by two-tailed Student t-test.

MD39- or ICOS-MD39 nanoparticles were subcutaneously injected into the CD4creBcl6fl/fl mice on week 0 and 2, and the blood was collected every two weeks (Fig. 6D). The HIV-1 envelop-specific IgG titer was 13-fold higher in ICOS-MD39 nanoparticle vaccination groups two weeks post-boost than that in MD39 nanoparticle group (Fig. 6E). Moreover, the ICOS group also exerted significant HIV-1 envelop-specific IgG antibodies over a long period (Fig. 6F). Interestingly, compared to MD39 nanoparticle, ICOS-MD39 nanoparticle vaccination produced 38% more high-avidity antibodies in the TFH-deficient mice in the urea-based assay (Fig. 6G). Furthermore, ELISpot assay showed more HIV-1 envelop-specific ASCs in the spleens from the CD4creBcl6fl/fl mice vaccinated with ICOS-MD39 nanoparticle at day 14 post-boost, compared to that in MD39 nanoparticle vaccination group (Fig. 6H). Taken together, all these data demonstrated that ICOS also potently enhanced humoral response of HIV-1 nanoparticle vaccine in the immunodeficient mouse model.

Discussion

Vaccine immunization is an ideal and practical approach to combat infectious diseases and has efficiently controlled some of the most threatening diseases, such as smallpox, measles, poliovirus52. However, the immunocompromise and immunodeficiency significantly impaired the vaccines efficacy. Herein, our findings showed that the stimulation of ICOS-ICOSL interaction by adding ICOS on the nanoparticle vaccine significantly substitute the function of TFH cells to support B cell response, which promoted the proliferation and survival of B cells through PKCβ and NF-κB pathway, leading to potent humoral responses. This strategy could be further incorporated into the development of various preventive or therapeutic vaccines for these immunocompromised individuals with defective function in TFH, which include the elder or HIV-1-infected individuals.

It is an interesting strategy to enhance immunogenicity by introducing immune regulatory molecules into the vaccine formulation. It was reported that CD40L could promote antibody production by linking with the immunogen to form a fusion protein53,54. In our screening system, it was the first time to show that ICOS empowered vaccine with potent immunogenicity in a nanoparticle manner, which imitated the T-B interaction to enhance antibody immunity independent on TFH cells. ICOS performed better enhancement on the vaccine efficacy compared to CD40L in nanoparticle system, possibly because CD40L required additional co-stimulation of interleukins or cytokines55. Nanoparticle vaccines have exhibited more potent capabilities in inducing antibody production and low side effects, compared to mRNA vaccines42. Therefore, the ferritin nanoparticle provided a useful platform to improve the efficacy by conjugating variants of molecules, such as various immunogens to assemble mosaic vaccines or immune regulators.

ICOSL expression on activated B cells plays a critical costimulatory role for T cell activation, however, the effect of ICOSL itself upon B cells has not well been studied. We showed that ICOS nanoparticles could also serve as an adjuvant to stimulate proliferation and reduce apoptosis of the B cells via ICOSL, which exhibited a similar phenomenon that LPS activated B cell antigen-independently through toll-like receptor.

It is reported that the activation of ICOSL on DC by ICOS could enhance pattern-recognition receptor-induced signaling and cytokines secretion16. It is possible that ICOS nanoparticle stimulates DCs for enhanced T cell activation, such as T helper cell polarization, resulting in an improved effectiveness of vaccines via cellular immunity in the absence of TFH cells, which merits being further explored that the mechanisms of ICOS/ICOSL in the APC in future. It is also reported that ICOS could help the fused antigen to be held near the cell membrane by ICOSL, and facilitate the endocytosis and antigen presentation56. However, the nanoparticle itself has the ability to enhance the antigen presentation, where we observed the enhanced cell proliferation and ASCs formation elicited by ICOS in our system.

In the primary or acquired T cell deficiency, or in the age-related immunocompromise individuals, vaccines were not able to elicit potent antibodies as in the healthy cohorts. mRNA vaccine-elicited neutralizing antibodies against SARS-CoV-2 were observed in the absence of TFH cells57, which indicated an alternative pathway for the induction of B cell responses to complement traditional GC-dependence. It is possible that elevated ICOSL expression on B cells under infection- or vaccination-mediated inflammation situations compensated for GCs compromise, as ICOSL expression was reported to be regulated by NF-κB signaling58. ICOSL deficiency indeed damaged the antibody production and ICOSL downregulation was also associated with poor clinical outcomes for cancer59. ICOS-ICOSL plays an important role in immunoregulatory, which could be a target for adjuvant design.

In this study, we first reported that ICOS could help SARS-CoV-2 RBD vaccines, when were assembled on the same nanoparticle, induce superior Ag-specific antibody-secreting cells proliferation and antibody formation. Importantly, these enhanced immune responses were observed in TFH deficient mouse model, which characterized by the increased percentages and number of memory B cells, plasmablasts, and ASCs. ICOS not only significantly improved the efficacy of RBD nanoparticle against Omicron EG.5 virus infection, but also facilitated HIV-1 MD39, germline-targeted immunogen, to induce potent antibodies in immunodeficient mice. These results suggested that a TFH-independent design of ICOS-adjuvanted nanoparticle has been successfully established. Our work expands the capability of self-assembled nanoparticle vaccines in the application in the immunocompromise or immunodeficiency, which elicits potent immune responses with significant high-avidity antibody production.

Methods

Cells and viruses

HEK293T and Vero E6 cells were obtained from ATCC. These adherent cells were cultured in DMEM supplemented with 1% penicillin-streptomycin (ThermoFisher) and 10% FBS (ThermoFisher). HEK293T expressing hACE2 (hACE2/HEK293T) was constructed. Expi293F suspension cells were stored in Union 293F (Union) containing 8 mM glutamine (ThermoFisher), and 1% penicillin-streptomycin (ThermoFisher) and placed on a polycarbonate vented conical flask shaker at 37 °C, 8% CO2, and a shaker at 120 rpm. All cells were regularly tested for mycoplasma DNA using PCR, and it confirmed that they were mycoplasma-free. The authentic SARS-CoV-2 virus, isolated and authorized by Guangdong Provincial Center for Disease Control and Prevention, was carried out in the BSL-3 facility at Sun Yat-sen University.

Animal models

The Bcl6fl/fl and CD4Cre mice were purchased from Jackson’s laboratory and genotypes were confirmed by PCR. Bcl6fl/fl mice were crossed with Cd4Cre mice to generate CD4creBcl6fl/fl mice. 6 to 8-week-old C57BL/6J mice of Specific Pathogen Free (SPF) were purchased from Guangdong Medical Laboratory Animal Center. The numbers of mice used for the C57BL/6 and CD4CreBcl6fl/fl strains were indicated in the figure legends. All mice were raised and vaccinated in the SPF facilities at the Laboratory Animal Center of Sun Yat-sen University. For animal welfare ethics, mice were euthanized by isoflurane inhalation followed by cervical dislocation according to predefined endpoints, i.e., at the end of the experimentation (the last timepoint for analysis post vaccination indicated in the figures), or when animals exhibited signs of cachexia, prolonged behavioral abnormalities, or physical impairment.

Protein Expression and Purification

The construction of the RBD nanoparticle vaccine was described in the manuscript. To further improve the binding efficiency of the nanoparticle vaccine, the Gv/Sd system was used to enhance the efficiency of the SARS-CoV-2 nanoparticle vaccine. Gv-HPF could be self-assembled to be a 24-mer nanoparticle and conjugated with RBD-Sd robustly. Gv-HPF was expressed and purified from prokaryotic expression of BL21 bacteria (Vazyme) induced by Isopropylthio-β-galactoside (IPTG). The bacterial cultures were harvested and lysed by sonication in Tris buffer (20 mM Tris, 50 mM NaCl, pH 7.5). The lysate supernatants were heated to 65 °C for 15 min to precipitate most of the Escherichia coli proteins. Sequences of mCD40L, mIL-21, mICOS were obtained by overlap PCR, and then connected with Sd-tag, to be inserted into pET28a vector using NocI and XhoI sitess. Similarly, the constructs were transformed into Rosetta (Solarbio) for expression and purification. The sequences of Delta-RBD, Beta-RBD, JN.1-RBD, and HIV-MD39 were cloned into the pcDNA3.1 vector, followed by expression and purification from Expi293F cells. The RBD or HIV MD39 were genetically fused with Sd-Tag at the N-terminus, which could be conjugated with Gv-HPF in the manner of an isopeptide bond. Seven days after transfection, the supernatants were collected and the cellular debris were discarded by centrifugation. The supernatants were passed through Ni-NTA agarose four times to enrich His-tagged target proteins and then eluted with Tris buffer containing imidazole. After centrifugation and concentration, the supernatants were loaded onto a Superose 6 Increase (GE Healthcare) size-exclusion column and eluted with the Tris buffer at a rate of 0.5 mL min−1. The concentration of Gvo-protein was determined by the BCA assay method. The bacterial endotoxins in nanoparticles were quantified by the Tachypleus amebocyte lysate test (≤ 10 EU per dose). All purified proteins were confirmed by gel electrophoresis of SDS-PAGE with Coomassie blue staining.

Animal vaccination

The CD4creBcl6fl/fl mice were subcutaneously immunized at days 0 and 14 with two doses of Alum-adjuvanted RBD, CD40L-RBD, IL-21-RBD, or ICOS-RBD nanoparticles, in equimolar of 6 ug equivalent of RBD nanoparticle, respectively. Serum was collected at days 0, 14, 28, 42, and 64. A verified anti-ICOSL antibody, clone HK5.3 (BioLegend 107412), was used for in vivo blockade. 500 ug of HK5.3 antibody was intravenously injected into each mouse to block ICOSL. Twenty-four hours later, ICOS-RBD or RBD nanoparticles were used to immunized the HK5.3 antibody-treated mice to determine the influence of ICOS/ICOSL on the vaccine efficacy. For HIV-1 vaccination experiments, the CD4creBcl6fl/fl mice were subcutaneously immunized at days 0 and 14 with two doses of Alum-adjuvanted MD39 or ICOS-MD39 nanoparticles, in equimolar of 6 ug equivalent of MD39 nanoparticle, respectively.

SARS-CoV-2 infection

The CD4creBcl6fl/fl mice, immunized with indicated vaccines, were challenged with authentic SARS-CoV-2 Omicron EG.5 strain in the BSL-3 facility. Mice were anesthetized with isoflurane and inoculated intranasally with 5 × 104 FFU of SARS-CoV-2 viruses. The lungs were collected at day 5 post-infection (d.p.i.).

Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR)

The lungs of challenged CD4creBcl6fl/fl mice were collected and homogenized with gentleMACS M tubes (Miltenyi Biotec, 130-093-236) in a gentle-MACS dissociator (Miltenyi Biotec, 130-093-235). RNAs were extracted using RNeasy Mini Kit (QIAGEN, 4104) according to the manufacturer’s instruction, followed by the qRT-PCR to determine the viral RNA copies of lung tissues utilizing a one-step SARS-CoV-2 RNA detection kit (PCR-Fluorescence Probing) (Da An Gene Co., DA0931). The SARS-CoV-2 nucleocapsid (N) gene was cloned into a pcDNA3.1 expression plasmid for standards to generate a standard curve. The indicated copies of N standards were ten-fold serially diluted from 109 to 102 and proceeded to qRT-PCR utilizing the same one-step SARS-CoV-2 RNA detection kit to obtain standard curves. The reactions were carried out on a BioRad CFX according to the manufacturer’s instructions. The viral RNA copies of each tissue were calculated into copies per µg total RNA and presented as a log10 scale.

Histopathology and immunohistochemistry

The CD4creBcl6fl/fl mice challenged with SARS-CoV-2 were euthanized in the BSL-3 facility. Lungs were collected and fixed in 4% paraformaldehyde buffer for 48 h at 4 ˚C, followed by embedding with paraffin. Longitudinal sections were performed on these tissues. The sections (3–4 µm) were stained with hematoxylin and eosin (H&E). For immunohistochemistry, lung sections of each mouse were incubated with rabbit anti-SARS-CoV-2 Nucleoprotein (N) at 1:1000 dilution, and the IHC was conducted as published before.

Enzyme-Linked Immunosorbent Assay (ELISA)

The recombinant RBD-Sd and MD39-Sd proteins were diluted to 5 µg mL−1 with coating buffer and coated on high-binding 96-well plates (Corning) overnight at 4 °C, respectively. After washing with PBS three times, the plates were blocked with 5% non-fat milk/PBS for 1 h at room temperature. The plates were washed three times with PBST (containing 1% Tween-20) again, and the animal serum was diluted serially in PBS, followed by incubating the plates for 1 h at 37 °C. After washing with PBST, HRP-conjugated goat anti-mouse IgG, goat anti-mouse IgG1, and goat anti-mouse IgG2c were added at a dilution of 1:4000, 1:3000, and 1:3000 to detect antigen-specific each isotype antibody in serum of mice. After incubating for another 1 h, the plates were washed with PBS/T. Subsequently, 50 µL HRP substrate TMB solution (eBioscience) per well was added under dark, and the reaction was terminated with a stop solution (Solarbio) after sufficient development. The absorption was measured at 450 nm. GraphPad Prism 9.0 software was used to perform non-linear regression analysis on the data to calculate the endpoint titer.

7M Urea Avidity ELISA

Coat ELISA plates overnight at 4°C with 5 μg/ml antigen protein in 0.1 M bicarbonate buffer (pH 9.3). After washing with PBS three times, the plates were blocked with 5% non-fat milk/PBS for 1 h at room temperature. The plates were washed three times with PBST (containing 1% Tween-20) again, and the animal serum was diluted serially in PBS, followed by incubating the plates for 1 h at 37 °C. After incubation with diluted serum, wash plates twice with PBS/T. Incubate the wells with PBST or PBST/7M Urea for 10 minutes at room temperature. After washing with PBS/T, HRP-conjugated goat anti-mouse IgG was added at a dilution of 1:4000 to detect antigen-specific antibody in the serum of mice. After incubating for another 1 h, the plates were washed with PBST. Subsequently, 50 µL HRP substrate TMB solution (eBioscience) per well was added under dark, and the reaction was terminated with stop solution (Solarbio) after sufficient development. The absorption was measured at 450 nm. GraphPad Prism 9.0 software was used to perform non-linear regression analysis on the data to calculate the endpoint titer. Calculate the avidity index for each serum sample by dividing readings from 7 M Urea treatment by readings from PBST-only treatment.

Pseudotyped virus neutralization assay

The generation protocol was described as published before.[4] In brief, the plasmid expressing the respective mutant pseudotyped virus spike protein of Delta, was constructed. HEK293T cells were co-transfected with the psPAX2 (Addgene) plasmid, the lentiviral plasmid expressing luciferase (Addgene), and the plasmid expressing the respective mutant spike by using polyethyleneimine (PEI, Sigma). 48 h after transfection, the culture supernatant was collected and filtered with a 0.20 µm filter and then stored at −80 °C. Virus titration was performed by serially diluting the viral infection of hACE2-293T cells, and the infectivity was measured by detecting luminescence. The serum of all immunized animals was serially diluted and incubated with a pre-titrated amount of pseudotyped SARS-CoV-2 virus at 37 °C for 1 h. Subsequently, the serum/virus mixture was added to the wells containing 2 × 104 hACE2-293T cells and incubated at 37 °C in 5% CO2 for 48 h. Then the cells were lysed with lysis buffer (Promega), and the lysate was measured by detecting the relative luminescence unit (RLU) in the photometer (Promega) to measure the luciferase activity. GraphPad Prism 9.0 software was used to analyze the serum-neutralizing antibody titers of the pseudotyped virus.

Cell isolation

The spleen or draining lymph nodes were collected and homogenized through a 70 µm strainer in PBS. The red blood cells (RBCs) were removed by adding ACK lysis buffer, followed by centrifuging to discard the supernatant and resuspended the cells with PBS. Mouse primary B cells were isolated by Mouse B Cell Separation Kit (RWD, K1305-20) according to the instruction.

Flow cytometry

Single-cell suspension was stained with fluorochrome-conjugated monoclonal antibodies for 20 min within PBS containing 0.5% BSA on ice. LIVE/DEAD Fixable Viability Dyes (Thermo) were used to gate live cells. The following indicated antibodies were used: CD3- FITC (17A2, Biolegend 100204) CD19-BV510 (6D5, Biolegend 115546), B220-APC/Fire™ 750 (RA3-6B2, Biolegend 103260), B220-APC (RA3-6B2, Biolegend 103212), CD38-PE (90, Biolegend 102707), CD38-PercP5.5(90, Biolegend 102721), IgD-PE(11-26c.2a, Biolegend 405705), IgD-APC (11-26 c.2a, Biolegend 405713), CD95-PE-Cy7 (SA367H8, Biolegend 152617), CD95-PE(SA367H8, Biolegend 152607), GL7-APC (GL7, Biolegend 144617), GL7-FITC (GL7, Biolegend 144604), CD138-BV421 (281-2, Biolegend 142507), CD138-PE(W20051E, Biolegend 142503) CD21-FITC(7E9, Biolegend 123407), CD23-PE(B3B4, Biolegend 101607), ICOSL-PE(HK5.3, Biolegend 107405), ICOSL-Biotin (HK5.3, Biolegend 107403) Streptavidin-BV421(Biolegend 405226), Bromodeoxyuridine (BrdU) (Biolegend, 423401), AnnexinV-PB(Biolegend, 640918), Brdu-AF488(Biolegend, 364106) The flow data were collected by BD LSRFortessa and analyzed with Flowjo 10.4.

Elispot

The Multiscreen filter plates of 0.45 um (Merck Millipore, MSIPS4W10) were coated sterilely with 2.5 ug mL-1 RBD or MD39 in ELISA coating buffer overnight at 4 ˚C, respectively. Plates were washed three times with PBS and blocked with complete DMEM for 2 h at 37 ˚C while spleens were being prepared. The single-cell suspensions from spleen cells were loaded onto the plates at a suitable density per well and left overnight at 37 ˚C. Antibody-secreting cells (ASC) of spleen cells were detected by All-IN-ONE mouse ELISPOT Accessory kit (Dakewe) according to the manufacturer’s protocol. Plates were dried overnight and antigen-specific spots were counted by ImmunoSpot software (Cellular Technology Ltd).

Western blot

The mouse primary B cells from spleens were treated with ferritin or ICOS-RBD nanoparticles for 36 hours, followed by cell lysis and protein immunoblotting. To perform WB, the cells were lysed with RIPA buffer with protease inhibitor cocktail (Roche). Equal amounts of cells lysates were resolved in SDS-PAGE gel and transferred to PVDF membrane (Millipore, IPVH00010). The membrane was blocked with 5% nonfat milk and incubated with appropriate primary and secondary antibodies. Akt (pan) (40D4) Mouse mAb (CST, 2920) (1:2000 for WB), Phospho-Akt (Ser473) (D9E) XP® Rabbit mAb (CST,4060) (1:500 for WB), Phospho-NF-κB p65 (Ser536) (93H1) Rabbit mAb (CST,3033) (1:1000 for WB), NF-κB p65 (D14E12) XP® Rabbit mAb (CST,8242) (1:1000 for WB), PKC-beta Rabbit mAb (Abclonal, A13628)(1:500 for WB), Phospho-PKC (pan) (βII Ser660) Rabbit mAb(CST, 9371) (1:1000 for WB), Cyclin D2 Rabbit pAb (ABclonal, A1773) (1:1000 for WB), Cyclin D3 Rabbit mAb (ABclonal, A3989) (1:1000 for WB), β-Actin Ab Mouse(Proteintech, 66009) (1:2000 for WB), Goat anti-Rabbit Ab(Licor, 926-32211) (1:10000 for WB), Goat anti-Rabbit Ab(Licor, 925-32280) (1:10000 for WB) The signal intensity of WB bands was quantified by Image J.

Statistical analysis

Pilot studies were used for the estimation of the sample size to ensure adequate power. No samples were excluded from the analyses. Data distribution was assumed to be normal but this was not formally tested. Statistical analyses were performed using a two-tailed unpaired Student’s t-test, or One/Two-way ANOVA with GraphPad Prism 9 as indicated in the Figure legends. Statistical differences with a P value of 0.05 or less were considered significant. All assays were performed at least two times. The exact value of n, which represents the number of mice used in the experiments, is indicated in the Figure legends. Details of statistical analyses and biological replicates are described in each figure legend.