C57BL/6J mice were acquired from Orient Bio in Korea. CD45.1⁺ SMARTA mice were kindly provided by Youn Soo Choi at Seoul National University College of Medicine. Specific-pathogen-free male mice (6-8 weeks of age) were used for experiments. All procedures involving animals were conducted in compliance with the protocols approved by the Institutional Animal Care and Use Committee of the College of Medicine at The Catholic University of Korea (CUMC-2023-0164).

Spleens from SMARTA mice were harvested, and naive CD4⁺ T cells were isolated using negative selection according to the manufacturer’s protocol (The EasySep™ Mouse CD4⁺ T Cell Isolation Kit, STEMCELL, #19852). The cells were counted, and the proportions of TCRVα2⁺ CD45.1⁺ CD4⁺ T cells were determined by flow cytometry. 5 x 10⁴ (for day 8) or 10 x 10⁴ (for day 15) naive SMARTA T cells (CD45.1⁺) were adoptively transferred into recipient mice (CD45.2⁺) via intravenous injection into the retroorbital sinus.

For immunization, 5 μg of SARS-CoV-2 spike RBD (sRBD), 10 μg of KLH-NP, or 10 μg of KLH-gp₆₁ were each mixed with Alum (Alhydrogel; Invivogen) alone or combined with 5 μg of 3′3′-cGAMP (Invivogen) to a final volume of 20 μL. These preparations were administered into each footpad of the mice. The chosen cGAMP concentration was based on its ability to induce B_GC and B_PC responses in dose-response trials using the KLH-NP antigen (1, 5, and 10 μg). A 5 μg cGAMP dosage reliably provoked B_GC and B_PC responses, with 10 μg showing no further impact. For IFN-β blockade, mice received 250 μg of anti-IFN-β blocking antibody or isotype control via intraperitoneal injection three times during the course of protein immunization.

Single-cell suspensions from popliteal or inguinal lymph nodes were stained with a panel of monoclonal antibodies for surface markers including CD4 (RM4-5, BV510), CD8 (53-6.7, BV605), SLAM1 (TC15-12F12.2, PepCP-Cy5.5), ICOS (C398.4A, FITC), IgD (11-26c.2a, PE), CD138 (281-2, APC), Streptavidin (BV421) (BioLegend), B220 (RA3-6B2, AF700), CD44 (IM7, PE.Cy7), PD1 (J43, PE), CD4 (RM4.5, APC-eF780), CD8 (53-6.7, APC-eF780), CD45.1 (A20, APC-eF780 or BV605), Fixable Viability Dye eF780 (eBioscience) and FAS (Jo2, BV510) (BD Biosciences) at 4°C for 30 min in PBS with 0.5% BSA.

For cytokine detection, the cells were in vitro cultured with 10 μg/mL gp_66-77 and 1 μg/mL brefeldin A for 5 hours. Intracellular staining for cytokines was performed with monoclonal antibodies for IL-4 (11B11, PE), IFN-γ (XMG1.2, AF700) (eBioscience), IL-2 (JES6-5H4, PE.Cy7) (BioLegend), and recombinant mouse IL-21 receptor Fc (R&D systems), followed by anti-human IgG (DyLight650) (Invitrogen). A Fixation/Permeabilization kit (BD Biosciences) was employed for permeabilization. Intracellular staining for T-bet transcription factor was performed with monoclonal antibody T-bet (4B10; PerCP.Cy5.5) using the Foxp3/Transcription Factor Staining Buffer Set (eBioscience). The cells were then analyzed using an LSR Fortessa (BD Biosciences) and FlowJo software v.10.8.

Flat-bottom immuno plates (MaxiSorp) were coated with 1 μg/mL of various antigens (sRBD, BSA-gp₆₁, OVAL-NP₍₁₇₎, BSA-NP₍₂₎ for high-affinity antibodies, and BSA-NP₍₃₆₎ for low-affinity antibodies) in PBS at 4°C overnight. Plates were washed with PBST (PBS + 0.1% Tween-20) and blocked with PBST-B (PBS with 0.05% Tween-20 and 0.5% BSA) for 90 min at room temperature. Mouse sera were then added in serial dilution and incubated for 90 min at room temperature. After washing, HRP-conjugated goat anti-mouse IgG (Thermo Fisher Scientific) was added in PBST-B (1:5,000), and the plates were incubated for another 90 min at room temperature. The enzymatic reaction was developed with a TMB solution (BioLegend). The reactions were stopped with 2N sulfuric acid, and the absorbance was measured at 450 nm with a microplate reader.

STING ligands, such as 2’3’-cGAMP, 3’3’-cGAMP, and their derivatives, have been shown to enhance protective immunity against various pathogens, including SARS-CoV-2 and influenza (9–11). However, the impact of STING ligands on the germinal center reaction has remained somewhat unclear. To investigate the effect of STING ligands on the differentiation of B_GC, B_PC, and the generation of antigen-specific antibodies, we conducted experiments employing 3’3’-cGAMP (cGAMP) in response to various antigens. Mice were immunized with the SARS-CoV-2 spike RBD protein (sRBD) as an antigen, either in Alum alone or Alum+cGAMP as an adjuvant ( Figure 1A ). Mice immunized with Alum+cGAMP adjuvant exhibited a significant increase in frequencies of B_GC (Fas^hiGL7^hi) cells and B_PC (IgD^loCD138^hi) compared to the mice immunized with Alum alone on day 8 after immunization ( Figure 1B ). The kinetics of B_GC and B_PC differed significantly. The increase in B_GC frequency induced by Alum+cGAMP persisted for 30 days with the peak response on day 15 after immunization. In contrast, the variance in B_PC frequency peaked on day 8 and gradually declined over one month post-immunization ( Supplementary Figure 1A ). Consistent with a previous report (9), sRBD-specific antibody titers were significantly higher in mice treated with Alum+cGAMP compared to those treated with Alum alone ( Figure 1C ).

We further investigated whether the effect of cGAMP on B_GC and B_PC is a general process in immunization with different antigens. Mice were immunized with keyhole limpet hemocyanin (KLH) conjugated with hapten 4-hydroxy-3-nitrophenyl acetyl (NP) (KLH–NP) as a model antigen, either in Alum alone or Alum+cGAMP ( Figure 1D ). Similarly, B_GC and B_PC responses were enhanced in mice treated with Alum+cGAMP in response to KLH–NP immunization ( Figure 1E ), along with an increase in NP-specific antibody production ( Figure 1F ). The increased frequency of both B_GC and B_PC was sustained for over one month when mice were immunized with Alum+cGAMP for priming (day0) and boosting (day14), suggesting that cGAMP treatment in primary and boost immunization can prolong the GC response and the differentiation of B_PC ( Supplementary Figures 1B,C ). Moreover, the magnitude of high-affinity (NP₍₂₎; NP₂-bound) antibody production was significantly greater in the Alum+cGAMP group compared to the control Alum alone group ( Supplementary Figure 1D ). These data demonstrate that cGAMP enhances the differentiation of GC B cells, plasma cells, and the production of antigen-specific antibodies.

Ligation of cGAMP to STING triggers the production of type I interferon (IFN) (13). To explore whether cGAMP has the potential to enhance the germinal center response through the type I interferon pathway, we immunized mice with KLH–NP in Alum+cGAMP and administered either anti-IFN-β blocking antibodies or isotype control antibodies before (D-1) and after (D1 and D3) immunization ( Figure 2A ). Interestingly, there were no significant changes in the frequency of B_GC in anti-IFN-β group. However, the frequency of B_PC was significantly reduced upon blocking IFN-β ( Figure 2B ). Consistent with the B_PC response, the production of NP-specific IgG was also decreased ( Figure 2C ). In summary, cGAMP plays a critical role in enhancing the differentiation of germinal center B cells and plasma cells, ultimately resulting in the production of antigen-specific antibodies. Specifically, cGAMP regulates plasma cell generation through the type I interferon pathway.

T_FH, a subset of CD4 T cells, plays a crucial role in GC formation, B cell differentiation and affinity maturation, and the production of high-affinity antibodies by producing plasma cells (1). We explored the potential of cGAMP to promote T_FH differentiation. Notably, cGAMP significantly enhanced the frequency of GC T_FH (CXCR5^hiPD1^hi) following sRBD immunization ( Figures 1A , 3A ). Furthermore, cGAMP markedly upregulated the expression of inducible T-cell costimulator (ICOS), a critical costimulatory molecule for T_FH differentiation and synapse formation with B cells ( Figure 3B ). This upregulation was also observed in mice immunized with KLH-NP antigen, where cGAMP increased the frequency of GC T_FH cells compared to immunization with Alum alone ( Figures 1D , 3C ). Importantly, the presence of GC T_FH cells was sustained for over a month when mice received prime-boost immunization with KLH-NP in Alum+cGAMP, unlike those immunized with Alum alone ( Figure 3D and Supplementary Figure 1B ), indicating the capacity of cGAMP to induce T_FH differentiation.

The effect of cGAMP on T_FH response to sRBD and KLH-NP antigens corresponds with the enhanced formation of B_GC and B_PC. Intriguingly, despite cGAMP’s known role in promoting T_H1 differentiation and IFN-γ production through STING activation in CD4 T cells, it can also augment T_FH differentiation (14). We previously noted an increased frequency of antigen-specific CXCR5^loBlimp1⁺ non-T_FH cells and a decreased CXCR5⁺ T_FH population upon immunization with KLH conjugated with lymphocytic choriomeningitis virus (LCMV) glycoprotein 61–80 peptide (KLH–gp₆₁) in Alum+cGAMP compared to Alum alone on day 7 (15). To explore whether cGAMP elicits variant CD4 T cell responses to different antigens, mice were immunized with the KLH-gp₆₁ antigen in Alum alone or Alum+cGAMP. Consistent with our previous findings (15), cGAMP substantially reduced the frequency of GC T_FH cells on day 8, which is considered the peak T_FH response ( Figure 3E and Supplementary Figure 2A ). Although the frequency of GC T_FH cells was lower in the Alum+cGAMP, the absolute number of GC T_FH cells was higher. Furthermore, B_PC differentiation post-KLH-gp61 immunization on day 8 was elevated compared to that with Alum alone ( Supplementary Figure 2B ). This indicates that cGAMP’s influence on B_PC generation may not solely rely on the T_FH to non-T_FH ratio but also on the number of T_FH cells present during this stage.

To determine whether the disparity in T_FH frequency in response to the KLH-gp₆₁ antigen was time-dependent, we evaluated the T_FH response on day 15 post-immunization. In this experiment, we performed adoptive transfer of SMARTA CD4 T cells to examine both antigen-specific and endogenous polyclonal CD4 T cell responses ( Supplementary Figure 2C ). On day 15, the frequency of CD45.1⁺ gp-specific GC T_FH was comparable between the Alum+cGAMP and Alum alone groups in the KLH-gp₆₁-immunized mice ( Figure 3F ; upper panel). Intriguingly, while Alum+cGAMP markedly increased the frequency of GC T_FH cells in response to KLH-NP on day 15 ( Figure 3C ), the frequency of GC T_FH in the polyclonal CD4 T cell population (CD45.1^-) in KLH-gp₆₁-immunized mice treated with Alum+cGAMP was similar to that in the Alum alone group ( Figure 3F ; lower panel). These findings suggest that different conjugations to carrier proteins may induce varied T-dependent immune responses. Although the frequency of GC T_FH was lower on day 8 or similar on day 15 in the Alum+cGAMP group, the total number of GC T_FH cells was significantly higher in Alum+cGAMP-treated mice than those given Alum alone ( Figures 3E, F ). These observations suggest that cGAMP may initially favor non-T_FH differentiation over T_FH differentiation, especially in response to the KLH-gp₆₁ antigen, yet it appears to support sustained T_FH responses over time, influencing both the proportion and absolute number of T_FH cells.

We further examined the characteristics of antigen-specific CD4 T cells on day 8 after immunization. The mice were adoptively transferred with SMARTA CD4 T cells, followed by immunization with KLH-gp₆₁ in Alum alone or Alum+cGAMP ( Figure 4A ). Similar to the observation from polyclonal CD4 T cell responses on day 8, frequency of CXCR5^hiPD1^hi GC T_FH cells decreased significantly in the Alum+cGAMP group compared to the Alum alone group ( Figure 4B ). Interestingly, the expression level of ICOS markedly increased in the Alum+cGAMP group, while those of CXCR5 was similar ( Figure 4C ). These suggest that although the frequency of GC T_FH was lower on day 8, the ICOS⁺ T_FH cells might have the potential for further differentiation into GC T_FH at a later time point. Moreover, Alum+cGAMP induced the production of IL-21, an important cytokine for the B_GC differentiation, in CXCR5⁺ T_FH cells ( Figure 4D ). Similarly, IL-2-producing T_FH cells also increased in the Alum+cGAMP group. In contrast, IL-4-producing T_FH cells were substantially decreased by cGAMP treatment, while the expression of IFN-γ was increased in T_FH population by cGAMP ( Figure 4D ). Furthermore, the majority population of CXCR5^low non-T_FH SMARTA cells was IFN-γ⁺ cells in mice treated with Alum+cGAMP, emphasizing that cGAMP enhanced T_H1 response and repressed T_H2 and T_FH responses at earlier immune response ( Figure 4E ). Correspondingly, the non-T_FH population expressed higher levels of the T-bet transcription factor in the Alum+cGAMP group ( Figure 4F ). Taken together, these results demonstrate that cGAMP downregulates GC T_FH differentiation in response to KLH-gp₆₁ on day 8, but it improves its regulatory function over time.