New MoDC-Targeting TNF Fusion Proteins Enhance Cyclic Di-GMP Vaccine Adjuvanticity in Middle-Aged and Aged Mice

Cyclic dinucleotides (CDNs) are promising vaccine adjuvants inducing balanced, potent humoral, and cellular immune responses. How aging influences CDN efficacy is unclear. We examined the vaccine efficacy of 3′,5′-cyclic diguanylic acid (cyclic di-GMP, CDG), the founding member of CDNs, in 1-year-old (middle-aged) and 2-year-old (aged) C57BL/6J mice. We found that 1- and 2-year-old C57BL/6J mice are defective in CDG-induced memory T helper (Th)1 and Th17 responses and high-affinity serum immunoglobulin (Ig)G, mucosal IgA production. Next, we generated two novel tumor necrosis factor (TNF) fusion proteins that target soluble TNF (solTNF) and transmembrane TNF (tmTNF) to monocyte-derived dendritic cells (moDCs) to enhance CDG vaccine efficacy in 1- and 2-year-old mice. The moDC-targeting TNF fusion proteins restored CDG-induced memory Th1, Th17, and high-affinity IgG, IgA responses in the 1- and 2-year-old mice. Together, the data suggested that aging negatively impacts CDG vaccine adjuvanticity. MoDC-targeting TNF fusion proteins enhanced CDG adjuvanticity in the aging mice.


INTRODUCTION
The current COVID-19 pandemic has highlighted the vulnerability of aging populations to emerging pathogens where people 45 and older account for ∼97% of COVID-19 deaths (1). Vaccination offers the most efficient and cost-effective method to stop infectious diseases. However, vaccine efficacy is substantially reduced with age due to the progressive age-related decline of innate and adaptive immune responses (immunosenescence) (2)(3)(4). Immunosenescence also posts a significant challenge for the development of immunotherapeutic agents. The number of Americans ages 65 and older will double and reach 80 million in 2040 (5). Improving the efficacy of vaccines and immunotherapeutic agents in the aging population is critical to public health.
The biggest challenge in vaccination response in the aged is the inadequate antibody response (6,7). Nevertheless, improving vaccine efficacy in the aged is achievable. For instance, while the efficacy of Prevnar13 R decreases with age (8) and is only ∼45% effective in adults 65 years or older (9), the Shingrix vaccine for shingles is 90% effective in people over 70. The underlying mechanism behind the successful Shingrix in the aged, however, is unclear. Current vaccine strategies for the elderly are empirical. They include higher antigen dose, adjuvanted vaccine, and alternative route of immunization (10)(11)(12). We reasoned that mechanism-guided vaccine design could aid in the development of effective vaccines for the aged.
In this study, we examined CDG adjuvanticity in middle-aged and aged C57BL/6J mice. We further developed a mechanismguided moDC-targeting strategy to enhance CDG adjuvanticity in the middle-aged and aged mice. Table   REAGENT

Antibody ELISA
A 96-well, flat-bottomed polystyrene plate was coated with 500 ng of H1N1-NP, PspA, or NP 2 BSA, NP 23 BSA, washed with 1 × PBS with Tween 20 (PBST), blocked with 2% bovine serum albumin (BSA) in 1 × PBS, and incubated at 37 • C (20). After wash, serum samples were serially diluted in PBS, added to the wells in triplicate, and incubated at 37 • C for 1 h. The plate was washed. Horseradish peroxidase (HRP)-conjugated anti-mouse IgG/IgG1/IgG2c/IgA was added at a dilution of 1:5,000 and incubated at 37 • C for 1 h. After washing, 3,3 ′ ,5,5 ′tetramethylbenzidine (TMB) substrate reagent was added and incubated for 20 min in the dark. The reaction was stopped using 50 µl of stop solution, and absorbance reading noted at 450 nm.

Ex-vivo Recall of Lung Cells and Splenocytes
Lung and spleen cells were harvested and digested as before (21). The cells were resuspended in RPMI media containing 10% fetal bovine serum (FBS) in a 96-well tissue culture plate at a concentration of 1 × 10 6 /well. The cells were stimulated with 1 µg H1N1-NP or 2 µg NP 6 CGG and incubated for 4 days at 37 • C and 5% CO 2 . Culture supernatants were collected, and cytokine levels were evaluated for interleukin (IL)-13, IL-5, interferon (IFN)-γ, and IL-17a by ELISA.

Flow Cytometry
Lungs were harvested and processed as discussed above. The single-cell suspension was prepared and analyzed by BD TM LSR II and FACScan flow cytometry (21).

Detection of Lung Tumor Necrosis Factor
Mice were intranasally immunized with either PBS or CDG (5 µg) and sacrificed after 5 h by CO 2 asphyxiation. Lungs were perfused with ice-cold PBS, removed and stored in 0.7 ml Tissue Protein Extraction Reagent (T-PER) containing protease inhibitors at −80 • C. Later, the lung was thawed on ice and homogenized with Minilys R (Precellys, 5,000 RPM for 30 s) using Precellys lysing kit. Lung homogenates were transferred to a 1.5-ml tube and harvested at 14,000 g for 30 min at 4 • C. The supernatant was collected and analyzed for TNF production by ELISA.

Experimental Design
Data exclusion was justified when positive or negative control did not work. All experiments will be repeated at least three times. All repeats are biological replications that involve the same experimental procedures on different mice. Experiments comparing different genotypes, adjuvant responses are designed with individual treatments being assigned randomly. Where possible, treatments will be assigned blindly to the experimenter by another individual in the lab. When comparing samples from different groups, samples from each group will be analyzed in concert, thereby preventing any biases that may arise from analyzing individual treatments on different days.
CDG immunization generates potent and long-lasting memory T helper (Th)1, Th2, Th17 response in adult mice (16,(20)(21)(22). However, ex vivo recall assay showed that 60 days post-immunization, aged mice had dramatically (10folds) reduced memory Th1 and Th17 responses in the lung ( Figure 1G). The lung memory Th2 responses were also reduced by half ( Figure 1G). CDG induced lung T follicular helper (TFH), which is critical for its adjuvant responses (22). We examined lung TFH cells on day 14 post-immunization. Lungs from immunized aged mice showed a 10-fold reduction of CXCR5 + PD1 + CD4 + TFH cells than the adult mice ( Figure 1H). Together, the data indicated that CDG adjuvanticity was impaired in the aged mice.
3 ′ ,5 ′ -Cyclic Diguanylic Acid Did Not Activate Lung TNFR2 + cDC2 or Monocyte-Derived Dendritic Cells in Aged Mice CDG activates lung DCs to mediate its adjuvanticity in vivo (21). The lung DCs are functionally heterogeneous. In particular, lung cDC2 and moDCs play a pivotal role in the CDG vaccine adjuvanticity (22). We found that aged mice had increased moDCs in the lung (Figures 2A,B). On the other hand, lung cDC2 was decreased in the aged mice, but the difference was not significant (Figures 2A,B). cDC2 consists of functionally distinct TNFR2 + and TNFR2 − cDC2 subsets (22). The TNFR2 + cDC2 subset (R2D2) mediates the cellular immune responses of CDG adjuvant, while the TNFR2 − cDC2 subset activates moDCs to mediate the humoral responses of CDG (22). Interestingly, lung R2D2 population was significantly decreased in the aged mice ( Figure 2C).
Next, we examined CDG-induced lung DC activation in the aged mice. Intranasal administration of CDG did not activate lung R2D2 or moDCs in the aged mice, as indicated by the lack of CD86 upregulation (Figures 2D,E). CDG activates lung cDC2 to produce lung TNF ( Figure 2F) that is critical for CDG adjuvanticity (19,20,22). We measured lung TNF in aged and adult mice. We first noticed that aged mice had elevated lung TNF levels compared to adult mice at steady state ( Figure 2F). Second, CDG did not increase lung TNF production in the aged mice ( Figure 2F), suggesting that lung cDC2 from the aged mice is defective in response to CDG.
Last, TNFR2 expression on moDCs is essential for CDG adjuvanticity in vivo (22). Consistently, CDG induced less lung TNFR2 + moDCs in the aged mice than the adult mice ( Figure S1A). TNFR2 signals via the non-canonical nuclear factor (NF)-κB pathway RelB. Again, CDG-induced pRelB was reduced in lung moDCs from the aged mice ( Figure S1B). Together, the data suggested that CDG-induced lung R2D2 or moDC activation was impaired in the aged mice.
3 ′ ,5 ′ -Cyclic Diguanylic Acid-Induced Th1 and Th17 Responses Were Impaired in the 1-Year-Old Middle-Aged Mice Adult mice (∼3 months old), not aged mice (∼18 months old), exhibit potent CDG-induced humoral and cellular responses (Figure 1). We asked at what age CDG adjuvant responses start to wane. We examined the 1-year-old mice, which are the equivalent of ∼42.5-year-old middle-aged humans. We immunized 1-yearold mice with CDG/NP 6 CGG twice at 2 weeks' interval and examined antibody and memory Th responses after 60 days. We chose antigen NP 6 CGG so that we can separate vaccineinduced low-affinity (anti-NP 23 ) and high-affinity (anti-NP 2 ) antigen-specific antibodies influenced by aging. High-affinity antibody formation is a hallmark of germinal center formation and provides rapid neutralization of extracellular pathogens.
Surprisingly, while CDG still induced antigen-specific Th2 responses in the lung of 1-year-old mice ( Figure 3B; Figure S2A), CDG barely induced lung memory Th1 or Th17 responses in the 1-year-old mice (Figures 3A,C). Consistently, CDG induced antigen-specific IgG1, but not Th1-related IgG2C in the 1-year-old mice (Figures S2B,C). Next, we examined the high-and low-affinity antibodies in CDG-immunized 1year-old mice. We found that CDG did not induce high-affinity lung IgA in the 1-year-old mice (Figures 3K,L). CDG still induced low-affinity lung IgA (Figures 3I,J). We reasoned that 1-year-old mice might be defective in CDG-induced antibody affinity maturation in the lung. In the serum, CDG induced low-affinity IgG, but the titers of high-affinity IgG was low (Figures 3E,F,G,H). The data, thus, suggested that CDG-induced memory Th1, Th17 responses, and high-affinity antibody production were impaired in the 1-year-old middle-aged mice.

Design a Monocyte-Derived Dendritic
Cell-Targeting TNF-Fusion Protein to Enhance 3 ′ ,5 ′ -Cyclic Diguanylic Acid Adjuvant Responses in Aging Mice The discovery that 1-year-old healthy mice already had waning CDG vaccine response was potentially significant in the translational research of CDNs. The median age for COVID-19 patients in the US is ∼63 years old (23). The median age for pneumococcal meningitis patients is 41 years old (24). The median age for a recent CDN cancer clinical trial was 61 years old (ClinicalTrials.gov NCT02675439). Here, we discovered that CDG adjuvanticity depends on age. To translate CDG research into the clinic, we need methods to improve CDG adjuvanticity in aging humans.
We generated TNF-Fc (IgG2A) and TNF D221N/A223R -Fc (IgG2A) fusion proteins to target solTNF or tmTNF to moDCs  Significance value, p, for (A-F) was calculated between adult and aged mice groups using two-way ANOVA followed by Tukey's multiple comparison test (A,C,E) and Student's t-test (B,D,F). Significance value, p, for (G,H) was calculated between adult and aged mice groups using one-way ANOVA followed by Tukey's multiple comparison test. Error bars represent mean ± SEM. Statistical significance is represented by p; **p < 0.01, ***p < 0.001, ****p < 0.0001. ( Figure 3D). The TNF D221N/A223R mutant mimics tmTNF that binds only to TNFR2, not TNFR1 (25). MoDCs express the highaffinity FcR, FcγRI (also known as CD64) that is not found on cDCs or lymphocytes (26). FcγRI bind the Fc of IgG2a with the highest affinity (10 8 M −1 ), more than 1,000-fold higher than its next binding partner IgG2b-Fc (27). We showed that intranasal administration of allophycocyanin (APC)-conjugated mouse IgG2A was taken up exclusively by CD64 + lung cells, including MHC II hi CD64 + moDCs (28). The CD64 + major histocompatibility complex (MHC) class II low/int macrophages were also targeted (28). However, macrophages are dispensable for CDG mucosal adjuvanticity in vivo (21,22). To test the efficacy of moDC-targeting TNF fusion proteins, we used the IRF4 fl/fl CD11c cre mice. CDG does not generate lung TNF or vaccine responses in the IRF4 fl/fl CD11c cre (22), thus facilitated the TNF fusion proteins complement experiment. Furthermore, IRF4 fl/fl CD11c cre mice only have cDC1 and moDCs. cDC1 is dispensable for CDG adjuvanticity (22). If and analyzed for NP-specific antibody titers by ELISA. IgG and IgA endpoint titers were determined by comparing the OD450 from mice immunized with NP 6 CGG. Significance value was calculated by two-way ANOVA followed by Tukey's multiple comparison test. Error bar represents mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, n.s., not significant. (M,N) One-year-old C57BL/6 mice were immunized (i.n.) twice with PspA, CDG/PspA, or CDG/PspA/(solTNF-Fc + tmTNF-Fc) on days 0 and 14. Serum samples were collected 120 days post last immunization, and anti-PspA specific antibody titer was determined by ELISA (M). Antibody titer was quantified by comparing the OD450 of immunized mice with PspA only immunized mice (N). n = 4 mice/group. Data are representative of three independent experiments. Significance value was calculated by two-way ANOVA followed by Tukey's multiple comparison test (M) or Student's t-test (N). Error bar represents mean ± SEM. **p < 0.01, ***p < 0.001. either solTNF-Fc or tmTNF-Fc restores antibody responses in the IRF4 fl/fl CD11c cre mice, it will strongly suggest that these TNF fusion proteins act through moDCs to enhance CDG adjuvanticity in vivo.
We found that intranasal administration of tmTNF-Fc (IgG2A) plus CDG generated high-affinity serum IgG antibody in the IRF4 fl/fl CD11c cre mice (Figure S3), suggesting that tmTNF-Fc (IgG2A) can enhance CDG-induced humoral response in the absence of cDC2. This is an important feature as we showed that aged mice had a decreased number of cDC2, especially R2D2 that is critical for CDG adjuvanticity. Thus, tmTNF-Fc (IgG2A) may be suitable to enhance CDG antibody responses in the aged mice where lung R2D2 is limited.
SolTNF-Fc (IgG2A) did not restore CDG-induced antibody responses in the IRF4 fl/fl CD11c cre mice ( Figure S3). This was consistent with the previous finding that TNFR2 expression of moDCs is important for CDG-induced lung TFH cells and GC (Germinal Center) B cell production (22). Last, as expected, neither tmTNF-Fc (IgG2A) nor solTNF-Fc (IgG2A) restored the memory Th responses in the IRF4 fl/fl CD11c cre mice (28) as we previously demonstrated that R2D2 is required for CDG-induced cellular immunity (22).
Last, we immunized (i.n.) 1-year-old mice with CDGadjuvanted mucosal pneumococcal protein subunit vaccine CDG/PspA with or without TNF-Fc fusion proteins twice at 2 weeks' interval. Serum anti-PspA IgG was determined 120 days post the last immunization. Remarkably, 1-year-old mice immunized with CDG/PspA plus TNF-Fc fusion proteins had stronger serum anti-PspA IgG 4 months post-immunization than the 1-year-old mice immunized with only CDG/PspA (Figures 3M,N). Together, the data indicated that the moDCtargeting TNF fusion proteins restored the memory Th1, Th17, and high-affinity antibody production in CDG-immunized 1year-old middle-aged C57BL/6J mice.
Intranasal administration of TNF elicits antibody and Th2 responses (29). However, intranasal administration of solTNF-Fc (IgG2A) or tmTNF-Fc (IgG2A) alone did not induce any antibody or memory Th responses (Figure S4), indicating that the moDC-targeting TNF fusion proteins enhanced CDG adjuvant responses, rather than producing vaccine responses de novo. We used a much lower dose of TNF-Fc (IgG2a) (100 ng) than previous TNF (2 µg) used as a mucosal adjuvant (29).
The induction of antigen-specific IgA response in the lung is highly desirable for preventing respiratory infections such as influenza, S. pneumococcal, and SARS-CoV infections. The vast majority of licensed vaccines do not generate lung mucosal IgA responses, even in healthy adults. CDG-adjuvanted vaccines induce antigen-specific lung IgA in the adult but not aged mice ( Figure 1C). We next examined if the moDC-targeting TNF fusion proteins restored lung IgA production in the aged mice. We collected BALF from the CDG/NP 6 CGG or CDG/NP6CGG + solTNF-Fc + tmTNF-Fc immunized (i.n.) 2year-old C57BL/6J mice 60 days post the last immunization and determined antigen-specific IgA titer including high-affinity IgA. A significant elevation in the IgA titer was observed in aged mice immunized with CDG/NP 6 CGG adjuvanted with solTNF-Fc (IgG2A) and tmTNF-Fc (IgG2A) as compared to the mice group immunized with only CDG/NP 6 CGG (p < 0.0001) (Figures 4G,H; Figure S4D). Further analysis showed CDG/NP 6 CGG elicited primarily low-affinity lung IgA titer while CDG/NP 6 CGG with moDC-targeting TNF fusion proteins generated high-affinity lung IgA (p < 0.0001) in the 2-year-old mice (Figures 4I,J).
Together, the data indicated that the moDC-targeting TNF fusion proteins markedly enhanced CDG-induced durable and high-affinity antibody responses, both in the serum and in the lung, in the 2-year-old C57BL/6J mice.
SolTNF-Fc (IgG2A) and TmTNF-Fc (IgG2A) Restores Systemic and Lung Memory Th1, Th17 Responses in 3 ′ ,5 ′ -Cyclic Diguanylic Acid-Immunized 2-Year-Old C57BL/6J Mice CDG-induced memory T cell response, primarily Th1 and Th17, is impaired in aged mice ( Figure 1D). Vaccine-induced Th1 responses are critical for the protection against intracellular pathogens, while Th17 responses, via enhancing neutrophil recruitment, are essential for clearing bacteria, e.g., Streptococcus pneumoniae infections. Next, we examined memory T cell responses in 2-year-old mice immunized with the moDCtargeting TNF fusion protein-adjuvanted CDG/NP 6 CGG vaccine 60 days post the last immunization by ex vivo recall assay. We examined memory T cell production in the lung (mucosa) and spleen (systemic) in the immunized mice.

DISCUSSION
In this report, we take a mechanism-guided rationally designed approach to improve vaccine efficacy in the middle-aged and aged mice. Our vaccine strategy is (i) to identify the in vivo mode of action of CDG adjuvant in the adult mice (20,22); (ii) to define the defects in the aged mice that compromises CDG adjuvanticity in vivo; (iii) to design new biologics to circumvent the defective component in CDG mode of action and restore CDG adjuvanticity in the aged mice. Our new vaccine strategy focuses on moDC targeting and TNF induction because (i) CDG adjuvanticity is mediated by two DC subsets: cDC2 and moDCs (22); (ii) CDG-induced cDC2 production of TNF is critical for CDG adjuvanticity (19,20,22); (iii) aged mice had decreased cDC2 but increased moDCs; and (iv) CDG did not induce lung TNF in the aged mice. We designed the TNF-Fc (IgG2A) to targeted activating moDCs by TNF. Our strategy successfully restored CDG adjuvanticity in the middle-aged and aged mice.
Aging leads to a progressive decline of the innate and adaptive immune system that likely causes the reduced vaccine efficacy in the aged. Our vaccine strategy intends to adapt rather than change or reverse the immunosenescence in the aged. The current vaccine strategies to enhance vaccine responses in the elderly are empirical due to our limited understanding of immunosenescence (30). The success of our targeted vaccine strategy depends on the knowledge of CDG adjuvant mode of action in the adult and aged mice. This new strategy does not require a deep understanding of immunosenescence; thus, it is likely more practical for the development of vaccines for the aged.
It remains to be determined if our moDC-targeting TNF fusion proteins will enhance non-CDN vaccines in the aged. Most vaccines likely will have a mode of action different from CDNs. To adapt this targeted vaccine strategy requires efforts to understand the mode of action for each vaccine, which may be challenging. To solve this limitation, we are exploring if our moDC-targeting TNF fusion proteins can enhance the efficacy of non-CDN vaccines in the aged.
CDNs are excellent adjuvants eliciting not only long-lasting and potent antibody responses but also Th1, Th2, Th17, and antitumor CD8 T cell response. There were tremendous interests in translating CDNs into the clinic. For example, in 2015, Aduro Biotech and Novartis announced a $250M-plus initiative to develop CDNs as cancer immunotherapies (31). Surprisingly, how aging affects CDN vaccine adjuvanticity was not well-addressed.
Vaccination offers the most efficient and cost-effective method of preventing infectious diseases. However, vaccines are less effective in the elderly. For instance, the influenza vaccine has an efficacy between 70 and 90% in children and adults but dropping to 30-50% for those over 65 years of age (4,34). Similarly, responses to pneumococcal polysaccharide and hepatitis B vaccines are compromised by the old age (35). In addition, ∼83% of pneumonia deaths in the US occurred in the elderly (36,37). Last, people 45 and older (middle-aged and aged) account for ∼97% of COVID-19 deaths (1). Methods to improve vaccine protection in the elderly are highly desirable. We used the model antigen NP 6 CGG to reveal a defect of high-affinity antibody and Th1/17 induction in the middle-aged and aged mice. We observed the aged-related defect using H1N1-NP and PspA antigens. TNF-Fc fusion protein enhanced the duration (120 days post-immunization) and the magnitude of CDG/PspA antibody production in 1-year-old mice. CDN-adjuvanted protein subunit vaccines protected adult mice from respiratory infections such as influenza (38,39), bacterial pneumonia (40), Mycobacterium tuberculosis (41), and anthrax (42). CDNs were also therapeutic for cancer in adult mice (13,14). Future studies are needed to determine if the TNF-Fc fusion proteins can improve protection in CDN-adjuvanted vaccines for infectious diseases and cancers in the elderly.
As a vaccine specific to the elderly, safety is paramount. Intranasally administered CDG (5 µg/mouse) does not cause acute toxicity in mice (20,43,44) and is proven safe in various animal models (38,40,(44)(45)(46)(47). Clinical trials on CDNs (ClinicalTrials.gov NCT02675439, NCT03010176, NCT03172936, NCT03937141, and NCT0414414) have established an excellent safety file in humans, including the elderly. The human body makes TNF. Low-dose (ng) targeted delivered TNF biologics are likely safe in the elderly as well. Nevertheless, a future study is needed to establish a safety profile for the CDG/TNF biologics vaccines for the elderly.
Last, the moDC-targeting TNF fusion proteins applied here did not have adjuvant activity when administered in the absence of CDG. Furthermore, TNF when used as a mucosal adjuvant mainly induces Th2 responses (29). Here, moDC-targeting TNF biologics enhanced memory Th1 and Th17-immunity of the CDG. We propose that TNF fusion proteins enhanced CDG adjuvanticity rather than acted as a separate adjuvant to replace CDG in aged mice. In all, we showed that aging negatively influences CDG adjuvanticity, especially CDG-induced memory Th1 and Th17 responses. A new moDC-targeting strategy can restore CDG adjuvanticity in middle-aged and aged mice.

DATA AVAILABILITY STATEMENT
All datasets presented in this study are included in the article/Supplementary Material.

ETHICS STATEMENT
The animal study was reviewed and approved by Institutional Animal Care and Use Committee, University of Florida.

AUTHOR CONTRIBUTIONS
SM and LJ conceived the research. LJ designed the experiments, wrote the manuscript, and supervised the research. HG, SM, DK, and LJ performed experiments and analyzed the data. HG drafted the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
This work was supported by NIH grants AI110606, AI125999, and AI132865 (to LJ).