Rescuing age-associated vaccine hyporesponsiveness in murine and nonhuman primate models with an immunostimulatory vaccine adjuvant Authors: Jill Maxwell, Jason A Joyner, James E Norton, Walter Knapp, I-Ming Wang, Gokul Swaminathan, Morgan A Monslow, Michael P Citron Published: 2026 ABSTRACT ABSTRACTAge plays a crucial role in controlling the outcomes of infection, from duration and consequence to effectiveness of active immunization. Attenuated responses to vaccines in the elderly are contrary to the significant benefits that vaccination provides for society. The decline in normal immune function, collectively known as immunosenescence, is often reflected in reduced vaccine response rates. To investigate the influence of age on systemic immunization, different vaccine modalities in aged animals were compared to young counterparts, first in mice and then in nonhuman primates. In this work, reduced antibody production following systemic immunization was the major age-associated biomarker observed in aged mice and monkeys. The lower vaccine antibody phenotype observed in these aged animals was reversed using a well-established immunostimulatory adjuvant that has demonstrated clinical effectiveness in humans. This finding offers a valuable model for exploring age-related changes in the immune system that are associated with immunosenescence and importantly translatable solutions to address the associated outcomes for vaccines targeting the elderly population.KEYWORDS: Animal models, vaccine hyporesponse (VHR), aging, immunosenescence, adjuvant Introduction Advancing age is associated with challenges in active immunization. Vaccines for Influenza,1,2 Pneumonia,3 and Varicella4,5 have shown inconsistent effectiveness in older adults. Various innovations have been employed to begin to address the phenomena of vaccine hyporesponsiveness (VHR) and manage the process of immunosenescence, the gradual decline of the immune system with age. These include increasing antigen doses,6 using potent adjuvants,7 creating novel formulations,8 and altering routes of administration.9 As global life expectancy rises and personalized vaccines for the elderly are developed,10 there will be a growing need for advanced model systems that accurately mimic human aging to evaluate vaccine efficacy and safety.11 Preclinical animal models have contributed significantly to the understanding of adaptive immunosenescence.12 However, notable differences exist between animals and humans, such as differences in lifespan and the maintenance of naive T cells.13,14 Here, we assess the suitability of rodents and nonhuman primates (NHPs), two commonly used species in vaccine research, across their lifespans. The similarities these models share with clinical observations in humans, especially with respect to the NHP, are highlighted. Biomarkers associated with age and VHR in NHP were explored, with a primary focus on the weakened antibody responses in the older animals. These findings demonstrate that both models reproduce immune dysregulation, primarily associated with reduced vaccine efficacy seen with age. These observations have the utility to evaluate strategies to overcome these challenges and explore possible mechanisms. Materials and methods Animal ethics statement All experiments involving laboratory animals were approved by the Institutional Animal Care and Use Committee (IACUC) at Merck & Co., Inc., Rahway, NJ, USA, Mannheimer Foundation, and the University of Louisiana at Layfette, New Iberia Research Center, and conducted in accordance with the Guide for the Care and Use of Laboratory Animals. Additionally, the studies adhered to the ARRIVE guidelines of the National Center for the Replacement, Refinement, and Reduction of Animals in Research.15 Human, mouse, monkey age relationship Mice and monkeys exhibit rapid developmental progression early in life compared to humans, and the equivalence ratio changes with age. The correspondence between these species is not simple, proportional, or linear across the lifespan and developmental stage. Studies and resources such as the Jackson Laboratory and other online calculators can estimate relative age across species using biomarkers of cellular and molecular senescence and other biological and structural changes.16–20 In general, 18–24-month-old mice are considered old and broadly comparable to humans aged 56–69 y, although this varies by strain and experimental context.21 For nonhuman primates, age equivalence estimates also vary, with several sources suggesting an approximate human-to-monkey ratio of about 1:3 in adulthood, depending on species, sex, the biological endpoint of interest, and the age passed the post-senescent stage.22,23 In our study, age groups were selected to balance biological relevance with practical and ethical considerations while enrolled into and over the course of a study. Mouse immunizations Female C57Bl/6J mice were purchased from Jackson Laboratories (Bar Harbor, ME, USA) and housed in microisolator cages in the animal facility at the research laboratories of Merck & Co., Inc., Rahway, NJ, USA (MRL). Food and water were provided ad libitum. Groups of eight mice, aged approximately 7 and 52 to 68 weeks, were immunized intramuscularly (I.M.) at weeks 0 and 3 with either Twinrix® Hepatitis A/B (GSK, Durham, NC, USA) using 1/10 human dose, Fluzone Quadrivalent (Sanofi Pasteur Inc.) at 1/5 human dose, OVA LIONEX Diagnostics and Therapeutics, Braunschweig, Germany) at 10ug dose or recombinant gE protein24 with adjuvant, AS01 GSK Durham, NC, USA) at 1/10 human dose. Blood was collected from the tail vein, 3 weeks following the first immunization and 2 weeks following the second immunization. Sera from whole blood were processed and stored frozen until analyzed. NHP Immunizations Domestically bred African green monkeys (Chlorocebus sabaeus) and rhesus macaques (Macaca mulatta) were maintained either at the Mannheimer Foundation, Homestead, Florida, or the New Iberia Research Center (NIRC) of the University of Louisiana at Lafayette, New Iberia, LA, USA. For the immunization studies, sedated monkeys were immunized intramuscularly (I.M.) with Twinrix® Hepatitis A/B at weeks 0, 4, and 8 using the human adult volume of 1 ml; tetanus toxoid (Zoetis, Inc. Kalamazoo, MI, USA) at weeks 0 and 4 using 0.5 ml; or varicella-zoster virus (VZV) in the form of Shingrix® (GSK) or gE protein,24 field mixed with AdjuPhos®, an aluminum phosphate wet gel suspension (Invivo Gen San Diego, CA, USA) or AS01B, administered in 1 ml at Day 0 and again 1 month later. For the study with both HepA/B and tetanus, the vaccines were administered concomitantly in alternative limbs. Blood via venipuncture was collected on sedated animals prior to the first immunization and intermittently after each immunization. Additionally, for the HepB study in NHP, animals were bled acutely post‑immunization for cytokines and for complete blood chemistry and in Paxgene tubes for nucleic acid isolation. Determination of serum IgG titers For antibody determinations in mice, either Hepatitis B surface Antigen (HBsAg) protein (2 μg/ml), Fluzone vaccine (2 μg/ml), OVA protein (10 μg/ml)from Chondrex, Inc., Woodinville, WA, USA, or recombinant His-tagged gE protein from varicella-zoster virus (VZV) was coated on 96-well Maxisorb Nunc Immunoplates® (Thermo Fisher Scientific, Rochester, NY, USA) in PBS at 4 C overnight. Plates were washed with PBS containing 0.05% Tween-20 (PBST) and blocked with 3% Nonfat milk in PBST for 1 h at room temperature. One hundred microliters of the antisera raised against the test immunogens was diluted starting at 1:20 (post-dose 1) or 1:200 (post-dose 2) in a 3-fold series in milk-PBST and added to each well. Plates were incubated for about 1.5 h at room temperature followed by washing with PBST. HRP-conjugated goat anti-mouse IgG (H + L) secondary antibody (Zymed, South San Francisco, CA) in milk-PBST (50 μl of a 1:2000 dilution) was added to each well and incubated at room temperature for 1 h. Plates were washed with PBST followed by development with 100 μl per well of the substrate SuperBlu-Turbo3,30,5,50-tetramethylbenzidine (TMB) solution (Virolabs, Chantilly, VA) and stopped after 3–5 min of development with 100 μl of Stop SolutionTM® (Virolabs Chantilly, VA, USA). Absorbance (450 nm) was measured, and the antibody titer was defined as the reciprocal of the highest dilution with absorbance 2 standard deviations above the mean of control samples for HepB, Flu, and Shingrix samples or extrapolated against standard curve for OVA samples. T-test between the geometric mean (GMT) of young and older adults was done via GraphPad Prism® 10.2.2 software. For antibody determinations in NHP, VZV gE antibodies were measured using the same method as in mice. For Tetanus antibody, however, anti-tetanus toxoid IgG1 levels from monkey serum were measured using a commercially available ELISA Kit (Abnova Tetanus Toxoid IgG1 Monkey ELISA Kit, Catalog #KA2442) and conducted, sample blinded, by ImQuest BioSciences Frederick, MD, USA. Serum was diluted 1:200 in assay diluent (1.5 μL of specimen to 298.5 μL of assay diluent). The 2.5 μg/mL working standard was prepared by diluting 8.5 μL of the reference standard in 491.5 μL of assay diluent. Two-fold serial dilutions were performed five times in assay diluent to prepare subsequent dilutions (1.25 μg/mL to 0.078 μg/mL). One-hundred microliters of each sample was added to the plates as a single replicate, and 100 μL of each concentration generating the standard curve was added to the plate in duplicate. Plates were covered with a sealer and incubated at room temperature on an orbital shaker (100 to 150 rpm) for 60 min (+/- 5 min). Following the incubation, the plates were washed manually five times with 400 μL of 1X wash solution. Plates were blotted to remove any residual wash solution. One-hundred microliters of enzyme conjugate was added to all wells on the plates. The plates were covered with a plate sealer and incubated for an additional 45 min at room temperature on an orbital shaker (100 to 150 rpm). Following the second incubation, the plates were washed similarly as described above. Following the washing, 100 μL of TMB substrate reagent was added to all wells, and the plates were incubated for 20 min at room temperature on an orbital shaker (100 to 150 rpm). One hundred microliters (100 mL) of stop solution was added to each well to terminate the reaction. The plates were gently tapped to ensure complete mixing before reading spectrophotometrically at 450 nm. The standard curve, ranging from 2.5 μg/mL to 0.078 μg/mL, was graphed using linear regression, and the antibody concentration for each sample was extrapolated from the standard curve. Titers and analysis were graphed using GraphPad Prism® 10.2.2 software. Composite baseline bioage score using RNA-seq To assess whether a NHP transcriptional signature is analogous to humans, we evaluated 72 NHPs, including 39 African green monkeys (AGMs) and 33 rhesus monkeys (RHMs) spanning a broad age range. Animals were stratified into chronological age cohorts; 30 animals were ≤2 y old and 42 were 15 or older. Among the older animals, the mean age was 23.0 y (median, 23.5 y; range, 15–31 y), corresponding to approximately 69.0 human years by a 1:3 monkey-to-human age conversion. Whole blood was collected into PAXgene® tubes, and total RNA, including microRNA, was isolated following ribosomal and globin RNA reduction using RiboZero and GlobinClear. RNA quantity, purity, and integrity were assessed by A260/A280 ratio, 28S/18S ratio, and Agilent TapeStation analysis. For RNA-seq library preparation, ribosomal and globin RNA-depleted RNA was fragmented and converted into indexed cDNA libraries, which were quantified by qPCR using primers specific for Illumina sequencing adapters. Libraries passing quality control were sequenced at EA Genomics (Morrisville, NC, USA) on an Illumina (HiSeq 2500) platform according to the manufacturer’s instructions. Raw FASTQ files were used for downstream analyses, including gene count quantification, differential expression analysis, variant calling, and fusion detection. Differentially expressed genes where assessed for Biological Process, Molecular Function, and Cell Components and manually identified, clustered, and annotated with OmicSoft (v10.1) [OmicSoft Corporation, Cary, NC, USA], as previously described.25 Cytokine and cellular markers A 23-plex cytokine/chemokine (Milliplex MAP Nonhuman Primate Cytokine Magnetic Bead Kit, EMD Millipore; Cat. #PCYTMG-40K-PX23) Luminex assay using a Curiox DropArray 96-well Bead Plate system was used to evaluate the following markers; G-CSF, GM-CSF, IFNγ, IL-1β, IL-1RA, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12/23 (p40), IL-13, IL-15, IL-17A, IL-18, MCP-1, MIP-1β, MIP-1α, sCD40L, TGFα, TNFα, VEGF. Results Hyporesponse and adjuvant rescue of the hyporesponse in mice Aged mice show a diminished vaccine effect for various vaccine modalities compared to young mice. Subunit vaccines, such as OVA protein Figure 1(A) and HepB surface antigen Figure 1(B), plus AdjuPhos®, an aluminum phosphate (AP) wet gel suspension (Invivo Gen San Diego, CA, USA), as well as inactivated such as seasonal influenza vaccine Fluzone Quadrivalent Figure 1(C), each had inferior responses against the vaccine antigens in ELISA in older mice versus their younger counterparts. The decreased responses in the aged mice are also observed after a single dose of gE protein formulated with the AS01 adjuvant system using 1/10 human dose of the commercially available vaccine Shingrix®. However, the lower antibody responses in the aged mice are ameliorated after immunization with the second dose Figure 1(D). Antibody titers are summarized in Table 1. Vaccine antibody titers in young and aged mice. Mean titers at pd1 and pd2 for both young (≤8 weeks) and older mice (≥52 weeks). Mice were immunized two times with (a) OVA protein formulated with Lipid Nanoparticle (LNP), (b) HBSAg formulated with both aluminum hydroxide and aluminum phosphate, (c) Fluzone® or (d) Shingrix®, 3 weeks apart. Serum for antibody determinations using ELISA was collected 3 weeks after the first immunization and 2 weeks after the second immunization. Significance is indicated (*) using One-way ANOVA with Sidak’s multiple comparisons test. Table 1.Vaccine antibody determinations in young and aged mice.VaccineMean Titer Young/OlderDifference Older/Young ±SEM95% Confidence IntervalP-valuePd1Mean Titer Young/OlderDifference Older/Young ±SEM95% Confidence IntervalP-value Post-dose 1Post-dose 2OVA+LNP14,908/4,999−9,909 ± 2055−14316 to −55020.0003817,435/423070−394365±−562084 to −2266460.0002HBSAg+AP780/410−370.0 ± 212.6−825.9 to 85.920.103746,170/7560−38610 ± 12860−66191 to −110290.0025Fluzone®10,000/3738−6263 ± 4684−16309 to 37840.202680,000/19813−60188 ± 23860−111361 to −90140.0244Shingrix®143,750/15250−128500 ± 13251−156339 to −100661<0.000111,718,750/7,343,750−4375000 ± 3747105−12247376 to 3,497,3760.2582Mean Titers at Post-dose 1 (pd1) and Post-dose 2 (pd2) for both young and older, the difference ± SEM, 95% Confidence Interval and the two-tailed P-value using unpaired t-test via GraphPad version 10.2.2, where P < .05 is significantly different between the young and older cohort for each vaccine. Vaccine antibody determinations in young and aged mice. Mean Titers at Post-dose 1 (pd1) and Post-dose 2 (pd2) for both young and older, the difference ± SEM, 95% Confidence Interval and the two-tailed P-value using unpaired t-test via GraphPad version 10.2.2, where P < .05 is significantly different between the young and older cohort for each vaccine. Aged monkeys have a unique peripheral blood RNA-sequencing signature In addition to evaluating a VHR in mice, the phenotype was investigated in NHPs. From a human experimental medicine study (EM131), stronger hepatitis B vaccine responses were linked to pre-vaccination B-cell/memory B-cell signatures, whereas inflammatory, innate, and erythrocyte-related signals predicted weaker responses, enabling a moderately accurate baseline predictor.26 The transcriptomics‑based bioage signature that was built and validated was derived from approximately 1,240 individuals aged 15–94 to identify genes associated with biological age.27 The approaches were used by other investigators to identify age-related gene signatures.28,29 Since NHPs have played a critical role in medical research, especially vaccine discovery and development,30,31 we wanted to first determine whether they exhibited baseline signatures using data sets employed in the human studies, that could align with HBV vaccine-naive NHPs. Therefore, we profiled monkeys that were less than 2 y of age and older animals. For the latter group, the minimum age was 15 y old, the group mean was 23.0 y (median, 23.5 y), corresponding to approximately 69.0 human years using the 1:3 monkey-to-human age conversion. As shown in Figure 2, both AGMs and RHMs exhibited a composite bioage signature that can be demarcated when stratified by chronological age. The bioage signature is preserved in NHPs. Composite Bioage Score correlates with the age of both Rhesus and African green monkeys, similar to humans in EM131. Young animals (mean age 24 months) and aged animals older than 18 y of age (mean age 22 y) were used to collect whole blood, and isolated RNA was sequenced by EA Genomics, Morrisville, NC 27560. FASTQ files were annotated with OmicSoft (v10.1) [OmicSoft Corporation, Cary, NC, USA]. The resulting gene sets were converted to bioage signatures from the EM131 microarray dataset. Monkeys have a hyporesponse to anti‑HBSAg antibody (IgG) titers Rhesus macaques with various ages (n = 20 older and n = 21 younger) Table 2 were rerolled and immunized each with Twinrix®, a combination vaccine targeting hepatitis A and B, twice at a month interval Figure 3(A). Younger monkeys had a mean age of 4.9 y old and median of 5. The older cohort mean was 25.2 and median age of 24 which is roughly 75.6 human years using the 1:3 monkey-to-human age conversion. Blood was collected prior to immunization (baseline) and after each immunization at monthly intervals. The systemic anti-HBSAg responses were assessed, as well as cytokine and chemokines in both cohorts. The focus on the hepatitis B antigen was chosen because of the positive association made between bioage and vaccine response in the EM131 human study. Table 2.Assessing HBV vaccine’s effect on young and older AGMs.GroupAgenMales/femalesAntigen1Young2120/1Hep B2Old203/17Hep BRhesus (twenty-one young = mean 4.9 y old; twenty older = mean 25.2 y old) were immunized on Day 0 and again on Day 56 with HBV vaccine. Assessing HBV vaccine’s effect on young and older AGMs. Rhesus (twenty-one young = mean 4.9 y old; twenty older = mean 25.2 y old) were immunized on Day 0 and again on Day 56 with HBV vaccine. Young and aged rhesus monkeys exhibit hyporesponsiveness to the HBV vaccine that mirrors the outcomes observed in humans. (a) Following two immunizations with Twinrix® in rhesus macaques: (b) Anti‑HBsAg IgG antibody titers were measured on Day 28 (not shown) and Day 56 (shown). The mean OD450 for older animals was 0.6966 versus 1.571 for young animals (p = .0001 by t-test). The difference between means ± SEM was 0.8743 ± 0.1080 (95% CI, 0.6558 to 1.093). Responders by group: 17 young and 2 older monkeys were high responders; 4 young and 6 older mid-responders; 12 low responders were all older. (c) CBCs were performed at baseline to determine white (young higher) and red blood (young lower) cell counts (p = .0053 and .0005, respectively, by t-test). (d) Baseline concentrations of selected chemokines and cytokines were also measured. IL-5, MCP-1, and IL-15 were significantly increased in older animals (Welch’s unpaired two-tailed t-test p-value .0254, .0471, and .01, respectively). (Continued). (Continued). (Continued). (Continued). Anti-HBSAg IgG in the older animals (mean OD450 = 0.6966) was significantly lower (p-value = .0001; Unpaired t-test) than the younger animals (mean OD450 = 1.571) Figure 3(B). Additionally, lymphocytes were decreased in the elderly animals (p-value .0053), contrary to RBCs which showed an increase in the older animals (p-value .0005) Figure 3(C). Lastly, an evaluation of cytokines and chemokines at baseline showed Th2/eosinophil-activating cytokine IL-5, proinflammatory recruiter or monocytes and macrophages, chemokine MCP-1, and NK activator cytokine IL-15 were upregulated in the elderly macaques prior to the first immunization Figure 3(D). IL-5, MCP-1, and IL-15 were significantly increased in the older animals, at baseline compared to the younger cohort (p-value .0254, .0471, and .01, respectively, using unpaired t-test). In the initial NHP immunogenicity study, the young animals were represented by a majority of males (20 males and 1 female), and the aged monkeys were mostly females (3 males and 17 females). To access the impact on gender, broaden the ages, and interrogate recall responses, the monovalent tetanus vaccine as well as the HepA/B vaccine in a subsequent rhesus study were assessed, concomitantly. Conducting studies at geographically different sites, using animals of the same species but with potential intraspecific genetic variation, may introduce other confounding host variables. Monkeys were stratified into one year old or less, 2–3, 4–7, 10–16, and 19–26 y old for both male and female. Compared with the first NHP study, the HBsAg antibody responses in this study were not consistent between the young adult and advanced age animals. In males, there was significant difference for all age groups except 10–16-y-old males (p-values = .0241 for <1 vs 2–3 y olds; 0.0157 for <1 vs 4–7 y olds; and 0.0160 for <1 vs 19–26 y olds (2way ANOVA Tukey’s multiple comparisons test). However, there was no significant difference between any females Figure 4(A,B). When genders are combined, a significant difference between the youngest and oldest group is evident (p-value = .0201) Figure 4(C). HBsAg-naïve rhesus monkeys were grouped by sex (□ = males and ◇ = females), age, and weight and immunized twice, 1 month apart, in alternate arms with 1 mL of Twinrix® and tetanus toxoid. Anti-HBsAg IgG titers were measured on Day 56 (OD value at 1:1600 dilution), and anti-tetanus toxoid IgG1 levels were measured (Avg IgG1 Concentration (ug/mL) on Days 28 and 56. Because the colony receives a primary two-dose tetanus immunization series and boosters approximately every 5 y, both primary and recall responses to tetanus were assessed. Infants were evaluated after the primary series, whereas adults and elderly animals were assessed after two additional boosters. In addition, an opportunity was taken to assess recall memory responses using the monovalent tetanus vaccine. This cohort of animals was vaccinated as tetanus-naive infants and approximately every 5 y thereafter. Therefore, infants represented a primary or de novo response group, whereas adults and elderly animals had received multiple boosts at the time of experimental vaccination. Infant animals showed no response after the first vaccination (at or below the lower limit of detection) and measurable (mean IgG of 109 µg/ml) antibody after the second. In contrast, adults exhibited a higher response after the first experimental immunization (mean IgG of 70 µg/ml), which increased further after the second dose (mean IgG of 213 µg/ml). This nearly 2x antibody response over the infant levels indicates a robust recall response. Elderly animals also responded just above the lower limit of detection after first dose and a mean IgG of 78 µg/ml after boosting, which is nearly equivalent to younger adult monkeys after only a single vaccination. The lower magnitude in the older animals suggests that the preserved recall responses were attenuated with age Figure 4(D). The reduced IgG response observed in aged mice was replicated in rhesus macaques vaccinated with 50 µg of VZV gE formulated with AP (250 µg) or with Shingrix® using the human dose of AS01b. The 0.5 ml injection volume of Shingrix® contains 50 µg of MPLA + 50 µg of QS‑21 in liposomes. Monkeys were grouped as young with a mean age of 5.7 and median of 5.8 y old (min 5, max 6.1). Older animals were 20 to 29 y old with a mean and median of 22.7 and 22, respectively. After one dose, aged macaques given gE + AP showed a 3-fold lower gE-specific IgG titer than young macaques; after two doses, the difference increased to approximately 4-fold. As a group, the antibody response was not significant, but there were clearly non-responders in the aged animals. Additionally, as shown in Table 3 and represented in Figure 5, young animals immunized with Shingrix® had a mean gE‑specific IgG of 529 versus 40 with AP (mean diff. 488.8; 95% CI − 799.8 to −177.7; p = .0014), while in older animals, the means were 52 (AS01b) and 13 (AP), not significantly different (mean diff. −38.42; 95% CI − 349.5 to 272.7; p = .9995). After two doses, both adjuvants boosted responses from the initial immunization, but AS01b produced substantially higher titers: in young animals, the means were 8,468 (AS01b) versus 732 (AP) (mean diff. −7,736; 95% CI − 11,838 to −3,635; p = .0002), and in aged animals 5,100 (AS01b) versus 186 (AP) (mean diff. −4,914; 95% CI − 9,016 to −812.2; p = .0146). Thus, AS01b markedly increases gE‑specific IgG magnitude in aged macaques relative to AP and largely reduces the age‑associated deficit; young vs aged: ~2‑fold with AS01b versus ~4‑fold with AP. Groups were analyzed using Šídák’s 2-way ANOVA multiple comparisons using GraphPad Prism Version 10.2.2. Table 3.Antibody responses in young and old NHPs immunized with immunostimulatory adjuvant.DoseAge groupAdjuvantMean titerComparison (adjuvant)Mean difference (AS01b/AP)95% CI of diff.p-value1YoungAS01b529AS01b vs AP488.8−799.8 to −177.7.00141YoungAP40 1AgedAS01b52AS01b vs AP−38.42−349.5 to 272.7.9995 (n.s.)1AgedAP13 2YoungAS01b8,468AS01b vs AP−7,736−11,838 to −3,635.00022YoungAP732 2AgedAS01b5,100AS01b vs AP−4,914−9,016 to −812.2.01462AgedAP186 Antibody responses in young and old NHPs immunized with immunostimulatory adjuvant. Reversing a vaccine hyporesponse in monkeys. Five rhesus macaques per group were included. Sex ratios (male:female) were 3:2 for the young group, 4:1 for the AS01 older group, and 3:2 for the AdjuPhos® (AP) older group. The mean age was 5.7 y for young animals and 22.1 y for aged animals. Animals received VZV gE protein plus AP or Shingrix at days 0 and 35. Serum antibody was assessed at days 21 and 49 (2 weeks after the second immunization) to determine VZV gE-specific antibody titers by ELISA. EC50 titers are shown. Statistical significance between groups was assessed by two-way ANOVA with Tukey’s multiple comparisons test. Discussion Recent advances have improved the ability to interrogate and predict VHR in vivo.32,33 A range of intrinsic, extrinsic, environmental, behavioral, and vaccine-related factors contribute to the immune response elicited by immunization.34–38 Age-associated decline in immune function, collectively termed immunosenescence, is a characteristic feature that often manifests as VHR.39–47 Age itself is a predictor of compromised responses to vaccination.48,49 Reductions in antibody quantity and the associated diminished protection against infection after vaccination have been documented.50–52 Strategies to enhance vaccine responses include increasing antigen dose, employing more potent adjuvants, and modulating host inflammation.53,54 Still, elucidating the mechanisms of immunosenescence and developing strategies to overcome it will become increasingly important as global life expectancy continues to rise. Accordingly, robust systems will be essential to investigate the interplay between physiology and immune interactions and to test hypotheses aimed at improving vaccine efficacy and host responses. Experimental in vivo systems have enabled detailed examination of host responses to vaccines in both healthy and dysregulated states.55 Beyond assessing antibody titers, novel methods have been applied to assess vaccines from conception through development.56 High-throughput transcriptomics, which analyzes RNA transcripts differentially expressed under specific host conditions, has been revolutionary in uncovering the complex networks among cells, tissues, organs, accessory molecules, and signals within the immune system. This systems biology approach is beginning to reveal predictors for personalized vaccination57–61 and for aging, more broadly.62,63 Its aim is to identify biomarkers that distinguish responders and non-responders. In a human experimental study (EM131), Fourati and colleagues established a Bioage signature using microarray-based transcriptional profiling. RNA from study participants’ peripheral blood was compared, and gene clusters were assigned and categorized into 20 modules across chronological age. Within this framework, a 15‑gene bioage signature emerged as a predictor of HBV vaccine responsiveness in humans. To bridge the gap from bench to bedside, we sought to establish a reverse‑translational model that phenocopies human findings in preclinical animal models. To this end, we employed a stepwise strategy to identify peripheral blood biomarker(s) to predict VHR in nonclinical, in vivo models. First, we verified reduced antibody responses to vaccines in aged mice. We then evaluated the performance of commercially available vaccines in aged nonhuman primates. Finally, we tested whether these models could recapitulate real‑world and experimental outcomes. Age-related differences in murine immune responses have been previously reported.64–70 Nonetheless, we sought to independently reproduce these findings as an important control for lab‑to‑lab variability. After demonstrating that two vaccine modalities, subunit protein and inactivated virus, elicited reduced antibody responses in aged mice compared to younger cohorts, we next sought to restore these diminished antibody levels. Since adjuvants have been shown to restore diminished antibody responses in aged mice,71–74 the AS01b adjuvant previously shown to contribute to the effectiveness of a herpes zoster vaccine (HZV) in older adults75 was used to demonstrate rescue of a weakened antibody response in aged mice. Notably, the mice used were just over one year old and at or just prior to the reproductive senescence.76 Yet, the antibody responses remained markedly lower than those of the younger animals. After identifying and rescuing a VHR in mice Figure 1, age-related differences in monkeys were next evaluated. Consistent with the human experimental study, whole-blood RNA-seq profiling demonstrated a difference in composite bioage scores between young animals (mean age 24 months) and aged animals (mean age 23 y), roughly equivalent to a 69 ‑year‑old human Figure 2. These differences in baseline expression across the lifespan provided confidence that we could distinguish antibody titers, and possibly molecular signatures associated with vaccine responsiveness, across age groups. Fourati and colleagues showed that in hepatitis-naïve humans immunized with the HBV vaccine, high vaccine responders were associated with gene signatures that enhanced B-cell responses and memory B-cell frequencies, whereas low responders had higher levels of transcripts associated with inflammation and increased erythrocytes and heme-containing cells. Based on these findings, we chose to immunize monkeys with the HBV vaccine, Twinrix®. Using HBV‑naïve monkeys, anti-HBSAg IgG binding antibodies were significantly lower in older immunized monkeys (mean age 25 y; approximately 75 human years based on a 1:3 monkey-to-human age conversion) than in young monkeys Figure 3(B). Trying to ascribe a predictor to the VHR in NHPs was less conclusive. Known associations between inflammation, immune dysfunction, and aging,77–79 as well as previously reported expression signatures, consistent with aged humans,29,80–82 were at best, modest, and conflicting. Luminex analysis revealed increase in three out of twenty-three cytokines in aged animals. IL5, IL-15, and Monocyte chemoattractant protein-1 (MCP-1) were higher in the older rhesus macaques prior to the first immunization. MCP-1, IL-5, and IL-15 have been associated with inflammation; however, their roles in normal aging, inflamm-aging, and immunosenescence have various degrees of involvement. In both humans and animals, the literature is conflicting, with reports describing both increased and decreased levels of each of these markers, plus discordant with other studies for the other markers.83 Additionally, lymphocytes were decreased in older animals, contrary to RBCs Figure 3(C). Although older animals showed lower WBC counts, consistent with aging-related immune changes,84 the higher RBC counts observed in older NHPs remain unclear. Others have shown higher hematocrit and hemoglobin which may reflect cohort-specific physiological differences such sex and immune status or differences in hydration and stress.85 Beyond the composite bioage score, which correlates with chronological age in NHPs, and some cellular hallmarks, we also observed the lack of lactobacillus in the older animals using 16S and metagenomic sequencing on stool samples (data not shown). Because reduced antibody production in older animals was the clearest distinction from younger animals, the immunoglobulin biomarker was the most attractive target to interrogate. A significant difference in HBsAg antibody levels between young and older monkeys (p-value = .0001) Figure 3(B) was demonstrated. However, the extraordinary result is confounded by an unbalanced sex distribution. The older cohort was predominantly female while the younger cohort was mostly male. Males, both humans and rodents, often exhibit lower post‑immunization antibody responses than females for some vaccines86–93 which could contribute to the observed difference. The differences are not clear-cut, since some studies indicate that vaccine efficacy or immune responses are not higher in females and more nuanced depending on several other cofactors.94,95 To address the potential sex-specific bias, a follow-up study with a more balanced gender distribution across different age groups was designed. The study included HBV‑naïve animals spanning a broad age range, from infants to older adults. We enrolled animals that were <1, 2–7, 8–15, or >15 y of age. Animals were randomly assigned to a group based on their chronological age, gender, and body weight. Each group had five males and five females. The <1‑year‑old male rhesus elicited antibodies against HBV that were significantly elevated compared to 2–3-y-olds (adjusted p-value = .0241), 4–7-y-olds (adjuvated p-value = .057), and the oldest cohort of >18-y-olds (adjusted p-value of .0160). There were no significant differences for any female cohorts. When males and females were combined, the <1-y‑olds had significantly higher antibodies than the oldest animals (adjusted p-value = .0201). Since females tend to have higher antibody levels than males, the dramatically lower antibody levels observed in the older cohort of the initial NHP study, which was composed primarily of females, would likely have resulted in an even larger difference between the extreme age groups if sex had been controlled.96–98 Interestingly, the phenocopy hyporesponse in NHPs remains evident despite sex disparity. Another key advantage of our second study was the ability to examine age effects in animals previously primed with a tetanus vaccine. This cohort had been immunized with the monovalent tetanus vaccine starting in infancy, receiving two primary doses followed by booster doses every 5 y. Thus, the younger animals were receiving their primary vaccination series at the start of the study, while the older animals were receiving booster doses. For anti-tetanus titers, significant differences (Fisher’s exact chi-square test, p-value = <.0001) were observed among young, adult, and elderly groups. Despite continued boosting, the oldest animals still had lower titers than both infants and adult animals. The reasons why HBV antibody titers did not stratify by age as clearly in this study as they did in the first study are unclear. Also unsettled, but interesting, in humans vaccinated with the Hepatitis subunit vaccine (HBsAg adjuvanted with Al(OH) and AIP04), the age-related difference is only observed after primary vaccination,99 unlike what was observed in the rhesus macaques. NHPs, similar to mice,100–102 have been shown to exhibit a VHR that can be mitigated with adjuvants.103–106 Monkeys, which share close phylogenetic and immunological similarities to humans,107–109 are not surprisingly frequently utilized to evaluate vaccines.110 We previously demonstrated VHR associated with differential changes in gut bacterial composition25 and now also in the context of aging. Age-dependent differences in vaccine responses have been documented previously in rhesus.111 Accordingly, it is fitting to determine whether a clinically relevant adjuvant system can enhance the attenuated vaccine responses observed in aged NHPs. AS01 is an immunostimulatory adjuvant containing 3-O-desacyl-4’-monophosphoryl lipid A (MPLA) and the saponin QS-21. The adjuvant is used in licensed human subunit vaccines against malaria, herpes zoster, and RSV.112–115 This adjuvant is thought to activate the innate immune pathways through TLR4 signaling, a key receptor that is known to be functional in young and aged immune cells, resulting in NF-kB activation and the production of proinflammatory cytokines and chemokines.116,117 In addition, it transiently activates dendritic cells, thereby enhancing antigen uptake, processing, and presentation.118,119 Together, these effects may help compensate for the impaired immune surveillance associated with senescent cells by mitigating reduced antigen presentation, altered cytokine signaling, and weakened T-cell priming, while promoting Th1-biased adaptive responses and thereby enhancing overall adaptive immunity.120,121 While the efficacy of this adjuvant is supported by real-world evidence,122 its mechanism of action, as well as the ability of adjuvants with comparable immunostimulatory properties to produce similar effects, remains to be documented. Work by Kofler and colleagues reported that a MPLA‑formulated beta-amyloid vaccine did not overcome reduced antibody levels in older macaques (a mix of rhesus, cynomolgus, and pigtails monkeys), with more than 80% of aged animals producing delayed, weaker, and more variable titers compared with young animals.123 Other adjuvants, such as the oil-in-water emulsion adjuvants MF59 and CpG oligodeoxynucleotides (CpG ODNs), which are Toll‑like receptor 9 (TLR9) agonists, have demonstrated enhanced humoral immunity with various antigens in older adults.112 To investigate if we can rescue the VHR in NHPs, we immunized rhesus, twice, using VZV gE protein with and without AS01 adjuvant in both young (mean 5.7 y old) and older animals (mean 22.1 y old). Although overall antibody responses did not differ significantly between young and older animals immunized with the aluminum-containing vaccine, considerable individual variability was evident, with some older animals clearly exhibiting a non-responder phenotype. After two doses, both the aluminum-containing vaccine and the TLR4 agonist elicited boosted antibody responses; the increase induced by the latter was substantial and generally mitigated the age-associated deficit between the age groups. After two doses, antibody levels for the AS01-adjuvanted vaccine were generally comparable between young and old animals. However, responses to the AS01-containing vaccine in older animals were substantially higher than in animals immunized with vaccine containing AP. Specifically, AS01‑containing vaccines used to immunize older animals were roughly 9-fold higher than in young animals given AP and about 86-fold higher than in older animals given the AP-containing vaccine. These data suggest that NHPs phenocopy outcomes observed in aged humans when compared with conventional alum adjuvants.124 A key limitation of this work is that the primary endpoint relied on serum antibody measurements, chosen for its ubiquity and comparability across labs. The studies lacked orthogonal functional (e.g., neutralization, avidity) or in vivo protection readouts that may capture expanded age-related effects on infection125–128 and vaccine outcomes. Functional activity of antibodies, compared with binding antibodies, which include both functional and non-functional antibodies, can provide valuable information. However, these measurements are not directly comparable across the antigen panel evaluated in this study, because the antigens do not share a single standardized functional readout. Influenza-specific responses, for instance, can be assessed using hemagglutination inhibition or microneutralization assays, whereas OVA, a research antigen, does not have a validated human correlate of protection. Similarly, hepatitis B vaccine responses have historically been assessed using protective hepatitis B surface antibody levels of ≥10 mIU/mL measured by Enzyme-linked immunosorbent assay (ELISA) or chemiluminescent immunoassay (CLIA). However, even these long standing correlates are being challenged.129 Accordingly, including heterogeneous functional assays would not necessarily eliminate confounding results in this study, given differences in assay formats and the use of nonhuman species, which may not align with established correlates of protection. Binding IgG titers, a widely accepted and highly reproducible measure, were used as the primary readout of humoral immunogenicity. Assessing functional activity may be an important additional investment and will be considered as a future follow-up. Although multiomic datasets such as metagenomic, metabolic, and transcriptomic were collected, cross-study comparisons were limited because of long study periods and gaps between them. Harmonized, comprehensive multi-omics profiling across cohorts (including matched analyses in the gender-controlled and original studies) could enable discovery of microbiome or transcription-driven correlates and clarify sex-specific mechanisms. Also notable were environmental and husbandry differences between sites, including facility microbiota, diet, water, pathogen exposure, handling, which could contribute to differences across studies. The inconsistencies in the overall magnitude of the antibody response observed in the two NHP studies may also reflect these factors, as well as others such as hormonal status, immune history, assay variability, or differences associated with specific HLA types as noted in humans.130 Taken together, the data suggest that any sex-related effect on age-associated VHR is likely modest and context-dependent, and further studies will be needed to clarify this relationship. To our knowledge, this is the first study to show that an adjuvant with established human efficacy can rescue the humoral response of an attenuated vaccine due to age, in NHPs. This investigation of vaccine hyporesponsiveness across multiple antigens, adjuvants, and model systems is perhaps the most comprehensive preclinical evaluation to date. Modeling human-age phenotypes in nonclinical animals creates a platform for testing interventions for an aging population. These models can provide a platform for targeted immunomodulation investigation and, in the near term, provide an empirical system to evaluate novel systems designed to boost vaccine responses in the elderly. Acknowledgments We want to thank The Mannheimer Foundation, the New Iberia Research Center, and the MSD Veterinary and Care staff in Laboratory Animal Resources for their expert animal care and technical support. Additionally, we thank Dr. Pablo Morales and the talented Veterinary and Care staff at The Mannheimer Foundation, Inc. in Homestead, Florida, USA as well as Jane Fontenot and the veterinary, medical research, and animal care staff at the University of Louisiana at Lafayette New Iberia Research Center, New Iberia, LA, USA. We are grateful for the exceptional care provided to all the animals that improve animal and human life. Moreover, we thank the team at ImQuest BioSciences, Frederick, MD, USA, for evaluating the tetanus titers. We thank Nancy Hoyt for graphics and figure design. The authors also would like to thank Dan Freed for serology, as well as Drs. Lan Zhang, Amy Espeseth, and Daria Hazuda for their leadership. Biography Michael P. Citron is a Principal Scientist in Infectious Disease and Vaccines Discovery at Merck Research Laboratories. He is widely recognized for his expertise in nonclinical models that mimic human disease outcomes. His research has focused on paramyxoviruses, pneumoviruses, orthomyxoviruses, filoviruses, flaviviruses, HIV, and other viral and bacterial pathogens. His translational insights into disease mechanisms have enabled the development of prophylactic and therapeutic interventions. Funding Statement This research was fully funded by Merck Sharp & Dohme LLC., a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. Disclosure statement Authors may be shareholders and are current or former employees of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. Data availability statement Data is stored in electronic notebooks and files maintained by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ USA. Samples will be archived and then discarded according to policies of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ USA. Abbreviations micro, nanometer, revolutions per minute, Immunoglobulin G, Complete Blood Count, plus/minus, Ribonucleic acid, RNA sequencing, Quantitative Polymerase Chain Reaction, hour; *<.05, ** = <.01, *** = <.001, **** = <.0001 References Associated Data Data Availability Statement Data is stored in electronic notebooks and files maintained by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ USA. Samples will be archived and then discarded according to policies of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ USA.