Viral correlates of protection and the multifarious interactions of B and T cells Authors: H Cody Meissner, Peter B Gilbert, Derek Eisnor, Daniel N Wolfe Published: 2025 ABSTRACT ABSTRACTValidated, reproducible methods to assess immunity to a specific virus following either infection or vaccination are important for understanding susceptibility to existing and emerging infectious diseases. Correlates of protection for a specific viral illness involve humoral and cellular responses that differ by pathogen and by the portal of entry into the host. Contemporary understanding of a correlate of protection primarily is based on serum neutralizing antibody concentrations because of difficulty in assaying the complex interactions of T cells. A more comprehensive understanding of susceptibility and protection is needed. The goal of this paper is to enumerate the multiple variables that complicate understanding of cellular contributions to the innate and adaptive immune responses.KEYWORDS: Correlates of protection, B cells, T cells, innate immunity, adaptive immunity, vaccines Numerous factors in the innate and the adaptive immune response including those in the humoral and cellular paradigms contribute to a host’s response to an invading virus. Determination of a readily measured metric that identifies a protective immune response generally is attempted during vaccine development trials and is referred to as a correlate of protection (CoP).1–3 A CoP identifies a gradient of an immune biomarker meaning that at higher concentrations a specific disease outcome is less likely to occur and at lower concentrations the risk of infection and severe disease increases.3 Often this biomarker is assessed in a vaccine efficacy trial by comparison of the immune response (such as binding or neutralizing antibody concentrations) among unprotected participants who experience a specific outcome versus participants who do not acquire the outcome.4–6 Currently understood CoPs generally are based on serum IgG concentrations but seldom are absolute, especially for viruses that enter the host through the respiratory tract and do not involve a viremic phase. As described by Plotkin and Gilbert, a CoP can be defined as either a mechanistic CoP (mCoP) or a non-mechanistic CoP (nCoP)3 (Table 1). A mCoP involves bioassays that evaluate immune functions for protection and are directly involved with viral inactivation. A nCoP, identifies components of the immune response that correlate statistically with a specific endpoint, but may not have a direct relationship with antiviral function and are less predictive of susceptibility.3 Determination of the cellular contribution to a CoP is complicated by the multiple functions involved in the immune response that ensure redundancy, indicating more than one component of the immune response may define a CoP for a specific virus.7 Table 1.Definitions of correlate of protection (from Gilbert and Plotkin3).Correlate of Protection (CoP)An immune response (biomarker) that is statistically correlated with protection; a biomarker measured at a specific time after vaccination that can be used to reliably infer vaccine efficacy against a specific endpoint; a CoP may be either mechanistic or non-mechanisticMechanistic Correlate of Protection (mCoP)An immune response that is directly responsible for protectionNon-mechanistic Correlate of Protection (nCoP)An immune response that substitutes for the true mechanistic correlate of protection when the mCoP is either unknown or not easily measurable but still can be convenient because it is statistically associated with protectionAbsolute CorrelateA specific level of response highly correlated with protection; a threshold above which protection is nearly certain Definitions of correlate of protection (from Gilbert and Plotkin3). Knowledge of a CoP is useful throughout the life cycle of vaccine development for reasons noted in Table 2. Because immunity peaks and wanes, a CoP often is defined at a specific time interval after infection or a vaccination series for the endpoint being evaluated. Exposure to a high viral inoculum may overcome a threshold of what otherwise is considered to be a CoP. The relative importance of humoral versus cellular immunity may be influenced by the duration of the incubation period. For example, the long incubation period of hepatitis B virus (range 6–20 weeks) enables recruitment of protective memory cellular and humoral immunity in an immunized person even though circulating antibody to HBsAg may not be detectable at the time of infection. The short incubation period of influenza virus (range 1–4 d) places emphasis on existing humoral immunity for the prevention of infection, although the role of T cells remains incompletely understood. Protection against a specific endpoint may not be the same immune response that determines recovery from infection.8 Table 2.Potential applications of CoP during vaccine development.1.To guide vaccine development for a specific antigen2.To facilitate dosing strategies throughout clinical development and marketing activities3.To demonstrate consistency and potency of vaccine lots for regulatory requirements4.To define the timing for booster doses after waning immunity5.To define an end point for vaccine expiration6.To evaluate second generation vaccines through immunobridging studies7.To assist with development of experimental vaccines using serocomparator studies rather than efficacy trials8.To determine vaccine efficacy among populations not included in early clinical trials (such as autoimmune or immunocompromised patients, pregnant, pediatric or elderly populations)9.To serve as an endpoint for licensing an experimental vaccine when randomized clinical trials are not possible because of low rates of disease or human challenge studies are not possible such as for agents of biological-threats10.To determine possible interference when two or more vaccines are co-administered11.To predict the risk of vaccine enhanced disease with an imbalance in Th1/Th2 ratios from an experimental vaccine12.To identify an adaptive immune profile that may be associated with risk for post-infectious, long-term complications Potential applications of CoP during vaccine development. Mutations in viral surface proteins responsible for inducing humoral immunity may render a previously protective immune response less effective. The mumps, measles and influenza viruses represent examples of viral antigen mutation and their different effects on immunity. The first mumps vaccine was licensed in 1967 based on the attenuated genotype A (Jeryl-Lynn strain) which still is used in today’s measles-mumps-rubella vaccine. Currently, a different mumps strain, genotype G, has become the dominant mumps virus genotype worldwide, at least partly explaining the lower efficacy of the mumps vaccine.9 In contrast to mumps virus, the measles virus has undergone antigenic mutation in protein epitopes that are not important for stimulating a protective immune response, enabling vaccine induced immunity to persist. The measles virus evolved into multiple distinct lines of wild type virus, resulting in 8 clades containing 24 genotypes based on nucleotide sequencing of the hemagglutinin and nucleoprotein genes. In spite of this evolution, conserved neutralization epitopes in the hemagglutinin protein (across genotypes) that are critical for binding to cellular receptors have resulted in continued efficacy of the initial measles vaccine, >60 y after licensure.10 The heterogenous characteristics of the T cell response including the polymorphic nature of the HLA locus that results in T cell receptor diversity contributes to difficulty in establishing a metric to define a protective T cell response. Variation in the quality and quantity of an individual’s immune response to a vaccine is modified by intrinsic factors including genetics, sex and age (older individuals have a smaller naïve T cell pool). Evidence suggests that the baseline status of the immune system prior to vaccination influences the characteristics of the immune response.11 For viruses that infect the respiratory mucosa such as influenza virus or respiratory syncytial virus, a waning serum antibody concentration alone does not consistently predict susceptibility to infection indicating that cellular immunity may work additively or synergistically with non-neutralizing antibodies to compensate for the lack of an effective neutralizing antibody response.12 Influenza viruses typically undergo genetic evolution in a constant but unpredictable pattern to improve viral fitness (the ability to spread despite elevated levels of population immunity.) Antigenic changes in the two major surface proteins (hemagglutination, neuraminidase) due to point mutations or reassortment of RNA segments results in variable degrees of mismatch between existing serum neutralizing antibody from previous infection or vaccination. Serum hemagglutination inhibition antibody concentration ≥1:40 (to the globular head of the hemagglutinin protein) equates to ~50% protection against symptomatic influenza infection.13 The coordination of preexisting cellular immunity, antibody responses, and innate immunity to facilitate influenza viral clearance after re-exposure to an antigenically distinct strain differs between vaccinated and unvaccinated adults. Evidence suggests that baseline antibody levels as well as cellular correlates, including natural killer cell populations, enhance prediction of susceptibility to influenza infection.14 Because of the complex response of T cell phenotypes to a coronavirus infection, much of the focus in vaccine development during the recent pandemic has been on neutralizing antibodies. Following SARS-CoV-2 immunization, the greater the concentrations of anti-spike IgG to the receptor binding domain and pseudovirus neutralizing antibodies shortly after a two-dose vaccine regimen, the greater the protection against symptomatic illness, identifying a CoP for related variants for a limited period of time.5 By several months after vaccination against COVID-19, neutralizing antibody levels decline and protection against less severe disease endpoints such as milder disease or emergency department visits for symptomatic infection wanes.15,16 In contrast, protection against severe disease endpoints, such as hospitalization, tends to continue indicating non-humoral based correlates (T cells, learned innate immunity) persist due to epitopes recognized by T cells that are less impacted by mutations occurring in surface proteins. These observations suggest that while neutralizing antibody concentrations may define one CoP, a metric that includes the contribution of T cells will enhance the accuracy of this metric against mild illness as well as more severe disease. The limited role of antibodies in control of viral replication is consistent with results from clinical trials with convalescent serum or passively administered neutralizing monoclonal antibody for therapy of hospitalized, actively SARS-CoV-2 infected patients.17,18 In addition, reports of SARS-CoV-2 infected patients with agammaglobulinemia and absent circulating B cells or patients with pharmacologic B cell depletion therapy do not experience disproportional rates of severe disease. T cells retain a capacity to identify distinct, conserved epitopes on the highly expressed, large surface spike trimeric glycoprotein (~1300 amino acids resulting in numerous epitopes) as well as on other viral encoded proteins such as nucleocapsid, matrix and ORF3a proteins that are not primarily located on the viral surface, are less likely to induce neutralizing antibodies and thus are not as driven to escape mutations as epitopes targeted by neutralizing antibodies.19–21 Identification of T cell-based vaccine responses requires collection of peripheral blood mononuclear cells with appropriate processing and storage of specimens in small volumes.19 In vitro assays for cell stimulation require either fresh blood specimens or viable cryopreserved cells able to produce cytokines or surface activation markers upon stimulation. Variability of T cell analyses can arise from variable methods of sample collection and logistics and the assays themselves. Efforts to simplify the process from sample acquisition through sample analysis will facilitate future evaluation of vaccine programs and is a critical area for development. Adding further complexity, the role of individual immune components in protecting against a viral infection depend on the anatomic location under consideration (e.g., circulation, tissue, mucosal surface of the upper or lower respiratory tract) as well as the endpoint being studied (blockage of acquisition of a viral infection, asymptomatic infection with shedding, symptomatic infection, or severe illness). The local immune response at the portal of viral entry (nasal passage, oropharyngeal cavity including the tonsils and salivary glands, lungs) is unlikely to be fully reflected in blood samples and may require assessment of specimens of saliva, nasal wash or bronchial lavage. Furthermore, the cellular and humoral immune responses generated by antigens in a vaccine are unlikely to present viral epitopes that elicit the same systemic and mucosal immune responses as a natural infection. In addition, different sites of vaccine administration (i.e., intramuscular versus respiratory mucosal or intradermal) are likely to elicit variable immune responses. Finally, a CoP may not be the same in an infection-naïve person as in a person with immunity from a previous infection or immunization who experiences an anamnestic response. This was demonstrated in the different immune responses to dengue virus vaccination with a tetravalent vaccine based on a yellow fever vaccine 17D backbone among seropositive and seronegative individuals which was associated with different clinical outcomes.22 Results from the COVID-19 Variant Immunologic Landscape (COVAIL) trial demonstrated differences in neutralizing antibody concentrations (a CoP) after mRNA vaccine boosting between infection naïve-, and non-naïve individuals. This was at least partly explained by the higher neutralizing antibody titer and more durable protection among non-naïve participants.23 Vaccines consisting of replication deficient viral vectors such as simian adenoviruses or less commonly encountered human adenovirus serotypes demonstrate potent T cell priming responses to recombinant antigens.24 Recombinant, replication-deficient poxviruses such as the attenuated vaccinia Ankara vector vaccine containing viral gene inserts have demonstrated effectiveness in generating or boosting a cellular response to a recombinant antigen. As the threat of a future pandemic continues, perhaps with a greater case fatality rate than occurred with the recent coronavirus pandemic, the development of safe and effective vaccines as a medical countermeasure is paramount, especially if a paucity of antiviral treatment continues. Large scale, randomized vaccine trials for demonstration of vaccine safety and efficacy will be critical to the development of future vaccines and the preparation of clinical practice guidelines to establish an optimal standard of care. Validated, standardized immunogenicity assays that provide a more thorough understanding of the host’s immune response encompassing both humoral and cellular responses following vaccination or infection are needed. This understanding will enable a more complete assessment of the role of T cells in the adaptive immune response but will require improved clinical sampling techniques and the development and validation of assays to measure cellular contributions. The rapidly evolving capacity of artificial intelligence and machine learning to identify patterns from analysis of large amounts of data is likely to enhance understanding of the complex interactions of the numerous components of the human immune response to a foreign antigen. This will enable the design of improved vaccines, permitting data-driven clinical guidelines and potentially, offer a pivot to other viral species in the same family when an unexpected, novel virus with pandemic potential emerges. Acknowledgments This project has been supported in whole or in part with federal funds from the Department of Health and Human Services. This work was authored as part of the Contributor’s official duties as an employee of the United States Government. The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Department of Health and Human Services or its components. Biography H. Cody Meissner is an Independent Contractor for Tunnell Government Services and serves as a Senior Vaccine & Biologic Development Analyst at BARDA/ASPR in the Department of Health and Human Services (HHS). He was a member of the Advisory Committee on Immunization Practices (ACIP) from 2014 to 2018 and was re-named as an ACIP voting member in May 2025. From 2018 to 2024, he was a member of the Vaccines and Related Biologics Products Advisory Committee (VRBPAC) for the FDA. He was Vice-Chair and then ex-officio member of the Committee on Infectious Diseases (COID) for the American Academy of Pediatrics for 20 y and an Associate Editor of the Red Book for several editions. He was a member of the National Vaccine Advisory Committee (NVAC) from 2017 to 2022. From 2018 to 2020, he was Chair of the Vaccine Injury Compensation Program for HHS. Dr Meissner is a Professor in the Section of Infectious Diseases and International Health at the Geisel School of Medicine at Dartmouth. Funding Statement The work was supported by the Biomedical Advanced Research and Development Authority. Disclosure statement No potential conflict of interest was reported by the author(s). Ethics statement The request for an ethics statement does not apply as no human or animal subjects were involved with this report References