(T1030-02-10) Modeling-Derived Immune Signature Explains Population-Scale Variability in Vaccine-Induced Immune Response

Purpose: Vaccines are pivotal in combating infectious diseases, saving millions of lives worldwide and vastly improving public health. Despite their success, the effectiveness of vaccines varies across individuals, posing a significant challenge in achieving community-wide protection. Several host factors such as pathogen pre-exposure, age, sex, genetics, and co-morbidities can influence the strength and longevity of the vaccine-induced immune response^1,2. This variability thus affects the population-scale effectiveness of vaccines and underlines the need for an approach to optimize the use of vaccines to improve outcomes. To tackle this challenge, we developed a mechanistic model that incorporates the key process of vaccine-induced immune response. Our aim is to elucidate the influence of various demographic factors on vaccine efficacy by rigorously quantifying immunological signature parameters across diverse cohorts. This approach is intended to unravel the complex interplay between individual demographic characteristics and their collective impact on immunological response to vaccination.

Methods: Based on our previous work³, we developed a model based on a system of ordinary differential equations to define the temporal evolution of the key immunological variables, including CD4+ T-cells, antibody-secreting plasmablasts, neutralizing antibodies, and cytokines (IL-2 and IFN-gamma) (Fig. 1A) . A closed form solution of the kinetics of the injected antigen governs the activation of CD4+ T-cells, which initiate the downstream immunological processes leading to antibody-induced immunity. We used extensive patient-level immunological data^4-10 (N ~ 3700 patients) from literature for COVID-19 vaccines in healthy adults to develop the model. Despite being healthy, these individuals exhibited significant variability in antibody response. We fitted our model to population-averaged immune response kinetics data (over 230 days) to estimate unknown model parameters (Fig. 1B-F). This was followed by global sensitivity analysis (GSA) to identify the key parameters influencing antibody response, using as an indicator of the immune signature shaping vaccine immune response, defined as Immunogenicity = (T_AT/T_RT)*(AUC_AT/AUC) where T_AT is the total time above the antibody threshold for protection¹¹,T_RT is the time to reach the threshold, AUC_AT is the area under the curve above the threshold, and AUC is the total area under the curve.

Results: The model fits were in good agreement with the data (Pearson correlation R > 0.9). Global Sensitivity Analysis (GSA) revealed that the duplication rate of plasma cells (P_PB), antigen sensitivity (K_Ant), death rate of plasma blasts (D_PB), and the production rate of antibodies per cell (P_Ab) were the four key parameters driving antibody responses (Fig. 1G). Individual patient-level fitting of the model for these four parameters provided a unique set of values for each subject. Using the AUC before the second vaccine dose as a sorting criterion, we observed significant influences of sex, age, and health status on the early immune response—females, younger individuals, and those in good health status exhibited more robust immune responses (Fig. 2A). We then examined the distribution of AUC before the second dose (Fig. 2B) and applied k-means clustering to categorize the population into high and low responders. Analysis of the demographic composition of these clusters revealed that these factors significantly influenced vaccine response outcomes (Fig. 2C-E). Notably, there were statistically significant differences in the four key parameters between high and low responders (Fig. 2F). These findings suggest that the variation in humoral response among demographically diverse individuals can be attributed to differences in these parameters, potentially forming a broader immune signature that shapes vaccine immunogenicity. These preliminary results underscore the importance of a more detailed and quantitative exploration of immune responses, highlighting the need for further research.

Conclusion: Our study presents a translational mechanistic model that elucidates the underlying variability in vaccine-induced immune responses. By identifying distinct immunological profiles within the population, we offer a novel lens through which vaccine efficacy can be optimized on a demographic basis through the optimization of dose, injection frequency, or vaccine combinations. This approach not only enhances our understanding of immune response dynamics but also holds the promise of refining vaccine distribution strategies to achieve more uniform community-wide immunity. Furthermore, the insights garnered from this research could be instrumental in streamlining the design of clinical trials or conducting in silico clinical trials, particularly in the rapid development of vaccines during pandemics when resource allocation and participant recruitment are constrained. Ultimately, our findings lay the groundwork for a more nuanced and effective deployment of vaccines, tailored to meet the specific needs of diverse population segments.

References: 1. Klein SL, Flanagan KL. Sex differences in immune responses. Nature Reviews Immunology. 2016/10/01 2016;16(10):626-638. doi:10.1038/nri.2016.90
2. Lynn DJ, Benson SC, Lynn MA, Pulendran B. Modulation of immune responses to vaccination by the microbiota: implications and potential mechanisms. Nature Reviews Immunology. 2022/01/01 2022;22(1):33-46. doi:10.1038/s41577-021-00554-7
3. Dogra P, Schiavone C, Wang Z, et al. A modeling-based approach to optimize COVID-19 vaccine dosing schedules for improved protection. JCI Insight. 05/25/ 2023;doi:10.1172/jci.insight.169860
4. Pérez-Alós L, Armenteros JJA, Madsen JR, et al. Modeling of waning immunity after SARS-CoV-2 vaccination and influencing factors. Nature Communications. 2022/03/28 2022;13(1):1614. doi:10.1038/s41467-022-29225-4
5. Figueiredo JC, Merin NM, Hamid O, et al. Longitudinal SARS-CoV-2 mRNA Vaccine-Induced Humoral Immune Responses in Patients with Cancer. Cancer Res. Dec 15 2021;81(24):6273-6280. doi:10.1158/0008-5472.Can-21-3554
6. Ebinger JE, Fert-Bober J, Printsev I, et al. Antibody responses to the BNT162b2 mRNA vaccine in individuals previously infected with SARS-CoV-2. Nature medicine. 2021;27(6):981-984.
7. Oyebanji OA, Wilson B, Keresztesy D, et al. Does a lack of vaccine side effects correlate with reduced BNT162b2 mRNA vaccine response among healthcare workers and nursing home residents? Aging Clin Exp Res. Nov 2021;33(11):3151-3160. doi:10.1007/s40520-021-01987-9
8. Collier A-rY, Yu J, McMahan K, et al. Differential kinetics of immune responses elicited by Covid-19 vaccines. New England Journal of Medicine. 2021;385(21):2010-2012.
9. Chang A, Akhtar A, Linderman SL, et al. Humoral responses against SARS-CoV-2 and variants of concern after mRNA vaccines in patients with non-Hodgkin lymphoma and chronic lymphocytic leukemia. Journal of Clinical Oncology. 2022;40(26):3020-3031.
10. Chang A, Akhtar A, Lai L, et al. Antibody Binding and Neutralization of Live SARS-CoV-2 Variants Including BA.4/5 Following Booster Vaccination of Patients with B-cell Malignancies. Cancer Research Communications. 2022;2(12):1684-1692. doi:10.1158/2767-9764.CRC-22-0471
11. Khoury DS, Schlub TE, Cromer D, et al. Correlates of protection, thresholds of protection, and immunobridging among persons with SARS-CoV-2 infection. Emerging Infectious Diseases. 2023;29(2):381.


Figure 1 A) Model description with key mechanism indicated by dotted lines for stimulation and continuous lines for production B-F) Results from the fit process, red line is the model prediction while the blue band is the confidence interval calculated. G) GSA of the model using immunogenicity as a variable, the insert is the local sensitivity analysis (LSA) of the top 10 parameters.


Figure 2 Analysis of the data obtained from the individual fit. A) Comparative analysis of the normalized median response profiles across different categories using an array of metrics. The metrics on the x-axis are in order: Antibody at day 4, Antibody at day 7, Plasma Blast at day 4, Plasma Blasts at day 7, Antibody AUC before the second dose, Antibody peak before the second dose, Antibody AUC over 720 days, Time spent above the protection threshold, Time needed to reach the threshold, Antibody AUC above the threshold, Antibody peak value over 720 days, Time to reach antibody peak, Immunogenicity,P_PB K_Ant, D_PB and P_Ab . The best performing individual are healthy and younger. B) The distribution of the log10 values of the antibody area under the curve (AUC) prior to the second dose is depicted, segmented into two clusters representing low and high responses. C-E) Analysis of the two clusters in function of Sex, Age and Health status. F) Violin plot of the parameters for the two clusters.