Infant Antibody Repertoires during the First Two Years of Influenza Vaccination

Abstract

The first encounter with influenza virus biases later immune responses. This "immune imprinting," formerly from infection within a few years of birth, is in the United States now largely from immunization with a quadrivalent, split vaccine (IIV4 [quadrivalent inactivated influenza vaccine]). In a pilot study of IIV4 imprinting, we used single-cell cultures, next-generation sequencing, and plasma antibody proteomics to characterize the primary antibody responses to influenza in two infants during their first 2 years of seasonal influenza vaccination. One infant, who received only a single vaccination in year 1, contracted an influenza B virus (IBV) infection between the 2 years, allowing us to compare imprinting by infection and vaccination. That infant had a shift in hemagglutinin (HA)-reactive B cell specificity from largely influenza A virus (IAV) specific in year 1 to IBV specific in year 2, both before and after the year 2 vaccination. HA-reactive B cells from the other infant maintained a more evenly distributed specificity. In year 2, class-switched HA-specific B cell IGHV somatic hypermutation (SHM) levels reached the average levels seen in adults. The HA-reactive plasma antibody repertoires of both infants comprised a relatively small number of antibody clonotypes, with one or two very abundant clonotypes. Thus, after the year 2 boost, both infants had overall B cell profiles that resembled those of adult controls. IMPORTANCE Influenza virus is a moving target for the immune system. Variants emerge that escape protection from antibodies elicited by a previously circulating variant ("antigenic drift"). The immune system usually responds to a drifted influenza virus by mutating existing antibodies rather than by producing entirely new ones. Thus, immune memory of the earliest influenza virus exposure has a major influence on later responses to infection or vaccination ("immune imprinting"). In the many studies of influenza immunity in adult subjects, imprinting has been from an early infection, since only in the past 2 decades have infants received influenza immunizations. The work reported in this paper is a pilot study of imprinting by the flu vaccine in two infants, who received the vaccine before experiencing an influenza virus infection. The results suggest that a quadrivalent (four-subtype) vaccine may provide an immune imprint less dominated by one subtype than does a monovalent infection.

Keywords: B cell memory; circulating antibodies; immune imprinting; influenza virus; viral immunity.

FIG 1

Study design and sample description. (A) Immunization and influenza infection history for the two infants enrolled in the study. Blood drop symbols indicate the blood draws analyzed in the study. HBV, hepatitis B vaccine; DTaP, diphtheria, tetanus, and pertussis vaccine; PCV, pneumococcal vaccine; IPV, polio vaccine; Hib, Haemophilus influenzae type B vaccine; Rota, rotavirus vaccine; IIV4, quadrivalent inactivated influenza vaccine; HAV, hepatitis A virus vaccine; MMR, measles, mumps, and rubella vaccine; VZV, varicella-zoster virus vaccine. (B) Summary of B cells analyzed in this study. (C) Plasma IgG binding titers against HAs included in the IIV4. Values indicate ED₅₀s from ELISA. H1 CA, A/California/2009 X181; H3 HK, A/Hong Kong/2014 X263B; B Bris, B/Brisbane/60/2008; B Phu, B/Phuket/3073/2013; LOQ, limit of quantitation.

FIG 2

Total B cell repertoire. (A and B) The IGHV SHM levels of non–class-switched (A) and class-switched (B) total postvaccination B cells. Median values are represented by horizontal lines, and mean values are included above each data set. Statistical significance was determined by Mann-Whitney U test (**, P < 0.01). (C and D) IGHV-IGHJ usage of postvaccination B cells across 2 years for infant 1 (C) and infant 2 (D). IGHV with >1% abundance in either infant was included in the data set.

FIG 3

Features of HA-specific Bmem. (A and B) Binding specificities of HA-specific non–class-switched (A) and class-switched (B) Bmem. Doughnut plots represent percentages of Bmem that were IAV specific, IBV specific, or cross-reactive to both (IAV+IBV). Numbers indicate sample size (number of HA-reactive Bmem) at each time point. (C and D) IGHV SHM levels of HA-specific non–class-switched (C) and class-switched (D) Bmem. (E and F) Avidity index values of HA-specific non–class-switched (E) and class-switched (F) Bmem. For panels C to F, median values are represented by horizontal lines, and mean (C and D) or geometric mean (E and F) values are included above each data set. Two adult subjects studied previously are designated KEL01 (male, age 39) and KEL03 (female, age 39). Red, blue, and purple dots represent IAV-specific, IBV-specific, and IAV+IBV Bmem, respectively.

FIG 4

Summary of IAV and IBV HA-reactive plasma IgG repertoire. (A and B) Compositions of IAV and IBV HA-reactive plasma IgG antibody repertoires of infant 1 (A) and infant 2 (B). Each bar represents a unique antibody clonotype, and their heights correspond to the relative abundance of each clonotype in the IAV or IBV HA-reactive IgG repertoire.