H3N2 influenza infection elicits more cross-reactive and less clonally expanded anti-hemagglutinin antibodies than influenza vaccination

Abstract

Background: During the recent H1N1 influenza pandemic, excess morbidity and mortality was seen in young but not older adults suggesting that prior infection with influenza strains may have protected older subjects. In contrast, a history of recent seasonal trivalent vaccine in younger adults was not associated with protection.

Methods and findings: To study hemagglutinin (HA) antibody responses in influenza immunization and infection, we have studied the day 7 plasma cell repertoires of subjects immunized with seasonal trivalent inactivated influenza vaccine (TIV) and compared them to the plasma cell repertoires of subjects experimentally infected (EI) with influenza H3N2 A/Wisconsin/67/2005. The majority of circulating plasma cells after TIV produced influenza-specific antibodies, while most plasma cells after EI produced antibodies that did not react with influenza HA. While anti-HA antibodies from TIV subjects were primarily reactive with single or few HA strains, anti-HA antibodies from EI subjects were isolated that reacted with multiple HA strains. Plasma cell-derived anti-HA antibodies from TIV subjects showed more evidence of clonal expansion compared with antibodies from EI subjects. From an H3N2-infected subject, we isolated a 4-member clonal lineage of broadly cross-reactive antibodies that bound to multiple HA subtypes and neutralized both H1N1 and H3N2 viruses. This broad reactivity was not detected in post-infection plasma suggesting this broadly reactive clonal lineage was not immunodominant in this subject.

Conclusion: The presence of broadly reactive subdominant antibody responses in some EI subjects suggests that improved vaccine designs that make broadly reactive antibody responses immunodominant could protect against novel influenza strains.

Figure 1. Characterization of peripheral blood plasmacytosis 7 days after TIV or EI.

A. Peripheral blood B cells (CD3/14/16/235a⁻ CD19⁺) stained for plasma cell markers (CD3/14/16/235a⁻ CD19⁺ CD20^−/lo CD27^hi CD38^hi); points shown are percentage of B cells that were plasma cells. TIV mean 2.75%±0.90%; EI mean 2.26±0.74%; two-tailed t test, p = 0.68. B. Human rmAbs derived from sorted plasma cells tested for reactivity. From TIV subjects, 252/404 rmAbs (62.4%) reacted with ≥1 influenza antigen (blue wedge), 152/404 (37.6%) did not react with any tested rHA or split-virus antigen (gray wedge). From EI subjects, 37/451 rmAbs (8.2%) reacted with ≥1 influenza antigen (tested vs. TIV, χ² = 279.5, p<0.0001), 414/451 (91.8%) did not react with any tested rHA or split-virus antigen. C. Human rmAbs membership in clonal lineages. From TIV subjects, 175/404 (43.3%) rmAbs were members of 46 clonal lineages (red wedge); from EI subjects, 28/451 (6.2%) rmAbs were members of 12 clonal lineages (χ² = 162.1, p<0.0001).

Figure 2. Characterization of influenza-specific mAbs from TIV or EI subjects.

A. Clonal lineages. From TIV subjects, 159/252 (63.1%) of influenza-specific rmAbs were members of 44 clonal lineages (purple wedge); from EI subjects, 6/37 (16.2%) were members of 2 clonal lineages (χ² = 28.9, p<0.0001). B, Multiple reactivity to influenza antigens. From TIV subjects, 159/252 (63%) of rmAbs were strain-specific (blue wedge); multiply reactive rmAbs were less common [two antigens 79/252 (31.4%) (yellow wedge); three antigens 11/252 (4.4%) (orange wedge); four antigens 2/252 (0.8%) (red wedge); five antigens 1/252 (0.4%) (black wedge)]. From EI subjects, 16/37 (43.2%) were strain-specific (χ² = 7.74, p = 0.0054); multiply reactive mAbs were more common [two antigens 8/37 (21.6%); three antigens 8/37 (21.6%); four antigens 3/37 (8.1%); five antigens 1/37 (2.7%); six antigens 1/37 (2.7%) (white wedge)]. C. Original antigenic sin rmAbs. From TIV subjects, 5/252 (2%) of influenza-specific rmAbs did not react with strains contained in the administered vaccine but only with previously circulating influenza antigens (striped wedge). From EI subjects, 7/37 (19%) of influenza-specific rmAbs did not react with the infecting strain but only with previously circulating antigens (χ² = 19.2, p<0.0001).

Figure 3. Clonal lineage 2569 from EI13.

Three of 4 members (75%) derived from IgM-expressing plasma cells, 1/4 (25%) derived from an IgA1 plasma cell. The highest affinity binding for all members was to H3 Jobg; high affinity binding to other H3 rHAs and H1 Bris was also observed. Three members were tested for HAI and neutralization and displayed similar breadth (Table 3). H1 SI = H1N1 A/Solomon Islands/03/2006; H1 Bris = H1N1 A/Brisbane/59/2007; H1 Cal = H1N1 A/California/04/2009; H3 Wisc = H3N2 A/Wisconsin/67/2005; H3 Bris = H3N2 A/Brisbane/10/2007; H3 Jobg = H3N2 A/Johannesburg/33/1994; H5 Indo = H5N1 A/Indonesia/05/2005; H5 Viet = H5N1 A/Vietnam/1203/2004.

Figure 4. Clonal lineage 641 from TIV01.

During screening, 15/18 rmAbs (83%) bound one influenza antigen (blue dots), 1/18 (6%) bound two antigens (yellow dot), 2/18 (11%) bound no antigen tested (crossed dots). Antibody 1270 bound rHA H1 SI with high affinity and H1 Bris with weak affinity. Two branches of the tree derived from IgA1-expressing plasma cells. Inferred intermediates (int) of one of these branches were consistent with affinity maturation (arrows pointing to circles on the tree indicate the position produced int rmAbs); int #15 bound with lower affinity to H1 SI than later int #3 or int #5. Branches of the tree near the bottom showed breadth. Int #10 bound only H1 SI; int #9 had higher affinity for H1 SI, bound H1 Bris and H1 Cal with moderate affinity, and weakly bound H3 Wisc. Int #11 bound H1 Cal more weakly and recovered rmAbs 676 and 1261 bound with a similar pattern. Recovered rmAb 2258 had the highest H1 SI affinity in this part of the lineage but lost cross-reactivity, retaining only weak reactivity to H1 Cal. Embedded tables show affinity measurements in nM for each rmAb; NB = no binding observed. H1 SI = H1N1 A/Solomon Islands/03/2006; H1 Bris = H1N1 A/Brisbane/59/2007; H1 Cal = H1N1 A/California/04/2009; H3 Wisc = H3N2 A/Wisconsin/67/2005; H3 Bris = H3N2 A/Brisbane/10/2007; H3 Jobg = H3N2 A/Johannesburg/33/1994.