Fc-optimized antibodies elicit CD8 immunity to viral respiratory infection

Abstract

Antibodies against viral pathogens represent promising therapeutic agents for the control of infection, and their antiviral efficacy has been shown to require the coordinated function of both the Fab and Fc domains¹. The Fc domain engages a wide spectrum of receptors on discrete cells of the immune system to trigger the clearance of viruses and subsequent killing of infected cells^1-4. Here we report that Fc engineering of anti-influenza IgG monoclonal antibodies for selective binding to the activating Fcγ receptor FcγRIIa results in enhanced ability to prevent or treat lethal viral respiratory infection in mice, with increased maturation of dendritic cells and the induction of protective CD8⁺ T cell responses. These findings highlight the capacity for IgG antibodies to induce protective adaptive immunity to viral infection when they selectively activate a dendritic cell and T cell pathway, with important implications for the development of therapeutic antibodies with improved antiviral efficacy against viral respiratory pathogens.

Fig. 1. Anti-HA stalk monoclonal antibodies engineered for increased FcγRIIa affinity exhibit improved protective activity.

a, Influenza virus antigens (HA and NA) and the names of the tested monoclonal antibodies. b, Fc variants with differential FcγR binding affinity were generated for anti-influenza monoclonal antibodies. WT, wild type. c, d, Fc variants of anti-HA stalk monoclonal antibodies FI6v3 (c) and FY1 (d) were administered intraperitoneally (4 mg kg⁻¹ for FI6v3 and 2 mg kg⁻¹ for FY1) to FcγR-humanized mice before challenge with influenza (H1N1 PR8) (n = 6 mice per group for PBS-treated; n = 10 for WT FI6v3, GA, ALIE, GRLR and WT FY1; n = 9 for FI6v3 GAALIE and FY1 V11; and n = 8 for FY1 GA, afucosylated (Afuc), and GAALIE in two independent experiments). Weight loss (left; mean ± s.e.m.) and survival curves (right) were compared to the corresponding wild-type human IgG1 antibody-treated group by two-way ANOVA (Bonferroni post hoc analysis adjusted for multiple comparisons) (c: **P = 0.0072, #P = 0.0029, *P = 0.014, ***P = 0.0035) and log-rank (Mantel–Cox) test, respectively (c: *P = 0.019654; **P = 0.006759; ***P = 0.000303; d: *P = 0.0494, **P = 0.0045). NS, not significant. Source data

Fig. 2. Anti-HA globular head and anti-NA antibodies depend on FcγRIIa to confer protective activity in vivo.

a–c, Fc variants with differential FcγR affinity (Fig. 1b) were generated for the anti-HA head monoclonal antibodies 4G05 (a) and 1A01 (b), and the anti-NA monoclonal antibody 3C05 (c). FcγR-humanized mice were administered intravenously with Fc domain variants of 4G05 (0.5 mg kg⁻¹) (n = 10 mice per group for antibody-treated groups; n = 5 for PBS group in two independent experiments) (a), 1A01 (2 mg kg⁻¹) (n = 8 mice per group for WT, GA, ALIE and GAALIE groups; n = 6 for GRLR; n = 4 for PBS in two independent experiments) (b), and 3C05 (20 mg kg⁻¹) (n = 12 mice per group for WT and GA groups; n = 10 for GRLR and ALIE; n = 9 for GAALIE; n = 6 for PBS in two independent experiments) (c) before lethal challenge with influenza (H1N1 Neth/09). Weight loss (left; mean ± s.e.m.) and survival curves (right) were compared to the corresponding wild-type unmodified IgG1-treated group by two-way ANOVA (Bonferroni post hoc analysis adjusted for multiple comparisons) (a: *P = 0.0034, #P = 0.0052; ^P = 0.02; ¶P = 0.03; b: *P = 0.03, **P = 0.02, #P = 0.01, ^P = 0.001, ##P = 0.005, §P = 0.002, ^^P = 0.006, ¶P = 0.04, ^#P = 0.009; c: *P = 0.03, ^P = 0.02, #P = 0.04) and log-rank (Mantel–Cox) test, respectively (a: **P = 0.0042, b: **P = 0.0032, *P = 0.01285, c: *P = 0.0167, **P = 0.0055). Source data

Fig. 3. Engagement of FcγRIIa by Fc-engineered monoclonal antibodies drives dendritic cell maturation and protective CD8⁺ T cell responses.

a–c, FcγR-humanized mice (n = 4 mice per group for all groups, except n = 3 for day 7 WT and GRLR) were treated with FI6v3 monoclonal antibody Fc variants (intraperitoneally 3 mg kg⁻¹), infected with H1N1 (PR8), and then lung-resident dendritic cells and T cells were analysed on days 4 (n = 4 mice per group) or 7 (n = 3 mice per group for WT and GRLR, n = 4 for GAALIE in two independent experiments) after infection. One-way ANOVA (Bonferroni post hoc analysis adjusted for multiple comparisons) was used to compare the frequency of mature conventional dendritic cell subsets (CD80^highCD86^high) on day 4 after infection (*P = 0.0384, **P = 0.0075, ***P = 0.0008, ****P = 0.0002) (a), CD40 expression in cDC1 subsets (*P = 0.0261, **P = 0.0139, ***P = 0.0061, ****P = 0.0007) (b), and the frequency of activated (CD44^hiCD69⁺) CD8⁺ and CD4⁺ T cells (day 4 after infection) (^P = 0.0412, *P = 0.0232, **P = 0.0018, ***P = 0.0003) (c). MFI, mean fluorescence intensity. d–f, Fc variants of the FY1 monoclonal antibody (mAb) were administered intraperitoneally (2 mg kg⁻¹) to FcγR-humanized mice before influenza challenge (H1N1 PR8). Isotype or anti-mouse CD8 monoclonal antibodies were administered on day 3 after infection (isotype-treated groups: n = 12 mice per group for WT, GAALIE and PBS groups, n = 8 mice for GA; anti-CD8-treated groups: n = 11 mice per group for WT, n = 8 for GA, n = 10 for PBS, and n = 12 for GAALIE in three independent experiments). e, f, Weight loss (mean ± s.e.m.) (e) and survival curves (f) were compared against the corresponding isotype-treated group by two-way ANOVA (Bonferroni post hoc analysis adjusted for multiple comparisons) (*P = 0.03, ^P = 0.0220) and log-rank (Mantel–Cox) test, respectively (**P = 0.00199, *P = 0.0477). Source data

Fig. 4. Evaluation of GAALIE variants of the FY1 monoclonal antibody in models of therapy or prevention of influenza infection.

a–c, FcγR-humanized mice (n = 6 mice per group for antibody-treated groups; n = 3 for PBS-treated in two independent experiments) were infected with influenza (H1N1 PR8), and FY1 Fc variants were administered intraperitoneally 3 days after infection at the indicated dose. b, c, Weight loss (mean ± s.e.m.) (b) and survival curves (c) of GAALIE-treated mice were compared with the PBS-treated group at the corresponding antibody dose by two-way ANOVA (Bonferroni post hoc analysis adjusted for multiple comparisons) (*P = 0.0456, ^P = 0.041, **P = 0.0014, ^^P = 0.0003, #P = 0.0005, ***P < 0.0001) and log-rank (Mantel–Cox) test, respectively (**P = 0.000911). d–h, The protective activity of LS and GAALIE–LS variants of FY1 was evaluated in a model of antibody-mediated prophylaxis. FcγR/FcRn-humanized mice (n = 8 mice per group for GAALIE–LS at 1.6 and 0.4 mg kg⁻¹; n = 10 for LS at 0.1 mg kg⁻¹, n = 3 for PBS, n = 9 for the remaining treatment groups in two independent experiments) were administered intravenously with the indicated dose of FY1 2 days before influenza challenge (H1N1 PR8). e, f, Weight loss (mean ± s.e.m.) (e) and survival curves (f) of GAALIE–LS-treated mice were compared against the LS group at the corresponding antibody dose by two-way ANOVA (Bonferroni post hoc analysis adjusted for multiple comparisons) (*P < 0.001, ^P = 0.01, #P = 0.02, §P = 0.006, ¶P = 0.002) and log-rank (Mantel–Cox) test, respectively (*P = 0.0283, **P = 0.0266, ***P = 0.00377, ****P = 0.000143). h, The enhanced potency conferred by the GAALIE–LS variant was quantified by plotting the maximum weight change after infection against the serum antibody concentration at the time of challenge. Data were fitted by nonlinear regression analysis (four-parameter, variable slope). Source data

Extended Data Fig. 1. Characterization of the FcγR binding profile and Fc effector activity of Fc domain variants.

Fc domain variants with differential FcγR binding affinity were generated through the introduction of amino acid substitutions at the hinge proximal region of the CH2 domain of human IgG1. a, The positions of the mutated residues for the different Fc domain variants are highlighted. b, The affinity of these human IgG1 Fc variants for the different human FcγRs was assessed by surface plasmon resonance and the dissociation constant (K_d) (in M) is presented. The following references are cited in the table: refs. ^,. c, d, HPLC analysis of Fc domain variants using size-exclusion chromatography (SEC) columns was performed to determine whether mutations at the Fc domain are associated with increased antibody aggregation. The SEC profiles (c: overlay; d: individual Fc variants) and the abundance (percentage) of monomeric IgG is presented for the different Fc variants. e, Fc variants for the anti-NP monoclonal antibody 3B62 were generated and their binding (monomeric for FcγRI or as IgG immune complexes using NP-BSA for all other FcγRs) to immobilized human FcγRI, FcγRIIa, FcγRIIb and FcγRIIIa was assessed by ELISA. Results are from one experiment performed in duplicates. The Fc effector activity of anti-influenza monoclonal antibody FY1 Fc variants was assessed in vitro using FcγRIIa-expressing (f: n = 2 independent experiments for wild type, n = 1 for other groups, i: n = 2 independent experiments) and FcγRIIIa-expressing (g (F¹⁵⁸ allele), j (V¹⁵⁸ allele); n = 1 for each variant, except for GAALIE–LS (n = 2 independent experiments)) NFAT reporter cell lines. Fc variants with enhanced affinity for FcγRIIa or FcγRIIIa demonstrated increased capacity to induce NFAT reporter activation. h, Similarly, FY1 Fc variants engineered for FcγRIIIa binding exhibited improved primary human natural killer cell-mediated ADCC activity against HA-expressing cells. Results are the mean from two independent experiments using different natural killer cell donors. Source data

Extended Data Fig. 2. Anti-HA and NA monoclonal antibody titration studies to determine the optimal dose required for protection against mouse influenza infection.

Given the differential epitope specificities and in vitro neutralization potency of the selected anti-influenza monoclonal antibodies, titration studies were performed to assess the capacity of these antibodies to protect mice against lethal influenza infection. Anti-HA and NA antibodies (a, FI6v3 (n = 5 mice per group for 8 mg kg⁻¹ group, n = 6 mice per group for all other groups); b, FY1 (n = 6 mice per group for 4 and 2 mg kg⁻¹, n = 5 mice per group for 1 and 0.5 mg kg⁻¹, n = 4 mice per group for PBS); c, 4G05 (n = 7 mice per group for 0.5 mg kg⁻¹ group, n = 6 mice per group for all other groups); d, 1A01 (n = 4 mice per group for 1 mg kg⁻¹ group, n = 5 mice per group for all other groups); e, 3C05 (n = 5 mice per group for 20 and 5 mg kg⁻¹, n = 4 mice per group for 10 mg kg⁻¹ and PBS groups)). All antibodies were expressed as human IgG1 and administered intraperitoneally (for FI6v3 and FY1) or intravenously (for 4G05, 1A01, and 3C05) at the dose indicated to mice (C57BL/6) 4 h before lethal challenge with influenza (five mLD₅₀; H1N1 PR8 for a, b; H1N1 Neth/09 for c–e). Survival was monitored for 14 days. Source data

Extended Data Fig. 3. In vitro characterization of the antigenic specificity, neutralization potency, and HAI activity of Fc domain variants of anti-influenza monoclonal antibodies.

a–l, To study the role of Fc–FcγR interactions in the antibody-mediated protection against influenza infection, Fc domain variants with differential FcγR affinities (Fig. 1b) were generated for antibodies that target distinct epitopes on influenza antigens (Fig. 1a). These antibodies include FI6v3 (a–c) and FY1 (d–f), which both recognize the stalk region of influenza HA and exhibit broad and potent neutralizing activity against group 1 and 2 influenza strains, 4G05 (g–i), which is a pan-H1 monoclonal antibody against the globular head of HA and exhibits potent neutralizing and HAI activity, 1A01 (j–l), which is a pan-H1 anti-globular head HA antibody with no neutralizing or HAI activity, and 3C05 (m–o), a broadly (pan-H1) protective antibody against NA. To confirm that changes in the Fc domain have no effect on the antigenic activity and Fab-mediated functions of these antibodies, Fc domain variants were characterized by ELISA (a, d, g, j, m; n = 1 experiment performed in duplicates) to assess their specificity against purified HA (H1N1 PR8 strain for FI6v3 and FY1; Cal/09 strain for 4G05, 1A01 and 3C05), by microneutralization assay (b, e, h, k and n; c, f, i, l and o for IC₅₀ values; n = 2 independent experiments; data were fitted with nonlinear regression analysis (four-parameter) to calculate IC₅₀ values) to evaluate their neutralizing activity against H1N1 (H1N1 PR8 strain for FI6v3 and FY1; Neth/09 strain for 4G05, 1A01, and 3C05), and by HAI assay (H1N1 PR8 strain for FI6v3 and FY1; Neth/09 strain for 4G05, 1A01, and 3C05; n = 2 independent experiments) to determine their HAI titre (c, f, i, l and o). Source data

Extended Data Fig. 4. In vivo half-life of Fc domain variants.

a, Fc variants of an anti-HIV antibody (3BNC117) were administered (intravenously; 100 μg) to FcγR-humanized mice and antibody serum levels were determined by ELISA at various time points after antibody administration. n = 4 mice per group in two independent experiments. Data are mean ± s.e.m. b–f, To ensure that the observed differences in the protective activity of Fc variants of anti-HA and NA antibodies (Figs. 1 and 2) were not due to differential in vivo antibody half-lives, serum was obtained from influenza-infected mice (day 3 after infection) and analysed by ELISA to quantify antibody levels. (b: n = 9 mice per group for GAALIE, n = 10 mice per group for all other groups; c: n = 10 mice per group for wild type, n = 8 mice per group for all other groups ; d: n = 10 mice per group; e: n = 6 mice per group for GRLR, n = 8 mice per group for all other groups; f: n = 12 mice per group for wild type and GA; n = 10 mice per group for GRLR and ALIE; n = 9 mice per group for GAALIE). g, To ensure that the observed differences in the protective activity of Fc variants of FY1 antibodies (Fig. 3e, f, Extended Data Fig. 10d, e) were not due to differential in vivo antibody half-lives, serum was obtained from influenza-infected mice (day 3 after infection) and analysed by ELISA to quantify FY1 antibody levels. n = 12 mice per group for wild type/isotype, GAALIE/isotype, and GAALIE/anti-CD8 groups, n = 11 mice per group for wild-type/anti-CD8, n = 8 mice per group for GA/isotype, GA/anti-CD4, and GA/anti-CD4 groups, n = 6 mice per group for wild-type/anti-CD4 group. h, Titration studies were performed in a mouse model of antibody-mediated prophylaxis of influenza infection (Fig. 4d) and serum was obtained at the time of virus challenge and analysed by ELISA to quantify FY1 antibody levels. n = 10 mice per group for LS 0.1 mg kg⁻¹ dose, n = 8 for GAALIE–LS at 1.6 and 0.4 mg kg⁻¹ doses, n = 9 mice per group for all other groups. Source data

Extended Data Fig. 5. Evaluation of the role of FcγRIIa and the contribution of neutrophils in the antibody-mediated protection against influenza infection.

a, b, To confirm the dependence of FcγRIIa engagement in driving the protective activity of the GA variant, mice expressing the human FcγRIIa transgene on an FcγR_null background (hFcγRIIa⁺; gating strategy (a) and representative flow cytometry histograms (b) of FcγRIIa expression in lung-resident leukocytes) or deficient for all classes of FcγRs (hFcγRIIa⁻) were administered with GA variants of FY1 (intraperitoneally 2 mg kg⁻¹) (n = 7 mice per group for FcγRIIa⁺, n = 6 mice per group for FcγRIIa⁻ in two independent experiments) or PBS (n = 4 mice per group for FcγRIIa⁺, n = 5 mice per group for FcγRIIa⁻ in two independent experiments) 4 h before lethal challenge with influenza (H1N1; PR8, 5 mLD₅₀). c, d, Weight loss (c) (mean ± s.e.m.) and survival (d) were monitored for 14 days and compared by two-way ANOVA (Bonferroni post hoc analysis adjusted for multiple comparisons) (c: *P = 0.02 §P = 0.03, ¶P = 0.006, ^P < 0.0001, #P = 0.0002, ^^P = 0.0006, **P = 0.001, ##P = 0.002 versus GA-treated FcγRIIa⁻) and log-rank (Mantel–Cox) test, respectively (d: *P = 0.0006 versus GA-treated FcγRIIa⁻). e, f, FcγR-humanized mice were administered with GA variants of FI6v3 (intraperitoneally 4 mg kg⁻¹) before challenge with PR8 (as described in c, d). To block the ligand binding activity of FcγRIIa, recombinant anti-FcγRIIa (clone IV.3) expressed as human IgG1 GRLR variant to abrogate FcγR binding or isotype control (anti-hapten (NP) monoclonal antibody; clone 3C13) was administered intranasally (80 μg) to mice 1 day after virus challenge (n = 7 mice per group for GA/IV.3-treated group, n = 5 mice per group for all other groups in two independent experiments) and weight loss (e) (mean ± s.e.m.) and survival (f) were monitored. g, To establish the efficiency of antibody-mediated neutrophil depletion, FcγR-humanized mice (n = 3 mice per group in one experiment) were injected intravenously with 150 μg anti-mouse Gr-1 monoclonal antibody (clone RB6-8C5) or isotype control (clone LTF-2). The abundance of neutrophils (Ly6G⁺) in peripheral blood was determined 2 days after antibody administration by flow cytometry. **P = 0.0053, two-sided unpaired t-test. h, i, To assess the contribution of neutrophils to the protective activity of FcγRIIa-enhanced variants of anti-influenza antibodies, GA variants of the anti-HA stalk antibody FY1 were administered intraperitoneally (2 mg kg⁻¹) to FcγR-humanized mice (n = 6 mice per group, except for PBS/Gr-1 treated (n = 5 mice per group) in two independent experiments) 4 h before lethal challenge with PR8. Isotype or anti-mouse Gr-1 monoclonal antibodies (150 μg intravenously) were administered on day 1 after infection and weight loss (h) (mean ± s.e.m.) and survival (i) were recorded. Source data

Extended Data Fig. 6. Abundance and FcγR expression profile of leukocyte populations in the lungs of influenza-infected FcγR-humanized mice at different time points afteer infection.

a–f, To determine the abundance and FcγR expression profile of lung resident and infiltrated leukocytes during the course of influenza infection, cohorts of FcγR-humanized mice were infected (intranasally with H1N1 PR8; 5 mLD₅₀) and euthanized at different time points after infection (day 0 to 6). Lungs were homogenized and analysed by flow cytometry (a: gating strategy) to determine the frequency (b, c) and FcγR expression profile (f) of innate effector leukocytes. n = 3 and n = 4 mice per time point for day 0–2 and day 3–6 time points, respectively. Influenza infection was associated with the recruitment of natural killer (NK) cells, neutrophils and monocytes, whereas the number of alveolar macrophages was reduced at the later stages of infection. Owing to the high degree of sequence similarity between FcγRIIa and FcγRIIb, expression of these FcγRs was assessed using antibody clones (d: IV.3 for FcγRIIa; e: 2B6 for FcγRIIb) that exhibit high specificity, as assessed by ELISA using recombinant FcγRs (n = 1 experiment performed in duplicates). Analysis of the FcγR expression profile (MFI) revealed that influenza infection had no effect on the levels of FcγRs expressed by the various leukocyte types. With the exception of natural killer cells, which expressed only FcγRIIIa, most innate effector leukocytes co-expressed multiple FcγRs, including the activating FcγRs, FcγRIIa and FcγRIIIa/b, as well as the inhibitory FcγRIIb. n = 3 and n = 4 mice per time point for day 0–2 and day 3–6 time points, respectively. Source data

Extended Data Fig. 7. Treatment of FcγR-humanized mice with GAALIE variants of anti-HA antibodies is associated with increased frequency of activated dendritic cells.

a–f, To determine the abundance and FcγR expression profile of dendritic cell subsets during the course of influenza infection, cohorts of FcγR-humanized mice were infected (intranasally with H1N1 PR8; 5 mLD₅₀) and euthanized at different time points after infection (day 0 to 6). Lungs were homogenized and analysed by flow cytometry (a: gating strategy) to determine the frequency (b) and FcγR expression profile (c: representative flow cytometry overlay; d–f: MFI) of the three major dendritic cell subsets identified: cDC1, cDC2 and tipDCs. Influenza infection was not associated with any major changes in the number of lung-resident cDC1 and cDC2, whereas tipDCs were almost absent at baseline, but their number increased markedly after infection. cDC1 and cDC2 expressed FcγRIIa and FcγRIIb, but they were negative for FcγRIIIa. By contrast, tipDCs expressed FcγRIIa and FcγRIIIa, along with the inhibitory FcγRIIb. Owing to the very low number of tipDCs at baseline, FcγR expression (MFI) was omitted. n = 4 mice per time point assessed. g–i, To investigate the effect of enhanced FcγRIIa engagement by GAALIE variants on the maturation status of dendritic cells, FcγR-humanized mice were treated with Fc domain variants of the anti-HA stalk antibody FI6v3, exhibiting differential FcγR affinity—wild type IgG1 (baseline FcγR affinity), GRLR (diminished binding to all classes of FcγRs), and GAALIE (increased FcγRIIa and FcγRIIIa affinity). Fc domain variants were administered intraperitoneally (3 mg kg⁻¹) to FcγR-humanized mice (n = 4 mice per treatment group in two independent experiments) 4 h before lethal challenge with H1N1 (PR8; 5 mLD₅₀). Mice were euthanized on day 4 and lung-resident dendritic cells were analysed by flow cytometry. The abundance of mature (defined as CD80^highCD86^high) cDC1 (g), cDC2 (h), and tipDCs (i) was compared between mice treated with the various Fc domain variants of FI6v3. Representative flow cytometry plots from data presented in Fig. 3a. In contrast to cDC1 and cDC2, no differences were observed in the maturation status of tipDCs among mice treated with different FI6v3 Fc variants (one-way ANOVA). j, k, In vitro differentiated monocyte-derived dendritic cells (n = 4 peripheral blood mononuclear cell donors performed in two independent experiments) were stimulated overnight with IgG immune complexes (anti-NP–NP-BSA immune complexes (j) and heat-aggregated IgG complexes (k)). The abundance (percentage) of mature dendritic cells (defined as CD80^highCD86^high) was assessed by multicolour flow cytometry and compared against the corresponding wild-type-treated group by one-way ANOVA (Bonferroni post hoc analysis adjusted for multiple comparisons) *P = 0.0417, **P = 0.0134, ***P < 0.0001, ^P = 0.0049. l–n, Cluster analysis of dendritic cell populations present in the lungs of influenza-infected mice treated with Fc domain variants of anti-HA monoclonal antibodies. Mice (n = 4 mice per group in two independent experiments) were treated as described in g–i, euthanized on day 4 after infection and dendritic cells (Lin⁺MHCII⁺CD11c⁺) were analysed by multicolour flow cytometry. Data were dimensionally reduced and visualized using the UMAP algorithm. l, UMAP plots of dendritic cells in mice treated with the different Fc domain variants are presented. Populations were identified by X-shift using KNN density estimation and assigned IDs (A-J). m, The abundance of the various dendritic cell clusters in the different treatment groups was plotted and populations that are enriched or reduced in GAALIE-treated mice were identified. n, Histogram plots of the expression of CD80, CD86, CD40 and MHCII in dendritic cell populations that are enriched (red, A; and orange, H) or reduced (cyan, D; purple, E; and blue, F) in GAALIE-treated mice. GAALIE treatment was associated with the enrichment of dendritic cell populations characterized by high levels of CD86 and CD40 expression. Results are from four mice per treatment condition in two independent experiments. Source data

Extended Data Fig. 8. Treatment of FcγR-humanized mice with GAALIE variants of anti-HA stalk antibodies is associated with increased activation of CD8⁺ and CD4⁺ T cells.

To investigate whether the observed increase in the frequency of mature dendritic cells in mice treated with GAALIE variants of anti-HA monoclonal antibodies was associated with enhanced T cell responses, the activation status of CD8⁺ and CD4⁺ T cells was analysed and compared between mice treated with anti-HA Fc domain variants with differential FcγR affinity (wild-type IgG1, GRLR and GAALIE). Fc domain variants of the anti-HA stalk antibody FI6v3 were administered (intraperitoneally 3 mg kg⁻¹) to FcγR-humanized mice before lethal challenge with H1N1 (PR8; 5 mLD₅₀). Mice (n = 4 mice per group in two independent experiments) were euthanized on day 4 after infection and T cell populations were analysed by multicolour flow cytometry. a, b, The frequency of activated (defined as CD44^hiCD69⁺) CD8⁺ (a) and CD4⁺ (b) T cells was compared between mice treated with the various Fc domain variants of FI6v3. Representative flow cytometry plots from data in Fig. 3c. In addition, cluster analysis of T cell populations present in the lungs of influenza-infected mice treated with Fc domain variants of anti-HA monoclonal antibodies was performed. Flow cytometry data were dimensionally reduced and visualized using the UMAP algorithm. c, UMAP plots of T cells in mice treated with the different Fc domain variants are presented. Populations were identified by X-shift using KNN density estimation and assigned IDs (A-M). d, Heat map of the abundance of the various T cell clusters in the different treatment groups. Populations that are enriched or reduced in GAALIE-treated mice were identified. e, Histogram plots of the expression of CD69, CD25, CD44 and CD62L in T cell populations that are enriched (red, G; green, E; orange, J) or reduced (cyan, C; blue, D; magenta, L) in GAALIE-treated mice. GAALIE treatment was associated with the enrichment of T cell populations characterized by high levels of CD69, CD44 and CD25 expression. Results are from four mice per treatment condition. Source data

Extended Data Fig. 9. FcγR expression analysis of T cells.

To determine whether human T cells express human FcγRs, human peripheral blood mononuclear cells were analysed by multicolour flow cytometry. a, Gating strategy for identifying human T cells. Monocytes (CD14⁺CD11b⁺) and B cells (CD19⁺CD11b⁻) were included as controls for FcγR immunofluorescence staining. b, g, Analysis of FcγR expression revealed that human T cells are negative for FcγR expression (b: representative flow cytometry histogram overlays; g: quantification of FcγR⁺ cells; n = 3 donors). c–f, h–l, Human FcγR expression on peripheral blood (c, d, h, i), spleen (e, f, j, k), or lung (l) T cells from naive (h, j) or influenza-infected (i, k, l; day 6 after infection (5 mLD₅₀; H1N1 PR8)) FcγR-humanized mice was assessed by multicolour flow cytometry. Gating strategy for identifying T cells in mouse blood (c), spleen or lungs (e). Myeloid cells (CD11b⁺) and B cells (B220⁺) were included as controls for FcγR immunofluorescence staining. Analysis of FcγR expression on T cells from naive or influenza-infected FcγR-humanized mice revealed minimal expression of FcγRs on T cell (d, f: representative flow cytometry histogram overlays; h–l: quantification of FcγR⁺ cells (n = 2 mice per group in one experiment) in the blood (h, i), spleen (j, k) and lungs (l) of naive or influenza-infected mice). Although a small fraction of B cells appears positive for FcγRIIa expression, this is probably due to the cross-reactivity of the anti-FcγRIIa antibody (clone IV.3) tο FcγRIIb (see Extended Data Fig. 6d). Source data

Extended Data Fig. 10. T cell depletion experiments and the effect of FcγRIIa-enhanced variants on anti-influenza IgG responses.

a, b, To establish the efficiency of antibody-mediated CD8⁺ or CD4⁺ T cell depletion, FcγR-humanized mice were injected intravenously with 150 μg anti-mouse CD8α monoclonal antibody (clone 2.43) or anti-mouse CD4 (clone GK1.5), or isotype control (clone LTF-2). The abundance of CD8⁺ (a) or CD4⁺ (b) T cells in peripheral blood was determined at various time points after antibody administration by flow cytometry. Baseline CD8⁺ or CD4⁺ T cell frequencies were determined in blood samples obtained before antibody administration. Results are expressed as the percentage of CD8⁺CD3⁺ or CD4⁺CD3⁺ T cells (n = 6 for anti-CD8 and anti-CD4-treated groups, n = 7 mice per group for isotype groups in two independent experiments). c–e, To determine the contribution of CD4⁺ T cells in the protective activity of FcγRIIa-enhanced variants of anti-influenza monoclonal antibodies, CD4⁺ depletion studies were performed. c, Wild-type or GA variants of the anti-HA stalk antibody FY1 were administered intraperitoneally (2 mg kg⁻¹) to FcγR-humanized mice 4 h before lethal challenge with influenza (5 mLD₅₀; H1N1 PR8). Isotype or anti-mouse CD4 antibodies (150 μg intravenously) were administered on day 3 after infection. Isotype-treated groups: n = 12 mice per group for wild-type and PBS groups, n = 8 mice per group for GA group; anti-CD4-treated groups: n = 6 mice per group for wild type, n = 8 mice/group for GA, n = 4 mice per group for PBS in two independent experiments. d, e, Weight loss (d) (mean ± s.e.m.) and survival (e) were monitored for 14 days. Data for isotype-treated mice are also in Fig. 3e, f. f–m, Wild-type or GAALIE Fc variants for the anti-HA globular head antibody 1A01 were administered (intravenously 4 mg kg⁻¹) to FcγR-humanized mice (n = 7 mice per group for wild-type group, n = 11 mice per group for GAALIE-treated group, n = 5 mice per group for PBS-treated in two independent experiments) 4 h before lethal challenge with influenza (5 mLD₅₀; H1N1 Neth/09). f, g, Weight loss (mean ± s.e.m.) and survival (g) were monitored for 14 days. h–k, IgG responses against HA (h, i) and NP (j, k) from homologous (H1N1; A/California/04/2009) or heterologous (H3N2; A/x31) strains were evaluated in surviving mice. l, Analysis of IgG titres against homologous or heterologous HA or NP revealed no differences among mice previously treated with wild-type or GAALIE Fc variants of 1A01 antibody. m, Similarly, comparable HAI titres were noted among the treatment groups. Source data