Impaired antibody response to influenza vaccine in HIV-infected and uninfected aging women is associated with immune activation and inflammation

Abstract

Background: Aging and HIV infection are independently associated with excessive immune activation and impaired immune responses to vaccines, but their relationships have not been examined.

Methods: For selecting an aging population we enrolled 28 post-menopausal women including 12 healthy volunteers and 16 HIV-infected women on antiretroviral treatment with <100 HIV RNA copies/ml. Antibody titers to trivalent influenza vaccination given during the 2011-2012 season were determined before and 4 weeks after vaccination.

Results: Seroprotective influenza antibody titers (≥ 1:40) were observed in 31% HIV(+) and 58% HIV-uninfected women pre-vaccination. Following vaccination, magnitude of antibody responses and frequency of seroprotection were lower in HIV(+) (75%) than in HIV(-) (91%) women. Plasma IL-21, the signature cytokine of T follicular helper cells (Tfh), and CD4 T cell IL-21R were upregulated with seroconversion (≥ 4 fold increase in antibody titer). Post-vaccine antibody responses were inversely correlated with pre-vaccination plasma TNFα levels and with activated CD4 T cells, including activated peripheral (p)Tfh. Plasma TNFα levels were correlated with activated pTfh cells (r=0.48, p=0.02), and inversely with the post-vaccination levels of plasma IL-21 (r=-0.53, p=0.02). In vitro TNFα blockade improved the ability of CD4 T cells to produce IL-21 and of B cells to secrete immunoglobulins, and addition of exogenous IL-21 to cell cultures enhanced B cell function. Higher frequencies of activated and exhausted CD8 T and B cells were noted in HIV(+) women, but these markers did not show a correlation with antibody responses.

Conclusions: In aging HIV-infected and uninfected women, activated CD4 and pTfh cells may compromise influenza vaccine-induced antibody response, for which a mechanism of TNFα-mediated impairment of pTfh-induced IL-21 secretion is postulated. Interventions aimed at reducing chronic inflammation and immune activation in aging, HIV-infected patients may improve their response to vaccines.

Figure 1. Antibody response to influenza vaccination.

Antibody titers specific for the whole 2011-2012 seasonal influenza vaccine were determined before (t0) and 4 weeks (wk4) after vaccination by hemagglutination inhibition assay in the plasma of 12 HIV^– and 16 HIV⁺ elderly women. Reverse cumulative distribution curves for the antibody titers prior to (A) and 4 weeks after (B) vaccination for HIV-uninfected (open circles) and HIV-infected (filled circles) donors were derived to illustrate immune responses. (C) Reciprocal of Ab titers before and after vaccination. (D) Ratio between the wk4 and the baseline reciprocal of Ab titer. P values were calculated with Student’s t-test or Mann-Whitney test as appropriate.

Figure 2. T cell activation and senescence are associated with decreased humoral response to influenza vaccination.

Frozen PBMC obtained before influenza vaccination were thawed, rested over night and stained with monoclonal antibodies for immunophenotyping of CD4 (ViViD^-CD3⁺CD4⁺), pTfh (ViViD^-CD3⁺CD4⁺CD45RO⁺CXCR5⁺) and CD8 (ViViD^-CD3⁺CD8⁺) T cell subsets. Correlations were established between the reciprocal of influenza Ab titer at week 4 and the frequency of activated (CD38⁺HLA-DR⁺), dividing (Ki-67⁺) and senescent (CD28^-CD57⁺) CD4 (A), pTfh (B) and CD8 (C) T cells for 10 HIV^– (open dots) and 15 HIV⁺ (filled dots) post-menopausal women. Pearson analysis was utilized to establish statistical correlation between variables.

Figure 3. Phenotypic alterations in B cell subsets in HIV-infected aging women are not associated with antibody response to influenza vaccination.

PBMC were stained with surface MoAbs to CD20, CD21, CD10, CD27 and FcRL4. Lymphocytes were gated based on forward and side scatter. CD3^-CD20⁺ cells were gated into CD21^hi and CD21^lo/neg cells and further divided based on the expression of CD27 and CD10 as: early transitional (CD21^lo/negCD27^-CD10⁺), late transitional (CD21^hiCD27^-CD10⁺), naïve (CD21^hiCD27^-CD10^–), resting memory (CD21^hiCD27⁺CD10^–), activated memory (CD21^lo/negCD27⁺CD10^–), and exhausted tissue-like (CD21^lo/negCD27^-CD10^–) subsets. (A) The percentage of cells in each B cell subpopulation was determined in 10 healthy controls (open dots) and 15 HIV-infected women (filled dots) before vaccination. (B) Expression of the exhaustion marker FcRL4 was evaluated in each B cell subset. (C) Correlations were established between the B cell populations that resulted significantly different between HIV- and HIV+ in A and B and the reciprocal of influenza Ab titer at week 4. Correlation between variables was established with Pearson analysis.

Figure 4. Association between circulating levels of TNFα, Ab response to influenza vaccination and T cell activation and senescence.

(A) Correlation between TNFα plasma levels at t0 and the reciprocal of vaccine-specific Ab titers at week 4 was established for 12 HIV^– and 14 HIV⁺ aging women. Plasma TNFα levels were measured using a customized MILLIPLEX™ Cytokine Human Ultrasensitive magnetic bead panel (EMD Millipore). (B-D) T cells were immunophenotyped using MoAbs specific for activation (CD38, HLA-DR), proliferation (Ki-67) and senescence (CD28, CD57) markers. Frequencies of CD4, pTfh and CD8 T cell subsets expressing these markers were correlated with pre-vaccination TNFα plasma levels. Statistical analysis was performed using Pearson correlation.

Figure 5. TNFα blockade promotes IL-21 secretion and IgG production in CD4-B cell co-cultures.

Memory CD4 T cells obtained from 4 healthy donors were cultured with autologous B cells at a 1:1 ratio in the presence of SEB and/or TNFα, IL-21, anti-anti-TNFα or anti-IL-21R blocking antibodies or the appropriate isotype control. (A) Phenotyping was performed on day 3 of co-culture in gated live CD4 and pTfh cells. (B) IL-21 levels measured by ELISA in day 3 culture supernatants. (C, D) IgM, IgG and IgA levels were evaluated by ELISA after 7 days of co-culture. In (D), IgG levels in samples where IL-21R-Fc was added were not measurable because of the cross-reactivity to the Fc portion of IL-21R-Fc. P values were calculated using paired Student’s t-test for comparisons between two groups, and repeated measures one-way ANOVA for comparisons between larger groups.

Figure 6. Increased plasma IL-21 and IL-21R expression in CD4 T cells is associated with humoral response to the influenza vaccine.

(A) Circulating levels of IL-21 were measured by ELISA at t0 and wk4 in the plasma of 11 HIV^– seroprotected (HIV^– SP), 12 HIV⁺ seroprotected (HIV⁺ SP) and 4 HIV⁺ non seroprotected (HIV⁺ NSP) post-menopausal women. (B) Correlation between TNFα plasma levels at baseline and the ratio between wk4 and baseline plasma IL-21. (C, D) IL-21R mean fluorescence intensity (MFI) was established in live CD4 (C) and CD8 (D) T cells of 8 HIV^– SP, 11 HIV⁺ SP and 4 HIV⁺ NSP aging women. Frozen cells obtained before and four weeks after vaccine administration were thawed, rested overnight, stained for surface markers and acquired by flow cytometry. IL-21R MFI was established for ViViD^-CD3⁺CD4⁺ or ViViD^-CD3⁺CD8⁺ events. Statistical analysis was performed using paired Student’s t-test for comparisons within groups, unpaired Student’s t-test for comparisons between groups, and Pearson correlation as appropriate.