Hepcidin-Mediated Hypoferremia Disrupts Immune Responses to Vaccination and Infection

Abstract

Background: How specific nutrients influence adaptive immunity is of broad interest. Iron deficiency is the most common micronutrient deficiency worldwide and imparts a significant burden of global disease; however, its effects on immunity remain unclear.

Methods: We used a hepcidin mimetic and several genetic models to examine the effect of low iron availability on T cells in vitro and on immune responses to vaccines and viral infection in mice. We examined humoral immunity in human patients with raised hepcidin and low serum iron caused by mutant TMPRSS6. We tested the effect of iron supplementation on vaccination-induced humoral immunity in piglets, a natural model of iron deficiency.

Findings: We show that low serum iron (hypoferremia), caused by increased hepcidin, severely impairs effector and memory responses to immunizations. The intensified metabolism of activated lymphocytes requires the support of enhanced iron acquisition, which is facilitated by IRP1/2 and TFRC. Accordingly, providing extra iron improved the response to vaccination in hypoferremic mice and piglets, while conversely, hypoferremic humans with chronically increased hepcidin have reduced concentrations of antibodies specific for certain pathogens. Imposing hypoferremia blunted the T cell, B cell, and neutralizing antibody responses to influenza virus infection in mice, allowing the virus to persist and exacerbating lung inflammation and morbidity.

Conclusions: Hypoferremia, a well-conserved physiological innate response to infection, can counteract the development of adaptive immunity. This nutrient trade-off is relevant for understanding and improving immune responses to infections and vaccines in the globally common contexts of iron deficiency and inflammatory disorders.

Funding: Medical Research Council, UK.

Keywords: T-cells; adaptive immunity; global health; hepcidin; hypoferremia; immunometabolism; infection; influenza virus; iron; vaccination.

Graphical abstract

Figure 1

Minihepcidin-Induced Serum Iron Deficiency Inhibits Immune Responses (A) Experimental design for investigation of the effect of combined minihepcidin injection and low iron diet on the primary immune response to immunization. Serum iron levels and liver hepcidin mRNA were measured 24 h after the last minihepcidin injection. Means ± SDs. Student’s 2-tailed t test, unpaired. (B) Left: representative flow cytometry plots showing frequency of OVA (SIINFEKL) tetramer-positive CD8 effector T cells (Tetramer⁺ CD44⁺ CD8 T cells) as a percentage of CD8s T cells. Right: number of splenic OVA-specific CD8 effector cells 7 days after immunization with AdHu5-OVA. Means ± SDs. Student’s 2-tailed t test, unpaired. (C) Number of splenic OVA-specific CD8 effector cells 7 days after OVA protein in adjuvant immunization. Means ± SDs. Student’s 2-tailed t test, unpaired. (D) Number of T follicular helper cells (PD-1⁺ CXCR5⁺ CD44⁺ CD4 T cells), germinal center (GC) B cells (IgD⁻ CD95⁺ B220⁺ B cells), plasma cells (CD138⁺ IgD⁻ cells), and anti-OVA IgG titer (a.u., arbitrary unit) 7 days after immunization with OVA in adjuvant. Means ± SDs. Student’s 2-tailed t test, unpaired. (E) Experimental design to determine whether parenteral iron supplementation can reverse the effect of minihepcidin injection; FAC = ferric ammonium citrate. Left: the frequency of antigen-specific splenic OT-I CD8 T cells; center: normalized relative mean fluorescence intensity (MFI) of surface CD71 expression by OT-I CD8 T cells 5 days after OVA and adjuvant immunization were detected with flow cytometry; right: liver hepcidin mRNA. Means ± SDs. One-way ANOVA with correction for multiple comparisons.

Figure 2

Minihepcidin-Mediated Serum Iron Deficiency Inhibits the CD8 T Cell Response to a Diverse Range of Immunizations and Impairs the Quality of CD4 T Cell Responses. (A) Experimental scheme to examine effect of minihepcidin injection on immunization. (B) Antigen-specific T cell populations quantified by flow cytometry 7 days after immunization. Left to right: number of splenic endogenous OVA-specific CD8 T cells after AdHu5-OVA immunization, frequency of endogenous OVA-specific CD8 T cells as a percentage of total CD8s in peripheral blood after AdHu5-OVA immunization, number of splenic OT-I OVA-specific CD8 T cells after MVA-OVA immunization, number of splenic OT-I OVA-specific CD8 T cells after OVA and adjuvant immunization, and frequency of endogenous vaccinia (B8R epitope)-specific CD8 T cells as a percentage of total CD8s, resolved by IFNγ production after ex vivo peptide restimulation, induced by MVA-OVA immunization. Means ± SDs. t test. Student’s 2-tailed t test, unpaired. (C) Left to right: relative MFI of IFNγ and TNF-α for OT-I effector cells producing the respective cytokine, MFI normalized to average of vehicle group; percentage of OT-I effector cells that secrete IL-2. Cytokine-producing cells resolved by intracellular cytokine staining after ex vivo restimulation of splenocytes from mice with SIINFEKL peptide 7 days after MVA-OVA immunization. Means ± SDs. t test. Student’s 2-tailed t test, unpaired. (D) Frequency of endogenous vaccinia-specific IFNγ, TNF-α, or IL-2 producing CD40L⁺ CD4 Th1 effector T cells as a percentage of total CD4s, resolved by intracellular cytokine staining after ex vivo restimulation of splenocytes with MVA-OVA-pulsed dendritic cells. Means ± SDs. t test. Student’s 2-tailed t test, unpaired. (E) Number of splenic OT-II T follicular helper cells induced by OVA and adjuvant immunization. Frequency of splenic TNF-α⁺ and IL-2⁺ OT-II effector cells induced by OVA and adjuvant immunization as a percentage of total OT-II CD4 T cells after ex vivo restimulation with peptide. Frequency of splenic GC B cells as a percentage of B cells after OVA and adjuvant immunization. All 7 days post-immunization. Means ± SDs. t test. Student’s 2-tailed t test, unpaired.

Figure 3

Iron Uptake via the Transferrin Receptor Is Cell-Intrinsically Essential for Immune Responses (A) Experimental design for establishing mixed bone marrow chimeras to investigate cell-intrinsic effect of Tfrc^Y20H/Y20H allele on immune response in vivo. Data from Tfrc^Y20H/Y20H ; WT chimeras are displayed in red, whereas data from WT;WT chimeras are in blue. (B) Ratio of the frequencies of CD45.2:CD45.1 cells within peripheral blood CD8 T cells, CD4 T cells, and B cells at 65 days after establishment of chimeras, determined by flow cytometry. Means ± SDs. Student’s 2-tailed t test, unpaired. (C) Comparison of the ratio of the frequencies of CD45.2:CD45.1 cells within naive and effector lymphocyte populations within each chimeric mouse, 72 days after establishment of chimeras and 1 week after MVA-OVA immunization. The data are displayed for naive (CD44⁻) and effector tetramer⁺ splenic CD8 T cells, dLN naive (CD44⁻) CD4 T cells and T follicular helper cells (PD-1⁺ CXCR5⁺ CD44⁺ CD4 T cells), and dLN naive follicular (IgD⁺, CD95⁻) and GC B cells (IgD⁻ CD95⁺ GL7⁺ B cells). Student’s 2-tailed t test, paired.

Figure 4

Cellular Iron Metabolism Is Remodeled upon T Cell Activation, and Iron Deficiency Disrupts T Cell Physiology (A) Left to right: Histograms of CD71 (Tfrc) expression on 4-OHT treated CD8 T cells from IRP1/2 floxed mice lacking (blue) or expressing Cre (red) at 24h (top) and 72h (bottom) after in vitro activation; relative CD71 MFI for CD8 T cells treated with 4-OHT measured by flow cytometry (72h after activation); relative CD71 MFI of 4-OHT treated compared to EtOH treated CD8 T cells (to control for effects of 4-OHT, 72h after activation). Mean+/- SD, Student's two tailed T-test, unpaired. (B) Percentage divided CD8 T cells of each genotype/EtOH or 4-OHT combination cultured with a titration of FeSO4, 72 h after activation. Percentage of divided determined as percentage of cells that had diluted out their CellTrace Violet signal once or more. Mean ± range. Two-way ANOVA, corrected for multiple comparisons, p values for difference between EtOH (control), and 4-OHT at each iron concentration. (C) Left to right: ATP content was measured using CellTiter Glo luminescent assay for equal numbers of WT CD8 T cells per condition and normalized to the control condition of each biological replicate. Total, glycolytic, and mitochondrial ATP production rates measured by Seahorse. Each data point is a biological replicate from a single mouse, representing the average of 2–5 technical replicate wells. Mitochondrial content (MitoTracker Green [MTG]) and mitochondrial inner membrane potential (MitoTracker Deep Red [MTDR]) were measured by flow cytometric analysis of 72-h activated divided cells and the ratio calculated. Mean ± range. One-way ANOVA, effect of iron, paired within each biological replicate.

Figure 5

CD8 T Cell Recall Responses Are Disrupted by Serum Iron Deficiency during the Primary Immune Response (A) Experimental design to investigate the effect of serum iron deficiency during the primary immune response on the magnitude of the secondary recall CD8 T cell response. (B) Number of endogenous OVA-specific secondary CD8 effector cells in the spleen detected by OVA tetramer binding and percentage of splenic OVA-specific CD8 effector cells detected by IFNγ⁺ production after ex vivo stimulation with SIINFEKL. Means ± SDs. Student’s 2-tailed t test, unpaired.

Figure 6

Iron Supplementation Normalizes Piglet Iron Status and Improves Vaccine Responses at 28 Days of Age Routine vaccinations are less efficacious in IRIDA patients. (A) Serum iron and hemoglobin of control and iron-supplemented piglets. Means ± SDs. Serum iron, mixed-effects model, p value is effect of iron supplementation. Hemoglobin, 2-way ANOVA, p value is effect of iron supplementation. (B) Antibody response against M. hyopneumoniae vaccination reported as improvement in percentage block at 28 days post-natal, relative to pre-immunization percentage block at 14 days after birth. Mean ± range. Student’s 2-tailed t test, unpaired. (C) Age of TMPRSS6 and WT control cohorts. Means ± SDs. Student’s 2-tailed t test, unpaired. (D) Antibody concentrations against rubella, Haemophilus influenzae b, and Streptococcus pneumoniae serotype 1 measured in IRIDA patients with TMPRSS6 mutations and non-anemic healthy controls without mutant TMPRSS6 alleles from the same clinic. Means ± SDs. Student’s 2-tailed t test, unpaired; Mann-Whitney U test for rubella. Right-most panel: analysis of the frequency of individuals for both genotypes with antibody concentrations against Streptococcus pneumoniae serotype 1 exceeding the protective threshold of 0.35 μg/mL (dotted line). Fisher exact test.

Figure 7

Serum Iron Deficiency Suppresses Adaptive Immune Responses to Influenza Infection, Worsening Lung Inflammation and Preventing Recovery of Weight (A) Experimental scheme for investigation effect of minihepcidin injection on the outcome of infection with 0.08 hemagglutination units of influenza virus A/X-31 (H3N2). (B) Representative plots of the frequency of influenza nucleoprotein (NP)-specific CD8 effector cells as a percentage of CD8 T cells in the spleens of mice 8 days post-infection. Numbers of NP-specific CD8 effector cells in the spleen and lung at 8 days post-infection. Number of splenic granzyme B⁺ CD8 T cells. Means ± SDs. Student’s 2-tailed t test, unpaired. (C) Mediastinal lymph node T follicular helper cell number (PD-1⁺ CXCR5⁺ CD44⁺ CD4 T cells) and frequency of CD44⁺ antigen-experienced CD4 T cells as a percentage of CD4 were identified by flow cytometry. Representative plots of the frequency of mediastinal lymph node GC B cells as a percentage of total B cells and the number of mediastinal lymph node GC B cells at day 8 post-infection (IgD⁻ CD95⁺ GL7⁺ B cells). Means ± SDs. Student’s 2-tailed t test, unpaired. (D) Half-maximal effective concentration (EC₅₀) of neutralizing antibody from serum on day 10 post-infection. LOD, limit of detection. Means ± SDs. Student’s 2-tailed t test, unpaired. (E) Influenza NP RNA in homogenates of a whole lung lobe taken at day 10. Means ± SDs. Student’s 2-tailed t test, unpaired. (F) Leder staining of representative lung sections from mice on day 10 post-influenza infection from each group; 200×, scale bar 20 μm. Polymorphonuclear cells stained with pink cytoplasm indicated by arrows. (G) Lung inflammation and neutrophil infiltration scoring. Means ± SDs. Kruskal-Wallis test with correction for multiple comparisons.