Pregnancy enables antibody protection against intracellular infection

Abstract

Adaptive immune components are thought to exert non-overlapping roles in antimicrobial host defence, with antibodies targeting pathogens in the extracellular environment and T cells eliminating infection inside cells^1,2. Reliance on antibodies for vertically transferred immunity from mothers to babies may explain neonatal susceptibility to intracellular infections^3,4. Here we show that pregnancy-induced post-translational antibody modification enables protection against the prototypical intracellular pathogen Listeria monocytogenes. Infection susceptibility was reversed in neonatal mice born to preconceptually primed mothers possessing L. monocytogenes-specific IgG or after passive transfer of antibodies from primed pregnant, but not virgin, mice. Although maternal B cells were essential for producing IgGs that mediate vertically transferred protection, they were dispensable for antibody acquisition of protective function, which instead required sialic acid acetyl esterase⁵ to deacetylate terminal sialic acid residues on IgG variable-region N-linked glycans. Deacetylated L. monocytogenes-specific IgG protected neonates through the sialic acid receptor CD22^6,7, which suppressed IL-10 production by B cells leading to antibody-mediated protection. Consideration of the maternal-fetal dyad as a joined immunological unit reveals protective roles for antibodies against intracellular infection and fine-tuned adaptations to enhance host defence during pregnancy and early life.

Extended Data Fig. 1:. Pregnancy enables antibody mediated protection against Lm infection.

(a, b) Bacterial burden (a) and anti-Lm IgG titer (b) in neonatal mice infected with virulent Lm born to WT naive mice or mice preconceptual primed with ΔActA Lm once or twice 2 weeks apart. (c) C. albicans fungal burden in neonatal mice born to WT female mice primed with attenuated ΔActA Lm one week prior to mating or naive control mice. Pups were infected with virulent C. albicans 3 days after birth, with enumeration of pathogen burden 48 hours post-infection. (d) Bacterial burden 72 hours post-infection with virulent Lm in adult WT, μMT−/− or CD8−/− mice with or without ΔActA Lm priming 4 weeks prior. (e) Anti-Lm IgG titer in adult CD8−/− mice 3 days after primary Lm infection compared with secondary challenge of ΔActA Lm-primed virgin female mice or preconceptually ΔActA Lm-primed CD8−/− female mice 3 weeks post-partum. Each symbol represents an individual mouse, with graphs showing data combined from at least 2 independent experiments each with 3-5 mice per group per experiment. Bar, mean ± standard error. P values between key groups are shown as determined by one-way ANOVA adjusting for multiple comparisons. Dotted lines, limit of detection.

Extended Data Fig. 2:

Anti-Lm antibodies acquire protective function during pregnancy through maternal FcγR-expressing cells. (a) Anti-Lm IgG titer in neonatal mice 72 hours after virulent Lm infection that were transferred sera from ΔActA Lm-primed virgin (vSera) or sera from late gestation/early post-partum (pSera) mice. (b) Cross-fostering schematic and anti-Lm IgG titer in each group of neonatal mice 72 hours after virulent Lm infection. (c) Anti-Lm IgG titer in neonatal mice nursed by WT, μMT−/−, or FcRγ−/− mice administered vSera on the day of delivery and 3 days later. (d) Anti-Lm IgG titers and bacterial burdens after virulent Lm infection in neonatal mice administered vSera that had been incubated with splenocytes from virgin or pregnant (E18) WT or FcRγ−/− mice for 48 hours. Pups were infected with virulent Lm 3-4 days after birth, 24 hours after antibody transfer, with enumeration of bacterial burden 72 hours post-infection. Each symbol represents an individual mouse, with graphs showing data combined from at least 2 independent experiments each with 3-5 mice per group per experiment. Bar, mean ± standard error. P values between key groups are shown as determined by one-way ANOVA adjusting for multiple comparisons. Dotted lines, limit of detection.

Extended Data Fig. 3:. Lm-specific IgG titer, subclass distribution, neutralization efficiency, and lectin staining profile comparisons in virgin compared with pregnant mice.

(a) IgG titers against UV-inactivated Lm bacteria or purified LLO protein for WT female pregnant/postpartum mice preconceptually primed with ΔActA Lm, ΔActA Lm-primed virgin mice, or naive control mice. (b) OD₄₅₀ for each antibody isotype or subclass in sera from ΔActA Lm-primed virgin or pregnant mice. All sera diluted 1:100. Background subtracted from staining using naive sera. (c) LLO-induced hemolysis of sheep erythrocytes in media containing anti-Lm IgG from virgin (vIgG) or late gestation/early postpartum pregnant (pIgG) mice. (d) Relative vIgG and pIgG binding to each lectin with defined carbohydrate-specificity for LLO-specific IgG. (e,f) SNA-agarose fractionation of sialylated (SNA Bound) compared with non-sialylated (SNA Unbound) anti-Lm vIgG. SNA lectin staining confirmed successful separation (e). Lm- and LLO-specific IgG was titered from each fraction (f). (g, h) Anti-Lm IgG titer in neonatal mice transferred each type of enzymatically-treated pIgG (g). Lectin staining to confirm de- and resialylation of pIgG: SNA signal indicates presence of terminal sialic acid and ECA signal indicates presence of galactose uncovered by terminal sialic acid removal (h). (i-k) Anti-Lm IgG titer in neonatal mice transferred glycoengineered vIgG (i). SNA and ECA Lectin staining to confirm de- and resialylatoin of vIgG with sialic acid variants (j). PToV virolectin probe for 9-O-acetylated sialic acid and PToV with mutated binding site (nonbinder) demonstrating specificity for vIgG resialylated with each sialic acid variant or deacetylated with SIAE treatment (k). Representative data from at least 2-3 independent experiments (a-f) or combined data from 3 independent experiments with 3-5 mice per group per experiment (g-k) are shown. Bar, mean ± standard error. P values between key groups are shown as determined by one-way ANOVA adjusting for multiple comparisons. Dotted lines, limit of detection.

Extended Data Fig. 4:. Deacetylated sialic acid on Lm-specific IgG enables antibody-mediated protection.

(a, b) Anti-Lm IgG titer 72 hours post-infection (a) and survival in neonatal mice transferred SIAE- or mock-treated anti-Lm IgG from ΔActA Lm-primed virgin mice (b). (c, d) Bacterial burden (c) and anti-Lm IgG titer in neonatal mice infected with virulent Lm transferred SIAE- or mock-treated anti-Lm IgG from preconceptual ΔActA Lm-primed late gestation/early postpartum mice (d). (e) CCA lectin detection of 9-O-Acetyl-Neu5Gc on LLO-specific IgG obtained 72 hours after transferring sera from virgin ΔActA Lm-primed mice (vSera) to WT or SIAE KO postpartum recipients on the day of delivery and 3 days later (top). Anti-Lm IgG titer in neonatal mice nursed by these females 72 hours post-infection with virulent Lm (bottom). (f) Anti-Lm IgG titers (top) and bacterial burdens (bottom) in virulent Lm infected neonatal mice nursed by WT or ST6Gal1 KO mice administered vSera as described in panel (e). Pups were infected with virulent Lm 3-4 days after birth, 24 hours after antibody transfer, with enumeration of bacterial burden 72 hours post-infection. Each symbol represents an individual mouse, with graphs showing data combined from at least 2 independent experiments each with 3-5 mice per group per experiment. Bar, mean ± standard error. P values between key groups are shown as determined by one-way ANOVA adjusting for multiple comparisons (a, c, d, f), or Log-rank (Mantel-Cox) test (b). Dotted lines, limit of detection.

Extended Data Fig. 5:. Oxonium ion searches for sialic acid modifications.

(a) Schematic for purification of Lm-specific IgG from ΔActA Lm-primed virgin and pregnant mice for mass spectroscopy analysis. (b) Oxonium ion m/z used to search for sialic acid variants. Diagnostic ions were filtered through the MS/MS spectra to select glycopeptides containing Acetyl-Neu5Gc (Neu5Gc,Ac). Example spectra shown for virgin Lm-specific IgG. (c) MS/MS fragmentation for the low mass region (150-500) demonstrating sialic acid variants. All presented m/z are z=1. Experiment was performed in triplicate with representative data shown.

Extended Data Fig. 6:. Glycoforms on IgG2 Fc conserved region N-glycans do not contain acetylated sialic acid.

Lm-specific IgG from ΔActA Lm-primed virgin and pregnant mice was subjected to trypsin or chymotrypsin digestion and then LC-MS/MS analysis. (a) Representative MS/MS fragmentation from the conserved N-linked glycosylation site at position 183 on the Fc region of IgG2b/c (glycopeptide EDYNSTIR). (b) Extracted ion chromatograph of EDYNSTIR with the observed sialic acid (Neu5Gc) glycoforms and the absence of Acetyl-Neu5Gc, which was confirmed for the entire length of the LC-MS run. (c, d) All glycoforms with observed masses for N-glycans on the conserved Fc region of IgG2b/c for virgin (c) and pregnant (d) Lm-specific IgG. Experiment was performed in triplicate with representative (a, b) and combined (c, d) data shown.

Extended Data Fig. 7:. Acetylated sialic acid localizes to the IgG Fab variable region.

(a) Extracted ion chromatograph showing m/z corresponding to Acetyl-Neu5Gc (Neu5Gc,Ac) and m/z of the same glycopeptides containing Neu5Gc. Examples of two individual glycopeptides are shown. (b) Representative MS/MS fragmentation of glycopeptide N₁E₂S₁L₁Y₁T₁A₂I₁K₁ demonstrating presence of Acetyl-Neu5Gc (Neu5Gc,Ac). (c) For Fab glycans, the exact amino acid position and sequence are unknown, and shown are potential amino acid compositions given the observed mass size. Experiment was performed in triplicate with representative (a, b) and combined (c) data shown .

Extended Data Fig. 8:. Fc-mediated functions are dispensable for antibody-mediated protection against Lm.

(a) Presence of IgG heavy (H) and light (L) chains or IgG2b Fc for pF(ab’)₂ confirming efficient Fc removal after pepsin digestion. (b) Lectin staining for pF(ab’)₂ compared with full-length pIgG for anti-LLO antibodies showing Fab glycosylation. SNA binds α2,6-linked sialic acid residues, while AAL binds α1,6-linked fucose. (c) SNA lectin blot comparing full-length IgG under reducing conditions to separate heavy and light chains, with purified F(ab’)₂ fragments under non-reducing conditions, demonstrating preserved SNA lectin staining. (d) Bacterial burden after virulent Lm infection in neonatal mice transferred sera from preconceptual ΔActA Lm-primed pregnant/postpartum (pSera) or naive control mice, along with anti-CD16/32 blocking or isotype control antibodies. (e, f) Bacterial burden after virulent Lm infection in neonatal WT, complement C1q-deficient mice (e) or complement C3-deficient (f) transferred pSera or sera from naive mice. Pups were infected with virulent Lm 3-4 days after birth, 24 hours after antibody transfer, with enumeration of bacterial burden 72 hours post-infection. Each symbol represents an individual mouse, with graphs showing data combined from at least 2 independent experiments each with 3-5 mice per group per experiment. Gel is representative of results from 3 independent experiments each with similar results (c). Bar, mean ± standard error. P values between key groups are shown as determined by one-way ANOVA adjusting for multiple comparisons. Dotted lines, limit of detection.

Extended Data Fig. 9:. Deacetylated anti-Lm antibodies protect via CD22-mediated suppression of B cell IL-10 production.

(a) Anti-Lm antibody titers in neonatal WT or CD22-deficient mice transferred sera from ΔActA Lm-primed pregnant/postpartum (pSera) or naive mice. (b, c) Bacterial burden (b) or anti-Lm IgG titers (c) after virulent Lm infection in neonatal mice transferred pSera or sera from naive mice along with anti-CD22 blocking or isotype control antibodies. (d, e) Bacterial burden (d) or anti-Lm IgG titers (e) after virulent Lm infection in neonatal mice transferred vIgG glycoengineered to express Neu5Ac. Neonates received either isotype control mAb or anti-CD22 mAb, or were CD22 deficient. (f, g) Bacterial burden (f) and anti-Lm IgG titers (g) after virulent Lm infection in WT or CD19−/− neonatal mice transferred vSera or sera from naive mice. (h) Anti-Lm IgG titers in neonatal WT or μMT−/− mice transferred sera from ΔActA Lm-primed virgin (vSera) or pSera. Each symbol represents an individual mouse, with graphs showing data combined from at least 2 independent experiments each with 3-5 mice per group per experiment. Bar, mean ± standard error. P values between key groups are shown as determined by one-way ANOVA adjusting for multiple comparisons. Dotted lines, limit of detection.

Extended Data Fig. 10:. Deacetylated anti-Lm antibodies protect via CD22-mediated suppression of B cell IL-10 production.

(a) Representative FACS plots showing GFP expression by B220⁺IgM⁺CD5⁺ splenocytes 72 hours after virulent Lm infection in neonatal IL10-eGFP reporter mice administered vSera or pSera along with anti-CD22 neutralizing or isotype control antibodies, compared with no infection control mice. (b) Representative FACS plots showing GFP expression by B220⁺IgM⁺CD5⁺ splenocytes from IL10-eGFP reporter mice stimulated with UV-inactivated Lm for 20 hours in the presence of vIgG or pIgG plus anti-CD22 neutralizing antibody or isotype control. (c) GFP expression by B220⁺IgM⁺CD5⁺ splenocytes from IL10-eGFP reporter mice after UV-Lm stimulation plus anti-CD22 neutralizing antibody or isotype control without anti-Lm IgG. (d) GFP expression by B220⁺IgM⁺CD5⁺ splenocytes from IL10-eGFP reporter mice stimulated with the TLR2 agonist Pam3CSK4 in the presence of vIgG or pIgG. (e) Anti-Lm antibody titers in neonatal mice transferred vSera or sera from naive mice along with anti-IL10 receptor blocking or isotype control antibodies. (f) Model of pregnancy-induced deacetylation of anti-Lm antibodies that unleash their protective function by overriding IL-10 production by B10 cells via CD22. Each symbol represents the data from cells in an individual well under unique stimulation conditions combined from 4-5 independent experiments (c, d), or individual mice (e) with graphs showing data combined from at least 2 independent experiments with 3-5 mice per group per experiment. Bar, mean ± standard error. P values between key groups are shown as determined by one-way ANOVA adjusting for multiple comparisons.

Fig. 1.. Anti-Lm antibodies acquire protective function during pregnancy.

(a-c) Anti-Lm IgG titers (a), survival (b) and bacterial burdens (c) in neonatal mice infected with virulent Lm born to WT, μMT−/− or CD8−/− female mice primed with attenuated ΔActA Lm one week prior to mating or naive WT control mice without preconceptual priming. (d) Bacterial burden in neonatal mice infected with virulent Lm born to WT female mice primed with ΔActA Lm or virulent Lm 3 weeks prior to mating. (e) Bacterial burden 3 days after secondary challenge of ΔActA Lm-primed virgin female CD8−/− adult mice or preconceptually ΔActA Lm-primed CD8−/− female mice 3 weeks post-partum. (f) Bacterial burden after virulent Lm infection in neonatal mice transferred sera from ΔActA Lm-primed virgin (vSera) or pregnant mice in late gestation through early postpartum (pSera). (g) Bacterial burden after virulent Lm infection in neonatal mice transferred sera from preconceptual ΔActA Lm-primed mice harvested during the indicated pregnancy or postpartum time points. (h) Bacterial burden after virulent Lm infection in neonatal mice cross-fostered within 12 hours after birth by naive or ΔActA Lm-primed dams. (i) Bacterial burden after virulent Lm infection in neonatal mice nursed by WT, μMT−/−, or FcRγ−/− mice administered vSera on the day of delivery and 3 days later. Pups were infected with virulent Lm 3-4 days after birth, with enumeration of bacterial burden 72 hours post-infection. Each symbol represents an individual mouse, with graphs showing data combined from at least 2 independent experiments each with 3-5 mice per group per experiment. Bar, mean ± standard error. P values between key groups are shown as determined by one-way ANOVA adjusting for multiple comparisons (a, c, d, f, g, h, i), unpaired t-test (e), or Log-rank (Mantel-Cox) test (b). Dotted lines, limit of detection.

Fig. 2.. SIAE deacetylates anti-Lm IgG sialic acid enabling antibody-mediated protection.

(a) IgG N-linked glycosylation schematic including sialic acid variants and SIAE activity. (b) IgG from naive (nIgG) or preconceptual ΔActA Lm-primed pregnant/postpartum (pIgG) mice was glycoengineered to remove existing sialic acid with neuraminidase (pIgG^Neur) followed by resialylation with ST6Gal1 plus CMP-Neu5Ac (pIgG^Neu5Ac). Bacterial burdens after virulent Lm infection in neonatal mice following transfer of indicated pIgG preparations. (c) IgG recovered from naive (nIgG) or ΔActA Lm-primed virgin (vIgG) mice, treated with neuraminidase, then resialylated with each sialic acid variant using ST6Gal1 plus CMP-sialic acid donors. Bacterial burdens after Lm infection in neonatal mice following transfer of indicated vIgG preparations. (d) PToV virolectin and CCA lectin detection of 9-O-Acetyl-Neu5Gc in virgin compared with pregnant LLO-specific IgG. (e) Relative Siae expression in splenocytes from virgin and late gestation (E18-20) pregnant mice, or in PBMCs from non-pregnant (NonPreg) versus pregnant women (12-32 weeks gestation). (f) Bacterial burden after Lm infection in neonatal mice administered SIAE- or mock-treated vIgG. (g) Bacterial burden after Lm infection in neonatal mice nursed by WT or SIAE−/− (black outline, exon3−/−; no outline, exon3/4−/−) mice administered vSera on the day of delivery and 3 days later. Pups were infected with virulent Lm 3-4 days after birth, and where indicated, 24 hours after antibody transfer, with enumeration of bacterial burden 72 hours post-infection. Each symbol represents an individual mouse, with graphs showing data combined from at least 2 independent experiments each with 3-5 mice per group per experiment (b, c, e-g), or representative data from >3 independent experiments (d). Bar, mean ± standard error. P values between key groups were determined by one-way ANOVA adjusting for multiple comparisons (b, c, f, g) or unpaired t-test (d, e). Dotted lines, limit of detection.

Fig. 3.. Acetylated sialic acid localizes to Fab N-glycans.

(a) Percentage IgG2 conserved Fc N-glycan sites with each indicated glycoform. (b) Relative mass peak area for Acetyl-Neu5Gc determined by tandem mass spectroscopy oxonium ion filtering on three unique Fab glycopeptides from Lm-specific IgG purified from virgin or pregnant mice. (c) Fold-reduction in Acetyl-Neu5Gc for each Fab glycopeptide in pregnant compared with virgin Lm-specific IgG described in panel b. (d) Fold increase in Acetyl-Neu5Gc after pepsin treatment for Fab glycopeptides relative to untreated full-length IgG. (e, f) Bacterial burdens (e) and anti-Lm F(ab’)₂ titers (f) after virulent Lm infection in neonatal mice transferred pepsin-treated IgG from: naive mice [nF(ab’)₂], pregnant/postpartum Lm-primed mice [pF(ab’)₂], or virgin Lm-primed mice [vF(ab’)₂] treated with SIAE to remove acetylation from sialic acid residues. Pups were infected with virulent Lm 4 days after birth, 24 hours after F(ab’)₂ transfer, with enumeration of bacterial burden 72 hours post-infection. Mass spectroscopy experiments were performed 3-4 times using purified IgG pooled from virgin or pregnant mice and data from each replicate experiment is shown (a-d). For infection and anti-Lm titers, each symbol represents an individual mouse, with graphs showing data combined from at least 2 independent experiments each with 3-5 mice per group per experiment (e, f). Bar, mean ± standard error. P values between key groups are shown as determined by one-way ANOVA adjusting for multiple comparisons (e, f) or unpaired t-test (c, d).

Fig. 4.. Deacetylated anti-Lm antibodies protect via CD22-mediated suppression of B cell IL-10 production.

(a) CD22+ cell distribution amongst neonatal splenocytes 72 hours after Lm infection compared with uninfected controls. (b) Bacterial burden after Lm infection of WT or CD22-deficient neonatal mice transferred sera from ΔActA Lm-primed pregnant/postpartum (pSera) or naive mice. (c) Bacterial burden after Lm infection of WT or μMT−/− neonates administered pSera or naive sera, along with anti-CD22 or isotype control antibodies. (d) Bacterial burden after Lm infection of WT or μMT−/− neonates administered sera from ΔActA Lm-primed virgin mice (vSera), pSera, or naive sera. (e) Serum IL-10 levels 72 hours after Lm infection in WT or μMT−/− neonates administered vSera or pSera, plus CD22 neutralization . (f) B220⁺IgM⁺CD5⁺ splenocyte GFP expression 72 hours after Lm infection in neonatal IL10-eGFP reporter mice administered vSera or pSera, plus CD22 neutralization, compared with uninfected mice. (g) GFP expression by B220⁺IgM⁺CD5⁺ splenocytes from IL10-eGFP neonates stimulated with UV-inactivated Lm in the presence of vIgG or pIgG, plus CD22 neutralization. (h) Bacterial burden after Lm infection in neonates administered vSera or naive sera, along with anti-IL10 receptor or isotype control antibodies. Neonatal mice were infected with virulent Lm 3-4 days after birth, 24 hours after antibody transfer, with enumeration of bacterial burden 72 hours post-infection. Pie charts show representative data from one of 3 independent experiments, averaged across 5 mice per group (a). Each symbol represents the data from an individual mouse (b-f, h) with graphs showing data combined from at least 3 independent experiments with 3-5 mice per group per experiment, or cells in an individual well under unique stimulation conditions combined from 4-5 independent experiments (g). Bar, mean ± standard error. P values between groups determined by one-way ANOVA adjusting for multiple comparisons. Dotted line indicates limit of detection.