Akkermansia muciniphila phospholipid induces homeostatic immune responses

Abstract

Multiple studies have established associations between human gut bacteria and host physiology, but determining the molecular mechanisms underlying these associations has been challenging^1-3. Akkermansia muciniphila has been robustly associated with positive systemic effects on host metabolism, favourable outcomes to checkpoint blockade in cancer immunotherapy and homeostatic immunity^4-7. Here we report the identification of a lipid from A. muciniphila's cell membrane that recapitulates the immunomodulatory activity of A. muciniphila in cell-based assays⁸. The isolated immunogen, a diacyl phosphatidylethanolamine with two branched chains (a15:0-i15:0 PE), was characterized through both spectroscopic analysis and chemical synthesis. The immunogenic activity of a15:0-i15:0 PE has a highly restricted structure-activity relationship, and its immune signalling requires an unexpected toll-like receptor TLR2-TLR1 heterodimer^9,10. Certain features of the phospholipid's activity are worth noting: it is significantly less potent than known natural and synthetic TLR2 agonists; it preferentially induces some inflammatory cytokines but not others; and, at low doses (1% of EC₅₀) it resets activation thresholds and responses for immune signalling. Identifying both the molecule and an equipotent synthetic analogue, its non-canonical TLR2-TLR1 signalling pathway, its immunomodulatory selectivity and its low-dose immunoregulatory effects provide a molecular mechanism for a model of A. muciniphila's ability to set immunological tone and its varied roles in health and disease.

Fig. 1. Structural and functional analysis of A. muciniphila PE.

a, Flow diagram for fractionation of A. muciniphila PE. Amounts in active fractions are shown in red. FA composition of PE fraction also shown. b, TNFα production by mBMDCs treated with A. muciniphila lipid extract fractions as measured by ELISA. The fraction indicated in red was used for structural characterization. Pam3CSK4 was used as a control agonist. Data are presented as mean values ± s.d. of technical replicates (n = 4). c, The structure of a15:0-i15:0 PE. d, The relative abundance of FAs in A. muciniphila PE. e, Dose response of TNFα production by mBMDCs treated with natural (Nat.) and synthetic (Syn.) a15:0-i15:0 PE lipids as measured by ELISA. Data are presented as mean values ± s.d. of technical replicates (n = 4). f, a15:0-i15:0 PE and complete PE (AmPE) trigger release of TNFα and IL-6 but not IL-10 or IL-12p70 from mBMDCs, as measured by flow cytometry. LPS was used as a control. Data are presented as mean values ± s.d. of technical replicates (n = 3). g, TNFα release is lost in TLR2 knockout mBMDCs but not in TLR4 knockout mBMDCs as measured by ELISA. Pam3CSK4 was used as a TLR2 control agonist, and LPS was used as a TLR4 control agonist. Data are presented as mean values ± s.d. of technical replicates (n = 4). All experiments were repeated independently at least twice with similar results. DMSO, dimethyl sulfoxide.

Fig. 2. Biosynthesis and laboratory synthesis of A. muciniphila PE.

a, Key genes involved in the putative biosynthetic pathway for A. muciniphila BAA-835 PE. b, Leucine or isoleucine feeding increases TNFα induction by A. muciniphila in a TLR2-dependent fashion as measured by ELISA. Pam3CSK4 and LPS were used as controls. Data are presented as mean values ± s.d. of technical replicates (n = 4). Unpaired t-test with two-tailed P value; ****P < 0.0001. c, Outline of synthetic scheme for a15:0-i15:0 PE and analogues. d, Overlay of mass spectrometric data from the natural and synthetic a15:0-i15:0. e, TNFα induction by natural and synthetic a15:0-i15:0 PE. a15:0-i15:0 PE induces production in mBMDCs, whereas n14:0-n14:0, n15:0-n15:0, n16:0-n16:0, a15:0-a15:0 and i15:0-i15:0 PE have no detectable TNFα induction, as measured by ELISA. i15:0-a15:0, the positional isomer, shows partial induction. Pam3CSK4 and LPS were used as controls. Data are presented as mean values ± s.d. of technical replicates (n = 4). All experiments were repeated independently at least twice with similar results. Ile, isoleucine; Val, valine; Leu, leucine.

Fig. 3. TLR2–TLR1 binding model and T cell activation by a15:0-i15:0 PE.

a, View of the TLR2–TLR1 complex from the Protein Data Bank (PDB ID 2z7x) with the modelled a15:0-i15:0 PE ligand in the ‘bridging’ conformation, showing the branches with C13 coloured green and C12 purple. b, An overview of the modelled TLR2–TLR1-a15:0-i15:0 PE complex in the surface representation. The dashed circle indicates the buried lipid head group. c, TLR1 and TLR2 are required for natural and synthetic A. muciniphila lipids to induce TNFα production in human monocyte-derived dendritic cells (MDDCs). The production of TNFα was measured by ELISA 18 h after adding natural or synthetic A. muciniphila lipids, Pam3CSK4, FSL-1 or LPS to cell culture media of human MDDCs following nucleofection. d, IL-23A and IL-12B induction by natural and synthetic a15:0-i15:0 PE lipids. e–g, Effects of treatment of human MDDCs with a15:0-i15:0 PE in combination with Pam3CSK4 or LPS. With long (18 h) delay times, low doses of a15:0-i15:0 PE suppress immune responses to Pam3CSK4 and moderate immune responses to LPS (e). Both effects disappear with shorter delay times (3 h in f or none in g). LPS and Pam3CSK4 were used at final concentrations of 100 ng ml⁻¹. Data in c (n = 3), d (n = 6) and e–g (n = 4) are representative of two independent experiments, showing mean values ±  s.d. P values in a were calculated by two-way analysis of variance. *P < 0.05; **P < 0.001; ****P < 0.0001; NS, not significant.

Extended Data Fig. 1. Characterization of a15:0-i15:0 PE.

Key COSY, HSQC, and HMBC correlations of a15:0-i15:0 PE. All experiments were repeated independently at least twice with similar results.

Extended Data Fig. 2. Human-cell activation by PE of A. muciniphila.

a15:0-i15:0 PE induces TNFα, IL-6, IL-10, and IL-12/23p40 from human myeloid cells as measured by flow cytometry. Data are presented as mean values ± s.d. of technical replicates (n = 3). Experiments were repeated independently at least twice with similar results.

Extended Data Fig. 3. Biosynthetic pathway of a15:0-i15:0 PE in A. muciniphila.

Proposed biosynthesis of a15:0-i15:0 PE in A. muciniphila BAA-835.

Extended Data Fig. 4. Biosynthetic pathways of branched-chain amino acids in A. muciniphila.

De novo biosynthesis of l-leucine, l-isoleucine, and l-valine in A. muciniphila BAA-835.

Extended Data Fig. 5. TNFα activation of straight chain fatty acids.

TNFα production by mBMDCs treated with single chain fatty acids (from C17:0 to C13:0) as measured by ELISA. LPS was used as a control agonist. Data are representative of at least two independent experiments and show mean values ± s.d. of technical replicates (n = 4).

Extended Data Fig. 6. Heat-map of human derived monocyte activation by natural and synthetic A. muciniphila lipids.

Monocytes were purified from peripheral blood by negative selection with magnetic beads. Cells were immediately placed in culture and stimulated with the indicated stimuli for 6 h. After stimulation, cells were lysed for mRNA extraction and library preparation by SmartSeq2 for RNA sequencing. Monocytes were processed from two independent donors in technical triplicate. The final concentration of bacterial lipids was 50 μg/ml. Pam3CSK4, FSL-1, and LPS at a final concentration of 100 ng/ml were used as controls.

Extended Data Fig. 7. Effects of treatment with synthetic A. muciniphila lipids in combination with Pam3CSK4 or LPS on human monocyte-derived dendritic cells.

With long (18 h) delay times, low doses of synthetic a–c, a15:0-i15:0 PE or d–f, i15:0-a15:0 PE suppress immune responses to Pam3CSK4 and moderate immune responses to LPS. Both effects disappear with shorter delay times (3 h or none). LPS and Pam3CSK4 were used at final concentrations of 100 ng/mL. Data are presented as mean values ± s.d. of technical replicates (n = 4). Experiments were repeated independently at least twice with similar results.