Engineering Mycoplasma pneumoniae to bypass the association with Guillain-Barré syndrome

Abstract

A non-pathogenic Mycoplasma pneumoniae-based chassis is leading the development of live biotherapeutic products (LBPs) for respiratory diseases. However, reports connecting Guillain-Barré syndrome (GBS) cases to prior M. pneumoniae infections represent a concern for exploiting such a chassis. Galactolipids, especially galactocerebroside (GalCer), are considered the most likely M. pneumoniae antigens triggering autoimmune responses associated with GBS development. In this work, we generated different strains lacking genes involved in galactolipids biosynthesis. Glycolipid profiling of the strains demonstrated that some mutants show a complete lack of galactolipids. Cross-reactivity assays with sera from GBS patients with prior M. pneumoniae infection showed that certain engineered strains exhibit reduced antibody recognition. However, correlation analyses of these results with the glycolipid profile of the engineered strains suggest that other factors different from GalCer contribute to sera recognition, including total ceramide levels, dihexosylceramide (DHCer), and diglycosyldiacylglycerol (DGDAG). Finally, we discuss the best candidate strains as potential GBS-free Mycoplasma chassis.

Keywords: Galactocerebrosides; Glycolipids; Glycosyltransferases; Immune response; Molecular mimicry; Mycoplasma pneumoniae.

Fig. 1

Glycolipids metabolism of M. pneumoniae and its essentiality. Enzymes (inside circles) are coloured according to their essentiality (red for essential, blue for fitness and green for non-essential genes). Gene identifier is also included when the protein activity has been linked to a specific gene in M. pneumoniae. Predicted protein activities (Lip: lipase, Pap: phosphatidate phosphatase) lack gene identifier and are coloured in black as their essentiality character is not known. A scheme depicting the structure of the most relevant compounds is shown close to their names. An asterisk (∗) highlights compounds demonstrated as precursors for galactolipids by M. pneumoniae. Compounds in the map: PC, Phosphatidylcholine; Gly-3-P, Glycerol-3-Phosphate; LysoPA, Lyso Phosphatidic Acid; PA, Phosphatidic Acid; Acyl-P, Acyl-phosphate; Glc-6-P, Glucose-6-Phosphate; Glc-1-P, Glucose-1-Phosphate; DAG, Diacylglycerol. Glycolipids: MGDAG, MonoglycosylDAG (GalDAG or GlcDAG); DGDAG, DiglycosylDAG; TGDAG, TriglycosylDAG; MHCer, Cerebrosides (GalCer or GlcCer); DHCer, Dihexosylceramide; THCer, Trihexosylceramide.

Fig. 2

Growth kinetics of the different strains engineered in this work. Growth was monitored by pH changes in the medium due to lactate and acetate secretion by M. pneumoniae. Graphs show growth kinetics (growth index corresponds to the ratio Abs430nm/Abs560nm) of (A) the three knock out strains Δ257, Δ483 and Δ028 compared to the M. pneumoniae M129 wild-type strain (WT); (B) the three strains Δ483::GT, Δ483::AG and Δ483::BS with mpn483 replacement by an orthologous gene compared to the WT and Δ483 strains; (C) the double mutant strains Δ257Δ483 and Δ257Δ483::prGT; compared to the WT strain. Mean values from three bio-replicates. Error bars indicate standard deviation (SD). Source and processed data provided in the suppl. File F1.

Fig. 3

Glycolipid profiling of the engineered strains by liquid chromatography-high resolution mass spectrometry (LC-HRMS). Quantification of diacylglycerol (DAG), monoglycosylDAG (MGDAG), diglycosylDAG (DGDAG), ceramide, total cerebrosides (MHCer) and dihexosylceramide (DHCer) in mycoplasma extracts from strains Δ028, Δ257, Δ483, Δ483::GT, Δ483::AG, Δ483::SB, Δ257Δ483::prGT, Δ257Δ483 and M. pneumoniae M129 wild-type strain (WT). Galactocerebroside (GalCer) and glucocerebroside (GlcCer) were also quantified individually. Graphs show lipid concentration in pmol equivalent per mg of protein in the mycoplasma extracts (pmol_e/mg_prot). Mean ± SD from a minimum of two bio-replicates. T-test to compare results to WT sample (values considered statistically significant with two-sided P-value<0.05 are indicated with an asterisk; P-values are listed in Suppl. Table T1). Source and processed data provided in the Suppl. file F3.

Fig. 4

Inhibition ELISA assay. (A) Scheme depicting the steps followed to evaluate cross-reactivity of IgG anti-GalCer positive MpGBS sera with the engineered strains using an inhibition ELISA assay. Sera enriched in GalCer antibodies (blue shapes) is incubated with WT or engineered strains. The unbound fraction of these incubations is later added to GalCer coated plates to determine the recognition profile of MpGBS sera to each strain (B) Cross-reactivity of IgG anti-GalCer with M. pneumoniae strains was assessed by incubating anti-GalCer reactive serum samples with increasing doses of whole-mycoplasma extracts of strains WT (M129), Δ257, Δ483, Δ483::GT, Δ483::AG, Δ483::SB, Δ257Δ483::prGT, Δ257Δ483 and two types of C. jejuni LOS (GB2 and GB19). Graph shows the mean ± SD percent inhibition compared to the reference (serum incubated with PBS without mycoplasma extract was defined as 0% inhibition) from a minimum of two biological replicas, obtained from two MpGBS patient serum samples (see Suppl. Fig. S7. T-test to compare results to WT sample (values considered statistically significant with two-sided P-value<0.05 are indicated with an asterisk; and with P-value<0.001 with a triangle; P-values are listed in Suppl. Table T1). Source and processed data provided in the suppl. File F4. (C) Correlation of results from inhibition ELISA assays with the abundance of the indicated glycolipid species in the different strains. Pearson's correlation coefficient (r) and R-squared coefficient of determination (R²) are shown for each linear regression calculated.