Bacterial expression of a designed single-chain IL-10 prevents severe lung inflammation

Abstract

Interleukin-10 (IL-10) is an anti-inflammatory cytokine that is active as a swapped domain dimer and is used in bacterial therapy of gut inflammation. IL-10 can be used as treatment of a wide range of pulmonary diseases. Here we have developed a non-pathogenic chassis (CV8) of the human lung bacterium Mycoplasma pneumoniae (MPN) to treat lung diseases. We find that IL-10 expression by MPN has a limited impact on the lung inflammatory response in mice. To solve these issues, we rationally designed a single-chain IL-10 (SC-IL10) with or without surface mutations, using our protein design software (ModelX and FoldX). As compared to the IL-10 WT, the designed SC-IL10 molecules increase the effective expression in MPN four-fold, and the activity in mouse and human cell lines between 10 and 60 times, depending on the cell line. The SC-IL10 molecules expressed in the mouse lung by CV8 in vivo have a powerful anti-inflammatory effect on Pseudomonas aeruginosa lung infection. This rational design strategy could be used to other molecules with immunomodulatory properties used in bacterial therapy.

Keywords: infection; interleukin; live biotherapeutics; mycoplasma; protein engineering.

Figure 1. Effects of human IL‐10 recombinant protein (hIL‐10r) or IL‐10 secreted by M. pneumoniae on macrophage primary cells on anti‐inflammatory (A) or pro‐inflammatory biomarkers (B)

1. A, B

   Data were normalised using the control group and are represented as fold‐change (FC) of the mean fluorescence intensity (MFI). Macrophage cells were obtained from human donors (n = 4 biological replicas) and incubated with medium (CON) or recombinant human IL‐10 (hIL‐10r), MPN wild‐type (WT) or MPN expressing IL‐10 ORF (WT_IL10). Data are represented as mean ± standard error of the mean (SEM). Statistical analysis was performed using one‐way ANOVA + post hoc Tukey's multiple comparison test (*P < 0.05; **P < 0.001, ***P < 0.0001).

Source data are available online for this figure.

Figure EV1. Analysis of SC mutants designed in this work

1. A, B

   Detection of phosphorylated Tyr 705 of STAT3 (p‐STAT3) and unphosphorylated STAT3 by Western blot (see Materials and Methods). (A) Data from macrophages isolated from two independent donors, showing unstimulated cells (control, CON) or cells incubated for 24 h with human IL‐10 recombinant protein (hIL‐10r), supernatant of MPN WT (WT) or of MPN expressing IL‐10 (WT_IL‐10 ORF). (B) Data from HAFTL murine B‐cell line after exposure for 20 min to the supernatant of MPN WT (WT) or MPN expressing IL‐10 (WT_IL‐10 ORF).

2. C

   HEK‐Blue™ reporter cell activation dose–response analysis by MutSC1 (linker NGGLD) and MutSC1_Gly (linker GGGGG) supernatants. The x‐axis shows the range of IL‐10 concentration analysed (Molar, M), and the y‐axis represents the mean ± SD of the absorbance at 630 nm. Data were generated in three independent assays with two technical replicas (n > 6).

3. D

   Western blot of p‐STAT3 activation after 20 min of induction with a fixed IL‐10 concentration (20 ng/ml) of supernatants from MPN WT (WT) or MPN expressing IL‐10 WT (WT_IL10 ORF), MutSC1 (WT_MutSC1), MutSC2 (WT_MutSC2) in two different cell lines: THP‐1 (human monocyte) and HAFTL (murine pre‐B cell line).

Figure 2. Schematics of all mutations generated in this work to increase IL‐10/R1/R2 receptor affinity

Details for each of the positions mutated to improve interactions with R1 are shown in the individual panels. Note that when aspartic acid‐28 is mutated to glutamate in IL‐10, an electrostatic displacement occurs over the arginine‐24, allowing it to form an H‐bond with glutamate‐145 from the receptor 1 (yellow double arrow in the subfigure indicates D28E).

Figure EV2. Multiple sequence alignment of IL‐10 from different vertebrate species performed with the ClustalX algorithm

The first residue corresponds to the first residue of IL‐10 in the crystal structure 1y6k.

Figure EV3. Superimposition of the crystal structures of IL‐10 with R1 (1yl6k) and of IL‐10 with both R1 and R2 (6x93)

Unstructured regions adopted a different conformation when interacting with R2. Gorby (9)‐mutated positions (N19, N92, K99) are zoomed‐in on the right side in the crystallographic superimposition with cryo‐EM (9), showing an appreciable backbone shifting upon binding to R2 in some of the positions.

Figure 3. Schematic depiction of the design of single‐chain (SC) IL‐10s

IL‐10L and IL‐10M monomers are respectively shown in blue and magenta. To avoid confusion, IL‐10M residue numbers are denoted by adding the prime (′) character.

1. Steps to generate two different sewing patterns: (1) deletion of fringe residues; (2) rewiring schema; (3) structural rearrangement after rewiring the corresponding regions; (4) final numbering in the SC, including numerical gaps long enough to host peptide bridges with different lengths (up to 20) during linker search or bridging. Monomer 2 (IL‐10M, in magenta) residue numbers are marked by '.

2. MutSC1 and MutSC2 built from the IL‐10 sewing pattern 1. Linker sequences are shown as red labels; post‐rewiring mutations are included for MutSC2 using monomeric IL‐10 numbering.

Figure 4. Expression levels and apparent dissociation constant of selected IL‐10 variants expressed by M. pneumoniae

Each point in the figure is a biological replica.

1. A

   EC‐50 (molar, M) for human IL‐10 recombinant protein (hIL‐10r) and IL‐10 WT (IL‐10 ORF) and different variants (MutM, hIL‐10r, Mut1, Mut2, Mut3, MutSC1 and MutSC2) expressed by M. pneumoniae. Data are represented as mean ± SD. Statistical comparison was done by one‐way ANOVA + post hoc Bonferroni multiple comparison test, using the IL‐10 ORF condition as a reference (*P < 0.05).

2. B

   Fold‐change (FC) in expression levels (fg/CFU) of IL‐10 variants Mut3 and MutSC1 secreted to the medium normalised by expression level of IL‐10 ORF (ORF). Statistical comparison of mean ± SD was performed by one‐way ANOVA + Tukey's post hoc test (*P < 0.05).

3. C, D

   Average and SD values for the FC in the relative EC‐50 values determined by flow cytometry analysis of phosphorylated STAT3 after a 20‐min exposure of the BlaER1 (C) or HAFTL (D) cell lines to IL‐10 ORF (reference) or the mutant MutSC1 or MutSC2 (see Materials and Methods). Numbers indicate the average FC ± SD.

Source data are available online for this figure.

Figure 5. Characterisation of CV8 chassis in vivo

C57Bl/6 were infected with 10⁷ CFU of WT or CV8 strains and sacrificed at 2 days post‐infection (dpi) or 4 dpi (n ≥ 3 biological replicas/group).

1. Average ± SD bacterial loads recovered of the WT (circle) and CV8 (square) strains in mouse lungs at 2‐ or 4‐dpi. Data are shown as log₁₀ CFU/lung homogenate.

2. Inflammatory profile of mouse lungs inoculated with WT (dark blue), CV8 (light blue) or PBS (grey). Gene expression was analysed by RT–qPCR using gapdh as an endogenous control (see Materials and Methods). Data are shown as average ± SD of fold‐change (FC) in mRNA expression (one‐way ANOVA + Tukey's post hoc test; *P < 0.05).

3. Histological findings of mouse lung samples at the analysed time points. Left, plots representing the quantitative evaluation (score: 0–5) of alveolar infiltrate (top panel) and peribronchial/solar infiltrate (FC; bottom panel; see Materials and Methods). Parameters were normalised using the average of the PBS group (FC = sample value/average PBS; FC control group, ~1). Data are shown as average ± SD of FC (one‐way ANOVA + Tukey's post hoc test; *P < 0.05). Right, representative images of lungs stained with haematoxylin–eosin (the line represents 250 μm).

4. IL‐10 secretion by the MPN WT or CV8 strain coding for hIL‐10 WT (ORF), MutSC1 or MutSC2. Data are expressed as the mean ± SD of IL‐10 supernatant concentration (μg/ml) by biomass (protein content in μg/ml; n = 3 biological replicas).

Source data are available online for this figure.

Figure EV4. Analysis of P. aeruginosa PAO1 infection of mice lungs

1. Pseudomonas aeruginosa PAO1 bacterial load obtained from mice infected with 10⁴ or 10⁵ CFU at 24‐ or 48 h post‐infection (hpi). Data are shown as mean ± SD of Log₁₀ CFU/lung homogenate of at least 3 mice per group (n < 3 biological replica). Statistical analysis was performed using one‐way ANOVA + Tukey's post hoc test (*P < 0.05).

2. Fold‐change in mRNA expression of different inflammatory markers in the lung of mice infected with PAO1 at 24 hpi.

Figure 6. Immunomodulatory effect of IL‐10 variants in vivo

Each point is a biological replica.

1. Schematic representation of the experimental design performed.

2. CFUs of PAO1 (left) or Mycoplasma strains (right) recovered from lung samples. Data are shown as mean ± SD of Log₁₀ CFU/lung homogenate. Two replicates were performed and individual symbols are shown (experiment 1, circle; experiment 2, square).

3. Fold‐change (FC) of mRNA gene expression of inflammatory markers determined by qPCR. The 2^−ΔΔCt method was used to normalise the values, using gapdh as endogenous control (see Materials and Methods). Data are shown as average ± SD of FC. Statistical analyses were performed using unpaired t‐test (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001). Two replicates were performed and individual symbols are shown (Experiment 1, circle; Experiment 2, square).

4. Quantitative analysis of immunochemistry (IHC) against neutrophil elastase (NE) of lung samples. Data are represented as the mean ± SD of percentage of positive cells, calculated as follows: % = 100 × (positive cells/positive cells + negative cells; one‐way ANOVA + Tukey's post hoc test, *P < 0.05).

5. Representative images of IHC staining of lung samples against NE are shown. Arrows indicate NE‐positive cells. Scale bar size: 50 μm.

Source data are available online for this figure.