Immunopeptidomics-based design of mRNA vaccine formulations against Listeria monocytogenes

Abstract

Listeria monocytogenes is a foodborne intracellular bacterial pathogen leading to human listeriosis. Despite a high mortality rate and increasing antibiotic resistance no clinically approved vaccine against Listeria is available. Attenuated Listeria strains offer protection and are tested as antitumor vaccine vectors, but would benefit from a better knowledge on immunodominant vector antigens. To identify novel antigens, we screen for Listeria peptides presented on the surface of infected human cell lines by mass spectrometry-based immunopeptidomics. In between more than 15,000 human self-peptides, we detect 68 Listeria immunopeptides from 42 different bacterial proteins, including several known antigens. Peptides presented on different cell lines are often derived from the same bacterial surface proteins, classifying these antigens as potential vaccine candidates. Encoding these highly presented antigens in lipid nanoparticle mRNA vaccine formulations results in specific CD8⁺ T-cell responses and induces protection in vaccination challenge experiments in mice. Our results can serve as a starting point for the development of a clinical mRNA vaccine against Listeria and aid to improve attenuated Listeria vaccines and vectors, demonstrating the power of immunopeptidomics for next-generation bacterial vaccine development.

Fig. 1. Identification of MHC class I immunopeptides presented on Listeria-infected HeLa and HCT-116 cells.

a Four replicates of Listeria-infected and uninfected HeLa or HCT-116 cells were dissolved using a mild lysis buffer and purified as described. One-half of the resulting immunopeptides were analyzed by label-free LC-MS/MS analysis on a Q Exactive HF mass spectrometer, while the other half was differentially labeled by TMT, pooled per cell line and fractionated into 12 fractions prior to LC-MS/MS analysis on a Fusion Lumos. Database searching was carried out using PEAKS Studio X+ ™ (version 10.5 build 20200219)^,. Created with BioRender.com. b Overview of identified and quantified immunopeptides per cell line and in total (Supplementary Data 1). c Immunopeptide length distribution representing the typical dominance of 9mers in the human MHC I-presented immunopeptidome. d 9mer peptide sequences were submitted to MixMHCp 2.1 for motif deconvolution using sequence logos for visualization of the modeled position weight matrices (PWMs) after unsupervised clustering^,. To this end, the tool assumes that all peptides are of length equal to core length, are naturally aligned and that defined positions at the beginning (first three amino acids) and end (last two positions) can be found in the peptide sequence, which is typically the case for HLA-I ligands. HLA binding motifs derived from experimental data demonstrated good overall matching with the expected cell line-specific NetMHCpan 4.1 reference motifs^,,. For HeLa cells, two out of three motifs could be fully restored, while the motif for HLA-C*12:03 was only partially recovered. For HCT-116, five out of six HLA motifs were fully reconstituted, while HLA-C*07:01 was only incompletely recovered, likely due to typically low expression levels of many HLA-C alleles. e Principal component analysis (PCA) using the immunopeptide intensities separated the uninfected samples from the Listeria-infected samples.

Fig. 2. Detection of high confident Listeria immunopeptides.

a Filtering of Listeria peptides for detection in at least two infected samples, higher overall abundance in infected samples and absence in human proteins resulted in 68 high confident bacterial peptides from 42 Listeria protein antigens (Supplementary Data 2). b Peptide binding affinity prediction using the NetMHCpan EL 4.1 algorithm demonstrated that the large majority of the high confidence Listeria peptides are indeed predicted to bind to at least one HLA class I allele of the respective cell line^,–. c A selection of 24 Listeria immunopeptide sequences was synthesized to compare their synthetic and experimental fragmentation spectra, confirming the bona fide identification of the four peptides shown here and in Supplementary Fig. 3. The correlation coefficient r is shown for each Listeria-synthetic peptide pair (see methods). d Volcano plots show that the high confident Listeria immunopeptides belong to the most upregulated immunopeptides upon infection in both cellular models. e Likewise, heat maps visualizing individual Listeria peptide intensities (z-scored) following label-free quantitation show that peptides are generally absent in the uninfected control samples, while TMT-labeling data demonstrates more low-level background intensities presumably due to ratio suppression^,. Immunopeptides identified by both label-free and TMT workflows are indicated by an asterisk (*), including the previously described VAYGRQVYL epitope from LLO. Two versions of each heat map are shown, with the left one indicating z-scored intensities after imputing empty values with low values around the detection limit (to allow t-testing). The heat maps on the right are identical, but show originally missing (unimputed) values in gray. f Venn diagrams showing the overlap between both cell lines in presented Listeria peptides (top) and their parent protein antigens (bottom). Only a single peptide (HLVPEFTNEV) was detected on both cell lines, while the overlap at the antigen level is substantially higher with seven Listeria proteins being presented by the two cellular models. Source data are provided as a Source Data file.

Fig. 3. Listeria antigens are often derived from the bacterial periphery.

a Subcellular localization prediction of Listeria antigens indicated that the majority of antigens are localized at the bacterial periphery as either cell surface-associated or extracellular proteins. b Clusters of Orthologous Groups (COG) terms assessment corroborated the importance of cell wall and membrane-associated proteins for presentation as antigens^,. c Histogram showing the number of identified immunopeptides for all 42 detected Listeria antigens (Supplementary Data 2). The seven most presented antigens without any known enzymatic or harmful activity to host cells were selected as vaccination targets (green bars). Antigens with multiple peptides identified but known enzymatic or toxic properties were excluded from further assessment as vaccine candidates (red bars). Most of these selected antigens were identified from both cell types and are predicted to be present at the bacterial surface. d Vaxign-ML scores were calculated for all 2847 Listeria monocytogenes EGD proteins and plotted according to scoring rank. The seven selected Listeria antigens were among the top scoring proteins, further supporting their selection as vaccine candidates. e Box plot showing the Vaxign-ML scores for all 2847 EGD proteins, the 42 identified Listeria antigens and the seven selected antigens. The latter showed the highest average score, followed by all identified Listeria antigens, both scoring significantly higher than the average score of all EGD proteins (two-sided Mann-Whitney nonparametrical testing, n = 2847 scores for all EGD protein, n = 42 scores for identified antigens, n = 7 scores for selected vaccine candidates). Box plots visualize the median, with the box bounds showing the 25th and 75th percentiles and the whiskers the 5th and 95th percentile. Values below the 5th and above the 95th percentile are visualized as individual data points. Asterisks indicate p values with *p  <  0.05 and ****p  <  0.0001. Source data are provided as a Source Data file.

Fig. 4. Highly presented antigens provide protection as mRNA vaccine candidates.

a C57BL/6J mice were vaccinated utilizing full length, mRNA-encoded Listeria antigens formulated in α-GC adjuvanted cationic LNPs. All seven antigen candidates were each tested in one independent experiment (experiments 1 and 2). Additional experiments for LMON_0149 and LLO_E262K are shown in Supplementary Fig. 5. Mice were vaccinated intravenously with cationic LNPs comprising 10 µg of Listeria antigen mRNA. As negative control and inter-experiment reference, 10 µg ovalbumin (OVA) mRNA was injected in both experiments. In experiment 1, a combination vaccine was administered containing 5 µg LLO_E262K and 5 µg LMON_2272 mRNA. In experiment 2, PBS injection was included as additional negative control, while low-dose Listeria monocytogenes EGD infection (1 × 10⁴ CFUs) served as positive control. Two weeks after prime vaccination, an identical booster was administered and 2 weeks later animals were challenged by i.v. injection of 7.5 × 10⁵ bacteria. Mice were euthanized 72 h post-challenge and bacterial loads in spleen and liver were assessed by CFU counting. Created with BioRender.com. b Bar charts depicting CFU counts in spleen (upper) and liver (lower) relative to the OVA negative control. All Listeria vaccines reduced the bacterial burden, while only LMON_0149, EF-Tu and the combination vaccine reached statistical significance in both organs (representative results from single experiments, two-tailed Mann-Whitney test, data are presented as mean values ± SD, n = 5 individual animals except in liver for LMON_2272 (n = 3), FtsI (n = 3), EF-Tu (n = 4), Listeria infection (n = 3), and in spleen for OVA (n = 4) and Listeria infection (n = 4), where plating of the excluded replicates did not yield CFUs). Pearson (c) and Spearman rank (d) correlations were calculated with GraphPad Prism 9.3 between the number of identified bacterial immunopeptides per vaccine candidate and vaccination efficacy expressed as % CFU reduction. For the combination vaccine, peptide numbers for both antigens were summed up. In both liver and spleen, a positive correlation between the number of presented immunopeptides and protective efficacy is indicated by positive r values, although without reaching statistical significance. Asterisks indicate p values with *p value < 0.05 and **p value < 0.01. Source data are provided as a Source Data file.

Fig. 5. Specific CD8⁺ T-cell responses upon LMON_0149 vaccination.

C57BL/6J mice were vaccinated with mRNA galsomes encoding a LMON_0149 or b OVA as control (10 mice/group). After 7 days, splenocytes were isolated and pulsed with a control OVA epitope or two synthetic peptide epitopes predicted from the LMON_0149 sequence using the IEDB resource tools NetMHCpan v4.1 and MHC-NP (Supplementary Data 4). a Mice vaccinated with LMON_0149 showed significantly higher levels of CD8⁺ T-cell responses to both LMON_0149 YSYKFIRV and QVFEGLYTL epitopes as compared to the OVA epitope SIINFEKL (representative results of a single experiment, Shapiro-Wilk test confirmed data normality, paired two-sided t-test applied, data are presented as mean values ± SD, n = 7 individual animals since splenocytes from three animals with <35% cell viability were excluded). b Conversely, OVA-vaccinated mice responded well to SIINFEKL but not to YSYKFIRV, confirming that mRNA vaccination with LMON_0149 elicits specific CD8⁺ T-cell responses against this Listeria antigen (representative results of a single experiment, Shapiro-Wilk test rejected data normality, Wilcoxon matched-pairs signed rank test applied, data are presented as mean values ± SD, n = 10 individual animals). Additional control experiments with OVA-vaccinated or PBS-injected mice are shown in Supplementary Fig. 6. Asterisks indicate p values with **p  <  0.01 and *****p  <  0.01. Source data are provided as a Source Data file.