Cytosolic Sensing of Intracellular Staphylococcus aureus by Mast Cells Elicits a Type I IFN Response That Enhances Cell-Autonomous Immunity

Mast cells are important effector cells of the innate immune system and contribute to the early host defense against pathogens (1–3). They are present in practically all tissues and are predominantly located at sites that interface with the external environment, such as mucosal surfaces, as well as in s.c. tissue in close proximity to blood vessels. Mast cells, therefore, may be among the first immune cells encountering invading pathogens and initiating the ensuing immune response. They are equipped with a variety of receptors, including TLRs, and several Fc and complement receptors that recognize specific bacterial components and enable them to tailor their response to the pathogen that they encounter (4, 5). Mast cells have been shown to be essential for containing pathogens at the sites of infection and prevent further dissemination (1, 2). They also play a major role in initiating both innate and adaptive immune responses to many bacterial pathogens (3).

A prominent feature of mast cells is the presence of abundant secretory granules in the cytoplasm, which contain large amounts of preformed mediators, including serotonin, histamine, heparin, TNF-α, and enzymes such as tryptase and chymase, and are rapidly released following activation (6). The release of preformed mediators initiates the recruitment and activation of effector immune cells to the sites of pathogen invasion (2). Mast cells can also release de novo synthesized mediators, such as proinflammatory leukotrienes, PGs, chemokines, and cytokines fitted to the specific pathogen (7). For example, although mast cells respond to dengue virus infection with the release of high amounts of CCL5 and low amounts of IL-1β and IL-6 (8), they produce large amounts of CCL20, IL-1α, IL-1β, CXCL8, and GM-CSF in response to Pseudomonas aeruginosa (9, 10). Furthermore, it has also been reported that, although mast cells can produce type I IFNs (IFN-Is) in response to viral infection, they elicit only proinflammatory cytokines, but not IFN-I responses, after infection with Gram-positive or Gram-negative bacteria (11). The authors argued that the lack of IFN-I responses was owed to the incapacity of bacterial pathogens to internalize within mast cells because signaling cascades leading to IFN-I production are triggered by receptors located in intracellular compartments (11). However, the incapacity to internalize into mast cells seems not to be a general phenomenon for all bacteria because we (12, 13) and others (14) have shown that the Gram-positive bacterium Staphylococcus aureus is capable of internalizing and surviving within mast cells. Mast cells are commonly found at sites of the body used as portals of entry by S. aureus, including the skin and the respiratory tract, and they will probably be one of the first cells of the innate immune system that sense and respond to this pathogen after invasion of the host. In previous studies, we reported that mast cells respond to S. aureus by releasing antimicrobial granule compounds, as well as extracellular trap in an attempt to kill the pathogen in an extracellular manner (12). However, S. aureus is able to subvert the extracellular antimicrobial mechanisms of the mast cells by promoting its internalization within these cells using α5β1 integrins expressed on the mast cell surface (12). In humans, mast cells harboring internalized S. aureus have been observed in nasal polyps isolated from patients with chronic rhinosinusitis (15). Because S. aureus internalization and intracellular survival could affect the ensuing response of the infected mast cells, the objective of this study was to investigate how the human mast cells HMC-1 and S. aureus respond to each other by assessing simultaneously gene expression changes taking place in the infected host cell and in the intracellular bacteria using a dual RNA sequencing (RNA-seq) approach (16). The results of this study highlight the plasticity of S. aureus to reprogram its transcriptional response to adapt to the intracellular environment and survive within HMC-1 cells. More importantly, we also show that intracellular viable S. aureus triggers the cytosolic DNA-sensing cGMP–AMP synthase (cGAS)–STING pathway within HMC-1 cells and leads to the production and release of IFN-I. Released IFN-Is act via the surface receptor, IFN-α/β receptor (IFNAR), in an autocrine fashion on the infected HMC-1 cells to enhance cell-autonomous host defenses and in a paracrine fashion to sensitize noninfected neighboring cells and thereby amplifying the immune response.

The human mast cell line HMC-1 was provided by J.H. Butterfield (Mayo Foundation for Medical Education and Research, Rochester, MN) (17).

The following S. aureus bacterial strains were used in this study: S. aureus strain SH1000 (18), S. aureus strain Newman (NCTC 8178), S. aureus strain 6850 (19), GFP-expressing S. aureus SH1000 (20), and S. aureus hla-deficient mutant strain (Δhla) (21). Salmonella enterica subsp. enterica serotype Typhimurium (NTCC 12023) was also used in this study. S. aureus strains were grown to midlog phase in brain-heart infusion medium (Roth) at 37°C with shaking (120 rpm), and Salmonella typhimurium was grown in lysogeny broth (Roth) also at 37°C with shaking. Bacteria were collected by centrifugation, washed with sterile PBS, and diluted to the required concentration.

For heat inactivation, bacteria were heated to 95°C for 2 h using an Eppendorf thermomixer.

HMC-1 cells were adjusted to 2 × 10⁶ cells/ml in IMDM (Life Technologies) supplemented with 5% FCS and infected with S. aureus at a multiplicity of infection (MOI) of five bacteria per one HMC-1 cell. After 2 h of infection, lysostaphin (2.5 µg/ml) (Sigma-Aldrich) was added and HMC-1 cells were incubated for 10 min to remove noninternalized extracellular bacteria. HMC-1 cells were then washed twice with sterile PBS and further incubated in medium containing 100 µg/ml gentamicin. At the indicated times, infected HMC-1 cells were centrifuged at 1500 × g for 5 min, and cells in the pellet were lysed by incubating them with 0.1% Triton X-100 in double-distilled H₂O for 5 min. The numbers of viable bacteria were determined by plating serial dilutions on blood agar plates. The cell culture supernatants were used for determination of IFN-α by ELISA.

In some experiments, HMC-1 cells were incubated 1 h before infection with 1 µg/ml of the irreversible STING inhibitor H-151 (InvivoGen) or with 100 nM for the TANK-binding kinase 1 (TBK1)/IKKε inhibitor BX-795 (Cayman Chemicals). Control HMC-1 cells were incubated with a similar concentration of vehicle DMSO. HMC-1 cells transfected with the retinoic acid–inducible gene I (RIG-I) ligand 5′ppp dsRNA using the transfection reagent LyoVec according to the manufacturer’s instructions (Invivogen) were used to confirm that H-151 (1 µg/ml) is specific for STING and does not affect RIG-I signaling.

For blocking the IFNAR, HMC-1 cells were incubated in the presence of 500 ng/ml anti-IFNAR Ab or isotype-matching IgG Abs as control (Sigma-Aldrich).

In stimulation experiments, 5 × 10³ IU/ml rIFN-α (Abcam) was added to HMC-1 cells 1 h before infection.

HMC-1 cells were infected with S. typhimurium at an MOI of 5:1 for 2 h. Gentamicin was then added at a concentration of 100 µg/ml to kill extracellular bacteria, and HMC-1 cells were further incubated at 37°C and 5% CO₂. After 24 h, HMC-1 cells were harvested, and supernatants were collected for determination of IFN-α.

HMC-1 cells were adjusted to 2 × 10⁶ cells/ml in IMDM supplemented with 5% FCS and infected with S. aureus strain SH1000-GFP for 2 h at an MOI of 5:1. After 2 h of infection, 2.5 µg/ml lysostaphin was added, and HMC-1 cells were incubated for 10 min to remove noninternalized extracellular bacteria. HMC-1 cells were then washed twice with sterile PBS and further incubated for 24 h in medium containing 100 µg/ml gentamicin. HMC-1 cells harboring intracellular S. aureus (GFP⁺) were separated from noninfected bystander HMC-1 cells (GFP⁻) by FACS using a BD FACSAria III (Becton Dickinson) and resuspended in RNAlater (Ambion). Sorted HMC-1 cells were centrifuged for 10 min at 1000 × g, washed twice with sterile prewarmed PBS, and carefully resuspended in 600 µl per 5 × 10⁶ cells of cell lysis buffer included in mirVANA miRNA Isolation Kit (Ambion). Cell lysates were then transferred to FastPrep 24 lysing matrix tubes (mechanical lysis with FastPrep at 5′, 1000 × g), and RNA was isolated following the recommendations provided in the mirVANA miRNA Isolation Kit (Ambion).

RNA integrity was determined using a 2100 Bioanalyzer and the RNA 6000 Nano kit (Agilent Technologies, Santa Clara, CA). RNA integrity values for all samples ranged from 8.5 to 10.0. In accordance with the manufacturer’s instructions, rRNA was depleted using Illumina’s RiboZero Epidemiology Kit (Illumina). In brief, rRNA-specific biotinylated DNA probes were added to the total RNA. After hybridization of the probes and the rRNA, magnetic beads were added that bind to the rRNA–DNA hybrids. By placing the samples on a magnetic stand, the rRNA–DNA hybrids that are bound to magnetic beads were pulled down. The rRNA-depleted RNA was then purified using RNA Clean & Concentrator 5 kit (Zymo Research) following the manufacturer’s protocol (manual version 2.2.1).

RNA was fragmented using NEB Next Magnesium RNA fragmentation module (New England Biolabs) following the manufacturer’s protocol. The following modifications were introduced in the protocol: Mg²⁺ was used to fragment RNA for 3 min at 94°C using ABI 9700 PCR System. The fragmented RNA was purified with the RNA Clean & Concentrator kit 5 (Zymo Research), and RNA quality was determined using a 2100 Bioanalyzer and the RNA 6000 Pico kit (Agilent Technologies). Prior to adapter ligation, RNA was dephosphorylated at the 3′ end and phosphorylated at the 5′ end using 10 U T4-PNK ± 10 mM ATP (New England Biolabs). RNA was then decapped twice using 5 U RppH (New England Biolabs) following the manufacturer’s protocol for eukaryotic cells and prokaryotic cells, respectively. RNA was purified with RNA Clean & Concentrator kit 5 (Zymo Research) after each enzymatic treatment as described earlier. cDNA synthesis was performed using NEBNext Small RNA Library Prep Set for Illumina (Illumina). In brief, RNA fragments were ligated to the 3′ SR and 5′ SR adapters prediluted 1:4 with nuclease-free water. PCR amplification to add Illumina adaptors and indices was performed for 15 cycles with 1:4 prediluted primers. Prior to sequencing, cDNA libraries were purified using the magnetic MagSi-NGS^PREP Plus beads (magtivio) at a 1.8:1 ratio of beads to sample volume and afterward quantified with the Qubit 2.0 Fluorometer using Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific). The libraries’ quality and size distribution were checked with a 2100 Bioanalyzer using HS DNA 7500 kit.

The amount of IFN-α was quantified in the culture supernatants using a human IFN-α Instant ELISA System according to the manufacturer’s instructions (Invitrogen).

Total RNA was isolated from HMC-1 cells at the indicated time points using the GeneJET RNA purification kit (Fisher Scientific). RNA samples were reverse transcribed and amplified using a SensiFAST SYBR No-ROX Kit (Bioline) following the manufacturer’s recommendations. The primers used for quantitative RT-PCR were for IFNA1 (IFN-α), forward [for.]: 5′-GTG AGG AAA TAC TTC CAA AGA ATC AC-3′, reverse [rev.]: 5′-TCT CAT GAT TTC TGC TCT GAC AA-3′; IFNB1 (IFN-β), for.: 5′-CGC CGC ATT GAC CAT CTA-3′, rev.: 5′-TTA GCC AGG AGG TTC TCA ACA ATA GTC TCA CTA-3′; and for the gene encoding β-actin (ACTB), for.: 5′-AAC TCC ATC ATG AAG TGT GAC G-3′; rev.: 5′-GAT CCA CAT CTG CTG GAA GG-3′. Thermal cycling conditions for IFNA1 and ACTB mRNA quantification consisted of reverse transcription for 20 min at 45°C, initial denaturation for 5 min at 95°C, followed by 40 cycles of 20 s at 95°C (denaturation), 20 s at 58°C (annealing), and 20 s at 72°C (elongation). Primers for RT-PCR quantification of selected IFN-I–induced genes mRNA and for RT-PCR quantification of TNFA (TNF-α) mRNA were purchased from OriGene and used following the conditions recommended by the manufacturer (OriGene). The following qPCR Primer Pairs were used: RSAD2 (Viperin) (NM_080657; catalog number [CAT#]: HP216708), IFN regulatory factor (IRF) 7 (NM_004031; CAT#: HP231979, IFI6 (NM_022873; CAT#: HP225644), IFI27 (NM_005532; CAT#: HP208651), MX2 (NM_002463; CAT#: HP206143), and TNF-α (TNF) (NM_000594; CAT#: HP200561). Data were normalized against the housekeeping gene β-actin. Fold change values were calculated by the Pfaffl equation, in which the expression ratio is estimated by (Etarget)ΔCt, target (control − experimental)/(Eref)ΔCt ref (control − experimental).

HMC-1 cells (2 × 10⁶ cells/ml) were preincubated for 1 h with 1 µg/ml anti–β1-integrin blocking Abs (Santa Cruz biotechnology) or for 30 min with 5 µg/ml cytochalasin D (Sigma-Aldrich). Control cells received medium alone. HMC-1 cells were washed to remove unbound Abs or cytochalasin D and infected for 2 h with S. aureus at an MOI of 5:1. Lysostaphin was added at a concentration of 2.5 µg/ml for 10 min to eliminate noninternalized extracellular bacteria, and HMC-1 cells were washed and used either to determine the amount of intracellular viable bacteria as described earlier or further incubated for 24 h in medium containing 100 µg/ml gentamicin to determine the concentration of IFN-α in the culture supernatant.

HMC-1 cells in suspension at a density of 1 × 10⁶ cells/ml in IMDM supplemented with 5% FCS were infected with S. aureus at an MOI of 10:1 for the immunofluorescence (IF) staining and MOI 20:1 for electron microscopy (EM). At different infection times, parallel samples of infected and uninfected HMC-1 cells were fixed for IF or EM. Fixation was performed in the IF samples by adding the same volume of a 6% paraformaldehyde solution in PBS and incubating during 20 min at room temperature. Cells were centrifuged at 1000 × g for 10 min, and the pellet was used for the IF labeling. Cells processed for EM were first centrifuged at 1000 × g for 10 min, washed with PBS and resuspended in PBS, and immediately fixed for field emission scanning EM or transmission EM.

Confocal microscopy examination of a total of 45 HMC-1 cells was used to calculate the percentage of HMC-1 cells harboring internalized S. aureus and the mean number of bacteria per cell.

The staining of HMC-1 cells in suspension was performed following a modified protocol (22). HMC-1 cells were fixed as mentioned earlier and transferred to microcentrifuge tubes where the labeling was performed. Generally, after each step, cells were washed with 1200 µl of PBS, centrifuged at 1000 × g, and the supernatant was discarded by aspiration. The different labeling solutions were added to the pellet and after mixing, each labeling was performed on an Eppendorf thermomixer set at 700 × g with the temperature control off. Cells were first washed with 900 µl of 10 mM glycine in PBS and after centrifugation were permeabilized with 0.1% Triton X-100 in PBS during 5 min and then washed twice with PBS. The pellet was resuspended in 100 ml of PBS, transferred to a fresh microcentrifuge tube, and 800 µl of 10% FBS-PBS was added for blocking during 45 min. After centrifugation, 120 µl of custom-produced anti–S. aureus rabbit serum diluted 1:100 was added to the cells and incubated during 1 h. After washing twice, HMC-1 cells were incubated with 1:500 secondary Ab Alexa Fluor 488–conjugated goat anti-rabbit (Thermo Fisher Scientific) for 45 min at room temperature. After washing twice, cells were stained with 10 µl of Alexa Fluor 633 phalloidin (Thermo Fisher Scientific) in 500 µl of PBS for 45 min and washed three times. ProLong Gold Antifade Mountant with DAPI (Thermo Fisher Scientific) was added to the pellet and carefully mixed. A total of 7 µl of sample was applied to the center of a 22 × 22-mm coverslip, and a microscope slice was placed on top. Mounted cells were allowed to dry overnight, and the edges of the coverslips were sealed before microscopic observation. Imaging was performed with a confocal laser-scanning upright microscope Leica SP5 equipped with an HC PL APO 63×/1,40 oil-immersion objective using three lasers, diode (405), argon (488 nm), and He-Ne (633 nm), and the LAS AF software. After the confocal laser-scanning upright microscope measurement, the image stacks were processed with Fiji-ImageJ.

HMC-1 cells were fixed with 4% paraformaldehyde, washed with TE buffer (20 mM Tris, 1 mM EDTA [pH 6.9]) and dehydrated after incubation with a graded series of ethanol (10, 30, 50, 70, 90, 100%) on ice for 15 min. HMC-1 cells were then critical-point dried with liquid CO₂ and covered with a gold film by sputter coating (SCD 40; Balzers Union). HMC-1 cells were examined in a field emission scanning electron microscope (Zeiss DSM 982 Gemini) using the Everhart Thornley SE detector and the inlens detector in a 50:50 ratio at an acceleration voltage of 5 kV.

HMC-1 cells were fixed with 2% glutaraldehyde and 3% formaldehyde in cacodylate buffer for 1 h on ice, washed with cacodylate buffer, and osmificated with 1% aqueous osmium for 1 h at room temperature. HMC-1 cells were then dehydrated with a graded series of acetone (10, 30, 50, 70, 90, and 100%) for 30 min at each step. The 70% acetone dehydration step was performed in 2% uranyl acetate overnight. HMC-1 cells were then infiltrated with an epoxy resin, and ultrathin 70-nm sections were cut with a diamond knife. Sections were counterstained with uranyl acetate and lead citrate and examined in a TEM910 transmission electron microscope (Carl Zeiss) at an acceleration voltage of 80 kV. Images were taken at calibrated magnifications using a line replica and recorded digitally with a Slow-Scan CCD-Camera (ProScan) with ITEM Software (Olympus Soft Imaging Solutions). Brightness and contrast were adjusted with Adobe Photoshop CS3.

Illumina reads were trimmed using cutadapt (version: 1.16) (23). Illumina’s TruSeq “Read 1” adapter sequence was removed from the 3′ end. Nucleotides with a Phred quality score <20 and their following downstream (5′–3′) bases were also cut off. Further filtering steps including read mapping and downstream analysis, such as gene quantification, generation of coverage files, and differential gene expression analysis, were made by the RNA-seq tool READemption (version: 0.4.3, doi: 10.5281/zenodo.250598) (24). Additional reads filtering included clipping of poly(A) sequences and discarding of reads that had a read length <20 nucleotides after performing the trimming steps. The read mapping was performed using the short read mapper segemehl (version: 0.2.0) (25), which is integrated into READemption. The mapping was performed with an accuracy of 95% and segemehl’s aligner lack (26). The human genome and annotation were obtained from GENCODE (version: 27, NCBI assembly name: GRCh38.p10) and the bacterial ones from NCBI’s RefSeq database (accession number: NC_007795.1, https://www.ncbi.nlm.nih.gov/nuccore/88193823; RefSeq assembly accession number: GCF_000013425.1, https://www.ncbi.nlm.nih.gov/assembly/GCF_000013425.1). The S. aureus annotation was extended with small RNAs (sRNAs) predicted by ANNOgesic (27). Transcripts that were not associated with any of RefSeq’s annotated features were determined as sRNA candidates based on their predicted folding energy. Candidates that had homologs in NCBI’s nonredundant protein database (https://www.ncbi.nlm.nih.gov/refseq/about/nonredundantproteins) were discarded, while candidates with homologs in sRNA database BSRD (28) were accepted. The gene quantification files (i.e., the number of reads overlapping with an annotated feature) and the coverage files in wiggle format (i.e., the number of reads overlapping with each base of the genome) were created using READemption. Afterward both file types were split up by species. The coverage was normalized by the total number of aligned reads of a given replicate and multiplied by 1,000,000. Differential gene expression analysis was performed with the R package DESeq2 (version: 1.20.0) based on raw read counting (29). Genes with an adjusted (Benjamini–Hochberg corrected) p < 0.05 were defined as differentially expressed.

Raw read files can be found at the European Nucleotide Archives under the project ID PRJEB43874 (https://www.ebi.ac.uk/ena/browser/view/PRJEB43874). The complete bioinformatical workflow is available at the Repository for Life Sciences (https://repository.publisso.de/resource/frl:6427216, doi: 10.4126/FRL01-006427216). A shell script can be executed step by step or in one go to reproduce the analysis. Singularity images are provided that contain all required programs.

Heatmaps, hierarchical clustering dendrograms, and principal-component analysis (PCA) plots were generated using the corresponding function of the platform MetaboAnalyst v.3.0 (30). Gene lists of all significantly expressed genes between the different conditions were used as input for the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway (S. aureus) or Reactome (HMC-1) analysis using DAVID (31). Comparisons between groups were made using a parametric ANOVA test with Tukey’s posttest or a t test. The p values <0.05 were considered significant.

In contrast with what has been previously reported for other Gram-positive bacteria (11), S. aureus is capable of internalizing within HMC-1 cells. The IF and EM photographs depicted in (Fig. 1A show the capacity of S. aureus to adhere to the surface of HMC-1 cells (Fig. 1A, upper panel) and internalize within the HMC-1 cells (Fig. 1A, lower panel). In our experimental setting, ∼66% of HMC-1 cells cells were found to harbor internalized S. aureus with a mean of 5.6 ± 6.5 bacteria per cell. Within HMC-1 cells, S. aureus could be found within membrane-bound vacuoles (Fig. 1B, lower panel, red arrows) or free in the cell cytosol (Fig. 1B, lower panel, insert). Evaluation of intracellular bacterial viability indicated that, although HMC-1 cells has the capacity to kill a proportion of the internalized S. aureus, a subpopulation of bacteria was capable of escaping the intracellular antimicrobial mechanisms of HMC-1 cells and remained viable after 24 h of infection (Fig. 1C).

A dual RNA-seq approach was then used to investigate the strategies used by S. aureus to survive within HMC-1 cells, as well as the functional consequences of harboring intracellular S. aureus for the HMC-1 cells responses. For this purpose, HMC-1 cells cells were infected with GFP-expressing S. aureus for 2 h, noninternalized bacteria were removed by lysostaphin treatment, and HMC-1 cells were further incubated for 24 h in the presence of antibiotics. HMC-1 cells harboring intracellular S. aureus (GFP⁺) were then separated from noninfected bystander HMC-1 cells (GFP⁻) cells by FACS and subjected to dual RNA-seq for parallel gene expression analysis of HMC-1 cells and intracellular S. aureus. The transcriptional response of uninfected HMC-1 cells and of S. aureus in the input infection inoculum were used as control for host and intracellular pathogen, respectively. We also determined the transcriptional response of noninfected bystander HMC-1 cells (GFP⁻) cells, as well as of HMC-1 cells cocultured with S. aureus in separated chambers using a permeable transwell system. The different infection settings are summarized in the scheme depicted in (Fig. 1D. Total RNA was isolated from the different samples and subjected to RNA-seq analysis. The distribution of RNA classes from HMC-1 cells indicated that between 40 and 50% of the HMC-1 cell reads mapped to coding sequences in the different samples (Fig. 1E). Regarding the RNA classes distribution from S. aureus, tRNAs were more represented in intracellular S. aureus than in S. aureus in the infection inoculum (Fig. 1F), probably suggesting a more active protein synthesis in the intracellular bacteria. The dual RNA-seq analysis of HMC-1 cells harboring S. aureus showed that ∼95% of the reads could be mapped to the human genome and 5% to the bacteria genome in each of the three replicates (Fig. 1G).

To gain a better understanding of the strategies used by S. aureus to survive and persist within HMC-1 cells, we compared the expression profile of protein coding genes from intracellular S. aureus with that of S. aureus in the infection inoculum. Hierarchical clustering (Fig. 2A), PCA (Fig. 2B), and heatmap of gene expression levels (Fig. 2C) showed a clear separation between the transcriptome datasets of S. aureus located within HMC-1 cells (intracellular) and S. aureus in the infection inoculum (inoculum). This indicated that S. aureus underwent profound remodeling of the transcriptional response on internalization within HMC-1 cells. We performed differential gene expression analysis on the RNA-seq datasets and focused on transcripts with differential expression of log₂ fold change > 2 (upregulated) or log₂ fold change < −2 (downregulated) (Benjamini-Hochberg adjusted p < 0.05) for further analysis. We found 143 genes upregulated and 126 downregulated by intracellular S. aureus in comparison with the bacteria in the inoculum, with many of them encoding hypothetical proteins (Fig. 2D, Supplemental Table I). KEGG pathway enrichment analysis of differentially expressed genes showed “ribosome,” followed by “galactose metabolism” and “monobactam biosynthesis” as the most predominant enriched pathway in genes upregulated by intracellular S. aureus (Fig. 2E). Indeed, many genes encoding components of the protein translation machine, such as ribosomal proteins and ribonucleoproteins, were expressed to a significantly higher extent by intracellular S. aureus than by S. aureus in the input inoculum (Supplemental Table I). Interestingly, the genes of the lactose operon lacABCD operon, which are implicated in the catabolism of lactose and d-galactose, as well as the cotranscribed genes lacFEG, which encode the proteins for transport, phosphorylation, and cleavage of these carbon sources (32), were expressed by intracellular S. aureus, but not by the bacteria in the infection inoculum (Supplemental Table I). These genes are inducible by lactose or galactose (33) and repressed in the presence of glucose (34), indicating that lactose or galactose, but not glucose, are the carbon sources available to the bacterium in the intracellular compartment. Furthermore, the gene encoding the ROK family protein, which is involved in the metabolism of the amino sugar N-acetylglucosamine and the sialic acid N-acetylneuraminate (35), was also induced by S. aureus in the intracellular environment (Supplemental Table I). In addition to these pathways, genes involved in the stress response, such as the genes encoding components of the classical chaperones DnaK/DnaJ and GroES/GroEL (dnaK, groES, groEL) (36), as well as those coding for Clp chaperones (clpB and clpC) (37) and genes encoding virulence factors, such as superantigen-like protein SSL6 (ssl6), coagulase (coa), fibronectin-binding proteins (fnbA, fnbB), the extracellular matrix, and plasma binding protein Emp (emp), were also upregulated by intracellular S. aureus (Supplemental Table I).

KEGG pathway enrichment analysis of genes exhibiting lower expression in intracellular S. aureus than in S. aureus in the infection inoculum identified high enrichment of pathways involved in “two-components system” followed by “cationic antimicrobial peptide (CAMP) resistance,” “carotenoid biosynthesis,” and “glucolysis/gluconeogenesis” (Fig. 2E). The genes dltB, dltD, and dltABCD, which encode factors involved in cationic antimicrobial peptide resistance (38) as well as vraF, which encodes part of the VraFG ABC transporter that potentially enhances export of cell wall/teichoic acid precursors (39) were also downregulated by intracellular S. aureus.

To investigate the consequences of harboring intracellular S. aureus for the HMC-1 cells responses, we compared the gene expression profile of HMC-1 cells harboring intracellular bacteria with the gene expression profile of either noninfected bystander HMC-1 cells, uninfected HMC-1 cells, or HMC-1 cells cocultured with S. aureus but separated by a transwell system. Hierarchical clustering of the transcriptome of HMC-1 cells in the different infection settings showed that S. aureus–harboring HMC-1 and noninfected bystander HMC-1 cell samples clustered together but away from uninfected HMC-1 and transwell HMC-1 cell samples (Fig. 3A). This clustering was also reflected by the heatmap depicted in (Fig. 3B showing the pattern of gene expression across the samples. The results of these analyses indicated that the transcriptional response of S. aureus–harboring HMC-1 cells was highly similar to that of noninfected bystander HMC-1 cells but significantly different from the transcriptional response of uninfected or transwell samples. The fact that the gene expression profile of HMC-1 cells in a transwell system, where they are separated from S. aureus by a permeable membrane, did not differ from that of uninfected HMC-1 cells excluded a potential effect of soluble factors released by S. aureus on the transcriptional response of HMC-1 cells.

To get further insights into the transcriptional changes taking place in HMC-1 cells harboring intracellular S. aureus and in noninfected bystander HMC-1 cells, we performed differential gene expression analysis of these samples in comparison with uninfected HMC-1 cells. The results of this analysis showed 59 genes with significantly higher expression (log₂ fold change > 2, adjusted p < 0.05) and 3 genes with significantly lower expression (log₂ fold change < −2, adjusted p < 0.05) in S. aureus–harboring HMC-1 cells with respect to uninfected HMC-1 cells (Fig. 3C, Supplemental Table II) and 66 genes with significantly greater expression and 6 genes with significantly lower expression in noninfected bystander HMC-1 cells in comparison with uninfected HMC-1 cells (Fig. 3C, Supplemental Table III). No differentially expressed genes with log₂ fold change > 2, adjusted p < 0.05 or log₂ fold change < −2, adjusted p < 0.05 were identified between S. aureus–harboring HMC-1 cells and noninfected bystander HMC-1 cells (Fig. 3C). Likewise, no differentially expressed genes with adjusted p < 0.05 were found between transwell HMC-1 cells and uninfected HMC-1 cells (Fig. 3C). Reactome pathway enrichment analysis performed in differentially expressed genes with greater expression in S. aureus–harboring HMC-1 cells than in uninfected HMC-1 cells indicated a robust transcriptional signature related to genes induced by IFN-I (Figs. 3D, 4A). A similar overlapping IFN-I–induced transcriptional response was observed in noninfected bystander HMC-1 cells in comparison with uninfected HMC-1 cells (Figs. 3E, 4A). The induction of IFN-I target genes in S. aureus–infected HMC-1 cells was confirmed by RT-PCR (Fig. 4B).

IFN-Is comprise a family of highly pleiotropic cytokines that includes IFN-α and IFN-β (40). Because the induction of IFN-I target genes in S. aureus–harboring HMC-1 and noninfected bystander HMC-1 cells was observed in the transcriptional analysis performed after 24 h of infection, we speculated that IFN-I proteins may already be present in the culture supernatant at this time of infection; consequently, the induction of the genes encoding IFN-I may take place at earlier times of infection. To investigate whether this is the case, we first determined the expression levels of the genes encoding IFN-α and IFN-β in HMC-1 cells at 2 and 4 h postinfection by RT-PCR. The results show that both genes were induced at 2 h postinfection and their level of expression substantially increased at 4 h postinfection, although the gene encoding IFN-α was expressed to a significantly greater extent than the gene encoding IFN-β (Fig. 4C). In addition to IFN-I, NF-κB target genes, such as TNF-α, were also upregulated by HMC-1 cells in response to S. aureus infection (Fig. 4D). At the protein level, significant amounts of IFN-α were detectable in the supernatant of S. aureus–infected HMC-1 cells at 24 h of infection, but not in the supernatant from uninfected HMC-1 cells, HMC-1 cells cocultured with S. aureus in a transwell system, or HMC-1 cells infected with Salmonella typhimurium, which have been previously reported to be incapable of eliciting IFN-I in human mast cells (11) (Fig. 4E). Bacterial viability was required for the production of IFN-I by HMC-1 cells because IFN-α was under detection levels in the supernatant of HMC-1 cells incubated with heat-inactivated S. aureus (Fig. 4E). We also demonstrated that IFN-I production by HMC-1 cells was not bacterial strain dependent because they produced significant amounts of IFN-α not only postinfection with S. aureus strain SH1000, which is the strain that has been used in all the earlier-described experiments, but also postinfection with S. aureus strain Newman and strain 6850 (Fig. 4F).

We next explored the signaling pathway leading to IFN-I induction in S. aureus–infected HMC-1 cells. Cytosolic signaling pathways, such as the cGAS–STING pathway that recognizes DNA (41) and RIG-I that recognizes RNA (42), have emerged as the major sensing systems driving IFN-I responses. Because these pathways are largely triggered by recognition of pathogen-derived nucleic acids in the cell cytosol, we first determined the requirement of bacterial internalization for IFN-I production by HMC-1 cells. Inhibition of S. aureus internalization using either the actin polymerization inhibitor cytochalasin D or β1-integrin blocking Abs prevented S. aureus internalization within HMC-1 cells (Fig. 5A) as previously reported (12, 13) and resulted in complete abrogation of IFN-α production (Fig. 5B). Furthermore, HMC-1 cells failed to produce IFN-α postinfection with a S. aureus mutant strain deficient in the production of α-hemolysin (Δhla), which has been reported to be impaired in its capacity to internalize and survive within mast cells (13) (Fig. 5C). Because S. aureus has been reported to activate IFN-I responses in macrophages via the cGAS–STING pathway (43), we next explored the relevance of this pathway in the production of IFN-I by infected HMC-1 cells. In the cGAS–STING pathway, pathogen-derived DNA present in the cell cytosol binds to the cGAS, resulting in conformational changes that induce enzymatic activity (44). Activation of cGAS leads to the production of the second messenger cGMP–AMP, which binds to the endoplasmic reticulum–localized adaptor protein STING. After activation, STING translocates from the endoplasmic reticulum to the Golgi, where it recruits kinases such as TANK-binding kinase 1 (TBK1), which phosphorylates IRF3 and triggers the expression of IFN-I (41). STING can also directly bind bacterial c-di-AMP in the host cytosol and induce an IFN-I response (45, 46). To determine the potential involvement of the cGAS–STING pathway in the production of IFN-I by S. aureus–infected HMC-1 cells, we blocked this pathway using the STING-specific inhibitor H-151 (47). As shown in (Fig. 5D, treatment with H-151 almost completely abrogated the production of IFN-α by S. aureus–infected HMC-1 cells. Treatment with H-151 did not affect the capacity of HMC-1 cells to produce IFN-α after stimulation of the alternative signaling pathway RIG-I with the agonist 5′ppp-dsRNA (Fig. 5E). These results corroborated the specificity of H-151 for STING inhibition.

Furthermore, we also demonstrated that treatment with BX795, a potent inhibitor of TBK1 (48), resulted in profound reduction of IFN-α production by S. aureus–infected HMC-1 cells (Fig. 5F). These results indicated that the cGAS–STING–TBK1 axis was involved in the induction of IFN-I in HMC-1 cells by S. aureus.

Although the concerted activation of IFN-I–stimulated genes is a key component of the innate immune response against viruses (49), it has become increasingly evident that they also play an important role in the control of intracellular bacterial pathogens (50). IFN-I molecules bind to a common surface receptor named IFNAR, which comprises two subunits, IFNAR1 and IFNAR2, forming a ternary complex that leads to the activation of the Jak tyrosine kinase 2 and Jak1 (51). After activation, these kinases propagate downstream signaling leading to the activation of transcription factors such as STAT1 and STAT2 that after dimerization translocate to the nucleus, where they assemble with IRFs and mediate the transcription of a large number of IFN-I–stimulated genes involved in cell-autonomous immunity (51).

To determine the relevance of IFN-I–induced response on the capacity of HMC-1 cells to control intracellular S. aureus, we disrupted IFN-I signaling by blocking IFNAR1 with specific Abs. Disruption of IFN-I/IFNAR1 signaling did not influence the amount of S. aureus internalizing within HMC-1 cells cells (Fig. 6A, 0 h), but reduced considerably the capacity of HMC-1 cells to control intracellular S. aureus because significantly higher numbers of intracellular S. aureus were detected in HMC-1 after inhibition of IFN-I/IFNAR signaling in comparison with untreated HMC-1 cells at 2, 4, and 24 h postinfection (Fig. 6A). To discard that the effect of blocking IFNAR1 on the capacity of HMC-1 cells to reduce intracellular S. aureus was due to an unspecific effect of the Ab, we determined the level of expression of a set of IFN-I–induced genes in S. aureus–infected HMC-1 cells treated with either anti-IFNAR1 Abs or with an isotype-matched (IgG) control Ab. As shown in (Fig. 6B, whereas the expression levels of the genes encoding IFI27, IFR7, and MX2 in S. aureus–infected HMC-1 cells treated with isotype control Abs were comparable to those observed in untreated S. aureus–infected HMC-1 cells (Fig. 4B), the expression levels of these genes were significantly lower in S. aureus–infected HMC-1 cells treated with anti-IFNAR1 Abs. These results corroborate the specific effect of anti-IFNAR1 blocking Abs.

We also investigated the effect of stimulating HMC-1 cells with rIFN-α prior to infection on their capacity to control intracellular S. aureus. As shown in (Fig. 6C, treatment with rIFN-α enhanced the capacity of HMC-1 cells to control intracellular S. aureus because they exhibited significantly lower numbers of intracellular viable bacteria than untreated HMC-1 cells.

Altogether, these results indicated that IFN-α released by HMC-1 cells harboring intracellular S. aureus signaled back in an autocrine manner resulting in the induction of IFN-I target genes and improved cell-autonomous host defenses. The observation that IFN-I target genes were also upregulated in bystander HMC-1 that did not harbor intracellular S. aureus indicated that IFN-I released by S. aureus–harboring HMC-1 cells signaled also in a paracrine manner to induce an IFN-I signature in these cells.

Mast cells are generally located at host sites used by S. aureus for invasion of the host; therefore, they may be among the first innate immune cells recognizing and fighting this pathogen. We have previously reported the capacity of murine and human HMC-1 cells mast cells to recognize extracellular S. aureus and respond by releasing extracellular traps and antimicrobial compounds in an attempt to immobilize and kill the pathogen (52). We have also reported that S. aureus was able to induce its own internalization within mast cells to escape the extracellular antimicrobial mechanisms of these cells (12, 13). In this study, we show that HMC-1 cells responded to S. aureus internalization by activating intracellular antimicrobial defense mechanisms that resulted in a significant reduction of intracellular bacteria within a few hours after bacterial internalization. However, a subpopulation of internalized S. aureus was capable of circumventing these antimicrobial mechanisms and survived within HMC-1 cells for long periods. Therefore, the interactions between S. aureus and HMC-1 cells during infection involve a series of events as each part deploys mechanisms of defense and survival. Because the outcome of these interactions can influence the ensuing immune response, we investigated in this study how the human mast cells HMC-1 cells and S. aureus respond to each other by assessing simultaneously gene expression changes taking place in the infected HMC-1 cells and in the harbored bacteria using dual RNA-seq analysis.

The results of the bacterial gene expression analysis indicated that, to survive within HMC-1 cells, S. aureus undergoes profound transcriptional reprogramming to readjust its metabolism to the nutritional changes and to counteract the stress conditions encountered in the intracellular niche. Thus, the genes encoding enzymes and transport systems involved in the galactose/lactose and d-tagatose-6-phosphate metabolic pathways were upregulated in intracellular S. aureus in comparison with the bacteria in the infection inoculum, whereas the genes associated with glycolysis were downregulated. This is interesting because S. aureus is one of the few microorganisms known to exclusively use enzymes of the d-tagatose-6-phosphate pathway to metabolize d-galactose, which is imported into the bacterial cell by a transport system encoded by genes lacFEG and metabolized by proteins encoded by the lactose operon, lacABCD (53). The lac operon has been shown to be inducible by the presence of d-galactose or lactose (33), suggesting that these sugars may be the carbon source available to the bacteria within the HMC-1 cells. Furthermore, the gene encoding the ROK family protein, which is involved in the metabolism of the amino sugar N-acetylglucosamine and the sialic acid N-acetylneuraminate by S. aureus (35), was also induced in the intracellular environment. We speculated that peptidoglycans, which are complex macromolecules comprising disaccharides, such as N-acetyl-glucosamine and galactose, and are abundant within mast cells because they play an important role in the tight packaging of compounds within secretory granules (54), could provide a source of galactose and amino sugars for intracellular S. aureus. Increased expression of the genes belonging to the heat shock stimulon, including the DnaK and GroESL chaperones, was also observed in intracellular S. aureus, most probably required for the bacteria to deal with the highly stressful conditions encountered within HMC-1 cells.

On the host cell side, we observed that HMC-1 cells produced IFN-I in response to internalized S. aureus. The production of IFN-I by HMC-1 in response to S. aureus contrasts with another study where the authors claimed that only viruses and not bacterial pathogens can induce an IFN-I response in mast cells because of the incapacity of bacteria to internalize into these cells (11). In that study, the authors used the Gram-positive Listeria monocytogenes and Streptococcus pyogenes and the Gram-negative Salmonella typhimurium in their mast cells infection assays (11). The results of our study indicate that this is not the case for all bacterial pathogens but probably only for those that fail to internalize within mast cells. Indeed, inhibition of S. aureus internalization within HMC-1 cells after treatment with cytochalasin or β1-integrin blocking Abs or infection of HMC-1 cells using a mutant S. aureus strain unable to internalize within HMC-1 cells (13) completely prevented the production of IFN-I. Our study, therefore, provides compelling evidence that mast cells can indeed produce IFN-I in response to bacterial infection and argues against the concept that mast cells can elicit IFN-I responses only to viral infections as previously reported (11). IFN-I produced and released by HMC-1 cells harboring intracellular S. aureus can then bind to the IFNAR either on the same infected HMC-1 cells cells in an autocrine fashion or on noninfected bystander neighboring MHC-1 cells in a paracrine way, resulting in the induction of a large number of IFN-I–stimulated genes. It has been reported that the product of these IFN-I–stimulated genes contributes to enhanced cell-autonomous host defense against intracellular pathogens in infected cells (55). In our study, the autocrine stimulation of IFN-I–stimulated response in infected HMC-1 cells seems to contribute, at least to some extent, to the proper control of internalized S. aureus because interfering with IFN-α/IFNAR signaling using blocking Abs significantly reduced the capacity of HMC-1 cells to kill intracellular S. aureus and resulted in much lower expression of IFN-I–induced genes. The transcriptional analysis also indicated that IFN-I signaled in a paracrine manner in noninfected bystander HMC-1 cells and induced IFN-I–stimulated genes, probably instructing them to enter a state of enhanced resistance toward S. aureus. Indeed, pretreatment of HMC-1 cells with rIFN-α increased the capacity of these cells to control intracellular S. aureus.

We also found that the cGAS–STING–TBK1 signaling pathway was involved in the recognition of intracellular S. aureus by HMC-1 cells and in the induction of IFN-I. The role of STING in detection of cytosolic DNA, such as those from viral or bacterial infections, is well known (56–58). In bacterial infections, STING-dependent induction of IFN-I has been reported for both intracellular and extracellular pathogens (58). In the particular case of S. aureus, cGAS–STING signaling activated an IFN-I response in macrophages after infection with live but not killed bacteria (43). This is in line with our data showing that HMC-1 cells incubated with heat-killed S. aureus failed to produce IFN-α. Activation of STING in infected HMC-1 cells can ensue either after recognition of bacterial DNA or most probably through its direct activation by c-di-AMP produced by S. aureus. In this regard, it has been reported that c-di-AMP released from S. aureus biofilms can activate STING and induce an IFN-I response in macrophages (59).

In summary, the results of this study provide a scenario where, after invasion of the host, mast cells recognize extracellular S. aureus, most probably via pattern recognition receptors on the cell surface or by sensing bacterial toxins such as δ toxin as previously reported (60), and respond by undergoing degranulation with the concomitant discharge of prepackaged antimicrobial compounds or by releasing extracellular traps to kill the extracellular bacteria (52). S. aureus, in turn, induces its own internalization within mast cells, most probably to escape their extracellular killing mechanisms, and establishes a survival niche within these cells. S. aureus–infected mast cells sense the intracellular bacteria by cytosolic receptors and produce IFN-Is that act in an autocrine manner to enhance cell-autonomous host defense in the infected mast cells and in a paracrine way to sensitize neighboring cells and amplify the immune response. Our study thus has provided important information about the strategy used by mast cells to recognize S. aureus and how they contribute to the induction and propagation of an antimicrobial immune response.

We thank Sabine Beyer (Infection Immunology Research Group/Helmholtz Centre for Infection Research), Ina Schleicher and Melanie Tillig (Central Facility for Microscopy/Helmholtz Centre for Infection Research), and Elena Katzowitsch (Core Unit Systems Medicine Würzburg) for excellent technical assistance.

The authors have no financial conflicts of interest.