Prior Exposure of Airway Epithelial Cells to Mycobacteria Reduces Subsequent Mycobacterium tuberculosis Infection and Resulting Inflammation

Abstract

Introduction

Repeated exposures to Mycobacterium tuberculosis (Mtb) and related species may influence host responses, which in turn may affect vaccine efficacy and could even render the host less or more susceptible to progression to active tuberculosis (TB) disease.

Methods

Using well-differentiated primary human bronchial epithelial cells (PBEC), we investigated the effect of a prior exposure of the epithelium to Mtb and Mycobacterium bovis vaccine strain Bacille Calmette-Guerin (BCG) on the intracellular infection efficiency of Mtb and Mycobacterium avium (Mav) during a second exposure and measured cytokine and antimicrobial peptide secretion.

Results

PBEC that were first exposed to BCG were significantly more resistant to subsequent infection with Mtb. A similar trend was observed in PBEC that were previously exposed to Mtb, although to a lesser magnitude compared to BCG pre-exposure. Furthermore, while the first exposure to mycobacteria induced inflammatory cytokine secretion by PBEC, cytokine secretion was dampened upon a secondary exposure to Mtb, most strongly in previously BCG-exposed cells. Secretion of the antimicrobial peptide hBD-2 was not affected by sequential exposures.

Conclusion

Repeated exposure of differentiated airway epithelial cells to mycobacteria reduced intracellular infection and inflammation.

Introduction

Tuberculosis (TB) is a serious infectious disease that mainly affects the respiratory tract and is caused by the bacillus Mycobacterium tuberculosis (Mtb). Per year, more than 10 million people continue to fall ill with TB [1]. In addition, cases of pulmonary disease caused by nontuberculous mycobacteria (NTM) such as Mycobacterium avium (Mav) are on the rise [2]. Infection with Mtb results most commonly from exposure of the respiratory mucosa to Mtb-containing aerosols expelled by individuals with active TB disease, while infection with Mav is most often acquired from environmental sources.

Due to the abundance of mycobacteria present in TB endemic areas, individuals are at increased risk of being repeatedly exposed to these pathogens, which in turn may affect host responses. For instance, multiple exposures to Mtb increased the risk of progression from Mtb infection to TB disease by 1.1 times per additional exposure in a Canadian cohort [3]. In contrast, altered host responses due to repeated exposure might also contribute to the resistance to TB and other infectious diseases found in high TB-endemic countries. For example, highly exposed household contacts from Ugandan patients with TB remained tuberculin skin test and interferon gamma release assay negative over long periods of follow-up [4]. Thus, there is no consensus on whether repeated exposures to mycobacteria are detrimental or beneficial, and on which factors this is dependent.

The Mycobacterium bovis strain Bacille Calmette-Guerin (BCG) is currently the only available vaccine against TB but offers only partial protection in adolescents and adults [5]. Previous studies have shown that the protection induced by BCG vaccination is at least in part due to trained immunity, the phenomenon of innate immunological memory characterized by either heightened or reduced host responses to pathogens following a recent exposure to (unrelated) pathogens or vaccines [6, 7]. In monocytes and macrophages, this can translate into a stronger inflammatory cytokine response and enhanced killing of bacteria, or alternatively a reduction in such responses, which indicates a tolerogenic response. A growing body of evidence suggests that BCG vaccination also confers cross-protection to other unrelated pathogens, again showcasing the induction of nonspecific protection characteristic of trained immunity. For example, countries with high BCG vaccine coverage had lower incidence of COVID-19 [8], although emergency BCG vaccination did not affect SARS-CoV-2 incidence or severity [9, 10]. Intriguingly, it has been suggested that exposure to NTM may influence the efficacy of the BCG vaccine against TB, either by already conferring partial protection due to their similarity to Mtb or by inducing regulatory T-cell responses which dampen the immune response [11–14]. Interestingly, individuals with a history of TB were found to be more susceptible to NTM infections as well [15]. Taken together, exposure to various species of mycobacteria, including BCG, can influence host responses to subsequent infections.

It has been demonstrated that mucosal administration of the BCG vaccine induced improved protection against Mtb compared to standard intradermal vaccination in rhesus macaques, which was correlated with increased frequencies of tissue-resident memory T cells in the mucosa [16, 17]. However, the roles of non-immune players like the lung epithelium have not been thoroughly investigated yet. As mucosal vaccination against TB is seriously considered, and the respiratory epithelium also plays a role in initiating host defenses against mycobacteria [18, 19], it will be relevant to further delineate epithelial responses to BCG.

Epithelial cells line the respiratory tract, and are in frequent contact with airborne microorganisms. As Mtb is transmitted via aerosols, airway epithelial cells are likely among the first cells encountering these pathogens. We and others have shown that airway epithelial cells can become infected by mycobacteria and mount a rapid host response [18, 20–22]. However, repeated exposures may affect the magnitude of the host response to pathogens. Innate immune memory is not exclusive to immune cells, and heightened responses to unrelated pathogenic stimuli have been described in various other cell types [23]. For example, pre-exposure of the human bronchial epithelial cell line BEAS-2B to flagellin either increased or reduced its inflammatory response, depending on the secondary stimulus given [24]. Another study using primary nasal progenitor cells demonstrated that these cells retain intrinsic memory of previous IL-4 and IL-13 exposures via epigenetic modifications [25]. Inflammatory memory was also identified in mouse skin epithelial stem cells, which maintained chromosomal accessibility to key stress response genes upon pre-exposure to a TLR7 agonist [26]. Although there are some indications that airway epithelial cells are affected by previous encounters, potential alterations in response to mycobacterial infections remain unknown. We hypothesized that there could be three possible outcomes when lung epithelial cells are repeatedly exposed to mycobacteria: first, repeated exposure might induce augmented responses and play a role in resistance to TB. A second possibility is that multiple exposures may reduce epithelial responses to mycobacteria, which may (adversely) affect the efficacy of mucosally delivered vaccines, including BCG [17]. And finally, multiple doses of mycobacteria may not alter epithelial responses at all.

Using air-liquid interface (ALI) cultures of differentiated primary human bronchial epithelial cells (PBEC) containing the major epithelial cell types, we investigated whether prior exposure to BCG or Mtb could alter infection percentage and intracellular load upon a second exposure to Mtb or Mav. In addition, we assessed the effect of multiple exposures on epithelial cell death, cytokine release, and antimicrobial secretion.

Methods

Source of Bronchial Epithelial Cells

Primary bronchial epithelial cells (PBEC) were isolated from macroscopically normal lung tissue obtained from patients undergoing resection surgery for lung cancer at the Leiden University Medical Center (LUMC), the Netherlands. Patients from which this lung tissue was derived were enrolled in the biobank via a no-objection system for coded anonymous further use of such tissue (www.coreon.org). Samples from this biobank were approved for research use by the LUMC Medical Ethical Committee, Approval No. BB22.006/AB/ab. Since September 1, 2022, patients are enrolled in the biobank using written informed consent in accordance with local regulations from the LUMC Biobank with approval by the LUMC Medical Ethical Committee, Approval No. B20.042/KB/kb. All experiments were performed in accordance with relevant guidelines and regulations. The study population included 7 individual donors, of whom 4 were female. Median age was 69 years (range 50–78). Of these donors, 2 were current smokers (28.6%), while the other 5 were ex-smokers. None of the donors were diagnosed with chronic obstructive pulmonary disease.

Bronchial Epithelial Cell Culture

PBEC were cultured as previously described [27]. In brief, PBEC (passage 1) of individual donors were expanded for 5 days in T75 culture flasks precoated with a mixture of 30 μg/mL PureCol (Advanced BioMatrix, Carlsbad, CA, USA), 5 μg/mL human fibronectin (PromoCell, Heidelberg, Germany) and 10 μg/mL BSA (Fraction V; Thermo Fisher Scientific, Carlsbad, CA, USA). Per experiment, cells of 3–4 donors were seeded at a density of 40,000 cells per donor on precoated 0.4 µm membrane pore size polyethylene terephthalate inserts (cellQART, Northeim, Germany) in 12-well plates. Cells were cultured in medium consisting of one part Bronchial Epithelial Cell Medium-basal (ScienCell, Carlsbad, CA, USA) and one part Dulbecco’s modified Eagle’s medium (StemCell Technologies, Vancouver, Canada) supplemented with Bronchial Epithelial Cell Growth Supplement (ScienCell), 12.5 mm HEPES (Gibco, Thermo Fisher Scientific), 100 U/mL penicillin, 100 μg/mL streptomycin (all from ScienCell), 2 mm glutaMAX (Thermo Fisher Scientific), further referred to as complete BD medium (cBD). In the submerged stage, cBD was supplemented with 1 nm EC 23 (Tocris, Bristol, UK). After reaching 100% confluence, apical medium was removed and cells were differentiated at the ALI in cBD supplemented with 50 nm EC 23. Culture medium was changed 3 times per week during which the apical side of the culture was carefully washed with 200 µL warm PBS for 10 min at 37°C to wash away mucus and cell debris. After 14 days of mucociliary differentiation at ALI, all dominant cell types were present and ciliary activity was observed. Cultures were subsequently used for experiments.

Bacteria

Mycobacteria were cultured in Difco Middlebrook 7H9 medium (BD Biosciences, Franklin Lakes, NJ, USA) containing 0.05% Tween-80 (Sigma-Aldrich, St. Louis, MO, USA), 0.2% glycerol (Sigma-Aldrich), and Middlebrook ADC growth supplement (750 mm BSA, 110 mm dextrose, 130 µm catalase in H₂O, produced in-house). A list of bacterial strains used and their origins can be found in online supplementary Table S1 (for all online suppl. material, see https://doi.org/10.1159/000550118). The BCG strain expressing GFP and the M. avium strain expressing Wasabi were cultured with 100 µg/mL hygromycin B (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). The M. tuberculosis H37Rv strain expressed Venus constitutively. BCG GFP and Mav Wasabi were diluted to OD₆₀₀ = 0.25 and Mtb Venus to OD₆₀₀ = 0.4 the day before experiments were performed, to ensure exponential growth of the bacteria.

Exposure of Airway Epithelial Cell Cultures to Mycobacteria

Before and during the first exposure, PBEC were cultured in medium supplemented with penicillin/streptomycin. Mtb Venus and BCG GFP were cultured as described above, and OD₆₀₀ of bacterial cultures was measured. Bacteria were HK for 30 min at 80°C. Then, a multiplicity of infection (MOI) of 100 was prepared accordingly in cBD containing 50 nm EC 23. The same batch of HK bacteria was used to perform all experiments. Mucus was washed from the apical surface of PBEC cultures with PBS. Immediately following mucus removal, 50 µL of HK bacterial suspension was added to the apical side of the cells. Control cultures received 50 µL medium without bacteria. Next, cultures were centrifuged briefly at 300 g for 2 min to spin bacteria down to the cell layer, and incubated at 37°C for 24 h [28]. After this initial exposure, PBEC were washed with PBS to remove stimuli. PBEC were incubated at 37°C for 5 days. At least 48 h before the second exposure, PBEC were cultured in medium without antibiotics. For a second exposure, live Mtb Venus and Mav Wasabi were prepared at MOI 100 as described above, and mucus was washed from the apical surface of PBEC with PBS. Following mucus removal, 50 µL of live bacterial suspension was added to the apical side of PBEC, while control cultures received medium without bacteria. PBEC were centrifuged briefly at 300 g for 2 min and incubated at 37°C for 24 h. PBEC were then washed with PBS and treated apically with 30 µg/mL gentamicin for 15 min at 37°C to kill extracellular bacteria. Cultures were then washed again with PBS and incubated another 24 h. For experiments using IFN-γ pre-exposure, PBEC were incubated with 10 ng/mL IFN-γ (Thermo Fisher Scientific) basally for 1 h prior to receiving their first exposure to HK Mtb or BCG. This concentration of IFN-γ was chosen based on concentrations used by others to stimulate PBEC, monocytes, and macrophages [29, 30]. Apical and basal samples were taken for Luminex and LDH assays (see below) 24 h after the first exposure (T1), after the resting period immediately before the second exposure (T2), and 24 h after the second exposure (T3). The cells were prepared for flow cytometry and the CFUs assay as described below. During incubations, PBEC were supplied basally with 1 mL new medium every 3 days and always 24 h before supernatant collection. During experiments, the integrity of the PBEC cell layer was evaluated every 2–3 days via light microscopy.

Flow Cytometry

Basal medium from PBEC cultures was replaced with calcium-free PBS and 200 µL was added to the apical side. Cultures were incubated 30 min at 37°C. PBS was removed, and 200 µL pre-warmed TrypLE Express (Gibco, Thermo Fisher Scientific) containing 10 µm Y-27632 (StemCell Technologies) was added onto the cell layer. Cultures were incubated for 15 min at 37°C. After pipetting up and down to break up the cell layer, single cells were collected from the inserts and transferred to tubes containing 800 µL PBS. Of each sample, 100 µL of the total volume (10%) was transferred to a separate tube for the CFUs assay (see below), and the remaining cells were used for flow cytometry. Samples were stained with Violet Fixable Live/Dead dye (Invitrogen, Thermo Fisher Scientific), 1:800 in PBS for 30 min at room temperature, then washed with PBS. Samples were fixed in 4% PFA for 10 min, then measured in fixative with a BD FACS Lyric 3L12C (BD Biosciences) cytometer.

CFU Assay

Single cell samples collected in tubes during sample preparation for flow cytometry were treated with 30 µg/mL gentamicin (Sigma-Aldrich) for 15 min at 37°C to kill extracellular bacteria, then washed with PBS. Cells were lysed in H₂O + 0.05% SDS (Invitrogen, Thermo Fisher Scientific) for 15 min at RT. Then those lysates were washed with PBS, centrifuged at 800 g, and resuspended in 100 µL PBS. Serial dilutions of the lysates were plated on Difco 7H10 agar (BD Biosciences) plates and incubated at 37°C. The number of colonies formed was determined by manual count.

LDH Assay

PBS washes from the apical side and cell culture medium from the basal side of repeatedly exposed PBEC were used to measure LDH release with the Cytotoxicity Detection Kit (LDH) from Roche (Basel, Switzerland). To create a positive control sample, 200 µL PBS + 2% Triton X-100 was added to the apical side of 2 nonexposed PBEC cultures for 5 min at RT. From the apical and basal sides, 100 µL was collected in a flat bottom 96-wells plate. As background controls, 100 µL cBD and 100 µL PBS were used. Then, samples were centrifuged 3 min at 805 g to spin down any cells and bacteria. Supernatants were transferred to a new flat bottom 96-wells plate. Reaction mixture from the kit was prepared according to manufacturer’s instructions and 100 µL was immediately added to each sample. Samples were left to incubate 28–30 min in the dark at RT, and then OD₄₈₅ and OD₆₉₀ were measured on an Envision Multimode Plate Reader (PerkinElmer). The percentage cytotoxicity was calculated with the following equation: $\text{Cytotoxicity}\left(\%\right)=\frac{Samplevalue-Background}{Positivecontrol-Background}\times 100$.

Luminex Assay

After the first exposure (T1), before the second exposure (T2), and after the second exposure to mycobacteria (T3), 200 µL apical supernatant and 1 mL basal medium were collected from each cell culture and stored at −20°C until processing. At all time points, PBEC had received new cBD medium and were washed apically with PBS 24 h before sample collection. Supernatant and medium were filtered in FiltrEX 96-well filter plates (Corning Costar) with a pore size of 0.2 µm to remove bacteria. Supernatants and medium were diluted 4 times with Luminex Assay buffer (Bio-Rad, Hercules, CA, USA). Following manufacturer’s instructions, the Bio-Plex Pro Human Cytokine 48-plex Assay (Bio-Rad) was performed. Samples were measured on a Bio-Plex 200 System (Bio-Rad). Per analyte, a lower and upper limit of detection was determined with standard curves. Concentrations measured below the assays’ detection limit were set to 1 pg/mL and those measured over the detection limit were set to the maximum quantifiable pg/mL per analyte. As apical wash was collected in a 5 times smaller volume than basal medium, concentrations retrieved in the groups were corrected for differences in sample volume before further analysis.

Antimicrobial Peptide ELISA

Levels of secreted antimicrobial peptides in apical supernatants from T1, T2, and T3 were assessed by ELISA. Samples were diluted 2 times to obtain sufficient volume to perform assays. Levels of beta-defensin 2 (hBD-2) were measured using the Human Beta-2 Defensin ELISA Development Kit (ABTS) (PeproTech, Thermo Fisher Scientific) according to manufacturer’s instructions. Levels of cathelicidin (LL-37) were measured using the LL-37 Human ELISA Kit (Hycult Biotech, Uden, The Netherlands) according to manufacturer’s instructions. Using standard curves, upper and lower limits of detection were determined.

Statistics

Data were analyzed in GraphPad Prism version 10.2.3. All data were determined to be non-parametric by Shapiro-Wilk normality testing. Statistical differences in infection percentage, CFU, LDH, and antimicrobial peptide secretion were determined by Friedman test with Dunn’s correction for multiple comparisons. Statistical differences in cytokine secretion were determined by generalized linear mixed model with Dunnett’s correction for multiple comparisons. Principal component analysis of cytokine data was performed in R Studio version 4.1.2. R packages used were as follows: caret, tidyverse, tidyr, ggplot2, ggpubr, ggbreak, reshape2, viridis, plotrix, ggrepel, mixOmics, scales, dplyr, and ctrlGene. Heatmap clustering was performed in Morpheus (https://software.broadinstitute.org/morpheus) using the One Minus Spearman Rank hierarchical clustering method. Threshold for statistical significance used was p < 0.05. Significance is indicated in figures as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Results

Prior Exposure to BCG Reduces Mtb Infection in Bronchial Epithelial Cells

To investigate the effects of prior exposure of airway epithelium to mycobacteria, we used PBEC differentiated at the ALI. This robust model has previously been characterized in depth and consistently displayed proper barrier function and cell differentiation, containing the major cell types such as basal, goblet, ciliated, and club cells [27]. PBEC were first exposed to heat-killed (HK) Mtb or HK BCG for 24 h, as a model of natural exposure or airway vaccination. HK bacteria were chosen to mimic a primary exposure but also ensure that no live bacteria from this exposure remained that could interfere with the secondary infection. Five days after this initial exposure, PBEC received a second exposure, this time to live Mtb or Mav for 24 h. As we wished to study intracellular infection only, this was followed by a short gentamicin treatment to kill extracellular bacteria [18], after which PBEC were cultured for an additional 24 h (Fig. 1a).

Fig. 1.

Repeated exposure of bronchial epithelial cells to mycobacteria reduces infection. a Timeline showing the experimental setup and timing of exposures. b Flow cytometry plots from a representative experiment showing percentages of infected cells in single exposed and repeated exposed PBEC. c, d Bar graphs of infection percentages of Mtb and Mav, respectively, in PBEC with multiple exposures to mycobacteria. Data points represent the average of 2 technical replicates per donor. e, f Bar graphs showing intracellular colony-forming units (CFU) recovered from repeatedly exposed PBEC. Percentages were calculated using PBEC with a single Mtb exposure as baseline. g Bar graphs of cytotoxicity as measured by LDH release assay in PBEC with repeated exposures. All data are shown as median with range. Statistics performed using Friedman test with Dunn’s correction for multiple comparisons. Statistical significance indicated as *p < 0.05. Ex, exposure; T, Mtb; B, BCG; A, Mav; HK, heat-killed. Blue bars indicate Mtb was used during secondary infection, and green bars indicate Mav was used.

The percentages of infected cells were measured by flow cytometry (Fig. 1b), and the gating strategy is shown in online supplementary Figure S1A (for all online suppl. material, see https://doi.org/10.1159/000550118). Infection percentages varied between Mtb and Mav, with Mtb having a slightly higher infection efficiency in most donors during the primary exposure (Fig. 1c, d). With prior exposure to BCG, the infection percentage of Mtb was significantly decreased in PBEC (Fig. 1c). Interestingly, marked inter-donor variability was observed, with some affected by prior BCG exposure while others were not (online suppl. Fig. S1B). Mtb infection percentages tended to be lower in PBEC that had received a prior Mtb exposure, although this was not statistically significant. Mav infection percentages also tended to be lower in PBEC that were previously exposed to Mtb, but this did not reach statistical significance (Fig. 1d).

The total intracellular bacterial load per PBEC culture was determined via colony-forming unit (CFU) assay. In BCG-exposed PBEC, the burden of Mtb was decreased by approximately 20% compared to PBEC that were not previously exposed to mycobacteria (Fig. 1e). This corresponded well with our flow cytometry observations. PBEC that were exposed twice to Mtb did not show a reduction in CFU (Fig. 1e) and the bacterial burden even increased in several donors compared to PBEC without prior exposure (online suppl. Fig. S2A). The bacterial burden of Mav did decrease by approximately 40% upon prior exposure to Mtb, although this reduction was not statistically significant (Fig. 1f; online suppl. Fig. S2A).

To determine whether the reduced Mtb infection percentage and CFU in BCG-exposed PBEC could be attributed to increased cell death, we performed a lactate dehydrogenase (LDH) release assay (Fig. 1g). There was no significant increase in LDH release on both the apical and basal side of PBEC cultures that received multiple exposures compared to PBEC that received a single exposure. Thus, we conclude that the observed decreased infection rate is not due to increased cell death upon repeated exposure.

Reduction of Mtb Infection due to Prior BCG Exposure Is Unchanged by Pre-Exposure to IFN-γ

We hypothesized that the induction of heightened host responses during sequential exposures may be partially mediated by signals from the innate or adaptive immune system, such as the effector molecule IFN-γ, hallmark of adaptive immunity, which plays a pivotal role in TB. For instance, IFN-γ can play a role in macrophage activation and contributes to Mtb killing [31, 32]. In addition, a recent study by Kain et al. [33] showed that exposure to IFN-γ could induce enhanced bacterial killing and metabolic rewiring in mouse bone marrow-derived macrophages upon Mav infection. As airway epithelial cells also express type II IFN receptors [34], we hypothesized that pre-exposure of PBEC to IFN-γ might modulate the capacity of these cells to respond to multiple exposures of mycobacteria. To mimic the presence of an adaptive immune response, PBEC were incubated with IFN-γ for 1 h, after which it was removed prior to receiving a first exposure to HK Mtb or BCG. After 5 days, PBEC received a second exposure to live Mtb or Mav for 24 h to investigate infection of the epithelial cells. Again, this was followed by a short gentamicin treatment, after which PBEC were incubated another 24 h (Fig. 2a).

Fig. 2.

Infection is reduced in repeatedly exposed epithelial cells despite pre-exposure to IFN-γ. a Timeline showing the experimental setup and timing of exposures. b Bar graph comparing infection percentages in conditions with and without IFN-γ pre-exposure. Striped bars show data as presented in Figure 1c, d. c, d Bar graphs of infection percentages of Mtb and Mav, respectively, in PBEC with multiple exposures to mycobacteria and prior stimulation with IFN-γ. Data points represent the average of 2 technical replicates per donor. e, f Bar graphs showing intracellular colony-forming units (CFUs) recovered from repeatedly exposed PBEC with prior stimulation with IFN-γ. Percentages were calculated using PBEC with a single Mtb exposure as baseline. g Bar graphs of cytotoxicity as measured by LDH release assay in PBEC with repeated exposures and prior stimulation with IFN-γ. All data are shown as median with range. Statistics performed using Friedman test with Dunn’s correction for multiple comparisons. Statistical significance indicated as *p < 0.05. Ex, exposure; T, Mtb; B, BCG; A, Mav. Blue bars indicate Mtb was used during secondary infection, and green bars indicate Mav was used.

IFN-γ pre-exposure itself did not influence infection percentages during the various exposures (Fig. 2b). With prior IFN-γ and BCG exposure, the infection percentage of Mtb was significantly decreased compared to cells that received IFN-γ and a single Mtb exposure (Fig. 2c), although the size of this effect was donor-dependent (Fig. S1b). PBEC receiving two Mtb exposures exhibited a small decrease in infected cells. Furthermore, Mtb exposure followed by Mav infection tended to reduce Mav infection percentages, although this was not statistically significant (Fig. 2d). CFU levels were compatible with the flow cytometry data, with a reduction in Mtb bacilli in BCG-exposed PBEC and a reduction in Mav intracellular load in Mtb-exposed PBEC of some donors (Fig. 2d–f; online suppl. Fig. S2B). Interestingly, although underpowered, in a post hoc analysis, we observed a trend of female donors responding more strongly to BCG exposure than male donors in the absence of IFN-γ, as shown by a stronger reduction in infection percentage (online suppl. Fig. S3A–D). However, when given an IFN-γ pre-exposure, cultures from male donors tended to control intracellular loads more strongly than cultures from female donors, which performed worse with addition of IFN-γ. There was no significant increase in LDH release upon IFN-γ stimulation followed by multiple exposures compared to single exposures (Fig. 2g; online suppl. Fig. S3E), indicating that the reduction in infection rate cannot be attributed to cell death. Taken together, these data show that the presence of IFN-γ, in the concentration and time window tested, did not influence the epithelial response to Mtb or Mav infection in PBEC upon repeated exposure.

BCG Induces a Stronger Inflammatory Response than Mtb during First Exposure

We have previously demonstrated that airway epithelial cells secrete numerous cytokines and chemokines in response to infection with mycobacteria [18]. Here, we sought to investigate differences in cytokine profiles induced by HK BCG and HK Mtb, as HK BCG pre-exposure reduced subsequent Mtb infection more strongly than HK Mtb pre-exposure. Apical wash and basal medium were collected at three time points: 24 h after the first exposure (T1), 6 days after first exposure (T2), and 24 h after second exposure (T3) (Fig. 3a). At T1, the cytokine response induced by HK BCG was stronger than the response induced by HK Mtb (Fig. 3b, c). This was particularly evident in apical secretions (Fig. 3b). Cytokines that were significantly upregulated by exposure to HK BCG included several chemoattractants like CCL2, CCL4, and G-CSF, as well as inflammatory cytokines like IL-8 and TNF-α (Fig. 3d). This corresponded well with observations from our previous study, even though those data were derived from mixed donor pools rather than single donors [18]. Cytokines downregulated by HK BCG exposure included SCGF-β and, surprisingly, IFN-γ. The same trends were often observed in secretions from PBEC exposed to HK Mtb, although the magnitude of the response was lower compared to HK BCG.

Fig. 3.

Exposure of PBEC to HK mycobacteria induces an inflammatory cytokine response. a Timeline depicting the time points at which supernatants were collected, indicated by blue arrows. Heatmaps of median log₂ fold changes in apical (b) and basal (c) cytokine secretion by PBEC 24 h after a first exposure with HK BCG or Mtb, compared to uninfected PBEC. d Bar graphs depicting secretion levels of cytokines that were significantly up- or downregulated by BCG or Mtb exposure in PBEC. Statistical significance determined by generalized linear mixed model with Dunnett’s correction for multiple testing and significance indicated as *p < 0.05.

Principal component analysis demonstrated that the largest variance in the data could not be explained by the various exposures, time points, or sex of the donors. Rather, the difference in apical and basal profiles caused the largest variance (online suppl. Fig. S4A). Hierarchical clustering of the median log₂ fold change induced by the different exposures compared to uninfected PBEC revealed that the different time points cluster mostly together in both apical and basal profiles (online suppl. Fig. S4B, C).

Intriguingly, at T2, PBEC that received HK BCG decreased apical cytokine secretion compared to uninfected PBEC, while those that received HK Mtb still had slightly increased cytokine secretion (online suppl. Fig. S5A). Basally, cytokine profiles between the groups were very similar at T2, and secretion was mostly decreased compared to uninfected PBEC (online suppl. Fig. S5B). Significantly downregulated cytokines at T2 included M-CSF and IL-1β (online suppl. Fig. S5C).

BCG-Exposed Epithelial Cells Decrease Secretion of Inflammatory Cytokines and Chemokines upon Secondary Exposure

After a second exposure to live Mtb, apical cytokine secretion was reduced in PBEC with prior exposure to HK BCG and HK Mtb compared to PBEC that received only the single live Mtb exposure, and this effect was stronger for HK BCG than for HK Mtb (Fig. 4a). Basally, previous exposure to HK BCG decreased cytokine secretion, while HK Mtb increased it (Fig. 4b). Overall, both inflammatory cytokines like IL-1β, IL-8, TNF-α, and TNF-β as well as chemokines such as CCL4, CCL11, and G-CSF were significantly downregulated by BCG-exposed PBEC (Fig. 4c). While Mtb-exposed cells often followed this trend of reduction, there are multiple cytokines for which repeated exposure to Mtb slightly increased expression either apically or basally, such as IL-1β and SCF. These results indicate that HK BCG and HK Mtb may partly modulate epithelial responses to subsequent exposures differently.

Fig. 4.

Prior exposure of PBEC to HK mycobacteria reduces inflammation upon live Mtb infection. Heatmaps of median log₂ fold changes in apical (a) and basal (b) cytokine secretion by PBEC 24 h after a second exposure with live Mtb, compared to PBEC that received only a single exposure to live Mtb. c Bar graphs depicting secretion levels of cytokines that were significantly up- or downregulated after second exposure. Statistical significance determined by generalized linear mixed model with Dunnett’s correction for multiple testing and significance indicated as *p < 0.05.

Repeated Exposures to Mycobacteria Do Not Upregulate Secretion of the Antimicrobial Peptide hBD-2

To examine whether the observed reduction in percentage infected cells upon prior exposure to HK bacteria was due to increased antimicrobial peptide secretion, the antimicrobial peptides beta-defensin 2 (hBD-2) and cathelicidin (LL-37) were measured in apical supernatants from T1, T2, and T3 (Fig. 5). At T1, hBD-2 secretion was slightly increased by stimulation with HK BCG, albeit not significantly. At T2 and T3, however, there were no differences in the secretion of hBD-2 between uninfected PBEC and PBEC stimulated with HK mycobacteria. LL-37 was not detected in secretions of PBEC in any of the conditions tested (data not shown). Thus, secretion of hBD-2 was not significantly affected by repeated exposures to mycobacteria and is therefore unlikely to be the mechanism of reduced bacterial loads.

Fig. 5.

Antimicrobial hBD-2 secretion is not altered by repeated exposure to mycobacteria. Bar graphs demonstrating secretion levels of beta-defensin 2 (hBD-2) in apical secretions of PBEC 24 h after first exposure (T1), before second exposure (T2), and 24 h after second exposure (T3). Statistical significance determined by Friedman test with Dunn’s correction for multiple comparisons. Ex, exposure; T, Mtb; B, BCG; A, Mav.

Discussion

In areas highly endemic for TB, individuals may be exposed to Mtb as well as NTM species multiple times in relatively short time spans, which may influence their risk of developing TB. It was unknown whether airway epithelial cells are affected by previous encounters with mycobacteria. In the present study, we demonstrate that bronchial epithelial cells modify their response in answer to prior exposure to mycobacteria, resulting in a reduction in the proportion of Mtb-infected cells and a trend toward reduced Mav-infected cells and reduced intracellular CFU. Furthermore, we show that the secretion of several cytokines was modulated differently by HK BCG and HK Mtb during a first exposure. The cytokine response to a subsequent exposure to live Mtb was dampened in PBEC with prior mycobacterial exposure, more strongly so in BCG-exposed PBEC. Finally, both cell death and secretion of antimicrobial peptides were not significantly affected by repeated exposures to mycobacteria. To our knowledge, this is the first report of airway epithelial cells exhibiting altered host responses after repeated exposure to mycobacteria.

In innate immune cells, exposure to BCG can induce trained immunity, a term coined by Netea et al. [35] to describe nonspecific innate immunological memory. In short, this encompasses a first exposure, leading to a heightened response toward a second unrelated stimulus, which may consist of enhanced pathogen killing, increased cytokine secretion, or both. These effects can be maintained for months via metabolic and epigenetic reprogramming. The reduced infection rates that we report suggest that epithelial cells also respond differently to consecutive infections and may display a form of trained mucosal immunity. Overall, inflammation levels decreased during a secondary infection with Mtb if PBEC had been pre-exposed to mycobacteria, which might be a direct consequence of decreased infection. A proof-of-concept study by Bigot et al. [24] found that levels of IL-8, a cytokine highly secreted by epithelial cells, were increased in bronchial BEAS-2B cells during secondary exposure. However, this was highly dependent on the combination of stimuli given. A study using mouse type 2 alveolar epithelial cells found that sequential stimulation with TLR ligands resulted in partial tolerance as defined by reduced cytokine expression, which was reversible by histone deacetylase inhibitors, indicating epigenetic modifications as drivers of this phenotype [36]. However, these studies did not measure infection efficiency of live pathogens in epithelial cells during repeated exposures.

The innate and adaptive immune response might be instrumental in inducing heightened host responses via IFN-γ, as it has been shown that macrophages stimulated overnight with IFN-γ had increased intracellular Mtb killing capacity [31]. When PBEC were exposed to IFN-γ (in this dose and time window), the percentage infected cells and intracellular CFU of both Mtb and Mav were neither increased nor reduced compared to PBEC that did not receive IFN-γ. In another study, PBEC pre-stimulated for 24 h with the same concentration of IFN-γ that was used here (10 ng/mL) reduced infection with respiratory syncytial virus, likely via epigenetic upregulation of RIG1 [29], suggesting that the concentration used here could be sufficient to effectively stimulate PBEC. However, a longer stimulation period might be required as we stimulated for only 1 h. However, murine skin epithelial cells were shown to maintain a prolonged inflammatory memory resulting in improved wound healing independent of CD45⁺ tissue-resident immune cells, suggesting that IFN-γ is not required to induce these effects [26]. Thus, the altered response of PBEC is likely not influenced by the hallmark adaptive immune signal IFN-γ, at least under the conditions tested. However, these experiments do not rule out an effect of IFN-γ in vivo, where repeated IFN-γ exposures or higher doses may result in a measurable impact on epithelial responses to mycobacteria.

Exposure of PBEC to HK BCG and HK Mtb led to similar outcomes, but at differing magnitudes. HK BCG-exposed cells controlled subsequent Mtb infection more strongly than HK Mtb-exposed cells and naïve cells. Interestingly, cells stimulated with HK BCG during a first exposure secreted higher levels of inflammatory cytokines and chemokines than cells stimulated with HK Mtb, indicating that these stimuli may differentially influence host responses to future infections. Although the genomes of BCG and Mtb are over 90% homologous, many important Mtb antigens and virulence factors are absent in BCG [37, 38]. Considering that we first stimulated PBEC using bacteria that were HK, it is likely that differences in cell wall-associated antigens are driving this effect. Apart from missing several virulence factors, BCG differs from Mtb also in the expression levels of cell wall proteins. A study comparing the membrane proteomes of BCG and Mtb found that nearly 300 membrane-associated proteins were significantly differentially expressed with at least two-fold difference in abundance between these species [39], which might in part explain the differences in PBEC responses. Interestingly, it was shown in mouse hematopoietic stem cells that BCG-induced trained immunity, while Mtb induced a tolerogenic response via type I interferon [40]. With regard to our data, while the reduction in cytokine secretion may fit a tolerant phenotype, the decrease in infection may be indicative of a trained immunity-like response.

Mucosal BCG vaccination offers increased protection against Mtb challenge compared to standard intradermal vaccination in macaques [17, 41]. Our results suggest that exposure of bronchial epithelial cells to BCG directs subsequent responses to Mtb infection, which may affect the efficacy of mucosally delivered BCG vaccines. First, intracellular Mtb bacilli in epithelial cells are residing in a relatively immune-sequestered niche, as epithelial cells are not as efficient at degrading pathogens and co-stimulation as “professional” immune cells like macrophages [42], which may hamper or delay antigen presentation. We show that pre-exposure to BCG reduced intracellular Mtb loads during a subsequent infection. This may be due to increased bacterial killing, either intracellularly or by secreted factors. For example, antimicrobials other than hBD-2, which was not differentially induced by repeated exposure, could enhance extracellular killing. Alternatively, the reduction in infection may be due to reduced phagocytosis, possibly via reduced pathogen sensing. Regardless of the underlying mechanism, mucosal BCG vaccination may reduce epithelial infection and thereby help curb Mtb infection more efficiently. Second, a reduction of inflammatory cytokines and chemokines may attract fewer immune cells to the site of infection. This could benefit the host by both limiting access of Mtb to macrophages, its preferred host niche, and limiting lung injury due to excessive inflammation [43]. Elevated levels of cytokines like IL-6, IL-8, and TNF-α in bronchoalveolar lavage fluid of patients with active TB disease correlated with increased granuloma cavitation, fibrosis, and bronchial wall thickening [44]. Thus, modulation of epithelial responses may be an additional consequence of mucosal vaccination, in addition to previously described effects on the immune system [17, 41]. Whether the beneficial effect of prior BCG exposure on epithelial responses to Mtb also occurs in vivo and with different vaccine strains remains to be investigated.

We observed different responses to repeated mycobacterial exposure in PBEC derived from different donors. In addition, some donors were readily infected by mycobacteria while others were more resistant, and within donors, the response to Mtb and Mav infection varied as well. What underlies this donor variation is still unclear. However, a reduction in Mtb infection percentage was observed in PBEC from all donors when previously exposed to BCG. A recent study demonstrated that in healthy Dutch adults, approximately a quarter of individuals have monocytes with the ability to control Mtb outgrowth prior to vaccination, while a quarter acquires this capability post-vaccination. Finally, half of the vaccines were not able to control outgrowth at all [45]. Such a preexisting ability to control mycobacterial outgrowth may also exist in epithelial cells and could be an explanation for several donors being resistant to infection with live Mtb or Mav. Another explanation might be that these donors have previously already been exposed to mycobacteria, as various NTM species are abundantly present in the environment and thus already exhibit altered responses. Another interesting observation we made was the difference in PBEC responses between female and male donors, although these groups were very small and we therefore did not statistically test these observations. Cells from female donors tended to respond to BCG pre-exposure by reducing Mtb infection more strongly than those from male donors. The efficacy of BCG vaccination was reported to be slightly higher in females than males in a South Korean study: 78.6% versus 65.8% [46]. This difference has also been observed in mice, where females are better protected against TB than males and had lower pulmonary bacterial burdens [47]. This is a fascinating observation, especially as it appears to be maintained in in vitro epithelial cell culture as well. However, our data need further confirmation and potential mechanisms underlying these differences have not yet been elucidated.

During first stimulations, HK bacteria were used instead of live bacteria to avoid the possibility of bacteria persisting in the cell cultures in between stimulations. However, HK bacteria may induce different responses than live bacteria. We have previously shown that HK bacteria induce largely the same gene expression and cytokine profiles in PBEC as live bacteria, albeit less potently [18]. While some minor differences exist, the influences of heat-killing on shaping the global host response are likely to be relatively small. In addition, the small number of donors used in this study is a limitation. Furthermore, we have used a supraphysiological dose of MOI 100 for stimulations of PBEC. This was demonstrated to result in a detectable number of Mtb-infected epithelial cells in an earlier publication [18]. It is unlikely that such concentrations of mycobacteria are present in the lungs during early infection, although high concentrations may occur locally upon bursting of granulomas. Finally, our in vitro model did not include immune cells, whose interaction with epithelial cells may also influence the outcome of repeated exposures.

In conclusion, we demonstrated that prior exposure of airway epithelial cells to different species of mycobacteria leads to increased resistance to subsequent infection, particularly when cells were previously exposed to BCG. This phenomenon was present in the absence of innate or adaptive immune cells. Furthermore, cytokine secretion was dampened by consecutive exposures, and the secretion of two major antimicrobial peptides was not impacted. These findings offer new insights into the roles of epithelial cells during mycobacterial lung diseases and suggest a potential reason for the highly variable reactions to such infections. Finally, our findings also likely have implications for further mucosal vaccine development.

Acknowledgments

The authors gratefully thank Daaf G. B. van der Geest for his contributions to protocol optimization and Prof. Dr. William R. Jacobs (Albert Einstein College of Medicine, Yeshiva University, New York City, USA) for kindly providing the Mtb strain constitutively expressing Venus.

Statement of Ethics

Lung tissue was derived from patients undergoing lung resection surgery for lung cancer in the Leiden University Medical Center (LUMC). These patients were enrolled in a biobank via a no-objection system for coded anonymous further use of such tissue (www.coreon.org). Samples from this biobank were approved for research use by the LUMC Medical Ethical Committee, Approval No. BB22.006/AB/ab. Since September 1, 2022, patients are enrolled in the biobank using written informed consent in accordance with local regulations from the LUMC Biobank with approval by the LUMC Medical Ethical Committee, Approval No. B20.042/KB/kb. All experiments were performed in accordance with relevant guidelines and regulations.

Conflict of Interest Statement

The authors declare that they have no competing interests or other interests that might be perceived to influence the results and/or discussion reported in this paper.

Funding Sources

This work was supported by a grant from the Leiden-Edinburgh Joint PhD program for Integrated One Health Solutions, awarded by the universities of Leiden and Edinburgh in December 2019.

Author Contributions

A.B. collected and analyzed data and drafted the article. D.N. performed all cell culture work to create the PBEC in vitro models. K.W. collected data for flow cytometry and colony-forming unit assays. P.H., T.O., A.D., and S.J. served as scientific advisors and critically reviewed the study design and article.

Funding Statement

This work was supported by a grant from the Leiden-Edinburgh Joint PhD program for Integrated One Health Solutions, awarded by the universities of Leiden and Edinburgh in December 2019.

Data Availability Statement

Flow cytometry, CFU, LDH, ELISA, and cytokine data have been deposited on Mendeley Data, doi: 10.17632/n6s69y3mcd.1. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request (Simone Joosten, S.A.Joosten@lumc.nl).