Abstract
Introduction:Haemophilus influenzae type a (Hia) has recently emerged as a cause of invasive disease in North American Indigenous children. Factors determining the outcomes of exposure to the pathogen, from asymptomatic carriage to fatal disease, are poorly understood. The role of innate immune activation in the pathogenesis of invasive Hia disease remains unexplored. We used clinical Hia isolates to determine whether innate immune responses depended on the presence of the capsule, strain genetic background, and abilities to cause invasive disease. Methods: Differentiated THP-1 cells and HL-60 neutrophil-like cells were stimulated with four Hia strains (invasive or noninvasive; encapsulated or nonencapsulated), in comparison to 1 invasive and 1 noninvasive non-typeable H. influenzae. Surface expression of ICAM-1 and CD64 and release of pro-inflammatory cytokines TNF-α and IL-1β were quantified. Results: In vitro Hia infection resulted in robust activation of inflammatory responses in terms of expression of ICAM-1 and release of TNF-α and IL-1β, irrespective of the presence or absence of the capsule, or abilities to cause invasive disease. Inhibition of TLR4 decreased TNF-α release by THP-1 cells stimulated by Hia. Conclusion: Powerful activation of pro-inflammatory responses induced by Hia may contribute to the pathogenesis of invasive Hia disease. As the activation of macrophages and neutrophils did not depend on encapsulation or source of Hia isolation, the functional abilities of phagocytic cells unlikely represent a limiting factor in host defenses. The development of invasive versus noninvasive disease may depend on the functional abilities of the adaptive immune system.
Introduction
Haemophilus influenzae is a human-restricted Gram-negative pathogen able to cause severe invasive infections such as sepsis and meningitis [1]. The most virulent strains express a polysaccharide capsule that protects the bacteria against immune defense mechanisms. The capsules of various strains differ in chemical structure: six serotypes (a-f) are distinguished based on the antigenic properties of capsular polysaccharides; nonencapsulated H. influenzae strains are termed non-typeable (NTHi) [1]. Experimental studies have established that among all H. influenzae, the serotype b (Hib) is the most virulent followed by serotype a (Hia); NTHi strains are less virulent than encapsulated ones [2, 3]. Prior to the introduction of the pediatric Hib-conjugate vaccines in the late 1980s to the early 1990s, Hib was the most common cause of meningitis in children globally [4]. Immunization against Hib resulted in a dramatic decrease of the incidence rates of invasive Hib disease [4]. However, in the post-Hib vaccine era, NTHi and non-type b encapsulated H. influenzae became the major causes of invasive disease in industrial countries with universal pediatric immunization programs [5].
Over the last two decades, Hia became the significant cause of severe invasive disease in several North American regions with a high proportion of Indigenous populations, particularly in the Arctic [6]. In some Indigenous populations, incidence rates of invasive Hia disease are close to invasive Hib prior to the introduction of immunization against Hib, with young children being most affected [7]. The invasive Hia disease can manifest as very severe clinical forms, such as meningitis, pneumonia, or septic arthritis, with high case-fatality rates [6, 8]. Yet, Hia infection is characterized by a wide clinical spectrum as Hia can cause local infections as well as asymptomatically colonize the upper respiratory tract [9‒11]. Our recent studies found that over 9% of healthy Indigenous children carried Hia in the nasopharynx [12].
Factors determining the outcomes of exposure of young children to the pathogen, in the range of asymptomatic carriage to fatal disease, are poorly understood. Molecular genetic studies of encapsulated H. influenzae have identified a strong association of bacterial virulence with mutations increasing the amount of capsular material, such as the IS1016-bexA deletion in the cap locus [13, 14]. Analysis of clinical Hib and Hia isolates revealed that bacteria exhibiting such mutations caused extremely severe inflammatory responses in vivo [15, 16], suggesting that the capsule may contribute to inflammation via innate immune activation. However, the mechanisms mediating an enhanced inflammation caused by mutant capsular forms of H. influenzae are unknown.
Although the role of adaptive immunity in host defense against encapsulated H. influenzae has been established, the role of innate immunity is still incompletely understood [17]. It is known that recognition of H. influenzae by the innate immune system involves certain pattern recognition receptors (PRRs), such as Toll-like receptors (TLR) 2 and 4 [18, 19]. H. influenzae, which is a typical extracellular pathogen, can nevertheless enter the cytosol and consequently reach intracellular NOD-like receptors [20‒22]. Some studies have demonstrated that TLR2, TLR4, and NOD1 are all required for mucosal clearance of capsulated (Hib), but not of nonencapsulated strains [23], suggesting that the capsule is recognized by PRRs, and the resulting signaling is essential for innate immune responses. However, the role of H. influenzae capsule in the activation of innate immune responses remains unexplored.
In this study, we used Hia strains isolated from patients with various forms of clinical disease, to determine whether in vitro innate immune responses depend on the presence of the capsule, strain genetic background, and abilities to cause invasive versus noninvasive disease. To normalize the outcome of the study, we applied standardized methods of stimulating two human cell lines with several bacterial strains under the same conditions.
Methods
Cell Culture Conditions
The human THP-1 monocytic leukemia cell line (ATCC, Manassas, VA, USA) was cultured in RPMI-1640 medium (Sigma-Aldrich, Oakville, ON, Canada) supplemented with 20% fetal bovine serum (FBS) (R&D Systems, Inc., Minneapolis, MN, USA) and 0.4% antibiotic-antimycotic (Gibco, Eugene, OR, USA) at 37°C and 5% CO2. The cells were passaged every 3–5 days, to maintain an approximate concentration of 1 × 106 cells/mL. Macrophage differentiation was induced adding 20 ng/mL phorbol myristate acetate (Sigma-Aldrich, Oakville, ON, Canada) to 2.5 × 106 cells/mL in 6-well plates (Corning Incorporated, Corning, NY, USA) followed by incubation for 24 h at 37°C and 5% CO2. The medium was then removed, cells were washed with RPMI-1640 medium, and 2 mL of RPMI-1640 supplemented with 10% FBS was added for another 48 h incubation at 37°C and 5% CO2. Macrophage differentiation was confirmed by examination under microscope.
The human HL-60 myeloblastic leukemia cell line (ATCC, Manassas, VA, USA) was cultured in RPMI-1640 medium supplemented with 20% FBS and 1% antibiotic-antimycotic at 37°C and 5% CO2. Cells were passaged every 2–3 days to maintain a concentration between 1 × 105 and 1 × 106 cells/mL. Differentiation was initiated by seeding in T-25 tissue culture flasks at a concentration of 1 × 105 cells/mL in RPMI-1640 medium supplemented with 20% FBS, 1.25% dimethyl sulfoxide (Fisher BioReagents, PA, USA), and 1% antibiotic-antimycotic. Medium was replaced every 2 days and cell concentration was maintained below 1 × 106 cells/mL. Differentiation was complete 9 days after initiation and confirmed microscopically.
H. influenzae Strains and Culture Conditions
H. influenzae serotype a (Hia) strains 11-139, 14-61, 13-0074, and 13-240 were provided by Dr. Raymond Tsang (National Microbiology Laboratory, Winnipeg, MB, Canada). Noninvasive NTHi 375, isolated from the middle ear of a child [24], and invasive NTHi 08-254 isolated from blood of a pediatric patient (Dr. Tsang’s collection) were also used (Table 1). Bacteria were grown on brain heart infusion (BHI) agar plates containing 10 μg/mL hemin chloride and 5 μg/mL nicotine adenine dinucleotide for 16 h at 37°C and 5% CO2. Isolated colonies were transferred to BHI broth containing growth factors and grown to mid-log phase at 37°C with shaking. Prior to their use for stimulation of differentiated THP-1 and HL-60 cells, bacterial suspensions were diluted with BHI broth containing growth factors to reach an OD600 of 0.1.
Characteristics of H. influenzae strains
Serotype . | Label . | ST1 . | Encapsulation . | Invasiveness . | Source of isolation . |
---|---|---|---|---|---|
A | 11-139 | 23 | yes | yes | From blood of an adult patient [25] |
A | 14-61 | 23 | yes | no | From the middle ear of a pediatric patient [25] |
A | 13-0074 | 23 | no | yes | From blood of a pediatric patient (R. Tsang, personal communication) |
A | 13-240 | 4 | yes | yes | From blood of a pediatric patient [25] |
NTHi2 | 375 | 3 | no | no | From the middle ear of a pediatric patient [24] |
NTHi2 | 08-254 | 599 | no | yes | From blood of a pediatric patient (R. Tsang, personal communication) |
Serotype . | Label . | ST1 . | Encapsulation . | Invasiveness . | Source of isolation . |
---|---|---|---|---|---|
A | 11-139 | 23 | yes | yes | From blood of an adult patient [25] |
A | 14-61 | 23 | yes | no | From the middle ear of a pediatric patient [25] |
A | 13-0074 | 23 | no | yes | From blood of a pediatric patient (R. Tsang, personal communication) |
A | 13-240 | 4 | yes | yes | From blood of a pediatric patient [25] |
NTHi2 | 375 | 3 | no | no | From the middle ear of a pediatric patient [24] |
NTHi2 | 08-254 | 599 | no | yes | From blood of a pediatric patient (R. Tsang, personal communication) |
1Determined by the multilocus sequence typing [25].
2Non-typeable H. influenzae.
Stimulation with H. influenzae
Immediately prior to stimulation, the culture medium in the wells containing differentiated THP-1 macrophages was replaced with 2 mL fresh RPMI-1640 medium supplemented with 10% FBS. Bacterial suspensions with OD600 of 0.1, washed twice and resuspended in sterile PBS, were used to achieve a multiplicity of infection (MOI) of 0.1, 1, 10, or 100. Cell number and viability were determined with a Bright-Line Hemacytometer (Hausser Scientific, Horsham, PA, USA) using a 1:1 dilution factor with 0.4% Trypan blue solution (Sigma-Aldrich, St. Louis, MO, USA). THP-1 cells were stimulated with Hia strains at indicated MOI, using sterile PBS for the negative control, and Escherichia coli lipopolysaccharide (LPS) (Invitrogen, Carlsbad, CA, USA) for the positive control. In some experiments, cells were stimulated with Hia lipooligosaccharide (LOS) isolated from Hia 11-139 (provided by Dr. Andrew Cox, The National Research Council, Ottawa, ON, Canada [25]). The quantity of LOS corresponding to the number of bacterial cells used for stimulation was calculated as previously described [26]. The plates were incubated at 37°C and 5% CO2 for 1 h, then 100 μg/mL gentamycin (United States Biological, Salem, MA, USA) was added, and incubation continued for an additional 17 h at 37°C and 5% CO2.
To determine the role of LOS in cellular responses to Hia, THP-1 cells were treated with TLR4 inhibitors using two methods: (1) the treatment with 1 μg/mL LPS from Rhodobacter sphaeroides (LPS-RS) (InvivoGen) immediately prior to stimulation with bacteria, with an equivalent volume of sterile PBS used as control; (2) pretreatment with 5 or 10 μg/mL anti-TLR4 mouse IgG (InvivoGen) 1 h prior to stimulation with bacteria, using mouse control IgG1 (InvivoGen) as an isotype control.
Following stimulation of THP-1 cells, supernatants were collected for use in ELISA and stored at −80°C. Cells were collected through cold shock and manual scraping and immediately used for immunostaining and flow cytometry analysis.
For HL-60 cell stimulation, differentiated cells were plated in 6-well plates at a concentration of 2.5 × 106 cells/mL in RPMI-1640 medium supplemented with 20% FBS. The cells were stimulated with Hia at MOI of 1, 10, or 100 as described above, using PBS as negative control and E. coli LPS as positive control. Plates were incubated at 37°C and 5% CO2 for 1 h, then 100 μg/mL gentamycin was added, and plates were incubated for an additional 71 h at 37°C and 5% CO2. Following incubation, both adherent and nonadherent cells were collected for immediate use in flow cytometry analysis.
Enzyme-Linked Immunosorbent Assays
Enzyme-linked immunosorbent assays (ELISA) were completed to quantify TNF-α and IL-1β in supernatants. Human TNF-α and human IL-1β-uncoated ELISA kits were used according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA, USA). Primary antihuman TNF-α or IL-1β and secondary (biotin-conjugated antihuman TNF-α or IL-1β) antibodies, standards, the enzyme (streptavidin-HRP concentrate), and substrate (tetramethylbenzidine) were provided in the ELISA kits. Supernatants from stimulation experiments were diluted 40× in the diluent provided in the ELISA kits, to achieve a range within the limits of the assay. Samples were run in duplicate on the ELISA plates and the means were used for analysis. Plates were read at 450 nm and 570 nm for wavelength subtraction.
Flow Cytometry Analysis
Collected cells were centrifuged and resuspended in sterile PBS supplemented with 10% FBS for immunostaining with 1 μg/mL phycoerythrin-conjugated mouse-antihuman ICAM-1 (intercellular adhesion molecule 1) (BD Biosciences, Mississauga, ON, Canada) or CD64 (Invitrogen) for 1 h at 4°C in the dark, then washed twice, resuspended with sterile PBS, and placed on ice. Immediately prior to flow cytometry analysis, cells were stained with 1 μg/mL propidium iodide (PI) (BD Biosciences) to test for cell death. Flow cytometry was completed on the SONY SA3800 spectral cell analyzer with SA3800 Software (SONY Corporation, CA, USA). The desired population was gated based on light scattering properties; 10,000 gated events were collected. Gates for PI were created to obtain the % PI-negative cells and acquire the mean fluorescence intensity (MFI) for the ICAM-1 or CD64-positive population. Phycoerythrin-conjugated isotype control (Invitrogen) was used to exclude nonspecific antibody binding from analysis.
Statistical Analysis
Data were expressed as a mean of at least 3 independent experiments. Statistical difference was determined with a one-way analysis of variance (ANOVA) with a Tukey post hoc test or Student’s t test for paired comparisons. A p value of <0.05 was reported as statistically significant. Statistical analysis was performed using Graph-Pad Prism 9 (GraphPad Prism Software Inc., San Diego, CA, USA).
Results
Stimulation of differentiated THP-1 cells with genetically and phenotypically diverse clinical Hia isolates resulted in a similar increase in cell surface expression of ICAM-1 and in TNF-α and IL-1β release. For the optimization experiment, we used Hia 11-139 (ST23), an encapsulated invasive clinical isolate, which was utilized for the production of the capsular polysaccharide in the development of a new Hia polysaccharide-protein conjugate vaccine [27]. Stimulation of differentiated THP-1 cells with Hia 11-139 for 18 h (the first hour without and 17 h with antibiotics) resulted in a dose-dependent increase in the surface expression of ICAM-1 (Fig. 1a, b). The stimulation/PBS control ratio varied from 7.9 ± 1.3 (MOI 0.1) to 12.3 ± 1.9 (MOI 10), p = 0.02 (online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000540582). The highest ICAM-1 expression was found on cells stimulated at MOI 10, and this MOI was used in the following experiments for comparison between different H. influenzae strains. ICAM-1 expression at MOI 10 was significantly higher compared to the LPS control (p = 0.04). In these experiments, cellular viability, determined by the Trypan blue exclusion assay, varied between 91% and 99% (data not shown).
Cell surface expression of ICAM-1 on differentiated THP-1 cells stimulated with Hia clinical isolate Hia 11-139 at MOI 0.1, 1, 10, 100, or with E. coli LPS at 100 ng/mL for 18 h (flow cytometry analysis). a ****p < 0.0001, statistical significance between MOI of 0.1, 1, 10, 100, or LPS and control (PBS) (n = 4 independent experiments). b Histograms showing ICAM-1-positive events in the PI-negative gate, with MFI shown (results of one representative experiment). PE, phycoerythrin; MFI, mean fluorescence intensity.
Cell surface expression of ICAM-1 on differentiated THP-1 cells stimulated with Hia clinical isolate Hia 11-139 at MOI 0.1, 1, 10, 100, or with E. coli LPS at 100 ng/mL for 18 h (flow cytometry analysis). a ****p < 0.0001, statistical significance between MOI of 0.1, 1, 10, 100, or LPS and control (PBS) (n = 4 independent experiments). b Histograms showing ICAM-1-positive events in the PI-negative gate, with MFI shown (results of one representative experiment). PE, phycoerythrin; MFI, mean fluorescence intensity.
Stimulation of differentiated THP-1 cells with diverse Hia strains, including encapsulated (11-139, 14-61, 13-240) and nonencapsulated mutants (13-0074), invasive (11-139, 13-0074, 13-240) or noninvasive (14-61) isolates at MOI 10, resulted in a similar significant increase in ICAM-1 cell surface expression (Fig. 2). The effect of two nonencapsulated NTHi strains (noninvasive 375 and invasive 08-254) was comparable to Hia. No difference in cellular viability following stimulation with different bacterial strains was noticed (data not shown). Under the same conditions, stimulation of cells with 106 ng of Hia 11-139 LOS (48 ng/mL) that was equivalent to the amount of LOS present in 5 × 106 bacterial cells (MOI 10) induced slightly lower ICAM-1 expression in comparison to the effect of Hia 11-139, p > 0.05 (Fig. 2).
Cell surface expression of ICAM-1 on differentiated THP-1 cells stimulated with Hia clinical isolates 11-139 (n = 5), 14-61 (n = 6), 13-0074 (n = 7), 13-240 (n = 7), non-typeable (NTHi) 375 (n = 3) and 08-254 (n = 3) at MOI 10, Hia LOS at 48 ng/mL (n = 2), or E. coli LPS at 100 ng/mL (n = 12) for 18 h as described in the Methods section (flow cytometry analysis). ****p < 0.0001, statistical significance between treatments and PBS; ##p < 0.01, between bacterial stimulation and LPS. MOI, multiplicity of infection; MFI, mean fluorescence intensity ± SD.
Cell surface expression of ICAM-1 on differentiated THP-1 cells stimulated with Hia clinical isolates 11-139 (n = 5), 14-61 (n = 6), 13-0074 (n = 7), 13-240 (n = 7), non-typeable (NTHi) 375 (n = 3) and 08-254 (n = 3) at MOI 10, Hia LOS at 48 ng/mL (n = 2), or E. coli LPS at 100 ng/mL (n = 12) for 18 h as described in the Methods section (flow cytometry analysis). ****p < 0.0001, statistical significance between treatments and PBS; ##p < 0.01, between bacterial stimulation and LPS. MOI, multiplicity of infection; MFI, mean fluorescence intensity ± SD.
Under the same conditions of stimulation, we observed a considerable release of pro-inflammatory cytokines TNF-α and IL-1β into cell culture supernatants, with no significant differences observed between different Hia and NTHi strains (Fig. 3a, b). In comparison to the effect of LPS, all the bacterial strains induced a significantly higher release of TNF-α, whereas the concentration of IL-1β was higher in supernatants from stimulation with Hia 14-61, 13-0074, and 13-240, but was not noticeably different from stimulation with Hia 11-139 or either NTHi strain (Fig. 3a, b).
Cytokine release from differentiated THP-1 cells stimulated with Hia clinical isolates 11-139, 14-61, 13-0074, 13-240, non-typeable (NTHi) 375 and 08-254 at MOI 10, or E. coli LPS at 100 ng/mL for 18 h as described in the Methods section. Concentrations of TNF-α (a) and IL-1β (b) were measured by ELISA (pg/mL). Data represent mean cytokine concentrations ± SEM (n = 7–11 independent experiments). *p < 0.05, ****p < 0.0001, statistical significance between treatments and PBS; #p < 0.5, ###p < 0.001, ####p < 0.0001, between bacterial stimulation and LPS.
Cytokine release from differentiated THP-1 cells stimulated with Hia clinical isolates 11-139, 14-61, 13-0074, 13-240, non-typeable (NTHi) 375 and 08-254 at MOI 10, or E. coli LPS at 100 ng/mL for 18 h as described in the Methods section. Concentrations of TNF-α (a) and IL-1β (b) were measured by ELISA (pg/mL). Data represent mean cytokine concentrations ± SEM (n = 7–11 independent experiments). *p < 0.05, ****p < 0.0001, statistical significance between treatments and PBS; #p < 0.5, ###p < 0.001, ####p < 0.0001, between bacterial stimulation and LPS.
Immunostimulatory Effect of Hia Partially Depended on LOS
Stimulation of differentiated THP-1 cells with different concentrations of purified Hia 11-139 LOS or E. coli LPS demonstrated the ability of both compounds to increase the expression of ICAM-1 (Fig. 4a, b) as well as TNF-α and IL-1β release (online suppl. Fig. 1) in a dose-dependent manner. To assess the impact of LOS-dependent signaling on pro-inflammatory responses to H. influenzae, we treated THP-1 cells with specific TLR4 inhibitors and quantified the release of TNF-α and IL-1β following stimulation with Hia 11-139. Using mouse monoclonal antibody specific to TLR4, we observed its dose-dependent inhibitory effect on the release of TNF-α induced by Hia 11-139 and Hia 11-139 LOS. Both 5 and 10 μg/mL anti-TLR4 concentrations had a significant inhibitory effect on TNF-α induced by E. coli LPS, without any impact of the isotype control antibody (Fig. 5a). Under the same conditions, no significant effect of anti-TLR4 on IL-1β release by Hia-stimulated cells was noted, while inhibition of TLR4 decreased the release of IL-1β induced by both LOS and LPS although in a smaller degree as compared to TNF-α (Fig. 5b).
Cell surface expression of ICAM-1 on differentiated THP-1 cells stimulated with Hia LOS at 1 ng/mL, 10 ng/mL, or 100 ng/mL (n = 3) (a) or E. coli LPS at 1 ng/mL, 10 ng/mL, or 100 ng/mL (n = 4) (b) for 18 h as described in the Methods section (flow cytometry analysis). **p < 0.01, ****p < 0.0001, statistical significance from 1 ng/mL treatment. MFI, mean fluorescence intensity ± SD.
Cell surface expression of ICAM-1 on differentiated THP-1 cells stimulated with Hia LOS at 1 ng/mL, 10 ng/mL, or 100 ng/mL (n = 3) (a) or E. coli LPS at 1 ng/mL, 10 ng/mL, or 100 ng/mL (n = 4) (b) for 18 h as described in the Methods section (flow cytometry analysis). **p < 0.01, ****p < 0.0001, statistical significance from 1 ng/mL treatment. MFI, mean fluorescence intensity ± SD.
Cytokine release from differentiated THP-1 cells stimulated with Hia clinical isolate 11-139 at MOI 1, E. coli LPS, or Hia 11-139 LOS at 10 ng/mL for 18 h following inhibition of TLR4 as described in the Methods section. Concentrations of TNF-α (a, c) and IL-1β (b, d) were measured by ELISA (pg/mL). Data represent mean cytokine concentrations ± SEM (n = 3–13 independent experiments). Statistical significance was determined using one-way ANOVA with a Tukey post hoc test. a, b THP-1 cells were pretreated with 5 (LB) or 10 μg/mL (HB) mouse anti-TLR4 IgG 1 h prior to stimulation; IgG Control, mouse control IgG1 at a concentration of 10 μg/mL. *p < 0.05, **p < 0.01, ***p < 0.001, ****p ≤ 0.0001 statistical significance between treatments and IgG control. c, d THP-1 cells were treated with 1 μg/mL LPS from R. sphaeroides (LPS-RS) immediately prior to stimulation, with an equivalent volume of sterile PBS used as control. *p < 0.05, ***p < 0.001, ****p ≤ 0.0001, statistical significance between LPS-RS treatment and control (PBS).
Cytokine release from differentiated THP-1 cells stimulated with Hia clinical isolate 11-139 at MOI 1, E. coli LPS, or Hia 11-139 LOS at 10 ng/mL for 18 h following inhibition of TLR4 as described in the Methods section. Concentrations of TNF-α (a, c) and IL-1β (b, d) were measured by ELISA (pg/mL). Data represent mean cytokine concentrations ± SEM (n = 3–13 independent experiments). Statistical significance was determined using one-way ANOVA with a Tukey post hoc test. a, b THP-1 cells were pretreated with 5 (LB) or 10 μg/mL (HB) mouse anti-TLR4 IgG 1 h prior to stimulation; IgG Control, mouse control IgG1 at a concentration of 10 μg/mL. *p < 0.05, **p < 0.01, ***p < 0.001, ****p ≤ 0.0001 statistical significance between treatments and IgG control. c, d THP-1 cells were treated with 1 μg/mL LPS from R. sphaeroides (LPS-RS) immediately prior to stimulation, with an equivalent volume of sterile PBS used as control. *p < 0.05, ***p < 0.001, ****p ≤ 0.0001, statistical significance between LPS-RS treatment and control (PBS).
In addition, treatment of THP-1 cells with LPS from the photosynthetic bacterium R. sphaeroides (LPS-RS), with reported abilities to act as TLR4 antagonist [28], resulted in a similar effect, i.e., decreasing TNF-α but not IL-1β release by Hia-stimulated cells (Fig. 5c, d). Interestingly, although LPS-RS completely inhibited both TNF-α and IL-1β induced by E. coli LPS stimulation, its inhibitory effect on TNF-α induced by LOS was much smaller (Fig. 5c). Moreover, LPS-RS increased the amount of IL-1β released by cells stimulated with LOS (Fig. 5d), suggesting that LPS-RS may interact with LOS via yet unidentified mechanisms.
Considering the significance of neutrophils as the major type of innate immune cells, in addition to the analysis of macrophage responses, we studied the expression of pro-inflammatory markers by differentiated neutrophil-like HL-60 cells stimulated with H. influenzae. Besides ICAM-1, we analyzed the surface expression of CD64, the high-affinity receptor for the Fc portion of IgG (FcγRI), following the stimulation with Hia 11-139. The MFI of ICAM-1 and CD64 significantly increased in a dose dependent manner from MOI 1 to MOI 100, with a higher expression level of ICAM-1 as compared to CD64 (Fig. 6a, b).
Cell surface expression of ICAM-1 (a) and CD64 (b) on differentiated HL-60 cells stimulated with Hia clinical isolate Hia 11-139 at MOI 1, 10, or 100 for 72 h (flow cytometry analysis). ***p < 0.001, **p < 0.01 compared to all other treatments (one-way ANOVA with a Tukey post hoc test). MFI, mean fluorescence intensity (n ≥ 3 independent experiments).
Cell surface expression of ICAM-1 (a) and CD64 (b) on differentiated HL-60 cells stimulated with Hia clinical isolate Hia 11-139 at MOI 1, 10, or 100 for 72 h (flow cytometry analysis). ***p < 0.001, **p < 0.01 compared to all other treatments (one-way ANOVA with a Tukey post hoc test). MFI, mean fluorescence intensity (n ≥ 3 independent experiments).
When the expression of ICAM-1 was compared between THP-1 and HL-60 cells, the increases in MFI with increasing MOI followed the same pattern, with highest MFI detected at MOI 100. However, the MFI values of ICAM-1 on HL-60 cells were about one order of magnitude lower than the corresponding values on THP-1 cells (Fig. 1, 6a, Fig. 7a, b). For example, stimulation of HL-60 cells with Hia 11-139 induced a 2.5- and 4.3-fold increase in ICAM-1 MFI over unstimulated cells at MOI 1 and 100, correspondingly (online suppl. Table 2). Under the same conditions of bacterial stimulation, ICAM-1 expression on THP-1 cells increased by 10.2- and 11.7-fold (online suppl. Table 1).
a ICAM-1 ell surface expression on differentiated HL-60 cells stimulated with Hia clinical isolates 11-139, 14-61, 13-0074, 13-240 (n = 3) and non-typeable H. influenzae (NTHi) 375 (n = 4) at MOI 1 and 100, E. coli LPS at 100 ng/mL (n = 20) and 1 μg/mL (n = 3) for 72 h. Flow cytometry analysis, ratio of stimulation (MFI)/PBS control (MFI) ± SD. Statistical significance between MOI 1 and MOI 100 for each isolate, and between E. coli LPS at 100 ng/mL and 1 μg/mL displayed (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001); one-way ANOVA with a Tukey post hoc test. b Histograms showing ICAM-1 expression on differentiated HL-60 cells stimulated with Hia 11-139 or LPS; flow cytometry. MFI, mean fluorescence intensity indicated (results of one representative experiment); PE, phycoerythrin.
a ICAM-1 ell surface expression on differentiated HL-60 cells stimulated with Hia clinical isolates 11-139, 14-61, 13-0074, 13-240 (n = 3) and non-typeable H. influenzae (NTHi) 375 (n = 4) at MOI 1 and 100, E. coli LPS at 100 ng/mL (n = 20) and 1 μg/mL (n = 3) for 72 h. Flow cytometry analysis, ratio of stimulation (MFI)/PBS control (MFI) ± SD. Statistical significance between MOI 1 and MOI 100 for each isolate, and between E. coli LPS at 100 ng/mL and 1 μg/mL displayed (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001); one-way ANOVA with a Tukey post hoc test. b Histograms showing ICAM-1 expression on differentiated HL-60 cells stimulated with Hia 11-139 or LPS; flow cytometry. MFI, mean fluorescence intensity indicated (results of one representative experiment); PE, phycoerythrin.
No significant difference in ICAM-1 expression on HL-60 cells stimulated by different isolates was noted at MOI 1. At MOI 100, NTHi 375 showed the highest stimulatory effect compared to each other isolate (p < 0.0001); Hia 13-240 had a higher immunostimulatory effect than Hia 11-139 (p < 0.05), Hia 14-61 (p < 0.05), and Hia 13-0074 (p < 0.01) (Fig. 7a).
Discussion
In this study, we have used two human cell lines, THP-1 cells differentiated to macrophages and HL-60 neutrophil-like cells, in an in vitro infection model to determine whether innate immune responses to clinical Hia isolates depend on their characteristics. We selected 4 genetically and phenotypically diverse Hia strains isolated from patients with different forms of clinical infection. These strains have been thoroughly characterized with regard to their molecular-genetic features, using whole-genome sequencing; their sequence types (STs) and phylogenetic relations based on amplification and sequencing of 7 enzyme genes (adk, atpG, frdB, fucK, mdh, recA) and detection of the superoxide dismutase gene (sodC) have been established [29]. In addition, the prevalence of various Hia STs in different geographic regions and populations as well as their association with invasive disease has been described [30]. We were interested in comparing cellular responses of encapsulated Hia bacteria to an Hia mutant which was able to cause an invasive disease despite the lack of the polysaccharide capsule, the major virulence factor of encapsulated H. influenzae [31]. In addition, we tested cellular responses to two nonencapsulated NTHi strains with different capabilities of causing invasive disease.
As a readout of the cellular responses, we have studied the cell surface expression of ICAM-1 and the release of two pro-inflammatory cytokines, TNF-α and IL-1β, assessed post one-hour long stimulation of innate immune cells with live bacteria followed by additional incubation in the presence of an antibiotic to kill extracellular bacteria. ICAM-1 is a transmembrane glycoprotein, member of the immunoglobulin superfamily, which mediates leukocyte-endothelial cells and leukocyte-leukocyte interactions crucial in the process of cellular trafficking during inflammatory responses. In addition, ICAM-1 expressed on antigen-presenting cells interacts with LFA-1 on T cells in the process of immunological synapse formation providing essential co-stimulatory signals for T-cell activation and hence represents a critical molecule involved in both innate and adaptive immune responses [32]. Previous studies identified ICAM-1 as a receptor for H. influenzae on airway epithelial cells mediating bacterial adhesion through the outer membrane protein P5 and the type IV pilus of NTHi [33, 34]. However, to the best of our knowledge, no studies assessed the role of ICAM-1 in interactions of encapsulated H. influenzae with leukocytes or identified any bacterial structures involved in this process. In the context of bacterial infections, most published studies addressed ICAM-1 expression on epithelial or endothelial cells, yet it was recently demonstrated that LPS stimulation induced ICAM-1 expression in macrophages that resulted in an enhanced phagocytosis [35]. When we tested whether stimulation of macrophage and neutrophil-like cell lines with Hia made an impact on ICAM-1 expression, we found a dose-dependent upregulation of this molecule on both cell lines. Although an increase in ICAM-1 expression on neutrophils was much smaller than on macrophages, the data suggested that both cell types became activated that could result in their enhanced functions including phagocytosis and antigen-presenting capacities. In addition, H. influenzae infection induced a significant release of TNF-α and IL-1β by THP-1 cells, further supporting their activation. The production of both cytokines and ICAM-1 is regulated by the transcription factor NF-κB, as a result of the pro-inflammatory signaling initiated by the recognition of pathogen-associated molecular patterns (PAMPs) by PRRs [36]. However, no differences in ICAM-1 expression or cytokine release induced by encapsulated versus nonencapsulated H. influenzae or invasive versus noninvasive clinical isolates were noted: all the tested strains produced similar responses. Intriguingly, in HL-60 cells, Hia 13-240 (ST4) exhibited a higher immunostimulatory effect than any ST23 Hia strains (Fig. 6a). The ST4 belongs to a genetically distinct H. influenzae clonal complex as compared to ST23 [29], and ST4 isolates have been associated with enhanced virulence and severe inflammation [15]. Although our observations call for further mechanistic studies, one may speculate that an excessive neutrophil activation caused by ST4 can contribute to the severity of inflammatory responses in invasive Hia disease [37]. To further study pro-inflammatory responses of neutrophil-like cells, we quantified the surface expression of the high-affinity receptor for the Fc portion of IgG (FcγRI) following infection of HL-60 cells with a representative invasive strain, Hia 11-139. At the higher bacterial load (MOI 100), the FcγRI expression level significantly rose over the baseline, indicating an increase in cellular capacities of binding IgG. In the real-life infectious process, the engagement of FcγRI via binding Hia-specific IgG antibodies would initiate immunoreceptor tyrosine-based activation motif-mediated intracellular signaling that would in turn result in enhanced neutrophil functional abilities, i.e., phagocytosis, production of superoxide and pro-inflammatory cytokines, all essential factors in host defense against H. influenzae infection [38, 39].
The role of the polysaccharide capsule as the major virulence factor in infections caused by encapsulated bacteria has been well established by earlier studies. It has been demonstrated that the capsule protects Hib against phagocytosis and the bacteria can only be effectively destroyed by phagocytes in the presence of opsonizing antibodies specific to the capsular antigens [31]. Whether H. influenzae capsule can activate innate immune responses or be recognized by PRRs remains uncertain, although the immunostimulatory effect of the capsule of another Gram-negative bacterium, Klebsiella pneumoniae, has been reported [40]. In this study, using 3 encapsulated Hia strains in comparison to a nonencapsulated Hia mutant, as well as 2 nonencapsulated NTHi, we found that the pro-inflammatory effect of H. influenzae did not depend on the presence or absence of polysaccharide capsule. Moreover, the development of invasive versus noninvasive clinical disease did not appear to directly depend on the presence of the capsule or the degree of the activation of pro-inflammatory responses. As further discussed below, our results indicated that LOS, rather than the polysaccharide capsule, had the major role in the activation of pro-inflammatory responses to H. influenzae infection.
Although NTHi less commonly causes invasive disease than encapsulated H. influenzae, and hence it is considered to be less virulent, recent studies suggest that the invasive capabilities may depend on some factors unrelated to the presence of the capsule. Indeed, encapsulated H. influenzae are commonly isolated from asymptomatic carriers, as well as from individuals with noninvasive infections; for example, in this study, the source of encapsulated Hia 14-61 isolation was the middle ear. We identified Hia as a cause of noninvasive disease, such as otitis media, in Northwestern Ontario, Canada [11]. In the same geographic region, Hia was isolated from the nasopharynx in 9% of healthy children [12]. High prevalence of Hia carriage among healthy individuals has been reported by other researchers; for example, in Alaska, 43% of individuals having contact with cases of Hia invasive disease were asymptomatically colonized by Hia [41]. Although biological characteristics associated with the ability of encapsulated H. influenzae to cause invasive rather than noninvasive disease have not been determined, in the case of NTHi, recent findings shed light on some potential mechanisms. Invasive NTHi isolates were found to be more resistant to complement-mediated killing than noninvasive NTHi; the resistance was associated with the capacity of bacteria to evade binding of IgM [42]. In addition, particular structural LOS characteristics due to phase-variation of certain LOS biosynthetic genes were found to be associated with abilities of NTHi to cause invasive versus noninvasive disease [43].
LOS is the major virulence factor of both encapsulated and nonencapsulated H. influenzae. Unlike LPS typical for most Gram-negative bacteria, LOS lacks the repeating O-antigen, but includes the tri-heptose oligosaccharide backbone covalently attached to a 3-deoxy-d-manno-oct-2-ulosonic acid moiety, that is the core region [44]. The core region is covalently linked to the lipid A, a TLR4 ligand, which has powerful abilities to activate innate immune and inflammatory responses [45]. In our experiments, exposure of THP-1 cells to a neutralizing TLR4 antibody [46] or TLR4 antagonist LPS-RS [28] resulted in a decrease in TNF-α production induced by Hia 11-139 or LOS isolated from Hia 11-139. These findings showed the significance of TLR4-mediated signaling in macrophage pro-inflammatory responses to H. influenzae. However, under the same experimental conditions, inhibition of TLR4 did not affect the release of IL-1β by cells stimulated with Hia. In contrast to the regulation of TNF-α expression, which is directly mediated by the TLR4 downstream signaling followed by the activation of NF-κB [47], IL-1β production depends on a 2-signal mechanism [48]. Whereas the primary signal, largely mediated by the NF-κB activation, leads to the expression of pro-IL-1β, the secondary signal is required for the activation of the inflammasome complex, being an essential factor for mature IL-1β production [49]. Therefore, the activation of LOS-induced pro-inflammatory signaling in the process of H. influenzae infection would be insufficient for IL-1β release. Yet, the ability of H. influenzae to activate the NLRP3 inflammasome has been demonstrated in an in vitro model of NTHi infection [50]. Indeed, H. influenzae express a plethora of immunostimulatory molecules recognized by innate immune receptors including those implicated in the inflammasome formation [51]. Considering that inhibition of TLR4 did not result in complete abrogation of TNF-α release following Hia stimulation, the role of H. influenzae immunostimulatory molecules beyond LOS could be considerable. Immunostimulatory effect of H. influenzae could be mediated by the recognition of peptidoglycan, porin, and DNA, by NOD1, TLR2, and TLR9, correspondingly [22, 52, 53]. In our infection model, we used stimulation with live H. influenzae followed by incubation with killed bacteria likely providing a sufficient RNA signal, which belongs to the special class of viability-associated PAMPs as described by Sander et al. [54].
Limitations
The major limitation in this study is the use of an in vitro model of bacterial infection. Such a model inevitably underestimates complex interactions among various innate immune cells and between innate and adaptive immune mechanisms. To accurately assess the mechanisms of host-pathogen interactions, including the role of different virulence factors in the development of the infectious process, it would be beneficial to use in vivo experimental models although any animal model would not necessarily reproduce the complexity of human immune responses. As using cell lines does not completely represent the response of circulating leukocytes, further studies on primary human monocytes and neutrophils would help to clarify the role of innate immune cell activation in the pathogenesis of invasive H. influenzae disease. While we selected several clinical isolates on the base of their distinct features, they do not represent the whole spectrum of H. influenzae strains, which cause various forms of clinical disease, especially in the case of NTHi, characterized by remarkable genetic heterogeneity.
Conclusion
We have established that in vitro infection of human innate immune cells with Hia leads to the powerful activation of pro-inflammatory responses that may potentially contribute to the pathogenesis of invasive Hia disease. However, presence or absence of the polysaccharide capsule was not associated with the strength of the pro-inflammatory responses. Likewise, abilities to cause invasive or noninvasive disease were not directly related to the immunostimulatory capacities of H. influenzae strains we used in this study. These findings suggest that the development of invasive versus noninvasive clinical disease may depend on the abilities of the adaptive immune system to control rapidly replicating bacteria via the action of antibodies specific to capsular and noncapsular antigens that will opsonize H. influenzae for phagocytosis. It appears that the functional abilities of phagocytic cells can be activated by PAMPs of H. influenzae regardless of the presence of the capsule, with LOS being the major factor responsible for the innate immune activation.
Acknowledgments
We thank Dr. Raymond Tsang (National Microbiology Laboratory, Winnipeg, MB, USA) for his kind donation of H. influenzae clinical isolates and Dr. Andrew Cox (National Research Council, Ottawa, ON, USA) for donation of Hia 11-139 LOS.
Statement of Ethics
The work was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. This study is exempt from ethical committee approval as all work was performed in vitro using cell lines and no primary human samples were used. According to the Tri-Council Policy Statement “Ethical Conduct for Research Involving Humans” (Government of Canada, TCPS 2, 2018), the reuse of de-identified human somatic cell lines does not require REB review. According to the Tri-Council Policy Statement “Ethical Conduct for Research Involving Humans” (Government of Canada, TCPS 2, 2018), written informed consent is not required for research that relies exclusively on the reuse of de-identified human somatic cell lines where the researcher does not know or have access to the identity of the participant.
Conflict of Interest Statement
The authors declare no conflicts of interest.
Funding Sources
This work was supported by the National Science and Engineering Research Council (NSERC) Discovery Grant (Ref. 1464208), the Canadian Immunization Research Network (Grant CT20ON31), and NOSM University Faculty Association Research Development Grant (MU). The funders had no role in the design, data collection, data analysis, and reporting of this study.
Author Contributions
Marina Ulanova conceived and designed the study and wrote the manuscript. Brenda Huska, Courtney Ferris, and Zaid Shahid carried out the experiments and analyzed the data. All the authors reviewed the manuscript prior to submission.
Data Availability Statement
All data generated during this study are included in this article and its online supplementary material files available without any restrictions. Further inquiries can be directed to the corresponding author.