Bacteria and their cell wall components uniformly co-activate interleukin-17-producing thymocytes

Abstract

Interleukin (IL)-17-producing T cells play a critical role in the immune response against microbial pathogens. Traditionally, experimental studies have focused upon understanding the activity of IL-17-producing T cells which differentiate from naive T cells in the peripheral immune system. However, we have demonstrated previously that IL-17-producing T cells are also present in the thymus of naive wild-type mice and can be co-activated there by microbial stimuli. Other studies have supported the concept that IL-17-producing thymocytes have a specific role in the immediate defence against microbial pathogens, which is independent from the development of an adaptive immune response. Given an important role of the thymus in systemic bacterial infection and sepsis, in this study we investigate the effect of a broad spectrum of bacteria and cell wall components on thymocyte cytokine production. Surprisingly, we find that all types of bacteria investigated (including non-pathogenic species) uniformly activate IL-17-producing thymocytes upon α-CD3 stimulation. In contrast, there is a heterogeneous effect on IL-6 and interferon (IFN)-γ-production with Gram-negative bacteria inducing far higher frequencies of IL-6- and IFN-γ-producing thymocytes than Gram-positive bacteria. We conclude that IL-17-producing thymocytes constitute a ‘first line of recognition’, but not a ‘first line of defence’ against bacteria in general. Their activity might lead to immune activation, but not necessarily to a pathological inflammatory disease condition. The difference between these two states might be determined by other immunological effector molecules, such as IL-6 and IFN-γ.

Introduction

Interleukin (IL)-17 is a potent proinflammatory cytokine produced by activated memory T cells, called T helper type 17 (Th17) cells. Th17 cells have been demonstrated to be an effector T cell lineage on their own, characterized by specific differentiation and activation requirements 1,2. In addition to induced IL-17-producing T cells which are primed in the immune periphery during the induction of an antigen-specific T cell response, IL-17-producing T cells are also present in the thymus of naive wild-type mice 3. Their development is shaped by thymic dendritic cells 4, they underlie positive selection and have been termed naturally occurring IL-17-producing T cells 5,6. These cells have the ability to provide effector functions such as cytokine secretion without the requirement for T cell receptor (TCR)-induced differentiation, as it is required for induced T cells in the immune periphery. They share these properties with other lymphocyte subpopulations with innate-like properties 7, and seem to have an important role in immediate reactivity against microbial agents, e.g. upon systemic bacterial infection 3,6,8. In humans, it has been demonstrated that all IL-17-producing cells originate from CD161⁺ naive CD4⁺ T cells of umbilical cord blood, as well as of the postnatal thymus 9,10.

Systemic bacterial infection and bacterial sepsis profoundly influence the thymus, leading to strong proinflammatory cytokine production by thymocytes followed by subsequent apoptosis and thymic atrophy 11,12. Toll-like receptor (TLR)-2 is a receptor recognizing a broad spectrum of different bacterial components. TLRs in general play an important role in the innate immune response as well as in the polarization of an adaptive immune response. As well as whole microbial organisms, they detect conserved microbial immune stimulants such as bacterial cell wall components (e.g. peptidoglycan, lipopeptides, lipopolysaccharide) or viral and bacterial DNA and RNA sequence motifs 13.

Peptidoglycan (PGN), also known as murein, is a polymer of alternating N-acetylglucosamine and N-acetylmuramic acid and exists substantially more highly in Gram-positive than in Gram-negative bacteria 14. Lipoteichoic acids (LTA) are linked to the cytoplasmic membrane of most Gram-positive bacteria and vary between different species 15. Another important part in the immune response involves inflammasome complexes. Inflammasomes recognize microbial products or endogenous molecules through direct binding of ligands as well as indirect mechanisms 16. Previous investigations have revealed that activation of the NOD-like receptor P3 (NLRP3) inflammasome complex influence IL-17 production by enhanced secretion of mature IL-1β 17. It seems likely that by reacting to bacterial microorganisms, viruses, fungi and parasites the inflammasome initiates major intracellular defence mechanisms and a broad range of immune responses 18,19.

We have demonstrated previously that TLR ligands such as cytosine–phosphate–guanosine (CpG) (TLR-9), Imiquimod (TLR-7), flagellin (TLR-5) or poly I:C (TLR-3) can potently co-activate IL-17-producing thymocytes 3,20,21. We therefore considered it relevant to investigate the effect of bacteria themselves on thymocyte cytokine production. A systematic study on the impact of bacterial microorganisms or parts of their cell wall on the activation of IL-17-producing T cells in the thymus is not available so far. Herein we study the influence of a broad (and heterogeneous) spectrum of different bacteria, cell wall components and inflammasome inducers on the pro- and anti-inflammatory cytokine production in thymocytes with a focus on IL-17. In particular, we intend to investigate whether or not different species characteristics (see Table 1) have an influence on the cytokine pattern.

Table 1.

Characteristics of the different bacteria included in our experiments with regard to the categories Gram staining [Gram-positive (+) or Gram-negative (−)], cell respiration (aerobic or anaerobic), pathogenicity (with typical disease if pathogenic) and localization (intra- or extracellular)

Bacterium	Abbr.	Gram staining	Cell respiration	Pathogenicity/disease pattern	Localization
Escherichia coli 0111:B4	HKEB	−	Facultative anaerobic	Non-pathogenic	Extracellular
Helicobacter pylori	HKHP	−	Micro-aerophilic	Pathogenic (peptic ulcers and gastritis)	Extracellular
Listeria monocytogenes	HKLM	+	Facultative anaerobic	Pathogenic (listeriosis)	Intracellular
Legionella pneumophila	HKLP	−	Aerobic	Pathogenic (legionellosis)	Intracellular
Lactobacillus rhamnosus	HKLR	+	Facultative anaerobic	Non-pathogenic	Extracellular
Mycoplasma fermentans	HKMF	No cell wall	Facultative anaerobic	Potentially pathogenic (arthritis)	Intra- and extracellular
Pseudomonas aeruginosa	HKPA	−	Aerobic	Pathogenic (pneumonia, osteomyelitis)	Extracellular
Porphyromonas gingivalis	HKPG	−	Anaerobic	Pathogenic (periodontal disease)	Extracellular
Staphylococcus aureus	HKSA	+	Facultative anaerobic	Normally non-pathogenic	Extracellular
Streptococcus pneumoniae	HKSP	+	Aerotolerant anaerobic	Pathogenic (pneumonia)	Extracellular

Material and methods

Animals

Wild-type female C57.BL/6J mice at age 6–8 weeks were purchased from Centre d'Elevage R. Janvier (CERJ, Le Genest-St-Isle, France) and maintained at the local animal facilities under special pathogen-free conditions. All animal experiments were fully approved by the local authorities for animal experimentation.

Cell preparation

After killing the animals with isoflurane euthanasia, thymi were prepared. Subsequently, each thymus was squeezed using the back of a syringe plunger to obtain single-cell suspensions. To separate out cell clusters the obtained suspensions were filtered through a 70-μm cell strainer for enzyme-linked immunospot (ELISPOT) analysis or a 40-μm cell strainer for fluorescence activated cell sorter (FACS) analysis (BD Biosciences, Heidelberg, Germany). The cells were counted by trypan blue exclusion and plated together with the respective stimulants at the cell numbers indicated in serum-free HL-1 medium (Lonza, Cologne, Germany).

Reagents

The following heat-inactivated bacterial strains were utilized in this study: heat-killed Escherichia coli 0111:B4 (HKEB), Helicobacter pylori (HKHP), Listeria monocytogenes (HKLM), Legionella pneumophila (HKLP), Lactobacillus rhamnosus (HKLR), Mycoplasma fermentans (HKMF), Pseudomonas aeruginosa (HKPA), Porphyromonas gingivalis (HKPG), Staphylococcus aureus (HKSA) and Streptococcus pneumoniae (HKSP). Furthermore, lipoteichoic acids from Bacillus subtilis (LTA-BS) and S. aureus (LTA-SA) as well as peptidoglycans from E. coli 0111:B4 (PGN-EB), E. coli K12 (PGN-EK) and S. aureus (PGN-SA) were used. For NLRP3 inflammasome induction, alum crystals, hemozoin, monosodium urate crystals (MSU) and calcium pyrophosphate dehydrate (CPPD) crystals were utilized. All reagents were obtained from InvivoGen (San Diego, CA, USA). When indicated, thymocytes cultures were stimulated with anti (α)-CD3 (clone 145-2C11; BD Pharmingen, San Diego, CA, USA) at a concentration of 1 μg/ml. For cultures stimulated with HKMF, 10⁷ bacteria/ml were utilized. Bacterial concentrations were increased to 5 × 10⁷ bacteria/ml for HKHP and HKLP and 10⁸ bacteria/ml for HKEB, HKLM, HKLR, HKPA, HKPG, HKSA and HKSP. Peptidoglycans were used from 1–10 μg/ml and lipoteichoic acids from 0·1 to 10 μg/ml. For inflammasome induction we applied concentrations from 0·5 to 500 μg/ml.

Cytokine measurement by ELISPOT and computer-assisted ELISPOT image analysis

ELISPOT assays were essentially performed as described previously 3. Briefly, MultiScreen_HTS 96-well filtration plates (Millipore, Schwalbach, Germany) were coated overnight with capture antibodies in sterile phosphate-buffered saline (PBS). The following coating antibodies were used: IL-5 (TRFK5), IL-6 (MP5-20F3) and IL-17 (TC11-18H10) were used at 2 μg/ml and interferon (IFN)-γ (P46–A2), IL-2 (JES6-1A12) and IL-4 (11B11) at 4 μg/ml. Antibodies were ordered from BD Pharmingen. Plates were blocked with sterile PBS/bovine serum albumin (BSA) 0·5% and washed with sterile PBS. Thymocytes (10⁶ per well) were plated in HL-1 medium (BioWhittaker, Walkersville, MD, USA) containing 1% glutamine and 1% penicillin/streptomycin, each in duplicate cultures. Thereafter cells were stimulated with different stimulation reagents and incubated for 20 h at 37°C, 5% CO₂. Plates were washed with PBS before adding the detection antibodies (BD Pharmingen) overnight in PBS/BSA 0·5%. Antibodies against IFN-γ (XMG1·2), IL-2 (JES6-5H4), IL-4 (BVD6-24G2), IL-5 (TRFK4) and IL-6 (MP5-32C11) were used at 2 μg/ml, against IL-17 (TC11-8H4) at 0·5 μg/ml. After washing the plates, streptavidin-AP (BD Pharmingen) in PBS/BSA 0·5% (1 : 1000) was added before plates were visualized using the AP Conjugate Substrate Kit (Bio-Rad Laboratories, München, Germany).

ELISPOT image analysis

Image analysis of ELISPOT assays was performed with the ImmunoSpot™ Analysis Software after scanning the plates with an Immunospot™ Analyzer (Cellular Technologies, Cleveland, OH, USA). In brief, digitized images of individual wells of the ELISPOT plates were analysed for cytokine spots, based on the comparison of experimental wells (containing immune cells and stimuli) and control wells (immune cells, no stimuli). After separating spots that touched or partially overlapped, non-specific ‘background noise’ was gated out by applying spot size and circularity analysis as additional criteria. Then, spots that fell within the accepted criteria were highlighted and counted. Single wells which could not be enumerated because of confluence phenomena were assessed by using the highest numbers of cytokine-producing cells which could be counted regularly in other wells in the same assay as an approximated estimate.

FACS analysis of extra- and intracellular staining

Thymocytes were prepared and stimulated for 20 h with 10⁷/ml HKEB, HKPA, HKSA and HKLR, 1 μg/ml for LTA-BS and LTA-SA and 5 μg/ml for PGN-EB, PGN-EK and PGN-SA. For intracellular cytokine staining thymocytes were additionally activated with α-CD3. For the last 10 h of culture, BD GolgiStop™ was added. After blocking with BD Fc Block™ (clone 2·4G2) cells were stained with phycoerythrin (PE)-labelled anti-CD4 (clone GK1·5) and allophycocyanin (APC)-H7-labelled anti-CD8a (clone 53-6·7). Intracellular staining was then performed using BD Cytofix/Cytoperm™ plus Fixation/Permeabilization Kit and PE-cyanin 7 (Cy7)-labelled anti-IFN-γ (clone XMG1·2). All antibodies were purchased from BD Biosciences. FACS analysis was performed on a Millipore Guava EasyCyte™ 8 using guavaSoft™ software version 2·2.2.

IL-17 secretion assay

For analysis of rare antigen-specific IL-17-producing cells we used the mouse IL-17 secretion assay (Miltenyi Biotech, Bergisch Gladbach, Germany). Stimulated thymocytes were incubated with an IL-17-specific catch reagent for a short time to allow IL-17 secretion. After binding of the secreted IL-17 to the catch reagent, cells were subsequently labelled with IL-17 detection antibody conjugated to biotin, followed by anti-biotin-PE antibody. After magnetically labelling with anti-PE microbeads, IL-17-secreting cells were enriched by magnetic-activated cell sorting (MACS) separation. Enrichment of IL-17-secreting leucocytes using MACS technology increases the analysis sensitivity and enables further characterization. For FACS analysis we used fluorescein isothiocyanate (FITC)-labelled anti-CD4 (clone H129·19) and APC-H7-labelled anti-CD8 (clone 53-6·7). FACS analysis was performed on a Millipore Guava EasyCyte™ 8 using guavaSoft™ software version 2·2.2.

Statistical analysis

For statistical analysis, one-way analysis of variance (anova) with Dunnett's two-tailed t-test (Instat, GraphPad version 3·00) was performed. Differences at P < 0·05 were considered statistically significant; values of P < 0·01 were considered statistically highly significant.

Results

Only P. aeruginosa induces IFN-γ in thymocytes by itself in low frequencies, whereas IL-6 is triggered by different heat-killed bacteria

Thymocytes were stimulated with heat-killed bacteria and ELISPOT assay was performed as described in Material and methods. As seen in Fig. 1a, statistically significant higher frequencies of IFN-γ [mean: 8·18, standard deviation (s.d.): 7·51] could be measured in thymocytes for HKPA. The other bacteria only showed a negligible effect on IFN-γ production. As indicated in Fig. 1d, statistically highly significant frequencies of IL-6-producing cells induced by bacterial stimulation alone could be measured for HKLP (mean: 36·2, s.d.: 26·2) and HKPA (mean: 37·7, s.d.: 26·8). HKEB (mean: 18·7, s.d.: 11·7), HKHP (mean: 13·0, s.d.: 8·56), HKLM (mean: 14·5, s.d.: 8·26), HKPG (mean: 13·9, s.d.: 7·84) and HKSA (mean: 12·6, s.d.: 9·27) showed higher, although not significant, frequencies of IL-6-producing cells in comparison to medium control (mean: 1·25, s.d.: 1·69). However, as pointed out in Fig. 1b,c,e,f, stimulation with heat-killed bacteria only did not trigger IL-2, IL-4, IL-5 and IL-17 production in thymocytes.

Fig. 1.

Frequencies of interferon (IFN)-γ- (a), interleukin (IL)-2 (b), IL-4- (c), IL-6- (d), IL-5- (e) and IL-17- (f) producing cells after single stimulation with heat-killed bacteria as measured by enzyme-linked immunospot (ELISPOT) assay. Thymocytes were plated at a density of 10⁶ cells per well. Data of eight [six for heat-killed Helicobacter pylori (HKHP) and Legionella pneumophila (HKLP)] mice tested individually in three independent experiments are shown. Measurements were counted in duplicate wells. Each data point indicates one individual animal; line represents the mean.*P < 0·05; **P < 0·01 by Dunnett's t-test.

Stimulation with heat-killed bacteria influences the cytokine response of α-CD3-activated thymocytes

Thymocytes were stimulated with heat-killed bacteria and additionally activated with α-CD3. As indicated in Fig. 2a, statistically highly significant frequencies of IFN-γ-producing cells after α-CD3 co-activation could be detected for HKPA (mean: 372·3, s.d.: 186·2), HKEB (mean: 141·6, s.d.: 104·7) and HKLP (mean 173·6, s.d.: 101·7), statistically significant higher frequencies for HKPG (mean: 125·2, s.d.: 100·3) in comparison to single stimulation with α-CD3 (mean: 3·43, s.d.: 4·79). As depicted in Fig. 2b, upon combined stimulation with α-CD3, a statistically highly significant response of IL-2 could be measured for HKHP (mean: 41·7, s.d.: 13·9), HKLP (mean: 73·7, s.d.: 16·8), HKPA (mean: 45·7, s.d.: 19·1) and HKPG (mean: 47·1, s.d.: 15·8). As displayed in Fig. 2c, statistically highly significant IL-4 responses for HKPA (mean: 55·8, s.d.: 24·4) and HKPG (mean: 52·0, s.d.: 25·3) after co-activation with α-CD3. As shown in Fig. 2d, combined stimulation with 1 μg/ml α-CD3 led to a strong up-regulation of IL-6 secretion, which was qualitatively similar to the induction pattern seen with bacterial stimulation alone. Statistically highly significant frequencies could be detected for HKLP (mean: 227·8, s.d.: 73·1), HKPA (mean: 202·0, s.d.: 76·7) and HKHP (mean: 108·4, s.d.: 91·2). Higher, but statistically not significant frequencies could be determined for HKPG (mean: 92·3, s.d.: 71·9), HKEB (mean: 69·1, s.d.: 52·9), HKLM (mean: 39·1, s.d.: 26·3) and HKSA (mean: 44·3, s.d.: 30·9) in comparison to single stimulation with α-CD3 (mean: 2·43, SD: 2·45). As seen in Fig. 2f, combined stimulation with α-CD3 induced statistically highly significant frequencies of IL-17-producing cells for HKEB (mean: 96·9, s.d.: 30·7), HKHP (mean: 103·0, s.d.: 34·3), HKLM (mean: 80·3, s.d.: 23·8), HKLP (mean: 97·5, s.d.: 32·6), HKPA (mean: 99·0, s.d.: 25·2), HKPG (mean: 117·5, s.d.: 23·5), HKSA (mean: 98·3, s.d.: 33·1) and HKSP (mean: 62·3, s.d.:18·1). For HKMF (mean: 55·1, s.d.: 19·7) and HKLR (mean: 44·6, s.d.: 14·6), statistically significant higher frequencies of IL-17-producing cells could be detected in comparison to single stimulation with α-CD3 (mean: 15·4, s.d.: 13·5). For the cytokine IL-5, no higher response after co-activation with α-CD3 could be detected for any bacterium.

Fig. 2.

Frequencies of interferon (IFN)-γ- (a), interleukin (IL)-2 (b), IL-4- (c), IL-6- (d), IL-5- (e) and IL-17- (f) producing cells after stimulation with heat-killed bacteria and additional co-activation with α-CD3 as measured by enzyme-linked immunospot (ELISPOT) assay. Thymocytes were plated at a density of 10⁶ cells per well. Data of eight [six for heat-killed Helicobacter pylori (HKHP) and Legionella pneumophila (HKLP)] mice tested individually in three independent experiments are shown. Measurements were counted in duplicate wells. Each data point indicates one individual animal, line represents the mean.*P < 0·05; **P < 0·01 by Dunnett's t-test.

Only PGN-EB induces low frequencies of IL-17 in thymocytes; upon co-activation with α-CD3, both PGNs and LTAs induce distinctly higher frequencies of IL-17-producing cells

Thymocytes were stimulated with cell wall components of microorganisms. While LTA-SA and LTA-BS did not enhance the IL-17 production in thymocytes after single stimulation, co-activation with α-CD3 triggered the IL-17-producing cells. As seen in Fig. 3d, statistically highly significant frequencies of IL-17 could be detected for 0·1 μg/ml (mean: 146·7, s.d.: 16·3) and 1 μg/ml (mean: 147·6, s.d.: 21·9) LTA-SA, and also for 0·1 μg/ml (mean: 207·4, s.d.: 17·2) and 1 μg/ml (mean: 224·7, s.d.: 19·3) LTA-BS (Fig. 3e). As pointed out in Fig. 3c, stimulation with PGN-EB alone led to statistically highly significant responses of IL-17 at 5 μg/ml (mean: 26·9, s.d.: 11·1) and 10 μg/ml (mean: 21·0, s.d.: 13·2) in comparison to medium control (mean: 0·9, s.d.: 1·25). Co-activation with α-CD3 induced a strong enhancing effect for all PGNs used. While for PGN-SA (Fig. 3a) statistically highly significant frequencies could be measured at all concentrations used, with the highest frequencies for 5 μg/ml (mean: 222·3, s.d.: 19·2), the PGNs from various E. coli species showed different responses in IL-17 induction. As seen in Fig. 3b, PGN-EK showed a statistically significantly higher response at 1 μg/ml (mean: 178·5, s.d.: 18·4) and a statistically highly significant response at 5 μg/ml (mean: 195·5, s.d.: 31·6), whereas PGN-EB led to statistically highly significant frequencies at all concentrations used (Fig. 3c), with a maximum at 5 μg/ml (mean: 251·3, s.d.: 37·5).

Fig. 3.

Frequencies of interleukin (IL)-17-producing cells in thymocytes after single stimulation with peptidoglycans Staphylococcus aureus (PGN-SA) (a), Escherichia coli K12 (PGN-EK) (b), E. coli 0111:B4 (PGN-EB) (c), lipoteichoic acids S. aureus (LTA-SA) (d) and Bacillus subtilis (LTA-BS) (e) with and without co-stimulation with α-CD3 as measured by enzyme-linked immunospot (ELISPOT) assay. Thymocytes were plated at a density of 10⁶ cells per well. Data of five mice tested individually in two independent experiments are shown. Measurements were counted in duplicate wells. Each data point indicates one individual animal, line represents the mean. *P < 0·05; **P < 0·01 by Dunnett's t-test.

Inflammasome inducers trigger IL-17 responses by themselves in thymocytes; co-activation with α-CD3 results in higher frequencies of IL-17-producing cells

Thymocytes were stimulated with NRLP3 inflammasome inducers. As demonstrated in Fig. 4b, single stimulation with CPPD crystals led to statistically highly significant frequencies of IL-17 at a concentration ranging from 10 μg/ml to 300 μg/ml, with a maximum at 300 μg/ml (mean: 34·7, s.d.: 6·30). Furthermore, statistically significantly higher frequencies could be detected after single stimulation with 250 μg/ml MSU crystals (mean: 21·4, s.d.: 10·8) in comparison to medium control (mean: 8·4, s.d.: 2·86). However, co-stimulation with α-CD3 and NLRP3 inflammasome inducers enhanced IL-17 production in thymocytes. As indicated in Fig. 4a, statistically highly significant frequencies of IL-17-producing cells after combined stimulation with alum crystals and α-CD3 could be measured at 1 μg/ml (mean: 113·0, s.d.: 42·8) and 10 μg/ml (mean: 132·8, s.d.: 32·2). For CPPD crystals (Fig. 4b) and MSU crystals (Fig. 4c), statistically highly significant frequencies could be detected after combined stimulation with α-CD3 at an equal concentration. Thereby, frequencies of IL-17-producing T cells were higher after stimulation with 100 μg/ml CPPD crystals (mean: 139·3, s.d.: 41·3) in comparison to 100 μg/ml MSU crystals (mean: 111·4, s.d.: 23·9). As demonstrated in Fig. 4d, statistically highly significant responses of IL-17-producing cells could be detected after combined stimulation with α-CD3 at 200 μg/ml (mean: 138·1, s.d.: 27·8) and 350 μg/ml (mean: 140·6, s.d.: 20·4) hemozoin. However, lower concentrations already showed statistically significant responses after combined stimulation in comparison to single stimulation with α-CD3 (mean: 72·4, s.d.: 18·2).

Fig. 4.

Frequencies of interleukin (IL)-17-producing cells in thymocytes after stimulation with alum crystals (a), calcium pyrophosphate dehydrate (CPPD) crystals (b), monosodium urate (MSU) crystals (c) and hemozoin (d) with and without co-stimulation with α-CD3. Thymocytes were plated at a density of 10⁶ cells per well. Data of seven mice tested individually in two independent experiments are shown. Measurements were counted in duplicate wells. Each data point indicates one individual animal; line represents the mean.*P < 0·05; **P < 0·01 by Dunnett's t-test.

CD4⁺-positive T cells are the main IFN-γ producers in the thymus after TLR-2 activation

To confirm the finding with the ELISPOT and determine the main IFN-γ-producing population, intracellular FACS staining was performed after stimulation with HKEB, HKPA or HKSA and additional co-activation with α-CD3. As seen in Fig. 5, stimulation with HKPA and co-activation with α-CD3 (mean: 3·76%, s.d.: 0·81%) enhanced the number of IFN-γ-producing cells in comparison to α-CD3 alone (mean: 1·43%, s.d.: 0·25%). Thereby, CD4⁺-positive cells, including CD4⁺ single-positive and CD4⁺CD8⁺ double-positive cells, were the main IFN-γ producers. Together with the ELISPOT data, we see that stimulation with heat-inactivated bacterial strains can trigger IFN-γ responses after co-activation with α-CD3.

Fig. 5.

Intracellular cytokine staining of interferon (IFN)-γ gated on lymphocytes (upper) and IFN-γ production from CD4⁺-positive thymocytes (below). Thymocytes were activated with α-CD3 and stimulated with heat-killed Escherichia coli (HKEB), Pseudomonas aeruginosa (HKPA) and P. aeruginosa (HKSA). Intracellular cytokine staining was performed as described in Material and methods. Representative data of three independent experiments with four mice are shown.

Gram-positive bacteria, PGNs and LTAs change the phenotypical distribution of subpopulations in the thymus, whereas Gram-negative bacteria induce no considerable alteration

As indicated in Fig. 6, stimulation with Gram-positive and Gram-negative heat-killed bacteria can cause an alteration of thymocyte subpopulations. While activation with α-CD3 and Gram-negative bacteria did not result in any effect, stimulation with Gram-positive bacteria (especially HKLR) resulted in higher percentages of CD4⁺CD8⁺ double-positive thymocytes. As depicted in Fig. 7, stimulation with LTAs led to distinctly higher percentages of CD4⁺CD8⁺ double-positive thymocytes, whereas PGNs only had a small effect. Additional co-activation with α-CD3 illustrated no changes for either stimulation with heat-killed bacteria or for cell wall components (data not shown).

Fig. 6.

Fluorescence activated cell sorter (FACS) analysis of surface markers of thymocytes after stimulation with heat-killed bacteria. Thymocytes were stimulated with heat-killed Escherichia coli (HKEB), Pseudomonas aeruginosa (HKPA), P. aeruginosa (HKSA) and Lactobacillus rhamnosus HKLR. Single-cell suspensions were made and extracellular staining was performed as described in Material and methods. The percentages of CD4⁺, CD8⁺, CD4⁺CD8⁺ and CD4⁻ CD8⁻ T cells were analysed by FACS. Representative data of three independent experiments with four mice are shown.

Fig. 7.

Fluorescence activated cell sorter (FACS) analysis of surface markers of thymocytes after stimulation with cell wall components. Thymocytes were stimulated with lipoteichoic acids Bacillus subtilis (LTA-BS), Staphylococcus aureus (LTA-SA), peptidoglycans Escherichia coli K12 (PGN-EK), E. coli 0111:B4 (PGN-EB) and Staphylococcus aureus (PGN-SA). Single-cell suspensions were made and extracellular staining was performed as described in Material and methods. The percentages of CD4⁺, CD8⁺, CD4⁺CD8⁺ and CD4⁻ CD8⁻ T cells were analysed by FACS. Representative data of three independent experiments with four mice are shown.

CD4⁺CD8⁺ double-positive T cells are the main producers of IL-17 in thymocytes after TLR-2 stimulation and additional co-activation

To confirm the ELISPOT findings and determine the main IL-17-producing thymocyte population, IL-17 secretion assay with MACS separation and FACS analysis was performed after stimulation from α-CD3-activated thymocytes with HKEB, HKPA, HKSA, HKLR, LTA-SA and PGN-SA. As seen in Fig. 8, a higher IL-17 response could be detected for heat-killed bacteria and cell wall components after α-CD3 co-stimulation in comparison to stimulation with α-CD3 only. Furthermore, we note that the largest amounts of IL-17-producing cells were CD4⁺CD8⁺ double-positive thymocytes.

Fig. 8.

Frequencies of interleukin (IL)-17-producing cells after stimulation with heat-killed bacteria and cell wall components and enrichment with magnetic-activated cell sorting (MACS) separation. Thymocytes were activated with α-CD3 and stimulated with heat-killed Escherichia coli (HKEB), Pseudomonas aeruginosa (HKPA), P. aeruginosa (HKSA), Lactobacillus rhamnosus (HKLR), lipoteichoic acids Staphylococcus aureus (LTA-SA) and peptidoglycans Staphylococcus aureus (PGN-SA). IL-17-secretion assay was performed and IL-17-secreting cells were enriched over a MACS separation. Total numbers of IL-17-producing cells (upper) and amount of CD4⁺ or CD8⁺-positive cells gated on IL-17 (below). Representative data of three different experiments with three mice are shown.

Discussion

IL-17-producing T cells have an important function at the interface of innate and adaptive immunity 2. The majority of IL-17-producing thymocytes represent a subpopulation of CD4⁺ T cells which is able to react immediately on environmental stimuli without a further priming phase 3,6–8, but innate immune cells can also produce IL-17 in consequence of early immune responses 22. Given the strong reaction of the thymus in bacterial sepsis 11,12, this study investigated how various bacteria affect cytokine production in the thymus both with and without additional TCR activation, with a particular focus on IL-17. In general, experimental studies have focused upon understanding the activity of induced IL-17-producing T cells, which differentiate from naive T cells of the peripheral immune system. However, we and others have demonstrated that IL-17-producing T cells are also present in the thymus of naive wild-type mice, and can be co-activated by different microbial stimuli 8. The CD3 T cell co-receptor is a protein complex assembled together with the TCR heterodimer. Normally, T cell activation is initiated by the interaction of the TCR with antigenic peptides complexed to major histocompatibility complex (MHC)-II molecules. For our experiments, we activated the thymocytes via the CD3/TCR complex with soluble α-CD3 and stimulated them additionally with TLR-2 ligands 23. For this purpose, we used bacteria which were categorized in several main ‘observation parameters’, among them pathogenicity for humans, Gram status, cell respiration mechanism and extra-/intracellular localization.

Surprisingly, our data indicate that the capability to activate IL-17-producing thymus cells does not depend upon the degree of pathogenicity of a bacterial organism. Non-pathogenic bacteria (such as L. rhamnosus), species existing in or on the human body without being pathogenic (such as E. coli or S. aureus) or bacteria only causing mild local pathology (such as P. gingivalis) co-activate IL-17 producing thymocytes in a way which is quantitatively not essentially different from highly pathogenic organisms (such as L. pneumophila or P. aeruginosa), causing highly inflammatory and often lethal disease conditions. For some bacteria, IL-6 production (IL-6 representing an important mediator of the acute phase protein response and systemic inflammation 24) correlates with the degree of pathogenicity: P. aeruginosa and L. pneumophila induce high frequencies of IL-6-producing cells, but S. pneumoniae does not. Conversely, primarily non-pathogenic bacteria such as E. coli can also trigger a strong IL-6 response. As a consequence, we conclude that pathogenicity is not necessarily mirrored by the cytokine pattern in the thymus upon bacterial encounter.

As another parameter of interest, we considered the Gram status of the different bacteria. A clear effect could be seen of the cell wall on the induction of IL-6 and IFN-γ production, but not on IL-17 production: Gram-negative bacteria (e.g. H. pylori, L. pneumophila or P. aeruginosa) induce far higher frequencies of IL-6- and IFN-γ-producing cells than Gram-positive bacteria (e.g. L. monocytogenes, L. rhamnosus or S. aureus). This effect is most pronounced regarding IFN-γ production: upon co-stimulation with α-CD3 only Gram-negative bacteria have a relevant enhancing effect on IFN-γ secretion. Furthermore, we demonstrate that the majority of IFN-γ is produced by CD4⁺-positive cells, including CD4⁺ single-positive and CD4⁺CD8⁺ double-positive T cells. These differences are not so pronounced with regard to IL-6, but Gram-negative bacteria also stimulate IL-6 production stronger than Gram-positive bacteria, which is in line with clinical observations in sepsis 25,26. Therefore, it seems as if the composition of the bacterial cell wall essentially influences proinflammatory cytokine production in the thymus, but not the frequencies of IL-17 secreting cells.

In order to support our findings, in this study we tested the effect of different cell wall components on thymocytes concerning their ability to induce IL-17 production. Our data show that lipoteichoic acids from Gram-positive bacteria lead to far higher frequencies of IL-17-producing cells. Recent studies have shown that stimulation with LTAs induces activation of nuclear factor (NF)-κB via TLR-2, which results in proinflammatory cytokine secretion 27. Peptidoglycans are also potent inducers of proinflammatory cytokines via activation of NF-κB, but not via TLR-2 activation. Their immune-stimulatory activity is triggered by other pattern recognition proteins such as NOD1 and NOD2 28, but our results show that stimulation with PGNs leads to a Gram-staining independent increase of the frequency of IL-17-producing thymocytes. This means that the IL-17 response to parts of the bacterial cell wall is not only mediated through TLR-2 activation, but other pathways can also influence the IL-17 response.

Looking at the cell respiration pattern, our data do not indicate a correlation between the activation of IL-17-producing thymocytes and whether a bacterium is aerobic, anaerobic or facultative anaerobic. However, aerobic bacteria (P. aeruginosa and L. pneumophila) tend to induce the highest frequencies of IL-6 and IFN-γ-producing thymocytes. For the cellular localization (extracellular/intracellular), no clear correlation could be detected for IL-17, IL-6 or IFN-γ. With regard to the correlation between IL-6 and IL-17, our data do not suggest a dependent relationship between these cytokines. This is a crucial difference to induced IL-17-producing T cells 29 and is in line with a previous study from our laboratory, which has demonstrated that generation and activation of IL-17-producing thymocytes is independent from IL-6 in general 30. As an important technical consideration, the highly different patterns of IL-6 and IFN-γ induction of the various bacterial preparations may serve as proof that the relatively homogeneous activation of IL-17-producing thymocytes is unlikely to be caused by an artificial factor of the heat-killed state of the bacteria, as they seem to retain individual immune-stimulatory properties.

In addition to TLR-2 stimulation, we tested the influence of NLRP3 inflammasome inducers with regard to IL-17 production in thymocytes. Our data show that stimulation with NLRP3 inflammasome inducers triggers the IL-17 response in thymocytes, an effect enhanced after α-CD3 co-activation. We hypothesize that there may be a connection between these two signalling pathways. Studies describe that TLR-2/myeloid differentiation primary response gene 88 (MyD88)/NF-κB signalling is required for NLRP3 gene expression. Activation of the TLR2/MyD88/NF-κB pathway represents the first signal sequence and enhanced expression of genes for NLRP3. The second signal, consisting of both reactive oxygen species (ROS) and potassium efflux, promotes inflammasome activation 31. For this reason, we assume that increased IL-17 responses after TLR-2 activation via heat-killed bacteria or cell wall components is in conjunction with the NLRP3 pathway.

Another interesting aspect is that stimulation with heat-killed bacteria or cell wall components influence the phenotypical distribution of thymocyte subpopulations. Based on our investigations, we conclude that Gram-positive bacteria, especially their cell wall components, enhance the percentage of the CD4⁺CD8⁺ double-positive phenotype. A possible explanation for the higher percentages of CD4⁺CD8⁺ double-positive cells is a depletion of CD4⁺ and CD8⁺ single-positive cells. However, this seems to have no major role with regard to IL-17 production, because we receive higher IL-17 responses after co-activation with α-CD3 independently of bacterial strain. We summarize that the phenotypical distribution of subpopulations is not critical for IL-17 production in the thymus. Rather, all subpopulations in the thymus (CD4⁺, CD8⁺, CD4⁺CD8⁺ and CD4⁻CD8⁻) seem to have the potential to produce IL-17 as quickly as possible in case of a bacterial infection and thus initiate an immune response against possible other infections. In addition, the largest percentage of IL-17-producing thymocytes are CD4⁺CD8⁺ double-positive T cells.

We conclude that (i) the induction of IL-17 in thymocytes is a uniform pattern by bacteria in general, which (ii) is not limited to pathogenic bacteria or bacteria characterized by a specific structural/metabolic characteristic and (iii) does not appear to depend upon simultaneous IL-6 induction. This suggests that other inflammatory mediators than IL-17 are likely to be responsible for uncontrolled inflammation in the thymus causing thymocyte apoptosis and atrophy in bacterial sepsis. IL-17-producing thymocytes seem to constitute a rapidly available ‘first line of recognition’ for bacterial organisms in general, rather than a combative ‘first line of defence’ against pathogenic bacteria. Specific bacterial characteristics do not seem to be necessary to trigger their activity. IL-17-producing thymocytes might lead to a systemic immune process with a ‘fine-tuning’ control and guidance function, but not necessarily to a pathological systemic inflammatory disease condition. Their activity does not indicate a host response against bacterial virulence. This is in line with the observation that IL-17-producing T cells from the thymus are not only able to promote, but also to limit inflammatory processes in the immune periphery 5.

Disclosure

None declared.