Toll-like receptor 2 induced cytotoxic T-lymphocyte-associated protein 4 regulates Aspergillus-induced regulatory T-cells with pro-inflammatory characteristics

Patients with cystic fibrosis, chronic obstructive pulmonary disease, severe asthma, pre-existing pulmonary lesions, and severely immunocompromised patients are susceptible to develop infections with the opportunistic pathogenic fungus Aspergillus fumigatus, called aspergillosis. Infections in these patients are associated with persistent pro-inflammatory T-helper (TH)2 and TH17 responses. Regulatory T-cells, natural suppressor cells of the immune system, control pro-inflammatory T-cell responses, but can also contribute to disease by shifting to a pro-inflammatory TH17-like phenotype. Such a shift could play an important role in the detrimental immunopathology that is seen in aspergillosis. Our study demonstrates that Aspergillus fumigatus induces regulatory T-cells with a TH17-like phenotype. We also demonstrate that these regulatory T-cells with a pro-inflammatory TH17-like phenotype can be reprogrammed to their “classical” anti-inflammatory phenotype by activating Toll-like receptor 2 (TLR2), which regulates the induction of cytotoxic T-lymphocyte-associated protein 4 (CTLA4). Similarly, soluble CTLA4 could reverse the pro-inflammatory phenotype of Aspergillus-induced regulatory T-cells. In conclusion, our results suggest a role for regulatory T-cells with a pro-inflammatory TH17-like phenotype in Aspergillus-associated immunopathology, and identifies key players, i.e. TLR2 and CTLA4, involved in this mechanism.


A. fumigatus induces regulatory T-cells with a pro-inflammatory T H 17-like phenotype. By
determining the kinetics of IL-17A and IL-10 production over a course of 7 days in PBMCs simulated with A. fumigatus conidia, we determined the optimal time point to detect T H 17-like pro-inflammatory T reg cells. Similar to previous studies with Candida albicans 24 , A. fumigatus-induced IL-17A release was detectable at peak levels in supernatants after 7 days (Fig. 1A). The production of IL-10, an anti-inflammatory cytokine required for T reg cell differentiation 25 , which is also produced by T reg cells, demonstrated a biphasic response (Fig. 1A). Expression of the transcription factor of T H 17 cells, RORγt, started to increase after 5 days exposure to conidia, similar to IL-17A release, whereas FoxP3 expression was already induced after 24 h and continued to increase in expression towards 7 days (Fig. 1B).
To detect A. fumigatus-induced T H 17-like pro-inflammatory T reg cells, PBMCs were stained and measured by flowcytometry after 7 days of stimulation with A. fumigatus conidia. T-cells were identified through CD4 (Fig. 1C). Within the CD4 + population, the number of T reg cells was quantified as the percentage of CD25 + FoxP3 + cells (Fig. 1D). T H 17 cells were quantified as RORγt + IL-17A + cells within the CD4 + population (Fig. 1E). Finally, the percentage of cells with T H 17 markers, i.e. RORγt / IL-17A, was determined within the T reg population, i.e. CD25 + FoxP3 + (Fig. 1F).
In order to assess the cytokine release by these different cell populations, IL-10 production was measured in the culture supernatant after 24 hours and 7 days, and IL-17A production was measured after 7 days. After 7 days of stimulation, production of both IL-10 and IL-17A was significantly increased (p = 0.0273 n = 15 and p < 0.0001, n = 17) ( Fig. 2E and F). However, this effect was not observed for IL-10 after 24 hours (Fig. 2D).

TLR2 regulates regulatory T-cells with a T H 17-like phenotype. Toll-like receptor (TLR)2 is involved
in the recognition of pathogen-associated molecular patterns (PAMPS) in the Aspergillus cell wall (reviewed in ref. 26), and is associated with the induction of T reg cells in response to fungi 27,28 . Based on the observation that naïve splenocytes of Tlr2 −/− mice produce higher levels of IL-17A than wild type (WT) mice (Supplementary Figure 1), we hypothesized that TLR2 could have a role in shaping the population of Aspergillus-induced regulatory T-cells.
To investigate the role of TLR2 in the induction of T reg cells with a T H 17-like phenotype, human PBMCs were stimulated with A. fumigatus conidia for 24 hours and 7 days while TLR2 was blocked with a neutralizing antibody.
As demonstrated previously 29 , blocking TLR2 before stimulating with A. fumigatus conidia resulted in a significant increase in IL-17A production (p = 0.0039 n = 9). However, no change in IL-10 production after 24 hours, and after 7 days was observed (Fig. 3A). Within the CD4 + population, the number of CD25 + FoxP3 + T reg cells significantly decreased with TLR2 blockade (p = 0.0117 n = 9), while a non-significant trend towards increased expression of T H 17 cell-characteristics, i.e. RORγt + IL-17A + within these cells was observed (p = 0.1875 n = 6) (Fig. 3B)

CTLA4 regulates the induction of Aspergillus-induced regulatory T-cells with a T H 17-like phenotype.
One of the most potent molecules that regulates the induction of pro-inflammatory T-cell subsets is CTLA4. To determine whether there is a role for CTLA4 in controlling the induction of IL-17A-expressing regulatory T-cells by Aspergillus, we investigated these cells in a patient with CTLA4 deficiency. The CTLA4 deficient patient showed slightly reduced induction of CD25 + FoxP3 + regulatory T-cells and a similar induction of IL-17A + T H 17 cells, compared to the healthy control. T H 17 cell-characteristics, i.e. IL-17A + , within CD25 + FoxP3 + T reg cells, were substantially more induced by the patient compared to the healthy control (Supplementary Figure 2). To assess whether the capacity of TLR2 to reduce the expansion of regulatory T-cells with a T H 17 phenotype was due to modulation of CTLA4, the capability to induce CTLA4 through TLR2 signalling was investigated. Activation of TLR2 by the agonist P3C resulted in a significant upregulation of CTLA4 expressing CD4 + T-cells (p = 0.0049 n = 12). Similarly, co-stimulation of TLR2 with A. fumigatus significantly upregulated the induction of CTLA4 expressing CD4 + T-cells, compared to stimulation with A. fumigatus (p = 0.0005 n = 12) (Fig. 4A). Soluble CTLA4-Ig (Abatacept) is known to inhibit the T-cell co-stimulatory molecule B7-1 (CD80), similar to endogenous CTLA4 30 . Using soluble CTLA4, we validated whether increased CTLA4 levels could reduce Aspergillus-induced CD25 + FoxP3 + regulatory T-cells with a T H 17 phenotype. Addition of CTLA4-Ig to human PBMCs resulted in a decreased IL-17A production after 7 days (p = 0.0234 n = 7) (Fig. 4B). Also, induction of the number of CD25 + FoxP3 + regulatory T-cells expressing a T H 17 phenotype, i.e. IL-17A + , CD25 + FoxP3 + regulatory T-cells and IL-17A + T H 17 cells significantly decreased upon addition of CTLA4-Ig (p = 0.0343, p = 0.0343 and p = 0.0469 n = 7) (Fig. 4D).

Discussion
In this study we investigated whether A. fumigatus-induced T reg cells exert a pro-inflammatory phenotype, and if this pro-inflammatory phenotype can be reprogrammed to the "classical" anti-inflammatory T reg phenotype. Human PBMCs exposed to A. fumigatus conidia showed induction of T reg cells with pro-inflammatory T H 17-like characteristics. Induction of classical (CD25 + FoxP3 + ) T reg cells and (IL-17A + RORγt + ) T H 17 cells was also significantly increased and related with increased production of IL-10 and IL-17A. These results indicate that A. fumigatus can indeed induce T reg cells with a pro-inflammatory T H 17-like phenotype, which potentially could contribute to detrimental IL-17-mediated immunopathology.
Although the IL-17A axis plays an important role in the protective immunity against fungal pathogens such as A. fumigatus 31 , it has been demonstrated that in some cases IL-17A mediated immune responses overwhelm this protective effect, promoting infection and impairing antifungal immunity 13,32 . Diseases such as CPA and ABPA are characterized by a persistent hyper inflammatory state with massive influx of neutrophils and eosinophils 12,15 , which may be attributed to exaggerated T H 17 responses. T reg cells are potent suppressors of T H 17 cells 17 , yet this suppressive effect is annulled when regulatory T-cells also acquire T H 17 characteristics. In our study, we observed that the reduction of T reg cells with a T H 17 phenotype, either through TLR2 stimulation or Abatacept, correlated with a reduction of the T H 17 cytokine IL-17. Since it is technically difficult to determine the origin of IL-17 in culture supernatants we cannot ascertain that this reduction of IL-17A is strictly due to a reduction of IL-17 + T reg cells.
TLR2, together with TLR4 and TLR9, is an important Toll-like receptor in the host defence against Aspergillus (reviewed in refs 26, 33). The observation that splenocytes of Tlr2 −/− mice have an increased A. fumigatus-induced production of IL-17A, directed us towards a possible suppressive role for TLR2 in the induction of IL-17 responses. In addition, many studies have demonstrated a critical role for TLR2 in the induction of regulatory T-cells 27,28,34 . We found that blocking TLR2, while stimulating with A. fumigatus conidia, exaggerated the induction of T reg cells with pro-inflammatory T H 17-like characteristics. Reversely, we observed that TLR2 co-stimulation not only increased the induction of regulatory T-cells, but also dampened the induction of regulatory T-cells with a T H 17 phenotype. We suggest that altered TLR2 signalling may be involved in the aberrant induction of T reg cells with pro-inflammatory T H 17-like characteristics. Whether alterations in TLR2 signalling, pathogen-or host-related, are actually involved in IL-17 mediated immunopathology needs further investigation. An attractive approach would be to investigate whether common genetic variations that slightly alter TLR2 function are associated with detrimental IL-17-mediated inflammation in aspergillosis.
TLR2 activating therapy is currently under consideration as a potential immunotherapy to increase the number of regulatory T-cells, reducing T H 2 mediated hypersensitivity 35 , and increasing anti-tumour capacity 36 . However, whether such therapies are attractive to employ for treatment of aspergillosis remains to be determined. Experimental evidence highlights that TLR2 plays an important role in the antifungal host response against Aspergillus. TLR2 deficient mice for example show an impaired recruitment of neutrophils to the site of infection 37 , and a higher lethality and fungal burden 33 . Interestingly, these mice also show an increased T H 2 responses compared to WT mice upon stimulation with A. fumigatus 37 . This T H 2 response, associated with ABPA, suppresses the protective T H 1 response 38 . In line with our data it is tempting to speculate that the TLR2 deficient mice are more susceptible to aspergillosis due to modulation of T reg cells, as TLR2 mediated induction of T reg cells is crucial for maintaining pro-inflammatory responses and promoting fungal clearance in aspergillosis 33,37 . Several studies have clearly demonstrated a protective role of T reg cells in fungal infection 39,40 , and that TLR2 plays a crucial role in maintaining this population 28,34 . It can however not be excluded that TLR2 has multiple lines of action in host defence against aspergillosis, such as induction of pro-inflammatory responses crucial for recruitment and activation of innate immune cells that are responsible for clearing the fungi from the lungs.
T reg cells partly exert their anti-inflammatory function through contact dependent mechanisms such as CTLA4 20 . PBMCs of one patient who was deficient for CTLA4 were stimulated with A. fumigatus. Interestingly, a higher induction of T reg cells with a pro-inflammatory T H 17-like phenotype was observed, compared to the healthy control. In the field of rheumatology, therapeutics such as Abatacept (CTLA4-Ig) are often administered to reduce pro-inflammatory responses in which the T H 17 -T reg balance shifts towards the T H 17 cell-response 41 . Therefore, such therapeutics might help reverse T reg cells with pro-inflammatory T H 17-like characteristics back to their "classical" anti-inflammatory phenotype. We examined the role of Abatacept in preventing the induction of T reg cells with T H 17-like characteristics, and a significantly decreased induction of T reg cells with T H 17-like characteristics was seen. This indicates that Abatacept has the potential to hamper the induction of T reg cells with pro-inflammatory T H 17-like characteristics. These results also indicate that CTLA4 might play a role in hampering the induction of T reg cells with pro-inflammatory T H 17-like characteristics, and warrants further investigation to explore whether Abatacept could be used as a targeted therapy in patients with aspergillosis that are suffering from infection-induced immunopathology.
We observed that TLR2 stimulation increases the number of CTLA4 expressing CD4 T-cells and the level of expression on these cells. Separately we demonstrated that either TLR2 stimulation or CTLA4 could reduce the induction of T reg cells with pro-inflammatory T H 17-like characteristics. The fact that TLR2 signalling boosts CTLA4 expression could suggest that TLR2 modulates the T H 17 phenotype of Aspergillus-induced T reg cells through CTLA4. It should be noted that Aspergillus itself did not significantly induce CTLA4 expression. It is therefore tempting to suggest that Aspergillus induces such a pro-inflammatory T reg phenotype due to the fact that it fails to induce CTLA4 expression, but does induce the inflammatory mediators required for induction of pro-inflammatory T-cell responses.
Collectively, our study demonstrates that A. fumigatus is capable of inducing regulatory T-cells with a pro-inflammatory T H 17-like phenotype. In addition, we found that TLR2 and CTLA4 play a role in regulating the induction of these cells. These findings could pave the way for novel therapeutic approaches that target T reg cells with pro-inflammatory T H 17-like characteristics in order to return them to their natural immune regulatory state.

Methods
Healthy volunteers and patients. Blood was collected from healthy volunteers and patients by venous blood puncture, after informed consent was obtained. One patient that was deficient for CTLA4 was included, and for those experiment the healthy control was age (±5 years) and sex-matched. All experiments were performed and conducted in accordance to Good Clinical practice, the Declaration of Helsinki, and the approval of the Arnhem-Nijmegen Ethical Committee (nr.2010/104).

Aspergillus fumigatus.
A. fumigatus V05-27, a previously characterized clinical isolate 42 , was used for all stimulations. Resting conidia were heat-inactivated, for 1 h at 65 °C and were checked for viability on Sabouraud agar. Conidia were stored at −80 °C until use. A concentration of 1 × 10 7 /mL was used in the experiments, unless otherwise indicated.

Isolation and stimulation of peripheral blood mononuclear cells (PBMCs). Blood was diluted in
phosphate buffered saline (PBS) (1:1) and fractions were separated by Ficoll (Ficoll-Paque Plus; GE healthcare, Zeist, The Netherlands) density gradient centrifugation. Cells were washed twice with PBS and resuspended in Roswell Park Memorial Institute medium (RPMI) 1640 Dutch modification culture medium (Life Technologies/ Invitrogen, Breda, The Netherlands) supplemented with 50 μg/mL gentamicin, 2 mM Glutamax, and 1 mM pyruvate (Life Technologies). Cells were counted using a particle counter (Beckmann Coulter, Woerden, The Netherlands) after which, the concentration was adjusted to 5 × 10 6 /mL. PBMCs were plated in 96-well round-bottom plates (Corning) at a concentration of 5 × 10 5 /mL in a total volume of 200 µL. The samples were stimulated with A. fumigatus heat inactivated conidia (1 × 10 7 /mL), with or without TLR2 blockers/ligands or Abatacept, or remained unstimulated for either 24 hours or 7 days at 37 °C with 5% CO 2 . After stimulation, supernatants were collected and stored at −20 °C until cytokine assays were performed. Cell pellets were used for flowcytometry.
Cytokine measurements. IL-17A (R&D Systems, Mineapolis, MN) and IL-10 (Sanquin, Amsterdam, the Netherlands) were measured using commercially available ELISAs according to the protocols that were supplied by the manufacturers. Mouse IL-17A was assessed in splenocyte stimulations using the Luminex multiplex platform (Millipore, Billerica, MA).
The cells were measured on a FC500 flow cytometer (Beckman Coulter) and the data were analysed using CXP analysis software v2.2 (Beckman Coulter). mice were bred and maintained in the St. Jude Children's Research Hospital, Memphis, TN, USA. Spleens were homogenized in 0,4 μM cell strainer (BD) and the cell number was adjusted to 1 × 10 7 /mL. The cell suspensions (500 μL/well) were placed in a 24 well plate (Corning) and incubated with culture medium or A. fumigatus conidia for 1 or 5 days at 37 °C and 5% CO 2 .

Ex vivo
Statistical analysis. Data of PBMC stimulation with either RPMI or A. fumigatus were analysed using the Wilcoxon signed rank test. P-values < 0.05 were considered statistically significant. Data are shown as scattered dot plots, columns with mean ± standard error of mean (SEM), or as a Venn diagram. All data were analysed using GraphPad Prism v5.0. The proportional Venn diagram was drawn using the eulerAPE application v2.0.3 43 .