Characterization of Specific Immune Responses to Different Aspergillus Antigens during the Course of Invasive Aspergillosis in Hematologic Patients

Several studies in mouse model of invasive aspergillosis (IA) and in healthy donors have shown that different Aspergillus antigens may stimulate different adaptive immune responses. However, the occurrence of Aspergillus-specific T cells have not yet been reported in patients with the disease. In patients with IA, we have investigated during the infection: a) whether and how specific T-cell responses to different Aspergillus antigens occur and develop; b) which antigens elicit the highest frequencies of protective immune responses and, c) whether such protective T cells could be expanded ex-vivo. Forty hematologic patients have been studied, including 22 patients with IA and 18 controls. Specific T cells producing IL-10, IFN-γ, IL-4 and IL-17A have been characterized through enzyme linked immunospot and cytokine secretion assays on 88 peripheral blood (PB) samples, by using the following recombinant antigens: GEL1p, CRF1p, PEP1p, SOD1p, α1–3glucan, β1–3glucan, galactomannan. Specific T cells were expanded through short term culture. Aspergillus-specific T cells producing non-protective interleukin-10 (IL-10) and protective interferon-gamma (IFN-γ) have been detected to all the antigens only in IA patients. Lower numbers of specific T cells producing IL-4 and IL-17A have also been shown. Protective T cells targeted predominantly Aspergillus cell wall antigens, tended to increase during the IA course and to be associated with a better clinical outcome. Aspergillus-specific T cells could be successfully generated from the PB of 8 out of 8 patients with IA and included cytotoxic subsets able to lyse Aspergillus hyphae. Aspergillus specific T-cell responses contribute to the clearance of the pathogen in immunosuppressed patients with IA and Aspergillus cell wall antigens are those mainly targeted by protective immune responses. Cytotoxic specific T cells can be expanded from immunosuppressed patients even during the infection by using the above mentioned antigens. These findings may be exploited for immunotherapeutic purposes in patients with IA.


Introduction
In mouse model of invasive aspergillosis (IA), previous studies have reported that adaptive immunity contribute to the defence of the host against fungi of the genus Aspergillus and that Aspergillusspecific T cells producing interferon-gamma (IFN-c) may result protective, while those producing interleukin-10 (IL-10) may result non-protective to the fungus [1][2][3]. In healthy subjects, it has subsequently been demonstrated that T cells may proliferate and produce different amount of IL-10, IFN-c, IL-4 and IL-17A in response to Aspergillus recombinant antigens and that the protective T cells may be expanded from their peripheral blood as possible source of adoptive therapy [4,5].
However, the emergence and changes of dynamics of Aspergillusspecific T cells have not yet been described in patients with IA and the above mentioned results could be barely applicable to immunosuppressed patients during the common clinical practice.
Thus, we have phenotypically and functionally characterized Aspergillus-specific T-cell responses against seven different Aspergillus recombinant antigens either in patients with IA, during the course of the infection, or in a comparable number of controls, and identified which antigens are most frequently targeted by protective immune responses. The results in patients with IA and controls have been compared with those obtained in healthy subjects (HS). Furthermore, Aspergillus-specific T cells have been expanded from the peripheral blood of patients with IA.

Ethics Statement
Written informed consent was obtained according to the Declaration of Helsinki, and after the local Ethical Committee's study approval (Comitato Etico Provinciale di Modena -protocol nu 2414-63/11).

Patients
Forty hematologic patients were studied: 22 patients with IA, 16 proven and 6 probable cases according to the current diagnostic criteria [6], and 18 patients with infectious complications other than IA or without infections (Table 1). Furthermore to validate the results in IA and control patients we have also studied the occurrence of specific immune response in 13 HS.

ELISpot for the Detection of Aspergillus-specific T Cells
The enzyme linked immunospot (ELISpot) assay has been performed to detect Aspergillus-specific T cell producing IL-10, IFN-c, IL-4 and IL-17A, on 101 peripheral blood samples, by using seven recombinant antigens of Aspergillus. Peripheral blood mononuclear cells (PBMCs) were obtained by Ficoll-Hypaque gradient centrifugation and co-cultured for 18-40 hours on a 96well bottom plates coated with anti-cytokine antibodies, namely anti-IL10, anti-IFN-c, anti-IL4 and anti-IL17A, as previously reported [7,8]. The viability of the analysed T cells has been determined by stimulation with anti-CD3 antibody. Aspergillusspecific T cells producing IL-10, IFN-c, IL-4 were evaluated in 13 out of 13 HS; while those producing IL-17A in 4 out of 13 HS.
Based on our previous studies, a sample was considered positive if all the followings were fulfilled: 1) the presence of at least 10 spot forming cells (SFCs); 2) the presence of more than 10 SFCs in the sample compared with the negative control; 3) the presence of a stimulation index of $20 (defined as the ratio of the number of SFCs in the positive control to that in the negative control) [7][8][9].
ELISpot results have been reported as median spot forming cells (SFCs)/10 ' 6 PBMCs with their 25 th and 75 th percentile values.

Cytofluorimetric and Functional Characterization of Aspergillus-specific T cell
In patients with IA, on peripheral blood samples resulted positive at the ELISpot assay, we have phenotypically and functionally characterized Aspergillus-specific T cells to the seven recombinant antigens by using the cytokine secretion assay (CSA) for IFN-c, IL-10, IL-4, IL-17A, according to the manufacturer's instructions (CSA Detection Kit; Miltenyi Biotec), as already reported [7,8], with few modifications. Briefly, 1610 6 PBMCs were stimulated with Aspergillus recombinant antigens, at a final concentrations of 5-7 mg/mL, from 17 to 40 hours, according to the cytokine studied. After stimulation, PBMCs were immuno- Table 1. Clinical characteristics of the patients. Caspofungin stained with IFN-c, IL-10, IL-4, or IL-17 catch reagent, and the phenotype of the cytokine producing T cells was directly assessed after sample counterstaining, as reported [7,8]. Unstimulated and PHA-stimulated PBMCs were used as negative and positive controls, respectively. Cells were acquired on a FACSCalibur flow cytometer (BD Biosciences) and analyzed by the use of CellQuest (BD Biosciences) and Summit softwares (Dako). Frequencies of antigen-reactive effector memory (EM) or central memory (CM) T cells were calculated as mean differences compared with unstimulated controls. CD8 + or CD4 + T cells were gated on CD3 + events after passing through a small lymphocyte gate. The phenotype of the cytokine producing cells was directly assessed after sample counterstaining with CD3 APC, CD8 FITC or PerCP, CD4 PerCP or APC, CD62L or CCR7 PE, allowing the identification of EM T cells (CD3+, CD8+ or CD4+, CD62L2/ CCR72) or CM T cells (CD3+, CD8+ or CD4+, CD62L+/ CCR7+). Results were expressed as mean percentages of either PB CD8 + T cells or CD4 + T cells 6 standard deviation (SD).

Aspergillus-specific T cells Expansion and Demonstration of Lytic Activity
Aspergillus-specific T cells were obtained by culturing PBMCs from 8 patients with IA: 1) alone, as control; 2) with a mixture of Aspergillus recombinant antigens (PEP1p, GEL1p, a1-3 glucan and b1-3 glucan) (8 ug/mL); and 3) with heat killed germinated conidia (100.000/ml), for 13 days in the presence of IL-2, IL-7 and IL-5. To further functionally characterize Aspergillus-specific T cells, we investigated the presence of specific-cytotoxic T cell subsets by means of the colorimetric assay with (2,3-bis[2methoxy-4-nitro-5-sulfophenyl]2H-tetrazolium-5-carboxyanilide) sodium salt (XTT; Sigma) plus coenzyme Q0 (2,3-dimethoxy-5methyl-1,4-benzoquinone; Sigma). Before the assessment of their lytic activity, cells from short term cultures were phenotypically characterized. An Aspergillus fumigatus isolate from one of the patients has been used as target. T-cell induced specific hyphal damage has been assessed as follows. Briefly, conidia of Aspergillus fumigatus (1.5610 4 per well) were plated in a 96-well flat bottom plate and incubated at 37uC for 16 hours to allow germination. On the following day, unstimulated and stimulated anti-Aspergillus T cells were added at an effector-to-target (E/T) ratio of 3:1 and 5:1 and incubated at 37uC with 5% CO 2 for 2 and 22 hours, respectively, and each experimental condition was performed in triplicate. Anti-hyphal activity was expressed as a specific hyphal damage and was calculated according to the formula: percent hyphal damage = [12((X2Y)/C)] 6100, where X is the absorbance of experimental wells with stimulated cells, Y is the absorbance of experimental wells with unstimulated cells and C is the absorbance of control wells with hyphae only. In three out of 8 patients, to evaluate whether anti-hyphal activity of Aspergillusspecific T cells was similar to those of antigen presenting cells (APCs) and polymorphonuclears (PMNs) and how it was accomplished, we compared the killing ability of anti-Aspergillus T cells, APCs, PMNs and the supernatant of T-cell cultures. Furthermore, we quantified the amount of cytokines (Granzyme B and IFN-c) produced by PBMCs after 24 h culture with the above mentioned antigens by enzyme linked immunosorbent assay (ELISA).

Phases of Invasive Aspergillosis
Based on the kinetics of radiologic signs of pulmonary IA on HRCT [12,13], the infection course has been divided into four phases, defined from t1 to t4, and corresponding to the number of days elapsed from the radiological diagnosis of IA (t1, from the radiologic disclosure of the infectious lesions until day +14; t2, Table 1. Cont. from day +15 until day +30; t3, from day +31 to day +45; t4, from day +46 to day 60 days from the radiological diagnosis of IA).
Only two patients had a sample collected more than 61 days after the radiological demonstration of IA and the respective results were shown as t4+.

Statistical Analysis
The Fischer's exact test has been used to determine if there were nonrandom associations between two categorical variables and the outcome of the patients; the association between the number of antigens targeted by specific T cells and the status of the subjects (HS or patients with IA). Chi-squared test has been used to compare the rate of samples with more than 100 SFCs/10 ' 6 PBMCs to all the antigens between HS and patients with IA. The paired T test has been used to compare the rates of fungal hyphae lysis between the single cell fractions and their associations. P values below.05 were considered significant. The results were obtained using the Stata Software (11.0, College Station, Texas, USA).

Identification of Aspergillus-specific T cells
A median of 2 time-points were analysed for each patient (range from 1 to 5) ( Table 1). All the patients with IA presented Aspergillusspecific T cells at least at one time point (Figures 1 and 2). Patients who never showed specific T cells producing IFN-c or demonstrated such cells against only one antigen (pts 16,17,18,19,20,21,22), presented more frequently an unfavourable outcome compared with patients presenting protective immune responses versus two or more antigens. Actually, 4 (pts 17, 19, 21, 22) out of 7 patients died of IA in the first group, while only 1 (pt nu1) out of 15 patients died of the infection in the second group (p = 0.021) (Figures 1 and 2).
Specific T cells producing IL-4 and IL-17A could be detected in only one patient (pt7; Figure 1) by the ELISpot assay.
In the 18 control patients the occurrence of Aspergillus-specific T cells could not be demonstrated (Table 1). The comparison of specific immune responses producing IL-10 and IFN-c between patients with IA and HS demonstrated that: 1) patients with IA presented higher number of Aspergillus-specific T cells producing IL-10 (100 or more SFCs/10 ' 6 PBMCs) to each of the seven recombinant antigens than HS (55.56% vs 2.2% of the tested samples, p = .0001); 2) patients with IA presented wider antigenic specificity (directed to 4 or more antigens) of Aspergillusspecific T cells producing either IL-10 (77.27% of the IA patients Blue columns represent the number of Aspergillus-specific T cells producing IFN-c; yellow columns represent the number of Aspergillus-specific T cells producing IL-10; green columns represent the number of Aspergillus-specific T cells producing IL-17A; red columns represent the number of Aspergillus-specific T cells producing IL-4; dark gray background represents T-cell responses in wells with anti-CD3 antibody. Vertical axis shows the number of spot-forming cells (SFCs) per million peripheral blood mononuclear cells (PBMCs); horizontal axis indicates the antigen, which the specific immune responses are directed to and the phase of IA calculated as the number of days elapsed from the radiological diagnosis of the infection (t1, from the radiologic disclosure of the infectious lesions until day +14; t2, from day +15 until day +30; t3, from day +31 to day +45; t4, from day +46 to day 60 days; t4+ .61 days). CRF1 = ortholog of Crh1p associated in b1,6 glucan-chitin linkages; PEP1 = aspartic protease; GEL1 = 1,3-b glucanosyltransferase; a1-3 glu = a1-3 glucan; b1-3 glu = b1-3 glucan; SOD1 = superoxide dismutase; and GM = galactomannan. Pt = patient. * = patients with unfavorable outcome. ' = patients with probable IA. The red dotted line divides the two groups of patients: from pt nu 1 to pt nu 15 with Aspergillus-specific T cells producing IFN-c to two or more antigens; from pt nu 16 to pt nu 22 without Aspergillus-specific T cells producing IFN-c or with such cells to only one antigen. doi:10.1371/journal.pone.0074326.g002 vs 15.38% of HS, p = .0001) or IFN-c (54.25% of the IA patients vs 0% of HS, p = .031).

Phenotypic and Functional Characterization of Aspergillus-specific T cells
The CSA analysis showed that antigen-specific T cells were: 1) either CD8+ or CD4+ T cells producing IFN-c (mean

Discussion
Here we report, for the first time, how Aspergillus-specific immune responses to different Aspergillus antigens may emerge and develop in patients with IA. The absence of such immune responses in patients without IA demonstrates that: a) they are specific to Aspergillus spp., b) unrelated to the administered treatments, c) and not in antithesis with previous reports and our data showing Aspergillus specific T cells in healthy subjects [4]. Indeed, intensive chemotherapy may cause the reduction of such cells at frequencies below the ELISpot threshold, while the occurrence of IA, with the antigen rechallenge, may increase again specific immune responses. Moreover, compared with the results obtained in HS, patients with IA demonstrated noticeably higher frequencies of Aspergillus-specific T cells producing IL-10 and Aspergillus-specific T cells producing IFN-c with wider antigenic specificity, implying that, given such a set of antigens, the occurrence of high number of the former cells and of a wide antigenic specificity of the latter may represent the immunologic signature of IA. These data may also provide the rationale for further studies to use the ELISpot as a novel diagnostic tool for IA. Of note, Aspergillus-specific T cells producing IL-17A may be detected more frequently and to more antigens in HS than in patients with IA, suggesting that the loss of such cells may be implicated in the pathogenesis of IA. However, further data are necessary to draw any firm conclusions about this.
Our study shows that although IA is predominantly characterized by the presence of specific T cells producing IL-10, protective immune responses may occur since the onset of and tend to increase during IA. The antigens eliciting the highest frequencies of specific T cells producing IFN-c are all involved in cell wall biosynthesis of Aspergillus, namely GEL1p and both glucans, partially in agreement with the results obtained in mice with IA, where GEL1p, CRF1p and a1-3 glucan resulted the antigens associated with a higher activation of type 1 responses [5]. The recognition of polysaccharides is probably mediated by specific T cells through the class II major histocompatibility complex (MHC) on APCs, as previously reported for other infectious agents showing glycoantigens, such as Staphylococcus aureus, Streptococcus pneumoniae and Bacteroides fragilis [14,15]. Moreover, in our patients, the association between T cells producing IFN-c to two or more antigens and a more favorable outcome may suggest that such protective T cells play a role in the resolution of the infection.
The functional and phenotypical characterization shows that Aspergillus-specific T cells are either CD4+ or CD8+ T cells, and that T cells producing either IL-4 or IL-17A may be detected to all the antigens, suggesting that they are present at very low frequencies in patients with IA. These findings confirm that CD8+ T cells contribute to the immune response to filamentous fungi [7,16], and are in line with those reports showing that Aspergillus fumigatus is a weak inducer of IL-4 and IL-17A responses [5,17].
The absence of concordance between the Elispot assay and CSA, in particular about the frequencies of specific T cells producing IL-10, IL-4 and IL-17A is not unexpected and reflects the different intrinsic characteristics of the two assays, being the ELISpot assay more suitable to detect low-level responses and to define such responses as either positive or negative, while the CSA more appropriate for a phenotypic discrimination of responding cells, as already reported [7,[16][17][18].
By expanding previous findings, so far obtained only in healthy donors [19][20][21], we have expanded Aspergillus-specific T cells from PB of patients with IA using a mixture of recombinant antigens different from CRF1p, and found that such cells may induce direct lysis of Aspergillus hyphae, comparable to those obtained by APCs and PMNs. By showing that the hyphal lysis obtained with the supernatant of anti-Aspergillus CTLs is lower than that registered with the whole CTLs culture (anti-Aspergillus CTLs+supernatant), and that the production of Granzyme B and IFN-c was higher in stimulated samples, our data suggest that the activity against Aspergillus is mediated directly by T cells, either by secreting cytokines or by cell-cell contact, and are consistent with previous reports [22,23]. The observation that protective immune responses in HS are almost exclusively elicited by CRF1p, while they are predominantly directed to Aspergillus cell wall antigens in patients with IA, should be taken into account when planning immunotherapeutic strategies.
In conclusion, specific immune responses to several Aspergillus antigens may be detected in patients with IA and those producing IFN-c directly mediate the lysis of Aspergillus hyphae, possibly contributing to the clearance of the fungus, being their presence also associated with a more favourable outcome. The identification of the antigens eliciting the strongest protective responses in patients with IA may spur further studies to define the complete repertoire of Aspergillus specific immune responses during the course of the infection to design therapeutic strategies of either vaccine or autologous cytotoxic cell infusions.
cultures. E/T = effector/target cell ratio. D. Rates of hyphal damage by Aspergillus-specific T cells expanded with Aspergillus either heat killed conidia (white columns) or recombinant antigens (GEL1, PEP1, a1-3 glucan, b1-3 glucan) (black columns); by the supernatant of cultures of Aspergillusspecific T cell lines only; by antigen presenting cells (APCs); by polymophonuclears (PMNs) and by combinations of the different three cell fractions from three further patients with IA, at two-hour cultures. * = P,.05. E/T = effector/target cell ratio. Results are expressed as mean+2Standard Deviation. doi:10.1371/journal.pone.0074326.g005