Identification and characterization of Aspergillus-specific immune responses to diagnose invasive aspergillosis in high risk patients: a multicenter study

Background. The mortality of Invasive Aspergillosis (IA) still affects from 27% to 55% of high risk hematologic patients. The reasons of such a poor outcome also rely on several drawbacks limiting the di- agnostic accuracy of non cultural based diagnostic methods (NCBDM) and hampering the opportunities for an early intervention. Studies in mice model of IA and in healthy subjects have shown that Aspergillus-specific T-cells producing interferon-gamma (IFN- gamma-T1) are protective, while Aspergillus-specific T-cells pro- ducing interleukin-10 (IL-10-T2) are non-protective to IA. Aims. We have investigated whether the identification of Aspergillus-specific IFN-gamma-T1 and/or IL-10-T2 through an ex-vivo enzyme linked immunospot (ELISPOT) assay may be effective in the diagnosis of IA in high risk patients. Furthermore, in the proven IA patients, we have functionally and phenotipically characterized such T cells through the cytokine secretion assay (CSA). Methods. 180 patients (168 hemato- logic and 12 solid organ transplant patients) have been enrolled. They were classified, according the revised EORTC/MSG criteria, as fol- lows: 18 proven, 35 probable, 17 possible IA cases and 110 controls. The control patients were divided in two groups: group 1 included 86 (78.2%) patients with hystological and/or cultural verified infec- tious/inflammatory/neoplastic diseases, but other than IA; group 2 in- cluded 24 (21.8%) patients without clinical and/or microbiological features of IA. ELISPOT has been performed, as described [Potenza et al. Leukemia 2007; 21: 578-81], by using as antigens Aspergillus either conidia or recombinant antigens, namely CRF1p, GEL1p, PEP1p, SOD1p, α1-3 glucan, β1-3 glucan and galactomannan (GM). Results. The patient and sample positivity rates were 94.4%/89.5% in proven, 45.7%/35.3% in probable, 35.3%/50% in possible IA cases and 1.8%/4.5% in the controls, respectively. The sensitivity and speci- ficity of ELISPOT for the diagnosis of IA resulted 94.4% (95% CI, 73%-99%) and 98.2% (95% CI, 93%-99%), respectively. The PPV of the test was 89.5% (95% CI, 67%-99%), the NPV was 99.1% (95% CI, 94%-100%) and the efficiency was 97.6% (95% CI, 92.3%- 99.4%). The positive likelihood ratio (LR) resulted 51.89, the negative LR was 0.06 (Table 1A,B). In proven IA patients, CSA demonstrated that Aspergillus-specific IL-10-T2 were predominantly central memory (CM) CD4+ T cells (median frequency 0.37%/0.22%), while Aspergillus-specific IFN-gamma-T1 were either CD4+ or CD8+ cells of either effector memory (EM) or CM phenotype (median frequen- cies 0.24%/0.20%). Also lower frequencies of Aspergillus-specific ei- ther CD4+ or CD8+ T cells producing IL-4 (0.11%/0.19%) of EM phenotype, and EM CD8+ cells producing IL-17 (0.18%), were de- tected. Moreover, although CRF1p, GEL1p, α1-3 glucan and SOD1p resulted the antigens eliciting the highest number of Aspergillus-spe- cific IFN-gamma-T1, only GEL1p and α1-3 glucan were those most constantly targeted by protective immune responses along the entire course of the IA. Conclusions. Our findings demonstrate the potential of ELISPOT in the diagnosis of IA, suggesting that it may comple- ment the other NCBDM, enabling a more consistent diagnosis of IA. Furthermore, this study describes for the first time the Aspergillus- specific immune responses in patients with proven IA, identifying also the antigens predominantly targeted by protective IFN-gamma- T1, with possible consequences in designing strategies of either adoptive cell infusion or vaccine therapies.