Hypersensitivity pneumonitis (HP) is a family of interstitial lung diseases that occur in susceptible individuals following exposure to a variety of inciting antigens. HP has been recently classified according to clinical, radiological, and pathological characteristics into two forms: fibrotic (chronic) and non-fibrotic (acute, predominantly inflammatory HP) (Raghu et al., 2020).

HP is an orphan disease, with only a minority of individuals exposed to antigens that potentially induce HP developing the disease (Girard et al., 2009). Genetic factors, such as MUC5B, have been proposed as risk factors to develop fibrotic HP (Ley et al., 2017). Tobacco smoking reduces the risk of HP, likely because nicotine attenuates the production of IL-1 and TNF and promotes phagocytosis by macrophages (Barnes et al., 2022). Smoking is also thought to affect the immune response to antigen exposure, steering the process toward fibrosis (Hamblin et al., 2022). Continuous contact with antigens can lead to permanent lung damage (Cormier, 2014). Agents capable of inducing HP are found in various settings, including the workplace, home, and recreational environments. It has been estimated that 12 to 28 % of HP cases are directly attributable to workplace exposure, the most common of which is farming (Barnes et al., 2022).

Farmer's lung disease (FLD) was the first form of hypersensitivity pneumonitis (HP) to be described and is still one of the most frequently occurring forms in the world (Pepys and Farmer's lung., 1994; Lacasse et al., 2003). FLD is due to regular exposure to high quantities of microbial antigens, especially during the handling of moldy hay, straw, or grain dust stored in conditions of high humidity (Pepys and Farmer's lung., 1994) (Gomes et al., 2021). Antigens responsible for FLD vary between countries and depend on climate, farming, and production conditions. In the East of France and Finland, the main etiological agents are actinomycetes (Saccharopolyspora rectivirgula (SR), Thermoactimomyces vulgaris (TV), and Saccharomonospora viridis (SV)) and molds (Eurotium amstelodami (EA), Lichtheimia corymbifera (LC), and Wallemia sebi (WS)) (Pepys and Farmer's lung., 1994; Lacasse et al., 2003; Dalphin et al., 1994; Bellanger et al., 2019; Fenoglio et al., 2007).

In the past, the serological diagnosis of FLD cases was carried out exclusively using immunoprecipitation (IP) techniques. The panel of antigens classically used by our laboratory for the serological diagnosis of FLD consists of the lyophilized somatic antigens of the six micro-organisms listed above. However, IP techniques are time consuming and require a high level of expertise. One such technique, electrosyneresis, was abandoned recently due to the lack of cellulose acetate strips on the market, although it had previously been validated as a confirmation technique (Fenoglio et al., 2007). The measurements of serum-specific immunoglobulin (SS-IgG) is a diagnostic test proposed for several forms of HP (Johannson et al., 2020). Establishing the diagnosis of FLD is crucial for the implementation of preventive measures or even the proposal of professional reconversion and for disease prognosis. There is a need to develop a new approach to serodiagnosis of FLD using standardized techniques that can be easily used by non-specialist laboratories. As early as 2005, Fink et al. highlighted the need to improve standardization of the antigens used for these serological assays (Fink et al., 2005). An alternative to somatic antigens are purified antigens (PAgs) and recombinant antigens (r-Ags). Somatic antigens are obtained from the microorganisms present in the material manipulated by patients. Somatic antigens are composed of a variety of antigenic fractions. Proteic purified antigens are obtained through the enzymatic lysis of cell wall polysaccharides, protein acid precipitation, and acetone purification. PAgs are more standardized than somatic antigens.

Proteomic analysis to compare antigenic proteins between FLD and exposed control populations allowed us to identify protein biomarkers from SR, LC, and EA (Barrera et al., 2014a; Millon et al., 2012; Rognon et al., 2016). In total, 17 recombinant proteins were produced for SR, six for LC, and five for EA (Barrera et al., 2014a; Millon et al., 2012; Rognon et al., 2016). Enzyme-linked immunosorbent assay (ELISA) tests showed good diagnostic performance with three r-Ags from SR (SR17, SR1FA, SR22) (Barrera et al., 2014a), two from LC (DLDH, Psubα), (Rognon et al., 2016), and two from EA (GLPV, G6PI) (Millon et al., 2012). Sensitivity (Se) using these r-Ags was between 72 % and 83 % and specificity (Sp) between 72 % and 94 %. The sensitivity of a test, noted as ‘Se’, corresponds to its ability to correctly identify individuals with the disease (true positives), while the specificity, noted as ‘Sp’, designates its ability to correctly identify individuals without the disease (true negatives) (Trevethan, 2017).

The main objective of the present study was to develop an effective global strategy using an ELISA for the serodiagnosis of FLD. The initial phase of our study focused on identifying a high performance combination of PAgs and r-Ags that could be used as a panel of antigens for an ELISA using sera from an initial group of patients (group 1). The second phase aimed to validate the most efficient strategy for FLD serodiagnosis combining screening by ELISA and confirmation using an IP technique on another group of FLD and non-FLD patients (group 2). To improve the performance of the serological tests, we used a mixture of PAgs and r-Ags, as used in certain commercially available ELISA kits (Dumollard et al., 2016).