Numbered sera of BD showed no reactivity to B. henselae proteins (Figs 1, 3 and 4). Table 2 shows the Se and Sp of clinical biomarkers based on the immunoreactivity of the B. henselae proteome against serum samples from CSD and IE patients. For these biomarker proteins, combining both IE and CSD, the sensitivity ranged Enzalutamide in vivo from 21% to 100%. The sensitivity for IE serum samples was higher than that of CSD serum samples, ranging from 43% to 100%, with the lowest and the highest reactivity percentage observed for Pnp and GroEL, respectively. Although ATPD, GroEL and FusA each showed a sensitivity of 100% for CSD serum samples, these proteins were also detected using the control group serum samples as shown in

Fig. 2. The proteins that showed 100% specificity for both diseases were GroES, HbpD, Pap31, PdhD2, SodB, Ppi and two proteins of a nonapplicable locus: BH11510 (OMP) and BH12180 (ABC IDH inhibitor transporter, periplasmic oligopeptide-binding protein). Based on these results, BH11510, BH12180, GroES, Pap31, PdhD2 and SodB were selected as biomarkers for IE, while BH02000 was selected as a biomarker for CSD. The aim of this study was to identify the serodiagnostic markers of CSD and endocarditis due to B. henselae and expand the number of biomarkers selected previously by others (McCool et al., 2008; Eberhardt et al., 2009). Indirectly, our study allowed cross-validation of some

protein targets for clinical application. The systemic infection leading to the massive infiltration of bacteria may be explained by the higher IFA titer serum samples obtained from patients suffering from IE compared with samples from CSD patients. The titers for the IE samples ranged from 1 : 400 to 1 : 6400 as reported previously (Fournier et al., 2002; Jacomo et al., 2002). The main problem with IFA is that cross-reactive antibodies between Bartonella species, especially B. henselae and B. quintana, prevent the identification of the bacteria at the species level (La Scola & Raoult 1996) (Table 1). Only serologically

sophisticated methods such Metalloexopeptidase as Western blots with cross-adsorption studies may help to eventually identify the causative agent at the species level (Houpikian & Raoult, 2003). Several diagnostic assays based on whole-cell detection have been used in clinics (Giladi et al., 2001; Herremans et al., 2007, 2009; Vermeulen et al., 2007). According to ELISA assays for B. henselae infection, recombinant proteins rGroES, rRplL, rBepA and rGroEL yielded high sensitivity >70%, but low specificity ≤59% (Table 3). Only the B. henselae r17-kDa protein has high sensitivity and specificity (Loa et al., 2006; McCool et al., 2008; Hoey et al., 2009) (Table 3). The application of an immunoproteomic strategy to identify the antigenic proteins associated with diseases has been used recently for B. quintana (Boonjakuakul et al., 2007) and B.