Abstract
Brucellosis is a widespread zoonotic disease and a potential agent of bioterrorism. Current serological assays for human brucellosis primarily detect agglutinating anti-LPS antibodies. To explore the broader landscape of antibody responses following B. melitensis infection, a protein microarray was developed containing 1,406 predicted B. melitensis proteins. This array was used to probe sera from experimentally infected goats and naturally infected humans from Peru. The study identified 18 antigens differentially recognized in infected goats and 13 serodiagnostic antigens in acute human patients. Notably, only two serodiagnostic antigens overlapped between the two species. A validation line blot using human serodiagnostic antigens confirmed the microarray's accuracy. These findings indicate that the immune response in a natural reservoir host differs fundamentally from that in an accidental human host, with significant implications for diagnostic and vaccine development.
Introduction
Brucella melitensis is the primary cause of human brucellosis, a major global zoonosis typically transmitted through the consumption of unpasteurized dairy products or direct contact with infected livestock. Beyond its public health impact, it is classified as a potential bioterrorism agent. Despite its prevalence, diagnostic capabilities remain limited. Traditional serological tests rely on the detection of antibodies against lipopolysaccharide (LPS). However, LPS-based assays often fail to distinguish between different stages of infection (acute vs. chronic) or between vaccinated and naturally infected animals. To overcome these limitations, researchers are shifting focus toward protein-based diagnostics. This study utilizes a systems biology approach to profile the antibody response across nearly half of the B. melitensis proteome in both humans and goats.
Methodology
The researchers constructed a high-throughput protein microarray representing approximately 44% of the B. melitensis genome (1,406 open reading frames). Proteins were produced using an E. coli-based in vitro transcription and translation system and printed onto nitrocellulose-coated glass slides.
The arrays were probed with serum samples from two distinct cohorts:
- Human Cohort: 42 culture-positive acute patients, 18 symptomatic but culture-negative individuals, and 57 healthy controls from Peru and the United States.
- Goat Cohort: Goats experimentally infected with the B. melitensis 16M strain, compared against uninfected and specific pathogen-free (SPF) controls.
Data analysis involved Bayesian regularized t-tests (Cyber-T) to identify significant antigens and Support Vector Machines (SVM) to determine the minimal set of proteins required for high-accuracy diagnosis. Findings were validated using a physical nitrocellulose line blot (immunostrip) containing the top 13 human antigens.
Results and Discussion
The most striking result was the minimal overlap in immune recognition between the two species. While goats recognized 18 diagnostic antigens and humans recognized 13, only two proteins—including the well-known marker Bp26 (BMEI0536)—were common to both. This "host gap" suggests that the immune signature of a natural reservoir host like the goat is not a reliable proxy for the human immune response.
Several factors may contribute to this divergence. The route of infection differs significantly: humans typically ingest the pathogen, whereas the goats in this study were infected via the conjunctival route. Additionally, the metabolic state of the bacteria during infection appears to influence antigenicity. Many identified antigens are involved in nutrient acquisition, such as ABC transporters, suggesting that Brucella is metabolically active and scavenging for resources within host cells.
The study also highlighted the impact of endemicity. Healthy individuals in Peru exhibited higher background antibody levels compared to U.S. controls, underscoring the challenge of developing universal diagnostics in regions with frequent subclinical exposure.
Diagnostic Performance
The protein microarray demonstrated exceptional diagnostic potential. Using the top five human antigens, the SVM model achieved 95% sensitivity and 96% specificity. The simplified immunostrip validation test maintained high performance (94% sensitivity, 89% specificity), proving that these markers are suitable for transition into practical clinical tools.
Conclusion and Future Directions
This large-scale profiling reveals that B. melitensis infection elicits distinct antibody profiles in humans and goats. While Bp26 remains a robust cross-species marker, the overall lack of overlap necessitates host-specific diagnostic strategies. Future work will expand the microarray to cover the full proteome and investigate whether these protein markers can differentiate between chronic and relapsing brucellosis, addressing a critical gap in clinical management.
Summary updated