Proteomic signatures of SARS-CoV-2 susceptibility in Mexican free-tailed bats and their application to viral surveillance

The increasing emergence of virulent pathogens necessitates novel approaches to predict and manage infectious disease risks. The importance of integrating observational and experimental approaches to studying host–pathogen interactions has long been recognized, as captive studies can mechanistically test hypotheses derived from field studies and identify causal factors shaping host susceptibility or tolerance of infection. However, captive experiments can also determine biomarkers of infection outcomes that could improve later interpretation of field data and identify at-risk hosts in wild populations. Such work could be especially useful in preempting or managing risks of pathogen spillover or spillback. SARS-CoV-2 emerged in humans in late 2019 and was rapidly followed by spillback into naïve wildlife, leading to both mortality events and novel enzootic cycles. Of special concern is whether SARS-CoV-2 could establish in bats in the Americas, given that sarbecoviruses coevolved with rhinolophid bats in the Eastern Hemisphere, and as coronavirus infection may exacerbate effects of white-nose syndrome. Here, we leverage residual plasma samples from a previous SARS-CoV-2 challenge study of Mexican free-tailed bats (Tadarida brasiliensis) to identify candidate protein biomarkers of susceptibility and test whether these can predict coronavirus risks in wild bats. We generated plasma proteomes from captive (n = 20; four resistant, five susceptible, 11 unchallenged) and wild (n = 15) bats using the S-Trap method and LC-MS/MS, identifying 475 proteins using data-independent acquisition and a species-specific genome annotation generated by the Bat1K Project. Receiver operator characteristic curves identified 27 potential biomarkers of SARS-CoV-2 susceptibility (AUC ≥ 0.8), and subsequent enrichment analyses of these proteins suggested downregulation of blood clotting and upregulation of complement activation and humoral immunity in susceptible bats. We then mined plasma proteomes from wild bats (sampled in 2022 from Bracken Cave Preserve, the largest known Mexican free-tailed bat population) to show that all candidate biomarkers were present in this population, with coefficients of variation ranging from 16 to 150% per protein. We detected coronaviruses in 20% of wild bats, with two cases of potential SARS-CoV-2 spillback. We demonstrate that at least four of these candidate susceptibility biomarkers classified bats with and without coronavirus infection in the wild. Our results inform the possible immune strategies underlying SARS-CoV-2 susceptibility in bats and give a preliminary example of how captive challenge studies can be coupled with field studies to inform zoonotic and conservation risks.