L929 mouse conditioned medium containing murine macrophage colony stimulating factor-1 (M-CSF) was successfully used to induce M. myotis macrophage growth. Even though a smaller proportion of M. myotis bone marrow cells were alive after seven days of differentiation compared to mouse bone marrow, M. myotis bone marrow-derived cells were indistinguishable in terms of size and shape to their murine counterparts. Treatment of M. myotis and M. musculus bone marrow derived macrophages with synthetic TLR ligands, LPS and Poly(I:C), resulted in the induction of pro- and anti-inflammatory cytokine genes. Both mouse and M. myotis macrophages mounted a potent antiviral and proinflammatory response, as demonstrated by induction of interferon - β (IFN-β), interleukin-1β (Il-1β) and tumor necrosis factor (TNF) by both treatments (Fig. 1A–C). Myotis myotis macrophages also elicited a rapid anti-inflammatory response, showing significantly higher and persistent induction of interleukin-10 (Il-10) compared to mouse. Notably, the mouse macrophages elicited a higher induction of Il-1β in response to potent TLR ligand LPS. Il-1β is an acute-phase signalling proinflammatory cytokine (Guarda et al., 2011). The ratio between fold induction of anti-inflammatory Il-10 and proinflammatory TNF or Il-1β was higher in M. myotis compared to mouse, especially after 24 hr of the LPS or Poly(I:C) treatment (Fig. 1E–F). Il-10/TNF mRNA induction ratios derived from previous literature studies of pig, human and mouse macrophages treated with LPS for 24 hr are also presented (Fig. 1G). Moreover, mouse, but not M. myotis macrophages, demonstrated increased production of NO, in response to LPS treatment (Fig. 1H). Consistently, qPCR analyses also showed strong induction of iNOS (inducible nitric oxide synthase) expression 24 hr after immune challenge only in the mouse macrophages (∆∆Ct = -6.87 ± 0.56 SEM for LPS; -3.93 ± 3.99 for Poly(I:C)), but not in M. myotis (∆∆Ct = -0.54 ± 0.56 for LPS; -0.28 ± 0.61 for Poly(I:C)).

We performed phylogenomic analyses and positive selection tests in 12 representative immune genes involved in sensing viral infection and activation of the innate antiviral and proinflammatory signalling cascade in 39 mammalian taxa spanning basal eutherian divergences (including six bat species — Supplementary Fig. S1). The function of proteins encoded by tested genes is summarised in Supplementary Table S3 and their position in the immune signalling pathway is highlighted in Fig. 2. Tests for adaptive evolution at codon sites across eutherian mammals revealed evidence of significant positive selection acting on all investigated genes except Il-10 and TBK1. At least one BEB site was reported for all positively selected genes with exception of TNF (Supplementary Table S4A). The Branch-Site test (A versus A1 model) revealed positive selection within the ancestral bat lineage acting on proinflammatory Il-1β (Supplementary Table S4B). We also detected vespertilionid-specific, Myotis -specific or Pteropus -specific non-synonymous amino acid substitutions in RIG-I, Il-1β, Il-18, NLRP3, STING and CASP 1 in the positively selected mammalian BEB sites (Supplementary Table S4C and Fig. 3).