We captured bats by using mist nets in Busia and Taita-Taveta counties in Kenya during 2019. In brief, we captured, euthanized, and dissected bats and collected tissue samples. We screened the samples for BOMV by using reverse transcription PCR (RT-PCR) as described previously (8). We placed bat carcasses that had remaining tissues immediately into 10% buffered formalin, transferred them to 80% ethanol after 24–48 h, and shipped those to the University of Helsinki, Finland, for histopathologic assessment. We analyzed 3 BOMV RT-PCR–positive and 6 BOMV RT-PCR–negative M. condylurus bats; the BOMV-negative bats were used as negative controls (8).

We trimmed and processed formalin-fixed tissue samples from the 9 bats for histologic evaluation. We routinely stained tissue sections with hematoxylin and eosin. After heat-induced antigen retrieval, we also immunolabeled sections from each tissue sample by using polyclonal rabbit serum against EBOV matrix protein VP40 (11) (Appendix). No BOMV isolates were available for this study, which necessitated finding a suitable, cross-reactive surrogate virus. EBOV is the most studied member in the ebolavirus family, and EBOV antibodies are readily available. EBOV and BOMV have ≈75% amino acid identity match within the VP40 protein (Appendix Figure); therefore, we considered EBOV VP40 to be a suitable antigen to detect BOMV VP40.

Figure 1

Figure 1. Representative tissue section from the lung of a bat in study of histopathologic changes in Mops condylurusbats naturally infected with Bombali virus, Kenya. We stained lung...

We observed mild perivascular and interstitial mononuclear infiltrates in all 3 BOMV RT-PCR–positive M. condylurus bat lung samples stained with hematoxylin and eosin (Table; Figure 1). We found similar nonspecific mononuclear infiltrates in lung tissue from 4/6 BOMV RT-PCR–negative bats. Overall, we did not observe substantial histopathologic changes in the lungs of any bats that suggested severe acute BOMV disease. We found that both BOMV RT-PCR–positive and RT-PCR–negative bats had mixed acute and chronic inflammatory lesions in other internal organs. Those lesions likely reflect natural disease in wild bats, and we interpreted those findings as incidental.

We did not find an association between immunologic detection of EBOV VP40 antigen, used as a surrogate for BOMV, and overt histopathologic findings in BOMV RT-PCR–positive M. condylurus bats. We previously showed that BOMV RT-PCR–positive bats developed antibodies reactive against ebolavirus proteins from Zaire ebolavirus–infected cells (6). In this study, both BOMV RT-PCR–positive and RT-PCR–negative bats displayed only mild mononuclear infiltrates in lungs and exhibited mild inflammatory changes in other tissues. Those changes indicate that BOMV infection is likely well-tolerated by M. condylurus bats, which is consistent with expectations of a reservoir host (12,13). However, our findings should be interpreted with caution. For example, we cannot exclude the possibility that the BOMV RT-PCR–positive M. condylurus bats we captured had less severe pathology than infected wild conspecifics or that negative control bats might have been previously infected but cleared any residual viral antigens.

Figure 2

Figure 2. Representative bat lung tissue showing Ebola virus (EBOV) cytoplasmic granules in study of histopathologic changes in Mops condylurusbats naturally infected with Bombali virus, Kenya. We labeled...

Positive immunolabeling with EBOV VP40 polyclonal antibody was observed in lungs from 2 of 3 BOMV RT-PCR–positive bats (Figure 2) but not in other tissues or in any tissues from BOMV RT-PCR–negative bats. Immunopositive-labeling occurred as granular cytoplasmic aggregates in interstitial macrophages. In contrast to their usual role in protective immunity, macrophages are directly implicated in the pathogenesis of several viruses as either viral repositories or sites of active replication and dissemination (14). In humans, pathogenic filoviruses rely heavily on macrophages, which are major replication sites (15). Our findings suggest that macrophages might provide similar chronic reservoirs for BOMV infections in bats. However, without data showing active replication, we cannot rule out that our findings in macrophages have a different interpretation.

Positive immunolabeling in bat lungs was consistent with previous RT-PCR results (8). Lungs are the most frequent tissue infected by BOMV and have markedly higher viral loads than other tissues (6–8). Of note, 1 lung sample, 2 spleen samples, and 1 liver sample that were BOMV RT-PCR–positive (8) had no detectible EBOV VP40 antigen. Those results are most likely because immunohistochemistry has lower sensitivity than PCR, and a non–BOMV-specific antigen was used.

We propose that the presence of BOMV in macrophages and absence of acute pathology typical of ebolaviruses (necrosis, thrombosis, hemorrhage, and edema) support bats as a chronic subclinical BOMV reservoir with the potential for infection and sporadic virus excretion. However, the small number of samples, absence of recovery of infectious virus, and lack of longitudinal data make interpretation of our results difficult. As for all wildlife-borne ebolaviruses, no direct evidence currently exists regarding the transmission route of BOMV between susceptible hosts. Our findings suggest the mode of transmission involves activation of replication in either macrophages or other cells and virus dispersal through saliva droplets or feces. The identification of BOMV infection in lungs and occasional virus presence in intestines further supports this hypothesis (6–8).

In summary, we showed EBOV-VP40 immunolabeling, acting as a cross-reactive surrogate for BOMV antigen, in pulmonary macrophages of wild BOMV RT-PCR–positive M. condylurus bats and an absence of substantial concurrent pathologic findings. Our results support M. condylurus bats as reservoir hosts for ebolaviruses and highlight the importance of bat surveillance to understand and mitigate potential emerging disease risks.