Genetic Diversity of Bartonella spp. in Cave-Dwelling Bats and Bat Flies, Costa Rica, 2018

To determine Bartonella spp. dynamics, we sampled bats and bat flies across 15 roosts in Costa Rica. PCR indicated prevalence of 10.7% in bats and 29.0% in ectoparasite pools. Phylogenetic analysis of 8 sequences from bats and 5 from bat fly pools revealed 11 distinct genetic variants, including 2 potentially new genotypes.

B artonella, the causative agent of bartonellosis, is a genus of gram-negative bacteria. Bartonellosis causes a range of symptoms from severe to life-threatening (e.g., endocarditis and meningitis). Clinical syndromes from Bartonella infections include trench fever (B. quintana), cat scratch disease (B. henselae), and Carrion's disease (B. bacilliformis) (1). Bats (Order Chiroptera) and their blood- 1 These authors contributed equally to this article. feeding ectoparasitic bat fl ies (Superfamily Hippoboscoidea) host a diversity of Bartonella species, awakening interest in their potential role as natural reservoirs for this pathogen (2,3). To learn more about this interplay, we examined the genetic diversity and geographic sharing of Bartonella spp. in diverse assemblages of bats and bat fl ies across Costa Rica.
In 2018, we nonlethally sampled 321 bats (18 species) by using hand nets, mist nets, and harp traps across 15 roosts throughout Costa Rica (Appendix Figure Each sequence is labeled with its GenBank accession number, the organism on which it was detected, and the country of origin. For species in this study, we included the specifi c site (accession numbers in Appendix To create global phylogenies, we trimmed obtained consensus sequences to 768 bp and aligned them to 45 genetic sequences: 28 from known Bartonella species, 12 Bartonella sequences from bats and bat flies in Costa Rica (8), 4 sequences from bats in Guatemala (9), and 1 sequence from Mexico (3). We used B. tamiae and Brucella melitensis as outgroups to root the tree. We created the alignment by using the multiple alignment program MAFFT (https://mafft. cbrc.jp/alignment/software), manually checked in MEGA X (https://www.megasoftware.net), and further refined with alignment refinement tool Gblocks version 0.91b (http://molevol.cmima.csic.es/castresana/Gblocks/Gblocks_documentation.html). We constructed the global phylogenetic tree by using Bayesian Markov chain Monte Carlo analyses (MrBayes 2.2.4, https://www.geneious.com) with 1 million generations and a burn-in fraction of 25% and determined the parameters for the nucleotide changes (MEGA X).
Bartonella prevalence from all samples, determined by PCR, was 14.3% (45/314), 10.7% (27/252) for bats and 29.0% (18/62) for ectoparasite pools (Table). Bartonella seems to be widespread and diverse in bats and bat flies in Costa Rica, where 6 of the 16 bat species and 9 of the 23 bat fly species were positive for the bacterium. Because of sequence quality, we included only 8 Bartonella sequences from bats and 6 from bat fly pools in phylogenetic analyses, which revealed 11 genetic variants, including 2 potentially new genotypes (93.2% similarity value; Figure; Appendix Table). These 11 genetic variants clustered into 9 clades of 96.0%-99.2% similarity.
Our results suggest that within Costa Rica variants are shared between bats and their flies in different parts of the country and in different years. For example, Bartonella sequences from Emus Cave (GenBank accession no. MW115627) and Túnel Arenal (GenBank accession no. MW115628) at opposite ends of the country (clade V; Appendix Figure) clustered together with sequences from a study conducted in Costa Rica in 2015 (8). In addition, Bartonella sequences from our study clustered with previously identified sequences from bats and bat flies from Guatemala (9) and Mexico (3), suggesting wide geographic distribution. We also found a high level of diversity of Bartonella variants within caves and species (Figure). For example, Bartonella sequences from different bats (of same and different species) in Emus Cave clustered in 4 distinct clades. In addition, Carollia perspicillata bats, the most sampled species in our study, carried Bartonella with sequences from 6 distinct clades. This finding suggests that >1 Bartonella strain is circulating within bat species, even within the same cave.
When assessing spillover risk to humans and domestic animals, we found that the Bartonella sequences we detected did not cluster with Bartonella species known to cause infection in humans and other animals and did not significantly overlap with sequences from any globally identified species (Figure). To fully assess potential for Bartonella spillover from bat and bat fly species to other animals and humans, further analyses should be conducted.
In conclusion, we found Bartonella species to be diverse, prevalent, and potentially widely shared among species of bats and bat flies in Costa Rica and Mesoamerica. We expanded existing scientific knowledge on the prevalence and diversity of Bartonella in bats and bat flies in Costa Rica by including species that were not previously tested and described as positive by PCR for these bacteria. We also described 2 new Bartonella genotypes through phylogenetic analysis. Information about the dynamics of Bartonella in its natural hosts can be used to predict and avert further Bartonella emergence.
To the Editor: Recently, Liu et al. (1) described the predictors of nonseroconversion after severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection (36.1% of cases), where nonresponders had significant higher cycle threshold (C t ) and were younger. Although a recent study showed that 1 dose of mRNA vaccine is sufficiently effective in previously infected persons (2), Reynolds et al. reported a previously infected vaccinee who never seroconverted (3). We report the case of a previously infected vaccinee who did not seroconvert and was subsequently reinfected.
In April 2020, a 55-year-old female nursing manager had mild SARS-CoV-2 pneumonia diagnosed that did not require admission, confirmed by weakly positive genes E and RNA-dependent RNA polymerase PCR testing (both C t >33, near the limit of detection using homemade techniques). Concomitantly, her husband experienced symptoms and also tested positive, supporting that the woman's case was not a false-positive. One month later, SARS-CoV-2 serology revealed no detectable antibodies to nucleocapsid or spike (S) proteins.
Despite a low risk for SARS-CoV-2 reinfection in a healthcare worker without underlying conditions (4) and having been vaccinated with 1 dose of mRNA BNT162b2 (Pfizer-BioNTech, https://www.pfizer. com) in April 2021, as recommended for previously infected persons, the woman was reinfected in September 2021 by the Delta variant. She had mild symptoms and a high estimated viral load (C t 26 for genes E and N2). Serologic testing at the time of the first detection of reinfection revealed a relatively low titer of 20 binding antibody units/mL of S antibodies, which then increased to 243 BAU/mL 1 month after reinfection. Testing to rule out immune deficiency (serum protein electrophoresis, quantitative immunoglobulin assay, and assessment for complement deficiency) detected no abnormalities.
Our findings support a 2-dose vaccine policy for previously infected persons, as applied in the United States. This cautious approach is even more relevant because neutralizing antibody titers are substantially reduced in patients infected with the Delta variant (5) and in light of efforts to promote a third dose of vaccine, to ensure a stable antibody level over time in