Breathing can be dangerous: Opportunistic fungal pathogens and the diverse community of the small mammal lung mycobiome

Abstract

Human lung mycobiome studies typically sample bronchoalveolar lavage or sputum, potentially overlooking fungi embedded in tissues. Employing ultra-frozen lung tissues from biorepositories, we obtained fungal ribosomal RNA ITS2 sequences from 199 small mammals across 39 species. We documented diverse fungi, including common environmental fungi such as Penicillium and Aspergillus, associates of the human mycobiome such as Malassezia and Candida, and others specifically adapted for lungs (Coccidioides, Blastomyces, and Pneumocystis). Pneumocystis sequences were detected in 83% of the samples and generally exhibited phylogenetic congruence with hosts. Among sequences from diverse opportunistic pathogens in the Onygenales, species of Coccidioides occurred in 12% of samples and species of Blastomyces in 85% of samples. Coccidioides sequences occurred in 14 mammalian species. The presence of neither Coccidioides nor Aspergillus fumigatus correlated with substantial shifts in the overall mycobiome, although there was some indication that fungal communities might be influenced by high levels of A. fumigatus. Although members of the Onygenales were common in lung samples (92%), they are not common in environmental surveys. Our results indicate that Pneumocystis and certain Onygenales are common commensal members of the lung mycobiome. These results provide new insights into the biology of lung-inhabiting fungi and flag small mammals as potential reservoirs for emerging fungal pathogens.

Keywords: Aspergillus fumigatus; Blastomyces; Coccidioides; Onygenales; Pneumocystis; Valley fever.

Figure 1

Relationship between frequency (percent of lung samples) and total reads for 762 OTUs. Taxonomic orders are given for those OTUs with a frequency of ≥50% plus two OTUs with high read counts, OTU1 (Pneumocystidales) and OTU12 (Thelephorales).

Figure 2

RaxML maximum likelihood phylogenies for Onygenales (A) and Pneumocystidales (B). Both trees have a mid-point root and bootstrap values (1000 replicates) are shown for branches with greater than 60% support. (A) Onygenalean fungi obtained through culturing (blue) and Illumina ITS2 sequencing (red) of lung tissues. Host species for which lung fungal cultures were obtained are in parentheses. Sequences in black type were acquired from GenBank. (B) Pneumocystidales OTUs were obtained from Illumina ITS2 sequencing. Each box to the right of an OTU represents one sample with >1,000 sequence reads for a given OTU. Boxes are color coded by genus with warm colors (red, orange, yellow) within Heteromyidae, cool colors (green, blue, purple) within Cricetidae, and black for Sciuridae. Branches with species names and accession numbers represent sequences obtained from GenBank. These trees are presented to illustrate the diversity of fungi from these two groups that were obtained from lung samples, together with associations among cultured fungi, ITS sequences, and mammalian species. Based on ITS sequences alone we were not able to resolve deeper evolutionary relationships unambiguously.

Figure 3

NMDS ordination of small-mammal samples from Kern County California only (A), Sierra County New Mexico only (B), and all 199 small-mammalian lung samples. (A) Samples collected within a 50 km radius in Kern County California suggest a difference in the lung mycobiome communities of species of Dipodomys and Perognathus (R²= 0.22905, p=0.001). (B) Samples within a 50 km radius in Sierra County New Mexico suggest a clustering of the lung mycobiome community within species of Chaetodipus but not differences from Neotoma or Peromyscus species (R²=0.07634, p=0.188). (C) Analysis of all 199 samples reveals no fungal community differences between those samples with Coccidioides and those without (R²=0.00539, p=0.351). (D) Analysis of all 199 samples demonstrated potential differences but also substantial overlap between those samples with OTU3 (top BLAST hits to strains of Aspergillus fumigatus) and those without (R²= 0.00883, p=0.005). A degree of clustering of samples with higher numbers of OTU3 reads (>10,000) was observed, however (indicated by the smaller turquoise oval). For the analysis shown, A. fumigatus sequence reads were removed to prevent the effect of abundant OTU3 on Bray-Curtis dissimilarity values. Results obtained when A. fumigatus reads were included in the analyses were in substance the same (R²= 0.01037, p=0.009, results not presented). Additional NMDS ordination results are presented for collection year, location by state, and host family and genus in Supplementary Figure 3 .

Figure 4

Bar plot of Coccidioides positive and negative samples by species (A) and by genus (B).

Figure 5

Sampling of small-mammalian lungs across the southwestern U.S. detect Coccidioides in wild animals. The number of samples sequenced per county is depicted in blue. Coccidioides positive trapping locations are shown with an orange dot with increasing size based on the number of positive samples in that specific location. GPS locations less than 1 km apart were merged. The insert indicates the endemic range for the genus Coccidioides, with darker shades of green indicating higher levels of incidence, as recognized by the U.S. Centers for Disease Control and Prevention (see Gorris et al., 2019).