The Microbiome of Temporal Arteries

Abstract

Objective: A role for microorganisms in giant cell arteritis (GCA) has long been suspected. We describe the microbiomes of temporal arteries from patients with GCA and controls.

Methods: Temporal artery biopsies from patients suspected to have GCA were collected under aseptic conditions and snap-frozen. Fluorescence in situ hybridization (FISH) and long-read 16S rRNA-gene sequencing was used to examine microbiomes of temporal arteries. Taxonomic classification of bacterial sequences was performed to the genus level and relative abundances were calculated. Microbiome differential abundances were analyzed by principal coordinate analysis (PCoA) with comparative Unifrac distances and predicted functional profiling using PICRUSt.

Results: Forty-seven patients, including 9 with biopsy-positive GCA, 15 with biopsy-negative GCA and 23 controls without GCA, were enrolled. FISH for bacterial DNA revealed signal in the arterial media. Beta, but not alpha, diversity differed between GCA and control temporal arteries (P = 0.042). Importantly, there were no significant differences between biopsy-positive and biopsy-negative GCA (P > 0.99). The largest differential abundances seen between GCA and non-GCA temporal arteries included Proteobacteria (P), Bifidobacterium (g), Parasutterella (g), and Granulicatella (g) [Log 2-fold change ≥ 4].

Conclusion: Temporal arteries are not sterile, but rather are inhabited by a community of bacteria. We have demonstrated that there are microbiomic differences between GCA and non-GCA temporal arteries, but not between biopsy-positive and biopsy-negative GCA.

Keywords: giant cell arteritis; microbiome; vasculitis.

Figure 1.

Distribution of bacterial DNA in temporal arteries. Tissue sections were probed with fluorescently labeled oligonucleotide probes against bacterial DNA (green). Sections were counterstained with DAPI (blue) and Concanavalin A (red) to delineate nuclei and glycoproteins, respectively. Sections were scanned by confocal microscopy. In a control temporal artery (A), bacterial DNA is scattered throughout the media, with select examples highlighted by green arrows. Notably, no/negligible bacterial DNA staining is apparent in the lumen or intima (A, bar graph). The green channel emitted from the internal elastic lamina is a result of autofluorescence. In a temporal artery with histopathological evidence of GCA (B), bacterial DNA is scattered throughout the media and at a higher mean intensity than control (bar graph). Arterial layers are more disorganized and less distinct compared to a control temporal artery, as evidenced by weak autofluorescence and less distinct internal elastic lamina. There is an absence of bacterial DNA at the external edge of a GCA-involved temporal artery specimen (C, bar graph).

Figure 2.

Microbiomes from TAs with biopsy-positive and biopsy-negative GCA cluster together but differently from those from control patients. Principal component analysis (PCoA) of TA microbiomes. (A) β-diversity (not α, data not shown), differs between GCA and control groups (P = 0.042). (B) There were no statistically significant differences between TA microbiomes in those with biopsy-positive GCA vs those with biopsy-negative/clinically positive GCA (P > 0.99).

Figure 3.

Most differentially abundant taxa in temporal artery biopsies from patients with GCA and from control patients. (A) Bar blot representation from DESeq2 showing the most over-represented (+) and under-represented (-) phyla in TAs from patients with GCA compared to TAs from controls. (B) Bar blot representation from DESeq2 showing the most over-represented (+) and under-represented (-) genera in TAs from patients with GCA compared to TAs from controls. (C) Heat map of bacterial communities in TA with GCA (“inflammatory” blue bar) compared to those without GCA (“noninflammatory” pink bar) based on the top dominant OTUs. Columns and rows represent samples and dominant OTUs, respectively. Row names on the right of the heat map include Green Genes ID followed by family and genus.

Figure 4.

Predicted functional pathways differentially represented in GCA TA compared to control (non-GCA) TA. Representation of PICRUSt DESeq2 analysis yielding relatively under-represented functional pathways in (A) GCA TA compared to control TA, (B) biopsy-positive GCA TA compared to control TA and (C) biopsy-negative GCA TA versus control TA.