Exploring the Link Between Infections and Primary Osteoarthritis: A Next-Generation Metagenomic Sequencing Approach

Abstract

This prospective pilot study examined the association between microorganisms and knee osteoarthritis by identifying pathogens in the synovial membrane, synovial fluid, and blood samples from two patients with primary bilateral knee osteoarthritis, using metagenomic next-generation sequencing (mNGS). Intraoperatively, during routine knee arthroplasty procedures, we collected the following 12 samples from each patient: two synovial membrane samples, two synovial fluid samples, and two venous blood samples. After DNA isolation and library construction, each sample was subjected to deep whole-genome sequencing using the DNBSEQT17 platform with the read length PE150 as the default. Metagenomic sequencing data were mapped to the NCBI NT database to determine species abundance. The predominant species in all samples tested were classified under the Enterobacterales order, the most abundant being Yersinia enterocolitica. The second and third most common microorganisms detected were Escherichia coli and autotrophic, Gram-negative bacteria Synechococcus sp., which is a bioaerosol component, indicating a risk of inhalation of the toxic metabolites of this latter microorganism. This article provides an initial exploration of mNGS use to study the etiopathogenetic mechanisms of knee osteoarthritis (OA). While our analysis identified bacterial DNA, particularly from Yersinia, further cross-sectional studies in larger populations with and without OA are needed to determine the role of these agents in OA pathogenesis.

Keywords: Escherichia coli; Synechococcus; Yersinia enterocolitica; primary knee osteoarthritis.

Figure 1

The species abundance columnar stack graph. The ordinate represents the relative abundance of the species; the abscissa represents the samples taken from 2 patients (patient A and patient B; b represents the blood sample, f—synovial fluid, t—tissue, synovial membrane). Two samples (1 and 2) of the same biological material were taken from each patient. The column color represents the taxon of the species.

Figure 2

Species abundance heatmap. Each row represents a species, and each column represents a patient (A and B). The top 30 species of microorganisms with the highest abundance are shown in the diagram. The species clustering tree is on the left. The closer the branches of the clustering tree, the more similar they are. The square’s color represents the species’ abundance in the samples. The scale bar (to the heatmap’s right) shows a particular species’ abundance.

Figure 3

The UpSetR diagram shows similarities and differences in species composition identified in samples taken from both patients. The bar chart on the left represents the number of all species in specimens from patients A and B. The bar chart on the right represents the number of species in specimens unique to individual patients (B; the first bar, A; the second bar) and the number of species in specimens common to both patients. Samples were analyzed using relative abundance filtering/mean abundance.

Figure 4

STAMP expanded histogram showing analysis of differences between patients in terms of species abundance. In the histogram on the left, the ordinate represents different species, the abscissa represents average group abundance (%), and the color of the columns indicates patients (A and B). The dots in the scatter plot on the right indicate test results with significant differences (the p-value), and FDR is the false discovery rate. The middle area shows the 95% confidence interval for the statistical test of the difference in abundance between the two patients. The position of the dots indicates the average value of the abundance differences, the color corresponds to that of the group with higher abundance, and the boundary of the line segment connected by dots represents the confidence limit.

Figure 5

Species alpha and beta diversity boxplots. Each boxplot represents a diversity index. The abscissa and different boxes indicate groups of samples; the ordinates indicate the index values. The upper and lower edges of the box represent the first quartile and lower quartile. The horizontal line within the box represents the median of distances, and the ends of the straight line above and below the box represent the maximum and minimum distances, respectively; p < 0.05 indicates a significant difference between the groups of samples; * indicates p < 0.05; ns; indicates p > 0.05). (a) Species alpha diversity boxplot. The Kruskal–Wallis H test was used to determine the variability between groups of samples. No substantial differences in Chao1, Simpson and Shannon indices were observed. (b) Species alpha diversity boxplot. The Wilcoxon test was used to determine the variability between samples taken from individual patients (A and B); comparison between patients A and B showed substantial differences in Simpson and Shannon indices. This indicates a significant diversity of microorganisms and the richness and uniformness of samples taken from patient B. (c) Species beta diversity boxplot. The Kruskal–Wallis H test was used to determine the distance between samples. Beta diversity results indicate a difference in species composition between blood samples and synovial fluid samples and between fluid samples and synovial tissue samples. (d) Species beta diversity boxplot. The Wilcoxon test was used to determine variability between samples taken from individual patients (A and B). No significant difference in beta indices between patients was observed.

Figure 6

Sample similarity heatmap. Each row and column represents a sample (12 samples), and the color of the square represents the similarity index (Bray–Curtis anisotropy index) value or distance between two samples. Magenta reflects a smaller dissimilarity coefficient, and cyan represents a larger dissimilarity coefficient (the minimum similarity) in the heatmap graph. The left and top are sample cluster maps; the closer the branches, the more similar the samples.

Figure 7

Pre-operative radiographs of the knees of patient A and patient B. Key: 1. Axial AP view of lower limbs, 2. AP view of left knee, 3. Lateral view of the left knee.