Tertiary Lymphoid Organs Are Associated With Bacterial Dysbiosis in Chronic Rhinosinusitis

Tertiary lymphoid organs (TLOs) are increasingly recognized as significant entities in many inflammatory processes [1]. The phenomenon of de novo and organized lymphocytosis in normally immunodeplete tissue allows effector cells to be aggregated within the vicinity of antigenic sources, thereby heightening the inflammatory response [2]. Chronic bacterial infections have been shown to induce TLO formation. Recently the presence of TLOs has been described within the mucosa of chronic rhinosinusitis (CRS) patients [3]. This subset of patients demonstrated severe, treatment-resistant disease necessitating multiple surgical interventions [4]. Unlike the distinct disease processes listed above, no single causative mechanism for CRS and subsequent TLO development has been delineated [5]. One potential mechanism is the interaction between the sinonasal microbiome and the underlying immune response, which appears to be central to the CRS paradigm. Research involving human tissue was approved by the Human Research Ethics Committees from both The Queen Elizabeth Hospital and The University of Adelaide. Microbiome swabs and sinonasal tissue were harvested intraoperatively. Samples were collected in a standardized fashion prior to any surgical alteration. Endoscopically guided microbial swabs, nasal polyps and mucosal tissue were taken from the middle meatus. TLOs were confirmed via immunofluorescence using a previous protocol described by Lau et al. [3]. Microbiome analysis was conducted utilizing a standardized protocol described in the International Sinonasal Microbiome Study (ISMS) [6], Bioinformatics was performed utilizing QIIME 2 (version 2018.11) software. “Analysis of compositions of microbiomes” (ANCOM) determined differentially abundant taxa between groups. Alpha (Phylogenetic Diversity) and beta (Weighted UniFrac) diversity metrics were calculated using the scikit-bio python package. Transcriptomic analysis was conducted as per Bassiouni et al. [7]. In summary, transcript quantification was done using Salmon (version 0.12.0) against the GENCODE human transcriptome version 29 (“gencode.v29.transcripts.fa.gz”). Feature/gene-level count estimates were calculated using the R package tximport. DESeq2's likelihood ratio test (LRT) was used to test differentially expressed genes in any of the four study groups simultaneously against a null model and Wald tests was used for pairwise detection of differentially expressed genes between study groups. In total, 117 consecutive patients were recruited. Characteristics of the cohort are described in Supporting Information Table S1. TLOs were identified in 35 diseased patients, with six (35%) in the CRSsNP group and 29 (45%) in the CRSwNP group. No TLOs were observed in the control group (Supporting Information Table S1). No significant differences existed between controls and the two CRS phenotypes (Figure 1A). When classified by TLO status (Figure 1B), Alpha diversity was significantly reduced among the TLO population (2.43 vs. 3.0 in TLO-positive and TLO-negative CRS patients respectively; p = 0.028). Examination of both the TLO-negative and TLO-positive CRS groups revealed a significant reduction in Corynebacterium species in the TLO-positive group (relative abundance 47.16% vs. 8.63% in TLO-negative vs. TLO-positive patients; p < 0.05; ANCOM; Figure 1). An increase in Staphylococcal species was also observed (relative abundance of 10.3% vs. 36.8% in TLO-negative vs. TLO-positive patients; p > 0.05; ANCOM). Twenty-four patients were selected for transcriptomic analysis (six controls; three CRSsNP without TLOs; five CRSwNP without TLOs; 10 CRSwNP with TLOs). Compared to the mean expression, there was a total of 1264 genes upregulated, and 599 genes downregulated, determined by LRTs (Figure 2A). Figure 2C demonstrates similar gene-expression between study cohorts. Gene ontology analysis between these two groups demonstrated a significantly increased expression in leukocyte chemotaxis genes among TLO-positive patients (GO:0030595; p < 0.05). Differentially expressed genes between TLO-positive and TLO-negative are listed in Supporting Information Table S2. A principal component analysis (PCA) reveals that while control patients cluster separately from CRS patients, there is overlap of all the CRS patients irrespective of phenotype or TLO status (Figure 2B). To our knowledge this represents the first investigation into TLO-related microbiomes in any human disease process. Our results demonstrate that, in the presence of TLOs, there is a significant depletion of Corynebacterium species within the sinonasal cavity among CRS patients. The reduction in Corynebacterium relative abundance occurs simultaneously with an increase in Staphylococcal representation. This is the first time that a distinct microbiome dysbiosis has been demonstrated in association with TLOs in any disease model. The significantly reduced microbial diversity results in a reduction of genetic potential in the mucosa-microbe organ. If a number of these microbially provided genes were critical for immunological balance within the sinuses, it could account for the particularly severe disease seen within this subset of patient. To further characterize the inflammatory consequences of TLOs in CRS, a global transcriptomic analysis was performed. When comparing TLO-positive and -negative CRS cohorts, an increase in leukocyte chemotaxis genes was observed in the former. To the best of our knowledge no other studies were identified which examine global transcriptional expression in human tissue specifically investigating TLO status in a disease setting. Interestingly, a PCA of our patients did not find any significant clustering or difference between CRSwNP patients with and without TLOs. As has been hypothesized previously [4], this could suggest that perhaps all CRS patients have the potential to develop TLOs, with only a subset going on to do so. Alternatively, TLOs could simply represent a marker of severe disease among the CRS population. Several limitations of note existed within this study. It is important to note that the present study was unable to analyze microbial organisms to a species level due to current limitations in short-read 16S gene sequencing technology. The findings are also purely associative in nature. Pathophysiological correlation between microbial dysbiosis and TLO formation needs further investigation. In conclusion we have highlighted a novel dysbiosis among CRS patient who harbor TLOs, represented by a significant depletion in Corynebacterium species and overabundance of Staphylococcal species. This is the first time that such a dysbiosis has been described for TLOs in any disease process. Sathish Paramasivan, Mahnaz Ramezanpour and Clare Cooksley were responsible for data collection and laboratory experimental work. Sathish Paramasivan, Ahmed Bassiouni, Nicholas A. Bokulich, and J. Gregory Caporasso performed statistical analysis. J. Gregory Caporasso, Peter-John Wormald, Alkis J. Psaltis, and Sarah Vreugde were responsible for experimental design, data collection, and supervision. All the authors contributed to manuscript creation and revision. This work was supported by a Garnett Passe and Rodney Williams Memorial Foundation Academic Surgeon Scientist Fellowship to Sathish Paramasivan. Open access publishing facilitated by The University of Adelaide, as part of the Wiley - The University of Adelaide agreement via the Council of Australian University Librarians. Ahmed Bassiouni, Nicholas A. Bokulich, Mahnaz Ramezanpour, Clare Cooksley, and Sarah Vreugde: No conflicts of interest relevant to this study. Peter-John Wormald: Receives royalties from Medtronic, Integra, and Scopis, and is a consultant for NeilMed. These are not relevant to this study. Alkis J. Psaltis: Consultant for Aerin Devices and ENT technologies and is on the speakers’ bureau for Smith and Nephew. Received consultancy fees from Lyra Therapeutics. These are not relevant to this study. Sathish Paramasivan: Supported by a Garnett Passe and Rodney Williams Memorial Foundation Academic Surgeon Scientist Research Scholarship and Fellowship. J. Gregory Caporaso: No conflicts of interest to declare. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.