Intervertebral disc microbiome in Modic changes: Lack of result replication underscores the need for a consensus in low-biomass microbiome analysis

Tamara Mengis; Natalia Zajac; Laura Bernhard; Irina Heggli; Nick Herger; Jan Devan; Roy Marcus; Florian Brunner; Christoph Laux; Mazda Farshad; Oliver Distler; Stefan Dudli

doi:10.1002/jsp2.1330

Intervertebral disc microbiome in Modic changes: Lack of result replication underscores the need for a consensus in low-biomass microbiome analysis

JOR Spine. 2024 Apr 4;7(2):e1330. doi: 10.1002/jsp2.1330. eCollection 2024 Jun.

Authors

Tamara Mengis^{1

2}, Natalia Zajac³, Laura Bernhard^{1

2}, Irina Heggli^{1

2}, Nick Herger^{1

2}, Jan Devan^{1

2}, Roy Marcus⁴, Florian Brunner², Christoph Laux⁵, Mazda Farshad⁵, Oliver Distler^{1

2}, Stefan Dudli^{1

2}

Affiliations

¹ Center of Experimental Rheumatology, Department of Rheumatology University Hospital, University of Zurich Zurich Switzerland.
² Department of Physical Medicine and Rheumatology, Balgrist University Hospital University of Zurich Zurich Switzerland.
³ Functional Genomics Center Zurich University and ETH Zurich Zurich Switzerland.
⁴ Department of Radiology, Balgrist University Hospital University of Zurich Zurich Switzerland.
⁵ Department of Orthopedics, Balgrist University Hospital University of Zurich Zurich Switzerland.

Abstract

Introduction: The emerging field of the disc microbiome challenges traditional views of disc sterility, which opens new avenues for novel clinical insights. However, the lack of methodological consensus in disc microbiome studies introduces discrepancies. The aims of this study were to (1) compare the disc microbiome of non-Modic (nonMC), Modic type 1 change (MC1), and MC2 discs to findings from prior disc microbiome studies, and (2) investigate if discrepancies to prior studies can be explained with bioinformatic variations.

Methods: Sequencing of 16S rRNA in 70 discs (24 nonMC, 25 MC1, and 21 MC2) for microbiome profiling. The experimental setup included buffer contamination controls and was performed under aseptic conditions. Methodology and results were contrasted with previous disc microbiome studies. Critical bioinformatic steps that were different in our best-practice approach and previous disc microbiome studies (taxonomic lineage assignment, prevalence cut-off) were varied and their effect on results were compared.

Results: There was limited overlap of results with a previous study on MC disc microbiome. No bacterial genera were shared using the same bioinformatic parameters. Taxonomic lineage assignment using "amplicon sequencing variants" was more sensitive and detected 48 genera compared to 22 with "operational taxonomic units" (previous study). Increasing filter cut-off from 4% to 50% (previous study) reduced genera from 48 to 4 genera. Despite these differences, both studies observed dysbiosis with an increased abundance of gram-negative bacteria in MC discs as well as a lower beta-diversity. Cutibacterium was persistently detected in all groups independent of the bioinformatic approach, emphasizing its prevalence.

Conclusion: There is dysbiosis in MC discs. Bioinformatic parameters impact results yet cannot explain the different findings from this and a previous study. Therefore, discrepancies are likely caused by different sample preparations or true biologic differences. Harmonized protocols are required to advance understanding of the disc microbiome and its clinical implications.

Keywords: Cutibacterium acnes; Modic changes; metagenomics; microbiome.