INTRODUCTION

Modic type 1 changes (MC1) are an inflammation of the vertebral bone marrow and an independent source of low back pain. Two etiologies have been described: (i) an autoinflammatory reaction of the bone marrow against disc cells and matrix and (ii) infection of the disc with bacteria, mainly with Cutibacterium acnes (C. acnes). The different etiologies require different treatments. Therefore, we need to be able to diagnostically distinguish the two etiologies and understand the etiology-specific pathomechanisms. The aim of this study was to identify etiology-specific pathomechanisms that could serve as basis for etiology-specific biomarkers.

METHODS

This study was approved by the local Ethics Commission and performed in accordance with the Declaration of Helsinki. Bone marrow aspirates were obtained from MC1 patients undergoing spinal fusion. Aspirates were taken prior to screw insertion through the pedicle screw trajectory. From each patient, a MC1 and an intra-patient control aspiration from the adjacent vertebral level was collected in K2-EDTA tubes. Plasma and cell fraction were separated and collected by centrifugation. Plasma (n=12+12) was analyzed for interleukin-1β (IL-1β), granulocyte-macrophage colony-stimulating factor (GM-CSF), epithelial-derived neutrophil-activating peptide 78 (ENA-78), tumor necrosis factor alpha (TNF-α), colony stimulating factor 1 (M-CSF), chemokine C-C motif ligand 2 (CCL2), interleukin-4 (IL-4), IL-6, IL-8, and IL-13 using MesoScale U-Plex. Different cytokine profiles were identified with Uniform-Manifold-Approximation-Projection (UMAP) analysis. Protein concentrations between two clusters were compared with t-tests. Neutrophils (n=8+8) were isolated from the cell fraction (StemCell neutrophile kit), RNA isolated (Qiagen RNeasy Kit) and sequenced (Novaseq). Differentially expressed genes (DEG; p-value < 0.01, log2fc > ± 0.5) in MC1 were compared to control bone marrow and analyzed with gene ontology (GO) enrichment in R (GOseq). Discs adjacent to MC1 (n=8) were tested with anaerobic bacterial culture and 16S qPCR for the presence of C. acnes.

RESULTS

UMAP dimensionality reduction identified two cytokine clusters (Figure a). IVDs of patients in cluster 2 had higher C. acnes genome copies (n = 6, median = 2848 copies/g) than IVDs in cluster 2 (n = 2, median = 351 copies/g) (Figure b). This suggests that patients of cluster 1 have an autoinflammatory etiology and patients of cluster 2 an infectious etiology. There was no difference in cytokine concentrations between MC1 and control. However, when stratifying by clusters, IL-8, ENA-78, M-CSF, and IL-1β were significantly higher in cluster 2 than in cluster 1 (Figure c). Since IL-8 and ENA-78 are potent neutrophil chemoattractants and activators, we compared the neutrophile transcriptome in MC1 bone marrow of cluster 1 and 2. Only 15 DEGs overlapped between cluster 1 (185 DEG) and cluster 2 (165 DEG) suggesting different pathomechanisms (Figure d). Six of the top 12 enriched biological processes in cluster 2 related to bacterial defense response. None of them were present in cluster 1.

DISCUSSION

Autoinflammatory (cluster 1) and infectious etiologies (cluster 2) of MC1 have different cytokine profiles and neutrophile activation profiles indicating different pathomechanisms. It is clinically relevant to be able to distinguish the two etiologies, because different pathomechanisms require different treatments.