Abstract
Category:
Basic Sciences/Biologics; Ankle
Introduction/Purpose:
Peroneal tendon disease (PTD) represents a spectrum of pathologies. Current literature focuses on pathoanatomy, but not the microenvironment of the peroneal complex on a histopathological cellular level. This information is crucial to understanding disease progression and future development of therapeutic strategies. This study aims to define the histological microenvironment and transcriptome of tenosynovium in patients with various disease stages. We further propose a classification system for PTD: (Type 0) Normal tenosynovium; (Type 1) Inflammatory tenosynovium with normal appearing tendons; (Type 2) inflammatory or degenerative tenosynovium with tendinosis or split tear involving less than 50% of tendon substance; (Type 3) Inflammatory or degenerative tenosynovium with tendinosis or split tear involving greater than 50% of tendon substance; (Type 4) Complete or near complete degenerative rupture.
Methods:
Tenosynovium was sampled from patients, classified for PTD, and then processed for RNA sequencing (RNA-seq; n = 3) and histology (n = 7). RNA-seq was performed using the Illumina NovaSeq6000 platform. Differential expression was determined using the R edgeR (3.22.5) package after adjusting via a scaling normalization factor and correcting P-values via the Benjamini Hochberg method. Tenosynovium samples were processed for hematoxylin and eosin (H&E) staining and evaluated using machine learning to classify cell types. Briefly, QuPath software was used to load images of H&E section4-means, fuzzy c-means and gaussian mixture. Cells were assigned to clusters and then quantified and visualized with density plots in QuPath.
Results:
UMAP with k-means clustering was the best data reduction and clustering technique. Total cellular density increased with type 3 and 4 PTD compared to type 1; cluster 3 cells were the most increased with types 3 and 4 compared to type 1 (Figure 1A). Cluster 3 cells appear in vascular structures and are elongated/oval with dark nuclear staining (Figure 1B). In type 3 and 4, cluster 3 cells appear in and around vascular structures and at the surface lining in type 4. RNA-seq identified 13 genes differentially expressed (p < 0.05; Figure 1B). Significant gene ontology terms include muscle contraction, myofibril assembly, and actin myosin filament sliding (Figure 1C). Significant reactome pathways include muscle contraction, fatty acyl-CoA biosynthesis, and activation of gene expression by SREBF.
Conclusion:
In this study, RNA-seq and machine learning cell classification in H&E staining yield findings that represent the cellular and molecular disease signature in PTD. Histological analysis demonstrated that in more severe classes of PTD (types 3-4), cell morphologies found in vascular structures and surface lining were increased and suggestive of fibroblasts/ myofibroblasts and smooth muscle cells. RNA-seq demonstrated an increased expression of genes, with muscle contraction and myofibril activity representing significant signatures. Overall, our study demonstrates potentially increased cell and gene pathways consistent with smooth muscle activity and fibrosis. These pathways present targets for better classifying and treating PTD pathology.