High-Resolution Profiling of Human Vocal Fold Cellular Landscapes With Single-Nuclei RNA Sequencing

Abstract

Introduction: The function of the vocal folds (VFs) is determined by the phenotype, abundance, and distribution of differentiated cells within specific microenvironments. Identifying this histologic framework is crucial in understanding laryngeal disease. A paucity of studies investigating VF cellular heterogeneity has been undertaken. Here, we examined the cellular landscape of human VFs by utilizing single-nuclei RNA-sequencing.

Methods: Normal true VF tissue was excised from five patients undergoing pitch elevation surgery. Tissue was snap frozen in liquid nitrogen and subjected to cellular digestion and nuclear extraction. Nuclei were processed for single-nucleus sequencing using the 10X Genomics Chromium platform. Sequencing reads were assembled using cellranger and analyzed with the scanpy package in python.

Results: RNA sequencing revealed 18 global cell clusters. While many were of epithelial origin, expected cell types, such as fibroblasts, immune cells, muscle cells, and endothelial cells were present. Subcluster analysis defined unique epithelial, immune, and fibroblast subpopulations.

Conclusion: This study evaluated the cellular heterogeneity of normal human VFs by utilizing single-nuclei RNA-sequencing. With further confirmation through additional spatial sequencing and microscopic imaging, a novel cellular map of the VFs may provide insight into new cellular targets for VF disease.

Level of evidence: NA Laryngoscope, 134:3193-3200, 2024.

Keywords: cellular architecture; single nuclei RNA sequencing; vocal fold.

Figure 1:. Tissue samples.

(A) Representative wound bed after excision of anterior VF. There is deep extension including vocal ligament and thyroarytenoid muscle. There is not extensive infraglottic extension. (B) Masson Trichrome stain of a transverse section of excised VF tissue. Epithelium, blood vessels, and muscle are in red, while collagen fibers are in blue. Section demonstrates all layers of VF are present, including deep submucosal glands at the infraglottic border. 2.5X magnification; scale bar 1mm. Abbreviations: E: Epithelium, VL: Vocal ligament, M: Muscle, AC: Anterior commissure, SLP: Superficial lamina propria.

Figure 2:. Global clustering and annotation.

(A) UMAP of 18 clusters identified after single nuclei sequencing. (B) z-score normalized gene expression for key defining genes in each annotated cluster. Each gene mean expression per cell type is scaled relative to the mean expression of that gene in all other cell types. (C) Suggested final annotation for the global dataset after further subcluster analysis.

Figure 3:. Epithelial subcluster analysis.

(A) UMAP of 12 clusters identified after subclustering to epithelial cell types. (B) Matrix gene expression of cytokeratins showing expression in the epithelial clusters. Clear differential expression patterns were seen. (C) Matrix gene expression patterns of seromucinous markers. (D) (B) z-score normalized gene expression for key defining genes in each annotated epithelial cluster. Each gene mean expression per cell type is scaled relative to the mean expression of that gene in all other cell types. (E) Final epithelial cluster annotation based on subanalysis.

Figure 4:. Immune cell subcluster analysis.

(A) UMAP of 13 clusters identified after subclustering to immune cell types. (B) z-score normalized gene expression for key defining genes in each annotated immune cluster. Each gene mean expression per cell type is scaled relative to the mean expression of that gene in all other cell types. (C) Gene expression overlays on UMAPs showing CD8A expression in most lymphocyte derived subclusters. (D) NK_cell_score overlays on UMAPs showing expression localized to cluster 2. (E) Final immune cell subcluster annotation.

Figure 5:. Fibroblast subcluster analysis.

(A) UMAP of 9 clusters identified after subclustering to connective tissue cells. (B) Dot plot matrix of highest expressing ECM components within the fibroblast subclusters. (C) z-score normalized gene expression for key defining genes in each annotated stromal cluster. Each gene mean expression per cell type is scaled relative to the mean expression of that gene in all other cell types. (D) Dot plot matrix of genes implicated in scar formation. PRRX1, YAP1, WWRT1, TEAD1 all showed co-expression in clusters 4, 6, and 7 with COL3A1.