Introduction: Regional lymph node metastasis is associated with poorer outcomes in patients and may contribute to systemic immunosuppression, facilitating distant metastasis. Despite its clinical significance, the mechanisms by which lymph node metastases mediate disease progression, particularly in the context of tissue organization, remain unclear. This study aims to elucidate the spatial and cellular dynamics involved in lymph node metastasis in head and neck squamous cell carcinoma (HNSCC). We hypothesized that a comprehensive spatial profiling of immune and cancer cell phenotypes within primary tumors and metastatic lymph nodes would reveal critical insights into the metastatic process.
Materials and
Methods: We analyzed 475 matched formalin-fixed paraffin-embedded (FFPE) samples from primary tumors, metastatic lymph nodes, and benign lymph nodes from 80 HNSCC patients (Fig. 1A). Using the Phenocycler technology, we performed high-dimensional, single-cell multiplexed imaging of 50+ antibodies and 100+ RNA molecules. A 59-marker antibody panel was designed to identify major immune cell types and cancer cell subphenotypes, such as epithelial-to-mesenchymal transition and hypoxic tumor cells (Fig. 1B). Additionally, a 100-target RNA panel was created to explore functional networks and communication pathways, including chemokine axes suggested by single-cell RNA sequencing.
We performed image processing, single-cell segmentation for quantified protein expression, single-cell marker intensity z-normalization, clustering using unsupervised leiden clustering, and annotations for cell types based on marker expression and location in tissue. We co-register paired CODEX imaging (resolution=0.5 m) with shared H&E following imaging (Fig. 1C). To analyze this single-cell data, used Delaunay triangulation for distance-based cell-cell interactions across cell pairs and permutation testing for neighborhood-neighborhood pairs. We identified multicellular neighborhoods by taking the 20 nearest neighbors, associate these into cellular vectors and cluster with k-means clustering to identify conserved compositions of cells. We similarly processed the cytokine fields by adding up a number of cytokine transcripts from cells that had been segmented in certain radii and clustering cytokine neighborhoods.
Results, Conclusions, and Discussions: Our high-plex protein imaging identified 27 distinct cell types across nearly 1.8 million cells, allowing for comprehensive spatial neighborhood and cell-cell interaction analyses (Fig. 1C). We observed significant associations CD4+ T cells in the tumor microenvironment and CA9+ hypoxic tumor cells associated with nodal metastasis (Fig. 1D, E). Comparison of the paired tumor microenvironment, metastatic lymph node, and benign lymph nodes established a number of trends of the increase of B, CD4+ T, and dendritic cells, while PD1+ CD8+ T, Treg, Macrophages, and CAFs all decreased respectively (Fig. 1F).
In higher-order spatial interactions, we observed significant changes in cell-cell interactions, particularly between tumor cells and CD4+ T regulatory cells (Fig. 1G). We also observe significant changes in abundance of multicellular neighborhoods such as the plasma cell enriched immune neighborhood within the primary tumor that was associated with not having distant recurrence of the cancer (Fig. 1H). Finally, single-cell RNA profiling further revealed critical chemokine axes, including CXCL9 and CXCL10 (Fig. 1I). Indeed, epithelial cells and epithelial cell neighborhoods from within the tumor secrete higher levels of these cytokines (Fig. 1J).
Our study combines high-plex protein and RNA spatial profiling to unravel lymph node metastasis processes in HNSCC. We identified specific cell-cell interactions and chemokine axes, such as CXCL10, providing insights into immune evasion and metastatic progression. The significant role of Tregs and other immune cells in the metastatic niche suggests potential therapeutic targets. Mapping the spatial organization of immune and cancer cells within metastatic lymph nodes offers a framework for understanding tissue organization’s influence on disease progression. Future studies should explore targeting identified molecular pathways to inhibit metastasis and improve patient outcomes. The integration of high-plex profiling connects molecular mechanisms with spatial tissue interactions, revealing active communication between cancer cells and the immune microenvironment. Our findings highlight the value of spatial profiling technologies in oncology, offering a comprehensive understanding of tumor biology and aiding in the development of precise strategies to combat metastasis and enhance clinical outcomes for HNSCC patients.
.jpg "John Hickey, PhD photo")