Spatial mapping of immune cell environments in NF2-related schwannomatosis vestibular schwannoma

Abstract

NF2-related Schwannomatosis (NF2 SWN) is a rare disease characterised by the growth of multiple nervous system neoplasms, including bilateral vestibular schwannoma (VS). VS tumours are characterised by extensive leucocyte infiltration. However, the immunological landscape in VS and the spatial determinants within the tumour microenvironment that shape the trajectory of disease are presently unknown. In this study, to elucidate the complex immunological networks across VS, we performed imaging mass cytometry (IMC) on clinically annotated VS samples from NF2 SWN patients. We reveal the heterogeneity in neoplastic cell, myeloid cell and T cell populations that co-exist within VS, and that distinct myeloid cell and Schwann cell populations reside within varied spatial contextures across characteristic Antoni A and B histomorphic niches. Interestingly, T-cell populations co-localise with tumour-associated macrophages (TAMs) in Antoni A regions, seemingly limiting their ability to interact with tumorigenic Schwann cells. This spatial landscape is altered in Antoni B regions, where T-cell populations appear to interact with PD-L1⁺ Schwann cells. We also demonstrate that prior bevacizumab treatment (VEGF-A antagonist) preferentially reduces alternatively activated-like TAMs, whilst enhancing CD44 expression, in bevacizumab-treated tumours. Together, we describe niche-dependent modes of T-cell regulation in NF2 SWN VS, indicating the potential for microenvironment-altering therapies for VS.

Fig. 1. Imaging mass cytometry (IMC) workflow and generation of the single-cell atlas in NF2 SWN-related vestibular schwannoma.

A Schematic of IMC workflow for investigating NF2 SWN-related VS: niche identification, tissue processing and staining, data acquisition and processing of raw data, single-cell segmentation, and downstream analyses. B Sequential haematoxylin and eosin (H&E) and IMC images of Antoni A and Antoni B regions, visualising the core Schwann cell (S100B; purple), macrophage (Iba1, red), T-cell (CD8α; green), vascular (SMA; cyan), and proliferation (Ki-67; white) markers. Scale bar representative of 100 μm. C Heatmap detailing expression pattern of all single-cell populations identified in IMC analysis via Leiden clustering. Scale bar indicating normalised expression (N.E.). D Comparison of case-averaged abundance of significantly different cell populations across Antoni A and Antoni B regions. The remaining non-significant population abundances by histopathology can be found in Fig. S4 (n = 13). E Comparison of alternatively activated-like TAMs versus classically activated-like TAMs across both Antoni A and Antoni B regions (n = 13). F Relative abundance of cell populations across ROIs, annotated by case, histopathology, and disease severity. SMA: smooth muscle actin. Statistical comparisons in D and E were made using two-tailed unpaired t-tests (for normally distributed data) or two-tailed Mann–Whitney U tests (result are the mean of the group + SD). Source data are provided as a Source Data file.

Fig. 2. Spatial omics highlight the interactomes between Schwann cell and Myeloid cell populations in NF2 SWN-related vestibular schwannoma.

A From left to right, IMC visualisation of Schwann cells and myeloid cells across Antoni A and Antoni B regions. Markers used: S100B (purple), SOX-10 (white), Iba1 (red), HLA-DR (green), and SMA (cyan). Next, single-cell spatial seaborn maps of all Schwann cell and myeloid cell populations within Antoni A and B regions are shown, with a composite of the 2 major cell groups with the vascular networks. The scale bar indicates 100 µm. B Cross-pair correlation functions (PCF) heatmaps showing the Schwann cell and myeloid cell interaction; scale bar indicates strength of significant cell pair correlates. C Spatial connectivity plots indicating the significant networks of Schwann cell and myeloid cell populations, by adjacency cell network (ACN) analysis, across Antoni A and Antoni B histomorphic niches, with representative images of each network. Node size (coloured circle) indicates mean abundance for each cell cluster across all ROIs; lines connecting each node shows significant cell networks between cell types informed by ACN analysis. Line thickness associates the gr20 value to each significant cell pair, where the thicker the line, the higher the co-localisation of the cell populations. scale bar indicates strength of connectivity. B, C *p < 0.05, using cross-PCF and ACN statistical analyses, see Methods (n = 13). Source data are provided as a Source Data file.

Fig. 3. Spatial omics demonstrate the T-cell interactome across different histomorphic niches in NF2 SWN-related vestibular schwannoma.

A From left to right, IMC visualisation of T-cells across Antoni A and Antoni B regions. Markers used: S100B (purple), Iba1 (red), CD8α (green), CD4 (white) and SMA (cyan). Next, single-cell spatial seaborn maps of all T-cell populations, and co-localisation to Schwann cell or TAM populations with the vascular networks across Antoni A and B regions. Scale bar indicates 100 μm. B Cross-pair correlation functions (PCF) heatmaps showing T-cell interaction partners with all other cell populations; scale bar indicates strength of significant cell pair correlates. C Spatial connectivity plots indicating the significant networks of T-cell populations, by adjacency cell network (ACN) analysis, across Antoni A and Antoni B histomorphic niches, with representative images of each network. Node size (coloured circle) indicates mean abundance for each cell cluster across all ROIs; lines connecting each node show significant cell networks between cell types informed by ACN analysis. Line thickness associates the gr20 value to each significant cell pair, where the thicker the line, the higher the co-localisation of the cell populations. the scale bar indicates strength of connectivity. B, C *p < 0.05, using cross-PCF and ACN statistical analyses, see Methods (n = 13). Source data are provided as a Source Data file.

Fig. 4. Receptor-ligand interaction analyses for CD8+ and CD4+ TEMRA cell populations within VS.

Receptor-ligand analysis between A CD8⁺ and B CD4⁺ TEMRA cells and statistically confirmed co-localised cell populations (cross-PCF analyses, Fig. 3). Receptor-ligand pairs were taken from CellPhoneDB. Transcriptomes for IMC populations were identified by matching IMC and scRNA-seq cells using MaxFuse. Only significant (p < 0.05 with false discovery rate in the Squidpy permutation test) interactions are shown. Values are interaction scores as calculated by Squidpy. In significant sender interactions (blue), the lymphocyte expresses the ligand, and the interacting cell expresses the receptor. In significant receiver interactions (green), the lymphocyte expresses the receptor, and the interacting cell expresses the ligand. Source data are provided as a Source Data file.

Fig. 5. Cellular neighborhood analysis identifies discrete microenvironmental regions across histomorphic niches in NF2 SWN-related vestibular schwannoma.

A Dot plot of the individually derived spatial clusters (referred to as cellular neighbourhoods, CN) for Antoni A regions and Antoni B regions. Optimal cluster stability for both groups was 10. Cells were clustered using Gaussian mixture modelling. Node size indicates relative cell type enrichment of each cell across generated CNs (red = high enrichment, grey = low enrichment) (n = 13). B Relative abundance of both Antoni A and Antoni B CNs across cases (n = 13). C Representative spatial mapping of CNs across Antoni A and Antoni B regions. Each image is representative on an independent case to highlight intertumoral heterogeneity (CN colour key as in A, B). D Significant correlations between CNs and tumour growth rate (n = 13). Correlations were assessed by either Pearson correlation coefficient or Spearman correlation. Line of best-fit was generated by simple linear regression analysis. E CN proximity heatmaps depicting the spatial proximity of each CN to every other CN. Y-axis indicates sender communication; x-axis indicates receiver communication. Diverging scale bar denotes proximity enrichment, with red specifying high proximity, and blue specifying low proximity. Source data are provided as a Source Data file.

Fig. 6. IMC analyses highlight the influence of bevacizumab treatment on the microenvironment in NF2 SWN-related vestibular schwannoma.

A (left) IMC visualisation of core tumour and immune-related markers across treatment naïve and bevacizumab-treated groups. Markers used: S100B (purple), Iba1 (red), CD8α (green), CD44 (white), HLA-DR (yellow) and SMA (cyan). (right) Spatial seaborn maps of all populations across treatment naïve and bevacizumab-treated groups. Scale bar indicates 100 µm. B Abundance of significantly different cell populations across treatment naïve (n = 13) and bevacizumab-treated (n = 3) groups, tested by Mann Whitney U test. The remaining non-significant population abundances by treatment status can be found in Fig. S8A-D. C Comparison of alternatively activated-like TAMs versus classically activated-like TAMs across treatment naïve (n = 13) and bevacizumab-treated (n = 3) groups. Treatment-naïve tested by two-tailed Mann Whitney U test, and bevacizumab-treated tested by two-tailed unpaired t-test. D Visualisation of bevacizumab-inclusive CNs across treatment naïve and bevacizumab-treated groups. Dot plot of spatial clusters can be found in Fig. S8E. E Mean proportion of CNs across treatment-naïve (n = 13) and bevacizumab-treated (n = 3) groups. F Comparison of CN Bev-9 between treatment-naïve and bevacizumab treated groups, tested by two-tailed Welch’s t-test. Bev = bevacizumab. (B, C, F result are the mean of the group + SD). Source data are provided as a Source Data file.