Authors: Jing Yi LEE, Bavani KANNAN, Jason Yongsheng CHAN

Cancer Discovery Hub, National Cancer Centre Singapore

Each individual rare cancer, as the term suggests, affects only a fraction of the world's population. Yet, its rarity accentuates the importance of further research in the area, to resolve the lack of understanding of disease etiology, effective treatment options and accurate diagnosis of these understudied diseases. 

Harnessing modern molecular profiling technologies, our group at the Cancer Discovery Hub, National Cancer Centre Singapore has been illuminating the pathobiology of rare soft tissue sarcomas as part of the STARLIGHT Initiative ("Steering Translational Rare Cancer Research into the Light”) with the aim to elucidate the molecular underpinnings of rare cancers and eventually improve clinical outcomes of this group of diseases with unmet clinical need. In our previous study on angiosarcoma, an Asian-centric soft tissue sarcoma, we uncovered distinct mutational signatures and gene expression profiles, as well as immune signatures specific to the cancer [1, 2]. Specifically, we noted that angiosarcomas of the head and neck region are more prevalent in Asian than Western populations, in keeping with other reports [1, 2]. Using NanoString profiling, we identified 3 distinct clusters mainly characterized by lack (clusters 1 and 2) or enrichment (cluster 3) of immune-related signaling and immune cells [1]. UV-related angiosarcomas as well as those arising from the head and neck were predominantly found in clusters 1 and 3 [1]. Notably, cluster 3 was distinctly represented by overexpression of inflammation- and immune-related signaling pathway genes, as well as strong enrichment for immune cells [1]. Unsurprisingly, cluster 3 had the highest tumor inflammation signature (TIS) score, as determined by NanoString profiling, suggesting that this subgroup of angiosarcomas may benefit from checkpoint immunotherapy more so than the other clusters [1]. Intriguingly, when compared with HHV-7^– tumors, HHV-7⁺ angiosarcomas were found to be associated with higher TIS scores and were overrepresented in cluster 3 [1]. These findings suggest the existence of angiosarcoma subtypes characterized by distinct etiological and biological phenotypes. We sought to further investigate the genomic, transcriptomic and immune landscapes of angiosarcoma in a larger cohort at a multidimensional level.

Encouraged by the findings above, we were interested in validating our findings in an expanded cohort and gain insights into the tumor immune microenvironment and its implications on immune surveillance. We studied a total of 81 angiosarcoma tumors, with a focus on Asian head and neck (AS-HN) cases. Whole genome sequencing revealed recurrent somatic mutations that included POT1 and TP53, in keeping with our previous study, and others' [1,2]. Tumor mutation burden (TMB) was significantly higher in UV-related AS-HN cases compared to non-UV-related ones. NanoString PanCancer IO360 panel profiling identified three clusters with distinct levels of immune-related signaling – immune-hot, -intermediate and -cold, consistent with our previous report [1]. Interestingly, based on NanoString gene expression data, HHV-7 positive tumors were relatively enriched for B-cells and mast cells and had higher tumor inflammation signature (TIS) scores compared to HHV-7 negative tumors. UV-positive tumors were relatively enriched for CD8+/cytotoxic T-cells but had comparable TIS scores to UV-negative tumors.

Spatial transcriptomic profiling of AS-HN sarcomas representing four putative pathogenic etiologies (source: [3])

Spatial transcriptomics is a powerful tool that allows for in situ cell typing, gene expression profiling, and the study of the tumor microenvironment (TME). By profiling AS-HN tumors, each representative of the 4 putative etiological subtypes (AS-17: UV-positive, HHV-7-positive, and immune hot; AS-52: UV-positive, HHV-7-negative, and immune intermediate; AS-03: UV-negative, HHV-7-positive, and immune intermediate; AS-07: UV-negative, HHV-7-negative, and immune cold) on the 10X Genomics Visium platform, we found that each sample showed distinct distributions of tumor cells and microenvironmental components, such as immune cells and fibroblasts. The tumors also had different TIS scores that could be visualized spatially. Interestingly, we showed that the immune-hot AS-HN tumor had immune cells that were largely excluded from the bulk of the tumor compartment, which mainly consisted of myeloid cell infiltrate. Immune-cold AS-HN tumors, on the other hand, may contain high levels of TIS gene expression in a focal pattern within the tumor compartment. How these observations are clinically relevant is a question that needs to be further investigated. These preliminary findings, however, highlight the potential of integrating factors other than high TMB and elevated PD-L1 expression as biomarkers to predict treatment response to immunotherapy in angiosarcoma, such as mutational signatures and TIS scores.

Jing Yi LEE presenting our work on AS-HN angiosarcoma at CTOS Annual Meeting 2022, Vancouver, Canada (photo credit: Dr Herbert Loong, The Chinese University of Hong Kong). 

So, what's next? We aim to expand our discovery efforts into the epigenetic landscape of angiosarcoma. Epigenetic variation is an important contributor to tumor heterogeneity. A recent study has demonstrated how the use of DNA methylation profiling of angiosarcoma revealed distinct clusters that correlated with clinical subtype, overall survival and chromosomal instability [4]. By analyzing the DNA methylation patterns of a patient's tumor, researchers can identify specific epigenetic modifications that are associated with the tumor's growth and progression and thus, tailor personalized treatment approaches that target the specific epigenetic modifications present in the tumor. Ultimately, a deep dive into the pan-omic landscape of angiosarcomas will hopefully present greater insights into this rare cancer [5]. 

References

1. Chan, J. Y., Lim, J. Q., Yeong, J., et al. (2020). Multiomic analysis and immunoprofiling reveal distinct subtypes of human angiosarcoma. Journal of Clinical Investigation, 130(11), 5833-5846. doi:10.1172/jci139080
2. Chan, J. Y., Tan, G. F., Yeong, J., et al. (2021). Clinical implications of systemic and local immune responses in human angiosarcoma. Npj Precision Oncology, 5(1). doi:10.1038/s41698-021-00150-x
3. Loh, J. W., Lee, J. Y., Lim, A. H., et al. (2023). Spatial transcriptomics reveal topological immune landscapes of Asian head and neck angiosarcoma. Communications Biology, 6(1). doi:10.1038/s42003-023-04856-5
4. Weidema, M. E., Van de Geer, E., Koelsche, C., et al. (2023). Data from DNA methylation profiling identifies distinct clusters in angiosarcomas. Clinical Cancer Research, 26(1), 93-100. doi: 10.1158/1078-0432.CCR-19-2180.doi:10.1158/1078-0432.c.6529385.v1
5. Lee, J. Y.,  Kannan, B., Lim, B. Y., et al. (2022). The multi-dimensional biomarker landscape in cancer immunotherapy. International Journal of Molecular Sciences, 23(14), 7839. doi: 10.3390/ijms23147839.