Sarcoma ecosystems : spatial characterization and prognostic significance

Abstract

Sarcoma is a highly heterogeneous disease with complex biological activities and unique tumor microenvironments (TME) in distinct subtypes. The limited treatment options and inadequate responses to current therapies necessitate a deeper understanding of sarcoma biology and personalized treatment strategies. Our research comprehensively explores the sarcoma TME through advanced spatial analysis and investigates sarcoma's molecular and genetic profiles through transcriptome and genome sequencing.

In paper I, we focused on undifferentiated pleomorphic sarcoma (UPS) using multiplex immunofluorescence (mIF) staining for in-depth spatial analysis of B cell populations and lymphocyte aggregates (LAs). LAs in UPS were found to be associated with longer overall survival (OS) and metastasis-free survival (MFS). Moreover, we unveiled distinct maturation profiles among B cell subsets, indicative of different phenotypes that contribute to functional ecosystems in TME. LA-positive tumors displayed a more well-differentiated B cell profile throughout the entire tumor section, not limited in LA regions. We introduced the B-index, an integrated measurement tool combining B cell abundance and maturity, which demonstrated predictive power for both MFS and OS. Using the TissUUmap tool, we identified B cell desert areas characterized by extremely low B cell infiltration. LA-positive tumors displayed smaller and more fragmented B cell desert areas.

In paper II, we performed double immunohistochemistry to study CD11c-positive antigen-presenting cells (APCs) and CD8- positive cells in 177 soft tissue sarcoma (STS) patients. We found that CD11c-CD8 interactions in the TME were associated with improved MFS and OS. Transcriptomic analysis in The Cancer Genome Atlas (TCGA) sarcoma cohort supported the prognostic significance of combining CD11c with CD8, irrespective of FOXP3 levels and in the presence of CD274 (PD-L1).

In paper III, we conducted transcriptome and targeted DNA sequencing in 91 synovial sarcomas, identifying three distinct Synovial Sarcoma Clusters (SSCs) mirroring histological subtypes. SSC-I was characterized by high cell proliferation and immune evasion with an unfavorable prognosis. SSC-II was dominated by vascularstromal components and correlated with better outcomes. SSC-III displayed biphasic differentiation, genomic complexity, and immune checkpoint-mediated immune suppression, leading to adverse outcomes, even after a good histological response to neoadjuvant treatment.

In paper IV, we analyzed Ewing sarcoma (ES) transcriptome signatures in four previously published cohorts and identified 29 prognostic RNA-binding protein (RBP) genes, from which we constructed and validated an RBP-associated prognostic risk model (RPRM). The RPRM demonstrated stable predictive value for prognosis, with NSUN7 emerging as an independent and favorable prognostic marker.

In summary, our research integrates spatial analysis of the sarcoma TME to identify unique immune features and prognostic markers. Moreover, we use transcriptomic and genomic analyses to categorize specific sarcoma types for more detailed survival stratification. This work provides a deeper insight into the sarcoma TME and suggests an improved grouping strategy, aiming to shape the development of personalized immunotherapy in the future.