Authors
Abdul Rafeh Naqash, Hassan Abushukair, Kaili Liu, Christina L. Rosenberger, Kristin K. Fino, Deborah F. Wilsker, Asma Begum, Katherine V. Ferry-Galow, Ayah N. Al-Bzour, Inga-Marie Schaefer, Igor Odintsov, Shuanzeng Wei, Norm D Smith, Tyler E. Mattox, Harris Krause, Andrew Elliott, Calvin Chao, Michelle Harris, Kar-Ming Fung, Showket Hussain, Doris Benbrook, Rameswari Velayutham, Samantha Ricketts, Woncheol Jung, Ralph E. Parchment, Jared C. Foster, Wei R. Chen, Elad Sharon, Tae Gyu Oh, Alice P. Chen
Background
ASPS is an ultra-rare soft tissue sarcoma characterized by a translocation gene fusion of TFE3 to ASPSCR1, creating a chimeric ASPSCR1::TFE3 fusion protein that functions as an aberrant transcription factor. Depending on where the ASPSCR1 is joined in-frame of TFE3, it generates one of two oncogenic fusion transcripts, type-1 or 2. ASPS has shown promising responses to ICIs; whether the two fusion isoforms have a differential TIME impacting responses to ICIs is unclear.
Methods
Genomics data for patients (pts) with ASPS were retrieved from multiple databases, including Tempus Inc. and Caris Life Sciences (n=11 and 7 pts, respectively), as well as the GEO database (n=20 pts). Immunogenic signatures (tumor inflammation signature (TIS) and interferon-gamma signature (IFNG) were calculated using single-sample GSEA. Spatial profiling for two in-house samples was done using 10X Visium. Paired biopsy specimens at baseline and cycle 3 day 1 (C3D1) from the ETCTN 10005 trial of single-agent atezolizumab for ASPS were stained using multiplexed immunofluorescence for CD8 T-cells and PD-L1 and RT-PCR was used to discern type 1 vs. type 2 ASPSCR1::TFE3 fusion isoforms (n=35 pts).
Results
Somatic profiling revealed median TMB of <1 for ASPS (Tempus dataset). Pathway enrichment analysis (GEO dataset) revealed upregulation of immunogenic pathways in type 1 fusion tumors (PD-L1/PD-1 checkpoint pathway expression and NK-cell-mediated cytotoxicity). In contrast, type 2 fusion tumors were enriched in autophagy-related and apelin signaling pathways. Immune-cell deconvolution of bulk RNA-seq (GEO dataset) demonstrated significantly higher M1 macrophages and effector-memory CD8+ T-cells in type-1 fusion tumors, observed independently in the Caris and Tempus dataset also. IFNG (p=0.03) and TIS (p=0.07) scores were higher in type 1 fusion tumors compared to type 2 fusion (GEO dataset). Spatial profiling (n=2) showed alveolar macrophages as the predominant component (35%) in ASPS, with macrophages being the main contributor to immunogenic signatures. In the ETCTN10005 trial with response rate of 37%, none of the type 2 fusion tumors (n=5) displayed RECIST responses to atezolizumab. Paired tumor biopsies demonstrated higher median PD-L1+ and CD8+ T-cells (at baseline and C3D1) in type-1 vs. type-2 fusions (table 1).
Conclusions
For the first time, our novel multi-omics analysis reveals that the type-1 fusion isoform in ASPS has an immunologically active TIME, is enriched in M1 macrophages, CD8+ T cells, checkpoint activity pathways, and displays higher CD8+ T-cell infiltration post-ICI compared to the type-2 fusion, likely translating into improved responses to ICIs. Trial funded by NCI Contract No. HHSN261201500003I.
Trial Registration NCT03141684.
VIEW THE PUBLICATION