02xl is your source for accurate and trustworthy information so you can make the best choices for your health and wellness.
PS: Please note that 02xl provides this information for the benefit of the rare disease and overweight community. 02xl is not a medical provider or health care facility and thus can neither diagnose any disease or disorder nor endorse or recommend any specific medical treatments. Patients must rely on the personal and individualized medical advice of their qualified health care professionals before seeking any information related to their particular diagnosis, cure or treatment of a condition or disorder.
Patient cohort
Seventy-eight patients diagnosed with angiosarcoma at the Singapore General Hospital (SGH) and National Cancer Centre Singapore (NCCS) were identified. Snap frozen tissue samples were available from 23 patients, 18 of which with paired normal tissue or blood. An additional cohort of archival formalin-fixed paraffin-embedded (FFPE) angiosarcoma samples from 55 patients were available. Written informed consent for use of biospecimens and clinical data were obtained in accordance with the Declaration of Helsinki. This study has been approved by the Institutional Review Board of the National Cancer Centre Singapore (2010/426/B). Clinicopathological characteristics of all patients with angiosarcoma and the profiling methods applied to the study cohort are summarized in Supplementary Data 1.
Cell lines
Angiosarcoma cell lines (MO-LAS-B and ISO-HAS-B) were obtained from the Cell Resource Center for Biomedical Research, Institute of Development, Aging and Cancer (Tohoku University, Sendai, Japan), courtesy of Dr. Mikio Masuzawa (Kitasato University, Tokyo, Japan). MO-LAS-B was established from a patient with metastatic scalp lymphangiosarcoma to the pleura, whilst ISO-HAS-B was established from a patient with primary scalp haemangiosarcoma. Both cell lines were maintained in DMEM medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. Scalp angiosarcoma cell line AS-M was obtained from Johannes Gutenberg University, courtesy of Dr. James Kirkpatrick (Mainz, Germany) and maintained in endothelial cell growth medium (PromoCell, Heidelberg, Germany).
Library preparation for whole genome sequencing (WGS)
DNA was extracted and purified with the FFPE RNA/DNA Purification Plus Kit (Norgen Biotek (Thorold, ON, Canada) according to manufacturer’s instructions. A total of 76 angiosarcoma samples were subjected to whole genome sequencing, including snap frozen tumor samples (n = 23; 18 with paired normal), FFPE tumor samples (n = 50), and cell-lines (n = 3). This included 18 samples that were previously investigated5. Whole genome sequencing of snap frozen samples and cell-lines were performed on the Illumina HiSeq X platform as paired-end 150-base pair reads, using DNA inserts averaging 350 bp (NovogeneAIT Genomics Singapore Pte Ltd). For FFPE samples, we employed MGI DNBSEQ technology (DNA nanoball sequencing platform by MGI Tech, China) for WGS. All 50 libraries were sequenced, based on 100 bp paired-end reads, on the MGI DNBSEQ-T7 platform (MGI Tech, China).
Somatic variant calling and generation of mutation signatures
Read pairs were aligned to the human reference genome NCBI GRC Build 37 (hg19) using Burrows-Wheeler Aligner (BWA MEM, http://bio-bwa.sourceforge.net/)36 and somatic mutations were identified by the Mutect237 variant caller with default parameters, following the standard GATK practices, including removal of PCR duplicates. Variants were subsequently annotated by Ensembl Variant Effect Predictor (VEP)38. TMB was estimated based on the proportion of nonsynonymous single nucleotide variants and short indels per coding megabase. Somatic mutational signatures were extracted using SigProfiler, an algorithm based on the 96 base substitution classification via nonnegative matrix factorization, and compared with COSMIC v3 set of signatures39.
NanoString gene expression profiling
We used the NanoString Pancancer IO360 panel (NanoString Technologies, Seattle, WA, USA) to interrogate gene expression, following manufacturer’s protocol using the nCounter platform. Briefly, RNA was extracted from five 10 micron sections on all samples with adequate tumor tissue available and purified with the FFPE RNA/DNA Purification Plus Kit (Norgen Biotek (Thorold, ON, Canada) according to manufacturer’s instructions, and analyzed using the 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA). After excluding samples with suboptimal RNA integrity and content, the remaining samples were included in the nCounter analysis. Including 38 samples that were previously investigated5, the final set of data passing QC (n = 67) were analyzed on the nSolver 4.0 Advanced Analysis module using default settings to derive differentially-expressed genes, pathway scores, and cell type scores. The panel was further analyzed using the Tumor Inflammation Signature (TIS) algorithm, which measures the level of immune infiltrate in a tumor and the tumor microenvironment13. This signature contains 18 genes related to antigen presentation, chemokine expression, cytotoxic activity, and adaptive immune resistance. A score is calculated as a weighted linear combination of the 18 genes’ expression values normalized to stable housekeeper gene expression. High TIS was defined as more than or equal to the median score in the cohort.
10X genomics visium platform
Samples were embedded in TissueTek O.C.T compound (Sakura Finetek USA) and flash frozen in isopentane (Sigma-Aldrich; Merck), placed in a liquid nitrogen bath. RNA was extracted from four micron sections cut from the sample blocks chosen for the assay with the RNeasy Mini Kit (Qiagen). RNA quality was assessed by RNA integrity number (RIN) measured using TapeStation 4150 (Agilent Technologies). Haematoxylin (Dako) and eosin (Dako) staining (H&E) was performed on the sections to determine the morphology of the tissues. Optimization of the permeabilization timings for each sample was performed using Visium Tissue Optimization Slides and Reagent Kits (10X Genomics) with 10 µM sections. For the gene expression assay, 10 micron sections from the respective sample blocks were placed onto a Visium gene expression slide (10X Genomics). H&E staining was done according to the manufacturer’s protocol. Permeabilization, reverse transcription, second strand synthesis and cDNA amplification was performed using the Visium Spatial Gene Expression Reagent Kit (10X Genomics). Dual indexed libraries were made with the Library Construction Kit (10X Genomics) and Dual Index Kit TT Set A (10X Genomics) according to the manufacturer’s protocol. The final libraries were assessed using the Agilent Bioanalyzer High Sensitivity DNA kit and chip (Agilent Technologies). Loupe Browser 6.0 was used to estimate the capture area covered by the tissue within each frame on the slide to calculate the sequencing depths required. The libraries were pooled and sent for paired-end dual-indexed sequencing on the Novaseq 6000 instrument (Illumina).
Analysis of spatial sequencing data
The reads were demultiplexed and mapped against the hg38 reference genome using 10X Space Ranger v.1.3.1 (10X Genomics, CA, USA) with the default parameters for automatic alignment. Using Seurat v4.0, spatial data were first loaded into count matrices, and spots which had less than 10% of transcripts mapping to mitochondrial genes were retained for the scaling and normalization of genes expression measurements using sctransform approach, before merging (merge, Seurat) them to do an integrative analysis of dimensionality reduction and clustering. UMAP embeddings were used to visualize the spots distribution using 30 precomputed principal components in FindNeigbors and RunUMAP as well as with default resolution of 0.8 in FindClusters. The celltype(s) of each spot was established with PanglaoDB Augmented 2021 database40, in addition to the gene expression profile of some cell type-specific markers (as indicated in Supplementary Fig. 3). Cell types with less than three spots, in each sample, were omitted from the downstream analysis. No batch-effect correction was performed to maintain the clustering architecture when overlaying on the tissue morphology of each sample.
Eighteen genes (PSMB10, CD276, PDCD1LG2, HLA-DQA1, NKG7, HLA-E, TIGIT, CD274, CXCL9, CMKLR1, LAG3, CD26, CXCR6, IDO1, CD8A, HLA-DRB1, CCL5, and STAT1) from the NanoStringIO360 panel were used to identify TIS signatures (AddModuleScore, Seurat) in characterizing immune hot and cold specimens. In addition, selected genes from the same panel (tumor cells: PECAM1 and ERG; fibroblasts: FBLN1, FAP, and DES; B-cells: CD79A and CD79B; CD8T-cells: CD8A and CD8B; NK-cells: KLRK1 and KLRD1; macrophages: CD14 and CD68; neutrophils: CSF3R and FPR1, and Dendritic cells: CD209 and CCL13) were used to annotate (AddModuleScore, Seurat) immune cells. Wilcoxon test, with Holm-Bonferroni correction, was conducted to statistically evaluate the differential expression of TIS genes in each cell type against the base mean of each sample.
Statistics and reproducibility
Comparisons of the frequencies of categorical variables were performed using Pearson’s Chi-squared tests or Fisher’s exact tests. Continuous variables were represented by Box–Whisker plots and their associations with categorical variables were evaluated by Mann–Whitney U tests. All statistical analyses were conducted assuming a two-sided test with significance level of 0.05 unless otherwise stated, and performed using MedCalc for Windows, version 18.2.1 (MedCalc Software, Ostend, Belgium).
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.