Patient cohort and selection criteria
A sarcoma patient sample series of 543 tissue samples from nine different sarcoma types (fibrosarcoma, gastrointestinal stromal tumour (GIST), liposarcoma (LPS), leiomyosarcoma, malignant peripheral nerve sheath tumour, myxofibrosarcoma, sarcoma with unspecified histotype, and synovial sarcoma) was collected from the archives of the Department of Pathology, of the Helsinki University Hospital (HUH). A series of 179 LPS tissue samples was collected from the archives the Department of Pathology, and corresponding clinical data were extracted from the patient registries of HUH. The ethics committee of Helsinki University Central hospital approved the study and VALVIRA, the national supervisory authority of health and welfare, provided permission to use patient tissue samples.
RNA extraction
Tissue sections (2 x 10 µm) from FFPE samples were collected to 2-ml micro tubes (Cat. No. 72.693.005, Sarstedt, NRW, Germany) and deparaffinized with 160 µl of deparaffinization solution (Cat. No. 939018, QIAGEN, NRW, Germany). Samples were mixed by vortexing and subsequently centrifuged and incubated at 56°C for 3 min. 120 µl of buffer PKD (Cat. No. 1034963, QIAGEN) was added, mixed by vortexing and centrifuged 13,000 × g for 1 min. 10 µl of proteinase K (Cat. No. 19133, QIAGEN) was added to the lower liquid phase of the samples. After a 1-hour incubation at 56°C in a tubeshaker, samples were spinned down and RNA was extracted with QIAsymphony RNA Kit (Cat. No. 931636, QIAGEN) using a QIAsymphony® SP instrument (QIAGEN). RNA quality and concentration was measured with 2100 Bioanalyzer instrument (Agilent, CA, USA).
Total RNA from cell lines was extracted with NucleoSpin® RNA (Cat. No. 740955.50, MACHEREY-NAGEL, NW Germany) following the manufacturer’s instructions.
RNA sequencing
Extracted RNA from patient samples were 3’ sequenced. After adding 1 µl of ERCC RNA spike-in control (1:1,000) to each sample, an RNA sequencing library was prepared with QuantSeq 3’ mRNA Library kit (Lexogen, Vienna, Austria) following the manufacturer’s instruction as previously described [42]. Sequencing was performed with HiSeq 2500 System (Illumina, CA, USA) using high-throughput mode and v4 chemistry.
Sequenced mRNA from 152 samples was processed with the data analysis pipeline BlueBee Genomics platform (BlueBee Holding BV, NB, the Netherlands) integrated in the QuantSeq 3’ mRNA Library kit following the manufacturer’s instructions.
Processing of Sequencing Data
Htseq-count files from Bluebee sequencing process were read into R [R Core Team (2022). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/] (version 4.1.2) and were matched to clinical data per sample. Gene annotation was retrieved from AnnotationHub (snapshotDate: 2021-10-20). Data were imported into DGEList object by using edgeR package [43] (version 3.36.0). Genes where further filtered by filterByExpr function (default parameters) from edgeR. log2 counts per million were calculated with cpm function (default parameters) from edgeR. Additional sample quality control was performed by excluding samples with median logcount with deviation > 10% from the median logcount over all samples (21 samples were excluded resulting in 131 samples for downstream analysis).
Identification of DEG between LPS subtypes
A differential expression test of genes between LPS subtypes was performed by using estimateDisp and exactTest function from edgeR with default parameters. DEGs were defined as those genes with a positive logFC > 1.5 (with P value < 0.05) between one against all other subtypes, resulting in three lists of DEGs. Altogether, we identified 97 DEGs between low-high-dediff versus the other two, 247 DEGs between myx/rc verses the other two, and 37 DEGs between pleo versus the other two.
IST reference data fetching and processing
Data from IST database was fetched into R (4.1.2) by using Dplyr extension for SQlite. We calculated mean expression for each gene (excluding genes with < 6 datapoints) per each healthy anatomical tissue type (n = 95) and per each cancerous tissue type (n = 106) [25]. For each tissue type (both healthy and cancerous), we then chose the top 500 expressing genes. The final IST reference gene list was calculated by choosing a unique list of genes where each gene did not appear in more than one healthy tissue top 500 lists and in more than three cancer tissue top 500 lists. This IST reference list contained in total 2754 genes, and was used as an exclusion list to eliminate non-LPS specific genes from LPS subtype DEG analyses. Thus, genes from this list were filtered away from DEG results.
Quantitative real-time PCR of LPS tissue RNA
PDE3A and SLFN12 mRNA levels were measured from 63 FFPE samples. RNA extracted from FFPE cell line blocks were used as references. cDNA was synthesized from RNA with SuperScript™ IV VILO™ Master Mix (Cat. No. 11756500, Invitrogen™, Thermo Fisher Scientific, MA, USA). Primers and hybridization probes (Universal ProbeLibrary Set, Human Probes #1 - #90, Cat. No. 04683633001 Roche, Basel, Switzerland) were designed using ProbeFinder version 2.53 for Human. The primers were as follows: PDE3A, F: 5’-AAA GAC AAG CTT GCT ATT CCA AA-3’, R: 5’-GTG GAA GAA ACT CGT CTC AAC A-3’; SLFN12, F: 5’-CTT TGT TCA ACA CGC CAA GA-3’, R: 5’-ATG CAG TGT CCA AGC AGA AA-3’, and YWHAZ, F: 5’-CGT TAC TTG GCT GAG GTT GC-3’, R: 5’-TGC TTG TTG TGA CTG ATC GAC-3’. Real-time qPCR was performed with LightCycler® 480 Probes Master Kit (Cat. No. 04887301001, Roche) under the following conditions: 95°C for 10 min; 45 cycles of 95°C for 30 s, 60°C for 30 s, and 72°C for 45 s with plate read (amplification); 72°C for 7 min; 65°C to 95°C (increment 0.01°C s-1, melt curve). PCR was performed in triplicates for each sample and primer. Relative PDE3A and SLFN12 mRNA expression was calculated using the ΔΔCt method and normalized with YWHAZ expression.
Drug – gene interaction
Drug Gene Interaction Database (DGIdb) available from DGIdb website (https://www.dgidb.org/) [44] was used to identify compounds that potentially target LPS subtype specific genes from the transcriptome data. The compound search was conducted 2022-09-30.
Immunohistochemical staining
Tumour tissue microarray (TMA) blocks were prepared in Helsinki Biobank. Blocks included two 1-mm diameter core samples of same tumour tissue block from representative tumour areas. TMAs were generated with an automated tissue microarrayer (TMA Grand Master, 3DHISTECH Ltd., Budapest, Hungary). TMA and whole section slides were generated by cutting 3.5 µm sections on positively charged slides.
Slides were first incubated at 56°C for 1 hour before deparaffinized with Tissue-Tek DRS 2000 Automated Slide Stainer (Sakura Finetek Japan, Tokyo, Japan). After blocking endogenous peroxidase, antigen retrieval was performed with EnVision FLEX Target Retrieval Solution Low pH (pH 6, Cat. No. K8005, Agilent) in a decloaking chamber (Biocare Medical, CA, USA) at 110ºC for 3 min. Slides were then washed and incubated with polyclonal rabbit anti-PDE3A (1:100, Cat. No. HPA014492, Sigma-Aldrich, MO, USA) in WellMed BrightDiluent for 30 minutes (Cat. No. BD09-500, ImmunoLogic, Duiven, The Netherlands). WellMed BrightVision (one component detection system, Goat Anti-Rabbit IgG HRP, Cat. No. DPVR500HRP, ImmunoLogic) was used for antibody detection and slides were stained with ImmPACT® DAB Peroxidase (HRP) substrate kit (Cat. No. SK-4105, Vector Laboratories, CA, USA). Tissues were counterstained with haematoxylin (Mayer’s, Cat. No. S3309, Agilent). After dehydration process and mounting, images were required with Pannoramic 250 Flash III (3DHISTECH), and analysed with Slide Viewer (3DHISTECH). Healthy colon tissue was used as a positive control for PDE3A staining. The intensity of PDE3A staining was scored as 0 (no staining in tumour cells), 1 (weak), 2 (moderate) or 3 (strong). Visual presentation of scoring in different sarcoma types is shown in Supplementary Fig. S1. In whole-section staining of LPS series, PDE3A expression was evaluated as H-scores (staining intensity (0–3) × percentage of positive cells), which was further divided into quantiles (0: 0–75; 1: 76–150; 2: 151–225; 3: 226–300). To analyse statistically significant differences of PDE3A expression in different LPS subtypes, samples were divided into two groups based on H-scores (0: 0–150, and 1: 151–300).
Cell culture
GOT3, MLS1765-92, and MLS402-91 cell lines were established and kindly provided by Prof. Pierre Åman (University of Gothenburg); GIST48, GIST882 and LPS141 cell lines were kindly provided by Dr. Jonathan Fletcher (Harvard Medical School); and the SA4 cell line was kindly provided by Dr. Kjetil Boye. The GIST-T1 cell line (Cat. No. PMC-GIST01C) was purchased from Cosmo Bio; the 93T449 (Cat. No. CRL-3043), 94T778 (Cat. No. CRL-3044), and SW872 (Cat. No. HTB-92) cell lines were purchased from ATCC. GIST882 and GIST48 cell lines were verified by sequencing to contain known mutations; these were KIT exon 13 p.K642E in GIST882, and KIT exon 11 p.V560D and exon 17 p.D820A in GIST48. SA4 cell line was verified not to be HeLa contaminated by western blot where HPV E7 protein was not detected. All cell lines were tested negative for Mycoplasma, and were cultured in a humidified, 5% CO2 atmosphere at 37°C. GIST-T1 cells were cultured in Dulbecco’s modified Eagle’s medium (Cat. No. 11965092, Life Technologies, Thermo Fisher Scientific) with 10% fetal bovine serum (FBS, Gibco, Thermo Fisher Scientific), 100 U/mL of penicillin, 100 U/mL of streptomycin, and 0.03 mg/mL of L-glutamine (Pen Strep Glut, Cat. No. 10378-016, Life Technologies, Thermo Fisher Scientific). Other cell lines were cultured in Gibco™ RPMI Medium 1640 (Cat. No. 11530586, Life Technologies, Thermo Fisher Scientific) with Pen Strep Glut and 5- (MLS402-91 and MLS1765-92), 10- (SA4, GOT3, 93T449, and 94T778), or 20% (LPS141, SW872, GIST48, and GIST882) FBS. Cell images were acquired with a CMEX-5Pro camera (euromex, Arnhem, The Netherlands) and DC.1355 lens adapter (euromex) through ImageFocus Alpha (euromex).
Cell blocks with plasma-thrombin method
After detaching, washing, and pelleting the cells, 2 ml of PBS was added on top of the cell pellet for 2 min. Cells were then resuspended and divided into two tubes and centrifuged at 9,600 × g for 1 min. Supernatant was removed and the pellet resuspended in 45 µl of human plasma (separated 2,000 × g for 10 min), after which 45 µl of thrombin (from bovine plasma, 50 NIH-U/mg, CAS No. 9002-04-4, Merck KGaA, HE, Germany) was added and mixed. After a 1-min incubation, the clot was detached with 1 ml of 5% formalin, transferred to a larger tube and 6 ml formalin was added on top. After two-hour fixation, clots were transferred to casettes and treated through dehydration process, and incubated overnight at + 60ºC. RNA was extracted from FFPE cell blocks with High Pure FFPE RNA Micro Kit (Cat. No. 04823125001, Merck KGaA) following the manufacturer’s instructions. Incubation with proteinase K was performed overnight at 55°C.
Cell-line RT-qPCR
Cell-line RNA was synthesized to cDNA with iScript cDNA Synthesis Kit (Cat. No. 1708891, Bio-Rad, CA, USA). RT-qPCR was performed with SensiFAST™ SYBR Lo-ROX Kit (Cat. No. BIO-94005, BIOLINE, London, UK) with the following primers: PDE3A, F: 5’-GCT CCG GAG CTC TCG GAA A-3’, R: 5’-CCA GCA GCG CCA GCA GAA A-3’; SLFN12, F: 5’-ATC TGG GCT TGC AAG AGA AC-3’, R: 5’-TTT TTG CCA GCT TCT GCT TT-3’, and GAPDH, F: 5’-TCA AGA TCA TCA GCA ATG CC-3’, R: 5’-CGA TAC CAA AGT TGT CAT GGA-3’. The RT-qPCR run was performed as follows: 95°C for 2 min; 40 cycles of 95°C for 5 s, 60°C for 10 s with plate read and 72°C for 10 s (amplification); 72°C to 95°C (increment 0.05°C s-1, melt curve). Three replicates per sample were used and results were obtained from two separate experiments. Relative PDE3A and SLFN12 mRNA expressions was calculated using the ΔΔCt method and normalized to that of GAPDH.
Western blotting
Cells were lysed in M-PER™ Mammalian Protein Extraction Reagent (Cat. No. 78501, Thermo Fisher Scientific) containing HALT™ protease and phosphatase inhibitor cocktails (Cat. No. 78429 and Cat. No. 78420, Thermo Fisher Scientific). When immunoblotting apopototic markers, floating cells in media were also collected. Denatured protein lysates were subjected to SDS-PAGE using Mini-PROTEAN® TGX™ Precast Gel (Cat. No. 456–1034, Bio-Rad) and blotted to PVDF membranes (Mini Format, Cat. No. 1704156, Bio-Rad) with a Trans-Blot® Turbo™ Transfer system (Bio-Rad). After blocking, membranes were incubated in primary antibodies at 4°C overnight. Polyclonal goat anti-rabbit IgG (1:10,000, Cat. No. 111-035-003, Jackson ImmunoResearch, PA, USA) was used as secondary antibody. Antigens were detected with SuperSignal™ West Pico PLUS chemiluminescent substrate (Cat. No. 34580, Thermo Fisher Scientific) on autoradiography films, and quantitated using Fiji ImageJ 1.53 (64-bit). PDE3A was quantified by summing the intensities of all three isoforms. Antibodies used for western blotting were polyclonal rabbit anti-PDE3A (1:1,000, Cat. No. HPA014492, Sigma-Aldrich), monoclonal rabbit anti-SLFN12 (1:1,000, Cat. No. ab234418, Abcam, CB, UK), polyclonal rabbit anti-caspase-3 (1:1,000, Cat. No. 9662, Cell Signaling TECHNOLOGY®, MA, USA) and polyclonal rabbit anti-PARP (1:1,000, Cat. No. 9542S, Cell Signaling TECHNOLOGY®). Target protein expressions was normalized to polyclonal rabbit anti-cytoskeletal Actin (1:150,000, Cat. No. A300-491A, Bethyl Laboratories, TX, USA).
Cell viability and cytotoxicity experiments
The following drugs were used in drug efficacy experiments: anagrelide hydrochloride (CAS No. 58579-51-4, Cat No. 735554, Lancrix, Shanghai, China), cilostazol (CAS No. 73963-72-1, Cat. No. C0737, Sigma-Aldrich), A-1155463 (CAS No. 1235034-55-5, Cat. No. 27369, Cayman Chemical, MI, USA), navitoclax (CAS No. 923564-51-6, Cat. No. S1001, Selleckchem, TX, USA), DNMDP (CAS No. 328104-79-6, Cat. No. HY-W028690, MedChemExpress, NJ, USA), and 3-hydroxy Anagrelide (CAS No. 733043-41-9, Cat No. sc-206647, Santa Cruz Biotechnology, TX, USA). All drugs were dissolved in DMSO, and a corresponding DMSO concentration in cell media was used as control treatment.
Cells were seeded at 3,500 (GIST-T1), 8,000 (GIST48), 10,000 (GIST882), 3,000 (SA4), 1,800 (GOT3) and 3,000 (MLS402-91) cells/well on a Falcon® 96 Well White/Clear bottom microplate (Cat. No. 353377, CORNING, NY, USA) and allowed to attach for 24 hours. The number of cells plated in wells differed between cell lines in order to attain 70–80% confluency in the control-treatment on the final measuring day. Cell viabilities were measured with CellTiter-Glo® Luminescent Cell Viability Assay (Cat. No. G7572, Promega, WI, USA) following the manufacturer’s instruction. Luminescence was recorded with Hidex Sense microplate reader (Hidex Oy, Turku, Finland). Medium backgrounds were excluded from results, and 8 well –replicates per treatment were always used.
The synergy results presented were obtained from two separate experiments. Luminescence-measurements were divided by the average of control treatment, indicating a relative cell viability change percentage. For sequential treatment, cells were treated with first drug for 48 hours, after which media with second drug was changed for 48 hours.
Cytotoxicity was measured with LIVE/DEAD™ Viability/Cytotoxicity Kit (Cat. No. L3224, Invitrogen™, Thermo Fisher Scientific) according to the manufacturer’s instructions. Images were acquired with automated Countess 2 FL cell counter (Thermo Fisher Scientific). Both attached cells and detached floating cells were collected for analysis. Fluorescence-positive cells were then automatically counted using Fiji ImageJ 1.53 (64-bit) of single-color images (no plugin required). Presented results were obtained from two separate experiments; control treatment averages were used as reference.
Drug Sensitivity and Resistance Testing
A high-throughput drug sensitivity and resistance test (DSRT) with 527 approved or investigational anticancer compounds was performed in eight LPS cell lines (Supplementary Table S2) as previously described [26]. Briefly, the compounds were dissolved in DMSO or milli-Q water and dispensed into Corning® 384-well plates (Cat. No. CLS3573, Merck KGaA) by an Echo 550 (Labcyte Inc., CA, USA) liquid handler covering a 10,000-fold concentration range. Plates were stored in pressurized StoragePods® (Roylan Developments Ltd., Sy, UK) prior to plating cells (1,000 cells/well). Cells were then incubated with the compounds for 72 hours at 37°C, after which cell viabilities were measured using CellTiter-Glo® and microplate reader PHERAstar® (BMG LABTECH, BW, Germany). Drug responses were evaluated by using a calculated Drug Sensitivity Score (DSS) [45]. Inhouse data analysis tool Breeze [46] was used to calculate DSS, IC50 and other values.
Drug combination testing with anagrelide and the 12 selected drugs was performed with a 8 × 8 concentration matrix including negative control (DMSO), and cell viabilities were measured with CellTiter-Glo® Luminescent Cell Viability Assay (Promega). Synergy was assessed using the web-application SynergyFinder [38] and zero interaction potency model (ZIP-score) with inhibition for readout and yes for baseline correction as parameters. The drug combination was considered synergistic when the ZIP synergy score is positive. Negative values indicated an additive interaction.
Statistical analysis
Statistical analyses were performed using IBM SPSS Statistics 28.0.0.0 (IL, USA). A significance level of 0.05 was used. Statistically significant differences of PDE3A immunohistochemical staining (H-score low (1 and 2) and high (3 and 4)) between the three main LPS subtypes (WD-/DDLPS, MLPS, and PLPS) were analysed with crosstabulation and Pearson χ2 test (asymp. sig., two-sided). Correlation between PDE3A IHC H-score and qPCR expression results was analysed by using Spearman’s Rank-Order Correlation. Statistical significance was analysed using independent-samples Kruskall-Wallis test to compare PDE3A and SLFN12 mRNA expression levels between LPS subtypes DDLPS, MLPS and PLPS. Differences in cell viability across treatments were analysed using two-tailed unpaired t-test.