Further, we clinically validated expression of E-Cadherin (E-cadh), TCF21 and PARP1 in Class1 (CCM/E) tumors and Slug, AnnexinA2 (Anxa2) and hyaluronic acid in Class2 (EMT/M) tumors. Class 3 (E/M hybrid) tumors were either double positive (DP) or double negative (DN) for these biomarkers13,15. Clinical validation also revealed 'class switching', involving phenotypic plasticity and altered marker expression in response in different tumor niches and/or treatment. These observations together was the rationale to initiate extensive histopathological and molecular assessments to understand the extent of phenotypic heterogeneity, plasticity and stemness in the ID8 syngeneic mouse model. This would assign the model a precise, subtype-specific perspective for interpretation of experimental data towards its application in therapeutic evaluation.
A4 cell line previously established from HGSC patient ascites was cultured as reported. OVCAR3 (from the NCCS cell repository) and OVCA420 (a kind gift from Prof. Clements, TRI, Australia) were obtained and grown in RPMI-1640 supplemented with 10% FBS. ID8 cells were purchased from ATCC and grown in Dulbecco's Modified Essential Medium supplemented with 10% FBS. All cell lines were maintained in a humidified incubator at 37 °C under 5% CO2 atmosphere.
Immunostaining was performed as described previously. Briefly, cells were fixed with 4% paraformaldehyde, permeabilized with chilled methanol, blocked with 10% goat serum, incubated with primary antibody, followed by secondary antibody (Table S1) performed in triplicates (three individual experiments with cells of different passages and three technical replicates with the cells of same passage). Images were acquired (60 images from 5 fields per replicate) on Olympus FV3000 confocal microscope at 63× magnification; mean fluorescence intensity (MFI/ area) calculated on ImageJ. Live cell imaging for wound healing was performed as described previously. Briefly, a scratch generated on a confluent cell layer was imaged for healing in serum-free medium at 37 °C, 5% CO2 atmosphere over 48 h using an inverted laser scanning confocal fluorescence microscope (LPSS5, Leica Microsystems), performed in triplicates (individual experiments with cells of different passages and 3 replicates with cells of same passage).
6-8-week-old C57BL6 female mice, provided by NCCS Animal Experimentation Facility, Pune were housed in IVCs (2-4 mice per cage) according to the guidelines provided by Institutional Animal Ethics Committee [IAEC; EAF/2023/B-436(I)]. Primary tumors were developed subcutaneously (ID8-sc: 5 × 10 ID8 cells on right and left flank, n = 3) or orthotopically (ID8-ortho: 2 × 10 ID8 cells intra-bursal on right ovary, n = 5). A latency of ~ 4 months from date of injection for both was noted; ID8-ortho further progressed to metastasis. Temporal harvesting of peritoneal tumor ascites was performed 3 times under aseptic conditions (ID8-asc tap1, ID8-asc tap2, ID8-asc tap3) using a syringe and needle. H and E-stained images of xenografts were acquired, for 4 different mice of the same batch on Zeiss Axio Imager Z2 microscope using a 0.5 × 0.5 cm transparent grid, selecting upper right corner section in each grid, at 10 × and 40 × magnification. Number of fields counted were variable for each sample because of tumor size (details in Additional File 2). Training of Qupath algorithm was achieved in pre-stratified HGSC patient tumors (13 patients) and tested on a set of unstratified patient tumors (8 patients) (details in Additional File 2). The trained Qupath-based stratification was then extended to annotation and quantification of ID8 tumors.
RNA extraction and library preparation from 3 biological replicates of ID8 cells (control) and tumors (sample) was performed by MedGenome Labs Ltd employing modified NEBNext RNA Ultra II directional protocol; total RNA sequencing performed on Illumina HiSeqX (60 M, 2 × 150 bp reads/sample). Gene Set Enrichment Analysis (GSEA software v 2023.2) was applied to detect enriched MSigDB mouse gene sets, ~ 1000 "gene set" type permutations were performed with significant differentially expressed genes (p < 0.05, Fold Change > 2& < - 2), which returned pathways displayed as enrichment maps (p < 0.05).
Trypsin-digested protein preparations from 3 biological replicates of ID8 cells (control) and tumors (sample) were used to acquire mass spectra in positive ionization mode on Orbitrap Fusion Mass Spectrometer (Thermo-Fischer Scientific), and data subjected to label-free quantification (LFQ) in triplicate (ID8 cells of same passages and tumors from 3 different mice of same batch), as described previously, using Max Quant (v1.6.17.0) with Mus musculus proteome (Uniprot ID- UP000000589). LFQ > ratio counts 1 were selected to return quantified peptides from different protein groups, for triplicates (ID8 cells of same passages and tumors from 3 different mice of same batch). These were subjected to Perseus (v1.6.14.0) in triplicate, to identify significantly expressed proteins based on LFQ intensity (student 't' test; p ≤ 0.05, > 2 unique + razor peptides with sequence coverage percentage > 10%). This analysis resolved common and exclusive proteins (expressed in both or either control or test samples respectively) as represented in volcano plots (log2 fold-change ≥ 2 or ≤ - 2 vs. log10 p > 1.3). Further pathway enrichment analysis was achieved using ClueGO (v2.5.10) as a Cytoscape (3.9.1) plugin with GO-biological process, GO (Cellular Component, Immune System Process, Molecular Function), KEGG, Reactome (Pathways and Reactions) considering minimum 5 genes per term, Kappa score 0.4 cutoffs, p ≤ 0.05 for triplicates..
Immune Cell Abundance Identifier-mouse tool was used to deconvolute normalized read counts of ID8 tumor transcriptomes into 36 immune cells against mouse reference expression profiles and marker gene sets for triplicates (ID8 cells of same passages and tumors from 3 different mice of same batch). For flow cytometry based estimations, tumors were dissociated enzymatically using collagenase while ascites was processed for RBC lysis. Staining was performed as described. Briefly, single cell suspensions were processed for live/dead staining, blocked for 20 min followed by surface staining for 30 min with fluorophore tagged antibodies. Intracellular staining samples fixed and permeabilized for 30 min using 1 × Fix/Perm buffer followed by intracellular staining for 60 min with fluorochrome tagged antibodies (Table S2) for triplicates (Tumors from different mice of same batch and no. of batches were 3). Samples were acquired on Cytek Aurora on the same day. Compensation was performed on SpectroFlov3.0. Gating and UMAP plots were prepared on OMIQ software from Dotmatics (www.omiq.ai, www.dotmatics.com).
ID8 cells were grown till confluency and irradiated at 50 Gy. After 3 days of recovery Ecadherin expression was quantified through flow cytometry in comparison with non-irradiated ID8 cells in triplicates (3 individual experiments with different cell passages and 3 replicates with cells of same passage).
All experiments were performed in triplicates (3 individual experiments with samples in triplicates for each experiment). Data significance was assigned based on unpaired Student 't' test and Mann-Whitney Test. Data was used to generate graphs, PCA plots and Heatmaps using (GraphPad Prism 8.0.2, SRPlot and MeV 4 9 0 tool_ multiple arrays with median values as mid-range for the Heatmap respectively) and assigned significance as *p < 0.05, **p < 0.01, ***p < 0.001.