Supplementary Figures
Supplementary Table. 1Cell Line Growth Conditions
Cell lines were maintained at 37°C and at 5% CO2 in the recommended growth media with the indicated supplements; Foetal Calf Serum (FCS), Non-essential amino acids (NEAA), L-glutamine (LG), Sodium Pyruvate (Na Pyr) and Oxaloacetic acid (OAA)
Supplementary Fig.1 18s quantification of stromal quantities of xenograft models
To enable tumour (Human) and stromal (mouse) mRNA normalisation species specific 18s ARMS primers were designed to differentiate between human and mouse 18s rRNA (A) Assessment of the individual species specific 18s rRNA primers using standard curves of known synthetic mixes of QPCR Human Reference Total RNA and QPCR Mouse Reference Total RNA. Plots are of experimentally calculated (described in materials and methods) and expected proportions of human and mouse mRNA masses of mixed species standard curve (B). Calculated % proportions of human (grey) and mouse (black) transcript for the xenografts profiled in this study. Each xenograft model has n=5 biological replicates excluding OVCAR which n=4
Supplementary Fig 2.Quantification of Neutrophil infiltrate of the xenograft panel
Tumour Micro Cores (TMAs) were generated for the indicated xenografts and stained forGr-1.Staining was quantified as described in materials and methods. Xenografts are ordered for each graph according to their EMT phenotype
Supplementary Fig 3.Characterisation ofvascular and stromal architecture
Tissue microarrays (TMAs) generated from human xenografts. Each tumour was represented by 12 x 1.6mm diameter cores; 4 independent xenografts per tumour model with 3 cores per tumour. CD31:α-SMA IF stained TMAs were scanned into MIRAX scan (Zeiss) and analysed manually using the MIRAX viewer v 9.0 (Zeiss). Red; -SMA, Green, CD31 and Blue is DAPI. Models are split into high (>15%), medium (5% to 15%) and low (<5%) αSMA stinging; M-Mesenchymal; E-Epithelial; I-Metastable. Bar is approximately 100m
Supplementary Table. 2Linear regression analysis of the quantified histological staining for αSMA, F4/80 and MVD.
Displayed are the R2 and p-value for the indicated correlations
Supplementary Fig 4: Detection of tumour (human) genes in vitro and in vivo.
Cell Lines mRNA was isolated and profiled on the Human TaqMan® Array Micro Fluidic Cards and relative expression calculated. Assays with at CT>35, the confidence limit of the technology, or not detected were classed as absent data. The shading indicates if assays were not detected, detected in vivo only, in vitro only or in both in vitro and in vivo in the indicated models,
Supplementary Fig.5PCA analysis of tumour component normalised to total 18s
Principle components analysis (PCA) analysis was performed on the of the tumour profile normalised to total 18s. Data points represent individual biological replicates and the colours identify EMT phenotype.
Supplementary Figure 6;Comparison of cell line and xenograft transcript profile
PCA analysis was performed on the cell line (A) and xenograft (B) transcript profile as described in materials and methods and coloured according to their invivo ‘Epithelial to Mesenchymal transition’status. Blue, Mesenchymal; Red, Epithelial; Green, Metastable; Brown, Not Assessed.
Supplementary Table. 3Linear regression analysis of human gene expression vs. histological parameters (normalised to human 18S rRNA)
Linear regression was used to correlate tumour (human) gene expression normalised to human 18S rRNA with the MVD, F4/80 and -SMA staining. The average expression of tumour transcripts in the individual xenografts models was used for this correlation; genes are listed in acceding order of p-value and normalised to human component
Supplementary Table.4Linear regression analysis of human gene expression vs. histological parameters (normalised to total 18S rRNA)
Linear regression was used to correlation tumour (human) gene expression with the MVD, F4/80 and -SMA staining. The average expression of tumour transcripts in the individual xenografts models was used for this correlation; genes are listed in acceding order of p-value and normalised to total 18S rRNA
Supplementary Fig. 7 Correlations of key tumour derived ligand transcripts with protein
Protein levels of the indicated soluble ligands were established for each xenograft as described in the supplementary methods.Displayed is thelinear regression analysis of tumour transcripts levels normalised to human 18s and protein expression levels within the same xenografts Protein and mRNA levels are averaged across 4-5 biology replicates
Supplementary Table. 5 Linear regression analysis of ratio of pro- and anti-angiogenic factors and their correlations to MVD.
The ratio of expression between the indicted anti-angiogenic factors and all the tumour derived transcripts was calculated. Linear regression was performed to determine if there were any significant correlations between the expression ratio and MVD. Displayed are the R2 and p-values for the individual analyses.
Supplementary Table. 6Linear regression analysis of mouse gene expression vs. histological parameters (normalised to total 18S rRNA)
Linear regression was used to correlation stromal (mouse) gene expression with the MVD, F4/80 and -SMA staining. The average expression of stromal transcripts in the individual xenografts models was used for this correlation normalised to total 18s; genes are listed in acceding order of p-value and normalised to total mRNA
Supplementary Table. 7Linear regression analysis of mouse gene expression vs. histological parameters (normalised to mouse 18S rRNA)
Linear regression was used to correlate stromal (mouse) gene expression normalised to murine 18S rRNA with the MVD, F4/80 and -SMA staining. The average expression of stromal transcripts in the individual xenografts models was used for this correlation; genes are listed in acceding order of p-value and normalised to the murine component
Supplementary Figure 8. PCA analysis of stromal (mouse) transcript profile normalised to total 18s
A. Stromal expression profiles were obtained usingmouse-specific assays onTaqMan® Array Micro Fluidic Cardsand normalised to the total 18S rRNA level. All models have n=5 biological replicates except for OVACR-3 xenograft which has n=4 and principal components analysis (PCA) analysis was performed. Data points represent individual biological replicates.The colours identify individual tumours of the same xenograft. B. PCA loading plot highlighting the genes contributing to the scores plot.
Supplementary Table. 8Xenograft response to cediranib treatment at 1.5mg/kg
Mice were dosed orally once daily with cediranib at 1.5 mg/kg and the average tumour growth inhibition is displayed for the indicated models.
Supplementary Figure 9: Detection of stromal (mouse genes) in vivo
Xenograft mRNA was isolated and profiled on the Mouse TaqMan® Array Micro Fluidic Cards and relative expression calculated. Assays with at CT>35, the confidence limit of the technology, or not detected were classed as absent data. The shading indicates if assays were not detected or detected in vivo
Supplementary methods
Cell Line mRNA extraction
Cell Lines were grown as described in materials and method and were tryposined at 37°C for 5 minuets. mRNA was isolated from 1x107 cells using RNeasy Mini Kit (Qiagen) with addition of QIAShredder (Qiagen) as per the manufactory’s instructions. On-column DNase digestion was performed using the RNase-free DNase Kit (Qiagen). RNA concentration was measured using the NanoDrop ND1000 (NanoDrop), RNA integrity (RIN) was assessed using the RNA 6000 Nano Assay (Agilent Technologies), RIN values for all sample were between 7 and 10.
MSD Analysis
Frozen tumours were homogenised in RIPA buffer (50mM Tris-HCL pH7.4, 1% NP-40, 0.25% Na-deoxycholate, 150mM NaCl, 1mM EDTA, 1mM NaF, phosphatase and protease inhibitors) and spun to remove debris at 12000rpm for ten minutes at 4C and protein concentration was equalised to 0.8mg/ml. MSD Human Growth Factor 4-plex, (Meso Scale Discovery; K11029C) measuring VEGF-A, FGF-2, PLGF and soluble VEGFR1 (sFlt-1), MSD Human Demonstration Cytokine 7-plex (Meso Scale Discovery, K15003A) measuring IL-6, IL-8, IL-1, TNF-α, IL-2, IL-5 and GM-CSF, mouse VEGF (Meso Scale Discovery, K110BMB), were performed as per the manufactures instructions. All plates were analysed on the MSD SECTOR 6000 plate reader.
-SMA and CD31 co-immunofluorescence
-SMA and CD31 co-immunofluorescence was performed as previously described (21). CD31:α-SMA IF stained TMAs were scanned into MIRAX Scan (Zeiss) and analysed manually using the MIRAX viewer v 9.0 (Zeiss).
1