A. Bar graph representing the number of gene sets/pathways from the GSEA archived database that were enriched in hemangiosarcoma samples from Golden Retrievers versus hemangiosarcoma samples from non-Golden Retrievers. Each gene on the x-axis was present in the number of GSEA gene sets indicated on the y-axis (of 77 where FDR<0.05). B. Graphical representation of genes (x-axis) present in each GSEA pathway/gene set (y-axis), where a filled box means the gene was present and enriched in that GSEA pathway. Increasing red intensity reflects higher enrichment scores. The genes enriched in the highest number of gene sets are identified by name.

The effect of three VEGFR inhibitors on proliferation and viability of hemangiosarcoma cells was tested in vitro. The selectivity and half maximal inhibitory concentrations for Drugs 1, 2, and 3 are listed in the Materials and Methods. Cells (10,000/well) were plated in duplicate in a 96-well microtiter plates and allowed to attach for 16 hr prior to addition of inhibitors at the indicated concentration. Cells were then cultured for 72 hr, and the number of viable cells was determined using the MTS assay. Absorption at 490 nm for each well was averaged, and data normalized to % viability where the mean of wells that received no treatment (0 nm) was considered = 100%. The mean of two independent experiments is shown at drug concentrations of 100 nM. P-values were calculated using Student's T-test.

A. Hierarchical clustering of 6 Golden Retriever (GR) hemangiosarcoma samples versus 3 non-Golden Retriever (nGR) hemangiosarcoma samples (GEO series record GSE15086). Increasing green intensity indicates increased gene expression, increasing red intensity indicates decreased gene expression, and black indicates no change. Bars represent groups that cluster together. B. Gene differences between Golden Retrievers (GR) and non-Golden Retrievers (nGR) were plotted according to their cytogenetic location along the 38 autosomes and the X chromosome. Tick marks represent individual genes that show differential regulation, with the color intensity (green to black to red) representing expression changes as described in part A. Genes plotted received a p-value<0.05 and were derived from ANOVA analysis of the global list of filtered genes. C. Quantitative expression analysis of 3 genes found in panel A, TSP-3, DLA88 (MHC), and SMARCA1 (SWI/SNF) that were differentially expressed between Golden Retrievers with hemangiosarcoma and other breeds with hemangiosarcoma. Samples were evaluated for gene expression changes by RT-PCR followed by qPCR. One sample originating from a non Golden Retriever dog (Dal-4) was normalized to 1.0 and used as a reference; gene expression is presented as fold change compared to the reference sample. The samples used for real time PCR analysis (in the order presented) include CHAD P9, Dal-4, Joey, DD1, CHAD G6, CHAD G4, and Frog. D. Schematic representation of gene expression changes between Golden Retrievers and non Golden Retrievers with hemangiosarcoma grouped by biological function using ONTO/express gene ontology program (vortex.cs.wayne.edu/Projects.html).

A. Hemangiosarcoma cells from 4 Golden Retrievers (in order from top to bottom, Frog, Veronica, Tucker, Emma) and from 2 non-Golden Retrievers (Dal-4 and Joey) were cultured in chamber slides and stained with antibodies against VEGFR1 and VEGFR2 as described in the methods. Staining was visualized using epifluorescence. Bar = 20 µm. B. Emma, Frog, Joey, and Dal-4 cells obtained during the log growth phase were used to quantify expression of VEGFR1, VEGFR2 and ß-actin by immunoblotting. Conditions were optimized for linearity. Densitometric band quantification was done using Image J 1.37. Data are normalized to ß-actin using the sample with the highest expression for each receptor as the calibrator. C. Dal-4 cells were cultured in complete media supplemented with VEGF (+), in the absence of serum and growth supplements (−), with or without Drug 3 (100 nM) or Drug 1 (1 µM) as indicated. The activation status of VEGFR1 and VEGFR2 was examined using modification-state (phosphospecific) antibodies directed against pVEGFR1-Tyr915 and pVEGFR2-Tyr875. ß-actin was used as a loading control, and HUVEC lysates were used as a specificity control for VEGFR1.