Identification and characterization of angiogenesis targets through proteomic profiling of endothelial cells in human cancer tissues

Abstract

Genomic and proteomic analysis of normal and cancer tissues has yielded abundant molecular information for potential biomarker and therapeutic targets. Considering potential advantages in accessibility to pharmacological intervention, identification of targets resident on the vascular endothelium within tumors is particularly attractive. By employing mass spectrometry (MS) as a tool to identify proteins that are over-expressed in tumor-associated endothelium relative to normal cells, we aimed to discover targets that could be utilized in tumor angiogenesis cancer therapy. We developed proteomic methods that allowed us to focus our studies on the discovery of cell surface/secreted proteins, as they represent key antibody therapeutic and biomarker opportunities. First, we isolated endothelial cells (ECs) from human normal and kidney cancer tissues by FACS using CD146 as a marker. Additionally, dispersed human colon and lung cancer tissues and their corresponding normal tissues were cultured ex-vivo and their endothelial content were preferentially expanded, isolated and passaged. Cell surface proteins were then preferentially captured, digested with trypsin and subjected to MS-based proteomic analysis. Peptides were first quantified, and then the sequences of differentially expressed peptides were resolved by MS analysis. A total of 127 unique non-overlapped (157 total) tumor endothelial cell over-expressed proteins identified from directly isolated kidney-associated ECs and those identified from ex-vivo cultured lung and colon tissues including known EC markers such as CD146, CD31, and VWF. The expression analyses of a panel of the identified targets were confirmed by immunohistochemistry (IHC) including CD146, B7H3, Thy-1 and ATP1B3. To determine if the proteins identified mediate any functional role, we performed siRNA studies which led to previously unidentified functional dependency for B7H3 and ATP1B3.

Figure 1. A) Schematic representation of ECs isolation from human kidney tissues.

B) For isolation of pure EC populations from collagenase dispersed tissues, the EpCAM negative, CD146 positive endothelial content of cell population were sorted, isolated and enriched by a MoFlo cell sorter. Cells were submitted for LC-MS and feature detection either directly or indirectly after expansion in culture.

Figure 2. Endothelial Cell Content of Tissues.

Single cells following tissue processing from tumor and normal adjacent from kidney, lung, and colon tissues were stained with anti-CD146 Ab and analyzed for EC presence.

Figure 3. Mass Spectrometry (MS) analysis and quality control.

Scatter plots of common features from normal and tumor samples and differential analysis using decoupled maps are shown. Relative expression levels are determined by aligning peptide ion features from different MS experiments (LC/MS maps). The log2 intensities of common features detected in (A) process replicates of the control sample; (B) process replicates of the tumor sample; (C) control vs. tumor samples are plotted. The corresponding box plot (in log2 scale) of ratio distribution for each dataset is shown next to the scatter plot. The box contains 50% of the features, where 95% of the features are within the horizontal bars. (D) Prevalence of well known EC surface proteins was investigated and representatives are shown.

Figure 4. Global peptide analysis identified over-expressed proteins in kidney tumor tissue endothelium.

The heat map presents the analysis of peptide intensities for the 233 peptides identified in kidney tumor endothelium and sorted so that the most differentially-expressed peptides are at the top. The display colors were determined for each row separately by assigning black to the median intensity in the row, green to the lowest intensity in the row, and red to the highest intensity.

Figure 5. Phenotypic characterization of Ex Vivo cultured endothelial cells.

(A) Intense punctate fluorescence demonstrating the uptake of DiI-Ac-LDL in normal lung microvascular ECs. (B) Cultured normal and tumor tissue-derived ECs from colon and lung were stained with EC specific markers CD146 and CD31, and subjected to flow cytometry.

Figure 6. MS analysis of CD146 in tumor ECs.

The top panel shows the extracted ion chromatograms of kidney ECs samples from normal and tumor samples. The mass spectra are shown in the lower panel. The red arrow shows the intensity level in normal tissue sample.

Figure 7. Overexpression of CD146, B7H3, Thy-1 and ATP1B3 proteins in tumor cells and/or endothelial cells is confirmed by IHC.

(A) IHC images of normal and carcinoma samples. Over-expression of CD146 is found in tumor cells while normal epithelium was uniformly negative in multiple oncology indications. (B) Images indicate significant over-expression of CD146 in ECs of carcinoma vessels compared to normal samples. (C) Significant over-expression of B7H3 in ECs of carcinoma vessels compared to normal samples is depicted. (D) Significant over-expression of Thy-1 in carcinoma vessels compared to normal is shown. (E) IHC images of ECs indicate significant over-expression of ATP1B3 in carcinoma vessels compared to normal samples.

Figure 8. Phenotypic characterization of cellular content of tumors.

(A) Over-expression of CD146 in kidney tumor tissue cells was confirmed by flow cytometry. (B) and (C) respectively confirm B7H3 and Thy-1 over-expression in kidney and lung tumor tissue endothelia by flow cytometry.

Figure 9. Functional activity of selected targets.

(A) mRNA knockdown of B7H3 inhibits proliferation in HUVECs in a titration dependent manner. A representative screen and a titration experiment with duplexes 2 and 3 in HUVECs are shown. Proliferation was monitored by using the Alamar Blue assay, and SPA [3H]. Thymidine uptake assay system. (B) mRNA knockdown of ATP1B3 inhibits proliferation and increases apoptosis in HMVECs in a titration dependent manner. A titration experiment with duplexes 2 in HMVECs is shown. Proliferation was monitored by using the Alamar Blue assay, and apoptosis was measured by caspase 3/7 activity.