Identification of kinases activated by multiple pro-angiogenic growth factors

Abstract

Antiangiogenic therapy began as an effort to inhibit VEGF signaling, which was thought to be the sole factor driving tumor angiogenesis. It has become clear that there are more pro-angiogenic growth factors that can substitute for VEGF during tumor vascularization. This has led to the development of multi-kinase inhibitors which simultaneously target multiple growth factor receptors. These inhibitors perform better than monotherapies yet to date no multi-kinase inhibitor targets all receptors known to be involved in pro-angiogenic signaling and resistance inevitably occurs. Given the large number of pro-angiogenic growth factors identified, it may be impossible to simultaneously target all pro-angiogenic growth factor receptors. Here we search for kinase targets, some which may be intracellularly localized, that are critical in endothelial cell proliferation irrespective of the growth factor used. We develop a quantitative endothelial cell proliferation assay and combine it with "kinome regression" or KIR, a recently developed method capable of identifying kinases that influence a quantitative phenotype. We report the kinases implicated by KIR and provide orthogonal evidence of their importance in endothelial cell proliferation. Our approach may point to a new strategy to develop a more complete anti-angiogenic blockade.

Keywords: cell growth; kinase inhibitors; kinome regularization; polypharmacology; proangiogenic growth factors; tumor vascularization.

FIGURE 1

Establishment of the Experimental System. (A). Demonstration of Dermal Microvascular Endothelial Cell (DMEC) labeling. Fluorescent image of nuclear-localized mCherry overlaid on a phase contrast image of confluent DMECs. (B). The proliferation of DMECs in full growth medium (dotted line) is exponential (solid line is a fitted exponential curve). (C). Cell cycle analysis of relative DNA content determined by labeling live cells with Hoescht and imaged with a microscope. Full growth medium (left) produces a familiar distribution of cells while 24 h in Basal Proliferation Medium (BPM) (right) produces far fewer cells in S- and G2/M-phases. (D). Demonstration of YOYO-1 dye to detect dead cells. (E).) Example of nuclei (closed circles) and dead cells (open circles) over time in BPM. (F). Cell death in BPM is countered by inhibition of caspase activity via 10 µM Q-VD-OPh.

FIGURE 2

Concentration-Dependent Population Dynamics of Dermal Microvascular Endothelial Cells (DMECs) in the Presence of Pro-Angiogenic Growth Factors. (A). The number of cells (N) over time for four concentrations of FGF2. (B). The number of dead cells (M) over time for the experiment shown in panel (A). (C). The proliferation rate, p, for the same experiment as in panels A,B. (D). The birth rate, B, as a function of time. (E). The death rate, D, as a function of time.

FIGURE 3

The dose-response proliferative behavior of Dermal Microvascular Endothelial Cells (DMECs) to Three Pro-Angiogenic Growth Factors. The proliferation rate (A), the birth rate (B), and the death rate (C) for FGF2, VEGFA, and HGF over a range of concentrations. Note that the y-axes share the same scale to facilitate comparison of the relative magnitude of each.

FIGURE 4

Identification of kinases important for the proliferation rate of Dermal Microvascular Endothelial Cells (DMECs) in each pro-angiogenic growth factor studied. (A). Magnitudes of the Kinome Regression (KIR) coefficients (i.e., influential kinases) for each identified growth factor. (B). Venn diagram revealing the exclusivity and commonalities in the set of kinases identified for each growth factor. (C). The list of the intersection of kinases implicated in all three growth factors. Kinases are ordered according to a rank obtained by summing each kinase’s ranks across all growth factors.

FIGURE 5

Orthogonal evidence supporting the role of kinases identified by Kinome Regression (KIR) Analysis. (A). Targeting growth factor receptors with 5 nM of siRNA reduces the proliferation rate of Dermal Microvascular Endothelial Cells (DMECs) in the presence of cognate growth factors while minimizing reduction in proliferation in other growth factors. (B). Targeting kinases with siRNA largely agrees with kinases implicated by KIR. Here, the effect is defined as the average over all three siRNAs used for each kinase and normalized to the no-siRNA control. Error bars are s. e.m. The asterisks indicate growth factors context in which the siRNA targeting a given kinase produced a statistically significant result ( $p<0.05$ from multiple comparison testing).