An engineered Axl 'decoy receptor' effectively silences the Gas6-Axl signaling axis

Abstract

Aberrant signaling through the Axl receptor tyrosine kinase has been associated with a myriad of human diseases, most notably metastatic cancer, identifying Axl and its ligand Gas6 as important therapeutic targets. Using rational and combinatorial approaches, we engineered an Axl 'decoy receptor' that binds Gas6 with high affinity and inhibits its function, offering an alternative approach from drug discovery efforts that directly target Axl. Four mutations within this high-affinity Axl variant caused structural alterations in side chains across the Gas6-Axl binding interface, stabilizing a conformational change on Gas6. When reformatted as an Fc fusion, the engineered decoy receptor bound Gas6 with femtomolar affinity, an 80-fold improvement compared to binding of the wild-type Axl receptor, allowing effective sequestration of Gas6 and specific abrogation of Axl signaling. Moreover, increased Gas6 binding affinity was critical and correlative with the ability of decoy receptors to potently inhibit metastasis and disease progression in vivo.

Figures

Figure 1

Engineering and characterization of receptor-based Axl antagonists. (a) Axl’s extracellular domain consists of two immunoglobulin-like (Ig) domains containing high and low affinity Gas6 binding sites, followed by two fibronectin type III domains. Binding of Gas6 to Axl leads to receptor dimerization and activation of downstream signaling. Axl decoy receptors sequester Gas6, preventing activation of the Axl signaling cascade. (b) Overlaid flow cytometry dot plots representing binding of yeast-displayed wild-type Axl Ig1 (red) and unsorted Axl Ig1 library (blue) to 10 nM Gas6 (y-axis), and expression levels on the yeast cell surface (x-axis). (c) Flow cytometry histograms of the initial Axl library and intermediate sort products compared to wild-type Axl Ig1 (gray), measuring binding to 0.5 nM Gas6 (top row) and persistent Gas6 binding after 30 h incubation with excess competitor (bottom row). MYD1 is also included for comparison. For clarity, only the gated population of yeast expressing Axl is shown. (d) Binding affinities of wild-type Axl Ig1, MYD1, and Axlnb to Gas6 as determined by KinExA. (e) Binding affinities to Gas6 of every permutation of the four mutations found in MYD1. Raw KinExA data and associated error values can be found in Supplementary Figs. 2 and 3.

Figure 2

Structural basis for high-affinity binding. (a) Gas6/MYD1 co-complex showing overall architecture and 2:2 stoichiometry. (b) MYD1 Ig1 (orange) and Gas6 LG1 (gray) domains showing the location of the four mutations in MYD1 with respect to the major binding site which lies at the interface of these two domains. (c) Analysis of the wild-type structure (PDB 2C5D) reveals steric crowding between the side chains of T457Gas6 and V92Axl. The V92A mutation alleviates crowding in the MYD1 co-complex and facilitates local reorganization of side chains around V92A, exemplified by R48 and Q94. This in turn creates an elongated groove on MYD1 at the binding interface that allows reorientation of T457 on Gas6. (e) Reorientation of T457 results in capping of the N-terminus of Helix A. The wild-type (green) and MYD1 (gray) structures are overlaid for comparison. (f) Capping stabilizes Helix A, as seen by B-factor analysis (see Methods).

Figure 3

Design and characterization of Axl Fc fusions. (a) Schematic representation of the panel of MYD1 Fc fusions generated. The major and minor Gas6 binding sites are located on Axl’s Ig1 and Ig2 domains, respectively. (b) Apparent binding affinities of Axl Fc fusions to human and mouse Gas6. Raw KinExA data and associated error values can be found in Supplementary Fig. 8. (c) Proposed model of multivalent binding, where both arms of the fulllength Axl Fc fusion contact Gas6. While a 1:1 complex is shown, the same multivalent binding may occur in a 2:1 Gas6:Axl Fc ratio reminiscent of the physiologic active complex.

Figure 4

MYD1 Fc inhibits Axl activation and downstream signaling in skov3.ip cells. (a) Wild-type Axl Fc and MYD1 Fc, but not Axlnb Fc, can inhibit Gas6-mediated Axl activation in vitro. (b) Inhibition of Axl activation leads to reduced levels of phosphorylated Akt and Erk1/2, and an increase in the epithelial marker e-cadherin. For full (uncut) blots, see Supplementary Fig. 11.

Figure 5

Sequestration of Gas6 by MYD1 Fc inhibits metastasis (a) Amount of free Gas6 in serum of mice 12 h after administration of a single dose of MYD1 Fc. (b) Kinetics of Gas6 sequestration (black) and MYD1 Fc clearance (red) following a 1 mg/kg dose of MYD1 Fc. (c) Two mice were analyzed for each data point in (b) and (c). Using the off-rates of the Gas6/Axl Fc interactions (Fig. 3b), dissociation of Gas6 bound to either wild-type Axl Fc (red) or MYD1 Fc (blue) is plotted over time. The in vivo clearance of the Axl decoy receptors as measured in (c) is overlaid in black. (d – f) Tumor burden in in vivo models of metastatic human ovarian cancer. The number of visible metastases in animals treated with Axlnb Fc, wild-type Axl Fc, or MYD1 Fc was counted in the skov3.ip (d) and OVCAR (f) tumor models. Representative images of mice from each treatment group in the skov3.ip model are shown, arrows indicate disease (e). In both models, animals were administered 10 mg/kg of the indicated protein twice weekly. (g) Lung metastases in the 4T1 luciferase breast cancer model as quantified by ex vivo bioluminescent imaging. Mice received IV injections of the indicated treatment twice weekly. (h) Representative bioluminescent images of lungs and spleens from each treatment group. Error bars represent ± s.d., n = 6 – 12 mice per group, *P < 0.05; **P < 0.01, ***P < 0.001; ****P < 0.0001.