Notch1 and Notch4 core binding domain peptibodies exhibit distinct ligand-binding and anti-angiogenic properties

Abstract

The Notch signaling pathway is an important therapeutic target for the treatment of inflammatory diseases and cancer. We previously created ligand-specific inhibitors of Notch signaling comprised of Fc fusions to specific EGF-like repeats of the Notch1 extracellular domain, called Notch decoys, which bound ligands, blocked Notch signaling, and showed anti-tumor activity with low toxicity. However, the study of their function depended on virally mediated expression, which precluded dosage control and limited clinical applicability. We have refined the decoy design to create peptibody-based Notch inhibitors comprising the core binding domains, EGF-like repeats 10-14, of either Notch1 or Notch4. These Notch peptibodies showed high secretion properties and production yields that were improved by nearly 100-fold compared to previous Notch decoys. Using surface plasmon resonance spectroscopy coupled with co-immunoprecipitation assays, we observed that Notch1 and Notch4 peptibodies demonstrate strong but distinct binding properties to Notch ligands DLL4 and JAG1. Both Notch1 and Notch4 peptibodies interfere with Notch signaling in endothelial cells and reduce expression of canonical Notch targets after treatment. While prior DLL4 inhibitors cause hyper-sprouting, the Notch1 peptibody reduced angiogenesis in a 3-dimensional in vitro sprouting assay. Administration of Notch1 peptibodies to neonate mice resulted in reduced radial outgrowth of retinal vasculature, confirming anti-angiogenic properties. We conclude that purified Notch peptibodies comprising EGF-like repeats 10-14 bind to both DLL4 and JAG1 ligands and exhibit anti-angiogenic properties. Based on their secretion profile, unique Notch inhibitory activities, and anti-angiogenic properties, Notch peptibodies present new opportunities for therapeutic Notch inhibition.

Keywords: Angiogenesis; Inhibitor; Notch; Notch core binding domain; Peptibody.

Fig. 1

Construction and expression of Notch decoys. A Notch decoys are composed of specific EGF-like repeats (10–14) of the Notch1 or Notch4 extracellular domain fused to the human Fc domain. B Amino acid alignment of EGFs 10–14 of the Notch1 and Notch4 receptor, respectively. C Coomassie stained TGX-Gel analysis. Purified proteins were loaded under reducing conditions. D Western blot analysis. 100 ng of purified protein were resolved on TGX gels under non-reducing or reducing conditions and immunoblotted with Fc-specific antibody

Fig. 2

Binding of Notch ligands and Notch decoys examined by co-immunoprecipitation. Co-immunoprecipitation assays show interaction of Notch ligands DLL4 and JAG1 with Notch peptibodies. A N1_10-14Fc and full-length DLL4-MYC or JAG1-FLAG were transiently co-transfected into HEK-293Ts. Protein A/G beads were used to immunoprecipitate N1_10-14Fc acting as the “bait” protein from whole cell lysates. Binding of N1_10-14Fc and Notch ligands were determined by immunoblot using anti-Fc, anti-FLAG, and anti-MYC antibodies. B N4_10-14Fc was evaluated similarly to panel A

Fig. 3

Binding of Notch ligands and Notch decoys examined by SPR. SPR measurements show binding affinity of N1_10-14Fc and N4_10-14Fc to DLL4. A Recombinant Fc-tagged hDLL4 was immobilized on the sensor chip using amine coupling and multi-cycle kinetic experiments were performed using increasing concentrations of either N1_10-14Fc or N4_10-14Fc. B Normalization of the N4_10-14Fc sensorgram to N1_10-14Fc. The resulting sensorgrams were normalized using Biacore sensorgram fitting algorithms and similarity scores

Fig. 4

N1_10–14Fc suppresses endothelial Notch signaling. A N1_10-14Fc inhibits DLL4-induced cleavage of the Notch1 receptor. HUVECs were plated onto 1 μg/mL of recombinant hDLL4 in the presence of N1_10-14Fc or IgG Fc isotype control for 24 h and quantitated for cleaved Notch1 by Western blot. (B, C) RT-qPCR analysis of Notch decoy-induced gene changes in HUVECs. Cells were treated with either Notch peptibody or IgG Fc at indicated concentrations for 24 h. B Expression of targets of canonical Notch signaling in HUVECs treated with N1_10-14Fc. C Expression of targets of canonical Notch signaling in HUVECs treated with N4_10-14Fc. For all figures, error bars represent standard error of mean (SEM) and ***P value < 0.001, **P value < 0.01, *P value < 0.05

Fig. 5

N1_10–14Fc affects endothelial viability and modulates angiogenesis. A Representative images of Fibrin bead assays (FiBA). HUVEC-coated beads were embedded in fibrin gel with increasing doses of human IgG Fc, N1_10-14Fc, or N4_10-14Fc for 12 days. After 12 days, N1_10-14Fc significantly reduced both sprout number and sprout length at dosages of 5 and 10ug/ml. No significant effect was seen with N4_10-14Fc at any dosage. B Quantification of mean sprout number per bead for treated HUVECs. Box-and-whisker plots show median, minimum, and maximum values. C Quantification of mean sprout length for treated HUVECs. D Cell viability assay of HUVECs treated with increasing dose of IgG Fc, N1_10-14Fc, or N4_10-14Fc for 72 h. IgG Fc-treated control group was set at 100% and was compared with that of peptibody-treated groups. E Identical experiments conducted with mLMVEC. **P value < 0.01, *P value < 0.05

Fig. 6

N1_10–14Fc inhibits retinal angiogenesis in murine neonates. C57BL/6 mice were injected intragastrically with 12.5 mg/kg of recombinant N1_10-14Fc peptibody or IgG Fc for three days postnatally (P1-P3). A Representative images and quantification of postnatal day 5 (P5) retinal vasculature stained with Isolectin B4 (red). Radial outgrowth and percent vascular coverage near the angiogenic front were reduced in N1_10–14Fc-treated mice (N = 7–8), while tip cell density and percent vascular coverage of the mature plexus were not statistically altered. Scale bars: 1000 μm. B Representative images and quantification of postnatal day 5 (P5) retinal vasculature stained with Isolectin B4 (endothelium, red) and α-SMA (vascular smooth muscle cells, green). No difference was observed in the percentage of smooth muscle coverage in control and N1_10-14Fc-treated mice (N = 4). Scale bars: 106 μm. Box-and-whisker plots show median, minimum, and maximum values. **P value < 0.01, *P value < 0.05. (Color figure online)