Figure 3: Design and assessment of protease-activated imaging probes based on CGRP.

(a) Protease sensor designs incorporate an N-terminal blocking moiety (green), a protease cleavage sequence (red, cleavage sites indicated by triangles), and a C-terminal CGRP moiety (blue). Structures of five protease sensors are shown, with cognate proteases labelled in red text: (1) The synthetic peptide AP-CGRP-amide detects dipeptidase activity of fibroblast activation protein (FAP). (2–4) Recombinant fusion proteins comprises cysteine-free GFP, a short linker, a protease site and non-amidated CGRP detect TEV protease (2), enterokinase (EK) (3) and caspase-3 (CASP3) (4). (5) The synthetic peptide (long-chain biotin)-SG-DEVD-CGRP-amide also detects CASP3 activity. (b) Dose–response curve for sensor (5) measured using a cell-based bioassay³² (Supplementary Fig. 2) following incubation with or without CASP3. Luminescence values were normalized, and error bars reflect s.d. from n=3 replicates. (c) Protease sensing by CGRP-based probes measured in vitro, using probe concentrations indicated. Error bars represent s.d. of n=3 replicates. Sensors were incubated with or without corresponding proteases: (1) with 5 ng μl⁻¹ of human FAP; (2) with 0.1 U μl⁻¹ of TEV protease; (3) with 2 pg μl⁻¹ of EK light chain; (4) with 23 ng ul⁻¹ and (5) with 11.5 ng μl⁻¹ of human CASP3. (d) Left, protease-dependent switching of 100 nM CGRP-based molecular imaging probe (5) induces contrast differences in MRI. Significant haemodynamic activation can be seen in the presence but not the absence of coadministered 1.15 ng μl⁻¹ CASP3. Right, bar graph showing peak signal change induced by uncleaved versus cleaved sensor (5). Error bars throughout this figure indicate s.e.m. values over n=4 animals.