There is a critical need for molecular imaging probes that specifically target receptors that are overexpressed on tumors, and allow noninvasive characterization of tumors for patient-specific cancer treatment and disease management. We engineered cystine knot (knottin) polypeptides as a new class of in vivo molecular imaging agents of integrin receptor-expressing tumors. The knottin family includes protease inhibitors, toxins, and antimicrobials. These polypeptides are small (30-40 residues), very stable in vitro and in vivo, and their disulfide-constrained loops tolerate much sequence diversity. These qualities are desirable for drug design; however, knottins do not naturally bind to receptors expressed on tumors, rendering them useless in cancer applications.

Integrins are a family of extracellular matrix adhesion receptors that non-covalently associate into alpha/beta heterodimers with distinct ligand binding specificities. Several integrins, including alphavbeta3, alphavbeta5, and alpha5beta1, have generated clinical interest due to their expression on the surface of cancer cells and tumor neovasculature, as well as their proposed role in mediating angiogenesis, tumor growth, and metastasis. Despite the prevalence of integrin-binding peptides and peptidomimetics in the literature, suboptimal tumor targeting efficacy and pharmacokinetics have limited their clinical translation as molecular imaging agents.

We used the Ecballium elaterium trypsin inhibitor II (EETI-II), a knottin from the squash family of protease inhibitors, as a molecular scaffold for engineering integrin binding polypeptides. We used directed evolution to redirect the function of EETI-II from a protease inhibitor to a high affinity integrin-binding polypeptide. Yeast surface display was used to engineer EETI-II to bind to alphavbeta3 and alphavbeta5 integrin receptors with high (nanomolar) affinity. In the process, we also discovered the first known polypeptide that binds with high affinity to alphavbeta3, alphavbeta5, and alpha5beta1 integrins. Since all three of these integrins are co-expressed on tumors and contribute to angiogenesis, this knottin polypeptide has potential as a broad spectrum cancer therapeutic or imaging agent. Next, we showed that the engineered knottin polypeptides strongly inhibited tumor cell adhesion to the extracellular matrix protein vitronectin, and in some cases fibronectin, depending on their integrin binding specificity.

Finally, we used optical imaging and positron emission tomography to demonstrate that integrin binding affinity plays a critical role in tumor uptake in living organisms. High affinity knottin polypeptides exhibited a significant increase in tumor uptake in mouse human tumor xenograft models compared to weaker binding polypeptides, including a cyclic pentapeptide that is currently under clinical development for imaging of integrin-expressing tumors. Thus, engineered knottin polypeptides show great potential as clinical diagnostics for a variety of cancers.

Funded by: the NIH NCI Howard Temin Award 5K01 CA104706 and the Mallinckrodt Faculty Scholar Award (both to JRC), NCI ICMIC P50 CA114747 and NIH 5R25 CA118681 (both to SSG), a Stanford Molecular Imaging Scholars postdoctoral fellowship (to RHK), and a Stanford Dean's fellowship (to AML).