Current diagnosis of most cancers waits until the tumor is big enough to become anatomically detectable as a palpable lump or an unusual density on an x-ray or magnetic resonance image. By this stage, the malignancy is fairly advanced and may be too dispersed for surgical removal. When surgery is attempted, it is often difficult to balance the trauma of removing excess healthy tissue vs. the risk of leaving behind malignant cells. Older nonsurgical treatments mostly administer radiation or systemically toxic chemicals in the hope that faster-dividing cancer cells are slightly more vulnerable than non-dividing healthy cells, but such therapies usually have severe side effects and offer little prospect for major improvements. Instead, we need to exploit the most significant biochemical differences between cancer and normal cells to develop highly targeted diagnostic and therapeutic agents. The two leading approaches are (1) monoclonal antibodies that recognize surface markers that happen to be more abundant on some cancer cells than on normal tissue and (2) drugs to inhibit biochemical signals mediating proliferation or new blood vessel formation. However, antibodies are large, expensive protein molecules that generally do not penetrate solid tumors easily. Because the surface markers are of low abundance and each copy can only bind one antibody at a time, only tiny amounts of imaging agent or drug can be delivered. This delivery is mainly to the outside of the cell, because antibodies do not enter the true interior of the cell, i.e., the cytoplasm and nucleus, where the cell is biochemically most vulnerable. Signal-inhibiting small molecules take enormous pharmaceutical industry resources to develop. Usually they slow down but do not kill their target cells, so that the drugs must be taken for the life of the patient, during which the tumor cells take advantage of the ample opportunity to evolve resistance mechanisms. Therefore, alternative general strategies would be valuable.

We recently developed a novel rational mechanism to accumulate contrast agents (optical, radioactive, and magnetic) and therapeutic drugs within tumor cells. Certain peptide sequences rich in positive charges were known to drag a wide variety of molecular cargoes into cells and tissues in vivo. The new discovery is that such uptake can be prevented simply by appending negatively charged sequences, which neutralize the positive charges and inhibit membrane adhesion and cellular uptake. However, the linker sequence between the sets of opposite charges is designed to be cleaved by proteases known to be abundant in invasive tumors but that are rare in healthy adult tissue. Such enzymes include matrix metalloproteinases (MMPs) or urokinase plasminogen activator (uPA). Tumors generally need MMPs to be invasive, i.e., to chew their way through healthy tissues, and uPA is important for activating MMPs. Cleavage of the linker jettisons the negative charges, freeing the positively charged peptide to carry the contrast agent or therapeutic agent onto and into the immediately adjacent cells. Thus, the cargo accumulates in invasive tumors in preference to normal tissues not expressing the proteases. This mechanism enables amplification because each enzyme molecule can activate many molecules of peptide. Cleavage-dependent contrast has now been demonstrated in cell culture, xenografts of human cancer cells in nude mice, spontaneous mammary tumors in transgenic mice, and slices from squamous cell carcinomas freshly resected from surgical patients.

The main goals of this award are to extend and optimize the above preliminary results to pave the way for clinical trials. The existing results were obtained with far-red fluorescent probes for experimental convenience, but we need to extend them to other forms of imaging with greater depth penetration such as magnetic resonance imaging (MRI) or positron emission tomography (PET). Also, we plan to attach therapeutic cargoes, i.e., agents that can preferentially accumulate in the tumor and directly kill it or sensitize it to radiation. Many molecular variants will have to be synthesized and tested to optimize their selectivity and sensitivity for malignant tumors relative to normal tissues and to minimize potential side effects.

Two classes of clinical application currently appear most attractive. The more immediately applicable (in about 3 years) would be as a real-time adjunct to surgery to remove a primary tumor. An infrared-fluorescent probe would be applied topically or injected systemically. An infrared-sensitive viewing system would highlight any remaining malignant clusters lurking just beyond the excision as bright spots. Such tumors would be excised and the process repeated. A more ambitious type of application (5 or more years from now) would be to detect and destroy inoperable tumors. Ideally, a combined imaging/radiation-sensitizing probe would be administered, then noninvasively imaged by MRI or PET. If nothing lit up, the patient would be clear. If everything lit up, the probe would have failed without causing any harm. If discrete foci were visible with adequate contrast, radiation would be aimed precisely at each locus, triggering the trapped sensitizer to kill the cells that had taken up the probe. Both the protease-activated uptake and the image-guided aiming of the radiation would help confine damage to the tumor.