Targeted-dye Trace Cancer Imaging During Prostatectomy

In this multicenter multidisciplinary project, we are developing a new contrast agent that allows detection and imaging of as few as 10 prostate cells in vivo in real time. This class of agents is based on a coupling of a fluorescent marker capable of emitting light detectable through 5 to 10 cm of tissue, and a monoclonal antibody targeted against a cell surface domain of the luminal prostate-specific membrane antigen (PSMA) receptor. We will synthesize the agent and then test it in tissue culture, a xenograft mouse model, and finally in humans with the ultimate objectives of:

• reducing positive surgical margins during prostatectomy; 
• improving follow-up; 
• improving the staging of prostate cancer; and 
• eventually applying these techniques to other cancers.

Examples of successful clinical application of monoclonal antibodies in vivo include a first-generation antibody targeted to an intracellular domain of prostate PSMA, used as the basis of the ProstaScint scan. In breast, epidermal growth factor (EGF)-like sites (Her-2/neu) are used to target therapy, while targeted lymphoscintigraphy after intravenous or local injection helps to identify patients with axillary involvement. Two monoclonal agents are approved for colorectal cancer. When used in addition to CT scanning, one of these (CEA-Scan) increased preoperative identification of resectable patients by 40 percent and identified twice as many unresectable patients as computerized tomography alone.

	[Enlarge Image] Labeled Cancer Noninvasively Optically Imaged. Ten million labeled tumor cells (~7 mm HeLa tumor) injected i.v., followed on days 0, 7, and 14. Day 7: all tumor in body combined fits in a 2-mm sphere; clearly seen on optical scan in real time. Image courtesy of C. Contag, Stanford University.

Optical markers can be imaged in intact animals, as has been demonstrated using cell lines genetically engineered to constitutively express luciferase (an enzyme that spontaneously generates light) or green fluorescent protein. In the figure to the right, HeLa tumor cells were labeled with luciferase (transforming each cancer cell into a tiny light source that glows continuously) and injected intravenously. At day 0, tumor is seen in the lungs; by day 7, the total tumor load is only 1 x 105 cells; by day 14, tumor foci are growing in brain and liver. Open labeling with luciferase has allowed quantitation of tumor load and response to chemotherapy in the intact animal.

A method is needed to label previously unlabeled cells in vivo without the need of genetic engineering and with the ability to detect trace amounts of cancer (< 10,000 cells) in real time (< 10 sec). We propose to achieve this by injecting a contrast agent targeted toward the cell of interest, thus permitting use in a clinical situation; we may also apply this technique when investigating unmodified animal hosts.

Within 6 months, we hope to produce a targeted optical contrast agent specific for human prostate cells of luminal origin by conjugating a fluorophore to a well-characterized humanized mouse anti-PSMA monoclonal antibody and demonstrate that this dye can be detected in tissue models at concentrations expected to be found in vivo while retaining tissue culture specificity after conjugation. Within 2 years, we expect to:

• demonstrate the specificity and sensitivity, with appropriate controls, of binding this complex in vitro to human prostate cancer cells in culture and in vivo to human prostate cancer xenografts in immunodeficient mice;
• produce a pharmaceutical-grade conjugated antibody-dye appropriate for administration to humans;
• test a state-of-the-art optical imaging system that produces real-time images of the location of trace amounts of cancer in real-time with an expected update time of under 1 second per image; and
• conduct clinical pilot human trials.

[Enlarge Image] Cy Dye Imaging Image of a Cy7 dye in a mouse model, showing that optical images of tumors (lightest area) can be collected in mammals. Image courtesy of B. Ballou, Carnegie Mellon University.

Our long-term goal is to develop a class of optical agents that will detect, locate, and image prostate cancer in vivo and in real time with better sensitivity than ultrasound, magnetic resonance imaging, magnetic resonance spectroscopic imaging, positron emission tomography, or computerized tomography. If successful, we will administer this agent to patients for real-time imaging for guidance in radical prostatectomy. Our initial application will be detecting extraprostatic cancer during radical prostatectomy, leading to fewer positive margins during surgery and more effective screening, staging, and follow-up after surgery.

• David Benaron, MD, Principal Investigator (system construction, tissue studies), Stanford University School of Medicine, Stanford, CA, dbenaron@stanford.edu
• Peter Scardino, MD, Co-Investigator (clinical urology, animal studies), Memorial Sloan-Kettering Cancer Center, New York, NY, scardinp@mskcc.org
• Alan Waggoner, PhD, Co-Investigator (fluorescent dyes), Carnegie Mellon University, Pittsburgh, PA, waggoner@andrew.cmu.edu
• Yair Talmi, MD, Co-Investigator, Spectros Corporation, Portola Valley, CA, ytalmi@spectros.com
• Neil Bander, MD, Co-Investigator (former Urologic Oncologic fellow; anti-PSMA antibody/medical reagents), Weill Medical College, Cornell University, New York, NY, nhbander@mail.med.cornell.edu