Targeted Imaging Agent Puts the Focus on Cancer
By: Carol A. Rouzer, VICB Communications
Published: October 12, 2011
Discovery of a fluorinated imaging agent targeted to cyclooxygenase-2 promises to increase the ability of PET imaging to detect early stage tumors.
Despite major advances in chemotherapy and radiation therapy, early detection remains our single most effective weapon in the fight against cancer. New imaging modalities, such as magnetic resonance imaging (MRI) and positron emission tomography (PET) are promising noninvasive techniques that can frequently identify cancers while they are easily curable. However, to fully realize the potential of these new technologies requires that malignant tissue can be easily distinguished from normal tissue. To this end, considerable effort is being devoted to identifying the unique biomolecular signatures of malignancy, so that small molecule imaging agents can be designed to highlight cancerous tumors for detection by imaging technology.
Figure 1. PET-scan of a cancerous lymph node from a patient treated previously for cancer of the tongue. Figure reproduced from Wikimedia Commons under the GNU Free Documentation License.
One potential target for cancer imaging is the enzyme cyclooxygenase-2 (COX-2), which along with its isoform COX-1, synthesizes a class of biologically active lipid signaling molecules known as prostaglandins. The prostaglandins play a role in modulating a range of physiological and pathophysiological processes as widely varied as blood pressure, uterine contractions, and the sensation of pain. The COX-1 isoform is widely expressed in most tissues under normal conditions, while COX-2 expression is induced by inflammatory and proliferative stimuli. COX-2 expression is elevated in most cancers, as well as in precancerous lesions, suggesting that imaging agents that selectively bind to COX-2 should be concentrated in early stage malignancies. Proof-of-concept for this approach was provided by Vanderbilt Institute of Chemical Biology (VICB) investigator Larry Marnett and his lab, who showed that fluorescent COX-2-selective inhibitors label COX-2-expressing tumors for detection in vivo (see http://www.vanderbilt.edu/vicb/mole/vicb_news_2010_autumn.pdf). Although highly promising, the use of fluorescent imaging agents is restricted to tumors that can be visualized by visual light irradiation. Now, building on this foundation, the Marnett lab reports a radiofluorinated COX-2-selective imaging agent that can be used for PET-based imaging of cancers throughout the body [M. J. Uddin et al. (2011) Can. Prev. Res., 4, 1536, and commentary by D. A. Ostrov and C. H. Contag (2011) Can. Prev. Res., 4, 1523].
PET imaging agents require the presence of a positron-emitting radiolabeled atom. 18F is used most frequently because its short half-life limits the patient’s exposure to radioactivity. The rapid half-life also means that the radiolabel must be incorporated into the molecule quickly right before use. Thus, a COX-2-targeted PET imaging agent must contain a fluorine atom that can be rapidly incorporated as the 18F isotope. The COX enzymes are the primary targets of the widely used nonsteroidal anti-inflammatory drugs (NSAIDs) such as ibuprofen, naproxen, and indomethacin, which inhibit both isoforms nonselectively. More recently, COX-2-selective inhibitors, such as celecoxib, have come to market. Consequently, a large wealth of prior chemical and pharmacologic data provided a foundation for the design of COX-2-targeted imaging agents. The Marnett lab had previously discovered that converting the carboxylic acid-containing NSAID indomethacin into ester and amide derivatives produced a wide variety of COX-2-selective inhibitors. Because of the flexibility of this approach, they initially used it to design a series of potent fluorine-containing COX-2-selective inhibitors (e.g., compounds 2 and 4, Figure 2). However, they soon discovered that these compounds were not stable under the conditions necessary to incorporate the 18F radiolabel, eliminating them as candidates for PET imaging.
Figure 2. (Top) Structures of the nonselective NSAID indomethacin and two fluorine-containing COX-2-selective inhibitors derived from the indomethacin scaffold (compounds 2 and 4). A wide variety of side chains can be connected to indomethacin by converting indomethacin’s carboxyl group (blue) to an amide. The fluorine atom necessary for PET imaging is shown in red. Unfortunately, compounds 2 and 4 were not stable under the conditions needed to rapidly incorporate the radiolabeled 18F for PET imaging. (Bottom) Structures of the clinically used COX-2-selective inhibitor celecoxib and the PET imaging agent derived from the celecoxib scaffold (compound 7). Note that celecoxib contains three fluorine atoms (blue), but these are not readily exchanged for 18F. In contrast, the single fluorine in compound 7 is amenable to exchange for 18F under conditions used to generate PET imagine agents.
The investigators achieved greater success by making small modifications to the structure of the COX-2-selective inhibitor celecoxib. These efforts led to compound 7, a potent and selective COX-2 inhibitor containing a fluorine that could rapidly be incorporated as 18F before use in PET imaging. Although slightly less potent than celecoxib in vitro, kinetic studies indicated that compound 7 was a tight-binding COX-2 inhibitor with sufficient affinity for in vivo use. Furthermore, its structural similarity to celecoxib suggested that it should be safe for clinical use.
To test the effectiveness of compound 7 as a PET imaging agent, the Marnett lab used a model of inflammation in the rat hind paw. Injection of carageenan into the paw induces an inflammatory response characterized by high levels of COX-2 expression. The untreated paw serves as a control. Indeed, when radiolabeled compound 7 was injected into rats following the carageenan treatment, PET imaging indicated that the radiolabel concentrated in the inflamed, but not the untreated paw. Concentration of compound 7 could be blocked by pretreating the rats with celecoxib to block the inhibitor binding sites of COX-2 (Figure 3).
Figure 3. PET imaging of carageenan-induced inflammation in the rat paw using compound 7. (A) The paw on the left was injected with carageenan, while the paw on the right serves as the control. The PET image was acquired following administration of radiolabeled compound 7. (B) Same as A, but the rat was treated with celecoxib prior to compound 7 administration. Figure provided by the Marnett lab. Copyright 2011.
Next, the investigators tested compound 7 in a cancer model. Mice were engrafted with a human head and neck squamous cell carcinoma that expressed COX-2 or a colon carcinoma that was COX-2 negative. Following injection of compound 7, PET imaging revealed a concentration of radioactivity in the COX-2-expressing tumor, but not the COX-2-negative tumor. Again, pretreatment of the animals with celecoxib prevented the accumulation of compound 7 in the COX-2-expressing tumor.
Figure 4. PET imaging of a COX-2-expressing human head and neck cancer (1483) xenograft in a mouse (A) following injection of radiolabeled compound 7. The signal intensity in the 1483 tumor is much higher than in the COX-2-negative HCT116 colon carcinoma xenograft (B). Radiolabel did not accumulate in the 1483 tumor when mice were pretreated with celcoxib prior to injection of compound 7 (data not shown). Figure provided by the Marnett lab. Copyright 2011.
These highly promising results provide further evidence that COX-2 can be used as a target for the imaging of cancer. The ability to use this approach with PET imaging expands the potential applicability to tumors in any part of the body. Further work will be required to determine how well COX-2-targeted imaging will detect early stage cancer, but this is clearly work that is well worth the pursuit.