Elsevier

Nuclear Medicine and Biology

Volumes 62–63, July–August 2018, Pages 9-17
Nuclear Medicine and Biology

Impact of structural alterations on the radiopharmacological profile of 18F-labeled pyrimidines as cyclooxygenase-2 (COX-2) imaging agents

https://doi.org/10.1016/j.nucmedbio.2018.05.001Get rights and content

Abstract

Introduction

Non-invasive imaging of COX-2 in cancer represents a powerful tool for assessing COX-2-mediated effects on chemoprevention and radiosensitization using potent and selective COX-2 inhibitors as an emerging class of anticancer drugs. Careful assessment of the pharmacokinetic profile of radiolabeled COX-2 inhibitors is of crucial importance for the development of suitable radiotracers for COX-2 imaging in vivo. The delicate balance between the selection of typical COX-2 pharmacophores and the resulting physicochemical characteristics of the COX-2 inhibitor represents a formidable challenge for the search of radiolabeled COX-2 imaging agents. Several pyrimidine-based COX-2 inhibitors demonstrated favorable in vitro and in vivo COX-2 imaging properties in various COX-2 expressing cancer cell lines. Here, we describe a comparative radiopharmacological study of three 18F-labeled COX-2 inhibitors based on a pyrimidine scaffold. The objective of this study was to investigate how subtle structural alterations influence the pharmacokinetic profile of lead compound [18F]1a ([18F]Pyricoxib) to afford 18F-labeled pyrimidine-based COX-2 inhibitors with improved COX-2 imaging properties in vivo.

Methods

Radiosynthesis of radiotracers was accomplished through reaction with 4-[18F]fluorobenzyl amine on a methyl-sulfone labeling precursor ([18F]1a and [18F]2a) or late-stage radiofluorination using a iodyl-containing labeling precursor ([18F]3a). Radiopharmacological profile of 18F-labeled pyrimidine-based COX-2 inhibitors [18F]1a, [18F]2a and [18F]3a was studied in COX-2-expressing human HCA-7 colorectal cancer cell line, including cellular uptake studies in HCA-7 cells and dynamic PET imaging studies in HCA-7 xenografts.

Results

Cellular uptake of radiotracers [18F]2a and [18F]3a in HCA-7 cells was 450% and 300% radioactivity/mg protein, respectively, after 90 min incubation, compared to 600% radioactivity/mg protein for radiotracer [18F]1a. Dynamic PET imaging studies revealed a tumor SUV of 0.53 ([18F]2a) and 0.54 ([18F]3a) after 60 min p.i. with a tumor-to-muscle ratio of ~1. Tumor SUV for [18F]1a (60 min p.i.) was 0.76 and a tumor-to-muscle ratio of ~1.5. Pyricoxib analogues [18F]2a and [18F]3a showed distinct pharmacokinetic profiles in comparison to lead compound [18F]1a with a significantly improved lung clearance pattern.

Replacing the 4-[18F]fluorobenzyl amine motif in radiotracer [18F]1a with a 4-[18F]fluorobenzyl alcohol motif in radiotracer [18F]3a resulted in re-routing of the metabolic pathway as demonstrated by a more rapid liver clearance and higher initial kidney uptake and more rapid kidney clearance compared to radiotracers [18F]1a and [18F]2a. Moreover, radiotracer [18F]3a displayed favorable rapid brain uptake and retention.

Conclusion

The radiopharmacological profile of three 18F-labeled COX-2 inhibitors based on a pyrimidine scaffold were evaluated in COX-2 expressing human colorectal cancer cell line HCA-7 and HCA-7 xenografts in mice. Despite the overall structural similarity and comparable COX-2 inhibitory potency of all three radiotracers, subtle structural alterations led to significantly different in vitro and in vivo metabolic profiles.

Advances in knowledge

Among all tested pyrimidine-based 18F-labeled COX-2 inhibitors, lead compound [18F]1a remains the most suitable radiotracer for assessing COX-2 expression in vivo. Radiotracer [18F]3a showed significantly improved first pass pulmonary passage in comparison to radiotracer [18F]1a and might represents a promising lead compound for the development of radiotracers for PET imaging of COX-2 in neuroinflammation.

Introduction

Cyclooxygenases (COXs) are responsible for the conversion of arachidonic acid to prostaglandins to mediate the body's response to inflammatory conditions. The COX enzyme family consists of isoforms COX-1 and COX-2, which are constitutively and inducibly expressed, respectively. Besides COX-1 and COX-2, also a COX-3 isoform has been described in the literature. However, there is still a debate on its individual role as a drug target for inflammation or if it only represents a splice variant of COX-1 [1]. COX-1 and COX-2 have a long history as important drug target for treatment of inflammatory conditions. The development of non-steroidal anti-inflammatory drugs (NSAIDs) such as aspirin, ibuprofen, paracetamol and diclofenac predates the discovery of the COX enzyme family [2,3]. Consequently, most NSAIDs are not selective or only minimally selective in inhibiting COX-1 over COX-2 or vice versa. Selective COX-2 inhibitors were developed only after the discovery of the inducible COX-2 isoform in 1991 [4]. The discovery resulted in an important scientific paradigm which postulated that COX-1 is responsible for the maintenance of normal physiological processes, while acute and chronic inflammatory conditions are mediated by COX-2 [5,6]. Recent research over the past years, however, has also analyzed the involvement of COX-1 during acute and chronic inflammation. Inhibition of housekeeping enzyme COX-1 is one of the main reason for observed severe gastrointestinal side-effects associated with long term NSAID use [7].

The discovery of the COX-2 isoform led to the development of numerous COX-2 selective inhibitors, so called “coxibs”, such as celecoxib, rofecoxib, valdecoxib and others. These drugs promised to deliver the same relief from pain and inflammation as NSAIDs, but without gastrointestinal side-effects.

Development of highly selective compounds has made COX-2 also an attractive target and biomarker for molecular imaging applications [8]. Moreover, COX-2 was also found to be involved in the development and progression of a variety of cancers and neurodegenerative diseases [9,10]. As a result, a large number of COX-2 selective radiotracers have been developed to visualize functional expression of COX-2 in cancer, inflammation and other diseases [[11], [12], [13], [14], [15], [16]].

Despite initial promise, the therapeutic and diagnostic value of coxibs as drugs and radiolabeled molecular probes remained highly challenging. The majority of coxibs in clinical use during the early 2000's had to be withdrawn from market following concerns over their cardiac toxicity profile. Although more than two dozen of PET and SPECT radiotracers have been developed and tested over the past decade, none of them have progressed past initial preclinical testing stages [7].

The challenges of developing a COX-2 specific radiotracer are manifold. Many of the PET and SPECT imaging agents that have been reported to date have failed on the basis of insufficient metabolic stability or lack of COX-2 specific binding in vivo. Other studies found the lack of a well characterized animal model that stably expresses COX-2 to be a major obstacle [7]. The slow progress in the field of COX-2 molecular imaging is partially due to the complex nature of the target itself. Both COX enzymes are intracellular targets, membrane bound on the inside of the endoplasmic reticulum, while the active site is located at the end of a long hydrophobic channel the entrance of which is located inside the membrane [17]. Most widely biomarkers for molecular imaging are extracellular or cross membrane-bound targets, while the targeting of intracellular targets remains a formidable challenge [18]. Efforts to image intracellular targets such as COX-2 using small molecules have struggled to find an ideal structural balance for a favorable pharmacokinetic profile of the radiotracers. Molecular candidates for intracellular imaging must be lipophilic enough to readily diffuse through membranes. However, in that case the inevitable non-specific background uptake of these radiotracers in adipose and muscle tissue is expected to lead to unfavorable target-to-background ratios.

The efforts of our research group in this area led to the development and evaluation of radiotracer [18F]1a [19], a radiolabeled COX-2 inhibitor based on a central pyrimidine scaffold. Despite the formidable challenges associated with the development of COX-2 selective radiotracers for in vivo PET imaging, [18F]1a showed favorable and COX-2 selective tumor uptake profile in a preclinical COX-2-expressing HCA-7 colorectal tumor model [19]. Previous and unsuccessful attempts of developing celecoxib-based 18F-labeled radiotracers for imaging COX-2 [16], designed around a central pyrazole scaffold. His further suggests the importance of proper structure selection and underlines the suitability of a pyrimidine scaffold as central motifs for COX-2 selective imaging agents. Although radiotracer [18F]1a displayed promising COX-2 specific tumor uptake in vivo, the radiotracer also exhibited a number of non-specific off-target interactions. [18F]1a showed high uptake in the muscle tissue, which leads to only modest tumor-to-background ratios after >1 h post injection. In addition, radiotracer [18F]1a also showed significant lung retention upon first pass, which leads to delayed distribution pharmacokinetics [19].

In light of these challenges, the delicate balance of pharmacokinetic properties of COX-2 inhibitors is of crucial importance in the search for suitable COX-2 imaging radiotracer.

Here, we report on the in vitro and in vivo evaluation of two structural analogues of previously reported COX-2 radiotracer [18F]1a [19]. The objective of this study was to investigate whether small structural alterations can improve the pharmacokinetic profile of radiotracer [18F]1a to provide an improved COX-2 selective radiotracer. The synthesis of analogues 2a and 3a, as well as their respective 18F-labeled versions ([18F]2a and [18F]3a) has been reported previously [20]. Compound 1a and its structural analogues 2a and 3a display high COX-2 inhibitory potencies in the low nanomolar range while possessing comparable lipophilicity (logP).

Section snippets

General

All regents and solvents were obtained from Sigma-Aldrich, unless otherwise stated and used without further purification. Column chromatography was conducted using Merck silica gel (mesh size 230–400 ASTM). Thin-layer chromatography (TLC) was performed on Merck silica gel F-254 aluminum plates, with visualization under UV light (254 nm). High performance liquid chromatography (HPLC) purifications and analysis were performed using a Phenomenex LUNA® C18 column (100 Å, 250 × 10 mm, 10 mm) on a

Chemistry and radiochemistry

We have previously reported the design and radiosynthesis of three 18F-labeled COX-2 inhibitors based on a pyrimidine scaffold [20] ([18F]1a, [18F]2a, [18F]3a; Fig. 1a). Radiotracer [18F]1a carries a methyl sulfone COX-2 pharmacophore, while radiotracer [18F]2a carries a sulfonamide COX-2 pharmacophore (Fig. 1a - highlighted in blue). Radiotracer [18F]3a contains an ether bridge, while [18F]1a contains a secondary amine linker (Fig. 1a - highlighted in red).

Lipophilicity, experimentally

Discussion

Efforts to develop a radiotracer for in vivo imaging of COX-2 are hampered by a lack of standardized and well-characterized models [6]. Emerging COX-2 radiotracers are studied in a variety of in vitro and in vivo models, generally without comparison to a previously evaluated tracer. The results of these studies are difficult to interpret because it is seldom clear whether the novel radiotracer or the biological model have an innate inadequacy that leads to the relative inability of the

Conclusion

Two highly potent and selective 18F-labeled COX-2 inhibitors were evaluated in a well characterized COX-2 expressing colorectal cancer model and compared to previously reported evaluation of radiotracer [18F]1a. Despite their structural similarities, the hepatobiliary and renal clearance profiles of the radiotracers varied significantly.

We have shown that small modifications such as changing the methylsulfone COX-2 pharmacophore to a sulfonamide group and the change from the amine to the ether

Acknowledgement

The authors would like to thank John Wilson, Dave Clendening and Blake Lazurko from the Edmonton PET Center for radionuclide production and excellent technical support. We thank Gail Hipperson and Dan McGinn from the Vivarium of the Cross Cancer Institute for supporting the animal work. We also gratefully acknowledge the Dianne and Irving Kipnes Foundation, the Canadian Institute of Health Research (CIHR) and the National Science and Engineering Research Council of Canada (NSERC) -

References (37)

  • M. Tanaka et al.

    Radiosynthesis and evaluation of 11C-labelled diaryl-substituted imidazole and indole derivatives for mapping cyclooxygenase-2

    Biol Pharm Bull

    (2006)
  • M.J. Uddin et al.

    Fluorinated COX-2 inhibitors as agents in PET imaging of inflammation and cancer

    Cancer Prev Res

    (2011)
  • E.V. Hersh et al.

    Update on cyclooxygenase inhibitors: has a third COX isoform entered the fray?

    Curr Med Res Opin

    (2005)
  • S. Moncada et al.

    An enzyme isolated from arteries transforms prostaglandin endoperoxides to an unstable substance that inhibits platelet aggregation

    Nature

    (1976)
  • J.R. Vane et al.

    Cyclooxygenase-1 and 2

    Annu Rev Pharmacol

    (1998)
  • R.N. Dubois et al.

    Cyclooxygenase in biology and disease

    FASEB J

    (1998)
  • J.M. Scheiman

    Unmet needs in non-steroidal anti-inflammatory drug-induced upper gastrointestinal diseases

    Drugs

    (2006)
  • O. Tietz et al.

    Radiotracers for molecular imaging of cyclooxygenase-2 (COX-2) enzyme

    Curr Med Chem

    (2013)
  • Cited by (6)

    • Development of brain PET imaging agents: Strategies for imaging neuroinflammation in Alzheimer's disease

      2019, Progress in Molecular Biology and Translational Science
      Citation Excerpt :

      Another tracer that has been evaluated in COX-2 expressing tumor-bearing mice is [18F]pyricoxib (IC50 COX-2 = 7 nM; COX-1 > 30 μM).88,89 [ 18F]pyricoxib is metabolically relatively stable with 60% of intact tracer left at 2 h p.i.,88 and shows good uptake in COX-2 expressing tumors.88,89 Interestingly, a minor modification of [18F]pyricoxib (replacement of benzylamine by benzyloxy moiety), that only had a small effect on affinity toward COX-2 (20 nM versus 7 nM for pyricoxib), had a big impact in brain kinetics, making [18F]3a the preferred ligand to study in neuroinflammatory conditions.89

    View full text