Why would you say your work is highly cited?

In the Cell paper, we showed 1) expression of COX-2 in early mouse intestinal polyps, 2) genetic evidence that knocking out the COX-2 gene in mice dramatically reduced the number and size of polyps, and 3) pharmacological evidence that we can mimic the genetic experiment above with a COX-2 inhibitor. We further showed expression of COX-2 in the polyp stromal cells rather than in the epithelium.

“Our paper brought the COX-2 inhibitors into the arena of clinical cancer prevention.”

These data clearly indicated the significant role of COX-2 in the polyp, the early preneoplastic lesion. At the same time, our pharmacological evidence showed that COX-2 inhibitors could be used for polyposis treatment to prevent them from becoming malignant cancer.

Therefore, we presented convincing direct evidence for the role of COX-2 in tumorigenesis, although it had been suggested also from other circumstantial data. At the same time, our work opened the new possibility of treating many preneoplastic and neoplastic lesions with COX-2 inhibitors. In other words, it gave an unambiguous answer to the basic question on the role of COX-2 in cancer, and led to direct clinical applications of COX-2 inhibitors in cancer chemoprevention.

  What are the circumstances which led you to your work?

Shortly after the APC gene was reported, and its mutations were implicated in familial adenomatous polyposis (FAP), we constructed its gene knockout mouse strain as a model for FAP, and we began our study on various factors that possibly affect polyposis. While surveying literature, I noticed several lines of circumstantial evidence that COX-2 might play important roles in colon cancer and polyposis. Namely, there were rodent chemical carcinogen studies, and epidemiological and FAP trials data with non-steroidal anti-inflammatory drugs (NSAIDs). Although I had never worked on COX-2 before, this information rang a bell for me immediately, because cyclooxygenase was familiar to me. When I was a medical student, spending extracurricular hours in the Department of Medical Chemistry of Kyoto University Graduate School of Medicine, some scientists were purifying cyclooxygenase (now called COX-1) with Prof. Hayaishi. In addition, when we constructed the polyposis mice at Banyu-Tsukuba Research Institute (Merck), COX-2 inhibitors were in development at Merck-Frost, Canada, whereas COX-2 gene knockout mice were constructed at DuPont-Merck. Accordingly, we initiated collaborations with the two groups, and obtained COX-2 inhibitor MF-tricyclic and COX-2 gene knockout mice. At the same time, we constructed another COX-2 gene knockout with a lacZ reporter. While I was still working on the project, I had an offer from the University of Tokyo Graduate School of Pharmaceutical Sciences, and assumed the professor position of the new Genetics Department. Luckily, I obtained an agreement from both Banyu-Merck and the University of Tokyo to continue the project, and I was driving between the two institutions every week for more than a year.

  How would you describe the significance of this work for your field?

Inflammation research has a long history, and the role of cyclooxygenase was suggested by the seminal work of late Sir John Vane, who showed in the 1970s that NSAIDs, such as aspirin and indomethacin, suppress the biosynthesis of prostaglandins. As stated above, COX-1 was purified by Prof. Hayaishi and colleagues in 1976. Following the discovery of COX-2 in 1990-91, the role of COX-2 in inflammation was established. Independently of the inflammation research, involvement of cyclooxygenase in colon cancer was suggested by studies on NSAIDs, as stated above. These pieces of information, however, were mostly associative or only circumstantial. There was also some confusion. Epidemiological studies implicated cancer-preventive effects of aspirin, which is essentially a COX-1 inhibitor. (Only recently, the significant role of COX-1 has been worked out in molecular terms in the context of prostaglandin synthesis and tumorigenesis.) Thus, our Cell paper that combined the biochemical, genetic, and pharmacological data in the mouse model of polyposis, gave the most convincing evidence regarding the role of COX-2 in tumorigenesis, as a friend of mine called it the "nail-in-the-coffin" paper.

At the same time, these data gave a firm rationale for clinical applications. Around that time, some clinical COX-2 inhibitors became available, because many pharmaceutical companies were developing them for treatment of arthritis. Prompted by our paper, I hear, clinical trials were initiated first with FAP, and studies with celecoxib showed its efficacy, leading to the FDA approval of coxibs for FAP. Our paper also implicated a similar situation with other preneoplastic lesions, and cancers of various other organs, followed by many clinical trials. So, I might as well conclude this section by saying that our paper brought the COX-2 inhibitors into the arena of clinical cancer prevention.

  How much has this research advanced since you first started publishing on it?

There has been a lot of progress in two major aspects. In the basic research, many scientists continued studies on the role of COX-2 expression in cancer and preneoplastic lesions, as exemplified by tumor angiogenesis stimulation and apoptosis suppression. At the same time, expression of COX-2 and COX-1 has been studied extensively in various human cancer and preneoplastic lesions, suggesting the possibility for chemoprevention by COX-2 inhibitors.

Accordingly, many clinical trials were initiated to assess whether COX-2 inhibitors can prevent cancer progression. It is therefore unfortunate that risk of heart attack increases significantly with rofecoxib (Vioxx), a clinical COX-2 inhibitor from Merck. Whether this side effect is mechanism-based, or a compound-specific phenomenon, is currently being hotly debated. Although this is a question that should be carefully assessed clinically, I can offer some personal views on this issue. Initially, COX-2 inhibitors were developed to reduce the gastrointestinal (GI) side effects of NSAIDs. In fact, in the US alone, more than 16,000 patients are dying every year due to the side effects of NSAIDs. Therefore, for the patients who are at risk for the GI side effects of NSAIDs, COX-2 inhibitors appear to be a wonderful blessing. However, for the patients at risk of heart attack, it is a serious life-threatening issue. While the ratio of prostacyclin and thromboxane appears to play an important role in the heart, there are additional effects of COX-2 inhibition that have a strong selectivity. Functionally, COX-2 and COX-1 collaborate and compensate each other. For example, COX-1 appears to play a role as important as COX-2 in some tissues, and for COX-2 autoregulation by PGE₂ to work properly, it is essential for the basal level PGE₂ to be secured by COX-1. Therefore, we are walking on the thin ice that is the delicate balance between COX-1 and COX-2. It may be wise to clinically assess this balance for each particular patient subpopulation. In other words, COX-2 inhibitors with the strongest selectivity may be used for patients with high GI risk, but without heart problems, whereas those with partial selectivity may be used for patients with heart risk.

  Where do you see this research going 10 years from now?

In basic research, many new findings will be made regarding the mechanisms of COX-2 induction, and the effects of its product prostanoids, as well as the roles of the respective prostanoid receptors. Another important area of research that is rapidly expanding is the relationship of inflammation and some types of cancer. Although this is an old theme with some twists historically (e.g., Nobel prize-winning work that turned out to be wrong later), recent reports show unambiguous molecular and genetic evidence, and I believe that much more progress will be made in the coming years, including the role of COX-2 in cancer associated with inflammation.

Clinically, I am afraid that the pendulum of COX-2 inhibitors has swung from one end of hoopla to the other end of pessimism for a while, affected by the cardiac side effects of Vioxx. However, I hope it will come back to the center before too long, with more balanced and elaborate protocols for the respective COX-2 inhibitors, targeting particular patient subpopulations, based on their risk assessments in the heart and GI tract, respectively. In doing so, we may be able to use this new kind of pharmaceutical for the benefits of patients who need them.

  What lessons would you draw from your work to share with the next generation of researchers?