Introduction

There is an increasing body of evidence proving that cancer is a multistage genetic and epigenetic disease. Cancers originate from a single, still cycling cell (stem cell or proliferative progenitor cell), due to a genetic alteration, and subsequently tumour progression occurs through a clonal selection of cells which have acquired multiple additional genetic lesions. These additional genetic lesions are associated with six principal characteristics that are essential for malignant cancer and are called the 'hallmarks of cancer" (Hanahan and Weinberg, 2000): evading of apoptosis, self-sufficiency in growth signals, insensitivity to anti-growth signals, tissue invasion and metastasis, limitless replicative potential, sustained angiogenesis.

Although a large number of genetic changes can occur during carcinogenesis, only a limited number of pathways are really involved. The abnormal expression of cyclooxygenase-2 (COX-2) which occurs in many cancers and leads to an increased production of prostaglandin E₂ (PGE₂) appears to affect many of these pathways and by that most of the hallmarks of cancer.

COX enzymes play an important role in the biosynthesis of prostaglandins. They catalyze the conversion of arachidonic acid to PGG₂ and of PGG₂ to PGH₂. It is the rate limiting step in the biosynthesis of biologically active and physiologically important prostaglandins. Two COX-enzymes can be distinguished, COX-1 and COX-2. Although COX-3 has been reported, it is thought to be a slice variant of COX-1. The two isoforms share important similarities at the protein level, however, COX-1 and COX-2 are derived from distinct genes located on two different chromosomes and are encoded by mRNA of different sizes. The COX-1 is expressed in most tissues. COX-2, however, is normally absent and only induced by cytokines during inflammation or by tumours. In humans COX-2 is induced by many tumours already in the early phase of carcinogenesis. This is especially seen in epithelial tumours, like colorectal cancer, bladder cancer, mammary tumours, and squamous cell carcinomas.

COX-2/PGE2 and Carcinogenesis

The precise mechanism by which COX-2 in cancer is upregulated is still not fully understood. COX-2 induction is regulated by many signalling pathways, among which one is hypoxia. COX-2 can be induced by the hypoxic microenvironment in tumours and is mediated by HIF-1. In squamous cell carcinomas ultraviolet irradiation has been found as one of the etiologic factors. UV irradiation, however, also induces COX-2 upregulation in keratinocytes possibly by its action on the tumour suppressor gene p53. The way COX-2 on its turn influences several important processes in carcinogenesis is multiple and also still not completely known.

Apoptosis (programmed cell death) plays a central role in the control of proliferative cells. Failure of cells to go into apoptosis is therefore one of the hallmarks of cancer. COX-2/PGE₂ has been linked to the suppression of apoptosis in several ways. Deregulation of growth signalling causes cells to grow uncontrollably and unregulated by environmental influences. Ras-MAPK and PI3K/AKT signalling pathways are important proliferation stimulating pathways and have been found abnormally stimulated by an abnormal activation of the COX-2/PEG₂ pathway. In addition to the influence on apoptosis and growth-stimulatory signals, COX-2/PGE₂ also has an influence on anti-growth signals. Overexpression of COX-2 has been reported to cause down-regulation of the TGFβ type II receptor. TGFβ blocks cell progression through the G1 phase of the cell cycle. A second method by which COX-2 evades the anti-growth influences is by blocking the normal differentiation programme by activating β-catenin (the WNT pathway). Stimulation of this WNT pathway may also be responsible for the tumour cells to maintain the stem cell/progenitor cell phenotype. Apart from these effects over-expression of COX-2/PGE₂ induces the production of angiogenetic factors, e.g., vascular endothelial growth factor (VEGF), which are absolutely necessary for the formation of new blood vessels. Tumours larger than 1 mm in size are dependent on the formation of new blood vessels in order to be able to grow further. In the last hallmarks of cancer, tissue invasion and metastases, COX-2/PGE₂ also play important promoting roles by promoting cytoskeletal reorganisation and increasing of cell migration and invasion. Additional characteristics to the six hallmarks of cancer characteristics may also be influenced by increased PGE₂ levels. As an example PGE₂ has been reported to suppress immune responses by shifting the production of cytokines, away from a Th1 profile, leading to a more reduced production and activation of antitumour cytotoxic CD8+ T cells.

COX-2 Expression in Canine and Feline Cancer

A large series of studies have been performed into the expression profile of COX-2 in different canine malignant tumours. In general, epithelial tumours like transitional cell carcinoma (TCC), squamous cell carcinoma (SCC), intestinal-, nasal-, and mammary carcinomas are most frequently COX-2 positive, while sarcomas and hematopoietic tumours are often COX-2 negative. Normal prostate epithelial cells only express COX-1 and COX-2 is expressed in a much lesser degree by the stromal cells. In prostate carcinoma 75-82% of the tumours expresses COX-2. In mammary tumours some variation in COX-2 expression has been found, mostly related with the histological subtype. Inflammatory, anaplastic carcinomas are more often and stronger positive in COX-2 expression than adenocarcinomas. A high expression of COX-2 has been related in mammary tumours with the risk on metastases and with an unfavourable survival. Also in dogs with osteosarcomas strong COX-2 expression was related with decreased survival.

In the dog squamous cell carcinomas (SCC) are the tumours with the highest percentage of COX-2 expression. In a study in dogs with SCC all 40 cases (24 located at the feet, 9 in the skin, 7 in oral cavity) were positive by immunohistochemistry, while normal skin was negative. Immunoblotting analysis confirmed the restricted expression of COX-2 in SCC only. In contrast, only faint COX-2 expression was found in normal skin and SCC. In dogs with oral SCC COX-2 was expressed in a lower number of cases (17 of 26). Only 5 nasal SCC have been reported, but all were COX-2 positive.

In cats a smaller number of studies have been performed. Although in general a similar expression pattern of COX-2 has been found in cats as in dogs, the number of COX-2 positive epithelial tumours in cats is lower. COX-2 expression in feline SCC is different from that in the dog. In one study no COX-2 immunoreactivity was detected in any of the 6 cutaneous SCC, while of the oral cavity SCC only 2 of the 21 had weak to moderate COX-2 expression. Although in a larger series of oral SCC 37 of the 55 cases had COX-2 expression, still the majority of these cases had less than 1% positivity of all tumour cells. These results are a bit in contrast to later publications where one third of the 34 feline oral SCC appeared to be strongly positive with immunohistochemistry. Differences in results between these studies might be related to the use of different staining antibodies. In addition, PGE2 measurement might be a more accurate method for future studies, as it represents a functional assessment of COX-2 bioactivity.

Role of COX-2 Inhibitors

COX enzymes can be inhibited by NSAIDs. Two classes of NSAIDs can be distinguished: classical, non-selective COX inhibitors (including aspirin) and more specific COX-2 inhibitors. All classical NSAIDs are able to inhibit both COX-1 and COX-2 with a predominant on COX-1, whereas COX-2 inhibitors more or less selectively bind to COX-2. In humans COX-2 inhibitors have been used in the prevention and treatment of several tumour types, including colorectal polyposis/carcinomas, gastric and esophageal cancer, breast and prostate cancer, bladder cancer and cutaneous squamous cell carcinomas. Potential future use of COX-2 inhibitors could also include interference with the correlation of COX-2 and drug resistance.

Based on the above listed information some effects of inhibition of COX-2/PGE2 pathway might be expected in the treatment of some of these tumours, especially the epithelial tumours. Piroxicam has been used in prostate cancer in the dog with different outcomes. So far, no definitive proof exist for anticancer effects in this tumour type. This in contrast with TCC of the bladder where piroxicam was able to result in partial remissions in 6/18 dogs. In combined results of studies of oral SCC in the dog 8/26 dogs had tumour remissions after piroxicam treatment. In the cat hardly any clinical trial with COX-2 inhibitors has been published. Although in one study some beneficial effect of NSAID treatment was reported, w/wo other treatment, in a series of 16 cats with oral SCC in our clinic no single tumour responded to meloxicam.

Based on the COX-2 expression in tumours NSAIDs have a potential antitumour activity, especially in carcinomas. As results of administration of COX-2 inhibitors are not necessarily related to COX-2 expression or PGE2 concentration, but could also be related to still unknown other targets, well set up clinical trials should be performed in both the dog and cat. So far, no published reports of such trials are available, only some case series as reported above.