Prostaglandin endoperoxide synthase (PTGS), also called cyclooxygenase (COX), is responsible for the conversion of arachidonic acid to prostaglandin H₂ (PGH₂), which is the common precursor for the biosynthesis of various prostanoids including prostaglandin D₂ (PGD₂), prostaglandin E₂ (PGE₂), prostaglandin I₂ (PGI₂) and thromboxane A₂ (TXA₂). Among them, PGE₂ and PGI₂ are known to be important inflammatory mediators. There are two isozymes of COX, COX-1 and COX-2. In general, COX-1 is a constitutive type, whereas COX-2 is an inducible type whose expression is increased by various stimuli (^{Smith et al., 1996}; ^{Funk, 2001}; ^{Rouzer and Marnett, 2008}; ^{Park and Kwon, 2011}).

Most non-steroidal anti-inflammatory drugs (NSAIDs) are COX inhibitors, which occupy the active site of COX to inhibit the enzyme's catalytic activity. Traditionally used NSAIDs including aspirin, ibuprofen and indomethacin inhibit the catalytic activity of both COX-1 and COX-2 non-selectively. In contrast, recently developed NSAIDs such as NS-398, celecoxib or rofecoxib inhibit the catalytic activity of COX-2 selectively (^{Flower, 2003}).

While underlying mechanisms of aging have not been fully elucidated, it has been proposed that the pro-inflammatory catalytic activity of COX-2 plays a crucial role in the aging process. Reactive oxygen species (ROS) generated in the process of normal metabolism or inflammation have been suggested to activate the redox-sensitive transcription factor NF-κB inducing the expression of pro-inflammatory genes such as COX-2, which in turn stabilizes chronic inflammatory status by producing ROS to cause tissue damage and aging (^{Chung et al., 2009}).

If the pro-inflammatory catalytic activity of COX-2 plays a causal role for aging, COX-2 inhibitors will have to inhibit the aging process. However, it has not yet been determined whether the catalytic activity of COX-2 is involved in the aging process and whether COX-2 inhibitors prevent aging. Actually, expression levels of COX-2 are observed to increase in heart and kidney of aged rats (^{Kim et al., 2000}; ^{Kim et al., 2001}), but not in brain of aged rats and endothelial cells of aged humans (^{Baek et al., 2001}; ^{Sanguino et al., 2004}; ^{Donato et al., 2008}). In addition, non-selective COX inhibitors such as acetylsalicylic acid, salicylic acid and indomethacin do not affect or rather shorten the life span in drosophila (^{Massie et al., 1985}). A recent report shows that celecoxib, a selective COX-2 inhibitor, extends the life span in C. elegans, which is mediated by a COX-2 catalytic activity-independent mechanism (^{Ching et al., 2011}).

In the present study, to elucidate whether the catalytic activity of COX-2 is a causal factor for aging, we analyzed the effect of COX-2 inhibitors on aging in the intrinsic skin aging model of hairless mice. Our data shows that NS-398 (a selective COX-2 inhibitor) inhibits skin aging, while celecoxib (a selective COX-2 inhibitor) and aspirin (a non-selective COX inhibitor) accelerate skin aging. In addition, the aging-modulating effect of the inhibitors was associated with the expression of p53, p16, type I procollagen and caveolin-1.

The pro-inflammatory catalytic activity of cyclooxygenase-2 (COX-2) has been proposed to be a crucial factor for individual aging as well as cellular senescence (^{Zdanov et al., 2007}; ^{Chung et al., 2009}). In the present study, we have investigated whether COX-2 inhibitors could prevent aging in the intrinsic skin aging model of hairless mice. Our data showed that among three COX-2 inhibitors studied, only NS-398 inhibits the skin aging while celecoxib and aspirin accelerate it (Figures 1^-3), suggesting that the catalytic activity of COX-2 does not mediate aging at least in skin. Consistent with these results, we had previously observed that among six COX-2 inhibitors studied, only NS-398 inhibited cellular senescence whereas two selective COX-2 inhibitors (celecoxib and nimesulide) and three non-selective COX inhibitors (aspirin, ibuprofen and flurbiprofen) accelerated the senescence in human dermal fibroblasts (HDFs) (^{Kim et al., 2008}).

It is still controversial whether cellular senescence is a source for individual aging. At least in skin, however, senescent fibroblasts have been detected in the dermis and their numbers have been found to increase with age in mice, baboons and humans (^{Dimri et al., 1995}; ^{Herbig et al., 2006}; ^{Wang et al., 2009}). According to our present and previous studies, the aging-modulating effect of COX-2 inhibitors in in vivo skin is consistent with the senescence-modulating effect of the inhibitors in dermal fibroblasts (Figures 1^-3; ^{Kim et al., 2008}). These findings support the idea that cellular senescence is a source for individual aging, and suggest that modulation of cellular senescence might be a good strategy to modulate individual aging.

Cellular senescence can be induced by diverse stimuli including dysfunctional telomeres, DNA damage, oncogenes and others, which ultimately activate either or both of the p53 and p16/pRb pathways. In general, the activity of p53 increases in the course of cellular senescence with or without an elevation of p53 expression (^{Atadja et al., 1995}; ^{Kulju and Lehman, 1995}). The p16 expression is usually enhanced in cellular senescence (^{Alcorta et al., 1996}). Compatible with these results, expression levels of p53 increases in brain of aged rats (^{Dorszewska and Adamczewska-Goncerzewicz, 2004}). In the case of heart, kidney, liver and lung, the total protein amount of p53 is not changed but the amount of acetylated p53 increases in aged rats (^{Braidy et al., 2011}). On the other hand, the p16 expression increases in most tissues of aged mice and rats (^{Krishnamurthy et al., 2004}). These studies suggest that the p53 and p16/pRb pathways are ultimate channels for individual aging, too. We observed here that the expression of p53 and p16 was changed by the treatment of COX-2 inhibitors, which was in agreement with the aging-modulating effect of the inhibitors (Figure 4). These data suggest that COX-2 inhibitors modulate skin aging by affecting both p53 and p16/pRb pathways and strongly support our conclusion at molecular levels that NS-398 inhibits intrinsic skin aging whereas celecoxib and aspirin accelerate it.

According to our data, aspirin did not alter the average wrinkle depth at the concentration of 5 mM, whereas 50 mM of aspirin markedly increased the average wrinkle depth (Figure 3C). However, skin fold thickness and epidermal thickness were decreased by the treatment of both 5 mM and 50 mM of aspirin, showing that aspirin treatment caused skin thinning at both concentrations (Figures 2C and 3C). In addition, expression levels of p53 and p16 were also elevated by the treatment of 5 mM of aspirin, verifying at molecular levels that 5 mM of aspirin accelerated the skin aging (Figure 4C). For the reason of this discrepancy, we assume that the thinning of skin precedes and wrinkle formation appears later in the course of intrinsic skin aging.

It is now uncertain how COX-2 inhibitors modulate the skin aging. Nonetheless, we have here suggested that COX-2 inhibitors modulate the skin aging by regulation of type I procollagen and caveolin-1 expression. We observed that NS-398 increased the type I procollagen expression while celecoxib and aspirin decreased it (Figure 5). Compatible with these results, it has been reported that NS-398 enhances type I procollagen expression in HDFs and promotes fibrosis in rat muscle (^{Han et al., 2004}; ^{Shen et al., 2005}). In addition, celecoxib reduces type I collagen expression in mouse mammary gland and rat liver (^{Paik et al., 2009}; ^{Lyons et al., 2011}). Aspirin has been reported to reduce collagen biosynthesis in HDFs (^{Surazynski et al., 2004}). Since type I collagen is the major constituent of the extracellular matrix (ECM) of skin and the reduction of its content is associated with degenerative changes as seen in the intrinsic skin aging (^{Rittié and Fisher, 2002}), our results suggest that COX-2 inhibitors might modulate the skin aging through regulation of type I procollagen expression. In addition, since type I collagen is the most abundant protein in the ECM of most tissues, there is a possibility that COX-2 inhibitors modulate aging not only in skin but also in other tissues. Further studies are required to elucidate the above possibility and the mechanism by which COX-2 inhibitors regulate the expression of type I procollagen.

Caveolae are flask-shaped vesicular invaginations in the plasma membrane and play central roles in receptor-mediated endocytosis. Caveolin-1 is the main component of caveolae, which has been proposed to be a key determinant for aging (^{Zou et al., 2011}). Actually, the expression of caveolin-1 is enhanced not only in cellular senescence but also in brain, heart, kidney, lung and spleen in aged rats (^{Park et al., 2000}). In addition, overexpression of caveolin-1 reduces cellular life span in mouse fibroblasts and siRNAs for caveolin-1 triggers reinitiation of DNA synthesis in senescent HDFs (^{Volonte et al., 2002}; ^{Cho et al., 2003}). It appears that this pro-aging effect of caveolin-1 is attributable to its capability to interact with other aging-modulating molecules in caveolae. Under the treatment of hydrogen peroxide, caveolin-1 interacts with the antioxidant enzyme thioredoxin reductase 1 (TrxR1) to inhibit its enzymatic activity, which activates the p53/p21 pathway and premature senescence in NIH3T3 cells (^{Volonte and Galbiati, 2009}). In addition, caveolin-1 interacts with the ubiquitin ligase Mdm-2 to sequester it to the caveolar membrane, resulting in the activation of p53 and premature senescence in human and mouse fibroblasts (^{Bartholomew et al., 2009}). We here reported that the aging-modulating effect of COX-2 inhibitors is closely associated with caveolin-1 expression levels (Figure 6), suggesting that the inhibitors might modulate the skin aging by regulation of caveolin-1 expression. Consistent with these findings, we have previously reported that among three selective COX-2 inhibitors studied, NS-398 inhibited caveolin-1 expression whereas the other inhibitors (celecoxib and nimesulide) enhanced the expression, which was closely associated with the senescence-modulating effect of the inhibitors in HDFs (^{Kim et al., 2008}). Further studies are required to elucidate the mechanism by which COX-2 inhibitors regulate the expression of caveolin-1.

In conclusion, the results of this study suggest that the catalytic activity of COX-2 is not a causal factor for intrinsic skin aging as evidenced by acceleration of the skin aging by celecoxib and aspirin among three COX-2 inhibitors studied. However, it still remains to be elucidated whether COX-2 catalytic activity is involved in the intrinsic aging in other tissues. Therefore, it is necessary to investigate continuously the function of COX-2 and effects of COX-2 inhibitors at individual levels.

This work was supported by the National Research Foundation of Korea (NRF) grant by the Korea government (MEST) through the Ageing and Apoptosis Research Center at Seoul National University (2008-0057303), Basic Science Research Program through the NRF funded by the Ministry of Education, Science and Technology (2010-0007188), and a grant from the National R&D Program for Cancer Control, Ministry for Health and welfare, Republic of Korea (1020420).