Pharmacogenomics of Open-Angle Glaucoma

treated with topical corticosteroids following photorefractive keratectomy, N363S heterozygotes were associated with an increased risk of elevated IOP following treatment with topical prednisolone acetate (Szabo et


Introduction
Pharmacogenomics is an evolving research discipline in medicine. Within ophthalmology, the earliest candidate gene investigations have studied primary and secondary open-angle glaucoma (OAG), as well as normal tension glaucoma (NTG). At this time, pharmacogenomics data are not generally used to make clinical decisions. However, as we collect data from clinical trials, the role of pharmacogenomics in the care of patients with glaucoma and allied diseases will become clearer. There are at least two potential future roles for applying pharmacogenomics in the treatment of OAG and allied disorders. First, more targeted therapy may lead to better treatment outcomes, with less exposure to medications in patients unlikely to respond to them. Second, pharmacogenomic research may lead to the development of novel therapies for these diseases.

Pharmacogenomics of open-angle glaucoma
Data from multiple randomized clinical trials (RCTs) have demonstrated that control of intraocular pressure (IOP) is generally effective in delaying progression of optic neuropathy and visual loss in patients with OAG, NTG, and ocular hypertension (OH) (Vass et al., 2007). At present, two major categories of medications used to treat these disorders include -adrenergic antagonists and prostaglandin analogs. Unfortunately, both classes of medications are associated with a number of patients who are nonresponders. For example, a secondary analysis of pooled data from phase 3 RCTs reported a nonresponse rate of 28% with the -blocker timolol maleate and 18% with the prostaglandin analog latanoprost (Camras and Hedman, 2003). Using current clinical examination techniques, there is no reliable way to differentiate responders from nonresponders prior to initiation of therapy. Unfortunately, this "trial and error" strategy leads to, in at least some patients, extra office visits and exposure to additional medications. The precise mechanisms of nonresponsiveness remain poorly understood, but a genetic component is suspected. It is hoped that pharmacogenomics may lead to earlier identification of nonresponders and more targeted treatment decisions (Moroi et al., 2009). At this time, genotype-phenotype correlations have been studied, primarily using a candidate gene approach, with both -adrenergic antagonists and prostaglandin analogs. In addition, corticosteroid treatments are frequently associated with secondary elevated IOP, and the pharmacogenomics of the steroid response have also been studied .
The clinical efficacy of topical -blockers may be affected by other genes. For example, timolol is metabolized by cytochrome P40 2D6 (CYP2D6). Polymorphisms in this gene are associated with reduced efficacy of oral timolol in patients with systemic hypertension (McGourty et al., 1985). In a series of 19 OAG patients and 18 volunteers, poor metabolizers of CYP2D6 demonstrated higher systemic concentrations of ophthalmic timolol, suggesting a potential safety concern in these patients (Nieminen et al., 2005). In a series of 133 OAG patients, systemic bradycardia following administration of topical timolol was significantly associated with the genotype at the CYP2D6 Arg296Cys polymorphism (Yang et al., 2009).

Prostaglandin analogs
Latanoprost is a highly selective agonist of the prostaglandin F 2 (FP) receptor (Stjernschantz et al., 1995). The FP receptor gene, located on chromosome 1p31.1, belongs to the family of G protein coupled receptors (Betz et al., 1999). In a prospective, nonrandomized clinical trial, 100 normal volunteers were treated with latanoprost for 1 week. Ten polymorphisms in the FP receptor gene, of which 2 were novel, were studied. The polymorphisms rs3753380 and rs3766355 showed statistically significant associations with the magnitude of response to latanoprost. The promoter assay revealed that the C allele of rs3766355 and T allele of rs3753380 were associated with lower transcriptional activity of the FP receptor gene, which was in agreement with the differences of IOP response to latanoprost based on genotypes of these polymorphisms (Sakurai et al., 2007). Using pathway analysis, the following polymorphisms were studied and found to have no statistically significant relationship with IOP reduction: T396A in prostaglandin transporter (Van Der Zwaag et al., 2002), P129T in fatty acid amide hydrolase (Sipe et al., 2002), -1607 insG in MMP-1 gene (Rutter et al., 2002), C-1306T in MMP-2 gene (Price et al., 2001), -1171 delA in MMP-3 gene (Ye et al., 1995), and C-1562T (Zhang et al., 1999) and CA repeats (-131~-90) in MMP-9 gene (St. Jean et al., 1995).

Corticosteroid-induced glaucoma
Some patients develop increased IOP and secondary OAG when exposed to corticosteroids. The etiology of this steroid response has never been fully explained, although a genetic determinant has been suspected for decades (Becker, 1965). Glucocorticoid receptors are present on the surface of trabecular meshwork cells, providing a possible mechanism for corticosteroid action on IOP (Weinreb et al., 1981). There are 6 well-known polymorphisms in the human glucocorticoid receptor gene (Tissing et al., 2005): at codon 766 (Koper et al, 1997); 5. a G→C substitution within intron 3 (Koper et al., 1997); 6. a G→T substitution within intron 4. (Koper et al., 1997). In a study of 102 patients treated with topical corticosteroids following photorefractive keratectomy, N363S heterozygotes were associated with an increased risk of elevated IOP following treatment with topical prednisolone acetate (Szabo et al., 2007).
Intravitreal triamcinolone acetonide (IVTA) is used as an off-label treatment of several retinal diseases. Clinically significant IOP elevation has been reported in about 40% of these patients (Smithen et al., 2004). In a pilot study of 52 patients (56 eyes) treated with IVTA for various retinal diseases, no statistically significant associations were detected between any of the 6 studied polymorphisms and IOP response following treatment (Gerzenstein et al., 2008). Other genes have been investigated for associations with the steroid response. The glucocorticoid receptor has multiple isoforms (Duma et al., 2006), and the expression of these isoforms is affected by the splicesome proteins SFRS9 (Xu et al., 2003) and SFRS5 (Yan et al., 2010), the immunophilins FKBP4 and FKBP5 (Zhang et al., 2008), and other proteins. In a series of 197 OAG patients, 107 steroid responders, and 400 controls, there were no statistically significant differences among the groups with respect to 48 polymorphisms in SFRS3, SFRS5, SFRS9, FKBP4, and the glucocorticoid receptor genes (Fingert et al., 2010).

Pitfalls in applying pharmacogenomics to glaucoma research
In pharmacogenomics, a positive result in one study may not be shown consistently in other studies. There are two main reasons for this discrepancy. First, different study populations may have very different baseline genetic characteristics. Second, other factors may influence drug efficacy in glaucoma patients.

Differences in study populations
Different study populations may have very different genetic characteristics. A polymorphism which is associated with nonresponsiveness to a certain medication may be insignificant in another study population because of the differences in the genetic backgrounds of subjects, or the polymorphism may not be informative due to a low minor allele frequency, or the population may be out of Hardy-Weinberg equilibrium. Therefore, interpretation of conflicting results of similar studies should be done with attention to the nature of the study populations.

Factors influencing drug efficacy
Other factors may influence drug efficacy in glaucoma patients. IOP fluctuation makes evaluation of IOP response problematic, because an IOP change involves both true pharmacological effect and spontaneous IOP fluctuations such as diurnal or day-to-day fluctuations. In addition, measurement errors in IOP are not negligible, especially when the magnitude of IOP reduction is small. Therefore, precise determination of drug efficacy is a key issue in pharmacogenomic studies of glaucoma.

Baseline IOP
Greater IOP reductions from topical medications have been reported to be associated with higher baseline IOP. For example, two studies examined the efficacy of latanoprost in patients with NTG (Rulo et al., 1996;Ang et al., 2004). Both studies showed that IOP reduction by latanoprost correlated significantly with baseline IOP. Currently, determinants of baseline IOP level in healthy subjects or in patients with NTG are largely unknown. Given that IOP fluctuates following a circadian rhythm, which is similar to the change in the activity of the sympathetic nervous system, the relationship between polymorphisms in adrenergic receptor genes and baseline IOP level was examined in two studies. In a US-based study of racially diverse patients, baseline IOP in normal subjects was reported to be significantly higher in Arg389 homozygotes of the 1-adrenergic receptor gene than in Gly389 carriers (Schwartz et al., 2005). In contrast, a study of untreated Japanese NTG patients reported that diurnal mean IOP was significantly higher for Ser49 homozygotes in the 1-adrenergic receptor gene than for Gly49 carriers, while the polymorphism at codon 389 was unrelated to the diurnal IOP level (Gao et al., 2010). In addition, two other polymorphisms, del 301-303 in 2B-adrenergic receptor gene and del 322-325 in the 2C-adrenergic receptor gene, were associated with the difference in diurnal IOP level. The conflicting results between the two studies regarding the SNP at codon 389 in the 1-adrenergic receptor gene may be attributed to different study populations, or to discrepancies between normal volunteers and untreated NTG patients, or to other factors.

Central corneal thickness
A thinner central cornea is a risk factor for the development of OAG, as reported by largescale clinical studies including the Barbados Incidence Study of Eye Diseases (Leske et al., 2008) and the Ocular Hypertension Treatment Study (OHTS) (Gordon et al., 2002). Furthermore, in OHTS participants, thicker corneas were associated with smaller IOP responses to -adrenergic antagonists and prostaglandin analogues than normal or thin corneas (Brandt et al, 2004). These findings were not explained by an applanation artifact from a thin cornea.

Race
Significant associations have been reported between race and drug efficacy of -adrenergic antagonists and prostaglandin analogs. Timolol has been reported to be less effective in black patients with glaucoma or OH than in nonblacks (Higginbotham et al., 2002). Travoprost was reported to be more effective, while timolol was reported to be less effective, in black patients with OAG or OH than in nonblack patients (Netland et al., 2003). However, other studies have reported no association between race and drug efficacy. For example, an analysis of OHTS participants reported no statistically significant differences in IOP response to nonselective -adrenergic antagonists or prostaglandin analogs between selfidentified African American and Caucasian individuals (Mansberger et al., 2007).

Other factors influencing evaluation of drug efficacy
A number of other factors, including IOP fluctuations, errors in IOP measurements, and differences in medication compliance may modify the post-treatment IOP value and make the true IOP response difficult to measure. Furthermore, the true IOP response may vary over time (Takahashi et al., 2008). A one-eye trial of glaucoma medication, where the untreated eye serves as a control to subtract the IOP fluctuations, has been advocated to assess the true IOP responses (Shields, 1998). However, clinical usefulness of the one-eye trial has been questioned by several reports due to asymmetrical IOP fluctuations, especially in glaucoma patients (Chaudhary et al., 2008;Realini, 2009). Also, a one-eye trial is not suitable for drugs with contralateral effects, such as -adrenergic antagonists. Therefore, IOP measurements at several time points before and after treatments are thought to be necessary to estimate the average of the true IOP responses for each patient.

Conclusions
There is some clinical evidence that polymorphisms in the 1 -adrenergic receptor gene and the FP receptor gene affect clinical response to, respectively, betaxolol and latanoprost in www.intechopen.com normal volunteers. The preliminary results with respect to betaxolol and latanoprost should be confirmed in patients with OH or OAG. The relationship between polymorphisms in the 1-adrenergic receptor gene and clinical response to nonselective -blockers, such as timolol, is as yet not fully determined. At this time, there is no convincing evidence of any pharmacogenomic relationship with respect to steroid-induced glaucoma following treatment with IVTA. However, these early findings are noteworthy and merit further investigation. Any potential pharmacogenomic association with steroid response might lead to a molecular drug target for future therapy of steroid-induced glaucoma, as well as a better understanding of the steroid response. Despite the recent advances in ophthalmic pharmacogenomics, there is still much that remains to be elucidated. For example, to our knowledge, there are currently no peerreviewed data regarding possible pharmacogenomic relationships affecting other medications used in the treatment of glaucoma, such as carbonic anhydrase inhibitors, 1agonists, and cholinergic agents. Even within the systems described here, there are many additional candidate genes and pathways for future association studies. Both the -adrenergic receptor and the FP receptor pathways utilize a second messenger system, interacting with a G-protein, a primary effector, a secondary messenger, and secondary effectors. Elements of these pathways, as well as their regulatory components, are reasonable candidates for future analysis.

Acknowledgements
This work is partially supported by NIH Center Grant P30-EY014801 and by an unrestricted grant from the University of Miami from Research to Prevent Blindness, New York, NY. SGS has received lecture fees from Bausch + Lomb and is co-holder of a patent pending entitled "Molecular targets for modulating intraocular pressure and differentiation of steroid responders versus non-responders."