UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)
Molecular mechanisms underlying primary open angle glaucoma
Janssen, S.F.
Publication date
2014
Document Version
Final published version
Link to publication
Citation for published version (APA):
Janssen, S. F. (2014). Molecular mechanisms underlying primary open angle glaucoma.
General rights
It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).
Disclaimer/Complaints regulations
If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.
Molecular mechanisms underlying
primary open angle glaucoma
primary open angle glaucoma
The research presented in this thesis was carried out at the Netherlands Institute for Neuroscience (NIN), an institute of the Royal Netherlands Academy of Arts and Sciences, Department of Clinical and Molecular Ophthalmogenetics, Amsterdam, the Netherlands. This work was supported by grants from the General Dutch Foundation Preventing Blindness (ANVVB), National Foundation for Blinds and Low Vision (LSBS), Foundation Blinden Penning, Foundation Glaucoomfonds, Rotterdam Foundation for the Blind and the Professor Mulder Foundation (all together coordinated by UitZicht, project # UitZicht2008-7).
Publication of this thesis was financially supported by: Academic Medical Centre Amsterdam (AMC)
Landelijke Stichting voor Blinden en Slechtzienden (LSBS) Stichting Blindenhulp
Rotterdamse Vereniging Blindenbelangen Alcon
Allergan Thea Farma Ursapharm
© 2014 S.F. Janssen, Amsterdam, the Netherlands
Cover: Nikki Vermeulen, Ridderprint BV, Ridderkerk, the Netherlands Lay-out: Nikki Vermeulen, Ridderprint BV, Ridderkerk, the Netherlands Print: Ridderprint BV, Ridderkerk, the Netherlands
primary open angle glaucoma
ACADEMISCH PROEFSCHRIFT ter verkrijging van de graad van doctor
aan de Universiteit van Amsterdam op gezag van de Rector Magnificus
prof. dr. D.C. van den Boom
ten overstaan van een door het college voor promoties ingestelde commissie,
in het openbaar te verdedigen in de Agnietenkapel op dinsdag 4 november 2014, te 14:00 uur
door
Sarah Frederique Janssen geboren te Amsterdam
Promotores: Prof. dr. A.A.B. Bergen Prof. dr. N.M. Jansonius
Co-promotor: Dr. T.G.M.F. Gorgels
Overige leden: Prof. dr. F. Baas
Prof. dr. C.M. Duijn Prof. dr. M. Kamermans Prof. dr. E.J. Meijers – Heijboer Prof. dr. A.C. Moll
Prof. dr. M.P. Mourits Prof. dr. A.H. Zwinderman
door te begrijpen Willem Frederik Hermans
The first aim of this thesis was to find potentially new POAG disease genes, using collaborative GWAS studies.
The second aim of this thesis was to select and give a comprehensive overview of disease genes truly (not) implicated in POAG so far, to understand their functional role in the etiology of POAG better, and to test their potential usefulness for (future) personalized medicine.
The third aim of this thesis was to understand the pathobiology of POAG better. Since the trabecular meshwork and optic nerve head of the eye have been studied extensively elsewhere, we decided to focus on the characterization of the ciliary body epithelium (in POAG).
The fourth aim of this thesis was to build a comprehensive pathobiological model of POAG, as starting point for further genetic (risk) studies, pharmacogenetic and functional studies.
Chapter 1 Introduction
The vast complexity of primary open angle glaucoma: disease genes, risks, molecular mechanisms and pathobiology.
Prog Ret Eye Res 2013;37:31-67
11
Chapter 2 Common genetic variants associated with open-angle glaucoma. Hum Mol Genet 2011;20(12):2464-71
95 Chapter 3 Common genetic determinants of intraocular pressure and primary
open-angle glaucoma.
PLoS Genet 2012;8(5):e1002611
111
Chapter 4 Gene expression and functional annotation of the human ciliary body epithelia.
PLoS One 2012;7(9):e44973
135
Chapter 5 In silico analysis of the molecular machinery underlying aqueous humor production: potential implications for glaucoma.
J Clin Bioinforma 2013;3:21
173
Chapter 6 Gene expression and functional annotation of the human and mouse choroid plexus epithelium.
PLoS One 2013;8:e83345
195
Chapter 7 Gene expression-based comparison of the human secretory neuroepithelia of the brain choroid plexus and the ocular ciliary body: potential implications for glaucoma.
Fluids Barriers CNS 2014;11:2
223
Chapter 8 General discussion 245
Chapter 9 Summary 251
Chapter 10 Nederlandse samenvatting 255
Chapter 11 List of publications 261
Chapter 12 Dankwoord 265
Chapter 13 PhD Portfolio 271
Chapter 1
The Vast Complexity of Primary Open Angle Glaucoma:
Disease Genes, Risks, Molecular Mechanisms
and Pathobiology
Sarah F. Janssen, Theo G.M.F. Gorgels, Wishal D. Ramdas, Caroline C.W. Klaver, Cornelia M. van Duijn, Nomdo M. Jansonius, Arthur A.B. Bergen
Abstract
Primary open angle glaucoma (POAG) is a complex progressive optic nerve neuropathy triggered by both environmental and genetic risk factors. Several ocular tissues, including the ciliary body, trabecular meshwork and optic nerve head, and perhaps even brain tissues, are involved in a chain of pathological events leading to POAG.
Genetic risk evidence for POAG came from family linkage-studies implicating a small number of disease genes (MYOC, OPTN, WDR36). Recent Genome Wide Association Studies (GWAS) identified a large number of new POAG loci and disease genes, such as CAV1, CDKN2B and GAS7. In the current study, we reviewed over 120 family and GWA studies. We selected in total 65 (candidate) POAG disease genes and proceeded to assess their function, mRNA expression in POAG relevant eye tissues and possible changes in disease state. We found that the proteins corresponding to these 65 (candidate) POAG disease genes take part in as few as four common functional molecular networks. Functions attributed to these 4 networks were developmental (dys)function, lipid metabolism, and inflammatory processes. For the 65 POAG disease genes, we reviewed the available (transgenic) mouse models of POAG, which may be useful for future functional studies. Finally, we showed that the 65 (candidate) POAG genes substantially increased the specificity and sensitivity of a discriminative POAG risk test. This suggests that personal risk assessment and personalized medicine for POAG are on the horizon. Taken together, the data presented are essential to comprehend the role of genetic variation in POAG, and may provide leads to understand the pathophysiology of POAG as well as other neurodegenerative disorders, such as Alzheimer’s disease.
1
1. Introduction
Primary open angle glaucoma (POAG) is an insidious, sight-threatening neurodegenerative disease of the optic nerve, which affects a large proportion of the (elderly) population. Despite tremendous research efforts, the etiology of POAG remains obscure - probably due to the heterogeneous and complex nature of the disease - both on a clinical and on a molecular level.
In the past decade, much effort has gone into elucidating the genetic causes and risk factors of POAG. Application of advanced genetic technology, such as Genome-Wide Association Studies (GWAS), large-scale gene expression studies and proteomics, has yielded a wealth of information. A vast amount of new (candidate) POAG disease genes has been identified, and insights into the molecular mechanisms underlying POAG are increasing. One obvious next step is to attempt to understand the function of these genes and how they act in the pathobiology of POAG. In our lab, and in collaboration with the Rotterdam Study, we recently identified new candidate POAG genes (CDKN2B, ATOH7, SIX1, borderline CDCD7/ TGFBR3 and SALL1 (Ramdas et al., 2010, 2011a); GAS7 and TMCO1 (van Koolwijk et al., 2012) and we investigated the gene expression levels in POAG-related tissues, such as ciliary body epithelia (Janssen et al., 2012).
In this review, we discuss a variety of aspects of the genetics and the molecular mechanisms underlying POAG and we postulate a model of the pathobiological events that lead to the disease. We discuss the different tissues involved in POAG, their pathobiological changes and their possible role in the disease. Then, we review all the (candidate) POAG loci and genes currently known, and select scientifically confirmed entries. We then review our own and the literature’s gene expression data and we determine if, and to what extent, the (selected candidate) POAG genes are expressed in the POAG relevant tissues. Subsequently, for a better understanding of the kind of pathways the POAG genes are involved in, we functionally annotate these (candidate) POAG genes. Finally, in order to gain an insight into the potential role of the disease genes in POAG, and to facilitate future investigations, we review the currently available transgenic mouse models related to the (candidate) POAG genes that developed features of glaucoma.
2. Common features of Primary Open Angle Glaucoma
2.1. Primary Open Angle Glaucoma (POAG)
POAG is a progressive optic neuropathy with characteristic changes to the optic nerve head and corresponding visual field defects that is not explained by other diseases (as opposed to secondary glaucomas), together with a normal open anterior chamber (as opposed to angle closure glaucoma) (Casson et al., 2012; Gupta and Weinreb, 1997; Quigley, 2005; Shields
and Spaeth, 2012). POAG usually has an onset in adulthood (as opposed to congenital and juvenile glaucoma). From a pathobiological point of view, POAG is characterized by progressive retinal ganglion cell (RGC) death. POAG is an insidious disease: for a long time after the start of the disease, the gradual visual loss will not hamper daily life and stays unnoticed. The visual field loss starts usually in the periphery, shows initially little overlap between the eyes, and is masked by “filling-in” of the assumed image by the brain. The disease will progress to irreversible blindness. Lowering of the intraocular pressure (IOP) slows down the progression and is the only treatment proven to be effective (Maier et al., 2005).
Clinical examination of the optic nerve shows a progressive optic disc cupping resulting in a large cup-to-disc ratio (Figure 1), as well as optic disc haemorrhages, acquired pits (Radius et al., 1978) and parapapillary atrophy (Buus and Anderson, 1989; Jonas et al., 1989, 1992; Kasner et al., 1989; Rockwood and Anderson, 1988). Characteristic visual field defects include a nasal step scotoma, inferior, superior or paracentral scotoma and generalized depression (reviewed in Weinreb and Khaw, 2004). The IOP is frequently increased, but may also be within the normal range (normal pressure glaucoma [NTG]). High and normal IOP are classically divided by the IOP cut-off level of 21 mmHg. It is currently unclear which IOP parameter is the most important in POAG: mean IOP, peak IOP, IOP fluctuation or pre-treatment IOP (see Wesselink et al., 2012 for discussion).
A B
C D
Figure 1: Optic disc cupping in POAG
(A) Glaucomatous optic disc of a human eye showing advanced cupping and a visible lamina cribrosa. (B) Corresponding visual field showing central and temporal islands of vision. (C) Normal human optic disc. (D) Normal visual field test showing the blind spot located 15 degrees temporally from fixation (all from left eyes).
1
2.2. Prevalence of POAG
In western countries, the prevalence of POAG in the elderly population varies between 1-3% (De Voogd et al., 2005; Wolfs et al., 2000). The prevalence in Asian and African countries is between 1-4% (George et al., 2010; Iwase et al., 2004; Liang et al., 2011; Zhong et al., 2012) and 2-8% (Leske et al., 2007; Ntim-Ambonsah 2004; Rotchford et al., 2003), respectively. The incidence increases with age, and men seem to be more at risk than women (Czudowska et al., 2010; De Voogd et al., 2005; Leske et al., 2007) although one study showed an increased risk for POAG in women compared to men (Bengtsson, 1989). Worldwide, POAG causes around 12% of the blindness, which makes the disease the second cause of irreversible blindness (reviewed in Chen, 2004). In African populations, POAG is even the primary cause of blindness, as it is responsible for one-third of all blindness cases (Rotchford et al., 2003).
2.3. Risk factors for POAG
POAG is a heterogeneous and multifactorial neurodegenerative disease. Although the pathophysiology is largely unknown, there is an extensive list of risk factors that are possibly or probably associated with POAG. These include ocular risk factors, systemic risk factors and environmental and (molecular) genetic risk factors.
Ocular risk factors include a higher IOP, a thinner central corneal thickness (CCT; Aghaian et al., 2004; Dueker et al., 2007; Herndon et al., 2004; Morita et al., 2012; Pakravan et al., 2007; Saenz-Frances et al., 2011; Shah et al., 1999; Singh et al., 2001; Wolfs et al., 1997) and myopia (Marcus et al., 2011), the latter possibly because of an unfavorable anatomical constitution of the eye that could ultimately effect optic nerve damage (Jonas and Dichtl, 1997).
A number of systemic features may be associated with POAG: An increased body mass index (BMI) appears to be associated with a decreased risk for POAG (Leske et al., 1995), especially in women (Pasquale et al., 2010; Ramdas et al., 2011b). Hearing loss appears to be more frequently present in patients with exfoliation glaucoma (Cahill et al., 2002; Papadopoulos et al., 2012; Samarai et al., 2012; Shaban and Asfour, 2004; Yazdani et al., 2008), possibly due to fibrils in the inner ear. The association between hearing loss and POAG is less consistent. Several studies showed that a proportion of POAG patients have impaired auditory function (Kremmer et al., 2004; O’Hare et al., 2012), but other studies did not confirm this relationship (Hayreh et al., 1999; Shapiro et al., 1997).
A number of cardiovascular risk factors were studied in POAG. The population-based data on blood pressure and perfusion pressure are inconclusive (Bonomi et al., 2000; Dielemans et al., 1995; Leske et al., 2002, 2009; Mitchell et al., 2004; Muskens et al., 2007; Ramdas et al., 2011b; Tielsch et al., 1995a), as is the role of atherosclerosis (Hewitt et al., 2010; de
Voogd et al., 2006a) and diabetes (Dielemans et al., 1996; Klein et al., 1994; Leske et al., 1995; Mitchell et al., 1997; Tielsch et al., 1995b; de Voogd et al., 2006b).
The relationship between POAG and lifestyle (e.g. dietary aspects or drug intake) is not clear. Use of glucocorticoids may be a risk factor for development of ocular hypertension in certain people, and a high percentage of POAG patients have an increased cellular sensitivity to glucocorticoids (so called ‘steroid responders’) (Armaly and Becker, 1965; Becker, 1965; for a recent overview see Marcus et al., 2012a). Glucocorticoids affect the trabecular meshwork (TM) outflow system (Armaly and Becker, 1965; Miller et al., 1965) and the aqueous humor (AH) production by the CB (Kimura and Honda, 1982; Southren et al., 1979). Long-term use of statins might reduce the risk of POAG (Marcus et al., 2012b; Stein et al., 2012), while use of anticoagulants or platelet aggregation inhibitors appear to have no effect (Marcus et al., 2012b). Low intake of retinol and vitamin B1 and high intake of magnesium may increase the risk of developing POAG (Giaconi et al., 2012; Ramdas et al., 2012). Serum levels of uric acid (Yuki et al., 2010) and homocysteine (Xu et al., 2012) are increased in POAG patients and levels of vitamin C (Yuki et al., 2010) are decreased compared to controls. Smoking (Edwards et al., 2008; Ramdas et al., 2011b; Renard et al., 2012), alcohol consumption (Kang et al., 2007; Ramdas et al., 2011b), caffeine use (Kang et al., 2008), and fat intake (Kang et al., 2004) are not associated with POAG risk.
Classic genetic risk factors for POAG include a positive family history (Czudowska et al., 2010; Le et al., 2003; Leske et al., 1995; Mitchell et al., 2002; Tielsch et al., 1994; Wolfs et al., 1998) and race (Tielsch et al., 1991). For example, the prevalence of POAG among African-Americans is much higher than among white Americans (Zhang et al., 2012). Obviously, DNA analysis recently confirmed and refined the genetic risk for POAG, which is extensively discussed in the sections below.
3. Tissues and media involved in Primary Open Angle Glaucoma
3.1. Cornea
The cornea, particularly in terms of its thickness and firmness, plays an important role in (the diagnosis of) glaucoma. Several studies have measured central corneal thickness (CCT) and determined the “normal” CCT in a population (Shah et al., 1999; Suzuki et al., 2005; Wolfs et al., 1997). In a meta-analysis of Doughty and Zaman (2000) an average normal CCT of 473-597 µm (95% CI, 474-596) was found. Aging may cause a decline in CCT (Brandt et al., 2008; Cho and Lam, 1999; Foster et al., 1998) although not all studies found this effect (Doughty and Zaman, 2000; Nomura et al., 2002; Suzuki et al., 2005; Wolfs et al., 1997).
1
Central corneal thickness and IOP measurement
There is a positive correlation between CCT and the error of applanation tonometry to measure IOP. This correlation is statistically highly significant (Doughty and Zaman, 2000; Ehlers et al., 1975; Kohlhaas et al., 2006; Suzuki et al., 2005; Wolfs et al., 1997). The IOP is thus frequently overestimated in eyes with thicker corneas and underestimated in thinner ones. In the clinic, one should correct for this phenomenon, since a patient with a “high IOP” and a thick cornea does not necessarily have a harmfully high IOP, whereas a patient with “normal IOP” and a thin cornea might suffer from an IOP that is actually high and harmful.
Cornea and glaucoma
The CCT may thus account for an error in IOP measurement. In addition, a thinner CCT may also be a risk factor for glaucoma and this may be due to structural changes of the cornea, making the eye less capable of restraining the IOP. Indeed, corneal resistance is lower in NTG (Grise-Dulac et al., 2012; Morita et al., 2012) and a thinner CCT at presentation is associated with worse cup-disc ratios (Congdon et al., 2006; Herndon et al., 2004; Jonas et al., 2005; Lesk et al., 2006; Viswanathan et al., 2013). Consequently, subjects with thinner corneas and OHT are more at risk to develop POAG compared to subjects with thicker corneas and OHT (Gordon et al., 2002). Overall, a thinner CCT seems to be associated with NTG and POAG (Aghaian et al., 2004; Morita et al., 2012; Saenz-Frances et al., 2011; Shah et al., 1999; Singh et al., 2001; Wolfs et al., 1997) and a thicker CCT with OHT (Ventura et al., 2001; Medeiros et al., 2003; Wolfs et al., 1997).
3.2. Ciliary body epithelia (CBE)
Anatomy
The ciliary body (CB) is a ring-shaped tissue located in the posterior ocular chamber between the iris and the ora serrata. The CB contains the ciliary muscle and a double layer of two, partly folded, neuro-epithelia, the non-pigmented and the pigmented epithelium (respectively, the NPE and PE) (Figure 2). The PE lies next to the vascular stroma “outside the eye” and the NPE faces the posterior chamber “inside” the eye. The ciliary muscle is responsible for accommodation of the lens through the zonular fibers and acts on the TM to promote AH outflow.
Main functionalities CBE
The NPE and PE work together to produce the AH (reviewed in Civan and Macknight, 2004) and are involved in the genesis and maintenance of the vitreous (reviewed in Bishop et al., 2002). The CBE, too, have been implicated in a number of additional functionalities, including neuro-endocrine properties (reviewed in Coca-Prados and Escribano, 2007), the
ocular immune privilege (reviewed in Sugita, 2009), and their still controversial, neuro-developmental function (Bhatia et al., 2009; Cicero et al., 2009; Gualdoni et al., 2010; Inoue et al., 2010; Martinez-Navarrete et al., 2008; Xu et al., 2007). Using whole genome gene expression studies and bioinformatics, we recently annotated the main functionalities of the NPE and PE. This analysis confirmed and specified the involvement of the (N)PE in developmental processes, endocrine and metabolic signaling, immunological functions and the neural nature of the tissue (Janssen et al., 2012).
Lens Posterior chamber
Iris
Anterior chamber Cornea
TM
CB
CM
Figure 2: Histology of anterior chamber and ciliary body
Light micrograph of the anterior chamber and ciliary body of a healthy human eye. The image was adapted and reprinted with permission from prof. Adriana Silva Borges-Giampani and prof. Jair Giampani Junior (Borges-Giampani and (Borges-Giampani Junior, 2013) Abbreviations: TM: trabecular meshwork; CB: ciliary body; CM: ciliary muscle.
Transport and AH production mechanism
The AH is not a simple passive filtrate of the plasma, but has a unique composition due to a combination of para-cellular filtration, active local trans-cellular (N)PE transport and
local (N)PE protein synthesis. Several well established ion-channels, including Na+-K+-2Cl-
symporters, Na+-K+-activated ATPase, K+ and Cl- channels and water channels (aquaporins),
form the key for AH production (reviewed in Civan and Macknight, 2004). The (N)PE expressed many other genes coding for specific transporters, for example amino acids, glutamate, GABA, iron, copper, zinc, thyroid hormone, vitamins, glucose, and fatty acids. The rate of AH formation fluctuates during the day, with the lowest rate during night and the highest rate in the morning (McLaren et al., 1990; Reiss et al., 1984).
1
Changes with aging and glaucoma
A number of histological changes were found in the aging CB: thickening of the basement membranes (Okuyama et al., 1993; Rohen and Zimmermann, 1970; Schlotzer-Schrehardt et al., 1990), increased density of collagen fibers, augmentation of degenerative elements including lipids, lipofuscin granules and lysosomes (Schötzer-Schrehardt et al., 1990), and increased diameter of the PE with plug-like protrusions in the stromal surface (Rohen and Zimmermann, 1970). POAG related CB changes were: (1) the ciliary muscle contained an increased amount of plaque material. The anterior tendons of the ciliary muscle appeared to be glued together and the fiber sheaths of neighboring tendons tended to merge (Lütjen-Drecoll et al., 1988), (2) hyalinization and atrophy of the ciliary muscle (Fine et al., 1981). This interferes with the muscle function and outflow facility via the uveoscleral route, which may result in decreased outflow of the AH and increased IOP; and (3) in a rat model for glaucoma, also the NPE and PE showed extensive hyalinization of the stroma (Funk and Rohen, 1985). This could interfere with AH production and composition.
3.3. Aqueous humor (AH)
AH composition
The AH is secreted by the CBE and exits the eye through the TM back into the venous blood system. The balance between AH production and outflow determines the IOP. The IOP is necessary for good eye shape and optimal optical properties. In healthy subjects, the IOP fluctuates during the day and night around the average of 15 mmHg.
The entries in the AH come from the CBE, partly by passive para-cellular filtration from the plasma, partly by trans-cellular active transport, and partly by production of the CBE itself. The AH contains about 20 times fewer proteins than plasma, which is crucial to maintain ocular transparency. On the other hand, the ascorbate concentration in the AH is 20 times higher than in plasma, probably to protect the eye against light-induced oxidative stress. Other compounds that are present in the AH are glucose (Davies et al., 1984), growth factors (Welge-Lussen et al., 2001), neuroendocrine signaling factors (reviewed in Coca-Prados and Escribano, 2007), albumin, transferrin, lactate, antitrypsin, antioxidants, immunoglobulins and anti-angiogenic proteins (Chowdhury et al., 2010; Richardson et al., 2009). These proteins reflect the various functions of the AH, namely: nourishment of the nonvascularized tissues in the eye (TM, cornea and lens), interference with innate and adaptive immunity maintaining the ocular immune privilege, (neuro-endocrine) signaling, involvement in catalytic and enzymatic process and structural functions (Chowdhury et al., 2010; Richardson et al., 2009; Sharma et al., 2009).
AH production and composition: aging and glaucoma
With age, the rate of AH production declines and the outflow resistance increases. The overall effect is that the IOP is largely independent of age in Caucasians. Some studies reported a small but significant increase of IOP with age in Caucasians (Giuffre et al., 1995; Klein et al., 1992) whereas others did not (Costagliola et al., 1990; Gaasterland et al., 1978; Rochtchina et al., 2002; Wolfs et al., 2000) and one study reported a negative association (Gonzalez-Meijome et al., 2006). For the Asian, and especially the Japanese population, the studies are more consistent and show a negative association between age and IOP (Fukuoka et al., 2008; Lee et al., 2002; Shiose 1984; Shiose and Kawase, 1986). With aging, the average turnover of AH decreases, and it’s composition changes. In the elderly, AH contained increased concentrations of albumin, alpha 1-acid glycoproteins and transferrin (Inada et al., 1988).
In glaucomatous eyes, the rate of AH production remains essentially the same (Larsson et al., 1995). However, in POAG, the composition of AH changed with an overall increase in protein concentration compared to controls (Duan et al., 2010; Izzotti et al., 2010a). We reviewed the literature on these changed protein levels in POAG AH and summarized the findings in Supplementary Table S1. Examples of proteins with increased levels in POAG AH are erythropoietin, soluble CD44, transthyretin, and transferrin. Functionally, these proteins appeared to be involved in oxidative and mitochondrial damage, neural degeneration and apoptosis (Izzotti et al., 2010a). Examples of proteins with decreased concentrations in POAG AH are hyaluronic acid (Navajas et al., 2005) and osteopontin (Chowdhury et al., 2011).
3.4. Trabecular meshwork (TM)
Anatomy
The TM is located in the angle between the cornea, sclera and iris (Figure 2) and is subdivided into four compartments. The anatomy of the TM and its compartments are extensively reviewed elsewhere (Tamm, 2009). In summary: The first compartment, closest to the anterior angle, is the uveal meshwork, which consists of connective tissue prolongations from the iris and CB. These thin, cord-like, trabeculae are covered with endothelial cells and do not offer much resistance to the AH outflow. Next, the AH flows through the corneoscleral meshwork, a stack of perforated laminar sheets, which provides the first resistance to the AH flow. In molecular terms, this meshwork does not only contain elastin and collagen fibers, but also laminin, glycoproteins, collagens, hyaluronic acid and myocilin. The third TM layer is the juxtacanalicular or cribriform meshwork, a dense extracellular matrix network full of glycoproteins with narrow intercellular spaces that is responsible for the majority of the outflow resistance. Finally, the last TM barrier for the AH are the endothelial cells of the Schlemm’s canal. Giant vacuoles and intracellular
1
pores in the endothelial cells of the Schlemm’s canal facilitate bulk flow of the AH. The giant vacuoles respond rapidly to changes in IOP by increasing their number and size and by forming intracellular pores. From here, the AH enters the episcleral veins.
Changes with aging and glaucoma
With aging, plaque-like deposits develop in the TM (Boldea et al., 2001). Furthermore, there is an increased amount of fibronectin (Babizhayey and Brodskaya 1989), progressive coalescence of collagens, increase in previously unrecognized matrix material (Gong et al., 1992), decreased amount of hyaluronic acid, accumulation of chondroitin sulfates (Knepper et al., 1996) and an overall decrease in cellularity (Alvarado et al., 1984). In the endothelial cells of the aging Schlemm’s canal there is a reduction of intracellular pores and giant vacuoles (Boldea et al., 2001). These changes in the TM and Schlemm’s canal can cause a decreased AH outflow.
Histological and functional changes were observed in POAG affected-TM, including (1) decline in number of TM cells (Alvarado et al., 1984), (2) increase and changed ECM components, including hyalinization and thickening of the trabeculae (Rohen et al., 1993), increased amount of fibrillar components (Lütjen-Drecoll et al., 1981), increase in fibronectin (Babizhayey and Brodskaya 1989), and increase in elastin (Umihari et al., 1994), (3) deposition of extracellular plaques (Gottanka et al., 1997; Rohen et al., 1993;) and (4) stiffening of the TM with decreased contractility force of the elastic fibers (Last et al., 2011). POAG-affected TM cells showed signs of stress (increased expression of heat-shock protein αB-crystallin; Lütjen-Drecoll et al., 1998), oxidative stress (increased expression of iNOS; Fernandez-Durango et al., 2008), mitochondrial defects (He et al., 2008; Izzotti et al., 2010b), and apoptosis (Baleriola et al., 2008). In a gene expression study of POAG-affected TM, Liton et al. (2006) found upregulation of genes coding for inflammation and acute-phase response, lipid metabolism, G-protein signaling and ionic transport compared to healthy control TM. Genes downregulated in POAG-affected TM included members of the vacuolar protein sortin-10 domain, several members of the solute carrier family, ceruloplasmin and paraoxonase 3 (Liton et al., 2006). Together, these intrinsic changes of the TM in POAG result in decreased AH outflow.
3.5. Retinal ganglion cells (RGCs)
Anatomy
The cell bodies of RGCs constitute the innermost cellular layer of the retina. They receive input from the photoreceptors via bipolar and amacrine cells. The unmyelinated axons of the RGCs exit the eye at the optic nerve head (ONH) supported by the lamina cribrosa (LC) (Figure 3). Outside the eye, the RGC axons become myelinated and this myelination results in an increase in diameter of the optic nerve of around 40% (Elkington et al., 1990).
The axons partly cross at the optic chiasm and run to the lateral geniculate body in the thalamus. Optic nerve LC Retina layers Axons of RGCs Figure 3: Histology of the optic nerve head
Light micrograph of a section through the optic disc of a human eye. The figure shows the optic nerve head (ONH), with the RGC axons, lamina cribrosa (LC) and the optic nerve. The retina layers are also visible. Figure is adapted from Jonas et al. (2003), with approval of the journal.
Changes with aging and glaucoma
With aging, there is a decreasing amount of RGCs (Curcio and Drucker, 1993; Dolman et al., 1980; Gao and Hollyfield, 1992) and optic nerve fibers (Balazsi et al., 1984), as well as a decreasing thickness of the retinal nerve fiber layer (RNFL) (Da Pozzo et al., 2006; Harwerth et al., 2008; Kergoat et al., 2001).
In POAG, RGCs die and RGC axon loss contributes to cupping of the optic nerve head (Wirtschafter, 1983). The mechanisms of RGC death in glaucoma have been discussed in a number of recent reviews (Almasieh et al., 2012; Burgoyne, 2011; Downs et al., 2008; Qu et al., 2010). In summary, damage to RGC axons may occur through a number of complementary mechanisms, including mitochondrial dysfunction and oxidative stress, activation of extrinsic and intrinsic apoptotic pathways, lamina cribrosa deformation, vascular dysfunction and ischemia, excitotoxicity, glial activation and auto-immunity pathways.
1
3.6. Lamina cribrosa (LC)
Anatomy
The lamina cribrosa is a fibrous network encapsulated by astrocytes through which the non-myelinated axons of the RGCs exit the eye. Astrocytes are the major cell type of the ONH and are vital for RGC health. Astrocytes can remodel ECM (by producing collagens and elastins for example) and remove of waste products. Alongside the astrocyte-surrounded axons, there are blood vessels and quiescent microglia. The LC can be subdivided in two portions, the anterior (choroidal) and posterior (scleral) portion (Ogden et al., 1988). The anterior portion of the LC is characterized by fibrous astrocyte processes surrounding bundles of non-myelinated axons of the optic nerve. Occasionally, small bundles of collagen fibers are present within the anterior portion of the LC. The posterior portion of the LC also has an intimate mixture of small bundles of axons and astrocyte processes, but these are embedded in a collagenous trabecular canal. Each canal is divided into many small compartments by a web of fine astrocyte processes, through which the axons pass (Elkington et al., 1990; Radius and Gonzales, 1981). The major types of collagen in the LC are type I and III (Morrison et al., 1989).
Changes with aging and POAG
Upon aging, the LC increases in thickness and becomes increasingly more collagenous (Albon et al., 1995; Hernandez et al., 1989; Kotecha et al., 2006; Ogden et al., 1988), which results in increased stiffness and decreased resilience (Albon et al., 2000). These changes in mechanical properties can make older subjects more susceptible to axonal damage. In glaucoma, the LC suffers from various histopathological changes, including disturbances in collagen composition (Hernandez et al., 1990; Tengroth and Ammitzboll, 1984), loss of elastin fibers (Hernandez 1992), compression of the LC sheets, and, in later stages, a backward bowing of the entire LC (Quigley et al. 1983). The initially circular LC pores become more elongated and increase in size and number compared to healthy controls (Fontana et al., 1998; Tezel et al., 2004). All these factors can distort the laminar canals through which the RGC axons pass and may disturb the retrograde intracellular transport in the axons and even directly crush the RGC axons (Knox et al., 2007).
Another histopathological feature of glaucomatous LC is the (re)activation of astrocytes and microglia (Knox et al., 2007; Neufeld 1999; Varela and Hernandez, 1997). The role of (re)activated astrocytes and microglia in glaucoma are extensively reviewed by Hernandez (2000) and Morgan (2000), respectively. Microglial cells are normally quiescent in the LC; they seem to be strategically positioned close to blood vessels and the blood-retinal barrier and are activated under the influence of IOP or neuronal injury. In glaucomatous eyes, activated microglia are found as clusters of large amoeboid cells in the compressed LC and around blood vessels. Reactive astrocytes disturb the ECM of the LC which will ultimately result in the mechanical changes of the LC.
3.7. Intracranial pressure dynamics
Although an increased IOP is a major risk factor for POAG, the response to a given IOP level shows a large variability with NTG (RGC damage without elevated IOP) (Perkins et al., 1959; Pickard 1931) and ocular hypertension (OHT; elevated IOP without RGC damage) (Hollows and Graham, 1966) as the most extreme phenotypes. Obviously, variability in susceptibility of RGCs plays a major role here, but, crucially, the IOP is only a proxy of the actual force that acts upon the ONH. This force depends on the difference between the IOP and the retrolaminar tissue pressure. The retrolaminar tissue pressure is essentially determined by the intracranial pressure (ICP) (Morgan et al., 1998). The balance between the IOP and the ICP determines the net translaminar pressure on the ONH (Figure 4: translaminar pressure). In the translaminar pressure hypothesis, NTG results from a normal IOP and a decreased ICP, yielding an increased posterior directed translaminar pressure, whereas OHT results from an elevated IOP and an elevated ICP, yielding a normal translaminar pressure. Support for this translaminar pressure hypothesis came from both clinical and experimental studies. Several clinical studies showed that patients with NTG (and sometimes POAG) had significantly lower cerebrospinal fluid (CSF) pressures compared to controls (Berdahl et al., 2008a, 2008b; Ren et al., 2010, 2011a; Wang et al., 2012a). On the other hand, patients with OHT often had increased ICPs (Berdahl et al., 2008b; Ren et al., 2011a, 2011b) (see also Perspectives and Outlook, Section 8). Experimentally, Yablonski et al. (1979) artificially lowered the ICP in cat brains, which resulted in glaucomatous optic neuropathy, whereas lowering both ICP and IOP in the same animals, had no effect on the optic nerve head.
In summary, it seems that the translaminar pressure ultimately determines the state of the ONH and the potential damage to the RGCs. Therefore, we must not only examine and review relevant ocular tissues and media but also relevant brain tissues and media. The tissues and media that determine the ICP-dynamics are the choroid plexus epithelium (CPE), cerebrospinal fluid (CSF) and arachnoid villi (AV). All are briefly described below. Choroid plexus epithelium
The CPE is localized in the lateral, third and forth ventricles of the brain (Figure 4). It consists of a single cell layer of neuroectodermal origin and forms a tight junction barrier between the systemic circulation and the CSF (Figure 5). The major function of the CPE is the production of CSF. Many recent reviews exist on the mechanisms of CSF production by the CPE (Brown et al., 2004; Praetorius, 2007; Redzic and Segal, 2004; Speake et al., 2001; Wolburg and Paulus, 2010). In summary, CSF production is an active and regulated
process with specific ion transporters, including Na+-K+ ATPase, K+ channels, Na+/2Cl-/
K+ cotransporters, Cl-/HCO3- exchangers, Na+/HCO3- and K+/Cl- cotransporters. Water is
1
Arachnoid villus Choroid plexus Cerebrospinal fluid Translaminar pressure ICP IOP ICP IOP ONHFigure 4: Determinants of the translaminar pressure
Schematic cross-section of brain and eye that illustrates the determinants of the translaminar pressure. In the brain, the pink/brown color denotes the neural tissue, blue indicates cerebrospinal fluid (CSF), and red the choroid plexus that produces the CSF. In the bottom right corner of the figure, we zoom in on the area of the optic nerve head (ONH) in order to illustrate the two pressures that act on this area: the posterior directed intraocular pressure (IOP) and the anteriorly directed intracranial pressure (ICP). The difference between the IOP and ICP constitutes the net translaminar pressure. The IOP is determined by aqueous humor production (ciliary body) and outflow (trabecular meshwork) (not shown here, see figure 2). The ICP is determined by CSF production (choroid plexus) and outflow (arachnoid villi) in the brain.
* LV LV Stroma CPE CPE Stroma Figure 5: Histology of the choroid plexus
Human choroid plexus (CP) tissue from the lateral ventricle (LV) of the brain. The micrograph of the cryo-section (Periodic Acid Staining) shows CP epithelium (CPE), stroma, and a vascular lumen of a blood vessel (*)
The major water channel in the CPE is AQP1 (MacAulay and Zeuthen, 2010). The CPE also transports glucose, amino acids, thyroid hormone, amyloid beta and heavy metals, such as iron (Crossgrove et al., 2005). Besides its transport function, the CPE produces molecules which are excreted in the CSF, such as growth factors (e.g. insulin-like growth factor II, transforming growth factor β1) and hormones (e.g. vasopressin, transthyretin, transferrin, melatonin, insulin and leptin) (Redzic and Segal, 2004; Zappaterra and Lehtinen, 2012). The CPE cells shrink with aging and there is an increased amount of psammoma bodies in the underlying vascular stroma (Babik, 2007). Psammoma bodies are round, laminated, onion-like calcified structures, with positive staining for collagen and T-cell markers (Jovanovic et al., 2007). Pascale et al. (2011) observed decreased concentration of amyloid-beta in the aging CPE, decreased expression of LRP-2, an influx transporter, as well as an increased expression of amyloid-beta efflux transporters LRP-1 and P-gp. Using a proteomics approach, Chen et al. (2012a) found decreased expression of, among others, transthyretin (pre-albumin), gelsolin, and alpha-1-antitrypsin, in aged CPE. All these changes may play a role in changes in CSF production, composition and/or turnover rate, observed in aged subjects compared to controls (Johanson et al., 2008).
There are no data available on the possible changes of the CPE in POAG. Cerebrospinal fluid – composition and drainage
The CSF fills up the brain ventricles, spinal canal and subarachnoid space with a total volume of ~140 ml in human adults. The CSF serves as a metabolic, nutritional, immunological and signaling fluid for the central nervous system and it protects the brain because it acts as a kind of hydraulic cushion. Zhang et al. (2005) studied the proteins in the human CSF and found that the total protein concentrations of CSF vary from 300 to 1200 µg/ml, with the largest quantities of protein being albumin and immunoglobulins (>50% and >15%, respectively).
The outflow of the CSF occurs via the arachnoid villi (Kida et al., 1988; Yamashima et al., 1986) and lymphatic pathways, one of which is situated behind the optic nerve head just outside the eye (Kapoor et al., 2008; Pollay, 2010). The arachnoid villi are situated along the superior sagittal sinus along the upper centre of the head as small protrusions of the arachnoid through the dura mater into the venous sinuses of the brain (Figure 4).
With age, production (May et al., 1990) and outflow of CSF decline (Albeck et al., 1998; Knepper et al., 1983)). The net ICP tends to usually remain unchanged. Moreover, neural brain volume decreases with age, probably due to neurodegeneration. Consequently, the total CSF volume in the brain increases with age, while the turnover of CSF decreases. Proteomic studies of CSF of old subjects revealed distinct protein changes with overall increased protein content compared to CSF of young subjects (Baird et al., 2012; Zhang et al., 2005). This changed composition could also influence outflow facilities of the CSF.
1
There are no data on the possible changes in CSF composition and outflow in POAG, but one might hypothesize that in a subgroup of aging patients, a decreased ICP develops from an imbalance between CSF dynamics and brain ventricle volumes. This subset of subjects could be at greater risk of developing POAG/NTG.
Overall, a comparison can be made, both in function and in morphology, between tissues and media involved in the pressure-dynamics of the eye (IOP) and the brain (ICP), more specifically between CBE-CPE, AH-CSF and TM-arachnoid villi. We hypothesize that, due to inborn errors, gene mutations and/or aging, alterations in the physiology of these systems may disturb normal homeostasis, net pressure over the optic nerve head and, eventually, may contribute to pathogenesis of POAG.
3.8. Other tissues
Other tissues, media or physiological processes in the eye also play a role in the pathobiology of POAG. These include the vitreous, and the uveal scleral outflow.
The vitreous is a highly hydrated (99% water) extracellular matrix filling up the posterior part of the eye. The composition and origin of the vitreous were extensively reviewed elsewhere (Halfter et al, 2008) and are briefly summarized here: The major components of the healthy vitreous are water, specific extracellular matrix components (collagens types II, IX and V/XI, opticin, vitrin, fibulin-1, fibronectin, nidogen-1, hyaluronan, and albumin) and transferrin. The predominant origin of these vitreous proteins is the ciliary body. In POAG, Doganay et al. (2012) and Dreyer et al. (1996) observed increased neurotoxic glutamate levels in the vitreous, although these findings were not confirmed by others (Honkanen et al., 2003; Kwon et al., 2005; Wamsley et al., 2005). The apparently contradictory data may be due to different studies investigating different types or stages of glaucoma.
Besides the conventional AH outflow through the TM, a minor amount of AH can also leave the eye by uveoscleral outflow. This passage of AH through the ciliary muscle into the supraciliary and suprachoroidal spaces was first described by Bill (Bill, 1965). Toris et al. (1999) found a reduced uveoscleral outflow during aging. Interestingly, in glaucomatous dogs Samuelson and Streit (2012) found distinct changes of the uveoscleral outflow, including an accumulation of melanophores and a decrease in smooth muscle bundles. It is not known whether the uveoscleral flow in human POAG-affected eyes is also changed. The uveoscleral route seems to be of special interest for prostaglandin analog-related POAG drugs that reduce IOP by increasing uveoscleral AH outflow (reviewed in Alm, 2000).
4. Model of pathobiological changes and disbalances leading to POAG:
A hypothesis
POAG is a multifactorial disease in which various tissues and mechanisms are involved. Based on the histopathological and molecular changes in the tissues involved in POAG described in the previous section (Section 3) we built a hypothetical model of POAG (Figure 6). It is not known where the sequence of events leading to POAG starts. The disease may even have multiple starting points, each setting in motion a chain of events that, in itself or through interaction downstream, contribute to the susceptibility of the eye to develop POAG. For this discussion, we chose to start with the RGCs.
Increased IOP Posteriorly directed increased translaminar pressure Intrinsic CB changes Intrinsic TM changes Changed AH composition Disturbed LC properties
Thinner CCT Decreased ICP
Intrinsic RGCs changes
Figure 6: Hypothetical model of pathobiological changes and disbalences leading to POAG
Schematic representation of the various pathobiological changes involved in the development of POAG. These changes can be classified into seven major groups, namely (1) intrinsic changes of RGCs, disturbed biomechanical properties of the (2) LC and (3) cornea, (4) intrinsic changes of TM, (5) intrinsic changes of CB affecting (6) AH composition (factors 4-6 can cause an increased IOP), and (7) an increase in translaminar pressure due to a decrease in ICP. On the background of all these pathways, genetic constitution, age and habitual risk factors per patients will protect or further increase the risk of POAG.
Abbreviations: POAG: primary open angle glaucoma; AH: aqueous humor; TM: trabecular meshwork; CB: ciliary body; IOP: intraocular pressure; RGCs: retinal ganglion cells; ICP: intracranial pressure; LC: lamina cribrosa; CCT: central corneal thickness
Intrinsic changes in the RGC itself can cause a higher vulnerability and increased cell death. These intrinsic RGC pathways include mitochondrial dysfunction and oxidative stress, activation of intrinsic and extrinsic apoptotic pathways, vascular hypoperfusion and ischemia, excitotoxicity, glial cell activation and (auto-)immunity. All these factors may contribute to POAG pathology.
1
The axons of RGCs pass through the LC when they leave the eye. Age-related changes in collagen or elastin fiber firmness, or reactivation of astrocytes in the LC might cause structural changes in the LC. These biomechanical changes of the LC itself may contribute to the actual stress on the axons for a given translaminar pressure level and to ganglion cell death and glaucoma.
The fluid pressure(s) on the optic nerve head, and more specifically the LC, are also involved in ischemic or mechanistic damage of the RGC axons. The translaminar pressure is the net difference between the IOP and the ICP. The IOP pushes the ONH in posterior direction; the ICP pushes it in opposite direction. Obviously, both IOP and ICP fluctuate and are determined by multiple parameters, including anatomical properties (angle, CCT, TM), biochemical properties (composition of fluids) and physiological properties (fluid production and removal).
The sieve function of the TM affects the IOP, and POAG-related ECM changes in the TM hamper normal AH outflow which causes an increased IOP.
The CB-POAG-related changes may cause disruptions in AH production and composition. Here, newly present or missing biomolecules in the AH subsequently affect the TM cells and its sieve function. Furthermore, the overall increased AH protein content may further obstruct the TM.
A thinner CCT is a risk factor for POAG which may be due to a cornea less able to withstand IOP and to an overestimation of the IOP.
All these factors in the pathobiological mechanisms of POAG are of course influenced by an individual’s genetic constitution, their age and their personal habitual risk factors. The wide variety of mechanisms helps to understand the many phenotypes of glaucoma, including NTG and ocular hypertension.
5. Genetics of Primary Open Angle Glaucoma
5.1. Familial linkage studies: genetic loci and causative POAG disease genes
POAG is a genetically complex trait, which occurs both sporadically and in families. POAG disease genes segregating in families can be found using linkage analysis. This usually results in the positional identification of a linked chromosomal region and, subsequently, of the causative POAG disease gene. So far, genetic linkage analysis yielded twenty defined chromosomal regions that are linked to POAG (Table 1). In five genomic loci, the POAG disease genes were actually identified using mutational analysis. The best known and studied loci are GLC1A on chromosome region 1q21-q24 containing the MYOC disease gene, the GLC1E locus on region 10p14-p15 with the OPTN gene, and locus GLC1G (region 5q22.1) spanning the gene WDR36. The locus GLC3A on chromosome 2p21 with candidate gene CYP1B1 was initially attributed to congenital glaucoma (Stoilov et al., 1997), but at
least one report implicated this gene also in POAG (Melki et al., 2004). Recently, Pasutto et al. (2012) discovered in the GLC1F locus on region 7q35-q36 the gene ABS10, which is likely to be the disease-causing gene. The other loci that are linked to POAG contain a variety of candidate disease genes, whose identification awaits further research.
Below, we review the molecular characteristics, ocular tissue expression and function of the familial POAG disease genes MYOC, OPTN, WDR36, CYP1B1, and ASB10.
Table 1: Associated loci from linkage analysis
Locus Region Trait Population Genes in region References
GLC1A 1q21-q24 POAG USA MYOC Locus: Sheffield 1993 Gene: Stone 1997 GLC1B 2cen-q13 POAG UK e.g. ADRA2B, PAX8, IL1A,
IL1B Stoilova 1996
GLC1C 3q21-q24 POAG USA e.g. CD10, IL12A, TF, CP Wirtz 1997 GLC1D 8q23 POAG USA e.g. EBAG9, TRHR,
COL14A1 Trifan 1998
GLC1E 10p14-p15 POAG UK OPTN Locus: Sarfarazi 1998 Gene: Rezaie 2002 GLC1F 7q35-q36 POAG USA ASB10 Locus: Wirtz 1999 Gene: Pasutto 2012 GLC1G 5q22.1 POAG USA WDR36 Locus&gene: Monemi 2005 GLC1H 2p15-p16 POAG UK XPO1, OTX1 Suriyapperuma 2007
GLC1I 15q11-q13 POAG USA e.g. PAR5, NPAP1, GABRB3 Allingham 2005; Woodroffe 2006 GLC1J 9q22 POAG USA e.g. FOXE1, ROR2, TGFBR1 Wiggs 2004 GLC1K 20p12 POAG USA e.g. BMP2, RRBP1,
SLC4A11 Wiggs 2004
GLC1L 3p21-22 POAG Australia e.g. VII1, LZTFL1, MAP4 Baird 2005 GLC1M 5q22.1-q32 JOAG China e.g. WDR36, NRG2 Pang 2006 GLC1N 15q22-24 JOAG China e.g. NR2E3, SMAD6, CLN6 Wang 2006a GLC1P 12q14 CODA USA e.g. GDF11, NEUROD4,
WIF1 Fingert 2007b
GLC1Q 4q35.1-q35.2 POAG UK e.g.LRP2BP, UFSP2, CYP4V2 Porter 2011 GLC3A 2p21 PCG+
POAG USA CYP1B1 Locus&gene: Stoilov 1997 2q33-34 POAG Africa e.g. CLK1, HSPE1, CASP8 Nemesure 2003 10p12-13 POAG Africa e.g. MLLT10, NRP1, MSRB2 Nemesure 2003 14q11 POAG Africa e.g. PRKD1, ZNF219, JH4 Nemesure 2003 Abbreviations: POAG=primary open angle glaucoma; JOAG=juvelinel open angle glaucoma; PCG=primary congenital glaucoma; CODA=Cavitary optic disk anomalies; e.g. = example given
Bold genes: POAG genes; italic genes: genes situated on the locus, need further research to identify the candidate POAG disease gene(s)
1
Myocilin (MYOC)
The MYOC gene, initially called TIGR (trabecular meshwork-induced glucocorticoid response gene) consists of three exons and was first discovered by Stone et al. (1997) as the gene that was associated with POAG on the GLC1A locus. MYOC mutations cause POAG in approximately 2-4% of the population (Fingert et al., 1999; Hulsman et al., 2002). So far, more than 180 DNA sequence changes of the MYOC gene have been described (see www.myocilin.com). Approximately 40% of these variants are disease-causing mutations. The overwhelming majority of the mutations found (~85%) appeared to be missense mutations in the third exon (Hewitt et al., 2008). The most common mutation is c.1102C>T (rs74315329) (Fingert et al., 1999).
Before discovery of the association with POAG, the TIGR/MYOC protein was first identified by Polansky et al. (1997) as a TM protein induced by glucocorticoid. Independently, Kubota et al. (1997) found this new protein in the retina and called it MYOC. Subsequently, MYOC was found to be expressed in multiple ocular tissues including the cornea, iris, TM, CB, vitreous, optic nerve and LC (Huang et al., 2000; Karali et al., 2000; Ortego et al., 1997). MYOC mRNA is highly expressed in the CB and TM and moderately in the LC (Supplementary Table S2) and also in other eye tissues, such as sclera, iris and ONH (Swiderski et al., 2000). MYOC is a secreted protein that is involved in extracellular matrix turnover and cytoskeleton function and MYOC can also be expressed within the cell, where it is involved in mitochondrial functions (reviewed in Resch and Fautch, 2009).
POAG-related mutations in MYOC result in misfolded MYOC proteins. Secreted MYOC proteins form abnormal aggregates that clog into the TM (Lütjen-Drecoll et al., 1998). Intracellular misfolded MYOC proteins form abnormal aggregates in the endoplasmatic reticulum (ER) and mitochondria, resulting in increased apoptosis of the TM cells (He et al., 2009; Liu and Vollrath, 2004). Together, the extra- and intracellular aggregation of misfolded MYOC proteins severely impair the sieve function of the TM, with decreased AH outflow and increased IOP as a consequence. To complicate matters, the expression of MYOC was not only high in the TM, but also in other eye tissues, such as cornea, CB, retina, optic nerve and LC. This may implicate that mutations in MYOC not only cause disruptions of the outflow system, but also affect AH production or composition, as well as the tissues that restrain the IOP and even the RGC axons themselves.
Optineurin (OPTN)
The OPTN protein was first discovered by Li et al. (1998) and was initially called FIP2. The gene contains three non-coding exons in the 5-region and 13 exons that code for a 577-amino acid protein (Rezaie et al., 2002). Rezaie et al. (2002) discovered the association between glaucoma and OPTN and found that mutations in OPTN could be responsible for 16.7% of the hereditary forms of NTG and 13.6% of the sporadic glaucoma cases with
predominantly normal IOP. The authors also renamed the FIP2 gene in “optineurin (OPTN)” which stands for “optic neuropathy inducing” protein. Subsequent investigations showed OPTN mutation prevalences between 0-6% in POAG (Aung et al., 2003; Caixeta-Umbelino et al., 2009; Hauser et al., 2006a; Jansson et al., 2005; Leung et al., 2003; Liu et al., 2008; Mukhopadhyay et al., 2005; Sripriya et al., 2006; Tang et al., 2003; Toda et al., 2004; Wiggs et al., 2003). Mutation frequency appeared to depend heavily on the glaucoma subtype (NTG-only or not) and the population investigated.
The two most common OPTN variants are c.148G>A (rs28939688) (prevalence 1-2% in NTG) and c.293T>A (rs11258194) (prevalence < 6-10 % in NTG; 0-4 % in POAG) (Alward et al., 2003; Aung et al., 2003; Melki et al., 2003; Sripriya et al., 2006). Patients with c.148G>A were relatively young at the age of diagnosis and had more advanced optic disc cupping at diagnosis compared to controls (Aung et al., 2005; Hauser et al., 2006a). POAG patients with the c.293T>A variant apparently had a lower initial IOP compared to controls (Melki et al., 2003).
The OPTN gene is highly expressed in TM and LC and moderately expressed in cornea, CB and RGCs (Supplementary Table S2). The protein is expressed in many human eye tissues, including TM, and CB (Rezaie et al., 2002).
OPTN is a Rab-binding, multifunctional protein that might utilize NFκB and TNFα-signaling to mediate mitosis, cellular morphogenesis, immune response, vesicular transport (mediated by, among other factors, Rab8), apoptosis, and protection against oxidative stress (Chalasani et al., 2009; De Marco et al., 2006; Nagabhushana et al., 2010; Rezaie et al., 2002).
Various functional studies were performed to elucidate the role of OPTN variants in POAG. OPTN E50K transfected cells were more vulnerable to oxidative stress and apoptosis and more cell death occurred compared with controls (Chalasani et al., 2007; De Marco et al., 2006). Furthermore, the E50K mutations disregulated vesicle transport pathways resulting in an altered interaction with Rab8 and an impaired uptake of transferrin (Chalasani et al., 2009; Nagabhushana et al., 2010).
In summary, mutations in OPTN can cause oxidative stress-mediated apoptosis in RGCs. At the same time, OPTN mutations may influence AH production, composition and outflow through, for example, disturbed vesicle mediated transport.
WD repeat domain 36 (WDR36)
WDR36 consists of 23 exons, which encode for 951 amino acids and a protein with multiple G-beta WD40 repeats. Initially, Monemi et al. (2005) identified four different mutations in WDR36 (c.1064A>G (rs118204022); c.1345G>A (rs35703638); c.1586G>A (rs116529882) and c.1973A>G (rs34595252)) that were implicated in around 6% of unrelated POAG patients and in none of the 200 healthy controls. Several replication studies confirmed
1
the linkage or association of WDR36 variants with POAG, on average in 3.7% of the POAG patients (Blanco-Marchite et al., 2011; Fan et al., 2009; Frezzotti et al., 2011; Lee et al., 2010; Mookherjee et al., 2011; Pasutto et al., 2008; Weisschuh et al., 2007). Other population based studies, however, could not replicate the link between WDR36 mutations and POAG (Fingert et al., 2007a; Hauser et al., 2006b; Hewitt et al., 2006; Kramer et al., 2006; Pang et al., 2006). Together, these findings suggest that WDR36 may be a minor disease causing gene or a modifier gene for POAG (Hauser et al., 2006b).
WDR36 is expressed in the lens, iris, sclera, CB, TM, retina and optic nerve (Monemi et al., 2005). WDR36 is a scaffold protein and a member of the large family of WD40 repeat-containing proteins, which are involved in many cellular processes, including cell cycle, signal transduction, apoptosis and gene regulation. Specifically, WDR36 is involved in G-protein-coupled receptor signaling (Cartier et al., 2011) and preRNA processing (Gallenberger et al., 2011). In a mouse model with Wdr36 mutations, there was an increased loss of RGCs, a reduction in peripheral retina thickness and a normal IOP (Chi et al., 2010a). All together, WDR36 mutations may be involved in RGC development and viability and/or connective tissue quality/quantity of the eye.
Cytochrome P450 family 1, subtype B, polypeptide 1 (CYP1B1)
CYP1B1 consists of three exons and two coding regions (Stoilov et al., 1997). Vincent et al. (2002) reported possible digenic mutations in MYOC and CYP1B1 in POAG families. Family members with a MYOC mutation alone had an age of onset of around 51 years, whereas carriers of both the MYOC and CYP1B1 mutations had an average age at onset of 27 years. Therefore, they concluded that CYP1B1 may be a modifier gene of MYOC (Vincent et al., 2002). Subsequently, Melki et al. (2004) investigated the association of CYP1B1 variants and POAG in a French population. They found 11 mutation carriers among 236 patients (4.6%) as opposed to one single carrier in 197 controls. Multiple different population studies replicated this association between CYP1B1 variants and POAG (Acharya et al., 2006, 2008; Bhattacharjee et al., 2008; Chakrabarti et al., 2007; Chen et al., 2011; Kumar et al., 2007; Lopez-Garrido et al., 2006; Pasutto et al., 2010; Patel et al., 2012), although a few did not (Fan et al., 2010; Wolf et al., 2009).
During development, CYP1B1 protein was found to be highly expressed in human cornea, CB, iris, retina and RPE, but not in the TM (Doshi et al., 2006). In adults, CYP1B1 is expressed in CB and iris (Doshi et al., 2006). The gene is highly expressed in adult human TM, CB and LC (Supplementary Table S2). CYP1B1 is involved in many reactions including synthesis of lipids and drug metabolism. Mutations in CYP1B1 are also associated with Peter’s anomaly, a congenital defect of the anterior chamber of the eye (Vincent et al., 2001). Mutations in CYP1B1 cause reduction in enzymatic activity and decreased protein stability (Lopez-Garrido et al., 2010). Kennedy et al. (2012) classified CYP1B1 among the 50 physiological
biomarkers of glaucoma, after they found that the K423E mutation in MYOC caused a significant downregulation of CYP1B1 expression in cultured human TM cells. In summary, CYP1B1 mutations may cause developmental abnormalities of the anterior ocular angle, or may invoke changes in TM and LC (possibly via CB), resulting in disturbed AH dynamics and increased IOP, and perhaps, in parallel, less resistance of the cornea and LC against the IOP. Ankyrin repeat and SOCS box containing 10 (ASB10)
ASB10 consists of six exons and three isoforms of the gene are known. In 2012, Pasutto et al. (2012) DNA sequenced ASB10 in a large family affected by POAG, and found that exon 3 was absent. Two subsequent case-control association studies (USA and Germany) revealed 26 amino acid changes in 70 patients (6%) and 9 amino acid changes in 13 control subjects (2.8%), which showed a significantly larger presence of variants in ASB10 in POAG patients compared to controls (Pasutto et al., 2012). Another case-control association study of Fingert et al. (2012) in an Iowa (USA) population, could not replicate the association of ASB10 and POAG.
ASB10 mRNA expression is high in the human iris, moderate in LC and optic nerve and low in TM, CB, retina and choroid (Pasutto et al., 2012; Supplementary Table S2). The protein was found in human TM cells (outer TM beams, juxtacanalicular region and Schlemm’s canal), and the pigmented epithelium of the CB, as well as the RGCs and inner nuclear layer (Pasutto et al., 2012).
ASB10 is a member of the ankyrin repeat and SOCS box-containing (ASB) family. Members of this family serve as suppressors of cytokine signaling and might induce degradation of specific proteins.
Pasutto et al. (2012) also studied the effect of ASB10 silencing on the outflow facility of the TM and found in an in vitro whole anterior segment model, that the outflow decreased after knockdown of the gene. This suggested that ASB10 mutations can cause IOP disregulations in POAG. However, since the protein is also expressed in the RGCs, mutations might also cause cellular disregulations there, with increased apoptosis and cell death as a consequence.
So far, the POAG disease genes discovered with familial linkage analysis can only explain up to 10% of all POAG cases.
5.2. Genetic association studies: genetic loci and genetic risk factors for POAG
Since POAG is a genetically complex disease, population-based studies are essential to supplement genetic information available from family studies. Population studies tend to yield POAG associated genetic risk factors and not directly POAG-linked causative disease genes (identified in family studies; see above). Obviously, it is sometimes possible for
1
different mutations in the same disease gene to be both POAG-associated in populations and POAG-causatively linked in families.
Two types of genetic association studies are frequently used to find gene variations associated with POAG: candidate gene polymorphism studies and genome wide association studies (GWAS). Recently, an endophenotype approach of GWAS was successfully used in various studies: a quantitative endophenotype trait of POAG, such as central corneal thickness (CCT), intraocular pressure (IOP) or vertical cup-disc ratio (VCDR) was used to find gene variations associated with these endophenotypes. Subsequently, the resulting associated genomic loci were often also tested for potential association with the full POAG phenotype. The results of the association and replication studies are not always consistent and sometimes conflicting. This, however, may be due to variations in sample size, diagnosis criteria and study designs, mistypings, different ethnicity, genetic heterogeneity between populations or coincidence (Baas et al., 2010, 2012).
Selection of candidate POAG disease genes
We searched the literature for possible candidate genes that are associated with POAG (search lasted until 31st December 2012), analyzed over 120 POAG association studies (both candidate gene association and GWA studies) and identified 66 loci with 76 (candidate) genes. From these 76 genes, we selected those genes that either showed a positive association in at least three independent candidate gene selection studies or yielded positive association in at least one GWAS discovery cohort and at least one independent replication cohort. This selection resulted in 29 (candidate) POAG loci containing 33 (most likely candidate) POAG genes. An overview of the literature study on these loci and 33 genes is outlined in Table 2. Together with the five familial POAG disease genes, we thus recognized 38 highly likely POAG genes (33+5). There were also genomic loci with candidate POAG genes that did not met these criteria and need further investigation. We selected these loci (in total 21) that contained 27 genes and we named these genes the less likely candidate POAG disease genes (Supplementary Table S3). We assumed that this approach would cover perhaps not all, but certainly the majority of disease genes from these (genome wide) association studies. The total numbers of loci and genes and the different subgroups (familial, highly likely associated and less likely associated) are presented in Table 3.
Table 2: Highly likely candidate POAG genes from association studies
Locus Gene morphismPoly- P O V C I Population Study type Study size(n total) Reference 15q24-25 AKAP13 rs6496932 + D: CR, SCL
R: CR, SCL GWAS D: 1445R: 824 Vitart 2010 rs6496932
rs1828481 + D: AsianR: Asian GWAS D: 7711R: 2681 Cornes 2012 rs6496932 - D: UK
R: UK GWAS D: 437R: 344 Gibson 2012 rs6496932
rs1828481 - - SNP: US-CauD: US-Cau GWASSNP SNP: 6469D: 1117 Ulmer 2012 19q13.2 APOE rs429358 rs7412 + + Japan SNP 489 Mabuchi 2005 rs429358 rs7412 + Japan SNP 681 Fan 2005 rs429358 rs7412 + China SNP 700 Lam 2006 rs429358
rs7412 + Saudi Arabia SNP 190 Al-Dabbagh 2009 rs429358 rs7412 + AU, NZ SNP 193 Vickers 2002 rs429358 rs7412 - UK SNP 212 Ressiniotis 2004 rs429358 rs7412 - Turkey SNP 194 Saglar 2009 rs429358 rs7412 - Estonia SNP 429 Zetterberg 2007 rs429358 rs7412 - UK SNP 504 Lake 2004 10q21.3-22.1 ATOH7 rs1900004 + + D: NL R: NL, UK GWAS D: 7360R: 4455 Ramdas 2010 rs3858145 + D: AU R: UK GWAS D: 1368R: 848 Macgregor 2010 rs1900004 + D: NL GWAS D: 23000 Axenovich 2011 rs7916697 + D: Asian R: NL GWAS D: 4445R: 9326 Khor 2011 rs3858145 - D: UK R: UK GWAS D: 437R: 344 Gibson 2012 rs7916697 + African* SNP 437 Cao 2012 rs1900004 + EU SNP 45998 Ramdas 2011a rs1900004 + Japan SNP 616 Mabuchi 2012 rs1900004 rs3858145 - African* SNP 437 Cao 2012 rs7916697 rs1900004 - China SNP 431 Chen 2012b rs7916697 - US-Cau SNP 875 Fan 2011 rs1900004 + US-Cau SNP 875 Fan 2011 rs1900004 rs3858145 - AU, NZ SNP 1759 Dimasi 2012 rs61854782 + China SNP 431 Chen 2012b rs3858145 - US-Cau SNP 875 Fan 2011
