In this chapter we review the impact of functional genomics on the understanding of the function of tissues which play a role in the development of the disease as well as on the elucidation of their response to glaucoma associated agents. We will briefly review the number of microarray studies on all tissues involved, and extend and focus on the genes and function of the trabecular meshwork, field of our expertise.

III.FUNCTIONAL GENOMICS: MICROARRAYS, PROTEOMICS AND PROTEIN MODIFICATION

When encountered with a given disease or insult, the cell of a certain tissue reacts by regulating the expression of its genes. This regulation occurs at several diVerent levels, which overall can be reduced to three: transcription of the gene’s DNA into mRNA, translation of the mRNA into protein, and modification of the protein molecule by a number of chemical reactions which would determine its folding, targeting for degradation and all its biochemical and physiological properties. The end point of all these levels of regulation is function. While transcription and translation of a gene could be normal, a hostile cell environment could promote generation of free radicals or favor deamination of an amino acid residue at a key location and consequently alter its normal physiological role. Analysis of gene regulation at each of the three steps is necessary to elucidate the disease mechanism and to be able to design rational drugs. The technology to investigate each of these regulatory levels has advanced considerably and is providing us with a wealth of information. If anything, we are now at the crossroads of trying to figure out the relationship between the observed changes on known genes, the new mechanisms revealed by the presence of genes previously unknown to the tissue, and what does it all mean for the overall physiology.

A. Genechips/Microarrays

The GeneChip/microarray system allows us to investigate the mRNA population of any cell (transcriptome) and analyze transcriptional diVerences between diVerent conditions at the level of the entire genome. GeneChip microarrays determine whether a gene is transcribed or not, and whether the entire transcriptome is modulated or altered by the condition (s) under study. Most commonly Genechip/microarrays are used to analyze diVerences between a control and experimental condition, however, they are increasingly being used to compare changes across multiple experimental conditions/ temporal changes. This GeneChip system consists of a collection of small

oligonucleotide DNA fragments which have been embedded in a glass chip by optical lithography technology. The oligonucleotide sequences correspond to a collection of genes extracted from the human or animal’s specific databases. In the chip, each gene of the database is represented by a number of diVerent spots (probe set) and genes corresponding to internal controls are also included. The sequences of all spots in the chip are selected in such a way that they will undergo hybridization under similar annealing conditions to the whole population of a cell’s RNA. Total RNAs extracted from a cell or tissue, aVected or not by a given condition, are fluorescently labeled and hybridized (usually at Universities’ core facilities) to a set of identical chips. Sophisticated software analysis packages allow determination of the degree of diVerential expression of each of the genes between both conditions. Although most investigators tend to confirm the data of a few genes of interest using real time PCR, the continuous improving quality of the chips together with new and upgraded versions of analysis programs may soon preclude that step.

A number of commercial entities are oVering GeneChips. These chips contain from thousands of genes representing close to the entire human genome, to a few hundred genes which are linked to a particular function or characteristic (pathway focused microarrays). Mechanisms such as osteogenesis, signal transduction, or extracellular matrix can be independently studied. Chips with DNA representing the whole genome from most common experimental organisms such as rat, mouse or drosophila are also available. The pathway or disease focused pathways arrays contain less than 400 genes and use nylon membranes rather than glass slides. The membranes have a high DNA capacity and are produced by binding specifically designed 60 mer oligonucleotides using a non contact printing technology that prevents membrane deformations. Amounts, as small as 100 ng of total RNA are biotin labeled, hybridized to the membrane, treated with a conjugated streptavidin molecule (usually alkaline phosphatase) and developed using chemiluminescence substrates. Special software analyzes the diVerent spot intensities of the membrane. In addition to being focused, this technology is less costly and can be performed with standard laboratory equipment. A later, upgraded procedure involves the use of microarray 96 well plates containing in each well SYBR Green labeled primer sets for the relevant genes of the pathway disease (http://www.superarray.com/PCRArrayPlate. php). Addition of the reverse transcriptase reactions from the treated and untreated samples to two identical plates can then be run in a standard real time PCR instrument, rather than in a special core facility. In addition, this latest procedure allows the use of RNA amounts as small as 5 nanograms, an important consideration when dealing with individual small tissue samples.

B. Proteomics and Protein Arrays

Proteomics refers to the identification of the protein population (proteome) of a cell at a given time and/or under a sought after condition. There are a few strategies to identify a protein population. The original, conventional procedure used two dimensional (2D) gel electrophoresis. This procedure involved the separation of proteins by two of their main characteristics, molecular weight and isoelectric point (pI). The separation of these properties (mass and charge) is conducted by gel electrophoresis in perpendicular directions. Proper staining at the end of the runs results in a number of spots spread along the surface of the gel. Because it is quite unlikely that two proteins would have two identical characteristics, the 2D separation between proteins is greatly improved from the one direction gel method. Proteins are subsequently identified by extracting them from the gel spots and subject them to peptide cleaving and sequencing by mass spectrometry. The cataloging of the proteins at any of the cell stages is of utmost importance to understand molecular mechanisms of a disease.

A later technology uses protein arrays. In a similar manner than the arrays used for nucleic acids, proteins are immobilized on solid surfaces that include glass, membranes, microtiter wells, mass spectrometer plates or beads. While the nucleic acid arrays are based on molecular hybridization and PCR, protein arrays are based on protein–protein interactions, which could be protein–antibody, protein–protein, protein–ligand or protein–drug and enzyme–substrate bindings. These arrays are however more complicated than those containing nucleic acids, especially because the complexity of the human proteome far exceeds that of the genome. Taken into account alternative gene splicing and post translational modifications, the number of diVerent protein molecules species in humans could be at least one order of magnitude higher than the number of genes, that is about 500,000 (http:// www.functionalgenomics.org.uk/sections/resources/protein_arrays.htm# research). The feasibility and extent of protein expression profiles will then depend on the number of capture reagents available, such as antibodies, for the proteins of interest.

C. Protein Modifications

In addition to alternative splicing, where a single gene can give rise to proteins containing diVerent domains, an extensive number of post translational modifications are responsible for the diversity of the human proteome. These post translational modifications are often key for maintenance of physiological conditions and/or development of diseases (Krueger

and Srivastava, 2006). Some of the post translational modifications associated with signaling and diseases include phosphorylation, glycosylation, ubiquitinization, prenylation, oxidation, and citrullination. To date, technology to detect protein modifications is limited to study one modification at a time. For example, in what is called phosphoproteomics, a number of antibodies developed to specific phosphorylated sites are incorporated into protein arrays. These arrays become very useful in cases where there is a knowledge of an altered particular pathway which include phosphorylated proteins associated with a given disease such as cancer, (Sheehan et al., 2005). Another example, albeit less targeted, is the use of a glycomic profiling approach by isolating the total population of N linked oligosaccharides proteins and comparing the glycan profiles between the serum of patients having the disease with normal controls (Moniaux et al., 2004). Still another is the use of an activity base protein profile (ABPP) which measures the activity of a particular class of enzymes. This approach uses active site directed probes which detect the functional state of the enzymes (Liu et al., 1999; Jessani and Cravatt, 2004; Speers and Cravatt, 2004). Binding of the probes to the arrays will label and identify active enzymes, but not their inactive precursor (Kidd et al., 2001; Saghatelian et al., 2004), or their inhibitor bound forms (Greenbaum et al., 2002; Jessani et al., 2002; Saghatelian et al., 2004). Thus these arrays will serve to detect disease associated changes which may occur in the absence of their transcriptional and/or translational abundance (Joyce et al., 2004; Jessani et al., 2005). A representative case could be seen in the study by Sieber et al. (2006) who created a library of chemical probes directed to metalloprotease’s activities in a biological system. Such a library of metalloprotease directed labeled probes identified the diVerent activity of more than twenty metalloproteases in human cancer cell lines from invasive and noninvasive carcinoma (Sieber et al., 2006). Innovative approaches similar to the one described to detect active sites of family of proteins are expected to be developed in the coming years.

IV. TISSUES INVOLVED IN THE DEVELOPMENT OF GLAUCOMA. SURVEY OF MICROARRAY STUDIES

Glaucoma is a complex disease and, as mentioned above, development of this optic neuropathy can occur as a result of the physiological dysfunction of more than one tissue. The final outcome of the disease, vision loss, could be triggered by a pathological secretion of aqueous humor (ciliary body), an altered resistance to aqueous humor outflow (trabecular meshwork), a pressure independent insult for the RGC or, by a loss of support to the optic nerve (ONH astrocytes).

The past few years have witnessed a number of functional genomics studies comparing either normal cells tissue versus glaucoma or, normal cell tissues versus treatments with glaucomatous associated insults. Most approaches have involved determination of diVerences at the transcriptional level. They have used high density oligonucleotides microarrays to find whether the global expression of thousands of genes has changed. In contrast, very few proteomic analysis and no post translational modifications or activity profiles are yet available for any of the glaucoma tissues.

A. The Ciliary Body

The ciliary body comprises several structures which play an important role in the physiology of the eye. Two of them are relevant to the regulation of aqueous humor outflow. The first is the ciliary muscle which runs longitudinally and is directly inserted into the trabecular meshwork. Contraction of this muscle pulls on the trabecular meshwork aVecting its shape and as a consequence, influencing the free passing of the aqueous humor through the tissue. The ciliary muscle is also the site of the uveoscleral outflow which constitutes the alternative route for the exit of aqueous humor of the eye and an important drainage pathway in some animal species (Bill, 1965; Bill and Hellsing, 1965). The second structure of the ciliary body is the ciliary epithelium which is responsible for the secretion of aqueous humor. This epithelium consists of two well diVerentiated layers, pigmented and nonpigmented, which have the same embryological origin than the pigmented epithelium and neuroretina layers of the retina.

There are numerous studies addressing individual gene expression, both directly on the microdissected ciliary processes tissue and on transformed cell lines originated from the two layers of the ciliary epithelium. These cell lines were carefully established by the separation of the pigmented and nonpigmented layers and have proven to be an excellent source for investigating potential new signaling mechanisms of this tissue (Coca Prados and Escribano, 2007). However, to date, there are very few microarray studies available. A review of the current literature found two microarray reports, one carried out on whole human ciliary body tissue (dissected together with the iris and compared to other eye compartments) (Diehn et al., 2005) and the second one conducted on human ciliary muscle cells exposed to analogs of glaucoma drugs (Zhao et al., 2003).

B. Trabecular Meshwork

The trabecular meshwork is an avascular tissue whose main function is the maintenance of IOP. It is located at the angle formed by the iris and the cornea and it exhibits a distinctive architecture. There are three

morphological and functionally diVerent regions in the trabecular meshwork which are lined by a monolayer of cells at the contact with the Schlemm’s canal. The uveal region, closer to the anterior chamber, is followed by the corneoscleral region, where cells are attached to trabecular beams, and by the juxtacanalicular region, whose cells lay relatively free but can connect to each other and to the cells lining the Schlemm’s canal. All trabecular meshwork cells are embedded in an extracellular matrix (ECM) of distinct properties, which gives the tissue a unique spongiform three dimensional architecture. The cells of the trabecular meshwork derive from the mesenchymal cells of the neural crest (Johnston et al., 1979; Matsuo et al., 1993; Baulmann et al., 2002; Cvekl and Tamm, 2004) while those lining the Schlemm’s canal are of vascular origin (Hamanaka et al., 1992; Krohn, 1999). In humans, the trabecular meshwork is the main route for the outflow of the aqueous humor (Bill and Phillips, 1971).

To study functional genomics of the trabecular meshwork scientists have relied mainly on cultures of primary cells and perfused organ cultures. A few studies from in vivo animals are also available. The human cell cultures are generated from the dissected trabecular meshwork of post mortem donors (Polansky et al., 1984; Polansky and Alvarado, 1994; Stamer et al., 1995; Vittitow et al., 2002). More recently, the tissue is obtained from discarded surgical rims left over after a cornea transplant, which provide an excellent source of healthy cells (Rhee et al., 2003). These primary cultures contain cells from both trabecular meshwork and inner wall of the Schlemm’s canal. A few viral transformed cell lines from single normal and glaucomatous individuals have also been generated (Pang et al., 1994). Although these cells lack some of the intrinsic characteristics of the primary cultures, such as myocilin induction by Dexamethasone (DEX), they are very useful for a number of diVerent purposes.

The perfused organ cultures, originally developed for physiological measurements (Johnson and Tschumper, 1987; Johnson, 1997), have been adapted by our laboratory for functional genomics studies (Borra´s et al., 1998, 2002; Gonzalez et al., 2000a,b,c; Vittitow and Borra´s, 2004; Comes et al., 2006). The culture procedure involves perfusion of human anterior segments from paired eyes of a single donor. The whole globes procured by the eye banks are dissected, set in culture by 24–40 hour post mortem and perfused with serum free medium at the physiological aqueous humor flow rate. This method revives the trabecular meshwork tissue and makes it amenable to gene expression studies. Because these cultures conserve the original architecture of the trabecular meshwork, they provide an opportunity for studying gene’s responses to mechanical strain under conditions closer to those occurring in vivo. A unique characteristic of the outflow pathway is the existence of a pressure decrease across the inner wall of the Schlemm’s canal, that is, the pressure in the trabecular meshwork proximal to the canal is higher

than the pressure in the canal’s lumen (Johnstone and Grant, 1973; Grierson and Lee, 1975; Johnstone, 1979; Epstein, 1997). The mechanical forces exerted by the flow of aqueous humor in vivo are not present in traditional cell cultures, where the bottom of the dish precludes the presence of the pressure drop. In addition, these cultures address the individual variability concern present in all comparative human studies. Since each experiment involves the use of eye pairs from the same individual, gene expression in the treated eye can be directly compared with that of its untreated contralateral eye and thus not be confounded by genetic diVerences. An increasing number of studies using both cells and perfused organ cultures are appearing in the literature and providing a first glance to relevant genes and mechanisms of the trabecular meshwork.

C. The Retinal Ganglion Cells

Retinal ganglion cells are the only retinal neurons that project their axons (through the optic nerve) to the brain. Death of the RGCs occurs mostly by apoptosis (Garcia Valenzuela et al., 1995; McKinnon, 2003) and it is the major hallmark of loss of vision in glaucoma. Elevated IOP plays a key role in the death of the RGCs, most likely by disrupting the axonal flow which results from the damage exerted on the optic nerve by mechanical compression forces (Pease et al., 2000; Osborne et al., 2001; Whitmore et al., 2005). However, the apoptotic death of the RGCs and degeneration of their axons can occur via diVerent mechanisms including deprivation of neurotrophic factors (Finn et al., 2000) and immune state of individuals (Tezel and Wax, 2004). A better knowledge of functional genomics of the RGC under normal and glaucoma conditions has the potential for providing a unique understanding on the genes and mechanisms leading to their death.

To study gene expression profiles and global diVerential expression of RGC under glaucomatous insults, scientists have conducted microarray studies using a variety of starting material. Because the RGC layer comprises only about 1% of the cells of the entire retina, a first gene sequencing study utilized rat RGC cDNA purified by immunopanning with Thy1 antibodies (Farkas et al., 2004). One study utilized the whole retina and compared an ischemia insult obtained by 1 h of elevated pressure in living rats (Yoshimura et al., 2003) while other used cynomolgus monkeys and laser induced glaucoma (Miyahara et al., 2003). Another compared the rat glaucoma model induced by unilateral saline injection in the rat episcleral veins (Ahmed et al., 2004a). Diehn et al. (2005) separated the macula from the peripheral retina and compared their expression profiles with those of the diVerent tissues of the human eye (Diehn et al., 2005). Steele et al. (2006) used the mouse

glaucoma model DBA/2J mouse and compared retinas between periods of 3 and 8 months (normal and elevated IOP) (Steele et al., 2006). Piri et al. (2006) conducted the study after inducing RGC degeneration by optic nerve transection (ONT) (Piri et al., 2006). Levkovitch Verbin et al. (2006) also compared rats retinas after ONT but used instead a more focused gene array containing 18 signal transduction pathways (Levkovitch Verbin et al., 2006). Ivanov et al. (2006) compared expression of immunopanned isolated rat RGC with that isolated from the whole retina (Ivanov et al., 2006). Naskar and Thanos (2006) compared the whole retina of the Royal College of Surgeons (RCS) rat model, which spontaneously develops elevated IOP, with aged matched controls (Naskar and Thanos, 2006). In a human study, Kim et al. (2006) used laser capture microdissection to compare the retinal ganglion cell layer with other inner and outer layers of the retina (Kim et al., 2006). At the time of this writing, the last report included a comparison of a transformed RGC 5 cell line under serum free conditions to induce apoptosis (Khalyfa et al., 2007). The summary of these studies is presented as part of Fig. 1.

D. Lamina Cribrosa: The Optic Nerve Supporting Tissue

There are two glial cell types characterized in the lamina cribrosa, the ONH astrocyte and the lamina cribrosa (LC) cells (Hernandez et al., 1988). Astrocytes are the non neuronal brain cells which provide nutrition and structural support to other cells of the central nervous system (CNS). They constitute the most abundant cell type of the CNS and express glial fibrillary acidic protein (GFAP), a marker that distinguishes them from the neurons. The LC cells do not express GFAP. At the nonmyelinated optic nerve head, where RGC axons leave the ocular globe, astrocytes provide support to the axons and physically separate them from the surrounding capillary bed. Astrocytes form an integral part of the lamina cribrosa, a sieve like structure in the posterior part of the sclera that allows the passage of the RGC axons and central retinal vessels leaving the ocular globe. The lamina cribrosa helps maintain the pressure gradient between the intraocular and extraocular space. Astrocytes contribute to the maintenance of the extracellular matrix of the lamina cribrosa as well as to the ion balance and extracellular pH of the intercellular space.

Under stress and injury conditions, quiescent normal astrocytes become ‘‘reactive’’ and stop supporting axon survival. This reactive hallmark is accompanied by the altered expression of a number of genes which influence among others ECM organization and secretion of cytokines and survival growth factors (Yang et al., 2004). Reactive astrocytes have been shown to play major roles in the pathogenesis of neurodegenerative diseases, including