Ординатура / Офтальмология / Английские материалы / Ocular Disease Mechanisms and Management_Levin, Albert_2010

Table 19.1  Ocular diseases treated by glucocorticoids

Blepharitis

Conjunctivitis

Keratitis

Scleritis

Uveitis (anterior and posterior)

Macular edema

Choroidal neovascularization associated with age-related macular degeneration

Optic neuritis

Endophthalmitis

Table 19.2  Glucocorticoids used for ocular therapy

Prednisolone acetate (topical ocular 0.125% and 1% suspensions)

Prednisolone sodium phosphate (topical ocular 0.125% and 1% solutions)

Dexamethasone (topical ocular 0.1% suspension)

Dexamethasone (intravitreal implant 350 and 700 mg)

Dexamethasone sodium phosphate (topical ocular 0.1% solution; 0.05% ointment)

Loteprednol etabonate (topical ocular 0.2% and 0.5% suspensions)

Rimexolone (topical ocular 1% suspension)

Fluorometholone (topical ocular 0.1% and 0.25% suspensions; 0.1% ointment)

Fluorometholone acetate (topical ocular 0.1% suspension)

Medrysone (topical ocular 1% suspension)

Triamcinolone acetonide (10 mg/ml and 40 mg/ml injectable)*

Fluocinolone acetonide (intravitreal implant 0.59 mg)

*Used off-label; currently not approved for ophthalmic use.

Box 19.1  Overview of steroid glaucoma

•Glucocorticoid-induced ocular hypertension is an important side-effect of glucocorticoid therapy

•Iatrogenic form of secondary open-angle glaucoma

•Clinically similar to primary open-angle glaucoma (POAG)

•Differences in individual susceptibility:

Approximately 40% of the general population are steroid responders

Almost all POAG patients are steroid responders

GC-induced ocular hypertension.5 Physician investigational new drug clinical studies suggest that a single anterior juxtascleral depot administration of AA lowers IOPs in patients with ocular hypertension due to intravitreal treatment with potent GCs5 and in POAG patients.6 Because of these positive preliminary studies, AA is currently in phase II clinical studies for both GC-induced ocular hypertension and in patients with POAG.

Etiology

Box 19.2  Pathology of steroid glaucoma

•Glucocorticoid-induced ocular hypertension due to decreased aqueous humor outflow

•Associated with morphological and biochemical changes in the trabecular meshwork

•Glucocorticoid-induced ocular hypertension occurs in multiple species (human, monkey, bovine, cat, dog, rat, mouse)

•Glucocorticoid-induced ocular hypertension occurs in isolated perfusion-cultured human eyes

Pathology

The elevated IOP caused by GC administration is associated with morphological and biochemical changes in the trabecular meshwork (TM), the tissue involved in impaired aqueous humor outflow (Box 19.2).3 There is increased deposition of extracellular material in the uveal meshwork and juxtacanalicular tissue of eyes with steroid glaucoma compared to age-matched control eyes. Some of this material has a characteristic “fingerprint” pattern.7 There is also a decrease in intertrabecular spaces and an apparent “activation” of trabecular cells (TM cells have a more extensive Golgi apparatus and rough endoplasmic reticulum).

Humans are not the only species to develop GC-induced ocular hypertension. Topical ocular administration of potent GCs can elevate IOP in rabbits8,9 and cats.10 Interestingly, topical ocular administration of DEX to cynomolgus monkeys elevated IOP by >5 mmHg in 40% of the dosed animals,11 very similar to the responder rate in normal human subjects. IOP lowered to normal levels after discontinuing steroid administration. The responder/nonresponder status remained the same when the animals were rechallenged with DEX administration. In addition, DEX-induced ocular hypertension has been shown in isolated (ex vivo) perfusion-cultured human eyes. An average 17 mmHg increase was seen in ~30% of the DEX-treated eyes compared to controls,2 a responder rate that mimics that seen in humans. This GC-mediated ocular hypertension was associated with thickening of trabecular beams, decreased intertrabecular spaces, and increased deposition of extracellular material in the juxtacanalicular connective tissue.

Etiology

Endogenous glucocorticoids and glaucoma

In addition to the ability of GCs to induce ocular hypertension, the endogenous GC cortisol has also been implicated in the development of POAG. There have been several reports of increased levels of the cortisol in the plasma12–14 and aqueous humor14 of POAG patients compared to agematched controls. However, others studies have not found this association. Diurnal and stress-induced changes in plasma cortisol levels further complicate potential disease associations. Early reports of increased lymphocyte sensitivity to GCs in POAG patients suggested an increased systemic GC sensitivity,15 although other studies failed to support this

finding. A new study showed that POAG patients were more sensitive to GC-induced cutaneous vasoconstriction compared to age-matched controls,16 and it will be interesting to see if this initial discovery can be independently verified. In addition, there have been several reports that steroid responsiveness is a risk factor for the development of POAG.17,18

Both the physiologic and pharmacologic effects of GCs are mediated by the GC receptor, which is a ligand-depend- ent transcription factor. It is therefore not surprising that GCs alter the expression of hundreds of TM cell genes.19–22 This altered expression includes both upregulated and downregulated genes in diverse categories and pathways, consistent with the pleotrophic effects of GCs on the TM. As previously discussed, the expression of certain genes involved in extracellular matrix (ECM) metabolism is altered by DEX treatment, including increased expression of ECM (FN1, COL8A, LUM) and proteinase inhibitor genes (SERPINA3), and decreased expression of proteinase genes (MMP1, TPA, ADAMTS5). A number of growth factor pathway genes are also altered, such as decreased expression of insulin-like growth factor (IGF1)-binding protein 2 (IGFBP2), IGF1, hepatocyte growth factor (HGF) and BMP2. Altered expression of several cytoskeletal genes (ACTA2, FLNB, NEBL) may be associated with GC-mediated reorganization of TM cell microfibrils and microtubules. In addition, DEX induced the expression of a number of stress-related genes (e.g., increased expression of SAA1, SAA2, methalothioneins, and ceruloplasmin).

Myocilin (MYOC) was first identified as a major GCresponsive gene and protein in the TM.23,24 This gene is one of the most abundantly expressed genes in human TM tissues and is also found in the aqueous humor. In addition to induction in cultured TM cells, GCs also increase MYOC expression in the TM of perfusion-cultured human anterior segments and in monkeys treated systemically with GCs.25 Although the MYOC promoter contains partial GC response elements (GREs), GC induction of MYOC requires additional RNA and protein synthesis, and therefore this induction is indirect.26 Although MYOC was originally proposed to be the major mediator of GC-induced ocular hypertension, there is still no compelling evidence showing that it plays any role in this GC-mediated event. In fact, genetically overexpressing or knocking out MYOC in mice had no effect on IOP.27,28 However, MYOC was the first glaucoma gene identified29 and is responsible for approximately 4% of POAG.11 Glaucomatous mutations in MYOC result in a gain-of-function phenotype, leading to nonsecretion30 and mistargeting of MYOC,31 which is normally a secreted glycoprotein. Expression of glaucomatous MYOC also induces the endoplasmic stress response in cultured TM cells.32

Pathophysiology

Effects of GCs on the trabecular meshwork (Box 19.3)

The GC-induced decrease in conventional aqueous outflow and GC-induced morphological changes in the TM point to the TM as being the target tissue involved in GC-induced

148

Table 19.3  Effects of glucocorticoids on the trabecular meshwork

Decreased ECM turnover (decreased MMPs and tPA; increased PAI-1 and TIMPs)

Cytoskeleton

Gene expression

ECM, extracellular matrix; FN, fibronectin; LM, laminin; MMPs, matrix metalloproteinases; tPA, tissue plasminogen activator; PAI-1, plasminogen activator inhibitor-1; TIMPs, tissue inhibitor of matrix metalloproteinase; CLANs, crosslinked actin networks.

Box 19.3  Glucocorticoid (GC) effects on the

trabecular meshwork (TM)

•TM cells have GC receptors and are targets of GC activity

•GCs alter the expression of hundreds of TM cells, genes, and proteins

•GCs alter TM cell functions (decrease proliferation and phagocytosis)

•GCs increase extracellular matrix deposition

•GC reorganize the actin cytoskeleton

•GCs alter cellular junctions

ocular hypertension. TM cells and TM tissues express GC receptors,33,34 which are essential for GC responsiveness. As seen in many other tissues, GCs have a wide variety of diverse effects on TM cells and TM tissues3 (Table 19.3). GCs alter several TM cellular functions, including proliferation, phagocytosis, and cell shape and size (Table 19.3). The potent GC DEX inhibited TM cell proliferation induced by a number of different growth factors, and at least part of this activity was mediated by DEX inhibition of growth factor receptor expression.35 DEX also inhibited TM cell migration36 and significantly increased TM cell and nucleus size.37 In addition, DEX inhibited TM cell phagocytosis in a perfusion culture ex vivo system.38 DEX also induced ultrastructural changes in cultured TM cells, including proliferation of the Golgi apparatus, stacking of the endoplasmic reticulum, and increased numbers of secretory vesicles, which provides morphological support for the increased ECM deposition seen after DEX treatment.37,39

One of the hallmarks of steroid-induced glaucoma is the deposition of ECM material in the TM. Aqueous humor outflow in the TM is regulated by the ECM.40 The overall increased deposition of ECM in the TM of steroid-induced ocular hypertensive eyes could be due to increased ECM

synthesis and/or decreased degradation. The synthesis of fibronectin,37,41,42 laminin,43 collagen,44 and elastin45 was increased in DEX-treated TM cells. GCs also affect ECM turnover. In addition to decreasing matrix metalloproteinase (MMP) and tissue plasminogen activator expression,46 GCs also increase the expression of plasminogen activator inhibi- tor-1 (Clark, unpublished observation) and tissue inhibitors of MMPs.47 GCs alter TM cell glycosaminoglycan (GAG) metabolism, decreasing hyaluronan and increasing chondr­ oitin sulfate and GAGase-resistant material.48

The TM cytoskeleton regulates aqueous humor outflow.49 The cytoskeleton also regulates a number of cell functions, including proliferation, migration, phagocytosis, and cell size/shape, all of which are affected in TM cells treated with GCs. GC treatment reorganizes the actin cytoskeleton to form cross-linked actin networks (CLANs) in cultured TM cells.36,37 CLANs are geodesic dome-like structures, and GC induction of CLANs is unique to TM cells, not occurring in a variety of other ocular and nonocular cells. The dose, time, and potency dependency for GC-induced CLAN formation are very similar to GC-induced ocular hypertension, and like GC-induced ocular hypertension, CLANs in TM cells are reversible after GC withdrawal.36 DEX-induced CLANs are also seen in perfusion-cultured eyes.42 Interestingly, very similar cytoskeletal changes, including CLANs, are present in cultured glaucomatous TM cells42 and in glaucomatous TM tissues.49 In addition to reorganizing the actin cytoskeleton, DEX treatment of cultured human TM cells alters microtubules to form microtubule tangles.50 However, we do not know whether DEX directly alters microtubules or whether this microtubule change is indirectly due to CLAN formation. These GC-induced changes in the TM cytoskeleton may make these cells more resistant to cytoskeletal disrupting agents. DEX-treated TM cells were more resistant to microtubule disrupting agent (i.e., ethacrynic acid and ethylene glycol tetraacetic acid)-induced cellular retraction.51

Glucocorticoid mechanism of action

The most widely accepted mechanism for transducing GC signals into cellular responses is via a cognate cellular GC receptor (GR) molecule. GR belongs to the family of intracellular ligand-inducible transcription factors termed the steroid/vitamin D/retinoic acid superfamily.52–54 The superfamily encompasses the steroid receptor family and the thyroid/retinoid/vitamin D (or nonsteroid) receptor family. The class of steroid hormone receptors includes GR forms α and ß, progesterone receptor (PR) forms A and B, mineralocorticoid receptor (MR), androgen receptor (AR), and estrogen receptor (ER) forms α and ß. Like other members of this receptor superfamily, GR protein is composed of structurally and functionally defined domains. The amino-terminal part of the protein contains a major transactivation domain responsible for gene activation, whereas the central part includes a highly conserved cysteine-rich DNA-binding domain, composed of two highly conserved zinc fingers. The zinc fingers consist of two zinc ions coordinated with eight cysteine residues to form two peptide loops, which bind cooperatively to half-sites in specific palindromic sequences in the promoter regions, known as GC response elements (GRE), and this specific DNA association induces receptor dimerization.52,55 The moderately conserved carboxy-

Pathophysiology

terminal includes ligand-binding domain, which possesses the essential property of hormone recognition and ensures both specificity and selectivity of the physiologic response.56–58 This region also contains sequences that are involved in nuclear translocation, receptor dimerization, heat shock protein (Hsp) 90 binding, and transactivation.59–62

The classic model for steroid/thyroid hormone action involves a ligand-induced conformational change in the receptor that allows the receptor–hormone complex to bind to its cognate hormone response element (HRE) in the promoter region of a target gene. The interaction of the activated receptor with the basal transcriptional apparatus alters transcription of hormone-sensitive genes.54,63 The ligand-binding domain can be thought of as a molecular switch that, upon binding ligand, shifts the receptor to a transcriptionally active state.

Alternative GR transcripts

The full-length human GR has two isoforms, GRα and GRβ, which originate from the same gene by alternative splicing of the GR primary RNA transcript.64–69 There is also alternative translation initiation from a downstream, in-frame ATG codon.70 Alternative translation initiation produces two GR protein products, the longer protein (777 amino acids), initiated from the first ATG codon (Met 1), termed as GR-A, and the shorter protein (751 amino acids), termed as GR-B. A and B receptor isoforms have been consistently identified for both GRα and GRß in various tissues and cell lines.70–73 When expressed in vitro in mammalian cells, the A and B forms are generated in approximately equivalent levels from a single cDNA. However, the GR-B form appears to be more susceptible to degradation and is more effective than GR-A in gene transactivation, but not in transrepression.70

GRα

GRα is the major GR transcript that has GC-binding activity.64 Because of its predominant expression, ligand-binding activity, and transcriptional function, most of the physiological and pharmacological effects of GCs are directly mediated by GRα. GRα is expressed in most human tissues and cell lines. Unlike most members of the steroid receptor superfamily, GRα resides predominantly in the cytoplasm of cells in the absence of ligand as a multiprotein heterocomplex that contains Hsp 90, Hsp70, immunophilin, and several other proteins.74–77 Hormone binding to GRα causes a conformation change and activation of GRα. The activated GRα can alter gene expression via GRE-dependent (classical) and GRE-independent (nonclassical) mechanisms.

GRE-dependent pathway

Activated GRα translocates to the nucleus via retrograde transport along microtubules. Once in the nucleus, GRα can bind to specific palindromic DNA sequences (GRE) as a homodimer on the promoter region of target genes, where it interacts with the basal transcription apparatus to induce transcription of the target genes.78 In addition, GRα also functions as a negative regulator of transcription in a specific subset of GC-responsive genes, which contain a negative GRE (nGRE).79,80

Section 3  Glaucoma Chapter 19  Steroid-induced glaucoma

GRE-independent pathway

There is an additional way for GRα to inhibit rather than activate gene expression. GRα can inhibit the expression of genes that do not contain nGRE. GCs are known to suppress the expression of proinflammatory cytokines, which are key regulators of the immune response. However, the majority of proinflammatory genes that are suppressed by GCs lack nGRE in their promoter regions.81–83 Instead, GRα physically interacts with other transcription factors to prevent them from binding to their response elements of genes. The powerful GC-mediated anti-inflammatory actions and immune suppression are mediated via this GREindependent pathway.84–86

antagonize the function of GRα.66,90 Increased GRß expression has been associated with a variety of GC-insensitive conditions.91–99 We reported that decreased expression of GRß in glaucomatous TM cells was associated with increased GC sensitivity in glaucoma.73,100 These reports suggest a potential physiological consequence to changes in GRß expression. In addition, GRß has been reported to bind a ligand and has the ability to regulate gene expression on its own,101 suggesting that GRß may regulate GC responsiveness beyond its ability to manipulate the function of GRα. GRß may compete with GRα for GRE binding because GRß has an intact DNA-binding domain. Alternatively, GRß may complex with activated GRα to form transcriptional impaired GRα-GRß heterodimers, as has been demonstrated in corti- costeroid-insensitive cells70,102 (Figure 19.1).

GRß

In contrast, GRß was thought to be a nucleus-localized orphan receptor lacking ligand-binding activity and gene transcription regulation and, hence, it was suggested that GRß was generally of little physiological importance. However, there is increasing evidence that this view is incor- rect.87–89 GRß can act as a dominant negative regulator to

GRß role in glucocorticoid-resistant diseases

GCs are routinely used as anti-inflammatory drugs, and GC resistance poses a serious clinical problem. A number of clinical studies have reported an association between GRß expression and GC insensitivity. Increased expression of GRß appears to be responsible for unresponsiveness

to anti-inflammatory GC therapy in diverse clinical conditions, including asthma, inflammatory bowel disease, rheumatoid arthritis, ulcerative colitis, chronic lympho­ cytic leukemia, and microbial antigen-induced GCinsensitivity.91,95,96,98,99,103,104 GRß may be a key modulator of the development of these immune-related GC-resistant diseases.105–108 In addition, increased expression of GRß is associated with disease pathogenesis. However, there is considerable controversy about the expression and physiological significance of GRß.109–113

GRß role in glaucoma and enhanced TM cell sensitivity to GC (Box 19.4)

There are differences in steroid sensitivity between the glaucomatous and normal populations14; however the molecular basis of the increase in IOP experienced by patients with glaucoma and subjects receiving GCs is not well understood. We have reported73 that glaucomatous TM cells derived from donors with a documented history of glaucoma had a relatively lower expression of GRß, particularly lower in the nucleus, than did normal TM cells derived from normal donors. Glaucomatous TM cells were more susceptible to DEX induction of transiently expressed GRE-luciferase construct than were normal TM cells. Transfection of glaucomatous TM cells with an GRß expression vector suppressed the GC induction of the luciferase reporter gene and decreased the GC induction of MYOC and fibronectin, two glaucomaassociated genes in TM cells. A decrease in phagocytic activity has been proposed in the pathogenesis of glaucoma,38,114 although one study did not find any difference in phagocytic activity between normal and glaucomatous eyes.38 GCs suppress TM cell phagocytosis in perfusion-cultured eyes.115 We have also reported100 different phagocytic abilities between cultured normal TM and glaucomatous TM cells, with a reduced phagocytosis in glaucomatous TM cells. The glaucomatous TM cells were more responsive to DEX-induced suppression of phagocytosis compared to normal TM cells. Increased expression of GRß decreased this DEX response, and consequently these cells retained their phagocytic activity even following the DEX challenge.

Several mechanisms may be responsible for higher levels of GRß found in normal compared to glaucomatous TM cells. Expression of GRß could be regulated by alternative splicing efficiency through variations in splice sites, altered expression of splicing factors, or the presence of functional polymorphisms in splicing factors. GRß mRNA stability appears to be controlled by a 3’ untranslated region (UTR) single nucleotide polymorphism (SNP) and this SNP may vary between normal subjects and patients with glaucoma.

Box 19.4  Potential role for GRβ

•GRβ = alternative splice form of the glucocorticoid receptor

•Dominant regulator of glucocorticoid activity

•Implicated in a variety of steroid-resistant diseases

•Decreased levels in glaucomatous trabecular meshwork cells make these cells more sensitive to detrimental effects of glucorticoids

Pathophysiology

Genotyping studies are currently in progress to determine whether there are disease-associated polymorphisms in any of these sites in patients with glaucoma. Differences in the nuclear translocation of GRß may also explain the differential expression in normal versus glaucomatous TM cells. The expression level of GRß in TM cells can regulate the cell’s ability to respond to GC, and GRß inhibit not only the DEXinduced gene expression but also DEX-suppressed phagocytic activity. The lower expression of GRß in glaucomatous TM cells could contribute to increased susceptibility of glaucoma patients to chronic endogenous cortisol exposure (years) or to exogenous pharmacological GC treatment, which may lead to the increased aqueous humor outflow resistance and elevated IOP.

Interestingly, recent reports showed that, in the absence of GC, GRα was sequestered in the cytoplasm, and nuclear GRß acted as a transcriptional repressor of cytokine genes through the recruitment of histone deacetylase complexes.116 GRß, expressed in the absence of GRα, can regulate the expression of a number of diverse genes.101 These studies suggest that GRß can act as an orphan receptor, representing new constitutive activator/repressor activity, along with being a repressor for ligand-activated GRα. In addition, RU-486 was able to bind to GRß and stimulate gene transcription, despite only having a partial ligand-binding domain in its carboxy terminus. We also observed that, in the presence of FK506, DEX treatment induced nuclear translocation of GRß as it does to GRα,117 indicating that DEX might be a ligand for GRß in the presence of FK506. Our understanding of the physiological significance of GRß is expanding.

Differences in tissue/cell responses to GCs

Although almost all mammalian cells and tissues contain GRs, the biological responses to GCs are often widely divergent between different cell types. Indeed, GCs have a broad array of life-sustaining functions and are also frequently used as therapeutic compounds. Variations in tissue sensitivity and responses are essential for maintaining tissue homeostasis under both basal and stress conditions. Numerous pathologic conditions can be caused by either GC resistance or GC hypersensitivity. In human, generalized GC resistance can be due to inactivating mutations of the GR gene, but this syndrome is rare.118 A number of factors contribute to tissuespecific GC responses, many of which involve specific steps of the GR signaling pathway, including: ligand availability, ligand-binding activity, receptor isoform expression, intracellular trafficking, promoter association, interactions with tissue-specific cofactors, and clearance of the receptor from the target genes.119,120

One regulatory mechanism to change cellular responses to GCs occurs via GR phosphorylation, which influences multiple functions of the GR protein, including affinity for ligand binding, intracellular trafficking, and transcriptional activity.121–125 Several protein kinases, such as the p38 mitogen-activated protein kinase (MAPK),126 cell cycle-dependent kinases (CDK), and other mitogenactivated kinases have been reported to phosphorylate the GR.124,127–130 These regulatory mechanisms appear to be in tissue-specific fashions; for example, CDK5 functions specifically in the central nervous system.120

Selective glucocorticoid receptor agonists – future directions

GCs are used in the management of a wide range of inflammatory diseases and immunosuppressive agents, including a number of conditions affecting the eye, as noted above. Their complex actions have been attributed to a variety of cellular mechanisms, including direct effects on plasma membranes, interactions with transcription factors, genomic effects mediated by the GC response element and, more recently, a membrane-bound GC receptor.131 The classical genomic effects can be attributed to two types of responses, transactivation or transrepression of selective genes. GCmediated receptor transactivation may explain many of their side-effects, whereas the repression of key inflammatory transcription factors, including activator protein 1 (AP-1) and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB), is linked to their anti-inflammatory activity.132 Therefore, selective agents that can distinguish the generepressive actions from the transcriptional activation pathways will be desirable. However, separation of the desirable effects from the undesirable effects is difficult since unwanted side-effects can be seen with both types of activity. Common ocular side-effects of prolonged GC therapy include steroidinduced ocular hypertension and posterior subcapsular cataracts. Moreover, depending on the dose and duration of therapy, one can get infections, disturbed wound healing, and hypertension.133 The search for novel formulations is still a major interest of the pharmaceutical industry, particularly those clinical uses that might separate the desired generepressive anti-inflammatory activity from the gene activation activity that appears to promote side-effects. Much has been

learned from using a mouse model where the GC receptor was mutated in the dimerization portion of the molecule, which prevented GRE transactivation activity but allowed the receptor to bind other transcription factors and repress gene expression.134 Treatment of these mice with a GC generated anti-inflammatory activity, as in wild-type mice. However, these mice were not without all potential sideeffects because chronic treatment with a GC resulted in the induction of GC-dependent osteoporosis.135 This suggests that some of the transactivation activity may not be the result of the dimerization of the receptor, and other mechanisms may still exist for gene activation. Therefore, this model may not be as useful to tease out the different pathways for drug development.

A number of compounds have been produced to separate potent anti-inflammatory activity from side-effects, and there appears to be a beneficial ratio of desired over undesirable actions.136 It remains uncertain if these differences reside in the different pathways of genomic transrepression versus transactivation or the result of differences in receptor affinities or differences in receptor expressions. It will also be important to determine what pathways and cellular mechanisms are responsible for the anti-inflammatory versus each of the side-effects following GC administration. Since GCs induce the expression of a wide variety of cellular proteins, some of which may still be important for antiinflammatory activity, it may be extremely difficult to separate out the undesirable actions totally. However, it is worth the effort to identify selective GC receptor agonists with much reduced side-effects. For the eye, the ability to retain the anti-inflammatory actions while reducing the potential for developing glaucoma or cataract will be ideal. Certainly, the search is still on for such a compound.

2.Clark AF. Steroids, ocular hypertension, and glaucoma. J Glaucoma 1995;4:354– 369.

3.Wordinger RJ, Clark AF. Effects of glucocorticoids on the trabecular meshwork: towards a better understanding of glaucoma. Prog Retin Eye Res 1999;18:629–667.

4.Clark AF. Preclinical efficacy of anecortate acetate. Surv Ophthalmol 2007;52(Suppl. 1):S41–S48.

11.Fingert JH, Clark AF, Craig JE, et al. Evaluation of the myocilin (MYOC) glaucoma gene in monkey and human steroid-induced ocular hypertension. Invest Ophthalmol Vis Sci 2001;42: 145–152.

17.Kitazawa Y, Horie T. The prognosis of corticosteroid-responsive individuals. Arch Ophthalmol 1981;99:819–823.

29.Stone EM, Fingert JH, Alward WL, et al. Identification of a gene that causes primary open angle glaucoma. Science 1997;275:668–670.

Clinical background

Lowering intraocular pressure (IOP) remains the only proven method of preventing the onset and progression of glaucoma, yet the role of IOP in the disease remains controversial. This largely arises from the wide spectrum of individual susceptibility to IOP wherein a significant number of patients with normal IOPs develop glaucoma (e.g., normotensive glaucoma), and other individuals with elevated IOP show no signs of the disease. It is therefore important to understand the relationship between IOP and glaucomatous optic neuropathy when IOP is only one of several factors that influence the disease (Box 20.1). IOP is, by definition, a mechanical entity – the force per unit area exerted by the intraocular fluids on the tissues that contain them. Glaucomatous optic neuropathy is a biologic effect – likely the result of an IOP-related cascade of cellular events that culminate in damage to the retinal ganglion cell (RGC) axons. One of the challenges of biomechanics is to understand how the mechanics are transduced into a biological response and/or tissue damage. How does this take place? Why is there such a wide range in susceptibility to IOP? Why does elevated IOP lead to that particular cascade of events and not another? What can we do to predict, detect, and stop the progression of glaucomatous damage? Unfortunately we do not have answers to these questions, but in recent years there has been considerable progress towards understanding of the role of IOP in glaucoma. In this chapter, we focus on two main themes: what is known about how IOP-related forces and deformations are distributed in the posterior pole and optic nerve head (ONH), and what is known about the response of the living system.

Pathology

From a biomechanical perspective, the ONH is a natural site of interest when studying IOP effects because it is a discontinuity in the corneoscleral shell. Such discontinuities are often weak spots in mechanically loaded systems because they give rise to significant stress concentrations.1 In addition, it is the ONH, the lamina cribrosa (LC) in particular, that is the principal site of RGC axonal insult in glaucoma.2 Nevertheless there is evidence both for3 and against4 direct,

Biomechanical changes of the optic disc

Ian A Sigal, Michael D Roberts, Michael JA Girard, Claude F Burgoyne, and J Crawford Downs

IOP-induced damage to the retinal photoreceptors, and it is likely that there are also important pathophysiologies within the lateral geniculate and visual cortex.5

Etiology

We have proposed that the ONH be understood as a biomechanical structure, and that the mechanical effects of IOP on the tissues of the ONH, namely forces and deformations, are central determinants of both the physiology and pathophysiology of the ONH tissues and their blood supply at all levels of IOP (Figure 20.1). Within this framework, the susceptibility of a particular patient’s ONH to IOP-related insult is a function of the biomechanical response of the constituent tissues and the resulting mechanical, ischemic, and cellular events driven by that response. Experienced over a lifetime at physiologic levels of IOP, these events underlie normal ONH aging. Hence, eyes with a particular combination of tissue geometry and stiffness may be susceptible to glaucomatous damage at normal IOP, while others may have a combination of ONH tissue geometry and stiffness that render them impervious to any deleterious effects of high IOP.

We believe that the mechanical and vascular mechanisms of glaucomatous injury are inseparably intertwined: IOPrelated mechanics determines the biomechanical environment within the ONH, mediating blood flow and cellular responses through various pathways. Reciprocally, the biomechanics depend on tissue anatomy and composition, which are subject to change through cellular activities such as remodeling. The interrelationship between mechanics and physiology could be particularly strong within the LC due to its complexity. The LC is composed of a threedimensional (3D) network of beams of connective tissue, many containing capillaries, that provides functional and structural support to the RGC axons. IOP-related forces within the LC could deform the beams containing capillaries, diminishing the blood supply to the laminar segments of the RGC axons. Conversely, primary insufficiency in the blood supply to the laminar region could introduce cellmediated connective tissue changes that could remodel the extracellular matrix (ECM), making the laminar beams more prone to failure, and limit the diffusion of nutrients to adjacent RGC axons.

Feedback

Box 20.1  Why study optic nerve head

biomechanics?

•Intraocular pressure (IOP) is the principal risk factor for the onset and progression of glaucoma

•IOP is a mechanical entity that induces glaucomatous damage through a variety of mechanisms

•Biomechanical analysis of the effects of IOP on the optic nerve head and sclera will help us understand how IOP leads to glaucoma, and why some individuals are more susceptible to IOP-induced glaucomatous damage than others

Before proceeding with a presentation of the mechanical effects of IOP on the eye and ONH, we review the basic concepts of mechanics relevant to this analysis.

Basic concepts in mechanics

The following are fundamental terms and concepts from mechanics that may not be familiar to clinicians (Box 20.2). The interested reader may pursue these ideas in greater depth by referring to appropriate textbooks.1

Stress is a measure of the forces transmitted through, or carried by, a material or tissue. Specifically, stress is the force divided by the cross-sectional area over which it acts. For example, pressure is a stress and can be expressed in pounds per square inch (psi).

Strain is a measure of the local deformation induced by an applied stress. It is computed as the change in length of a material divided by its resting length, and is often expressed as a percentage. For example, a wire that was originally 10 mm long that is stretched an additional 1 mm exhibits 10% tensile strain.

In addition to tension and compression, a material can undergo shear. Tension, compression, and shear are

154

Box 20.2  Basic concepts in mechanics

•Strain is a measure of the local deformation

•Stress is a measure of the forces per unit area

•Stress and strain are related to each other through the material properties

•Ocular tissues are complex materials that are nonlinear, anisotropic, viscoelastic, and inhomogeneous

•A nonlinear material varies in stiffness as it deforms

•An anisotropic material exhibits different stiffness in different directions

•A viscoelastic material exhibits higher stiffness when loaded quickly rather than slowly

often referred to as the three modes of stress and strain.6 However, these three modes are not independent, as shown in Figure 20.2.

It has been established that the biologic response of tissues and cells depends strongly on the mode of the strain stimulus (tension, compression, or shear), as well as on their magnitudes and temporal profiles.7,8 It is therefore of interest to determine which modes of strain and stress the tissues of the ONH are exposed to as IOP is elevated. Note that strains are generally not homogeneous. When the LC deforms, some regions could be highly strained in different modes, whereas others remain largely unaffected. This is important because the biological effects on cells are likely to be strongly dependent on the local levels of strain or stress rather than on global levels.

We would also like to emphasize that mechanical stress, which represents forces, is not synonymous with notions of stress typically used in physiologic or metabolic contexts (e.g., ischemic or oxidative stress). Mechanical stress cannot be measured directly, and we believe it is strain that damages tissues. However, stress is often used to predict the sites of

Figure 20.2  Schematic illustration of tissue straining in two dimensions.

(A) A square tissue region abcd is deformed by forces that act normal to the faces of the square, represented by the black arrows. The tissue experiences tensile stretching in one direction and compression in another.

Superimposed on the square tissue region is a circle that deforms with the tissue. (B) The same tissue region is deformed by shearing forces that act tangential to the faces of the square. However, as this region deforms it not only experiences shear, but also extension and compression, as can be verified by noting how the distances between ac and bd change. (Adapted from Sigal IA, Flanagan JG, Tertinegg I, et al. Predicted extension, compression and shearing of optic nerve head tissues. Exp Eye Res 2007;85:312–322.)

and failure in engineering structures1 and has been correlated to damage in tissues,9 so it may be that, while strain is causing the damage, stress is a better predictor of the sites of that damage.10

Stress and strain (i.e., forces and deformation) in a material are related to each other through material properties, and this constitutive relationship is intrinsic to each material. For a given load, a material that exhibits large strains is thought of as compliant. Conversely, a material that exhibits small strains for the same load is termed stiff. A stiff tissue such as sclera can have high stress but low strain, while an equal volume of compliant tissue, like retina, might have high strain even at low levels of stress. The simple description above does not account for many of the complexities in material properties that occur in soft biologic tissues, such as anisotropy, nonlinearity, and viscoelasticity (Box 20.3). These complexities are likely to be fundamental to understanding ocular mechanics and will be discussed in the context of scleral biomechanics below.

Scleral biomechanics

From a mechanical perspective, the eye is a pressure vessel, on which IOP produces deformation, strain, and stress. Through computational modeling, the sclera has been shown to have a strong influence on how the LC deforms when IOP changes (Box 20.4). Therefore, understanding the mechanical behavior of the sclera is essential to understanding IOP-induced LC deformations. One of the mechanisms

Figure 20.3  Influence of scleral mechanics on optic nerve head mechanics. Intraocular pressure (IOP) induces large deformations on a compliant sclera (top); these deformations are transmitted to the scleral canal, resulting in a large scleral canal expansion that pulls the lamina cribrosa (LC) taut despite the direct posterior force of IOP on the LC. Conversely, a stiff sclera deforms little under IOP (bottom), with small scleral canal expansions and little lateral stretching of the LC, thus allowing the LC to be displaced posteriorly by the direct action of IOP on its anterior surface.

Box 20.3  Optic nerve head (ONH) biomechanics are

not trivial

•ONH biomechanics are complex, with the tissues simultaneously subject to tension, compression, and shear

•At present, it is difficult to measure the response of the lamina to changes in intraocular pressure (IOP), and therefore computational modeling is being used

•These models suggest that IOP-related stresses and strains within the connective tissues of the ONH are substantial, even at normal IOP

Box 20.4  Importance of the sclera

•The sclera is the main load-bearing tissue of the eye and deformations of the sclera are transmitted to the optic nerve head at the scleral canal wall

•As such, the mechanical properties and behavior of the sclera have a strong influence on how the optic nerve head responds to changes in intraocular pressure

by which scleral biomechanics can influence the response of the LC to IOP is illustrated in Figure 20.3.

Alas, describing the mechanical behavior of soft tissue such as the sclera is a formidable undertaking that demands extensive experimental and mathematical efforts. The first step in characterizing scleral material properties is the development of an experimental mechanical test to measure the deformation of the tissue subjected to loads (e.g., uniaxial, biaxial, or pressurization tests). The second step is the development of a constitutive model (i.e., relationship between

stresses and strains) which describes the tissue mechanical behavior as observed in the experiment and provides a mathematical representation of the tissue’s material properties.

Historically, the sclera has been described as a thinwalled, spherical pressure vessel obeying the analytical equation known as Laplace’s law. Laplace’s law is useful to estimate the state of stress in nonbiological pressure vessels, but it is inadequate for describing many aspects of the eye’s mechanical response to variations in IOP. The sclera has been shown to exhibit several properties that violate the assumptions of Laplace’s law. First, the eye is a pressure vessel of nonuniform thickness.11–13 Second, in terms of its material properties, the sclera is nonlinear,14–16 anisotropic,16,17 and viscoelastic.18–20 These concepts are fundamental to understanding scleral and ONH biomechanics, and below we present them in greater detail.

Nonlinearity is a property exhibited by most soft tissues, often as a consequence of collagen fibers within the tissue1 (Figure 20.4). In a pressure vessel with linear material properties, the stiffness would remain constant during pressureinduced deformation, whereas a nonlinear pressure vessel would experience either softening or stiffening as it deforms. Recent experiments have shown that the sclera exhibits a considerable increase in stiffness, at least fivefold, when exposed to an acute elevation of IOP from 5 to 45 mmHg.15 This dramatic change shows the substantial impact that collagen fibers can have on scleral stiffness and deformation.

Anisotropy, as opposed to isotropy, is the property by which materials exhibit different stiffness in different directions. For thin biological tissues, anisotropy is primarily dictated by the organization of their fibrous structure, which is confined within the plane of the tissue, as illustrated in Figure 20.4. Unlike the cornea, scleral collagen fiber orientation has not been fully characterized. However, it has been shown through computational modeling that collagen fiber orientation and distribution are major determinants of scleral deformation.21 Further experimental work is needed to characterize better scleral anisotropy and its effects on LC biomechanics.

Viscoelastic materials, such as the sclera, exhibit higher resistance to deformation when loaded quickly rather than slowly. Downs and coworkers characterized the viscoelastic material properties of normal rabbit and monkey peripapillary sclera18,22 and found that the material properties of peripapillary sclera are highly time-dependent (viscoelastic). This behavior protects ocular tissues from large deformations during short-term spikes in IOP, which occur during blinks, eye rubbing, or high-speed impacts.

These three aspects of the sclera’s material properties – nonlinearity, anisotropy, and viscoelasticity – affect IOPinduced scleral deformations, but they are not the only important determinants of the eye’s mechanical response. Material properties may be combined with thickness and shape to define another useful concept, that of structural, or effective, stiffness.

Studies of the scleral thickness of human13,23 and monkey11,12 eyes show that, on average, the human sclera is about twice as thick as the monkey sclera. The sclera is thinnest near the equator (as thin as 100 m in both species) and thickest in the peripapillary region (average of 1000 m in the human and 450 m in the monkey). Large variations

Figure 20.4  (A) Nonlinearity is a property that sclera exhibits, in which the relationship between loading and deformation is not linear. At low intraocular pressure (IOP), collagen fibers are initially crimped, which makes the sclera more compliant. As IOP increases, the scleral collagen fibers uncrimp and eventually become straight, resulting in a dramatic increase in scleral stiffness (an increase in the amount of IOP elevation necessary to produce the same deformation). (B) Schematic illustration of the various degrees of planar anisotropy present in thin soft tissues. Skin has a highly disorganized arrangement of collagen fibers and therefore resists loads similarly in many directions, a property known as isotropy. In contrast, tendons have well-organized collagen fibers running principally in the longitudinal direction, and therefore these tissues sustain loads differently along and across their length, a property known as anisotropy. The sclera is thought to have a collagen fiber alignment that is between those of skin and ligaments.

in peripapillary scleral thickness occur naturally and in pathologic conditions (e.g., myopia24), and have been hypothesized to be an important determinant of individual susceptibility to IOP.25,26 Figure 20.5 illustrates how IOPrelated stress is distributed in the peripapillary sclera in two situations: homogeneous thickness with a circular scleral canal, and inhomogeneous thickness with an elliptical scleral canal.

Optic nerve head and lamina   cribosa biomechanics

Models of the optic nerve head

Initial experimental studies of ONH biomechanics were often designed to examine and quantify a posterior deformation of the LC in response to an acute increase in IOP. Unfortunately, it is difficult to take measurements of the LC directly because it is fragile and relatively inaccessible, and