The adult mammalian retina is a highly differentiated tissue containing eight distinct layers plus the RPE, 10 major types of neurons, and a Müller glial cell (Fig. 17-1). The eight layers of the neural retina, which originate from the cells of the inner layer of the embryonic optic cup, are the nerve fiber layer (NFL), GCL, IPL, INL, OPL, ONL, rod and cone photoreceptor inner segment layer (RIS; CIS), and the rod and cone photoreceptor outer segment layer (ROS; COS). The RPE, which originates from the cells of the outer layer of the embryonic optic cup, is a single layer of cuboidal epithelial cells that lies on Bruch membrane adjacent to the vascular choroid. Between the RPE and photoreceptor outer segments lies the subretinal space, which is similar to the brain ventricles. The 10 major types of neurons are the rod and cone photoreceptors, (depolarizing) ON-rod and ON-cone bipolar cells, (hyperpolarizing) OFF-cone bipolar cells, horizontal cells, numerous subtypes of amacrine cells, an interplexiform cell, and ON-RGCs and OFF-RGCs. The Müller glial cell (MGC) is the only glial cell in the retina. The somas of the MGCs are in the INL. The end feet of the MGCs in the proximal or inner retina along with a basal lamina form the internal limiting membrane of the retina, which is similar to the pial surface of the brain. In the distal retina, the MGC end feet join with the photoreceptors and zonula adherens to form the external limiting membrane, which is located between the ONL and RIS/CIS. The interested reader is referred to the excellent references in the Introduction section as well as to numerous outstanding websites devoted exclusively to the retina (http://webvision.med.utah.edu; http://cvs.anu.edu.au; http://retina.anatomy.upenn.edu) for basic information on the anatomic, biochemical, cell and molecular biological, and physiological aspects of retinal structure and function.
The mammalian retina is highly vulnerable to toxicant-induced structural and/or functional damage due to (1) the presence of a highly fenestrated choriocapillaris that supplies the distal or outer retina as well as a portion of the inner retina; (2) the very high rate of oxidative mitochondrial metabolism, especially in photoreceptors (Ahmed et al., 1993; Medrano and Fox, 1995; Braun et al., 1995; Winkler, 1995; Shulman and Fox, 1996); (3) high daily turnover of rod and cone outer segments (LaVail, 1976); (4) high susceptibility of the rod and cones to degeneration due to inherited retinal dystrophies as well as associated syndromes and metabolic disorders (van Soest et al., 1999; Jones and Marc, 2005); (5) presence of specialized ribbon synapses and synaptic contact sites (Wässle, 2004; tom Dieck and Brandstätter, 2006); (6) presence of numerous neurotransmitter and neuromodulatory systems, including extensive glutamatergic, GABAergic and glycinergic systems (Kalloniatis and Tomisich, 1999; Wässle, 2004); (7) presence of numerous and highly specialized gap junctions used in the information signaling process (Bloomfield and Völgyi, 2009); (8) presence of melanin in the choroid, RPE, uvea, and iris (Meier-Ruge, 1972; Potts and Au, 1976); (9) a very high choroidal blood flow rate, as high as 10 times that of the gray matter of the brain (Nickla and Wallman, 2010); and (10) the additive or synergistic toxic action of certain chemicals with light (Dayhaw-Barker et al., 1986; Backstrom et al., 1993; Roberts, 2001, 2002; Glickman, 2002).
The retina is also an excellent model system for studying the effects of chemicals on the developing and mature CNS. Its structure–function relations are well established. The histogenic steps of development of the neurons and glial components are well characterized. The development of the CNS and most retinal cells occurs early during gestation in humans (Hendrickson, 1992; Hendrickson and Drucker, 1992) and continues for an additional seven to 14 days postnatally in the rat (Dobbing and Sands, 1979; Raedler and Sievers, 1975). Therefore, toxicological effects in the rodent retina have relevance for chemical exposure during the early gestation period in humans as well as during early postnatal development. The retina contains a wide diversity of synaptic transmitters and second messengers whose developmental patterns are well described. Moreover, the rodent retina is easily accessible, it has most of the same anatomical and functional features found in the developing and mature human retina, and the rat rod pathway is similar to that in other mammals (Finlay and Sengelbaub, 1989; Chun et al., 1993). Finally, rat rods have similar dimensions, photochemistry, and photocurrents as human and monkey rods (Baylor et al., 1984; Schnapf et al., 1988). These general and specific features underscore the relevance and applicability of using the rodent retina to investigate the effects of chemicals on this target site as well as a model to investigate the neurotoxic effects of chemicals during development.
Each retinal layer can undergo specific as well as general toxic effects. These alterations and deficits include, but are not limited to visual field deficits, scotopic vision deficits such as night blindness and increases in the threshold for dark adaptation, cone-mediated (photopic) deficits such as decreased color perception, decreased visual acuity, macular and general retina edema, retinal hemorrhages and vasoconstriction, and pigmentary changes. The list of drugs and other chemicals that cause retinal alterations is extensive, as evidenced by an examination of Table 17-1, Grant's Toxicology of the Eye (Grant and Schuman, 1993), Bartlett and Jaanus (2008) discussion of the adverse retinal effects of therapeutic systemic drugs and a review by Wolfensberger (1998) that addresses the adverse effects of drugs on the RPE. The overall aim of this section is to discuss in detail several chemicals, solvents, and drugs: (1) that are used as drugs or that are environmentally relevant neurotoxicants; (2) whose behavioral, physiological, and/or pathological effects on the retina are known; and (3) whose retinal site(s) and/or mechanism of action are well characterized.
The chemical- and drug-induced alterations in retinal structure and function are grouped into two major categories. The first category focuses on retinotoxicity of systemically administered therapeutic drugs. Six major drugs are discussed in detail: chloroquine/hydroxychloroquine, digoxin/digitoxin, indomethacin, sildenafil, tamoxifen, and vigabatrin. The second category focuses on well-known neurotoxicants that produce retinotoxicity: inorganic lead, methanol, selected organic solvents, and organophosphates. See Chap. 16 and 23 for information on the effects of lead on the brain and other target organs and Chap. 24 for additional information on methanol and the organic solvents discussed below.
Retinotoxicity of Systemically Administered Therapeutic Drugs
Ocular toxicity is a common side effect of cancer chemotherapy (Imperia et al., 1989; Schmid et al., 2006). Symptoms include blurred vision, diplopia, decreased color vision, decreased visual acuity, optic/retrobular neuritis, transient cortical blindness, and demyelination of the optic nerves (Imperia et al., 1989; Schmid et al., 2006). The retina due to its high metabolic activity and choroidal circulation (vide infra) appears to be particularly vulnerable to numerous cytotoxic drugs such as the alkylating agents cisplatin, carboplatin, and carmustine; the antimetabolites cytosine arabinoside, 5-fluorouracil, and methotrexate; and the mitotic inhibitors such as docetaxel. The ocular toxicity of different drugs is dependent upon the dose, duration of dosage, and route of administration. However, if not detected at an early stage of toxicity, the ocular complications are often irreversible even after chemotherapy is discontinued (Imperia et al., 1989; Schmid et al., 2006). One strategy to avoid such retinal complications is to conduct prospective ophthalmological examinations as well as scotopic and photopic ERG testing prior to the onset and during chemotherapy. The ocular side effects of tamoxifen, an estrogen antagonist used in oncology, are discussed below.
Two of the most extensively studied retinotoxic drugs are chloroquine (Aralen) and hydroxychloroquine (Plaquenil). The first case of chloroquine-induced retinopathy was reported more than 40 years ago (Bartlett and Jaanus, 2008). These 4-aminoquinoline derivatives are used as antimalarial and anti-inflammatory drugs. The low-dose therapy used for malaria is essentially free from toxic side effects; however, the chronic, high-dose therapy used for rheumatoid arthritis, and discoid and systemic lupus erythematosus (initially 400–600 mg/day for four to 12 weeks and then 200–400 mg/day; Ellsworth et al., 1999) can cause irreversible loss of retinal function. Chloroquine, its major metabolite desethylchloroquine, and hydroxychloroquine have high affinity for melanin, which results in very high concentrations of these drugs accumulating in the choroid and RPE, ciliary body, and iris during and following drug administration (Rosenthal et al., 1978). Prolonged exposure of the retina to these drugs, especially chloroquine, may lead to an irreversible retinopathy. In fact, small amounts of chloroquine and its metabolites were excreted in the urine years after cessation of drug treatment (Bernstein, 1967). Approximately, 20% to 30% of patients who received high doses of chloroquine exhibited some type of retinal abnormality, whereas 5% to 10% showed severe changes in retinal function (Burns, 1966; Shearer and Dubois, 1967; Sassaman et al., 1970; Krill et al., 1971). Hydroxychloroquine is now the drug of choice for treatment of rheumatic diseases because it has fewer side effects and less ocular toxicity. Doses less than 400 mg/day appear to produce little or no retinopathy even after prolonged therapy (Johnson and Vine, 1987).
The clinical findings accompanying chloroquine retinopathy can be divided into early and late stages. The early changes include (1) the pathognomonic “bull's-eye retina” visualized as a dark, central pigmented area involving the macula, surrounded by a pale ring of depigmentation, which is surrounded by another ring of pigmentation; (2) a diminished EOG; (3) possible granular pigmentation in the peripheral retina; and (4) visual complaints such as blurred vision and problems discerning letters or words. Late-stage findings, which can occur during or even following cessation of drug exposure, include (1) a progressive scotoma, (2) constriction of the peripheral fields commencing in the upper temporal quadrant, (3) narrowing of the retinal artery, (4) color and night blindness, (5) absence of a typical retinal pigment pattern, and (6) very abnormal EOGs and ERGs. These late-stage symptoms are irreversible. Interestingly, dark adaptation is relatively normal even during the late stages of chloroquine retinopathy, which helps distinguish the peripheral retinal changes from those observed in patients with retinitis pigmentosa (Bernstein, 1967).
In humans and monkeys, long-term chloroquine administration results in sequential degeneration of the RGCs, photoreceptors, and RPE and the eventual migration of RPE pigment into the ONL and OPL. In addition, in the RPE there is a thickening of the RPE layer, an increase in the mucopolysaccharide and sulfhydryl group content, and a decrease in activity of several enzymes (Ramsey and Fine, 1972; Rosenthal et al., 1978). Although the molecular mechanism of action is unknown, it has been suggested that the primary biochemical mechanism is inhibition of protein synthesis (Bernstein, 1967).
The cardiac glycosides digoxin and digitoxin are digitalis derivatives used in the treatment of congestive heart disease and in certain cardiac arrhythmias. As part of the extract of the plant foxglove, digitalis was recommended for heart failure (dropsy) over 200 years ago. Digitalis-induced visual system abnormalities such as decreased vision, flickering scotomas, and altered color vision were documented during that time (Withering, 1785). Approximately, 20% to 60% of patients with cardiac glycoside serum levels in the therapeutic range and 50% to 80% of the patients with cardiac glycoside serum levels in the toxic range complain of visual system disturbances within two weeks after the onset of therapy (Robertson et al., 1966; Haustein et al., 1982; Duncker et al., 1994). Digoxin produces more toxicity than digitoxin due to its greater volume of distribution and plasma protein binding (Haustein and Schmidt, 1988). The most frequent visual complaints are color vision impairments and hazy or snowy vision, although complaints of flickering light, colored spots surrounded by bright halos, blurred vision, and glare sensitivity also are reported. The color vision disturbances have been confirmed with the Farnsworth-Munsell 100 Hue Test (Haustein et al., 1982; Haustein and Schmidt, 1988; Duncker and Krastel, 1990). Clinical examinations show that these patients have decreased visual acuity and central scotomas but no funduscopic changes. ERG analysis revealed reduced rod and cone amplitudes, increased rod and cone implicit times, and elevated rod and cone thresholds (Robertson et al., 1966; Alken and Belz, 1984; Duncker and Krastel, 1990; Madreperla et al., 1994). Taken together, these ophthalmological, behavioral, and electrophysiological findings demonstrate that the photoreceptors are a primary target site of the cardiac glycosides digoxin and digitoxin.
The above results suggest that cone photoreceptors are more susceptible to the effects of cardiac glycosides than rod photoreceptors. Electrophysiological experiments by Madreperla et al. (1994) showed that cones were ∼50 times more sensitive to digoxin and were impaired to a greater degree at the same digoxin concentration than the rods. Following short-duration saturating light flashes, rods appear to recover faster and more completely than cones. However, neither recover to their dark-adapted baseline response level. This latter finding correlates with the slow recovery of the ERG seen in patients following termination of digoxin exposure (Duncker and Krastel, 1990; Madreperla et al., 1994) and is most likely due to the high affinity and slow off-rate of digoxin binding to the cardiac glycoside site located on the extracellular side of the catalytic α-subunit of the Na+,K+-ATPase enzyme (Sweadner, 1989).
Digitalis glycosides, like ouabain, are potent inhibitors of retinal Na+,K+-ATPase (Winkler and Riley, 1977; Fox et al., 1991b; Shulman and Fox, 1996). Digoxin-binding studies show that the retina has the highest number of Na+,K+-ATPase sites of any ocular tissue, even higher than those of brain (Lissner et al., 1971). There are three different isoforms of the α subunit of Na+,K+-ATPase (ie, α1, α2, and α3), and they differ significantly in their sensitivity to cardiac glycoside inhibition (Sweadner, 1989). In the rat retina, the α1-low and α3-high ouabain affinity isoforms of the enzyme account for ≥97% of the Na+,K+-ATPase mRNA. The α3 isoform is localized to rat photoreceptors, horizontal cells, and bipolar cells. Photoreceptors predominantly express the α3 mRNA (approximately 85%), a small amount of α1 mRNA (approximately 15%), and almost no detectable α2 mRNA. Electron microscopic immunocytochemistry studies reveal that the α3 isoform is localized exclusively to the plasma membrane of the rat photoreceptor inner segments (McGrail and Sweadner, 1989; Schneider et al., 1991). The α3 isozyme accounts for most of the rod Na+,K+-ATPase activity (Shulman and Fox, 1996). The rat rod photoreceptor Na+,K+-ATPase specific activity is approximately threefold higher than whole retinal (Fox et al., 1991b; Shulman and Fox, 1996) or whole brain values (Marks and Seeds, 1978). This also is reflected in the two- to threefold greater ouabain-sensitive oxygen consumption in the dark-adapted outer retina relative to the whole or inner retina, respectively (Medrano and Fox, 1995; Shulman and Fox, 1996).
Indomethacin is a nonsteroidal anti-inflammatory drug with analgesic and antipyretic properties that is frequently used for the management of arthritis, gout, and musculoskeletal discomfort. It inhibits prostaglandin synthesis by inhibiting cyclooxygenase. The first cases of indomethacin-induced retinopathy were reported approximately 30 years ago (Bartlett and Jaanus, 2008). Chronic administration of 50 to 200 mg/day of indomethacin for one to two years has been reported to produce corneal opacities, discrete pigment scattering of the RPE perifoveally, paramacular depigmentation, decreases in visual acuity, altered visual fields, increases in the threshold for dark adaptation, blue–yellow color deficits, and decreases in ERG and EOG amplitudes (Burns, 1966, 1968; Henkes et al., 1972; Koliopoulos and Palimeris, 1972). Decreases in the ERG a- and b-wave amplitudes, with larger changes observed under scotopic dark-adapted than light-adapted conditions, have been reported. Upon cessation of drug treatment, the ERG waveforms and color vision changes return to near normal, although the pigmentary changes are irreversible (Burns, 1968; Henkes et al., 1972). The mechanism of retinotoxicity is unknown; however, it appears likely that the RPE is a primary target site.
Sildenafil citrate (Viagra) is a cGMP-specific phosphodiesterase (PDE) type 5 inhibitor that is utilized in the treatment of erectile dysfunction (Corbin et al., 2002). Sildenafil is also a weak cGMP PDE type 6 inhibitor, which is present in rod and cone photorecepotrs (Corbin et al., 2002; Zhang et al., 2005). Transient visual symptoms such as a blue tinge to vision, increased brightness of lights and blurry vision as well as alterations in scotopic and photopic ERGs have been reported following sildenafil usage (Laties and Zrenner, 2002; Jagle et al., 2004). More recently, sildenafil has been associated with the occurrence of nonarteritic anterior ischemic optic neuropathy (NAION) in at-risk patients (ie, those with small cup-to-disc ratios and/or arteriosclerotic risk profiles) within minutes to hours after the ingestion of the drug (Fraunfelder et al., 2006). However, available data suggest that the risk of occurrence of NAION in patients taking sildenafil or tadalafil is not significantly different from the general population (Fraunfelder et al., 2006; Gorkin et al., 2006; Laties, 2009).
Tamoxifen (Nolvadex, Tamoplex), a triphenylethylene derivative, is a nonsteroidal antiestrogenic drug that competes with estrogen for its receptor sites. It is a highly effective antitumor agent used for the treatment of metastatic breast carcinoma in postmenopausal women. Tamoxifen-induced retinopathy following chronic high-dose therapy (180–240 mg/day for approximately two years) was first reported 20 years ago (Kaiser-Kupfer et al., 1981). At this dose, there is widespread axonal degeneration in the macular and perimacular area, as evidenced by the presence of different sized yellow–white refractile opacities in the IPL and NFL observed during fundus examination. Macular edema may or may not be present. Clinical symptoms include a permanent decrease in visual acuity and abnormal visual fields, as the axonal degeneration is irreversible (Ah-Song and Sasco, 1997; Bartlett and Jaanus, 2008). Several prospective studies, with sample sizes ranging from 63 to 303 women with breast cancer, have shown that chronic low-dose tamoxifen (20 mg/day) can result in a small but significant increase in the incidence (≤10%) of keratopathy (Pavlidis et al., 1992; Gorin et al., 1998; Lazzaroni et al., 1998; Noureddin et al., 1999). In addition, these studies showed that retinopathy is much less frequently observed than with high-dose therapy and, except for a few reports of altered color vision and decreased visual acuity, there were no significant alterations in visual function. Following cessation of low-dose tamoxifen therapy, most of the keratopathy and retinal alterations except the corneal opacities and retinopathy were reversible (Pavlidis et al., 1992; Gorin et al., 1998; Noureddin et al., 1999).
Vigabatrin is an inhibitor of GABA-transaminase that is used to treat refractory complex partial seizures and infantile spasms (Tolman and Faulkner, 2009). After 20 years of use in Europe, it was approved in 2009 by the FDA for use in the USA despite its risk of retinopathy characterized by irreversible bilateral, concentric peripheral visual constriction, and decreased retinal nerve fiber thickness (Maguire et al., 2010; Clayton et al., 2011; Plant and Sergott, 2011). Clinical and meta-analysis studies reveal that the prevalence of the asymptomatic visual field loss is ∼50% (Maguire et al., 2010; Clayton et al., 2011; Plant and Sergott, 2011). Onset of the visual field loss is variable as it has been observed as soon as six weeks of exposure, but generally requires a couple of years (Maguire et al., 2010; Clayton et al., 2011; Plant and Sergott, 2011). In addition, rod and cone ERGs as well as flicker responses are altered, indicating that retinal damage also occurs (Daneshvar et al., 1999; Wang et al., 2008). Recent dose–response studies in mice revealed disorganization of the photoreceptors and OPL retinal synaptic following one month of vigabatrin exposure (Wang et al., 2008). The drug is recommended only for epileptic patients with no alternative choices.
Retinotoxicity of Known Neurotoxicants
Inorganic lead is probably the oldest known and most studied environmental toxicant. For almost 100 years, it has been known that overt lead poisoning (mean blood lead [BPb] ≥80 μg/dL) in man produces visual system pathology and overt visual symptoms (Otto and Fox, 1993; Fox, 1998). Clinical manifestations include amblyopia, blindness, optic neuritis or atrophy, peripheral and central scotomas, paralysis of eye muscles, and decreased visual function. Moderate to high level lead exposure produces scotopic and temporal visual system deficits in occupationally exposed factory workers and developmentally lead-exposed monkeys and rats (Bushnell et al., 1977; Betta et al., 1983; Signorino et al., 1983; Campara et al., 1984; Fox and Farber, 1988; Fox et al., 1991a; Fox and Katz, 1992; Otto and Fox, 1993; Lilienthal et al., 1994; Rice, 1998). Early work in monkeys exposed to moderate to high levels of lead during and following gestation reveal that this lead exposure regimen produces irreversible retinal deficits (Lilienthal et al., 1988, 1994; Kohler et al., 1997). A prospective epidemiological study in seven- to 10-year-old children revealed that low-level gestational lead exposure produces long-lasting scotopic supernormal ERG deficits (Rothenberg et al., 2002). Similar results were found in a rodent model of low-level gestational lead exposure (Fox et al., 2008). Recently, a case report with similar findings was published (Nagpal and Brodie, 2009). However, relatively little effort has been made to understand the impact of lead-induced alterations on retinal and central visual information processing on learning and memory in children. These types of visual deficits can adversely affect learning and memory as well as experimental procedures used to assess these cognitive parameters (Anger et al., 1994; Hudnell et al., 1996; Walkowiak et al., 1998; Cestnick and Coltheart, 1999).
Studies in Occupationally Exposed Lead Workers
Clinical and electrophysiological studies in lead-exposed factory workers have assessed both the site of action and extent of injury. Several cases of retrobulbar optic neuritis and optic nerve atrophy have been observed following chronic moderate-level or acute high-level lead exposure (Baghdassarian, 1968; Baloh et al., 1979; Karai et al., 1982). Most of these cases presented with fundus lesions, peripheral or paracentral scotomas whereas the most severe cases also had a central scotoma. Generally, the scotomas were not observed until approximately five years of continuous lead exposure. Interestingly, the earliest observable scotomas were not detected under standard photopic viewing conditions but became evident only under scotopic or mesopic (rod- and cone-mediated) viewing conditions. These ophthalmological findings correlate directly with the ERG data observed in similarly exposed lead workers. No alterations in the critical flicker fusion threshold (ie, temporal resolution) were observed when the test was conducted under photopic conditions or when using red lights. However, consistent decreases in temporal resolution were observed when the test was conducted under scotopic conditions or when green lights were used (Cavelleri et al., 1982; Betta et al., 1983; Signorino et al., 1983; Campara et al., 1984). Moreover, in occupationally lead-exposed workers with or without visual acuity deficits or no observable alterations following ophthalmological examination, the sensitivity, and amplitude of the a-wave and/or b-wave of the dark-adapted ERG were decreased (Scholl and Zrenner, 2000). In other lead-exposed workers, one funduscopic study noted the presence of a grayish lead pigmentary deposit in the area peripheral to the optic disc margins (Sonkin, 1963).
In addition to the retinal deficits, oculomotor deficits occur in chronically lead-exposed workers who have no observable ophthalmological abnormalities. Results from three independent studies, including a follow-up, show that the mean accuracy of saccadic eye movements is lower in lead-exposed workers and the number of overshoots is increased (Baloh et al., 1979; Spivey et al., 1980; Specchio et al., 1981; Glickman et al., 1984). In addition, these studies also revealed that the saccade maximum velocity was decreased. Moreover, one study also observed abnormal smooth pursuit eye movements in lead-exposed workers (Specchio et al., 1981). Although the site and mechanism of action underlying these alterations are unknown, they most likely result from CNS-mediated deficits.
In summary, these results suggest that occupational lead exposure produces concentration- and time-dependent alterations in the retina such that higher levels of lead directly and adversely affect both the retina and optic nerve, whereas lower levels of lead appear to primarily affect the rod photoreceptors and their pathway. Interestingly, these latter clinical findings showing preferential lead-induced rod-selective deficits in sensitivity and temporal resolution are observed in both in vivo and in vitro animal studies (see below). Furthermore, these retinal and oculomotor alterations were, in most cases, correlated with the blood lead levels and occurred in the absence of observable ophthalmological changes, CNS symptoms, and abnormal performance test scores. Thus, these measures of temporal visual function may be among the most sensitive for the early detection of the neurotoxic effects of inorganic lead.
In Vivo and in Vitro Animal Studies with Lead
Lead exposure to adult animals and postnatally developing animals produces retinal damage and functional deficits. The degree and extent of these alterations depends upon the dose, age, and duration of lead exposure. High-level lead exposure to adult rabbits for 60 to 300 days (Hass et al., 1964; Brown, 1974; Hughes and Coogan, 1974) and to newborn rats for 60 days (Santos-Anderson et al., 1984) resulted in focal necrosis of the rod inner and outer segments, necrosis in the INL and Müller cells, and lysosomal inclusions in the RPE. In addition, high-level lead exposure to mice and rats from birth to weaning resulted in hypomyelination of the optic nerve and a reduction in its diameter; but, interestingly, there were no changes in the sciatic nerve (Tennekoon et al., 1979; Toews et al., 1980). Newborn monkeys exposed to high levels of lead for six years had no changes in optic nerve diameter or myelination, although visual cortex neuronal volume and branching were decreased (Reuhl et al., 1989). Rhesus monkeys exposed prenatally and postnatally to moderate or high levels of lead for nine years, followed by almost two years of no lead exposure, had decreased tyrosine hydroxylase immunoreactivity in the large dopaminergic amacrine cells and a complete loss of tyrosine hydroxylase immunoreactivity in small subset of amacrine cells (Kohler et al., 1997). These results suggest that long-term lead exposure produces a decrease in tyrosine hydroxylase synthesis, a finding consistent with other studies (Lasley and Lane, 1988; Jadhav and Ramesh, 1997), and/or a loss of a subset of tyrosine hydroxylase-positive amacrine cells, a finding consistent with recent in vitro work (Scortegagna and Hanbauer, 1997). In contrast to these studies, six weeks of moderate-level lead exposure to adult rats (Fox et al., 1997) and three weeks of low- or moderate-level lead exposure to neonatal rats from birth to weaning produced rod- and bipolar cell-selective apoptotic cell death (Fox and Chu, 1988; Fox et al., 1997). Moreover, recent results reveal that brief (15 minutes) exposure of isolated adult rat retinas to nanomolar to micromolar Pb2+, concentrations regarded as pathophysiologically relevant (Cavalleri et al., 1984; Al-Modhefer et al., 1991), resulted in rod-selective apoptosis (He et al., 2000, 2003). By extension, these results suggest that the triggering event (initiating phase) and the execution phase of rod and bipolar cell death share common underlying biochemical mechanisms.
Results from several studies suggest that an elevated level of rod photoreceptor Ca2+ and/or Pb2+ plays a key role in the process of apoptotic rod cell death in humans and animals during inherited retinal degenerations, retinal diseases and injuries, chemical exposure, and lead exposure. These include patients with retinitis pigmentosa and cancer-associated retinopathy (Thirkill et al., 1987; van Soest et al., 1999), mice with retinal degeneration (rd) (Chang et al., 1993; Fox et al., 1999), rats injected with antirecoverin monoclonal antibodies (Adamus et al., 1998), rats with hypoxic–-ischemic injury (Crosson et al., 1990), rats with light-induced damage (Edward et al., 1991), and lead-exposed rats (Fox and Chu, 1988; Fox et al., 1997, 1999). In addition, moderate level Pb2+ exposure produces apoptotic neuronal cell death in primary cultured cells (Oberto et al., 1996; Scortegagna and Hanbauer, 1997). In vivo and in vitro data suggest that Pb2+ produces a dose (concentration)-dependent inhibition of rod cGMP PDE, a resultant elevation of rod cGMP (Fox and Farber, 1988; Fox et al., 1991a; Srivastava et al., 1995a,b; Fox et al., 1997), which gates the nonselective cation channel of the rod photoreceptor outer segments (Yau and Baylor, 1989), and an elevation of the rod Ca2+ concentration (Fox and Katz, 1992; Medrano and Fox, 1994; He et al., 2000, 2003). Detailed kinetic analysis revealed that picomolar Pb2+ competitively and directly inhibits rod cGMP PDE relative to millimolar concentrations of Mg2+ (Srivastava et al., 1995a,b). In addition, nanomolar Pb2+ can elevate the rod Ca2+ (and Pb2+) concentration via its competitive inhibition of retinal Na+,K+-ATPase relative to MgATP (Fox et al., 1991b). Once inside the rod, both Ca2+ and Pb2+ enter the mitochondria via the ruthenium red-sensitive Ca2+ uniporter and induce mitochondrial depolarization, swelling, and cytochrome c release (He et al., 2000, 2003). The effects of Ca2+ and Pb2+ were additive and blocked completely by the mitochondrial permeability transition pore inhibitor cyclosporin A. Following cytochrome c release, caspase-9 and caspase-3 are sequentially activated. There was no evidence of caspase-8, oxidative stress or lipid peroxidation in this model. These results demonstrate that rod mitochondria are the target site for Ca2+ and Pb2+. This is consistent with numerous studies from different tissues demonstrating that lead is preferentially associated with mitochondria and particularly with the inner membrane and matrix fractions (Barltrop et al., 1971; Bull, 1980; Pounds, 1984). Taken together, the results suggest that Ca2+ and Pb2+ bind to the internal divalent metal binding site of the mitochondrial permeability transition pore (Szabo et al., 1992) and subsequently open it, which initiates the cytochrome c-caspase cascade of apoptosis in rods (He et al., 2000, 2003).
In vitro extracellular and intracellular electrophysiological recordings in isolated whole retinas or photoreceptors reveal that nanomolar to micromolar Pb2+ selectively depress the amplitude and absolute sensitivity of the rod but not cone photoreceptor potential (Fox and Sillman, 1979; Sillman et al., 1982; Tessier-Lavigne et al., 1985; Frumkes and Eysteinsson, 1988). These electrophysiological results are similar to the ERG alterations observed in occupationally lead-exposed workers (Cavelleri et al., 1982; Betta et al., 1983; Signorino et al., 1983; Campara et al., 1984) and in adult rats exposed to low and moderate levels of lead only during development (Fox and Farber, 1988; Fox and Rubinstein, 1989; Fox et al., 1991a; Fox and Katz, 1992). In addition, these postnatally lead-exposed rats exhibit rod-mediated increases in dark and light adaptation time, decreases in critical flicker fusion frequency (ie, temporal resolution), decreases in relative sensitivity, and increases in a- and b-wave latencies (Fox and Farber, 1988; Fox and Rubinstein, 1989; Fox et al., 1991a; Fox and Katz, 1992) and decreases in the temporal response properties of both sustained (X-type) and transient (Y-type) RGCs, such as decreased optimal temporal frequency and temporal resolution (Ruan et al., 1994). By extension, these results suggest that there is a common underlying biochemical mechanism responsible for these rod-mediated deficits. In vivo and in vitro data suggest that lead-induced inhibition of cGMP PDE and resultant elevation of rod Ca2+ underlies the ERG deficits (Fox and Katz, 1992; Medrano and Fox, 1994; Fox et al., 1997; He et al., 2000, 2003). Finally, rod-mediated alterations in dark adaptation and b-wave amplitude are also observed in adult rats and monkeys with prenatal and lifetime moderate- and high-level lead exposure (Hennekes et al., 1987; Lilienthal et al., 1988, 1994). In the gestationally and postnatally lead-exposed monkeys and children, the amplitude of the scotopic b-wave was increased (Lilienthal et al., 1988, 1994; Rothenberg et al., 2002; Nagpal and Brodie, 2009): an effect hypothesized to result from the loss of dopaminergic amacrine cells or their processes (Kohler et al., 1997). If rods and blue-sensitive cones in humans exhibit the same sensitivity to a lead-induced inhibition of cGMP-PDE as they do to the drug-induced inhibition of cGMP-PDE (Zrenner and Gouras, 1979; Zrenner et al., 1982; Fox and Farber, 1988) predicted that blue-cone (short wavelength or S-cones) color vision deficits as well as scotopic deficits may be found in adults and children exposed to lead. S-cone deficits have been observed in an occupationally lead-exposed worker (Scholl and Zrenner, 2000).
Methanol is a low-molecular-weight (32 Da), colorless, and volatile liquid that is widely used as an industrial solvent; a chemical intermediate; a fuel source for picnic stoves, racing cars, and soldering torches; an antifreeze agent; and an octane booster for gasoline. The basic toxicology and references can be found in a thorough review (Eells, 1992). Briefly, methanol is readily and rapidly absorbed from all routes of exposure (dermal, inhalation, and oral), easily crosses all membranes, and thus is uniformly distributed to organs and tissues in direct relation to their water content. Following different routes of exposures, the highest concentrations of methanol are found in the blood, aqueous, and vitreous humors, and bile as well as the brain, kidneys, lungs, and spleen. In the liver, methanol is oxidized sequentially to formaldehyde by alcohol dehydrogenase in human and nonhuman primates or by catalase in rodents and then to formic acid. It is excreted as formic acid in the urine or oxidized further to carbon dioxide and then excreted by the lungs. Formic acid is the toxic metabolite that mediates the metabolic acidosis as well as the retinal and optic nerve toxicity observed in humans, monkeys, and rats with a decreased capacity for folate metabolism (Murray et al., 1991; Eells, 1992; Lee et al., 1994; Garner et al., 1995a, 1995b; Eells et al., 1996; Seme et al., 1999).
Human and nonhuman primates are highly sensitive to methanol-induced neurotoxicity due to their limited capacity to oxidize formic acid. The toxicity occurs in several stages. It first occurs as a mild CNS depression, followed by an asymptomatic 12- to 24-hour latent period, then by a syndrome consisting of formic acidemia, uncompensated metabolic acidosis, ocular and visual toxicity, coma, and possibly death (Eells, 1992). The treatment of methanol poisoning involves both combating acidosis and preventing methanol oxidation, but it is not discussed further here. Experimental rats were made as sensitive to acute methanol exposure as primates by using two different, but related, procedures that effectively reduce the levels of hepatic tetrahydrofolate. One study acutely inhibited methionine synthase and reduced the level of hepatic tetrahydrofolate (Murray et al., 1991; Eells et al., 1996; Seme et al., 1999), whereas the other fed rats a folate-deficient diet for 18 weeks (Lee et al., 1994). Administration of methanol to rats with a decreased capacity for folate metabolism resulted in toxic blood formate concentrations of 8 to 16 mM (Murray et al., 1991; Lee et al., 1994; Garner et al., 1995a,b; Eells et al., 1996; Seme et al., 1999). Permanent visual damage occurred in humans and monkeys when the blood folate levels exceeded 7 mM (Eells, 1992).
Acute methanol poisoning in humans, monkeys, and experimental rats resulted in profound and permanent structural alterations in the retina and optic nerve and visual impairments ranging from blurred vision to decreased visual acuity and light sensitivity to blindness. Ophthalmological studies of exposed humans and monkeys reveal varying degrees of edema of the papillomacular bundle and optic nerve head (Benton and Calhoun, 1952; Potts, 1955; Baumbach et al., 1977; Hayreh et al., 1980). Histopathological and ultrastructural investigations in methanol-exposed monkeys and folate-modified rats showed retinal edema, swollen and degenerated photoreceptors, degenerated RGCs, swollen retinal pigment epithelial cells, axonal (optic nerve) swelling, and mitochondrial swelling and disintegration in each of these cells but especially in the photoreceptors and optic nerve (Baumbach et al., 1977; Hayreh et al., 1980; Murray et al., 1991; Seme et al., 1999). Considering the differences in species, methanol exposures, time course of analysis, and procedures utilized, the overall data for the acute effects of methanol on the ERG are remarkably consistent. Following methanol exposure, the ERG b-wave amplitude in humans, monkeys, and folate-modified rats starts to decrease significantly when the blood formate concentration exceeded 7 mM (Potts, 1955; Ruedeman, 1961; Ingemansson, 1983; Murray et al., 1991; Lee et al., 1994). These ERG b-wave alterations, as well as flicker-evoked ERG alterations (Seme et al., 1999), occur at lower formate concentrations than those associated with structural changes in the retina and optic nerve, as discussed above. Decreases in the a-wave amplitude are delayed, relative to the b-wave and occur when blood formate concentrations further increase (Ruedeman, 1961; Ingemansson, 1983; Murray et al., 1991; Eells et al., 1996). In addition, it has been shown that intraretinal metabolism of methanol is necessary for the formate-mediated alterations in the ERG (Garner et al., 1995a), although intravenous infusion of formate in monkeys does induce optic nerve edema (Martin-Amat et al., 1978). Finally, in the folate-modified rats, it appears that photoreceptors that respond to a 15-Hz flicker/510-nm wavelength mesopic–photopic stimulus (ie, rods and middle wavelength-sensitive [M] cones) are more sensitive to methanol than the ultraviolet-sensitive (UV) cones (Seme et al., 1999).
The retinal sources of the ERG a-wave and b-wave were previously discussed. Thus, the data from the ERG b-wave methanol studies suggest that the initial effect of formate is directly on the ON-type rod bipolar cells, MGCs, and/or synaptic transmission between the photoreceptors and bipolar cells. A well-designed series of pharmacological, ERG, and potassium-induced Müller cell depolarization studies using several controls and folate-modified rats revealed a direct toxic effect of formate on MGC function (Garner et al., 1995a,b). These studies also provided evidence that formate does not directly affect depolarizing rod bipolar cells or synaptic transmission between the photoreceptors and bipolar cells. Formate also appears to directly and adversely affect the rod and cone photoreceptors as evidenced by the markedly decreased ERG a-wave and flicker response data (Ruedeman, 1961; Ingemansson, 1983; Murray et al., 1991; Eells et al., 1996; Seme et al., 1999).
Although there are no direct data on the underlying molecular mechanism responsible for the toxic effects of formate on MGCs and photoreceptors, several findings suggest that the mechanism involves a disruption in oxidative energy metabolism. First, the whole retinal ATP concentration is decreased in folate-deficient rats 48 hours following methanol exposure, the time point when the b-wave was lost (Garner et al., 1995b). Second, both formate (10–200 mM) and formaldehyde (0.5–5 mM) inhibited oxygen consumption in isolated ox retina, and formaldehyde was considerably more potent (Kini and Cooper, 1962). Third, similar concentrations of formaldehyde inhibited oxidative phosphorylation of isolated ox retinal mitochondria, with greater effects observed using FAD-linked than NADH-linked substrates (Kini and Cooper, 1962). Unfortunately, the effects of formate were not examined. Fourth, and consistent with the above results, formate inhibits succinate-cytochrome c reductase and cytochrome oxidase activity (Ki = 5–30 mM), but not NADH-cytochrome c reductase activity in isolated beef heart mitochondria and/or submitochondrial particles (Nicholls, 1976). Fifth, ultrastructural studies reveal swollen mitochondria in rat photoreceptor inner segment and optic nerve 48 to 72 hours after nitrous oxide/methanol exposure (Murray et al., 1991; Seme et al., 1999). To date, there are no such studies conducted on the MGCs. Taken together, these results suggest formate is a mitochondrial poison that inhibits oxidative phosphorylation of photoreceptors, MGCs, and optic nerve. The evidence for this hypothesis and establishment of subsequent steps resulting in retinal and optic nerve cell injury and death remain to be elucidated.
n-Hexane, Perchloroethylene, Styrene, Toluene, Trichloro-ethylene, Xylene, and Mixtures
The neurotoxicity of organic solvents is well established. In addition, exposure to organic solvents and other volatile hydrocarbons are associated with deficits in color vision, contrast sensitivity and visual-motor performance. Similar deficits have been reported after occupational, residential and recreational exposures: the latter among inhalant drug abusers. However, there is a paucity of mechanistic studies on the adverse effects of organic solvents on the retina and visual system.
Loss of color vision (acquired dyschromatopsia) and contrast sensitivity have been reported in workers exposed to organic solvents and related compounds such as alcohols, n-hexane, toluene, trichloroethylene, styrene, xylene, and solvent mixtures (Mergler et al., 1987, 1988, 1991; Iregren et al., 2002; Paramei et al., 2004; Benignus et al., 2005; Attarchi et al., 2010). Workers in microelectronic plants, print shops, and paint manufacturing facilities, and painters who were exposed to concentrations of solvents that exceeded the threshold limit values, had acquired dyschromatopsia as assessed by the Lanthony D-15 desaturated color arrangement panel (Mergler et al., 1987, 1988, 1991; Iregren et al., 2002). These workers had no observable clinical abnormalities as assessed by biomicroscopy, funduscopy, and peripheral visual field tests. The color vision losses were characterized initially as an increase in blue–yellow confusion errors, although more severe red–green deficits were reported with extended duration or higher concentrations of exposure. As a general rule, acquired blue–yellow losses may result from lens opacification or outer retinal alterations, whereas red–green losses are traditionally associated with retrobular, or central visual pathway alterations (Porkony et al., 1979). In addition, many of the occupationally exposed workers also exhibited lower contrast sensitivity at intermediate spatial frequencies, which likely reflects alterations in neural function (Mergler et al., 1991; Broadwell et al., 1995). The data from Mergler et al. (1987, 1988, 1991) appear to show gender differences in these adverse visual effects. A study of female workers, where the Lanthony D-15 desaturated test was used to assess color vision, showed a trend toward increased prevalence of color vision impairment following exposure to low to moderate concentrations of toluene (Zavalic et al., 1996). The neuropathic compounds, n-hexane, its metabolite 2,5-hexanedione and carbon disulfide have been associated with macular changes (Raitta et al., 1974, 1978, 1981) and rod and cone degeneration (Backstrom and Collins, 1992; Backstrom et al., 1993). Photoreceptor damage and retinal pathology have not been reported for other hydrocarbon solvents. Clearly more detailed, well-designed, and well-executed studies are needed to determine the (1) specific solvents that cause alterations in color vision, (2) vulnerability of spatial and temporal contrast sensitivity, (3) dose (concentration)–response relations between exposures and effects, (4) possible gender differences, (5) potential reversibility of deficits if exposure is terminated, and (6) pathophysiological basis for these changes. Prospective occupational studies would be particularly helpful given the difficulties in obtaining appropriately matched control populations in cross-sectional study designs.
Deficits in visual function such as contrast sensitivity have been observed in residents of neighborhoods containing dry cleaners using perchloroethylene (Altmann et al., 1995) and in residents of apartment buildings with co-located dry cleaners (Schreiber et al., 2002). For both of these residential studies, the atmospheric concentrations of perchloroethylene were below those typical of occupational settings. Laboratory experiments with human subjects exposed to perchloroethylene for four hours for four days revealed increased peak latency delays in the N75, P100, and N150 of the VEP as well as decreases in contrast sensitivity at low and intermediate spatial frequencies (Altmann et al., 1990). Acute perchloroethylene inhalation exposure also reduced pattern VEP amplitudes in rats (Boyes et al., 2009). Moreover, gestational exposure to perchloroethylene produced clinical red–green color loss and decreased visual acuity in the children of occupationally exposed mothers (Till et al., 2003). Recently, a study by the New York State Department of Health reported mild, but statistically significant, deficits of contrast sensitivity in children living in apartment buildings with co-located perchloroethylene drycleaners in comparison to residents of control buildings without drycleaners (Storm et al., 2011). The above results reveal that perchloroethylene is toxic to both the developing and adult visual system.
Six independent studies report that workers exposed to mean atmospheric concentrations of styrene ranging from 20 to 70 ppm exhibit concentration-dependent alterations in color vision (Gobba et al., 1991; Fallas et al., 1992; Chia et al., 1994; Eguchi et al., 1995; Campagna et al., 1995; Iregren et al., 2005). A combined data analysis from two of the above studies (Gobba et al., 1991; Campagna et al., 1995) suggests that the threshold for color visual impairments is ≤4 ppm styrene (Campagna et al., 1996). This was well below the threshold limit value-time-weighted average (TLV–TWA) value for any country: range 20 to 50 ppm. The similarity of the styrene-induced blue–yellow color vision deficits observed by five independent groups demonstrates the reproducibility of these color vision deficits. In addition, two meta-analysis studies of solvent exposure and color vision are consistent with these results. Paramei et al. (2004) evaluated the color confusion index (CCI) scores from 15 studies of workers exposed to toluene, styrene, or mixed solvents. They concluded that while 13 of the original 15 studies reported CCI values indicative of impaired performance in the solvent-exposed group, the large variations among the effect sizes across studies obscured an overall statistically significant association between exposure and CCI values for all the substances except styrene, which was significantly associated with impaired color discrimination. Benignus et al. (2005) focused on only reports of styrene exposure from the six independent studies and also observed a statistically significant relationship between cumulative styrene exposure and increased CCI scores. The slope of the function observed suggested that the magnitude of decrement in CCI scores for a worker exposed to 20 ppm styrene for eight years would be equivalent to an additional 1.7 years of aging (Benignus et al., 2005). The reversibility of these impairments has not been thoroughly studied, although in one study no recovery was found after a one-month period of no exposure (Gobba et al., 1991). Rats exposed to styrene for 13 weeks showed lower retinal dopamine content and fewer tyrosine hydroxylase immunoreactive retinal amacrine cells (Vettori et al., 2000). In summary, there is a concordance of evidence that styrene exposure is associated with retinal toxicity in experimental animals and color vision deficits in occupationally exposed workers.
Toluene is one of the most widely used substances in industry and commerce. It is used as a solvent, degreaser, constituent of products such as paints and glues, and is a substantial component of gasoline and other fuels. Among the neurotoxic consequences of occupational toluene exposure are impaired color vision (Campagna et al., 2001; Cavalleri et al., 2000) and reduced visual contrast sensitivity (Donoghue et al., 1995). Acute toluene exposure impaired oculomotor function (Niklasson et al., 1995) and reduced pattern VEP amplitudes in proportion to estimated concentration of toluene in the brain (Boyes et al., 2007). Moreover, toluene has a relatively low irritancy and a high euphoric potential making it a favored selection among inhalant drug abusers; many are young and expose themselves repeatedly to very high toluene concentrations. Abuse of toluene is associated with poor visual acuity, altered or unrecordable pattern VEPs, optic neuropathy, abnormal MRI signals in visual cortex, decreased perfusion of thalamus and cerebral cortex, and visual hallucinations (Ryu et al., 1998; Kamran and Bakshi, 1998; Kiyokawa et al., 2005).
The neurotoxicity of organophosphates is well established (see Chap. 16); however, the link between organophosphate exposure and retinotoxicity is unresolved. Clinical studies conducted in Japan, report on ocular toxicity from laboratory animals exposed to organophosphates, and reports to the EPA by pesticide manufacturers suggest that various organophosphates produce retinotoxicity and chronic ocular damage (Ishikawa, 1973; Dementi, 1994). However, many of the early clinical reports were poorly designed and remain unconfirmed. The evidence for organophosphate-induced retinal toxicity is strongest for fenthion (dimethyl 3-methyl-4-methylthiophenyl phosphorothionate) (Imai et al., 1983; Misra et al., 1985; Boyes et al., 1994; Tandon et al., 1994). Two epidemiological studies of licensed pesticide applicators and their spouses did not find a statistically increased risk of retinal degeneration from use of organophosphate insecticides as a class, but risks were increased for some individual organophosphate chemicals (Kamel et al., 2000; Kirrane et al., 2005). Interestingly, both studies identified an increased risk of retinal degeneration in individuals exposed to fungicides. There were also reports that Japanese children exposed to organophosphates had a high incidence of myopia (Ko et al., 1988). Experimentally, the visual control of ocular growth, which is cholinergically mediated, was impaired in the eyes of chicks exposed to the organophosphate insecticide chlorpyrifos (Geller et al., 1998). Embryonic chick retinal cells did not develop normally when exposed to diazinon in vitro (Paraoanu et al., 2006). Currently, the mechanisms of ocular toxicity, sites of action, and whether the effects are restricted to some specific organophosphates such as fenthion or are more general to the chemical class are unknown. The use of organophosphate insecticides has been restricted (but not eliminated) in North America and Europe in recent years, but continues in much of the developing world. Thus, it is important to resolve the potential risks of ocular toxicity for this class of agents.