Nyctalopia

Difficulty with vision at night (night blindness) is one of the two hallmark symptoms of RP. Patients with typical RP usually attribute the beginning of night vision difficulties to the first or second decade of life. In some patients, especially those living in an urban setting, night vision problems may not be apparent until ocular disease is at an advanced stage. Tanino and Ohba¹² noted that the onset of symptoms of RP, most commonly nyctalopia, occurred at a median age of 10.7 years in autosomal recessive disease and 23.4 years in autosomal dominant disease. The symptom of nyctalopia should not be confused with the symptom of blurred vision with night myopia or uncorrected refractive error. People with RP have a narrowing of the visual field in the dark and may describe getting easily disoriented on dimly lit evenings when others are able to see adequately. Becoming accident-prone, especially at night, is a highly suggestive symptom.

Night blindness is not pathognomonic of RP and can be a feature of other retinal disorders, such as congenital stationary night blindness and age-related macular degeneration. Deteriorating night vision can also be a prominent feature of other ocular disease, such as high myopia.

Visual field loss

The second hallmark feature or symptom of RP is an insidious, progressive loss of peripheral visual field. In some patients, especially those with severe disease beginning in childhood, this may be manifested as a progressive contraction of the visual field (see also Perimetry under Psychophysical findings). Loss of peripheral vision can be detected in early disease with small, dim test targets (Fig. 40.1). For many types of RP, field deficits are usually found first, and are most severe, in the superior visual field (Fig. 40.2). This reflects the early involvement of the inferior retina in RP. Occasionally the visual field loss will be greater temporally, nasally, or inferiorly (Fig. 40.3). See the section on Perimetry for specifics about testing the visual field.

Fig. 40.1 (A) Goldmann kinetic visual fields in a 27-year-old woman with autosomal dominant retinitis pigmentosa. Note that the visual field is full to the IV-4-e test target but that isopters are contracted for smaller and dimmer test targets. No scotomas were plotted. (In this and all subsequent kinetic perimetric visual fields, although not specified on the illustration, all isopters were measured with the e setting.) (B) Octopus static perimetric fields for the same patient (combined programs 23 and 34). Note the overall decrease in retinal sensitivity in the mid and far periphery, with patchy areas of greater loss of sensitivity. The electroretinogram was profoundly abnormal with no detectable rod-mediated responses and very small (10–15 µV) prolonged residual cone responses. The 40-minute dark-adapted retinal threshold was elevated 2.8 log units above the normal mean.

Fig. 40.2 Goldmann kinetic perimetry showing visual field loss, which is greatest in the upper field, in a 27-year-old man with high myopia and X-linked retinitis pigmentosa. Note that the smaller test target isopters are contracted inferiorly as well, suggesting relative decreased retinal sensitivity in the inferior fields. The electroretinogram (ERG) showed subnormal, prolonged rod responses with even more subnormal, prolonged cone responses, suggesting a cone–rod ERG pattern.

Fig. 40.3 Goldmann kinetic perimetry for an 18-year-old man with autosomal recessive retinitis pigmentosa associated with posterior column ataxia (same patient as Fig. 40.15A and brother of patient in Fig. 40.21). Visual acuity was 20/30, J1, right eye, and 20/50, J2, left eye.

In a group of patients with several types of RP, Berson et al.¹³ found that overall about 4.6% of the remaining visual field was lost per year. Massof et al.^14,15 found that, in most forms of generalized RP, the kinetic visual field shrank approximately 50% over 4.5 years. Massof and Finkelstein¹⁵ argue that the time course for field loss in all major genetic and functional classes of RP is nearly identical if one corrects for the critical age, which is defined as the age when visual field loss is first detected by a test target of given size and brightness. In conjunction with this concept, they have proposed a two-stage hypothesis for the natural course of RP. The first stage is a predisposition for retinal degeneration, such as genetic or environmental factors, which would be different for the separate types of RP and possibly among individuals. The second stage involves the actual retinal degeneration, which proceeds on its own internal time course and which would be common to most forms of RP.

In general, there is a strong tendency for the visual field loss to be symmetric between the two eyes.^14,15 A notable exception to this is the phenotype expressed in female carriers of X-linked RP. In this situation, the distribution of mutant photoreceptors in the retina is determined by lyonization (X-chromosomal inactivation). This is a random event determining which genes of the two X chromosomes (the normal or mutant copy) are expressed in a particular cell. This leads not only to unusual, irregular patterns of visual field loss in individual eyes but also to quite striking differences in field loss between the two eyes.¹⁶ However, even in these cases the visual fields usually correspond quite closely to the fundus appearance, with the greater visual field loss in the eye with greater pigmentary abnormalities.

In typical RP the rate of progression of visual field loss is usually slow and relentless, with changes best appreciated when observed over many years or even decades. However, the visual fields may change dramatically over a few months or years. A patient may not notice what may be a striking loss of peripheral visual field if the central field remains clear. As the visual field reaches the stage of “tunnel vision,” the patient usually becomes acutely aware of subsequent change with time. This often leads the patient to the conclusion that the rate of degeneration is accelerating. In the USA, statutory legal blindness results when the remaining central visual field measures 20° or less horizontally in the better eye, using the III-4-e test target on a Goldmann perimeter.¹⁷ In the UK, most RP patients, with normal acuity, would be registered as partially sighted (“gross field defect”). Legal blindness occurs when fields become “very contracted” – a stage at which acuity loss has also usually occurred (www.dh.gov.uk/assetRoot/04/07/48/48/04074848.doc).

Central vision loss

There is a common misconception in RP (documented in older texts) that the central vision will remain good until most, if not all, the peripheral field is lost. Central visual function can, however, be seriously affected early in typical RP, while significant peripheral field remains.¹⁸ Cystoid macular edema (CME),¹⁹ diffuse retinal vascular leakage,²⁰ macular preretinal fibrosis²¹ (Fig. 40.4), and retinal pigment epithelial (RPE) defects in the macula or fovea^22,23 can occur early in the disease. Macular edema is the topic of a section on treatment of RP later in this chapter.

Fig. 40.4 Posterior pole of the left eye of a 33-year-old man with Bardet–Biedl syndrome from homozygosity for the common Met390Arg mutation of BBS1, demonstrating preretinal fibrosis with wrinkling of the internal limiting membrane. The acuity was 20/400. The maculae developed atrophic lesions over the next 5 years with further reduction of visual acuity.

To a considerable extent, the likelihood of retaining “good vision” (central acuity) to a given age in life depends on the specific inheritance type of RP. Patients with regional (“sector”) RP may retain good visual acuity all their life. Patients with autosomal dominant RP (adRP) are more likely than patients with autosomal recessive (arRP) or X-linked RP to retain good acuity beyond 60 years of age.¹³ Marmor, in his study of visual loss in adRP and arRP,²⁶ found that when visual acuity began to fail, it generally dropped to 20/200 within 4–10 (median 6) years. Patients with X-linked RP are usually blind (acuity 20/200 or worse) by 30–40 years of age.^13,24,25 A good prognostic sign of retention of central acuity is electroretinograph (ERG) amplitudes remaining quite large (e.g., greater than 100 µV).²⁶

Color vision defects

In general, color vision in typical RP remains good until the visual acuity is 20/40 or worse. Color vision may fail early in cases where central cones appear to be abnormal from the beginning. In such cases, pericentral scotomas encroach very close to fixation early in the disease. These patients, even though they show a rod–cone loss on ERG, may show impressive color vision defects even when peripheral vision is relatively good.

Photopsia and other symptoms

Many patients with RP at some time during the course of their disease have light flashes, or photopsias. These are reported as occurring in the midperipheral field of vision, often adjacent to areas of relative or absolute scotomas. These photopsias are described as tiny, blinking or shimmering lights or as a coarse, sparkling graininess to vision. The phenomenon is similar to that reported by patients with ophthalmic migraine except that, although retinal disease involvement expands over the years, the photopsias are generally stationary within the field. Also, unlike ophthalmic migraine, the photopsias may be continuous rather than episodic. As the scotomas become denser over the years, the photopsias decrease and finally disappear. In a retrospective survey of symptoms and findings in 500 patients with RP, Heckenlively et al.²⁷ reported light flashes in 170 (35%). Since 8 patients in this series had retinal detachments, the symptom of light flashes must be considered an indication for careful fundus examination.

The cellular or tissue correlates that underlie photopsias in RP are unknown but may include photoreceptor dysfunction, neurite sprouting, aberrant synapse formation, and secondary remodeling of the retina, all of which occur as sequelae of photoreceptor degeneration (see section on Cell and tissue biology: Histopathology, for further discussion).

Fundus appearance

The classically described fundus appearance of RP includes attenuated retinal vessels, mottling and granularity of the RPE, bone spicule intraretinal pigmentation, and optic nerve head pallor (Fig. 40.5). In general, when ophthalmoscopically detectable abnormalities of the fundus are present, there is a high degree of symmetry between the two eyes and fundus pigmentation is often greater in the midperiphery (Fig. 40.6). In advanced disease, atrophy of the RPE and choriocapillaris leads to fundus pallor and larger choroidal vessels become visible (Fig. 40.7). In some forms of RP and allied disorders, sufficient atrophy of the retina and choroid will occur during the late stages of the disease that the fundus appearance overlaps with certain of the primary choroidal atrophies (Fig. 40.7).²⁸ As the disease advances to late stages, vessel attenuation can become so severe that the retinal vessels appear thread-like (Fig. 40.8).

Fig. 40.5 (A) Fundus appearance of the left eye of a 13-year-old girl with autosomal recessive retinitis pigmentosa. Visual acuity was 20/20 but visual field was constricted to 20° diameter with the III-4-e test target. (B) Advanced autosomal dominant retinitis pigmentosa in the left eye of a 57-year-old man from the Pro347Leu mutation of rhodopsin. Visual acuity was 20/100 right eye and 20/200 left eye and visual field constricted to 18° in both eyes with the III-4-e target.

Fig. 40.6 Fundus appearance of the right eye of a 40-year old patient with rhodopsin Pro23His retinitis pigmentosa. Areas of fundus changes correlated well with kinetic visual fields, performed at 39 years of age, as shown in Figure 40.19.

Fig. 40.7 Fundus appearance of the left eye of a 75-year-old patient with advanced rhodopsin Pro23His retinitis pigmentosa. This patient is the father of the patient whose visual field is shown in Figure 40.19 and whose fundus appearance is shown in Figure 40.6.

Fig. 40.8 Thread-like retinal vessels, pale, waxy optic discs, epiretinal fibrosis, and mottling of the retinal pigment epithelium in the left eye of a 10-year-old girl with advanced retinal degeneration in Alström disease.

(Reproduced with permission from Millay RH, Weleber RG, Heckenlively JR. Ophthalmologic and systemic manifestations of Alström’s disease. Am J Ophthalmol 1986;102:482–90.)

Almost all forms of RP go through a stage where the retina appears either normal or nearly normal, even though relative scotomas may be evident on visual field testing or areas of decreased retinal threshold may be present on static perimetry. Patients who have very early RP without fundus pigmentary abnormalities are often diagnosed as having RP sine pigmento or paucipigmentary RP. This is no longer considered a specific subtype of RP but a stage through which many, if not most, patients with RP pass. The sine pigmento stage may exist for decades before typical RP signs appear. For patients with high myopia and RP, the typical fundus features of high myopia may delay the appreciation of other retinal abnormalities (Fig. 40.9).

Fig. 40.9 Fundus of a 27-year-old man with high myopia and X-linked retinitis pigmentosa (same patient as in Fig. 40.2). Sparse but definite bone spicules were present in the inferior mid and far periphery. Discs showed a myopic tilt with peripapillary chorioretinal atrophy. The retinal pigment epithelium was diffusely hypopigmented.

When fundus abnormalities are evident, the earliest features are attenuation of retinal vessels and the appearance of fine mottling or granularity of the RPE in the mid and far periphery (Fig. 40.10). Spectral-domain optical coherence tomography (SD-OCT) imaging demonstrates that retinal thickness is decreased, particularly for outer layers, outside the macula and in areas of abnormal pigmentation with bone spicules.^29,30 The macular region may exhibit increased luster, abnormal highlights, or wrinkling, suggesting either macular edema or early epiretinal membrane formation or preretinal fibrosis.^19,22 In advanced stages, patients with RP may exhibit atrophic macular lesions (Fig. 40.11) that simulate what some authors have called (perhaps inappropriately so) “macular coloboma.”

Fig. 40.10 Fundus findings of early simplex retinitis pigmentosa at age 7 years (A), simplex retinitis pigmentosa at age 9 years (B), and autosomal dominant retinitis pigmentosa at age 14 years (C). Note midperipheral pigment dispersion in A and C and wrinkling of internal limiting membrane in B.

Fig. 40.11 Early-phase fluorescein angiogram frame of the left eye of a 44-year-old woman with autosomal dominant retinitis pigmentosa and bilateral atrophic macular lesions. Visual acuity was 20/300 right eye and 20/200 left eye.

Intraretinal, bone spicule pigment formations represent migration of pigment into the retina from disintegration of RPE cells with accumulation in the interstitial spaces surrounding retinal vessels. This process occurs most prominently at the junctions of vessels producing perivascular pigmentary cuffing and spicule-shaped deposits. The loss of pigment within the pigment epithelial cells often produces an overall gray, desaturated appearance to the retina with greater visibility of underlying choroidal vessels through the more transparent pigment epithelium. Eventually, the normal salmon-pink color of the entire mid and far peripheral fundus is replaced by dense bone spicule pigmentary formations. Heckenlively³¹ has described several patients with advanced RP with preservation of the RPE in a para-arteriolar distribution. Porta and colleagues³² described two additional cases. Van den Born et al.³³ reported a large consanguineous family with preserved periarteriolar RP and confirmed the autosomal recessive inheritance. Almost all these patients were hyperopic. The gene for this form of RP (termed RP12) has been found to be the human homolog CRB1 of the Drosophila crumbs gene.

In some patients abnormal pigmentation and atrophy are confined to one area of the retina. Such an appearance is termed “sector RP,” a diagnosis that undoubtedly represents a heterogeneous group of disorders characterized by marked regionality of disease. Sector RP is discussed in further detail in the section Subdivision by distribution of retinal involvement or fundus appearance, below. However, since many patients with generalized disease present with localized pigmentary changes with regional functional loss, sector RP must always remain a provisioinal diagnosis until an adequate period of observation has occurred.

A yellowish-white metallic tapetal-like reflex or sheen can occasionally be observed in women who are carriers for the X-linked form of RP³⁴ (Fig. 40.12) and rarely in young males with early disease^35–37 (Fig. 40.13). With time, regional or diffuse mottling and atrophy of the RPE eventually replace the tapetal-like sheen. At least one form of dominant RP can show a tapetal-like reflex limited to the foveal and parafoveal region (Fig. 40.14). No histopathologic assessment has yet identified the source of a tapetal-like reflex. Cideciyan and Jacobson³⁸ suggested that abnormal cones are the sources of the tapetal reflex. Berendschot et al.,³⁹ using spectral fundus reflectance, confirmed that outer segments of peripheral photoreceptor cells are the source of the abnormal reflex but that this abnormality was not confined to cones. The carrier state for X-linked RP appears to be a slowly progressive retinal dystrophy in its own right, with a patchy distribution of affected retina as determined by lyonization.^40,41 A number of studies have reported that the ERG is abnormal in most X-linked RP carriers.^16,42–45 A tapetal-like reflex is not pathognomonic of X-linked RP but has also been reported in occasional cases of X-linked retinoschisis,⁴⁶ and Oguchi disease.⁴⁷

Fig. 40.12 Golden granular tapetal-like sheen or reflex in 20-year-old (A) and 44-year-old (B) carriers for X-linked retinitis pigmentosa.

Fig. 40.13 Tapetal-like yellow-white reflex in a 13-year-old boy with X-linked retinitis pigmentosa.

Fig. 40.14 Tapetal-like golden foveal reflex in an 11-year-old boy with autosomal dominant retinitis pigmentosa. Left (A) and right (B) stereo pair of right macula. In these and all stereo pairs for both eyes, the bright sheen was only visible in the photograph obtained through the nasal half of the pupil. This marked orientation dependency is similar to that seen in crystalline structures and may result from deformations of the outer photoreceptor segments.

The optic disc may be normal in early RP, show a waxy fullness with hyperemia, or appear waxy and pale (Fig. 40.15A). A “golden ring” or yellowish-white halo can often be seen surrounding the optic disc in early RP (Fig. 40.15B). As disease progresses, this golden ring is replaced with peripapillary mottling, hyperpigmentation, and atrophy of the RPE (Fig. 40.15C). The optic nerve head cup-to-disc ratio has been reported to be significantly smaller in RP patients of all types (0.19 compared to 0.35 in normal subjects).⁴⁸ In advanced disease, dense optic nerve head pallor results, in part from optic atrophy and in part from gliosis overlying the discs.²¹ Drusen-like globular excrescences may develop on the optic nerve or adjacent retina (Fig. 40.16). These have been mistaken for astrocytic hamartomas⁴⁹ and are consistent with disc drusen due to aberrant axoplasmic transport.^50,51 In a survey of 117 RP patients,⁵² 10% were found to have clinically apparent optic nerve head drusen. Edwards et al.⁵³ suggested that this phenomenon is particularly common in RP associated with Usher syndrome type 1.

Fig. 40.15 Optic discs in an 18-year-old man with autosomal recessive retinitis pigmentosa as part of posterior-column ataxia and retinitis pigmentosa syndrome (A) (same patient as in Fig. 40.3), a 32-year-old man (B), and his 55-year-old father (C), both of whom have autosomal dominant retinitis pigmentosa. The disc in A shows peripapillary chorioretinal pallor, or the “golden ring” sign, that is characteristic of early retinitis pigmentosa. Note that the disc is nearly normal in the son (B), but shows severe waxy pallor in the father (C).

Fig. 40.16 Optic discs in the right eye of a patient with type 1 Usher syndrome (retinitis pigmentosa and prelingual deafness) at age 16 (A) and 25 (B), demonstrating development of peripapillary drusen.

Flynn et al.⁵⁴ found that the presence or absence of macular RPE defects is of significant prognostic importance with regard to retention of visual acuity over the next 5 years. Specifically, the absence of macular lesion was associated with only one line of acuity loss over the 5-year period, whereas bull’s-eye or geographic atrophy lesions were associated with a predicted acuity loss of three to four lines. The authors emphasize the importance of this type of information for prognostic counseling of patients with RP regarding the likelihood of preservation of visual acuity. Other maculopathies associated with RP include cellophane maculopathy, macular hole (complications of CME),⁵⁵ and a single report of RP with central serous retinopathy.⁵⁶

Vitreous abnormalities

The most common abnormality of the vitreous in RP is the presence of fine, dust-like pigmented cells released from degeneration of the RPE. In patients with RP, complete posterior detachment of the vitreous, “cotton-ball” opacities, interwoven filaments in the retrocortical space, and spindle-shaped vitreous condensations are seen more frequently than in normal subjects.^57,58 A single case is reported of vitreous opacity so dense that vitrectomy was required.⁵⁹ Vitreous abnormalities related to peripheral retinal ischemia and preretinal neovascularization have been reported in rare cases of adRP and arRP.⁶⁰ Uncommonly, a peripheral Coats-like retinal vasculopathy with lipid exudations and serous retinal elevation can be seen in RP^61–63 (Fig. 40.17). Unlike idiopathic Coats syndrome, the retinal vasculopathy seen in RP is usually bilateral, shows no sex predilection, and typically occurs in older patients. Coats-like disease is seen in 3.6% of RP cases⁶⁴ and has been reported with adRP,^62,63 arRP³³ but not X-linked inheritance⁶³ (although see Fig. 40.17). Coats-like changes that occur in children^64,65 may change over a relatively short period of months to years, must be monitored carefully, and may need to be considered for laser or cryocoagulation. Vitreous hemorrhage related to preretinal neovascularization has also been described in rare cases.^60,66 Retinal edema in some cases may be caused by inflammation resulting from the degeneration of outer retina.^60,67

Fig. 40.17 Coats disease in the right eye (A, temporal retina; B, posterior pole; C and D, inferotemporal retina) of a 33-year-old man with retinitis pigmentosa that is probably X-linked. Note also peripapillary retinal drusen nasal to the disc. Visual acuity was 20/30−, J8, right eye, and 20/25−, J3, left eye.

Anterior-segment abnormalities

Cataracts are frequent anterior-segment complications of RP. The prevalence of cataracts in RP patients aged 20–39 years varied with the inheritance type, being roughly 52% for autosomal dominant, 39% for recessive, and 72% for X-linked cases.⁶⁸ The incidence of cataracts in RP is highly age-dependent.^68,69 The most frequent type of cataract is a posterior subcapsular lens opacity, which occurs in 35–51% of patients.^69,70 One study has suggested that posterior subcapsular opacities are most commonly associated with autosomal dominant inheritance (51%),^69,70 whereas another has suggested a more common association with X-linked inheritance.⁷¹ Although keratoconus has been claimed to be more frequent among patients with RP,⁷² the occurrence of keratoconus in typical juvenile- or adult-onset RP is extremely rare in our experience. Glaucoma, especially primary open angle glaucoma, is also allegedly more frequent in RP.⁷³ Sector RP appears to be associated with hyperopia and angle closure glaucoma.⁷⁴ Because of the other abnormalities of the disc and fundus and because of the already abnormal visual fields, the diagnosis and management of glaucoma are quite challenging in RP patients.

Refractive status

High myopia and astigmatism are frequently associated with RP.^68,75

Perimetry

Assessment of visual field is an important aspect of the management of RP. This is an invaluable tool in the diagnosis of RP, as well as being the most reliable method of quantifying real change in visual deficit. It should be remembered when assessing the results of visual field testing that the reproducibility of visual fields is significantly less in patients with RP than in normal subjects. Kinetic visual fields tested on separate days by the same examiner varied by 5–6% in normal subjects but by 11–13% in patients with RP (intraobserver variation); variability between examiners for the same patient was 4–6% for normal subjects but 10–16% for patients with RP (interobserver variation).⁷⁶

In the great majority of cases of RP, the earliest defects of the visual field, on kinetic perimetry, are relative scotomas in the midperiphery, between 30 and 50° from fixation (Fig. 40.18). These enlarge, deepen, and coalesce to form a ring of visual field loss. These midperipheral depressions of retinal sensitivity are readily documented as decreased retinal thresholds on static perimetry. As ring scotomas enlarge toward the far periphery, islands of relatively normal visual field remain, usually temporal but occasionally inferior or nasal (Fig. 40.19). The peripheral islands or remnants of visual field are often lost before the central field contracts sufficiently to qualify the individual as legally blind.

Fig. 40.18 Goldmann kinetic perimetry for a patient with autosomal dominant retinitis pigmentosa from the Pro23His mutation of rhodopsin at age 41 (A) and 47 years (B). (C) Octopus static (combined programs 23 and 34) perimetry using stimulus size III for the patient at age 47. Note enlargement and deepening of midperipheral scotomas and overall correlation of static visual field with detailed kinetic perimetry.

Fig. 40.19 Goldmann kinetic perimetry for a 36-year-old patient with retinitis pigmentosa from the Pro23His mutation of rhodopsin. Dense ring scotomas have broken through to the periphery superiorly, leaving a central field of 25° in diameter and only an incomplete annulus of peripheral vision in each eye. The fundus appearance of this patient at 40 years of age is shown in Figure 40.6.

Kinetic visual field testing has traditionally been performed with the manual Goldmann perimeter. Recent advances have occurred in both kinetic and static perimetry. Semiautomated kinetic perimetry,⁷⁷ as implemented on the Octopus 101 and 900 perimeters (Haag-Streit, Köniz, Switzerland) (Fig. 40.20) automatically documents the testing and enables quantification of field by measurement of isopter area. A new fast thresholding algorithm, known as German adaptive thresholding estimation (GATE),⁷⁸ has enabled another advance in perimetry, that of full-field static perimetry to assess the entire visual field in a time frame comparable to that required to test the central 30° field with the Standard 4–2-1 strategy (Fig. 40.20) (Weleber et al., 2012, in preparation). The static threshold data from such testing can be modeled as sensitivity values into a three-dimensional hill of vision from which the magnitude and extent of the total visual field can be determined as a volumetric measurement (Weleber et al. 2011, in preparation).

Fig. 40.20 Semiautomated kinetic perimetry in a 27-year-old woman with retinitis pigmentosa (top). The shaded areas depict the normal range for the isopters for the test targets presented. Three-dimensional (3D) model of the hill of vision for full-field static perimetry for this same subject (middle in tilted view, below in face view) using the size V test target and the German adaptive thresholding estimation thresholding algorithm.⁷⁸ The volume of this model, which was 40.2 dB-sr OD and 28.0 DB-sr OS (compared to a normal 112.5 dB-sr for a normal), reflects the magnitude and extent of the visual field for this subject. Thus, this patient has 35.7% OD and 24.9% OS of expected field volume for his age. Note that the 3D model provides much more topographical information on the location of the defects in the field of vision than does the kinetic perimetry. This is the same patient whose fundus autofluorescence image for the left eye is shown in Figure 40.29A.

If an RP patient is driving a motor vehicle, regular visual field assessments are mandatory and should be undertaken at least every 2 years. Most centers perform kinetic visual fields using a Goldmann perimeter for clinical evaluation and for assessment for driving and for disability. Almost all patients with RP have to restrict their driving at night and eventually stop driving altogether. Periodic assessment of full-field kinetic perimetry helps to provide patients with knowledge of their visual limitations and when to restrict and eventually stop driving.

Two-color static perimetry has been shown to be a useful tool in assessing the dark-adapted visual fields of patients with RP. Using this technique, Massof and Finkelstein^79–81 found that their patients with RP could be divided into two groups: (1) type I RP, which is associated with early diffuse loss of rod sensitivity relative to cone sensitivity and childhood-onset nyctalopia; and (2) type II RP, associated with regional and combined loss of both rod and cone retinal sensitivity and adult-onset nyctalopia. No patients have been observed to change from one type to the other.⁸² No families have been reported that contain both types I and II in the same sibship.¹⁵ An excess of type II RP over the type I form has been reported in simplex disease. Analysis of subgroups within simplex and multiplex RP suggest that a sizable proportion of type II patients may represent sporadic or acquired disease, and not true RP.⁸³

Arden et al.⁸⁴ studied a series of cases of adRP both with ERGs and psychophysically with scotopic static thresholds. They found that the preservation of sizable rod responses on the ERG correlated perfectly with the Massof type II classification. Absence of rod ERG did not always indicate type I disease, and both types were seen in this group. Lyness et al.⁸⁵ reported clinical, psychophysical, and ERG findings in 104 patients (from 44 families) with adRP. Subgroup D (13 patients in four families) had diffuse loss of rod function and night blindness before the age of 10. Subgroup R (28 patients in 13 families) had regional loss of rod function, and most of these patients were unaware of night blindness until after the age of 20. Both groups had regional loss of cone functions. In the R group the rod-mediated ERG was usually present and substantial, whereas in the D group the rod ERG was undetectable. Although not identical, the D group bears similarities to Massof type I, and the R group bears similarities to type II. These studies suggest that the two types or groups may represent different pathophysiologic subtypes of RP. It should be emphasized, however, that a large proportion of patients (27 of the 44 families studied by Lyness et al.⁸⁵) could not be incorporated into the D- versus R-type classification, partly because many patients with RP have almost undetectable ERG responses at presentation, which is quite common.

The techniques of two-color scotopic static perimetry have been further developed to evaluate rod and cone retinal sensitivities in different regions of the retina.^86,87 Such testing allows the definition of which class of photoreceptor (rod or cone) is the determinant of retinal threshold in a particular region or area of the retina. This has allowed the characterization of the relative losses of rods and cones in various types of RP associated with specific mutations.^88,89 Equally important to the definition of early retinal abnormalities in retinal degenerations, two-color perimetry can be used for monitoring disease therapy or for testing hypothesis of disease pathophysiology. For example, the abnormal, thickened Bruch’s membrane, which acts as a diffusion barrier between the retina and choroid, has been proposed as a major factor leading to night blindness in Sorsby’s fundus dystrophy.⁹⁰ The return of rod function after oral vitamin A supplements (50 000 IU/day) was elegantly recorded using these techniques.⁹⁰

Dark adaptometry

Normal subjects, when placed in the dark after exposure to a strong adapting light, will rapidly reduce their retinal psychophysical threshold using their cone system, reaching a plateau in about 5 minutes. Thereafter rod adaptation slowly increases, and rods determine retinal thresholds for another 3 log units of sensitivity before a second plateau occurs at about 30 minutes of dark adaptation. Patients with RP, when tested with dark adaptometry, may show elevation of the cone segment, the rod segment, or both, to varying degrees (Fig. 40.21). Also, in some patients there may be a delay in reaching what eventually for them is a relatively good final dark adaptation rod threshold⁹¹ (Fig. 40.22).

Fig. 40.21 Dark adaptometry showing marked elevation of both cone and rod thresholds with little or no cone–rod break in an 18-year-old woman, broken solid curve (A), with autosomal recessive retinitis pigmentosa and large-fiber sensory peripheral neuropathy (sister of patient shown in Figs 40.3 and 40.15A). Mild elevation of cone threshold with greater elevation of rod threshold in a 15-year-old young woman with Bardet–Biedl syndrome, broken solid curve (B). Borderline normal cone and rod segments in a 15-year-old young woman with type 2 Usher syndrome, broken solid curve (C). Dashed curve (M) is mean for normals aged 10–20 years.

Fig. 40.22 (A) Delayed achievement of final dark adaptation threshold in a patient with autosomal dominant retinitis pigmentosa. Filled circles are data from the patient; open triangles, circles, and squares are from three normals. The time constant for reduction of the rod retinal threshold by a factor of 1/2.718 of the threshold at the beginning of the curve was 8.5 minutes for the normal data and 101.6 minutes for the patient. Measurements were taken from a point approximately 45° temporal to the fovea. (B) Retinal thresholds measured at different points of retinal eccentricity for the same patient as in A, showing zones of elevated thresholds in nasal and temporal midperiphery with lower thresholds centrally and more peripherally. The shaded regions represent the range of absolute thresholds for the long- (R) and middle-wavelength (BG) test flashes for 5 normal subjects. (The point used for A is noted with an arrow in B.)

(Reproduced with permission from Alexander KR, Fishman GA. Prolonged rod dark adaptation in retinitis pigmentosa. Br J Ophthalmol 1984;68:561–9.)

Time course analysis of dark adaptometry has been performed in patients with adRP and prolonged dark adaptation is associated with rhodopsin mutations Thr17Met, Pro23His, Gly106Arg, and Thr58A.^88,89,92 The most characteristic feature of the Pro23His genotype was prolonged dark adaptation, which was present in all patients regardless of their stage of disease.⁸⁹ Jacobson et al.⁸⁸ also reported prolonged rod dark adaptation in patients with Thr58Arg and Thr17Met mutations of rhodopsin. Indeed, prolonged rod dark adaptation appears to be a characteristic finding of virtually all rhodopsin mutations. These studies suggest that each of these mutations produces specific abnormalities in the rate of reactions within the rods that limit the recovery of scotopic sensitivity. For patients with regional forms of RP from other causes, zones of more normally functioning retina may allow reasonably good final rod thresholds (Fig. 40.22). Minor abnormalities in cone thresholds are more likely to result in complaints of poor dark adaptation than mild to moderate elevation of rod thresholds. One study has indicated that abnormalities in the normal interactions between rods and cones may underlie the symptoms of poor night vision in some patients rather than isolated rod deficits.⁹³

Retinal densitometry (fundus reflectometry)

Retinal densitometry is a technique whereby, to deduce effective photoreceptor photopigment density, measurements are made of the difference between light shone into the eye and light reflected out of the eye. These measurements can be made and compared at different retinal loci, at different times, or against a retinal standard. From these data, estimates can be made of the rates of photopigment regeneration. Retinal densitometry has been performed on patients with RP by several investigators using a variety of techniques.^94–97 Rhodopsin levels were found to be reduced in all patients, but rhodopsin bleaching and regeneration kinetics have been normal. Van Meel and van Norren⁹⁸ found subnormal cone pigment levels in all patients studied and found prolonged cone pigment regeneration in eight of 12 patients with RP.

The losses of retinal sensitivity in RP in some patients are believed to relate to loss of quantum catch by the reduction of rhodopsin levels in the retina.^94,95 A reduction of rhodopsin level to one-tenth of normal, for example, would be expected to be associated with a 10-fold elevation in the rod threshold. These calculations are only estimates, however. In vitamin A deficiency, the elevation of the rod threshold is much greater than predicted by the reduction in rhodopsin levels.⁹⁵ Perlman and Auerbach⁹⁹ described a group of patients with RP who, when studied with retinal densitometry, were found to have elevations of rod thresholds greater than what would be expected purely on the basis of reductions of rhodopsin levels. Kemp et al.¹⁰⁰ found that 4 patients with type II (regional) adRP had elevations of retinal threshold that could be explained solely by loss of quantum catch from reductions of rhodopsin levels in the retina. However, 5 patients with type I (diffuse) adRP had elevations of threshold that were much greater than expected on the basis of loss of rhodopsin and quantal absorption for the reduction of rhodopsin levels observed.¹⁰⁰ These studies provide evidence suggesting that the two classifications represent different pathophysiologic entities. Others have studied patients with defined mutations of rhodopsin. Kemp et al.⁸⁹ have shown that, for patients with the Pro23His mutation of rhodopsin, rhodopsin levels are decreased below normal by amounts that indicate that rod sensitivity loss is determined by the reduced ability to absorb light. Similar findings were reported for patients with rhodopsin Thr17Met, Thr58Arg, and Gln344ter.⁸⁸

Fulton and Hansen⁹⁷ have reported reflectometry results on a mother and three daughters whose RP appeared to have a sectoral distribution. Elevations of retinal threshold were greater than those predicted by reductions of rhodopsin seen on densitometry measurements, suggesting that the dysfunction was central to the photoreceptor outer segments.

Electrophysiology

Karpe¹⁰¹ in 1945 first reported that the ERG was “extinguished” in RP. The standing, or resting, potential of the eye was first described by DuBois-Reymond in 1849. Riggs¹⁰² in 1954 was the first to report that the resting potential was decreased in pigmentary degeneration of the eye. Arden et al.¹⁰³ in 1962 developed the currently widely used technique for measurement of the light-induced rise of the resting potential of the eye (the electro-oculogram (EOG)), which is considered a function of the RPE. Fast oscillations of the resting potential can be recorded, and these offer another measure of RPE function.¹⁰⁴ Gouras¹⁰⁵ in 1970 based clinical ERG on the use of Ganzfeld stimulation. Standardized conditions and assessment protocols have now been established for electrodiagnostic investigations.^106,107

Electrodiagnostic responses in an RP patient can range from normal to undetectable. In general, more sizable responses are seen with younger patients or at earlier stages of disease. Most patients with advanced RP have undetectable responses (typically less than 10 µV) to single-flash, nonaveraged techniques.¹⁰⁵ Computer averaging, however, can usually detect small ERG responses in even moderately advanced stages.^13,108 In studying the natural course of RP, Berson et al.¹³ found that patients lost an average of 16–18.5% per year of remaining ERG amplitude to bright white flashes (a mixed rod–cone response).

Many studies have compared ERG responses in the different RP inheritance types. Tanino and Ohba¹² and Berson,¹⁰⁹ using conventional techniques, found that the ERG is more likely to be recordable in older patients with adRP than arRP. Berson et al.¹³ have also shown that progressive ERG loss of amplitude over a 3-year period is least in adRP. Birch and Fish¹¹⁰ found that the rate of progression of rod loss, as indicated by the half-saturation coefficient of the rod stimulus–response curve, varied among the inheritance groups. This ranged from 0.3 log unit/year in elevation of rod threshold in X-linked RP to 0.12 log unit/year elevation in adRP.

Agreement exists that the cone-mediated implicit times can be significantly prolonged in various forms of RP. Berson¹¹¹ has also stressed the importance of delays in the implicit time for cone-mediated 30-Hz flicker responses. Berson and colleagues^112,113 found that the cone-mediated b-wave implicit time in their patients was characteristically prolonged in adRP with reduced penetrance but not in completely penetrant adRP. Sandberg et al.¹¹⁴ suggested that this was due to an abnormal contribution of extrafoveal cones in adRP with incomplete penetrance. This may be due to abnormal cone light adaptation or a disturbance of normal rod–cone interactions.

Histopathologic study has indicated that, in RP, a postreceptor deficit may contribute to macular dysfunction.¹¹⁵ ERG study has confirmed this. Falsini et al.¹¹⁶ have assessed 8–10-Hz focal ERGs in 22 simplex and recessive RP cases. They found a loss of fundamental and second harmonic components and an increase in the fundamental-to-second harmonic ratio. This was interpreted as suggesting that both inner and outer retinal abnormalities contribute to visual deficit. Cideciyan and Jacobson¹¹⁷ have documented electronegative b-waves in some cases of RP. This finding was seen with abnormalities of on/off components of cone ERGs and reduced oscillatory potentials¹¹⁷ and was interpreted as indicative of significant inner retinal abnormalities in these RP cases.

Models of analysis of the ERG have been developed that attempt to isolate from the ERG the photoreceptor and bipolar contributions, specifically the original components of Granit’s model of the cat ERG, termed P3 and P2¹¹⁸ (Fig. 40.23A). These analysis methods have been developed to quantify the visual deficit attributable to outer and inner retinal disease (Fig. 40.23B–D). P3 is the corneal-negative response of photoreceptor outer segments that contributes to the a-wave of the Ganzfeld ERG. The analysis involves subjecting the leading edge of the a-wave of the ERG at a range of intensities to a mathematical curve-fitting algorithm^119–121 (Fig. 40.23C) using equations that describe the kinetics of rod phototransduction activation.¹²² The method of Hood and Birch¹²³ allows for estimation of two variables: S, which is a sensitivity factor, and R_mp3, which is the maximum amplitude. Such analysis can determine whether phototransduction is affected in a given disease.

Fig. 40.23 (A) Simplified illustration of the retinal elements of the rod electroretinogram (ERG: upper left) based on Granit’s model of the cat ERG. The b-wave of the corneal ERG (solid line) is the summation of a negative P3 generated by photoreceptors and a positive P2 generated in the inner layers of the retina. (B) The solid curves (upper right) are the scotopic ERG responses elicited by a high-intensity series of flashes from a normal subject. The dashed curves are the predicted P3 responses from a model of the receptor response fitted to the leading edge of the a-wave. (C) The solid curves (lower left) are records from a normal subject to a series of brief flashes, up to about 2 log scotoma td-s in energy, with the underlying P3 (rod receptor) responses shown as the dashed curves. (D) The curves (lower right) are the derived P2 responses obtained by computer subtraction of the P3 responses from the ERGs. INL, inner nuclear layer.

(Modified with permission from Hood DC, Birch DG. A computational model of the amplitude and implicit time of the b-wave of the human ERG. Visual Neurosci 1992;8:107–26.)

P2 is the bipolar cell-derived component of the rod-isolated b-wave. The P2 component of the rod b-wave, derived from bipolar cells, can be isolated from the full-frequency ERG, using an algorithm whereby the P3 is first calculated for a given series of intensities, after which the P3 is subtracted from the series of rod responses^124,125 (Fig. 40.23D). This gives a much better estimate of the ON bipolar contribution to the b-wave, especially for patients in whom the ON bipolar cells are more affected than the photoreceptors.¹²⁴

Application of these techniques to 15 patients with RP has shown that, for P3 analysis, all patients had significantly reduced values of R_mp3, indicating effective loss or damage of rods, and a wide spectrum of values for S.¹²⁶ Three patients had normal values for S. Four RP patients had subnormal S values but normal local retinal psychophysical thresholds in some regions of their retina. The rest had abnormal S values and abnormal visual fields with globally reduced thresholds. The rod b-wave amplitudes to a range of intensities for these RP patients were fitted with the Michaelis–Menten equation, generating the parameters K_bw (the semisaturation intensity, which is a sensitivity function) and V_max (the maximum amplitude). For all patients, the changes in K_bw and V_max followed the changes seen in S and R_mp3.¹²⁶ Hood and Birch¹²³ studied rod phototransduction in 11 patients with RP and four with cone–rod dystrophy (CRD). Their findings, which were similar to those above, supported the conclusion that the activation stage of rod phototransduction is affected in some forms of RP and CRD but not all.

Supernormal and delayed derived P2 responses were found by Hood et al.¹²⁷ when they applied modeling analysis to an unusual retinal dystrophy, first reported by Gouras et al.¹²⁸ in 1983 as “cone dystrophy, nyctalopia, and supernormal rod responses.” Although Gouras et al.¹²⁸ initially attributed this dystrophy to an abnormality in receptor cyclic guanosine monophosphate (cGMP) activity, P3 and P2 modeling by Hood et al.¹²⁷ showed that the delays in the rod and cone b-wave were not caused by the speed or amplification of phototransduction but resulted from delays beyond the outer segments involving a delay in activation of the inner nuclear layer activity. This disorder was subsequently clinically and electrophysiologically defined as the enhanced S-cone syndrome (ESCS)^129,130 and was later shown to result from mutation of the gene NR2E3.¹³¹

Paradigms have been developed to study the kinetics of rod transduction deactivation.¹³² This technique uses a bright conditioning flash followed by a probe flash. The recovery of the amplitude of the response to the probe, which is presented at varying intervals after the conditioning flash, provides a measure of the return to baseline of the response to the conditioning flash and, hence, the kinetics of transduction deactivation. The authors conclude that the lifetime of activated rhodopsin in normal human rods is 2.3 seconds. This technique has been applied to humans with the Pro23His rhodopsin mutation, demonstrating a reduction of the gain of activation and marked slowing of deactivation of rhodopsin (by a factor of at least 5).¹³² Animal models of RP, for example, transgenic mice expressing the mutant Pro23His rhodopsin gene, have also demonstrated increases in the half-maximal recovery period, T_50%, indicating abnormally prolonged recovery from the conditioning flash.¹³³

Analysis of the a-wave leading edge has also recently shown that in RP, the efficiency of cone phototransduction is affected very early, even in some patients in whom the sensitivity of rod phototransduction is normal. X-linked inheritance seemed to be associated with greater phototransduction abnormalities in cones and rods at a younger age than seen in other modes of inheritance.¹³⁴

Multifocal ERG techniques have been developed whereby, through patterned stimulation of the retina cross-correlated with recording of the ERG responses from a corneal electrode, very small ERG responses can be obtained from precisely localized regions of the retina^135–137 (Fig. 40.24). Multifocal ERGs can be seen in advanced RP where the 30-Hz flicker response is barely detectable¹³⁸ (Fig. 40.25). Hood and Li¹³⁷ have shown that, in RP, amplitudes of ERG responses do not correspond as well as b-wave implicit times with the retinal sensitivity as measured by static perimetry (Fig. 40.26). This is most evident with the regional or sectoral types of RP. Thus the b-wave implicit time appears to reflect early involvement of the disease process in regions of the retina that still have relatively good retinal sensitivity. Robson et al.¹³⁹ compared pattern ERG changes in RP patients with abnormalities of fundus autofluorescence (FAF). In such patients, pattern ERG responses suggested that visual loss was attributable to dysfunction rather than cell death.

Fig. 40.24 Schematic of multifocal paradigm. The stimulus consists of a pattern of 103 hexagons contained within a 50° diameter area. The subject fixates on the center of the pattern, while the hexagons undergo changes from black to white and from white to black according to a mathematical sequence that allows the stimulation of various local areas of the retina to be cross-correlated with the electroretinogram (ERG) continuously recorded from the cornea. Within a few minutes (typically 4–8 minutes), one can record tiny submicrovolt ERGs corresponding to the changing pattern of each hexagon of the stimulus pattern.

(Modified with permission from Hood DC, Seiple W, Holopigian K, et al. A comparison of the components of the multifocal and full-field ERGs. Visual Neurosci 1997;14:533–44.)

Fig. 40.25 Full-field clinical electroretinogram (ERG) protocol (left) and multifocal ERG (right) from a patient with retinitis pigmentosa. The full-field ERG showed no discernible rod or light-adapted cone response to single flash and a barely detectable 0.2-µV response to 30-Hz flicker response. However, sizable ERGs were detected within the central fields using the multifocal ERG technique.

(Courtesy of D.C. Hood and D.G. Birch.)

Fig. 40.26 (A) Multifocal electroretinograms (ERGs) from the right eye of a patient (P1) with autosomal recessive retinitis pigmentosa. The amplitude of the responses was decreased throughout the retina. However, this subject’s disease was regional in distribution with marked preservation of function centrally and inferotemporally. (B) ERG amplitude loss field (in sd from normal) for the same subject calculated by subtracting the mean trough-to-peak amplitude for the controls from the trough-to-peak amplitude for the patient’s response at each location. These differences were rounded and expressed in units of 100 nV. The clear regions indicate a decrease in amplitude of less than 162 nV (< −2 sd); the light-gray regions signify a decrease between 162 and 324 nV (−2 to −4 sd); and the darkest gray regions signify a decrease greater than 324 nV (> −4 sd). The black hexagons indicate that the response amplitude did not meet the criterion value of 90 nV. (C) ERG delay field for same subject, showing that regions that had subnormal amplitudes could be either normal or delayed in implicit time. The numbers in the hexagons were calculated by subtracting the mean implicit time for the controls from the implicit time for the patient’s response at each location. The numbers in these fields are rounded to the nearest millisecond for presentation. The black hexagons indicate that the response amplitude did not meet the criterion value of 90 nV. The clear regions signify that the delay was less than 1.7 ms (< +2 sd); the light-gray regions signify that the delay was between 1.7 and 3.4 ms (+2 to +4 sd); and the darkest gray regions signify that the delay was greater than 3.4 ms (> +4 sd). (Note that, since the numbers in these figures have been rounded, the same delay, for example, 2 ms, can appear as clear or as light gray depending on whether it was less than or greater than 1.7 ms.) (D) Modified Humphrey visual field thresholds for the corresponding hexagons using a 40-minute size target on a background luminance of 10 cd/m². The number at each point is the log of the ratio of the patient’s threshold to the mean threshold of the control group for that point. The clear regions signify that the patient’s threshold at that point was within 0.5 log unit (< +2 sd) of the mean; the light-gray regions signify values between 0.5 and 1.0 log unit (+2 to +4 sd); and the dark-gray signify values greater than 1.0 log unit (> 4 sd). The regions designated NaN include the central point (which the Humphrey system does not allow to be measured with a custom program) and two points that fall within the blind spot of normals. Note that ERG delay field was a better predictor of visual field threshold loss than the ERG amplitude field.

(Reproduced with permission from Hood DC, Holopigian K, Greenstein V et al. Assessment of local retinal function in patients with retinitis pigmentosa using the multi-focal ERG technique. Vis Res 1998;38:163–79.)

In general, the EOG is abnormal in diffuse hereditary rod–cone degenerations of the retina.¹⁴⁰ In typical RP, the slow and fast light-induced oscillations of the resting potential are usually decreased simultaneously early in the disease (Fig. 40.27). In some patients the fast oscillations of the resting potential are diminished earlier than the slow oscillation.¹⁰⁴ In others the fast oscillations of the EOG are more preserved than the slow oscillation, a finding that also occurs in Best vitelliform macular dystrophy.^104,141 In most cases of RP the EOG is abnormal when the ERG is abnormal.

Fig. 40.27 Electro-oculogram (EOG) in (A) a normal subject; (B) a 24-year-old man with autosomal dominant retinitis pigmentosa; (C) a 25-year-old woman with autosomal recessive retinitis pigmentosa (case 3); (D) a 30-year-old man with retinitis punctata albescens (case 12); and (E) a 35-year-old woman with autosomal recessive pericentral retinitis pigmentosa. The event marker line below the tracings indicates “on” (up) or “off” (down) in regard to the 68 cd/m² background light in the Ganzfeld stimulator. The ordinate is the indirectly measured amplitude of the corneofundal or standing potential of the eye (in microvolts per degree of fixation shift), as generated by alternating fixation between two red light-emitting diode fixation lights 30° apart. During the first 15 minutes of testing, the light alternates between dark and light periods of 75 seconds each to stimulate the so-called fast oscillations of the standing potential of the eye. The dark trough (DT) is the lowest point in the standing potential during the second 15 minutes of total darkness. During the third 15-minute period the background light is on continuously, stimulating a slow light-induced rise in the standing potential to a light peak (LP) 7–8 minutes later. As retinitis pigmentosa progresses, the light-induced rise of the resting potential (DT-LP), especially as indicated by the light-to-dark ratio (L/D), decreases. Note that, in all patients except E, the slow oscillation, as evidenced by the L/D ratio, is more preserved than the fast oscillations.

(Reproduced with permission from Weleber RG. Fast and slow oscillations of the electro-oculogram in Best’s macular dystrophy and retinitis pigmentosa. Arch Ophthalmol 1989;107:530–7.)

Fundus photography/fluorescein angiography

Documentation by fundus photography can assist in monitoring changes in patients with RP. Fluorescein angiography in patients with RP will demonstrate hyperfluorescence in areas of RPE atrophy and can highlight areas of CME (Fig. 40.28). However, fluorescein angiography has largely been supplanted by OCT for detecting cystoid maculopathy. In addition, concerns about light exposure accelerating certain forms of RP in animal models have prompted many ophthalmologists to exercise caution in obtaining excessive photographs.¹⁴²

Fig. 40.28 Fundus appearance (A) and fluorescein angiograms (B and C) of the left eye of a patient with presumed autosomal recessive retinitis pigmentosa at 26 years of age. Note the vascular abnormality superotemporal to the disc, vascular leakage, and cystoid macular edema on the fluorescein angiogram.

In many cases of mild to moderate RP, fluorescein angiography reveals transmission defects of the RPE with later diffuse leakage.¹⁴³ In 78 of 82 patients with RP, Newsome¹⁴⁴ found extravasation of dye from perifoveal capillaries in only a few patients but frequent abnormalities of the blood–retinal barrier at the level of the RPE. Such angiographically significant CME can often be seen at an early stage of disease in young patients who still retain good vision (20/25 or better). Macular edema is a significant cause of early loss of visual acuity in RP.^144,145

Autofluorescence

FAF utilizes a scanning laser ophthalmoscope to stimulate intrinsically autofluorescent molecules of lipofuscin to visualize the RPE.¹⁴⁶ Studies from patients with RP have shown that lack of signal on FAF correlate well with areas of RPE atrophy, while areas of increased FAF can be seen in areas with persistent macular edema as well as within areas of surviving retina.¹⁴⁷ Most RP patients demonstrate a perifoveal ring (Fig. 40.29) of increased autofluorescence within the macula, which denotes the border between functional and dysfunctional retina (see Chapter 4, Autofluorescence).^{139,147–149} The border of the parafoveal ring of increased autofluorescence has been shown to correlate with functional status as measured by pattern ERG, multifocal ERG, scotopic fine matrix mapping, and microperimetery.^150,151 In addition, areas outside the ring have been correlated with the loss of outer nuclear layer thickness and disruption of the inner-segment–outer-segment (IS/OS) junction on OCT.^152,153 Near-infrared autofluorescence (NIA) has also been used to image melanin present in the apical tips of the RPE.¹⁵⁴ Similar to FAF, rings with increased NIA are seen in patients with RP.¹⁵⁵ Combined NIA and blue light FAF imaging suggest that the presence of NIA may correlate better with preserved cone function, while blue light FAF indicates only preservation of RPE cells.¹⁵⁵

Fig. 40.29 Fundus autofluorescence image of left eye of patient with early retinitis pigmentosa (A) and more advanced retinitis pigmentosa (B) showing in each eye a central ring of hyperfluorescence and a region of mottled hypofluorescence in the region along the arcades.

Optical coherence tomography

OCT has become one of the most utilized imaging modalities for studying retinal disease in the past several years. Ultrahigh-resolution OCT and SD-OCT have been used to study retinal structures in patients with RP.^156,157 In these patients, such studies have demonstrated decreased thickness of the outer nuclear layer and loss of the external limiting membrane and IS/OS junctions. Loss of outer nuclear layer thickness or of the IS/OS has been correlated with visual defects measured by visual fields, microperimetry, or multifocal ERG.^29,158–160 SD-OCT is especially useful for detecting CME (Fig. 40.30) or epiretinal membranes, which are common features in patients with RP.¹⁶¹ The ability to detect CME by OCT ring often eliminates the need for fluorescein angiography.

Fig. 40.30 Spectral-domain optical coherence tomography of a patient with retinitis pigmentosa without cystoid macular edema (A) and with cystoid macular edema (B). Note the loss of outer retinal structures, such as the inner-segment–outer-segment junction outside the fovea in both patients.

Adaptive optics scanning laser ophthalmoscopy

Traditional imaging modalities cannot resolve individual retina cells due to the optical limits imposed by the cornea and lens which create higher-order aberrations resulting in image blur.¹⁶² A combination of adaptive optics with flood illumination or scanning laser ophthalmoscopy can compensate for these factors and provide imaging of individual cone photoreceptors.¹⁶³ Adaptive optics studies from patients with both CRD and RP have demonstrated increased cone spacing as well as qualitative areas where cone profiles could not be identified.^164–166 Recent advances have now enabled imaging of rod photoreceptors and foveal cones.¹⁶⁷ The ability to quantify cone photoreceptors has already been studied in one treatment trial for RP¹⁶⁸ and will undoubtedly play a role in future studies.

Subdivision by inheritance type

The most useful subclassification of RP both in clinical and research setting is still subdivision on the basis of mode of inheritance. Typical RP can be inherited as an arRP, adRP, or X-linked recessive trait (X-linked RP). Although mitochondrial inheritance is associated with pigmentary retinopathy, typical nonsyndromic RP has yet to be reported with this type of inheritance. The percentage of each inheritance type has been found to vary from author to author and with country of origin of the study. Despite the fact that typical RP is always genetic, a lack of family history of retinal disease is often reported. Studies around the world have found that no other affected family member can be identified in 15–63% of cases. Such cases are labeled “simplex RP.” Averaging results from different studies suggests that 35% of RP patients in the USA, 42% in the UK, and 48% in China are simplex. It is assumed that large proportions of these cases represent recessive inheritance. Jay¹⁶⁹ has estimated that no more than 70% of simplex cases are autosomal recessive.

Subdivision by age of onset

Early-onset RP may be subdivided into congenital and childhood-onset forms. Timing of onset of blindness, by the patient’s parent(s), and the occurrence of nystagmus (usually suggestive of a congenital disease) can be used to differentiate between congenital and early-onset cases. Occasionally, a member of a family with otherwise typical RP may present in late infancy or early childhood, while other affected members present anywhere from the end of the first decade to the third decade. arRP is usually more consistent in the age of presentation among affected siblings.

The most common age for presentation of symptoms and subsequent diagnosis of RP is in the first three decades of life – juvenile-onset and early adult-onset RP. All three inheritance types may present in such a fashion. Often, children with juvenile-onset disease function quite well at home but have great difficulties navigating strange environments. Reliable testing of visual fields may be possible in some children as young as 6 or 7 years of age. Perimetry indicative of progression or improvement of visual field deficits should, however, be interpreted with caution in this age range.

Adult-onset and late-onset forms of RP are not uncommon but often go unrecognized as a retinal dystrophy. Some of these retinal degenerations may have a nongenetic basis, but those that are genetic are usually autosomal recessive.

Subdivision by molecular defect

The expanding discovery of gene mutations associated with forms of RP is leading to an ever-increasing understanding of these entities at the molecular level. We now recognize that a mutant allele for a gene can behave in different ways, depending on where the sequence change resides within the gene and the status of the other allele. Gene mutations that produce no gene product (so-called null alleles) may exhibit autosomal recessive inheritance, if one good copy of the gene is sufficient to produce enough product to maintain normal function. Null alleles can also be associated with dominant inheritance from haploinsufficiency if one good copy of the gene cannot produce enough product to maintain normal function. Dominant inheritance can be seen with “dominant-negative” alleles, where multiple products of the wild-type gene must normally interact to form a multimeric protein complex or supramolecular structure. Missense mutations can also behave as a dominant trait by producing a “toxic gain of function” whereby the mutant protein disables normal gene regulation to downregulate gene expression of the normal copy. Dominant inheritance can also occur if the mutant gene product eithers fails to bind or binds too tightly to another gene product, disabling normal regulation pathways or important biochemical systems. Eventually, information on both the gene and the specific mutation or combination of mutations at each allele will be essential for optimum care. Additionally, as more information is discovered on modifiers of genetic diseases, molecular information will be needed on the status of these genes as well.

Since patients rarely present with information on the specific gene mutation for their disease, a classification based on molecular genetic defect will not entirely supplant the need to consider subdivisions of RP on clinical or psychophysical grounds. However, such molecular information will aid immensely in defining the true spectrum and natural history of specific types of RP. This refinement of classification will be particularly useful for prognostic counseling. The ability to detect the presence of the gene defect by examining DNA taken from a blood sample will allow earlier and more accurate diagnosis, facilitate genetic counseling, open the way for prenatal diagnosis, and eventually guide the patient toward specific gene defect-related therapies. Schemes for molecular classification of RP are given later in this text.

Subdivision by distribution of retinal involvement or fundus appearance

A number of recognizable fundus appearances have been seen in certain cases of RP. RP sine pigmento – RP without signs of intraretinal pigmentation – is one. In almost all cases these represent early RP. In the early stages of RP, fine, whitish, punctate lesions in the mid and far periphery at the level of the RPE can be seen. This fundus appearance is similar to that seen in retinitis punctata albescens.¹⁷⁰ White lesions or dots deep in the retina can be seen with RP in younger members of families with older affected members who have typical findings of RP. A myriad of tiny, irregularly shaped, gray-white, deep retinal lesions (Fig. 40.31) associated with lifelong stationary night blindness, minimally attenuated retinal vessels, and absence of pigment clumps or bone spicules are characteristic of fundus albipunctatus. Bilateral central scotomas can be present in later years. History of slow progressive visual loss with the macular atrophic degeneration is suggestive of the fleck retinal degeneration known as retinitis punctata albescens.¹⁷¹ However, differentiation of fundus albipunctatus from retinitis punctata albescens can be difficult. Macular atrophic lesions have been reported by Miyake and colleagues to occur also in fundus albipunctatus.^172–174

Fig. 40.31 Temporal raphe (A) and posterior pole (B) of fundus of the patient with either fundus albipuncata or retinitis punctata albescens at age 30. Myriads of discrete white dots are present throughout the retina and the macula has a bull’s-eye atrophic lesion.

Sector and sectoral retinitis pigmentosa

Sector RP, first described by Bietti¹⁷⁵ in 1937, refers to a specific subtype of RP. This is characterized by pigmentary changes limited to one or two quadrants, visual field defects usually only in the regions of retinal pigmentation, relatively good ERG responses, and minimal or no extension of the retinal area involved with time.

Patients may be minimally symptomatic. The area of retinal involvement is usually an arcuate swathe of retina just below the macula. In later years this involved region of the fundus may show almost a total regional atrophy of choroid and retina.¹⁷⁶ Occasionally the nasal retina¹⁷⁷ or inferior and nasal retina is involved.¹⁷⁶ Rarely, sector RP has been reported as affecting the temporal or superior quadrants.¹⁷⁸ True sector RP can be either autosomal dominant or autosomal recessive.¹⁷⁶ Although sector RP has been reported with mutations of the rhodopsin gene^179,180 and with mutations of USH1C,¹⁸¹ sporadic or isolated cases of sector retinal degeneration are common and may possibly result from nongenetic causes.

Massof and Finkelstein⁷⁹ have shown in autosomal dominant sector RP that the absolute retinal thresholds are elevated throughout the retina, including the fovea. Rods and cones appear to be equally affected. Over a period of years, the visual field defects worsen. Overall, however, visual prognosis is good. Using a combination of testing modalities, Fleckenstein et al.¹⁵¹ studied the FAF associated with various forms of retinal dystrophy, including one case with sector RP. Microperimetry disclosed that the ring of hyperfluorescence sharply delineated the areas of severe impairment of sensitivity. In cases of true sector RP, the ERG demonstrates relative preservation of amplitudes, with mild to moderate subnormalities of both rod- and cone-mediated responses with normal implicit times.¹⁷⁷ One form of sector RP appears to be associated with angle closure glaucoma.⁷⁴

Most cases of RP that begin or present in a sectoral distribution are, in fact, merely the sectoral presentation of what will become with time a more diffuse disease. One notable example of this is the brother of the patient shown in Figures 40.3 and 40.15A, who presented with a strikingly sectoral phenotype in association with rhodopsin Pro23His RP. Although he had poor night vision from age 17, he began to notice areas of blindness in his upper visual fields at age 30 years. The diagnosis of RP was made at age 42 years. At age 50, his best-corrected visual acuity was 20/20 J1 in each eye. His visual fields (Fig. 40.32) showed dense superior loss but good preservation of inferior field. Fundus appearance (Fig. 40.33) showed inferior and nasal pigmentary changes in the right eye and well-demarcated inferior sectoral changes in the left eye. The Ganzfeld ERG at age 51 showed small but measurable rod responses, markedly subnormal scotopic bright flash responses, and modestly subnormal photopic cone responses; rod and cone implicit times were normal. The 45-minute dark-adapted rod psychophysical threshold was elevated about 0.5 log unit above normal in each eye. His acuity remains normal, and he was able to drive a car both during the day and at night until his mid-60s. This patient illustrates the phenotypic variability that can be evident with the age of onset of symptoms, varying from early adulthood to the fifth decade of life. The normal implicit times for rod and cone responses for his case are unusual but are consistent with the sectoral nature of the expression of his disease.

Fig. 40.32 Goldmann perimetry for a 51-year-old patient with autosomal dominant sectoral retinitis pigmentosa from the Pro23His mutation of rhodopsin. Note that the major visual field defect is superior.

Fig. 40.33 Fundus appearance of the right (A) and left (B) eyes of the 51-year-old patient with autosomal dominant sectoral retinitis pigmentosa from the Pro23His mutation of rhodopsin. Note the discrete border between affected and normal-appearing retina and how well the fundus appearance correlates with the visual field in Figure 40.32.

Pericentral retinitis pigmentosa

Pericentral RP is a special phenotype whereby the loss of visual field typically occurs between 5 and 15° (Fig. 40.34) from fixation rather than between 20 and 40° from fixation, as is seen more commonly. Retinal diseases may commence inferiorly (Fig. 40.35), with commensurate superior field loss leading to the misconclusion that this is a form of sector RP. Pericentral RP is an important subtype because, as the areas of depressed field deepen, coalesce, and enlarge, they encroach more on the central region of seeing field and, thus, create greater disability at an earlier stage of disease.¹⁸² Eventually, as the central region of retina becomes progressively smaller or if the macula develops cystoid edema or atrophic changes, the visual acuity can decrease rapidly from relatively good acuity to less than 20/400. Pericentral RP can occur in many genetic forms of RP and can even be seen in relatives whose other affected family members have a different pattern or a more limited form of disease (Fig. 40.36), suggesting that modifying genes, other ocular conditions (e.g., high myopia), or environmental factors may contribute to this phenotype. Pericentral RP can occur as an autosomal recessive or dominant trait. Selmer et al.¹⁸³ reported a Norwegian family with autosomal dominant pericentral retinal dystrophy associated with a novel mutation of the gene TOPORS.

Fig. 40.34 Goldmann kinetic perimetry of a woman with pericentral retinitis pigmentosa at 51 (A) and 54 years of age (B) and Octopus static (combining programs 23 and 34) perimetry (C) at 54 years of age. Although the defect was greater in the superior visual field, the nasal scotoma extends into the inferior field. By 65 years of age a dense, complete ring scotoma was present with a small window of central field remaining, which within a year thereafter was lost, reducing her visual acuity to 20/400.

Fig. 40.35 Fundus appearance (A) and fluorescein angiogram (B) of the right eye of patient at age 51 years, and right (C) and left (D) fundi at 54 years.

Fig. 40.36 Right (A) and left (B) fundi of the 81-year-old father of the patient presented in Figures 40.29 and 40.30. His condition was considered as a form of sector retinitis pigmentosa but his daughter’s disease fit the phenotype for pericentral retinitis pigmentosa.

Unilateral or extremely asymmetrical retinitis pigmentosa

The vast majority of cases of so-called unilateral RP are acquired rather than genetic unilateral disease. The most common form of unilateral pigmentary retinopathy that is refered to as unilateral RP is diffuse unilateral subacute neuroretinitis or DUSN. This will be covered in the section on Differential diagnosis: pseudoretinitis pigmentosa.

Extremely asymmetrical RP of genetic etiology can occur in two instances. One is the carrier state for X-linked RP. Lyonization, or X-chromosomal inactivation,⁴⁰ occurs close in time to lateralization during embryogenesis. Thus, if the number of cells undergoing inactivation of the X-chromosomes that contain the normal gene for RP is uneven at the time of lateralization and, by chance occurrence, a greater number of those cells are directed to one side of the developing embryo, the carrier will express an extremely asymmetrical phenotype with asymmetrical field loss (Fig. 40.37) and pigmentary changes (Fig. 40.38). The second mechanism by which unilateral RP can occur as a genetic trait is through somatic mosaicism of a dominant gene for RP. This mechanism has been reported as the cause of unilateral RP in a patient with somatic mosaicism of RP1.¹⁸⁴

Fig. 40.37 Goldmann kinetic perimetry for the 64-year-old carrier of X-linked retinitis pigmentosa, whose fundus appearance is shown in Figure 40.38. She had mild, lifelong difficulty with night vision that was stable. The vision in her left eye began to fail as a young adult. At 64 years of age, the visual acuity was 20/25 OD and 20/60 OS, with −5.00 +1.00 axis 120° OD and −9.75 +0.75 axis 105° OS. Color vision was normal on the right and markedly disturbed on the left (AOHRR plates). Note the inferior arcuate defect in the right eye and severe irregular superior field loss in the left eye. The electroretinogram was moderately subnormal (OD) and severely abnormal (OS).

Fig. 40.38 Posterior pole and superotemporal retina of the right eye (A) and posterior pole and temporal retina of the left eye (B) of a 64-year-old carrier of a 2-bp insertion in codon 99 of the RP3 gene RPGR. Note the areas of retinal pigment epithelium (RPE) atrophy around the discs and superotemporally in the right eye and RPE thinning inferiorly and nasally with bone spicule pigmentation temporally in the left eye.

(Reproduced with permission from Weleber RG, Butler NS, Murphey WH, et al. X-linked retinitis pigmentosa associated with a two base-pair insertion in codon 99 of the RP3 gene RPGR. Arch Ophthalmol. 1997;115:1429–35.)

Infantile Refsum disease

IRD was first reported in 1982 by Scotto et al.³⁰⁰ in infants presenting with craniofacial malformation, severe hypotonia, psychomotor retardation, bleeding episodes, liver dysfunction by 6 months of age, and severe deafness within the first year of life. The ophthalmic manifestations were reported in 1984²⁶² to include nystagmus, poor vision, retinal degeneration with white flecks evident in the midperipheral retina (leopard spots) that fade and are replaced by coarse pigment clumping, optic atrophy, and eventually cataract. The ERG is profoundly abnormal early in the disease and can be electronegative.^262,263 Abnormalities of general peroxisomal biogenesis and function, similar to that seen with Zellweger syndrome and neonatal adrenoleukodystrophy, were identified in IRD and it is now classified as a milder variant of Zellweger syndrome (reviewed by Pennesi and Weleber²⁶³). The disease is often fatal by the second or third decade of life.^262,263

Adult-onset Refsum disease

This autosomal recessive disease, also called heredopathia atactica polyneuritiformis, was first described in 1946³⁰¹ with the ophthalmic features reviewed by Refsum in 1977.³⁰² This disease has been recently reviewed by Pennesi and Weleber.²⁶³ The earliest symptoms in adult-onset Refsum syndrome are ataxia, weakness in the extremities, and nyctalopia presenting in later childhood. Progressive peripheral neuropathy and peripheral muscle wasting usually follow this and cardiac conduction defects occur in early adulthood. Other common findings include paresthesia, anosmia, deafness, dry skin and ichthyosis, epiphyseal dysplasia, spondylitis, and kyphoscoliosis.

Ocular features include cataract, miosis with poor pupil dilation, and retinopathy. Nyctalopia in the second decade of life is followed by peripheral field defects. A pigmentary retinopathy is not evident until the third decade. ERG responses are severely abnormal or unrecordable at all ages.

Phytanic acid levels in blood and urine are always very high due to a deficiency of phytanic acid oxidation. Protein levels in the cerebrospinal fluid are also characteristically elevated. Rather than a deficit of peroxisome biogenesis, as is seen in IRD, a specific peroxisomal enzyme, phytanoyl-coenzyme A hydroxylase, is deficient and inactivating mutations of the gene PAHX, also called PHYH, as well as mutations of the PTS2 receptor have been identified in Refsum syndrome patients.^303,304

Restriction of dietary phytanic acid (which is present in dairy products and ruminant fat) can limit progression of disease and often improve the ichthyosis, neurologic deficits, and cardiac disease. There is limitation without improvement of visual or auditory deficits.^305,306

Rubella retinopathy

The most common ocular manifestation of congenital rubella, rubella retinopathy,³³⁵ is occasionally confused with RP. This confusion is especially likely in children with congenital deafness and a pigmentary retinopathy that are erroneously thought to represent both congenital rubella retinopathy and deafness rather than Usher syndrome. Uncommonly, the converse occurs, and the pigmentary disturbance of rubella retinopathy in a child with deafness triggers the suspicion or even the presumptive diagnosis of RP. Rubella retinopathy is one of the most characteristic manifestations of congenital rubella.³³⁶ Rubella retinopathy may take any of several forms. In some, the macula is the only site of abnormality, with a speckling of fine pigment granules. In others, pigmentary changes can extend further into the peripheral retina. Usually the correct diagnosis can be established by a combination of clinical features and ERG, which is only mildly abnormal in rubella retinopathy but will be, almost invariably, severely abnormal in Usher syndrome. Rubella retinopathy is a disease that is capable of mild progressive increase in pigmentary changes³³⁷ or the development of clinically significant subretinal neovascularization.³³⁸

Syphilis

Congenital or acquired syphilis can present as a pigmentary retinopathy that may appear similar in some respects to advanced RP,³³⁹ but careful examination usually establishes the correct diagnosis. Interstitial keratitis is commonly seen in congenital syphilis, and the pigmentary retinal changes are more varied and patchy. Usually the pigment deposits are clumps or large patches of black pigment associated with chorioretinal scars; typical bone spicule pigment formations are uncommon. The posterior pole may be as involved as the periphery. There may be evidence of past or present overlying vitreous reaction or uveitis. Acquired syphilis can also present with a diffuse retinal involvement.³³⁹

Infectious retinitis

Rarely, toxoplasmosis or herpes infections of the retina may leave a pigmentary retinopathy that poses a diagnostic challenge. These entities usually produce extensive full-thickness chorioretinal scars but may in certain areas produce only mottling and granularity of the RPE. Usually, randomly arranged patches of severe retinopathy exclude a diagnosis of RP. ERG is usually only minimally affected in postinfectious retinopathies, unless extensive areas of retina are severely damaged, in which case the ERG will be accordingly decreased in amplitude.

Thioridazine

This phenothiazine has been linked to severe, blinding retinal toxicity. In general, phenothiazines bind to melanin and concentrate in the uveal tract and RPE.^352,353 The retinal toxicity of thioridazine is believed to result from the presence of a piperidyl group. Thioridazine can cause a pigmentary retinopathy that can be confused with RP (in its early stages) or choroideremia (in its later stages).³⁵⁴ In most patients with thioridazine retinopathy the drug has been taken at doses higher than 800 mg/day.³⁵⁵

The earliest symptoms include blurring of central vision, poor night vision, and a brownish discoloration to the vision. Early toxicity produces fine, deep retinal pigment posterior to the equator, which becomes coarser as the condition progresses. Large “signal” plaques of pigment deposition occur late. The characteristic appearance of advanced thioridazine retinopathy is diffuse, patchy atrophy of choriocapillaris, RPE, and overlying retina, simulating the fundus appearance of early choroideremia.³⁵⁶ The ERG may show decreased amplitudes of both cone- and rod-mediated responses; responses of both a-waves and b-waves are subnormal.

Chlorpromazine

This drug has been reported to produce a pigmentary retinopathy when taken at high doses for prolonged periods.³⁵⁷ The pigmentary retinopathy does not significantly affect retinal function, and the retinopathy appears to regress when the drug is stopped.

Chloroquine

Similar to the phenothiazines, chloroquine, if taken over a prolonged period, binds to melanin and causes a toxic retinal degeneration. Very few cases have been reported with patients taking a total dose of less than 300 g.³⁵⁸ Early in the course of toxicity the ERG, EOG, and visual fields can be normal, although in late, severe toxicity, they can become markedly abnormal.^359,360 In advanced chloroquine retinopathy, unlike RP, the dark adaptometry may be either normal or only minimally abnormal.³⁵⁹ The typical picture of advanced chloroquine toxicity is that of a bull’s-eye maculopathy, although pigmentary changes with bone spicule formations can occur in the mid and far periphery.^279,360

Hydroxychloroquine

This drug has mostly supplanted chloroquine usage for rheumatoid arthritis and systemic lupus erythematosus. However, toxicity can occur with chronic usage and the current recommendation is periodic ophthalmological examination with static permetric visual fields (Humphrey field analyzer 10–2 or 24–2) when the duration of usage is over 5 years, particularly if the dosage is greater than 6.5 mg/kg per day.³⁶¹ The multifocal ERG appears to be a useful diagnostic tool to help differentiate pericentral visual field loss due to hydroxychloroquine from that due to optic nerve or other causes.³⁶²

Quinine

Retinopathy due to oral ingestion of quinine is usually a consequence of large doses taken either during a suicide attempt or to abort a pregnancy. Often, the diagnosis is aided by the history of acute loss of vision and the finding of characteristic pallor and edema of the retina early, with later development of attenuation of retinal vessels and optic nerve head pallor.^363,364 At this stage, quinine toxicity may be misdiagnosed as RP sine pigmento.³⁶³ The ERG in quinine toxicity is usually of a “negative” configuration, with far greater depression of the scotopic b-wave than a-wave.^365,366

Molecular genetics

Retinal dystrophy research in the past decade has been the subject of many important discoveries in molecular genetics. This occurred because of the well-publicized advances in the methodology of investigating the human genome. Especially important were the development of molecular genetic markers (DNA polymorphisms identified throughout the genome), polymerase chain reaction methodology, and ever easier DNA sequencing for mutation screening.

Using molecular genetic analysis, 14 specific genes have been associated with adRP, 29 with arRP (with three more loci), two specific genes associated with X-linked RP (along with three more loci), and one with digenic inheritance. Further genetic mutations are now being identified. For those interested in the most recent gene assignments and localizations, a list of cloned and mapped genes causing retinal degenerations or allied disorders has been compiled by Dr. Stephen P. Daiger and is accessible on the internet at the following address: www.sph.uth.tmc.edu/Retnet/.

Autosomal dominant RP genes

In rod photoreceptors, rhodopsin is the light-absorbing, conjugated photopigment found in outer-segment discs. It is composed of an apoprotein opsin molecule covalently bound to 11-cis-retinal (Figs 40.45 and 40.46).³⁹⁴ Incident light induces isomerization of the retinal and decay through a sequence of spectrally defined photoproducts. These light-induced changes result in a number of conformational changes in the opsin molecule, exposing G-protein-binding sites. A cascade of reactions (the phototransduction cascade; Fig. 40.47) is then set in motion and results in closing cGMP-dependent cation channels and relative transmembrane hyperpolarization.³⁹⁵

Fig. 40.45 Diagrammatic representations of rhodopsin protein. Left, Positioning of rhodopsin within the rod outer-segment disc membrane. Right, Three-dimensional representation of rhodopsin illustrating the central localization of the retinaldehyde chromophore.

(Courtesy of Dr. D. Farrens.)

Fig. 40.46 Diagrammatic representation of rhodopsin. Amino acids mutated in retinitis pigmentosa are highlighted in red. There is a disulfide bridge important for the tertiary structure between the cysteines (C) at codons 110 and 187; palmitoylation sites exist at the cysteines at codons 322 and 323. The lysine (K) at codon 296 is the site of binding of retinal to rhodopsin; the glutamic acid (E) at codon 113 is the counter ion for the lysine at codon 296.

Fig. 40.47 Diagrammatic representation of the retinitis pigment epithelium (RPE)–rod complex. Phototransduction and the role of rds/ROM-1 protein complexes in maintaining outer-segment discs are illustrated. Rho, rhodopsin; A, arrestin; RK, rhodopsin kinase; RetGC, retinal guanylate cyclase; GCAP, guanylate cyclase-activating protein; GMP, guanosine monophosphate.

McWilliam et al.³⁹⁶ were the first to detect linkage of adRP with an anonymous marker on the long arm of chromosome 3q. Shortly thereafter, Dryja et al.³⁹⁷ reported a mutation of codon 23 (Pro23His) of the rhodopsin gene mapping to chromosome 3q (Fig. 40.45). A mutation in the third exon of rhodopsin, Met207Arg, has been found in the Irish family originally linked to chromosome 3q.³⁹⁸

It has been estimated that rhodopsin mutation accounts for approximately 25% of adRP.³⁹⁹ In the USA, rhodopsin Pro23His is by far the most common of all rhodopsin mutations, being found in 12–15% of all American families with adRP.⁴⁰⁰ It is interesting that the Pro23His mutation has rarely been reported anywhere else in the world.⁴⁰¹ In fact, the mutation rate of the rhodopsin gene appears to vary with ethnic background. For instance, frequencies of RHO mutations in Asian populations, such as Japanese (5.9%),⁴⁰² Chinese (2.0%–5.6%),⁴⁰³ Indian (2.0%),⁴⁰⁴ and Korean (2.0%),⁴⁰⁵ are lower than that in the USA,⁴⁰⁰ and Europe.⁴⁰⁶

From the occurrence of a rare, intron 1 polymorphism in all patients with Pro23His adRP, it has been suggested that all persons with this mutation may have originated from a single-mutation founder.⁴⁰⁷ The phenotype associated with rhodopsin Pro23His is best described as Massof type II⁷⁹ or R-type RP.^84,85 Most rhodopsin adRP phenotypes have in fact been described as type II or R-type disease,^408–410 with a minority of mutations associated with type I or D-type disease.^411,412

Pro347Leu is the next most frequent mutation of rhodopsin, accounting for approximately 5% of 148 adRP families in one series.⁴¹³ This form of adRP is a more severe, diffuse phenotype, with comparatively smaller visual fields with more ERG deficit than that seen in cases of Pro23His adRP.⁴¹⁴ The ophthalmologic findings of patients with Gly106Arg mutation of the rhodopsin gene⁴⁰⁸ appear to be variable in expression, with patients rarely having normal cone and low-normal rod ERGs, a finding quite unusual among patients with RP. Sieving et al.⁴¹⁵ have reported the evaluations of seven affected members of a large family with a Gly90Asp rhodopsin mutation and described the condition for this family as autosomal dominant congenital nyctalopia, to distinguish it from typical RP.

Sandberg et al.⁴¹⁶ have reported clinical details on patients with 27 different dominant rhodopsin mutations. They suggested that severity of disease, measured by visual acuity, visual field diameter, dark-adapted sensitivity, and ERG amplitudes, increased with increasing codon number (Fig. 40.46). Berson et al.⁴¹⁷ have suggested that the rate of progression of RP from mutation of rhodopsin is correlated with the site of mutation and that other modifying factors, genetic or environmental, may influence phenotypic expression. A “constant equivalent light” model has been proposed for the severe adRP phenotype associated with rhodopsin codon 296 mutation. Lys296 is the chromophore-opsin attachment site and is also involved in holding the opsin in an inactive conformation. Theoretically this would lead to overstimulation of the phototransduction cascade, a situation similar to constant light exposure. This situation is known to lead to photoreceptor cell death.⁴¹⁸ This explanation has been questioned, however, because the opsin appears to be inactivated by phosphorylation and arrestin binding in Lys296Glu transgenic mice.⁴¹⁹

The peripherin/rds gene was first associated with RP in 1991,^420,421 and over 39 sequence variants have been associated with a range of autosomal dominant phenotypes, including RP, macular dystrophy, CRD, central areolar choroidal dystrophy, fundus flavimaculatus, and pattern dystrophy.⁴²² Most are missense mutations, but it is unlikely that peripherin/rds mutations account for more than 5% of adRP.⁴²³ These reports do however represent the first instance of which an animal model of RP has been found to have a directly correlated human disease counterpart. Much of what we understand to be the role of peripherin/RDS in human photoreceptors (Figs 40.48–40.50) comes from initial studies on this naturally occurring mouse model (rds), which has a 10-kb insert in codon 230 of the disease gene, creating a null allele.⁴²⁴ From this work and more recent human studies, it has been found that the peripherin/rds molecule accumulates in the rims of photoreceptor outer-segment discs and plays an important role in disc morphogenesis and stability.⁴²⁵

Fig. 40.48 Diagrammatic representation of peripherin/RDS protein. Amino acids mutated in retinitis pigmentosa are highlighted in red. Both the carboxyl-terminus and acetylated amino-terminus are cytoplasmic. Morphogenesis of the disc rims appears to involve interactions among tetramers of the large second intradiscal loop.

Fig. 40.49 rds/ROM-1 interactions at the disc rim, showing the homodimers of rds and ROM-1 forming tetramers by noncovalent bonds. These tetramers, which may represent the extracellular margin templates of Corless and Fetter, appear to interact across the intradiscal space at the rim margin to form the terminal loop complex (upper inset).

(Reproduced with permission from Weleber RG. Phenotypic variation in patients with mutations in the peripherin/RDS gene. Digit J Ophthalmol 1999;5:1–8.)

Fig. 40.50 Schematic illustration of the template theory of disc morphogenesis proposed by Corless and Fetter. Disc morphogenesis begins as a plate-like evagination from the plasma membrane in the region of the connecting cilium. The extracellular margin templates (MT) become aligned and assemble to form the terminal loop complexes that Corless and Fetter believe represent the primary morphogen of the disc rims. The asterisk denotes the leading edge of disc rim closure. CC, connecting cilium; POS, photoreceptor outer segment; PIS, photoreceptor inner segment; PL, plasmalemma.

(Reproduced with permission from Weleber RG. Phenotypic variation in patients with mutations in the peripherin/RDS gene. Digit J Ophthalmol 1999;5:1–8.)

Bessant et al.⁴²⁶ reported a large family with adRP previously linked to chromosome 14q11 that is associated with a Ser50Thr mutation of NRL. The gene product NRL is a retinal-specific DNA-binding transcription factor that interacts with the cone–rod homeobox transcription factor CRX to promote and regulate transcription of rhodopsin and other retinal genes. The Ser50Thr NRL mutation, when coexpressed with CRX in transfection experiments, produced a marked increase in transactivation of the rhodopsin promoter, producing excessive transcription of rhodopsin. Other families with adRP have been reported with NRL mutations.⁴²⁷

A rare but remarkable regional (type II) phenotype with graded disease severity (variable expressivity) has been reported associated with mutations in PIM1, on chromosome 7p.⁴²⁸ This encodes a protein kinase involved in transcriptional control. The gene FSCN2, which encodes a photoreceptor-specific paralog of the actin-bundling protein fascin, with presumptive cytoskeletal formation functions, has been reported with autosomal dominant RP.⁴²⁹ A common FSCN2 mutation, 208delG, appears to account for 3.3% of adRP in Japan.⁴²⁹

Digenic inheritance and RP

In rod photoreceptors, two peripherin/rds molecules form homodimers that bind noncovalently to homodimers of the glycoprotein ROM-1^455,456 (Fig. 40.49). In cones, peripherin/rds homodimers form tetramers with homodimers of either ROM-1 or a homolog.⁴⁵⁷ These tetramers may represent the “extracellular marginal templates” described by Corless and Fetter⁴⁵⁸ (Fig. 40.50) that are crucial to the creation and structural integrity of the disc rim.

In a few families, three rds point mutations – Leu185Pro, Arg13Trp, and Leu45Phe⁴⁵⁹ – have been seen in association with ROM-1 mutations. The affected compound heterozygotes had a severe RP phenotype. This type of inheritance has been termed digenic and highlights the important structural relationship between RDS and ROM-1 in the architecture of the rod photoreceptor. Despite studies around the world, very few RP families have been found to exhibit this phenomenon. Putative mutations of ROM-1 in isolation have been reported in some small families.⁴⁶⁰ As yet, isolated ROM-1 mutations are not convincingly associated with retinal disease.⁴²³

Usher syndrome molecular genetics

Molecular genetic linkage analysis has shown that not only are the three clinical subtypes of Usher syndrome genetically distinct but also that there is notable genetic heterogeneity. Currently 15 chromosomal loci have been linked to Usher syndrome and 12 of the associated disease genes have been identified (Table 40.1). The most severe and most common subtype, USH1B, has been associated with mutations of myosin VIIA (MYO7A).⁴⁶¹ Mutation of MYO7A can also cause nonsyndromal deafness of either recessive (DFNB2) or dominant (DFNA22) type.

Four other genes have been discovered for type 1 Usher syndrome. The gene for USH1C encodes harmonin, a novel PDZ domain containing protein that functions in the organization and assembly of larger protein complexes for cell adhesion, signal transduction, and intracellular transport.⁴⁶² Mutations in the ear-specific exons of the gene for harmonin also cause nonsyndromic recessive deafness (DFNB18).⁴⁶³ The gene for USH1D is CDH23, which encodes cadherin 23, a novel member of a class of intercellular adhesion proteins.⁴⁶⁴ Mutations in CDH23 also cause autosomal recessive nonsyndromic deafness (DFNB12).⁴⁶⁵ CDH23 is expressed in the cochlea and retina and mutations in the mouse ortholog cause disorganization of inner-ear stereocilia and deafness in the waltzer mouse animal model.⁴⁶⁶ Two of the harmonin PDZ domains interact with MYO7A and CDH23, indicating that these three proteins form a complex important for organization, maintenance, and function of stereocilia in the cochlea.^467,468 The gene for USH1F is PCDH15, which encodes protocadherin 15.⁴⁶⁹ Mutations in PCDH15 have also been reported in nonsyndromic recessive deafness (DFNB23).⁴⁷⁰ PCDH15 localizes to stereocilia in inner-ear hair cells and to the photoreceptors of the retina and is presumed to be essential for homeostasis. Mutation of the murine homolog causes vestibular dysfunction and deafness in the Ames waltzer mouse.⁴⁷¹ USH1G encodes the protein SANS, which is a novel scaffold protein gene that is expressed in retina and cochlea and is associated with harmonin within stereocilia.⁴⁷² Thus, it appears that many of the genes for type 1 Usher syndrome (MYO7A, harmonin, CDH23, and SANS) function together in a complex essential for determining and maintaning cohesion of the stereocilia of the cochlea.⁴⁷²

The commonest gene association with type 2 Usher syndrome (USH2A) has recently been identified,⁴⁷³ and encodes a protein called usherin that has motifs that show homology to thrombospondin, laminin epidermal growth factor, and fibronectin type III. This indicates that the gene product may function as a cellular adhesion molecule or as a component of the basal lamina and extracellular matrix. Although discovered first in patients with type 2 Usher syndrome, a Cys759Phe mutation of USH2A was found in approximately 4.5% of 224 patients with arRP without hearing loss.

The gene for type 3 Usher syndrome (USH3A) was first reported as encoding a 120-amino-acid protein in 2001 by Joensuu et al.⁴⁷⁴ A common termination mutation, Y100X, accounts for the majority of mutant alleles in Finnish cases.⁴⁷⁴ The gene product, clarin-1, is believed to have a role in hair cell and photoreceptor cell synapses.⁴⁷⁵ Pennings et al. have reported that type USH3A mutations can produce disease that mimics either type 1 or 2 Usher syndrome and that the fundus appearance may mimic retinitis punctata albescens or RP sine pigmento.⁴⁷⁶

Abnormal pre-mRNA splicing

Proteins encoding three RP genes, PRPF8 (RP13), PRPF31 (RP11), and PRPF3 (RP18), have been associated with RNA splicing. RNA splicing is an essential process that removes intron sequences from pre-mRNA. This is carried out by the spliceosome, a high-molecular-weight ribonucleoprotein complex.⁴⁸⁵ The vast majority of pre-mRNA introns (designated the U2 type) are spliced by the major spliceosome, which is composed of five uridine-rich small nuclear ribonucleoproteins, or snRNPs, termed U1, U2, U4, U5, and U6. Each snRNP is the central component of a protein complex that also contains seven Sm or Sm-like proteins, which are assembled by the so-called SMN (survival of motor neurons) protein. In addition to the Sm and Sm-like proteins, snRNPs also contain 3–10 particle-specific proteins.⁴⁸⁶ The major spliceosome also contains a poorly defined number of other non-snRNP protein factors. The minor spliceosome is responsible for splicing a rare class of pre-mRNA introns, called the U12 type.⁴⁸⁵ Minor spliceosomes contain the U5 snRNP and four nonabundant snRNPs: U11, U12, U4atac, and U6atac.

The PRPF8 gene (chromosome 17p13.3) encodes precursor RNA processing factor 8 (PRPF8). The yeast homolog, Prp8p, is a U5 snRNP factor that is first required for assembly of the U4/U6 and U5 tri-snRNP⁴⁸⁷ and then for facilitating the binding of the tri-snRNP to the 50 and 30 splice sites.⁴⁸⁸ It functions as the catalytic core of the spliceosome, either by facilitating formation of the core and/or by stabilizing RNA interactions.⁴⁸⁹ Seven different missense mutations of PRPF8, all clustered within a 14-codon stretch in the last exon, have been identified in five RP13-linked families.⁴⁹⁰ PRPF8 mutations are associated with early-onset adRP with diffuse retinal involvement and a severe prognosis in comparison with other forms of adRP.^490,491 Individuals experience night blindness beginning between 4 and 10 years of age and constricted visual field as teenagers.⁴⁹¹ Classic midequatorial bone spicule pigmentation is seen by middle age.⁴⁹¹ Typically, individuals are registered as blind or partially sighted by age 30.⁴⁹⁰ PRPF31 (chromosome 19q13.4) encodes protein 61K, the putative ortholog of the yeast pre-mRNA splicing factor Prp31p.⁴⁹² Protein 61K is specific to the U4/U6 snRNP and is required for pre-mRNA splicing in vitro.⁴⁹³ Protein 61K participates in the formation of the tri-snRNP, possibly by physically tethering the U5 snRNP to the U4/U6 snRNP.

Haploinsufficiency appears to be the mechanism underlying the photoreceptor degeneration in the majority of alleles of the PRPF31 gene associated with RP because most of these alleles are predicted to be protein truncation mutations. Families with PRPF31 mutations are unique in showing bimodal expressivity of the phenotype, with asymptomatic carriers who have both affected parents and affected children.^494,495 Symptomatic individuals experience night blindness and loss of visual field in their teenage years and are typically registered as blind by their 30s.⁴⁹⁴ Evans et al. suggested that the bimodal expressivity seen in these pedigrees may be explained by a second allelic genetic factor influencing phenotype. This idea was taken further by McGee et al.,⁴⁹⁶ who confirmed that penetrance of the PRPF31 mutations may be influenced in trans by otherwise silent PRPF31 alleles or by a closely linked locus on the wild-type chromosome, because a statistically significant correlation was found between the development of RP in a carrier and the inheritance of the region around RP from the noncarrier parent.⁴⁹⁶ Vithana et al.⁴⁹² have demonstrated that asymptomatic patients inherit a different wild-type allele to the one inherited by their symptomatic siblings. The third of the three known pre-mRNA splicing factors associated with adRP is encoded by the PRPF3 gene (chromosome 1q21.1).⁴⁹⁷ PRPF3 protein associates with the U4/U6 snRNP during pre-mRNA splicing.⁴⁹⁸ Mutations in its yeast ortholog, Prp3p, cause instability of the U4/U6 snRNP and prevent assembly of the tri-snRNP, thereby preventing splicing.⁴⁹⁹ Two different missense mutations have been identified in three individuals and three families.⁴⁹⁹ Affected individuals suffer night blindness toward the end of the first decade of life.⁴⁹⁷ Visual field defects are seen by the fourth decade of life but the central field remains relatively spared. A few patients progress to total blindness by age 80.

Several mechanisms have been proposed to account for the fact that mutations in these three, ubiquitously expressed, splicing factor genes only cause adRP and not a more widespread abnormality. The loss of one functional copy of these splice factor genes may affect only rods because of their unusually high requirement for protein synthesis and, consequently, for pre-mRNA splicing.⁵⁰⁰ Alternatively, the adRP-associated mutations may decrease the rate at which spliceosome activation occurs,⁵⁰¹ making activation the rate-limiting step in rod photoreceptors but not other cells. Finally, it has also been suggested that the retina-specific phenotype may occur as a result of an unidentified retina-specific splicing cofactor that interacts with all three of these splicing factor proteins.⁴⁹⁶

RPGR interactome

An interactome is a complex representation of functional interactions between molecules either within a cell or within the organism as a whole. Often such interactomes reveal important interactions between molecules that at first would not appear to be functionally related. In this way, the functional consequences of genetic mutation can be predicted without the need for lengthy experimentation. An RPGR interactome has also been proposed.⁵⁰² RPGR is a component of centrioles, ciliary axonemes, and microtubular transport complexes, although its precise function is unknown. It colocalizes with RPGRIP1 at the axonemes of connecting cilia in rod and cone photoreceptors⁵⁰³ by binding to RPGRIP1. This localization is lost in Rpgrip1 knockout (KO) mice. Rpgr KO mice develop a slow retinal degeneration, with features resembling a cone–rod degeneration – cone photoreceptors degenerate faster than rods and there is partial mislocalization of cone opsins.⁵⁰⁴ Some residual Rpgr^ORF15 expression has been reported in this model.⁵⁰⁵ RPGR has been shown to co-immunoprecipitate in retinal extracts with a number of different axonemal, basal body, and microtubular transport proteins.⁵⁰⁵ These include nephrocystin-5 and calmodulin, which localize to photoreceptor-connecting cilia; the microtubule-based transport proteins, kinesin II (KIF3A, KAP3 subunits), dynein (DIC subunit), SMC1, and SMC3; and two regulators of cytoplasmic dynein, p150Glued and p50-dynamitin, which tether cargoes to the dynein motor. Inhibition of dynein by overexpressing p50-dynamitin abrogates the localization of RPGR^ORF15 to basal bodies. RPGRORF15 can be co-immunoprecipitated from retinal extracts with other basal body proteins, including NPM, IFT88, 14–3-3ε, and γ-tubulin.^{502, 505} RPGR^ORF15 and RPGRIP1 co-localize at centrosomes in a wide variety of nonciliated cells and at basal bodies in ciliated cells. Both proteins are core components of centrioles and basal bodies.⁵⁰² In summary, RPGRORF15 appears to have a role in microtubule-based transport to and from the basal bodies and within photoreceptor axonemes, perhaps concerned with movement of cargoes between IS and OS. RPGRORF15 is predominantly expressed in photoreceptor connecting cilia and basal bodies but expression has also been reported in OS in some species,⁵⁰⁶ although this has been disputed.⁵⁰³ RPGR is also expressed in the transitional zone of motile cilia in the epithelial lining of human bronchi and sinuses (RPGR^ex1–19 only) and within the human and monkey cochlea.^{503,507–509}

Ush interactome

The USH genes encode proteins of different classes and families, including motor proteins, scaffold proteins, cell adhesion molecules, and transmembrane receptor proteins. In vitro direct interaction studies between USH proteins and also localization studies in mouse models have however suggested functional relationships between these proteins such that an “USH interactome” has been proposed.⁵¹⁰ In these studies, it should be noted, however, that although all the USH mouse models show severe hearing loss and vestibular dysfunction, only the Ush2A^–/– mouse, shows signs of RP.⁵¹¹ From such work, a “multiprotein scaffold complex” model has been proposed for harmonin, whirlin, and sans. There is also evidence that harmonin and whirlin can bind all other components of the USH network, including cadherin 23, protocadherin 15, usherin, VLGR1, and myosinVIIA.^467,507,512

Myosin VIIa has also been found to interact with sans 11 and protocadherin 15.⁴⁶⁸ Such findings suggest that the USH proteins form a transmembrane network that regulates hair bundle morphogenesis, particularly the growing stereocilia or the kinocilium, and may also have a role in the mechanoelectrical signal transduction and synaptic function of mature hair cells.⁵¹³ In the USH interactome model, it is proposed that the extracellular interstereocilia links are anchored intracellularly by the scaffold proteins, harmonin and whirlin, through direct binding to the actin cytoskeleton or through other proteins, including myosin VIIa.^{467,507,514,515}

Additionally, it has been shown that the association of other proteins with the USH complex (through harmonin) may function in determining cell polarity and cell–cell interactions.⁵¹⁶ Also Myo7a is believed to use long filaments of actin as tracks along which to transport other USH complex molecules.⁵¹⁷ In photoreceptors, proteins of the USH interactome are localized in the connecting cilium and may participate in the ciliary transport, but are also arranged at the interface between the inner segment and the connecting cilium where they control cargo passage. USH protein complexes may also provide mechanical stabilization to membrane calycal processes and the connecting cilium.⁵¹⁸ Studies of coexpression of USH1 and USH2 proteins also show accumulation at synaptic endings of photoreceptor cells, indicating that they are organized in an USH protein network there.⁵¹⁹

Bardet–Biedl syndrome and the “BBSome”

Recently it has been proposed that many of the proteins encoded by Bardet–Biedl genes form complexes, e.g., BBS1, BBS2, BBS4, BBS5,BBS7, BBS8 and BBS9 – the “BBSome.”⁵²⁰ The complex is important in the function of primary cilia, nonmotile projections found in numerous cell types. In particular, BBSome complexes are thought to play a central role in vesicular trafficking of membrane proteins within cilia. The BBSome binds to Rab8, a GTP/GDP exchange factor involved in docking and fusion of rhodopsin carrier vesicles in the connecting cilium of photoreceptors.⁵²¹ It is also known that other Bardet–Biedl genes, e.g., BBS3 (Arl6), form functional relationships with this BBSome.⁵²² Others appear to function as chaperones, e.g., BBs6, BBS10, and BBS12, and in protein ubiquitination (BBS11/TRIM32).⁵²³

Abnormal intracellular trafficking

An increasingly important consequence of retinal disease gene mutation has been the concept of abnormal intracellular trafficking.^400,524 For instance, another rod damage in rhodopsin mutants is often associated with mutant rhodopsin accumulating in the outer segment.^400,525,526 It is thought that this interferes with normal phototransduction and triggers apoptotic cell death. It has been shown in transgenic mice that mutant rhodopsin is only transported to the rod outer segment when the C-terminal region of the molecule is intact.⁵²⁷ This observation has been corroborated in human eyes.⁵²⁸

Cell death pathways

A prevalent proposition in the last decade has been that, since many different gene mutations result in a similar clinical outcome (RP), there must be a “final common cell death pathway” precipitated by these mutations.⁵²⁹ Apoptosis, the first cell death biochemical pathway linked to RP, characteristically appears histologically as a rounding up of the cell, reduction of cellular volume, chromatin condensation, and finally engulfment by resident phagocyte. Apoptosis is triggering of intrinsic and (to some extent) extrinsic pathways involving activation of caspases which can be subclassified into both initiator (2, 8, 9, 10) and effector (3, 6, 7) molecules. Studies with caspase inhibitors, however, have often produced only marginal neuroprotection in RP models.^530,531 Other studies have highlighted that caspase-independent inducers of cell death such as apoptosis-inducing factor, calpains, and poly(ADP-ribose) polymerases 1 are activated during retinal degeneration.⁵³²

Previously, necrosis had not been considered a significant part of retinal degeneration. This passive, unregulated form of cell death is now considered to be inducible by regulated signal transduction pathways such as those mediated by RIP kinases.⁵³² Such necroptosis may play a role in RP.^532,533 Apoptosis and necroptosis mechanisms may act together or may form alternate pathways.

Histopathology

Early histopathologic studies found that the first degenerative abnormalities in RP occur in the photoreceptors. As more and more animal model histopathologic correlates of specific human genetic mutations become available, our knowledge of both outer and inner retinal changes in RP is leading to new concepts for therapy.^477,534

Photoreceptor abnormalities

The earliest histologic sign of retinal dystrophy is shortening of rod outer segments (Fig. 40.51).⁵³⁵ With progression of disease, rod outer segments become more severely shortened, and eventually whole cells are lost. This is reflected in reduced numbers of nuclei in the outer nuclear layer. Rod cell loss is usually seen initially in the midperiphery. In certain cases cell loss is initially seen in the inferior retina. This is a typical finding in cases with rhodopsin mutation Pro23His.¹⁷⁹ In such cases there is a corresponding altitudinal field defect. Interestingly, a similar pattern of photoreceptor loss is seen in transgenic Pro23His mutant mice. In this model, if mutant mice were dark-reared, the resultant retinal degeneration was more uniform across the retina and was significantly milder.⁵³⁶ This suggests that limiting light exposure in patients with this mutation may moderate the progression of disease.

Fig. 40.51 Light micrograph from a normal eye (A) and an eye with retinitis pigmentosa (B). Richardson methylene blue/azure II stain. Note marked loss of outer segments in retinitis pigmentosa. ILM, internal limiting membrane; NFL, nerve fiber layer; GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; OLM, outer limiting membrane; IS, inner segments and OS, outer segments of rod and cone layer; RPE, retinal pigment epithelium (×60).

(Courtesy of A.H. Milam.)

Using immunocytochemistry and antirhodopsin antibodies, extensive studies have been undertaken of cytopathologic abnormalities in photoreceptors. In normal rods, rhodopsin is concentrated in the outer segments. In a number of rhodopsin-mutant animal models, abnormal localization of mutant rhodopsin has been seen,⁵³⁷ and this has also been seen in human eyes.^538,539 Such observations have led to interesting hypotheses as to how mutant protein may damage photoreceptors.⁵⁴⁰ It has been proposed that certain mutant rhodopsin molecules accumulate in the endoplasmic reticulum or Golgi apparatus in the inner segment and lead to cell death by interfering with the function of these organelles. Clearly, though, there are species differences in the cytopathologic effects of particular mutations. In transgenic Pro347Ser mice, rhodopsin is seen to accumulate in extracellular vesicles attached to rod inner segments.⁵⁴¹ Alternatively, in the transgenic Pro347Ser pig, rhodopsin accumulates in the inner segments.^542,543 Such aberrant accumulation of rhodopsin in the inner segment has not, however, been seen in human cases of Pro347Ser.

Clarke et al.^544,545 studied 11 different animal models of outer retinal dystrophy. These authors noted that the kinetics of photoreceptor death fit an exponential model. This in turn suggests three important features: (1) the risk of death is constant throughout the lifetime of each mutant photoreceptor; (2) each mutant photoreceptor is at the same risk of death as every other mutant photoreceptor (other things being equal); (3) the time at which death will occur is random and independent of the time of death of other photoreceptors. The regional variation in photoreceptor death in RP is not inconsistent with the feature of equal risk because local geographic factors may be superimposed on the equal risk of death of all the photoreceptors, resulting in a higher risk in certain parts of the retina. In the animal model studies, the same region of the retina was always examined, eliminating any effect of regional variation. The identification of these three characteristics of outer retinal dystrophies is important for at least two reasons. First, exponential kinetics refutes the major alternative model of photoreceptor death, the cumulative damage model, in which the risk of death increases with time (akin to the risk of death in an aging population of people) and the cell death kinetics are sigmoidal, not exponential. Second, the recognition that mutant photoreceptors are at a constant risk of death requires that biochemical mechanisms proposed to underlie the photoreceptor death must explain both the exponential kinetics as well as other major characteristics of the disease. For example, this may mean that, for all the different gene mutations, probably only a few pathways triggering apoptotic cell death are relevant to photoreceptor degeneration.⁵⁴⁶

Outer retinal disease

After photoreceptor cell death, the RPE becomes detached from Bruch’s membrane and migrates into the neurosensory retina. The subretinal space is abolished. Accumulation of RPE cells in a cuff around retinal vessels leads to bone spicule pigmentation.⁵⁴⁷ Bruch’s membrane thickening has also been recorded under degenerate RP retina. In late-onset RP, material has been seen to accumulate between the RPE and the inner collagenous layer of Bruch’s membrane.^548–550 This material was periodic acid–Schiff-positive and contained lipid, calcium, and iron. Choriocapillaries underlying areas of debrided RPE undergo atrophy.⁵⁵¹

Inner retinal pathology

Reactive changes have been reported in all cell types in the inner retina subsequent to photoreceptor cell death. Müller cells undergo reactive gliosis.^552,553 In advanced RP, entangled, thickened Müller cell processes result in extensive scarring in the remaining retina. Glial fibrillary acidic protein-positive astrocytes undergo reactive hyperplasia. This growth of astrocytes contributes to the pallor of the optic nerve head and epiretinal membrane formation.^554,555 Microglia migration into the outer retina has been documented in the RCS rat model of retinal dystrophy,⁵⁵⁶ but the human and pathophysiologic significance of this is unknown.

Previously it has been suggested that RP has little effect on ganglion cells.^557,558 Stone et al.¹¹⁵ studied 41 RP retinas and found significant reductions in ganglion cell numbers in each inheritance type of RP. The most significant losses were in X-linked RP and adRP. Transneuronal degeneration^534,559 has been the explanation for this phenomenon. Compression of cell bodies by retinal vessels⁵⁶⁰ and inner retinal ischemia has also been proposed as causes for ganglion cell loss. Inner retinal ischemia, a consequence of pigment epithelium perivascular cuffing,⁵⁴⁷ may in part also explain the optic nerve head pallor that is observed clinically.⁵⁶¹

More recent publications have expanded on the details of the anomalous remodeling and the misguided attempts at repair that take place within the retina in retinal degenerations following death and loss of the photoreceptors.^562,563 The changes appear in response to the stress of injury and cell death and progress through three stages: (1) the primary insult to the photoreceptors, which results in their dysfunction (and produces visual symptoms); (2) death of photoreceptors that ablates the sensory retina, by initial photoreceptor stress, phenotype deconstruction, and cell death of other cell types, through bystander effects or loss of trophic support; and (3) a protracted period of global remodeling and large-scale reorganization of the remaining neural retina.⁵⁶⁴ These attempts at repair result in three major tranformations: (1) the reactive hypertrophy of Müller cells and the formation of a distal glial seal that essentially isolates the remaining neuroretina from the surviving RPE/choroids; (2) apparent neuronal migration along glial surfaces to ectopic sites; and (3) the establishment of aberrant rewiring through a complex formation of neurite fasicles, new synaptic foci, and aberrant connections within the retina.⁵⁶¹ It is also now known that bipolar and horizontal cells in particular undergo significant morphological changes and changes in neurotransmitter receptor expression in response to outer retinal degeneration.⁵⁶⁴ RP triggers permanent loss of bipolar cell glutamate receptor expression, though spontaneous iGluR-mediated signaling by amacrine and ganglion cells. It has been proposed that rod bipolar cell dendrite switching likely triggers new gene expression patterns and may impair cone pathway function.⁵⁶⁵ These changes may place constraints on the concepts of cell or tissue transplantation or implantation of artificial retinas for regaining sight in advanced retinal degenerations.

Cellular remodeling and vascular changes

In the slowly degenerating dystrophic retina, peripheral rods sprout long, axon-like processes that can extend as far as the inner limiting membrane.^535,566 These neurites have been found in rhodopsin mutant eyes and in cases of X-linked RP. They are found in areas of marked cell death but have yet to be seen in the macula. Neurite processes bypass bipolar and horizontal cell synapses to become closely associated with hypertrophied Müller cell processes (Fig. 40.52). Neurites aggregate to form fascicles that arborize at the inner plexiform layer to form beaded processes under the inner limiting membrane. Although numerous vesicles similar to those seen in synapses are found in neurites, true synaptic connections have yet to be identified. Immunocytochemistry has shown that rhodopsin photopigment accumulates within neurites. It is still unknown whether these processes fulfill any functional role. Rod neurite sprouting does not seem to occur in dystrophic mouse models of RP. This suggests that the process occurs in retinas that undergo protracted degeneration over years, rather than rapid degeneration over weeks to months that is typical of mouse models of retinal dystrophy. Such models therefore may not be suitable in retinal transplantation studies ultimately intended to justify human studies. The environment in which it is intended to re-establish synaptic contact between transplanted rods and second-order neurons must be very different in dystrophic mice as compared with humans.⁵⁶⁶ Thus, models of RP in larger animals, such as the transgenic pig,^542,567 are necessary.

Fig. 40.52 Rhodopsin immunofluorescence in the retina in a normal (A) and a Q64ter rhodopsin mutant retinitis pigmentosa (B) subject. In the normal retina, rhodopsin immunolabeling is confined to the rod outer segments (asterisk). In the retinitis pigmentosa eye, rhodopsin is seen delocalized to processes (rod neurites) extending into the inner retina (arrows). R, Retinal pigment epithelium; P, photoreceptor somata (×230).

(Courtesy of A.H. Milam.)

Cone degeneration may occur early or late in RP. The evolution of cone changes is similar to that seen in rods. Cytoplasmic densification, axon elongation, and eventual reduction in number of cell bodies indicative of cell death follow initial cone outer-segment shortening. In the peripheral retina, shortened cone outer segments become surrounded by pyramidal-shaped RPE processes.⁵⁶⁸ Cones do not, however, undergo neurite sprouting to the extent seen in rods. In some RP eyes, peripheral cones occasionally evolve limited projections.^535,566 In an RP eye with a rhodopsin 64ter mutation, enlarged cone axons that extended into the inner plexiform layer were noted.⁵³⁵ If cell death has been extensive, a monolayer of cone somata is still evident in the macula. A cell survival factor specifically elaborated by rods but important for cones has been discovered.⁵⁶⁹ This factor, termed rod-derived cone viability factor, is a truncated thioredoxin-like protein that holds promise for the generation of potential new treatment possibilities for RP.

Retinal vascular abnormalities can be a prominent feature in cases of RP. Perivascular cuffing⁵⁴⁷ and arteriolar attenuation are common features of advanced disease. Retinal blood flow decreases in advanced RP.⁵⁷⁰ This may be caused by this perivascular cuffing, which constitutes migrated pigment epithelium and deposited extracellular matrix, or it may be a phenomenon secondary to reduced metabolic need. Reduced cell density in the degenerating retina may lead to a reduced metabolic demand and therefore reduced flow.⁵⁷⁰

In a few RP patients, changes in retinal permeability can lead to exudative retinal detachment (Coats-like disease).⁶³ This can be associated with peripheral retinal telangiectasia.⁵⁷¹ Peripheral retinal ischemia and preretinal neovascularization have also been reported.⁶⁰ Histopathologic evaluations of these vascular changes in humans however have only been reported from eyes that were already at the endstage of disease.