Retinal degeneration is an early consequence of the lysosomal storage diseases that are collectively referred to as the neuronal ceroid lipofuscinoses (NCLs) and the fatty acid oxidation disorder long-chain 3-hydroxyacyl CoA dehydro-genase (LCHAD) deficiency. This review summarizes recent developments that have been made in diagnosing and understanding the molecular bases of these disorders. The first symptoms of many of the NCLs often relate to visual loss from retinal degeneration. The resulting decrease in vision is typically evident at an early age, and the ophthalmologist may be the first specialist to examine the patient. Since the fundus of a young patient may be normal or not diagnostic of specific disease, the eventual diagnosis of the NCL often must come from additional testing. Since the retina is a readily accessible portion of the central nervous system, tests of retinal function have potential value not only for diagnosis, but also for future treatment trials in NCLs and LCHAD deficiency as rational interventions become available. This review will summarize ERG findings in young children with these metabolic disorders. Because the degeneration is often severe, even at an early age, specialized techniques developed for patients with retinitis pigmentosa are also necessary for analyzing the very small (submicrovolt) electroretinograms (ERGs).^1,2

The neuronal ceroid lipofuscinoses (NCLs) are a group of progressive neurodegenerative disorders characterized by the accumulation of complex storage material within lysosomes. As a class, the NCLs are the most common neurodegenerative disorders affecting children. In a recent survey of the causes of intellectual and neurological deterioration in childhood, the NCLs represented the largest category with 16% of cases. All storage diseases combined accounted for 63% of cases. The worldwide incidence is 1:12,500 live births.¹⁴ The disease is characterized by severe psychomotor deterioration that progresses to a vegetative state, seizures, visual failure from retinal degeneration, and premature death.^5,8,18 Four classical forms exist—three childhood-onset forms, which are all autosomal-recessive, and one adult-onset form, which may be autosomal-recessive or -dominant:

In addition, as many as 15 atypical forms have been described, some of which may be allelic to certain of the classical forms. One of the variant forms (vLINCL, CLN5) occurs essentially only in the Finnish population and shows linkage to a site (13q22) distinct from the three classic forms of childhood NCL.¹⁶ In Europe, the term Batten's disease is often used collectively for all forms of NCL.

All forms of NCL show accumulation of storage material that is autofluorescent, sudanophilic, and PAS-positive within lysosomes in neurons and other cells. Because of its osmophilic nature and appearance on light microscopy, the storage material resembles ceroid and lipofuscin but actually is a complex mixture of lipoproteins and other hydrophobic peptides. The lipoprotein deposits within cells on electron microscopy take on characteristic patterns that are used for diagnosis and classification. Granular inclusions are seen in INCL, Kufs’ disease, and some atypical forms of JNCL. Curvilinear inclusions predominate in classic LINCL. Variant forms of LINCL often show a mixture of curvilinear and fingerprint profiles. Fingerprint inclusions are seen in JNCL (with occasional to rare curvilinear inclusions). Historically, the diagnosis of this group of disorders has been established by looking for inclusion bodies in cells from brain biopsy or full-thickness rectal biopsy. More recently, skin or bulbar conjunctival biopsies have supplanted these more invasive surgical procedures. Buffy coat leukocytes can be used but may include a wider range of inclusions that may represent other storage disorders, such as the mucopolysaccharidoses. Muscle biopsy appears to be the only tissue suitable for diagnosis for ANCL or Kuf's disease.

The defective gene CLN1 for INCL encodes the enzyme palmitoyl-protein thioesterase-1 (PPT-1), an enzyme that removes long-chain fatty acids, mostly palmitate residues, from S-acylated proteins. As such, this enzyme is necessary for the reversible palmitoylation-depalmitoylation cycles used by signal transport proteins. Patients with INCL accumulate fatty acid esters of cysteine in their cells.^7,18 The most common mutation is R122W, which accounts for 98% of disease chromosomes in Finland but is rare in other parts of the world. In the United States, R151X is the most common CLN1 mutation, accounting for about 40% of mutant alleles. The gene CLN2 for LINCL encodes a pepstatin-insensitive lysosomal peptidase (TPP-1), which cleaves tripeptides from the N-terminus of small proteins before their degradation by other lysosomal proteases.⁶ The gene for the Finnish variant form of LINCL (CLN5) has been found to be a transmembrane protein that shows no homology to previous proteins and is distinct from the proteins defective in the other forms of NCL. The gene CLN3 for JNCL has been cloned and mutations have been defined, although the function of the gene is not known. The most frequent mutation for JNCL is a 1.02-kb deletion that is present in 90% of abnormal alleles in Finland and in 81–85% of abnormal alleles worldwide.¹⁸

Vision loss in the three classic childhood forms (INCL, LINCL, and JNCL) typically involves central vision initially and eventually results in profound visual loss, often with complete blindness, within a few years after the onset of symptoms. The ERG becomes abnormal early in all forms of the disease and within a few years is usually totally abolished to standard single-flash recording techniques. Functional testing of patients with retinal degeneration involves both psychophysical and electrophysiological measures. Among psychophysical measures, visual acuity and visual fields quantify the degree of visual impairment from the disease and are important for determining the necessity of, and eligibility for, a variety of low-vision services. Determinations of legal blindness (20/200 or worse, or field diameter less than 20 degrees, in the better eye) also rely on these two measures. While acuity is typically measured with Snellen eye charts in the clinic, treatment trials for retinal disease often employ standardized measures of acuity based on the Bailey-Lovie eye charts.⁴ These charts have a number of advantages for clinical trials, such as a constant number of letters on each line and a logarithmic progression between lines. Similarly, while Goldmann perimetry has historically been used to quantify field loss, clinical trials are increasingly utilizing the additional quantification available with automated static perimetry. Among the earliest complaints in patients with retinal degeneration is night blindness. Devices such as the Goldmann-Weekers dark adaptometer (Haag Streit AG, Berne, Switzerland) have traditionally been used to measure the full time course of dark adaptation, but such measures are time consuming and laborious for both the patient and the examiner. An alternative is to measure the final dark-adapted threshold. Typically, this can be accomplished in less than 5 minutes after patching one eye of the patient for 45 minutes. Smaller and less expensive alternatives to the Goldmann-Weekers dark adaptometer, such as the SST-1 (LKC Technologies, Gaithersburg, MD), are now available for this purpose.¹³

The primary electrophysiological test for patients with retinal degeneration is the full-field ERG. The core of the full-field ERG protocol is a set of responses adhering to the International Society for the Clinical Electrophysiology of Vision (ISCEV) standards established in 1989.¹² The standard specifies stimulus conditions and recording parameters to ensure that responses are comparable among test centers. Standardization has been a key development in ensuring that reports can be readily transferred and interpreted at centers around the country (or world) when a patient moves. It is also crucial for planning and implementing multicenter trials as rational therapeutic intervention becomes available.

The ISCEV standard specifies four responses of particular relevance to hereditary retinal degeneration (figure 80.1). The rod response is recorded following 45 minutes of dark adaptation, utilizing a flash (either blue or dim white) that is below the threshold for eliciting a cone ERG. Rods are affected at an early age in many forms of RP and allied retinal degenerations, so it is not unusual for the response to be nondetectable even in a young patient. To obtain a response that can be followed over time, the standard specifies a maximal response to a specified achromatic flash. The maximal response is a mix of rod-mediated and cone-mediated components; in a normal subject, approximately 70% of the amplitude is generated by rods. Two stimulus conditions are used to isolate the cones. An achromatic stimulus flickering at 30 Hz exceeds the flicker fusion frequency of rods; that is, only cones can respond. Similarly, an achromatic background of 34 cd/m² (lower right panel) saturates the rods; cones alone mediate stable responses following 10 minutes of light adaptation. These four responses should be incorporated into any protocol designed to assess patients with hereditary retinal degeneration. When the patient is dark adapted ahead of time, either by patching the eye or by sitting in total darkness, the core protocol takes less than 20 minutes, allowing ample time for additional, more specialized testing.

Figure 80.1.  

Computer-averaged ERGs to ISCEV standard protocol in patients with NCL. Top row, Rod responses to blue flash of −0.1logscottds. Second row, Maximal response to standard achromatic flash (2.0logphottds). Third row, 30-Hz flicker response to standard achromatic flash. Spikes are superimposed markers for stimulus onset. Bottom row, Light-adapted (1.5logcd/m² background) cone response to standard achromatic flash.

The rate of progression of the retinal degeneration in patients with NCL is extremely rapid in comparison to typical forms of RP. As shown in figure 80.1, ERG responses may be significantly reduced in amplitude in patients as young as 2 years of age (#4702). This young girl was subsequently found to have an active epileptic focus and diagnosed with juvenile NCL at age 4. The ERGs shown in the second column were obtained from a 5-year-old boy (#5370) with JNCL. Rod responses at this age are barely detectable, and the cone response to 31-Hz flicker is reduced by 80%. The patient tested at age 7 (#5099) had a nondetectable rod response and a cone response that was less than 1.0µV in amplitude.

Specialized recording techniques, including the selective filtering of responses to periodic stimuli through narrow-band amplification, can resolve ERG signals in the submicrovolt range.¹ The need is particularly acute within the population of patients with RP and NCL. The requirements of following these patients and conducting clinical studies in RP and NCL have led to unique approaches to recording small signals. These techniques have evolved in conjunction with the availability of powerful but inexpensive computers to acquire and process the signals. Selective filtering of responses to periodic stimuli through narrowband amplification shares many advantages with Fourier analysis but is generally more commercially available. A key property of any system for acquiring submicrovolt signals is that the analysis be conducted on-line so that the quality of the recording can be evaluated before the patient leaves. Another is the utilization of an artifact-reject window. Narrowband filtering removes the high-frequency components of blinks and the low-frequency components of movement. With this prefiltering, the artifact-reject window can be narrowed to two to three times the stimulus amplitude, further eliminating those components of noise at the stimulus frequency. With the techniques used here, signals greater than 0.05µV can be reliably distinguished from noise.²

The ERGs in patients with INCL, LINCL, and JNCL (figures 80.2, 80.3, and 80.4) have been found to be abnormal early in the course of all three disease types.¹⁷ For a patient with INCL (figure 80.2), rod responses were severely subnormal; the ISCEV standard rod and bright-flash ERG showed a normal a-wave and a profoundly subnormal b-wave, indicating that the earliest manifestations of this disease appear not to directly affect phototransduction. The electronegative ERG was interpreted as evidence for an effect on neurotransmission from proximal photoreceptors to ON bipolar cells. This appeared to occur at one of three possible sites: a disturbance of proximal photoreceptor function that interfered with presynaptic neurotransmission, a disturbance of the post-synaptic plate region, or some other effect on the bipolar cells, with subsequent reduction of the generation of the b-wave.

Figure 80.2.  

Computer-averaged ERGs, using intravenous propofol sedation, to a modified ISCEV protocol in a patient with infantile NCL from the Arg151 stop mutation of the CLN1 gene that encodes PPT1. The tracings from the right and left eyes are shown in black; the red tracings show the average of both eyes from a normal subject age 1.6 years. The scotopic blue and red flash stimuli were matched in normal control subjects to produce equal rod amplitudes. Note the sizable rod a-wave and profoundly subnormal rod b-wave for the blue flash, the electronegative configuration of the scotopic ERG to the bright white flash, and the subnormal, prolonged photopic cone response. (Reproduced with permission from Weleber RG: The dystrophic retina in multisystem disorders: The electroretinogram in neuronal ceroid lipofuscinosis. Eye 1998; 12:580–590.)

Figure 80.3.  

Computer-averaged ERGs to modified ISCEV protocol in three patients with late infantile NCL. The tracings from the right and left eyes are shown in black; the red tracings show the average of both eyes from an age-similar normal subject. Note the sizable but delayed rod responses, the prolongation of the scotopic oscillatory potentials, and the subnormal, prolonged cone responses. (Reproduced with permission from Weleber RG: The dystrophic retina in multisystem disorders: The electroretinogram in neuronal ceroid lipofuscinosis. Eye 1998; 12:580–590.)

Figure 80.4.  

Computer-averaged ERGs to a modified ISCEV protocol in three patients with juvenile NCL from mutation of the CLN3 gene. The tracings from the right and left eyes are shown in black; the red tracings show the average of both eyes from an age-similar normal subject. All responses were elicited using the same Ganzfeld stimulator, but because a different computer system was used for recording the responses for Case 6, a different normal is shown. Note the profoundly subnormal rod responses, the electronegative configuration of the scotopic ERG to the bright white flash for Cases 5 and 6, and the subnormal photopic responses, which were greater for the b-wave than the a-wave for Case 5. (Reproduced with permission from Weleber RG: The dystrophic retina in multisystem disorders: The electroretinogram in neuronal ceroid lipofuscinosis. Eye 1998; 12:580–590.)

The ERGs of young patients with LINCL (figure 80.3) had mildy abnormal rod amplitudes, mildly prolonged rod implicit times, and severely subnormal, prolonged cone responses.¹⁷ Patients with more advanced stages of LINCL also had a greater loss of b-wave than a-wave, again consistent with loss of signal transmission from photoreceptor inner segments to bipolar cells. Unlike the ERG in either INCL or in JNCL, the rod responses in early LINCL were only mildly subnormal and prolonged but with much more preserved amplitude, even though cone responses were severely subnormal and delayed.

Patients with JNCL invariably showed severe ERG abnormalities when first tested (figure 80.4), with essentially no rod-mediated activity and marked loss of a-wave amplitudes.¹⁷ They showed even greater loss of b-wave amplitudes, creating electronegative configuration waveforms. Greater loss of b-wave than a-wave amplitude for patients with JNCL would be consistent with the inner retinal localization of the gene product for CLN3.¹⁶

Patients with inherited long-chain fatty acid oxidation disorders, such as long-chain 3-hydroxyacyl-CoA dehydrogenase (LCHAD) deficiency, are deprived of an essential source of energy during fasting or metabolic stress when carbohydrate stores become depleted. The patients are typically treated with a modified diet consisting of medium-chain triglycerides or simply through restriction of dietary fat (low-fat/high-carbohydrate diet). These treatments dramatically reduce the progressive deterioration of cardiac, muscular, hepatic, and neurologic function associated with this disorder. Without treatment, patients with LCHAD deficiency have severe disease that usually results in death during the first two years of life. Now that patients are living longer with dietary interventions, it has become apparent that retinal degeneration often is associated with LCHAD.

The LCHAD activity resides in the mitochondrial trifunctional protein (MTP). Enzyme activities of subunits of this protein are responsible for distinct steps within the β-oxidation cycle. Genes for both subunits of MTP have been localized to the p23 region of chromosome 2.^10,19 The MTP deficiency can result from a mutation in either subunit, whereas LCHAD deficiency has only been reported with mutations in the α-subunit.^3,9,11

The fundus appears to be normal in LCHAD deficiency at birth. Between the ages of 4 months and 5 years, some patients develop a granular appearance to the retinal pigment epithelium. This can occur with or without pigment clumping within the retina (figure 80.5).¹⁵ The patients subsequently show vessel attenuation and retinal atrophy (figure 80.6). ERGs in this subset of patients are characteristic of severe retinal degeneration (figure 80.7). Other patients with LCHAD deficiency do not seem to develop retinal degeneration and may retain entirely normal ERGs. Whether the presence or absence of retinal degeneration is related to the particular genetic mutation is currently under investigation.

Figure 80.5.  

Fundus appearance in 4-year-old patient with LCHAD deficiency and early retinal degeneration. Note the characteristic dark brown spot in the fovea, the early thinning and atrophy of the retinal pigment epithelium (RPE), and the early pigment dispersion with fine clumping. The ERG was still normal at this stage.

Figure 80.6.  

Fundus appearance in a patient with later stage LCHAD deficiency and retinal degeneration. Note the more extensive atrophy of the RPE and choroid in the posterior pole.

Figure 80.7.  

Computer-averaged ERGs to ISCEV standard protocol in a normal subject (first column) and a patient with LCHAD deficiency (second column). Rod responses are severely reduced in amplitude, while cone responses have delay characteristic of progressive retinal degeneration.