History

Rett’s syndrome was first reported in 1966 by Dr. Andreas Rett [Rett, 1966]. This initial case report was followed in 1978 with a publication about Japanese female patients with a particular pattern of symptoms, including mental retardation and stereotypical hand-wringing [Ishikawa et al., 1978]. It was not until 1983, when Hagberg et al. [1983] published a case report of 35 female patients, that Rett’s syndrome gained international attention. Intense investigation of Rett’s syndrome over the past 20 years led to identification of the genetic defect in 1999 [Amir et al., 1999], 2005 [Borg et al., 2005; Scala et al., 2005; Kalscheue et al., 2003; Evans et al., 2005], and 2008 [Ariani et al., 2008].

Clinical Description

The diagnosis of Rett’s syndrome is based on a set of clinical observations accompanied by changes in various laboratory test results. The clinical criteria for classic Rett’s syndrome were established in the 1980s [Hagberg et al., 1985; Trevathan and Moser, 1988], and include loss of speech, seizures, mental retardation, and classic motor (specifically hand) movements. Criteria for atypical Rett’s syndrome were reported in 1993 [Hagberg and Gillberg, 1993]. More than 75 percent of patients have classic Rett’s syndrome, whereas 25 percent have atypical Rett’s syndrome variants [Hagberg, 2002].

In classic Rett’s syndrome (Table 41-1), the newborn initially appears developmentally normal. This period is followed by deceleration of head growth, loss of purposeful hand movements, development of stereotypic hand movements, and gait dyspraxia. These five criteria must be met for the diagnosis of classic Rett’s syndrome [Hagberg, 1995]. The chronology of these symptoms is critical for the diagnosis. Normal development is typical for the first 3–6 months of life. Deceleration in the rate of head growth occurs between 3 months and 4 years. Patients lose the ability to use their hands in a purposeful manner between 9 months and 2.5 years. Stereotypic hand movements appear between 1 and 3 years of age, and a dyspraxic gait manifests between 2 and 4 years of age if the patient is ambulatory.

Table 41-1 Obligatory Criteria for the Diagnosis of Rett’s Syndrome

(Data from Hagberg B. Clinical manifestations and stages of Rett syndrome. Ment Retard Dev Disabil 2002;8:61–65, and from Percy AK. Clinical trials and treatment prospects. Ment Retard Dev Disabil 2002;8:106–111.)

Clinical manifestations of classic Rett’s syndrome are grouped into four stages: early onset (3–6 months of age), regression (1–4 years of age), stabilization, and late motor impairment (after the age of 3 years) [Jellinger, 2003]. The early-onset stage is characterized by developmental delay, deceleration of head growth [Neul and Zoghbi, 2004; Schultz et al., 1993], onset of autistic-like behavior, and classic hand-wringing [Jellinger, 2003]. Weight and height percentiles for age also decrease; the median values fall below the fifth percentile by age 7 years [Percy, 2002]. Although Rett’s syndrome can manifest earlier, clear signs of a central nervous system abnormality are usually not evident until 6 months of age [Akbarian, 2003]. The second stage of the syndrome is characterized by cognitive decline and regression [Hagberg, 2002]. Loss of speech and purposeful hand movements, emergence of stereotypic movements, seizures, breathing irregularities, other signs of autonomic instability, inattentive behavior, and hypotonia appear [Hagberg, 2002; Jellinger, 2003; Kerr et al., 2001]. The third stage consists of stabilization of symptoms; this stage differentiates Rett’s syndrome clinically from other pediatric neurodegenerative disorders. Sometimes there is a return of communication skills, with preservation of remaining ambulatory skills. This stage is also known as the pseudostationary stage because slow neuromotor regression continues [Hagberg, 1995]. In patients older than 3 years, bradykinesia and rigidity set in [Fitzgerald et al., 1990]. Stabilization can last years to decades. Late motor impairment begins when ambulation ceases; this signals the end of the stabilization or pseudostationary stage. This final stage of Rett’s syndrome is characterized by nonambulation and severe disability. The length of late motor impairment is variable and can last decades.

Common clinical manifestations of classic Rett’s syndrome include stereotypic hand movements, intense staring, breathing irregularities, bruxism, sleep disturbances and night laughter, scoliosis, lower limb spasticity and dystonia, seizures, swallowing dysfunction, constipation, gastroesophageal reflux, and small, bluish or red feet. The stereotypic hand movements occur while the individual is awake. These gestures are individualized, but they typically include continuous and repetitive twisting, wringing, knitting, and clapping motions. The intense eye communication may be compensatory for the loss of speech. This eye pointing has been observed in many individuals with Rett’s syndrome. The breathing irregularities are of two types: hyperventilation and breath-holding. Typically, they occur only while the individual is awake, but can also occur during sleep [d’Orsi et al., 2009a].

Periods of apnea can last 30–40 seconds, and they disrupt stretches of hyperventilation. Most individuals with Rett’s syndrome experience sleep disturbances. It has also been reported that up to 90 percent of young children interrupt sleep with night laughter.

Although it is not pathognomonic, many individuals with Rett’s syndrome have early growth retardation of their feet. The nails and skin demonstrate trophic changes, and the skin is cool to touch and discolored with a blue–red color. Autonomic dysregulation may produce some of these changes. The scoliosis in Rett’s syndrome is a double-curve deformation that develops during the first decade of life. Most commonly, the double curve has a longer upper curve and a shorter lower curve. The incidence of scoliosis increases with age, occurring in 8 percent of preschool patients and 80 percent of patients older than 16 years. Abnormalities of the lower extremities in Rett’s syndrome include asymmetric distal dystonia, mild spasticity, and feet that tend to orient in a flexed and supinated position.

Seizures are reported in 30–80 percent of individuals with Rett’s syndrome. The electroencephalogram (EEG) is always abnormal after the age of 2 years. Early-onset seizures were reported in patients with CDKL5 [Artuso et al., 2010] and Netrin G1 gene mutations [Borg et al., 2005]. Prevalence of drug-resistant epilepsy in RTT patients with MeCP2 mutations was 16 percent. No significant relationship was found between clinical severity of drug-resistant epilepsy and quantitative or qualitative EEG scores. In addition, no significant relationship was found between the drug-resistant epilepsy and the RTT genotype category, or a specific MECP2 genotype [Buoni et al., 2008]. Myoclonic status had been misdiagnosed as a movement disorder of gait impairment [Pelc and Dan, 2009; d’Orsi et al., 2009b]. Infrequent clinical manifestations of classic Rett’s syndrome include bloating, violent screaming, abnormal nociception, pain insensitivity [Devarakonda et al., 2009], hyperkalemic distal renal tubular acidosis [Assadi et al., 2006], and cardiac arrhythmias [Acampa and Guideri, 2006]. Bloating or air swallowing is generally mild, but 5–10 percent of individuals with Rett’s syndrome demonstrate severe bloating. Massive gastric dilatation, with total necrosis and perforation due to bloating, has been reported [Baldassarrea et al., 2006]. The gastrointestinal disturbances are attributed to changes within the autonomic nervous system. Screaming typically is encountered in teenage patients. The screaming may be associated with ill-defined pain, but no known pathology can be found. Occasionally, patients have abnormally prolonged responses or insensitivity to pain. Children and adults with Rett’s syndrome are at substantially increased risk of fracture. The lower limbs, especially the femur, are particularly susceptible and patients with the R270X and R168X mutations genotype are especially vulnerable. The presence of epilepsy also increased fracture risk [Downs et al., 2008]. Although a decreased life span is characteristic for this syndrome, many patients survive into adulthood [Sekul and Percy, 1992], with 50 percent remaining alive in their 30s [Akbarian, 2003].

The atypical Rett’s syndrome variants include a forme fruste variant, early seizure type variant, late childhood regression variant, preserved speech variant, and congenital Rett’s syndrome. Diagnosis of atypical Rett’s syndrome is complex. The criteria for variants of classic Rett’s syndrome, as outlined by Hagberg and Skjeldal [1994], are especially helpful (Box 41-1). Forme fruste is the most common atypical variant, accounting for about 80 percent of nonclassic Rett’s syndrome. There is a wide variability of function in forme fruste; it is a milder variant. It is seldom diagnosed before 8–10 years of age, and it is usually suspected in older individuals who are just beginning to develop symptoms of Rett’s syndrome. The early-onset seizure variant is linked to mutations in CDKL5 and Netrin G1 gene and manifests with early epilepsy onset between the first week and 5 months, hand stereotypies, severely impaired psychomotor development, and severe hypotonia [Artuso et al., 2010; Borg et al., 2005]. The late childhood regression form is characterized by a normal head circumference and by a more gradual and later onset (late childhood) of regression of language and motor skills. The preserved speech variant was first described in 1992 [Zappella, 1992]. It is characterized by the preservation of speech, but preserved head size and obesity are also common features [Zappella, 1992; Zappella et al., 2001]. There is some debate about whether the preserved speech variant is part of the autistic spectrum disorders, as well as the Rett’s syndrome spectrum [Percy et al., 1990]. Congenital Rett’s syndrome is rare. It differs from classic Rett’s syndrome because of the absence of the 3- to 6-month period of normal development [Hagberg and Skjeldal, 1994]. It is linked to mutations in FOXG1 gene [Ariani et al., 2008].

Box 41-1 Defining Variants of Rett’s Syndrome

Inclusion Criteria

1. A girl of at least 10 years of age with mental retardation of unexplained origin

2. And three of the six primary criteria defined below

3. And six of the eleven supportive manifestations defined below

Primary Criteria

1. Partial or subtotal loss of acquired fine finger skill in late infancy or early childhood

2. Loss of acquired single words, phrases, or nuanced babble

3. Stereotypic Rett’s syndrome hand movements, with hands together or apart

4. Early deviant communicative ability

5. Deceleration in head growth of two standard deviations below the mean (even if head growth or circumference is still within normal limits)

6. Rett’s syndrome profile

Stage I – early-onset stagnation

Stage II – period of regression

Stage III – recovery of some contact and communicative abilities after stage II. Slow neuromotor regression that lasts through school age and adolescence

Stage IV – late motor deterioration. Slow neuromotor regression that lasts through school age and adolescence

Supportive Manifestations

1. Breathing irregularities (e.g., hyperventilation, breath-holding)

2. Bloating or air swallowing

3. Teeth grinding

4. Gait dyspraxia

5. Neurogenic scoliosis or high kyphosis (in ambulant girls)

7. Development of abnormal lower limb neurology

8. Small, blue or cold, impaired feet; autonomic or trophic dysfunction

9. Abnormal electroencephalogram, consistent with Rett’s syndrome

10. Unprompted laughing or screaming spells

11. Impaired or delayed nociception

12. Intensive eye communication (i.e., eye pointing)

Clinical Diagnostic Tests

Pathology

Genetics

Rett’s syndrome is an X-linked dominant disorder that has been mapped to the Xq28 locus [Ellison et al., 1992; Sirianni et al., 1998]. Although most cases of Rett’s syndrome are sporadic, genetic mapping was possible because familial inheritance does occur, and there is concordance in monozygotic twins [Jellinger, 2003]. Mutations within the methyl-CpG binding protein 2 gene (MECP2) cause 70–80 percent of reported cases of Rett’s syndrome in females [Auranen et al., 2001; Van den Veyver and Zoghbi, 2002]. This gene was identified in 1999 [Amir et al., 1999]. Most mutations in males lead to fetal demise. Although DNA mitochondrial mutations are found in some cases of Rett’s syndrome, there is no indication that mitochondrial DNA plays a part in the development of this syndrome [Nielson et al., 1993; Colantuoni et al., 2001]. Mutations within MECP2 have also been linked to childhood-onset schizophrenia, Angelman’s syndrome, and mild mental retardation [Watson et al., 2001].

The MECP2 protein has three known functional domains: an amino-terminal methyl-CpG binding domain [Lewis et al., 1992], a transcriptional repressor domain, and a carboxyl-terminal domain [Chandler et al., 1999]. The MECP2 protein binds to methylated CpG dinucleotides by the methyl-CpG binding domain [Nan et al., 1993]. The transcriptional repressor domain interacts with various co-repressor complexes and disrupts transcription [Nan et al., 1996, 1997]. The nuclear localization signal (NLS), consisting of amino acid residues 265–271, is contained within the transcriptional repressor domain. The biochemical function of the carboxyl-terminal region is unknown [Kriaucionis and Bird, 2003]. Seventy percent of the mutations within MECP2 are in eight hotspots affecting translation of the following amino acids: R106, R133, T158, R168, R255, R270, R294, and R306. Seven of these eight mutation hotspots affect arginine, which contains a CpG in its codon. These mutations may result from unrepaired deamination of 5-methylcytosine. This mechanism is thought to cause one-third of all point mutations that lead to human genetic disease [Cooper and Youssoufian, 1988]. Eighty percent of females with classic Rett’s syndrome have nonsense or frameshift mutations within the MECP2 gene [Van den Veyver and Zoghbi, 2002].

Cell Biology

Because Rett’s syndrome is caused primarily by mutations within the MECP2 gene, it is necessary to understand the function and interactions of the MECP2 protein. The MECP2 gene encodes a protein with three known functional domains: the methyl-CpG binding domain, the transcriptional repressor domain, and the carboxyl terminus. Human MECP2 has 48 amino acids (about 80 kDa) [Akbarian, 2003]. The methyl-CpG binding domain contains 85 amino acids [Nan et al., 1993] and binds to single- and double-methylated CpG dinucleotides [Bird and Wolfe, 1999; Lewis et al., 1992]. There is a correlation between the capability of the methyl-CpG binding domain to bind to pericentromeric heterochromatic regions of DNA and the ability of the protein to repress methylated promoters [Kudo et al., 2003]. Residues R111, R133C, and R134C within the methyl-CpG binding domain are thought to come into contact with methylated cystines. Mutations affecting R111 cause the MECP2 protein to lose its binding ability to heterochromatic DNA and its capability to repress transcription. Mutations resulting in R133C or R134C affect neither of the aforementioned protein properties. The MECP2 protein associates with chromatin remodeling complexes and aids in the regulation of the structure and function of chromatin.

The transcriptional repressor domain is 100 base pairs (bp) long, and it interacts with various co-repressor complexes [Nan et al., 1997]. One of these complexes is the Sin3A co-repressor complex. This complex contains histone deacetylases 1 and 2, which remove acetyl groups from histones and create a compressed form of chromatin that then inhibits or represses gene expression [Nan et al., 1998]. The action of the transcriptional repressor domain is partially reversed by trichostatin A, a histone deacetylase inhibitor. This finding suggests that repression by means of the transcriptional repressor domain is caused by histone deacetylation. MECP2 recruits these histone deacetylases and other chromatin remodeling complexes to methylated CpG dinucleotides. This leads to chromatin condensation that interferes with the binding of transcription complexes [Akbarian, 2003]. The transcriptional repressor domain has also been seen to bind to TFIIB (also designated GTF2B), SKI (a proto-oncogene), DNMT1 (which codes for a DNA methyltransferase), and SUV39H1 (which codes for a histone methyltransferase), although the importance of these interactions is unknown [Fuks et al., 2003; Kaludov and Wolffe, 2000; Kimura and Shiota, 2003]. Repression can be mediated in other ways, because the MECP2 protein binds to general transcription factors and interferes with the binding of transcription complexes. The transcriptional repressor domain is able to repress transcription when bound as far as 2000 bp from the transcription initiation site. The carboxyl terminus is thought to be involved in the binding of MECP2 to naked and nucleosomal DNA. Specifically, the carboxyl-terminal region of the MECP2 protein binds to DNA that is coiled around histone octamers [Chandler et al., 1999].

The MECP2 protein is mostly located within the nucleus of cells, and a small portion is seen within the perikarya [Kaufmann et al., 1995]. Although MECP2 binds throughout chromosomes, binding is most dense around pericentromeric heterochromatic regions of DNA. Forty percent of methyl-CpGs (i.e., binding sites for the methyl-CpG binding domain) are found within pericentromeric heterochromatic DNA. The immunoreactivity of MECP2 is increased around centromeric and perinucleolar heterochromatin. MECP2 does not associate with ribosomal DNA, despite its many methylations. MECP2 distribution is regulated by unknown factors and does not simply distribute to where methylated CpGs are found. The MECP2 protein is thought to play a role in the maintenance of neuronal nuclei in the later stages of development and within the mature brain [Akbarian, 2003]. It is hypothesized that MECP2 makes chromatin more stable and less accessible to transcription factors by anchoring chromatin fibers into the nuclear matrix.

There are reduced levels of dopamine, serotonin, and their metabolites, homovanillic acid and 5-hydroxy-indoleacetic acid, in Rett’s syndrome [Lekman et al., 1990]. Some researchers have noticed a decreased density of postsynaptic D₂ receptors in older patients with Rett’s syndrome [Dunn, 2001], whereas others describe increased specific binding at D₂ receptors. The latter finding implies that the decreased levels of dopamine are causing increased levels or density of postsynaptic receptors [Chiron et al., 1993; Dunn et al., 2002]. The D₁ receptor density is unchanged [Wenk, 1995]. Jellinger et al. [1990] proposed that the different densities of postsynaptic receptors within the dopaminergic pathways might be age-specific. Increased choline concentrations and decreased choline acetyltransferase levels are thought to result from problems within the cholinergic system in the forebrain [Gokcay et al., 2002]. The frontal cortex and striatum have decreased levels of ferritin [Sofic et al., 1987]. Decreased levels of binding protein for the benzodiazepine receptor in the frontotemporal, parietal, and occipital regions of the brain also have been reported [Yamashita et al., 1998]. Significantly increased oxidative stress markers (intraerythrocyte non-protein-bound iron, plasma non-protein-bound iron, free F2-isoprostanes, esterified F2-isoprostanes, total F2-isoprostanes, and protein carbonyl concentrations) were evident in Rett’s syndrome subjects and associated with reduced arterial oxygen levels compared to controls. Biochemical evidence of oxidative stress was related to clinical phenotype severity and lower peripheral and arterial oxygen levels. Pulmonary mismatch was found in the majority of the Rett’s syndrome population. These data identify hypoxia-induced oxidative stress as a key factor in the pathogenesis of classic Rett’s syndrome [De Felice et al., 2009].

Animal Models

In a model of MECP2-null mice, males and females were affected. Homozygous female mice and heterozygous male mice were developmentally normal for the first several weeks of life, but they died soon after neurologic symptoms appeared [Chen et al., 2001; Guy et al., 2001]. This model replicates the genetic component of Rett’s syndrome, but it was difficult to study because of the rapid deterioration and early death of the animals after symptoms appeared. Shahbazian et al. [2002] developed a mouse model that has a truncation mutation within the MECP2 gene with a less severe phenotype than the MECP2-null mice. The truncated protein mimics a commonly observed human mutation, and it is partially functional, containing the methyl-CpG binding domain and transcriptional repressor domain. This group observed that the mice appeared developmentally normal up to 6 weeks after birth. After this period of normal development, the mice developed neurologic symptoms, including tremors, motor impairment, hypoactivity, seizures, kyphosis, and the classic forearm movements associated with human Rett’s syndrome [Shahbazian et al., 2002b]. Random X-inactivation causes a variety of phenotypes due to a single genotype in females, which makes analysis of the mouse model difficult [Young and Zoghbi, 2004]. One solution is to use male mice, because they are not subject to random X-inactivation.

Luikenhuis et al. [2004] overexpressed MECP2 in postmitotic neurons of homozygous MECP2-null mice. These mice did not display any neurologic symptoms, and they were developmentally equivalent to the control population. In normal mice, MECP2-encoded RNA is not expressed until about 10 days after conception, and it reaches adult levels by 16 days after conception. It has been postulated that the defect in Rett’s syndrome involves neuronal maintenance and maturation, and therefore affects developmental stability. Ballas et al. found that the loss of MeCP2 occurs not only in neurons but also in glial cells of Rett brains, and that mutant astrocytes from a Rett mouse model, and their conditioned medium, failed to support normal dendritic morphology of either wild-type or mutant hippocampal neurons [Ballas et al., 2009]. This suggests that astrocytes in the Rett brain carrying MeCP2 mutations have a non-cell autonomous effect on neuronal properties, probably as a result of aberrant secretion of soluble factor(s).

Pathogenesis

The timeline of MECP2 expression suggests that it is needed in the later phase of cortical development in neonates and after maturation in adults. MECP2 is expressed in normal fetal brains until 20 weeks’ gestation, after which it disappears from the cerebellum. It disappears from the brainstem after the perinatal period. MECP2 does not reappear in the brain until after adolescence. Lack of MECP2 during any of these developmental periods may lead to synaptic and neuronal dysfunction of the catecholaminergic neurons in patients with Rett’s syndrome [Itoh and Takashima, 2002]. Mice with MECP2 overexpressed in postmitotic neurons are rescued from the Rett’s syndrome phenotype. A decrease in MECP2 in postmitotic neurons during the later stages of development is sufficient to cause Rett’s syndrome [Chen et al., 2001], but MECP2 deficiency in neuronal precursors is probably not a major contributor to the pathogenesis in Rett’s syndrome.

MECP2 causes transcriptional repression, and loss of function of this protein may cause an imbalance between transcription and gene silencing, leading to dysregulated gene expression and pathologic changes. Some groups have found no evidence for MECP2 transcriptional repression in neurons or glia. Gene expression studies found an increase in glial transcription, in contrast to the predicted decrease. Levels of presynaptic proteins, however, were decreased. MECP2 may be causing perturbations within the presynaptic signal transduction pathway [Colantuoni et al., 2001]. The critics of this theory point out that these studies were performed on postmortem brains, suggesting that relevant time points in gene expression may have been missed. Additional evidence argues against MECP2 deficiency unsilencing transcription and causing Rett’s syndrome. Affymetrix GeneChip analysis of MECP2-deficient human fibroblasts did not demonstrate any large-scale dysregulation of gene expression. There were only small differences in gene expression in presymptomatic, early symptomatic, and late symptomatic MECP2-deficient mouse brains [Tudor et al., 2002].

If MECP2 does not act as a transcriptional repressor, it may play a maintenance role during development. MECP2-encoded mRNA is undetectable in the mouse forebrain during mid-gestation [Coy et al., 1999]. Immunohistochemical studies in nonhuman primates and mice demonstrate MECP2 expression in neuronal nuclei correlates with neuronal maturity, and levels of expression are highest in the adult cerebral cortex [Akbarian et al., 2001]. In human cerebral cortex, MECP2 is seen only in Cajal–Retzius cells, the earliest maturing neurons [Marin-Padilla, 1998], at 14 weeks’ gestation. At 26 weeks’ gestation, MECP2-immunoreactive neurons are seen in the deeper, more differentiated cortical layers. MECP2 neuronal immunoreactivity increases with age. Only about 10 percent of cells are immunoreactive during the third trimester of pregnancy, but approximately 80 percent of neurons demonstrate immunoreactivity for MECP2 at 10 years of age [Shahbazian et al., 2002a]. These findings are also seen in rodents and nonhuman primates. The observation that MECP2 expression is decreased in immature neurons and elevated in mature neurons suggests a role for MECP2 in neuronal maintenance. MECP2 expression studies show high levels of the protein in mature neurons but not in glia or astrocytes.

MECP2 may also play a role in cell division. Removing pericentromeric heterochromatin or disrupting heterochromatin silencing by inhibition of histone deacetylases in Drosophila and yeast reduced chromosome transmission during cell division [Henikoff, 2001]. Because MECP2 associates with pericentromeric heterochromatin, it may influence cell division.

Treatment

Rett’s syndrome has no cure. The treatments available have been empirically derived and are designed to combat specific symptoms. Antiepileptic agents include carbamazepine, valproic acid, and lamotrigine. The use of l-DOPA and dopamine agonists to increase motor ability in Rett’s syndrome patients is controversial [Zappella, 1990]. A study of the treatment of Rett’s syndrome with folate and betaine did not find any objective evidence of improvement [Glaze et al., 2009]. Zinc sulfate, lithium, and antidepressants have been demonstrated to increase central brain-derived neurotropic factor (BDNF) levels or signaling in human as well as animal studies. Thus, it is proposed that these agents could have therapeutic potential for RTT subjects [Tsai, 2006]. The breathing irregularities associated with Rett’s syndrome can be treated with naltrexone, an opiate antagonist. Naltrexone (1–3 mg/kg/day) can reduce disorganized breathing and increases oxygen saturation levels [Percy, 2002]. Use of high-fat, high-calorie diets is recommended in the late stages of the disease, and it has been suggested that individuals with Rett’s syndrome require a higher protein intake [Motil et al., 1999]. Feeding by means of a gastrostomy tube is sometimes indicated [Jellinger, 2003]. Constipation in Rett’s syndrome has been managed with high-fiber foods, enemas, mineral oils, milk of magnesia, and polyethylene glycol or MiraLax, with varying degrees of success. Orthopedists and physical therapists routinely see individuals with Rett’s syndrome. The goals are to improve balance, enhance flexibility, and strengthen atrophying muscles. Bracing for scoliosis is necessary when a 25° curvature exists, and surgery is advised when the curvature exceeds 40°. Speech and occupational therapy are occasionally used to improve communication.

Small-scale clinical trials using l-carnitine and the ketogenic diet have been completed. l-Carnitine has been reported to improve the respiratory features of Rett’s syndrome [Ellaway et al., 2001], and the ketogenic diet may reduce seizure frequency during the first 3 months, but no long-term clinical trials have been reported. A comprehensive, life-span approach to the management of scoliosis in Rett’s syndrome is recommended that takes into account factors such as physical activity, posture, and nutritional and bone health needs. Surgery should be considered when the Cobb angle is approximately 40–50° and must be supported by specialist management of anesthesia, pain control, seizures, and early mobilization. Evidence- and consensus-based guidelines were successfully created and have the potential to improve care of a complex comorbidity in a rare condition and stimulate research to improve the current limited [Downs et al., 2009].

History

The story of the discovery of Menkes’ disease begins in the 1930s, when veterinary physicians in Australia noticed the importance of copper for the normal development of sheep [Bennetts and Chapman, 1937]. They observed that mothers that grazed in copper-deficient pastures had offspring with cerebral demyelination, ataxia, and porencephaly. They concluded that copper deficiency in sheep was associated with ataxia and a demyelinating disease. The disease was described in 1962 by John Menkes [Menkes et al., 1962]. He described five males of English–Irish heritage who had “peculiar” hair, failure to thrive, and a neurodegenerative disorder. The syndrome was initially called Menkes’ kinky hair syndrome in reference to the unique appearance of the hair of these patients. The basis for Menkes’ disease was not known until the association was made with the illness in sheep [Danks et al., 1972, 1973]. The distinctive hair found in the copper-deficient sheep and in patients provided the necessary link. Serum testing revealed that patients who had this distinctive hair also had decreased serum copper and ceruloplasmin levels. Danks et al. [1972, 1973] then concluded that Menkes’ disease was a human example of a neurodevelopmental disorder caused by copper deficiency. In 1993, three groups used positional cloning to discover the ATP7A gene [Chelly et al., 1993; Mercer et al., 1993].

Clinical Description

Menkes’ disease typically occurs in males. Its occipital horn syndrome variant is also known as X-linked cutis laxa [Danks, 1995]. In classic Menkes’ disease, there is a period of normal development that typically lasts for 2–3 months [Kaler, 1994]. Developmental regression follows, with seizures, hypotonia, and failure to thrive [Kaler, 1998]. Patients exhibit severe mental retardation with symptoms of neurodegeneration [Mercer, 1998]. Most individuals with classic Menkes’ disease die between the ages of 7 months and 4 years [Bankier, 1995]. Occipital horn syndrome is characterized by a less severe genetic mutation that results in a connective tissue disorder.

The characteristic findings in classic Menkes’ disease are related to the patient’s hair. Commonly described as steel-woolish, the hair on the scalp and eyebrows is short, sparse, coarse, and twisted. The amount of hair is decreased, and it is generally shorter on the sides of the head. The color is often light, with white, silver, and gray being common. Light microscopy of hair reveals three characteristic findings [Kaler, 1998; Moore and Howell, 1985]:

Hypopigmentation is common but not the rule, and is caused by a deficiency of catechol oxidase. Patients typically have large jowls, sagging cheeks, large ears, and a high-arched palate. The skin appears loose at the nape of the neck, in the axillae, and on the trunk. Delayed tooth eruption and pectus excavatum are common, and the incidence of umbilical hernias is increased. Nephrocalcinosis and chronic renal failure have also been reported [Balestracci et al., 2009].

Classic Menkes’ disease has several distinctive neurologic findings. These patients exhibit truncal hypotonia with poor head control, hyperactive deep tendon reflexes, impaired visual fixation or tracking, and cortical adducted thumbs. There is increased appendicular tone, and asymmetric growth failure that appears shortly after neurodegeneration begins. EEGs are moderately or severely abnormal, and hypsarrhythmia occurs frequently [Venta-Sobero et al., 2004]. Ophthalmologic findings common in Menkes’ disease include myopia, strabismus, and problems with visual fixation and tracking.

Connective tissue disorders are common in the three variants. Pelvic ultrasound and cystograms reveal diverticula of the urinary bladder [Daly and Rabinovitch, 1981; Harke et al., 1977]. Associated vascular disorders include lumbar and iliac artery aneurysms [Adaletli et al., 2005]. Skull and skeletal radiographs are notable for Wormian bones in the skull, metaphyseal spurring of the long bones, and anterior flaring or multiple fractures of the ribs [Adams et al., 1974; Capesius et al., 1977; Koslowski and McCrossin, 1979; Stanley et al., 1976]. Congenital skull fracture, related to global osteopenia, subdural hematoma, intrauterine growth delay, and lethal outcome, is reported in neonatal Menkes’ disease [Veit-Sauca et al., 2009]. MRI demonstrates white matter changes and impaired myelination, cerebral blood vessel tortuosity, ventriculomegaly, and diffuse cerebral atrophy [Faerber et al., 1989; Johnsen et al., 1991]. White matter lesions localized in the deep periventricular white matter in the absence of diffuse cortical atrophy [Lee et al., 2007], and transient temporal lobe lesions related to vasogenic and cytotoxic edema [Ito et al., 2008] were reported.

Patients with occipital horn syndrome have hyperelastic skin and may develop other connective tissue disorders, including aortic aneurysms, hernias, bladder diverticula, and skeletal abnormalities, which most likely result from lysyl oxidase deficiency. The characteristic occipital horns are symmetric exostoses protruding from the occipital bone and pointing down. These may be present around 1–2 years of age, but are usually detected only around 5–10 years of age and continue to grow up to early adulthood [Tümer and Moller, 2009]. These individuals also have mild mental retardation and autonomic dysfunction that manifests as syncope, hypothermia, and diarrhea [Byers et al., 1980]. The severity of mild Menkes’ disease falls between that in the classic form and the much milder occipital horn syndrome [Procopis et al., 1981]. It is important for patients with milder variations of Menkes’ disease to have frequent vision examinations because ophthalmologic problems can greatly impair functioning of affected individuals.

Clinical and Biochemical Diagnoses

The clinical diagnosis of Menkes’ disease is supported by specific laboratory findings [Poulsen et al., 2002]. Early diagnosis of affected newborns is necessary for the institution of appropriate therapy and survival. A high index of suspicion based on clinical grounds is essential because supportive laboratory findings may be problematic in the first few months of life.

Initially, few or no neurologic manifestations occur [Gunn et al., 1984]. Kinky hair with light hair pigmentation is most suggestive of Menkes’ disease. The pili torti pathognomonic for Menkes’ disease are usually seen only in older patients. The diagnosis should be considered for a neonate born after premature labor and delivery with large cephalohematomas, unexplained hypothermia or hypoglycemia, jaundice requiring phototherapy, pectus excavatum, and inguinal or umbilical hernias [Kaler, 1998].

The classic biochemical markers of Menkes’ disease are low serum levels of copper and ceruloplasmin. Decreased intestinal copper absorption leads to low levels of copper in the plasma, liver, and brain. In contrast, copper stores are increased in the duodenum, kidney, spleen, pancreas, and skeletal muscle [Heydorn et al., 1975; Horn, 1984; Williams and Atkins, 1981].

Early laboratory diagnosis of neonates is complicated by the fact that copper and ceruloplasmin concentrations are normally low in healthy newborns, and these values can overlap with those typically found for older patients with Menkes’ disease [Lockitch et al., 1986, 1988]. Copper egress assay in cultured fibroblasts is the definitive diagnostic study at this age, but the test is lengthy and requires several weeks of cell culture. Rapid diagnosis of Menkes’ disease can be achieved by measurement of plasma catecholamines or polymerase chain reaction (PCR) detection of known deletions or point mutations in the ATP7A gene. The copper deficiency in Menkes’ disease affects the function of many enzymes requiring copper. Dopamine mono-oxygenase is one of the cuproenzymes that is dysfunctional in Menkes’ disease, resulting in abnormal plasma catechol concentrations in newborns and fetuses [Kaler et al., 1993a, 1993b, 1993c]. The plasma catecholamine profile is considered to be the most rapid and reliable way to diagnose Menkes’ disease during the neonatal period. Patients with Menkes’ disease have high plasma dopamine and low norepinephrine levels. Considered alone, neither dopamine nor norepinephrine levels have perfect sensitivity, whereas the ratio of dopamine to norepinephrine is high in all affected patients. Analogously, levels of the dopamine metabolite, dihydroxyphenylacetic acid, and the norepinephrine metabolite, dihydroxyphenylglycol, were imperfectly sensitive, whereas the dihydroxyphenylacetic acid to dihydroxyphenylglycol ratio is high in all patients. Plasma dihydroxyphenylalanine and the ratio of epinephrine to norepinephrine levels are high in affected neonates [Goldstein et al., 2009]. Increased urine ratios of homovanillic acid/vanillylmandelic acid (HVA/VMA) have been proposed as a screening tool for Menkes’ disease also [Matsuo et al., 2005]. PCR methods are helpful in the diagnosis of partial deletions and point mutations when they are already identified for a specific family. The DNA-based technologies used for screening ATP7A for point mutations are chemical cleavage mismatch detection [Das et al., 1994], single-strand conformational polymorphism analysis [Tumer et al., 1997], and dideoxy fingerprinting [Moller et al., 2000]. PCR-based methods for screening ATP7A for large partial deletions include multiplex PCR, genomic PCR, and reverse transcriptase PCR [Poulsen et al., 2002]. These PCR methods are useful in neonatal and prenatal diagnosis, and they are helpful for carrier screening.

Prenatal diagnosis and identification of carrier status are important for families that are at risk for Menkes’ disease. Assays looking for increased levels of copper in cultured fibroblasts also can be used for prenatal diagnosis and for testing of potential carriers [Goka et al., 1976; Poulsen et al., 2002; Tümer et al., 2003], but random X-inactivation renders carrier testing using this technique uninformative when negative. The only definitive tests that can exclude this disease are DNA-based assays.

An important component of all variants of Menkes’ disease is connective tissue involvement. Deoxypyridinoline is a cross-linking residue of type I collagen and is a good marker for lysyl oxidase activity, the cuproenzyme deficiency that is responsible for the connective tissue disorders observed in these patients. Deoxypyridinoline has been proposed as a marker for the presence of connective tissue disorders associated with Menkes’ disease [Kodama et al., 2003].

Genetics

Menkes’ disease is an X-linked disease, and one-third of cases are thought to represent new mutations [Haldane, 1935]. The gene responsible for Menkes’ disease, ATP7A, was discovered in 1993 [Chelly et al., 1993; Mercer et al., 1993; Vulpe et al., 1993]. The gene encodes the copper transporting ATPase known as ATP7A [Mercer, 1998]. ATP7A is part of a highly conserved family of cation-transporting ATPase proteins [Odermatt et al., 1993; Pederson and Carafoli, 1987]. This protein family includes the protein WD that is defective in Wilson’s disease. The ATP7A gene (formerly known as MNK) is located at Xq13.2–q13.3 [Tümer et al., 2003]. It has 23 exons and is about 140 kb long [Dierick et al., 1995]. ATP7A is 8.5 kDa. About 357 different mutations have been identified and are distributed as follows: nonsense mutations 18 percent, splice-site mutations 16 percent, missense mutations 17 percent, partial insertions 4 percent, minus mutations 5 percent, partial deletions 17 percent, chromosomal mutations 1 percent, and small indels 22 percent [Møller et al., 2009].

Occipital horn syndrome results from point mutations (75 percent), chromosomal rearrangements (about 1 percent), and large or partial deletions (about 15 percent) in the ATP7A gene [Liu et al., 1999; Tümer et al., 1999]. Severe classic Menkes’ disease RNA contains low levels of active ATP7A mRNA, whereas mild variants of Menkes’ disease have higher levels of active ATP7A mRNA. Occipital horn syndrome has 20–35 percent residual ATP7A mRNA in cultured cells, whereas classic Menkes’ disease has lower levels of mRNA in the cells [Kaler, 1994, 1998]. Decreased mRNA levels result from premature stop codons, missense mutations, frame shifts, and other deletions or mutations that affect RNA splicing [Das et al., 1995]. Patients with the identical deletion or mutation may have different clinical outcomes [Tümer et al., 2003]. This finding suggests that other modifying genes or proteins must play a role in the pathogenesis of this disease. Alternatively, other pathways for copper transport may exist at the cellular level.

Alternate splice products of ATP7A exist. One of the splice variants lacks exon 10, but the product is in-frame. This exon encodes transmembrane domains 3 and 4. This protein may act as a copper transporter, but it may not be able to function as a copper-transporting ATPase. It is also possible that alternate exons exist within intronic sequences.

Biochemistry

The genetic mutation of the ATP7A gene leads to problems with copper transport within the body. Normally, ATP7A allows efflux of copper from gut epithelium into the portal circulation, transport of copper across the blood–brain barrier, and transfer of copper that is reabsorbed by the kidney back into circulation. ATP7A protein deficiency results in an inability of the body to absorb copper from the gastrointestinal tract in amounts required to satisfy nutritional needs, as well as in impaired use and handling of the copper that is absorbed from the intestine. Copper is mobilized from the cytoplasm of cells so that it can be incorporated into secretory pathways [Tümer et al., 2003]. The mutation of a copper-transporting ATPase causes impaired cellular copper efflux, leading to increased intracellular copper concentrations. Patients with Menkes’ disease have high concentrations of copper in gut epithelial cells, and they absorb little copper from their diet. Copper accumulates in kidney tubules. At high levels, copper causes lipid peroxidation, protein cleavage, enzyme inhibition, and DNA damage. Normally, basal intracellular stores are maintained at low levels [Rae et al., 1999; Voskoboinik and Camakaris, 2002]. The disease is caused by a decreased amount of the ATP7A protein, or it can result from alterations to the protein that impair its ability to transport copper.

ATP7A has 6–8 transmembrane domains, and it transports Cu²⁺ ions using energy from adenosine triphosphate (ATP) hydrolysis [Vulpe et al., 1993]. The protein has several known motifs or domains within it, such as the ATP binding domain. The phosphorylation motif (DKTG) contains an aspartic acid that becomes phosphorylated during the protein’s cycle (present in all P-type ATPases). The cation transduction motif (CPC) features a conserved proline, which plays a role in the conformation changes that occur with cation transport [Silver et al., 1989]. There are six metal binding sites (MBSs) in the amino-terminal region of ATP7A. The consensus sequence of the MBS is GMxCxxC. Although human ATP7A has six different forms of MBS, microbial cells show that only one or two are necessary for a functional protein [Odermatt et al., 1993; Solioz et al., 1994]. Human ATP7A does not need an MBS, demonstrated by the fact that mutations in all six MBSs do not stop ATP7A from functioning [Forbes et al., 1999; Payne and Gitlin, 1998; Tsivkovskii et al., 2002]. The MBS may act as a sensor for intracellular copper levels, and it may play a regulatory role when concentrations are low [Goodyer et al., 1999; Strausak et al., 1999; Voskoboinik and Camakaris, 2002]. ATP7A also features a magnesium-binding motif (TGE), the “hinge” domain of the protein, and a phosphorylation motif (DKTG).

ATP7A-encoded mRNA is found is most cell types, but it is missing in liver. ATP7B, the protein mutated in Wilson’s disease (Table 41-2), transports copper in the liver [Vulpe et al., 1993; Paynter et al., 1994]. ATP7A and ATP7B are members of the P-type ATPases group IB family, as are bacterial heavy metal transporters [Tsivkovskii et al., 2002; Voskoboinik et al., 2001]. ATP7A has been localized to the trans-Golgi compartment under basal conditions [Petris et al., 1996]. This is consistent with the ability of ATP7A to supply cuproenzymes (i.e., lysyl oxidase) that are in secretory pathways. There is continuous recycling of ATP7A between the plasma membrane and the trans-Golgi compartment [Petris and Mercer, 1999]. ATP7A traffics to the plasma membrane with increased extracellular copper concentrations and is endocytosed back to the trans-Golgi compartment after extracellular levels decrease to normal. Copper-dependent vesicular trafficking moves ATP7A from the plasma membrane to the trans-Golgi network and back. Lower organisms have two copper ATPases, but this is unnecessary in humans because ATP7A traffics between the two areas where such ATPases are needed [Yuan et al., 1997]. Cu(I) may be the type of copper used by ATP7A as its substrate. It is unknown whether Cu(I) becomes Cu(II) before or after it is released in the Golgi lumen.

Table 41-2 Comparison of Menkes’ Disease and Wilson’s Disease

Menkes’ disease is caused by mutations that result in a substitution of highly conserved amino acids or those in highly conserved motifs. Any mutation that affects the structure and function of the ATP7A gene will lead to disease [Guy et al., 2001]. Mutations that induce Menkes’ disease include those that abolish the Mg²⁺ binding domain [Seidel et al., 2001] and those that change the cation transduction motif. Classic Menkes’ disease is usually caused by a premature truncation, typically occurring before the first transmembrane domain and resulting in loss of all catalytic activity. Occipital horn syndrome and mild Menkes’ disease usually result from missense or splice mutations. It is unknown how much catalytic activity is needed to result in a milder phenotype (i.e., ATP7A still able to absorb sufficient amounts of intestinal copper and enable its delivery to the requisite enzymes) [Tümer et al., 1997]. A case study of a patient with occipital horn syndrome showed that the splice mutation allowed 2–5 percent of ATP7A transcripts to be produced. This amount of protein was sufficient to allow partial absorption of copper from the gut epithelium and partial transport across the blood–brain barrier [Møller et al., 2000]. There was, however, too little protein for lysyl oxidase to function correctly. Between 2 and 5 percent of ATP7A activity is the proposed amount of protein necessary to decrease the severity of the Menkes’ disease phenotype.

Most of the clinical manifestations of Menkes’ disease can be explained by understanding which cuproenzymes are affected. Intracellular copper is necessary for oxidative reactions. Most clinical symptoms are caused by dysfunction of dopamine mono-oxygenase, peptidylglycine mono-oxygenase, cytochrome c oxidase, lysyl oxidase, and Cu/Zn superoxide dismutase (SOD), tyrosinase, and sulfhydryl oxidase.

Dopamine mono-oxygenase, also known as dopamine β-hydroxylase, is part of the catecholamine biosynthetic pathway. In Menkes’ disease, there is complete or partial deficiency of dopamine β-hydroxylase. This deficiency causes abnormal plasma and cerebrospinal fluid patterns. The degree of deficiency can be evaluated by looking at norepinephrine concentrations or the ratio of dihydroxyphenylalanine to dihydroxyphenylglycol. Cases of Menkes’ disease with deficient dopamine β-hydroxylase exist that have normal plasma and cerebrospinal fluid concentrations of norepinephrine. It is unknown why compensatory mechanisms are active in some patients but not in others. In the mouse model of Menkes’ disease, normal concentrations of norepinephrine are observed in certain areas of the brain [Prohaska and Bailey, 1994]. The dopamine β-hydroxylase deficiency causes temperature instability, hypoglycemia, eyelid ptosis, and loss of sympathetic adrenergic function [Biaggioni et al., 1990; Robertson et al., 1986].

Peptidylglycine mono-oxygenase is a cuproenzyme necessary for the removal of the carboxyl-terminal glycine residue from neuroendocrine precursors, including gastrin, cholecystokinin, vasoactive intestinal peptide, corticotropin-releasing factor, calcitonin, vasopressin, and thyrotropin-releasing hormone [Eipper et al., 1983, 1992]. When this enzyme is deficient, these neuroendocrine factors have 100- to 1000-fold decreased bioactivity. Affected animal models of Menkes’ disease have decreased peptidylglycine mono-oxygenase activity within their brains [Prohaska and Bailey, 1995]. The decreased activity of these neuroendocrine factors contributes to the phenotype of Menkes’ disease.

Cytochrome c oxidase is a copper-dependent enzyme that is deficient in Menkes’ disease. The decreased cytochrome c oxidase activity causes a subacute necrotizing encephalomyelitis without the severe lactic acidemia that is generally associated with complex IV defects [DiMauro et al., 1990; Robinson, 1989; Robinson et al., 1987]. Peripheral hypotonia and muscle weakness in patients with Menkes’ disease is partially caused by decreased activity of this enzyme.

The normal function of lysyl oxidase (protein lysine 6-oxidase) is to deaminate lysine and hydroxylysine during the first step of collagen cross-link formation [Siegel, 1979]. ATP7A may be needed to transport copper into the trans-Golgi for use by lysyl oxidase. Deficiency of lysyl oxidase decreases the strength of connective tissue in certain tissues or organs and leads to a host of connective tissue disorders that usually are associated with Menkes’ disease, including vascular tortuosity [Royce and Steinmann, 1990], bladder diverticula [Daly and Rabinovitch, 1981; Harke et al., 1977], and gastric polyps. ATP7A deficiency affects lysyl oxidase more profoundly than any of the other cuproenzymes [Gacheru et al., 1993]. It is possible that other cuproenzymes can acquire copper from cytoplasmic carriers without using ATP7A (an intermediate carrier), and this may explain why some mutations of ATP7A lead to the less severe occipital horn syndrome.

Cu/Zn SOD is also deficient in this disease [Rohmer et al., 1977]. Lowered levels of Cu/Zn SOD can increase susceptibility to damage by oxygen free radicals. It is not known whether decreased Cu/Zn SOD causes any developmental regression, because animal models without Cu/Zn SOD have normal development [Reaume et al., 1996] Postmortem studies of patients with Menkes’ disease have found increased levels of manganese SOD, which may be a compensatory change for the decreased levels of Cu/Zn SOD [Shibata et al., 1995]. It is unknown whether the decrease in this enzyme contributes to the phenotype observed in Menkes’ disease. Tyrosinase and sulfhydryl oxidase act on pigment formation and cross-linking of keratin, respectively; hence the manifestations of hypopigmentation, abnormal hair, and dry skin [Horn and Tümer 2002].

Pathology

Pathogenesis

Defective copper trafficking explains most changes seen in Menkes’ disease. The body’s nutritional needs for copper are crucial in the first 12 months of life. Brain growth and motor development are most rapid during this phase. Neurodevelopment and motor development are processes that require copper [Berg, 1994]. Expression of BCL2, an anti-apoptotic protein, is decreased in children with Menkes’ disease. Decreased levels of copper lead to mitochondrial damage that causes decreased levels of BCL2 [Rossi et al., 2001]. These decreased levels of BCL2 may explain the increased number of apoptotic cells seen in the brain of the animal model, as well as the mechanism behind neurodegeneration in Menkes’ disease.

Phenotypic differences in animal models and in humans appear to depend on the amount of functional ATP7A-encoded mRNA. During different stages of development, the various organs and tissues have different copper requirements. If the requirement is not met, the organ or tissue undergoes developmental failure.

Animal Models

Complete deficiency of the Menkes’ protein in mice causes lethality before birth, whereas humans can survive for a limited time without ATP7A [Mercer, 1998]. The murine homolog of ATP7A is Atp7a [Levinson et al., 1994; Mercer et al., 1994]. Just as in humans, there are a variety of observed phenotypes, with most of them reflecting their human counterparts. The severity of the murine phenotypes depends on the amount of ATP7A-encoded mRNA. Four murine genotypes are used to study Menkes’ disease:

Treatment

Treatment of Menkes’ disease is limited to supportive therapy and supplementation with parenteral copper histidine. Unlike nutritional copper deficiency, Menkes’ disease cannot be cured by copper replacement. Copper histidine is more effective in patients with milder phenotypes. A small subset of patients can achieve normal neurodevelopment. Early detection and intervention are critical for the copper histidine treatment to have an effect. Daily copper injections may improve the outcome in Menkes’ disease if started within days after birth [Kaler et al., 2008]. Many individuals with Menkes’ disease fare poorly, with little or no developmental improvement despite early intervention. Older patients with neurologic signs of Menkes’ disease demonstrate little improvement with copper histidine replacement. Therapy may, however, reduce irritability and allow for calmer sleeping patterns and minor improvements in personal and social development [Kaler, 1998]. These minor responses to therapy may help lessen the burden placed on caretakers. Copper histidine therapy helps the neurologic signs and symptoms of Menkes’ disease, but it has no effect on connective tissue disorders associated with this disease.

Early copper histidine replacement therapy is most effective in patients with milder phenotypes who have some residual copper transport activity. These milder variants result from mutations that do not affect the regions of ATP7A necessary for catalytic activity and allow for residual levels of copper transport. The beneficial response to copper histidine therapy can be predicted by PCR-based analysis of mutations.

Future Directions

Improved therapeutic strategies are critical for better outcomes. Biochemical approaches to bypass the block in copper absorption from the intestine and allow requisite copper absorption are needed. Affected individuals must be identified early to initiate treatment before manifestations of neurologic signs and symptoms occur. This is the period when treatment has the most potential. ATP7A is necessary for the transport of copper through the blood–brain barrier. To prevent neurologic developmental regression, other ways of delivering copper stores to the brain must be identified. The symptoms of Menkes’ disease result from the dysfunction of multiple cuproenzymes, and copper must be made available to these enzymes to reduce the non-neurologic symptoms experienced by these patients [Kaler, 1998]. Although advances have been made in early detection of affected individuals, there has been little progress in finding ways to deliver copper to the necessary cuproenzymes or through the blood–brain barrier. Because classic Menkes’ disease is the result of the partial or complete absence of the traditional copper transport pathway, identification of alternate routes may allow the delivery essential for normal development.

History

While visiting a colleague’s laboratory in Hamburg, American neuropathologist Bernard Alpers conducted a detailed study of a 4-month-old girl who died after a month of intractable seizures. The patient had a normal birth and normal development until age 4 months. There was necrosis of the deep gray matter nuclei and diffuse loss of ganglion cells in the third layer of the cortex. Alpers believed that the degeneration had a toxic cause. Many similar case reports have been made since then. Huttenlocher et al. [1976] were first to report that the neuronal degeneration was associated with liver failure. The pattern of inheritance has been postulated to be autosomal-recessive [Sandbank and Lerman, 1972]. A mitochondrial defect leading to Alpers’ disease has also been hypothesized [Naviaux et al., 1999]. The disease most likely has several causes, accounting for subtypes such as hepatocerebral or myopathic Alpers’ disease.

Clinical Description

Alpers’ disease is a clinical diagnosis that is documented on MRI but confirmed on postmortem examination. The disease is a diagnosis of exclusion, but specific clinical findings suggest its presence. Characteristic manifestations are liver failure, refractory seizures, and psychomotor retardation [Wefring and Lamvik, 1967]. Patients with Alpers’ disease are developmentally normal initially. Onset can occur between 1 month and 25 years of age [Harding et al., 1995]. Onset is more common during a patient’s infancy and adolescence, with most cases showing initial symptoms in infancy, usually before the age of 5. Death occurs between 3 months and 12 years of age, with most patients dying before the age of 3 years. The course of Alpers’ disease is variable, alternating between periods of development, degeneration, and mild recovery or stasis. Rare cases of Alpers’ disease have been reported in individuals as old as 25 years. Because of the variable age of onset, Alpers’ disease has been categorized as having juvenile, infantile, and prenatal forms [Frydman et al., 1993; Harding et al., 1995; Montine et al., 1995; Simonati et al., 2003; Worle et al., 1998]. Prenatal Alpers’ disease has been described in one family, and it is characterized by microcephaly, intrauterine growth retardation, retrognathia, joint limitations, and chest deformity. The infantile form manifests with early onset, a slowly progressive course, and late-occurring severe signs and symptoms. The juvenile form of Alpers’ disease is identical to the infantile form, but it is notable for a peripheral ataxia resulting from central and peripheral sensory axonopathy.

In addition to the classic triad of symptoms (i.e., psychomotor retardation, intractable seizures, and liver failure), patients with Alpers’ disease may have other manifestations. Hypotonia may occur initially [Egger et al., 1987]. Ataxia, febrile illness, and cortical blindness are less common [Naviaux and Nguyen, 2004]. A progressive ataxia that involves the sensory pathways has been described. This occurrence is similar to other sensory neuropathies caused by mitochondrial disorders [Fadic et al., 1997].

Liver failure is a complication that manifests late in the course of the disease [Smith et al., 1996], and it is usually the cause of death. Most cases of liver failure in these patients were attributed to hepatotoxicity caused by antiepileptic drugs, specifically valproic acid. However, some cases of liver failure and identical hepatic histology have occurred in patients who did not receive antiepileptic drugs. Valproic acid may cause hepatotoxicity in some cases of Alpers’ disease, but it cannot explain all cases. Orthotopic liver transplantation has not been helpful in cases with liver failure. Liver transplantation in patients with Alpers’ disease has been associated with neurologic deterioration [Delarue et al., 2000; Kayihan et al., 2000]. Complications of neuronal degeneration can lead to death from respiratory failure or primary hypoventilation.

Clinical Diagnostic Tests

Initial laboratory test results can be normal. Liver function test results can be elevated initially, although this finding is rare [Egger et al., 1987]. Values are elevated in the later stages of the disease because of liver cirrhosis. There are no specific in vivo serum or cerebrospinal fluid markers for Alpers’ disease, and the diagnosis must rely on neuropathologic evaluation. Lactate levels may be raised in the blood, and the protein and lactate levels may be elevated in the cerebrospinal fluid [Worle et al., 1998]. Positive cerebrospinal fluid oligoclonal bands, and very high immunoglobulin (Ig) G synthesis rate and IgG index were also reported [Bao et al., 2008]. Elevated cerebrospinal fluid neopterin, interleukin (IL)-6, IL-8, interferon (IFN)-c, reduced cerebrospinal fluid 5-methyltetrahydrofolate (5-MTHF), and increased serum, as well as cerebrospinal fluid folate receptor blocking autoantibodies, are present [Hasselmann et al., 2009].

Cranial CT scans demonstrate progressive atrophy and low densities in the occipital and temporal lobes. There is involvement of the cortex and the white matter [Kendall et al., 1987; Flemming et al., 2002]. Generalized atrophy is common throughout the brain in the later stages of the disease. Multiple findings are apparent on conventional MRI scans. These include diminished white matter and cortical thinning of the frontal, posterotemporal, and occipital lobes [Barkovich et al., 1993]. Lesions of the thalamus also have been reported. Occipital lobe atrophy is widespread. Proton MR spectroscopy reveals increased cerebrospinal fluid levels of lactate [Charles et al., 1994] and a reduced N-acetylaspartate to creatine ratio. The increase in lactate marks the switch from oxidative to anaerobic metabolism. N-acetylaspartate is an indicator of neuronal viability [Neumann-Haefelin et al., 2000].

EEG findings correlate with the lesions seen on MRI that are most apparent in the occipital lobes. The EEG of Alpers’ disease is described as slow (<1 Hz) with high-amplitude activity (0.2–1.0 mV). Lower-amplitude polyspikes are also recognized on the EEG. This pattern is seen in 75 percent of patients, but it may be present only transiently in periodic bursts [Martinez-Mena et al., 1998]. These bursts increase in number and duration as the disease progresses. Unilateral occipital rhythmic high-amplitude delta with superimposed (poly)spikes (RHADS) was described in convulsive status epilepticus in Alpers’ patients [Wolf et al., 2009]. Mild axonal sensory neuropathy was reported, as was central-peripheral sensory axonopathy in a juvenile case of the syndrome [Simonati et al., 2003]. There can be a loss of visual-evoked potentials [Martinez-Mena et al., 1998]. The triad of requisite clinical symptoms and MRI and EEG findings suggests the diagnosis of Alpers’ disease.

Pathology

Alpers originally described the neuronal pathology as cortical lesions with reactive gliosis, demyelination, nerve cell loss, spongy degeneration, and accumulation of neutral lipids. The occipital cortices are usually involved, and involvement can be symmetric or asymmetric. Patchy cerebral cortical destruction usually is worst within the striate cortex. Striate cortex involvement is a hallmark of Alpers’ disease [Harding, 1990; Harding et al., 1986]. Because of the destruction of the visual cortices, patients often have cortical blindness [Charles et al., 1994; Dietrich et al., 2001; Parsons et al., 2000]. Multiple case studies have described the cortical destruction as neuronal loss, spongiosis, astrocytosis, and gliosis. Milder changes are seen in the parietal cortices [Montine et al., 1995]. Necroses in the hippocampi, lateral geniculate nuclei, amygdala, substantia nigra, and dorsal columns may be evident. In patients with severe liver failure in the late stages of Alpers’ disease, Alzheimer’s disease type II astrocytes are seen. The white matter is only minimally affected.

Premortem liver biopsies reveal lobular disarray, microvesicular steatosis, and inflammation with acute and chronic hepatocyte necrosis [Narkewicz et al., 1991]. Common hepatic findings at autopsy are fibrosis, regenerative nodules, hepatocyte dropout, bile duct proliferation, fatty changes, and bile stasis. Pancreatitis is infrequently observed in these patients. Muscle biopsies may contain ragged red fibers and cytochrome c oxidase-negative fibers when Alpers’ disease is caused by a mitochondrial defect. A subset of patients with Alpers’ disease lacks any detectable energy metabolism defect and has normal hepatic histology [Frydman et al., 1993].

Biochemistry

The clinical manifestations of Alpers’ disease have several causes. Abnormalities within the citric acid cycle of leukocytes, cultured fibroblasts, and hepatocytes due to a defect in pyruvate metabolism or mitochondria have been suggested [Gabreels et al., 1984; Prick et al., 1981]. These defects result in deficiencies of cytochrome c oxidase, pyruvate cocarboxylase, and mitochondrial electron transport chain complex I. Mitochondrial DNA (mtDNA) polymerase gamma activity is less than 5 percent of normal in muscle and liver cells [Naviaux and Nguyen, 2004]. Immunoreactive subunits of the mtDNA polymerase are still present in patients. The deletion that causes the reduced amount of mtDNA polymerase is small because detectable levels of protein are still present. Southern blot analysis of two infants with Alpers’ disease revealed decreased mtDNA in muscle, liver, brain, and fibroblasts. The loss of mtDNA may be tissue-specific [Tesarova et al., 2004].

Genetics

The mode of transmission of Alpers’ disease is understood, and it is consistent between families. Case studies and biochemical evidence support autosomal-recessive inheritance and maternal or mitochondrial inheritance patterns. Cases with autosomal-recessive inheritance patterns have no mitochondrial deficiencies [Harding, 1990]. Simonati et al. [2003] observed that, although the disease affects both genders, there appears to be a mild male predominance. The same group also observed that children with juvenile Alpers’ disease do not have depletion or significant mutations in mtDNA.

Two mutations within the polymerase gamma gene (POLG) are associated with Alpers’ disease: G2899T and G1681A. G2899T causes a premature stop codon (Glu873Stop). Some Alpers’ disease patients are heterozygous for the POLG mutation Glu873Stop. It has been theorized that this stop codon is incomplete and allows a small number of ribosomes to be read. The number of ribosomes that are able to bypass the premature stop codon may be regulated in an age- and tissue-dependent manner, allowing Alpers’ disease to manifest at different ages and in different tissue types. In Drosophila melanogaster, proteins are regulated in a tissue-specific manner by changing the ratio of short to long forms of the protein [Robinson and Cooley, 1997]. The premature stop codon in POLG may change the ratio and cause tissue-dependent regulation.

The G1681A causes an Ala167Thr substitution. This corresponds to the linker region of the POLG protein. POLG is located on chromosome 15q24–26 [Zullo et al., 1997]. No other polymerase can be substituted for decreased POLG activity. POLG is the only polymerase of the 15 known DNA polymerases that has a mitochondrial import signal. Mutations within POLG are found in other mitochondrial diseases: progressive external ophthalmoplegia (autosomal-recessive and dominant forms) [Van Goethem et al., 2002, 2003], ophthalmoparesis, sensory ataxia, neuropathy, dysarthria, and male infertility. Children homozygous for the G2899T mutation are affected with Alpers’ disease, whereas control groups consisting of patients with neuromuscular diseases or patients with other mitochondrial diseases do not have this mutation. Many patients with Alpers’ disease are heterozygous for G1681A, but the mutation is not specific. Heterozygous G1681A mutations are also seen in ophthalmoparesis, both types of progressive external ophthalmoplegia, and in the asymptomatic parents of children with Alpers’ disease.

Pathogenesis

Alpers’ disease was originally believed to be a complication of concomitant hypoxia and epilepsy. Comparisons with status epilepticus cases are not valid because the cortical lesions in the two diseases differ significantly. Others have proposed that the underlying defect is metabolic. Alpers’ disease most likely can be caused by multiple metabolic defects. The cerebral destruction of the striate cortex in Alpers’ disease may result from a novel form of hepatocerebral toxicity, which is different from that seen in Wilsonian hepatocerebral degeneration. The possibility of a viral infection has been raised after brain tissue from a patient who died from Alpers’ disease was injected into hamsters, and caused a spongiform encephalopathy [Rasmussen et al., 2000].

Management

There is no known treatment for Alpers’ disease. It is important to recognize this disease early and avoid the use of valproic acid, because this increases the incidence of hepatotoxicity in these patients. Plasma alanine aminotransferase (AAT) can be used as an index of liver cell damage. Vigabatrin, which suppresses AAT activity, is better avoided in Alpers’ patients [Williams et al., 1998]. Although some patients may respond to antiepileptic medications early in their disease course, the seizures become refractory to treatment. Treatment with oral leucovorin (5-formyl-tetrahydrofolate) in a patient with increased serum as well as cerebrospinal fluid folate receptor-blocking autoantibodies was initiated at 0.25 mg/kg b.i.d., and later increased to 4 mg/kg twice daily; this resulted in improvement of seizure frequency and communicative abilities [Hasselmann et al., 2009]. Ketogenic diet is reported to cause clinical and EEG improvement [Joshi et al., 2009].

History and Terminology

The first clinical description of neuronal ceroid-lipofuscinosis was that of the juvenile form by Stengel [1826]. This was soon followed by clinical and pathologic descriptions by Batten, Mayou, Spielmeyer, Vogt, and Sjögren [Batten and Mayou, 1915; Mayou, 1904; Spielmeyer, 1923; Vogt, 1909]. The CLN3 gene responsible for the juvenile form (JNCL) was discovered in 1995 [Lerner et al., 1995]. The late infantile form (LINCL) was described by Jansky and Bielchowski in 1908 and 1913 [Jansky, 1908]. These two variants were previously referred to as Batten’s disease, a term that now refers to all variants. The adult form, or Kufs’ disease, an early-onset dementia with seizures and absence of visual findings, was described in 1925 [Dom et al., 1979]. The adult form (ANCL) has been described in sporadic cases and familial cases, with some families suggesting a dominant pattern of inheritance. Chromosomal location of the CLN4 gene or genes responsible for the adult disease remains unknown.

The term neuronal ceroid-lipofuscinosis was introduced by Zeman and Dyken in 1969 as a descriptive term referring to the autofluorescent, waxy, dusky lipid accumulating in neuronal endosomes, reminiscent of lipofuscin, the aging pigment [Zeman et al., 1970]. The infantile form (INCL) was described by Hagberg and then by Haltia and Santavuori in 1973 [Hagberg et al., 1968; Haltia et al., 1973a, 1973b; Santavuori et al., 1973]. The CLN1 gene was identified in 1995, followed by the CLN2 gene responsible for LINCL [Sleat et al., 1997]. Other types have been described since then, including variant late infantile forms and early juvenile forms due to defects in the CLN5, CLN6, and CLN8 genes [Gao et al., 2002; Ranta et al., 2001; Savukoski et al., 1998; Wheeler et al., 2002]. A CLN9 form is also described that is clinically similar to the juvenile form [Lin et al., 2001; Schulz et al., 2004]. The CLN9 gene remains to be characterized.

The terminology is confusing because it was established before many variants or clinical forms were defined and before any of the genes were identified. The terms INCL, LINCL, JNCL, and ANCL were initially chosen to separate the forms according to age of onset. This holds true for the main classic variants described. Since discovery of the various genes, many atypical cases have been described with variable ages of onset. It has, therefore, been decided to refer to these diseases by a unified nomenclature, illustrated by the following example: CLN1 disease infantile; CLN1 disease late infantile; and so on, as referred to in Table 41-3 [Mole et al., 2010].

The decision to name the genes CLN, as opposed to neuronal ceroid-lipofuscinosis or NCL genes, is most unfortunate, because it causes confusion with yeast cyclin genes, especially in scientific and medical literature searches. The decision to refer to all forms as Batten’s disease, although historically incorrect, is a simple and practical one, because of the length and wordiness of neuronal ceroid-lipofuscinoses. The term Batten’s disease has been universally adopted and accepted by family groups, private foundations, and U.S. government agencies.

Major Neuronal Ceroid-Lipofuscinosis Clinical Types or Syndromes

The clinical features, laboratory tests, pathology, biochemistry, and genetics are summarized for each of the major types (see Table 41-3).

CLN9-Deficient Juvenile-like Variant

Two German brothers and two American sisters of Serbian descent had been clinically diagnosed with the JNCL variant before identification of the CLN3 gene [Lin et al., 2001; Schulz et al., 2004]. When DNA from these cases was examined, it was determined that they had no defects in the CLN3 gene, and they had normal levels of CLN3 mRNA. Analysis of cDNA from CLN3-, CLN1-, CLN2-, and CLN6-deficient variants, together with cDNA from these unknown cases, using Affymetrix GeneChips, revealed a distinctive gene profile that grouped them together as a separate variant (Figure 41-1). Results of enzyme assays and molecular tests for all other known neuronal ceroid-lipofuscinosis variants were normal.

Fig. 41-1 Partial dendrogram depicting gene expression pattern for CLN1-, CLN2-, CLN3-, CLN6-, CLN9-deficient and normal fibroblast RNA.

Upregulated genes are red, downregulated genes are green, and no change from control is black. Notice the similarity of gene expression in the CLN9(1) and CLN9(2) patients. CLN, ceroid-lipofuscinosis, neuronal.

Clinical description

The clinical course for CLN9-deficient juvenile-like variant is almost identical to that of JNCL, with decreased vision occurring at age 4 years, cognitive decline at age 6 years, and ataxia and rigidity by age 9 years. These patients develop dysarthria and scanning speech, and are mute by age 12 years. One of the German brothers developed hallucinations and behavior problems. All four patients developed intractable seizures in their early to middle teens, which eventually diminished. Retinitis with pigmentary changes was documented in one of the two brothers. One of the brothers died of pneumonia at age 15 years, and the other died at age 19 years after suffering from swallowing difficulties and many seizures. The two sisters remain alive. One is in her early 20s and has been bedridden since she was 14 years old. She has well-controlled seizures and requires a feeding tube. The younger sister is now 12 years old and is following a similar course.

Clinical diagnostic tests

The ERG documented diminished wave amplitudes in one of the brothers with JNCL variant. The EEGs in all four patients demonstrated slowing and frequent polyspike discharges.

Electron micrographs of lymphocytes indicate numerous membrane-bound inclusions containing a mixture of electron-dense storage material and fingerprint patterns. CT and MRI scans showed progressive cerebral and cerebellar cortical atrophy. Abnormal signal intensity was observed in the periventricular white matter. A diagnostic brain biopsy from one of the sisters depicts neurons that contain granular osmiophilic inclusions and curvilinear bodies.

Pathology

The brain weight of the older brother at the time of death was 1140 g. Neurons were ballooned with fine, granular material. Large neurons remaining in the cerebral cortex and the deep gray structures were ballooned. The substantia nigra and nuclear thalami had moderate to severe astrogliosis, as did the spinal cord. The storage material was stained gray with Sudan black, and it had a yellow autofluorescence. The ultrastructure of neurons reveals GRODs and curvilinear bodies (Figure 41-2). Neurons also demonstrated immunoreactivity to subunit C or 9 of mitochondrial ATP synthase.

Fig. 41-2 Electron micrographs of CLN9-deficient frontal cortex.

A, Notice the secondary lysosomes with curvilinear bodies (magnification × 18,400). B, Granular osmiophilic deposits (GRODs) are denoted by arrows, and curvilinear bodies are denoted by a framed arrow (magnification × 45,400). C, Enlarged inset from B (magnification × 96,000).

Genetics

Chromosomal location of the underlying CLN9 gene remains unknown, but candidate genes include those encoding for proteins that impact the ceramide synthetic and catabolic pathways. The small number of affected cases has precluded traditional linkage analysis studies.

Biochemistry and cell biology

Although the gene remains unidentified, the biochemistry and cell biology of this variant are well understood. CLN9-deficient cells have a distinctive morphology and phenotype. Cells are small and rounded rather than elongated, and they have prominent nucleoli, as seen by filipin staining. They have rapid growth rates because of increased DNA synthesis, but they have an increased sensitivity to apoptosis. They have an adhesion defect, and a number of the genes involved in cell adhesion and apoptosis are dysregulated, as determined by gene profiling [Schulz et al., 2004]. Gene expression of cyclins A2, B1, C, E2, G1, and T2 were increased. Cyclin D1, encoded by a proto-oncogene involved in malignant transformation of breast tissue, was significantly downregulated, as was member 1A of the tumor necrosis factor superfamily.

Expression of subunits of cytochrome c oxidases and glutathione S-transferase was also increased. Ceramide, sphingomyelin, lactosylceramide, ceramide trihexoside, and globoside levels were decreased by 60–100 percent. The key regulating enzyme in the ceramide de novo synthetic pathway, serine palmitoyl transferase, was three- to fourfold upregulated, suggesting a block further downstream. The low levels of glycosphingolipids can explain the defect in cell adhesion observed in CLN9-deficient fibroblasts.

Management and treatment

Treatment is similar to that outlined for the JNCL variant. Flupirtine used empirically in the two sisters has led to stabilization of the course and better control of seizures. After the entire sphingolipid pathway and its perturbations are better defined for this variant, it will become possible to use drugs to correct defects in the pathway. CLN9 protein may be a regulator of dihydroceramide synthase. 4-hydroxyphenyl retinamide (4-HPR), an activator of the enzyme, could be developed as a treatment for CLN9-deficient patients [Schulz et al., 2006].