Neuropathology of Cockayne syndrome: Evidence for impaired development, premature aging, and neurodegeneration

Abstract

Global growth and development failure, premature, accelerated, pathologic aging, and neurodegeneration characterize Cockayne syndrome (CS) and the cerebro-oculo-facial-skeletal and xeroderma pigmentosum/CS syndromes which overlap CS partially in their genetic, somatic, and neuropathologic features. Mutations of CSA or CSB genes jeopardize transcription-coupled repair of damaged nuclear and mitochondrial DNA and resumption of replication and transcription. Resultant defective proteins or gene silencing eventuate in profound dwarfism and micrencephaly, cachexia, vasculopathy, and neurodegeneration. Cellular effects are highly selective. Purkinje cells may die by apoptosis and have grossly dystrophic dendrites. Neuronal death and axonal spheroids indexing neuronal pathology predominate in, but are not limited to, the cerebellum. Progressive loss of retinal, cochlear, and vestibular sensory receptors foster degeneration of ganglion cells and transneuronal brain degeneration. Some proliferating astrocytes are multinucleated and bizarre. Primary damage of oligodendrocytes and Schwann cells may – or may not – explain severe patchy myelin loss (“tigroid leukodystrophy”) and segmental demyelinating peripheral neuropathy. Age-related changes are minor in the brain, although precocious severe athero- and arteriolosclerosis are responsible for occasional strokes. Vasculopathology may contribute to myelin loss and to dystrophic mineralization of neurons and vessels, especially in basal ganglia and cerebellum. Understanding the genetics, biochemical, and cellular pathophysiology of these disorders remains fragmentary.

Introduction

Cockayne syndrome (CS) and related disorders constitute a complex of rare diseases characterized by deficient nucleotide excision-repair (NER). The pathophysiology of deficient NER for cell replication is now quite well understood, whereas the detail of the consequences of deficient NER for transcription, which plays the dominant role in CS, is less complete. One of the reasons is that there is a dearth of well-studied human cases, compared to a now rich literature on the molecular biology of NER disorders and its biochemical consequences in animal models and tissue culture.

Over the past 30 years, we and our colleagues have had the opportunity to study four cases of CS in clinical and pathological detail. We review the neuropathology of these cases and compare them to all those we were able to locate in the literature in order to illuminate major gaps in available information that future studies will need to fill. We point out pathologic features that support currently accepted pathophysiologic mechanisms and others that remain unexplained.

Cockayne syndrome (CS) is a progressive, autosomal recessive disorder that is usually present in infancy or early childhood. Most Cockayne neonates appear well formed but lag progressively behind age-peers because of severe postnatal growth failure and loss of fat. Affected children have sunken eyes, small poorly reactive pupils, sharp noses, carious teeth, and they sunburn easily. Increasing ataxia, spasticity, and a demyelinating neuropathy interfere with or preclude ambulation. Their motor, language, and cognitive development is quite delayed. Most are minimally verbal or nonverbal, but they typically have engaging personalities and few of them have seizures. Their hearing and eyesight eventually deteriorate due to cochlear pathology, corneal opacities, cataracts, retinitis pigmentosa, and optic atrophy. The disorder is slowly progressive. Dementia supervenes and most children die in late childhood, profoundly cachectic, of an intercurrent illness or, occasionally, hypertensive vascular disease. The nervous system, retina, inner ear, vascular, bony, and adipose tissues are disproportionately affected, whereas the function and structure of other organ systems are generally unimpaired despite their miniature sizes. The CS phenotype bespeaks a complex process of defective development, premature aging, and selective cell-type degeneration (Mizuguchi and Itoh, 2005).

In their comprehensive review of the clinical features of CS, Nance and Berry (1992) divided 140 published cases into three types: Type I, the most prevalent classical childhood disorder; Type II, an infrequent, very severe congenital or infantile variant that a recent review (Pasquier et al., 2006) suggests may not be quite as rare as originally suggested; and Type III, atypical later onset cases with prolonged survival.

The molecular basis of CS, which does not map systematically onto the clinical phenotypes, includes recessive mutations in either the CSA (CKN1 or ERCC8) gene (OMIM 216400) which accounts for some 20% of cases, or the CSB (CKN2 or ERCC6) gene (OMIM 133540) responsible for most of the remainder. The severity of the phenotype varies, including a UV sensitive freckled Japanese man who made no CSB protein but did not have the CS phenotype (Horibata et al., 2004). A minority of cases arises from mutations in one of the three xeroderma pigmentosum (XP) genes [XPB (ERCC6, OMIM 610651), XPD (ERCC3, OMIM 278730), or XPG (ERCC5, OMIM 278780)] in individuals who have clinical features of both disorders (Rapin et al., 2000). In some individuals XPB/CS is a relatively less severe disorder with a CS Type III phenotype, as at least three individuals survived to mid-adult life, but others have severe disease, depending in part on the location of the mutation and amount of protein produced (Oh et al., 2006). XPD/CS and, especially, XPG/CS are very severe lethal diseases of young children, some of whom have CS Type II clinical features (Lindenbaum et al., 2001). Rarely, mutation of CSB results in the cerebro-oculo-facial syndrome (COFS) (OMIM 214150), which shares many features with Type II CS or may be severe CS II. COFS may also arise from mutation of the XPD (Graham et al., 2001) or the ERCC1-XPF genes (Jaspers et al., 2007), and perhaps others. Like those with CS Type II, COFS infants have intrauterine or very early growth failure, with microcephaly, enophthalmos, cataracts, and other eye anomalies; they are severely dwarfed and mentally retarded, and die in early childhood. COFS infants differ from CS II in that they may not all develop cachexia and they have somewhat different dysmorphic facial features (Laugel et al., 2008). The arthrogryposis with which they are born denotes intrauterine lack of movement, which testifies to the severity and precocity of their neurologic involvement.

CS also overlaps with trichothiodystrophy (TTD—OMIM 601675). The TTD phenotype can be due to mutations of XPB, XPD, and two other genes TTDA and TTDN1 (Faghri et al., 2008). TTD, which is characterized in virtually all cases by sulfur deficient brittle hair and in some cases by ichthyosis, shares with CS growth failure, dysmorphism, photosensitivity, cataracts, cognitive impairment with an outgoing personality, and lack of susceptibility to cancer. Infants with TTD are more likely than those with CS to be born prematurely and to have experienced intrauterine growth retardation. Although children with TTD with cachexia have a 20-fold risk of dying before age 10 years of overwhelming infection, most are not cachectic. They do not have the progressive neurologic degeneration and premature aging of CS, but they do have mild white matter changes on imaging. If TTD is associated with intracranial calcification, MRI shows neither the dense calcifications of CS nor its dramatic atrophy of the white matter. As no autopsy report of TTD seems available, how much its neuropathology overlaps that of CS is unknown.

Overlapping clinical phenotypes in CS variants and related disorders preclude a fully satisfactory classification of cases reported without genetic characterization, especially as the same mutation of a gene does not lead reliably to a uniform phenotype (Oh et al., 2006). Among the factors that contribute to genotype–phenotype unpredictability (Boyle et al., 2008) is heterozygous mutations in different parts of the same gene, with differing consequences for transcription of messenger RNAs and amounts of protein synthesized (Nishiwaki et al., 2008). Double heterozygosity also affects differentially the non-coding RNAs that control the timing of gene expression.

Mutations in the CS and XP genes interfere differentially with the nucleotide excision-repair (NER) pathway because they code for many of the 28 proteins involved (Nouspikel, 2008). In addition, many of these proteins play additional complex metabolic roles. CSA and CSB proteins signal the need for NER of damaged DNA strand (or strands) so that accurate transcription can resume (Kraemer et al., 2007). CSB protein (less is known about CSA protein) also plays a critical role in the resumption of transcription following oxidative damage to nuclear and mitochondrial DNA, as well as in chromatin remodeling (Stevnsner et al., 2008). Signs of inadequate NER include excessive sensitivity to sun exposure, carcinogenesis, and the accumulation of repair proteins at sites of unrepaired DNA damage (Boyle et al., 2008).

Global genomic repair (GGR) hones in on single or double stranded bulky DNA-distorting lesions anywhere in the genome. It is much slower than transcription-coupled repair (TCR) of single strand errors in actively transcribing genes. Individuals with most XP mutations in whom GGR is compromised develop severe sunburn on minimal exposure, freckles, dyskeratoses, and skin cancers due to the accumulation of unrepaired mutations or gene rearrangements. In contrast, TCR is affected selectively in CS patients in whom GGR is competent and protects them from the skin and other malignancies of XP. A diagnostic hallmark of CS is extremely delayed recovery of RNA synthesis in cultured cells after irradiation with UV light, a finding used as a screening diagnostic test and for prenatal diagnosis of CS (Kleijer et al., 2006, Lehmann et al., 1993).

The four personal cases we review and compare to the available reports from the literature we were able to access include two typical Type I childhood cases, viz., Cases 2 (Soffer et al., 1979) and 3 (Gandolfi et al., 1984) studied some three decades ago; Case 4 was an adult survivor (Type III CS) (Rapin et al., 2006) and Case 1 a severely affected infant with XPG/CS (Lindenbaum et al., 2001, Moriwaki et al., 1996, Rapin et al., 2000). Our review convinced us that, besides differences in severity, there is no fundamental difference in pathology that precludes their being considered as a group.