Diagnosis of Xeroderma Pigmentosum and Related DNA Repair-Deficient Cutaneous Diseases

Xeroderma pigmentosum and related repair-deficient diseases

Xeroderma pigmentosum (XP) is a rare, human, autosomally inherited skin and neurodegenerative disease [1] that is associated with a very high incidence of skin and mucous membrane cancers due to exposure to normal sunlight. These cancers include squamous and basal cell carcinomas and melanomas, and are predominantly caused by exposure to ultraviolet B (UVB) radiation, although UVA cannot be excluded [2,3]. UVB (280–320 nm) is the shorter wavelength radiation in sunlight that is responsible for most sun-induced cancers in the general population, as well as in XP patients. The relative incidence of the various forms of skin cancers in XP patients is similar to that in the general population [4].

XP was the first nucleotide excision repair (NER)-related disease to be identified [1]. Currently, the complete family of NER-related diseases includes XP, XP with neurological complications, the XP variant (XPV), Cockayne syndrome (CS), cerebro-oculo-facio-skeletal syndrome (COFS), a mild UV-sensitive syndrome (UV^S), trichothiodystrophy (TTD), and the combinations XP/CS and XP/TTD [5–8]. These diseases show overlapping symptoms associated with cancer, developmental delay, immunological defects, neuro-degeneration, retinal degeneration, and premature aging.

The NER-related diseases are associated predominantly with failures in DNA repair or replication in cells that contain UV-induced photoproducts in their DNA. These photoproducts include the (5–5) and (6–6) cyclobutane pyrimidine dimers (CPDs) and the (6–4) pyrimidine pyrimidinone dimers (PDs) that can involve T and C pyrimidines [9]. When these photoproducts are unrepaired due to NER deficits, cytosine deamination and replication errors lead to characteristic C→T mutations (especially CC→TT mutations), which are found at high frequencies in p53 and other genes in solar-induced skin cancers of XP patients and others [10–16].

The diagnostic problem

Approximately one person per million suffers from XP, except for in a few locations that show founder effects [17–20]. The rarity of the disease means that diagnosis is not frequently required; hence, robust diagnostic methods have not been developed. Many who have engaged in basic research into the genetic and molecular basis of XP have provided informal services for patient and prenatal diagnosis using specialized DNA repair techniques [21–23]. Diagnosis has rarely been rapid, and many cases have required extensive research when the patient proved to have novel or rare mutations [17,24,25].

Several procedural as well as technical issues have precluded the development of routine diagnostic services for XP patients, including the size of the potential market, the technologies required, and licensing problems. The potential market size is small, which inhibits costly investment in novel technological approaches such as sequencing arrays. When the current author’s laboratory was carrying out diagnostic services, there would be a request for approximately one new diagnosis per month, each of which took approximately 6 weeks to complete [22,26]; this number is too few to support a commercial, fee-for-service diagnosis. Sequencing the candidate gene would be the ultimate standard for positive identification of an XP patient, but this approach can only be applied effectively once the gene in question is identified. More than 90% of all XP cases are accounted for by mutations in one of four genes: XPA, XPC, XPD, and XPV [27]. The need to diagnose a disease that has numerous underlying genes presents a technical challenge, especially since the information base is small, the number of mutations that occur in these genes has not been saturated, and new alleles continue to be identified.

A further complication, unique to the American market, is that the Clinical Laboratory Improvement and Accountability Act of 1988 regulates the laboratories that can perform patient-specific diagnostic tests. This has precluded the ability of research laboratories from using their specialized techniques for patient diagnosis and obtaining reimbursement. One approach that has been successful is for a dedicated diagnostic laboratory to provide generic DNA sequencing services for a large number of rare diseases, thus increasing their potential market. However, XP has too many potential alleles to be diagnosed cost-effectively in this manner. Therefore, at present, most methods for diagnosis depend on functional assays of UV sensitivity and DNA repair to delimit the sequencing required.

A rapid and robust routine diagnostic method is needed for confirmation of a clinical diagnosis of XP; and one that also provides essential information that could be employed in the dermatological clinic or commercial sector without the need for specialized technical approaches. In subsequent sections of this article, there will be a brief review of current knowledge of the biochemistry and genetics of XP, as well as a discussion of diagnostic techniques. An approach that the current author’s laboratory is developing and validating will also be presented.

Biochemistry

The NER system recognizes and repairs DNA damage that consists of UV-induced photoproducts and large DNA adducts [9]. Although the CPDs and (6–4)PDs are the major damage of concern in XP pathology, chemical adducts that are also substrates for NER include those produced by carcinogens or chemotherapy agents, such as N-acetoxy-N-acetyl aminofluorene (AAAF), benzo(a)-pyrene, aflatoxin, photoactivated psoralens, and cis-platinum. The neurological degeneration seen in some patients may be due to similar kinds of unrepaired damage generated endogenously by reactive oxygen species leaking from mitochondria or other sources [28,29].

The NER system consists of a series of reactions by which DNA damage is recognized in nuclear DNA within different functional domains (Figure 1) [30,31]. Damage in transcriptionally inactive regions is detected by the damage DNA-binding protein complex DDB1/DDB2 (DDB2 is also known as XPE) and by XPC/homologous recombination protein 23B (HR23B)/centrin2. Damage in transcriptionally active regions is detected through arrest of the transcriptional machinery involving RNA polymerases I and II, and requires the CSA and CSB proteins that are mutated in the CS disorder. The damaged site is then remodeled through a series of preincision complexes [31–33]. XPA, replication protein A (RPA), XPC, and the transcription DNA repair factor IIH (TFIIH) assemble in a random but cooperative order on the damaged site. They form an unstable preincision complex that is stabilized once the DNA is unwound by the ATPase activity of XPB and the ATPase/helicase activity of XPD in TFIIH [31–33]. TFIIH is a 10-component transcription factor containing XPB and XPD. XPC recruits XPG and is displaced from the complex. Cleavage then occurs on both sides of the damaged site, firstly by the XPG 3′ nuclease and then by the XPF/excision repair cross-complementing protein 1 (ERCC1) 5′ nuclease. The nucleases are anchored by the XPA/RPA complex, which serves to define the cleavage sites and strand specificity. Once the damaged oligonucleotide is removed, a patch is resynthesized by the proliferating cell nuclear antigen, the polymerases delta, epsilon, or kappa (δ, εor κ), and a ligase enzyme [30,34]. In quiescent cells, ligation involves X-ray repair cross-complementing protein 1 (XRCC1) and ligase III; in proliferating cells, ligation involves ligase I [35].

Figure 1.

Pathways of nucleotide excision repair and diagnostic methods. The proteins indicated are the main ones involved in XP; binding partners and other components are omitted for clarity. Left column: schematic of NER. The initial damage (top) is recognized by the XPE and XPC DNA-binding proteins. The damaged site is then remodeled through a series of preincision complexes, indicated within the hatched box [31–33]. XPA, RPA, and XPC/TFIIH form an initial preincision complex that is stabilized once the DNA is unwound by the ATPase activity of XPB and ATPase/helicase activity of XPD in TFIIH [31–33]. XPC recruits XPG and is displaced from the complex. Cleavage then occurs on both sides of the damaged site, firstly by the XPG 3′ nuclease and then by the XPF/ERCC1 5′ nuclease. Once the damaged oligonucleotide is removed, a patch is resynthesized and is completed by polymerases and ligases. Right column: steps in NER that have been addressed for diagnosing repair by (top to bottom) damage detection, protein expression by immunohistochemistry, strand breakage, damage removal, and unscheduled DNA synthesis.

BrdUrd: bromodeoxyuridine; dThd: tritiated thymidine; ERCC1: excision repair cross-complementing protein 1; HRP: horseradish peroxidase counter stain for XPC antibody; NER: nucleotide excision repair; RPA: replication protein A; TFIIH: transcription/DNA repair factor IIH; UV: ultraviolet; XP: xeroderma pigmentosum.

NER can remove DNA damage before DNA replication begins and, consequently, plays a major role in reducing the amount of damage that becomes fixed as mutations during replication [36]. Specialized polymerases are required to replicate DNA photoproducts because the normal DNA polymerases – alpha, delta, and epsilon (α, δ, and ε) – cannot accommodate large distortions such as DNA photoproducts or adducts in their active sites [10,37]. These damage-specific polymerases have relaxed substrate specificity, and the most important is the low-fidelity polymerase Pol eta (η) [38]. This is mutated in the XPV condition, which is often clinically indistinguishable from NER-deficient XP [39].

Genetics of XP

The total number of genes directly involved in NER is estimated to be approximately 40 [40]. Only eight are known to be associated with XP, two with CS, and four with TTD. Many of the other genes are apparently essential and would be lethal if mutated. For example, complete loss of function in both alleles is not seen in XPB or XPD patients because the proteins are essential components of TFIIH, which has 10 component proteins [41–44]. Mutations in ERCC1 have been reported in only one patient (with COFS), which was neonatally lethal [24]. Some patients may have such mild clinical disease that they blend into the population of other sun-sensitive individuals, and a number of these patients have been defined as having UV^S, which can overlap with CS [45–47]. Novel genes could yet be found to cause these diseases. There may also be new variant diseases that could be candidates for genetic analysis.

Diagnostic approaches

A direct sequencing approach is challenging for XP as it requires the sequencing of up to a total of 25 kb of coding sequence scattered over >100 exons of eight genes; this requires multiplex polymerase chain reaction amplification and independent cloning and sequencing in most cases. However, CS usually involves only one of two genes, CSA or CSB; direct sequencing of these genes for a new patient would be feasible. CS, especially when associated with XP-like symptoms, can also be caused by mutations in XPB, XPG, and particularly XPD; therefore, there remains a role for functional cellular assays in some cases of this disease [48]. Advances in multiplex amplification of large numbers of exons simultaneously, together with the falling costs of high-volume DNA sequencing, should eventually lead to a cost-effective approach of screening for mutations, even in rare diseases [49].

Many of the stages of NER can be exploited in order to assay for the presence of mutations that would be diagnostic for XP (Figure 1); however, not all of these methods can be conveniently transferred into the diagnostic context. The majority require access to a specialty laboratory with a calibrated UV source, and methods for analyzing DNA metabolism and photoproducts as well as transfer of complementing genes.

Specialized laboratory tests that the current author and others have used at various times include (Figure 1): UV sensitivity and plasmid transfection with or without cotransfection of wild-type genes to complement the defect [21,25], photoproduct formation and excision [50], single-strand break production during excision [22,51,52], and assay of incorporation of new bases during the repair synthesis phase (unscheduled DNA synthesis) [53,54]. These tests have all been used in various contexts for diagnosis of patients and for prenatal diagnosis [22,23,25,26,55]. To identify the gene that is mutated, the methods require a panel of characterized XP cells with which to compare a candidate cell line by cell-fusion techniques [53], or cDNA expression systems with the panel of XP genes for co-transfection [25].

These approaches all necessitate cells from a patient to be established in culture and their sensitivity to UV light determined, relative to that of cells from normal individuals. Fibroblasts from the skin or lymphoid cells from the blood have proven to be equally usable [17]. Detection of increased sensitivity to UV light will only identify a patient as belonging to one of the XP groups (A through to G). Group E represents mild disease with small repair deficit; its classification has proved difficult and misidentification has occurred in the past [56,57].

Cells from patients with XPV are only slightly sensitive to UV and their identification requires an additional sensitization step by growth in caffeine (a kinase inhibitor with a broad range of activity) after UV irradiation [58,59]. This sensitization is due, in part, to inhibition of an ataxia telangiectasia and Rad3-related protein (ATR)-dependent phosphorylation pathway in Pol η-deficient cells, but the complete explanation remains unclear. The sensitization is sufficient to identify a patient with XPV, after which sequencing of the Pol η gene would confirm the diagnosis [39]. If cells are not sensitive to UV light, with or without caffeine, this would be sufficient to exclude a diagnosis of XP and no further test would be required.

Immunohistochemistry as a route to identification of mutated XP genes

Many research groups have used the methods described above for identifying new XP patients (Figure 1); however, these are complex, specialized technical approaches that are not routinely available in dermatopathology laboratories. The current author’s laboratory has explored the potential of using a technique that is more readily available – that of immunohistochemistry (de Feraudy and Cleaver, unpublished observations). Most XP patients come to the attention of clinicians because of photosensitivity and incipient sun damage in the form of excessive freckling, keratoses, or early skin cancers. The majority of dermatological examinations of potential XP patients include biopsy of suspicious lesions that are then fixed and embedded in paraffin blocks for histological examination. These blocks are archived and can be accessed for subsequent analysis of XP protein expression without subjecting the patient to further invasive procedures.

Mutations in XP genes encompass the complete spectrum of loss-of-function changes, including stop codons, frame shifts, and splice-site and missense mutations. The majority of these result in reduced or absent protein levels [17,25,27,39,44,60]. Mutations in XPB and XPD, of which at least one allele is missense, only have small effects on the protein levels [44,61,62], unlike mutations in the p8 component of TFIIH that destabilize protein levels [63,64]. Therefore, it is likely that most mutations in XPA, XPC, XPG, XPF, and XPV, but less so for XPB and XPD, will cause a reduced amount of protein detectable by immunohistochemistry of fixed tissue with the appropriate antibody or panel of antibodies. Many of these antibodies are now available from commercial sources. Preliminary studies carried out in the present author’s laboratory have demonstrated that XPC is highly expressed in the basal layer of the skin, with decreasing levels as cells migrate through the epidermis. This laboratory is assessing the suitability of this method as a routine dermatopathology procedure, since immunohistochemistry is an accepted method covered by most US medical insurance plans. However, an evaluation of the sensitivity and precision of a range of commercial antibodies for detecting XP mutations, and the accuracy of predictions, is still needed. While the method should prove useful for identifying XPA, XPC, and XPV patients, its utility for XPB and XPD patients may be much less as these patients do not have a significant reduction in protein levels [61]. In addition, mild cases caused by leaky mutations with significant levels of protein expression may also be difficult to identify unequivocally. Standards consisting of paraffin blocks of skin biopsies from known patients, or pellets of tissue culture cells from XP patients would aid in validation of the method. A further development of this approach could be antibody arrays comprising a panel of XP antibodies on which extracts of peripheral lymphocytes could bind.

Diagnosis and prognosis

Any new patient requires an initial confirmation, by assessment of UV sensitivity, that they genuinely do have XP; the gene involved must then be identified. The next question is what use is to be made of the information since there are no cures, only palliative procedures, to offer the prospective patient? At its simplest, a diagnosis of XP dictates a lifetime of stringent protection from sun exposure for patients and their families. However, many patients ask for further information, including which gene and mutation is involved and what the future portends. Here, we face a dilemma because available data sets do not allow the generation of confident predictions of specific mutations leading to specific changes in later life. It is known that if sun exposure is not avoided, cancer is likely to develop in most cases; however, if there has been significant sun exposure before diagnosis, it is not clear how the residual DNA damage will affect later clinical symptoms. In general, most cases of XPC and XPV do not lead to later neurodegenerative symptoms, but this is not an absolute certainty [65,66]. Most cases of XPA, XPB, XPD, and XPG do lead to neurodegenerative conditions, but not always, and a wide range of disease severity has been reported for a given XP gene due to the specific mutations or genetic backgrounds [25]. Although it might seem reasonable to provide patients with as much detail as possible, much of the information gleaned from their DNA sequence may not aid them in terms of healthcare and lifestyle changes; thus, it remains debatable as to what should be included in a comprehensive diagnostic evaluation.