Chapter 8. Genetics in Relation to the Skin

In the 30 years since the first human gene, placental lactogen, was cloned in 1977, huge investments in time, money, and effort have gone into disclosing the innermost workings of the human genome. The Human Genome Project, which began in 1990, has led to sequence information on more than 3 billion base pairs (bp) of DNA, with identification of most of the estimated 25,000 genes in the entire human genome.1 Although a few relatively small gaps remain, the near completion of the entire sequence of the human genome is having a huge impact on both the clinical practice of genetics and the strategies used to identify disease-associated genes. Laborious positional cloning approaches and traditional functional studies are gradually being transformed by the emergence of new genomic and proteomic databases.2 Some of the exciting challenges that clinicians and geneticists now face are determining the function of these genes, defining disease associations and, relevant to patients, correlating genotype with phenotype. Nevertheless, many discoveries are already influencing how clinical genetics is practiced throughout the world, particularly for patients and families with rare, monogenic inherited disorders. The key benefits of dissection of the genome thus far have been the documentation of new information about disease causation, improving the accuracy of diagnosis and genetic counseling, and making DNA-based prenatal testing feasible.3 Indeed, the genetic basis of more than 2,000 inherited single gene disorders has now been determined, of which about 25% have a skin phenotype. Therefore, these discoveries have direct relevance to dermatologists and their patients. Recently, studies in rare inherited skin disorders have also led to new insight into the pathophysiology of more common complex trait skin disorders.4 This new information is expected to have significant implications for the development of new therapies and management strategies for patients. Therefore, for the dermatologist understanding the basic language and principles of clinical and molecular genetics has become a vital part of day-to-day practice. The aim of this chapter is to provide an overview of key terminology in genetics that is clinically relevant to the dermatologist.

Normal human beings have a large complex genome packaged in the form of 46 chromosomes. These consist of 22 pairs of autosomes, numbered in descending order of size from the largest (chromosome 1) to the smallest (chromosome 22), in addition to two sex chromosomes, X and Y. Females possess two copies of the X chromosome, whereas males carry one X and one Y chromosome. The haploid genome consists of about 3.3 billion bp of DNA. Of this, only about 1.5% corresponds to protein-encoding exons of genes. Apart from genes and regulatory sequences, perhaps as much as 97% of the genome is of unknown function, often referred to as “junk” DNA. However, caution should be exercised in labeling the noncoding genome as “junk,” because other unknown functions may reside in these regions. Much of the noncoding DNA is in the form of repetitive sequences, pseudogenes (“dead” copies of genes lost in recent evolution), and transposable elements of uncertain relevance. Although initial estimates for the number of human genes was in the order of 100,000, current predictions, based on the essentially complete genome sequence, are in the range of 20,000 to 25,000.1 Surprisingly, therefore, the human genome is comparable in size and complexity to primitive organisms such as the fruit fly. However, it is thought that the generation of multiple protein isoforms from a single gene via alternate splicing of exons, each with a discrete function, is what contributes to increased complexity in higher organisms, including humans. In addition to protein-encoding genes, there are also many genes encoding untranslated RNA molecules, including transfer RNA, ribosomal RNA, and, as recently described, microRNA genes. MicroRNA is thought to be involved in the control of a large number of other genes through the RNA inhibition pathway. Very recently, it has emerged that tracts of the genome are transcribed at low levels in the form of exotic new RNA species, including natural antisense RNA and long interspersed noncoding RNA. These transcripts are emerging as key regulatory molecules. Thus, a much greater proportion of the genome is actively transcribed than was previously recognized and this trend is likely to continue in the current “postgenome” era of human genetics.

The draft sequence of the human genome was completed in 2003. Subsequently, small gaps have been filled, and the sequence has now been extensively annotated in terms of genes, repetitive elements, regulatory sequences, polymorphisms, and many other features recognizable by in silico data mining methods informed, wherever possible, by functional analysis. This annotation process will continue for some time as more features are uncovered. The human genome data, and that for an increasing number of other species, is freely available on Web sites (Table 8-1). Some regions of the genome, particularly near the centromeres, consist of long stretches of highly repetitive sequences that are difficult or impossible to clone and/or sequence. These heterochromatic regions of the genome are unlikely to be sequenced and are thought to be structural in nature, mediating the chromosomal architecture required for cell division, rather than contributing to heritable characteristics.

Given the size and complexity of the human genome and other genomes now available, analysis of these enormous datasets in any kind of meaningful way is heavily reliant on computers. Even storage and retrieval of the sequence data associated with mammalian genome require considerable computer power and memory, and even the assembly of the raw sequence of any mammalian genome would have been unfeasible without computers. Many Web browsers for accessing genome data are available and the most useful of these are listed in Table 8-1. Each of these interfaces, which are the ones which the authors find most useful and user-friendly, contains a wide variety of tools for analysis and searching of sequences according to keyword, gene name, protein name, and homology to DNA or protein sequence data.

The main source of historical, clinical, molecular, and biochemical data relating to human genetic diseases is the Online Mendelian Inheritance in Man (OMIM) (see Table 8-1). All recognized genetic diseases and nonpathogenic heritable traits, including common diseases with a genetic component, as well as all known genes and proteins, are listed and reviewed by OMIM number with links to PubMed.

Human chromosomes share common structural features (Fig. 8-1). All consist of two chromosomal arms, designated as “p” and “q.” If the arms are of unequal length, the short arm is always designated as the “p” arm. Chromosomal maps to seek abnormalities are based on the stained, banded appearance of condensed chromosomes during metaphase of mitosis. During interphase, the uncondensed chromosomes are not discernible by normal microscopy techniques. Genes can now be located with absolute precision in terms of the range of bp that they span within the DNA sequence for a given chromosome. The bands are numbered from the centromere outwards using a system that has evolved as increasingly discriminating chromosome stains, as well as higher resolution light microscopes, became available. A typical cytogenetic chromosome band is 17q21.2, within which the type I keratin genes reside (see Fig. 8-1).

The ends of the chromosomal arms are known as telomeres, and these consist of multiple tandem repeats of short DNA sequences. In germ cells and certain other cellular contexts, additional repeats are added to telomeres by a protein–RNA enzyme complex known as telomerase. During each round of cell division in somatic cells, one of the telomere repeats is trimmed off as a consequence of the DNA replication mechanism. By measuring the length of telomeres, the “age” of somatic cells, in terms of the number of times they have divided during the lifetime of the organism, can be determined. Once the telomere length falls below a certain threshold, the cell undergoes senescence. Thus, telomeres contribute to an important biological clock function that removes somatic cells that have gone through too many rounds of replication and are at a high risk of accumulating mutations that could lead to tumorigenesis or other functional aberration.5

The chromosome arms are separated by the centromere, which is a large stretch of highly repetitious DNA sequence. The centromere has important functions in terms of the movement and interactions of chromosomes. The centromeres of sister chromatids are where the double chromosomes align and attach during the prophase and anaphase stages of mitosis (and meiosis). The centromeres of sister chromatids are also the site of kinetochore formation. The latter is a multiprotein complex to which microtubules attach, allowing mitotic spindle formation, which ultimately results in pulling apart of the chromatids during anaphase of the cell division cycle.

The majority of chromosomal DNA contains genes interspersed with noncoding stretches of DNA of varying sizes. The density of genes varies widely across the chromosomes so that there are gene-dense regions or, alternately, large areas almost devoid of functional genes. An example of a comparatively gene-rich region of particular relevance to inherited skin diseases is the type I keratin gene cluster on chromosome band 17q21.2 (see Fig. 8-1). This diagram also gives an idea of the sizes in bp of DNA of a typical chromosome and a typical gene located within it. This gene cluster spans about 900,000 bp of DNA and contains 27 functional type I keratin genes, several genes encoding keratin-associated proteins, and a number of pseudogenes (not shown). Because chromosome 17 is one of the smaller chromosomes, Fig. 8-1 starts to give some idea of the overall complexity and organization of the genome.

Protein-encoding genes normally consist of several exons, which collectively code for the amino acid sequence of the protein (or open reading frame). These are separated by noncoding introns. In human genes, few exons are much greater than 1,000 bp in size, and introns vary from less than 100 bp to more than 1 million bp. A typical exon might be 100 to 300 bp in size. The KRT14 gene encoding keratin 14 (K14) protein is one of the genes in which mutations lead to epidermolysis bullosa (EB) simplex (see Chapter 62) and is illustrated in Fig. 8-1. KRT14 is contained within about 7,000 bp of DNA and consists of eight modestly sized exons interspersed by seven small introns. Although all genes are present in all human cells that contain a nucleus, not every gene is expressed in all cells of tissues. For example, the KRT14 gene is only active in basal keratinocytes of the epidermis and other stratified epithelial tissues and is essentially silent in all other tissues. When a protein-encoding gene is expressed, the RNA polymerase II enzyme transcribes the coding strand of the gene, starting from the cap site and continuing to the end of the final exon, where various signals lead to termination of transcription. The initial RNA transcript, known as heteronuclear RNA, contains intronic as well as exonic sequences. This primary transcript undergoes splicing to remove the introns, resulting in the messenger RNA (mRNA) molecule.6 In addition, the bases at the 5′ end (start) of the mRNA are chemically modified (capping) and a large number of adenosine bases are added at the 3′ end, known as the poly-A tail. These posttranscriptional modifications stabilize the mRNA and facilitate its transport within the cell. The mature mRNA undergoes a test round of translation which, if successful, leads to the transport of the mRNA to the cytoplasm, where it undergoes multiple rounds of translation by the ribosomes, leading to accumulation of the encoded protein. If the mRNA contains a nonsense mutation, otherwise known as a premature termination codon mutation, the test round of translation fails, and the cell degrades this mRNA via the nonsense-mediated mRNA pathway.7 This is a mechanism that the cell has evolved to remove aberrant transcripts, and it may also contribute to gene regulation, particularly when very low levels of a particular protein are required within a given cell.

Splicing out of introns is a complex process. The genes of prokaryotes, such as bacteria, do not contain introns, and so mRNA splicing is a process that is specific to higher organisms. In some more primitive eukaryotes, RNA molecules contain catalytic sequences known as ribozymes, which mediate the self-splicing out of introns without any requirement for additional factors. In mammals, splicing involves a large number of protein and RNA factors encoded by several genes. This allows another level of control over gene expression and also facilitates alternative splicing of exons, so that a single gene can encode several functionally distinct variants of a protein. These isoforms are often differentially expressed in different tissues. In terms of the gene sequences important for splicing, a few bp at the beginning and at the end of an intron, known as the 5′ splice site (or splice donor site) and the 3′ splice site (or splice acceptor site) are crucial. A few other bp within the intron, such as the branch point site located 18–100 bp away from the 3′ end, are also critical. Mutations affecting any of the invariant residues of these splice sites lead to aberrant splicing and either complete loss of protein expression or generation of a highly abnormal protein.

The mRNA also contains two untranslated regions (UTR): (1) the 5′UTR upstream of the initiating ATG codon and (2) the 3′UTR downstream of the terminator (or stop codon, which can be TGA, TAA, or TAG). The 5′ UTR can and often does possess introns, whereas the 3′UTR of more than 99% of mammalian genes does not contain introns. The nonsense-mediated mRNA decay pathway identifies mutant transcripts by means of assessing where the termination codon occurs in relation to introns. The natural stop codon is always followed immediately by the 3′UTR, which, in turn, does not normally possess any introns. If a stop codon occurs in an mRNA upstream of a site where an intron has been excised, this message is targeted for nonsense-mediated decay. The only genes that contain introns within their 3′UTR sequences are expressed at extremely low levels. This is one of the ways in which the cell can determine how much protein is made from a particular gene.

Gene complexity is widely variable and not necessarily related to the size of the protein encoded. Some genes consist of only a single small exon, such as those encoding the connexin family of gap junction proteins. Such single exon genes are rapid and inexpensive to analyze routinely. In contrast, the type VII collagen gene, COL7A1, in which mutations lead to the dystrophic forms of EB (see Chapter 62), has 118 exons, meaning that 118 different parts of the gene need to be isolated and analyzed for molecular diagnosis of each dystrophic EB patient. The filaggrin gene (FLG) on chromosome 1, recently shown to be the causative gene for ichthyosis vulgaris (see Chapter 49) and a susceptibility gene for atopic dermatitis (see Chapter 14), has only three exons. However, the third exon of FLG is made up of repeats of a 1,000-bp sequence and varies in size from 12,000 to 14,000 bp among different individuals in the population. This unusual gene structure makes routine sequencing of genes such as COL7A1 or FLG difficult, time consuming, and expensive.

Each specific gene is generally only actively transcribed in a subset of cells or tissues within the body. Gene expression is largely determined by the promoter elements of the gene. In general, the most important region of the promoter is the stretch of sequence immediately upstream of the cap site. This proximal promoter region contains consensus binding sites for a variety of transcription factors, some of which are general in nature and required for all gene expression, others are specific to particular tissue or cell lineage, and some are absolutely specific for a given cell type and/or stage of development or differentiation. The size of the promoter can vary widely according to gene family or between the individual genes themselves. For example, the keratin genes are tightly spaced within two gene clusters on chromosomes 12q and 17q, but these are exquisitely tissue specific in two different ways. First, these genes are only expressed in epithelial cells, and therefore their promoters must possess regulatory sequences that determine epithelial expression. Therefore, these regulatory elements are specific for cells of ectodermal origin. Second, these genes are expressed in very specific subsets of epithelial cells, and so there must be a second level of control that specifies which epithelial cell layers express specific keratin genes. This is best illustrated in the hair follicle, where there are many different epithelial cell layers, each with a specific pattern of keratin gene expression (see Chapter 86).8

Transcription factors are proteins that either bind to DNA directly or indirectly by associating with other DNA-binding proteins. Binding of these factors to the promoter region of a gene leads to activation of the transcription machinery and transcription of the gene by RNA polymerase II. The transcription factor proteins are encoded by genes that are in turn controlled by promoters that are regulated by other transcription factors encoded by other genes. Thus, there are several tiers of control over gene expression in a given cell type, and the intricacies of this can be difficult to fully unravel experimentally. Nevertheless, by isolation of promoter sequences from genes of interest and placing these in front of reporter genes that can be assayed biochemically, such as firefly luciferase that can be assayed by light emission, the activity of promoters can be reproduced in cultured cells that normally express the gene. Combining such a reporter gene system with site-directed mutagenesis to make deletions or alter small numbers of bp within the promoter can help define the extent of the promoter and the important sequences within it that are required for gene expression. A variety of biochemical techniques, such as DNA footprinting, ribonuclease protection, electrophoretic mobility shift assays, or chromatin immunoprecipitation, can be used to determine which transcription factors bind to a particular promoter and help delineate the specific promoter sequences bound. Expression of reporter genes under the control of a cloned promoter in transgenic mice also helps shed light on the important sequences that are required to recapitulate the endogenous expression of the gene under study. Keratin promoters are unusual in that, generally, a small fragment of only 2,000 to 3,000 bp upstream of the gene can confer most of the tissue specificity. For this reason, keratin promoters are widely used to drive exogenous transgene expression in the various specific cellular compartments of the epidermis and its appendages for experiments to determine gene, cell, or tissue function.9

Some promoter or enhancer sequences act over very long distances. In some cases, sequences located millions of bp distant, with several other genes in the intervening region, somehow influence expression of a target gene. In some genetic diseases, mutations affecting such long-range promoter elements are now emerging. These types of mutations appear to be rare, but since they occur so far away from the target gene and are therefore very difficult to find, this class of mutation may, in fact, be more common than is immediately obvious. In general, relatively few disease-causing mutations have been shown to involve promoters, but this class of defect is probably greatly underrepresented because the sequences that are important for promoter activity are poorly characterized. Prediction of transcription factor binding sites by computer analysis is an area for further study. Although these undoubtedly exist, there are relatively few examples so far of pathogenic defects in microRNA or other noncoding regulatory RNA species.

In establishing the molecular basis of an inherited skin disease, there are two key steps. First, the gene linked to a particular disorder must be identified, and second, pathogenic mutations within that gene should be determined. Diseases can be matched to genes either by genetic linkage analysis or by a candidate gene approach.10 Genetic linkage involves studying pedigrees of affected and unaffected individuals and isolating which bits of the genome are specifically associated with the disease phenotype. The goal is to identify a region of the genome that all the affected individuals and none of the unaffected individuals have in common; this region is likely to harbor the gene for the disorder, as well as perhaps other nonpathogenic neighboring genes that have been inherited by linkage disequilibrium. Traditionally, genome-wide linkage strategies make use of variably sized microsatellite markers scattered throughout the genome, although for recessive diseases involving consanguineous pedigrees, a more rapid approach may be to carry out homozygosity mapping using single nucleotide polymorphism (SNP) chip arrays. By contrast, the candidate gene approach involves first looking for a clue to the likely gene by finding a specific disease abnormality, perhaps in the expression (or lack thereof) of a particular protein or RNA, or from an ultrastructural or biochemical difference between the diseased and control tissues. Nevertheless, the genetic linkage and candidate gene approaches are not mutually exclusive and are often used in combination. For example, to identify the gene responsible for the autosomal recessive disorder, lipoid proteinosis (see Chapter 137), genetic linkage using microsatellites was first used to establish a region of linkage on 1q21 that contained 68 genes.11 The putative gene for this disorder, ECM1 encoding extracellular matrix protein 1, was then identified by a candidate gene approach that searched for reduced gene expression (lack of fibroblast complementary DNA) in all these genes. A reduction in ECM1 gene expression in lipoid proteinosis compared with control provided the clue to the candidate gene because there were no differences in any of the other patterns of gene expression. Ultrastructural and immunohistochemical analyses can also provide clues to underlying gene pathology. For example, loss of hemidesmosomal inner plaques noted on transmission electron microscopy and a complete absence of skin immunostaining for the 230-kDA bullous pemphigoid antigen (BP230) at the dermal–epidermal junction, led to the discovery of loss-of-function mutations in the dystonin (DST) gene, which codes for BP230, in a new form of autosomal recessive epidermolysis bullosa simplex.12

Having identified a putative gene for an inherited disorder, the next stage is to find the pathogenic mutation(s). This can be done by sequencing the entire gene, a feat which is becoming easier as technologic advances make automated nucleotide sequencing faster, cheaper, and more accessible. However, the large size of some genes may make comprehensive sequencing impractical, and therefore initial screening approaches to identify the region of a gene that contains the mutation may be a necessary first step. There are many mutation detection techniques available to scan for sequence changes in cellular RNA or genomic DNA, and these include denaturing gradient gel electrophoresis, chemical cleavage of mismatch, single stranded conformation polymorphism, heteroduplex analysis, conformation sensitive gel electrophoresis, denaturing high-performance liquid chromatography and the protein truncation test.13

The most critical factor that determines the success of any gene screening protocol is the sensitivity of the detection technique. In addition, when choosing a mutation screening strategy using genomic DNA, the size of the gene and its number of exons must be taken into account. The sensitivities of these methods vary greatly, depending on the size of template screened. For example, single-stranded conformation polymorphism has a sensitivity of >95% for fragments of 155 bp, but this is reduced to only 3% for 600 bp. Once optimized, denaturing gradient gel electrophoresis has a sensitivity of about 99% for fragments of up to 500 bp, and conformation sensitive gel electrophoresis is expected to have a sensitivity of 80% to 90% for fragments of up to 600 bp. Chemical cleavage of mismatch, on the other hand, has a sensitivity of 95% to 100% for fragments >1.5 kilobases (kb) in size and is ideal for screening compact genes where more than one exon can be amplified together using genomic DNA as the template. All these techniques detect sequence changes such as truncating and missense mutations as well as polymorphisms; however, the protein truncation test screens only for truncating mutations and is predicted to have a sensitivity of >95% and can be used for RNA or DNA fragments in excess of 3 kb.

Whichever approach is taken, having identified a difference in the patient's DNA compared with the control sample, the next stage is to determine how this segregates within a particular family and also whether it is pathogenic or not. Very recently, great advances have been made in DNA sequencing technology, with the emergence of “next generation sequencing” (NGS) technology. Currently, it is quite feasible to carry out whole exome sequencing in an individual using NGS, i.e., sequencing of all the protein-encoding exons in the genome, in a matter of days and for only a few thousand dollars. It is expected that whole genome sequencing, at a cost of $1,000 or less will be a commonplace in 2–3 years. This incredible new technology is set to revolutionize human genetics once more, and in particular, will facilitate identification of mutated genes in small kindreds that are not tractable by genetic linkage methods. These advances will also impact on diagnosis—in the near future it may be faster and cheaper to sequence a patient's whole genome rather than to do targeted sequencing of specific genes or regions.

Within the human genome, the genetic code of two healthy individuals may show a number of sequence dissimilarities that have no relevance to disease or phenotypic traits. Such changes within the normal population are referred to as polymorphisms (Fig. 8-2). Indeed, even within the coding region of the genome, clinically irrelevant substitutions of one bp, known as SNPs, are common and occur approximately once every 250 bp.14 Oftentimes, these SNPs do not change the amino acid composition; for example, a C-to-T transition in the third position of a proline codon (CCC to CCT) still encodes for proline, and is referred to as a silent mutation. However, some SNPs do change the nature of the amino acid; for example, a C-to-G transversion at the second position of the same proline codon (CCC to CGC) changes the residue to arginine. It then becomes necessary to determine whether a missense change such as this represents a nonpathogenic polymorphism or a pathogenic mutation. Factors favoring the latter include the sequence segregating only with the disease phenotype in a particular family, the amino acid change occurring within an evolutionarily conserved residue, the substitution affecting the function of the encoded protein (size, charge, conformation, etc.), and the nucleotide switch not being detectable in at least 100 ethnically matched control chromosomes. Nonpathogenic polymorphisms do not always involve single nucleotide substitutions; occasionally, deletions and insertions may also be nonpathogenic.

A mutation can be defined as a change in the chemical composition of a gene. A missense mutation changes one amino acid to another. Mutations may also be insertions or deletions of bases, the consequences of which will depend on whether this disrupts the normal reading frame of a gene or not, as well as nonsense mutations, which lead to premature termination of translation (see Fig. 8-2). For example, a single nucleotide deletion within an exon causes a shift in the reading frame, which usually leads to a downstream stop codon, thus giving a truncated protein, or often an unstable mRNA that is readily degraded by the cell. However, a deletion of three nucleotides (or multiples thereof) will not significantly perturb the overall reading frame, and the consequences will depend on the nature of what has been deleted. Nonsense mutations typically, but not exclusively, occur at CpG dinucleotides, where methylation of a cytosine nucleotide often occurs. Inherent chemical instability of this modified cytosine leads to a high rate of mutation to thymine. Where this alters the codon (e.g., from CGA to TGA), it will change an arginine residue to a stop codon. Nonsense mutations usually lead to a reduced or absent expression of the mutant allele at the mRNA and protein levels. In the heterozygous state, this may have no clinical effect [e.g., parents of individuals with Herlitz junctional EB are typically carriers of nonsense mutations in one of the laminin 332 (laminin 5) genes but have no skin fragility themselves; see Chapter 62], but a heterozygous nonsense mutation in the desmoplakin gene, for example, can result in the autosomal dominant skin disorder, striate palmoplantar keratoderma (see Chapter 50). This phenomenon is referred to as haploinsufficiency (i.e., half the normal amount of protein is insufficient for function).

Apart from changes in the coding region that result in frameshift, missense, or nonsense mutations, approximately 15% of all mutations involve alterations in the gene sequence close to the boundaries between the introns and exons, referred to as splice site mutations. This type of mutation may abolish the usual acceptor and donor splice sites that normally splice out the introns during gene transcription. The consequences of splice site mutations are complex; sometimes they lead to skipping of the adjacent exon, and other times they result in the generation of new mRNA transcripts through utilization of cryptic splice sites within the neighboring exon or intron.

Mutations within one gene do not always lead to a single inherited disorder. For example, mutations in the ERCC2 gene may lead to xeroderma pigmentosum (type D), trichothiodystrophy, or cerebrofacioskeletal syndrome, depending on the position and type of mutation. Other transacting factors may further modulate phenotypic expression. This situation is known as allelic heterogeneity. Conversely, some inherited diseases can be caused by mutations in more than one gene (e.g., non-Herlitz junctional EB; see Chapter 62) and can result from mutations in either the COL17A1, LAMA3, LAMB3, or LAMC2 genes. This is known as genetic heterogeneity. In addition, the same mutation in one particular gene may lead to a range of clinical severity in different individuals. This variability in phenotype produced by a given genotype is referred to as the expressivity. If an individual with such a genotype has no phenotypic manifestations, the disorder is said to be nonpenetrant. Variability in expression reflects the complex interplay between the mutation, modifying genes, epigenetic factors, and the environment and demonstrates that interpreting what a specific gene mutation does to an individual involves more than just detecting one bit of mutated DNA in a single gene.

There are approximately 5,000 human single-gene disorders and, although the molecular basis of less than one-half of these has been established, understanding the pattern of inheritance is essential for counseling prospective parents about the risk of having affected children. The four main patterns of inheritance are (1) autosomal dominant, (2) autosomal recessive, (3) X-linked dominant, and (4) X-linked recessive.

For individuals with an autosomal dominant disorder, one parent is affected, unless there has been a de novo mutation in a parental gamete. Males and females are affected in approximately equal numbers, and the disorder can be transmitted from generation to generation; on average, half the offspring will have the condition (Fig. 8-3). It is important to counsel affected individuals that the risk of transmitting the disorder is 50% for each of their children, and that this is not influenced by the number of previously affected or unaffected offspring. Any offspring that are affected will have a 50% risk of transmitting the mutated gene to the next generation, whereas for any unaffected offspring, the risk of the next generation being affected is negligible, providing that the partner does not have the autosomal dominant condition. Some dominant alleles can behave in a partially dominant fashion. The term semidominant is applied when the phenotype in heterozygous individuals is less than that observed for homozygous subjects. For example, ichthyosis vulgaris is a semidominant disorder in which the presence of one or two mutant profilaggrin gene (FLG) alleles can strongly influence the clinical severity of the ichthyosis.

In autosomal recessive disorders, both parents are carriers of one normal and one mutated allele for the same gene and, typically, they are phenotypically unaffected (Fig. 8-4). If both of the mutated alleles are transmitted to the offspring, this will give rise to an autosomal recessive disorder, the risk of which is 25%. If one mutated and one wild-type allele is inherited by the offspring, the child will be an unaffected carrier, similar to the parents. If both wild-type alleles are transmitted, the child will be genotypically and phenotypically normal with respect to an affected individual. If the mutations from both parents are the same, the individual is referred to as a homozygote, but if different parental mutations within a gene have been inherited, the individual is termed a compound heterozygote. For someone who has an autosomal recessive condition, be it a homozygote or compound heterozygote, all offspring will be carriers of one of the mutated alleles but will be unaffected because of inheritance of a wild-type allele from the other, clinically and genetically unaffected, parent. This assumes that the unaffected parent is not a carrier. Although this is usually the case in nonconsanguineous relationships, it may not hold true in first-cousin marriages or other circumstances where there is a familial interrelationship. For example, if the partner of an individual with an autosomal recessive disorder is also a carrier of the same mutation, albeit clinically unaffected, then there is a 50% chance of the offspring inheriting two mutant alleles and therefore also inheriting the same autosomal recessive disorder. This pattern of inheritance is referred to as pseudodominant.

In X-linked dominant inheritance, both males and females are affected, and the pedigree pattern may resemble that of autosomal dominant inheritance (Fig. 8-5). However, there is one important difference. An affected male transmits the disorder to all his daughters and to none of his sons. X-linked dominant inheritance has been postulated as a mechanism in incontinentia pigmenti (see Chapter 75), Conradi–Hünermann syndrome, and focal dermal hypoplasia (Goltz syndrome), conditions that are almost always limited to females. In most X-linked dominant disorders with cutaneous manifestations, affected males may be aborted spontaneously or die before implantation (leading to the appearance of female-to-female transmission). Most viable male patients with incontinentia pigmenti have a postzygotic mutation in NEMO and no affected mother; occasionally, males with an X-linked dominant disorder have Klinefelter syndrome with an XXY genotype.

X-linked recessive conditions occur almost exclusively in males, but the gene is transmitted by carrier females, who have the mutated gene only on one X chromosome (heterozygous state). The sons of an affected male will all be normal (because their single X chromosome comes from their clinically unaffected mother) (Fig. 8-6). However, the daughters of an affected male will all be carriers (because all had to have received the single X chromosome from their father that carries the mutant copy of the gene). Some females show clinical abnormalities as evidence of the carrier state (such as in hypohidrotic ectodermal dysplasia; see Chapter 142); the variable extent of phenotypic expression can be explained by lyonization, the normally random process that inactivates either the wild-type or mutated X chromosome in each cell during the first weeks of gestation and all progeny cells.15 Other carriers may not show manifestations because the affected region on the X chromosome escapes lyonization (as in recessive X-linked ichthyosis) or the selective survival disadvantage of cells in which the mutated X chromosome is activated (as in the lymphocytes and platelets of carriers of Wiskott–Aldrich syndrome; see Section “Mosaicism”).

Aberrations in chromosomes are common. They occur in about 6% of all conceptions, although most of these lead to miscarriage, and the frequency of chromosomal abnormalities in live births is about 0.6%. Approximately two-thirds of these involve abnormalities in either the number of sex chromosomes or the number of autosomes; the remainder is chromosomal rearrangements. The number and arrangement of the chromosomes is referred to as the karyotype. The most common numerical abnormality is trisomy, the presence of an extra chromosome. This occurs because of nondisjunction, when pairs of homologous chromosomes fail to separate during meiosis, leading to gametes with an additional chromosome. Loss of a complete chromosome, monosomy, can affect the X chromosome but is rarely seen in autosomes because of nonviability. A number of chromosomal disorders are also associated with skin abnormalities, as detailed in Table 8-2.

Structural aberrations (fragility breaks) in chromosomes may be random, although some chromosomal regions appear more vulnerable. Loss of part of a chromosome is referred to as a deletion. If the deletion leads to loss of neighboring genes this may result in a contiguous gene disorder, such as a deletion on the X chromosome giving rise to X-linked ichthyosis (see Chapter 49) and Kallman syndrome. If two chromosomes break, the detached fragments may be exchanged, known as reciprocal translocation. If this process involves no loss of DNA it is referred to as a balanced translocation. Other structural aberrations include duplication of sections of chromosomes, two breaks within one chromosome leading to inversion, and fusion of the ends of two broken chromosomal arms, leading to joining of the ends and formation of a ring chromosome. Chromosomal anomalies may be detected using standard metaphase cytogenetics but newer approaches, such as SNP arrays and comparative genomic hybridization arrays, can also be used for karyotyping. Array-based cytogenetic tools do not rely on cell division and are very sensitive in detecting unbalanced lesions as well as copy number-neutral loss of heterozygosity. These new methods have become commonplace in diagnostic genetics laboratories. A further possible chromosomal abnormality is the inheritance of both copies of a chromosome pair from just one parent (paternal or maternal), known as uniparental disomy.16 Uniparental heterodisomy refers to the presence of a pair of chromosome homologs, whereas uniparental isodisomy describes two identical copies of a single homolog, and meroisodisomy is a mixture of the two. Uniparental disomy with homozygosity of recessive alleles is being increasingly recognized as the molecular basis for several autosomal recessive disorders, and there have been more than 35 reported cases of recessive diseases, including junctional and dystrophic EB (see Chapter 62), resulting from this type of chromosomal abnormality. For certain chromosomes, uniparental disomy can also result in distinct phenotypes depending on the parental origin of the chromosomes, a phenomenon known as genomic imprinting.17,18 This parent-of-origin, specific gene expression is determined by epigenetic modification of a specific gene or, more often, a group of genes, such that gene transcription is altered, and only one inherited copy of the relevant imprinted gene(s) is expressed in the embryo. This means that, during development, the parental genomes function unequally in the offspring. The most common examples of genomic imprinting are Prader–Willi (OMIM #176270) and Angelman (OMIM #105830) syndromes, which can result from maternal or paternal uniparental disomy for chromosome 15, respectively. Three phenotype abnormalities commonly associated with uniparental disomy for chromosomes with imprinting are (1) intrauterine growth retardation, (2) developmental delay, and (3) reduced stature.19

In addition to the 3.3 billion bp nuclear genome, each cell contains hundreds or thousands of copies of a further 16.5-kb mitochondrial genome, which is inherited solely from an individual's mother. This closed, circular genome contains 37 genes, 13 of which encode proteins of the respiratory chain complexes, whereas the other 24 genes generate 22 transfer RNAs and two ribosomal RNAs used in mitochondrial protein synthesis.20 Mutations in mitochondrial DNA were first reported in 1988, and more than 250 pathogenic point mutations and genomic rearrangements have been shown to underlie a number of myopathic disorders and neurodegenerative diseases, some of which show skin manifestations, including lipomas, abnormal pigmentation or erythema, and hypo- or hypertrichosis.21 Mitochondrial DNA mutations are very common in somatic mammalian cells, more than two orders of magnitude higher than the mutation frequency in nuclear DNA.22 Mitochondrial DNA has the capacity to form a mixture of both wild-type and mutant DNA within a cell, leading to cellular dysfunction only when the ratio of mutated to wild-type DNA reaches a certain threshold. The phenomenon of having mixed mitochondrial DNA species within a cell is known as heteroplasmy. Mitochondrial mutations can induce, or be induced by, reactive oxygen species, and may be found in, or contribute to, both chronologic aging and photoaging.23 Somatic mutations in mitochondrial DNA have also been reported in several premalignant and malignant tumors, including malignant melanoma, although it is not yet known whether these mutations are causally linked to cancer development or simply a secondary bystander effect as a consequence of nuclear DNA instability. Indeed, currently there is little understanding of the interplay between the nuclear and mitochondrial genomes in both health and disease. Nevertheless, it is evident that the genes encoded by the mitochondrial genome have multiple biologic functions linked to energy production, cell proliferation, and apoptosis.24

For Mendelian disorders, identifying genes that harbor pathogenic mutations has become relatively straightforward, with hundreds of disease-associated genes being discovered through a combination of linkage, positional cloning, and candidate gene analyses. By contrast, for complex traits, such as psoriasis and atopic dermatitis, these traditional approaches have been largely unsuccessful in mapping genes influencing the disease risk or phenotype because of low statistical power and other factors.25,26 Complex traits do not display simple Mendelian patterns of inheritance, although genes do have an influence, and close relatives of affected individuals may have an increased risk. To dissect out genes that contribute and influence susceptibility to complex traits, several stages may be necessary, including establishing a genetic basis for the disease in one or more populations; measuring the distribution of gene effects; studying statistical power using models; and carrying out marker-based mapping studies using linkage or association. It is possible to establish quantitative genetic models to estimate the heritability of a complex trait, as well as to predict the distribution of gene effects and to test whether one or more quantitative trait loci exist. These models can predict the power of different mapping approaches, but often only provide approximate predictions. Moreover, low power often limits other strategies such as transmission analyses, association studies, and family-based association tests. Another potential pitfall of association studies is that they can generate spurious associations due to population admixture. To counter this, alternative strategies for association mapping include the use of recent founder populations or unique isolated populations that are genetically homogeneous, and the use of unlinked markers (so-called genomic controls) to assign different regions of the genome of an admixed individual to particular source populations. In addition, and relevant to several studies on psoriasis, linkage disequilibrium observed in a sample of unrelated affected and normal individuals can also be used to fine-map a disease susceptibility locus in a candidate region.

In recent years, advances in the identification of many millions of SNPs across the entire genome, as well as major advances in gene chip technology that allows up to 2 million SNPs to be typed in a given individual for a few hundred dollars, coupled with high powered computation, have led to the current era of genome-wide association studies (GWAS).27 This has become the predominant technology for tacking complex traits, with GWAS having already been performed for psoriasis, atopic eczema, vitiligo, and alopecia areata. GWAS for other dermatological complex traits are underway. A typical GWAS design involves collecting DNA from a well-phenotyped case series of the condition of choice, preferably from an ethnically homogenous population. Normally, 2,000 or more cases are required versus 3,000 ethnically matched random population controls. Correct clinical ascertainment of the cases is paramount and so GWAS represents a great opportunity for close cooperation between physicians and scientists. These 5,000 or more individuals are genotyped for 500,000 to 2 million SNPs, generating billions of data points. For each SNP across the genome, a statistical test is performed and a P value derived. If an SNP is closely linked to a disease susceptibility gene, then a particular genotype will be greatly enriched in the case series compared to the general unselected population. The P values are plotted along each chromosome (“Manhattan plot”) and where disease susceptibility loci exist, there are clusters of strong association. Typically, P values of 10−10 or lower are indicative of a true locus, although this generally has to be replicated in a number of other case-control sets for confirmation. Although SNP-based GWAS is currently the weapon of choice in complex trait genetics, it has limitations. If a causative lesion in a susceptibility locus is very heterogeneous, i.e., if there are multiple mutations or other changes that cause the susceptibility, then the locus is poorly identified by GWAS. Furthermore, across the entire field of complex trait genetics, relatively few causative genes have emerged (the role of the filaggrin gene in atopic dermatitis, below, being a notable exception). In the majority of cases, there is currently little clue about what defect the associated SNPs are linked to that actually causes the disease susceptibility.

However, recently, a conventional genetics approach has revealed fascinating new insight into the pathophysiology of one particular complex trait, namely atopic dermatitis (eczema). This finding emanated from the discovery that the disorder ichthyosis vulgaris was due to loss-of-function mutations in the gene encoding the skin barrier protein filaggrin (see Chapters 14 and 49).28 To dermatologists, the clinical association between this condition and atopic dermatitis is well known, and the same loss-of-function mutations in filaggrin have subsequently been shown to be a major susceptibility risk factor for atopic dermatitis, as well as asthma associated with atopic dermatitis, but not asthma alone.4 This suggests that asthma in individuals with atopic dermatitis may be secondary to allergic sensitization, which develops because of the defective epidermal barrier that allows allergens to penetrate the skin to make contact with antigen-presenting cells. Indeed, transmission–disequilibrium tests have demonstrated an association between filaggrin gene mutations and extrinsic atopic dermatitis associated with high total serum immunoglobulin E levels and concomitant allergic sensitizations.29 These recent data on the genetics of atopic dermatitis demonstrate how the study of a “simple” genetic disorder can also provide novel insight into a complex trait. Therefore, Mendelian disorders may be useful in the molecular dissection of more complex traits.30

The presence of a mixed population of cells bearing different genetic or chromosomal characteristics leading to phenotypic diversity is referred to as mosaicism. There are several different types of mosaicism, including single gene, chromosomal, functional, and revertant mosaicism.31 Multiple expression patterns are recognized.32

Mosaicism for a single gene, referred to as somatic mosaicism, indicates a mutational event occurring after fertilization. The earlier this occurs, the more likely it is that there will be clinical expression of a disease phenotype as well as involvement of gonadal tissue (gonosomal mosaicism); for example, when individuals with segmental neurofibromatosis subsequently have offspring with full-blown neurofibromatosis (see Chapter 141). However, in general, if the mutation occurs after generation of cells committed to gonad formation, then the mosaicism will not involve the germ line, and the reproductive risk of transmission is negligible. Gonosomal mosaicism refers to involvement of both gonads and somatic tissue, but mosaicism can occur exclusively in gonadal tissue, referred to as gonadal mosaicism. Clinically, this may explain recurrences among siblings of autosomal dominant disorders such as tuberous sclerosis or neurofibromatosis, when none of the parents has any clinical manifestations and gene screening using genomic DNA from peripheral blood samples yields no mutation. Segmental mosaicism for autosomal dominant disorders is thought to occur in one of two ways: either there is a postzygotic mutation with the skin outside the segment and genomic DNA being normal (type 1), or there is a heterozygous genomic mutation in all cells that is then exacerbated by loss of heterozygosity within a segment or along the lines of Blaschko (type 2). This pattern has been described in several autosomal dominant disorders, including Darier disease, Hailey–Hailey disease (see Chapter 51), superficial actinic porokeratosis (see Chapter 52), and tuberous sclerosis (see Chapter 140).

The lines of Blaschko were delineated over 100 years ago; the pattern is attributed to the lines of migration and proliferation of epidermal cells during embryogenesis (i.e., the bands of abnormal skin represent clones of cells carrying a mutation in a gene expressed in the skin).33 Apart from somatic mutations [either in dominant disorders, such as epidermolytic ichthyosis (formerly called bullous congenital ichthyosiform erythroderma) leading to linear epidermolytic ichthyosis (epidermal nevus of the epidermolytic hyperkeratosis type) (see Chapter 49), or in conditions involving mutations in lethal dominant genes such as in McCune–Albright syndrome], mosaicism following Blaschko's lines is also seen in chromosomal mosaicism and functional mosaicism (random X-chromosome inactivation through lyonization). Monoallelic expression on autosomes (with random inactivation of either the maternal or paternal allele) is also feasible, and probably underdocumented.34 Chromosomal mosaicism results from nondisjunction events that occur after fertilization. Clinically, this is found in the linear mosaic pigmentary disorders (hypomelanosis of Ito (see Chapter 75) and linear and whorled hyperpigmentation). It is important to point out that hypomelanosis of Ito is not a specific diagnosis but may occur as a consequence of several different chromosomal abnormalities that perturb various genes relevant to skin pigmentation, which has led to the term “pigmentary mosaicism” to describe this group of disorders.

Functional mosaicism relates to genes on the X chromosome, because during embryonic development in females, one of the X chromosomes, either the maternal or the paternal, is inactivated. For X-linked dominant disorders, such as focal dermal hypoplasia (Goltz syndrome) or incontinentia pigmenti (see Chapter 75), females survive because of the presence of some cells in which the X chromosome without the mutation is active and able to function. For males, these X-linked dominant disorders are typically lethal, unless associated with an abnormal karyotype (e.g., Klinefelter syndrome; 47, XXY) or if the mutation occurs during embryonic development. For X-linked recessive conditions, such as X-linked recessive hypohidrotic ectodermal dysplasia (see Chapter 142), the clinical features are evident in hemizygous males (who have only one X chromosome), but females may show subtle abnormalities due to mosaicism caused by X-inactivation, such as decreased sweating or reduced hair in areas of the skin in which the normal X is selectively inactivated. There are 1,317 known genes on the X chromosome, and most undergo random inactivation but a small percentage (approximately 27 genes on Xp, including the steroid sulfatase gene, and 26 genes on Xq) escape inactivation.

Revertant mosaicism, also known as natural gene therapy, refers to genetic correction of an abnormality by various different phenomena including back mutations, intragenic crossovers, mitotic gene conversion, and second site mutations.35,36 Indeed, multiple different correcting events can occur in the same patient. Such changes have been described in a few genes expressed in the skin, including the keratin 14, laminin 332, collagen XVII, collagen VII, and kindlin-1 (fermitin family homolog 1) genes in different forms of EB (Fig. 8-7; see Chapter 62). The clinical relevance of the conversion process depends on several factors, including the number of cells involved, how much reversal actually occurs, and at what stage in life the reversion takes place. Attempts have been made to culture reverted keratinocytes and graft them to unreverted sites,37 a pioneering approach that may have therapeutic potential for some patients.

Apart from mutations in nuclear DNA, mosaicism can also be influenced by environmental factors, such as viral DNA sequences (retrotransposons) that can be incorporated into nuclear DNA, replicate, and activate or silence genes through methylation or demethylation. This phenomenon is known as epigenetic mosaicism; such events may be implicated in tumorigenesis but have not been associated with any genetic skin disorder.

Disease phenotypes reflect the result of the interaction between a particular genotype and the environment, but it is evident that some variation, for example, in monozygotic twins, is attributable to neither. Additional influences at the biochemical, cellular, tissue, and organism levels occur, and these are referred to as epigenetic phenomena.38 Single genes are not solely responsible for each separate function of a cell. Genes may collaborate in circuits, be mobile, exist in plasmids and cytoplasmic organelles, and can be imported by nonsexual means from other organisms or as synthetic products. Even prion proteins can simulate some gene properties. Epigenetic effects reflect chemical modifications to DNA that do not alter DNA sequence but do alter the probability of gene transcription. Mammalian DNA methylation machinery is made up of two components: (1) DNA methyltransferases, which establish and maintain genome-wide DNA methylation patterns, and (2) the methyl-CpG-binding proteins, which are involved in scanning and interpreting the methylation patterns. Analysis of any changes in these processes is known as epigenomics.39 Examples of modifications include direct covalent modification of DNA by methylation of cytosines and alterations in proteins that bind to DNA. Such changes may affect DNA accessibility to local transcriptional complexes as well as influencing chromatin structure at regional and genome-wide levels, thus providing a link between genome structure and regulation of transcription. Indeed, epigenome analysis is now being carried out in parallel with gene expression to identify genome-wide methylation patterns and profiles of all human genes. For example, there is considerable interindividual variation in cytosine methylation of CpG dinucleotides within the major histocompatibility complex (MHC) region genes, although whether this has any bearing on the expression of skin disorders such as psoriasis remains to be seen. New sensitive and quantitative methylation-specific polymerase chain reaction-based assays can identify epigenetic anomalies in cancers such as melanoma.40 DNA hypermethylation contributes to gene silencing by preventing the binding of activating transcription factors and by attracting repressor complexes that induce the formation of inactive chromatin structures. With regard to melanoma, such changes may impact on several biologic processes, including cell cycle control, apoptosis, cell signaling, tumor cell invasion, metastasis, angiogenesis, and immune recognition. A further but as yet unresolved issue is whether there is heritability of epigenetic characteristics. Likewise, it is unclear whether environmentally induced changes in epigenetic status, and hence gene transcription and phenotype, can be transmitted through more than one generation. Such a phenomenon might account for the cancer susceptibility of grandchildren of individuals who have been exposed to diethylstilbestrol, but this has not been proved. However, germ line epimutations have been identified in other human diseases, such as colorectal cancers characterized by microsatellite instability and hypermethylation of the MLH1 DNA mismatch repair gene, although the risk of transgenerational epigenetic inheritance of cancer from such a mutation is not well established and probably small. Over the course of an individual's lifespan, epigenetic mutations (affecting DNA methylation and histone modifications) may occur more frequently than DNA mutations, and it is expected that, over the next decade, the role of epigenetic phenomena in influencing phenotypic variation will gradually become better understood.41

Human leukocyte antigen (HLA) molecules are glycoproteins that are expressed on almost all nucleated cells. The HLA region is located on the short arm of chromosome 6, at 6p21, referred to as the MHC. There are three classic loci at HLA class I: (1) HLA-A, (2) HLA-B, and (3) HLA-Cw, and five loci at class II: (1) HLA-DR, (2) HLA-DQ, (3) HLA-DP, (4) HLA-DM, and (5) HLA-DO. The HLA molecules are highly polymorphic, there being many alleles at each individual locus. Thus, allelic variation contributes to defining a unique “fingerprint” for each person's cells, which allows an individual's immune system to define what is foreign and what is self. The clinical significance of the HLA system is highlighted in human tissue transplantation, especially in kidney and bone marrow transplantation, where efforts are made to match at the HLA-A, -B, and -DR loci. MHC class I molecules, complexed to certain peptides, act as substrates for CD8+ T-cell activation, whereas MHC class II molecules on the surface of antigen-presenting cells display a range of peptides for recognition by the T-cell receptors of CD4+ T helper cells (see Chapter 10). Therefore, MHC molecules are central to effective adaptive immune responses. Conversely, however, genetic and epidemiologic data have implicated these molecules in the pathogenesis of various autoimmune and chronic inflammatory diseases. Several skin diseases, such as psoriasis (see Chapter 18), psoriatic arthropathy (central and peripheral), dermatitis herpetiformis, pemphigus, reactive arthritis syndrome (see Chapter 20), and Behçet disease (see Chapter 166), all show an association with inheritance of certain HLA haplotypes (i.e., there is a higher incidence of these conditions in individuals and families with particular HLA alleles). However, the molecular mechanisms by which polymorphisms in HLA molecules confer susceptibility to certain disorders are still unclear. This situation is further complicated by the fact that, for most diseases, it is unknown which autoantigens (presented by the disease-associated MHC molecules) are primarily involved. For many diseases, the MHC class association is the main genetic association. Nevertheless, for most of the MHC-associated diseases, it has been difficult to unequivocally determine the primary disease-risk gene(s), owing to the extended linkage disequilibrium in the MHC region. However, recent genetic and functional studies support the long-held assumption that common MHC class I and II alleles themselves are responsible for many disease associations, such as the HLA cw6 allele in psoriasis. Of practical clinical importance is the strong genetic association between certain HLA alleles and the risk of adverse drug reactions. For example, in Han Chinese and some other Asian populations, HLA-B*1502 confers a greatly increased risk of carbamazepine-induced Stevens–Johnson syndrome and toxic epidermal necrolysis. Therefore, screening for HLA-B*1502 before starting carbamazepine in patients from high-risk populations is recommended or required by regulatory agencies.42

The National Society of Genetic Counselors (http://www.nsgc.org) has defined genetic counseling as “the process of helping people understand and adapt to the medical, psychological and familial implications of genetic contributions to disease.” Genetic counseling should include: (1) interpretation of family and medical histories to assess the chance of disease occurrence or recurrence; (2) education about inheritance, testing, management, prevention, resources, and research; and (3) counseling to promote informed choice and adaptation to the risk or condition.43

Once the diagnosis of an inherited skin disease is established and the mode of inheritance is known, every dermatologist should be able to advise patients correctly and appropriately, although additional support from specialists in medical genetics is often necessary. Genetic counseling must be based on an understanding of genetic principles and on a familiarity with the usual behavior of hereditary and congenital abnormalities. It is also important to be familiar with the range of clinical severity of a particular disease, the social consequences of the disorder, the availability of therapy (if any), and the options for mutation detection and prenatal testing in subsequent pregnancies at risk for recurrence (one useful site is http://www.genetests.org).

A key component of genetic counseling is to help parents, patients, and families know about the risks of recurrence or transmission for a particular condition. This information is not only practical but often relieves guilt and can allay rather than increase anxiety. For example, it may not be clear to the person that he or she cannot transmit the given disorder. The unaffected brother of a patient with an X-linked recessive disorder such as Fabry disease (see Chapter 136), X-linked ichthyosis (see Chapter 49), Wiskott–Aldrich syndrome (see Chapter 143), or Menkes syndrome (see Chapter 88) need not worry about his children being affected or even carrying the abnormal allele, but he may not know this.

Prognosis and counseling for conditions such as psoriasis in which the genetic basis is complex or still unclear is more difficult (see Chapter 18). Persons can be advised, for example, that if both parents have psoriasis, the probability is 60% to 75% that a child will have psoriasis; if one parent and a child of that union have psoriasis, then the chance is 30% that another child will have psoriasis; and if two normal parents have produced a child with psoriasis, the probability is 15% to 20% for another child with psoriasis.44 Ongoing discoveries in other diseases, such as melanoma genetics, can also impact on genetic counseling. The identification of family-specific mutations in the CDKN2A and CDK4 genes, as well as risk alleles in the MC1R and OCA2 genes and other genetic variants, allow for more accurate and informative patient and family consultations.45

In recent years, there has been considerable progress in developing prenatal testing for severe inherited skin disorders (Fig. 8-8). Initially, ultrastructural examination of fetal skin biopsies was established in a limited number of conditions. In the late 1970s, the first diagnostic examination of fetal skin was reported for epidermolytic hyperkeratosis and Herlitz junctional EB (see Chapter 62).46,47 These initial biopsies were performed with the aid of a fetoscope to visualize the fetus. However, with improvements in sonographic imaging, biopsies of fetal skin are now taken under ultrasound guidance. The fetal skin biopsy samples obtained during the early 1980s could be examined only by light microscopy and transmission electron microscopy. However, the introduction of a number of monoclonal and polyclonal antibodies to various basement membrane components during the mid-1980s led to the development of immunohistochemical tests to help complement ultrastructural analysis in establishing an accurate diagnosis, especially in cases of EB.48 Fetal skin biopsies are taken during the midtrimester. For disorders such as EB, testing at 16 weeks’ gestation is appropriate. However, for some forms of ichthyosis, the disease-defining structural pathology may not be evident at this time, and fetal skin sampling may need to be deferred until 20 to 22 weeks of development.

Nevertheless, since the early 1990s, as the molecular basis of an increasing number of genodermatoses has been elucidated, fetal skin biopsies have gradually been superseded by DNA-based diagnostic screening using fetal DNA from amniotic fluid cells or samples of chorionic villi; the latter are usually taken at 10 to 12 weeks’ gestation (i.e., at the end of the first trimester).49,50 In addition, advances with in vitro fertilization and embryo micromanipulation have led to the feasibility of even earlier DNA-based assessment through preimplantation genetic diagnosis, an approach first successfully applied in 1990, for risk of recurrence of cystic fibrosis.51 Successful preimplantation testing has also been reported for severe inherited skin disorders.52 This is likely to become a more popular, though still technically challenging, option for some couples, in view of recent advances in amplifying the whole genome in single cells and the application of multiple linkage markers in an approach termed preimplantation genetic haplotyping.53 This approach has been developed and applied successfully for Herlitz junctional epidermolysis bullosa.54 For some disorders, alternative less invasive methods of testing are now also being developed, including analysis of fetal DNA or RNA from within the maternal circulation and the use of three-dimensional ultrasonography.

In the current absence of effective treatment for many hereditary skin diseases, prenatal diagnosis can provide much appreciated information to couples at risk of having affected children, although detailed and supportive genetic counseling is also a vital element of all prenatal testing procedures.

The field of gene therapy can be subdivided in different ways.55 First, there are approaches aimed at treatment of recessive genetic diseases where homozygous or compound heterozygous loss-of-function mutations lead to complete absence or complete functional ablation of a vital protein. These types of diseases are amenable to gene replacement therapy, and it is this form of gene therapy that has tended to predominate because it is generally technically more feasible than treatment of dominant genetic conditions.56 In dermatology, these include diseases such as lamellar ichthyosis (see Chapter 49), where in most cases, there is hereditary absence of transglutaminase-1 activity in the outer epidermis, or the severe Hallopeau–Siemens form of recessive dystrophic EB, where there is complete absence of type VII collagen expression due to recessive mutations.57

The second form of gene therapy, in broad terms, is aimed at treatment of dominant-negative genetic disorders and is known as gene inhibition therapy. Here, there is a completely different type of problem to be tackled because these patients already carry one normal copy of the gene and one mutated copy. The disease results because an abnormal protein product produced by the mutant allele, dominant-negative mutant protein, binds to and inhibits the function of the normal protein produced by the wild-type allele. In many cases, it can be shown from the study of rare recessive variants of dominant diseases that one allele is sufficient for normal skin function, and so if a means could be found of specifically inhibiting the expression of the mutant allele, this should be therapeutically beneficial. However, finding a gene therapy agent that is capable of discriminating the wild-type and mutant alleles, which can differ by as little as one bp of DNA, is challenging and, until recently has resulted in little success. A typical dominant-negative genetic skin disease is EB simplex (see Chapter 62), caused by mutations in either of the genes encoding keratins 5 or 14. The vast majority of cases are caused by dominant-negative missense mutations, changing only a single amino acid, carried in a heterozygous manner on one allele.58

Gene therapy approaches can also be broadly subdivided according to whether they involve in vivo or ex vivo strategies.55 Using an in vivo approach, the gene therapy agent would be applied directly to the patient's skin or another tissue. A disadvantage of the skin as a target organ for gene therapy is that it is a barrier tissue that is fundamentally designed to prevent entry of foreign nucleic acid in the form of viruses or other pathogenic agents. This is an impediment to in vivo gene therapy development but is not insurmountable due to developments in liposome technology and other methods for cutaneous macromolecule delivery.59 In an ex vivo approach, a skin biopsy would be taken, keratinocytes or fibroblasts would be grown and expanded in culture, treated with the gene therapy agent, and then grafted onto or injected back into the patient. The skin is a good organ system for both these approaches because it is very accessible for in vivo applications. In addition, the skin can be readily biopsied, and cell culture and regrafting of keratinocytes can be adapted for ex vivo gene therapy.

Gene replacement therapy systems have been developed for lamellar ichthyosis (see Chapter 49) and the recessive forms of EB (see Chapter 62), among other diseases. These mostly consist of expressing the normal complementary DNA encoding the gene of interest from some form of gene therapy vector adapted from viruses that can integrate their genomes stably into the human genome. Therefore, such viral vectors can lead to long-term stable expression of the replacement gene.60 Early studies tended to use retroviral vectors or adeno-associated viral vectors, but these have a number of limitations. For example, retroviruses only transduce dividing cells and therefore fail to target stem cells; consequently, gene expression is quickly lost due to turnover of the epidermis through keratinocyte differentiation. Furthermore, there have been some safety issues in small-scale human trials for both retroviral and adeno-associated viral vectors. Lentiviral vectors, derived from short integrating sequences found in a number of immunodeficiency viruses, have the advantage of being able to stably transduce dividing and nondividing cells, with close to 100% efficiency and at low copy number. These may be the current vectors of choice, but they also have potential problems because their preferred integration sites in the human genome are currently ill defined and may lead to concerns about safety. However, with a wide variety of vectors under development and testing, it should become clear in future years which vectors are effective and safe for human use. Ultimately, like all novel therapeutics, animal testing can only act as a guide because the human genome is quite different and may react differently to foreign DNA integration, so that phase I, II, and III human trials or adaptations thereof will ultimately have to be performed to determine efficacy and safety. Currently, small-scale clinical trials are ongoing for junctional EB and are planned for a number of other genodermatoses, mainly concentrating on the more severe recessive conditions.

Treatment of dominant-negative disorders has recently started to receive a great deal of attention with the discovery of the RNA inhibition pathway in humans and the finding that small synthetic double-stranded RNA molecules of 19 to 21 bp, known as short inhibitory RNA (siRNA), can efficiently inhibit expression of human genes in a sequence-specific, user-defined manner.58,61 There is currently a great deal of attention being focused on development of siRNA inhibitors to selectively silence mutant alleles in dominant-negative genetic diseases, such as the keratin disorders—EB simplex and pachyonychia congenita (PC). Currently, the big challenge in this rapidly evolving new field is finding an effective, noninvasive method to get siRNA through the stratum corneum and into keratinocytes or other target cells. A number of groups are working on means of delivering siRNA to skin and other organ systems, and there is currently much optimism about these developing into clinically applicable agents in the near future.

In particular, a great deal of rapid progress has been made in PC in recent years. Following development of reporter gene methodology to rapidly screen many different siRNA species, two siRNAs were identified that could specifically and potently silence mutant keratin K6a mRNA differing from the wild-type mRNA by only a single nucleotide, i.e., these siRNAs represent allele-specific gene silencing agents. Following a battery of preclinical studies in cells and animal models to show efficacy, the K6a mutation-specific siRNA was manufactured to Good Manufacturing Practice standards and was shown to have an excellent toxicity profile in rodents, as per a small molecule drug. This facilitated FDA approval for a double blind split body Phase 1b clinical trial in a single volunteer with PC. The trial was successful, with a number of objective measures showing statistically significant clinical improvement. This study, funded by the patient advocacy organization PC Project (www.pachyonychia.org), was the first in human siRNA trial using a mutation-specific gene silencing approach and only the fifth siRNA trial in humans. This personalized medicine strategy gives hope for patients with incurably dominant genodermatoses and future trials in EB simplex are currently in the planning stages.