Supervisors: Niklas Dahl, Professor, M.D.
Anders Vahlquist, Professor, M.D.
Uppsala University
Uppsala, Sweden
Faculty opponent: Juha Kere, Professor, M.D.
Karolinska Institute
Stockholm, Sweden
Review board: Göran Andersson, Professor University of Agricultural Sciences
Uppsala, Sweden
Mona Ståhle, Professor, M.D.
Karolinska Institute
Stockholm, Sweden
Kerstin Lindblad-Toh, Professor
Uppsala University
Uppsala, Sweden; &
Broad Institute of Harvard and MIT Cambridge, MA, US
Chairman: Marie-Louise Bondeson, Associate Professor
Uppsala University
Uppsala, Sweden
List of Papers
This thesis is based on the following papers, which are referred to in the text by their Roman numerals.
I Dahlqvist, J.*, Klar, J.*, Hausser, I., Vahlquist, A., Pigg, M., Gedde-Dahl Jr, T., Dahl, N. (2007) Congenital ichthyosis: Mu- tations in Ichthyin are associated with specific structural abnor- malities in the granular layer of epidermis. J Med Gen, 44(10):615-20.
II Dahlqvist, J., Westermark, G., Vahlquist, A., Dahl, N. Ichthyin localizes to keratins and desmosomes in epidermis and is in- volved in lipid metabolism. Manuscript.
III Dahlqvist, J., Klar, J., Tiwari, N., Schuster, J., Törmä, H., Bad- hai, J., Pujol, R. , van Steensel, M.A.M., Brinkhuijzen, T., Gi- jezen, L., Chaves, A., Tadini, G., Vahlquist, A., Dahl, N. (2010) A single-nucleotide deletion in the POMP 5’ UTR causes a transcriptional switch and altered epidermal proteasome distri- bution in KLICK genodermatosis. Am J Hum Genet, 9;86(4):596-603.
IV Dahlqvist, J., Törmä, H., Badhai, J., Dahl, N. siRNA silencing of proteasome maturation protein (POMP) expression activates the unfolded protein response and constitutes a model for KLICK genodermatosis. Manuscript submitted.
*These authors contributed equally to the work.
Reprints were made with permission from the respective publishers.
Contents
Introduction ... 11
The human genome ... 11
Genetic variation ... 13
Repetitive sequence ... 13
Structural variants ... 13
Single nucleotide polymorphisms ... 14
Disease-associated variants ... 14
The transcriptome ... 15
mRNA ... 15
Structural and regulatory RNA ... 16
Epigenetics ... 17
Protein turnover ... 17
The proteasome ... 17
The unfolded protein response ... 19
Human genetic disease ... 20
Monogenic disorders ... 20
The human skin ... 22
Autosomal recessive congenital ichthyosis ... 23
KLICK syndrome ... 24
Methods in genetic and molecular studies ... 25
Mapping of disease genes ... 25
Identification of disease-causing mutations ... 26
Gene expression analyses ... 27
Protein expression analyses ... 29
Present investigations ... 31
Aims ... 31
Autosomal recessive congenital ichthyosis (Paper I-II) ... 32
Subjects and samples ... 32
Results and discussion ... 32
KLICK syndrome (Paper III-IV) ... 37
Subjects and samples ... 37
Results and discussion ... 37
Concluding remarks and future perspectives ... 43
Svensk sammanfattning av avhandlingsarbetet ... 46
Genetiska och molekylära studier av ärftliga hudsjukdomar ... 46
Acknowledgements ... 48
References ... 50
Abbreviations
A Adenine
ABCA12 ATP-binding cassette transporter 12 ALOX12B Arachidonate 12-lipoxygenase ALOXE3 Arachidonate lipoxygenase 3
ARCI Autosomal recessive congenital ichthyosis ATF4 Activating transcription factor 4
ATF6 Activating transcription factor 6 BiP Immunoglobulin binding protein bp Base pairs
C Cytosine
cDNA Complementary DNA CHOP C/EBP homologous protein
CYP4F22 Cytochrome P450, family 4, subfamily F, polypeptide 22 DNA Deoxyribonucleic acid
EM Electron microscopy ER Endoplasmic reticulum FATP4 Fatty acid transport protein 4 FBS Fetal bovine serum
G Guanine
HGMD Human gene mutation database HIV Human deficiency virus
kb Kilobases
KID Keratitis-ichthyosis-deafness
KLICK Keratosis linearis with ichthyosis congenita and keratoderma LI Lamellar ichthyosis
LOD Logarithm of odds
MW Mann-Whitney
Mb Megabases
Mg Magnesium
miRNA Micro ribonucleic acid
mRNA Messenger ribonucleic acid
NCIE Non-bullous congenital ichthyosiform erythroderma ncRNA Non-protein coding ribonucleic acid
nt Nucleotides
OMIM Online Mendelian Inheritance in Man ORF Open reading frame
OS One-sample PCR Polymerase chain reaction PEST Penicillin-streptomycin POMP Proteasome maturation protein
qPCR Quantitative real-time reverse transcriptase PCR RACE Rapid amplification of cDNA ends
RNA Ribonucleic acid rRNA Ribosome ribonucleic acid SD Standard deviation siRNA Small interfering ribonucleic acid SNP Single nucleotide polymorphism SV Structural variant T Thymine TGM1 Transglutaminase 1 tRNA Transfer ribonucleic acid TSS Transcription start site
UPR Unfolded protein response
UTR Untranslated region
Introduction
A genetic component is involved in most human diseases, either as a direct cause of a disease or as a factor that increases the susceptibility to a particu- lar disease. Substantial work worldwide has been, and is being dedicated to the investigation of genetic involvement in human disease in order to im- prove both the diagnosis and treatment of patients.
The present thesis is based on studies of two rare hereditary skin disord- ers; autosomal recessive congenital ichthyosis and KLICK syndrome. Al- though the number of patients affected by such rare disorders is low, much can be gained from the identification of genes and molecular pathways that are involved in the disorders. These new insights increase the understanding of normal human physiology and can also provide clues to the molecular pathways involved in more common human diseases.
The human genome
When Gregor Mendel, in 1865, described the laws of inheritance based on experiments on pea plants he initiated the era of modern genetics.
1Since then the field has expanded enormously and much effort has been directed towards elucidating the characteristics of the human genome.
The main part of the human genome is organized as chromosomes and is
located within the nucleus of almost all cells of the body (Figure 1). There
are 23 pairs of chromosomes with one chromosome in every pair being inhe-
rited from each parent. The genome is built up of DNA consisting of four
organic bases: adenine (A), thymine (T), cytosine (C) and guanine (G). The
bases are bound to each other through a backbone of sugar and phosphate,
forming nucleotides. The nucleotides are arranged pairwise, with each pair
bound to the next in a long chain coiled as a helix.
2The chain of base pairs
(bp) is wrapped around proteins (histones), forming nucleosomes. These are
packed together either densely in regions called heterochromatin, or more
loosely in regions called euchromatin.
3Figure 1. The nuclear DNA is stored in chromosomes, which are DNA strands tightly wound around histones into densely packed nucleosomes. Adapted from Access Excellence, National Healthy museum;
http://www.accessexcellence.org/RC/VL/GG/chromosome.php.
An important event in the history of human genetics occurred in 2001 when the first draft of the entire human genome sequence was published.
4; 5Three years later the sequence was almost completely described, and it is now clearly understood that the haploid human genome consists of approx- imately three billion base pairs, comprising 20,000-23,000 protein coding genes.
6Initially it was considered surprising that the human genome does not contain more genes than a “simple” nematode and that the protein coding genes constitute only 1-2% of the DNA.
7It soon became clear, however, that the non-protein coding part of the genome may also be of functional impor- tance, as gene regulatory regions, and with a vast majority of the genome being transcribed. The greater part of all these transcripts is never translated into amino acids, but is thought to exercise a function in its own right.
Hence, the old dogma of “one gene makes one transcript makes one protein”
no longer holds true as the reality proves to be much more complex. It has
been suggested that it is the extent and complexity of the non-protein coding
DNA and RNA that distinguishes human beings as higher organisms from lower organisms.
8-11Genetic variation
The human genome holds a considerable amount of genetic variation be- tween individuals. This variation contributes both to the diversity of human biology and to disease. The extent of inter-individual genomic differences is becoming increasingly clear as the entire genome of more individuals is sequenced.
6; 12-14There are several different kinds of genetic variants, with various frequencies in different ethnic populations. A genetic variation is defined as polymorphic if it occurs at a frequency of ≥ 1% in a certain popu- lation.
15As man carries two copies of each chromosome, two copies of each polymorphism are also carried, hence with the presence of two possible va- riants, or alleles.
Repetitive sequence
Approximately 46% of the human genome consists of repetitive sequence, which appears in two different subtypes. The first category, interspersed repeats, originates from transposons (moveable genomic elements) that are scattered throughout the genome with little variation between individuals.
The second category, tandem repeats, are repeats of a DNA sequence of 2-60 nucleotides (nt) that follow each other in tandem. The number of repetitions can vary greatly, from 10 to 100, between different individuals. Repeats with units of 9 nt or fewer are referred to as microsatellites whilst repeats of more than 9 nt are named minisatellites.
16Structural variants
The knowledge regarding duplications, deletions and inversions of short or
long sequences of DNA in the human genome has increased rapidly over the
past few years. The acceleration in discovery of these alterations, called
structural variants (SV’s), started with the development of a new technology
based on genome-wide BAC arrays (array-comparative genomic hybridiza-
tion), following the publication of the complete human genome.
17As the
technologies have advanced towards dense oligonucleotide arrays and mas-
sively parallel sequencing, the number of identified SV’s has increased and
smaller SV’s (indels) can now be identified. Gradually the concept of SV’s
and repetitive sequence as being different entities is being questioned and
they are now often considered as one category.
18; 19There are currently more
than 66,000 SV’s (of which ~34,000 consist of 100-1000 bp) reported in the
Database of Genomic Variants (http:// projects.tcag.ca/variation), a collec-
tion of genomic information of healthy individuals. This is probably only the beginning of an era of recognizing SV’s, since more SV’s will be identified and the consequences of such variants elucidated as technologies improve.
Single nucleotide polymorphisms
Single nucleotides that vary between individuals are scattered throughout the genome; they are defined as single nucleotide polymorphisms (SNP’s). A SNP can either be a base substituted for another base, or the deletion of or insertion of a base. Ten years ago between 1.4 and 2.1 million different SNP’s had been identified within the human genome, a number that has now increased to almost 38 million according to NCBI SNP database (dbSNP;
http://www.ncbi.nlm.nih.gov/projects/SNP/ ; build 132).
4; 5In 2002, the In- ternational HapMap project was initiated with the aim to genotype and cha- racterize SNP’s and SV’s in a large group of individuals of different geo- graphical origin (http://hapmap.ncbi.nlm.nih.gov/).
20The international colla- boration has resulted in an extensive map of the interrelations of common and rare genetic variants, greatly influencing the current knowledge of genet- ic diversity and studies of genetic disease.
21; 22Disease-associated variants
As the amount of human sequence data accumulates, it becomes increasingly clear that even gross genetic alterations do not necessarily result in disease, but can just simply be “normal” variants. The fine line between genetic vari- ations that promote disease and those that do not is blurred. In order to iden- tify genetic alterations that cause, or increase the risk of, disease it is impor- tant to first determine the common and rare variations that occur in (see- mingly) healthy individuals and to investigate their biological, or phenotyp- ic, effects. This is the purpose of The 1000 Genomes Project, in which the genome of 1000 individuals from different populations is being sequenced using high-throughput technologies.
14The complexity of human genetic variation in relation to health and disease is illustrated by the findings of deletions and duplications of more than one megabase (Mb), encompassing various genes, in healthy individuals.
17; 18Additionally, according to current estimations each individual differs from a reference human genome se- quence at about 10,000 non-synonymous sites, meaning that each individual has 10,000 amino acid changes or other disruptions in protein composition compared with the reference.
14All of the different genetic variations mentioned above can, however, also
cause or increase the risk of disease. In repeat expansion disorders tandem
repeats have a pathological increase in the number of repeat units, for exam-
ple, in Huntington’s disease where an increase in a CAG repeat results in a
neurotoxic protein.
23SV’s have been shown to both directly cause disease,
as in the case of a duplicated LMNB1 gene in adult-onset autosomal domi- nant leukodystrophy, and to affect susceptibility to disease, for example with different copies of a certain chemokine gene affecting the susceptibility to infection with HIV.
24; 25Common SNP’s can affect the susceptibility to complex disorders, such as Crohn’s disease and systemic lupus erythemato- sus.
26-28In addition to these large or common pathogenic variants, there are rare mutations of one or a few bases of DNA that cause disease. There are cur- rently almost 73,000 mutations, of 20 bp or less, registered in the Human gene mutation database (HGMD; http://www.hgmd.cf.ac.uk/ac/index.php).
The most common mutations in protein-coding genes are base substitutions (e.g. missense and nonsense mutations), or deletions or insertions of one or several nucleotides. These (and SNP’s likewise) may occur: in a protein coding sequence, thereby modulating the amino acid sequence; in splice sites, resulting in exon skipping or intron retention; in untranslated regions affecting transcript stability and translation; or, in promoters affecting gene expression. Disease causing mutations may also occur in non-protein coding parts of the genome, for instance, in non-protein coding RNA’s or in gene regulatory regions.
The transcriptome
During the past decade extensive studies of genomic expression have shown that the human genome is pervasively transcribed, from both sense and anti- sense DNA strands. Different transcripts overlap, as do regulatory DNA regions that control the expression of different transcripts.
9In general, there is more active transcription taking place in euchromatin than in heterochro- matin, as the DNA is more loosely packed and more accessible to transcrip- tion factors in those regions. In addition, transcription varies according to cell type, cell differentiation stage and environmental triggers. The tran- scripts are roughly divided into different categories according to their func- tions: messenger RNA’s (mRNA’s), structural RNA’s and regulatory RNA’s.
mRNA
Protein coding genes consist of a sequence of exons and introns that are tran-
scribed into pre-mRNA. The introns are then spliced out and the exons
joined together. The mature mRNA’s are transported to the cytoplasm where
they are translated by ribosomes into amino acids, forming a peptide. The
first (5’) and the last (3’) exon of an mRNA are normally untranslated re-
gions (UTR’s), meaning that the sequences are not translated into amino
acids. 5’ UTR’s have been shown to affect the translation rate of mRNA’s
by different means: regulatory proteins can bind to specific sequences in the UTR; the 5’ UTR can form secondary structures that affect binding of ribo- somes or proteins; and the UTR can contain upstream open reading frames (upstream ORFs). These are short sequences that are translated into amino acids by the ribosome, thereby blocking translation of the “real” mRNA ORF. Long 5’ UTR’s are more likely to contain inhibitory secondary struc- tures or upstream ORF’s and generally cause a subsequent decrease in trans- lation rate compared with short 5’UTR’s.
29; 30The 3’ UTR determines tran- script stability and is important for the transportation of the mRNA to the correct cellular location.
31; 32One gene can have several different 5’ UTR and 3’ UTR variants, thus increasing the possibilities of gene regulation.
A major part of the mRNA’s, perhaps as much as 95%, is subject to alter- native splicing, in that different exons can be included in transcripts from one and the same gene.
33This gives rise to different variants of a protein, or even to different proteins, and often occurs in a tissue specific manner.
9; 34New data indicate that alternative exons can be located hundreds of kilobas- es (kb) away from the annotated transcription start site (TSS) of a gene and, moreover, that transcripts can be formed by fusion of exons from what has been regarded as different genes.
9For transcription of an mRNA to take place several transcription factors and RNA polymerase are required to cooperate. The transcription factors bind to a regulatory promoter region, often located in the vicinity of the gene TSS, and mediate binding of the RNA polymerase to the DNA.
9The enzyme uses one DNA strand as a template and transcribes a complementary strand of RNA. A gene can have multiple promoters and can share a promoter with another gene.
35Furthermore, DNA regulatory regions located considerably distant from the gene can affect gene expression. Enhancers are regions bound by proteins that can interact with promoter bound factors and activate gene transcription, while insulators are regions that can block such interac- tion.
36Structural and regulatory RNA
Structural and regulatory RNA’s exercise functions without being translated into proteins. There are two sub-types of structural RNA’s: ribosomal RNA (rRNA) and transfer RNA (tRNA). Both are important for protein translation as rRNA’s constitute substantial parts of the ribosomes and tRNA’s carry amino acids used for the peptide synthesis. rRNA and tRNA genes exist in multiple copies in the genome as their expression cannot be increased by augmented translation.
Regulatory RNA’s are non-protein coding RNA’s (ncRNA’s) that regu-
late, for instance, transcription, transcript modification and translation. The
ncRNA’s are commonly divided into small (<200 nt) and long (>200 nt)
ncRNA’s. A well-studied subclass of small ncRNA’s is classified as micro
RNA’s (miRNA’s).
10These RNA’s are processed and incorporated into miRNA-induced silencing complexes, from where they bind mRNA’s with complementary sequences, leading to translational repression or cleavage of the mRNA. miRNA’s have been shown to be important regulators of all known cellular processes, such as differentiation and proliferation.
32There are several other types of small ncRNA’s and new sub-classes are conti- nuously being identified.
10; 37; 38Long ncRNA’s, ranging in size from 200 nt to more than 100,000 nt, are less well understood. There is, however, con- vincing evidence for their involvement in regulation of, e.g., transcription and splicing.
39; 40Epigenetics
The epigenome can be defined as the combination of all modifications of the chromatin in a specific cell type that regulates the gene expression pattern.
The main chromatin modifications are DNA methylation and histone mod- ifications such as acetylation and methylation. The epigenetic characteristics are dynamic and change according to endogenous and environmental trig- gers. The chromatin modifications modulate the accessibility of the DNA to, e.g., transcription factors and regulatory RNA’s and affect cell differentia- tion and genomic imprinting among other things.
41Protein turnover
The mRNA sequences are translated into amino acid chains by ribosomes, often located at the membrane of the endoplasmic reticulum (ER). The nas- cent peptide is transported over the membrane into the ER, where it is folded into its mature shape by chaperones. After addition of post-translational modifications, such as glycosylation, the peptide is either integrated into the ER, or secreted into another cell compartment or to the extra-cellular space.
Once a protein is no longer of use it is degraded in order for the cell to maintain a balance between proteins in use and the recycling of cellular components. This is important, for example, during cell cycle progression, cell differentiation and nutrient starvation.
42; 43There are two main systems for protein degradation in the cell: autophagy and the proteasome. In auto- phagy double-membrane vesicles (autophagosomes) engulf proteins or orga- nelles to be degraded and then fuse with lysosomes. The content is degraded by acidic lysosomal hydrolases.
42; 44The proteasome
The proteasome is a large protein complex that degrades redundant and mis-
folded proteins. Proteins to be degraded are tagged by several copies of the
protein ubiquitin. These are recognized by specific parts of the proteasome, the two 19S regulatory particles, after which the proteolysis occur in the 20S core particle.
45Additionally, the 20S particle can recognize and degrade oxidized proteins.
46Each 19S particle comprises approximately 20 subunits of which some are responsible for recognizing and cleaving polyubiquitin chains and others for protein unfolding.
43The 20S particle is formed by four rings of subunits, bonded together into a barrel-shape (Figure 2). There are seven subunits in each ring; the two outermost rings, containing α subunits, are identical, as are the innermost rings with β subunits. The assembly of the 20S particle is initiated by the step-wise formation of an α-ring with α subunit 1-7, a process chaperoned by complexes PAC1-PAC2 and PAC3-PAC4. Next, subunits β1-7 are recruited and assembled precisely into a ring on the α-ring, forming a hemiproteasome. This recruitment is orchestrated by proteasome maturation protein (POMP), which is also thought to mediate binding of the premature proteasome to the ER membrane. When a hemiproteasome is complete, POMP mediates its dimerization with another hemiproteasome, at which the catalytical sites of the 20S particle (β1, β2, β5) are activated through autolysis. Finally, binding of the 19S particles to the 20S particle opens the α–rings for entry of substrates into the proteasome.
45; 47-50The β1, β2 and β5 subunits are associated with peptidylglutamyl-peptide
hydrolyzing, trypsin-like and chymotryptic-like enzymatic activities, respec-
tively. These activities can be inhibited by different agents. During the past
few years there has been a development of anti-cancer drugs that irreversibly
inhibit the chymotryptic-like proteasome activity. These drugs are thought to
induce apoptosis in malignant cells by intra-cellular aggregation of unfolded
and misfolded proteins as well as a stabilization of p53.
43; 51Figure 2. Model of 20S proteasome assembly. Alpha subunits (green) are assembled into a ring by chaperone complexes PAC1-PAC2 and PAC3-PAC4 (yellow). POMP (orange) mediates the binding of -rings to the ER membrane and the assembly of subunits (blue) to the complex. One complete -ring and -ring constitute a hemi- proteasome. POMP facilitates the dimerization of two hemiproteasomes into a ma- ture 20S proteasome and POMP and PAC1-PAC2 are the first substrates to be de- graded. Adapted with permission from Macmillan Publishers Ltd: EMBO reports,
48copyright 2007.
The unfolded protein response
Glucose deprivation, changes in calcium balance, ischemia and mutations that impair protein folding are all factors that can perturb ER function, re- sulting in accumulation of unfolded and misfolded proteins.
52Protein aggre- gates are injurious to the cell, for example, by disturbing functional proteins and by blocking lysosomal protein degradation.
53The accumulation of un- folded proteins is called ER stress and an extensive cell system, the unfolded protein response (UPR), has evolved to restore ER homeostasis. Immunog- lobulin binding protein (BiP) is a chaperone in the ER that binds to nascent, unfolded proteins and guides them into the ER. Under physiological condi- tions it is also bound to three so-called transducers of ER stress located in the ER membrane. When unfolded proteins accumulate, BiP dissociates from the ER stress transducers permitting them to be activated and to induce the UPR. The UPR pathways aim to normalize protein levels in the ER by different means: (i) the expression of ER chaperones is increased in order to remove unfolded proteins from the ER; (ii) general protein translation is hampered, minimizing entrance of new peptides into the ER; and (iii) ex-
1 2
4 3
5 6
7 PAC1-PAC2 PAC3-PAC4 -ring
POMP
ER
membrane
Hemiproteasome
20S proteasome
ER lumen 1
2 4
3
6 7
5
pression of chaperones that guide proteins towards proteasomal degradation is increased. If these actions are insufficient to restore ER balance apoptosis is induced through the activation of C/EBP homologous protein (CHOP) by transcription factors ATF4 and ATF6.
5452; 55Cells with substantial protein secretion, such as plasma cells and differen- tiated keratinocytes, have a physiological up-regulation of the UPR, in order to handle the extensive protein production.
56; 57The UPR can be actively induced by proteasome inhibiting drugs as these lead to an increased aggre- gation of unfolded and misfolded proteins.
56Human genetic disease
Genetic diseases are traditionally grouped into the following categories based on the genetic aberrations causing or influencing disease: chromosom- al disorders; mitochondrial disorders; monogenic disorders; and complex or multifactorial disorders. Chromosomal disorders are due to an atypical num- ber of chromosomes or to large structural abnormalities on chromosomes, whereas mitochondrial disorders result from mutations in mitochondrial DNA.
Complex disorders are caused by a combination of environmental factors and genetic variants that increase the risk of disease. The most common dis- orders in the Western world, e.g., high blood pressure and asthma, belong to this category. There are often many genetic variants involved in such disord- ers, each contributing slightly to the risk of disease. During the past five years thousands of genes have been associated with different complex dis- orders, through large genome-wide association studies of SNP’s in healthy and affected individuals.
27Monogenic disorders
The work leading to this thesis concerns Mendelian, or monogenic, disord- ers, which are due to mutations in a single gene. The Mendelian disorders are categorized according to the kind of chromosome that carries the muta- tion, and whether one or both copies of a gene are required to be mutant in order for an individual to be affected. Autosomal dominant and autosomal recessive disorders are caused by mutations on chromosome 1-22 (the auto- somes). Dominant inheritance means that only one gene copy needs to be mutant in order for an individual to be affected, whereas recessive disorders only cause disease if there are two mutated alleles. X-linked disorders are caused by mutations on the X-chromosome; Y-linked disorders, with muta- tions on the Y-chromosome, are extremely rare.
A complicating factor when studying the inheritance of a disorder is the
issue of penetrance. Some disorders are characterized by reduced penetrance,
meaning that not all individuals carrying the disease-associated genotype become symptomatic. In addition to this, many diseases show variability in expression with different symptoms, or severity of symptoms, in different individuals with the same mutation. This can be due to other genetic compo- nents that modify the phenotype or to environmental factors. On the other hand, one and the same phenotype can be caused by mutations in different genes and commonly by different mutations in a single gene.
Mendelian disorders, particularly autosomal recessive traits, are rare. For an offspring to be affected by these traits both parents need to be carriers of a disease allele, unless a mutation occurs de novo in the germline. The preva- lence of autosomal recessive disorders is increased in geographical areas where consanguineous marriages are common, as the parents in such fami- lies share a larger proportion of genetic variants than parents in non-related marriages.
Online Mendelian Inheritance in Man (OMIM;
http://www.ncbi.nlm.nih.gov/Omim) is a database of genetic disorders in
which there are currently more than 6,500 different monogenic diseases
(phenotypes) described. A major goal of the database is to link known phe-
notypic data to identified genes and molecular pathways.
The human skin
The surface of the skin, the epidermis, consists mainly of squamous epitheli- al cells, keratinocytes, organized into different histological layers: the basal, spinous, granular and horny layer (Figure 3). It is a resilient tissue able to resist physical insults and to keep microorganisms out. The epidermis re- news itself by division of proliferating cells located in the basal layer upon a basement membrane. After division some of the cells detach from the base- ment membrane and start to differentiate, moving outwards as this happens.
58During differentiation there is a switch in gene expression with an increase in factors important for formation of the outermost horny layer.
59The expression of specific types of keratin (cytoskeletal filaments) is in- creased as these are important for the strength of epidermis. At the stage of the granular cell layer the cells acquire keratohyaline granules that store different proteins, such as profilaggrin, and lamellar bodies, which are tubul- ous formations containing lipids and proteins. Filaggrin, the mature form of profilaggrin, is thought to aggregate keratin filaments into strong bundles, making the cells collapse into a flattened shape.
60The degradation products of filaggrin may also function as hydrating factors in the horny layer.
61Li- pids (e.g., ceramides) are synthesized, packaged into lamellar bodies and released into the inter-cellular space of the horny layer where they form la- mellae. A cornified cell envelope consisting of cross-linked proteins is formed around the cells and is further connected to intra-cellular keratins and inter-cellular lipids. As the cells move to the horny layer they constitute flat cell remnants, corneocytes, and are subsequently shed from the skin surface.
The cornified cell envelopes and the large amount of lipids in the horny layer function as an important barrier to water loss from the organism.
58; 62Terminal epidermal differentiation requires not only formation of large amounts of proteins and lipids de novo but also extensive proteolysis. Cer- tain cornified envelope proteins and profilaggrin require processing for acti- vation, and nuclei and mitochondria are degraded as corneocytes are formed.
Adhesion junctions and desmosomes organize inter-cellular adhesion in epi-
dermis and need to be degraded for proper desquamation. Only a proportion
of the proteolytic enzymes involved in this vast proteolysis is known.
58; 60; 63Figure 3. The human skin is divided into dermis, consisting of, e.g., fibroblasts and collagen, and epidermis, consisting mainly of stratified squamous epithelium. The cells of the basal layer are proliferating, whereas the cells of the spinous and granu- lar layer are differentiating in order to form the horny layer. The horny layer is the body’s barrier against water loss.
Autosomal recessive congenital ichthyosis
There is a wide range of skin disorders that are caused by disturbances in the differentiation of keratinocytes or formation of the epidermal horny layer.
Non-syndromic autosomal recessive congenital ichthyosis is a heterogeneous group of skin disorders that can be subdivided into two major clinical groups: one that exhibits widespread skin scaling of large brown scales and one that features generalized redness of the skin and fine white scaling.
There is, however, a large variation in symptoms between these groups and the phenotype can vary even within families.
64; 65Electron microscopy (EM) of the epidermis can be used to delineate distinct sub-phenotypes.
66Several mutant genes (TGM1, ABCA12, ALOXE3, ALOX12B, CYP4F22, FATP4, Ichthyin) have been identified in autosomal recessive congenital ichthyosis but the association between phenotype and genotype is poorly understood.
67-74Mutations in TGM1 can, for example, result in skin symp- toms of either clinical sub-group.
75Dysfunction of either of these proteins results in a disruption of the water barrier of the skin, and subsequently de- hydrated skin with extensive scaling.
76TGM1 is an enzyme essential for cross-linking of the cornified cell envelope and for anchoring ceramides of the horny layer to the cell envelope.
72; 77The gene products of ALOXE3, ALOX12B and FATP4 are important for specific lipid synthesis whereas ABCA12 is crucial, among other functions, for the transport of ceramides into lamellar bodies.
73; 78-81Knock-out mice have been generated for Tgm1, Alox12b and Abca12. The knock-out homozygous pups of all three kinds die within a few hours after birth due to extensive water loss through a defective
DERMIS Spinous layer Granular layer Horny layer
EPIDERMIS
Basal layer
water barrier.
79; 82; 83When examined by EM, the Tgm1 -/- mice display ab- normal keratohyaline granules and an absence of cornified cell envelopes in the epidermis.
82Alox12b knock-out results in numerous vacuoles and aber- rant vesicle-like lamellar bodies in the granular layer.
79Abca12 -/- mice show dislocalization of ceramides and absence of lamellar bodies and lipid lamellae.
83; 84Both Alox12b -/- and Abca12 -/- mice show abnormal filaggrin solubility or processing, underlining the complexity of interacting factors in terminal epidermal differentiation.
The exact functions of CYP4F22 and ichthyin in the skin remain un- known, but CYP4F22 is suggested to function as an enzyme in lipid syn- thsis.
70Ichthyin has been predicted to include transmembrane domains, sug- gesting it would be located in a membrane, and a previous study of magne- sium (Mg) transporters proposes that Ichthyin serves as a membrane bound cation transporter.
85At present there is no causal therapy available for ichthyosis. Topical treatments with moisturizing and keratolytic ointments, including salicylic acid or retinoids (vitamin A acid), are often used as these reduce the amount of scales. Oral retinoids can be administered in severe cases.
76KLICK syndrome
Keratosis Linearis with Ichthyosis Congenita and Keratoderma (KLICK) syndrome is a very rare skin disorder. The main manifestations are ichthyo- sis, thickened skin (hyperkeratosis) particularly on the palms and soles, kera- totic papules in a linear formation on flexure sides of wrists, elbows and knees, circular constrictions around fingers and flexion deformity of fingers.
Histology of the skin reveals wide granular and horny layers and EM analy- sis shows an increased amount of enlarged keratohyaline granules in the granular layer. KLICK is believed to follow an autosomal recessive mode of inheritance but the precise genetic cause has, as yet, remained elusive.
86-89A few similar disorders, such as Vohwinkel’s syndrome and Keratitis- Ichthyosis-Deafness (KID) syndrome, have been genetically characterized.
In both disorders mutations have been identified in the connexin26 gene, resulting in ichthyosis, hyperkeratosis and hearing loss, as well as constric- tions around fingers with auto-amputation and hyperkeratotic striae (Voh- winkel’s syndrome) and keratitis (KID syndrome).
90; 91Connexins form gap junctions, i.e., channels in the cell membrane allowing the passage of small molecules between cells. Vohwinkel’s syndrome without deafness can also be caused by mutant loricrin, an important component of the cornified cell envelope.
92The treatment for KLICK syndrome, Vohwinkel’s syndrome and KID
syndrome is usually oral or topical retinoids.
87; 91Methods in genetic and molecular studies
Mapping of disease genes
Linkage analysis
Positional cloning is often used as an approach to identify the mutant gene in Mendelian disorders and it was applied in Paper I and Paper III of this thesis.
This method identifies the gene by its position in the genome without any assumptions about its function.
93Genetic polymorphisms have been widely used for this, based on the principle of genetic linkage; genetic loci that are located close together on a chromosome tend to be inherited from parent to offspring as a block (haplotype). During meiosis homologous chromosomes exchange genetic material, in a process known as recombination. The greater the distance is between two loci, the greater the probability of a crossover between them, thereby breaking up the haplotype. Based on this, genetic polymorphisms with known genomic locations, such as microsatellites and SNP’s, can be used as genetic markers. Markers close to a disease mutation will be inherited with the disease more often than expected by random se- gregation and are, thus, linked to the disease locus. Genome wide analysis of marker alleles in families segregating disease allows for the mapping of a disease gene.
In linkage analysis the likelihood of a marker locus being linked to a dis- ease locus is statistically calculated. The probability of genetic linkage is given as logarithm of odds (LOD) score, which stands for the log
10ratio between the likelihood that two loci are linked rather than unlinked. For a monogenic trait a LOD score >3 is generally regarded as significant evi- dence of linkage, as this indicates 1000 times greater odds that the loci are linked than unlinked.
The advantage of using microsatellites as genetic markers is that they are highly polymorphic, making them suitable for linkage and haplotype analy- sis. They are, however, quite sparse (occurring on average once every 2kb), which can make it difficult to finely map a genomic region.
4; 94SNP’s are usually only bi-allelic, but are on the other hand frequent in the genome. The genotypes of SNP’s can be analyzed in microarrays. During the last five years this technology has developed immensely, from arrays analyzing 10,000 SNP’s in 2005 to arrays of more than 1 million SNP’s today.
In association studies the association between specific alleles and disease
are analyzed. A disease associated allele may either be causative itself, or be
in allelic association, linkage disequilibrium, with the disease-causing muta-
tion. Genome wide association studies of complex disorders analyze the
genotypes of up to a million SNP’s in thousands of disease cases and con-
trols, in order to identify association between disease and specific alleles.
Homozygosity mapping
A child of a consanguineous marriage affected with an autosomal recessive disease is likely to carry homozygous disease alleles, as the related parents are expected to carry the same mutation. As a consequence the haplotypes spanning the mutation locus are also expected to be homozygous by descent in the offspring, as recombination events are unlikely to occur between prox- imate loci. This is also the case for affected individuals from a restricted population that have inherited mutations (founder mutations) from a com- mon ancestor. Based on this, disease genes can be found by homozygosity mapping (called autozygosity mapping when the genomic regions are iden- tical by descent), which means searching for regions that are consistently homozygous in all affected individuals included in the study.
95This can be achieved using either microsatellites or, as in this thesis (Papers I and III), SNP’s as genetic markers.
Identification of disease-causing mutations
Depending on what kind of mutation is expected in a disorder different ap- proaches can be used to identify it. In the work leading to this thesis homo- zygosity mapping was used to delineate a candidate region, followed by gene sequencing of DNA from affected individuals to identify the disease- causing mutations (Papers I and III).
Polymerase chain reaction
In order to analyze a specific DNA sequence it needs to be selected from the complete genome and then amplified to reach a sufficient amount for analy- sis. In the early 1980’s a revolutionizing technique, polymerase chain reac- tion (PCR), was developed to fulfill these requirements in a time-efficient manner.
96In this method, a DNA segment of interest is amplified using DNA polymerase, free nucleotides and oligonucleotides (primers), which match flanking sequences of the segment. Using different temperatures the DNA strands are separated, allowing the oligonucleotides to hybridize to both single-strands and the polymerase to “build” new complementary DNA strands. This process is repeated, often up to 35 times, which gives an expo- nential increase in the selected DNA segment. PCR is currently one of the most widely used molecular techniques and has inspired to several new technologies.
DNA sequencing
Once the DNA region of interest has been selected and amplified by PCR the
exact DNA sequence can be analyzed. This has traditionally been performed
using Sanger sequencing, which is similar to the PCR. Complementary
strands are built up on single-stranded DNA by DNA polymerase, using only
one oligonucleotide. The free nucleotides used for the new strand are a mix- ture of deoxynucleotides and dideoxynucleotides that are randomly added; as the latter are incorporated into the new strand no more nucleotides can be added and the strand synthesis is terminated. This process gives rise to DNA fragments of all possible sizes and these are separated by size using gel or capillary electrophoresis. Since the dideoxynucleotides are labeled with dif- ferent fluorophores, a laser can excite a specific light for each fluorophore and “read” the DNA sequence.
97Although not used in the projects contributing to this thesis the new me- thods of massively parallel sequencing are worthy of mention. Whilst se- quencing a large amount of genetic sequence using Sanger sequencing is both time and DNA consuming, new sequencing technologies can sequence a whole genome in a matter of days. The most widely used platforms are Roche 454 genome sequencer, Illumina genome analyzer and Life Tech- onologies SOLiD system and they are all based on the principle of fragmen- tation of the DNA to be sequenced and ligation of known adapter sequences to the DNA fragments. The technologies differ in the methods of DNA am- plification, using either emulsion or solid phase technologies, and produce DNA reads of different length and error rate.
98It is becoming increasingly common to sequence the whole genome or exome (all protein coding exons) of individuals affected by monogenic disorders, in order to identify muta- tions. There are also great expectations that these new technologies will re- veal rare variations associated with complex disorders.
Gene expression analyses
Quantitative real-time reverse transcriptase PCR
Gene transcripts can be analyzed with respect to levels and sequence. Since
RNA is a nucleic acid easily degraded by RNase enzymes, it is normally
transcribed into its complementary DNA (cDNA) using the enzyme reverse
transcriptase. The cDNA can then be amplified by PCR and analyzed to
identify sequence aberrations or alternative splicing. To study levels of a
specific RNA quantitative real-time reverse transcriptase PCR (qPCR) is
used. qPCR functions as a normal PCR but with the addition of a fluorescent
dye that binds to double-stranded DNA (PCR-products). After each PCR
cycle the levels of fluorescence is measured by a detector, hence quantifying
the increase in the selected cDNA segment. The gene expression levels are
usually related to the levels of a housekeeping gene (a constitutive gene ex-
pressed at a relatively constant rate) in order to exclude differences induced
by methodological handling when two samples are compared. qPCR was
used in all papers of this thesis.
5’ rapid amplification of cDNA ends
Many genes have several alternative variants of the 5’ UTR, most of them not yet annotated in any databases.
33To analyze the 5’ UTR variants of an mRNA 5’ rapid amplification of cDNA ends (RACE) can be used. The prin- ciple of this method is to first remove all RNA that is not mRNA and then to ligate a known oligonucleotide to the 5’ end of all mRNAs. After reverse transcription the cDNA of interest can be amplified by PCR using one pri- mer that binds to the ligated oligonucleotide and one primer further 3’ in the transcript. The cDNA is sequenced from the 3’ end, resulting in different fragments depending on the different variants of 5’ UTR. This approach was used in Paper III.
Gene silencing by RNA interference
The cell’s own system of miRNA-mediated transcript degradation can be mimicked in vitro by introducing synthetic small interfering RNA (siRNA) into cultured cells (Figure 4). siRNA duplexes are designed towards a target sequence of a transcript of interest and are injected into cells using, for ex- ample, lipofection (membrane vesicles that fuse with the cell membrane).
The siRNA is then, like miRNA, converted into single-stranded RNA, incor- portated into an RNA-induced silencing complex and by binding comple- mentary to a target sequence induces its degradation by the catalytic part of the complex. This technique was used in Paper IV.
Figure 4. siRNA-induced degradation of target mRNA. siRNA duplexes are trans- ported into cells by liposomes, converted into single-stranded siRNA and incorpo- rated into the RNA-induced silencing complex (RISC). The siRNA binds comple- mentary to the target mRNA, leading to mRNA degradation by RISC.
Liposome
Cel l m
em bra ne
RNA duplex RISC
mRNA
Fragmented mRNA
Protein expression analyses
Western blot analysis
A method to detect a certain protein or to measure protein levels in a sample is Western blot (used in Paper II-IV of this thesis). Protein lysate is purified from the sample and separated in a polyacrylamide gel by electrophoresis according to its molecular weight. The proteins are then transferred by an electrical current to a stable membrane where specific proteins can be asso- ciated with primary antibodies. The primary antibodies can be visualized on the membrane by different methods. Using secondary antibodies labeled with fluorescent compounds the protein of interest can be detected and quan- tified using the Li-Cor® Odyssey imaging system. As for qPCR, protein levels are quantified by reference to a housekeeping gene product.
Light microscopy
Light microscopy can be used to visualize the expression and distribution of a protein in tissue. A biopsy is taken from a tissue of interest and is fixed either by immediate freezing and cold acetone or using formaldehyde, in order to cross-link proteins and preserve them from degradation. After em- bedding of the biopsy material it is cut into thin (usually ~5 µm) sections and mounted onto glass slides. Proteins are detected by interaction with primary antibodies and biotinylated secondary antibodies. An avidin-biotin enzyme complex that binds to the biotin on the secondary antibodies is added and reacted with the substrate DAB, which gives a brown color when oxidized.
The consequent immunostaining is analyzed using a light microscope in which visible light is transmitted through or reflected by the sample and passes through appropriate lenses to produce a magnified image visible by eye or that can be recorded on a computer. Light microscopy was used in Papers II, III and IV of this thesis.
Confocal microscopy
Different types of samples, such as cultured cells or biopsy sections (pre-
pared as above), can be analyzed using confocal microscopy (used in Papers
II-III). Common applications are analyses of protein expression and distribu-
tion in cells and tissue, as well as the analysis of co-localization between
different proteins and/or cell organelles. Unlike a light microscope, that il-
luminates the whole sample, a confocal microscope uses point illumination
that scans over the sample and an emission filter pinhole to eliminate light
that is not in focus. This gives the microscope a higher resolution and makes
it possible to capture 3D images of the specimen. For this technique second-
ary antibodies conjugated with fluorescent compounds are used. The fluo-
rescent compound emits light of a specific wavelength upon excitation by
light from a laser source. It is possible to detect several proteins at the same
time, using secondary antibodies labeled with compounds that emit light of different wavelengths.
Electron microscopy
An electron microscope uses an electron beam, instead of light, to create an
image. The electron beam passes through the specimen and components
within the sample either allow electrons to pass through unhindered or to be
deflected. Detection of the deflected and non-deflected electrons makes up
the captured image. The shorter wavelength of electrons compared with visi-
ble light gives the microscope a much higher magnification capability, mak-
ing it useful for detailed studies of cell and tissue morphology. The sample
to be analyzed is fixed, e.g., in formaldehyde or glutaraldehyde, embedded
in polymer and cut into ultrathin sections. If a specific protein is to be ana-
lyzed the sections are stained with primary antibodies and secondary antibo-
dies couple with heavy metals, such as lead and gold, thereby enhancing the
contrast between scattered electrons. Electron microscopy was used in Pa-
pers I and II.
Present investigations
Aims
The aims of the present thesis were:
to identify mutant genes in a sub-group of patients with autosomal reces- sive congenital ichthyosis (ARCI) and in those with KLICK syndrome;
to investigate the physiological characteristics of the identified genes and gene products in skin tissue and cells; and
to characterize the pathological effects of the mutations at the molecular
level
Autosomal recessive congenital ichthyosis (Paper I-II)
Subjects and samples
Twenty-seven Scandinavian patients (from 18 families) with ARCI were selected for genetic analysis based on the specific ultrastructural abnormali- ties of their epidermis, i.e., regular elongated perinuclear membranes, vesi- cular complexes and vacuoles (with or without vesicle content) in the granu- lar and/or horny layer (Figure 5). The clinical manifestations of the patients varied from moderate to severe ichthyosis often with a reticulate pattern, mild to moderate erythema, keratoderma of the soles and some patients showed ectropion. Blood samples for isolation of DNA were collected from all patients, while skin punch biopsies for keratinocyte culture, immunohis- tochemical/fluorescent analysis and ultrastructural analysis were collected from a subset of patients.
Additionally, 18 ARCI patients (from 16 families) with unknown genetic cause, varying clinical manifestations and ultrastructural aberrations not consistent with those mentioned above were included in the genetic analysis as a control group.
Figure 5. EM image of the interface between the granular and the horny layers of the epidermis from a patient with ARCI. Characteristic aberrations are vacuoles, both empty (black dashed arrow) and containing vesicles that are believed to be abnormal lamellar bodies (black arrow). Normal lamellar bodies (white arrow) are also seen.
Results and discussion
To identify the mutant gene(s) in the 27 selected patients with ARCI approx- imately 10,000 SNPs were genotyped using SNP array analysis in a subset of the patients. An 873-kb homozygous region was identified on chromosome 5q33 using homozygosity mapping in six individuals. The region includes the Ichthyin gene, previously reported as being mutated in a clinically hete- rogeneous group of ARCI patients.
69The gene was sequenced in all patients
Horny layer
Granular
layer
and 23 out of the 27 individuals were shown to harbor two mutant Ichthyin alleles and another two individuals to harbor one mutant allele. A total of two different missense mutations and two splice site mutations were identi- fied (Table 1). One of the patients from the control ARCI group of 18 pa- tients was shown to harbor homozygous Ichthyin mutations. These findings demonstrate that mutant Ichthyin is strongly, although not completely, asso- ciated with specific structural abnormalities in the epidermis of patients with ARCI. Since it has been difficult to associate any particular ARCI symptom with mutations in a specific gene,
65; 99the present findings represent an im- portant step towards a better understanding of correlations between pheno- type and genotype in this group of disorders. The finding of Ichthyin muta- tions in 93% of patients featuring a specific ultrastructural phenotype sug- gests that EM, in combination with genetic analysis, provides a tool for the accurate diagnosis of ARCI patients.
Table 1. Distribution of Ichthyin mutations in 27 ARCI patients with specific ultra- structural aberrations in epidermis. Four different mutations were identified in 25 of the patients.
Ichthyin mutations
Allele 1 Allele 2 Number of
families Number of patients
c. 527C>A c. 527C>A 9 15
c. 688G>A c. 688G>A 1 1
c. 527C>A c. 688G>A 3 5
c. 688G>A Unknown 1 2
c. 527C>A c.464-1G>A 1 1
c. 527C>A c.772+1G>A 1 1
Unknown Unknown 2 2
Total 18 27
Little is known about the Ichthyin gene and its protein product. To investi-
gate the physiological characteristics of ichthyin as well as the pathological
effects of the mutations, gene expression analyses were performed using
keratinocytes cultured ex vivo. Initially, the Ichthyin mRNA of a patient with
compound heterozygosity for an Ichthyin missense (c.527C>A) and intron
five splice site mutation (c.772+1G>A) was studied. The splice site mutation
was shown by reverse transcriptase PCR to cause an inclusion of 68 intronic
nucleotides into exon 5, resulting in a premature stop codon at amino acid
position 273. Next, Ichthyin mRNA levels in cultured non-differentiated
keratinocytes from five patients homozygous for the c.527C>A mutation and
from the patient with compound heterozygosity were analyzed using qPCR,
but neither of the mutations resulted in altered transcript levels compared
with healthy controls (Paper I). Finally, qPCR and Western blot analysis
were performed in keratinocytes subject to Ca
2+-mediated differentiation
from control subjects and the patient with compound heterozygosity. How-
ever, these analyses did not reveal any significant differences in Ichthyin transcript or protein levels between patient and controls, although there was a tendency for lower protein levels in differentiated cells from the patient (Figure 6; Paper II). Ichthyin transcript levels, but not protein levels, in- creased in differentiated cells compared with non-differentiated cells in both patient and controls (transcript levels: controls p=0.0017, patient p=0.054).
The finding that Ichthyin transcript levels increased after differentiation pro- vides evidence for ichthyin being an important factor for formation of the horny layer in the skin as previous studies have shown that genes important for terminal epidermal differentiation are up-regulated during keratinocyte differentiation.
59Moreover, Ichthyin transcript levels of patients were equal to those of controls, suggesting that the mutations disrupt the function or localization of the protein rather than transcript stability.
Figure 6. Ichthyin mRNA (A) and protein (B) expression in keratinocytes. Cells from three healthy controls and one ichthyosis patient were cultured according to standard protocols (Non-diff.) or with supplement of 2 mM CaCl
2for 72h for diffe- rentiation (Diff.). mRNA and protein levels were analyzed by qPCR and Western blot, respectively. Results are presented as mean of three separate experiments and error bars signify standard deviation (SD) of the experiments. ** = p<0.01.
To study the subcellular localization of ichthyin in epidermal tissue, skin sections from patients homozygous for c.527C>A and c.688G>A missense mutations, respectively, and controls were stained for immunofluorescent detection of ichthyin (Paper II). Confocal imaging revealed that ichthyin is expressed strongly at cell borders and diffusely in cytoplasms of keratino- cytes in all epidermal layers in both patient and control skin. To further de- fine the subcellular localization of ichthyin, skin sections of a control and patient with homozygous Ichthyin missense mutations (c.527C>A) were analyzed using EM and immunodetection of ichthyin. The ichthyin immuno- labeling was clearly restricted to keratinocyte keratins and desmosomes in the basal, spinous and granular layers and to hemidesmosomes in the basal
0 0,005
0,01 0,015 0,02 0,025
Non-diff. Diff.
Transcript levels relative to beta actin
Controls Patient
**
0 0,02 0,06 0,1 0,14 0,18
Non-diff. Diff.
Protein levels relative to GAPDH
mRNA Protein
A B
layer of both patient and control epidermis. The immunolabeling was most prominent in the basal layer, and more abundant in patients, which also showed more aggregated ichthyin labeling (Figure 7). These findings indi- cate that ichthyin executes its function in strong interaction with the keratin filament network and desmosomes. Furthermore, it appears as if mutant ich- thyin is misfolded and aggregates in patients. This may be an important me- chanism in the pathogenesis of ARCI due to Ichthyin mutations, as protein aggregation deteriorates the protein function and is known to disturb normal cellular processes.
53; 100Figure 7. EM analysis of subcellular localization of ichthyin in a patient with Ich- thyin mutations. Ichthyin expression (arrows) is restricted to keratin bundles (dark areas) and forms aggregates in the patient. The keratin bundles are located in sepa- rate keratinocytes on either side of the inter-cellular space (IC).
Several ARCI associated genes are implicated in epidermal lipid metabol- ism.
77; 78; 80; 83In order to investigate the effects of mutations in Ichthyin on lipid distribution in epidermis, skin sections of controls and a patient with homozygous Ichthyin mutations (c.527C>A) were analyzed using Nile red neutral lipid staining. The patient epidermis showed a normal staining of keratinocyte cell membranes and lipid layers in the horny layer, but, addi- tionally, an abnormal staining of irregular structures within the cells of the granular and horny layers not seen in control epidermis. These findings sug- gest that ichthyin is involved in epidermal lipid metabolism and that ichthyin deficiency results in an incomplete transport or synthesis of lipids.
The exact function of ichthyin remains obscure. The vesicular complexes and vacuoles with vesicles, visualized by EM in patients with Ichthyin muta- tions, are interpreted as abnormal lamellar bodies and their content. This suggests that ichthyin is involved in the transport or function of lamellar bodies and, hence, affects the transport of lipids and enzymes to the inter- cellular space of the horny layer. The restricted localization of ichthyin to
2 uM
IC
keratins and desmosomes strongly indicates that the function of ichthyin is dependent on these structures. Mutations in keratin 1 and keratin 10 have been shown to cause not only keratinocyte fragility, resulting in skin blisters, but also a defect in the epidermal water barrier leading to ichthyosis.
101Pa- tients with keratin 1 and keratin 10 mutations have a perturbed secretion of lamellar bodies from cells of the granular epidermal layer to the inter- cellular space, resulting in reduced lipid lamellae in the horny layer.
102This finding, in combination with the results of the present study indicate that the keratin filament network, the desmosomes and ichthyin together are essential for adequate processing and secretion of lamellar bodies.
From the combined results of the present studies it is concluded that mu-
tations in Ichthyin induce ichthyosis and result in abnormal lamellar bodies,
lipid vacuoles and membranous structures in epidermis. Ichthyin is physio-
logically localized to keratins and desmosomes in epidermis and mutations
in Ichthyin do not affect transcript levels, but possibly cause protein aggrega-
tion and thereby deficient protein function.
KLICK syndrome (Paper III-IV)
Subjects and samples
Samples of DNA were collected from 12 individuals, from eight non-related European families, affected with KLICK syndrome. Skin punch biopsies for keratinocyte cell culture and immunohistochemical analyses were obtained from two and three patients, respectively, and from healthy controls. The patients all showed the KLICK associated skin manifestations of ichthyosis, thickened skin on palms and soles, hyperkeratotic plaques on knees, wrists and in axillae, circular sclerotic constrictions around fingers, flexural defor- mities of fingers, and linear hyperkeratotic papules on flexural surfaces of wrists, elbows, and knees.
Results and discussion
To identify a candidate gene associated with KLICK syndrome, the DNA of six affected individuals was analyzed using 10k SNP array and homozygosi- ty mapping (Paper III). A common homozygous region of 1.5Mb, including ten protein coding genes, was identified on chromosome 13q. DNA sequenc- ing of all the gene exons revealed a homozygous single-nucleotide deletion located in the 5’ region of the POMP gene (c.-95delC) in all 12 patients, but was not found in 280 control chromosomes. To elucidate whether the c.- 95delC locus is transcribed and is part of the POMP mRNA, 5’ RACE was performed on patient keratinocyte mRNA. The analysis revealed POMP transcripts with several different lengths of the 5’ UTR where a majority had a 5’ UTR shorter than 95 nucleotides, but a proportion of the transcripts had a longer 5’ UTR, thereby including the deletion site.
To investigate whether the deletion affects transcript levels of POMP, qPCR was run on mRNA from non-differentiated keratinocytes and kerati- nocytes differentiated in vitro, from patients and controls. Two distinct fragments of POMP were amplified; one representing both long and short 5’
UTR transcripts (forward primer at c.-63) and one specific for long 5’ UTR
transcripts (forward primer at c.-116). The total amount of POMP mRNA,
represented by the first amplified fragment, did not differ between patients
and controls. However, the proportion of transcripts with a long 5’ UTR
differed significantly between patients and controls. Long 5’ UTR transcripts
constituted 18% of all POMP transcripts in non-differentiated cells, increas-
ing to 83% in differentiated cells (p=0.00369; Student’s t-test) in patients,
whereas the proportions in controls were 1.3% and 2.6%, respectively
(p=0.116; Figure 8). In patients, the amount of long 5’ UTR transcripts was
increased almost 18-fold in non-differentiated cells (p=0.00017) and 29-fold
in differentiated cells (p=7.82x10
-6), compared with controls.
Figure 8. qPCR analysis of POMP transcript levels of non-differentiated keratino- cytes (A) and keratinocytes differentiated in vitro (B), from patients and controls.
Two different fragments of POMP were amplified and results are presented as the proportion of all transcripts that have long and short 5’ UTR, respectively.
Long 5’ UTR’s tend to reduce translation rate compared with short 5’
UTR’s.
30Western blot analysis of POMP protein levels did not, however, reveal any differences between patient and control non-differentiated cells and cells differentiated in vitro (Paper III). It was hypothesized that a transla- tional effect caused by the switch in POMP transcripts would not take place until later in the differentiation process of keratinocytes, a stage that was not reached in cell cultures. In order to find further support for a reduced transla- tion rate of long POMP 5’ UTR’s three different full length POMP cDNA constructs in GFP-vectors were transfected into HeLa and HaCaT cell lines (Paper III & IV). The constructs differed only in the 5’ UTR: two of the con- structs had an 81-nt-5’ UTR and a 181-nt-5’ UTR, respectively, with wild type sequence, and one construct had a 181-nt-5’ UTR including the KLICK associated deletion (c.-95delC). POMP-GFP fusion protein levels were measured for each construct using Western blot. A reduction in protein le- vels of 52% (HeLa) and 40% (HaCaT) was observed for the wild type 181- nt-5’ UTR construct compared with the 81-nt-5’ UTR construct (HeLa p=0.0011, HaCaT p=0.0137; One-Sample t-test; Figure 9) (Paper III and unpublished data). The protein levels of the mutant 181-nt-5’ UTR construct showed a strong tendency for a reduction by 23% in HeLa cells and by 31%
in HaCaT cells (HeLa p=0.0801, HaCaT p=0.0894; Paper IV), compared with the 81-nt-5 ‘UTR construct. These results support the hypothesis of reduced POMP protein levels in late differentiated patient keratinocytes due to an increased proportion of long 5’ UTR transcripts.
0%
20%
40%
60%
80%
100%
Paents Control subjects
0%
20%
40%
60%
80%
100%
Paents Control subjects
A B
Proportion of all POMP transcripts Proportion of all POMP transcripts
Short 5’ UTR transcripts Long 5’ UTR transcripts
Nondifferentiated cells Differentiated cells