A Phenotype Map of the Mouse X Chromosome: Models for Human X-linked Disease
Abstract
The identification of many of the transcribed genes in man and mouse is being achieved by large scale sequencing of expressed sequence tags (ESTs). Attention is now being turned to elucidating gene function and many laboratories are looking to the mouse as a model system for this phase of the genome project. Mouse mutants have long been used as a means of investigating gene function and disease pathogenesis, and recently, several large mutagenesis programs have been initiated to fulfill the burgeoning demand of functional genomics research. Nevertheless, there is a substantial existing mouse mutant resource that can be used immediately. This review summarizes the available information about the loci encoding X-linked phenotypic mutants and variants, including 40 classical mutants and 40 that have arisen from gene targeting.
Mammalian X-linked traits are easily recognized by their inheritance patterns and their mode of expression. Whereas hemizygous males carry one copy of X-linked loci and suffer from the full effect of any mutation, heterozygous females carry two copies and have a phenotype that reflects the relative expression, as determined by X-inactivation status, of the mutated and normal copies of the gene (Lyon 1999). The X chromosome is also unusual in that X linkage of genes is almost totally conserved in eutherian mammals (Ohno 1973). Therefore, disorders that are X linked in man are also X linked in the mouse, which leads to the ready identification of mouse models of human X-linked disease. The existing mouse mutant resource, which comprises well over 1000 different stocks and strains, has been exploited to investigate gene function and disease pathogenesis associated with X-linked and autosomal loci (Paigen 1995; Bedell et al. 1997)5. The past decade has seen a revolution in the ability to deliberately introduce mutations into mouse genes by homologous recombination (Fisher 1997; Müller 1999; Roths et al. 1999) and, as a result, the number of mouse X-linked traits has doubled. Although there are earlier reviews (Davisson 1987; Miller 1990), no comprehensive summary of existing mouse X-linked phenotypes has been published recently. The primary aim of this review is to describe the current status of the phenotype map of the mouse X chromosome for those working in the field of genome research with an interest in X-linked disease.
Analyzing mouse mutants at the molecular and phenotypic level is one of the most powerful ways of understanding gene function in mammals. Apart from a few notable exceptions, in which mutations exist in mouse genes whose homologs are known to be responsible for human disease, there are considerable phenotypic similarities between the mouse and human disorders, and the mutant mouse provides an animal model for understanding disease pathogenesis and for assessing therapeutic regimes. Therefore, it is likely that most of the remaining mouse phenotypes will provide a valuable resource for identifying the molecular basis of a homologous human disease.
Comparative Map of the Human and Mouse X Chromosomes
The initial stage of assessing and identifying mouse models for human disease from the existing mouse mutant resources involves careful characterization of the phenotype associated with the mutant locus and predicting the position of its human homolog on the man–mouse comparative map.
The mapping of >130 conserved loci on the X chromosomes of both mouse and man (Boyd et al. 1998, 1999) has confirmed the prediction that X linkage of genes is preserved in mammals (Ohno 1973). However, when the relative positions of loci on the human and mouse X chromosomes are compared, it can be seen that subchromosomal blocks of homologous loci have been rearranged with respect to each other during the 80 million years of evolutionary time that separate the two species. It is important to understand these rearrangements fully, because an identical comparative map position is an important criterion for identifying and confirming mouse models for human genetic disease. More than a decade ago, five distinct homologous blocks of loci or conserved segments were acknowledged as sharing homology on both the human and mouse X chromosomes (Searle et al. 1987, 1989; Amar et al. 1988; see Fig.1 and Table 1 for the current status of the comparative map). The human X chromosome long arm (Xq) was recognized as being split into two major blocks on the mouse X chromosome. There are only two minor modifications to this Xq comparative map, the identification of a 600-kb inversion around Xist (Rougelle and Avner 1996) and the mapping of the synaptobrevin like locus (Sybl1) to the proximal region of the mouse X chromosome (D'Esposito et al. 1997). In contrast, it soon became apparent that the three conserved segments thought to comprise the human X chromosome short arm (Xp) could be apportioned into five major and four minor conserved segments on the mouse X chromosome (Laval and Boyd 1993; Blair et al. 1994, 1995,1998b; Blaschke and Rappold 1997). Twelve conserved segments have now been identified on the man–mouse X chromosome comparative map (Fig.1), and additional regions of homology may be defined as conserved genes are mapped to a higher resolution (Ehrmann et al. 1998). Nevertheless, for most of the X chromosome, once the position of a locus is known on the human X chromosome, its position on the mouse X chromosome can be predicted with reasonable accuracy and vice versa. The evolutionary breakpoint regions remain the only areas of uncertainty and these, because they have been subjected to multiple rearrangements during evolution, can be expected to have a complex structure and this has been borne out by recent mapping data (Dinulos et al. 1996; Blair et al. 1998b; Disteche et al. 1998). Over 130 genes and conserved loci have been regionally mapped on both the human and mouse X chromosomes and form the framework for constructing the comparative map (Table 1). This map has been invaluable in identifying human diseases and candidate gene loci for many of the classical mutants that have been recovered from mouse colonies over the years.
List of X-Linked Genes and Conserved Sequences
Spontaneous and Induced X-Linked Mouse Mutants and Variant Traits
Thirty-eight X-linked mouse-independent visible phenotypes covering a wide range of traits are reported in the literature (Table 2; Fig.2). Most of these phenotypes have arisen spontaneously in mouse colonies or among the large numbers of mice used in mutagenesis experiments (George et al. 1994). The paucity of induced X-linked mutations is probably due to that fact that most novel mutations have been sought in the F1 progeny of mutagenized males and as a result, only those X-linked mutations that have a phenotypic effect in heterozygous females are identified. Some attempts were made to obtain sex-linked recessive lethals by identifying abnormal sex ratios in the offspring of F1 females produced in classical mutagenesis experiments (Searle et al. 1964). Recessive mutations, such as sex-linked fidget, arose in experiments like these (Lyon et al. 1981b; Phippard et al. 2000).
Spontaneous and Induced X-Linked Mouse Traits
The comparative phenotype map of the human and mouse X chromosomes. Each conserved segment, in which the order of loci is the same in mouse and man, is indicated by a colored rectangle. The loci that define each segment are given in Table 1; note that the order of segments1, 2, and 3 has not been established and is arbitrary. The segments are numbered from 1–12 from the centromere to the telomere on the mouse X chromosome and the order of loci indicated by an arrow alongside each block. The hatched region within segment 9 represents the 600-kb region that is inverted around the Xist locus (see text). The centromeric regions are indicated by black and white hatched rectangles; there is, as yet, no known evidence for evolutionary conservation of centromeric sequences between mouse and man. The Xp pseudoautosomal region, which has a complex evolutionary history (Blaschke and Rappold 1997), is not included because it is outside the scope of this review, as loci in this region do not exhibit X-linked inheritance. Classical mutants carrrying spontaneous and induced mutations and variants (Table 2) are indicated on the mouse X chromosome, and targeted mutations (Table 3) are positioned on the human X chromosome for clarity. Names in black text are (1) known to have an X-linked inheritance pattern but have not been positioned on the X (e.g., Ie, It), or (2) have not been subjected to high-resolution mapping (e.g., Bw1, the black line indicates the probable region in which the locus lies), or (3) have been mapped to evolutionary breakpoint positions and therefore the position of the human homolog cannot be defined (e.g.,exma, Bhd, wf). For those traits marked with an asterisk (*), the gene responsible is not known; for all other traits the underlying lesion has been defined.
a(Ir) Immune response genes that are X linked (see Table 2 footnote).
Examples of spontaneous and induced X-linked traits in the mouse. (A,B) Males carrying the greasy (Gs) (A) and harlequin (Hq) (B) mutations; neither of the genes responsible has been cloned. (C) A heterozygous female carrying the broad-headed (Bhd) mutation, which is associated with a craniofacial anomaly, note the unusually short and broad snout. Comparative mapping has shown that Bhdcannot be a model for FGD1 or ATRX, but it remains an interesting skeletal mutant in which males have multiple ossification anomalies and die shortly after birth (V. Reed and Y. Boyd, in prep.). (D) A heterozygous female carrying a mutation at one of the mottled (Mo) alleles associated with death between birth and weaning of affected males (Atp7aMo-10H ). The mosaic coat can be clearly seen, in which hypopigmented areas represent populations of cells carrying the mutant allele and the normally pigmented (brown) areas represent areas of cells carrying the wild-type allele. Details of the phenotypes and references describing each of these mutants are given in Table 2.
For most of the recovered mutations, only a single mutant allele is available for analysis, limiting the value of the mouse as a model for X-linked human disease, which is often associated with a high number of different new mutations (Rossiter and Caskey 1991). However, where several alleles exist, the mouse provides an ideal tool for studying the phenotypic effects of different mutations in the same gene on an identical genetic background. Over 20 alleles, associated with differing phenotypes have been recovered at the mouse mottled locus and these provide a useful paradigm for using the mouse to study human disease (Cunliffe 1999).
Mottled: A Paradigm for Mouse Models for Human X-Linked Disease
For many years, the mottled mouse has been recognized as a model for Menkes disease (MD), and in both species, a range of mutations has been found in the gene encoding the copper transporter ATP7A(Cecchi et al. 1997; Reed and Boyd 1997; Tümer et al. 1997). Affected mice and human patients show a similar and variable course of disease with the main features being growth retardation, neurological and connective tissue abnormalities, peculiar hair and hypopigmentation (Danks 1986; Tümer and Horn 1997; Fig. 2D). Mutations that have a mild phenotypic effect in the mouse, such as the splice-site lesion leading to the production of both normal and aberrant transcripts in mottled blotchy, seem to be of a similar type to those associated with occipital horn syndrome (OHS, a mild allele of MD). In both man and mouse, this type of mutation is associated mainly with connective tissue problems (Das et al. 1995). However, there are significant differences between man and mice in the type of molecular lesions associated with the more severe phenotypes. The classical MD phenotype is mainly associated with nonsense or frameshift mutations and there is also a substantial proportion (∼20%) of patients with gross deletions covering varying proportions of the ATP7A-coding region. In the mouse, the largest known deletion, which covers exons 11–14 (Cunliffe 1999), is in frame, and no nonsense or frameshift mutations have been reported to date. This fact demonstrates that caution must be taken when assessing potential therapies with mottled mice as animal models. The most striking phenotypic difference is that MD patients, with what might be expected to be null mutations at ATP7A, survive until birth, whereas over half of the mottled alleles are associated with prenatal lethality of males sometime after mid gestation. It has long been thought that the mottled mutations causing early postnatal death are the most appropriate model for classical MD in which affected boys die in the first few years of life. However, recent evidence has shown that in the mouse, Atp7a mutations most similar to those seen in classical MD cause prenatal death (Cunliffe 1999; P. Cunliffe, V. Reed, and Y. Boyd, in prep.). Therefore, at the level of cellular copper processing, mottled alleles associated with prenatal lethality may provide a better model. This is particularly important in light of the varied response of MD patients to copper histidine treatment (Tümer and Horn 1997).
Using the Mouse to Identify the Genes Responsible in X-Linked Dominants Associated with Intrauterine Male Lethality
A further interesting picture has emerged from studies on mouse mutants as possible models for X-linked dominant disease, which, because they are associated with prenatal lethality of males, are difficult to position accurately on the human X chromosome. Disorders associated with the lethality of males in utero are probably caused by lesions in genes that have important developmental roles and that may be conserved in man and mouse. In the mouse, the relative ease of high-resolution mapping renders the genes responsible amenable to positional cloning. Recently, an interesting association between sterol biosynthesis and skeletal defects has been revealed by the identification of the underlying molecular lesions in the Bpa,Str, and Td mutants, which are all associated with skeletal and skin anomalies in heterozygous females (Table 2). Mutations in the 3β-hydroxysteroid dehydrogenase Nsdhlwere first found in several independently derived Bpa andStr mutants, which were shown to be allelic, although there are differences in the severity of the phenotype (Liu et al. 1999). TheBpa mouse was suggested originally as a mouse homolog of CPDX2 as both display-striated hyperkeratosis and skeletal abnormalities including short stature, rhizomelic shortening of the limbs, epipyhseal stippling, and craniofacial anomalies (Happle 1983). However, a more detailed phenotypic analysis of Td has revealed that it also has many of these features and Td was discovered to be a model for CPDX2 when mutations in the sterol isomerase Ebp were discovered inTd and CPDX2 patients (Braverman et al. 1999; Derry et al. 1999). Thus, the Bpa/Str alleles and Td mice provide tools to investigate the relationship between sterol biosynthesis, intrauterine death of males, and the skin and skeletal defects seen in heterozygous females.
X-Linked Phenotypes Associated with Mutations Introduced by Gene Targeting
There are now as many mouse X-linked phenotypes introduced deliberately by gene targeting than have been recovered over the years in mouse colonies. Approximately one-half of the targeted genes are implicated in overt human disease and have been ablated to create models for understanding gene function and disease pathogenesis (Table3). Some of these genes have shown comparable phenotypes in the mouse, such as the targeted disruption of genes encoding factor VIII and IX, which have provided excellent mouse models for studies on haemophilia A and B (Bi et al. 1995; Wang et al. 1997). Others exhibit some, but not all, aspects of the corresponding human disease; for example, there is an impaired humoral immune response in CD40 ligand-deficient mice, but they do not develop full-blown hyperIgM syndrome (Renshaw et al. 1994). In the future, mouse models carrying defined molecular lesions identical to those found in human disease can be provided by gene-targeting technology. In conjunction with the production of a series of different mutations in the same target gene, as has been done at the autosomal Fgf8locus by Meyers et al. (1998), the availability of engineered mouse models relevant to human disease can only increase.
Targeted Mutations at Mouse X-Linked Loci
Other mutants have been created to investigate the potential functions and the phenotypic consequence of a deficiency in the targeted gene. Some of these studies have revealed interesting clinical effects in mice with potential applications for studying complex human disorders, for example the demonstration that a lack of biglycan leads to osteoporosis (Xu et al. 1998). Other gene disruptions have also proven to be highly informative in understanding gene function, for example, the targeting of the Xist locus demonstrated its vital role in the X-inactivation process (Penny et al. 1996; Marahens et al. 1997).
Until stable female embryonic stem cell lines are widely available and can be transmitted to the germ line with a high frequency, one problem that remains with the targeted disruption of X-linked genes is the study of those genes that are potentially lethal in males. Genes may be dispensable in stem cells but essential for embryonic development, as is the case with the methyl-CpG-binding protein Mecp2 (Tate et al. 1996). The Mecp2 knockout mice are particularly interesting in light of the recent revelation that the X-linked dominant neurological disorder Rett syndrome is associated with MECP2 mutations (Amir et al. 1999). The problem of early lethality preventing the recovery of mice carrying targeted mutations of X-linked genes can be partially circumvented by the production of conditional knockouts, in which ablation or modification of the gene of interest is engineered on a temporal or tissue-specific basis (for review, see Müller 1999;Roths et al. 1999). Tarutani et al. (1997) used this approach to demonstrate that the X-linked Piga gene plays an important role in skin development.
X Inactivation and Analysis of the Phenotypic Variation Associated with Heterozygous Females
Mice carrying X-linked mutations can also be used as tools to investigate the influence of cellular mosaicism associated with X-inactivation patterns in different tissues on the phenotype. X inactivation occurs early in mammalian female development and results in the transcriptional silencing of one of the two X chromosomes, at random, in every somatic cell (Lyon 1999). As a consequence, females heterozygous for X-linked coat mutations such as mottled have a variegated coat pattern resulting from the two cell populations, one expressing the mutant allele and the other expressing the wild-type allele (Fig. 2D). Because the number of cells at the time of X inactivation is small, and the choice of which X chromosome is inactivated is usually random, considerable differences are seen between females in the relative sizes of the two cell populations; this manifests as a variation in the severity of the phenotype between female mice carrying the same mutation. For some X-linked mutations, cellular mosaicism is a benefit with the population of cells expressing the normal allele, providing enough of the normal gene product to permit a normal phenotype to develop, for example, in the sfor spf mutations (DeMars et al. 1976; Lyon et al. 1990). Another possibility is that growth competition between the two cell populations may result in the virtual elimination of the cells expressing the mutant allele (Ogura et al. 1998). In this situation, the resultant skewed X-inactivation pattern provides heterozygous females with a normal, or mild, phenotype.
Traditionally, X-inactivation patterns in mice have been investigated in the coat by the observation of doubly heterozygous female mice produced when the mutation is bred to well-characterized coat mutants, such as the mottled blotchy or tabby mutants. In humans, approaches that exploit the differences in DNA methylation between the inactive and active X chromosomes are most widely used to study X-inactivation patterns in heterozygous females (Belmont 1996). In the mouse, extensive studies of the X-inactivation patterns can be achieved because of the availability of tissues from several replica heterozygotes at a range of times in development on a constant genetic background. Techniques such as the single nucleotide primer extension (SNuPE) assay have been used to quantitatively measure the relative expression from the two X chromosomes in females heterozygous for X-linked mutations (Greenwood et al. 1997; Ogura et al. 1998). Further, the X-linked LacZ transgenic mice created by Tam and Tan (1992) have great potential for the study of X-inactivation patterns in heterozygous females as the transgene is subject to the inactivation process. When the lacZ reporter is present on only one of the X chromosomes of a female heterozygous for an X-linked mutation, the β-galactosidase activity is limited to cells with only that X chromosome active. Therefore, the distribution of cells expressing the mutant allele in a heterozygous female can be studied and insights into the mode of action of the normal X-linked gene product can be provided by the analysis of the phenotype and the X-inactivation pattern of heterozygous females. In our laboratory, such studies are under way on the Li, Stpy, and Td mutants.
Future Progress
The 80 X-linked phenotypes that have been reported in the laboratory mouse correspond to mutations in approximately half of the cloned X chromosomal genes. Most of these traits are associated with mutations in known genes and are of immediate value as animal models for human X-linked diseases. However, they represent <5% of the probable number of X-linked genes and the phenotype map of the mouse X chromosome is still sparse. Alternative methods will be needed to provide the detailed phenotype maps worthy of complementing the encyclopedias of mouse genes that are being developed (Marra et al. 1999). Although several laboratories are attempting to increase the size of the mouse mutant resource (Justice et al. 1997; You et al. 1997; deAngelis and Balling 1998; Schimenti and Bucan 1998; Zheng et al. 1999), no plans appear to have been made to screen systematically the progeny of F1 females carrying mutagenized X chromosomes for X-linked traits. Without such a specific effort, it will be interesting to see whether the number of X-linked mutant stocks will increase to meet the needs and interests of groups trying to understand the roles of X-linked genes in health and disease.
Acknowledgments
We thank Kevin Glover for help in producing Figure 2. This work was supported by the MRC of Great Britain.
Footnotes
-
Present addresses: 1Institute for Animal Health, Compton, Newbury RG20 7NN, UK; 2Department of Physiological Sciences, Medical School, Framlington Place, University of Newcastle, Newcastle Upon Tyne NE2 4HH, UK; 3Department of Medicine, Manchester Royal Infirmary, Oxford Road, Manchester M13 9WL, UK.
-
↵4 Corresponding author.
-
E-MAIL yvonne.boyd{at}bbsrc.ac.uk; FAX 00 44 1635 577237.
-
↵5 A comprehensive list of international mouse strain resources, developed jointly by the UK MRC laboratory at Harwell and the USA Jackson Laboratory (Eppig and Strivens 1999), and a list of strains available at the newly established European Mouse Mutant Archive (EMMA) in Italy, can be accessed via the World Wide Web athttp://ismr.har.mrc.ac.uk; http://www.emma.cnr.it.
REFERENCES
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
