» » »

hmg

Human Molecular Genetics 2

2nd

Tom Strachan1 and Andrew P Read2

1University of Newcastle, Newcastle-upon-Tyne, UK

2University of Manchester, Manchester, UK

BIOS Scientific Publishers Ltd1-85996-202-51999

genetics

Chapter 21: Genetic manipulation of animals

A2702

21.1. An overview of genetic manipulation of animals

Experimental animals have been used in biomedical research for decades. In many cases, aspects of physiology and biochemistry have been investigated, and artificial manipulations have often been confined to examining the effect of altering the animal's environment or some aspect of its phenotype. Some animals, notably Drosophila and mice, have been particularly amenable to genetic analyses and traditional genetic manipulation of animals has involved carefully selected breeding experiments or exposure of animals to powerful chemical or radioisotopic mutagens. A new era in animal research was ushered in during the early 1980s when successful experiments designed to genetically modify animals by inserting foreign DNA were first reported. These new methodologies were expected to have many advantages for research but two major areas have benefited:

Gene function. While the use of cultured cells and cell extracts can be extremely valuable in studying gene expression and function, the ability to insert genes into whole animals or to selectively delete or alter single predetermined genes in an animal provides enormous power in studying gene function.
Animal models of disease. Nature has provided some animal models of disease and some have been generated by random mutagenesis programmes, but not in a predetermined way. The new technologies held the promise of altering at will even single genes within a living animal in such a way as to mimic mutations faithfully in an analogous gene in humans, thereby providing a higher chance of resembling human disease phenotypes.

In order to create genetically modified animals, it is necessary to modify the DNA of germline cells so that the modified DNA is heritable. As a result, certain cells that have the capacity to differentiate into the different cells of an adult animal (or at least to give rise to germ line cells) were considered to be the optimal targets for introducing foreign DNA. The fertilized oocyte is one such cell, being totipotent. Other target cells are cells of very early stage embryos, including embryonic stem (ES) cells. Although such cells are postzygotic they represent a stage in development where there has been incomplete separation of the soma and the germline. Such cells are therefore capable of giving rise to both somatic and germline cells.

When a foreign DNA molecule is artifically introduced into the cells of an animal, a transgenic animal is produced. The foreign DNA molecule is called a transgene and may contain one or many genes. By inserting a transgene into a fertilized oocyte or cells from the early embryo, the resulting transgenic animal may be able to transmit the foreign DNA stably in its germline. Many different types of transgenic animals have been created including transgenic Drosophila, transgenic frogs, transgenic fish and a variety of transgenic mammals including mice, rats and various livestock animals. The technology of transgenesis and its applications are considered in Section 21.2.

Although transgenes often integrate into the host chromosomes without affecting the expression of any endogenous genes, occasionally the integration event alters endogenous gene expression (insertional mutation), producing a recognizable phenotype. This constitutes a form of in vivo mutagenesis, albeit at an unselected target gene. Gene targeting was developed as a method of in vivo mutagenesis in which the mutation is introduced into a preselected endogenous gene. This can be achieved in somatic cells, but gene targeting in cultured ES cells is particularly powerful because it can lead to the construction of an animal in which all nucleated cells contain a mutation at the desired locus (see Section 21.3 below). In mammals, gene targeting has been possible only in mice but research on ES cells from other species may extend the capacity for gene targeting in the near future. Note that in some cases gene targeting is used to produce a subtle mutation and as a result the ES cells used for blastocyst implantation do not contain any foreign sequences. The resulting mice may be described as genetically modified but not as transgenic.

The ability to produce transgenic mice and particularly the ability to perform specific changes in a predetermined gene by gene targeting has permitted the design of many new animal models of human disease (Section 21.4). Another experimental approach involving genetic manipulation of animals has had a major impact recently. In 1997 a new era in mammalian genetics was heralded when the procedure of somatic cell nuclear transfer permitted ‘cloning’ of an adult mammal for the first time. This involved transfer of the nucleus from an adult cell into an enucleated oocyte and the technology has subsequently been applied as an alternative route to generating transgenic animals (Section 21.5).

21.2. The creation and applications of transgenic animals

Of the different transgenic animals that have been made thus far, transgenic Drosophila, transgenic frogs, and transgenic fish have been very important for understanding aspects of gene function and development in these species. Transgenic sheep and other transgenic livestock animals have been produced largely to serve as bioreactors, whole-animal expression cloning systems in which introduced genes are expressed to give large amounts of therapeutic or commercially valuable gene products (see Section 22.1.2). But it has been transgenic mice that have been the most useful to biomedical research, both in providing animal models of disease and in permitting the most useful analyses of mammalian gene function.

21.2.1. Transgenic animals can be produced following transfer of cloned DNA into fertilized oocytes and cells from very early stage embryos

Transgenesis involves transfer of foreign DNA into totipotent or pluripotent embryo cells (either fertilized oocytes, cells of the very early embryo or cultured embryonic stem cells) followed by insertion of the transferred DNA into host chromosomes. If the foreign DNA integrates into the chromosomes of a fertilized oocyte, the developing animal will be fully transgenic since all nucleated cells in the animal should contain the transgene. If chromosomal integration occurs later, at a postzygotic stage, the animal will be a mosaic, with some cells containing the transgene and some others lacking it. If the transgene is present in germline cells it can be passed through sperm or egg cells into some of the animal's progeny, and PCR-based tests can be used to quickly screen for the presence of the transgene. Progeny that are transgene positive can be expected to be fully transgenic; all their cells should have developed from a fertilized oocyte containing the transgene.

Pronuclear microinjection

Figure 21.1
Construction of transgenic mice by pronuclear (more...)

An external file that holds a picture, illustration, etc., usually as some form of binary object. The name of referred object is ch21f1.jpg.

Figure 21.1

Construction of transgenic mice by pronuclear microinjection

Very fine glass pipettes are constructed using specialized equipment: one, a holding pipette, has a bore which can accommodate part of a fertilized oocyte, and thereby hold it in place, while the microinjection pipette has a very fine point which is used to pierce the oocyte and thence the male pronucleus (because it is bigger). An aqueous solution of the desired DNA is then pipetted directly into the pronucleus. The introduced DNA clones can integrate into chromosomal DNA at nicks, forming transgenes, usually containing multiple head-to-tail copies. Following withdrawal of the micropipette, surviving oocytes are reimplanted into the oviducts of pseudopregnant foster females (which have been mated with a vasectomized male; the mating act can initiate physiological changes in the female which stimulate the development of the implanted embryos). Newborn mice resulting from development of the implanted embryos are checked by PCR for the presence of the desired DNA sequence (Gordon, 1992).

To obtain transgenic mice by this route, females are superovulated, mated to fertile males and sacrificed the next day. Fertilized oocytes are recovered from excised oviducts. The DNA of interest is then microinjected using a micromanipulator into the male pronucleus of individual oocytes (Figure 21.1

). Surviving oocytes are reimplanted into the oviducts of foster females and allowed to develop into mature animals (see Gordon, 1992).

During this procedure, the microinjected DNA (transgene) randomly integrates into chromosomal DNA, usually at a single site, although rarely two sites of integration are found in a single animal. Individual insertion sites typically contain multiple copies of the transgenes integrated into chromosomal DNA as head-to-tail concatemers (it is not unusual to find 50 or more copies at a single insertion site). As a result of chromosomal integration, the transgenes can be passed on to subsequent generations in mendelian fashion: if the foreign DNA has integrated at the one-cell stage, it should be transmitted to 50% of the offspring.

Transfer into pre- or postimplantation embryos

Cells from very early stage embryos may be totipotent or at least pluripotent and can provide a route for foreign DNA to enter the germline. DNA can be transferred into unselected cells of very early embryos, as described in this section or into cell lines derived from embryonic stem cells, as described in Section 21.2.2.

One method that allows foreign DNA to integrate into the chromosomes of the target cells uses retroviruses, RNA viruses which naturally undergo an intermediate DNA form prior to integrating into cellular genomes (see Figure 18.2). Infection of preimplantation mouse embryos with a retrovirus such as Moloney murine leukemia virus or injection of the retrovirus into early postimplantation mouse embryos results in mosaic offspring. Retroviruses should integrate rarely and at random into accessible cells, and the use of replication- defective retroviruses provides heritable markers for clonal descendants of the target cell (unlike wild-type viruses which spread from cell to cell). This approach has been used, therefore, for studying cell lineage using reporter genes.

In Drosophila, efficient chromosomal integration is possible by using sequences from a Drosophila transposable element known as the P element to permit insertion of single copies of a gene at random in the genome. The gene or DNA fragment to be inserted is first manipulated so as to be flanked by the two terminal sequences of the P element. The modified DNA is then microinjected into a very young Drosophila embryo along with a separate plasmid containing a gene encoding a transposase. In the presence of the transposase the terminal P element sequences allow the intervening DNA fragment to transpose and as a result of the ensuing transposition events, the injected DNA often enters the germline in a single copy.

21.2.2. Cultured embryonic stem (ES) cells provide a powerful route to genetic modification of the germline

Figure 21.2
Genetically modified ES cells as a route for transferring (more...)

An external file that holds a picture, illustration, etc., usually as some form of binary object. The name of referred object is ch21f2.jpg.

Figure 21.2

Genetically modified ES cells as a route for transferring foreign DNA or specific mutations into the mouse germline

Cells from the inner cell mass were cultured following excision of oviducts and isolation of blastocysts from a suitable mouse strain (129). Such embryonic stem (ES) cells retain the capacity to differentiate into, ultimately, the different types of tissue in the adult mouse. ES cells can be genetically modified while in culture by insertion of foreign DNA or by introducing a subtle mutation. The modified ES cells can then be injected into isolated blastocysts of another mouse strain (e.g. C57B10/J which has a black coat color that is recessive to the agouti color of the 129 strain) and then implanted into a pseudopregnant foster mother of the same strain as the blastocyst. Subsequent development of the introduced blastocyst results in a chimera containing two populations of cells (including germline cells) which ultimately derive from different zygotes (normally evident by the presence of differently colored coat patches). Backcrossing of chimeras can produce mice that are heterozygous for the genetic modification. Subsequent interbreeding of heterozygous mutants generates homozygotes.

The microinjection of foreign DNA into fertilized oocytes is technically difficult and not suited to large-scale production of transgenic animals or to sophisticated genetic manipulation. A popular alternative, but one which has so far been restricted to the construction of genetically modified mice, involves transferring the foreign DNA initially into cultured embryonic stem (ES) cells. Mouse ES cells are derived from 3.5–4.5 day postcoitum embryos and arise from the inner cell mass of the blastocyst (see Figure 21.2

). The ES cells can be cultured in vitro and retain the potential to contribute extensively to all of the tissues of a mouse, including the germline, when injected back into a host blastocyst and reimplanted in a pseudopregnant mouse.

The developing embryo is a : it contains two populations of cells derived from different zygotes, those of the blastocyst and the implanted ES cells. If the two strains of cells are derived from mice with different coat colors, chimeric offspring can easily be identified (see Figure 21.2

). Use of genetically modified ES cells results in a partially transgenic mouse. Because the injected ES cells can form all or part of the functional germ cells of the chimera, it is possible to derive fully transgenic mice. This is usually accomplished by screening the offspring of matings between chimeras (usually males) and mice with a coat color recessive to that of the strain from which the ES cells were derived (see Figure 21.2

The big advantage of ES cells is that they can be grown readily in culture. This means that a variety of genetic manipulations can be conducted in cultured ES cells. Importantly, the desired genetic modification can be verified in tissue culture, before injecting the genetically modified cells into a blastocyst prior to implantation. For example, the desired gene can be ligated to a marker gene, such as the neo gene, enabling a positive selection for cells that have been successfully transfected (see Box 10.1). The presence of the desired gene can also be verified quickly by a PCR-based assay. ES cells also offer the huge advantage of gene targeting by homologous recombination, a method which can permit a programed selective alteration of a single predetermined gene and also highly specific ways of chromosome engineering. Such approaches are extremely powerful for understanding gene function (Section 21.3) and also for creating animal models of disease (Section 21.4).

Box 21.1

Isolation and manipulation of mammalian embryonic stem

(more...)

The ES cell approach to constructing transgenic mice was made possible by the successful establishment in the early 1980s of stable cell lines from isolated mouse ES cells. ES cells were not so readily identifed in other mammals, although there have been some important recent successes (see Box 21.1).

21.2.3. Transgenic animals have been used for a variety of studies investigating mammalian gene expression and function

Transgenic animals have been extremely important for analyzing human genes (Hanahan, 1989; Camper et al., 1995; Theuring, 1995), and have helped greatly in our understanding of a variety of fundamental biological processes, notably in immunology, neurobiology, cancer and developmental studies. The following list is far from exhaustive but is intended to illustrate some major types of application:

Investigating gene expression and its regulation. Although evidence for cis-acting regulatory elements is often inferred initially from studies using cultured cells, they need to be validated in whole animal studies. Transgenes consisting of the presumptive regulatory sequence(s) coupled to a reporter gene, such as lacZ, provide a sensitive method of detecting gene expression and a powerful way of investigating regulation of gene expression. Long-range control of gene expression is often investigated using YAC transgenes, see Section 21.2.4.
Investigating gene function by targeted gene inactivation. Specific genes can be inactivated by a gene targeting procedure to introduce a transgene into the target gene (insertional inactivation). The effect on the phenotype of creating a null mutation in the gene of interest can provide powerful clues to gene function.
Investigating dosage effects and ectopic expression. In some cases, valuable information can be gained by over-expressing a transgene (e.g. Schedl et al., 1996) or by expressing it ectopically (the transgene is coupled to a tissue-specific promoter which causes expression in cells; the phenotypic consequence may provide valuable clues to function).
Cell lineage ablation. Transgenes can be designed consisting of a tissue-specific promoter coupled to a sequence encoding a toxin, for example diphtheria toxin subunit A or ricin. When the promoter becomes active at the appropriate stage of tissue differentiation, the toxin is produced and kills the cells. Thus, certain cell lineages in the animal can be eliminated (cell ablation) and the phenotypic consequences monitored.
Investigating gain of function. In principle, any mammalian gene that produces a dominant negative effect or gain of function can be investigated by introduction of an appropriate transgene. In some cases, this can provide proof of a suspected biological function. A classical example concerns the Sry gene. A variety of different genetic analyses had implicated this gene as a major male-determining gene but convincing proof was obtained using a transgenic mouse approach. The experiment consisted of transferring a cloned Sry gene into a fertilized 46,XX mouse oocyte. As a result of this artificial intervention, the resulting mouse, which nature had intended to be female, turned out to be male (Koopman et al., 1991).
Modeling human disease. Insertional inactivation is often used to model loss of function mutations whilst gain of function mutations can often be modeled by inserting a mutant transgene (see Section 21.4).

21.2.4. The use of YAC transgenes and inducible promoters has greatly extended the applications of transgenic animals

Inducible promoters

For many applications it is desirable to have a transgene expressed under the control of a tissue-specific promoter or an inducible promoter (see Section 8.2.4). In the former case, genetic engineering can be used to splice known tissue-specific promoters from cloned genes to the gene of interest. In some cases, coupled regulatory elements can confer both position-independent and tissue-specific expression as with sequence elements from the β-globin locus control region (Grosveld et al., 1987).

Various attempts have also been made to create inducible transgenic mice (e.g. by using heavy metal ions to induce expression of an integrated gene which has a coupled metallothionein promoter, etc.). Generally, the use of inducible promoters has been hampered by ‘leakiness’ in gene expression and by relatively low levels of induction, and they have often been applicable to a limited range of tissues. More recently, however, more promising systems have been developed. For example, methods employing tetracycline-regulated inducible expression have permitted construction of both highly inducible transfected cells (with much greater efficiency than the constitutive system) and transgenic mice (Shockett et al., 1995). In the latter case, the expression of a reporter gene, such as the luciferase gene, can be controlled by altering the concentration of tetracycline in the drinking water of the animals.

YAC transgenics

Early studies of gene expression and regulation in transgenic animals involved transfer of small genes. However, expression of small transgenes often fails to follow the normal temporal and spatial patterns of expression or match the expression level of the endogenous homolog. An increasing number of human genes are known to be very large (see Figure 7.7). Even in the case of small genes, important regulatory elements that are required for correct expression may be located many kilobases upstream of the coding sequence (see Figure 8.23). In order to be able to study the expression and regulation of a human gene under the control of its own cis-acting regulatory elements, it was therefore necessary to establish transfection conditions which would allow the transfer of large DNA clones.

A major breakthrough in transgenic studies was the development of so-called YAC transgenics (Lamb and Gearhart, 1995). The first report to be published described transfer of a 670 kb YAC containing the human HPRT (hypoxanthine phosphoribosyltransferase) gene into mouse ES cells (Jakobovits et al., 1993). This was accomplished by spheroplast fusion (i.e. fusion of ES cells with YAC-containing yeast cells that have been stripped of the hard cell wall; see Section 4.3.4). Fragments from the yeast genome can integrate at the same time, however, and so alternative methods have sought to purify an individual YAC by size-fractionation on a preparative gel using pulsed-field gel electrophoresis (assuming that the YAC migrates at a position in the gel that is different from any yeast chromosome). The purified YAC can be inserted into a fertilized oocyte by pronuclear microinjection (see above). This method is, however, limited to small YACs: the DNA of large YACs is more likely to fragment following microinjection with very fine micropipettes. Alternatively, purified YACs have been transferred into ES cells by using liposomes, artificial lipid vesicles that are used to transport molecules into a cell following fusion of the lipid coat with the plasma membrane of the recipient cell (see Figure 22.6).

Figure 21.3
Use of YAC transgenesis to construct a mouse with a (more...)

An external file that holds a picture, illustration, etc., usually as some form of binary object. The name of referred object is ch21f3.jpg.

Figure 21.3

Use of YAC transgenesis to construct a mouse with a human antibody repertoire

YACs containing Ig sequences are obtained by screening YAC libraries with suitable Ig probes. The recovery of YACs with comparatively small inserts meant that there was a need to artificially construct larger YACs by homologous recombination in yeast cells. Spheroplasts are created by treating the yeast cells so that the outer cell wall is stripped off, making the cells more amenable to cell fusion. See Mendez et al. (1997) for further details of the procedures used to construct these mice.

YAC transgenics have permitted study of large genes such as the 400 kb human APP gene, a gene known to contribute to Alzheimer's disease, encoding the amyloid protein precursor. A YAC transgene containing APP showed tissue- and cell type-specific expression patterns closely mirroring that of the endogenous mouse gene (for this and other examples, see Lamb and Gearhart, 1995). Long-range gene regulation mechanisms (locus control regions, imprinting and other chromatin domain effects; see ) can be modeled. An interesting application has been in the production of fully human antibodies in the mouse by transfer of human YACs containing large segments of the human heavy and kappa light chain immunoglobulin loci into mouse ES cells, and thence the creation of transgenic mice able to produce human antibodies (see Mendez et al., 1997 and Figure 21.3

). Finally, YAC transgenics may also find a role in modeling disease caused by large-scale gene dosage imbalance (see Section 21.4.4).

Transchromosomic animals

Ultimately, even YACs have upper limits for the size of foreign inserts that can be transferred. Mammalian artificial chromosomes have also been generated, including first generation human artificial chromosomes (Harrington et al., 1997; Ikeno et al., 1998). Such systems will have the capacity of transferring hundreds and possibly thousands of genes into transgenic animals, although a preferred route may be by using nuclear transfer technology (Section 21.5) rather than ES cells. Recently, however, transfer of whole chromosomes or chromosome fragments into ES cells has been possible by microcell-mediated chromosome transfer (see Sections 10.1.1 and 10.1.3). Using this approach Tomizuka et al. (1997) were able to transfer human chromosomes or chromosome fragments derived from normal fibroblasts into mouse ES cells. The resulting chimeric transchromosomic mice were viable, and the chromosome fragments appeared to show functional expression and could be transmitted through the germline.

21.3. Use of mouse embryonic stem cells in gene targeting and gene trapping

21.3.1. Gene targeting by homologous recombination in ES cells can be used to produce mice with a mutation in a predetermined gene

Gene targeting involves engineering a mutation in a preselected gene within an intact cell. It can therefore be viewed as a form of artificial site-directed in vivo mutagenesis (as opposed to the various methods of site-directed in vitro mutagenesis described in Section 6.4). The mutation may result in inactivation of gene expression (a ‘knock-out’ mutation), or altered gene expression, and so can be useful for studying gene function (see below). In addition, the same method can be used to ‘correct’ a pathogenic mutation by restoring the normal phenotype, and so has therapeutic potential (see Section 22.3).

Gene targeting typically involves introducing a mutation by homologous recombination. A cloned gene (or gene segment) closely related in sequence to endogenous target gene is transfected into the appropriate cells. In some of the cells, homologous recombination occurs between the introduced gene and its chromosomal homolog. Gene targeting by homologous recombination has been achieved in some somatic mammalian cells, such as myoblasts. However, the most important application involves mouse ES cells: once a mutation has been engineered into a specific mouse gene within the ES cells, the modified ES cells can then be injected into the blastocyst of a foster mother and eventually a mouse can be produced with the mutation in the desired gene in all nucleated cells (Capecchi, 1989; Melton, 1994).

Homologous recombination in mammalian cells is a very rare occurrence (unlike in yeast cells, for example, where it occurs naturally at high frequencies, enabling sophisticated genetic manipulation). The frequency of homologous recombination is increased, however, when the degree of sequence homology between the introduced DNA and the target gene is very high. As a result, the introduced DNA clone is a mouse sequence which should preferably be isogenic (derived from the same mouse strain as the strain of mouse from which the ES cells were derived). Even then, the frequency of genuine homologous recombination events is very low and may be difficult to identify against a sizeable background of random integration events.

To assist identification of the desired homologous recombination events, the targeting vector (often a plasmid vector) contains a marker gene, such as the neo gene (see Box 10.1), which permits selection for cells that have taken up the introduced DNA. PCR assays are used to screen for evidence of a homologous recombination event (by using a marker-specific primer plus a primer derived from a sequence present in the target gene but absent from the introduced homologous gene segment). The targeting construct is transferred into cultured mouse ES cells by electroporation, a method in which pulses of high voltage are delivered to cells, causing temporary relaxation of the selective permeability properties of the plasma membranes. Two basic approaches have been used:

Figure 21.4
Gene targeting by homologous recombination can inactivate (more...)

Figure 21.4
.

Gene targeting by homologous recombination can inactivate a predetermined chromosomal gene within an intact cell
(A) Insertion vector method. The introduced vector DNA (blue) is cut at a unique site within a sequence which is identical or closely related to part of a chromosomal gene (black). Homologous recombination (X) can occur, leading to integration of the entire vector sequence including the marker gene (M). Note that the letters do not represent exons but are simply meant to indicate linear order within the gene. (B) Replacement vector method. In this case, the marker gene is contained within the sequence homologous to the endogenous gene, and the vector is cut at a unique location outside the homologous sequence. A double recombination or gene conversion event (X X) can result in replacement of internal sequences within the chromosomal gene by homologous sequences from the vector, including the marker gene.

Insertion vectors target the locus of interest by a single reciprocal recombination, causing insertion of the entire introduced DNA including the vector sequence (Figure 21.4A). This is the most reliable way of causing a knock-out mutation.
Replacement vectors are designed to replace some of the sequence in the chromosomal gene by a homologous sequence from the introduced DNA (Figure 21.4B). This can occur as a result of a double reciprocal recombination or by gene conversion. The replacement method can inactivate a gene when the introduced sequence contains one or more premature termination codons or lacks critical coding sequences. It can also be used to correct a pathogenic mutation.

Figure 21.5
Double replacement gene targeting can be used to introduce (more...)

An external file that holds a picture, illustration, etc., usually as some form of binary object. The name of referred object is ch21f5.jpg.

Figure 21.5

Double replacement gene targeting can be used to introduce subtle mutations

Both the methods in Figure 21.4

result in introduction of a substantial amount of exogenous sequence within the endogenous gene. To introduce a subtle mutation without leaving residual exogenous sequence, a double replacement method with positive and negative selection can be used (Melton, 1994). Exons in the endogenous gene are represented as numbered large boxes, and introns as long thin boxes. In order to introduce a subtle mutation, such as a single nucleotide substitution in exon 8, a replacement knock-out vector is used with a marker gene (e.g. the HPRT gene) flanked by homologous sequences from introns 7 and 8, and a second marker, such as the herpes simplex thymidine kinase (TK) gene outside the homologous region. Gene conversion, or double crossover within the flanking intron sequences, can lead to replacement of exon 8 by the HPRT gene, and can be selected for if a mutant HPRT^- ES cell is used. A positive-negative selection system can be used: selection in the first step is for HPRT⁺ TK^- cells. Cells containing random vector integrations will contain the TK gene and can be killed with the thymidine analog gancyclovir (see ). The second replacement involves introducing an altered exon 8 with a point mutation (*) to replace the HPRT gene and can be screened by identifying HPRT^- cells. Note that mice engineered in this way cannot be described as transgenic because of the lack of foreign sequences in the germline.

The replacement vector approach, as well as the insertion vector method, often leaves foreign sequences at the target locus. In some cases, however, a more subtle mutation is required. For example, it may be desirable to investigate the effect of changing a single codon. Various two-step recombination techniques can be used to accomplish this method, and the resulting mouse, although genetically modified lacks any foreign sequences and so can no longer be described as transgenic (see Figure 21.5

and Melton, 1994).

Gene targeting in mice is popularly used for producing artificial mouse models of human disease (Section 21.4). In addition, it provides a powerful general method of studying gene function. The gene in question is selectively inactivated, producing a ‘knock-out’ mouse, and the effect of the mutation on the development of the mouse is monitored carefully. Sometimes there is little or no phenotypic consequence after inactivating a gene that would be expected to be crucially important, such as some genes which encode a transcription factor known to be expressed in early embryonic development. The lack of a phenotype in such cases is often thought to be due to genetic redundancy (another gene is able to carry out the function of the gene that has been knocked out). As a result, in some cases double or even triple gene knock-outs have been carried out to analyze gene function, as in the case of some of the Hox genes (see, for example, Manley and Capecchi, 1997).

Figure 21.6
The knock-in method replaces the activity of one chromosomal (more...)

An external file that holds a picture, illustration, etc., usually as some form of binary object. The name of referred object is ch21f6.jpg.

Figure 21.6

The knock-in method replaces the activity of one chromosomal gene by that of an introduced gene

The En-1 gene shown at top has two exons and coding sequences are shown by filled boxes. Its promoter (P) and polyadenylation site (pA) are also shown. The gene targeting vector (‘knock-in vector’) contains cloned En-1 gene sequences comprising an upstream sequence a which contains the En-1 promoter, and an internal segment b which spans the 3′ end of exon 1, the single intron and the 5′ coding sequence of exon 2. Separating these two sequences is the coding sequence of the En-2 gene. Two marker cassettes include a thymidine kinase (TK) gene and a neomycin-resistance gene (Neo) both driven by a phosphoglycerate kinase promoter (chosen because phosphoglycerate kinase is expressed in ES cells). The Neo gene is flanked by loxP sequences. The targeting procedure results in replacement of endogenous sequences a and b but the 5′ coding sequence of the En-1 gene is deleted. The knocked-in En-2 gene comes under the control of the En-1 promoter (Hanks et al., 1995). Note that the term ‘knock-in’ has also been applied to any procedure where an endogenous gene is inactivated by insertion of a new gene which is then expressed, even if the latter is only meant to serve as a reporter gene such as lacZ.

A useful example of investigating functional redundancy concerns studies of the mouse Engrailed genes, En-1 and En-2. Both of these genes are homeobox genes which had been considered to play crucial roles in brain formation. En-1 knock-outs have serious abnormalities but surprisingly En-2 knock-outs have only minor problems. Expression of the En-1 gene is switched on 8–10 hours before that of the En-2 product, suggesting that perhaps the En-1 product can compensate for the lack of En-2 product in En-2 knock-outs. To test for the possibility of functional redundancy, Hanks et al. 1995 used a variant of the knock-out procedure known as the ‘knock-in’ technique. In this case the transgene used to knock-out the target endogenous gene is itself designed to be expressed under the control of the cis-acting elements of the knocked-out gene. A transgene containing an En-2 coding sequence was used to knock-out the endogenous En-1 gene. In so doing, the introduced En-2 sequence came under control of the En-1 regulatory sequences and was expressed before the endogenous En-2 gene was switched on (see Figure 21.6

). The resulting En-1 knock-out mouse had a normal phenotype, demonstrating that the knocked-in En-2 gene was functionally equivalent to En-1 (Hanks et al., 1995).

21.3.2. Site-specific recombination systems, notably the Cre-loxP system, extend the power of gene targeting

Figure 21.7
Structure of the loxP recognition sequence

An external file that holds a picture, illustration, etc., usually as some form of binary object. The name of referred object is ch21f7.jpg.

Figure 21.7

Structure of the loxP recognition sequence

Note that the central 8 bp sequence which is flanked by the 13 bp inverted repeats is asymmetric and confers orientation.

Several site-specific recombination systems from bacteriophages and yeasts have been characterized and are promising tools for genome engineering (Kilby et al., 1993). Thus far, the Cre-loxP recombination system from bacteriophage P1 has been the most widely used. The natural function of the Cre (causes recombination) recombinase is to mediate recombination between two loxP sequences that are in the same orientation (the loxP sequence consists of 34 bp and comprises two inverted 13 bp repeats separated by a central asymmetric 8 bp spacer; Figure 21.7

). As a result of recombination, the intervening sequence between the two loxP sites is excised (see ). Using gene targeting, loxP sequences can be stitched into a desired gene or chromosomal location, and the subsequent provision of a gene encoding the Cre product can result in an artificially generated site-directed recombination event (see Chambers, 1994). As described below, several applications can be envisaged (Lobe and Nagy, 1998).

Tissue- and cell type-specific knock-outs

Some genes are vital to early development and simple knock-out experiments are generally not helpful because death ensues at the early embryonic stage. To overcome this problem, methods have been developed to inactivate expression of the target gene in only selected, predetermined cells of the animal (conditional knock-out). The animal can therefore survive and the effect of the knock-out can be studied in a tissue or cell type of interest.

Figure 21.8
Gene targeting using the Cre-loxP recombination (more...)

An external file that holds a picture, illustration, etc., usually as some form of binary object. The name of referred object is ch21f8.jpg.

Figure 21.8

Gene targeting using the Cre-loxP recombination system can be used to inactivate a gene in a desired cell type

(A) Illustration of a standard homologous recombination method using mouse ES cells, in which three loxP sites are introduced along with a marker M at a target locus A (typically a small gene or an internal exon which if deleted would cause a frameshift mutation). Subsequent transfection of a Cre recombinase gene and transient expression of this gene results in recombination between the introduced loxP sites to give different products. Type I recombinants are used to generate mice in which the target locus is flanked by loxP sites. Such mice can be mated with previously constructed transgenic mice (B) which carry an integrated construct consisting of the Cre recombinase gene linked to a tissue-specific promoter. Offspring which contain both the loxP-flanked target locus plus the Cre gene will express the Cre gene in the desired tissue type, and the resulting recombination between the loxP sites in these cells results in tissue-specific inactivation of the target locus A.

An early example of a conditional knock-out sought to study the role of DNA polymerase β (an enzyme that is essential for embryonic development) in T lymphocytes (Gu et al., 1994). The gene targeting procedure replaced part of the endogenous gene by an introduced homologous gene segment flanked by loxP sequences. Mice carrying this targeted mutation were then mated with a strain of mice which carried a Cre transgene gene under the control of a T cell-specific promoter. Offspring with the loxP-flanked pol β sequences plus the Cre transgene were identified and survived to adulthood. The Cre product was expressed only in T cells, leading to inactivation of the target gene in these cells by excision of DNA polymerase β gene segment between the two loxP sequences (see Figure 21.8

for the general method).

Tissue- and cell type-specific gene activation

This approach is the opposite to that described above: it involves selective activation of a gene in certain cells of the animal to switch on a foreign gene only in predetermined cells of the animal (Barinaga, 1994).

Chromosome engineering

Figure 21.9
Chromosome engineering can be accomplished using (more...)

An external file that holds a picture, illustration, etc., usually as some form of binary object. The name of referred object is ch21f9.jpg.

Figure 21.9

Chromosome engineering can be accomplished using cre-loxP systems

(A) Use of targeted insertion of loxP sites to facilitate a chromosomal translocation. See Smith et al. 1995 for a practical example. (B) Use of targeted insertion of loxP sites to permit modeling of a microdeletion by intrachromosomal recombination. See Ramirez-Solis et al. 1995 for a practical example, and for other examples of intrachromosomal recombinations.

Another important recent development is a strategy for chromosome engineering in ES cells which relies on sequential gene targeting and Cre-loxP recombination. Gene targeting is used to integrate loxP sites at the desired chromosomal locations and, subsequently, transient expression of Cre recombinase is used to mediate a selected chromosomal rearrangement (Ramirez-Solis et al., 1995; Smith et al., 1995; Figure 21.9

). Chromosome engineering strategies of this type offer the exciting possibility of creating novel mouse lines with specific chromosomal abnormalities for genetic studies.

The multiple targeting and selection steps in ES cells used in the above chromosome engineering methods can be avoided using the novel approach of Herault et al. 1998. This targeted meiotic recombination method takes advantage of the homologous chromosome pairing that occurs naturally during meiosis at the first cell division. A transgene is designed to express Cre recombinase under the control of a Sycp1 promoter (the Sycp1 gene encodes the SCYP1 protein which is part of the synaptonemal complex that facilitates crossing over). As a result Cre recombinase is produced in male spermatocytes during the zygotene to pachytene stages when chromosome pairing occurs.

21.3.3. Gene trapping in mouse ES cells allows an efficient approach to functional analysis

Large-scale approaches to investigating animal gene function have principally relied on exposing animals to high doses of radiation, or to potent chemical mutagens such as ethylnitrosurea (ENU) or ethyl methylsulfonate (EMS). In mouse mutagenesis programs, for example, males are typically exposed to high levels of a suitable mutagen to induce a high frequency of mutation in sperm DNA. The progeny of irradiated mice are then screened for obvious phenotypic abnormalities. Such mutagenesis screens have been very useful in producing novel mutants, but a major problem is that the mutations occur essentially at random. Identification of the structural change(s) in the DNA of a single mutant animal may therefore require a laborious positional cloning approach.

A mutation that is induced by inserting a known foreign DNA sequence (transgene) into the mouse genome has a major advantage over one induced by chemical mutagens or X rays: it leaves a sequence tag at the locus which is mutated. As a result, rapid molecular characterization of the mutated locus is possible. Mouse ES cells provide a way of introducing such mutations into the germ line, and so random insertional mutagenesis in ES cells using transgenes was considered a useful way of producing a large number of mouse mutants which could quickly be characterized at the molecular level. Unlike gene targeting, the transgene should not be related to endogenous sequences, but should integrate essentially at random by nonhomologous recombination. However, because large sections of the mouse genome are noncoding, many random insertion events may not result in gene disruption.

In order to improve the efficiency of recovering mutations that are likely to have a phenotypic effect (by altering a gene or its expression) the gene trap approach was devised (see Evans et al., 1997). The underlying principle is that the transgene which is inserted into ES cells contains a defective reporter gene or marker gene which lacks some component needed for gene expression. The reporter or marker gene is designed to be expressed only after it inserts into a gene (within an intron or exon) or at a promoter. When it integrates at such positions it can acquire the expression element that it lacks.

Figure 21.10
Gene trapping uses an expression-defective transgene (more...)

An external file that holds a picture, illustration, etc., usually as some form of binary object. The name of referred object is ch21f10.jpg.

Figure 21.10

Gene trapping uses an expression-defective transgene to select for chromosomal integration events that occur within or close to a gene

A specimen host cell gene with 4 exons (E1–E4) is shown at top together with its promoter (P) and polyadenylation signal (pA). Two possibilities for a gene trap vector are shown. Transgene 1 includes a reporter gene which lacks a promoter. It has two exonic sequences a and b, a polyadenylation signal (pA) and an upstream sequence which contains a splice acceptor sequence (SA). In this example, transgene 1 integrates into intron 1 of the host cell gene. During transcription from the endogenous promoter the splice acceptor sequence will help the transgene exons to be spliced to the sequence from the first exon of the endogenous gene, producing a fusion transcript. Selection is based on this transcript having a functional reporter activity (see Evans et al., 1997). Transgene 2 has a marker gene (drug-resistance, e.g. puromycin N-acetyl transferase) coupled to a promoter which works in ES cells (usually a phosphoglycerate kinase promoter) and a downstream splice donor sequence, but lacks a polyadenylation signal. Again, integration is intended to permit expression but this time from the transgene's promoter and the splice donor sequence helps the RNA transcript to be spliced to downstream host exons. The location of the integrated transgene can be identified by RACE-PCR (see Figure 20.1). See Zambrowicz et al. 1998 for a detailed example.

Different gene trap strategies are possible. In some cases, the reporter lacks a functional promoter and so relies on chance integration next to an appropriate cis-acting sequence element that can activate its transcription. Other approaches have used a marker gene coupled to a suitable promoter but lacking a downstream polyadenylation signal. Here, the marker gene is designed to be expressed after integrating into a host cell gene such that a fusion RNA product is made that utilizes 3′ host sequences in order to acquire a poly (A) tail (see Figure 21.10

). Already, this type of approach has been applied on a large scale, with a recent report describing disruption and sequence identification of 2000 genes in mouse embryonic stem cells (Zambrowicz et al., 1998).

21.4. Creating animal models of disease using transgenic technology and gene targeting

Table 21.1
Classes of animal models of disease

Table 21.1

Classes of animal models of disease

Origin	Type	Comments
(A) Spontaneous	Germline mutation → inherited disorder	See Table 20.4.
	Somatic mutation → cancer
(B) Artificial intervention or artificially generated	Selective breeding to obtain strains that are genetically susceptible to disease
	Infect animal strain with relevant microbial pathogen
	Manipulate environment to induce disease without causing mutation	e.g. injection of pristane, a synthetic adjuvant oil, into the dermis of rats has produced amodel of arthritis.
	In vivo mutagenesis using a strong mutagen such as X-rays or powerful chemical mutagens such as ethyl nitrosurea (ENU)	Large chemical mutagenesis programs have been established to produce mutant mice and zebrafish (Section 21.4.1)
	Genetic modification of fertilized egg cells or cells from the early embryo and subsequent animal breeding (transgenic and gene targeting technologies)	Sections 21.2 and 21.3

Animal models of human disease are crucially important to medical research. They allow detailed examination of the physiological basis of disease. They also offer a front-line testing system for studying the efficacy of novel treatments before conducting clinical trials on human subjects. Although many individual human disorders do not have a good animal model, animal models exist for some representatives of all the major human disease classes: genetically determined diseases, disease due to infectious agents, sporadic cancers and autoimmune disorders (see Leiter et al., 1987; Darling and Abbott, 1992; Clarke, 1994; Bedell et al., 1997). Some animal models of human disease originated spontaneously; others have been generated artificially by a variety of different routes (Table 21.1).

Until recently, the great majority of available animal disease models were ones which arose spontaneously or had been artificially induced by random mutagenesis using exposure to high doses of mutagenic chemicals or X-rays (next section). More recently gene targeting and transgenic technologies have provided direct ways of obtaining animal models of disease, and targeted mutations in the mouse have been particularly valuable. Interestingly, it has become increasingly clear that disease phenotypes due to comparable mutations in human and mouse gene homologs often show considerable differences (Section 21.4.5).

21.4.1. Animal models of disease occurring spontaneously or as a result of exposure to chemical mutagens or radiation may be difficult to identify

Spontaneous animal disease models

Mutant human phenotypes, especially those associated with obvious disease symptoms, are subject to intense scrutiny: most individuals who suffer from a disorder seek medical advice. If they present with a previously undescribed phenotype their case may well be referred to experts who often will document the phenotype in the medical literature. Given the motivation of both affected individuals and their families, physicians and interested medical researchers, and the large population size for screening (current total global population is ~6 billion individuals), there is a remarkably effective screening process for mutant human phenotypes. In contrast, many animal disease phenotypes will go unrecorded. Only a small percentage of the animal population is in captivity, and recording of spontaneous mutant phenotypes is largely dependent on examination of animal colonies bred for research purposes and, to a lesser extent, livestock and pet populations. Only mutants with obvious external anomalies are likely to be noticed.

Table 21.2
Examples of spontaneous animal mutants

Table 21.2

Examples of spontaneous animal mutants

Animal mutant	Phenotypic features and molecular pathogenesis
NOD mouse	Diabetic, but without being obese. Mimics human insulin-dependent diabetes mellitus
mdx mouse	X-linked muscular dystrophy due to mutations in mouse dystrophin gene. The original mdx mutant has a nonsense mutation but phenotype is much milder than Duchenne muscular dystrophy (DMD)
Hemophiliac dog	Missense mutation in canine factor IX gene causes complete loss of function. Human homolog is hemophilia B
Watanabe heritable hyperlipidemic (WHHL) rabbit	Hyperlipidemic as a result of a deletion of four codons of the low density lipoprotein receptor gene (LDLR); human homolog is familial hypercholesterolemia
Atherosclerotic pigs	Marked hypercholesterolemia. Normal LDL receptor activity, but variant apolipoproteins, including apolipoprotein B
Splotch mouse	Abnormal pigmentation; phenotypic overlap with Waardenburg syndrome suggested that it was an animal model for this disease, and confirmed by identification of mutations in homologous human and murine PAX genes
NF damselfish	Extensive neurofibromas suggest that it could be a homolog of human neurofibromatosis type I

Despite the difficulty in identifying spontaneous animal mutants, a number of animal phenotypes have been described as likely models of human diseases (see Table 21.2 for some examples). In some cases, the animal mutant phenotype closely parallels the corresponding clinical phenotype, but in others there is considerable divergence because of species differences in biochemical and developmental pathways (see Erickson, 1989; Erickson, 1996; Wynshaw-Boris, 1996; see also Section 21.4.5). Additionally, phenotypic differences may result because of different classes of mutation at orthologous loci.

Random mutagenesis using chemicals and irradiation

Box 21.2

The potential of animals for modeling human disease

(more...)

Classical methods of producing animal mutants have involved controlled exposure to mutagenic chemicals, notably ethyl nitrosurea (ENU) and ethyl methylsulfonate (EMS) or to high doses of X-rays. Large numbers of Drosophila and mouse mutants have been obtained by this method and, very recently, efficient large-scale mutagenesis screens have been conducted on both mice and also zebrafish (which offers some advantages as an animal model; see Box 21.2). A major problem with chemically-induced and irradiation-induced mutations, however, is that they are generated essentially at random. In order to identify a mutant phenotype of interest, a laborious screen for mutants needs to be conducted by close examination of the phenotypes following mutagenesis. The mutant phenotypes which have been described in these studies, as well as for spontaneous mutants, show a clear bias towards phenotypes with obvious external abnormalities, simply because of the ease of identifying them. Nevertheless, several important models of human disease have been created using such methods.

21.4.2. Mice have been widely used as animal models of human disease largely because specific mutations can be created at a predetermined locus

Spontaneous and artificially produced disease phenotypes have been described in a wide range of animal species with differing potentials for modeling human disease. In some cases, the species may be too evolutionarily remote from humans to provide useful disease models. For example, numerous Drosophila mutants have been generated and studies of some developmental mutants have enabled the identification of human genes that are important in development, but they cannot serve as useful models of human disease. Other species, such as the zebrafish, which are also evolutionarily remote from us, offer some advantages as model organisms and models have been reported for some human diseases (Box 21.2). Mammals would be expected to provide better disease models but, for a variety of reasons, our closest relatives, the great apes, have not been very useful in providing disease models. Instead, other mammals, notably mice, have been used widely to model human disease (see Box 21.2).

Table 21.3
Examples of transgenic or gene-targeted mouse models (more...)

Table 21.3

Examples of transgenic or gene-targeted mouse models of human disease

Human disease or abnormal phenotype	Gene	Method of constructing model
Cystic fibrosis	CFTR	Insertional inactivation by gene targeting
β-Thalassemia	HBB (β-globin)	Insertional inactivation by gene targeting
Hypercholesterolemia and atherosclerosis	Apolipoprotein genes, e.g. APOE	Insertional inactivation by gene targeting
Gaucher's disease		Insertional inactivation by gene targeting
Fragile-X syndrome	FMR1	Insertional inactivation by gene targeting
Gerstmann-Sträussler-Scheinker (GSS) protein gene with missense mutation syndrome	Prion protein gene (PRNP)	Integration of mutant mouse prion gene
Spinocerebellar ataxia type 1 (SCA1)	SCA1 (ataxin)	Integration of mutant human ataxin gene with expanded triplet repeat
Alzheimer's disease	APP (β-amyloid precursor protein)	Integration of mutant full-length APP cDNA under control of a platelet-derived growth factor promoter

In the case of animal disease models which are artificially induced by exposure to mutagenic chemicals or radiation, or which originate spontaneously, there is little or no artificial control over the resulting phenotype and frequently, the identification of an animal disease model is serendipitous. The great advantage of transgenic/genetargeted mouse models of disease is that specific disease models can be constructed to order. Provided that the relevant gene clones are available, including mutant genes in some cases, mice can be generated with a desired alteration in a chosen target gene. All the major classes of disease, inherited disorders, cancers, infectious diseases and autoimmune disorders can be modeled in this way (Table 21.3; Smithies, 1993; Clarke, 1994; Bedell et al., 1997). In most cases, the transgenic/gene targeting approaches have been used to model single gene disorders but, increasingly, attempts are being made to produce mouse models of complex genetic diseases, such as Alzheimer's disease, atherosclerosis and essential hypertension effects.

21.4.3. Single gene disorders resulting from loss of function and gain of function mutations can be conveniently modeled by gene targeting and by integration of mutant genes respectively

Modeling loss of function mutations by gene targeting in mice

Many disease phenotypes, including those of essentially all recessively inherited disorders and many dominantly inherited disorders, are thought to result from loss of gene function. The simplest way of modeling the disease for single gene disorders of this type is to make a knock-out mouse. The first step is to isolate the orthologous mouse gene and to use a segment of it to knock out the endogenous gene in mouse ES cells using gene targeting. Following injection of the genetically modified ES cells into the blastocyst of a foster mother, and continued development, founder mice are obtained with the targeted mutation in a sizeable proportion of their germ cells. These mice can be interbred and the offspring can be screened for the presence of the desired mutation, and for the presence of the wild-type allele using PCR assays of cells collected from tail bleeds.

The gene targeting event is intended to create a null allele (where there is complete absence of gene expression), but sometimes the result may be a ‘leaky’ mutation and the mutant allele retains some gene expression. This may explain why gene targeting has produced mouse models of a disease with divergent phenotypes. For example, the mouse model of cystic fibrosis described by Snouwaert et al. 1992 had a severe phenotype while that reported by Dorin et al. 1992 had a mild phenotype because of a ‘leaky’ mutation. Differences in phenotype may also occur because of modifier genes using different mouse strains (see Section 21.4.5).

Modeling gain of function mutations by insertion of a mutant gene

This general experimental design has been used frequently in conjunction with the pronuclear microinjection technique of gene transfer. The disease to be modeled must be one where the presence of an introduced DNA is itself sufficient to induce pathogenesis, and can include inherited gain of function mutations, oncogenes, etc. To model such disorders, it is necessary to clone a mutant gene or, if necessary, design one by in vitro mutagenesis. The mutant gene is then simply inserted as a transgene, e.g. by microinjection into fertilized oocytes. Because there is no requirement for the introduced mutant gene to integrate at a specific location, human mutant genes will suffice although, in some cases, mouse mutant genes have been used. The two examples below illustrate this approach.

An early example was intended to assess whether a leucine substitution found at codon 102 in a prion protein gene in a patient with Gerstmann-Sträussler-Scheinker (GSS) syndrome was pathogenic. An analogous mutation was artificially designed in a cloned mouse prion protein gene and the mutant gene was then injected into fertilized oocytes to produce transgenic mice. The mice went on to develop spontaneous neurodegeneration, reminiscent of that found in the human syndrome (Hsiao et al., 1990). A variety of other experiments using prion protein transgenes have been very helpful in understanding prions (Gabizon and Taraboulos, 1997).
Expanded triplet repeats causing neurodegenerative disorders (see Box 16.7) constitute another type of gain of function mutation. Spinocerebellar ataxia type 1 (SCA1) is a dominantly inherited disorder which results from unstable expansion of a CAG triplet repeat in the ataxin gene. It is characterized by degeneration of cerebellar Purkinje cells, spinocerebellar tracts and some brainstem neurons. Transgenic mice were produced by introduction of one of two transgenes driven by a Purkinje cell-specific promoter: the normal human ataxin gene (SCA1), and a mutant ataxin gene containing an expanded CAG repeat. Both types of transgene were stable in parent to offspring transmissions, but only those with the expanded allele developed ataxia and Purkinje cell degeneration, confirming the gain of function hypothesis (Burright et al., 1995).

21.4.4. Considerable effort is currently being devoted to constructing mouse models of cancers and other complex genetic disorders

Modeling human cancers

(See Ghebranious and Donehower, 1998; Macleod and Jacks, 1999.)

Gain of function. Disease due to inappropriate inactivation of a proto-oncogene can be modeled by constructing a transgenic mouse: the appropriate oncogene is introduced into the mouse genome by simple transgene integration.
Loss of function. Disease due to inactivation of tumor suppressor genes can be modeled by constructing knock-out mice through gene targeting. For example, several models have been generated by inactivating the mouse homologs of the TP53 and RB1 genes but the phenotypes show only broad similarity to the homologous human phenotypes, respectively Li-Fraumeni syndrome and retinoblastoma (see also Section 21.4.5).

Modeling chromosomal disorders

Existing mouse models for human chromosomal disorders are sparse. In some cases this is due to insufficent conservation of synteny between the two species. Taking the example of Down syndrome (trisomy 21), human chromosome 21 shares a large region of genetic homology with mouse chromosome 16 (see Figure 14.23), but trisomy 16 (Ts16) mice are not good models because they die in utero. The Ts16 mouse could never be expected to be a good model of human trisomy 21: it is not trisomic for all human chromosome 21 genes (the genes in the distal 2–3 Mb of human chromosome 21 have orthologs on mouse chromosomes 17 and 10) and is trisomic for some genes which have human orthologs on chromosomes other than chromosome 21.

In order to produce a better Down syndrome model in the mouse, attention has focused on the Down syndrome critical region at 21q21.3-q22.2 (deduced from observing the phenotypes of rare Down syndrome patients with partial trisomy 21). Within this region, the human minibrain gene at 21q22.2 may be an important contributory locus to the associated learning defects: transgenic mice in which a 180 kb YAC containing the 100 kb human minibrain gene but apparently no other gene, develop learning defects (Smith et al., 1997). A segmental trisomy 16 mouse, Ts65Dn, obtained by standard methods of irradiating mice was shown by Reeves et al. 1995 to have learning and behavior deficits. Other more recently produced models are discussed by Kola and Hertzog (1998).

Future efforts in modeling chromosomal disorders can be expected to take advantage of gene targeting using Cre-loxP. As described in Section 21.3.2 this system offers tremendous potential for genome engineering and can be used to engineer chromosome translocations at defined positions on preselected chromosomes. YAC transgenics can also be expected to be important in investigating over-expression of genes in other chromosomal disorders and in disorders resulting from aberrant gene dosage for large regions, such as Charcot-Marie-Tooth type 1A, which is due to overexpression as a result of a 1.5-Mb duplication in the PNP22 gene region (see Figure 16.7).

Modeling complex diseases

Increasingly, the focus in human genetics is moving towards understanding the pathogenesis of complex genetic diseases such as atherosclerosis, essential hypertension, diabetes, etc. Such disorders have a complex etiology with multiple genetic and environmental components. Some valuable animal models have been produced for some of these disorders but gene targeting approaches are expected in the future to provide additional badly needed models (Smithies and Maeda, 1995). As long as suitably promising genes can be identified as being involved in the pathogenesis, then breeding experiments can be used to bring different combinations of disease genes together, and the effect of different genetic backgrounds in different strains of mice and of different environmental factors can be assessed. This approach may not be so daunting as it sounds because, increasingly, many complex disease phenotypes are considered to be due mostly to the combination of only a very few major susceptibility genes. For example, a digenic model of spina bifida occulta was generated serendipitously in offspring obtained by crossing a mouse heterozygous for the Patch mutation (pdgfrb; platelet-derived growth factor receptor) with a mouse homozygous for the undulated mutation (pax-1) (see Helwig et al., 1995).

21.4.5. Mouse models of human disease may be difficult to construct because of a variety of human/mouse differences

It is not uncommon for spontaneous or artificially generated mouse models of disease to show phenotypes that are considerably different from the homologous human disorders. For example, gene targeting to inactivate several mouse tumor suppressor genes has often produced disappointing mouse models, as in the case of TP53 and RB1 (retinoblastoma) knock-outs. There may be problems in achieving the desired type of mutation. For example, in gene targeting intended null mutations may be offset in some cases by exon skipping or some other form of ‘leaky’ transcription, or the expression of a transgene may be affected by various parameters causing an unexpected phenotype. Setting aside these possibilities, there are several areas where differences between mice and humans could be expected to result in divergent disease phenotypes for mutations in orthologous genes (Erickson, 1989; Erickson, 1996; Wynshaw-Boris, 1996):

Biochemical pathways. Although biochemical pathways in mammals are generally well conserved, some differences are known between the pathways of humans and mice. The human retina appears to depend heavily on the accurate function of the Rb gene product, but other vertebrate retinas do not. As a result, spontaneous retinoblastoma mouse mutants have not been described, and retinoblastoma is not a feature of Rb1 knock-out mice. Another example may be provided by ganglioside degradation pathways and while deficiency of the hexosaminidase gene HEXA results in the severe Tay-Sachs disorder, inactivation of the mouse homolog Hexa results in abnormal accumulation of ganglioside in neurons but without motor deficits or learning deficits (Wynshaw-Boris, 1996).
Developmental pathways. The differences in human and mouse developmental pathways are not well understood but are expected to be significant for some organ systems, such as in brain development.
Absolute time. Because of the huge difference in the average lifespans of mice and humans, certain human disorders in which the disease is of late onset may possibly be difficult to model in mice.
Genetic background: the importance of modifier genes. Most human populations are outbred. Laboratory strains of mice, however, are very inbred. Often a particular phenotype can vary considerably in different strains of mice because of differences in their alleles at other loci (modifier genes), which can interact with the locus of interest.

A useful example of the importance of genetic background is the Min (multiple intestinal neoplasia) mouse which was generated by ENU mutagenesis and results from mutations in the mouse Apc gene. Mutations in the orthologous human gene, APC, cause adenomatous polyposis coli and related colon cancers and the Min mouse has been regarded as a good model for such disorders. The phenotype of the Min mouse is, however, dramatically modified by the genetic background. For example the number of colonic polyps in mice carrying APC^Min is strikingly dependent on the strain of mouse. Similar phenotypic variability is found in human families where different members of the same family may have strikingly different tumor phenotypes although they possess identical mutations in the APC gene. Some of the variability could be due to environmental factors, but the involvement of modifier genes had been strongly suspected. The Min mouse provides a well-defined genetic system for mapping and identifying modifier genes (Dietrich et al., 1993; MacPhee et al., 1995).

21.5. Manipulating animals by somatic cell nuclear transfer

21.5.1. Principles and practice of animal cloning

The term clones indicates genetic identity and so can describe genetically identical molecules (DNA clones), genetically identical cells or genetically identical organisms. Animal clones occur naturally as a result of sexual reproduction. For example, genetically identical twins are clones who happened to have received exactly the same set of genetic instructions from two donor individuals, a mother and a father. A form of animal cloning can also occur as a result of artificial manipulation to bring about a type of asexual reproduction. The genetic manipulation in this case uses nuclear transfer technology: a nucleus is removed from a donor cell then transplanted into an oocyte whose own nucleus has previously been removed. The resulting ‘renucleated’ oocyte can give rise to an individual who will carry the nuclear genome of only one donor individual, unlike genetically identical twins. The individual providing the donor nucleus and the individual that develops from the ‘renucleated’ oocyte are usually described as ‘clones’, but it should be noted that they share only the same nuclear DNA; they do not share the same mitochondrial DNA, unlike genetically identical twins.

Nuclear transfer technology was first employed in embryo cloning, in which the donor cell is derived from an early embryo, and has been long established in the case of amphibia. However, it was only comparatively recently when McGrath and Solter reported nuclear transplantation in the mouse embryo and paved the way for modern mammalian cloning (see Fulka et al., 1998). Subsequently, nuclear transplantation was conducted successfully in the eggs of domestic animals, including sheep and cows.

Unlike embryo cloning the prospect of cloning adults had seemed remote. During their development from (ultimately) the fertilized oocyte, adult somatic cells undergo an extensive series of cell division and differentiation steps. Until recently it was thought that these processes were accompanied by irreversible modifications to the genome. Even in frogs transplantation of an adult cell nucleus had never been reported to give rise to an adult animal; instead, the renucleated embryos underwent early development but thereafter failed to develop to term. A variety of early experiments in mice were also unsuccessful before the landmark study of Wilmut et al. 1997 reported successful cloning of an adult sheep. For the first time, an adult nucleus had been reprogrammed to become totipotent once more, just like the genetic material in the fertilized oocyte from which the donor cell had ultimately developed.

Figure 21.11
A sheep called Dolly was the outcome of the first successful (more...)

An external file that holds a picture, illustration, etc., usually as some form of binary object. The name of referred object is ch21f11.jpg.

Figure 21.11

A sheep called Dolly was the outcome of the first successful attempt at animal cloning

(A) Experimental strategy used by Wilmut et al. (1997). During culture in serum-depleted medium the normal cell cycle is interrupted and cells move from the G₁ phase of cell growth to a quiescent stage, G₀ where there is no cell division. This quiescent state was thought to facilitate successful nuclear transfer (see text). Note that in the Wilmut et al. (1997) study the nuclear transfer occurred by cell fusion, but in other cases different strategies have been used. For example, in the successful cloning of adult mice reported by Wakayama et al. (1998), a very fine needle was used to take up the donor cell nucleus with minimal contamination by donor cell cytoplasm. The donor cell was quickly, but very gently, microinjected into the enucleated oocyte. (B) Dolly with her first born, Bonnie. Original photo kindly provided by the Roslin Institute.

In the Wilmut et al. study, the donor cells were derived from a cell line established from adult mammary gland cells and were fused to an enucleated metaphase II-arrested oocyte (Figure 21.11A

). The donor cells were deprived of serum before use, forcing them to exit the cell cycle into a quiescent stage, G_o (Stewart, 1997). A certain degree of gene silencing is a characteristic feature of the nuclei of G_o cells. As egg cells are normally fertilized by transcriptionally inactive sperm cells, G_o cells may be more amenable to full genetic reprogramming. Another consideration is the degree of chromosome condensation and of access to chromatin ‘remodeling factors’ such as transcription factors in the oocyte. In any event, the cloning was extremely inefficient: out of a total of 434 oocytes that were submitted to the procedure, only 29 developed to the transferable stage and of these only one developed to term, being born as the now famous Dolly (Figure 21.11B

). Subsequent doubts about the exact origin of the donor cell and whether Dolly really was an adult clone (as opposed to a contaminant fetal cell) have been allayed by genetic testing of Dolly and the adult mammary gland donor cells. Importantly, successful animal cloning has also been achieved by other groups with comparatively high success in cloning of adult mice (Wakayama et al., 1998) and cows (Kato et al., 1998).

21.5.2. The successful cloning of an adult animal has major implications for research, medicine and society

The report by Wilmut et al. (1997) has generated enormous attention, in the scientific and general press, both because of its novelty and the significance for future work. In particular, the possible extrapolation to cloning of humans has generated a great deal of controversy.

Basic research

Successful cloning of adult animals has forced us to accept that genome modifications once considered irreversible can be reversed and that the genomes of adult cells can be reprogammed by factors in the oocyte to make them totipotent once again. Research investigations into the control of gene expression during development and basic processes of somatic differentiation, somatic mutation, aging and repair processes will undoubtedly benefit from animal cloning, especially cloning of mice. Other more recent studies are now forcing us to reconsider the potency of other cells. For example, adult mouse neural stem cells transplanted into an irradiated host animal have very recently been shown to develop into a variety of blood cell types (myeloid, lymphoid and early hematopoietic cells) and so the developmental potential of stem cells is not restricted to the differentiated elements of the tissue in which they reside (Bjornson et al., 1999).

Cloning of livestock and transgenic animals

The successful cloning of adult sheep and cows is clearly attractive to people who wish to perpetuate prized livestock, racehorses, pets and endangered species. In addition, transgenic animals can be cloned. The traditional route for making a transgenic animal is by pronuclear microinjection (Section 21.2.1). But this may be rather inefficient. Transgenic sheep and other livestock have been produced to serve as bioreactors, sources of medically valuable products such as human insulin (see Section 22.1.2). However, in the case of transgenic sheep, for example, only 2–4% of the founder animals born by implanting eggs which have been microinjected with a transgene turn out to be transgenic. Producing founder transgenic animals by nuclear transfer should be more efficient and will allow more sophisticated genetic modifications. An early success was achieved by Dr Wilmut's group who used fetal sheep cells containing a factor IX transgene as donor cells to generate transgenic sheep and this has been followed by cloning of transgenic cattle (Pennisi, 1998).

Human cloning

The most contentious issue in cloning animals is, of course, the potential extrapolation to cloning humans (Shapiro, 1997; Johnson, 1998). Clearly, the technology is still poorly developed and the comparatively high incidence of spontaneous abortions, perinatal losses and anomalous births observed in animal cloning would make the prospect of human cloning unappealing at present. In many countries, existing legislation would also preclude attempts at human cloning. For example, in the UK it is a criminal offence to experiment with human embryos without a licence, which will not be granted under any circumstances for experiments with embryos more than 14 days old.

Technological improvements in animal cloning will undoubtedly occur, however, and if the procedure were eventually to become both efficient and comparatively risk-free, there could be considerable pressure to apply nuclear replacement technology to human cells. Some applications need not involve human reproductive cloning. For example, nuclear replacement could be used to avoid transmission of inherited diseases derived from the mitochondria. Here, an unfertilized egg taken from an individual with mitochondrial disease could act as the donor with the nucleus being transferred into an enucleated egg from a donor containing normal mitochondria. The reconstructed egg could then be fertilized in vitro.

The use of nuclear transfer technology for human reproductive cloning is, inevitably, more contentious. For some infertile couples or women, for example, it could provide a welcome method of having children. However, the expectation that could be placed on such a child could be damaging to that individual because the parent(s) and later the child may be especially conscious of genetic identity between individuals whose ages are quite different. Unlike identical twins whose development proceeds in parallel, for example, a cloned child could be only too aware of how he/she might develop in later life by observing a parent who was essentially genetically identical. Against this, many would argue that a person's character and capability is not determined exclusively be his/her genetic endowment; the environment also has a powerful role to play.