  
|
  ![[help]](/images/buttons/help.gif) |
|
|
|
|
|
P&S Medical Review: Oct 1994, Vol.2, No.1
Genetically Engineered Mice as Models for Human Disease
Jeff D Hardin, Ph. D.
Columbia University College of Physicians and Surgens, New York, NY
Introduction
The development of recombinant DNA technology has allowed the introduction of exogenous genetic material first into bacteria, then eukaryotic cells in culture, and most recently whole animals.1,2 These advances have revolutionized the study of genes and have given birth to the exciting new technology of genetically engineered animals. The majority of this work has been performed in the mouse because this animal is small and easy to manipulate and maintain. This has led to the development of many of the reagents and techniques needed to do genetic engineering (e.g., micromanipulation and injection of embryos and embryonic stem cells). Moreover, the mouse is usually sufficiently biologically similar to humans to provide valid models of human disease.
This review describes the methodology involved in generating genetically modified mice, gives several examples of diseases that have been reproduced in engineered mice, and argues for the value of this technology in understanding and treating human illness. In general, of the several diseases that have been modeled so far, most bear striking resemblance to their human counterpart. However, due to differences between the human and mouse species, and the type of genetic alteration engineered, some mouse strains have limitations to their suitability as models of human disease (see cystic fibrosis, below). These differences can at times limit the usefulness of the animal models, but the models still can be used to understand mechanisms of resistance to disease in mice. These results can then be applied to design treatments for the human disease counterpart. Other animals (i.e., rats, rabbits, and pigs) have been used for modeling disease but they will not be discussed here.
Construction of Genetically Modified Mice
There are currently two methods used to genetically engineer mice: 1) conventional transgenic mice and 2) gene-targeted, or gene-disrupted mice (Figure 1). In the former, foreign genes are introduced into the mouse chromosomes by microinjection of DNA into single-cell zygotes.3 The DNA integrates randomly in the mouse chromosome, resulting in one or multiple copies of the inserted gene. This technique usually results in a mouse that has a new gene (e.g., a human gene) or that has an altered expression pattern of an existing mouse gene (e.g., increased levels of gene product or expression in different tissue types). As such, transgenic mice are usually "gain of function" mutants in that they dominantly express a new function or trait.
Gene-targeted mice are technically more demanding to generate and require the production of totipotent embryonic stem (ES) cells with a designed alteration of a particular gene present in the mouse chromosomes.4 The process relies on homologous recombination, a type of DNA rearrangement whereby two similar sequences of DNA trade places, to target the mutation to precisely the chromosome region of interest. This differs from the transgenic technique which generates random chromosomal integration sites and gene copy numbers. The correctly gene-targeted ES cells are microinjected into embryos to produce mice that are chimeric for both the normal and mutated cells. Next, the chimeric mice are cross-bred to generate mice homozygous for the genetic change. This is dependent on the presence of the genetic change in the chimera's gonads and usually requires two or more generations of breeding. The chimeric mice are first bred with wild-type mice to generate F1 (first-generation) mice which are heterozygous for the mutation. The F1 heterozygotes are then bred to each other to generate homozygotes (F2) at Mendelian frequencies. In gene-targeted mice the mutation is often a disruption of the gene, and hence, a "loss of function" mutation in that the gene no longer functions like its wild-type counterpart. Therefore, these gene-targeted mice usually express recessive traits.
Both transgenic and gene-targeted mice require the generation of DNA constructs (i.e., pieces of engineered DNA carrying entire genes or parts of the genes being studied) and the microscopic manipulation of early mouse embryos. Southern blotting and/or the polymerase chain reaction (PCR) is used to verify the presence of the genetic alteration in the resulting mice. After a new strain of mouse is generated by one of the two techniques, its suitability as a model of human disease is assessed using clinical and pathological laboratory techniques. So far, several hundred targeted and transgenic strains have been made and analyzed. Table 1 gives a partial list of existing models.
Cystic Fibrosis
Cystic fibrosis (CF) is an autosomal recessive disease that affects 1 in 2000 births in Caucasians and is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. The gene was cloned in 1989 and was found to code for a cyclic AMP (cAMP) activated chloride channel present in the apical membrane of certain secretory epithelial cells.5 In CF patients, mutations in the gene result in ion and water imbalances in several secretory tissues. The net effect is the accumulation of thickened secretions in the lung, the exocrine pancreas, the hepatobiliary system, and in reproductive epithelia. These abnormal secretions lead to the characteristic pathology of CF - pulmonary obstructive disease and infections, pancreatic insufficiency, infantile meconium ileus, and biliary stones and obstruction. Attempts to pharmacologically correct the molecular defect underlying cystic fibrosis have so far been limited, and CF patients have an average life expectancy of 30 years.
Four groups have reported the generation and characterization of mice that have targeted mutations in their CFTR genes.6-9 The gene-targeted mice have mutations in different regions of the CFTR gene and display varying degrees of disease. The CF mouse strains have a high post-natal mortality (i.e. in one strain, most die within 30 days after birth).9 This is caused by an obstructional meconium ileus that results in bowel rupture and peritonitis. In contrast, only 5-10% of CF patients suffer from this complication. In addition, CF mice have ion permeability defects in airway and respiratory epithelia and have the characteristic histopathology of CF in respiratory, intestinal, hepatobiliary, and male reproductive epithelia, yet have little dysfunction in these organs. Most humans with CF, however, have gross morbidity in these organs and eventually succumb to recurrent and severe pulmonary infections, which have not been observed in the CF mice. The differences in cystic fibrosis between mice and humans may be a consequence of species variation in the importance of the CFTR gene product and of the relative importance of a second chloride channel that may adequately compensate for the CFTR mutation in mice.
 |
| Figure 1. Transgenic and gene-targeted ("knock-out") mice are made by the in vitro manipulation of normal mouse development (1). Gene-targeted mice (2) are produced by several steps: i) the gene under study is mutated in totipotent ES cells using homologous recombination; ii) these altered ES cells are microinjected into a mouse embryo that is at the blastocyst stage; iii) the embryo is returned to a female mouse and develops to term; iv) the resulting mouse is a chimera of normal and altered cells and is directly studied or cross-bred to produce a mouse that is homozygous for the introduced mutation. Transgenic mice (3) are produced by microinjecting copies of the transgenic construct DNA into fertilized eggs (zygotes). The altered eggs are placed back into a female mouse where they resume development and are subsequently studied prenatally or after birth. |
Despite these differences, the animals are already proving useful in the study of the disease and the development of treatment strategies. CF mice provide an unlimited supply of live tissue with which to study the underlying pathophysiology of the defect.10 For example, the evaluation of drugs such as the secretagogue amiloride can be effected more safely and inexpensively in the CF mice.
Perhaps the greatest use of this new tool in the study of CF has come from two recent reports of successful gene therapy in the CF mice.11,12 Both groups developed liposomes carrying copies of a normal CFTR gene that was introduced into the respiratory and gastrointestinal epithelia of the CF mice. In both cases, the CF defect was safely reversed. These studies showed that even low levels of CFTR transgene expression can correct the ion transport deficiency in the lungs and intestines of CF mice, and are serving as the foundation for human trials of CF gene therapy.
| Disease Mode | Gene | Technique | Reference |
|---|---|---|---|
| Cystic Fibrosis | CFTR | Gene-Targeted | 6-9 |
| Atherosclerosis | Apo E, apo (a), Apo A-II | Gene-targeted, Transgentic | 21,22,24,36 |
| anti-Atherosclerosis Gene Therapy | Apo AI, Ape E, LDLR | Transgenic | 23,27,37 |
| B-Thalassemia | ß -globin | Gene-Targeted | 28 |
| Sickle Cell Anemia | ßs (and variants) | Transgenic | 28 |
| Inflammatory Bowel Disease | Interleukine-2, Interleukin-10 and T-cell Receptor ,ß ; MHC II | Gene-Targeted | 14-17 |
| Severe Combined Immunodeficiency Disease | Rag-1, Rag-2 | Gene-Targeted | 38,39 |
| Muscluar dystrophy Gene Theraphy | Dystrophin | Transgenic | 40 |
| Alzheimers disease | ß -amyloid | Transgenic | 41,42 |
| Amyotrophic lateral sclerosis (ALS) | neurofilament heavy chain | Transgenic | 43 |
| Insulin Dependent Diabetes Mellitus | interferon- | Transgenic | 44 |
| Cancer | many oncogenes and tumor supressor genes | Transgenic and Gene-Targeted | 45 |
Inflammatory Bowel Disease
Recently, the targeted disruption of several genes coding for proteins of the mouse immune system has led to an unexpected breakthrough in understanding the etiology of inflammatory bowel disease (IBD). IBD in humans is a chronic process involving patchy inflammation throughout the entire gastrointestinal system in Crohn's disease, and continuous inflammation limited to the colon in ulcerative colitis. These two diseases are believed to have an autoimmune etiology associated both with genetic and environmental influences.13 Mice with disruptions of the T cell receptor (TCR), the major histocompatibility complex (MHC) type II, the interleukin 2 (IL-2), or the interleukin 10 (IL-10) genes all acquire an age-dependent disease that is clinically and histologically similar to the two forms of human IBD.14-16 This work is surprising in that it demonstrates that mutations in genes having diverse functions in the immune system can lead to a common defect, and that despite the ubiquitous function of the immune system throughout the body, only the gut manifests disease. These findings have generated the proposal that a mutation in any of several different molecules involved in the immune system can lead to a loss of tolerance to the foreign antigens present in the gut (food and microorganisms). The result is a local invasion of gut mucosa by inflammatory cells and the production of a chronic immune response that destroys the epithelium. Thus, different genetic alterations under the influence of environmental factors leads to a final common pathway of autoimmune destruction of the intestinal epithelium.17
In humans, several genetic loci associated with IBD have been identified, but the genes that cause the disease are still unknown.18 It remains to be determined if subsets of patients with IBD have mutations in the immune system genes known to cause the mouse disease. The IBD mice have serendipitously contributed to the understanding of a long-studied autoimmune disease, providing a model to investigate the disease mechanisms and therapeutic options. Current therapy in IBD utilizes non-specific anti-inflammatory and immunosuppressive agents that are only partially effective and have serious side effects. However, the disease has been cured in IL-10 gene-targeted mice by the intravenous administration of IL-10.15 Also, the disease was markedly reduced or eliminated by removing gut organisms in IL-2 and IL-10 targeted mice, giving hope that similar modification of flora or diet will be beneficial in IBD patients.15,16 Further studies on the IBD mice should help reveal the complex immune interactions underlying the disease and provide novel treatments.
Atherosclerosis
In developed countries, the leading cause of mortality is associated with the clinical manifestations of atherosclerosis: heart disease and stroke. Epidemiological and molecular biological studies show that an individual's genetic predisposition, combined with environmental factors (chiefly diet and smoking), influence the plasma lipid levels which correlate with atherosclerosis.19 Plasma lipid is processed by a series of lipoproteins and their receptors present on liver and endothelial cells. Mutations within some of the genes coding for these proteins cause some rare heritable diseases characterized by increased levels of serum cholesterol and a predisposition to early development of atherosclerotic lesions, culminating in an early death.20 It is believed, that in persons without disease, the lipid processing machinery can be overloaded by the high fat content of the Western diet, resulting in accelerated lipid deposition in arteries. To facilitate the understanding of the role of lipid transport in atherosclerosis, several strains of mice have been engineered to have altered levels of the proteins involved in these processes (Table 1).
Apolipoprotein E (Apo E) is found in several lipid particles and is important in the post-prandial clearance of dietary lipids. This protein, present in chylomicrons, functions as a ligand for the low-density lipoprotein (LDL) receptor on the surface of liver cells. In humans, a mutation or absence of Apo E results in Type III hyperlipoproteinemia, characterized by increased plasma triglycerides and cholesterol, early onset atherosclerosis, and skin xanthomas.20 Apo E gene-targeted mice have been generated and have an 800% increase in plasma cholesterol as compared to control mice on the same low fat diet.21,22 Within 3 months the Apo E disrupted mice develop atherosclerotic lesions in the proximal aorta and coronary and pulmonary arteries. At 8 months, complete coronary artery occlusion can be observed. These mice are proving to be valuable tools in understanding the connection between lipid levels and atherogenesis, and are helping to develop ways to modulate hypercholesterolemia in patients. The effect of drugs and gene therapy on the disease present in the Apo E deficient mice is being evaluated. Mice transgenic for the human Apo E gene, for example, have significantly decreased levels of serum cholesterol compared to Apo E gene-targeted or normal mice.23
Even on a high fat diet most strains of mice are resistant to atherosclerosis. Mice normally lack apolipoprotein (a) [apo (a)], a subunit of the LDL particle present in humans that is associated with an increased risk for atherosclerosis. However, transgenic mice carrying the human apo (a) gene develop atherosclerotic lesions within 3 months.24 The apo (a) mice are being used to develop drugs that lower serum apo (a) levels by interfering with transcription of the gene. Nicotinic acid, in combination with neomycin, decreases apo (a) but has unacceptable side effects. Safer, more specific transcription inhibitors and apo (a) anti-sense drugs may selectively lower apo (a) levels and can be efficiently evaluated in the apo (a) mice.
The LDL receptor (LDLR) has been shown to be an essential factor in determining lipid clearance in humans by its feedback on the rate-determining enzyme of the cholesterol synthesis pathway, HMG CoA reductase.25 Mutations in the LDLR gene in humans result in familial hypercholesterolemia, a disease which is characterized by increased morbidity and mortality due to accelerated atherosclerosis.26 Transgenic mice that have increased levels of LDLR on their livers are protected from the atherosclerotic effects of a high fat diet.27 Studies using these mice support the role of LDLR gene therapy in the reduction of human atherosclerosis. Taken together, the mouse models of atherosclerosis demonstrate that modification of one of several genes can lead to altered levels of serum lipids and, ultimately, atherosclerotic disease. Although variation in the inherent susceptibility to atherosclerosis between mice and humans may limit the applicability of new discoveries, the study of these modified mice promises further understanding of lipid regulation.
The hemoglobinopathies
Hemoglobin is a tetramer of two a-globin-like and two b-globin-like chains. In the embryo and adult there is a regulated pattern of expression of globin genes giving rise to hemoglobin tetramers with different biochemical properties important to the function of hemoglobin at that developmental stage. In adult mice there are two -globin genes: major and minor accounting for 80% and 20% of -globin levels, respectively. Adult humans, on the other hand, have only one -globin gene giving rise to the major hemoglobin type called hemoglobin A (HbA) but 3% of hemoglobin has the -globin gene product and is called hemoglobin A2 (HbA2).
Thalassemia refers to a group of heritable diseases present in people of Mediterranean and African descent. These diseases stem from mutations that cause the reduction or absence of -globin in red blood cells (RBCs). This results in -globin aggregation that disrupts the RBC membrane and causes hemolysis. Hallmarks of this disease include a chronic microcytic anemia, intramedullary hemolysis, extramedullary hematopoiesis, iron overload and resultant organ damage.
b-major gene-targeted mice have a severe form of thalassemia and all die perinatally.28 A naturally occurring mouse strain exists that carries a different mutation in the b-major gene.29 These animals, however, display a milder form of disease due to a compensatory increase of b minor gene expression by 3-4 fold. This increase in b minor is not seen in the gene-targeted mice. Comparison of the two strains of mice has led to the proposal that promoter competition between major and minor genes determines the severity of disease. This model promises the development of drugs that may modulate promoter competition and increase the levels of minor.
Since normal hemoglobin function depends on a 1 to 1 ratio of - and -globins, an alternate approach to ameliorating the severity of thalassemia is to decrease the levels of -globin to be equivalent to the reduced -globin levels. Accordingly, mice with both the -globin and -globin thalassemia genes are healthier than mice with only thalassemia or thalassemia. This approach would not work in patients with severe thalassemia who make little or no -globin and hence would have insufficient levels of the hemoglobin tetramer.
Drugs such as 5-azacytidine, hydroxyurea, and sodium butyrate function by increasing the levels of fetal hemoglobin (Hb F) and are useful in some thalassemia patients. The thalassemia mice models should be useful in further understanding the mechanism of these drugs and in developing new agents that modulate globin gene expression, ultimately providing a model for gene therapy.
Several groups have modeled sickle cell anemia (SCA) in transgenic mice.30-34 In humans, this disease results from a heritable point mutation that causes the substitution of valine for glutamic acid in -globin. This s-globin gene codes for an abnormal protein (called Hb S) that polymerizes, causing sickling of RBCs in the deoxygenated environment of many of the body's tissues. Sickling results in a sequela of chronic anemia as well as painful vaso-occlusive crises caused by small vessel occlusion. Typically, these events occur in muscle, lung (acute chest), kidney, joints (avascular necrosis), spleen (splenic sequestration), and brain (stroke).
The different SCA mice carry transgenes with the human globin locus control region (LCR) and various human and s genes. The transgenes are placed in a mouse strain that has reduced levels of its own normal -globin (i.e., the mouse thalassemia strain). These SCA models express many aspects of the human disease but require hypoxic conditions to achieve sickling and organ damage,35 a difference that may be due to the low levels of minor present in the SCA mice (analogous to sickle/ thalassemia (S/+) patients who have one copy of the s globin gene and one copy of a thalassemia gene and who have a milder disease than SCA patients). Alternatively, species differences such as RBC size and cell membrane properties may provide resistance to sickling in the mouse. The limitation posed by this species-specific restricted phenotype may be overcome by the generation of strains of mice that have their own - and -globin genes disrupted but have human - and s-globin transgenes. These animals may better mimic human SCA disease and serve as more effective tools in the endeavor to understand and treat SCA.
Conclusions
The use of mouse models is rapidly revealing the role of genes in the development and homeostasis of animals and is altering our view of biology. For each new gene cloned, it has become almost routine to further study the gene by producing transgenic and ES chimeric mice. Today, the field of medicine is poised to reap the most practical benefit from this new technology. The secrets of enigmatic diseases such as IBD, CF, atherosclerosis, and the hemoglobinopathies are slowly yielding to the scrutiny afforded by animal modeling. Many mouse strains are becoming commercially available, allowing access by researchers who lack the resources to generate their own engineered animals. Genetically engineered mice offer an inexpensive, safe, and powerful resource for studying the pathophysiology of disease, and for designing and evaluating novel therapeutic interventions. Indeed, as the molecular basis of many diseases is revealed, these designer mice hold the promise of a bold revolution in the understanding and treatment of human illness. However, as greater numbers of engineered mice are produced, the sea of information generated from the models runs the risk of becoming overwhelming and confusing. The issue is further complicated by the biological differences, some subtle and some extreme, between humans and mice. To be sure, the nascent field of mouse genetic modeling is off to a hopeful beginning, holding many promises for the advancement of knowledge and treatment of illness. But, now more than ever, planning and foresight are required.
References