The genetic basis of the phenotypic variability observed in patients can be studied in mice by generating disease models through genetic or chemical interventions in many genetic backgrounds where the clinical phenotypes can be assessed and used for genome-wide association studies (GWAS). This is particularly relevant for rare disorders, where patients sharing identical mutations can present with a wide variety of symptoms, but there are not enough number of patients to ensure statistical power of GWAS. Inbred strains are homozygous for each loci, and their single nucleotide polymorphisms catalogs are known and freely available, facilitating the bioinformatics and reducing the costs of the study, since it is not required to genotype every mouse. This kind of approach can be applied to pharmacogenomics studies as well.
- precision medicine
- inbred mouse strains
precision medicine aims at providing tailored interventions for individual patients or as an initial step to subgroups of patients based on their genomics (18). A classical way to approach this problem is by genome-wide association studies (GWAS), which is an unbiased tool for investigating the genetic architecture of diseases. These studies assess common genetic variability between large groups of cases and controls. Additionally, GWAS can be used to determine genetic signatures correlating with phenotypes of groups of patients that present with different disease severity/symptoms (34). This is particularly challenging for rare genetic disorders, where patients do present with a broad spectrum of symptoms, but there are not enough available patients to ensure statistical power of the genomic analysis (14). Therefore, alternative strategies are required: such as mapping in model organisms followed by validation of the candidate modifier genes in a smaller cohort of patients. In this sense, inbred mouse strains are an extraordinary source of genetic diversity that can be used to uncover “modifiers of disease progression.” Inbred strains are generated by mating siblings for at least 20 consecutive generations, and as a consequence they are homozygous in each locus, isogenic (genetically identical), phenotypically uniform, and genetically stable (genetic drift is slow and due only to de novo mutations) (2).
The origins of the classical inbred mouse strains can be tracked to the early 20th century, when William Castle was studying the inheritance of color in mice (6). Castle and his student Clarence Little derived the first inbred strain, the DBA stock, in 1909. Concomitantly Abbie Lathrop began a huge mouse breeding program, which had as many as 11,000 animals, for sale as pets and for research. Many of the strains that we use today, including the C57-derived lines, were originated by Lathrop (2). The classical strains are a valuable source of genetic diversity to study disease outcomes, although they present with a limited range of genetic variation. To overcome this problem novel mouse stocks have been created to better randomize the genotypes of the classical strains. The Collaborative Cross (CC) has eight founder strains including three wild-derived strains to maximize genetic diversity (7). Furthermore, the Diversity Outbred (DO) stock, which is derived from the same eight founder strains, are maintained under randomized breeding conditions designed to ensure a balanced mixture of the founder genomes and avoid allelic loss. Thus, every DO mouse is genetically unique (8). Other stocks include recombinant inbred (RI) strains, which have been created by crossing two classical inbred strains until new inbred strains were generated (28).
Mouse strains have great potential for mapping genes. Thus, significant efforts have been made to catalog dense single nucleotide polymorphism (SNP) maps for commonly used mice strains. Many of the SNPs libraries are freely available at http://www.sanger.ac.uk/sanger/Mouse_SnpViewer/rel-1505 (21); thus the relationship between genotype and phenotype nowadays can be readily explored. To date, genetic studies in mice have identified genes contributing to behavior, physiology, and more (3, 27, 33); however, little progress has been achieved in terms of genetic diseases and pharmacogenetics, although this seems to be changing recently.
Gaucher disease (GD) (OMIM # 230800, 230900, and 231000) represents a prototype monogenetic disease where patients harboring identical genetic mutations can develop a broad spectrum of symptoms. A subset of patients show a very aggressive neurological disease ultimately leading to premature death, while others present with a visceral disease of which they can live for several decades and have a relatively normal life with current treatments. To uncover the genetic basis of phenotypic variability in this disorder, GD was systematically induced in several inbred mice strains by daily injections of conduritol B-epoxide (CBE), an irreversible inhibitor of the β-glucosidase, the enzyme mutated in this disease. This inducible GD model presents with diverse phenotypes depending on the mouse strain, ranging from severe neurological pathology and short life spans to others without evident signs of neurodegeneration and longer survival times, thus successfully recapitulating the human disease phenotypes. Using survival rates a GWAS was carried out and a small collection of candidate loci underlying the variable strain response to CBE was identified, subsequently leading to the uncovering of a combination of SNPs that could successfully predict the life expectancies of other inbred strains upon GD induction. Whether changes in the same genes are able to prognosticate the disease severity in humans remains to be validated. Among the top associated markers were several SNPs in genes associated to glutamate excitotoxicity, including one in the subunit B of the N-methyl-D-aspartate receptor (NMDA) receptor. Importantly, an FDA-approved NMDA receptor antagonist, memantine, delayed the progression of the motor deficiencies in these mice and extended the life expectancies of the short-lived strains by three- to fourfold (22). This opens the possibility that memantine could be used as therapeutic option in patients suffering neuropathic GD, although it has to be carefully evaluated.
This approach has several advantages: 1) development of new animal models to study the biology of the disease. Since the mice tissues are available, changes in pathways between animals with different phenotypes can be explored, facilitating the discovery of therapeutically relevant targets and biomarkers (including microbiome studies); 2) replication of the measurements in genetically identical individuals, minimizing environmental effects; 3) it is not necessary to genotype every mouse since the SNPs profiles are freely available, reducing the costs of the study. Furthermore, the SNP catalogs for inbred strains are dense, allowing for high-resolution genome mapping; 4) this bioinformatics analysis is easier and faster than for wild mice, because inbred mouse strains are homozygous for each loci, although to avoid false associations it is necessary to test strains from different phylogenetic origins and correct for the population structure due to the relatedness origins of the strains. Some limitations of this approach include the large number of strains that must be tested to ensure statistical power and the fixed genetic diversity of inbred strains (10). Other recommendations to avoid false associations include combining GWAS with linkage analysis using classical genetic crosses (25), although it requires genotyping the offspring. Alternative options include inducing the disease in other stocks. The RI strains are ideal to further refine an initial GWAS if the parental strains do show diverse phenotypes after disease induction. The rich genetic diversity of the CC and DO panels can be exploited for this purpose as well. To date, they have been mainly used for studying the genetic basis of viral and bacterial infections, in addition to the response to several toxic agents (11, 23, 30).
Of utmost importance is whether studies in mouse models provide valuable knowledge concerning human pathophysiology. Although mice and humans share the majority of its protein-coding genes (5), they have been under different evolutionary pressures. This has led to discrepancies between the functioning of some systems, such as the innate and adaptive immune responses (26).
In the future this kind of strategy can be applied to other genetic diseases. This is particularly interesting for diseases caused by loss of function mutations in enzymes or transporters or others, which can be achieved by inducing them with chemicals. Some examples are listed on Table 1. Other means include transferring genetic mutations by backcrossing transgenic or knockout mice with inbred strains. In this line, a recent study examined the behavioral responses of F1 crosses between C57BL/6J mice bearing null alleles on genes implicated in psychiatric disorders and 30 classical strains. The phenotypic data dramatically varied between the offspring, in some cases supporting opposite conclusions (31). This report highlights the relevance of genetic background for the reproducibility of results. Generalization of conclusions must be careful limited when only a single strain is considered. With the current advances on genome-editing technologies such as Crisper/Cas9 and others, a mutation of interest can be introduced in multiple strains for precision medicine studies. This type of approach can be used for toxicological and pharmacogenetics studies, as well for analyzing the responses to drugs in the full animal or in cells derived from different strains (12). This strategy can help to further our understanding of why subsets of patients are refractory to certain treatments. Furthermore, before designing clinical assays, loci associated to specific drug responses can be mapped, which could help to identify the correct subcohorts of patients for clinical trials while reducing the number of unsuccessful drugs on their paths to the clinics. To summarize, the rich genetic variability of inbred mouse strains will help us to move forward toward individualized medicine, which is on track to happen sooner rather than later.
The author thanks the Cure & Action for Tay-Sachs (CATS) Foundation and the Telethon Foundation for support.
No conflicts of interest, financial or otherwise, are declared by the author(s).
A.D.K. conceived and designed research; A.D.K. interpreted results of experiments; A.D.K. drafted manuscript; A.D.K. edited and revised manuscript; A.D.K. approved final version of manuscript.
The author thanks Dr. Dustin Bagley for helpful discussions.
- Copyright © 2017 the American Physiological Society