Single nucleotide polymorphisms (SNPs) are variations in DNA sequence that occur when a single nucleotide in the genome is altered. These seemingly small variations can have a major impact on how humans respond to disease, environment, and drugs.

Gene targeting in mice has allowed the analysis of varied aspects of gene function in mammals. During the past decade, thousands of null, hypomorphic, and conditional alleles have been constructed. Gene targeting can also be used to generate point mutations in mice for those genes in which human SNPs have been identified. This approach, however, has not yet been widely used due, in part, to the labor-intensive procedures involved in building the complex targeting vectors required. Recent advances in ‘recombineering’ of bacterial artificial chromosome vectors have streamlined this process (Yu et al, 2000; Lee et al, 2001), making the use of ‘knock-in’ mice a natural progression for the development of mouse models to investigate human disease-related SNPs.

Neuropsychiatric diseases are dependent on multiple genetic and environmental determinants, and thus represent some of the greatest challenges for animal modeling. However, a few genes harboring specific SNPs have emerged as promising candidates. Among these, common SNPs in brain-derived neurotrophic factor (Bdnf), the μ-opioid receptor (Oprm1), and catechol-O-methyltransferase (COMT) have been modeled in mice using three unique approaches.

A common SNP in the BDNF gene (Val66Met) is associated with anatomical (hippocampal volume) and behavioral (performance in memory tasks) impairments in humans. To recapitulate the equivalent variant in mice, we made a point mutation (G196A) to change valine 66 to methionine. In addition to the expected phenotypes of decreased hippocampal volume and impaired context-dependent memory, these mice revealed a novel anxiety phenotype that had not yet been reported in humans (Chen et al, 2006).

A large number of studies have examined the OPRM1 gene as a candidate for genetic contribution to the risk for substance dependence. The best-characterized polymorphism in this gene is a missense mutation in exon 1, involving an A–G substitution at position 118. Owing to the high sequence similarity between mouse and human at the nucleotide (86.9%), and amino-acid level (92.3%), a knock-in mouse was developed that possessed the mouse-equivalent SNP of the human A118G SNP in the murine Oprm1 gene (Mague et al, 2009). In a complimentary approach, a second mouse line for this SNP was generated that expressed humanized receptors with and without the A118G variant (Ramchandani et al, 2010). Both models recapitulated some phenotypes observed in humans, clarified discrepancies regarding functional aspects of the receptor, and identified novel phenotypes.

A third approach to model human SNPs takes advantage of a biochemical phenotype associated with a polymorphism in the COMT gene (COMT-Val), which results in higher protein levels and enzyme activity compared with individuals expressing COMT-Met. However, several other common haplotypes in the COMT gene have been associated with similar biochemical effects. Therefore, to model this phenotype and clarify the specificity of the COMT-Val SNP, was generated a transgenic mouse that overexpressed the COMT-Val gene in the continued presence of the mouse Comt1 gene (Papaleo et al, 2008). Phenotypes related to cognitive and stress reactivity in these transgenic mice were analogous to those reported in humans.

Modeling human SNPs in mice is important for a variety of reasons. In some cases the rationale might be to clarify inconsistencies associated with in vitro data, in others to provide more precise information on the specificity of the SNP, or to explore novel phenotypes. In all cases, these mice now allow for the determination of molecular mechanisms that mediate the behavioral consequences of these SNPs, and as such contribute to a better understanding of their significance in human disease.