“Fierce” mice, homozygous for the deletion of nuclear receptor 2E1 (NR2E1), show abnormal brain-eye development and pathological aggression. To evaluate functional equivalency between mouse and human NR2E1, we generated mice transgenic for a genomic clone spanning the human NR2E1 locus and bred these animals to fierce mice deleted for the corresponding mouse gene. In fierce mutants carrying human NR2E1, structural brain defects were eliminated and eye abnormalities ameliorated. Excitingly, behavior in these “rescue” mice was indistinguishable from controls. Because no artificial promoter was used to drive transgene expression, promoter and regulatory elements within the human NR2E1 clone are functional in mouse. Normal behavior in rescue animals suggests that mechanisms underlying the behavioral abnormalities in fierce mice may also be conserved in humans. Our data support the hypothesis that variation at NR2E1 may contribute to human behavioral disorders. Use of this rescue paradigm with other genes will permit the direct evaluation of human genes hypothesized to play a causal role in psychiatric disease but for which evidence is lacking or equivocal.
The study of mental illnesses is exceedingly difficult. Despite strong evidence for genetic modulation of disease risk (Tandon and McGuffin, 2002), positional cloning has shown only limited success in the identification of psychiatric disease genes. Genetic studies in humans are hampered by the incomplete penetrance of disease features and extensive phenotypic variability among affect individuals, resulting in part from polygenic inheritance (Owen et al., 2004) and gene-environment interactions (Caspi et al., 2002). Moreover, genetic heterogeneity between individuals undermines both population-based strategies and the replication of results by independent groups. Mouse genetics can address these concerns, ensuring near-identical environments and genetic make-up between subjects. Determining the subset of observations from mice that generalize to humans, however, remains challenging because of differences between the species. From this perspective, we reasoned that establishing a paradigm in which one could functionally evaluate the ability of a human gene to shape behavior and simultaneously retain the advantages associated with a model system would be powerful. Toward this end, we generated transgenic mice expressing the human form [Tg(Hum)] of orphan nuclear receptor 2E1 (NR2E1; previously Mtll, Tailless, Tll, and Tlx) and evaluated the ability of this human gene to modulate the behavior of mutant mice, which in its absence would have demonstrated pathological violence.
Nr2e1 null mice show hypoplasia of the cortex, olfactory bulb, and retina. Structural abnormalities in mutants are caused by abnormal neuronal proliferation during development as opposed to increased cell death (Stenman et al., 2003b; Miyawaki et al., 2004; Roy et al., 2004). Although the aggression observed in “fierce” is complex, the term pathological appropriately captures several important features. For one, the abnormal behavior is directly attributable to the loss of Nr2e1 (Monaghan et al., 1997; Young et al., 2002). Second, the aggression is habitual, starting at ∼5 weeks and persisting through adulthood (Young et al., 2002). Third, the aggression is maladaptive, with the males often killing their intended mate (Young et al., 2002). Mechanisms underlying behavioral abnormalities remain elusive but may involve developmental abnormalities of the amygdala (Stenman et al., 2003a), olfactory bulb (Monaghan et al., 1997; Yu et al., 2000; Young et al., 2002), or loss of Nr2e1 in adulthood. Independent genome scans for bipolar disorder (Dick et al., 2003; Middleton et al., 2004) find significant linkage with the specific interval to which NR2E1 maps (human 6q21-22).
Our results show that the presence of a transgene spanning human NR2E1 is sufficient to eliminate structural brain defects, ameliorate eye abnormalities, and restore normal behavior to fierce mice. Because no artificial promoter was used to drive transgene expression, promoter and regulatory elements within human NR2E1 are functional in mouse. Along with amelioration of developmental abnormalities, demonstration of human NR2E1 in the brains of transgenic animals supports appropriate transcription by the human transgene. The correction of fierce brain-behavior abnormalities by human NR2E1 supports conserved underlying mechanisms for behavior modulation. Moreover, use of this same approach with other genes may prove useful in the genetic dissection of discrete features or behaviors that contribute to psychiatric disease (endophenotypes) as well as the systematic evaluation of candidate genes for which involvement in disease is controversial.
Materials and Methods
Transgenesis. Pronuclear injection of a 141 kb PAC clone spanning human NR2E1 (pacEMS1; clone identification number, dJ429G5; GenBank accession number AL078596.7) was performed as described previously (Abrahams et al., 2003) to produce eight transgenic founders from which five independent strains were established successfully. Two strains were used here [C57BL/6J.Cg-Tg(NR2E1pacEMS1B)10Ems and C57BL/6J.Cg-Tg(NR2E1pacEMS1D)11Ems] with mice at backcross generation 6 (N6) (98.4% C57BL/6J) to N10 (99.9% C57BL/6J) before use. These strains will be represented by C57BL/6J [+/+, Tg(Hum)/+], or Tg(Hum). No differences between the two characterized lines, either before or after having been bred to fierce (see below), were observed for animals of any genotype. Moreover, similar results were obtained when the two strains were analyzed either separately or together. For simplicity, data reported here have been pooled across the two strains.
Breeding. Two strains bearing the fierce mutation were used in the breeding strategy (described below) with the transgenics (described above), C57BL/6J.Cg-Nr2e1frc (C57BL/6J; fierce/+, +/+) and 129S1/SvImJ.Cg-Nr2e1frc (129S1/SvImJ; fierce/+, +/+). Both strains had been backcrossed >10 times (to C57BL/6J and 129S1/SvImJ, respectively) and thus were >99.9% inbred before use. Because males homozygous for Nr2e1frc (referred to as fierce) kill their intended mates and homozygous fierce females are poor mothers (Young et al., 2002), experimental mice were obtained by a two-generation mating scheme involving the pairing of Nr2e1frc heterozygotes. C57BL/6J (fierce/+, +/+) mice were crossed to C57BL/6J [+/+, Tg(Hum)/+] mice to obtain C57BL/6J [frc/+, Tg(Hum)/+] females. These females were then crossed to 129S1/SvImJ (fierce/+, +/+) males to produce experimental animals. Thus, experimental mice were first generation C57BL/6J × 129S1/SvImJ (B6129F1) offspring. Production of F1s ensured (1) large litters (Banks et al., 2003), increasing yield per mating and as a consequence improving ability to control for gene-environment effects, (2) hybrid vigor, although hydrocephalus is frequent among B6 fierce homozygotes, for example, the same is not true of F1s (Young et al., 2002), and (3) genetically matched experimental animals, inheriting one B6 and one 129 chromosome, thereby differing only at the loci under investigation (Silva et al., 1997). Adult mice (12-24 weeks of age) were used for all analyses.
Genotyping. Three separate PCR assays were used to genotype each individual. oEMS650 (5′-GGCGGAGGGAGCTTAAATAG-3′) and oEMS1368 (5′-GATTCATCCTATTCCACAAAGTCA-3′) span the fierce deletion region and give a product only in its absence. oEMS1859 (5′-CTGGGCCCTGCAGATACTC-3′) and oEMS1860 (5′-GGTGGCATGATGGGTAACTC-3′) detect mouse but not human NR2E1. oEMS800 (5′-CCCAGCAGCTGCGGTTTTGC-3′) and oEMS801 (5′-GCAGCGCTCCAGGCAGGAC-3′) detect human but not mouse NR2E1. Reactions were put through 30 cycles of the PCR as described previously (Banks et al., 2003) but with the addition of 10% DMSO for oEMS1859/60 and 800/801.
Histology. Entire brains were dissected after intracardial perfusion of Avertin-anesthetized mice with formalin. Sectioned brains were prepared from a separate set of animals, killed by cervical dislocation. Brains were rapidly removed, embedded in OCT (VWR Scientific, Delta, British Columbia, Canada), and frozen on dry ice. Cryosections were prepared at 15 μm, then stained with cresyl violet and luxol fast blue.
Funduscopy. Direct funduscopy was performed between 18 and 24 weeks with a Kowa Genesis small animal fundus camera (Pacific Medical, Delta, British Columbia, Canada) in conjunction with a Volk 90D lens (Topcon Canada, Calgary, Alberta, Canada) as described previously (Hawes et al., 1999). Eyes were dilated with 25% atropine 30 min before examination, at which time mice were lightly sedated with Avertin.
Behavior testing. Mice were housed under a reverse light/dark cycle and tested in the dark (Hossain et al., 2004). Because of the aggressive nature of the fierce animals, approved procedures require they be housed singly. Thus, to control for the possible confound of environmental effects (Van Loo et al., 2001) and gene-environment interactions (Ouagazzal et al., 2003) on behavior, all experimental mice [wild type (Wt), fierce, Tg(Hum), and rescue] were housed individually from weaning at 18 d until they were killed for histological analyses in adulthood.
Struggle was measured in males and females at 13 weeks for 3 min using a PHM-300TSS mouse tail suspension system (Med Associates, St. Albans, VT). The apparatus was calibrated to offset animal weight before testing, and system settings for struggle and gain were 15 and 4, respectively. Percentage of time struggling was calculated by expressing the number of milliseconds during which force exceeded the struggle threshold (set to 15) as a proportion of test time. Force-exerted struggling was calculated by summing the force applied to the system across the testing period. Significant correlations (ranging from 0.96 to 0.97) between experienced human raters and the device we used have been reported recently (Crowley et al., 2004).
Spontaneous activity was measured in mice of both sexes at 12 weeks using a 3 min open-field paradigm. The apparatus we used (Med Associates) was Digiscan photocell-equipped to permit automated data capture. Ambulatory counts equaled the number of beam breaks recorded during the test period. Ambulatory time represented the amount of time (in seconds) spent in motion during the test period. Number of ambulatory episodes was calculated by summing the number of individual episodes (defined as a series of sequential ambulatory counts separated in time by no more than 500 ms).
A 4-10 min resident-intruder paradigm (Young et al., 2002) was used to assess aggression in 16- to 18-week-old experimental males. Wild-type mice matched to experimental animals for genetic background, sex, and weight served as intruders. Video analysis software (Observer Video Pro; Noldus Information Technology, Leesburg, VA) was used to score frequency and duration of tail rattling (rapid quivering or thrashing of the tail), wrestling (close contact that escalates to rolls and tumbling), attack (biting of the opponent mouse), sniffing (sniffing the head or snout of the partner), and anogenital investigation (exploration of the intruder's hindquarters). Tail rattling, wrestling, and attack behaviors were combined to give an index of aggression. Aggression per minute and percentage of time aggressive were calculated by expressing the number of aggressive behaviors and time engaged in aggressive behaviors over minutes of testing and as a proportion of total test time, respectively. Sniffing and anogenital investigation were combined to give an index of social behavior. Percentage of time social was calculated by expressing the amount of time engaged in social behaviors as a proportion of total test time.
Statistical analyses. All analyses were performed using JMP 4.0 (SAS Institute, Cary, NC). Continuous data were analyzed by group for normality and equal variances, then compared by parametric (ANOVA for genotype followed by Tukey-Kramer test for pair-wise comparisons when p < 0.05) or nonparametric statistics (Wilcoxon/Kruskal-Wallis test) as necessary. Significant Wilcoxon/Kruskal-Wallis results (p < 0.05) were attributed to a single group when results became nonsignificant (p > 0.05), after the removal of that group from the original data set. Nominal data (radial asymmetry and mottling) were analyzed using the Pearson χ2 test, with p < 0.05 taken as the cutoff for significance.
Reverse transcriptase-PCR. Total RNA was isolated from adult mouse brain using TriZol (Invitrogen, Burlington, Ontario, Canada) or in the case of fetal human brain, purchased from Origene (Rockville, MD). Five micrograms of DNaseI-treated RNA were subsequently concentrated using MicronYM-100 columns (Millipore, Nepean, Ontario, Canada) and used to generate oligo-dT primed cDNA with Superscript II according to the manufacturer's instructions (Invitrogen). A reverse transcriptase-negative (RT-) control was generated in parallel to each cDNA produced. Intron-spanning PCR assays specific for NR2E1 in human or mouse were developed. oEMS1633 (5′-GCATGAATACTGACAACACAGAC-3′) and oEMS1637 (5′-GCTAATTGACCGTAAAGCTGGT-3′) gave a 324 bp product only when mouse Nr2e1 was present. oEMS749 (5′-TTCCTGAAGGCTACACATTCC-3′) and oEMS825 (5′-AGAGGTGGTGGCTCGATTTA-3′) gave a 475 bp product positive only in the presence of human NR2E1. Each primer set was carried through two rounds of PCR, whereby 2 μl from a 25-cycle first reaction was used as template for a 30-cycle second reaction. An additional primer set against human β-actin was used to confirm the presence and absence of cDNA in all RT+ and RT- samples, respectively, across 30 cycles of PCR. oEMS1646 (5′-ATTGGCAATGAGCGGTTCCGC-3′) and oEMS1647 (5′-CTCCTGCTTGCTGATCCACATC-3′) give a 336 bp product against either mouse or human cDNA. Parameters for denaturation (94°C for 30 s), annealing (55°C for 60 s), and extension (72°C for 60 s) were kept constant across all assays. Each assay was repeated in a minimum of four animals from each genotype.
To test the functionality of human NR2E1 in mice, we established transgenic mice harboring a genomic DNA clone spanning the human NR2E1 locus. We then crossed transcription-positive strains (data not shown) to fierce mice shown previously to carry a deletion that spans Nr2e1 but leaves neighboring genes intact and transcriptionally active (Kumar et al., 2004). Previously identified abnormalities that make up core features of Nr2e1 mutants were then used to make blinded comparisons between mice from each of the following four groups: wild-type (+/+, +/+), fierce (fierce/fierce, +/+), transgenic [+/+, Tg(Hum)/+], and rescue [fierce/fierce, Tg(Hum)/+].
Human NR2E1 corrects structural brain abnormalities seen in the fierce mouse
Multiple brain abnormalities in fierce but not transgenic or rescue mice (Fig. 1) were present at the gross and microscopic level. Compared with wild-type controls, fierce brains showed gross hypoplasia of the olfactory bulbs (Fig. 1a, arrow) and anterior cortex, leaving midbrain colliculi exposed (Fig. 1a, between dashed lines). Quantitative analysis of surface areas for the olfactory bulbs (Fig. 1b) and cortex (Fig. 1c) confirmed significant reductions in both regions (49.1 and 38.9%, respectively) for fierce relative to each of the other genotypes. In contrast, brains of wild-type, transgenic, and rescue mice were phenotypically indistinguishable.
Human NR2E1 corrects abnormalities observed in sections from fierce forebrain
We observed characteristic abnormalities in sectioned brains from every fierce mouse examined but no abnormalities in any transgenic or rescue animals (Fig. 2). Sections through the fierce forebrain showed an overall reduction in size relative to wild-type controls (Fig. 2a). An unusually shaped cingulum, poorly defined piriform cortex, and smaller anterior commissure were also evident in fierce. At higher magnification, abnormal cortical lamination was noted (Fig. 2b), consistent with the previous observation (Land and Monaghan, 2003) that Nr2e1 is required for proper formation of layers II and III. A reduction in striatal volume was also observed in fierce (Fig. 2c) as a result of misspecification of the ventral lateral ganglionic eminence from which it develops (Stenman et al., 2003b). Fierce mice showed an enlargement of corticostriatal fibers that pass through the striatum (Fig. 2c, arrows), with clustering toward its medial border. In addition, the external capsule of the corpus callosum was thin in fierce relative to wild type (Fig. 2c).
Human NR2E1 ameliorates multiple aspects of the fierce eye
Abnormalities observed in the fierce eye by funduscopy were each corrected or ameliorated in transgenic and rescue animals (Fig. 3a). Asymmetry of the radial vasculature, mottling of the retinal pigment epithelium (Fig. 3a, arrow), and a reduction in retinal vessel number were characteristic features of the fierce eye. All fierce mice showed abnormal radial symmetry, and thus radial symmetry was significantly reduced relative to wild-type, transgenic, and rescue mice (Fig. 3b). Similarly, only fierce showed mottling of the retinal pigment epithelium (Fig. 3c). Vessel number was lowest in fierce and significantly reduced relative to rescue mice, which showed the next lowest vessel number (Fig. 3d, dagger). However, for vessel number only, rescue animals were reduced significantly relative to wild-type and transgenic mice (Fig. 3d, double dagger). Nevertheless, because rescue eyes were essentially normal both by qualitative and quantitative analyses, we conclude that the presence of human NR2E1 had a corrective influence on the structural development of the fierce eye. The slight reduction in vessel number may reflect the observation by others that development of the mouse eye is exquisitely sensitive to changes in gene dosage (Schedl et al., 1996; Brown et al., 1998; Chang et al., 2001).
Human NR2E1 corrects behavior abnormalities in the fierce mouse
Multiple behavior tests identified fierce mice as distinct from the three other groups, whereas no differences were found between wild-type, transgenic, and rescue animals. We used a tail-suspension test to quantify the hard-to-handle phenotype observed previously in fierce mice (Monaghan et al., 1997; Young et al., 2002).
In fierce mice, percentage of time struggling and force exerted struggling were significantly increased relative to mice of other genotypes (Fig. 4a,b). Similarly, fierce showed significant increases in spontaneous activity relative to the other groups as measured by ambulatory counts (Fig. 4c), ambulatory time (Fig. 4d), and ambulatory episodes (Fig. 4e). Additional measures obtained during this test also showed significant differences between fierce and all other genotypes (increased ambulatory distance and decreased resting time; data not shown). Fierce mice showed significantly heightened aggression in the resident-intruder test relative to the other genotypes (Fig. 4f,g). The proportion of time fierce mice engaged in social behavior was reduced relative to mice from each of the other genotypes (Fig. 4h). Most strikingly, unlike fierce males who attack, wound, and kill female mates (Young et al., 2002), male rescue animals were not aggressive toward female partners and mated successfully. Together, these results indicate that the presence of human NR2E1 renders the behavior of fierce mice indistinguishable from that of wild-type controls.
Human NR2E1 message is present in the brains of adult transgenic and rescue mice
NR2E1 is normally expressed during embryogenesis and in the adult brain of mice and humans. Thus, although embryonic expression would be essential, the question arose whether rescue of the fierce phenotype had occurred in the absence or presence of adult expression of the human transgene. cDNA prepared from human brain and the brains of adult mice corresponding to each of the genotypes under investigation were assayed for the presence of mouse Nr2e1 (Fig. 5a), human NR2E1 (Fig. 5b), and β-actin (Fig. 5c). RT+ and RT- aliquots of individual samples were run in adjacent lanes to demonstrate that signal was RT dependant. Human brain (Fig. 5, RT+, lane 2) gave no signal for mouse Nr2e1 (top) but was positive for each of human NR2E1 (middle) and β-actin (bottom). cDNA from wild-type (Fig. 5, RT+, lane 4) was positive for mouse Nr2e1 (top), negative for human NR2E1 (middle), and positive for β-actin (bottom). cDNA corresponding to the fierce mutant (Fig. 5, RT+, lane 6) was negative for each of mouse Nr2e1 (top) and human NR2E1 (middle) but positive for β-actin (bottom). Transgenic samples (Fig. 5, RT+, lane 8) were positive for each of mouse Nr2e1 (top), human NR2E1 (middle), and β-actin (bottom). Rescue mice (Fig. 5, RT+, lane 10) were negative for mouse Nr2e1 (top) but positive for each of human NR2E1 (middle) and β-actin (bottom). None of the RT- samples (lanes 3, 5, 7, 9, or 11) gave rise to any product for any of the three assays. Thus, there was adult expression of the human transgene in the rescue mice, further strengthening the conclusion of functional homology of the coding and regulatory sequences of human and mouse NR2E1 but not allowing for temporal dissection of NR2E1 function.
We show here that the presence of human NR2E1 was sufficient to eliminate structural brain abnormalities while ameliorating eye abnormalities in fierce mice deleted for mouse Nr2e1. We further demonstrate that the presence of human NR2E1 was able to correct the pathological aggression and other behavioral abnormalities normally present in fierce. Our data demonstrating the presence of transgene-derived human NR2E1 in the brains of adult transgenic and rescue mice provides additional support for shared function between the mouse and human gene homologs. Together, these data suggest that NR2E1 protein and regulatory sequences are comparable between mice and humans. Our data also demonstrate that conservation is sufficient to maintain the function of both upstream and downstream signaling pathways, at least in mice. Finally, we describe an experimental system in which to test the functional consequences of specific human NR2E1 alleles on behavior. Such a system may also be useful in understanding the specific role of NR2E1 in risk for psychiatric illness and the manner by which this risk interacts with changes in the environment.
Mechanisms underlying NR2E1 behavior modulation conserved to human
Transgenic rescue experiments in which a human gene is studied in a mouse background have proven useful previously in the analysis of gene function in vivo (Schedl et al., 1996; Chen et al., 2002; Cheung et al., 2004). Although increasingly common in disciplines outside of psychiatry, only a handful of experiments within the discipline have used the approach. Those that do have shown correction of embryonic lethality by a human gene (Hodgson et al., 1996; Pook et al., 2001) but not abnormal behavior (Peier et al., 2000). Thus, before the studies described here, it was unclear whether a human gene could correct abnormal behavior in mutant mice. By extension, the conservation of genetic mechanisms underlying the modulation of behavior was similarly unknown.
From mouse behavior to human disease
The function of NR2E1 itself is also complex and the subject of active investigation (Stenman et al., 2003a,b; Miyawaki et al., 2004; Roy et al., 2004; Shi et al., 2004). However, an emerging hypothesis suggests that the primary function of this nuclear receptor is to suppress the differentiation of neuronal stem cells (Younossi-Hartenstein et al., 1997; Nguyen et al., 1999; Shi et al., 2004). Consistent with this hypothesis, Nr2e1 is transcribed in the embryonic and adult mouse within regions important for neurogenesis (Monaghan et al., 1995; Shi et al., 2004). Moreover, although mice deleted for Nr2e1 show a complex series of phenotypic abnormalities (Monaghan et al., 1997; Yu et al., 2000; Young et al., 2002), the developmental hypoplasia observed within the forebrain, eye, and olfactory bulbs may be secondary to a reduced proliferative capacity. Thus, at the cellular level, NR2E1 may normally act to suppress differentiation, thereby enabling an expansion of the neuronal stem cell pool ultimately available for neurogenesis. Our data support this hypothesis but do not speak to whether the abnormal behavior observed in fierce mice is the result of abnormal development or the absence of NR2E1 signaling in adulthood. Because the human transgene, like mouse Nr2e1, appears to be expressed during development and adulthood, it is not possible, with the strains we have established, to examine the extent to which signaling at each stage may be required for normal behavior. Conditional ablation of the endogenous gene in adulthood or conditional rescue of expression through development alone would be informative in this regard. Furthermore, although our data also support the conservation of mechanisms underlying behavior abnormalities between mouse and human, the clinical presentation of a common cellular defect may be phenotypically distinct between species. Thus, the data we present here do not demonstrate that a human homozygous for a deletion at NR2E1 would be phenotypically similar to the fierce mouse, although they are consistent with this hypothesis. Although speculative, it is also possible that more subtle variation in human NR2E1 may predispose toward abnormal behavioral phenotypes. This hypothesis is bolstered by evidence for linkage with bipolar disorder at human 6q21-22 (Dick et al., 2003; Middleton et al., 2004); however, NR2E1 is only one among many good candidates within this interval. Of additional interest is that NR2E1-interacting genes (Kobayashi et al., 2000; Yu et al., 2000; Stenman et al., 2003a; Shi et al., 2004) are themselves implicated in mental illness (Goodman, 1998; Krezel et al., 1998; Heyman et al., 1999; Stober et al., 1999; Buervenich et al., 2000; Chen et al., 2001; Davis et al., 2002; Iwayama-Shigeno et al., 2003; Rothermundt et al., 2004). If and when association studies implicate specific human NR2E1 alleles in psychiatric disease, the functional evaluation of these candidates in our system would be of particular interest.
Rescue of fierce mouse behavior unexpected
We were surprised at obtaining complete correction of brain and behavior abnormalities in the fierce mouse despite only four amino acid differences between the mouse and human NR2E1 proteins. First, it is well established that subtle structural differences within gene orthologs from closely related species can give rise to striking phenotypic variation (Enard et al., 2002). Second, Nr2e1 transcript distribution in mouse is both spatially complex and temporally dynamic. It is first seen at approximately embryonic day 8 (E8) in a few adjacent neuroepithelial cells at the anterior limit of the prosensencephalon, but by E8.5, the transcript has spread to the presumptive diencephalon and is also present in the newly formed optic and olfactory evaginations (Monaghan et al., 1995). Transcription peaks at E13.5 in ventricular and subventricular zones and although almost undetectable in the perinatal brain (Monaghan et al., 1995), is seen later in the adult brain at high levels in the dentate gyri and subventricular zones and at lower levels scattered throughout the cortex (Shi et al., 2004). Because data from human are limited (Jackson et al., 1998), it was unclear whether the expression of the molecule is entirely conserved either in place or across development. Third, NR2E1 acts in more than one cell type (Shi et al., 2004), and at least one of its functions appears to be cell-type specific (Kobayashi et al., 2000). Given the ability of the human transgene to ameliorate developmental anomalies in brain and eye structure, however, we would suggest that the embryonic distribution of transgene-derived human NR2E1 must at least approximate that of the wild-type allele. Moreover, the presence of human NR2E1 in the brains of adult transgenic and rescue mice provides additional support for functional equivalency between the mouse and human gene homologs.
Human rescue paradigm: a useful tool for psychiatric research
Our data represent the first example of a human gene correcting mouse behavior. Whereas cell-based endpoints are suitable for some kinds of disease, others, particularly psychiatric disorders, require in vivo assay systems. Moreover, as the number of coding variants, and most importantly, noncoding variants implicated in disease increases, so will the need for systems in which to assess their function. We have established an experimental paradigm in which to functionally evaluate the role of human NR2E1 alleles in the modulation of whole-animal behavior. Experiments similar to those we have described will be useful to clarify the role of other genes and alleles in human behavior and psychiatric disease.
B.S.A. was supported by Doctoral Research Awards from both the Canadian Institutes of Health Research and the Scottish Rite Charitable Foundation of Canada. S.M.H. was supported by a Jack and Doris Brown Foundation Award. E.M.S. holds a Canada Research Chair in Genetics and Behavior and was supported by the Canada Foundation for Innovation New Opportunities Fund/British Columbia Knowledge Development Fund, the Canadian Institutes of Health Research, and the National Institute of Health Mental Health. We are grateful to Dr. Bruce P. Connop, Dr. Wim E. Crusio, Dr. Liisa A. Galea, Dr. Candace E. Hofmann, Ravinesh A. Kumar, Simon Warby, Bibiana K. Y. Wong, and Dr. Yuanyun Xie for helpful discussion, to Slavita Bohacec for technical assistance to Brent Gowen for histological work, to Boris Kuzeljevic and Tracey D. Weir for assistance with statistical analyses, to Wendy Wen for initial help with funduscopy, and to Veronica Yakoleff for aid in manuscript preparation.
Correspondence should be addressed to Elizabeth M. Simpson at the above address. E-mail:.
Copyright © 2005 Society for Neuroscience 0270-6474/05/256263-08$15.00/0