Genetic Analysis of the Hypothalamic Corticotropin-Releasing Factor System
http://www.100md.com
内分泌学杂志 2005年第5期
Department of Psychiatry and Behavioral Sciences, Emory University School of Medicine, Atlanta, Georgia 30322
Address all correspondence and requests for reprints to: Steven J. Garlow, M.D., Ph.D., Mood and Anxiety Disorders Program, Department of Psychiatry and Behavioral Science, Emory University School of Medicine, 1841 Clifton Road, 4th Floor, Atlanta, Georgia 30329. E-mail: sgarlow@emory.edu.
Abstract
The goal of this study was to use BxD recombinant inbred mice to search for genes that control the hypothalamic corticotrophin-releasing factor (CRF) system. The specific phenotype that was measured was abundance of transcripts that encode CRF, CRF receptor (Crf-R1), CRF binding protein, and arginine vasopressin (AVP) in total hypothalamic RNA. The strain distribution patterns for the transcript abundances for each target were continuously distributed, consistent with these being quantitative traits. Marker regression and interval mapping revealed associations with quantitative trait loci (QTL) for CRF transcript abundance on chromosome 1 (at 89.2 cM), chromosome 12 (between 54–58 cM), and chromosome 13 (between 26–30 cM); for Crf-R1 transcript abundance on chromosome 7 (at 1.5 cM), chromosome 12 (at 37 cM), and chromosome X (at 30 cM); for CRF binding protein transcript abundance on chromosome 7 (at 48.5 cM), chromosome 8 (at 65 cM), and chromosome 12 (at 19 cM); and for AVP transcript abundance on chromosome 7 (at 1 cM), chromosome 12 (at 13 cM), and chromosome 13 (at 45 cM). The transcript abundance QTL were not linked to their respective structural genes. Interval mapping on chromosome 7 reveals substantial overlap between QTL that control AVP and Crf-R1 transcript abundance and on chromosome 12 for QTL that control CRF and Crf-R1, which may indicate loci that coordinate regulation of the CRF system. There are QTL for all four targets on chromosome 12. There are a number of neurodevelopmental genes in very close proximity to the transcript abundance QTL that are potential candidate genes.
Introduction
CORTICOTROPIN-RELEASING FACTOR (CRF) plays a preeminent role in the central nervous system (CNS), as the major physiological mediator of the stress response. Hypothalamic CRF secretion regulates the activity of the hypothalamic-pituitary-adrenal axis, and both hypothalamic and extra-hypothalamic CRF circuits act in concert to coordinate responses to stress. Dysfunction of hypothalamic and extra-hypothalamic CRF circuits has been demonstrated in the CNS of depressed patients and has been posited to be integral in the pathophysiology of major depression and posttraumatic stress disorder (1, 2, 3, 4). Adverse events early in life such as childhood physical or sexual abuse and neglect have been shown to result in persisting alterations in the regulation of the CRF system and hypothalamic-pituitary-adrenal axis, with the consequence being an increased vulnerability to mood and anxiety disorders in adulthood (5). However, not all individuals exposed to a particular stressor either in early life or adulthood develops a mood or anxiety disorder. There is considerable interindividual variability in vulnerability to stress, and genetic regulation of stress response systems may be one mechanism by which this differential vulnerability is conferred. Although the Crh gene itself is not the likely substrate for these differences, genes that control the CRF system may contribute to risk of developing mood and anxiety disorders in response to stressful or adverse events.
The BxD series of recombinant inbred mice has emerged as one of the principal methods to study the genetic architecture of complex neurobehavioral phenotypes. This system has been used to nominate and map quantitative trait loci (QTL) that contribute to many different complex traits including behavioral, pharmacological, neuroanatomical, and neurochemical phenotypes (6, 7, 8, 9, 10, 11, 12, 13, 14). The major question posed in this study is whether the hypothalamic CRF system is under the control of one or more QTL and whether the expression of the various components of the system, CRF, Crf-R1, CRF binding protein (Crf-BP), arginine vasopressin (AVP) are controlled by the same set of genes, or by separate genes that regulate the expression of each target. The phenotype measured, transcript abundance for each target in total hypothalamic RNA, was chosen because it is not constrained to any particular molecular mechanism. QTL that control biological processes at many different levels, from specific transcriptional mechanisms through cellular differentiation and stress responsivity could be detected with this phenotype.
Materials and Methods
Animals and tissue preparation
Foundation stock for 28 BxD lines and the C57BL6 and DBA2J parent lines were acquired from The Jackson Laboratory (Bar Harbor, ME) and used to establish a breeding colony. Breeding pairs consisted of one male and two female mice. Dams were individually housed while pregnant and in the postnatal period through weaning. Animals were weaned at postnatal d 21 and then group housed three per cage by same sex and strain. Ambient colony conditions were temperature at 23 ± 2 C and relative humidity between 30 and 60%, with a 12-h light, 12-h dark cycle (lights out at 1800 h) and animals were allowed ad libitum access to standard laboratory chow and water. Animals were housed in a mouse-only facility. Animals were killed by cervical dislocation at postnatal d 75–90, under low stress conditions. Brains were rapidly dissected on ice into constituent anatomical regions and stored at –80 C until used.
RNA quantification
Total hypothalamic RNA was isolated with the RNeasy Mini Kit (Qiagen Inc., Valencia, CA) by manufacturer’s method. Five-microgram aliquots of total RNA from individual animals were converted to cDNA with the SuperScript First Strand Synthesis System for Real Time PCR (Invitrogen Life Technologies, Carlsbad, CA). Target and 18S transcript abundances were determined with specific real time RT-PCR assays, using TaqMan detection chemistry (Applied Biosystems, Foster City, CA). Amplification primers for CRF (accession no. NM_031019) were designed with the Primer Express computer program (Applied Biosystems) and were forward 5'-CAGCCGTTGAATTTCTTGCA-3', corresponding to nucleotides 114–133 of the CRF open reading frame and reverse 5'-TCACCCATGCGGATCAGA-3' corresponding to nucleotides 166–184 and the detection oligonucleotide was 5'-(6-carboxyfluorescein)-CGGAGCAGCCCCAGCAACCTC-(5-(and 6)-carboxytetramethylrhodamine)-3'corresponding to nucleotides 139–156. The amplification and detection primers for the other target assays Crf-R1 (accession no. NM_007762; ABI assay no. Mm00432669), Crf-BP (accession no. NM_139183; ABI assay no. Rn00594854), and AVP (accession no. NM_009732; ABI assay no. Mm00437761) were purchased from ABI Assays on Demand. The 18S transcript was quantified with the commercially available amplification-detection set from the manufacturer (Applied Biosystems). Aliquots of cDNA equivalent to 0.3 μg of input RNA were assayed in an ABI 7000 real time PCR system, in the 96-well format, using the manufacturer’s specified PCR cycle parameters, through 40 cycles. This is a relative abundance assay, and target transcript abundance is expressed as the detection threshold cycle (Ct), which is in the middle of the exponential phase of the PCR amplification. As detection cycle is the readout parameter, a higher Ct value corresponds to lower transcript abundance, because more PCR cycles were required to amplify the target. A dilution series positive control and no transcript and no reverse transcriptase negative controls were included in each assay. To assure assay stability, the entire experiment was performed in duplicate, with identical results.
Statistical analysis and genetic mapping
Strain mean transcript abundances were determined from six individual male animals per line (Fig. 1). Statistical analysis of transcript abundance data were conducted with the JMP-5 computer program (SAS Institute Inc., Cary, NC) on Macintosh G-4 computers. Descriptive statistics were calculated for each line and comparison of strain mean values was by ANOVA as implemented by JMP-5 (15). Genetic mapping was performed with the Macintosh version of QTX (http://www.mapmanager.org/mmQTX.html), run on Macintosh G-4 computers (16) and with WebQTL (http://www.genenetwork.org/home.html) (17). Both programs were used for marker regression analysis and interval mapping of the CRF system transcript abundance data and gave essentially identical results; hence the results from QTX are presented. The BxD genetic marker set used with QTX was from Dr. Robert Williams at the University of Tennessee Health Science Center, School of Medicine (Memphis, TN) (18). This marker set consisted of 975 error-checked and nonredundant loci. Results of genetic mapping with QTX are expressed as likelihood ratio statistic (LRS) and statistical significance is the probability of a type I (false positive) error, or the probability of obtaining by chance an LRS value as great as that observed (19). Marker regression was used to detect potential QTL, and these calculations were carried out at increasingly stringent point (2) significance levels, from P < 0.001 through P < 10–6. During the marker regression operation, QTX calculates a 95% confidence interval (CI) for the LRS peaks, expressed in centiMorgans, and the width of the CI are inversely proportional to the strength of a QTL at that location (20). Interval mapping was used to localize QTL on the relevant chromosomes (21, 22, 23). Critical LRS values for interval mapping operations for each target transcript were empirically set with the permutation test run through 10,000 iterations at a 1-cM interval (24). The critical LRS values are defined as suggestive which corresponds to the 37th percentile or P < 0.63, significant at the 95th percentile or P < 0.05, and highly significant at the 99.9th percentile or P < 0.001 (see Table 3) (25). Chromosomal locations of candidate and target structural genes are from the Mouse Genome Database, Mouse Genome Informatics (v3.1) web site (http://www.informatics.jax.org) at The Jackson Laboratory queried on December 20, 2004 (26). The WebQTL suite of programs was used to carry out exploratory correlation analyses of strain mean CRF-system transcript abundance data with the UTHSC Brain mRNA U74Av2 RMA mass gene expression database (March 2004 data freeze), and the INIA M430 brain RMA database (October 2004 data freeze). CRF-system transcript abundance data were also compared with the WebQTL BxD phenotypes database (17).
FIG. 1. SDP for transcript abundance data for gene products of the CRF system across lines of BxD mice and parent strains C57BL/6 (b6) and DBA/2J (d2). Values are mean of transcript abundance measurement from six individual male animals, with SD. Data are expressed as Ct, so higher Ct values correspond to lower transcript abundances. A, CRF transcript abundance (F = 4.315, df = 29, 147, P < 0.0001). B, CRF-R1 transcript abundance (F = 6.614, df = 27, 139, P < 0.0001). C, Crf-BP transcript abundance (F = 3.151, df = 27, 136, P < 0.0001). D, AVP transcript abundance (F = 55.978, df = 27, 138, P < 0.0001).
TABLE 3. Critical LRS values for interval mapping operations determined by permutation test run through 10,000 iterations for each set of target data set
This research project has been reviewed and approved by the Emory University Institutional Animal Care and Use Committee and complied with all relevant regulations.
Results
The strain distribution pattern (SDP) for the mean hypothalamic CRF transcript abundance was calculated for 28 BxD and both parent lines. Transcript abundance for the other three targets (Crf-R1, Crf-BP, and AVP) was determined from both parent and 26 BxD lines, due to poor breeding performance of lines 1 and 30. Strain mean transcript abundance for each target, as represented by Ct, differed significantly between the BxD and parent lines (Table 1). The strain mean transcript abundances for each target are distributed continuously, which is consistent with these being quantitative traits (Fig. 1). The narrow-sense heritabilities for the transcript abundance phenotypes are virtually identical to the model-corrected R2 values for effect of strain, and suggest no meaningful epistatic interactions (27). Ct is the output measure and higher Ct values correspond to lower transcript abundance as more PCR cycles were required to detect the target, and for all target transcripts scrutinized in this experiment (CRF, Crf-R1, Crf-BP, and AVP), the D2 parent strain has a lower Ct (higher transcript abundance) than the B6 line. The SDPs for each of the four targets were transgressive, with high and low expressing strains bracketing the parent lines.
TABLE 1. Statistical analysis of CRF system transcript abundance data in strains of BxD mice
Marker regression analysis of strain mean transcript abundance was carried out for each target at the point significance thresholds of P < 0.001 and P < 0.0001 (Table 2). At least one marker association for each of the four targets was detected at the P < 0.001 level and for three of four targets (CRF, Crf-R1, and AVP) associations were detected at the P < 0.0001 levels. The CRF transcript abundance (CRFta) associations on chromosome 12 at D12Msw102 and D12Mit280 and the Crf-R1 transcript abundance (Crf-R1ta) association on chromosome X at SXGnf55.520 continued to be detected at 2 significance of P < 10–5 and the AVP transcript abundance (AVPta) association on chromosome 7 at S07Gnf001.580 was detected at P < 10–6. Comparison of results of suggestive-level marker regression of CRF to Crf-R1 reveals associations of both phenotypes with a cluster of genetic markers on chromosome 12 (Fig. 2C). Similar comparison of Crf-R1 to AVP reveals an area of potential shared genetic regulation on chromosome 7 (Fig. 2B). Marker associations were detected for all four targets on chromosome 12 (Fig. 2C).
TABLE 2. Results of marker regression analysis of CRF-system transcript abundance data
FIG. 2. Results of interval mapping data of CRF system transcript abundance data on five mouse chromosomes. Critical LRS values for mapping were determined with permutation test run through 10,000 iterations for each target (Table 3); CRF (orange), Crf-R1 (green), Crf-BP (red), AVP (blue). A, Chromosome 1; B, chromosome 7; C, chromosome 12; D, chromosome 13; E, chromosome X.
Interval mapping was used to localize the linkage signals detected in the marker regression analysis (Table 2 for chromosomal map location) onto the respective chromosomes (Fig. 2). The critical LRS values for each data set were determined with the permutation test (Table 3). Interval mapping of CRFta revealed LRS peaks on chromosome 1 at P1Ehs1, on chromosome 12 at D12Mit280, and on chromosome 13 between D13Mit63 and S13Gnf043.620. Interval mapping of Crf-R1ta data revealed suggestive LRS peaks on chromosome 7 at D7mit68 and chromosome 12 at D12Mit5, and a significant peak on chromosome X at SXGnf55.520. Interval mapping of Crf-BPta data revealed suggestive LRS peaks on chromosome 7 at D7Mit301 and on chromosome 12 between D12Mit154 and D12Mit235. Interval mapping of AVPta data on chromosome 7 revealed a highly significant LRS peak at S07Gnf001.580, and significant peaks on chromosome 12 at D12Mit136 and chromosome 13 at D13Mit126.
The QTL detected in this analysis do not appear to be linked to the respective structural genes. The structural gene that encodes CRF (designated Crh) is on chromosome 3 (8 cM), Crf-R1 (Crhr1) on chromosome 11 (62 cM), Crf-BP (Crhbp) on chromosome 13 (52 cM), and AVP (Avp) on chromosome 2 (73.2 cM). All of these QTL must be acting in trans to exert an influence on the expression of the target genes, given that they are on different chromosomes. The hypothalamus is a sexually dimorphic organ, and all of the included analyses were performed in males. Thus the influence of gender on these QTL cannot be ascertained until the experiment is repeated in females.
Discussion
The distributions of strain mean transcript abundance values for each of the four targets, CRF, Crf-R1, Crf-BP, and AVP were continuous across the panel of BxD strains, consistent with these phenotypes being quantitative traits. The SDP were transgressive with extreme high and low expressing strains flanking the values of the parent lines. At least one suggestive linkage association was detected with both marker regression and interval mapping for all four targets, and significant associations were detected for CRF, Crf-R1, and AVP transcript abundance. The LRS peaks detected with both methods for these three targets were well above the significance levels set by the permutation test, so at least one QTL controlling hypothalamic CRF-system transcript abundance can be declared for each of these three target genes and there were others for each target in the suggestive range. There are QTL that apparently control expression of only one target and others that potentially impact expression of more than one target gene. Inspection of the QTL detection (Table 2) and localization (Fig. 2) results suggests that expression of CRF and Crf-R1 share elements of genetic regulation on chromosome 12 and AVP and Crf-R1 share elements on chromosome 7. Interestingly, this analysis did not reveal any potential shared genetic regulation between CRF and Crf-BP transcript abundances.
QTL are probabilistically defined entities, and the relationship between these genes and those that have been identified through sequencing of genomic DNA or expressed sequences remains obscure. The pool of potential candidate genes for a QTL is defined by those sequence-defined transcription units coincident with LRS peaks. One of the criteria considered central to connecting a specific candidate gene to a QTL is that biological evidence connecting the activity of the candidate gene to the biological process captured by the quantitative trait must be established (28). However, different phenotype measurements that define quantitative traits could detect the same QTL, if both were involved in the same fundamental process. The CRF system regulates and coordinates stress responses, so QTL that impact these types of behaviors could be considered as potential candidate genes. Proposing candidate genes for these QTL at this point is premature, as any of these LRS peaks could contain 10s to 100s of actual genes, nevertheless, there are a number of intriguing genes that map in the same locations as CRF system transcript abundance QTL.
The QTL Hipp1A, which controls hippocampal weight and cellularity, is located at map position 89 cM, in very close proximity to the CRFta LRS peak at 89.2 cM (11). Comparison of the CRFta data to the WebQTL phenotypes database revealed modest, uncorrected correlations of CRF-Ct values to a number of hippocampal phenotypes such as hippocampal weight (r = 0.599, P < 0.0003), pyramidal cell layer volume (r = 0.529, P < 0.003), total hippocampal volume (r = 0.507, P < 0.005), and 5-bromo-2'-deoxyuridine incorporation into adult hippocampal neurons (r = 0.7552, P < 0.003) (29). These correlations indicate an inverse relationship between hypothalamic CRF transcript abundance and measures of hippocampal volume and cellularity. The structural gene for the brain-specific isoenzyme of creatine kinase (CKB) is at map position 55 cM on chromosome 12, coincident with the most significant CRFta LRS peak detected in this analysis, and disruption of the CKB gene results in a complex phenotype including increased hippocampal mossy fiber field size and impaired habituation and spatial learning (30). The QTL power in slow-wave sleep 2 (Dps2) also maps at this location on chromosome 12 (31). The CRFta LRS peak on chromosome 13 is broad, but does contain the Shc3 gene, which encodes the neuron-specific ShcC protein that participates in intracellular signaling of receptor tyrosine kinases (32, 33).
There appears to be a region of shared genetic regulation for AVP and Crf-R1ta on chromosome 7, but there are few potential candidate genes in this region. The QTL maternal performance (Mprf) does map at 0.5 cM on chromosome 7 and AVP has been implicated in contributing to maternal behaviors (34, 35, 36). The AVPta LRS peak on chromosome 13 at 45 cM is coincident with the Nr2f1 gene which encodes the chicken ovalbumin upstream promoter-transcription factor orphan steroid/ thyroid hormone receptor that is involved in differentiation and development of the CNS (37). The QTL power in slow-wave sleep 1 (Dps1) also maps at this location (31). There are a number of potential candidate genes within the Crf-R1 LRS peak at 30 cM on the X chromosome, including Avpr2, which encodes the type-2 AVP receptor.
There are several limitations to this study. The first is that it was conducted in one set of segregating animals, that being the BxD recombinant inbred set, so the linkage findings need to be confirmed in a separate, independent analysis. These experiments are ongoing, but the statistical significance of some of the associations, in particular for CRFta on chromosome 12, AVPta on chromosome 7, and Crf-R1ta on the X chromosome, argues against these being false-positive findings. That different targets had unique and shared regions of linkage argues in favor of these QTL being elements of the genetic system that controls the hypothalamic CRF system. Another limitation is that all of the animals in this experiment were male, so no conclusions can be drawn as to the genetic regulation of the CRF system in females. The resolution of the mapping algorithms employed, and the density of recombinations in the strain set are not sufficient to engage in fine mapping, so linkages in smaller candidate intervals cannot be detected, thus limiting the resolution of the mapping results. Fine scale mapping of these QTL will be dependent on additional analysis of other sets of animals (segregating F2 cross, advanced intercross recombinant inbred set, etc.) and only with high resolution fine mapping will genuine candidate genes become apparent as the target interval shrinks.
In conclusion, we present evidence that the hypothalamic CRF system is controlled by a series of QTL, some of which are specific to only one target gene and some may exert a more global influence on multiple genes in the CRF system. Transcript abundances of CRF, Crf-R1, Crf-BP, and AVP in total hypothalamic RNA are quantitative traits and we have detected at least one QTL that controls the abundance of each of these transcripts in total hypothalamic RNA.
Acknowledgments
We thank Dr. Robert Williams and Dr. Kenneth Manly of the University of Tennessee Health Science Center, School of Medicine for advice and guidance in conducting these experiments.
References
Nemeroff CB 1996 The corticotropin-releasing factor (CRF) hypothesis of depression: new findings and new directions. Mol Psychiatry 1:336–342
Nemeroff CB, Widerlov E, Bissette G, Walleus H, Karlsson I, Eklund K, Kilts CD, Loosen PT, Vale W 1984 Elevated concentration of CSF corticotropin-releasing factor-like immunoreactivity in depressed patients. Science 226:1342–1344
Bremner JD, Licinio J, Darnell A, Krystal JH, Owens MJ, Southwick SM, Nemeroff CB, Charney DS 1997 Elevated CSF corticotropin-releasing factor concentrations in posttraumatic stress disorder. Am J Psychiatry 154:624–629
Arborelius L, Owens MJ, Plotsky PM, Nemeroff CB 1999 The role of corticotropin-releasing factor in depression and anxiety disorders. J Endocrinol 160:1–12
Heim C, Newport DJ, Heit S, Graham YP, Wilcox M, Bonsall R, Miller AH, Nemeroff CB 2000 Pituitary-adrenal and autonomic responses to stress in women after sexual and physical abuse in childhood. J Am Med Assoc 284:592–597
Belknap JK, Phillips TJ, O‘Toole LA 1992 Quantitative trait loci associated with brain weight in BxD/Ty recombinant inbred mouse strains. Brain Res Bull 29:337–344
Miner LL, Marley RJ 1995 Chromosomal mapping of the psychomotor stimulant effects of cocaine in BxD recombinant inbred mice. Psychopharmacology 122:209–214
Tarricone BJ, Hingtgen JN, Belknap JK, Mitchell SR, Nurnberger JI 1995 Quantitative trait loci associated with behavioral response of BxD recombinant inbred mice to restraint stress: a preliminary communication. Behav Genet 25:489–495
Tarricone BJ, Hwang WG, Hingtgen JN, Mitchell SR, Belknap JK, Nurnberger JI 1995 Identification of a locus on mouse chromosome 17 associated with high-affinity choline uptake using BxD recombinant inbred mice and quantitative trait loci analysis. Genomics 27:161–164
Gora-Masalak G, McClearn GE, Crabbe JC, Phillips TJ, Belknap JK, Plomin R 1991 Use of recombinant inbred strains to identify quantitative trait loci in psychopharmacology. Psychopharmacology 104:413–424
Lu L, Airey DC, Williams RW 2001 Complex trait analysis of the hippocampus: mapping and biometric analysis of two novel gene loci with specific effects on hippocampal structure in mice. J Neurosci 21:3503–3514
Mayeda AR, Hofstetter JR, Belknap JK, Nurnberger JI 1996 Hypothetical quantitative trait loci (QTL) for circadian period of locomotor activity in BxD recombinant inbred strains of mice. Behav Genet 26:505–511
Talbot CJ, Radcliffe RA, Fullerton J, Hitzemann B, Wehner J, Flint J 2003 Fine scale mapping of a genetic locus for conditioned fear. Mamm Genome 14:223–230
Zhang S, Gershenfeld HK 2003 Genetic contributions to body weight in mice: relationship of exploratory behavior to weight. Obes Res 11:828–838
2002 JMP statistics and graphics guide, version 5. Cary, NC: SAS Institute Inc.
Manly K, Cudmore JRH, Meer JM 2001 Map Manager QTX, cross-platform software for genetic mapping. Mamm Genome 12:930–932
Wang J, Williams RW, Manly KF 2003 WebQTL: web-based complex trait analysis. Neuroinformatics 1:299–308
Williams RW, Gu J, Lu L 2001 The genetic structure of recombinant inbred mice: high-resolution consensus maps for complex trait analysis. Genome Biol 2:1–18
Haley CS, Knott SA 1992 A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. Heredity 69:315–324
Darvasi A, Soller M 1997 A simple method to calculate resolving power and confidence interval of QTL map location. Behav Genet 27:125–132
Zeng ZB 1993 Theoretical basis for separation of multiple linked gene effects in mapping quantitative trait loci. Proc Natl Acad Sci USA 90:10972–10976
Zeng ZB 1994 Precision mapping of quantitative trait loci. Genetics 136:1457–1468
Lander ES, Botstein D 1989 Mapping Mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121:185–199
Churchill GA, Doerge RW 1994 Empirical threshold values for quantitative trait mapping. Genetics 138:963–971
Lander ES, Kruglyak L 1995 Genetic dissection of complex traits-Guidelines for interpreting and reporting linkage results. Nat Genet 11:241–247
Blake JA, Richardson JE, Bult CJ, Kadin JA, Eppig JT, Mouse Genome Database Group 2003 The Mouse Genome Database. Nucleic Acids Res 31:193–195
Belknap JK, Mitchell SR, O’Toole LA, Helms ML, Crabbe JC 1995 Type I and type II error rates for quantitative trait loci (QTL) mapping studies using recombinant inbred mouse strains. Behav Genet 26:149–160
Abiola O, Angel JM, Avner P, Bachmanov AA, Belknap JK, Bennett B, Blankenhorn EP, Blizard DA, Bolivar V, Brockmann GA, Buck KJ, Bureau JF, Casley WL, Chesler EJ, Cheverud JM, Churchill GA, Cook M, Crabbe JC, Crusio WE, Darvasi A, de Haan G, Dermant P, Doerge RW, Elliot RW, et al. 2003 The nature and identification of quantitative trait loci: a community‘s view. Nat Rev Genet 4:911–916
Kempermann G, Gage FH 2002 Genetic determinants of adult hippocampal neurogenesis correlate with acquisition, but not probe trial performance in the water maze task. Eur J Neurosci 16:1129–1136
Jost CR, Van Der Zee CE, In’t Zandt HJ, Oerlemans F, Verheij M, Streijger F, Fransen J, Heerschap A, Cools AR, Wieringa B 2002 Creatine kinase B-driven energy transfer in the brain is important for habituation and spatial learning behaviour, mossy fibre field size and determination of seizure susceptibility. Eur J Neurosci 15:1692–1706
Franken P, Chollet D, Tafti M 2001 The homeostatic regulation of sleep need is under genetic control. J Neurosci 21:2610–2621
O‘Bryan JP, Songyang Z, Cantley L, Der CJ, Pawson T 1996 A mammalian adaptor protein with conserved Src homology 2 and phosphotyrosine-binding domains is related to Shc and is specifically expressed in the brain. Proc Natl Acad Sci USA 93:2729–2734
Troglio F, Echart C, Gobbi A, Pawson T, Pelicci PG, De Simoni MG, Pelicci G 2004 The Rai (Shc C) adaptor protein regulates neuronal stress response and protects against cerebral ischemia. Proc Natl Acad Sci USA 101:15476–15481
Peripato AC, De Brito RA, Vaughn TT, Pletscher LS, Matioli SR, Cheverud JM 2002 Quantitative trait loci for maternal performance for offspring survival in mice. Genetics 162:1341–1353
Pedersen CA, Caldwell JD, Johnson MF, Fort SA, Prange A 1985 Oxytocin antiserum delays onset of ovarian steroid-induced maternal behavior. Neuropeptides 6:175–182
Pedersen CA, Ascher JA, Monroe YL, Prange A 1982 Oxytocin induces maternal behavior in virgin female rats. Science 216:648–650
Qiu Y, Pereira FA, DeMayo FJ, Lydon JP, Tsai SY, Tsai MJ 1997 Null mutation of mCOUP-TFI results in defects in morphogenesis of the glossopharyngeal ganglion, axonal projection and arborization. Genes Dev 11:1925–1937(Steven J. Garlow, Ericka )
Address all correspondence and requests for reprints to: Steven J. Garlow, M.D., Ph.D., Mood and Anxiety Disorders Program, Department of Psychiatry and Behavioral Science, Emory University School of Medicine, 1841 Clifton Road, 4th Floor, Atlanta, Georgia 30329. E-mail: sgarlow@emory.edu.
Abstract
The goal of this study was to use BxD recombinant inbred mice to search for genes that control the hypothalamic corticotrophin-releasing factor (CRF) system. The specific phenotype that was measured was abundance of transcripts that encode CRF, CRF receptor (Crf-R1), CRF binding protein, and arginine vasopressin (AVP) in total hypothalamic RNA. The strain distribution patterns for the transcript abundances for each target were continuously distributed, consistent with these being quantitative traits. Marker regression and interval mapping revealed associations with quantitative trait loci (QTL) for CRF transcript abundance on chromosome 1 (at 89.2 cM), chromosome 12 (between 54–58 cM), and chromosome 13 (between 26–30 cM); for Crf-R1 transcript abundance on chromosome 7 (at 1.5 cM), chromosome 12 (at 37 cM), and chromosome X (at 30 cM); for CRF binding protein transcript abundance on chromosome 7 (at 48.5 cM), chromosome 8 (at 65 cM), and chromosome 12 (at 19 cM); and for AVP transcript abundance on chromosome 7 (at 1 cM), chromosome 12 (at 13 cM), and chromosome 13 (at 45 cM). The transcript abundance QTL were not linked to their respective structural genes. Interval mapping on chromosome 7 reveals substantial overlap between QTL that control AVP and Crf-R1 transcript abundance and on chromosome 12 for QTL that control CRF and Crf-R1, which may indicate loci that coordinate regulation of the CRF system. There are QTL for all four targets on chromosome 12. There are a number of neurodevelopmental genes in very close proximity to the transcript abundance QTL that are potential candidate genes.
Introduction
CORTICOTROPIN-RELEASING FACTOR (CRF) plays a preeminent role in the central nervous system (CNS), as the major physiological mediator of the stress response. Hypothalamic CRF secretion regulates the activity of the hypothalamic-pituitary-adrenal axis, and both hypothalamic and extra-hypothalamic CRF circuits act in concert to coordinate responses to stress. Dysfunction of hypothalamic and extra-hypothalamic CRF circuits has been demonstrated in the CNS of depressed patients and has been posited to be integral in the pathophysiology of major depression and posttraumatic stress disorder (1, 2, 3, 4). Adverse events early in life such as childhood physical or sexual abuse and neglect have been shown to result in persisting alterations in the regulation of the CRF system and hypothalamic-pituitary-adrenal axis, with the consequence being an increased vulnerability to mood and anxiety disorders in adulthood (5). However, not all individuals exposed to a particular stressor either in early life or adulthood develops a mood or anxiety disorder. There is considerable interindividual variability in vulnerability to stress, and genetic regulation of stress response systems may be one mechanism by which this differential vulnerability is conferred. Although the Crh gene itself is not the likely substrate for these differences, genes that control the CRF system may contribute to risk of developing mood and anxiety disorders in response to stressful or adverse events.
The BxD series of recombinant inbred mice has emerged as one of the principal methods to study the genetic architecture of complex neurobehavioral phenotypes. This system has been used to nominate and map quantitative trait loci (QTL) that contribute to many different complex traits including behavioral, pharmacological, neuroanatomical, and neurochemical phenotypes (6, 7, 8, 9, 10, 11, 12, 13, 14). The major question posed in this study is whether the hypothalamic CRF system is under the control of one or more QTL and whether the expression of the various components of the system, CRF, Crf-R1, CRF binding protein (Crf-BP), arginine vasopressin (AVP) are controlled by the same set of genes, or by separate genes that regulate the expression of each target. The phenotype measured, transcript abundance for each target in total hypothalamic RNA, was chosen because it is not constrained to any particular molecular mechanism. QTL that control biological processes at many different levels, from specific transcriptional mechanisms through cellular differentiation and stress responsivity could be detected with this phenotype.
Materials and Methods
Animals and tissue preparation
Foundation stock for 28 BxD lines and the C57BL6 and DBA2J parent lines were acquired from The Jackson Laboratory (Bar Harbor, ME) and used to establish a breeding colony. Breeding pairs consisted of one male and two female mice. Dams were individually housed while pregnant and in the postnatal period through weaning. Animals were weaned at postnatal d 21 and then group housed three per cage by same sex and strain. Ambient colony conditions were temperature at 23 ± 2 C and relative humidity between 30 and 60%, with a 12-h light, 12-h dark cycle (lights out at 1800 h) and animals were allowed ad libitum access to standard laboratory chow and water. Animals were housed in a mouse-only facility. Animals were killed by cervical dislocation at postnatal d 75–90, under low stress conditions. Brains were rapidly dissected on ice into constituent anatomical regions and stored at –80 C until used.
RNA quantification
Total hypothalamic RNA was isolated with the RNeasy Mini Kit (Qiagen Inc., Valencia, CA) by manufacturer’s method. Five-microgram aliquots of total RNA from individual animals were converted to cDNA with the SuperScript First Strand Synthesis System for Real Time PCR (Invitrogen Life Technologies, Carlsbad, CA). Target and 18S transcript abundances were determined with specific real time RT-PCR assays, using TaqMan detection chemistry (Applied Biosystems, Foster City, CA). Amplification primers for CRF (accession no. NM_031019) were designed with the Primer Express computer program (Applied Biosystems) and were forward 5'-CAGCCGTTGAATTTCTTGCA-3', corresponding to nucleotides 114–133 of the CRF open reading frame and reverse 5'-TCACCCATGCGGATCAGA-3' corresponding to nucleotides 166–184 and the detection oligonucleotide was 5'-(6-carboxyfluorescein)-CGGAGCAGCCCCAGCAACCTC-(5-(and 6)-carboxytetramethylrhodamine)-3'corresponding to nucleotides 139–156. The amplification and detection primers for the other target assays Crf-R1 (accession no. NM_007762; ABI assay no. Mm00432669), Crf-BP (accession no. NM_139183; ABI assay no. Rn00594854), and AVP (accession no. NM_009732; ABI assay no. Mm00437761) were purchased from ABI Assays on Demand. The 18S transcript was quantified with the commercially available amplification-detection set from the manufacturer (Applied Biosystems). Aliquots of cDNA equivalent to 0.3 μg of input RNA were assayed in an ABI 7000 real time PCR system, in the 96-well format, using the manufacturer’s specified PCR cycle parameters, through 40 cycles. This is a relative abundance assay, and target transcript abundance is expressed as the detection threshold cycle (Ct), which is in the middle of the exponential phase of the PCR amplification. As detection cycle is the readout parameter, a higher Ct value corresponds to lower transcript abundance, because more PCR cycles were required to amplify the target. A dilution series positive control and no transcript and no reverse transcriptase negative controls were included in each assay. To assure assay stability, the entire experiment was performed in duplicate, with identical results.
Statistical analysis and genetic mapping
Strain mean transcript abundances were determined from six individual male animals per line (Fig. 1). Statistical analysis of transcript abundance data were conducted with the JMP-5 computer program (SAS Institute Inc., Cary, NC) on Macintosh G-4 computers. Descriptive statistics were calculated for each line and comparison of strain mean values was by ANOVA as implemented by JMP-5 (15). Genetic mapping was performed with the Macintosh version of QTX (http://www.mapmanager.org/mmQTX.html), run on Macintosh G-4 computers (16) and with WebQTL (http://www.genenetwork.org/home.html) (17). Both programs were used for marker regression analysis and interval mapping of the CRF system transcript abundance data and gave essentially identical results; hence the results from QTX are presented. The BxD genetic marker set used with QTX was from Dr. Robert Williams at the University of Tennessee Health Science Center, School of Medicine (Memphis, TN) (18). This marker set consisted of 975 error-checked and nonredundant loci. Results of genetic mapping with QTX are expressed as likelihood ratio statistic (LRS) and statistical significance is the probability of a type I (false positive) error, or the probability of obtaining by chance an LRS value as great as that observed (19). Marker regression was used to detect potential QTL, and these calculations were carried out at increasingly stringent point (2) significance levels, from P < 0.001 through P < 10–6. During the marker regression operation, QTX calculates a 95% confidence interval (CI) for the LRS peaks, expressed in centiMorgans, and the width of the CI are inversely proportional to the strength of a QTL at that location (20). Interval mapping was used to localize QTL on the relevant chromosomes (21, 22, 23). Critical LRS values for interval mapping operations for each target transcript were empirically set with the permutation test run through 10,000 iterations at a 1-cM interval (24). The critical LRS values are defined as suggestive which corresponds to the 37th percentile or P < 0.63, significant at the 95th percentile or P < 0.05, and highly significant at the 99.9th percentile or P < 0.001 (see Table 3) (25). Chromosomal locations of candidate and target structural genes are from the Mouse Genome Database, Mouse Genome Informatics (v3.1) web site (http://www.informatics.jax.org) at The Jackson Laboratory queried on December 20, 2004 (26). The WebQTL suite of programs was used to carry out exploratory correlation analyses of strain mean CRF-system transcript abundance data with the UTHSC Brain mRNA U74Av2 RMA mass gene expression database (March 2004 data freeze), and the INIA M430 brain RMA database (October 2004 data freeze). CRF-system transcript abundance data were also compared with the WebQTL BxD phenotypes database (17).
FIG. 1. SDP for transcript abundance data for gene products of the CRF system across lines of BxD mice and parent strains C57BL/6 (b6) and DBA/2J (d2). Values are mean of transcript abundance measurement from six individual male animals, with SD. Data are expressed as Ct, so higher Ct values correspond to lower transcript abundances. A, CRF transcript abundance (F = 4.315, df = 29, 147, P < 0.0001). B, CRF-R1 transcript abundance (F = 6.614, df = 27, 139, P < 0.0001). C, Crf-BP transcript abundance (F = 3.151, df = 27, 136, P < 0.0001). D, AVP transcript abundance (F = 55.978, df = 27, 138, P < 0.0001).
TABLE 3. Critical LRS values for interval mapping operations determined by permutation test run through 10,000 iterations for each set of target data set
This research project has been reviewed and approved by the Emory University Institutional Animal Care and Use Committee and complied with all relevant regulations.
Results
The strain distribution pattern (SDP) for the mean hypothalamic CRF transcript abundance was calculated for 28 BxD and both parent lines. Transcript abundance for the other three targets (Crf-R1, Crf-BP, and AVP) was determined from both parent and 26 BxD lines, due to poor breeding performance of lines 1 and 30. Strain mean transcript abundance for each target, as represented by Ct, differed significantly between the BxD and parent lines (Table 1). The strain mean transcript abundances for each target are distributed continuously, which is consistent with these being quantitative traits (Fig. 1). The narrow-sense heritabilities for the transcript abundance phenotypes are virtually identical to the model-corrected R2 values for effect of strain, and suggest no meaningful epistatic interactions (27). Ct is the output measure and higher Ct values correspond to lower transcript abundance as more PCR cycles were required to detect the target, and for all target transcripts scrutinized in this experiment (CRF, Crf-R1, Crf-BP, and AVP), the D2 parent strain has a lower Ct (higher transcript abundance) than the B6 line. The SDPs for each of the four targets were transgressive, with high and low expressing strains bracketing the parent lines.
TABLE 1. Statistical analysis of CRF system transcript abundance data in strains of BxD mice
Marker regression analysis of strain mean transcript abundance was carried out for each target at the point significance thresholds of P < 0.001 and P < 0.0001 (Table 2). At least one marker association for each of the four targets was detected at the P < 0.001 level and for three of four targets (CRF, Crf-R1, and AVP) associations were detected at the P < 0.0001 levels. The CRF transcript abundance (CRFta) associations on chromosome 12 at D12Msw102 and D12Mit280 and the Crf-R1 transcript abundance (Crf-R1ta) association on chromosome X at SXGnf55.520 continued to be detected at 2 significance of P < 10–5 and the AVP transcript abundance (AVPta) association on chromosome 7 at S07Gnf001.580 was detected at P < 10–6. Comparison of results of suggestive-level marker regression of CRF to Crf-R1 reveals associations of both phenotypes with a cluster of genetic markers on chromosome 12 (Fig. 2C). Similar comparison of Crf-R1 to AVP reveals an area of potential shared genetic regulation on chromosome 7 (Fig. 2B). Marker associations were detected for all four targets on chromosome 12 (Fig. 2C).
TABLE 2. Results of marker regression analysis of CRF-system transcript abundance data
FIG. 2. Results of interval mapping data of CRF system transcript abundance data on five mouse chromosomes. Critical LRS values for mapping were determined with permutation test run through 10,000 iterations for each target (Table 3); CRF (orange), Crf-R1 (green), Crf-BP (red), AVP (blue). A, Chromosome 1; B, chromosome 7; C, chromosome 12; D, chromosome 13; E, chromosome X.
Interval mapping was used to localize the linkage signals detected in the marker regression analysis (Table 2 for chromosomal map location) onto the respective chromosomes (Fig. 2). The critical LRS values for each data set were determined with the permutation test (Table 3). Interval mapping of CRFta revealed LRS peaks on chromosome 1 at P1Ehs1, on chromosome 12 at D12Mit280, and on chromosome 13 between D13Mit63 and S13Gnf043.620. Interval mapping of Crf-R1ta data revealed suggestive LRS peaks on chromosome 7 at D7mit68 and chromosome 12 at D12Mit5, and a significant peak on chromosome X at SXGnf55.520. Interval mapping of Crf-BPta data revealed suggestive LRS peaks on chromosome 7 at D7Mit301 and on chromosome 12 between D12Mit154 and D12Mit235. Interval mapping of AVPta data on chromosome 7 revealed a highly significant LRS peak at S07Gnf001.580, and significant peaks on chromosome 12 at D12Mit136 and chromosome 13 at D13Mit126.
The QTL detected in this analysis do not appear to be linked to the respective structural genes. The structural gene that encodes CRF (designated Crh) is on chromosome 3 (8 cM), Crf-R1 (Crhr1) on chromosome 11 (62 cM), Crf-BP (Crhbp) on chromosome 13 (52 cM), and AVP (Avp) on chromosome 2 (73.2 cM). All of these QTL must be acting in trans to exert an influence on the expression of the target genes, given that they are on different chromosomes. The hypothalamus is a sexually dimorphic organ, and all of the included analyses were performed in males. Thus the influence of gender on these QTL cannot be ascertained until the experiment is repeated in females.
Discussion
The distributions of strain mean transcript abundance values for each of the four targets, CRF, Crf-R1, Crf-BP, and AVP were continuous across the panel of BxD strains, consistent with these phenotypes being quantitative traits. The SDP were transgressive with extreme high and low expressing strains flanking the values of the parent lines. At least one suggestive linkage association was detected with both marker regression and interval mapping for all four targets, and significant associations were detected for CRF, Crf-R1, and AVP transcript abundance. The LRS peaks detected with both methods for these three targets were well above the significance levels set by the permutation test, so at least one QTL controlling hypothalamic CRF-system transcript abundance can be declared for each of these three target genes and there were others for each target in the suggestive range. There are QTL that apparently control expression of only one target and others that potentially impact expression of more than one target gene. Inspection of the QTL detection (Table 2) and localization (Fig. 2) results suggests that expression of CRF and Crf-R1 share elements of genetic regulation on chromosome 12 and AVP and Crf-R1 share elements on chromosome 7. Interestingly, this analysis did not reveal any potential shared genetic regulation between CRF and Crf-BP transcript abundances.
QTL are probabilistically defined entities, and the relationship between these genes and those that have been identified through sequencing of genomic DNA or expressed sequences remains obscure. The pool of potential candidate genes for a QTL is defined by those sequence-defined transcription units coincident with LRS peaks. One of the criteria considered central to connecting a specific candidate gene to a QTL is that biological evidence connecting the activity of the candidate gene to the biological process captured by the quantitative trait must be established (28). However, different phenotype measurements that define quantitative traits could detect the same QTL, if both were involved in the same fundamental process. The CRF system regulates and coordinates stress responses, so QTL that impact these types of behaviors could be considered as potential candidate genes. Proposing candidate genes for these QTL at this point is premature, as any of these LRS peaks could contain 10s to 100s of actual genes, nevertheless, there are a number of intriguing genes that map in the same locations as CRF system transcript abundance QTL.
The QTL Hipp1A, which controls hippocampal weight and cellularity, is located at map position 89 cM, in very close proximity to the CRFta LRS peak at 89.2 cM (11). Comparison of the CRFta data to the WebQTL phenotypes database revealed modest, uncorrected correlations of CRF-Ct values to a number of hippocampal phenotypes such as hippocampal weight (r = 0.599, P < 0.0003), pyramidal cell layer volume (r = 0.529, P < 0.003), total hippocampal volume (r = 0.507, P < 0.005), and 5-bromo-2'-deoxyuridine incorporation into adult hippocampal neurons (r = 0.7552, P < 0.003) (29). These correlations indicate an inverse relationship between hypothalamic CRF transcript abundance and measures of hippocampal volume and cellularity. The structural gene for the brain-specific isoenzyme of creatine kinase (CKB) is at map position 55 cM on chromosome 12, coincident with the most significant CRFta LRS peak detected in this analysis, and disruption of the CKB gene results in a complex phenotype including increased hippocampal mossy fiber field size and impaired habituation and spatial learning (30). The QTL power in slow-wave sleep 2 (Dps2) also maps at this location on chromosome 12 (31). The CRFta LRS peak on chromosome 13 is broad, but does contain the Shc3 gene, which encodes the neuron-specific ShcC protein that participates in intracellular signaling of receptor tyrosine kinases (32, 33).
There appears to be a region of shared genetic regulation for AVP and Crf-R1ta on chromosome 7, but there are few potential candidate genes in this region. The QTL maternal performance (Mprf) does map at 0.5 cM on chromosome 7 and AVP has been implicated in contributing to maternal behaviors (34, 35, 36). The AVPta LRS peak on chromosome 13 at 45 cM is coincident with the Nr2f1 gene which encodes the chicken ovalbumin upstream promoter-transcription factor orphan steroid/ thyroid hormone receptor that is involved in differentiation and development of the CNS (37). The QTL power in slow-wave sleep 1 (Dps1) also maps at this location (31). There are a number of potential candidate genes within the Crf-R1 LRS peak at 30 cM on the X chromosome, including Avpr2, which encodes the type-2 AVP receptor.
There are several limitations to this study. The first is that it was conducted in one set of segregating animals, that being the BxD recombinant inbred set, so the linkage findings need to be confirmed in a separate, independent analysis. These experiments are ongoing, but the statistical significance of some of the associations, in particular for CRFta on chromosome 12, AVPta on chromosome 7, and Crf-R1ta on the X chromosome, argues against these being false-positive findings. That different targets had unique and shared regions of linkage argues in favor of these QTL being elements of the genetic system that controls the hypothalamic CRF system. Another limitation is that all of the animals in this experiment were male, so no conclusions can be drawn as to the genetic regulation of the CRF system in females. The resolution of the mapping algorithms employed, and the density of recombinations in the strain set are not sufficient to engage in fine mapping, so linkages in smaller candidate intervals cannot be detected, thus limiting the resolution of the mapping results. Fine scale mapping of these QTL will be dependent on additional analysis of other sets of animals (segregating F2 cross, advanced intercross recombinant inbred set, etc.) and only with high resolution fine mapping will genuine candidate genes become apparent as the target interval shrinks.
In conclusion, we present evidence that the hypothalamic CRF system is controlled by a series of QTL, some of which are specific to only one target gene and some may exert a more global influence on multiple genes in the CRF system. Transcript abundances of CRF, Crf-R1, Crf-BP, and AVP in total hypothalamic RNA are quantitative traits and we have detected at least one QTL that controls the abundance of each of these transcripts in total hypothalamic RNA.
Acknowledgments
We thank Dr. Robert Williams and Dr. Kenneth Manly of the University of Tennessee Health Science Center, School of Medicine for advice and guidance in conducting these experiments.
References
Nemeroff CB 1996 The corticotropin-releasing factor (CRF) hypothesis of depression: new findings and new directions. Mol Psychiatry 1:336–342
Nemeroff CB, Widerlov E, Bissette G, Walleus H, Karlsson I, Eklund K, Kilts CD, Loosen PT, Vale W 1984 Elevated concentration of CSF corticotropin-releasing factor-like immunoreactivity in depressed patients. Science 226:1342–1344
Bremner JD, Licinio J, Darnell A, Krystal JH, Owens MJ, Southwick SM, Nemeroff CB, Charney DS 1997 Elevated CSF corticotropin-releasing factor concentrations in posttraumatic stress disorder. Am J Psychiatry 154:624–629
Arborelius L, Owens MJ, Plotsky PM, Nemeroff CB 1999 The role of corticotropin-releasing factor in depression and anxiety disorders. J Endocrinol 160:1–12
Heim C, Newport DJ, Heit S, Graham YP, Wilcox M, Bonsall R, Miller AH, Nemeroff CB 2000 Pituitary-adrenal and autonomic responses to stress in women after sexual and physical abuse in childhood. J Am Med Assoc 284:592–597
Belknap JK, Phillips TJ, O‘Toole LA 1992 Quantitative trait loci associated with brain weight in BxD/Ty recombinant inbred mouse strains. Brain Res Bull 29:337–344
Miner LL, Marley RJ 1995 Chromosomal mapping of the psychomotor stimulant effects of cocaine in BxD recombinant inbred mice. Psychopharmacology 122:209–214
Tarricone BJ, Hingtgen JN, Belknap JK, Mitchell SR, Nurnberger JI 1995 Quantitative trait loci associated with behavioral response of BxD recombinant inbred mice to restraint stress: a preliminary communication. Behav Genet 25:489–495
Tarricone BJ, Hwang WG, Hingtgen JN, Mitchell SR, Belknap JK, Nurnberger JI 1995 Identification of a locus on mouse chromosome 17 associated with high-affinity choline uptake using BxD recombinant inbred mice and quantitative trait loci analysis. Genomics 27:161–164
Gora-Masalak G, McClearn GE, Crabbe JC, Phillips TJ, Belknap JK, Plomin R 1991 Use of recombinant inbred strains to identify quantitative trait loci in psychopharmacology. Psychopharmacology 104:413–424
Lu L, Airey DC, Williams RW 2001 Complex trait analysis of the hippocampus: mapping and biometric analysis of two novel gene loci with specific effects on hippocampal structure in mice. J Neurosci 21:3503–3514
Mayeda AR, Hofstetter JR, Belknap JK, Nurnberger JI 1996 Hypothetical quantitative trait loci (QTL) for circadian period of locomotor activity in BxD recombinant inbred strains of mice. Behav Genet 26:505–511
Talbot CJ, Radcliffe RA, Fullerton J, Hitzemann B, Wehner J, Flint J 2003 Fine scale mapping of a genetic locus for conditioned fear. Mamm Genome 14:223–230
Zhang S, Gershenfeld HK 2003 Genetic contributions to body weight in mice: relationship of exploratory behavior to weight. Obes Res 11:828–838
2002 JMP statistics and graphics guide, version 5. Cary, NC: SAS Institute Inc.
Manly K, Cudmore JRH, Meer JM 2001 Map Manager QTX, cross-platform software for genetic mapping. Mamm Genome 12:930–932
Wang J, Williams RW, Manly KF 2003 WebQTL: web-based complex trait analysis. Neuroinformatics 1:299–308
Williams RW, Gu J, Lu L 2001 The genetic structure of recombinant inbred mice: high-resolution consensus maps for complex trait analysis. Genome Biol 2:1–18
Haley CS, Knott SA 1992 A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. Heredity 69:315–324
Darvasi A, Soller M 1997 A simple method to calculate resolving power and confidence interval of QTL map location. Behav Genet 27:125–132
Zeng ZB 1993 Theoretical basis for separation of multiple linked gene effects in mapping quantitative trait loci. Proc Natl Acad Sci USA 90:10972–10976
Zeng ZB 1994 Precision mapping of quantitative trait loci. Genetics 136:1457–1468
Lander ES, Botstein D 1989 Mapping Mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121:185–199
Churchill GA, Doerge RW 1994 Empirical threshold values for quantitative trait mapping. Genetics 138:963–971
Lander ES, Kruglyak L 1995 Genetic dissection of complex traits-Guidelines for interpreting and reporting linkage results. Nat Genet 11:241–247
Blake JA, Richardson JE, Bult CJ, Kadin JA, Eppig JT, Mouse Genome Database Group 2003 The Mouse Genome Database. Nucleic Acids Res 31:193–195
Belknap JK, Mitchell SR, O’Toole LA, Helms ML, Crabbe JC 1995 Type I and type II error rates for quantitative trait loci (QTL) mapping studies using recombinant inbred mouse strains. Behav Genet 26:149–160
Abiola O, Angel JM, Avner P, Bachmanov AA, Belknap JK, Bennett B, Blankenhorn EP, Blizard DA, Bolivar V, Brockmann GA, Buck KJ, Bureau JF, Casley WL, Chesler EJ, Cheverud JM, Churchill GA, Cook M, Crabbe JC, Crusio WE, Darvasi A, de Haan G, Dermant P, Doerge RW, Elliot RW, et al. 2003 The nature and identification of quantitative trait loci: a community‘s view. Nat Rev Genet 4:911–916
Kempermann G, Gage FH 2002 Genetic determinants of adult hippocampal neurogenesis correlate with acquisition, but not probe trial performance in the water maze task. Eur J Neurosci 16:1129–1136
Jost CR, Van Der Zee CE, In’t Zandt HJ, Oerlemans F, Verheij M, Streijger F, Fransen J, Heerschap A, Cools AR, Wieringa B 2002 Creatine kinase B-driven energy transfer in the brain is important for habituation and spatial learning behaviour, mossy fibre field size and determination of seizure susceptibility. Eur J Neurosci 15:1692–1706
Franken P, Chollet D, Tafti M 2001 The homeostatic regulation of sleep need is under genetic control. J Neurosci 21:2610–2621
O‘Bryan JP, Songyang Z, Cantley L, Der CJ, Pawson T 1996 A mammalian adaptor protein with conserved Src homology 2 and phosphotyrosine-binding domains is related to Shc and is specifically expressed in the brain. Proc Natl Acad Sci USA 93:2729–2734
Troglio F, Echart C, Gobbi A, Pawson T, Pelicci PG, De Simoni MG, Pelicci G 2004 The Rai (Shc C) adaptor protein regulates neuronal stress response and protects against cerebral ischemia. Proc Natl Acad Sci USA 101:15476–15481
Peripato AC, De Brito RA, Vaughn TT, Pletscher LS, Matioli SR, Cheverud JM 2002 Quantitative trait loci for maternal performance for offspring survival in mice. Genetics 162:1341–1353
Pedersen CA, Caldwell JD, Johnson MF, Fort SA, Prange A 1985 Oxytocin antiserum delays onset of ovarian steroid-induced maternal behavior. Neuropeptides 6:175–182
Pedersen CA, Ascher JA, Monroe YL, Prange A 1982 Oxytocin induces maternal behavior in virgin female rats. Science 216:648–650
Qiu Y, Pereira FA, DeMayo FJ, Lydon JP, Tsai SY, Tsai MJ 1997 Null mutation of mCOUP-TFI results in defects in morphogenesis of the glossopharyngeal ganglion, axonal projection and arborization. Genes Dev 11:1925–1937(Steven J. Garlow, Ericka )