| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the *Department of Anesthesia and Perioperative Care, University of California, San Francisco, CA; The Waggoner Center for Alcohol and Addiction Research, University of Texas, Austin; Department of Biostatistics, University of California, San Francisco, CA; and Department of Genetics and Genomics, Roche Palo Alto, Palo Alto, CA.
Address correspondence to Dr. Sonner, Department of Anesthesia and Perioperative Care, S-455i, University of California, San Francisco, CA 94143-0464. Address e-mail to sonnerj{at}anesthesia.ucsf.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: To accomplish this, two inbred mouse strains differing in isoflurane MAC, the C57BL/6J and LP/J mouse strains, were bred through two generations to produce genetic recombination. These animals were genotyped for microsatellite markers. We also applied an independent, computational method for identifying QTL-regulating differences in isoflurane MAC. In this approach, the isoflurane MAC was measured in a panel of 19 inbred strains, and computationally searched for genomic intervals where the pattern of genetic variation, based on single nucleotide polymorphisms, correlated with the differences in isoflurane MAC among inbred strains.
RESULTS AND CONCLUSIONS: Both methods of genetic analysis identified a QTL for isoflurane MAC that was located on the proximal part of mouse chromosome 7.
|
Introduction |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The goal of this study was to detect a QTL underlying the difference in MAC between two inbred mouse strains. QTLs can be detected and mapped using linkage analysis, which tests the cosegregation of the trait of interest (MAC, in this case) and a genetic marker locus of known position on a chromosome (2). We used microsatellites (3) as our genetic markers. Microsatellites are repeats of short sequences. They are valuable for mapping purposes because they are plentiful, polymorphic, codominant, presumably neutral with respect to most traits and reproductive fitness, and easily typed via the polymerase chain reaction (PCR).
MAC is easily measured in inbred mouse strains. We have reported the variation in MAC for three conventional inhaled anesthetics, including isoflurane, among 15 inbred mouse strains (4). At the time of that investigation, genetic maps were available for nine of these inbred strains (5). To identify genetic regions responsible for differences between strains in isoflurane MAC, we prepared backcross progeny from two inbred strains, C57BL/6J and LP/J, that had significant differences in isoflurane MAC (mean MAC ± sd was 1.46% ± 0.06% atmospheres for C57BL/6J (n = 24 mice), and 1.74% ± 0.06% atmospheres for LP/J (n = 24 mice) (4).
By breeding female C57BL/6J mice with male LP/J mice, we generated F1 mice, and then backcrossed F1 male mice to female C57BL/6J mice to produce a set of 90 BC1 (first backcross generation) mice for analysis. The isoflurane MAC was measured in the BC1 mice, and offspring with higher or lower MACs were genotyped for a panel of microsatellite markers to identify QTL regulating the difference in isoflurance MAC. In addition to this experimental backcross, we also used a computational (in silico) method for genetic mapping (6). In this approach, the isoflurane MAC was measured in a panel of 19 inbred strains, and computationally searched for genomic intervals where the pattern of genetic variation based on single nucleotide polymorphisms (SNPs) correlated with the differences in isoflurane MAC among inbred strains (6). Both methods of genetic analysis identified a QTL for isoflurane MAC that was located on the proximal part of mouse chromosome 7.
|
METHODS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Isoflurane MAC was measured in 19 strains of mice, using previously published methods (4).
Female C57BL/6J and male LP/J mice (Jackson Labs, Bar Harbor, ME) were bred together to produce F1 offspring. F1 males were then backcrossed to female C57BL/6J mice. MAC to isoflurane was measured in triplicate in 90 backcross progeny, and one time in F1s and the parental (LP/J and C57BL/6J) mice (Fig. 1). At least 1 wk separated each MAC determination in a given mouse.
|
The 20 mice with the highest and the 20 with the lowest mean isoflurane MACs were chosen for genotyping. DNA was extracted from mouse livers using Invitrogen's Easy DNA kit (Carlsbad, CA).
Table 1 lists the microsatellites chosen for genotyping (for primer sequences, see: http://informatics.jax.org/javawi2/servlet/WIFetch?page=markerQF). The choice of microsatellites was based on the distribution of the markers throughout the mouse genome and their presumed ease of genotyping (i.e., the markers had to differ by at least two base pairs in the C57BL/6J and LP/J strains). PCR genotyping was performed using primers purchased from Research Genetics (now Invitrogen, Carlsbad, CA). Tfl DNA polymerase (from the thermophilic bacterium Thermus flavus), enhancer, buffers, and dNTPs were purchased from Epicenter (Madison, WI). Reactions used 1.20 µL MgCl2, 1.50 µL enhancer with betaine, 0.75 µL PCR buffer, 0.38 µL dNTPs, 0.30 µL Tfl polymerase (concentration 1 U/µL), 0.45 µL forward primer (concentration 6 µM), 0.45 µL reverse primer (concentration 6 µM), 6.00 µL water, and 4 µL genomic DNA. We used a touchdown PCR method in which cycles of annealing and extension were started above the calculated annealing temperature (7), to minimize nonspecific priming. The thermal cycler protocol (using a MJR model PC100 thermal cycler) started at 94°C for 2 min followed by two cycles of annealing and extension at a temperature 8°C above the calculated annealing temperature for the primer pair. The annealing temperature was then decreased by 1 degree and two PCR cycles performed at this lower temperature. This was continued until the annealing temperature was 5°C below the calculated annealing temperature for the primer pair, at which point 30 PCR cycles were performed to amplify the microsatellite.
|
PCR genotyping products varied in length from 96 to 346 bp. The C57BL/6J and LP/J microsatellites differed by 2–66 bp, but primers were typically chosen where possible to produce products that differed by 10–20 bp, to improve the ease of distinguishing C57BL/6J from LP/J microsatellites. To resolve DNA sequences as small as 2 bp, PCR products were run on 3.5% MetaPhor agarose (Fisher Scientific) gels in TBE buffer, then stained with ethidium bromide for visualization. The genotype of the microsatellite marker in question was determined to be homozygous for C57BL/6J if only one band was present by electrophoresis at the expected mobility, which was inferred by comparison with molecular weight markers and PCR products from C57BL/6J genomic DNA. The genotype of the mouse was heterozygous at a given marker if two bands (i.e., one for C57BL/6J and one for LP/J) were present. Replicate samples were run as needed to resolve ambiguity in genotyping.
A statistical analysis was used to correlate microsatellites with MAC values. The identification of statistically significant correlations indicates that the chromosome regions associated with those microsatellites harbor genes that explain some of the variation in MAC. In this study, QTLs were mapped in the experimental backcross using the statistical methods of Sen and Churchill (8). Significance was determined by a permutation test with 1000 permutations. A permutation test (9) is an empirical method, based on the experimental data, for estimating threshold LOD values for QTL studies (LOD values are the log base 10 of the odds, which is commonly used in genetics to estimate linkage of two loci). With the permutation test, MAC values are randomly assigned to the genotypes for each mouse. Then the LOD score is calculated. This process is repeated with a new, random assignment of MAC values to genotypes each time. The random assignment assures that there is no association between MAC values and genotypes. In the end, the distribution of the LOD score under the null hypothesis that there is no association between MAC and genotype is obtained. LOD scores more extreme than the 950th highest score are at or above the genome-wide 5% threshold. Those above the 990th highest score are at or above the genome-wide 1% threshold.
QTLs were mapped in silico using the strain MACs by previously described methods (6). The linkage prediction program which implemented this method scanned the SNP database and correlated the genetic distance (from SNPs) between strain pairs with MAC distances between strains, thereby identifying SNP markers linked to strain differences in MAC.
|
RESULTS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Genotyping the 20 backcross mice with higher isoflurane MACs and the 20 with lower MACs yielded three loci with a LOD scores exceeding 1 on chromosomes 4, 7, and 9. The highest LOD score was 1.7 for marker D7MIT246 at 15.0 cM on chromosome 7, which corresponds to a genome position of 30.0 Mb in NCBI mouse build 36 coordinates. In this study, the genome-wide 5% threshold was 1.34 and the 1% threshold was 1.64. The chromosome 7 QTL is statistically significant, with P < 0.01. It should be emphasized that this is the genome-wide Type I error rate corrected for genome-wide multiple comparisons. Based on this significance cutoff, the peaks on chromosome 9 and 4 were not significant.
The chromosome 7 microsatellite marker linked to the QTL in the C57BL/6J (i.e., lower MAC) strain segregated in the backcross offspring with higher MACs, while the markers for the LP/J (higher MAC strain) segregated in the backcross progeny with lower MACs.
A second, independent approach was used to identify chromosomal regions affecting isoflurane MAC in mice. The measured isoflurane MAC for 19 strains was used for computational genetic analysis. A QTL for isoflurane spanning a region from approximately 10 to 50 Mb on mouse chromosome 7 was identified, which was highly significant (the z score = number of standard deviation was 6.76, calculated from SNP allele frequencies). Of note, this computationally identified region encompasses the marker identified through analysis of the backcross progeny.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This produced the genetically diverse population that was required. We measured MAC in these offspring, and genotyped the offspring with extreme MAC values. Using statistical methods to identify the association of MAC with molecular markers, we identified a QTL on the proximal part of chromosome 7 influencing the difference in MAC in the parental strains.
In the second method, we measured MAC in 19 inbred strains for which dense SNP maps had been made. Using a computational method, we correlated the MAC difference between strains with the SNP maps. This detected the same QTL on chromosome 7 as the experimental backcross and gave an approximate map position for the QTL.
Genes identified as important to anesthesia in mice are likely to be important to anesthesia in humans because there is a high degree of homology between the genes of mice and humans. This, together with the fact that large numbers of mice can be bred and studied, provides a powerful rationale for using mice as a model system in studies of anesthetic mechanisms. However, to our knowledge, all studies of MAC in mice have used reverse genetic approaches, in which the function of a candidate gene is altered, a genetically modified animal is made, and the effect on the phenotype of interest (e.g., MAC) is evaluated (10). In this study, we chose to use a different but complementary forward genetic approach: the analysis of natural variation via QTL methods. All forward genetic methods start with the phenotype, rather than a candidate gene, to identify the genetic basis for observed phenotypic differences. In contrast to genetic approaches in which mutations are made, the QTL approach deals with the entire system as it occurs in nature. It is an intrinsically multivariate method which analyzes the effect of many genes, albeit with smaller effects than those which are usually genetically engineered into mice. Those genes are, however, the ones which are actually involved in producing the phenotype in nature.
Our primary goal in identifying QTLs for isoflurane MAC is to aid in finding genes responsible for the immobilizing effect of isoflurane. Typically, many genes are located in QTLs; consequently, locating the gene or genes responsible for the quantitative variation described by the QTL can be difficult. The chromosome 7 QTL we have identified contains many genes which might plausibly mediate the effects of isoflurane. They include a variety of ion channels, which are often considered the mediators of inhaled anesthetic action. Whether one of these, or a gene that does not code for an ion channel, or some other genetic element underlies our QTL, is the focus of continuing investigation.
Of interest, the chromosome 7 QTL we identified shows dominance in the opposite direction from that expected. That is, the QTL from the lower MAC strain (C57BL/6J) appears in the BC1 progeny with the higher MACs, and the QTL from the higher MAC strain (LP/J) appears in the BC1 progeny with the lower MACs. This is not an uncommon observation in QTL studies. It occurs because the dominance predicted from the F1 phenotype is the sum total dominance of all genes acting together. That is, the F1 phenotype resembles that of the LP/J parent, and thus the QTL from the LP/J parent might be expected to be present in the BC1 progeny with the higher MACs. However, the MAC of these progeny is the result of all QTLs acting together, not simply the chromosome 7 QTL. The observation that the chromosome 7 QTL from the LP/J parent is present in the BC1 mice with lower MACs is of note because it indicates the presence of other QTLs with counterbalancing effects.
This is the first QTL study examining MAC for an inhaled anesthetic. Other studies examining loss of righting reflex for various anesthetics have been conducted. It is of interest that QTLs on chromosome 7 have been identified in several of these studies. For instance, a strong QTL for propofol sensitivity, assessed using duration of loss of righting reflex in long sleep and short sleep mice, is located between 71.4 and 89.7 Mb on chromosome 7 (approximately 44 cM on the genetic map). (11) QTLs linked to this same locus, again using latency to loss of the righting reflex in long sleep and short sleep mice, have been identified for sensitivity to alcohol (12) and chloral hydrate (13). Given the map position, and the observation that isoflurane righting reflex does not differ in these mice (14), this probably represents a different QTL than that found here. A QTL at 29.0 cM on chromosome 7 was found using latency to loss of righting reflex to 4% ether in a backcross of C57BL/6J amd MSM/Ms mice (15). The gene underlying this QTL may possibly be the same as ours.
Our experimental plan of attack for identifying genes underlying QTLs will use methods that have been successfully applied for many traits (16–20). Now that a QTL has been identified, the next step is to analyze more backcross or intercross progeny to map the QTL to a manageable interval on chromosome 7. This will reduce the number of candidate genes in the QTL; in addition, it will probably lead to the identification of QTLs on other chromosomes. The number of candidate genes can be restricted further by determining which genes in the QTL are expressed in the central nervous system. Since isoflurane MAC is predominantly mediated by the spinal cord (21), analysis of gene expression in the spinal cord is probably the preferred approach. This analysis may reveal that differential expression of a gene is the basis for the QTL. Alternately, there may be a qualitative difference in the function of genes in the LP/J and C57BL6/J strains which influences isoflurane MAC in the BC1 mice. This may initially be revealed as coding sequence differences between the strains, requiring confirmation by functional evaluation of the product of the candidate gene. Finally, new computational methods, based on haplotype mapping (22) combined with detailed SNP maps, hold the promise of identification of genes underlying QTLs based on the association of isoflurane MAC and haplotypes, provided that SNP maps and MAC are determined in a large enough number of strains.
In summary, we report the identification of a QTL for isoflurane MAC on the proximal part of mouse chromosome 7 using an experimental backcross in C57BL/6J and LP/J mouse strains, and independently using an in silico method.
|
ACKNOWLEDGMENTS |
|---|
|
Footnotes |
|---|
This work was supported in part by a grant from the UCSF Academic Senate (to JMS) and by NIGMS P01 GM47818 (to EIE and JMS) and R01 GM069379 (to JMS).
Dr. Eger is a paid consultant to Baxter Healthcare Corp.
Reprints will not be available from the author.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
gene expression. Science 2004;306:690–5
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
