INTRODUCTION

TOP ABSTRACT INTRODUCTION REFERENCES

Asthma is a disease characterized by airway inflammation, airway hyperresponsiveness, and episodic airway obstruction (1). It is now well established (2) that one of the critical phenotypic traits of human asthma, and an important feature of animal models, is airway hyperresponsiveness (AHR). Airway hyperresponsiveness is defined as a significant narrowing of the airways in response to a provoking stimulus that would be innocuous to an unaffected individual (1). Although AHR is an important phenotype per se, this trait does not constitute clinical asthma; AHR occurs in some normal subjects (6). Individuals exhibiting heightened airway responsiveness, however, are more likely to display the clinical phenotype of asthma. This variation in airway responsiveness could result from either environmental or genetic influences; the relative roles of these differences have been studied in the mouse (7). Examples of environmental influences include activation of mast cells by anti-IgE and exposure to irritant chemicals (8) or antigens (9). The genetic influences mediating AHR are less well understood than environmental influences, and those genetic influences have become the focus of recent investigations in the inbred mouse.

The inbred mouse is an ideal animal to use for studying complex genetic traits. Unlike humans, all members of an inbred strain demonstrate homogeneity at every genetic locus across the entire genome. A mouse strain is regarded as inbred when it has been mated brother to sister for 20 or more consecutive generations; after this time all the loci in the genome are homozygous. The principal objective of inbreeding is to generate mice in which heterozygosity has been effectively eliminated at every genetic locus. Therefore, each mouse in a given inbred strain carries the same allele at a given locus on homologous chromosomes. This feature, i.e., a virtually endless supply of "identical twins," offers distinct advantages for the study of heritable traits over outbred strains. For example, if multiple genes are interacting with the environment to control a phenotypic trait in a given population, the use of inbred strains is better than outbred strains because it eliminates the genetic makeup of animal as a variable. This is not possible in an outbred population where genetic heterogeneity will result in chromosomal regions cosegregating with the disease in some individuals but not in others. Furthermore, when inbred strains are phenotyped, any observed differences in the trait can be directly attributable to environmental variability. The use of inbred mouse strains to study genetic traits is further warranted given the availability of high-resolution chromosomal maps (10) and the significant homology of the murine and human genomes (11). The homology between the human and mouse genome has been used successfully to find disease traits in humans (12). Thus, inbred strains of mice are ideally suited for studying the genetic basis of airway responsiveness.

	AIRWAY RESPONSIVENESS AMONG INBRED MOUSE STRAINS

It has been demonstrated that there is considerable variation, on the order of 300-fold, in airway responsiveness in outbred populations such as the dog (13), rat (14), and guinea pig (15). In contrast, several investigators have systematically shown in inbred strains of mice (3, 5, 16) variation on the order of tenfold. Indeed, the intrastrain variability in the airway responsiveness phenotype among inbred strains is significantly less than the observed interstrain variability. This finding is consistent with the hypothesis that airway responsiveness is a heritable trait. Levitt and co-workers assessed airway responsiveness in inbred mice by measuring the airway pressure-time index (APTI) following an intravenous bolus of either acetylcholine or serotonin. They demonstrated a significant difference in airway responsiveness among nine common inbred strains of mice (4, 5, 17). They determined that the APTI method of phenotyping inbred mice provided a sensitive and reproducible index for separating strains by airway reactivity. Among the nine strains examined (Figure 1), the AKR/J and A/J strains showed the greatest degree of airway responsiveness, while the C57BL/6J, SJL/J and C3H/HeJ strains were the least responsive (17). There was a sixfold difference in airway responsiveness for acetylcholine between the most divergent inbred strains (A/J and C3H/HeJ) (4). Konno and co-workers studied airway responsiveness to methacholine, a congener of acetylcholine, and reported similar strain distribution patterns (3). Bronchial reactivity to methacholine and serotonin was studied in A/J, DBA/2, WBB6F1-W/Wv. WBB6F1-+/+, C3H/ HeN, and C57BL/6 mice. The C57BL/6 and C3H/HeN strains were classified as "low-responder" and the A/J and DBA/2 as "high-responder" strains (3). Recently, strain-related differences in airway reactivity following intravenous challenge with acetylcholine was reported in eight strains of mice by Longphre and Kleeberger (16), who reported a strain distribution pattern similar to that observed previously by Levitt and Mitzner (4). The strains were grouped into hyperresponder (A/J, AKR/ J), middle-responder (C57L/J, 129/J) or hyporesponder (C3H/ HeJ, C57BL/6J) groups according to their bronchial responsiveness to acetylcholine. An investigation into the mechanism(s) underlying the difference in airway responsiveness between the A/J and C3H/HeJ strains suggests that differences in muscarinic receptor signal transduction may explain the enhanced airway hyperreactivity in the A/J strain (18).

View larger version (44K): [in this window] [in a new window]   Figure 1.   Airway pressure-time index response to either acetylcholine (250 nmol/kg) or 5-hydroxytryptamine (50 nmol/kg, intravenously) in different strains of inbred mice. Each symbol represents APTI response for an individual mouse. Bars indicate mean ± SE. (Reprinted from Reference 17 with permission.)

When airway reactivity was assessed using the APTI index in the same nine strains of inbred mice with the bronchoconstrictor serotonin, Levitt and co-workers (5, 17) reported a rank order of airway responsiveness similar, but not identical, to that observed with acetylcholine (Figure 1). Specifically, while the CBA/J, C57BL/6J, SJL/J, and C3H/HeJ strains exhibited a diminished APTI response and the A/J strain exhibited a substantial APTI response, the AKR/J, DBA/2J and WBA/2J, and WB/ReJ reacted differently to the two agonists (5). For example, Levitt and Mitzner reported that the AKR/J strain in particular is among the most reactive to acetylcholine but among the least reactive to serotonin (5). The authors attributed the dissimilarity in responsiveness to serotonin and acetylcholine to different genes that are inherited independently (4, 5). In a more recent study, Konno and co-workers examined bronchial reactivity in A/J, DBA/2, C3H/HeN, C57BL/6, WBB6F1-W/Wv, and WBB6F1-+/+ mice to both methacholine and serotonin and found no difference in the rank order of responsiveness to both agonists (3). Several factors may account for these differences, including methods of assessing airway reactivity and housing conditions.

In a more recent study, Levitt and Ewart evaluated the biological variability of airway responsiveness to atracurium in A/J, AKR/NCr, DBA/2, BALB/CAnNCr, CBA/JCr, C57BL/ 6NCr, SJL, C3H/HeJ, and SWR/J mice (19). Significant phenotypical variability among strains was observed, indicating a genetic predisposition to atracurium-induced bronchoconstriction. While the A/J, AKR/NCr, and DBA/2 strains had a hyperresponsive phenotype, the BALB/CAnNCr, CBA/JCr, C57BL/ 6NCr, SJL, C3H/HeJ, and SWR/J mice reacted minimally to atracurium (Figure 2). The maximal mean APTI index was twentyfold greater in the DBA/2 strain than in the SJL strain. When the strain distribution pattern for atracurium was compared to that of serotonin, there was a close resemblance in the bronchoconstrictor responses. The differences in bronchoconstriction between strains was attributed to strain-specific intrinsic differences in the lung response to atracurium (19).

View larger version (28K): [in this window] [in a new window]   Figure 2.   Airway pressure-time index responses to atracurium (10 mg/kg, intravenously) in nine inbred strains of mice. Each symbol represents the APTI response from an individual mouse. The range bars represent the mean ± SEM. (Reprinted from Reference 19 with permission.)

	MENDELIAN MODELS OF AIRWAY RESPONSIVENESS

Based on segregation analysis of airway responsiveness in a cross between A/J and C3H/HeJ progenitor strains, Levitt and Mitzner (4) proposed that airway responsiveness was inherited as a simple Mendelian trait in a recessive fashion. If this is true, one could pursue a genetic analysis of progeny derived from a cross between hyporesponsive and hyperresponsive mouse strains. The production and phenotyping of F1 (first filial) progeny derived from the parental strains, F2 intercross progeny produced by breeding the F1 progeny (F1 × F1), and backcross (B1) offspring produced by crossing the F1 progeny back to either parental strain (F1 × parent) is indicated for such an analysis. The backcross and F2 offspring are regarded as segregating progeny because genetic loci will segregate independently at meiosis, resulting in a different array of progenitor alleles on each chromosome. The variance of the phenotype in the nonsegregating F1 population represents (among genetically identical individuals) environmental variance, while the variance in the segregating F2 intercross and backcross progeny reflects both genetic and environmental variance. If the ratio of affected versus unaffected progeny in the segregating generations follows a Mendelian mode of inheritance, one can conclude that the genetic trait of interest is controlled by a single gene.

As previously mentioned, Levitt and co-workers carried out a systemic genetic dissection of airway responsiveness to acetylcholine in A/J mice, which exhibit intrinsic airway hyperresponsiveness, and C3H/HeJ mice, which are relatively hyporesponsive (4). In order to phenotype the mice for airway responsiveness, the investigators measured the time-integrated change in peak airway pressure, the APTI, in response to an intravenous infusion of acetylcholine. Based on the APTI assessment of airway reactivity to acetylcholine, they identified two distinct populations among the inbred and hybrid mice. Animals were arbitrarily designated as hyporesponse or hyperresponsive based on whether they achieved a given threshold value in the APTI in response to a single fixed dose of acetylcholine. Although the A/J (designated "A") and C3H/HeJ (designated "C") strains were hyperreactive and hyporeactive, respectively, the F1 progeny (CAF1 and ACF1) were uniformly hyporeactive. When C3H/HeJ mice were crossed with CAF1, the backcross progeny were exclusively hyporeactive. When the F2 intercross (ACF1 × CAF1) were phenotyped using the APTI measurement, two phenotypic groups were easily distinguishable, with an observed 3 to 1 ratio of hyporeactive to hyperreactive progeny (Figure 3); based on these data, the authors concluded that a single autosomal recessive locus regulated acetylcholine-mediated airway reactivity in the C3H/HeJ and A/J cross.

View larger version (22K): [in this window] [in a new window]   Figure 3.   Animals from simple genetic crosses of A/J and C3H/HeJ mice were challenged with 25 µg/kg acetylcholine given intravenously. The relationship of APTI from individual animals of each cross is illustrated. The cross-bars represent the mean ± SEM for each group of animals. Inbreds: A/J, C3H/HeJ. Abbreviations: CAF1 = C3H/HeJ female × A/J male; ACF1 = A/J female × C3H/HeJ male; C3H/HeJ BX = C3H/ HeJ female × CAF1 male; and F2 = ACF1 × CAF1. (Reprinted from Reference 4 with permission.)

In a subsequent study, Levitt and co-workers conducted the same type of segregation analysis using serotonin as the bronchoconstrictor agonist (5). Both DBA/2J (hyperresponsive) and C57BL/6J (hyporesponsive) mice and the progeny of crosses between them were studied. Although the nonsegregating F1 progeny exhibited a phenotype similar to the hyporesponsive C57BL/6J parent, analysis of the segregating progeny led the investigators to conclude that the inheritance of airway hyperreactivity to serotonin could be attributed to a single autosomal recessive gene (5). Although a similar mode of inheritance was found for both agonists, the observation of distinct strain distribution patterns for airway hyperresponsiveness to both acetylcholine and serotonin in segregating F2 mice led the authors to conclude that the genes do not cosegregate, are not closely linked, and therefore are inherited independently. Based on their segregation analysis of acetylcholine- and serotonin-induced bronchoconstriction, Levitt and co-workers concluded that airway hyperresponsiveness in the mouse is controlled by more than one gene and is agonist-specific (5). The exact mechanism underlying the serotonin- induced bronchoconstriction is unclear. Although two previous studies reported inhibition of serotonin-induced bronchoconstriction with atropine, suggesting the involvement of cholinergic pathways (4, 7), a more recent study reported complete inhibition of serotonin-induced bronchoconstriction with a selective serotonin receptor antagonist (20).

	NONMENDELIAN MODELS OF AIRWAY RESPONSIVENESS

While earlier studies suggested that airway responsiveness to acetylcholine and serotonin follows a Mendelian mode of inheritance in inbred mice, more recent studies have provided evidence that airway responsiveness is a complex quantitative trait that does not follow simple inheritance patterns (21, 25). Several advances in genetic analysis have greatly simplified our ability to dissect quantitative traits into discrete genetic factors. In particular, the identification of a comprehensive set of microsatellite markers distributed across the mouse genome that are polymorphic among various inbred strains has made it possible to define linkages to defined areas of the genome by using quantitative trait locus (QTL) linkage analysis (10, 22, 23). This whole genome search employs interval mapping, which uses both phenotypic and genotypic data to estimate the probable genotype and the most likely QTL effect across the whole genome. The identification of regions of the genome linked to the expression of a particular phenotypic trait has been successfully exploited to identify the genetic loci of importance in animal models for hypertension (23), diabetes (22), atherosclerosis (24), and more recently, airway responsiveness (21).

This approach was used to successfully define QTL linked to airway responsiveness in the mouse (21, 25). The two studies published have employed different methods to assess airway responsiveness in the mouse (3, 4, 21, 25). De Sanctis and co-workers used an in vivo measurement of pulmonary resistance (R_L) that provides a sensitive and reliable measure of airflow obstruction. The airway responsiveness phenotype (7, 21) is measured following the intravenous infusion of geometrically increasing doses of methacholine, a congener of acetylcholine not subject to degradation by acetylcholinesterase. From the relationship between the dose of methacholine administered and the measured pulmonary resistance, the effective dose of methacholine required to cause a given increase for example a doubling of pulmonary resistance, which would be termed ED₂₀₀ RLis determined by log-linear interpolation. This procedure is carried out in the A/J (hyperresponsive) and C57BL/6J (hyporesponsive) parental inbred strains and progeny derived from hybrid crosses. The ED₂₀₀ R_L index is derived from the dose-response curve and provides an accurate means of discriminating both extreme and intermediate responders. It is not subject to the quantitative bias, which fails to distinguish among extreme strains, that the single-dose assessment of APTI is subject to.

An examination of the mode of inheritance of the airway hyperresponsiveness phenotype in progeny of a cross between A/J and C57BL/6J mice (21) indicated that the trait was inherited in a dominant fashion (Figure 4). In order to control parental effects on the airway hyperresponsiveness phenotype, F1 mice were obtained from reciprocal crosses (A/J females with C57BL/6J males and C57BL/6J females mated with A/J males). There were no significant differences in mean ED₂₀₀ RL values between the reciprocal crosses, and the mean ED₂₀₀ RL for the F1 progeny was not significantly different from that observed in the A/J parental line. Additional support for airway hyperreactivity being inherited as a dominant rather than a recessive trait comes from a recent study by Longphre and Kleeberger (16). In this study, F1 progeny derived from the hyperreactive AKR/J and hyporeactive C3H/HeJ parental strains were uniformly hyperreactive. In contrast, Levitt and Mitzner (4) reported that the F1 progeny in their cross (A/J × C3H/ HeJ) were hyporeactive and that there were significant differences in the mean APTI values between their reciprocal F1 crosses, suggesting significant parental influences on the phenotype. The different findings may in part be explained by the differences in genetic background of the crosses. When the F1 (C57BL/6J × A/J) progeny were intercrossed and the log ED₂₀₀ RL index analyzed by normal quantile analysis, the airway responsiveness phenotype in the segregating progeny failed to segregate as a simple Mendelian trait; rather, it segregated continuously in a unimodal normal distribution (21). These findings are consistent with airway responsiveness being a polygenic trait. Interestingly, airway responsiveness has also been reported to be normally distributed in human populations (6). De Sanctis and co-workers also gathered further support for a normal distribution of airway responsiveness when they examined 19 different AXB recombinant inbred (RI) strains--inbred strains produced by repeated brother-sister matings of offspring for 20 or more generations from an A/J × C57BL/6J cross. Because recombination can occur at each meiosis during the generation of the RI strains, each strain inherits a unique set of recombinant alleles that differs from the parental strains. Therefore, phenotypes that are under the influence of a single major locus will segregate according to the phenotypes of each parental strain and will be distributed in a bimodal fashion among the RI strains. The observation that airway responsiveness was not bimodally distributed among the 19 RI strains in this study (as would be predicted for a Mendelian trait) supports the hypothesis that airway responsiveness is a polygenic trait (21).

View larger version (13K): [in this window] [in a new window]   Figure 4.   Scatterplot of log ED200 RL from individual animals in parental, F1, F2, and backcross mice. *Significantly different from C57BL/6J mice (p < 0.05). Abbreviations: A = A/J; B = C57BL/6J. (Reprinted from Reference 21 with permission.)

In order to localize QTLs affecting airway responsiveness, De Sanctis and co-workers performed a comprehensive analysis of airway responsiveness in 321 backcross progeny derived from a (A/J × C57BL/6J) F1 × C57BL/6J cross (21). In these animals airway responsiveness was normally distributed, with 50% of the variance attributable to genetic variance and the rest to environmental variance. Given that airway responsiveness to methacholine did not segregate as would be expected with a single locus, De Sanctis and co-workers used Wright's approach (26) to estimate the number of loci responsible for regulating airway responsiveness. This technique assumes that all loci contribute equally to the expression of the phenotype (26). Using this preliminary analysis, De Sanctis and colleagues estimated that two loci were involved in regulating airway responsiveness in the mouse (21).

Selective genotyping of the phenotypically extreme backcross progeny with 157 simple sequence-length polymorphic (SSLP) markers spaced approximately 9 cM apart was first carried out to identify regions of interest (21). The three regions so identified were subsequently genotyped in all 321 backcross progeny with 31 additional markers. Genetic markers that are informative will cosegregate with the phenotype and be localized within a short distance of the QTLs; the closer a marker is to the trait locus, the fewer recombinations it has undergone, and the more informative it is. A correlation between genotypic and phenotypic data revealed three QTLs influencing airway responsiveness. These three loci, designated Bhr1, Bhr2, and Bhr3 for bronchial hyperresponsiveness, mapped close to the murine homolog of several candidate genes that have been implicated in asthma (Figure 5). The respective lod score for the three QTLs were 3.0, 3.7, and 2.83 for Bhr1, Bhr2, and Bhr3 (21). An analysis of variance including cross-terms for two-way and three-way interactions revealed significant evidence of epistatic interactions between the QTLs, pointing to the possibility of important biological interactions among the genes. Thus, using QTL analysis, De Sanctis and co-workers demonstrated for the first time in a mammalian species that airway responsiveness is a polygenic trait.

View larger version (17K): [in this window] [in a new window]   Figure 5.   Calculation of lod scores from genotypes of murine SSLP polymorphic markers on 321 backcross progeny [(C57BL/6J) × C57BL/6J] on chromosomes 2, 15 and 17. (Reprinted from Reference 21 with permission.)

Further support for a complex mode of inheritance comes from a study examining the genetics of airway responsiveness in AKR/J (hyperresponsive) and C3H/HeJ (hyporesponsive) strains and progeny derived from hybrid crosses (16). As mentioned above, segregation analysis in this study demonstrated that airway hyperresponsiveness to acetylcholine was inherited as a dominant trait in the F1 animals. When the investigators tested the hypothesis that airway reactivity to acetylcholine is determined by a single gene, the observed ratios of hyperreactive and hyporeactive phenotypes in the segregating progeny failed to fit the predicted Mendelian proportions (16); this finding was consistent with that reported by De Sanctis and co-workers (21), i.e., that airway responsiveness is not inherited as a simple Mendelian trait.

In contrast to the earlier work by Levitt and Mitzner (4), in which a Mendelian mode of inheritance was reported in the A/J and C3H/HeJ background, Ewart and co-workers (25) recently reported segregation analysis data in the same genetic background, suggesting that acetylcholine-induced airway hyperresponsiveness is not controlled by a single gene. In this study the APTI and respiratory resistance (Rrs) response to a single dose of acetylcholine was used to study the genetics of acetylcholine-induced airway responsiveness in the parental, F1, and segregating progeny (25). Using these two methods of phenotyping, Ewart and co-workers observed discordance in the frequency distributions of the F2 and backcross progeny. They concluded that the two methods of phenotyping measured different physiological parameters. Quantitative trait locus analysis was performed using a panel of polymorphic markers between the parental strains on 42 A/J backcross mice; this yielded a single QTL on mouse chromosome 6 (lod score = 3.11) using the APTI phenotype, but no QTLs when the Rrs phenotypic measurement was used. When multiple inheritance models were tested for good fit, the researchers concluded that a mixed general model, in which the inheritance pattern is determined by one major locus along with a polygenic component, best fit their segregation data.

While there are no straightforward reasons to explain the varying modes of inheritance indicated by the different studies, factors such as genetic background of the mice (i.e., different crosses), gender mixing, ages of the mice, viral antigen- free status of the animals, and method of assessing the airway responsiveness phenotype may be important. Several recent studies all support the notion that airway responsiveness is not inherited as a single Mendelian trait (16, 21, 25). The issues that remain contentious are the specific mode of inheritance and the genetic loci of importance.