The complicated behaviors that are modulated by the opioid system are undoubt- edly polygenic. Thus, the genetic context in which opioid pathways are studied is likely to have a significant influence on these behaviors and, consequently, exper- imental results. Different strains of mice diverge for their basal sensitivities in nociceptive assays, their analgesic sensitivities to opioid drugs, and their propen- sity to self-administer opioid drugs. Since these behaviors are influenced by
multiple neurotransmitter systems, it is expected that the source of strain variation will not be found in only one gene. Experiments using targeted genetic mutations must be designed carefully so that a genetically equivalent wild-type control is used, and interpretations of experimental results must be in the context of the particular mouse strain used. We also must be aware of the potential role genetic background can have in determining phenotypes and some of the idiosyncrasies of specific mouse strains. Of utmost importance is the adoption of standards for maintaining mutant strains of mice.
A flawed, but commonly used, technique is to maintain a line of mutant mice and a parallel line of wild-type mice by inbreeding F2 homozygous littermates. Super- ficially, this is an enticing method because it eliminates the need for genotyping and is economical. However, this method is flawed because the genetic background of such inbred mice is not known and cannot be reproduced. In fact, after 20 generations of brother–sister matings, new recombinant inbred lines will be generated whose genetic makeup, in addition to the targeted locus, can differ drastically. These differences arise because of random segregation and fixation of polymorphic alleles, including genes linked to the targeted locus, that are unique to the wild-type or the targeted line. Thus, it is preferable to avoid this scheme at all costs. If this method has been used, then the targeted mutation should be transferred to standard inbred backgrounds, which can be done in 10 generations of backcrosses.
3.8.1 BACKCROSSING TECHNIQUES
The method the Banbury Conference recommends is maintaining a mutant mouse line on a known genetic background. For example, chimeras are mated to both a 129 strain (origin of the ES cells and mutant locus) and a C57BL/6 strain (origin of the blastocyst).123 Heterozygous mice are then repeatedly backcrossed in parallel on the inbred C57BL/6 and 129 lines, such that congenic heterozygotes are created on both backgrounds. Heterozygotes from the two congenic strains are then crossed to produce F1 hybrid mice in Mendelian proportions of 1 wild-type:2 heterozygotes:2 homozygotes for experiments. F1 hybrid mice have the advantages of reproductive vigor and the absence of deleterious recessive traits from both of the background strains. Unfortunately, this breeding procedure is often not possible because some behaviors to be studied are only present in certain inbred strains, and the procedure is prohibitively expensive for most laboratories. For example, studying a double mutant by this method would require massive numbers of mating pairs because double mutants would only be 1/16 of the F1 progeny of double heterozygous parents.
A less expensive alternative is to backcross the chimeric mice onto only one inbred strain and use littermates from heterozygote matings as the experimental subjects.
However, this method requires some knowledge of the strain to be used and should not be done without a comprehensive literature review and pilot study.
Ten backcrosses are required to produce 99.9% congenic mice. However, after only 5 generations the genome is nearly 97% congenic to the background strains.
These mice could be used, at the least for pilot studies, although the backcrossing procedure should continue indefinitely. It is possible to expedite this process to as little as 3 to 4 generations with a procedure known as speed congenics.124,125 This
process requires PCR genotyping with strain-specific genetic markers. Genotyping assists in selecting for mice that are represented by the maximum amount of chromo- somes from the recipient strain. Approximately 150 different markers were used resulting in a map density of about 10 cM.124 In addition, we have postulated that the typical interval between generations can be shortened. Heterozygote 25-day-old females are immediately genotyped and superovulated by standard methods.18 They are then mated with wild-type, mature males from the recipient strain and mating plugs are checked. Parallel females of an outbred strain are mated with vasectomized males to use as host mothers. On the fourth day following coitus, blastocyst stage embryos can be harvested and transferred to the pseudopregnant females via inter- uterine transfer. We estimate that the average time between generations can be shortened by at least 2 weeks. Additionally, if C57BL/6 mice are the recipient strain, one will find that the first pregnancy is often lost or not nursed. The use of host mother from an outbred strain increases the chances of a healthy litter.
3.8.2 PHENOTYPIC DIFFERENCES OF SOME INBRED STRAINS
Many studies of null mutations have reported learning deficits, but many inbred lines have phenotypic abnormalities in learning themselves. For example, several 129 lines have spatial learning deficits,126 and most 129 and DBA strains show poor hippocampal-dependent learning.127,128 Mice from the C57BL/6 line become deaf to certain frequencies at an early age,129 and they are poor avoidance learners.130 Studies that are especially relevant to experiments on the opioid system have shown that the C57BL/6 strain is considerably more tolerant to the analgesic effects of mu-selective opioids such as morphine and DAMGO compared to other strains.131,132 These strain differences are not limited to performances on behavioral tests. For example, kainic acid-induced seizures occur in the absence of hippocampal neuronal degeneration in C57BL/6 and Balb/C mice,133,134 and the corpus callosum is defective in many mice from BALB and 129 substrains.135–137 An F1 hybrid cross of inbred lines often eliminates some of these abnormalities because homozygous recessive genes are often present in the inbred strains. For example, of all inbred lines tested, the C57BL/6 is the best performer on the Morris water maze, yet all F1 hybrid lines performed better than the C57BL/6 mice.128 A null mutation of the Thy-1 gene was made in ES cells from both C57BL/6 and 129/Sv/Ev and resultant outbred lines of all combinations showed that genetic background strongly influenced initial learning of the Morris water maze.41 If the Morris water maze is an intended assay, then an F1 hybrid might be the preferred genetic background for studying the effects of single gene deletions that are expected to decrease spatial learning.
A fundamental problem of each of the breeding strategies described is the phenom- enon of “hitchhiking genes” closely linked to the targeted genetic locus.138 These alleles are derived only from the genome of the ES cells and even after 12 generations of backcrossing to produce a congenic line they may represent as much as 16 cM of the ES genome, or 1% of the total genome containing 50 to 100 genes.123 Virtually all reliable ES cells are derived from 129 substrain mice, so these closely linked hitchhiking genes are 129 alleles. In addition to these closely linked genes the substrains of 129 can vary significantly.40 Standardized behavioral testing of these substrains is incomplete
and many are no longer commercially available. Moreover, certain substrains are noto- rious for inherited neurological disorders126 as well as low fecundity. There are at least two instances in which phenotypic differences between induced mutant mice and wild- type control mice have been attributed to closely linked gene alleles originating from the 129 background strain, rather than the targeted mutation itself.133,134,139 Remedies for this problem are not easy. The most desirable, but still not available, solution will be the development of a new series of ES cells derived from a panel of different inbred strains that can be used for gene targeting and are easily available.