If you want a mutant mouse, of course, you can get it if you really want, although the extent to which you must try, try and try (again) will vary. Over the last decade, embryonic stem cell technology has moved from the domain of a few laboratories to that of the scientific community at large. Mutant mice afford unprecedented opportunities to experimentally address gene function in vivo in a malleable experimental context rarely provided by humans with analogous mutations. Consequently, announcement of a 'disease' gene often spawns multiple new research projects in many laboratories that will then compete to be the first to generate, analyse and report on a mouse strain missing the disease-gene homologue. Where a 'hot' gene is concerned, the number of groups at the starting line may be more than a dozen. While this duplication of effort might seem wasteful, the variety of alleles generated by parallel efforts is frequently beneficial in elucidating a gene's function. There are some excellent examples of this; for instance, hypomorphic alleles of tumour suppressors Brca2 (Refs 1, 2) and Brca1 (ref. 3) have been as important as null alleles in offering insight into gene function. Competing groups often examine the effect of a null allele in different genetic backgrounds, which can give dramatically different phenotypes4,5, illustrating the effects of modifier alleles and pointing to genetic interactions which might go undetected if only one group had generated a targeted mutation.
The relative ease with which mutant mice can now be made contrasts with the difficulty of making sense of the results. Phenotypic studies can take a long time, and it is not unusual to be left with the unsatisfying conclusion that a given mutation does not alter the phenotype6. This often reflects experimental limitations in knowing what assays to perform on a highly complex multicellular organism. Analysis of embryonic lethal phenotypes is also challenging, and it is not unusual for competing groups to disagree on the embryological cause of a developmental defect. Recently, independent knockout studies of the tumour suppressor Pten reached different conclusions on the embryonic function of this gene. While both groups agreed that Pten is essential for embryonic development, Antonio Di Cristofano and colleagues7 concluded that Pten is required for the differentiation of endoderm, mesoderm and ectoderm while Akira Suzuki and co-workers8 reported that Pten is required for patterning of the cephalic and caudal regions of the embryo and for placental development. What is the basis for these different conclusions? Failing to observe mutant embryos after embryonic day (E) 7.5, Di Cristofano et al. performed detailed examination of the differentiation of Pten-deficient ES cells in vitro. By way of contrast, Suzuki et al. detected a normal number of mutant embryos at E7.5 (ref. 8). Although these embryos were developmentally retarded, they had gastrulated and all three germ layers were present. More than half of them appeared viable at E8.5, at which time considerable differentiation had occurred. Faced with this discrepancy, one should keep in mind the limitations of using ES cell differentiation as a primary tool to make meaningful conclusions about early embryonic lethal phenotypes without reference to the morphology of the embryo. The Pten mutation described by Suzuki et al. deletes more of the Pten locus (compared with the mutant allele generated by Di Cristofano et al.), and would possibly be expected to give rise to a more severely affected mouse, but it seems likely that both disruptions generate null alleles. While it is possible that modifiers in different genetic backgrounds may explain the milder embryological phenotype described by Suzuki et al., more detailed analyses are required to discern the cause.
This is a preview of subscription content, access via your institution