Evaluation of Targeted Mutations
Pfizer Central Research
Our mission at Pfizer is different from the one Dr. Johnson described in that we most often are trying to look at targeted mutations. We have tried to combine our internal expertise to be able to phenotype genetically altered mice. I think if you survey the literature for genetically altered mice, you will find that often phenotyping is inadequate or incomplete. We have attempted to avoid repeating errors of the past by formulating a systematic team approach toward phenotyping.
The necessary speed of analysis in this very competitive industry and the complexity of the science involved in both the production and analysis of these mice have been the two primary driving forces in the development of this team approach. The approach has also been the outcome of a deliberate increase in the working relationship between our discovery group; our core genetic facility that produces transgenic, knock-in, and knock-out mice; our pathologists; and our laboratory animal clinicians, who are responsible for the care of these mice.
A typical core team consists of a principal investigator (a biologist from any area in drug discovery), one or more molecular biologists (responsible for making the constructs), a pathologist and/or clinical pathologist (depending on the genetic alteration), a laboratory animal clinician, and other specialists (depending on the gene being targeted). For these teams to work quickly and efficiently, they must be able to communicate effectively. Each member must maintain a basic understanding of the technologies used in creating these mice. Team members speak with each other because technologies will always affect phenotype. Keeping up with all of the technologies being produced is becoming increasingly a challenge.
In the beginning, we had only transgenic technology, which affected the philogenetic background of the mice (quite rarely on FNB). In addition, posi
tional effects and even insertional mutagenesis produced phenotypic surprises quite often. With knock-in and knock-out mice, we were able to avoid those surprises, but positional effects, as with transgenic mice, are present throughout the ontogeny of the animal. So we still must deal with the problem of embryonic and fetal lethals.
Today we have a whole array of tissue-specific gene promoters, heterologous recombination systems. These systems allow us to avoid many of the developmental problems, but they create new problems. Many of the inducible systems, for instance, use drugs such as tetracycline or dexamethasone, which can produce effects in bone or in the thymus, both in the morphology and in the function.
For our mouse plan to progress smoothly, we develop our phenotyping plan in advance. When possible, we want to customize each plan to fit the particular mouse project and the circumstances. We try to encourage the principal investigators to assemble all of the team members to solicit their input so that we are assured of having the required reagents and of identifying the necessary techniques. With this approach, we also can begin breeding and husbandry plans, ensuring that we have adequate numbers of mice for later analysis. Finally, we make every effort to utilize the best scientific practices for obtaining accurate phenotypes.
We customize the phenotyping plans because we often want very targeted, specific models. We want to be able to focus the evaluations, to narrow the scope of the evaluations to identify the critical determinants of that model. Then, as soon as a mouse is produced, we confirm the functional success of the alteration by molecular phenotyping. Our methods of choice are, as usual, Northern blots or reverse transcription polymerase chain reaction. For our primary characterization of changes in gene expression, we use Western blots and enzyme-linked immunosorbent assay primarily for analyzing the changes in protein expression. These methods are non-slide based. The more advanced ones we use are in situ hybridization and immunohistochemistry, which is slide based and performed by a pathologist. We are increasingly finding that gene microarrays are useful when the team is interested in determining which secondary changes have been induced by the targeted genetic alteration.
We complete the molecular phenotyping before we schedule any additional analysis. We follow our primary molecular phenotyping with pathology, including gross examination pathology and clinical pathology in most of these plans. For a full validation of any phenotype, we perform a comprehensive pathology evaluation so that we can identify anticipated as well as unanticipated changes in phenotype. Often the secondary changes will render these models useless for us.
Timing is an important aspect of the evaluations. We first try to determine the timing for evaluation of the mice rationally, and if that is not possible, we use clinical signs. If we are not able to determine the best time for evaluation, we use a periodic default timing, which targets puberty, adulthood, and full maturity, as
needed, to characterize the phenotype. When we are looking for targeted models, we perform full gross examinations. We select the organs we will evaluate for routine histology. If it is intended to be a model for Alzheimer's disease, we will look (perhaps only) at the brain. We may include major organs if those are not among the target organs. Then we select either routine or special clinical chemistry and hematology tests, again, depending on our knowledge of the targeted gene.
Once the model has been shown to have the critical determinants, we proceed to the comprehensive pathology evaluation. In instances in which the purpose of the analysis is simply to characterize the phenotype (which it often is), when we are exploring the function of a newly discovered gene, we will go directly into comprehensive pathology. We then perform full gross examination to obtain histopathology on a full set (about 40) tissues. We run a routine clinical chemistry panel, hematology, and other special tests as needed, such as hormones or cytokines, again depending on the target gene.
The best scientific practices for accurate phenotyping require experienced molecular laboratories. Laboratories should be accustomed to working with RNA and particularly the immunohistochemistry that now utilizes custom anti-bodies for analyzing these mice, which can often pose particular challenges.
Pathologists who perform the analysis should also be experienced in rodent pathology and in situ electrotechniques. It is necessary to understand the genetic background of these mice and be familiar with the spontaneous as well as agerelated lesions that can occur. Pathologists must also be able to trouble shoot the in situ techniques and understand the procedures and the common problems that can arise. We often encounter embryonic or fetal lethals in which cases we use specialists (either developmental biologists or pathologists) who are trained in murine development. Finally, we always insist on adequate sample sizes, appropriate control mice, and an environment controlled for feed, light, and housing (particularly if they are models for the study of cancer or endocrinology).
Sometimes teams can be very awkward vehicles for solving problems. However, we have found that at Pfizer, we are able to phenotype mice quickly, while making the best use of our resources, by working in teams. For us, working in teams is really the only way to proceed.