by
Abstract
Once it was thought impossible to create meaningful models of human disease using mere mice. Today, mice are used routinely to develop new therapeutic strategies to treat illness. In fact, they are among the most exciting models for medical research, as discussed in the "Conditional Genetic Technologies in the Mouse" conference in Cold Spring Harbor, New York, held from August 31 to September 2, 1998. Researchers gathered to discuss the limitations of, recent developments in, and exciting information gained using site and time-specific mutations in the mouse.
Imagine altering mammalian DNA in such a way as to make one think harder, sleep less, or eat more. But why imagine? We already have, at least in mice.
To date, thousands of mice (or mice-to-be) have had portions of their genes
mutated, moved, augmented, and knocked out, leading to offspring with
functions altered, novel, increased, or lost. In some instances, the
expression of extra DNA inserted into a mouse inhibits the function of its
own endogenous genes.
This revolutionary manipulation of the mouse genome began in 1974, when direct injection of DNA into fertilized mouse eggs (blastocysts) began efficiently producing genetically altered (transgenic) mice for study. But DNA integration cannot be controlled: the foreign, inserted DNA usually incorporates at a single but not predictable genetic site, or locus. Occasionally, the DNA will enter several loci, disrupting critical genes in the mouse genome and causing multiple undesired effects, or even death.
Mouse geneticists solved this problem by transplanting exogenous
DNA-containing embryonic step cells (ES cells) into blastocysts.
Blastocysts are pluripotent (can differentiate into any kind of mouse
cell) and are derived from ES cells themselves. Researchers first
transplanted donor ES cells from one blastocyst to another in 1984, where
they contributed to the formation of all kinds of host tissues. In 1986,
they transplanted ES cells that had incorporated viral DNA. Mammalian
DNA-containing ES cells soon followed.
ES cells allowed researchers to select for correctly incorporated donor DNA prior to transplantation. For any gene to be expressed, it must be in the correct orientation. Insert DNA cannot incorporate upside-down or backward, and it must follow directly behind a specific type of DNA sequence called promoters. Promoter regions "promote" or drive the expression of a gene or group of genes. Researchers can microinject into ES cells a normal or mutant gene and marker without a promoter, such as an an antibiotic resistance gene. Only genes that have inserted themselves directly behind compatible promoter regions express the gene of interest and marker. ES cells in culture that survive treatment with antibiotics (i.e., show expression of the marker gene) are used for injection into mouse host blastocysts. It was an ingenious method of selection.
Researchers had developed the ability to add fragments of exogenous DNA to
specific sites in the mouse genome. Why is this so important? Capitalizing
on species differences, one could examine the gene regulation and expression
of inserted non-mouse DNA. For example, when researchers inserted the
luciferase gene from fireflies into mammalian cells, they lit up. This
confirmed that luciferase protein is responsible for the glowing properties
of the firefly. Perhaps most important, one could develop model systems of
human disease. Foreign DNA in the mouse genome can lead to the
overproduction of normal or mutant protein, which occurs in conditions such
as Down's syndrome and cystic fibrosis.
Despite all of these advances, there were many remaining limitations for Daniel Metzger of the Institut de Génétique et de Biologie Moléculaire et Cellulaire in Strasbourg, France, to outline at the Cold Spring Harbor symposium. Sometimes expression of the marker DNA one uses for ES cell selection interferes with the genetic characteristic under investigation. For example, several researchers use the lacZ marker gene. This gene expresses a protein that stains cells dark blue. This would have been a poor marker for examining the expression of the firefly luciferase gene. Also, since mammalian development is characterized by redundant function to ensure maturation and viability, exogenous normal or mutant DNA might not result in a discernible phenotype if some compensation inherent in the developmental process in the mouse somehow masks its effects. Finally, insertion of multifunctional DNA can lead to confusing results in the host mouse.
Metzger and others proposed, as a solution, the use of spatiotemporally controlled mutations. Concomitantly controlling the space (organ system) and time (stage of development) in which a particular DNA insert is expressed could overcome many of these and other problems.
For example, spatial control could discern those effects that are
independent of the cell type in which the DNA is expressed - those whose
effects vary depending on where the insert functions. A mutant form of the
insulin gene making altered protein in non-pancreatic cells is likely to
have an effect on the mouse because insulin travels to where it is needed
regardless of where it is made. On the other hand, mutant non-functional
insulin receptor DNA may not effect the mouse if it is expressed in the
wrong cell types. Temporal control could prevent the early death of
transgenic mice from lethal effects of inserting or knocking out a gene
activity, and avoid the confusing compensation effects of normal
development.
So how is this done? The most common way of achieving spatiotemporal gene regulation requires creating and then crossing two separate transgenic mouse lines, which is clever but complicated. One parent lineage contains a recombinase gene under the control of a promoter that can be turned on or off at will. When it is turned on, the recombinase gene will make recombinase enzyme that recognizes specific DNA sequences, recombinase target sites (RTSs), and in the presence of those sites, cuts the DNA. Therefore, if a gene is surrounded by two RTSs, recombinase will cut the DNA at those two sites, remove the gene, and re-ligate DNA. Recombinase and RTSs are found only in bacteria, so putting the recombinase gene into a mouse will not cut up its own DNA. This transgenic mouse line is called the "deleter line" for its capacity to delete.
The other parent lineage is called the reporter line. A reporter mouse
contains a ubiquitous promoter followed by a marker that is surrounded by
RTSs and the insert DNA of interest. Because reporter mice do not express
recombinase, the marker will not be removed and it will get expressed. The
DNA insert of interest doesn't get expressed because the promoter is too far
away. Promoters can be very picky about driving the expression of a gene.
They won't drive just any gene. Some will drive only one gene (like this
one), and some will drive more than one at a time, as previously mentioned.
Some are low-level expressors, and some high. Some drive expression only in
certain tissues, and others at certain times (e.g., development, disease).
So these two parents are "silent." But their offspring contains both transferred genes. In the progeny, the recombinase enzyme is active and it recognizes the RTSs, cutting out the intervening marker DNA. The promoter can then turn on the activity of the gene of interest. Because the recombinase gene is under the control of an inducible promoter, the promoter can be turned on in a dose-dependent manner, expressing the gene of interest at any time.
One can also use tissue-specific promoters in the deleter line or, even
better, use an inducible promoter in the deleter line and a tissue-specific
promoter in the reporter line. Now one can turn a gene on and off in a
particular tissue at will. Formerly, using only DNA microinjection and ES
cell transplantation technologies, normal and mutant DNA inserts could
result in multiple deleterious effects, or even death because the gene was
expressed in the wrong place and/or at the wrong time, and in the wrong
amount - it was all up to the one promoter.
One researcher described inducible gene expression as the difference between turning off a night-light with a hammer or a switch. Both accomplish the task, but only one will let you turn on the light again later. Spatial control is having night-lights on in some rooms, but not in others, and dose-dependent induction is like using a dimmer switch.
Several attendees at the conference demonstrated how these subtle,
conditional, and/or tissue-specific mutations could mimic human disease.
Dimitris Kontoyiannis of the Hellenic
Pasteur Institute in Greece controls the expression of mutant tumor
necrosis factor (TNF) in first-generation mouse progeny using the
recombinase/target system. Altered temporal and spatial patterns of mutant
TNF expression results in a severe inflammatory gastrointestinal disorder
resembling Crohn's disease. This model can be used to study the mechanisms
of inflammatory cytokines in gastrointestinal illness.
Ann Puech of the Albert Einstein College of Medicine in New York is using the recombinase/target system to generate mouse models of human genetic disease. Her team's goal is to create a panel of spatiotemporally controlled mutations in the mouse homologue (same gene) of human chromosome region 22q11. Large deletions in this region in the human are found in velocardiofacial and DiGeorge syndromes, which are characterized by heart defects, facial defects, and learning disabilities. This will be a tedious process, yet most likely a very rewarding one.
Mark Mayford of the University of California at San Diego discussed how his laboratory uses
regulated transgene expression to study memory. They created a breed of
double-transgenic mice that can be induced to express mutant enzyme
(calcium-independent CAMKII) in specific regions of the brain. Altered
enzyme expression in the striatum and lateral amygdala during the
consolidation phase of learning leads to confusion and loss of emotional
conditioning. This shows that these regions of the brain need normal CAMKII
enzyme during the learning process.
Members of the Skirball Institute of Biomolecular Medicine and New York University, along with members of Beth Israel Hospital in Boston and Jackson Laboratories in Bar Harbor, Maine, combined their efforts to achieve conditional expression of the tumor gene TLS-CHOP, which is found in all human malignant liposarcomas. Recombination systems such as these are emerging as powerful tools in the study of human cancers.
The meeting's final discussions concerned the availability of reagents and
mice to allow these studies to continue. Dupont Pharmaceuticals
carries the patent on Cre-loxP technology (Cre is the recombinase and loxP
the target sequence). Cre-loxP is the most commonly used and, rumor has it,
the most successful recombination system. On August 19, only days before the
conference, Dupont Pharmaceuticals signed an agreement with the National Institutes of Health (NIH) to allow
the free transfer of Cre mice and loxP DNA sequence between nonprofit
research labs. NIH director Harold Varmus
described this as a milestone in the cooperative relationship between
academia and industry.
Hermann Bujard of the University of Heidelberg in Germany, stated: "You [Dupont] came around to a situation that is normal."
Klaus Rajewski of the University of
Cologne in Germany, one of the symposia's chairpersons, agreed.
"When you publish a paper and someone asks for it [a reagent], you have
to send it out. That's sufficient. True, in academia that has always been
the case, but pharmaceutical companies are not governed by the same rules.
Dupont's agreement with NIH to allow the transfer of their patented
technology without additional cost demonstrates that free distribution of
critical technologies can also occur between industry and academia. It is
the hope that other nonacademic institutions will follow suit."
Janet Rossant of the Samuel Lunenfeld Research Institute in Toronto, Canada, summed up the discussion by emphasizing the need to distribute reagents widely, and to establish central repositories containing mice that express both ubiquitous and tissue-specific recombinase. With all of the promise these technologies hold - and the enthusiasm displayed by the attendees at Cold Spring Harbor - demand should make both of these wishes a reality.
Danielle M. Kerkovich is a visiting student at the University of Connecticut Health Center. Her thesis work focuses on the developmental regulation of heterotrimeric GTP-binding protein, Gza, in fetal human and rodent cortical neurons.
Andrzej Krauze is an illustrator, poster maker, cartoonist, and painter who illustrates regularly for HMS Beagle, The Guardian, The Sunday Telegraph, Bookseller, and New Statesman.
Web Resources for Model Organisms - an HMS Beagle In Situ review of Internet resources for the study of model organisms, with tips on how to interface with the available genomic information.
Model Systems - an HMS Beagle discussion of the selection and use of model systems in biological studies.
Jackson Laboratory: Mouse Genome Informatics - extensive sources of information on the genetics and biology of the laboratory mouse.
Mouse and Rat Research: Genome Informatics - a comprehensive collection of databases, chromosome maps, DNA/RNA resources, inbred strains, and transgenic animals.
Mouse Knockout and Mutation Database - a fully searchable database of mouse knockouts and mutations.
The Mighty Mouse - a short essay on the importance of mice as models for human disease. Produced by the U.S. Department of Energy as an overview of the Human Genome Project.
Web sites mentioned in this column: