Potential of Transgenics
Potential of Transgenics: With Great Power Comes Great Responsibility by Bradley Monk
A review of the review – Gene targeting: technical confounds and potential solutions in behavioral brain research. By Robert Gerlai. In Behavioural Brain Research 125 (2001) 13–21.
INTRODUCTION All organisms have DNA that encode molecular instruments for survival and reproduction. Strings of DNA called genes, code for proteins that play a functional role in the organism. Proteins have specific functions (e.g. synthesize acetylcholine) and also work in concert to allow for more complex processes (e.g. long term memory formation). The differential expression of genes over time (due to developmental or environmental factors) leads to alternative states of being. These states can manifest themselves as a subset of physiological and behavioral responses to particular stimuli. If this definition of a gene seems convoluted, it is only because it is necessarily so; currently we have only a general understanding of the interaction between genes, environment, and behavior. Yet, this issue is at the forefront of behavioral neuroscience. An effort is underway to understand and catalog the functional role of genes from the molecular to behavioral level. The ability to alter the genetic makeup of an organism and observe the resulting sequela is among the most powerful tools in the field of behavioral neurobiology. However, issues have been raised concerning the utility transgenic methods for studying behavior. These concerns are highlighted in a 2001 review by Robert Gerlai published in Behavioural Brain Research titled “Gene targeting: technical confounds and potential solutions in behavioral brain research”. Gerlai outlines both theoretical and technical confounds of transgenic research, with an emphasis on the interpretation of results. In the decade since this review was published there have been major advances in transgenic research methods, and the number of studies that employed a transgenic method is legion. This provides an excellent opportunity to appraise the extent to which issues raised by Gerlai have been addressed, and whether the theoretical or technical concerns levied more attention.
MAJOR POINTS OF CONTENTION
The Gerlai (2001) review critically examines gene targeting studies in the field of behavioral brain research. From his point of view, the primary issues of transgenic research in this field are such:
The problems can be divided into two main categories, both of which have general importance in gene targeting experiments concerned with brain function and behavior. The first is a cluster of problems associated with compensatory mechanisms. This problem is rather difficult and there is no general solution to avoid it. The second problem is associated with genetic background and linkage (the so called flanking region problem). [pp. 13]
The first of the two problems above is addressed in some detail in a section of the article labeled “Compensatory mechanisms.” Gerlai suggests that gene targeting aims to reveal a gene’s in vivo function; but raises the concern that a genetic mutation results in a barrage of compensatory processes. Therefore, observed phenotypical changes may not be related to the gene of interest in the way we perceive.
Clearly, a null mutant organism may not only lack the product of a single gene but may also possess a number of developmental, physiological, or even behavioral processes that have been altered to compensate for the effect of the null mutation. [pp. 14]
The second set of problems with gene targeting experiments outlined by Gerlai is explored in a section labeled “Genetic background: an important confounding factor.” Knockout mice are usually the F2 offspring of two mouse strains. Recombinant mice therefore create issues in dealing with gene linkage background. Here it is laid out the issues of producing null-mutant mice through strain-crossing mating procedures.
A polymorphism in the genetic background will not allow one to conclude with certainty that a particular phenotypical change observed in a null mutant animal was indeed due to the null mutation or to the genetic background… Null mutant mice of hybrid origin are genetically different from their control littermates not only at the locus of the targeted gene but at other loci as well [pp. 15-16] Gerlai goes into a fair amount of detail describing these concerns and couples them with examples of how results of such studies could give rise to misinterpretation. There is also section that describes a methodological work-around to the flanking region problem. However, the underlying message of the review is one of concern for misattribution of phenotypic observations to the gene target of the study.
EVALUATION AND RESPONSE
The issues raised by Gerlai are important indeed, establishing a connection between a gene-protein product and a behavioral alteration are of interest to a diverse crowd (e.g. pharmaceutical, genetics, and psychological research). Therefore, any claim of such a link should not be taken lightly. However, according to the criteria outlined by Gerlai, we might never understand the direct phenotypic impact of a particular gene. It’s my contention that molecular-level compensatory changes are simply part of a phenotype that should be assayed through a variety of techniques. To further explore this idea, I will use this example throughout the remainder of the essay:
A group of researchers are interested in the underlying mechanisms of learning and memory in mice. In particular, they have chosen to focus on a subset of memory processes involved with fear conditioning. The behavioral aspect of the experiment is carried out as such: a mouse is taken from its home cage and placed into the novel environment of the fear conditioning apparatus. After 60 seconds a audible high frequency tone is played, followed 1 second later by a flashing green light, and then 5 seconds later a mild foot shock is delivered through the floor. This experience is salient enough to causes freezing behavior in mice after a single trial. That is, a mouse reintroduced to the apparatus after 24 hours will freeze for a short time upon being placed in the apparatus, and again after hearing the tone. Now suppose these research found that AMPA receptors were greatly upregulated in the auditory cortex of mice that were shocked compared to controls. Specifically, this upregulation was found in neurons that respond to the particular frequency of the tone that was played prior to the foot shock, but not surrounding neurons. No such change was found in the visual cortex. One of the researches suggests looking into the AMPA trafficking protein PKMz as a possible regulator of this apparent AMPA-related memory formation.
There are currently a variety of ways for examining a possible link between PKMz and learning-dependent trafficking of AMPA receptors. Transgenic manipulation of the PKMz gene might help us understand the mechanisms that underlie LTP, the brain regions involved in the memory formation process, the sensory systems that encoded the stimuli, and the behavior of the subject. Starting with a classical gene knock-out experiment, we can define the limits of what we can explain using a particular transgenic assay.
Transgenic knock-out The PKMz gene is excised from an embryonic stem cell and introduced into a blastocyst. This embryo is raised to adulthood and through two generations of selective mating, a resulting F3 offspring have heterozygous null mutant mice, homozygous null mutant mice, and wild type mice. If we were to test the fear conditioning experiment in these mice, we might find there are no learning deficits in the null mutant mice. However, PKMz may still play an important role in learning and memory; we do not know whether this result was due to some behavioral compensation (e.g. the mice were trained to rely more heavily on PKMz independent learning mechanisms), or a physiological compensation (e.g. during development, the absence of PKMz caused a drastic upregulation of CREB which was able to adequately support those learning processes). This assay is limited in all the technical ways Gerlai describes above.
It’s obvious there is a need for more powerful transgenic techniques to answer these questions. Newer methods improve upon the transgenic knockout model, and are able to reveal things like, (a) new genes required for a process to occur, (b) the order in which gene products act in the process, and (c) whether the proteins encoded by different genes interact with one another (Stieger, et al. 2009; Webber, et al. 2011). Here is a summary of some novel transgenic methods:
Conditional vectors Using site-specific recombinase systems, conditional vectors irreversibly inactivate gene function in a local region (Cre-Lox gene excision). This can be used to address questions that are confounded by global expression changes.
Inducible vectors Tetracycline regulated systems can be used to activate/inhibit gene expression in a reversible manner. Inducible vectors provide temporal and spatial control of gene expression.
Multifunctional vectors Multifunctional vectors are designed to generate point mutation insertions and conditional deletions within the same gene.
BAC Transgenes Transgene expression from a knock-in models is usually weak or undetectable. BAC transgenes provides all regulatory sequences needed for the appropriate transcriptional machinery, resulting in elevated transgene expression.
Promoter Transgenes Promoter transgenes can drive expression of a reporter or other gene of interest. These can be used to selectively express designer genes like channelrhodopsin (ChR2) in a tissue-specific, activity-dependent manner.
These techniques address many of the technical concerns in the Gerlai article. It’s at least safe to say the “flanking-region” problem is handled by using lentivirus vectors to cause transgenic mutation in F0 generation (i.e. there is no need to breed for the mutation). Thus we can focus on the other primary concern in the article, “Null mutant organism may not only lack the product of a single gene but may also possess a number of developmental, physiological, or even behavioral processes that have been altered to compensate for the effect of the null mutation.” Say for instance, a mouse has learned to behaviorally compensate for a sensory limitation caused by traditional transgenic gene KO as it progressed through development with this limitation. Using tet-inducible constructs we are able to cause the suppression of a gene only at the specific time-point of interest. Therefore, a mouse can act as its own control subject with regard to behavior (Sun, et al. 2007). Multifunctional vectors causing point mutations and synthetic lethal mutations can reveal genes whose encoded proteins function in redundant pathways (compensatory reaction pathways). These techniques can also reveal the functional role of a protein and how it may interact with a substrate or other effector and regulatory proteins. In our PKMz example, multifunctional vectors can be used to test whether this protein is normally upregulated in response to a conditioned stimuli, but due to the point mutation, cannot transport AMPA that necessitates LTP and learning. Conversely, we can examine if a point mutation in a regulatory domain of PKMz causes aberrant over-transportation of AMPA and hinders learning in this regard. Promoter transgenes allow for the expression of ChR2 under the control of PKMz upregulation. We can then target cells that were ostensibly active during the learning process (high PKMz activity) using fiberoptic techniques. If shining light onto these ChR2-expressing neurons elicits a freezing response, we can glean that PKMz activity is among the components in the fear conditioning pathway. Thus, a number of transgenic tools are now available to examine a purported link between a particular gene and behavior. This does not mean these techniques are being used in an ideal manner; the extent to which they augment the understanding of experimental results continues to impinge upon mindful interpretations of the experimenters.
THE ARTICLE IN CONTEXT OF CENTRAL LESSONS FROM OUR COURSE THEMES: NEUROBIOLOGY, DEVELOPMENT, ECOL-EVOLUTIONARY
The Gerlai article has a very interesting statement that I believe deserves further cross-discipline appraisal:
One may argue that when it comes to the question of phenotypical effects, genes may not be the units and the definition of their function may be more complicated. One may suggest that clusters of genes defined by higher organizational level phenomena, including developmental, physiological, or even behavioral, may represent the functionally relevant unit.
This idea is self-evident, depending upon how the concept of phenotype is defined. Any attempt to describe the substrates of behavior should consider the developmental, ecological, and evolutionary history of an organism. Defining a functionally relevant unit of a behavior is a philosophical task in its own right, and I don’t suspect transgenic and recombinant methods alone are equipped to provide answers in an ecological and evolutionary context. However, these perspectives can add meaningful insight to the results of transgenic experimentation. They can also help in the planning stages of a transgenic experiment to help define a more appropriate model for tackling systems-level research questions (Lazarus et al. 2011). While a stringent methodological approach can help to behaviorally phenotype transgenic mice (Brown, 2007), some types of behavior may be too complex to acquire a meaningful genetic definition. For example, the complex behavioral phenotype that characterizes schizophrenia or autism spectrum disorders might only be meaningfully defined on a developmental or ecological level. Indeed, there is evidence that even the currently most advanced transgenic methods have not been very fruitful in defining meaningful aspects of schizophrenia (Pletnikov, 2009). It is therefore important that we recognize the limitations and explanatory power of the genetic approach. While many subsets of behavior are regulated by protein interactions, the complex epiphenomena of the brain should not be considered reducible to the molecular level on all accounts; at least in a way that can be meaningfully interpreted.
Transgenic methods offer a powerful tool for assessment of the functional roles of proteins. There is no doubt that we will continue to advance the technologies that give us control over gene manipulation. The utilization of transgenic methods will continue to aid our pursuit in understanding behavior, but should not be considered a liberator from critical scrutiny of experimental design and results interpretation. One thing is for certain, genes do not act independently of the system from which they are expressed, and should not be treated as such. These technologies are merely a tool in the scientific arsenal for examining various components of existence.
Weber, Fussenegger (2011). Emerging biomedical applications of synthetic biology. Nature Reviews Genetics, advance online publication, Published online 29 November 2011, doi:10.1038/nrg3094
Stieger, Belbellaa, Le Guiner, Moullier, Rolling (2009). In vivo gene regulation using tetracycline-regulatable systems. Advanced Drug Delivery Reviews, 61, 527-541.
Lazarus, Saper, Fuller (2011). Genetic dissection of neural circuitry regulating behavioral state using conditional transgenics. Sleep and Biological Rhythms, 9 (Suppl. 1): 78–83.
Pletnikov (2009). Inducible and conditional transgenic mouse models of schizophrenia, Progress in Brain Research, 22.
Brown (2007). Behavioural phenotyping of transgenic mice. Canadian Journal of Experimental Psychology/Revue. 61(4), 328-344.
Molecular Cell Biology, 6th Edition (2008). W.H. Freeman and Company, New York, NY.