Project 1: The Mouse Brain Library Project 2: Internet Microscopy (iScope) Project 3: Neurocartographer and Segmentation of the MBL Project 4: The Neurogenetics Tool Box














Principal Investigator/Program Director Williams, Robert W.

  Investigating genetic variation

Normal variation can be impressive. Numbers of neurons in the human neocortex vary from 15 billion to 32 billion (Pakkenberg and Gundersen 1997). The volume and cell number of the human visual cortex vary three-fold, as does the density of cones in the foveal pit (Gilissen and Zilles 1996; Curcio et al. 1990). Numbers of ocular dominance columns within the primary visual cortex of rhesus monkeys vary more than 50% (Horton and Hocking 1996). These robust differences are caused not by mutations but by the cumulative action of many normally variable genes and by the action of numerous developmental and environmental factors. In the long run, normal genetic polymorphisms are the most critical source of variance: they are the substrate for evolutionary and developmental modification of brain size and cellular architecture (Williams and Herrup 1988; Lipp 1989; Williams et al. 1993).

In the late 1960s, Thomas Roderick, John Fuller, Douglas Wahlsten, and Richard and Cynthia Wimer began an ambitious program to manipulate neuroanatomical traits in mice by selective breeding (Roderick 1979). Their aim was to explore correlated changes in behavior. They gave the rapidly expanding field of behavioral neurogenetics a rigorous foundation in quantitative and statistical neuroanatomy (Wimer et al. 1969; Fuller and Geils 1972; Wahlsten 1975; Roderick et al. 1976; Fuller 1979; Wimer 1979; Wimer and Wimer 1985). Rather than relying on mutants, they exploited the substantial variation among standard inbred strains of mice. This work led to some important breakthroughs and some brick walls. One of the breakthroughs was successfully selecting for substantial differences in brain weight over less than 20 generations (Fuller 1979). An obvious limitation, highlighted by Roderick (1976), was that it was not possible to map gene loci responsible for the remarkable quantitative variation in CNS size, regional architecture, or behavior.

The research tools available to neuroscientists have changed radically in the past decade (Lander and Botstein 1989; Plomin et al. 1991; Belknap et al. 1992; Johnson et al. 1992; Tanksley 1993; Frankel 1995). Computational methods and molecular reagents—particularly the polymerase chain reaction (PCR) method—have become so powerful and economical that it is now practical to systematically dissect complex polygenic traits such as brain weight into sets of single well-defined QTLs. Virtually any heritable trait in mice, whether structural, physiological, pharmacological, or behavioral, can be targeted for analysis.

Recent examples in mice include epilepsy (Rise et al. 1991), effects of ethanol and haloperidol (Belknap et al. 1993; Plomin and McClearn 1993; Hitzemann et al. 1994; Kanes et al. 1996; Buck et al. 1997), patterns of sleep and activity (Toth and Williams 1998), and the mouse equivalent of anxiety (Flint et al. 1995). As illustrated in the work of Belknap and colleagues (1992), it is now feasible to continue the systematic genetic dissection of the mouse CNS begun in the late 1960s and to start identifying genes that underlie heritable variation in CNS size and structure.

QTLs are conventional genes that have two or more alleles that contribute to quantitative variation of specific traits (Roff 1997; Lynch and Walsh 1998). A trait may be a concentration or number; a size, weight, or density; an activity or behavior; a severity index; or an age of onset. QTLs are often contrasted with Mendelian loci that have discontinuous effects on phenotypes and predictable segregation patterns. Individual QTLs usually have more modest effects on a particular phenotype and are associated with phenotypes in a probabilistic way. A QTL might account for as little as 2% or as much as 50% of the total variance of a phenotype.

The clinical potential of QTL mapping in mice.  

The analysis of complex polygenic traits is an important facet of human genetics. Any recent issue of the Journal of the American Society of Genetics will include 5 to 10 reports on gene loci affecting disease susceptibility, onset, or severity in different human populations. These "susceptibility" genes are in fact QTLs. The genetics of alcoholism, schizophrenia, depression, heart disease, glaucoma, myopia, several dementias, breast cancer, colon cancer, and autoimmune disease all are benefiting enormously from the application of these methods.

Systematic histology of the huge human brain in a genetic context is daunting, to say the least, but in mice, the brain is the size of a plump pistachio. A 1-in-10 series of sections through the entire brain fits on a single 50 x 75 mm slide. Because so many diseases of the CNS have complex genetic and environmental etiologies (Parkinson's, schizophrenia, depression, Alzheimer’s, macular degeneration, etc.) it is vital to build up resources for a systematic genetic assault on complex CNS traits. We can investigate the genetic basis of many traits in the mouse, then look for corresponding features in humans.

We already know that morphometric CNS traits are often highly variable and highly heritable. For example, brain weight in the DBA/2J strain of mice is about 420 mg, and in C57BL/6J it is about 500 mg. Morphometric traits tend to have relatively high heritabilities compared to behavioral and fitness traits (Roff, 1997). Most heritabilities that we have computed (either broad or narrow sense) have been well above 0.3, with an average of about 0.5.

QTLs can be mapped routinely.  

Information on our set of over 5000 mice and over 100 strains is available online at . We have used this large data set to test the feasibility of mapping brain weight QTLs. The results have been extremely successful, and we have now mapped several QTLs that specifically modulate brain weight (see Appendix material). We have recently gone one step further to show that it is also entirely practical to map QTLs that affect specific subdivisions of the CNS. We have mapped QTLs that modulate the weight of the cerebellum (Airey et al. 1998, 1999) and one QTL that has an especially marked effect on the size of the hippocampus (Lu et al. 1999).

Probability of success in QTL mapping studies.  

What is the probability of successfully mapping one or more QTLs? For CNS traits that have heritabilities above 50%, it will usually be possible to map several QTLs in even a small cross consisting of 200 to 400 individuals. Behavioral traits such as open-field activity tend to have relatively low heritabilities (<30%), yet they have been successfully dissected into sets of QTLs (Flint et al. 1995). For example, Le Roy, Roubertoux, and colleagues (Le Roy et al. 1999) have successfully mapped more than a dozen QTLs that modulate the development of several behavioral traits in preweanling mice.

The main constraint in this work is the number of animals that can be phenotyped and genotyped. We have mapped four QTLs affecting retinal ganglion cell number (Williams et al. 1998a; Strom 1999), eye weight (Zhou and Williams 1999b; Williams and Zhou 1999), brain weight (Strom and Williams 1997; Strom 1999), and cerebellar weight (Gilissen and Williams 1997; Airey et al. 1998). In each case, approximately 150 animals were phenotyped per QTL. If multiple polymorphic loci are clustered near one another, then a small intercross may detect a single poorly localized QTL, in essence a polygenic QTL. In contrast, a large high-resolution cross, especially an advanced intercross of the type that will be a major resource in this program project, will separate the individual loci (Darvasi 1998; Williams 1998).

Although it is satisfying to decompose variation in brain weight into sets of individual QTLs, we still need to determine what parts of the CNS and what cell populations are most and least affected. For example, is there a subset of QTLs that specifically modulates the size of the cerebellum or the proliferation of granule cells? As part of this program project, our whole group is assembling the resources needed to carry out systematic stereological studies of individual nuclei and cell types. The main impediment for an individual investigator is the massive effort needed to section, stain, and count particular nuclei or regions in hundreds of cases. If a single cross is shared among many neuroscientists and then used to analyze many different CNS structures, then the effort required to map each QTL is reduced substantially. This, of course, is an important benefit of the program project: other investigators will be able to use the resources we develop to map multiple QTLs, yet these other groups will not have to process any tissue or genotype any animals.

Cloning QTLs.  

Mapping QTLs is the reconnaissance stage in a systematic effort to explore mechanisms that modulate the development of the CNS. The next step is matching each QTL with a single gene and its alternative alleles. QTLs will generally need to be mapped with a precision of 1 to 2 cM, a chromosomal interval that typically harbors 50100 genes. Achieving this level of accuracy is not impractical, although it will often require an analysis of 1000 or more animals (Darvasi 1997, 1998). A small subset of positional candidate genes can then be chosen for further analysis on the basis of expression patterns, known function, and differences in DNA sequence among strains. The efficiency of the candidate gene approach will improve greatly in the next decade. The genome of C57BL/6J will have been sequenced within several years, and it is also likely that the utility of this code will be enhanced with sequence data from other major inbred strains such as 129, A, BALB/c, C3H, DBA/2, CAST/Ei, and SPRET/Ei. Once sequence data have been combined with expression maps for different parts of the mouse brain, candidate genes can be winnowed to a very short list.

Even before it has been cloned, a QTL can be used to study mechanisms of brain development or function. For example, we wanted to determine whether Nnc1 modulates neurogenesis or cell death. To answer this question, we counted the cell population after neurogenesis but before the onset of cell death. We showed convincingly that the bimodality is produced by a fundamental difference in the total production of retinal ganglion cells (Strom and Williams 1998). Nnc1 must influence the proliferation of retinal ganglion cells, for example, through variation in the progenitor pool size, pathways of cell differentiation, or cell cycle parameters. The effect is robust and different strains carry alternative phenotypes and alleles, so it should be possible to explore the relative importance of these processes and define more precisely how Nnc1 modulates neurogenesis.
  Generic Resources versus Specific Goals  
  The External Advisory Board  
  Institutional Support for this Program Project