Mouse vs. mouse
In Chinese culture, 2011 is the year of the rabbit. In the autism research field, it seems to be the year of the mouse model, with five new mouse strains that carry the same genetic glitches as some people with the disorder.
By my count, that means researchers now have access to more than a dozen mice carrying genetic signatures linked to autism.
The idea is that by altering an individual candidate gene or genetic region in a mouse, scientists can better understand how it causes particular behaviors. These studies are exciting, but I think it's also important to reflect on the mutant mice that, despite carrying an autism-linked genetic variation, don't show any autism-like behaviors.
Case in point: a mouse model lacking a working version of SEMA5A, described in the November issue of Behavioural Brain Research. After performing ten standardized behavioral tests on this mouse, the researchers concluded that it has no autism-specific characteristics.
This surprised me because SEMA5A was, not so long ago, a pretty high-profile gene. In 2009, a huge research consortium reported in Nature that individuals with autism are significantly more likely to carry a common variant near the gene than are those without the disorder. What's more, the team found lower levels of the SEMA5A protein in postmortem brain tissue from people with autism compared with controls.
There's little doubt that the gene is important for brain development. Many studies have shown that it's involved in positioning developing axons, the skinny projections of neurons. In 2005, researchers made mice missing both copies of the gene and found that all of the mice die in utero.
When the gene is knocked out of a mouse strain with a different genetic background, however, the mice survive with no major problems. The new study looked for features of autism in this mouse strain.
The researchers tested social interactions, communication, repetitive behaviors, memory, anxiety, clumsiness and reactions to sensory stimuli — in other words, all of the recommended assays for testing autism-like behaviors. They found nothing out of the ordinary, except for a bit more activity and anxiety on certain tests.
These sorts of negative studies don't get much attention, but I think they should. Here we have an example of two animal models of the same interesting autism candidate gene. When carried by one kind of mice, it causes early death, and by another, it does little at all.
There are at least five different mouse models of SHANK3 and these, too, show intriguing differences. The fragile X mutation also causes a wide variety of behaviors depending on which mouse strain carries it.
Now that the field has created such an impressive library of mouse models, perhaps it's time to stop growing the library and begin checking out its fascinating books. 2012 could be a year of comparing existing mouse models against different genetic backgrounds or against each other.
Comments
Studying the consequences of genetic syndromes cannot ignore origins. All males and females generate sperm and egg mutations. Molina et al (2011) studied sperm mutations in healthy male volunteer donors focusing on three mutations identified in children with a genetic syndrome with high ASD risk. All three of these mutations have been reported in articles on genetically engineered mice in SFARI articles published within the last month. The three sperm mutations that were specifically examined that are associated with genetic disorders with high rates autism risk were: 7q11.23 (Williams Syndrome), 15q11-13 (Prader-Willi Syndrome), DiGeorge/velo-cardio-facial (22q11 deletion) syndrome. The mutations were sporadically found in the sperm of the volunteer donors. Almost all cases of genetic and epigenetic mutations associated with these three genetic syndromes with high autism risk are de novo mutations in contrast to their being inherited events and in stark contrast to the oft made claim that autism is the most heritable of the developmental disorders. If the frequency of sperm or egg mutations increases with advancing parental age then environmental factors have to be considered central in disentangling the complex mechanisms involved in mutagenesis as far as the genetic syndromes with high rates of co-occurring ASD is concerned. The question that the SFARI researchers have to ask themselves, where did the heritability go in the genetic syndromes with high autism risk? These sperm or egg mutations are random events that can strike any family at any time apparently without any reason other than pure chance (Molina et al 2011 )
REFERENCE
Molina O, Anton E, Vidal F, Blanco J. Hum Genet. Sperm rates of 7q11.23, 15q11q13, 22q11..2 deletions and duplications: a FISH approach. 2011 Jan;129(1):35-44. Epub Oct 8.
http://www.ncbi.nlm.nih.gov/pubmed/20931230
This is a good post discussing the important issue of how to think about potential mouse models of ASD, especially those that do not give a strong (or any) phenotype when candidate genes are mutated. The case of SEMA5A is a tricky one. While it's true that the signal reported for the SNP flanking SEMA5A reached genome-wide significance, it just barely cleared the bar, leaving open at least the possibility that it was a false positive result (the accompanying expression data were nice, but I don't think add a lot of weight to the result). We also know of examples where intergenic SNPs such as this one affect genes that are far distant, and not those that are closest to them. Finally, given that the putative SEMA5A-affecting SNP had a small effect size--as do almost all common variants--it's possible that in a mouse model it would just serve as a modifier of risk in the presence of another highly penetrant mutation. For example, it would be interesting to know if the 'autism-like' phenotypes of the NRXN1 or SHANK3 mutant mice are exacerbated by the presence of a SEMA5A mutation.