Hair of the dog

By Leah Eisenstadt, Broad Communications, September 11th, 2008
The dissimilar appearance of these two varieties of the Chinese crested dog breed has been traced to a difference in the FOXI3 geneImages courtesy of iStockphoto

When it comes to dogs, looks can be deceiving. The distinctive hairless Chinese crested looks so different from its coated counterpart, a fluffy pooch dubbed “powderpuff,” that the two appear to be distinct breeds. In fact, hairless and powderpuff are varieties of a single breed, the Chinese crested. Now, a team of researchers from the Broad Institute of MIT and Harvard and elsewhere has revealed the genetic mutation behind this breed’s dichotomous looks, underscoring the power of dog DNA in the search for critical disease and trait genes in canines as well as humans. The findings appear in the September 12 issue of Science.

A team of European scientists, led by Tosso Leeb of the University of Berne in Switzerland, had been searching for years for the genetic roots of the Chinese crested’s hairlessness and its abnormal dentition. (In addition to sparse tufts of hair on the head, feet, and tail, hairless Chinese cresteds often have missing or misshapen teeth.) After testing all suspect genes, however, the team came up empty-handed. Thinking that an unbiased scan of the roughly 20,000 genes in the dog genome might be a more successful approach, the scientists contacted Broad researcher Kerstin Lindblad-Toh, who co-directs the Broad’s Genome Sequencing and Analysis Program and also led the 2005 effort to decode the dog genome.

The canine genome has some unique features that make it ideal for genetic mapping, due to the selective pressures humans have put on dogs. By first domesticating them from wolves and later creating individual breeds enriched for desirable traits, humans inadvertently created two bottlenecks in the evolution of the species. In all organisms, DNA is inherited in chunks, known as haplotypes. Most dog breeds are less than a few hundred years old, so the DNA has strayed little from those bottleneck points, when the first desirable dogs were selected to create a breed. Consequently, dogs of the same breed share longer chunks of DNA than do those of different breeds. Indeed, selective breeding has led to haplotypes in the dog genome that are 100 times larger than those found in the human genome.

Taking advantage of this unique genomic structure, Lindblad-Toh and colleagues developed a two-stage gene-mapping strategy that requires fewer animals and fewer genetic markers than are necessary in human studies. They also used data generated from the dog genome sequencing project to develop powerful tools for analyzing dog DNA. With these tools, Lindblad-Toh and her team have been able to locate genes underlying traits such as the raised hair ridge of Rhodesian ridgebacks associated with a neural tube defect called dermoid sinus, coat color in boxers, and dayblindness in dachshunds. Eager for more evidence for the dog genome’s utility as a research model, Lindblad-Toh joined with Leeb’s team to begin an unbiased, genome-wide search for the elusive hairless gene.

To narrow their search, the team first gathered DNA samples from 20 hairless Chinese crested dogs and 20 powderpuffs in Switzerland and Finland. With the help of the Broad’s Genetic Analysis Platform and researchers Elinor Karlsson and Michele Perloski, they performed a genome-wide association scan, which searches the dog genome for markers found in hairless dogs but not in powderpuffs. The scan produced one strong signal, pointing to a roughly million-letter-long section of chromosome 17. To zoom in on the causal gene even further, the researchers compared this section of DNA in Chinese cresteds to that of two other hairless breeds, the Mexican and Peruvian hairless dogs, which also feature both hairless and coated varieties. They found that hairless dogs in all three breeds shared a 102,000-letter-long chunk of DNA that contains only two genes.

One of these genes, named FOXI3, is similar to a family of mammalian genes with roles in development, suggesting that it could harbor the hairless mutation. “No one knew anything about this gene, FOXI3,” said Karlsson, adding that it hadn’t even been studied much in mice, a common model in genetic research. Performing their own studies, the team discovered that the mouse Foxi3 gene was indeed active in developing hair and teeth, supporting its possible role in hair and teeth abnormalities in hairless dogs.

To identify the exact mutation in FOXI3, the researchers examined the precise sequence of DNA in the 102,000-letter stretch shared by the hairless Mexican, Peruvian, and Chinese crested dogs. One peculiarity was found in the DNA of all 140 hairless dogs and in none of 87 coated dogs tested: seven letters of repeated DNA in the FOXI3 gene. Because the genetic code operates by a rule of three — three letters of DNA encode a single amino acid in a protein — the addition of seven letters to the template completely shifts how the DNA is made into protein. Even though they come from different parts of the world, all three breeds share exactly the same seven-letter DNA change, suggesting they are all are descended from the same, ancestral hairless dog.

Like a sentence with spaces in all the wrong spots, the shifted, mutant protein is biological jibberish compared to the normal one, and it is likely unable to perform the same role in the cell. Though it’s possible that the mutant protein functions by actively interrupting normal hair and tooth development, it’s also possible that less normal protein is made in dogs with the mutation. Dogs with both copies of their gene mutated, who make no normal protein, usually do not survive until birth. The researchers found that all living hairless dogs in the three breeds they studied are actually heterozygous for the mutation. It seems that having one normal version of the gene may give the hairless dogs sufficient normal protein to survive, but not enough to generate a full coat of hair.

The researchers don’t yet know exactly how the FOXI3 mutation leads to hairlessness, though their continuing work may yield an explanation soon. “Clearly this gene is critical to development, since we know that a double-dose of the mutated gene is lethal during early embryonic stages,” said Karlsson. “We just don’t know the mechanism yet.”

Dog hairlessness is similar to a human condition known as anhidrotic ectodermal dysplasia, resulting in abnormal hair, teeth, and sweat glands. Lindblad-Toh explains that there may be some connection between the FOXI3 mutation and the genetic cause of the human dysplasia, though more work is needed to test that hypothesis. “The hairlessness study is primarily a proof of principle for using the dog genome to dissect dog traits and shed light on human illness,” she said.

Lindblad-Toh and her team are also in the midst of similar studies on diseases more complex than some of the traits mapped in dogs so far. “This year, we’ve primarily published studies on simple traits to prove that this method can work,” said Lindblad-Toh. “However, we also have discovered regions in the dog genome linked to several cancers, autoimmune disease, and other complex illnesses, and we hope to find the precise, causal mutations in the coming year. The goal then will be to look at the same genes in human patients.”

Paper(s) cited: 

Drogemuller C, Karlsson EK, Hytonen MK, Perloski M, Dolf G, Sainio K, Lohi H, Lindblad-Toh K, Leeb T. Mutations in hairless dogs implicate FOXI3 in ectodermal development. Science. September 12, 2008. DOI: 10.1126/science.1162525.

Karlsson et al. (2007) Efficient mapping of mendelian traits in dogs through genome-wide association. Nature Genetics DOI:10.1038/ng.2007.10.

Wiik AC, Wade C, Biagi T, Ropstad EO, Bjerkas E, Lindblad-Toh K, Lingaas F. A deletion in nephronophthisis 4 (NPHP4) is associated with recessive cone-rod dystrophy in standard wire-haired dachshund. Genome Research. DOI:10.1101/gr.074302.107.