Arexis AB

Arexis is using specially designed animal models to unravel the genetics of metabolic disease.

Using model organisms to unravel the genetics of metabolic and inflammatory disease

• Krokslätts Fabriker 30

• Mölndal SE-43137, Sweden

• Phone: [phone]

• Fax: [phone]

• E-mail/Web Site: [email protected]

• Contact: Vidar Wendel-Hansen, CBO

• Industry Segment: Drug Discovery

• Business: Drugs for metabolic and inflammatory disease

• Founded: 1999

• Founders: Vidar Wendel-Hansen, MD, PhD (Karolinska Institute t); Prof. Leif Andersson, PhD (Swedish University of Agricultural Sciences); Prof. Rikard Holmdahl, MD, PhD (University of Lund); Holger Luthman, PhD (Karolinska Institute t)

• Employees: 5

• Financing to Date: $5.4 million

• Scientific Advisory Board: Ulf de Faire, MD, PhD (Karolinska Institute t); Ulf Pettersson, MD, PhD (Uppsala University); Paul Siegel, PhD (Virginia Polytechnic Institute and State University)

Sweden's Arexis AB is one of many players in the post-genomics era investigating the genetic basis of metabolic disease. Not unusually, it uses model organisms to test hypotheses about the involvement of different genes in multifactorial conditions such as diabetes and obesity.

But for Arexis, "model" refers to something much more carefully defined than just knockout mice or transgenic rats. The company's founding scientists have spent the last ten years developing "congenic" animal models. Congenic animals are bred to display specific, carefully chosen characteristics. For example, a rat selected for a certain feature such as high blood sugar might be crossed with another from a line that has never shown any instances of diabetes or high blood sugar. In this way the company creates a population of animals, over a number of generations, with a controlled genetic background. "A congenic animal retains just a small segment of the genome of the sick rat, while the rest of the genome is derived from the healthy parental line," explains Arexis' business development officer Vidar Wendel-Hansen, MD, PhD. Such models allow scientists to isolate specific genes or chromosomal regions of interest and to study their precise role in creating the phenotype.

The advantage of this approach over other techniques for finding gene-derived drug targets, such as expression profiling (comparing gene expression levels between diseased and healthy animals), is that the company gains a detailed insight into the biological processes associated with a particular gene, and what precise role that gene plays in the disease process. Expression profiling may generate hundreds of hypothetical targets, but, says Wendel-Hansen, "each of these targets will have much less information value. You may find that animals with disease may have high expression of a gene in a certain tissue. But this doesn't give more detailed information about how you should affect that gene in order to develop a new drug." Nor does it always tell you which of all these differentially expressed genes are actually relevant to the phenotype.

Arexis intends to achieve such precision—and thereby a more powerful system for generating drug targets—by applying its models in combination with methods like expression profiling. Mapping over- or under-expressed genes to the chromosomal regions that Arexis' models have identified as important means its models effectively "can be used as a filter to save time and find which of these hundreds of genes are the real culprits," explains Wendel-Hansen.

Just as animal models have their limitations, so studying susceptibility genes among humans has proved elusive. Genetic heterogeneity among human populations means it is difficult to prove meaningful, statistically supported associations between a particular region of the genome and a given phenotype. But here too, Arexis aims to use its models alongside other techniques. The company will need to validate its genes in humans, just as human genetics companies such as Iceland's deCode genetics Inc. will likely have to rely on model systems for early-stage drug testing. Arexis already collaborates with the University of Sassari in Sardinia, through which it has access to a very old, relatively tightly contained human population, and is looking for more such partnerships.

Arexis' models are useful not only for generating targets, but also further downstream, to do proof-of-concept studies on the drug candidate that is selected. This allows the company, at an early stage, to prove and measure the drug's effect on the particular signaling pathway towards which it is directed.

The company has assembled—both internally, and through acquisition—a range of congenic animal models across various species, which it uses in a comparative fashion. Arexis has set up a chicken cross, for example, based on work carried out at the Virginia Polytechnic Institute and State University, where one of Arexis' SAB members, Paul Siegel, PhD, has selected and bred lines of chickens with divergent growth rates. "What we now have is a number of chicken lines that represent different features of a whole panorama of metabolic phenotypes, ranging from insulin resistance to anorexia," explains Wendel-Hansen.

It was in a pig breed, however, that founding scientist Leif Andersson, PhD, uncovered one of Arexis' most important gene targets. A specific mutation in this particular gene, which codes for a novel gamma chain of the AMP-activated protein kinase, was found to be associated with high levels of glycogen in the skeletal muscle of pigs. Such a phenotype is exactly the opposite of that of diabetics, whose ineffective use of insulin slows the conversion of blood sugar into muscle glycogen. Arexis has patented both the gene—since cloned in humans—and its utility as a potential target for therapies in metabolic disorders.

The company is building its patent portfolio around such gene candidates (including a second one, also in the metabolic area, discovered in a congenic rat), as well as around the congenic animal models themselves. In this way, "even before we find specific genes, we are able to protect the animals in which they reside," says Wendel-Hansen. Arexis has five of its own patents pending and has in-licensed two issued patents, of which one is in inflammation, the company's second area of focus.

In the near term, Arexis will secure revenues by giving other companies access to its model populations as a tool for validating drug candidates. It already has an undisclosed partner in the field of rheumatoid arthritis.

Another of the tools Arexis hopes to exploit short-term is its genetic database for storing, analyzing and interpreting data stemming from its models and clinical collaborations. The company has to date identified over 20 chromosomal regions involved in metabolic disease or inflammatory disorders. Wendel-Hansen says the company is considering whether to spin off the database—which has already generated considerable interest—within a daughter company, or to collaborate with a bioinformatics group to make it into a more substantial business.

Longer term, Arexis will seek partnerships with clinical development companies to take its projects as far downstream as possible and so turn itself into a discovery and development company. In anticipation of this, Arexis last May appointed Lennart Hansson, PhD, as CEO and Björn Löwenadler, PhD, as CSO, both from AstraZeneca PLC .

Arexis is in the process of building up its own protein chemistry and bioinformatics capabilities and plans to have more than ten employees by year-end. It hopes to sign its first alliance within 12 months and to have at least four others within 3-5 years.—MAS