DeveloGen Diabetes Pax4 Gene Therapy Enters Late Preclinicals

A gene therapy approach to treat diabetes has passed through several validation screens and is now being tested in higher mammals including primates, DeveloGen VP-Corporate Affairs Edward Stuart, PhD, told the Ernst & Young/Atlas Venture European Life Sciences Conference April 17 in Amsterdam.

The Goettingen, Germany-based functional genomics company has discovered the beta-cell master control gene, Pax4, believed by the company to stimulate production of new beta-cells in the pancreas, which could allow diabetes patients to produce proper amounts of insulin with the newly grown cells.

Pax4 "is not expressed in any other cell population. When we knock it out we lose beta-cells specifically [and] insulin-producing cells specifically," Stuart said.

"When we return Pax4 into these pancreases we are able to regenerate beta-cells. So we know not only that we can regenerate beta-cells with Pax4, but in the absence of Pax4, regeneration cannot occur." The company has novel viral vectors and other delivery systems to introduce the Pax4 gene into adult pancreatic cells.

The Pax4 gene was discovered using the company's functional genomics and gene discovery DeveloScreen technology platform, which incorporates bioinformatics with four in vivo models - the Drosophila fly, and three vertebrates: the zebrafish, chick and mouse.

The company seeks to find developmental control genes, which govern how undifferentiated cells become various cell types based on when the developmental control genes are activated or inactivated, Stuart said. Finding how the control genes give rise to the functions of certain cells will help create better drug targets as well as cell-based and gene therapy, he continued.

Gene sequence and protein function are generally conserved along animal species, Stuart noted. Thus, human diseases can be examined by creating mutants in the model species, isolating the mutants that express the phenotype similar to the human disease, and then searching for the genetic mutations responsible.

Using the zebrafish screen, DeveloScreen was able to identify 50 different mutations that affected the pancreas endocrine system. Then the screen was repeated in the chick system, which involves a more complex pancreas development, and about 50 additional novel genes associated with pancreas development were identified.

In the mouse model, they were able to "perform reverse genetics as well as other gene discovery tools to further analyze what happened within the regenerated beta-cells," Stuart said.

A Pax4-based therapy would likely differ from current conceptions of gene therapy, Stuart said. "What we are doing is really targeting a [pancreatic stem] cell population and activating the master control gene either by introducing an exogenous copy or by activating the endogenous gene."

This process "is not a classical gene therapy approach because development control genes work on the principle that you only need to activate them transiently and they kick off a whole domino effect within the cell. So you just need to activate the process and the body takes care of everything else itself."

DeveloGen is also looking into other regenerative medicine areas, including liver and central nervous system gene discovery and validation.

In addition, an obesity program in Drosophila flies has resulted in the identification of a gene that causes a naturally occurring adipose fly, and a homologue of the protein it encodes is seen in other species, including humans. Coupling proteins associated with that protein are now being identified.

Like DeveloGen, Cologne-based functional genomics firm Artemis is collaborating with the Max-Planck Institute in Goettingen. With the university, as well as University College, London, the Howard Hughes Institute, Children's Hospital in Boston and the University of Heidelberg, Artemis announced the launch of the "Tuebingen 2000" zebrafish screen.

The Tuebingen project includes 17 in vivo screens that will try to correlate the phenotypes of mutant zebrafish with specific single-gene mutations, Business Development Head Paul Rounding, PhD, told the conference April 17. The zebrafish can be bred in very large numbers and the screens, planned for 17 mil. mutant zebrafish larvae, is the largest vertebrate genetics screen attempted, he said.

"We introduce a chemically induced gene mutation in male adult zebrafish," Rounding explained. Using a complex breeding schedule, "by the third generation we know...that these larvae contain only one gene mutation, a point mutation," he said.

The larvae, which are transparent and thus easier to assess, are screened "through constructed functional assays looking for particular phenotypic changes [such as] changes in the function of particular organs."

Next, "we do a series of secondary screens, designed to isolate those genetic mutations which we would subsequently like to clone, and we isolate the cloning gene responsible. What we are getting from this are novel drug targets clearly well-functionally validated with in vivo systems, and also novel secreted proteins," Rounding said.

The Tuebingen screen is slated to take a year. Artemis will retain commercial rights to discoveries made from the seven in-house screens, and will likely obtain exclusive rights for the other 10 screens.

The screens will focus on finding therapeutic targets in the cardiovascular and bone and cartilage systems, and will seek about 1,000 different gene mutants, by phenotype. Screen designs extend from the smaller scale "Tuebingen 1" screen, conducted from 1996-1997.

The earlier screen was looking for developmental defects but did yield some therapeutic discoveries. For example, Artemis found mutations associated with decreased cardiac pumping force, which would provide gene targets for chronic heart failure, Rounding said.

The advantage of the secondary in vivo screen is that mutants with the gene identified for the CHF phenotype can then be observed for blood pressure or cardiac compressibility characteristics, for example, or ion channel differences in cardiac myocytes can be examined, Rounding observed.

"These secondary studies will give us detail to allow us to identify those reasons which we would subsequently wish to develop genes," he concluded.

The Tuebingen 1 screen also yielded "loss-of-function" mutations associated with proteins necessary to form cartilage. The findings could produce a therapeutic protein or a small molecule drug target.

Artemis has found that transplantation of wild-type genes into the cartilage-deficient mutants will rescue the zebrafish's ability to produce cartilage, even in distant cells. The results suggest the mutation is in the secreted protein, making discovery of a therapeutic protein that will rebuild cartilage in arthritis patients likely, Rounding suggested.

A secondary high-throughput screen is being run with 30 zebrafish larvae per well in 48-well plates. If a secreted protein is identified that may be therapeutic, or a potential drug target is found, the company would seek a development partner, Artemis indicated.

The company also has an inducible knockout mouse model system, which can be used to validate discoveries from the zebrafish models, or to take targeted genes from pharma companies and "give them a functional basis for starting their screens," Rounding said.

While conventional knock-out mice are born without a designated gene, the Conditional mouse allows a selected gene to be knocked out after the mouse has reached adulthood. One technique to accomplish this is incorporating an estrogen receptor/recombinase fusion protein that will knock out a gene upon exposure to tamoxifen, Rounding explained.

Artemis was co-founded by South San Francisco-based Exelixis, and launched in September 1998 (¹ (Also see "Artemis Initial Financing Will Broaden Genomics Platform With Exelixis" - Pink Sheet, 5 Oct, 1998.)). Exelixis uses Drosophila flies and C. elegans nematode worms for model screens; those organisms already have their genomes entirely sequenced and are known to have well-conserved disease pathways with higher organisms, particularly in cancer and metabolic diseases.

However, the zebrafish genome is almost completely sequenced. Artemis estimates that by the end of the year, as many polymorphic markers will be known for the zebrafish as are known for the mouse and the fruit fly. This system is better for CV and bone/cartilage modeling, the company indicated, and it has 90%-95% homology with human genes.

Another firm that presented at the E&Y conference, Ghent, Belgium-based DevGen, bases its high-throughput screens on the nematode worm. CEO Thierry Bogaert, PhD, pointed to the three-day lifecycle of the organism, its transparency, the detailed knowledge surrounding the nematode genome, and about 75% homology with the human genome as ideal for "assayable" and "druggable" targets.

DevGen intends to "generate animals which mimic a disease state and cure them back to normal either by finding a mutation or a compound," Bogaert said. About 20%-30% of the compounds found this way in the nematodes will have activity in mammalian systems, he claimed.