This Small Business Innovation Research (SBIR) Phase I project proposes to develop a highly predictive model for assessing the potential of compounds to cause birth defects in the developing human embryo using human embryonic stem (hES) cells as the test substrate. The project proposes to use hES cells and metabolomics to understand the impact of drugs and other chemicals on the development of the human embryo. These cells have the ability to differentiate into any cell in the body and as such, offer the opportunity to study defects in development in a way never available prior to the isolation of hES cells from an embryo.
The broader impacts of this research are to increase the safety of compounds and to prevent birth defects resulting from exposure to drugs or other chemicals during pregnancy by more accurately predicting the potential for compounds to cause birth defects. Compound exposure is responsible for 4-5% of all birth defects, yet this is the most preventable type of birth defect. Currently, animal models are used to predict birth defects, however these tests are costly, time-consuming, and are only 60% predictive of the effect on human development. These animal models are the same tests that have been used for more than fifty years to predict the effect of drugs like Thalidomide and Accutane which have caused numerous birth defects in humans. More accurate screens are needed to predict if exposure to specific environmental chemicals or drugs will be hazardous to development.
Birth defects are a leading cause of postnatal deaths and pediatric disorders affecting approximately 3% to 5% of children born in the United States. Any chemical that is known to cause birth defects is called a teratogen. Between 2% and 3% of birth defects can be classified as teratogen-induced malformations. Teratogens include a wide range of substances, including pharmaceuticals, environmental contaminants and radiation. Currently, live animals are tested and sacrificed to determine if drugs and other chemicals may be teratogens. These tests are only accurate about 60% of the time. The most striking illustration of this problem occurred when thalidomide was released into the market in 1954 for the treatment of morning sickness in pregnant women. Thalidomide had been tested in rodents, and declared safe for pregnant women; however by 1962, when it was removed from the market, it had caused more than 10,000 cases of birth defects. In addition, there is a significant global push to reduce the use of animals in toxicity testing. We are developing a teratogenicity test using human embryonic stem (hES) cells that produces more accurate human results than animal tests, given that hES cells originate from the developing human embryo. This test will reduce the testing costs and time of chemical testing, as well as increasing predictability thereby helping to decrease the incidence of birth defects. The use of non-human embryonic stem cells to test developmental toxicity of chemical substances, such as the mouse embryonic stem cell test (EST), has been previously established. However, the biological mechanisms of the EST are still poorly understood, the test is not specific to human response and the actual measurement (the number of beating heart cells from the embryoid bodies) is somewhat subjective. We use an analytical technique called liquid chromatography mass spectrometry (LC-MS), a standard analytical approach in metabolomics, to detect how the cellular small molecules change after treatment with the chemicals being tested for teratogenicity. Metabolomics delivers sensitive and quantitative biomarkers, while also providing additional information about biological mechanisms and pathways that are affected, unlike most other developmental toxicity tests which can only deliver a "yes, no or maybe" answer to the question of a compound’s teratogenicity. We analyze the cell media outside of the cells because molecules released by the cells may also be measurable in other biofluids such as serum, urine or amniotic fluid. One of our future goals is to examine the biomarkers in biofluids using procedures that are less invasive than tissue biopsies to detect metabolic signatures associated with disease. In our original study, hES cells were cultured in 6-well plates and dosed with a test set consisting of 20 drugs of known teratogenicity. Statistical analyses performed on the resulting data to group and compare the data sets to determine changes in the levels of the molecules. This allows us to determine a metabolic signature from the cells. Next we identified a set of candidate biomarker molecules from the cells that consistently indicate potential teratogenicity. A statistical model that was developed and tested in two blinded studies correctly predicted the teratogenicity for seven out of the eight drugs. Subsequently, in our NSF SBIR Phase I study, we made our test faster and automatable by changing from our original 6-well cell culture plate to a 96-well plate format. We treated the hES cells cultured in the 96-well format with a test set and identified chemicals that the cells released in response to this treatment. We also increased the robustness of the assay be performing testing a total of 11 non-teratogens and 12 teratogens on two additional cell lines. We discovered that no metabolites were unique to a single cell line and few differences in their levels were observed, concluding that the general metabolic makeup of the cell lines is similar, however results suggest that the cell lines did respond differently to drug treatments. This study has helped to provide us with a set of biomarkers that are biologically relevant across diverse genetic backgrounds. By doing so, we are arriving at a set of biomarkers of teratogenicity that will be repeatable in genetically independent lines and more likely to mirror actual conditions in a pregnant woman. We used this set of biomarkers to develop new predictive models of toxicity. In a blinded study, these models successfully predicted the teratogens with up to 90% accuracy. We also validated the chemical identities of several candidate biomarkers by performing additional mass spectrometric analyses. This provides increased confidence in our data, biochemical pathway determinations and models. The culmination of this work is the development of a commercial assay to test for teratogenicity with the ultimate goal of reducing human birth defects. Our assay, "DevTOX" is now ready for commercial application and we are grateful to the NSF for the funding provided to perform this ground breaking work.
