For statistical analysis, the cell population was considered positive for PS externalisation

try to exactly reproduce any studies previously performed, but rather to extend the testing in an independent laboratory, using methods that were modeled as closely as possible on the human clinical trial experience. For example, we used glucose tolerance testing to screen NOD mice for “late prevention” studies, modeled on Type 1 Diabetes TrialNet’s approach to risk screening. Overall, the performance of the contract laboratory was sufficient to accomplish most of the testing 11 Efficacy Testing in Rodent Models of T1D goals. The laboratory also had the capability to perform mechanistic and pharmacokinetic studies for those agents that showed preliminary efficacy. Mechanistic studies were rarely done in advance of a demonstration of preclinical efficacy. Although it would have been preferable, drug dosing and pharmacokinetics could not be evaluated for every agent tested in the program. In most cases, since we selected drugs that had shown some prior efficacy, we used doses that had been shown previously to be effective. In one case, dosing and effects of drug were tested in a pilot prior to efficacy testing, because of a change in drug manufacturing. As reported here, the drug did not induce the expected acute effects, and further studies confirmed that the drug was not as active as expected. In other cases, drugs supplied by investigators were lead compounds and pharmacokinetic assays were unavailable, so it would not have been possible to determine whether drug failure was due to dosing or bioavailability. Tregitope, for example, is a drug in development for which dose, delivery, and route of administration have yet to be optimized. Any one of these drug attributes, or combinations 8832224 thereof, could impact the ability to demonstrate efficacy. However, agents such as the Tregitopes were chosen for testing because there was preliminary data that demonstrated at least some prior efficacy in T1D in at least one rodent model, whether or not PK and dosing 16392774 were optimized. When an agent had no prior testing in a T1D model, but for which PK was available, we did perform a dosing PK experiment. This program also illustrated some of the challenges in preclinical testing. It turned out to be a challenge to get agents into the program for testing. Some companies were hesitant to supply agents, given the expectation that all results, including negative results, would be made public. Some academic researchers were interested in the program initially, but declined to proceed with testing when they learned that they would not receive funds to perform the testing in their own laboratory under their own direction. We found that negotiation of a materials transfer agreement could be a barrier to participation for some companies, but progress was made on developing a document that NIDDK and most collaborators could accept. The methods presented here are an improvement in the state of the art, but our results also point to further opportunities for improvement. For example, adding standard pharmacokinetic and mechanism of action biomarker testing would BIRB796 site improve the interpretation of negative efficacy results. Improved preclinical testing does not always equate to standardization and optimization of methods, although those things are important. Preclinical testing results will perhaps be most robust when similar results are reported from different laboratories, using different models, and different but rigorous methods, all thoroughly reported. Since rodent m