Meeting Notes - December 16, 2008

Summary of main points of discussion at first face to face meeting of GLUE investigators. December 16th 2008.

1. Translational Therapeutics

(a). Humans.
Emphasis in human studies initially on detailed deep phenotyping of drug response involunteers. This will be particularly true of attempts to define differences in efficacy. Recognition that relief of "pain" moves from evoked pain responses through models like dental pain to chronic pain like OA. Recognition that gene variability impact on analgesic effect is most easily defined in the first rather than the last, where overlay from disease variability may obscure the phenotype/genotype relationship.

(b). Mice.
Exploit and define the variability in mice already noted in phenotypic perturbations of this pathway – e.g. variability in blood pressure response to NSAIDs as a function of genetic background. Exploit interaction with the Collaborative Cross, an international consortium making multiple strain crosses – e.g. to identify genetic modulators of the BP/NSAID interaction that is highly variable in people. Focus on gender dependent differences in phenotype.

One might take a strain-based approach, or use recombinant inbred lines, which are commercially available from the Jackson laboratories. F2 studies would need to be confined to a single site, as exchanging individual mice between sites would be tricky at best. Suggestion of Wiltshire (UNC) or Schadt.
A challenge will be to establish ground rules for (i) sharing mice and (ii) unusual phenotyping facilities – e.g. imaging. Altman suggestion of common MOU based on PharmGkB an early implementation step to facilitate this process. This has been drafted pending informal review by NIGMS.

An unusual and unique repertoire of mice available to probe this pathway. Besides KO and knock down mice for both COXs, COX mutants that segregate COX from POX activity, COX-1 under the COX-2 promoter, mice lacking HuR, a major RNA binding protein that regulate COX-2 and other inflammatory response genes, multiple cell specific deletions of COX -2 and mPGES-1, we have the capacity to define biochemically COX-2 selective regimens using urinary PG metabolites. The possible importance of substrate rediversion can also be addressed with mice lacking 5- LOX or FLAP and the role of sphingolipids by mice lacking relevant receptors. Opportunity to apply increasingly sensitive broad spectrum metabolomic approaches and make SILAC mice. Soluble epoxide hydrolase KOs available to probe EET pathway. The development of more comprehensive lipidomics strategies opens the possibility of performing flux studies as a prelude to such analyses in drug treated humans.

( c ). Fish.
Perhaps ideally suited to probing variability e.g. the prior variability that we observed with COX-1 morpholino driven phenotypes. Simple phenoclustering of zebrafish developmental phenotypes with a range of COXs may be useful for exploration of variability in off target effects. Many of the deep phenotypes from human studies can be adapted for the fish, and if scalable, will be useful for HTS to identify genetic determinants. Besides translucency, we know that we can measure relevant lipids in this system and see consequences of NSAID administration that mirror human experience. Have started to explore quantitative proteomics approaches in this system and plan SILAC style experiments. These and other lower throughput, but high resolution fish phenotypes can be used in rapid POC approaches for hypotheses deriving from systems approaches or epidemiology prior to mouse of human studies. It may be possible to systematically address environmental modifiers including dietary and nutraceutical exposures using parallel small molecule screens.

Nice model for RNA manipulation POCs. Can also introduce RNA here as in single cells to probe functional relevance of large dynamic variability observed in systems analyses.

Atherosclerosis model (gain of function mutations in LRP6) in fish being characterized – relevant to studies probing the question of emerging risk in patients at low initial CV risk with drug exposure in RCTs of coxibs.

2. Population Pharmacology.

Epidemiological databases probed to explore further the gender / drug response issue prompted by the suggestion in mice that COX-2 derived PGI2 contributes to estrogen dependent cardioprotection and by congruent data in an observational study of the HRT-NSAID interaction.

MOU with FDA permits access to data from RCTs which is currently being pursued.
Oxford led collaboration has access to individual patient level data from RCTs to address the CV question. Majority of these trials have (i) too few events to permit evaluation of covariates of risk of serious CV or GI events and (ii) have no blood or urine collections or insufficient DNA for large scale analyses. Another approach is to follow events backwards from plaintiff lawyers to seek stable genetic determinants of risk. However, this approach relies on highly selected cases, so one must be cautious. Little is known even of the factors that influence variability in susceptibility to GI risk, such as duration of dosing, time of dosing and gender. Age is an established determinant of GI risk, although it does not modify the impact of NSAIDs on GI risk. Same could be said of CV risk.

The risk of long term exposure of premenopausal females with juvenile arthritis to NSAIDs, especially those specific for COX-2 and the issue of whether NSAIDs foster accumulation of plaque burden are two intermediate term RCT possibilities. However, the longer term goal – most likely to initiate in the second cycle are outcome studies in refined populations including a repertoire of genetic, biochemical and physiological markers that seem likely to segregate benefit from risk.

Although much of the RCT activity may occur in the second cycle it is timely to initiate the development of information systems that will inform the design of those trials; harmonize the IRB submission structure and engage other relevant bodies – FDA,
AHRQ and CTSAs.

Dissemination strategies will be developed as the rationale for these markers accumulates to enable FDA, industry, institutes etc to access this information so that they may incorporate