"Evidence-based innovation in biomedicine: potential and pitfalls" Mark M. Rasenick Distinguished University Professor, Physiology & Biophysics and Psychiatry. U. Illinois Chicago College of Medicine Chicago Illinois USA raz@uic. edu
You can’t always get what you want. • But in biomedical research, you often do just that.
Requirement of Hippocampal Neurogenesis for the Behavioral Effects of Antidepressants Luca Santarelli, 1* Michael Saxe, 1* Cornelius Gross, 1 Alexandre Surget, 2 Fortunato Battaglia, 3 Stephanie Dulawa, 1 Noelia Weisstaub, 1 James Lee, 1 Ronald Duman, 4 Ottavio Arancio, 3 Catherine Belzung, 2 Rene´ Hen 1† Science 8 August 2003 2611 citations • Since then, at least 8 papers published demonstrating neurogenesis has no causative role in antidepressant action. And just about everything you do to an animal increases neurogenesis. • Fewer than 1200 citations, cumulatively. • But wait, it gets worse
Only 28% of published data were reproducible by pharmaceutical industry. Inconsistencies included inconsistent drug targets and actions (literature vs, inhouse studies). Prior to this study, it was “common knowledge” that only about 50% of academic studies were able to be validated by Pharma. Prinz et. al.
Selective Data Collecting: Selective Data Reporting • Non-random data recording (blind microscopy)—looking where the light is, • Publish the “good results” • Nobody wants to know about negative results. (clinical trials. gov/PLOSone)
Increased protein complexes? Depends upon where you look
Selective Data Collecting: Selective Data Reporting • Non-random data recording (blind microscopy)—looking where the light is, • Publish the “good results” • Nobody wants to know about negative results. (clinical trials. gov/PLOSone)
Gs is distributed preferentially in lipid raft fractions in post mortem brain of depressed subjects * P < 0. 05 Total Gs is not significantly different between suicide and control samples Raft/non-raft Gq is not significantly different between suicide and control samples Only the Raft distribution of Gs is variable Donati et. al, 2008
Gs is distributed preferentially in lipid raft fractions in post-mortem brain of depressed subjects ** P < 0. 02 * P < 0. 05 Total Gs is not significantly different between suicide and control samples Raft/non-raft Gq is not significantly different between suicide and control samples Only the Raft distribution of Gs is variable Donati et. al, 2008
Problems with clinical studies • Placebo response: somatic disease. • Placebo response: Psychiatry • Volunteers: some do it for the money: some just enjoy volunteering. This is a problem unique to psychiatry—much more difficult to fake a somatic ailment.
How do we fix this? • Biomarkers can: • 1) screen volunteers for presence of disease. • 2) provide a standardized hallmark of sickness or wellness.
Sensitivity and Specificity of Gs Biomarker • • Sensitivity=. 93 Specificity=. 78 PPV=. 88 (LR+ = 4. 20) NPV=. 88 (LR- = 0. 086) Note: these data are obtained by differential detergent extraction; not preparation of rafts by density gradient sedimentation. LR+ /LR- : likelihood ratio
Lipid Rafts and caveolae act as inhibitory domains, separating Gs from the GPCRs that activate it. b 2 AR Adenylyl Cyclase Lipid Rafts b g as GTP as as GTP ATP c. AMP as as as Chronic antidepressant treatments liberate Gs from caveolae/raft membrane domains and foster coupling to adenylyl cyclase (Zhang and Rasenick, 2010). Depression has the opposite effect, as Gs is enriched in lipid raft fractions prepared from suicide brains (Donati et al. , 2008) So, can raft localization of Gsa be a biomarker for Depression and antidepressant response?
Fluorescence Recovery After Photobleaching (FRAP) provides reproducible measurements to study the mobility of GFP-Gαs in C 6 glioma cells.
Lipid Rafts and caveolae act as inhibitory domains, separating Gs from the GPCRs that activate it. b 2 AR Adenylyl Cyclase Lipid Rafts b g as GTP as as GTP ATP c. AMP as as as Chronic antidepressant treatments liberate Gs from caveolae/raft membrane domains and foster coupling to adenylyl cyclase (Zhang and Rasenick, 2010). Depression has the opposite effect, as Gs is enriched in lipid raft fractions prepared from suicide brains (Donati et al. , 2008) So, can raft localization of Gsa be a biomarker for Depression and antidepressant response?
Factors modifying Gs FRAP
Randomization helps, but not all questions can be randomized…. ) • Question: does being physically disciplined as a young child increase the risk of being incarcerated as an adult? • Without randomization, it is difficult to discern causality from correlation • those who are disciplined more might be those who are already anti-social! • Yet one could never randomly assign children to different amounts of physical discipline! • Besides the ethical issues, the mere act of administering physical punishment in an arbitrary manner changes the entire dynamic of the situation that you want to study.
Sometimes controlled studies are difficult • Consider the putative therapeutic/palliative effects of marijuana. • The subjects who do not experience psychoactive effects know that they are the controls (although some of them still get “high”)
Animal studies don’t have these drawbacks • Animal models exist for many diseases • Genes are relatively easy to manipulate, even in mammals. But what have we learned from the genome? • Most strains are syngenic, little genetic variation—but that’s also a drawback– Humans are quite variable. Also, some genes have multiple roles during the course of development (conditional k/o addresses this). • Do mice have mood? Monkeys do. • Yeast to humans, successive approximation. • Human stem cells, including those derived from blood and skin (i. PSC) allow us to compare neurons from normal vs. disease.
But, at the end of the day: • We can all get some satisfaction, knowing that we are a bit closer to the truth.
Скачать презентацию Evidence-based innovation in biomedicine potential and pitfalls Mark
6d7ebac145229cbe58c7d6c32c897c05.ppt