Thread by @nicoleacrowley

#SOBP2019 update of the famous ACNP panel, animal models of neuropsychiatry

from @EricJNestler best established risk factor for depression is still stress (chronic, repeated, emotional)

& should we be looking for treatment resistance? isn't that a better model?

We are not paying adequate attention to validating animal models (i.e. post-mortem brain tissue)

”from the grapevine” → program officers have been (overly) critical of specific stress models. They are stewards of public money, but have been over-reaching and dictating experiments → younger faculty are most vulnerable to this

we shouldn’t be funded if we’re submitting the same thing over and over again. Obligation on faculty to innovate.

up next @TheBaleLab "when we overstate our findings... we are doing a disservice to ourselves... over-promising and not delivering to patients. clean up language"

If your sex differences in animal models don’t replicate human literature, what is that telling you?

sex differences in disorder may only present in different stages in life (i.e., mood disorders not different before pubertal onset)

inter-generational transmission is also sex-specific (i'll plug @nila_lab_PSU here)

not all sex differences are related to hormones. They contribute but are not everything! Sex chromosomes = interesting (x,y, linked genes)

change in GABAergic activity during pregnancy (@jmaglab's work being discussed) = first FDA approved treatment for PPD

brain is not binary, brain can be "more male like or more female like"

"NIH is not asking everyone in this room to study sex differences, but consider sex as a biological variable like any other variable"

"& when did male become the control???"

Next up Scott Thompson from University of Maryland (on twitter? yes no?)

predictability: what models will best lead us to compounds and approaches to treat human disease? How come after 50-60 years of basic neuroscience we only have a few new targets? Are we using the right models/levels of rigor?

-our obligations as preclinical investigators: limit use of the term “depression” to humans, not mice
-only call a drug antidepressant if it works as an antidepressant in humans

(theme of session: rigor in language)

Transition from DSM5 to RDoC breaks down symptoms into individual behavioral components, this improves science and rigor in preclinical laboratories.

you should have some idea how a drug is acting on the circuits/synapses/in the black box. What’s the target? Why should that target change behavior? Need to have a mechanistic hypothesis

-cheap, easy, quick tests of antidepressant efficacy fail in a lot of ways (i.e. SSRIs in many mouse behaviors will produce an immediate response, ie FST, even though that’s not how SSRI response in humans works)

32 compounds in clinical FDA pipeline in 2018, 2 have made it through since ACNP (ketamine, brexanolone)

up next is Kerry Ressler. areas of psychiatry with most progress: conserved reward circuitry / addiction (ventral striatum, reward processing, etc) & neuro circuitry of threat

how do we dissect subcircuits more and more? High throughput single cell sequencing (@gstuber )

next generation of human phenotyping: digital phenotyping (i.e. EEG, active/passive data on phones) enormous amounts of data, unbiased. treat the human as an unbiased subject

Up next "headliner" (Kerry's words, not mine!) @NIMHDirector

.@NIMHDirector starts by reminding history of this session: all this was taking place at a closed meeting (i.e. ACNP), needed to be repeated at an open meeting

traditionally used screening approaches to find new therapies (mice, FST, some decently predictive assays) or screening approaches in humans & just tried them. But lots of money wasted with that approach (lots of failures)

-we can’t use these models to capture the complexity of the illness, but we can use animal models to study the underlying processes and mechanisms. Let’s figure out some level of behavior which is translatable and allows us to get to underlying mechanism...

... to identify targets (typically molecular, not always). Screen TARGETS not BEHAVIOR.

(when looking at grants) is the question impactful? Will the study answer the question? Is the model the right one to answer the question? (efficient and ethical use of resources, feasibility, evolutionary conservation of mechanism)

cross species, strain, sex comparison are incredibly important. Experiments motivated by modeling disease not as informative as those testing mechanisms

& now panel taking questions/open discussion, but that's a wrap of my live coverage!