Hi Deborah - it has always been more of surprise to me that so many people have used imipramine as a positive control and found it to be effective - given that it is only poorly effective in the clinic.

This paper argues that if imipramine has any antidepressant activity at all this is likely to be mediated by its effects on stimulating social behaviour rather than a direct antidepressant effect per se

http://www.academia.edu/1215804/Short-term_individual_housing_induced_social_deficits_in_female_Mongolian_gerbils_Attenuation_by_chronic_but_not_acute_imipramine

This second paper argues that models such as FST and TST are merely screens for monoamine activity rather than actual models of depression and that the poor antidepressant efficacy of the MAOI reputake inhibitors in the clinic means there is no real purpose in screening for more.

The paper also outlines the logical flaw that undermines these models and why this means that they were never actually fit for purpose - and hence why no progress has been made using them.

http://www.academia.edu/2240421/The_failure_of_the_antidepressant_drug_discovery_process_is_systemic

Dear Collin, thank you for your most helpful answer. We used FST and TST as screening for antidepressant effect of an as yet untested compounds. But I agree with you that these are less than ideal models. Thanks again

Dear Deborah, many factors may affect the antidepressant-like response in the FST. You may give a look to our paper and references cited therein (Cervo et al 2005 J. Neurosci 25(36)). 8165

Hi Professor Deborah (we met two years ago at the III INF Summer School in USP-RP) - we're also trying to use the FST as a screening model, but we'll use fluoxetine (FLX) as this is more effective than other drugs both in the clinic and in experimental tests, and its mechanism of action involves mainly serotonergic pathways (we're testing some possible serotonergic compounds). I'm not sure if there has been developed any other test (apart from the TST) for screening antidepressant compounds, and although I agree with Dr. Collie's reply, I believe that until up to now the FST and TST are the best tests for such a purpose. On the other hand, the TST shows a greater sensitivity -at least to some compounds- than the FST in mice (Cryan et al., 2005), so this would be a better choice. Good luck Professor.

Dear Abel - Apologies for repeating the point but the FST and TST are logically flawed - they are models of monoamine activity only - this would not be an issue if compounds that influenced mononamine activity were effective antidepressants but we now know they are not - the most recent meta-analyses show that monoamine reuptake inhibitors are no more effective as antidepressants in the clinic than placebo.

The consequence is that we can no longer justify using these ethically highly questionable procedures on the grounds that they will lead to better medicines - they will not - and indeed never could - since the logical flaw that undermines them means they are capable only of identifying drugs that work in the same way as the drugs we already have (and that we so desperately need to replace).

These points are further explained in the second paper I linked above and this paper also suggests ways in which we can move forward from this very difficult situation.

Dear Angelo - thanks for your interesting comments.

These raise three questions in my mind which I hope you don't mind me asking

Firstly, you say that the FST is a test of at least 5-HT and NA - this is very much an observation built around the monoamine hypothesis. Hence I was wondering what clinical evidence there was to support this hypothesis (i.e. what clinical studies are there to show that 5HT/NA levels are low in patients with depression and that these are increased by treatment with antidepressants in a way that is correlated with clinical improvements)?

Secondly, the assumption behind models such as FST and TST is that currently prescribed antidepressants are more effective in the clinic than placebo. Indeed, their sensitivity to these drugs is the very basis of their definition as models of depression. However much of the evidence over the past 10 years or so that also incorporates the results of unpublished clinical trials very much fails to support this assumption. Therefore I was wondering if you thought that the lack of clinical efficacy of the drugs these models have been built to detect undermines their usefulness?

The third question is to do with Ketamine. There has of course been much hype surrounding this but to the best of my knowledge the key studies still need to be done - that is double blind evidence including another hallucinogen as a control is still missing and we do not know the effects of chronic treatment in this context.

Hence I was wondering if you thought enough work had been done on Ketamine yet for us to be confident that this won't be just another new 'wonder drug' that fails once it gets into the clinic'?

The issues as I see them are to do with efficacy over appropriate control and chronic treatment - the potential for tolerance to develop of course but there is clear reason to have concerns about toxicology since the drug is already known to have severe adverse effects in street users.

Hi  Deborah, everyone here raises excellent points about these models, what they mean, and if we should use them.  But going back to your original question, I have seen imipramine not work in these models.  This could be due to strain differences (e.g., Witkin et al. 2014 reports that the CD1 strain is resistant to imipramine).  I've also sometimes had lower doses  (10 mg/kg or below) fail to work, depending on who was running the experiment, whereas high doses (20 - 30 mg/kg) more consistently work.  In addition, in our hands, colder water temperatures seems to weaken the efficacy of imipramine in mouse FST.  Finally, if you're using a camera software system to analyze the animals, the settings need to be fine tuned to correlate with what a blind human observer is recording.  If the wrong settings are being used or the animal is being poorly tracked by the cameras, you can fail to see effects.