Cotcha Yankinov surely knows his work well. I can just imagine him leaning down to pick something off the floor of the lab only for an NMR sample tube dropping out of his top pocket. I bet this happened once. Then there is always someone who is a passion fingers. Is it an apocryphal story of a sub-graduate putting a sample directly into the NMR, not in a tube? I keep hearing it. I forget the name but a guy in NZ selling a tiny NMR setup that used solid-state cooling. I forget the company name but it was the size of a small (European) fridge and was silent in operation. I would have loved one but didn't have the $750,000 on me at the time... or ever.
-Check for gross hazards so standard banks of tests (hERG, Ames, Membrane Potential and so on) carried out by people like Huntingdon Life Sciences. hERG, Ames and other required data.
-Check that the compound fits Lipinski's rule of 5. A low cLogP means a drug won't concentrate in brain tissue. If the pKa at plasma pH means 100% of the material carries a formal charge, it won't cross the BBB. How is it metabolized? What physiological effects do the metabolites display?
-Check for the affinity of the compound. How many classes and subtypes of receptor will it interact with? That known, what action does it have at all of those receptors? We are getting much better, but 'silver bullets' are rare and the low-hanging fruit has been harvested.
-Check existing papers and patents. Reaxys is wonderful but costly. If you can access from an educational establishment then do so, it will save you a lot of pain. TGN1412 almost killed 6. A Previous CD28 ligand had resulted in severe reactions and been reported. Nobody checked!
-Now you need to know what action the material actually has. Phases of clinical research uncover first acute and then chronic side-effects. Gaboxadol is a good example. Failed to reach clinical endpoints so someone bought the rights to try it for other medicinal action.
Many CNS agents exhibit neurotoxicity. It seems more like the rule than the exception. It is the level of toxicity and the bodies ability to restore the damage that defines how toxic it is. Sometimes, CNS compounds damage other organs. I won't detail alcohol but both aminorex and fenfluramine exhibited 5HT2b affinity which causes cardiac valvular fibrosis. Bupropion causes seizures. From the gray market, the ortho substituted n-methylaryl PEA class 5HT2a agonists (commonly referred to as NBOMes) exhibit toxic symptoms often enough for a lot of papers. Clin Toxicol (Phila). 2013 Mar; 51(3): 174–177. Clin Toxicol (Phila). 2015 Feb;53(2):85-92. Journal of Analytical Toxicology 2015;39:602 –606. There are many more but the cardinal symptoms are aggression, agitation, hallucinations, hypertension, seizure and tachycardia. Apparently seizure and Serotonin Syndrome (leading to malignant hyperthermia and Long-QT, itself becoming Torsades de Pointes).
What I do find a little worrying is that some medicinal chemists are a little too gung ho for my liking. If people are interested in the life of a stage 1 subject, there is even a jobzine for the work.
https://www.guineapigzero.com Now is it just me that feels that this model is unethical? If nothing else, the subjects have have other novel drugs in their system when they do a trial. Sooner or later the tailings of a statin will interact with the next novel compound? While thalidomide, Amobarbital & Propoxyphene are famous examples, rofecoxib, valdecoxib & trovafloxacin have destroyed many more lives. Even pronethalol, the first beta-blocker on the market was withdrawn because of it's carcinogenicity. That is why drug designers look at fused aromatic rings askance. That drug showed the risks of epoxides. Even that Chinese guy who named his product 'Eric 1' was selling a napthalene analogue of pyrovalerone! This even after a second benzene ring was shown to be a bioisostere of a propyl moiety (at least in SNDRI designs based on pyrovalerone).
Sorry.... just rambling.... sorry for wasting everyone's time.