Most of the high troughput testing selected molecules don't work in in vivo systems, and almost never in humans. My prediction however is that that the rule of 5 will not solve that problem. Novel drugs usually don't work because we overestimate the knowledge we have on the etiology of the disease and on the behavior of drugs in vivo. Some ideas: 1) Using screening tests on living animals in which the disease is acurately modeled eg by genetics (correctly modelling is an extreme challenge). 2) Carefully look at side-effects of existing drugs used in humans and get inspiration for new drugs based on that. 3) Accept that you don't understand all details of the disease. Knowledge often blinds the scientists. Coincidence rules the world. Let's use it intelligently!

Lipinski's Rule of 5 in early stage of selection of molecules has no theoretical basis and hence is a limiting factor for drug development. Chemistry and especially medicinal chemistry are not like physics and has many such rules. Basic flaw with them is they are usually empirical and so are based on data available at that time and naturally loose value with passage of time with accumulation of more data and more exception to the rule. As drug development matures such rules will come and go with time. there is no harm in using them just like another tool and not depending on them beyond a point and this may vary person to person depending on the availability of other tools of drug development. SAR is one such most commonly used by medicinal chemist tools without much theoretical basis. In fact existing of many so many empirical methods tells again and again that drug-development is far from maturity. Drug-development requires in depth understanding of forces responsible for the formation of drug-receptor complex which itself is developing area and comes in the domain of supramolecular chemistry. In my opinion a modern medicinal chemist involved in new drug-development has to be both chemist and supramolecular chemist.

Do not throw away compound if you have found new pharmacophore fragment. Physico-chemical properties can be tuned in lead optimization. Best way is to contact with chiefs of screening programs in corporation.

Lipinski Rule of 5 is guide but not dogma - one point of view see at http://onlinelibrary.wiley.com/doi/10.1002/jhet.5570370324/abstract

Kamlakar, I agree with what you say about Ro5 and SAR, would anybody else like to comment?

Thanks Andrei that gives me some train reading!

As Andrei suggested, don't through the potent molecules, it might be possible to improve the properties of such molecules at later stage to bring them under the rule of 5. From my experience in the screening of chemical compounds, I've seen poorly molecules, with higher potency can go to later stage post modification.

Often the ro5 is just to narrow the number of compounds to eventually screen in a VS campaign. When working with a relatively small dataset and having sufficient computational resources, I often ignore Lipinski's ro5

The rule of 5 was an empirical observation. The analysis also showed that compounds were, in general, getting larger and more lipophilic. This has been tied to poorer drug developability and lower success in getting molecules through to the market. As in all things, you can use past experience to help understand the weaknesses and risks associated with the particular drug discovery campaign you are engaged on. My experience is that the Ro5 and other metrics (ligand efficiency and LLE) are useful in highlighting whether you are making genuine progress or merely gaining potency by adding lipophilicity. It also removes the single minded focus on potency at the expense of developability. I would suggest that, in the best scenario, we would have multiple series and be able to prioritise on the basis of molecules with the best inherent developability properties (particularly if we are aiming for a CNS target). In the event of having a single series, the metrics tell you how much risk you carry with you and should alter the order with which you carry out the experiments to address the risks.

Final point is that the rule of 5 is for oral drugs - other routes may have very different 'sweet spots' in terms of the physicochemical properties most likelyto deliver successful drugs.

... just to follow on this, there have been several study in the past highlighting the fact that often optimised compounds fall outside the lipinski chemical space. Just as an example:

http://www.ncbi.nlm.nih.gov/pubmed/16533713

At times is better to consider lipinski rules as a warning rather than a filter.

David and Fabio: Thanks, you have confirmed at least some of my suspiscions reagrding the Ro5.

Here's another question then:

Can anybody point me towards any papers that show direct evidence for passive diffusion through the plasma membrane of a living cell for any drug. It's a horrendous thing to try and trawl the literature for.

Difficult to find clear references - there is evidence of molecules getting in and PAMPA assays can suggest that they get in through passive diffusion but I dont know of any definitive proof. Nearest I have seen was some work done with Raman microscopy. Again, they show that molecules get in but I think there is evidence that the rate is consistent with passive diffusion. I will check for the reference.

Book chapter with originals citation:

http://www.pharmpress.com/files/docs/molbiopharmaceutics_sample.pdf

Most of the high troughput testing selected molecules don't work in in vivo systems, and almost never in humans. My prediction however is that that the rule of 5 will not solve that problem. Novel drugs usually don't work because we overestimate the knowledge we have on the etiology of the disease and on the behavior of drugs in vivo. Some ideas: 1) Using screening tests on living animals in which the disease is acurately modeled eg by genetics (correctly modelling is an extreme challenge). 2) Carefully look at side-effects of existing drugs used in humans and get inspiration for new drugs based on that. 3) Accept that you don't understand all details of the disease. Knowledge often blinds the scientists. Coincidence rules the world. Let's use it intelligently!

During hit search oral bioavailability is not a priority. Therefore instead of lipinsky rule use filters to eliminate highly reactive groups, protein chelators. When hits are obtained try to enhance potency by lead modification when moving from in vitro to in vivo a proper balance should be made between potency and bioavailability only if target and treatment duration demands it. However lipinsky rule is not hard and sole rule about 15 % of drugs doesnot obey this rule. More over if your compound is potent enough then it can be given by non oral route if oral bioavailability is an issue. Ultimately oral bioavailability should be considered only when other routes are not viable. Ex antidiabetic drug which have to be taken daily for long duration needs strict oral bioavailability while antibacterial, antiviral drug need not require such criteria.