Home » The Rule of Two

The Rule of Two

by Darryl B McConnell
7 minutes read

In contrast to molecular weight and H-bond acceptor count, H-bond donor count and lipophilicity remains constant over time for approved oral drugs. This highlights that H-bond donor count and lipophilicity are tighly linked to drug-likeness.

recap of the rule of 5

Lipinski’s seminal paper outlining the rule of 5 brought molecular properties into the limelight of drug design. Specifically, molecular size (although a metric for weight rather than size was used), lipophilicity, H-bond acceptors and donors were highlighted as predictors for orally bioavailable drugs. This has clearly had a massive impact on the scientific community with now over 25000 citations.

Rule of 5 Citations per Year

This analysis led to a simple mnemonic called the ‘rule of 5’. The rule of 5 focussed not on druglikeness but rather drug unlikeness, focussing on cutoffs that defined the 90th percentile of approved oral drugs. eg only 10% of approved oral drugs (at the time) had a value greater than each of the four rules.

The ‘rule of 5’ states that: poor absorption or permeation are more likely when:

  • There are more than 5 H-bond donors (expressed as the sum of OHs and NHs);
  • The MWT is over 500;
  • The Log P is over 5 (or MLogP is over 4.15);
  • There are more than 10 H-bond acceptors (expressed as the sum of Ns and Os).

Compound classes that are substrates for biological transporters are exceptions to the rule.

If two of the above four criteria are exceeded then a compound is said to be Ro5 non-compliant.

the properties that change over time

A drug-like property is a molecular property that predicts drugs and hence is arguably the most important aspect of drug design. The predictive value of aspects of Lipinski’s rule of 5 has long been debunked by Schulz who demonstrated that the Rule of 5 did not hold over time. A more recent review of the Rule of 5 is available from Hartung. While a lot of the scientific discourse has focused on the “Rules” and their predictive value, Shultz made the important observation that only two of Lipinski’s four properties didn’t changed over time. If a property is truly a drug-like property then it should not change over time. Schulz’ examination of oral drug parameters approved before and after the original Ro5 analysis demonstrated that molecular weight and H-bond acceptor count increased substantially over the ensuing 20 years. This highlights that the much talked about property molecular weight and also H-bond acceptor count are not drug-like properties (see below).

Molecular Weight of Approved Oral Drugs Over Time (1960 to 2020)

H-Bond Acceptor Count of Approved Oral Drugs Over Time (1960 to 2020)

the properties that don´t change over time

In contrast to molecular weight and H-bond acceptor count, Schulz highlighted that H-bond donor count and lipophilicity (clogP) remained essentially constant over time. This highlights that H-bond donor count and lipophilicity are tighly linked to drug-likeness.

H-Bond Donor Count of Approved Oral Drugs Over Time (1960 to 2020)

Lipophilicity of Approved Oral Drugs Over Time (1960 to 2020)

the rule of two

Clearly with the advent of good predictve models for important drug properties such as metabolic stability and permeability the need for simple rules of thumb in medicinal chemistry is diminishing. However, given the didactic power of simple rules a “rule of two” based on the two properties which oral approved drugs maintain over time, HBD count and lipophilicity, is proposed (first disclosed at the EFMC Zagreb 2023). In keeping with Lipinski´s tradition, the 90th percentile was used as the upper bound for inclusion for oral approved drugs. Analyzing the properties of the 1733 approved drugs from the Chembl database at the time of writing highlights that 90% of all historical approved oral drugs have a H-bond donor count of 4 or less and 90% have a ClogP of between -1 and 5.5.

Thus the Rule of Two states: In order to maximize your chances of discovering an oral drug, don´t go over 4 for H-bond donor count & stay between -1 and 5.5 for ClogP. The Rule of 2 zone and approved drugs mapped on to this space is illustrated in the figure below.

Rule of Two: Don´t go over 4 for H-bond donor count & stay between -1 and 5.5 for ClogP

The Rule of Two for Oral Drugs (including the Rule of 2 Zone)

rule of two properties versus MW

Typically, during a small molecule optimization program, the molecular weight of drug candidates increases on going from hit to lead to a development candidate. As such, it is interesting to analyse the trends for HBD count and logP with increasing MW of approved drugs, the trends of which can be used a surrogate guidline for compound growing during optimization. Interesting, the average H-bond donor count of approved drugs does not change with MW. This highlights the importance of HBDs regardless of size thus the optimization process should be strickly limited. (hence the Ro2 upper limit of 4). This also highlights the importance of minimizing H-bond donors when selecting hits and including a minimum HBD count for inclusion in compound screening libraries.

H-Bond Donor Count of Approved Drugs versus Molecular Weight (1960 to 2020)

Equally interesting, is the fact that lipophilicity increases with increasing MW within approved oral drugs over time. This suggests a less strick approach could be taken with respect to lipophilicity during compound optimization. This is reflected in the relatively wide ClogP range given on the Rule of 2 of -1 to 5.5.

It is worth highlighting that the ubiquitously used logP(octanol-water) measurements and calculations have the limitation of “not seeing” H-bond donors. This because the oxygen atoms in octanol and water are equally good acceptors and thus compete for H-bonding with H-Bond donor a of solutes. It is intruiging that it is precisely the molecular parameter that ClogP does not measure which make up the pair of drug-like properties of the Rule of Two. More on this when we get into the asymmetry of H-Bond donors and acceptors.

Lipophilicity of Approved Drugs versus Molecular Weight (1960 to 2020)

 

Related Articles

2 comments

DChen 09/01/2024 - 3:22 am

Like your writings and appreciate the sharing of MedChem knowledges. As PROTAC and peptide oral drugs advancing to the market, I am wonding if the trend of the two parameters still holds unchanged in the next ten years.

Darryl 09/01/2024 - 9:39 am

Thanks! The current oral, clinical PROTACs are Rule of 2 compliant. I’ll do a post later in the year on the Ro2 and big compounds (eg PROTACs and macrolides)

Comments are closed.