You need to know the rules to break them. More eloquent versions of this adage have been attributed to visionaries as varied as Pablo Picasso and the Dalai Lama. But is an adage that works for famous artists and spiritual leaders holding medicinal chemists back? For more than 25 years, scientists have been grappling with Lipinski’s rule of 5, which is four guidelines based on multiples of five (hence the name). It was devised to help chemists design drugs that can be taken orally—the largest category of medicines.
For a medicine to be successful when it’s swallowed, it must dissolve in the stomach’s acidic environment and traverse the cell membrane in the gut. In 1997, Christopher A. Lipinski and colleagues working at the pharmaceutical giant Pfizer published guidelines on how to make compounds that can do just that by striking a balance between solubility and permeability (Adv. Drug Delivery Rev., DOI: 10.1016/S0169-409X(96)00423-1). Since its publication, the paper introducing Lipinski’s rule of 5 has been cited more than 24,000 times, according to Google Scholar, and has been republished twice. It has become both dogmatic and divisive.
For years, the rule of 5 was taken by many as medicinal chemistry canon. It helped speed up drug discovery and helped drug designers focus their efforts. But as more molecules that break the rule of 5 have become oral drugs, researchers have begun to question its value. Some chemists say the rule of 5 still provides useful boundaries for drug designers. Others argue that the rules limit the creativity of chemists and that it’s time to abandon them altogether.
It was during the summer of 1995 that Lipinski devised the rule of 5. Never a fan of meetings or other interruptions, Lipinski says he found summers to be “a good time to think and also just to have scientific fun.” As his Pfizer coworkers went on vacation and the relentless interruptions eased, he gained intellectual free time.
That’s when a colleague introduced him to the graphing software JMP. Lipinski used the program to make plots of Pfizer compounds that cleared the clinical trial hurdle between Phase 1 and Phase 2. “It had nothing to do with any projects I was involved in,” he recalls.
Looking at the graphs, Lipinski noticed some trends. Most of the molecules that made the leap from Phase 1 to Phase 2 fit within four key guidelines, all coincidentally related to the number five.
Their molecular masses were 500 Da or less; they had 5 or fewer hydrogen-bond donors, based on N–H and O–H groups; they contained no more than 10 hydrogen-bond acceptors, based on oxygen and nitrogen atoms; and their logP, a measure of lipophilicity, was 5 or less.
At Pfizer in the mid-1990s, drug discovery and development were separate enterprises, Lipinski says. Many compounds coming out of discovery were getting stuck in development. Lipinski suspected that their physicochemical properties were to blame. So he showed his rule-of-5 findings to his boss, Beryl W. Dominy.
It was too bad that Pfizer couldn’t incorporate the guidelines into the registration system where Pfizer chemists logged molecules that would be sent on for biological testing, Lipinski said. But Dominy, who ran the registration system, decided to include them. Chemists would enter the four measurements for any molecule intended for biological testing, and those that were outside the rule of 5 were flagged.
Adding the rule of 5 to the compound registration system “really made a difference in terms of the quality of the compounds that were being made,” Lipinski says.
The rule of 5 was never meant to explicitly exclude compounds that don’t fit the guidelines. Lipinski just wanted chemists to be aware they might encounter problems if the compounds are outside the prescribed ranges. “The rule of 5 would never stop someone nominating a compound” for biological evaluation, he says. “But what it did do is it raised the bar on the evidence that you had to present.”
After it was published, the rule of 5 quickly gained traction outside Pfizer. That’s no surprise, says Jan Kihlberg of Uppsala University, who was head of medicinal chemistry at AstraZeneca from 2003 to 2009. In the 1990s, it was in vogue to build combinatorial libraries of thousands, even tens of thousands, of compounds. But many of the compounds in those libraries weren’t amenable to being drugs. They were too large and too lipophilic, Kihlberg says. The rule of 5 got medicinal chemists to rethink their strategies.
The rule of 5 became a catchy mnemonic, and it shaped the thinking of a generation of medicinal chemists, says Ingo Hartung, global head of medicinal chemistry at Merck Healthcare KGaA. “It put thinking about compound properties really high on the priority list,” he says. It taught chemists to ponder properties like size and lipophilicity before they even began to imagine a molecule.
Number of hydrogen-bond donors:
Number of hydrogen-bond acceptors:
David DeGoey, a medicinal chemist at AbbVie, agrees that the rule of 5 “ushered in an era of property-based drug design,” spawning additional rules based on rotatable bonds and aromatic rings, among other properties. But, DeGoey says, “for a long time there was an overinterpretation of these rules and an overreliance on these rules.”
In 1997, DeGoey was a junior medicinal chemist at Abbott Laboratories (now AbbVie) working on developing drugs for infectious diseases. When DeGoey learned about the rule of 5, he says, he remembers thinking that the rules could never apply to the targets he was trying to address. “Many of the medicines for infectious diseases wouldn’t exist if you dogmatically follow the rule of 5,” he says.
Twenty years later, DeGoey led a team at AbbVie that analyzed the successes and failures in the company’s compound collection that broke the rule of 5, known as beyond-rule-of-5 compounds. At the outset, the researchers were dubious that any trends would emerge, but they found that even the successful rule breakers had to have the right balance of physicochemical properties to be oral drug candidates (J. Med. Chem. 2017, DOI: 10.1021/acs.jmedchem.7b00717).
Michael Shultz, a medicinal chemist at the Novartis Institutes for BioMedical Research, first heard of the rule of 5 during a job interview at Pfizer in the late 1990s. He was discussing some of the molecules he’d made as a graduate student when his interviewers told him those compounds would never work as drugs because they violated the rule of 5. “I was this young kid coming out of grad school,” he says, and he felt ashamed that he’d never heard of the rule of 5.
Years later, Shultz noticed that excluding compounds that broke the rule of 5 from Novartis’s screening libraries was eliminating many hits. He decided to do a retrospective analysis of oral drugs approved by the US Food and Drug Administration to see if the rule of 5 still held.
His thinking went like this: drug-like properties should be immutable—they shouldn’t change over time. But he noticed that while drugs still met certain rule-of-5 guidelines—for example, drugs still had 5 or fewer hydrogen-bond donors, and calculated logPs were still 5 or less—they flouted other limits. The molecular masses of many oral drugs approved after 1997 exceeded 500 Da, and they contained more than 10 hydrogen-bond acceptors (J. Med. Chem. 2018, DOI: 10.1021/acs.jmedchem.8b00686).
“If these properties are moving, we shouldn’t hold them as a rule,” Shultz says. “We’re in the business of science, so it’s a scientific debate. There should be hypotheses; there should be challenges; there should be revisions.” Medicinal chemists should still focus on what the rules call attention to—permeability and solubility. But, he says, scientists now have much better methods to predict and verify those properties.
Before the rule of 5 was even established, there were oral drugs that broke it—typically natural products. The fungal metabolite cyclosporin A, for example, has a mass of 1,202 Da and 17 hydrogen-bond acceptors. The FDA approved it as an immunosuppressant in 1983.
But chemists like Uppsala’s Kihlberg say natural products may not be the best examples of oral drugs that break the rule of 5. “I usually say that it took nature probably millions of years to come up with cyclosporin, and we don’t have millions of years,” he says.
“If you go to outside of the rule of 5, it’s an advantage if your compound can behave as a molecular chameleon” and fold to hide its hydrogen-bond donors and acceptors, Kihlberg adds. This is what cyclosporin A does.
Kihlberg also notes that some proteins will naturally bind only to larger drugs. While small-pocket binding sites on proteins work well with drugs that are small, proteins with flat or elongated, groove-shaped binding sites probably require larger molecules. He points to AbbVie’s leukemia and lymphoma treatment venetoclax as an example. The drug, which has a mass of 868 Da, blocks the B-cell lymphoma 2 protein by binding within a groove in the protein. “If they’d have been able to do that with a small, 400 or 500 Da compound, of course they would have done it,” Kihlberg says.
Merck’s Hartung points to the red-hot field of targeted protein degradation. Researchers were concerned that the clinical candidates they were making, which feature two protein-binding molecules linked together, would not be able to be dosed orally because of their size. But it turns out that most of the molecules in this class that are now in clinical trials, such as Arvinas’s bavdegalutamide, can be given as pills.
While Kihlberg now studies drugs that exist beyond the rule of 5, he says it’s important to have a rationale for making such molecules. “If you’re going to spend more time making a compound with a long synthetic route with low yields, you need to know that there is a reason,” he says. It must be because you cannot modulate your target with a smaller compound. “If you could do it with molecular weight 400, why should you do it with a molecular weight 800? More complicated is not better. Simpler is better.”
John Link, who was vice president of medicinal chemistry at Gilead Sciences and worked at the company from 2006 to 2020, says that while he understands the attraction to the rule of 5, it’s time to abandon the guidelines for good. Link worked on four drugs that break the rules: the hepatitis C virus treatments ledipasvir, velpatasvir, and voxilaprevir and the HIV drug lenacapavir. “We did not look at rules at all during the discovery of any of these drugs,” Link says. “This was not part of the conversation.”
Link was a junior medicinal chemist when the rule of 5 was introduced in 1997. Before then, he says, there were really no guidelines for medicinal chemists. Nevertheless, he’s philosophically opposed to the idea of a rule-based approach to drug discovery. “Rules provide easy answers, but there really are no easy answers in drug discovery,” Link says. “Creativity is at the core of drug discovery, and rules limit possibilities.”
Darryl McConnell, a medicinal chemist and senior vice president at Boehringer Ingelheim, says the rule of 5 has “not really allowed us to get to the bottom of the science” of creating drugs. If there is more scientific understanding of how and why certain molecules work as oral drugs, he says, “then I think we’re going to be better and faster at discovering drugs.”
Hartung and colleagues recently published a perspective about the rule of 5 called “Rules Were Made to Be Broken” (Nat. Rev. Chem. 2023, DOI: 10.1038/s41570-022-00451-0). To write it, he collected input from chemists on Twitter. “I would say there are 10 times more people out there who believe that the rule of 5 does not help drug discovery chemists than the number of people who think that this is a good thing,” he says. “Most are concerned that it limits the creativity of chemists.”
But there are those who don’t think we should disregard the rules entirely. Mary Mader, vice president of molecular innovation at Indiana Biosciences Research Institute, says most chemists think of the rule of 5 as guardrails rather than rules. Designing an orally bioavailable drug is complicated, she says.
“To try to reduce it to only a handful of properties is pretty simplistic, and yet it still is useful,” she says. “The sweet spot for any chemical series will differ by the scaffold you’re working on. And yet there is a generally true phenomenon that if you want the properties of solubility, absorption, and metabolic stability, you’re often still working in that space” defined by the rule of 5.
Even Lipinski counts himself among those that don’t think the rule of 5 should be hard and fast. “I am still slightly taken aback by how some people want strict guidelines without consideration of nuance,” says the chemist, who retired from Pfizer in 2002 and is now a consultant. “At heart, I am an experimentalist, and I believe that generally, experimentation trumps calculation. So one should try hard to measure the properties of an interesting compound even if the calculations on that compound may not look too promising.”
Hartung is also in this camp, saying he believes that the rules can be ignored if done intentionally. Echoing Picasso and the Dalai Lama, he says, “I think you can only break the rule if you have understood why there is a rule.”