Printer Friendly

Stronger protection for new drugs.

Filing patents on new chemical entities (NCEs) and series using Markush structures provides no protection from structurally unrelated bioisosteres that are active at the same target. This leaves companies exposed to 'fast-follower' compounds and reduces their market exclusivity periods. New technologies that can predict structurally diverse bioisosteres provide the basis to protect a much more specific set of chemistry that encompasses more of the structural diversity that can be active at the target of interest. This will provide stronger protection and increased returns on R&D.

The use of Markush structures to protect chemical series in patent applications has evolved significantly from its first adoption in the 1920s. (1), (2) Markush patterns in some Composition of Matter filings have exploded to the point where it is effectively impossible to verify the millions of potential structures protected. As early as 1935, the United States Patent and Trademark Office (USPTO) noted that the misuse of Markush structures was "like a fire which had spread beyond control." (3) Widespread over claiming of this type led to several initiatives to overcome the administrative challenges that they cause, not least shifting the burden of proving claimed activity between the patent examiner and the patent filer.


At the same time, there is a more fundamental problem with the use of Markush structures--they do not provide very effective protection for an activity, and have failed to prevent fast-follower compounds from reaching the market ever quicker. The average market exclusivity period for new compounds has been eroded during the last 20 years from approximately 5 years in 1987 to less than 6 months in 2007 (4). In short, the protection offered by Markush structures is simultaneously too broad to be examined properly and too narrow to include large numbers of alternate compounds that have the same biological effect.

Me-Too Compounds

There are many reasons for the increase in 'me-too' compounds, not least the changing ways that we go about discovering new drugs. The traditional iterative, evidence-led drug discovery process gave way in the 1990s to automated screening and data led processes. Screening provides molecular starting points for drug development where the potential 'hits' have already been envisioned and encapsulated in a corporate compound collection. These collections typically include a few hundreds of thousands or millions of compounds. (5) Because much of the available chemistry is very similar, there is a significant degree of overlap between companies' chemistries. As new targets become available they are quickly screened by multiple companies, resulting in a number of similar new compounds coming to market within months of each other. This in turn reduces the return from those compounds and ultimately threatens the on-going R&D budget. A key reason for this is the limited protection that a Markush structure can offer. The way in which we protect our chemical inventions has lagged behind developments in the R&D process.

Markush structures protect a virtual set of structurally related compounds, which share a common scaffold, but have different R-groups at specified positions selected from a group consisting of a closed listing of radical substituents. The basis of the use of Markush structures is that any member of the virtual set is considered equivalent and to have 'unity of invention.'

Whilst this may help protect the main lead series and even the known backups, it is only the tip of the iceberg.

All medicinal chemists know from repeated experience that many compounds share the same biological activity and properties as a lead compound even though they are structurally unrelated. Such compounds are called bioisosteres. (6) The development of structurally unrelated (and, therefore, unprotected) bioisosteres is routine in drug discovery. Active 'patent-busting', working around existing filings to preserve or even improve activity at a given target, has become a legitimate R&D activity in its own right. To protect against this kind of fast-follower strategy, tools that can routinely detect structurally diverse bioisosteres are required.

Detecting Bioisosteres

To find such bioisosteres, we have to rethink our notions of what mediates the biological activity of a compound. Medicinal chemists know from experience that this cannot be the 2D structure of a compound, because frequently we see that small changes (a methyl--ethyl switch, for example) can have significant impact on the activity of a compound, whilst at the same time, highly diverse bioisosteres with no structural similarity exist. Markush structures, which are linked to the 2D structure of a compound, are inherently limited by this simple fact.

Deeper consideration tells us that activity is mediated by the interaction between the molecular fields (surfaces) of the target protein with those of the ligand in their respective 3D binding conformations. The protein and the ligand both present a range of fields including positive and negative electrostatics, hydrophobic and steric, and the interactions between these fields determine whether a ligand is a good fit for the target's active site. (7) Unfortunately, it still takes a lot of computer power to directly compare 3D surfaces, so we have to use a highly compressed representation of the fields to compare molecules. The most important regions of the fields (the maxima, where molecular interactions are strongest) are each represented by a field point as shown in Figure 1. As the target protein's active site conformation and fields have been highly constrained by evolution, it follows that any compound capable of presenting a complementary set of field points (in a 3D conformation that is accessible under physiological conditions) is likely to have the same biological activity and properties as the natural ligand.


As fields can be computed from the structures of ligands alone, they can be used to compare the fields of potential lead structures with those of known 'hits' or natural ligands, even when the X-ray crystal structure of the target protein is unknown. Knowledge of the field patterns associated with one or more active lead compounds and/or the target protein's active site can be used to search databases of millions of compounds to identify those that are likely to be bioisosteres. These searches will identify all of the known structurally diverse forms that a bioisostere with the same activity as a lead compound might take. The results from this type of search are highly specific. They will exclude many structures that are closely related in 2D structure to the original lead if they cannot present the same pattern of fields, whilst including multiple diverse structures that can.

An example of the kind of results that field-based tools such as Cresset's FieldScreen generate is shown in the 11b-hydroxysteroid dehydrogenase type 1 (11[beta]HSD-1) example (Figure 2). A local role for 11[beta]HSD-1 in the control of visceral fat deposition is established in the literature. (8) Although originally thought to predominate as an oxidase, converting cortisol to cortisone, this enzyme has been shown to act as a reductase in vivo for cortisone. This strongly suggests that the inhibition of 11[beta] HSD-1 and the concomitant decrease of active cortisol could be important in the control of obesity, insulin resistant diabetes and cognition. (9), (10)


In the absence of specific X-ray data at the time of the project, the only information available was known ligands. The fields around these ligands were used as a seed to search a field database, which then contained 2 million commercially available compounds. The search results were ranked by field similarity to the seed structure.

A list of the best 500 hits was compiled and 408 of these compounds were purchased. When tested, 10 of these were active at <10 [micro]]M, one was active at 470 nM and the best was active at 170 nM. Three of the most active compounds are shown (Figure 3), illustrating the diversity of chemotypes that field-based tools are designed to find. It can be seen that these structures bear little resemblance either to each other or to the natural ligands of the enzyme.


At the time, the tetrazole chemotype (rightmost above) was pursued and patents based on this scaffold were filed. Other drug discovery companies subsequently submitted independent patent filings for the benzfuranones and piperazines.

Had the client chosen to pursue and protect all of these specific chemotypes, they would have been able to create a stronger patent position and licensing opportunities around the 11[beta]HSD-1 activity. To do so, they would have had to create three much more specific and examinable Markush structures, each covering one of the chemical scaffolds of interest. These filings, whilst covering a smaller number of compounds, would demonstrably protect all of the relevant chemistry and at the same time, include many fewer inactive compounds. Aside from making it easier to examine these patents this would also make it harder for fast-followers to patent-bust against this activity and consequently create new partnering opportunities for the series that the inventor chose not pursue.


As the crunch on pharmaceutical R&D budgets continues, and the regulatory and market pressures caused by the preponderance of 'me-too' compounds in the pipelines mount, we need to embrace technologies that can provide greater returns for the R&D dollar. Field-based tools that augment and enhance the protection offered by current NCE patenting practices may be a simple, cost-effective and timely way to achieve this.


(1.) USPTO Federal Register 72, 154 44992-45001 (2007).

(2.) Ex parte Markush, 1925 Dec. Comm'r Pat. 126, 128 (1924).

(3.) V.I. Richard, "Claims Under the Markush Formula," 17 J. Patent Office Soc. 179, 190 (1935).

(4.) The Pharmaceutical Market Outlook to 2010, Business Insights.

(5.) J. Harris, "Letting the Target Determine your Compound Acquisition Strategy," Drug Discov. World (2009).

(6.) T. Cheeseright, "The Identification of Bioisosteres as Drug Development Candidates," Innovat. Pharmaceut. Tech. 28, 22-26 (2009).

(7.) T. Cheeseright, et al., "Molecular Field Extrema as Descriptors of Biological Activity," J. Chem. Inf. Model 46(2), 665-676 (2006).

(8.) R.H. Stimson, et al., "Effects of Proportions of Dietary Macronutrients on Glucocorticoid Metabolism in Diet-Induced Obesity in Rats," PLoS One 5(1) e8779 (2010).

(9.) K.E. Chapman and R.J. Seckl, "11beta-HSD-1, Inflammation, Metabolic Disease and Age-Related Cognitive (Dys)Function," Neurochem. Res. 33(4) 624-636 (2008).

(10.) J.L. Yau, et al., "Enhanced Hippocampal Long-Term Potentiation and Spatial Learning in Aged 11Beta-Hydroxysteroid Dehydrogenase Type 1 Knock-Out Mice," J Neurosci. 27(39) 10487-10496 (2007).

For more information

Steve Gardner is Chief Operating Officer at Cresset-BMD Ltd and has built a number of innovative life science informatics products and companies.

Andy Vinter is Chief Scientific Officer at Cresset-BMD Ltd. He worked in the pharmaceutical industry as an organic and computational chemist and in 2001 started Cresset with a grant from the Wellcome Trust.
COPYRIGHT 2010 Via Media Ltd.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2010 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Gardner, Steve; Vinter, Andy
Geographic Code:1USA
Date:May 1, 2010
Previous Article:Addressing the challenge of poorly soluble drugs in early development: expediting formulation development and clinical screening in humans using...
Next Article:Have your business risks gone global?

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |