Irreversible inhibition of a protease central to hepatitis C infection; New HCV Protease AVL-192: Study (full text; published pdf attached)
Download the PDF here
Mon, Nov 29, 2010
A new study has demonstrated that irreversible covalent inhibition can increase selectivity, potency and duration of action, broadens applications for targeted covalent drugs to the protease gene family.
Avila Therapeutics Inc., a biotechnology company developing novel targeted covalent drugs, has demonstrated the first-ever selective irreversible inhibition of a viral protease using a targeted covalent drug.
Avila has used its proprietary Avilomics platform to design covalent irreversible protease inhibitors that are highly selective, potent and with superior duration of action as compared to conventional protease inhibitors.
The research has demonstrated that covalent drugs can be designed and targeted to irreversibly and covalently bond to molecular domains specific to proteases.
"This research elevates covalent drug design to a fundamentally new level. By creating extremely selective protease inhibitors with their platform, Avila is showing the remarkable therapeutic potential of irreversible covalent drugs to address a broad spectrum of drug targets," said Simon Campbell , a renowned scientist.
"This approach can make a difference to patients living with HCV infection, and we expect to make an impact in other important areas such as cancer and inflammatory disease," said said Juswinder Singh, co-author of the paper.
In order to maximize selectivity and minimize off-target effects, the irreversible covalent inhibitors of HCV protease were designed to covalently target a unique structure in the HCV protease not found in human proteases. Key findings include:
A representative irreversible covalent inhibitor designed, by Avila, was shown to inhibit the HCV protease (also known as "NS3") in cells at a concentration of 6 nM.
Specific covalent bond formation between the drug and target protease was demonstrated through use of mass spectrometry and also x-ray crystallography.
The findings were published in the journal Nature Chemical Biology. (ANI)
Avila Presents New Data on its Novel, Orally-Available Targeted Covalent Drug, AVL-192
Covalent Inhibition Achieves Superior Potency Against Drug-Resistant HCV Mutants
BOSTON and WALTHAM, MA - November 2, 2010 - Avila TherapeuticsTM, Inc., a biotechnology company developing novel targeted covalent drugs, presented results today of preclinical studies that demonstrate its orally-available targeted covalent drug candidate, AVL-192, achieves superior potency against drug-resistant mutations of the Hepatitis C Virus (HCV). These new data were presented today at the Annual Meeting of the American Association for the Study of Liver Diseases (AASLD) international meeting in Boston, Massachusetts.
HCV protease (also known as NS3) is a promising target of intervention for the treatment of hepatitis C infection. However, medicines currently in late stages of clinical development are vulnerable to drug- resistant mutations. AVL-192 is a novel, orally available compound that can rapidly and completely silence the HCV protease through highly selective, irreversible covalent bonding to the target protein. Preclinical data have demonstrated that AVL-192 achieves very high potency and selectivity for NS3 and also potently and effectively inhibits the drug-resistant mutations observed clinically.
Avila's covalent approach to silencing the NS3 protein has resulted in a product candidate with a potential best-in-class profile due to the ability to retain potency against clinically-arising resistance mutations, and potential breadth of activity across HCV genotypes with anticipated once-per-day dosing.
In a poster presentation at the meeting, entitled, "Second Generation of Covalent Irreversible Inhibitors Have Superior Potency Across Genotypes and Drug Resistant Mutants," data were presented from preclinical studies that evaluated the efficacy of AVL-192 in biochemical and cell culture studies. Highlights of the data demonstrate:
· AVL-192 has a time-dependent mode of action that delivers potent and rapid inhibition of WT NS3/4A and retains high potency against drug-resistant mutant NS3/4A proteases;
· AVL-192 is able to inhibit the protease long after the compound is removed, offering the benefit of less frequent dosing;
· AVL-192 as monotherapy can be curative in the replicon clearance assay;
· AVL-192 is highly selective and spares host proteases; and
· AVL-192 has high plasma exposure following oral administration in rats and dogs.
"These new data reinforce our belief that our targeted covalent drug candidate AVL-192 has the potential to be a best-in-class, pan-genotype HCV therapeutic due to its unique mechanism of action," said Juswinder Singh, Ph.D., Avila's Founder and Chief Scientific Officer.
About Avila TherapeuticsTM, Inc.
Avila focuses on design and development of targeted covalent drugs to achieve best-in-class outcomes that cannot be achieved through traditional chemistries. This approach is called "protein silencing". The company's product pipeline has been built using its proprietary AvilomicsTM platform and is currently focused on viral infection, cancer and autoimmune disease. Avila is funded by leading venture capital firms: Abingworth, Advent Venture Partners, Atlas Venture, Novartis Option Fund, and Polaris Venture Partners. For additional information, please visit http://www.avilatx.com
Selective irreversible inhibition of a protease by targeting a noncatalytic cysteine
Nature Chemical Biology, 28 November 2010
Margit Hagel, Deqiang Niu, Thia St Martin, Michael P Sheets, Lixin Qiao, Hugues Bernard, Russell M Karp, Zhendong Zhu, Matthew T Labenski, Prasoon Chaturvedi, Mariana Nacht, William F Westlin, Russell C Petter & Juswinder Singh
1Avila Therapeutics, Inc., Waltham, Massachusetts, USA. 2Millenium: The Takeda Oncology Company, Cambridge, Massachusetts, USA. *e-mail: email@example.com
Designing selective inhibitors of proteases has proven problematic, in part because pharmacophores that confer potency exploit the conserved catalytic apparatus. We developed a fundamentally different approach by designing irreversible inhibitors that target noncatalytic cysteines that are structurally unique to a target in a protein family. We have successfully applied this approach to the important therapeutic target HCV protease, which has broad implications for the design of other selective protease inhibitors.
The fundamental challenge in designing protease inhibitors is to achieve potency without sacrificing selectivity. This problem arises frequently because the typical protease inhibitor achieves potency through covalent interactions with the catalytic apparatus, yet such pharmacophores also confer affinity for other proteases in the same mechanistic family1. This is a significant challenge for protease drug design, because over 500 proteases exist in the human genome. Achieving selectivity while targeting the catalytic machinery is thus particularly difficult2.
Covalent irreversible drugs that form persistent, nonlabile covalent bonds yield unique therapeutic benefits including rapid onset of inhibition, greater potency, longer duration of drug action and potent and persistent activity against mutations that would otherwise lead to drug resistance3. There are many examples of drugs that work through irreversible covalent bonding that have proven to be safe and successful therapies for a wide variety of indications4. Despite their prevalence, to date covalent drugs have largely been discovered serendipitously, and general methods to facilitate their deliberate discovery and design have yet to be described.
HCV NS3/4A viral protease (HCVP) activity is essential for viral replication5 and has been recently validated as a clinical target6, 7, 8, 9, 10, 11, 12, 13, 14, 15. Protease inhibitors such as telaprevir exemplify the challenges of covalent targeting of the catalytic binding site; upon binding to HCVP, the α-ketoamide forms a reversible covalent linkage with the catalytic serine that is conserved within proteases6, 10. Indeed, telaprevir inhibits some host serine proteases at concentrations that may be achieved in a therapeutic setting14, 16.
The aim of this study was to achieve potent inhibition of viral proteases through covalent bond formation without compromising selectivity of the inhibitors. Our new design strategy used structural bioinformatics to create a structural alignment between viral proteases and host proteases to identify nucleophilic amino acids in the binding site that were unique to the viral proteases. Our structural alignment revealed that current covalent inhibitors such as telaprevir target a catalytic residue that is common across the protease family and therefore susceptible to selectivity issues. In contrast, we identified a nucleophilic amino acid, cysteine 159 (Cys159), in the substrate-binding site that could be targeted for covalent bonding. Importantly, Cys159 is strictly conserved across all 919 HCV NS3 sequences in a database of all known HCVP sequences, including all HCVP subtypes and genotypes sequenced to date (Supplementary Fig. 1), allowing design of a pan-genotype HCVP inhibitor. Although HCVP shows structural similarity with host proteases, Cys159 is structurally unique to HCVP and therefore was an ideal target for achieving selectivity between HCVP and host proteases.
Structure-based drug design was used to create a peptidomimetic inhibitor (1) (Fig. 1a), designed to form canonical reversible interactions with the S2-S1-S1' pockets of HCVP similar to those observed with other reversible peptidomimetic inhibitors12, 17, 18. Further molecular modeling was used to evaluate structures that positioned a low-reactivity Michael acceptor close enough to Cys159 to form a covalent bond (2) (Fig. 1a). 2 was prepared by installation of an acrylamide using a simple glycine linker. Linking the electrophilic acrylamide via a D-alanine linker provides the more conformationally constrained inhibitor 3 (see Supplementary Methods). Our expectation was that propanamide 4, the reversible congener of 3, would bind weakly to the HCVP, as potent inhibitors reported to date typically possess functionality that forms extensive nonbonding interactions with the S3 and S4 pockets.
As predicted, 1 and 4 showed weak inhibition of the wild-type HCVP (half-maximal inhibitory concentration (IC50) of 1 = 2,458 nM, 4 IC50 = 1,147 nM) whereas 2 and 3 were very potent inhibitors (2 IC50 = 4 nM, 3 IC50 = 2 nM) (Supplementary Table 1). To further support the importance of covalency in conferring activity of 3, we tested the activity of 3 against a mutant NS3 protein in which the target cysteine is changed to a serine (C159S). The C159S protease is comparable in enzymatic activity to wild-type protease (Supplementary Fig. 2); however, mutation of the amino acid required for bond formation results in a sharp decrease in potency of the covalent inhibitor (IC50 = 1,782 nM; Supplementary Table 1). In further support of the mechanism, 3 shifted the mass of HCVP by 685 Da, consistent with the formation of a covalent complex between 3 and the protease, but was unable to covalently bond to HCVP with the C159S mutation (Supplementary Fig. 3). We also confirmed with X-ray crystallography that 3 was covalently linked to the side chain of Cys159 (Fig. 1b, Supplementary Fig. 4 and Supplementary Table 2).
The selectivity of 3 was further demonstrated using a panel of host proteases. As expected, 3 showed no notable inhibition of host proteases, whereas telaprevir inhibited multiple host proteases, when each inhibitor was tested at 10 µM (Fig. 2a). Moreover, 3 showed no significant nonspecific reactivity toward glutathione (Supplementary Fig. 5). These data highlight the value of covalent bonding to a noncatalytic residue as a means of achieving HCVP selectivity while minimizing the potential for nonspecific reactivity with other thiols such as glutathione.
Huh-7 wild-type (1b) replicon cells were used to demonstrate that 3 can potently inhibit HCVP activity in cells, leading to decreased replication of viral RNA. Luciferase activity was greatly reduced in cells treated with 3 (half-maximal effective concentration (EC50) = 6 nM) (Fig. 2b and Supplementary Table 1). In contrast, the reversible congener, 4, did not inhibit luciferase activity (EC50 > 3000 nM), demonstrating that covalent bonding greatly enhances potency for this class of compounds. Importantly, 3 does not inhibit proliferation of Huh-7 wild-type replicon cells, nor does it affect growth of other cell lines tested (Supplementary Table 3), strongly suggesting that replicon inhibition is because of specific viral protease inhibition. 3 was inactive in the C159S mutant replicon cells (EC50 > 3,000 nM) and, as expected, the activity of 4 and telaprevir were unchanged by the C159S mutation, as they are not dependent on the cysteine for their mechanism of action (EC50 > 3,000 nM, EC50 = 623nM, respectively) (Fig. 2b and Supplementary Table 1). Of note, the C159S mutant replicon cells showed fitness similar to that of wild-type replicon cells (Supplementary Fig. 6).
Numerous NS3 mutations have been reported that render HCVP resistant to the current protease inhibitors. Thus, activity against drug-resistant clinical mutations is important for effective antiviral therapeutics19. 3 was able to inhibit and bond to HCVP proteins of clinically relevant NS3 variants (Supplementary Table 1 and Supplementary Fig. 7). Furthermore, the selectivity conferred by Cys159 also allows for binding and inhibition of HCVP from multiple genotypes (Supplementary Table 1 and Supplementary Fig. 8), suggesting that this approach will lead to potent and selective pan-genotype HCVP inhibitors.
To demonstrate direct inhibition of HCVP activity using our irreversible covalent drug, we developed an assay using the internal self-cleavage activity of HCVP20 (Supplementary Fig. 9). We confirmed the necessity of HCVP activity in the proteolytic cleavage of NS3/4A by expressing wild-type HCVP or a protease-dead mutant, NS3/4A-S139A (Supplementary Fig. 10). This autoproteolytic cleavage activity was used to directly measure HCVP activity in replicon cells in the presence and absence of the covalent inhibitor. When HCVP activity is inhibited, self-cleavage is abolished, leaving only the full-length holoenzyme. 3 demonstrated inhibition of HCVP internal self-cleavage activity, and the inhibition was sustained for 8-24 h after compound removal. In contrast, HCVP self-cleavage activity had completely returned by 30 min after removal of telaprevir (Supplementary Fig. 11a,b).
A unique advantage of an irreversible covalent molecule is that it allows the investigation of target occupancy in a time- and dose-dependent manner. We designed a biotinylated irreversible covalent probe that bonds to NS3/4A protease (Supplementary Scheme 4), enabling the quantitative analysis of NS3 occupancy with 3, and found that inhibitory activity and NS3-occupancy closely correlate (Fig. 3a,b). Following treatment with 3, there is little or no free NS3 available to bind to the biotinylated probe for at least 8 h after 3 has been removed (Fig. 3b). This indicates that essentially all of the NS3 protein was bound by 3, and newly synthesized protein is being detected at 8-24 h. Return of self-cleavage activity is concomitant with the detection of newly synthesized protease. The biotinylated covalent probe compound is also an indicator of the selectivity of 3, as the two compounds share structural similarities and electrophiles. Only the full-length NS3 protein and NS3/4A cleavage products were detected as having been labeled by the biotinylated probe, indicating that it is specific for NS3/4A under these conditions (Fig. 3c).
A targeted covalent design approach21 has been applied to kinases, several of which are currently in clinical testing with encouraging evidence of efficacy and safety22. This study describes the first successful example of applying targeted covalent inhibition to the protease family. Our data indicate that the electrophile on 3 must be brought into close proximity to a nucleophilic thiol via specific affinity-driven binding to enable covalent bond formation between the small molecule and the targeted HCVP . This strategy enables us to selectively inhibit HCVP and minimize the potential for toxicity through reactivity with off-target proteins. 3 is an excellent prototype HCVP inhibitor but has a number of important limitations as a drug candidate; these properties have been optimized in our current development compounds, AVL-181 and AVL-192, which have excellent pharmacokinetics and bind potently to wild-type HCVP as well as to multiple other genotypes and mutant forms of HCVP, including C159S, but only covalently modify when Cys159 is present23. The successful design of a highly selective targeted covalent inhibitor of HCVP suggests that this approach can be broadly applied to other protease family members and indeed to a wide range of protein families.