A McGill University-led team has shown that Vertex Pharmaceuticals Inc.'s lumacaftor and other cystic fibrosis transmembrane conductance regulator corrector compounds target only one of two sequential steps required for proper folding of the mutated protein.1 The findings suggest there is a clear rationale for developing combinations of correctors to treat patients with cystic fibrosis, three of which have been detailed at recent scientific meetings.

Cystic fibrosis is caused by mutations that reduce the function of cystic fibrosis transmembrane conductance regulator (CFTR), an anion channel that helps keep the lung and intestinal epithelium hydrated and prevents mucous buildup that leads to airway obstruction and infection.

Kalydeco ivacaftor (VX-770), a small molecule CFTR potentiator from Vertex that increases chloride transport through the channel, is the only marketed disease-modifying treatment for CF. The drug was approved in January 2012 and is indicated to treat only the 4% of patients with CF that have the G551D CFTR gating mutation, which decreases ion transport through the channel but does not impair its localization to the cell surface.

About two-thirds of patients with CF inherit a different mutation, the DF508 CFTR allele, which encodes for a misfolded version of the protein that is degraded and does not reach the surface of the cell. Because Kalydeco can only improve the function of CFTR that reaches the cell surface, corrector compounds are needed to repair the folding defect caused by the DF508 mutation.

Two correctors are in clinical development. Lumacaftor (VX-809) is in Phase III testing in combination with Kalydeco to treat patients with DF508 CFTR. VX-661, a compound in the same chemical class as lumacaftor, is in Phase II testing in the same population.

In separate Phase II trials, lumacaftor or VX-661 in combination with Kalydeco significantly increased lung function compared with placebo. The Kalydeco-lumacaftor combination has received breakthrough therapy designation from the FDA.

However, in vitro studies have suggested that combining Kalydeco plus lumacaftor can at best restore up to about 25% of CFTR function, leaving open the question of whether further correction of CFTR folding could lead to additional clinical improvements for patients with CF.2

Last year two groups, one led by McGill researchers and one from The University of Texas Southwestern Medical Center, used a combination of biophysical studies and mutational analyses to show that restoring normal function to DF508 CFTR requires correcting two distinct folding steps.3,4 The first step consists of the folding of CFTR's nucleotide binding domain 1 (NBD1), in which F508 is located, and the second step concerns the proper folding of CFTR through interactions between NBD1 and other distinct structural regions within CFTR.

The researchers involved in those studies told SciBX at the time that the next logical advance would be to use the knowledge to guide the identification of compounds that act on each step.5

Now, the McGill-led team has taken a step in that direction by providing the most complete analysis to date of the mechanisms of action for available corrector compounds.

The team, led by McGill professor of physiology Gergely Lukacs, took advantage of previously identified suppressor mutations that restore DF508 CFTR function by correcting either of the two steps. Combining both sets of suppressor mutations leads to a synergistic and dramatic increase in DF508 CFTR function, bringing it to wild-type CFTR levels.

Thus, by testing corrector compounds on DF508 CFTR variants with suppressor mutations that correct one of the two steps, the researchers could deduce which folding step the compounds act upon.

In cell lines expressing DF508 CFTR variants carrying suppressor mutations that improved NBD1 interaction with other regions of CFTR, lumacaftor and other corrector compounds increased the amount of DF508 CFTR at the cell surface, but the total amount was still only about one-third of wild-type CFTR levels.

In cell lines expressing DF508 CFTR variants carrying suppressor mutations that improved the folding of NBD1, the compounds restored DF508 CFTR cell surface localization to levels comparable to those in wild-type cell lines. This suggested existing corrector compounds do not act to improve NBD1 folding but instead stabilize the interaction between NBD1 and other regions of CFTR.

To probe the mechanism of action for these corrector compounds in more detail, the authors performed in vitro experiments in which DF508 CFTR variants were reconstituted in an artificial lipid bilayer. In this assay, lumacaftor and a related compound directly acted on CFTR, and a combination of computational modeling, mutational analysis and biophysical experiments pinpointed the interface between NBD1, membrane spanning domain 1 (MSD1) and MSD2 as their likely site of action.

According to Lukacs, this is the most detailed public description of lumacaftor's mechanism of action to date and complements previously published work by Vertex. "Without photocrosslinkable variants in hand, we don't have absolute evidence of direct drug binding and its location. However, the artificial lipid bilayer study is the closest evidence we have that the compound is directly interacting," he said.

He added that this is in line with data published earlier this year by a group at The University of North Carolina at Chapel Hill School of Medicine that suggested the compound was acting to stabilize the interaction of NBD1 with other regions of CFTR.6

The McGill team went on to characterize additional correctors that are structurally unrelated to lumacaftor and found that some were in fact mechanistically distinct from the compound and likely stabilized interactions between the NBD1 and NBD2 regions of CFTR. No correctors repaired the NBD1 folding defect.

"It was disappointing that we did not find any compounds that corrected the folding defect in NBD1," said Lukacs.

Finally, the team tested how clinically relevant combinations of correctors could be developed by combining them with chemical chaperones that stabilize NBD1 folding, such as glycerol and myo-inositol. These chemical chaperones are not drug-like but served as tools to demonstrate proof of concept for functional DF508 correction.

In cultured human bronchial cell lines and in a recently developed organoid model system,7 lumacaftor plus the chemical chaperones led to substantial increases in DF508 CFTR function compared with either lumacaftor or chaperones alone. The effect was enhanced when a third compound was added that stabilized NBD1-NBD2 interactions, supporting the idea that multiple compounds that act on distinct folding steps could be combined to restore DF508 function.

Results were published in Nature Chemical Biology.

Cocktail mixing

Philip Thomas, professor of physiology at UT Southwestern Medical Center and a cofounder of Reata Pharmaceuticals Inc., told SciBX that the study was the natural follow-on to the two-step folding hypothesis described by his lab and the Lukacs lab last year.

"This work underlines what was in the two earlier papers in Cell, but what is important is that it shows there are no compounds in the set people have been looking at that are effective at correcting the first step, namely NBD1 folding," he said.

Fred Van Goor, head of biology for Vertex's CF research program, agreed with Thomas. "This is a follow-on to the earlier work that set the foundation by showing that just deleting F508 causes multiple structural defects. It provides a mechanistic rationale for why two CFTR correctors could be additive for each other."

Only one corrector of the NBD1 folding defect has been reported so far. According to Thomas, a screen carried out by Reata, with screening technology licensed from the UT Southwestern Medical Center and developed in Thomas' lab, identified a compound that corrected NBD1 folding and synergized with an analog of lumacaftor. Data were presented in March at the European Cystic Fibrosis Society Basic Science Conference by Andre Schmidt, a member of Thomas' lab. Reata did not respond to interview requests.

At last week's European Cystic Fibrosis Society Conference, Vertex presented data on a corrector that synergizes with Kalydeco and lumacaftor to improve chloride transport in DF508 human bronchial epithelial cells. Vertex did not provide details of the development status or mechanism of action for this compound but said the company has an active research program to identify second-generation correctors for use in future combination regiments. Vertex also said it hopes to have a second-generation corrector in clinical development by the end of 2014.

Proteostasis Therapeutics Inc. has also disclosed data on small molecule proteostasis modulators that show that the compounds can synergize with lumacaftor to improve chloride transport in human bronchial epithelial cells. The data were presented at the 26th Annual North American Cystic Fibrosis Conference last October and the EMBO meeting on May 21. The compounds are currently in lead optimization.

Proteostasis' approach is distinct from that of the other companies in that it does not specifically target CFTR but rather goes after cellular protein trafficking mechanisms.

"Many of the current corrector approaches seek compounds directly interacting with CFTR itself. I think of them as molecular staples that in some way enhance folding and/or correct a folding deficit that allows CFTR to pass some of its quality control checkpoints," said Peter Reinhart, president and CSO of Proteostasis.

According to Reinhart, his company has "a fundamentally different approach, which is to identify modulators of the cells' endogenous quality control machinery, which will ultimately handle CFTR by enhancing its folding and trafficking."

Last May, the Cystic Fibrosis Foundation announced it would collaborate with Proteostasis to develop therapies to treat patients with DF508 CFTR. The foundation also is collaborating with Vertex, Pfizer Inc. and Sanofi's Genzyme Corp. unit. CFF, Pfizer and Sanofi did not respond to interview requests.

David Thomas, a professor in the Department of Biochemistry at McGill, told SciBX that the McGill work provides a template for characterizing compounds as they emerge from screening efforts. "You can think of it as a funnel. People have found lots of correctors, and now they have a nice way to identify which step they are involved in."

Thomas was not involved in the studies by Lukacs.

In 2011, Thomas and McGill University colleagues began collaborating with GlaxoSmithKline plc to characterize the functions of correctors identified in high throughput screens, and Thomas and John Hanrahan, professor of physiology at McGill, are developing some of the correctors from the screens. In addition, the pair have recently founded Traffic Therapeutics Inc. to develop CFTR correctors.

Earlier this year, GSK also began collaborating with a team at The Hospital for Sick Children and the University of Toronto to discover CFTR correctors.8

Lukacs said his lab now plans to identify compounds that act to correct the NBD1 folding defect in DF508 CFTR using structural defect-targeted, new high throughput screening assays based on monitoring the channel's biochemical appearance at the cell surface.

He added that ultimately, additional structural information about CFTR will be needed to move the field beyond phenotypic screening and toward rational drug design.

"The interesting question is how we can rationally design correctors for NBD1 stabilization," he said. "What has been done is a random search, and to make the process more successful it should be more targeted than a random chemical library screen. The question is how best to do that."

Results from the study are not patented.

Cain, C. SciBX 6(24); doi:10.1038/scibx.2013.589
Published online June 20, 2013

REFERENCES

1.   Okiyoneda, T. et al. Nat. Chem. Biol.; published online May 12, 2013; doi:10.1038/nchembio.1253
Contact: Gergely L. Lukacs, McGill University, Montreal, Quebec, Canada
e-mail: gergely.lukacs@mcgill.ca

2.   Van Goor, F. et al. Proc. Natl. Acad. Sci. USA 108, 18843-18848 (2011)

3.   Rabeh, W.M. et al. Cell 148, 150-163 (2012)

4.   Mendoza, J.L. et al. Cell 148, 164-174 (2012)

5.   Cain, C. SciBX 5(8); doi:10.1038/scibx.2012.192

6.   He, L. et al. FASEB J. 27, 536-545 (2013)

7.   Dekkers, J.F. et al. Nat. Med.; published online June 2, 2013; doi:10.1038/nm.3201

8.   Osherovich, L. SciBX 6(10); doi:10.1038/scibx.2013.230

COMPANIES AND INSTITUTIONS MENTIONED

Cystic Fibrosis Foundation, Bethesda, Md.

EMBO, Heidelberg, Germany

European Cystic Fibrosis Society, Karup, Denmark

Genzyme Corp., Cambridge, Mass

GlaxoSmithKline plc (LSE:GSK; NYSE:GSK), London, U.K.

The Hospital for Sick Children, Toronto, Ontario, Canada

McGill University, Montreal, Quebec, Canada

Pfizer Inc. (NYSE:PFE), New York, N.Y.

Proteostasis Therapeutics Inc., Cambridge, Mass.

Reata Pharmaceuticals Inc., Irving, Texas

Sanofi (Euronext:SAN; NYSE:SNY), Paris, France

Traffic Therapeutics Inc., Montreal, Quebec, Canada

The University of North Carolina at Chapel Hill School of Medicine, Chapel Hill, N.C.

The University of Texas Southwestern Medical Center, Dallas, Texas

University of Toronto, Toronto, Ontario, Canada

Vertex Pharmaceuticals Inc. (NASDAQ:VRTX), Cambridge, Mass.