From CFTR biology toward combinatorial pharmacotherapy: expanded classification of cystic fibrosis mutations.

More than 2000 mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) have been described that confer a range of molecular cell biological and functional phenotypes. Most of these mutations lead to compromised anion conductance at the apical plasma membrane of secretory epithelia and cause cystic fibrosis (CF) with variable disease severity. Based on the molecular phenotypic complexity of CFTR mutants and their susceptibility to pharmacotherapy, it has been recognized that mutations may impose combinatorial defects in CFTR channel biology. This notion led to the conclusion that the combination of pharmacotherapies addressing single defects (e.g., transcription, translation, folding, and/or gating) may show improved clinical benefit over available low-efficacy monotherapies. Indeed, recent phase 3 clinical trials combining ivacaftor (a gating potentiator) and lumacaftor (a folding corrector) have proven efficacious in CF patients harboring the most common mutation (deletion of residue F508, ΔF508, or Phe508del). This drug combination was recently approved by the U.S. Food and Drug Administration for patients homozygous for ΔF508. Emerging studies of the structural, cell biological, and functional defects caused by rare mutations provide a new framework that reveals a mixture of deficiencies in different CFTR alleles. Establishment of a set of combinatorial categories of the previously defined basic defects in CF alleles will aid the design of even more efficacious therapeutic interventions for CF patients.

INTRODUCTION

Cystic fibrosis (CF), caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR), is characterized by a multiorgan pathology affecting the upper and lower airway, gastrointestinal and reproductive tracts, and endocrine system (Riordan ; Collins, 1992; Rowe ; Cutting, 2015). CF is one of the most common lethal autosomal-recessive diseases, with a prevalence of one in 3500 in the United States and one in 2500 in the European Union (Farrell, 2008; Pettit and Fellner, 2014). Lack of functional CFTR expression at the apical membrane of secretory epithelia results in defective Cl− and bicarbonate secretion, coupled to enhanced Na+ absorption and mucus secretion, which in airway epithelia leads to dehydration and acidification of the airway surface liquid (Tarran ; Chen ; Derichs ; Pezzulo ). As a consequence, impaired mucociliary clearance provokes recurrent infection and uncontrolled inflammation culminating in lung damage, which is the primary cause of morbidity and mortality in CF (Ratjen and Doring, 2003; Boucher, 2007; Stoltz ). CFTR is member of the ATP-binding cassette (ABC) subfamily C (ABCC7) (Kerr, 2002). It consists of two homologous halves, each containing a hexa-helical membrane-spanning domain (MSD1 and MSD2) and a nucleotide-binding domain (NBD1 and NBD2) that are connected by an unstructured regulatory domain (Riordan, 1993; Riordan ).

BIOLOGY OF CFTR MUTATION: TRADITIONAL CLASSIFICATION

CF is caused by ∼2000 mutations in the CFTR gene with a wide range of disease severity (www.genet.sickkids.on.ca/home.html; www.cftr2.org; Sosnay ), which is further influenced by modifier genes (Collaco and Cutting, 2008; Cutting, 2010) and by the environmental and socioeconomic status of patients (Schechter ; Barr ; Taylor-Robinson ; Kopp ). The first classification of CF mutations into four classes according to their primary biological defect was proposed by Welsh and Smith in a landmark paper (Welsh and Smith, 1993). Currently, six major classes are distinguished (Rowe ; Zielenski and Tsui, 1995) (Figure 1).

FIGURE 1:

Traditional classification of CF mutations based on their cellular phenotype. Class I: protein synthesis defect; class II: maturation defect; class III: gating defect; class IV: conductance defect; class V: reduced quantity; and class VI: reduced stability. ER, endoplasmic reticulum; TGN, trans-Golgi network.

Class I encompasses frameshift, splicing, or nonsense mutations that introduce premature termination codons (PTC), resulting in severely reduced or absent CFTR expression.

Class II mutations lead to misfolding, premature degradation by the endoplasmic reticulum (ER) quality-control system, and impaired protein biogenesis, severely reducing the number of CFTR molecules that reach the cell surface.

Class III mutations impair the regulation of the CFTR channel, resulting in abnormal gating characterized by a reduced open probability.

Class IV mutations alter the channel conductance by impeding the ion conduction pore, leading to a reduced unitary conductance (Sheppard ; Hammerle ).

Class V mutations do not change the conformation of the protein but alter its abundance by introducing promoter or splicing abnormalities (Highsmith , 1997; Zielenski and Tsui, 1995).

Class VI mutations destabilize the channel in post-ER compartments and/or at the plasma membrane (PM), by reducing its conformational stability (Haardt ) and/or generating additional internalization signals (Silvis ). This results in accelerated PM turnover and reduced apical PM expression (Haardt ; Silvis ).

For many of the identified mutations, the disease liability is unknown, but efforts are under way to assess their functional consequence and clinical severity (www.cftr2.org; Sosnay ).

MUTATION CLASS–SPECIFIC PHARMACOTHERAPY

Defining the cellular and molecular pathology of CFTR mutations proved to be invaluable for development of small-molecule compounds targeting the underlying defect(s) in CF. The fact that some CFTR variants carrying class III or IV mutations can be expressed at the apical membrane of secretory epithelia at a density similar to that of the wild-type protein, although they are functionally impaired (e.g., G551D), led to the development of gating potentiators that increase the open probability and thereby the PM chloride conductance (Yang ). VX-770 (ivacaftor) is the first potentiator drug to be U.S. Food and Drug Administration approved for CF treatment; it directly targets the gating defect of the class III mutation G551D-CFTR (Van Goor ). This compound was developed by Vertex Pharmaceuticals in conjunction with Cystic Fibrosis Foundation Therapeutics, Inc. (CFFT), and shows remarkable clinical benefit in patients carrying the mutation in either one or two alleles (Van Goor ; Accurso ; Ramsey ). The approval of VX-770 was extended to eight additional class III mutations (G178R, S549N, S549R, G551S, G1244E, S1251N, S1255P, and G1349D) (Yu ; Vertex, 2014a) and recently to the class IV mutation R117H (Vertex, 2014b).

The prototypical class II mutation, ΔF508-CFTR (Phe508del), elicits a complex folding defect that compromises both NBD1 stability and the channel’s cooperative domain assembly (Du and Lukacs, 2009; Du ; Mendoza ; Rabeh ). For many years, large-scale efforts have been under way to isolate correctors that act as pharmacological chaperones by directly binding to and promoting the biogenesis of class II CFTR mutations. The most promising corrector compound at present, VX-809 (lumacaftor), partially reverts the ΔF508-CFTR functional expression defect by stabilizing the NBD1-MSD1/2 interface (Farinha ; Loo ; Okiyoneda ; Ren ), leading to a marked correction from 3 to 15% of wild-type channel activity in vitro (Van Goor ). A clinical trial, however, failed to observe significant clinical benefit in homozygous ΔF508-CFTR patients (Clancy ). Acute addition of VX-770 to VX-809-corrected ΔF508-CFTR doubled the PM activity in vitro (Van Goor ), and the combination therapy showed modest but significant clinical improvement (Boyle ; Wainwright ). Based on these results, the combination treatment has been approved for CF patients 12 years and older with two copies of the ΔF508 mutation (Vertex, 2015). Other class II mutations that can be corrected by VX-809 in vitro include E56K, P67L, E92K, R170G, L206W, V232D, F508G, and A561E (Caldwell ; Okiyoneda ; Ren ; Veit ; Awatade ).

Ribosomal read-through allows synthesis of full-length CFTR carrying class I mutations. To this end, ataluren (PTC124) was developed as a drug that promotes near-cognate aminoacyl-tRNA incorporation at PTCs (Lentini ; Welch ). Ataluren partially restores G542X-CFTR (class I) expression in a mouse model and modestly corrects CFTR function in nasal epithelia in patients with class I mutations (Du ; Sermet-Gaudelus ; Wilschanski ). In a recent phase 3 clinical trial, however, ataluren treatment failed to produce significant clinical benefit, perhaps due to an adverse drug–drug interaction with tobramycin, which is a commonly administered, inhaled antibiotic used to treat lung infections in CF patients (Kerem ).

LIMITATIONS OF CF MUTATION CLASSIFICATION

The efficacy of available monotherapies for some mutant alleles, which have been designated as class I, class II, or class III/IV mutations, is currently limited. This could be partly explained by the pleiotropic molecular defects caused by single mutations. Thus comprehensive mapping of the multiple molecular defects caused by a single or combination of mutant alleles could offer considerable advantage for improving therapeutic interventions and for future development of drug combinations. In the following list, we present a subset of mutations that display combinatorial molecular defects.

ΔF508: The most prevalent class II mutation impairs CFTR conformational maturation and leads to its targeting for premature ER-associated degradation (Cheng ; Cyr, 2005; Kim and Skach, 2012; Lukacs ). However, ΔF508-CFTR molecules that either constitutively or following rescue procedures escape the ER quality control and accumulate at the PM of airway epithelia exhibit a channel-gating defect, which is a hallmark of class III mutations (Dalemans ), as well as accelerated turnover in post ER compartments and at the PM, a class VI mutation characteristic (Lukacs ). Unless the folding and conformational dynamics of the rescued ΔF508-CFTR are fully restored to that of the wild-type protein by pharmacological treatment, this mutation remains partially defective and requires correction of its gating and/or peripheral stability defect. Rescue of the gating defect can be achieved with potentiators (e.g., VX-770) (Van Goor ). Peripheral stabilization of the ΔF508-CFTR could be attained by 1) the peptide inhibitor iCAL36 (Cushing ), 2) preventing post-Golgi ubiquitination (Fu ; Okiyoneda ), 3) restoring autophagosome formation (Luciani ), or 4) modulating cellular protein homeostasis (Hutt ). Thus the most common mutant has multiple defects that extend beyond the features of a class II mutation.

W1282X: This PTC represents a class I mutation, though recent studies suggest a more complex phenotype. First, the level of the W1282X transcript is reduced by nonsense-mediated RNA decay (Hamosh ; Linde ). Second, the PTC deletes part of the NBD2, which likely compromises NBD1-NBD2 dimerization and W1282X-CFTR folding and activity. Moreover, if the primary defect is corrected either with spontaneous or drug-induced read-through, some of the fully translated channel will contain nonconservative amino acid substitutions. These missense mutations may cause structural defects (class II characteristic), as suggested by the phenotype of CF patients with a missense mutation at the W1282 residue (Faucz ; Ivaschenko ; Visca ), as well as a gating defect (class III characteristic), which can be inferred based on W1282X-CFTR channel activation after exposure to VX-770 (Xue ).

P67L: P67L is a mild class II mutation that results in attenuated CFTR biogenesis, as indicated by the reduced ratio between post-ER complex–glycosylated (band C) and ER-resident core-glycosylated protein (band B) (Ren ; Sosnay ; Van Goor ). Treatment with the corrector VX-809 increases the abundance of the complex-glycosylated form and PM density to nearly the level of WT-CFTR (Ren ; Veit ). However, the mutant channel is also sensitive in vitro to potentiator treatment (a class III characteristic), both in the presence and absence of corrector (Van Goor ; Veit ). Accordingly, treatment with VX-770 ameliorated the CF lung disease in a heterozygous P67L/ΔF508 patient (Yousef ).

R117H: This mutation in conjunction with the 5T variant in the polythymidine tract in intron 8 was originally categorized as a class IV mutation, but it also exhibits a gating defect (class III trait) that, at least in part, can be rectified by VX-770 treatment (Sheppard ; Van Goor ). The R117H mutation also results in reduced complex-glycosylated CFTR expression, which is a class II characteristic (Fanen ; Sheppard ). This potentially explains the limited success of VX-770 treatment in patients carrying this mutation (Char ; Moss ).

AN EXPANDED CLASSIFICATION OF MUTANT CFTR BIOLOGY

We propose a modification of the current classification scheme, which would entail permutations of the traditional class I–VI CF mutations. This expanded classification of the major mechanistic categories (Welsh and Smith, 1993; Zielenski, 2000; Rowe ) accommodates the unusually complex, combinatorial molecular/cellular phenotypes of CF alleles. It consists of 31 possible classes of mutations, including the original classes I, II, III/IV, V, and VI, as well as their 26 combinations, as depicted in the Venn diagram shown in Figure 2. For the sake of simplicity, class III and IV mutations, representing functional (gating and conductance, respectively) defects, are combined. For example, according to the expanded classification, G551D will be designated as a class III mutation as before (Welsh and Smith, 1993), while ΔF508 will be classified as class II–III–VI, W1282X as class I–II–III–VI, P67L as class II–III, and R117H as class II–III/IV, reflecting the composite defects in mutant CFTR biology (Figure 2 and Table 1).

FIGURE 2:

Refined classification of CF mutations accounting for complex phenotypes of major CFTR cellular defects. The Venn diagram indicates all combinations of mutation classes with selected examples. Possible combinations without identified mutation are indicated in gray.

TABLE 1:

Examples for CF mutations with complex or classical cellular phenotypes.

Refined classification	Mutation	I	II	III/IV	V	VI	Model	Reference
I–II–III–VI	W1282X	X1,2,3	X2,5	X2,4,5		X5	1HNE	1Hamosh et al., 1992
							2HBE	2Cyr lab, unpublisheda
							3CFBE	3Frizzell lab, unpublishedb
							4CFBE	4Xue et al., 2014
							5CFBE	5Lukacs lab, unpublishedc
II–III	M1V		X6	X6			6FRT	6Van Goor et al., 2014
II–III	E56K		X5,6	X6			5CFBEd	5Lukacs lab, unpublished
							6FRT	6Van Goor et al., 2014
II–III	P67L		X3,6,7,8,9,10	X6,7,10			3CFBE	3Frizzell lab, unpublished
							6FRT	6Van Goor et al., 2014
							7CFBE	7Veit et al., 2014
							8Hek293	8Ren et al., 2013
							9HeLa	9Sosnay et al., 2013
							10FRT	10Sorscher lab, unpublishede
II–III	R74W		X6	X6			6FRT	6Van Goor et al., 2014
II-III	E92K		X3,5,6,8,16	X5			3CFBE	3Frizzell lab, unpublished
							5CFBE	5Lukacs lab, unpublished
							6FRT	6Van Goor et al., 2014
							8HEK293	8Ren et al., 2013
							16HEK293	16Brodsky lab, unpublished
II–III	P99L		X11	X11			11HeLa	11Sheppard et al., 1996
II–III	D110H		X6	X6			6FRT	6Van Goor et al., 2014
II–III	R117C		X6	X6			6FRT	6Van Goor et al., 2014
II–III	R117H		X2,3,12,13	X2,6,12			2HBE	2Cyr lab, unpublished
							3CFBE	3Frizzell lab, unpublished
							6FRT	6Van Goor et al., 2014
							12FRT, HeLa	12Sheppard et al., 1993
							13HeLa	13Fanen et al., 1997
II–III	R170G		X7,14	X7			7CFBE	7Veit et al., 2014
							14BHK	14Okiyoneda et al., 2013
II–III	E193K		X5	X5,6			5CFBE	5Lukacs lab, unpublished
							6FRTf	6Van Goor et al., 2014
II–III	P205S		X11	X11			11HeLa	11Sheppard et al., 1996
II–III	L206W		X5,6,8	X5,6			5CFBE	5Lukacs lab, unpublished
							6FRT	6Van Goor et al., 2014
							8HEK293	8Ren et al., 2013
II–III	V232D		X15	X15			15HEK293	15Caldwell et al., 2011
II–III	R334W		X2,3,5	X2,5,6,12			2COS-7	2Cyr lab, unpublished
							3CFBE	3Frizzell lab, unpublished
							5CFBE	5 Lukacs lab, unpublished
II–III	R334W		X2,3,5	X2,5,6,12			2COS-7	2Cyr lab, unpublished
							3CFBE	3Frizzell lab, unpublished
							5CFBE	5 Lukacs lab, unpublished
							6FRTf	6Van Goor et al., 2014
							12HeLa	12Sheppard et al., 1993
II–III	I336K		X6	X6			6FRT	6Van Goor et al., 2014
II–III	T338I		X5,6	X5,6			5CFBE	5Lukacs lab, unpublished
							6FRT	6Van Goor et al., 2014
II–III	S341P		X5,6	X5,6			5CFBE	5Lukacs lab, unpublished
							6FRT	6Van Goor et al., 2014
II–III	A455E		X3,6,16, 17	X6			3CFBE	3Frizzell lab, unpublished
							6FRT	6Van Goor et al., 2014
							16HEK293	16Brodsky lab, unpublishedg
							17FRT, HeLad	17Sheppard et al., 1995
II–III	S549R		X3,5,18	X5,18			3CFBE	3Frizzell lab, unpublished
							5CFBE	5Lukacs lab, unpublished
							15FRT	18Yu et al., 2012
II–III	D579G		X5,6	X5,6			5CFBE	5Lukacs lab, unpublished
							6FRT	6Van Goor et al., 2014
II–III	R668C		X6	X6			6FRT	6Van Goor et al., 2014
II–III	L927P		X6	X6			6FRT	6Van Goor et al., 2014
II–III	S945L		X6	X6			6FRT	6Van Goor et al., 2014
II–III	S977F		X6	X6			6FRT	6Van Goor et al., 2014
II–III	L997F		X6	X6			6FRT	6Van Goor et al., 2014
II–III	H1054D		X6	X6			6FRT	6Van Goor et al., 2014
II–III	R1066H		X6	X6			6FRT	6Van Goor et al., 2014
II–III	A1067T		X6	X6			6FRT	6Van Goor et al., 2014
II–III	R1070Q		X6	X6			6FRT	6Van Goor et al., 2014
II–III	R1070W		X6,14	X6			6FRT	6Van Goor et al., 2014
							14BHK	14Okiyoneda et al., 2013
II–III	F1074L		X6	X6			6FRT	6Van Goor et al., 2014
II–III	D1270N		X6	X6			6FRT	6Van Goor et al., 2014
II–VI	S492F		X3,5,6			X5	3CFBE	3Frizzell lab, unpublished
							5CFBE	5Lukacs lab, unpublished
							6FRT	6Van Goor et al., 2014
II–III–VI	R347P		X3,5,6	X5,6,12		X5	3CFBE	3Frizzell lab, unpublished
							5CFBE	5Lukacs lab, unpublished
							6FRT	6Van Goor et al., 2014
							12HeLa	12Sheppard et al., 1993
II–III–VI	ΔF508		X19	X20		X21	19COS	19Cheng et al., 1990
							20Vero	20Dalemans et al., 1991
							21CHO	21Lukacs et al., 1993
II–III–VI	A561E		X6,22,23	X23		X23	6FRTd	6Van Goor et al., 2014
							22HBEd	22Awatade et al., 2015
							23BHK	23Wang et al., 2014
II–III–VI	L1077P		X3,6,16,24	X24		X24	3CFBE	3Frizzell lab, unpublished
							6FRT	6Van Goor et al., 2014
							16HEK293	16Brodsky lab, unpublished
							24CHO	24Sheppard lab, unpublishedh
II–III–VI	N1303K		X2,3,5,6,16,22,24	X24		X5	2HBE	2Cyr lab, unpublished
							3CFBE	3Frizzell lab, unpublished
							5CFBE	5Lukacs lab, unpublished
							6FRT	6Van Goor et al., 2014
							16HEK293	16Brodsky lab, unpublished
							22HBE	22Awatade et al., 2015
							24CHOi	24Sheppard lab, unpublished
III–VI	Q1411X			X25		X26	25BHK	25Gentzsch et al., 2002
							26Cos, BHK	26Haardt et al., 1999
II	A46D		X6				6FRT	6Van Goor et al., 2014
II	G85E		X3,6,16,24				3CFBE	3Frizzell lab, unpublished
							6FRT	6Van Goor et al., 2014
							16HEK293	16Brodsky lab, unpublished
							24CHO	24Sheppard lab, unpublished
III	R352Q			X5,6			5CFBE	5Lukacs lab, unpublished
							6FRTf	6Van Goor et al., 2014
II	L467P		X6				6FRT	6Van Goor et al., 2014
II	V520F		X3,6,16				3CFBE	3Frizzell lab, unpublished
							6FRT	6Van Goor et al., 2014
							16HEK293	16Brodsky lab, unpublished
II	A559T		X6				6FRT	6Van Goor et al., 2014
II	R560S		X6				6FRT	6Van Goor et al., 2014
II	R560T		X3,6,16				3CFBE	3Frizzell lab, unpublished
							6FRT	6Van Goor et al., 2014
							16HEK293	16Brodsky lab, unpublished
II	R560K		X3				3CFBE	3Frizzell lab, unpublished
II	Y569D		X6				6FRT	6Van Goor et al., 2014
II	D614G		X3				3CFBE	3Frizzell lab, unpublished
II	L1065P		X3,6				3CFBE	3Frizzell lab, unpublished
							6FRT	6Van Goor et al., 2014
II	R1066C		X3,6				3CFBE	3Frizzell lab, unpublished
							6FRT	6Van Goor et al., 2014
II	R1066M		X6				6FRT	6Van Goor et al., 2014
II	H1085R		X6				6FRT	6Van Goor et al., 2014
II	M1101K		X6				6FRT	6Van Goor et al., 2014
III	D110E			X6			6FRT	6Van Goor et al., 2014
III	G178R			X18			18FRT	18Yu et al., 2012
III	R347H			X6,7			6FRT	7Veit et al., 2014
							7CFBE	7Veit et al., 2014
III	S549N			X18			18FRT	18Yu et al., 2012
III	G551D			X27, 28			27CHO	27Bompadre et al., 2007
							28L	28Yang et al., 1993
III	G551S			X18			18FRT	18Yu et al., 2012
III	F1052V			X6			6FRT	6Van Goor et al., 2014
III	K1060T			X6			6FRT	6Van Goor et al., 2014
III	D1152H			X6			6FRT	6Van Goor et al., 2014
III	S1235R			X6			6FRT	6 Van Goor et al., 2014
III	G1244E			X18			18FRT	18 Yu et al., 2012
III	S1251N			X18			18FRT	18Yu et al., 2012
III	S1255P			X18			18FRT	18Yu et al., 2012
III	G1349D			X18,27			18FRT	18Yu et al., 2012
							27CHO	27Bompadre et al., 2007

Superscript numbers refer to references in far-right column.

aS.A.H. and D.M.C., unpublished observations

bK.W.P. and R.A.F., unpublished observations.

cR.G.A., H.Xu, and G.L.L., unpublished observations.

dDoes not exhibit a gating or conductance defect in this cell model.

eJ.S.H. and E.J.S., unpublished observations.

fDoes not exhibit a biogenesis defect in this cell model.

gA.N.C. and J.L.B., unpublished observations.

hZ.C. and D.N.S., unpublished observations.

iDoes not exhibit a peripheral stability defect in this cell model.

A recent study by Vertex Pharmaceuticals successfully demonstrated that 24 of 54 tested missensse mutations display both a processing (class II) and gating (class III) defect in the Fischer rat thyroid epithelial expression system (Van Goor ). Characterization of several rare CF mutations is ongoing in laboratories of the CFTR2 Consortium, the CFTR Folding Consortium, CFFT, Vertex Pharmaceuticals, and many others (Caldwell ; Yu ; Sosnay ; Harness-Brumley ; Hong ; Van Goor ; Wang ; Awatade ). This work will likely provide further examples of combinatorial mechanistic defects exhibited by CF mutants.

THERAPEUTIC SUSCEPTIBILITY OF CF MUTATIONS WITH COMPLEX BIOLOGICAL DEFECTS

In-depth analysis of the biology of CF mutants distinguishes them according to their complex molecular pathology and suggests drug combinations for treatment of different patient populations. This process, called “theratyping” (Cutting, 2015), will pave the way to personalized medicine in CF. However, reliable prediction of the responsiveness of a mutant phenotype to pharmacotherapy could be challenging and is dependent on the cellular model system (Pedemonte ).

Emerging evidence also suggests that the efficacy of approved and preclinical drugs may vary with different mutations within the same class. For example, while nearly complete processing correction of P67L- and R170G-CFTR (class II) was achieved with VX-809 treatment (Okiyoneda ; Ren ; Veit ), VX-809 only partially reversed the folding defect of some other class II mutants; for example, N1303K and ΔF508 (Okiyoneda ; Awatade ). This differential susceptibility to correction is attributed to the nature of the primary folding/structural defect. According to one hypothesis, robust folding correction of ΔF508-CFTR requires corrector combinations to avert its NBD1-MSD1/2 interface and NBD1 stability defects (Mendoza ; Rabeh ; He ; Okiyoneda ). The N1303K mutation in NBD2 was not rescued by VX-809, and only modest processing was observed by targeting both the NBD1/MSDs and NBD2 interfaces with C4 and C18 (a VX-809 analogue) (Okiyoneda ; Rapino ).

Some of the class III mutations also respond differently to the gating potentiator VX-770. Although R347H- and T338I-CFTR cause severe functional defects with no or modest loss of protein expression, only R347H-CFTR is potentiated by VX-770 to near wild type–like conductance (Van Goor ). Likewise, the P5 potentiator activates ΔF508-CFTR, but it has no effect on G551D-CFTR chloride permeation (Yang ). Thus identification of mutation-specific novel potentiators or their combinations may further optimize channel rescue for specific class III/IV mutations. Additive enhancement of G551D-CFTR activity by the combination of the potentiators genistein and curcumin supports the feasibility of combining potentiators (Yu ). Likewise, we envision that mutation-specific read-through drugs will ultimately need to be combined with other correctors and potentiators, based on the pleiotropic defects associated with this class of mutations (as illustrated for W1282X above).

CONCLUDING REMARKS

The ultimate goal of theratyping is to achieve optimal correction of a specific mutant defect by selecting the most efficacious CFTR modulator(s), including correctors(s), potentiator(s), and/or read-through drugs, or a combination of these drugs. Based on accumulating observations, however, mechanistic subdivisions of some of the major classes of mutations (classes I, II, and III) may be necessary to further improve the success of drug-selection strategies. This will facilitate the theratyping of CF alleles and their combinations and expedite the identification and approval process for combination therapies. Theratyping has already proven successful in identifying class III mutations that are responsive to VX-770 (Yu ), leading to the approval of this drug for eight rare mutations besides G551D (Vertex, 2014a). In fact, the results of large-scale theratyping could be overlaid as a third dimension on the Venn diagram presented in Figure 2.

Thus, during the 22 years following the initial classification of CF mutations (Welsh and Smith, 1993), our understanding of the molecular complexity of CF alleles has evolved remarkably, establishing the need for an advanced mutation classification scheme in conjunction with personalized CF therapy.