The response of rare CFTR mutations to specific modulator combinations

Abstract

Background

The combination of the cystic fibrosis transmembrane conductance regulator (CFTR) modulators elexacaftor (VX-445)–tezacaftor (VX-661)–ivacaftor (VX-770) (ETI) enables the effective rescue of CFTR function in people with the F508del mutation and 177 other US Food and Drug Administration-approved alleles. However, the effect of modulator combination treatment on many rare CFTR mutations is mostly unknown. Furthermore, rare CFTR mutations may not require all ETI components to reach maximal correction. This can be studied in intestinal organoids derived from patients carrying rare mutations.

Methods

Intestinal organoids were generated from six patients carrying the Q1100P and/or K163E alleles, not receiving ETI. Measurements of the response to ETI or combination of its components were performed in three-dimensional organoids by forskolin-induced swelling and in two-dimensional monolayers by short-circuit current. Based on these results, patients initiated off-label ETI treatment. Clinical data before and after treatment were collected.

Results

Functional measurements showed that both mutations are responsive to ETI. The results further showed that VX-445 had a dramatic effect on K163E function. Both Q1100P and K163E mutations achieved clinically significant CFTR activity levels with VX-661+VX-445, without benefit from VX-770. Following these results patients initiated off-label ETI treatment, resulting in significant and sustained clinical improvements, in all patients, in lung function (forced expiratory volume in 1 s and lung clearance index), body mass index and sweat chloride.

Conclusion

Our results support that CFTR function measurements in patient-derived intestinal organoids carrying rare CFTR alleles can detect potential responders to modulator treatment and serve as the basis for drug approval by health providers. Furthermore, this approach allows for patient-specific optimisation of modulator combinations, minimising unnecessary exposure to ineffective treatments.

Shareable abstract

CFTR function measurements in patient-derived intestinal organoids carrying rare mutations can detect responders to modulator treatment and enable optimisation of patient-specific modulator combinations, minimising exposure to ineffective treatments https://bit.ly/41QA6ou

Introduction

Cystic fibrosis (CF) is a lethal autosomal recessive disease caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene, which encodes a chloride ion channel in epithelial cell membranes. The disease is characterised by chronic progressive lung disease, pancreatic insufficiency, gastrointestinal malabsorption, malnutrition, diabetes and male infertility [1]. Over 2100 CFTR sequence variants were identified [2], of which 804 are disease-causing [3].

Substantial progress has been made in the development of CFTR-modulating therapies [4]. The latest, highly efficient treatment for mutations affecting CFTR processing and folding consists of the correctors elexacaftor (VX-445) and tezacaftor (VX-661), which improve the folding, processing and stability of the F508del protein [5–7], and the potentiator ivacaftor (VX-770), which increases channel opening [8]. This combination is referred to as ETI. Clinical trials have shown that ETI significantly improves lung function in CF patients carrying at least one F508del allele [9, 10]. ETI, known as Trikafta in the USA and Kaftrio in Europe, was approved for patients with the F508del mutation, affecting >85% of patients worldwide. The US Food and Drug Administration (FDA) extended approval to 177 additional mutations based on in vitro studies in Fischer rat thyroid (FRT) cells [11].

However, ∼10–15% of CF patients carry rare mutations not approved for ETI. While complementary DNA-based platforms such as FRT cells may help extend available drugs to such subjects, they do not reflect the CFTR native context and might not be a suitable model for all variants, such as splicing and nonsense mutations, dependent on intronic sequences. Ex vivo models, such as intestinal organoids [12], enable measurement of the CFTR modulator response within the individual genetic background. Furthermore, it has been shown that some rare CFTR mutations may not require all ETI components to reach maximal correction [6, 13, 14]. This can be studied in organoids derived from patients carrying rare CFTR mutations.

Intestinal organoids are sensitive to modulator effects and provide a dynamic CFTR-dependent functional readout, measured by the forskolin-induced swelling (FIS) assay [15]. Importantly, in vitro CFTR modulator responses in organoids correlate with lung disease severity and sweat chloride levels, biomarkers of in vivo CFTR function [16, 17].

Here, we studied the response of organoids derived from patients carrying two rare CFTR mutations, Q1100P (c.3299A→C) and K163E (c.487A→G), to ETI and to each of its components. The results show that both mutations are responsive in intestinal organoids to ETI. Q1100P was responsive to VX-445 treatment alone, and K163E responded to VX-661 alone. VX-445 alone had a dramatic effect on the function of the K163E allele. Both mutations achieved clinically significant CFTR activity levels with VX-661+VX-445, without benefit from VX-770. We suggest that patient-derived models such as intestinal organoids can enable the optimisation of mutation-specific modulator combinations to maximise efficacy and minimise lifelong drug exposure of patients.

Following these results ETI treatment was initiated, resulting in significant and sustained clinical improvements in all patients.

Materials and methods

Rectal biopsies and organoid generation

We were approached by CF centres to test the response to ETI of six patients carrying the Q1100P and/or the K163E alleles (table 1). They were invited to participate in the study and informed consent was obtained. Genotypes were defined by exome sequencing, revealing no additional mutations or polymorphisms within or outside the coding exons (gene by gene). Patients were not receiving modulator treatment at the initiation of this study. Rectal biopsies were obtained at Hadassah Medical Centre (Jerusalem, Israel) and organoids were successfully derived from all patients according to a published protocol [15]. Organoids were seeded in Matrigel drops (Corning Inc.), cultured in IntestiCult (Stemcell Technologies) and passaged weekly.

TABLE 1.

Patient characteristics

Patient	Organoid culture	Allele 1	Allele 2	Age (years)	Sex	FEV1	Sweat chloride, mmol·L−1	PS/PI	CFRD
1	CF1	Q1100P	C225X	15	Male	79	102	PI	–
2	CF2	Q1100P	N1303K	40	Female	53		PI	–
3#	CF3	Q1100P	K163E	18	Male	53	72	PS	–
4#	CF4	Q1100P	K163E	17	Male	75	73	PS	–
5##	CF5	K163E	K163E	18	Male	76	48	PS	–
6##	CF6	K163E	K163E	8	Female	99	70	PS	–

Patients marked by ^# are siblings to each-other, as are patients marked by ^##. ^# and ^## are first cousins. –: Patient does not have cystic fibrosis-related diabetes; CFRD: cystic fibrosis-related diabetes; FEV₁: forced expiratory volume in 1 s; PI: pancreatic insufficient; PS: pancreatic sufficient.

CFTR function measurement in intestinal organoids by FIS

Organoids were incubated with or without 3 μM VX-445 and/or VX-661 (Selleck Chemicals LLC) for 24 h before the assay. FIS was conducted as previously described [15]. The total organoid-area increase was analysed using an in-house Python script in FiJi (see the supplementary material) [18].

Organoid-derived monolayers and short-circuit current (I_sc) recordings

Monolayers were cultured following a published protocol [19] except for our use of Matrigel 1:40 for filter coating. Electrophysiological measurements were performed using the Ussing chamber, 6–8 days after seeding, according to the standard operating procedure (ICM_EU001 of the European Cystic Fibrosis Society Diagnostic Network Working Group, version 2.7), with modifications for monolayers (see supplementary material). Non-CF values were determined in monolayers derived from four non-CF individuals who had biopsies taken while undergoing routine colonoscopies. The median of the four non-CF individuals was used for calculation.

CFTR protein levels

Protein extracts were prepared in radioimmunoprecipitation assay buffer separated on 6% polyacrylamide gels and transferred to a nitrocellulose membrane. Antibody hybridisation and chemiluminescence were performed according to standard procedures, using mouse anti-CFTR 596 (CFF) and rabbit anti-calnexin (Sigma) primary antibodies. Secondary antibodies were obtained from Jackson Immunoresearch Laboratories.

Statistical analysis

FIS: the collected data were analysed using Prism10 (GraphPad Software, LLC). Results are presented as mean+sem of 4–12 biological repeats. Difference sbetween treatments were determined by one-way ANOVA (p<0.05).

I_sc recordings: results are presented as median+min–max range of 4–11 biological repeats, with differences between treatments determined by t-test or one-way ANOVA (p<0.05).

Results

Genotypes and clinical characteristics of patients participating in the study

We studied the response to the current available modulators in organoids derived from six patients carrying at least one copy of the rare CFTR mutations Q1100P or K163E. Patients 1 and 2 carry Q1100P on one allele and a minimal function mutation on the other allele (table 1).

Q1100P was reported in Spain, Portugal and Brazil [20–23] and was identified in various population screenings [24–29], although it does not appear in the CFTR2 database. K163E has not been reported in CFTR1 nor CFTR2.

The triple combination VX-661+VX-445+VX-770 restored CFTR function in organoids with the Q1100P and/or K163E mutations

To assess CFTR function following treatment with the triple combination VX-661+VX-445+VX-770, organoids were incubated with increasing forskolin concentrations (0.02–5.0 μM). No baseline swelling was detected in any of the organoids, indicating minimal function genotypes (figure 1). Organoids were incubated for 24 h with 3 μM VX-445 and VX-661, and acutely stimulated with forskolin and 3 μM VX-770 at time=0. As a control for a clinically relevant response, we used the CFTR activity level in organoids from a patient heterozygous for F508del and the S1251N gating mutation, treated with VX-770 (figure 1). In order to enable comparison among organoids, we used FIS levels reached at 0.32 μM forskolin, which allowed for the best differentiation between responses to treatment. All organoids showed a significant response to VX-661+VX-445+VX-770, ranging from 54 to 157% of control levels (table 2). Notably, some functional rescue was observed at the lowest forskolin concentration (0.02 µM) in organoids carrying at least one K163E allele, indicating high susceptibility of this mutation to correction (figure 1).

FIGURE 1.

The triple combination tezacaftor (VX-661)+elexacaftor (VX-445)+ivacaftor (VX-770) restored cystic fibrosis transmembrane conductance regulator (CFTR) function measured by forskolin-induced swelling (FIS) in intestinal organoids carrying the Q1100P and/or K163E mutations. a–f) FIS of patient-derived intestinal organoids carrying the Q1100P and/or K163E mutations, treated for 24 h with 3 μM VX-661+VX-445 or dimethyl sulfoxide (DMSO), induced by different forskolin concentrations, with or without acute addition of 3 μM VX-770, expressed as the area under the curve (AUC). g) Control: FIS of intestinal organoids derived from a patient compound heterozygous for the S1251N and F508del CFTR mutations, treated with 3μM VX-770. 0.32 μM forskolin (dotted line) was chosen for further comparison. *: p<0.05. **: p<0.01. ***: p<0.001. ****: p<0.0001.

TABLE 2.

Forskolin-induced swelling summary

Organoid culture	Basal	VX-661	VX-445	VX-661+ VX-445	VX-770	VX-661+VX-770	VX-445+VX-770	VX-661+VX-445+VX-770
CF1	0	0	14	93	0	0	31	98
CF2	0	0	3	43	0	0	8	54
CF3	0	12	93	152	0	51	126	119
CF4	0	21	109	168	1	66	169	157
CF5	0	10	53	118	1	31	74	115
CF6	0	50	125	153	7	128	155	135

Results are presented as % of control at 0.32 µM forskolin. VX-445: elexacaftor; VX-661: tezacaftor; VX-770: ivacaftor.

The double combination VX-661+VX-445 restored CFTR function in organoids with the Q1100P mutation, without added benefit from VX-770

It has been shown that there are rare CFTR mutations, eligible for ETI, that reach maximal correction with a combination of two modulators only [6, 13, 14]. Thus, we investigated the response of Q1100P and K163E to each modulator separately and in combination. Patient 1 is a compound heterozygote for the missense mutation Q1100P and the nonsense mutation C225X. C225X introduces a premature termination codon in exon 6 triggering transcript degradation by the nonsense-mediated mRNA decay pathway leading to low levels of truncated CFTR proteins. Consequently, C225X cannot contribute to CFTR function. As can be seen in figure 2, VX-661 alone had no effect on CFTR function in organoids derived from patient 1 (CF1). VX-445 alone partially restored CFTR function and addition of VX-770 had no effect. Combined treatment of VX-445 and VX-661 further enhanced CFTR function to 93% of control. Acute addition of VX-770 to the VX-661+VX-445 combination had no beneficial effect (figure 2 and table 2).

FIGURE 2.

The double combination tezacaftor (VX-661)+elexacaftor (VX-445) restored cystic fibrosis transmembrane conductance regulator (CFTR) function measured by forskolin-induced swelling (FIS) in intestinal organoids carrying the Q1100P mutation, without added benefit from ivacaftor (VX-770). a) and b) FIS of intestinal organoids treated for 24 h with 3 μM CFTR modulator combinations as indicated, induced by 0.32 μM forskolin, with or without acute addition of 3 μM VX-770. c) Same experimental conditions as in a) and b), except for additional analyses of acute addition of 3 μM VX-445 (VX-445a). Data are expressed as mean+SEM of the area under the curve (AUC). Dotted line represents CFTR activity from S1251N/F508del intestinal organoids treated with 3 μM VX-770, for a clinically relevant response. DMSO: dimethyl sulfoxide; NS: nonsignificant. *: p<0.05. **: p<0.01.

Patient 2 is a compound heterozygote for Q1100P and the minimal function mutation N1303K. N1303K was previously shown by us and others to respond to the triple combination in patient-derived organoids and primary respiratory epithelial cells [6, 30–33]. The response pattern to modulator treatment in CF2 reflected that of CF1. VX-445 alone had a small effect on CFTR function, which was significantly enhanced with the combined treatment VX-661+VX-445, without additional benefit from VX-770 (figure 2 and table 2). In comparison, organoids from a patient homozygous for N1303  required all three modulators for restoration of CFTR function (figure S1), indicating that Q1100P determines the response pattern to modulator treatment in CF2. Maximal CFTR activity in CF2 was moderate, reaching 54% of control.

VX-445 is known to act both as a corrector and a potentiator of wild-type (WT) as well as CFTR proteins with F508del or other rare mutations [32, 34]. To investigate whether the rescue of CFTR function of Q1100P by VX-445 involves potentiation, we examined the effect of acute VX-445 addition on CFTR function in CF1, with and without prior treatment with modulator combinations. Similar to VX-770, no additive effect from acute treatment with VX-445 was seen following treatment with VX-661 alone or in combination with VX-445 (figure 2c). This suggests Q1100P causes a processing/folding defect corrected by VX-661+VX 445. However, since VX-445 is chronically administered, the results do not enable to differentiate between the corrector and potentiator functions of VX-445, raising the possibility that part of the VX-445 effect results from its function as a potentiator of the corrected CFTR proteins.

To evaluate the impact of modulator treatment on mature CFTR protein production, we performed Western blots alongside functional data analysis. The CFTR protein appears as a fully glycosylated band C and an incompletely glycosylated band B. Figure S2 shows VX-661 alone did not affect CFTR maturation in CF1. VX-445 enabled band C production, confirming its function as a corrector for the Q1100P mutation. Band C was further enhanced by VX-661+VX-445. Weaker signals in CF2 correlated with lower functional rescue. Remarkably, even low CFTR maturation levels led to significant CFTR activity restoration.

VX-445 alone significantly restored CFTR function in organoids with the K163E mutation

CF3 and CF4 are compound heterozygous for the Q1100P/K163E mutations, thus their response may arise from both alleles. CF5 and CF6 are homozygous for K163E. VX-661 alone partially restored CFTR function in organoids from all patients (CF3–CF6) and acute VX-770 addition enhanced this CFTR activity (figure 3 and table 2). VX-445 alone significantly restored CFTR function reaching control levels in CF3, CF4 and CF6. The response of CF6 reached its maximum effect already with VX-445 alone. The addition of VX-661 further enhanced the response to VX-445 in CF3–CF5, exceeding control levels. In all organoids, VX-770 addition provided no additional benefit. This is a different response pattern than the one observed for Q1100P, which did not respond to VX-661 alone (figure 2). These results suggest that K163E determines the response pattern in all cultures carrying this mutation.

FIGURE 3.

Elexacaftor (VX-445) alone significantly restored cystic fibrosis transmembrane conductance regulator (CFTR) function measured by forskolin-induced swelling (FIS) in intestinal organoids with the K163E mutation. a–d) FIS of intestinal organoids treated for 24 h with 3 μM CFTR modulator combinations as indicated, and induced by 0.32 μM forskolin, with or without acute addition of 3 μM ivacaftor (VX-770). e) Same experimental conditions as in a–d), except for additional analyses of acute addition of 3 μM VX-445 (VX-445a). Data are expressed as mean+SEM of the area under the curve (AUC). Dotted line represents CFTR activity from S1251N/F508del intestinal organoids treated with 3 μM VX-770, for a clinically relevant response. DMSO: dimethyl sulfoxide; NS: nonsignificant; VX-661: tezacaftor. *: p<0.05. **: p<0.01. ***: p<0.001. ****: p<0.0001.

To investigate whether the rescue of CFTR function of K163E by VX-445 involves potentiation, we examined the effect of acute addition of VX-445 on CFTR function on K163E. The analysis was performed in CF5, homozygous for the mutation. Acute VX-445 addition significantly increased CFTR activity after chronic VX-661 treatment (figure 3e), indicating VX-445 acted as a potentiator of the VX-661-corrected K163E protein. However, this potentiation resulted in a lower activity than chronic treatment with VX-445 alone. Altogether, each corrector alone benefited from the addition of a potentiator, but the highest CFTR activity was achieved by the combination of the two correctors VX-661+VX-445, suggesting K163E causes a processing/folding defect. However, since VX-445 is chronically administered, the results raise the possibility that part of the VX-445 effect results from its function as a potentiator of the corrected CFTR proteins.

Analysing the effect of modulator treatment on CFTR protein maturation by Western blots of CF3-CF6 showed VX-661 alone partially produced mature CFTR proteins, and VX-445 alone had a stronger effect, in all organoids, but to a lesser extent in CF3. This confirms that VX-445 effectively corrected the processing defect induced by K163E. VX-661+VX-445 further increased CFTR maturation (figure S2). Despite variations in CFTR maturation levels, CFTR function in all organoids reached or exceeded control levels, indicating that significant function restoration can be achieved even by partial maturation.

The double combination VX-661+VX-445 restored CFTR function in intestinal monolayers, without added benefit from VX-770

Our results using FIS show a restoration of CFTR activity similar or higher than the response of gating organoids treated with VX-770, suggesting clinical relevance. For further validation of the response pattern of Q1100P and K163E, we analysed CFTR response to the modulators by I_sc, using two-dimensional cultures from the same organoids, enabling a direct measurement of CFTR function as well as comparison of the response to non-CF levels. Measurements in CF1 (Q1100P/C225X) monolayers showed no response to VX-661, partial response to VX-445 and the response was further enhanced by VX-661+VX-445, reaching ∼40% of the non-CF response (figure 4a and a representative trace in figure S3A). This response pattern was similar to that observed in FIS. Measurements in CF4 (Q1100P/K163E) and CF5 (K163E/K163E) monolayers showed partial response to VX-661, significant response to VX-445 alone, reaching 79% and 84% of non-CF level respectively, and the response was further enhanced by VX-661+VX-445, reaching above non-CF range (106% and 118% of non-CF) (figure 4b,c and representative traces in figure S3B,C). In these monolayers, acute VX-770 addition had no beneficial effect. This response pattern was also similar to that observed in FIS. It is important to note that there is a high correlation between the responses in FIS and I_sc to each treatment (figure S4), indicating that Q1100P does not benefit from VX-770 addition and K163E is primarily responding to VX-445. Furthermore, the I_sc measurements, which reached non-CF levels, support the clinical potential of modulator treatment.

FIGURE 4.

The double combination tezacaftor (VX-661)+elexacaftor (VX-445) restored cystic fibrosis transmembrane conductance regulator (CFTR) function measured by short-circuit currents (I_sc) in intestinal monolayers carrying the Q1100P and/or K163E mutations, without added benefit from ivacaftor (VX-770). a–c) I_sc of intestinal organoid monolayers treated for 24 h with 3 μM CFTR modulator combinations as indicated and induced by 10 μM forskolin, with or without acute addition of 3 μM ivacaftor (VX-770). Results are presented as the median+min–max range of Δforskolin. Grey shading: non-CF range; dotted line: non-CF median. d–f) Western blot analysis of CFTR protein expression in intestinal organoid monolayers after 24-h treatment with 3 μM CFTR modulator combinations as indicated, or dimethyl sulfoxide (DMSO). Calnexin is a loading control. Band B: immature, core-glycosylated CFTR; Band C: mature, complex-glycosylated CFTR. NS: nonsignificant. *: p<0.05. **: p<0.01. ***: p<0.001. ****: p<0.0001.

Analysis of CFTR protein maturation in CF1 monolayers showed VX-661 alone did not affect CFTR maturation, VX-445 increased band C production, further enhanced by VX-661+VX-445 (figure 4), similarly to the pattern observed in three-dimensional (3D) organoids (figure S2). Analysis of CFTR protein maturation in CF4 and CF5 showed VX-661 alone induced CFTR maturation, VX-445 further increased band C production, and treatment with VX-661+VX-445 restored protein levels to non-CF levels. The higher level of protein maturation observed after modulator treatment in monolayers compared to 3D organoids might be explained by the exposure to the medium of both apical and basal side of cells in monolayers, compared to the 3D medium exposure from the basal side only, resulting in higher modulator uptake and efficiency. Altogether, the CFTR response pattern measured in monolayer organoids validated the pattern found in the 3D model.

ETI treatment had a significant and sustained clinical benefit in patients with the Q1100P and/or the K163E mutations

All patients showed significant clinical improvement after 1 month of treatment with ETI: a 24.2±14.7% mean increase in forced expiratory volume in 1 (FEV₁), a 2.0±2.5 kg mean increase in weight, a 0.8±0.4 mean increase in body mass index (BMI) among the five adult patients (1–5), a −4.0±1.6 mean decrease in lung clearance index and a −23.3±13.1 mmol·L⁻¹ mean decrease in sweat chloride (table 3). In four out of five patients, sweat chloride levels were reduced to normal or borderline levels. All patients had long-term follow-up (at two time-points) in pulmonary function, BMI and weight, showing sustained improvement in FEV₁ and continued increase in BMI and weight (figure 5 and table S1).

TABLE 3.

Clinical data before and after 1 month of treatment with elexacaftor–tezacaftor–ivacaftor

Patient	FEV1, % predicted			Sweat chloride, mmol·L−1			LCI			BMI			Weight (kg)
	Before	After 1 month	Change	Before	After 1 month	Change	Before	After 1 month	Change	Before	After 1 month	Change	Before	After 1 month	Change
1	79	113	+34	102	86	−16	10.7	7.5	−3.2	19.7	20.3	+0.6	60.2	62.5	+2.3
2	53	99	+46							19.0	20.0	+1.0	54.2	56.9	+2.7
3	53	74	+21	72	42	−30	16.4	11.5	−4.9	20.9	22.2	+1.3	61.8	68.0	+6.2
4	75	90	+15	73	64	−9	13.1	7.5	−5.6	27.1	28.2	+1.1	87.0	87.3	+0.3
5	76	101	+25	48	22	−26	13.7	9.7	−4	25.0	24.9	−0.1	76.6	75.5	−1.1
6	99	103	+4	70	32	−38	8.3	6.1	−2.2	16.6	16.8	+0.2	28.0	29.8	+1.8
Average			+24.2±14.7			−23.8±11.5			−4±1.3			0.7±0.5			2±2.5

BMI: body mass index; FEV₁: forced expiratory volume in 1 s; LCI: lung clearance index.

FIGURE 5.

Long-term effect of treatment with elexacaftor–tezacaftor–ivacaftor (ETI) in patients carrying the Q1100P and/or K163E mutations. Patients were treated with ETI and monitored for up to 26 months after initiation of treatment. At long-term follow-up meetings, patients were assessed for a) forced expiratory volume in 1 s (FEV₁) (% predicted), b) weight (kg) and c) body mass index (BMI).

Discussion

Here, we demonstrated the ability of VX-661, VX-445 and VX-770 to restore CFTR activity in organoids derived from patients carrying the ultra-rare mutations Q1100P and K163E (figure 1).

In organoids carrying Q1100P (CF1 and CF2), VX-445 alone increased Q1100P function, which was further augmented by VX-661. Acute addition of VX-770 to the VX-661+VX-445 combination had no added beneficial effect (figure 2 and table 2). Q1100P, located in transmembrane (TM) helix 11, possibly disrupts the α-helix structure due to the bulky side chain of proline and its inability to form hydrogen bonds. The same response pattern to modulators was previously shown for M1101K [6], a missense mutation on an amino acid adjacent to Q1100P, which has been approved for ETI based on FRT cell results [11]. Recently, structures of the CFTR protein bound to each of the ETI components were established through cryo-electron microscopy, explaining how each modulator rescues the CFTR structure and function [5, 7]. VX-445 stabilises TM10 and 11 [5], strengthening the transmembrane domain 1 (TMD1)–nuclear binding domain 1 (NBD1) interface, crucial for CFTR folding and processing [35]. VX-445 interacts with TM11 through residues adjacent to Q1100 and M1101 (W1098 and R1102 [7]), potentially correcting their defects. Interestingly, both mutations do not disrupt VX-445 binding to TM11, enabling their correction.

In all organoids carrying K163E (CF3–CF6), VX-661 partially restored CFTR activity (figure 3 and table 2). K163 is located in TMD1 and is part of a stretch of conserved lysine residues at the TM2–intracellular loop 1 (ICL1) boundary [36], a region participating in the ICL1–NBD1 interphase, which is essential for correct folding of the CFTR protein. This region is stabilised by both VX-809 [35, 37, 38] and its structural analogue VX-661 [7]. Interestingly, VX-445 alone achieved a high restoration of CFTR activity in all K163E organoids, which was further enhanced by VX-661+VX-445, without added benefit from acute VX-770 addition.

The Q1100P and K163E proteins were maximally corrected by VX-445+VX-661 and did not benefit from the acute addition of VX-770 (figures 2, 3 and 4). The reason for this lack of effect is not clear but could result from the activity of VX-445 as a potentiator. Previous studies have shown that VX-445 can act both as a corrector and a potentiator [32, 34]. In the case of K163E (CF5), acute VX-445 or VX-770 addition potentiated VX-661-corrected K163E; however, chronic treatment with VX-445 in combination with VX-661 was more efficient in restoring K163E CFTR activity (figure 3e). In the case of Q1100P (CF1), acute VX-445 treatment following VX-661 alone or in combination with VX-445 did not increase the CFTR response (figure 2c) and the maximal effect was achieved with chronic VX-661+VX-445 treatment. Thus, in both mutations, part of the VX-445 effect may also result from its function as a potentiator of the corrected CFTR proteins, reaching maximal function that could not be further enhanced by VX-770.

In this work, we assessed CFTR function in patient-derived intestinal epithelial cells using two assays, FIS and I_sc. FIS is a relatively simple assay, measuring CFTR function indirectly by coupling ion secretion to fluid transport. I_sc in monolayers enable for a direct measurement of CFTR activity and a direct comparison to non-CF activity but require complex culture and measurement methods. The unique response pattern of each of the rare, studied CFTR mutations was observed in both CFTR functional assays. We show in the two assays a significant correlation between the CFTR responses to the different modulator combinations (figure S4), further validating FIS as a reliable assay for CFTR activity.

Based on our in vitro results, the medical providers approved off-label treatment with ETI for all patients presented here. The clinical benefits were immediate and substantial. After the first month of treatment, the mean FEV₁ increased by 24.2+14.7% (table 3). Patients had long-term follow-up in pulmonary function, weight and BMI, showing sustained improvement (figure 5 and table S1). Thus, our results further support that results from organoids can serve as a basis for drug approval for patients carrying rare CFTR mutations not included in clinical trials.

In a large study, Bihler et al. [39] analysed the response of 655 rare CFTR variants to ETI, using transient transfections of CFTR variants into FRT cells. In this study, the response of the Q1100P allele was assessed and surprisingly it was defined as a nonresponder. The threshold defined in that large study was 10% of WT; however, as the authors mentioned, this threshold should not be considered as a rigid cut-off. We suggest that borderline responses should be assessed in patient-derived systems if possible, or by a clinical assessment of ETI responsiveness.

ETI was initially developed for the prevalent F508del mutation [5] and was approved by the FDA as a triple combination drug with a significant effect compared with the previous drug combination tezacafor/ivacaftor (Symdeko) [9]. The triple combination was further approved for an additional 177 rare CFTR mutations [11]. The results presented in this study show that there may be rare mutations such as K163E that can be substantially corrected by treatment with VX-445 alone. Additionally, our results concur with previous results in patient-derived intestinal organoids and nasal epithelial cells showing that the rare CFTR mutations M1101K, P67L, L206W and A559T can be substantially corrected by the combined use of only the two modulators, VX-661+VX-445, without additional effect of VX-770 [6, 13, 14]. The highly effective treatments with the current available modulators were shown to have side effects in some patients. VX-770 has been shown in some patients to cause headaches, rash, elevation in liver enzymes [40] and some patients treated with ETI experienced several side-effects [9, 10], including serious psychiatric side-effects in ∼10% of patients after prolonged treatment [41]. All these results suggest that the use of patient-derived models can help optimise modulator combinations for rare mutations, limiting unnecessary patient exposure to ineffective compounds. Importantly however, clinical trials with specific modulator combinations are required before any change in clinical treatment could be implemented.

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material

01308-2024.SUPPLEMENT