EFFECT OF ELEXACAFTOR/TEZACAFTOR/IVACAFTOR ON MUCUS AND MUCOCILIARY CLEARANCE IN CYSTIC FIBROSIS

Abstract

Background:

The cystic fibrosis transmembrane conductance regulator (CFTR) modulator elexacaftor/tezacaftor/ivacaftor (E/T/I) is highly effective clinically for those with at least one F508del-CFTR allele. The effects of E/T/I on mucociliary clearance (MCC) and sputum properties are unknown. We, therefore, sought to characterize the effects of E/T/I on in vivo MCC and sputum characteristics hypothesized to impact mucus transport.

Methods:

Forty-four participants ≥12 years of age, were enrolled into this prospective, observational trial prior to initiation of E/T/I and had baseline measurement of MCC and characterization of induced sputum and exhaled breath condensate (EBC) samples. Study procedures were repeated after 1 month of E/T/I treatment.

Results:

Average age was 27.7 years with baseline forced expiratory volume in 1 second (FEV₁) of 78.2% predicted. 52% of subjects had previously been treated with a 2-drug CFTR modulator combination. The average whole lung MCC rate measured over 60 minutes (WLAveClr60) significantly improved from baseline to post-E/T/I (14.8 vs. 22.8%; p = 0.0002), as did other MCC indices. Sputum % solids also improved (modeled mean 3.4 vs. 2.2%; p<0.0001), whereas non-significant reductions in sputum macrorheology (G’, G”) were observed. No meaningful changes in exhaled breath condensate endpoints (sialic acid:urea ratio, pH) were observed.

Conclusions:

E/T/I improved the hydration of respiratory secretions (% solids) and markedly accelerated MCC. These data confirm the link between CFTR function, mucus solid content, and MCC and help to define the utility of MCC and mucus-related bioassays in future efforts to restore CFTR function in all people with CF.

1. INTRODUCTION

Highly effective CFTR modulator therapies are rapidly changing the lives of people with CF (pwCF). Elexacaftor/tezacaftor/ivacaftor (E/T/I) is approved for those with at least one copy of the common F508del-CFTR mutation (~85% of pwCF) and has been shown to improve lung function, exacerbation frequency, respiratory symptoms, and nutritional indices [1, 2]. The real-world effects of E/T/I were confirmed in the PROMISE observational study [3]. Multiple PROMISE sub-studies were designed to more deeply investigate the biological and clinical effects of CFTR restoration with E/T/I[4, 5].

CFTR controls the movement of chloride and bicarbonate across epithelial surfaces. In the lung, CFTR dysfunction manifests as mucus dehydration and airways obstruction, which promote chronic bacterial infection and airways inflammation. Restoration of CFTR activity with ivacaftor in those with the G551D-CFTR mutation led to significant improvement in mucociliary clearance (MCC) within 1 month of beginning treatment[6]. In contrast, the combination of lumacaftor and ivacaftor in F508del homozygous individuals did not have meaningful effects on MCC[7]. In this study, we sought to determine whether the potent effects of E/T/I on CFTR activity translated to improved MCC.

The link between mucus abnormalities and airways obstruction remains incompletely defined. We hypothesized that measurable mucus properties would respond to restoration of CFTR activity, thus explaining improvements in MCC. To test this hypothesis, induced sputum samples were collected to allow assessment of sputum hydration and rheology. Exhaled breath condensate (EBC) samples were also collected to determine whether restoration of CFTR function led to measurable changes in EBC pH and sialic acid:urea concentration, an index of mucin concentration[8].

2. METHODS

2.1. Population and Study Design

The PROMISE sub-study was performed at 4 specialized sites that were participating in the core PROMISE study (NCT04038047)[3]. Eligible participants were ≥12 years of age, had at least one copy of the F508del-CFTR mutation, and were prescribed E/T/I by their physician. People with CF treated with ivacaftor monotherapy for a highly responsive mutation were not included in this substudy. Exclusion criteria included a FEV₁ <30% predicted at screening, addition of a new chronic therapy within 28 days, or acute treatment with antibiotics or corticosteroids within 14 days of baseline MCC measurement. Participants were required to withhold hypertonic saline and dornase alfa use for at least 12 hours before each MCC scan, but otherwise were asked to continue all non-modulator treatments throughout the study. A minimum sample size of 32 subjects as determined to provide >80% power to detect a significant change in MCC and sputum % solids, based upon prior data[6].

After enrollment, a baseline visit was conducted before initiating E/T/I, and a post-treatment visit was performed approximately 30 days after initiating E/T/I. At each visit, clinical assessments included: spirometry, height, weight, Cystic Fibrosis Questionnaire-Revised (CFQ-R) respiratory domain[9] and sweat chloride measurement. MCC measurement by gamma scintigraphy, exhaled breath condensate (EBC) collection, and sputum induction were also performed.

2.2. MCC Measurement

A standard operating procedure was utilized as described previously[6]. A planar Co⁵⁷ source was used to delineate the lung contour and define regions of interest (ROI) (i.e., peripheral and central lung regions). Technetium-99m–labeled (Tc^99m-labeled) sulfur colloid particles were delivered in a liquid aerosol to the subject with a DeVilbiss 646 nebulizer (DeVilbiss Healthcare GmbH; Mannheim. Germany), Pulmoaid compressor (DeVilbiss Healthcare GmbH; Mannheim, Germany), and SPIRA dosimeter (Respiratory Care Centre; Hämeenlinna, Finland). An electronic flow meter and a metronome were used to target a defined, height-adjusted inspiratory flow rate (flow in ml/s = 3 x height in cm, rounded to nearest 100 ml/s value) and breathing pattern (1 second inspiration; 1 second expiration). Serial 2-minute images were then obtained for 64 minutes. The participant was then asked to forcefully cough through a peak flow device 10 times, every 10 minutes for an additional 30 minutes (30 total coughs) to assess cough-assisted clearance (CC). An additional static image at 6 hours after inhaling radioisotope was collected to assess particle retention at this time point.

2.3. Analysis of regional MCC and CC

The retention of radioactive counts in each lung ROI (whole lung, WL; central lung, CL; peripheral lung, PL) at sequential time points was corrected for isotopic decay and background radiation and then expressed as a fraction of the initial counts. Percent clearance at designated 10-minute intervals throughout the 94-minute period was determined using the sum of two 2-minute images that followed each 10-minute interval. Cough-assisted clearance between 64 and 94 minutes of the MCC scan was described by linear slopes of each clearance curve after normalizing the initial timepoint to account for the varying amount of tracer available to be cleared during this clearance interval. All image analyses were performed at the University of North Carolina.

2.4. Sputum Induction

A standard operating procedure developed by the CF Therapeutic Development Network (TDN) was followed for sputum induction, as previously described[10]. Induced sputum specimens were immediately placed on ice and then shipped cold (not frozen) overnight for central lab analyses at the University of Alabama at Birmingham (UAB).

2.5. Percent solids measurement

Measurements were performed as previously described[11]. Up to 3 technical replicates of 10 μL sputum aliquots were weighed using a Mettler Toledo UMX2 Ultra-microbalance to determine wet weight. Samples were left overnight in a dry oven for liquid evaporation and weighed again for dry weight. Wet and dry weights were used to calculate the solid content fraction; technical replicates were averaged for a single point estimate in statistical analyses. All measurements were conducted centrally (UAB) by operators blinded to visit number.

2.6. Sputum cone and plate rheology

Cone and plate rheology measurements were based upon published methods[12, 13]. On arrival samples were gently vortexed and divided into three 250 μL aliquots using a positive displacement pipet and maintained at 4°C. Each aliquot was then incubated at 37°C for 30 minutes before transferring 200 μL to a cone and plate rheometer (TA Instruments Discovery Series HRII Rheometer) to measure its viscoelastic properties. Oscillatory rheometry was used to measure storage (elastic) modulus (G’), loss (viscous) modulus (G”) and dynamic viscosity (η). After a 10-minute conditioning period, a flow ramp procedure began with an initial stress of 1.0 X 10⁻³ increased to 10 Pa in log mode over 600 seconds. Five points were observed per decade with oversampling of controlled stress. An oscillation frequency procedure immediately followed with 5% strain, a logarithmic sweep from 0.05 to 20 Hz, 10 points observed per decade, continuous direct controlled strain and 3-second conditioning time with 3-second sampling time data acquisition parameters. The elasticity values were recorded from the oscillation frequency procedure. A flow ramp procedure then followed with a logarithmic sweep from 0.02 to 1000 s⁻¹, recording 5 points per decade, followed by a 5-second equilibration period and a 10-second averaging time. The controlled rate was set by automated motor mode. The viscosity values were recorded from the flow sweep procedure. Rheologic values measured nearest the 1Hz frequency were analyzed using up to 3 replicates from each sample. All measurements were conducted centrally (UAB) by operators blinded to visit number.

2.7. Exhaled Breath Condensate Collection

The RTube system (Respiratory Research Inc, Charlottesville, VA) was utilized to non-invasively collect EBC samples according to manufacturer instructions using a 10-minute collection period. The condensed liquid was divided into 500 μl aliquots and stored at −80degC. Frozen samples were shipped as a single batch and analyzed centrally at the University of North Carolina (UNC).

2.8. EBC pH and sialic acid:urea ratio measurement

EBC pH was measured after sample de-aeration by bubbling with argon for 5 minutes as previously described [14], with samples maintained on ice during measurements.

The EBC sialic acid:urea ratio was analyzed by mass spectrometry as previously described [15]. Briefly, a stable isotope internal standard was added to EBC samples, which were then treated with 1% formic acid and incubated at 80°C for two hours to release terminal sialic acid from mucins. The samples were lyophilized and reconstituted in 15-fold less volume of water, then sialic acid, urea, and internal standards measured via liquid chromatography-tandem mass spectroscopy.

2.9. Statistical Analysis

The prespecified primary MCC endpoint was the average rate of whole right lung MCC over 60 minutes (AveClr60 WL). Secondary endpoints included average rates of MCC from defined regions of interest (i.e., central and peripheral lung ROI) and other time domains (i.e., 90 minutes, 6 hours). Descriptors of the initial isotope deposition pattern (skew; central/peripheral ratio, C/P) were assessed as covariates that impact MCC rates. Paired t-tests of MCC and clinical endpoints were performed as primary statistical tests. Linear mixed models were also constructed to determine if changes in isotope deposition descriptors impacted observed changes in MCC with treatment.

The primary respiratory biospecimen endpoint was the % solids content. Secondary biospecimen endpoints included sputum macrorheology indices, EBC sialic acid/urea ratio, and EBC pH. Sputum specimen data were log transformed before analysis using a repeated measures model (GEE) that incorporated all available data. Unless otherwise indicated, data were back transformed to the original data scale for presentation in tables. A secondary analysis using a paired t-test was performed to examine treatment effects only in those able to produce sputum at both timepoints. Paired EBC sample data were analyzed with the Wilcoxon Signed Rank test. Spearman’s ρ correlation coefficients were calculated to assess the association between selected endpoints of interest.

3. RESULTS

3.1. Baseline demographics

46 subjects were consented and 45 initiated E/T/I therapy. Baseline characteristics are described in Table 1. Approximately half of the participants were using a 2-drug CFTR modulator (lumacaftor/ivacaftor or tezacaftor/ivacaftor) prior to enrollment; the remainder were naïve to CFTR modulator therapy. The median time between E/T/I initiation and the post-treatment study visit was 32.5 days (IQR 27–50.8); median time between the baseline and post-treatment study visits was 38.0 days (IQR 31–53.8).

Table 1:

Baseline demographics

Age, years (range)	27.7 (12 – 54)
Gender
Female (%)	29 (63%)
Male (%)	17 (37%)
FEV1%pred (range)	78.2 (27– 122)
% with FEV1 <70% pred	41%
% with FEV1 <50% pred	16%
BMI (range)	22.4 (16.4 – 29.7)
Genotype:
F508del/F508del	28 (61%)
F508del/minimal function mutation	16 (35%)
F508del/other	2 (4%)
Prior Modulator Use
Lumacaftor/ivacaftor	5 (11%)
Tezacaftor/ivacaftor	19 (41%)
None	22 (48%)
Sweat Cl (mmol)	8 6.0±14.3
CFQ-R -respiratory domain score	68.9±17.4
* Pseudomonas infection category (n=42)
Persistent	45%
Intermittent	7%
Pseudomonas-free	48%
Cycling or continuous inhaled antibiotic	47.8%
Chronic azithromycin	56.5%
Chronic hypertonic saline	76.1%
Chronic dornase alfa	87.0%

Mean, SD, % and/or range, as indicated.

^*Categorization of Pseudomonas infection status within the year prior to baseline. There must have been at least 1 quarter with culture(s) done within year prior to baseline to be non-missing.

3.2. Clinical treatment effects

Notable improvement in lung function, sweat chloride, symptoms (CFQ-R resp domain), and nutrition were observed over the 1-month treatment period (Table 2). The magnitude of these changes is consistent with those observed in the larger core PROMISE cohort[3].

Table 2:

E/T/I effect on clinical endpoints

FVC%pred change (N=40)	7.3 (5.0 – 9.6); p < 0.0001
FEV1%pred change (N=40)	8.9 (6.6 – 11.3); p < 0.0001
FEF25–75%pred change (N=40)	13.4 (7.7 – 19.1); p < 0.0001
BMI change (N=40)	0.41 (0.18 – 0.64); p = 0.001
Weight, kg (N=40)	1.2 (0.57 – 1.88); p = 0.0005
CFQ-R resp domain change (N=23)	14.5 (5.7 – 23.3); p = 0.0024
Sweat chloride change, mmol/L (N=35)	−39.6 (−34.6 – −44.6); p < 0.0001

Mean, 95%CI and p value from paired t-test shown.

3.3. Effect of E/T/I on MCC

At baseline, WL MCC rates were significantly associated with aerosol deposition indices (C/P ratio, skew) but not with any clinical characteristics (e.g., age, gender, FEV₁pp, genotype group, prior modulator use) or mucus descriptors (e.g., percent solids, G’, G”). Large improvements in MCC in all lung compartments (WL, CL, PL) were observed following treatment with E/T/I (Fig 1; Supplemental Table 1). A similar absolute change in MCC was noted in all lung compartments, including the peripheral lung where MCC deficits are most notable in CF[16]. Although indices of initial isotope deposition (skew, C/P ratio) suggested a slightly more heterogeneous and central deposition pattern after E/T/I treatment, these differences were not statistically significant (Supplemental Table 1). Incorporation of skew or C/P ratio as covariates into repeated measures models to assess the impact of E/T/I treatment on MCC endpoints did not meaningfully alter the magnitude or significance of the effect of E/T/I treatment on MCC endpoints. The effect of E/T/I on MCC was independent of the participant’s genotype group (homozygote vs. heterozygote for F508del), gender, age, baseline lung function, chronic Pseudomonas aeruginosa infection status, and whether they previously were treated with a CFTR modulator. The change in MCC (WLAveClr60) after E/T/I did not significantly correlate with the associated change in FEV₁%pred (Spearman’s ρ = 0.05; p = 0.75), CFQ-R (Spearman’s ρ = 0.22; p = 0.30 or sweat chloride (Spearman’s ρ = 0.20; p = 0.26), nor with changes in sputum or EBC endpoints.

Figure 1: Whole lung mucociliary clearance.

A. Mean +/− SEM of whole lung (WL) clearance over time before and after elexacaftor/tezacaftor/ivacaftor treatment. B. Waterfall plot of individual changes in whole lung AveClr60 (%). C. Central lung (CL) clearance. D. Peripheral lung (PL) clearance.

3.4. Effect of E/T/I on Respiratory Biospecimens

Sputum induction and EBC collections were used to assess the characteristics of respiratory biospecimens before and after E/T/I. Fewer participants were able to produce sputum after E/T/I initiation (N=43 vs 23) despite sputum induction. When compared to persistent sputum producers, those unable to produce sputum after E/T/I had significantly better baseline lung function (FEV1%pred 89.0 vs. 67.5%; p = 0.0015) and were less likely to have been prescribed hypertonic saline (65% vs. 87%; p = 0.08) or azithromycin (43% vs. 70%; p= 0.07) at baseline. EBC was successfully collected from participants at each visit.

3.4.1. Sputum percent solids

The percent solids content of sputum collected after initiation of E/T/I was significantly reduced from baseline (modeled mean 3.4 vs 2.2%; p<0.0001). In a post hoc sensitivity analysis restricted to those with paired specimens, a statistically significant change in percent solids was again observed (Table 3, Figure 2). p<0.0001). Those who were unable to produce sputum after E/T/I initiation had a somewhat lower baseline % solids value (3.6 vs. 4.7%; p=0.18).

Table 3:

Biospecimen endpoints

Sputum analyses	Baseline (n=43)	Post-E/T/I (n=23)	P value
Percent solids content	3.4 (2.8 – 4.2)	2.2 (1.9 – 2.6)	<0.0001
G’ (elastic modulus) [Pa]	5.9 (2.9 – 12.1)	2.8 (1.6 – 5.1)	0.123
G” (viscous modulus) [Pa]	1.7 (0.8 – 3.9)	0.6 (0.3 – 1.1)	0.054
Dynamic viscosity (η)	1.2 (0.7 – 1.9)	1.0 (0.5 – 1.7)	0.469
EBC analyses	Baseline (n=45)	Post-E/T/I (n=41)	P value
EBC SA:urea	1.31 (0.86 – 1.92)	1.14 (0.77 – 1.65)	0.47
EBC pH	7.68 (7.09 – 8.01)	7.68 (7.06 – 7.99)	0.91

For sputum analyses, modeled mean, 95%CI and p values from repeated measures model of log-transformed data are presented. Data were converted back to original scale for display in table. Pa = Pascals; η = P•Sec. For EBC analyses, nedian (IQR) values are presented; p values from Wilcoxon Signed Rank test on paired values are presented.

Figure 2: Sputum % solids.

A. Scatter plot of % solids in paired samples before and after elexacaftor/tezacaftor/ivacaftor (E/T/I); p < 0.0001. B. Waterfall plot of the change in % solids after treatment with E/T/I.

3.4.2. Sputum rheology

Sputum rheology (G’, G”, dynamic viscosity) endpoints were quite variable, and no significant differences were observed (Table 3; Supplemental Figure S1). Trends toward reduced G’ and G”, but not dynamic viscosity, were observed. In a secondary analysis using only paired samples, no differences were observed (G’, p=0.71; G”, p=0.59). However, significant correlations between % solids and each rheology endpoint (Spearman’s ρ 0.37–0.60; each p ≤ 0.0022) were noted, reflecting the biophysical relationship between mucus hydration and rheologic properties at the individual specimen level.

3.4.3. EBC mass spectrometry

Mucin molecules are decorated with carbohydrates, including sialic acid (SA). Measurement of sialic acid, normalized to urea as an estimate of dilution, has been shown to be an index of airway mucus hydration in animal and human studies [8, 17]. Therefore, we hypothesized that EBC sialic acid:urea could provide a non-invasive mucus hydration endpoint that would respond to E/T/I treatment. At baseline, sialic acid:urea varied widely and within subject correlation was poor (Spearman ρ = −0.10; p = 0.56). No difference was observed between baseline and post-E/T/I measurements (Table 3; Supplement Figure S2). Notably, EBC SA:urea values also did not correlate with sputum % solids in samples collected at the same visit (Spearman ρ = 0.06; p = 0.64).

3.4.4. Exhaled breath condensate pH

We hypothesized that restoration of CFTR function, including bicarbonate transport, would yield a measurable alkalinization of EBC pH. EBC pH values were highly variable at baseline and within subject values were poorly correlated (Spearman’s ρ = 0.07; p 0.68). No change was noted after E/T/I treatment (Table 3; Supplement Figure S2).

3.4.5. Biospecimen Endpoint Correlations

We explored whether biospecimen (sputum, EBC) endpoints were related to clinical endpoints (FEV₁%pred, CFQ-R Respiratory Domain) or other clinical descriptors (genotype group, prior modulator use, pseudomonas status). At baseline, a moderate correlation between baseline % solids and CFQ-R respiratory domain score (r = −0.53, p = 0.001) was observed, while the correlation between the changes in %solids and CFQ-R was not statistically significant (r = −0.42; p = 0.13). All other correlations between biospecimen endpoints and clinical variables were weak (r < 0.4) or non-existent.

4. DISCUSSION

PwCF treated with E/T/I experience greatly improved respiratory symptoms, lung function, and fewer pulmonary exacerbations. Participants in this study demonstrated improvements in symptoms and lung function that are quantitatively similar to prior reports[1–3]. The improvement in MCC after E/T/I was also similar in magnitude to that observed after ivacaftor in individuals with the G551D-CFTR mutation[6, 18], where significant clinical benefits were achieved. In contrast, no improvement in MCC was noted after lumacaftor/ivacaftor[7] despite the well demonstrated, though modest, clinical improvements associated with this treatment[19]. Together, these observations suggest that improvements in MCC might not be observed until a minimum threshold of CFTR restoration has been achieved, but that significant clinical benefits would be expected with any novel CFTR-restoring therapy able to measurably improve MCC. As a result, measurement of MCC may be valuable in early trials of novel CFTR-restoring therapies, including genetic therapies, where study populations may be limited in size, making it more difficult to demonstrate lung function improvement and protection against pulmonary exacerbations. Similarly, MCC may also be a useful tool to characterize other pulmonary therapies that target the mucociliary clearance apparatus.

The link between CFTR function and mucus abnormalities has been speculated upon based on in vitro and animal studies but has never been clearly demonstrated in pwCF. Here we examined markers of mucus viscoelasticity and hydration/concentration at specialized centers where these measurements were possible. Indeed, E/T/I therapy was associated with a substantial reduction in mucus concentration (% solids) and improvement in MCC. These findings substantiate the hypothesized link between CFTR function, mucus hydration, and MCC while also providing a likely mechanism for the observed improvements in lung function, symptoms, and exacerbation frequency. Rheologic endpoints, however, despite significant correlation with % solids measurements were unable to demonstrate significant improvement due to higher measurement variability. Post-hoc estimates of the sample size needed to demonstrate a significant difference in G’ and G” using data from this study with a paired t-test were 85 (G”) and 121(G’) subjects. Data from EBC pH and SA/urea measurements, on the other hand, suggest limited utility as endpoints for CFTR modulation studies, either because of excessive variability or minor relevance to the restoration of CFTR function.

Measurement of sputum % solids is technically easy to perform, requires small sample volumes, is reasonably insensitive to storage (e.g., freeze/thaw) and shipping conditions. Sputum % solids measurements reflect respiratory secretion hydration, and therefore may be useful as a tool to screen for treatment effects of other CFTR restoring or mucus hydrating therapies. Of note, sputum % solids also correlated with respiratory symptoms (CFQ-R), adding further evidence for the relevance of this endpoint. Successful utilization of sputum macrorheology in future trials clinical trials will likely require that we overcome variability issues. Changes within the mucin network occur progressively over time post-collection and with freeze-thaw cycles, accentuating intrinsic sampling variability issues related to use of sputum. Rapid measurement on fresh samples at individual sites could potentially reduce variability and improve the performance of this assay. Emerging technologies that allow simple, rapid rheology measurements could overcome the current need for considerable expertise to make accurate measurement with traditional cone and plate instruments[20]. Importantly, mucus endpoints require the ability to expectorate sputum (or invasive sampling). This can be problematic after CFTR function has been restored, particularly in those with milder lung disease at baseline, as our data demonstrates. Sputum induction may only partially solve this shortcoming, and 52% of our subjects were unable to produce an adequate sample after E/T/I treatment. Sputum induction also imparts a dilution effect on analytes, that is estimated at 15–20% in a prior study of % solids and mucin concentration measurements[21].

Identification of a biomarker that could predict improvement in clinical endpoints (e.g., FEV₁, exacerbation rate) in an individual would be of significant value. As often observed with other in vivo biomarkers, intrasubject correlations between MCC and clinical endpoints (FEV₁, sweat chloride) were not observed in this study. This does not lessen the importance as a tool for drug discovery if, like sweat chloride, MCC helps to accurately discriminate between treatment groups to the same degree as changes in clinical endpoints, but with smaller sample sizes. The results we report here suggest that MCC and sputum % solids, if validated in other cohorts, may prove to be very useful biomarkers for the study of other CFTR restoring therapies directed to lung (mRNA, gene therapy, gene editing, oligonucleotide splice switching) where extrapulmonary markers of CFTR activity (i.e., sweat chloride) will not be useful.

Highlights.

• Elexacaftor/tezacaftor/ivacaftor increases sputum hydration in cystic fibrosis.

• Elexacaftor/tezacaftor/ivacaftor (E/T/I) markedly improves mucociliary clearance.

• Exhaled breath condensate pH and sialic acid levels were unchanged after E/T/I.