If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Cystic Fibrosis Research Laboratory, Stanford University, Stanford, CA, USADepartment of Pediatrics, Stanford University School of Medicine, Stanford, CA, USA
Corresponding author at: Center for Excellence in Pulmonary Biology, Stanford University Medical School, 770 Welch Rd Suite 380, Palo Alto, CA 94304, USA.
To determine in vivo effects of CFTR modulators on mutation S945L.
Methods
We measured effects of CFTR modulators on CFTR-dependent sweating (‘C-sweat’) in two pancreatic sufficient cystic fibrosis (CF) subjects. S1 (S945L/G542X) took ivacaftor and S2 (S945L/F508del) took ivacaftor + tezacaftor. Sweating was stimulated pharmacologically to produce sequentially both CFTR-independent (methacholine stimulated) M-sweat and C-sweat; and the ratio of these was compared. Sweat secretion was measured with two methods: real time secretory rate quantitative recording and by optically measuring the growth of sweat bubbles under oil from multiple identified glands.
Results
Using the quantitative recorder, we saw zero C-sweat secretion off-drug, but when on-drug the C-sweat responses for both subjects were comparable to those seen in carriers. The on-drug response was further quantified using the sweat bubble method. Each subject again showed robust C-sweat responses, with C-sweat/M-sweat ratios ~ half of the ratio determined for a cohort of 40 controls tested under identical conditions.
Conclusion
These in vivo results, consistent with prior in vitro findings, indicate that the drug treatments restore near-normal function to S945L-CFTR, and support the use of ivacaftor as a treatment for CF patients who carry this allele.
Cystic fibrosis (CF) is caused by mutations in the gene for CFTR, an anion channel. Certain mutations reduce the average open probability (PO) and hence the flow of anions through the channel. For secretory epithelia that depend upon anion flow through CFTR to mediate fluid secretion, this results in reduced or lost secretion, which underlies most of CF pathophysiology. Ivacaftor in vitro increases PO in CFTR channels affected by a wide variety of gating mutations [
Analysis of the 27 exons and flanking regions of the cystic fibrosis gene: 40 different mutations account for 91.2% of the mutant alleles in southern France.
]. The reduced function often results in pancreatic sufficient CF when paired with another mutation (61% of subjects with mean age of 22, CFTR2). Van Goor and colleagues [
] measured ivacaftor effects on CFTR function electrophysiologically using a panel of Fischer rat thyroid (FRT) cells expressing 54 missense mutations and found that ivacaftor potentiated multiple mutant forms of CFTR, including S945L. The effect on S945L function was striking: cells expressing it and stimulated with 10 μM forskolin displayed 6.0 ± 1.9% of Wild Type (WT) CFTR function before ivacaftor treatment and 75.8 ± 23.3% after ivacaftor treatment, with an EC50 of 181 ± 36 nM. Informed by these results, we were interested in evaluating CFTR function in vivo in response to ivacaftor in CF subjects carrying S945L. Since the original submission of this manuscript, in the US the approved use of Kalydeco was extended in May of 2017 to the treatment of subjects carrying S945L.
2. Materials and methods
2.1 In vivo CFTR function assays
We used two independent methodologies to quantify the rates of CFTR-dependent sweat secretion (C-sweat) which, when expressed as a ratio of their CFTR-independent sweating (M-sweat, thus C/M ratio), provides a near-linear readout of CFTR function (Fig. 1). Methacholine acts on muscarinic receptors in the sweat gland to elevate cytosolic calcium and activate calcium-activated chloride channels and basolateral potassium channels, which remain patent in CF subjects. This allows for the M-sweat to serve as a metric against which C-sweat can be compared. As we and others have reported, the C/M-sweat ratio gives a near-linear readout of CFTR function over a wide range, with the ratio for CF subjects being zero, while subjects carrying one CFTR mutation (hereafter ‘carriers’) have, on average, C/M ratios that are one half those of healthy controls (hereafter ‘controls’) [
Fig. 1Typical Sweat rate measurements with two systems. A-C illustrate responses of the Q-sweat metohd employed in the CFTR-sweat detection paradigm, showing CFTR independent responses (white arrows) from a control (A), a CF carrier for F508del (B), and a CF subject homozygote for F508del (C). A CFTR-dependent response (asterisk) follows intradermal injection of a β-adrenergic cocktail (blue arrows) in (A). This response is reduced by about half in carriers (B) and is absent in CF subjects (C). (D-F) CFTR-dependent sweat bubbles imaged under an oil layer with a method that allows individual, identified glands to be followed across experiments. Responses are for control (D), carrier of F508del (E) and CF subject heterozygote for F508del/R117H-5 T (F). (G) CFTR-independent M-sweat bubbles from the same CF subject and area shown in F. Scale bar = 1 mm.
The Q-Sweat quantitative sweat measurement system (WR Electronics Medical Co., Maplewood, MN, USA) is a sealed-capsule device that uses room air drawn across a desiccant to collect moisture from the skin surface. This moisture is transported by airflow to temperature and humidity sensors to determine secretory rates [
]. The device is capable of highly sensitive, real-time direct measurement of secretory rates with immediate results given in nL/min/cm2, and robust normative data has been established for secretory rates in response to cholinergic stimulation [
]. The Stanford CF Research Team developed a testing protocol that takes advantage of the sensitivity of the Q-Sweat device to low secretory rates. To conduct measurements an iontophoresis delivery device and a sealed capsule were tightly placed against the volar surface of the forearm. After obtaining a background recording, cholinergic stimulation was performed by iontophoresis of 1% acetylcholine for 5 min monitoring the secretion rate in real time and recording the peak secretory rate. This was then followed by a single subcutaneous injection under the measurement capsule, of 0.20 mL β-adrenergic cocktail solution containing 140 μM atropine to block cholinergic activity, 10 mM aminophylline to block phosphodiesterase activity, and 80 μM isoproterenol for β-adrenergic stimulation. Secretion was again monitored in real time for an additional 10 min, or until a steady plateau was reached, making note of the peak secretory rate.
2.1.2 Ratiometric measurement of sweat secretion from identified individual glands
The ‘bubble test’ optically tracks the growth of sweat bubbles produced by individually identified sweat glands as they are formed within an oil layer on the skin [
]. In brief, a region on the volar forearm was sequentially injected intradermally with 0.05 ml of a 1 μM solution of methacholine to stimulate CFTR-independent sweating (M-sweat) and then with 0.1 mL of a cocktail of 160 μM isoproterenol, 20 mM aminophylline, and 280 μM atropine in lactated Ringer's to block M-sweating and produce C-sweating. M-sweat was monitored for 10 min and C-sweating for 30 min. Sweat bubbles from single glands were captured in an oil layer, visualized by oblique lighting or dye-partitioning, and digitally imaged. Individual glands were identified by location. For each gland average M- and C-sweat rates were calculated by dividing the final sweat volumes by 10 or 30 min respectively.
2.2 Subjects
Subject 1 (S1) with CFTR genotype S945L/G542X, is a 43 year old male who is pancreatic sufficient, infertile, has FEV1 values 66% predicted, bronchiectasis, chronic infection with mucoid Pseudomonas aeruginosa and Achromobacter xylosoxidans, and sweat chloride of 86 mmol/L. S1 was considered to be a candidate for treatment with ivacaftor because of the evidence stated above that S945L responded well to ivacaftor when expressed in FRT cells [
Subject 2 (S2) with CFTR genotype S945L/F508del, is a 39 year old male who is pancreatic sufficient, infertile, has FEV1 values 62% predicted, bronchiectasis, chronic infection with mucoid Pseudomonas aeruginosa and intermittent infection with Staphylococcus aureus, and sweat chloride of 102 mmol/L. S2 was expected to respond to ivacaftor as well, but because his other mutation was F508del, additional benefit might be expected by using the CFTR corrector tezacaftor (VX-661), which acts in a manner similar to lumacaftor (VX-809) [
]. Therefore, S2 was enrolled in a clinical study of the drug combination tezacaftor + ivacaftor (ClinicalTrials.gov NCT02565914) and participated in our study during the open label treatment phase of the clinical trial.
2.3 Controls and modulator dosing
For Q-sweat measurements, subjects' off-drug responses were compared with their on-drug responses and with typical responses from control, carrier and other PI CF subjects. For the sweat bubble measurements, subjects' on-drug responses were compared with the average response of a cohort of 40 adult subjects recruited for a separate study of variations of CFTR-mediated sweating among controls (manuscript in preparation). S1 was on ivacaftor for 3 weeks prior to the on drug sweat rate quantitative rate test and for the off drug test was not taking ivacaftor for 3 weeks. For bubble tests he was on ivacaftor for the preceding 6 months. S2 was off ivacaftor/tezacaftor for 27 days prior to the off drug test and was on ivacaftor/tezacaftor for 7 months prior to the on drug sweat rate tests. There were no adverse events. Adherence was assessed by self-report, with both subjects reporting consistent intake of the medications.
2.4 Reagents
Methacholine Chloride, (Methapharm, Ontario, Canada), Isoproterenol HCl, Aminophylline, lactated Ringer's (Hospira, Lake Forest, IL) and Atropine Sulfate (American Reagent) were obtained from Stanford University Hospital Pharmacy. Heavy mineral oil was from EMD Chemicals, Gibbstown, NJ, and was water-saturated before use as previously described [
]. Erioglaucine disodium salt (CAS No. 3844-45-9) was from Sigma.
2.5 Statistics
We used a within-subjects study design for the Q-sweat data, and a between subjects design for sweat bubble results, where the magnitude of on-drug C/M ratios were compared with the mean of 40 control subjects. Pearson's r was computed for linear regressions of single gland C-sweat rates on M-sweat rates or C/M ratios on M-sweat rates and regression was done with Excel (Microsoft Inc., Redmond WA) and the statistical significance determined with an online p-value calculator (http://www.socscistatistics.com/pvalues/pearsondistribution.aspx) for correlation coefficients. A p value < 0.01 was taken as significant.
2.6 Study approval
The study was approved by the Institutional Review Board of Stanford University. After written informed consent, subjects were studied with the sweat secretion quantitative measurement system and/or the sweat bubble measurement system.
3. Results
Off-drug CFTR-dependent Q-sweat responses for both subjects were zero (Fig. 2A, E) while on-drug responses were similar to responses observed in carriers (Fig. 2B, F), therefore statistical comparisons are not meaningful. For sweat rate bubble measurements, the average on-drug C/M ratio for S1 was 0.09 ± 0.04, n = 59 glands, and for S2 was 0.12 ± 0.03, n = 45 glands. For 40 controls, the mean C/M ratio was 0.19 ± 0.07, range 0.06–0.36. The C/M ratios of the two S945L subjects on-drug were approximately half of the WT mean, and within the lower bounds of the WT distribution which overlaps the carrier range [
]. To place these findings in full context, we present in Fig. 1 typical sweat secretory responses. Fig. 1, A–C shows Q-sweat tracings for a control (A), a CF F508del carrier (B), and a CF subject homozygous for F508del (C). In each trace, a large secretory response in the order of 300–400 nL/min/cm2 follows the iontophoresis of 1% acetylcholine (white arrows). Subsequent intradermal injection of the β-adrenergic cocktail (blue arrows) blocked the cholinergic, CFTR-independent response and initiated CFTR-dependent C-sweating (A). C-sweating was about half as large (105 vs 202 nL/min/cm2) in the carrier (B), and zero C-sweating occurred in the CF subject (C). Fig. 1, D-F shows corresponding C-sweat bubble responses for a control (D), a CF carrier of F508del (E), and a CF subject carrying F508del/R117H-5T (F), each stimulated after an initial cholinergic response (not shown). Fig. 1G shows methacholine-stimulated sweat bubbles from the same area of the CF subject shown in 1F.
3.1 Subject S1, S945L/G542X
Fig. 2A shows Q-sweat tracings for S1 off ivacaftor. Following iontophoresis of 1% acetylcholine (started at white arrow) we observed a large secretory response with a peak rate of 402 nL/min/cm2. Upon injection of the β-adrenergic cocktail (blue arrow) with blockade of the cholinergic activity, unlike controls it failed to induce any increase in C-sweating as indicated by the flat Q-sweat trace (asterisk) following the injection. Fig. 2B shows Q-sweat tracings for S1 after starting ivacaftor. The cholinergic response was again large at 337 nL/min/cm2, but now the β-adrenergic cocktail produced a clear increase in CFTR-dependent sweating of 108 nL/min/cm2 (asterisk). Fig. 2, C and D show sweat bubble results for S1 while on ivacaftor. Fig. 2C shows the response of S1 to the cholinergic stimulus and 2D to the β-adrenergic stimulus. Each of the identified glands that responded to the cholinergic stimulus also responded to the β-adrenergic stimulus. (For example, corresponding M- and C-sweat bubbles from 9 identified glands were connected with dashed yellow polygons.)
Fig. 2(A-D) Q-Sweat and bubble responses for S1, S945L/G542X. (A) Responses off-drug, same convention as in Fig. 1. Note absence of a response to β-adrenergic cocktail (*). (B) Responses on ivacaftor, note presence of a response to β-adrenergic cocktail (*). (C) While on ivacaftor CFTR-independent M-sweating induced by an injection of methacholine. (D) Also while on ivacaftor CFTR-dependent C-sweating following a β-adrenergic cocktail. Dashed outline connects bubbles produced by the same 9 identified glands. (E-H) Q-Sweat and bubble responses for S2, S945L/F508del. (E) Responses off-drug, same convention as in A-E. (F) Responses on ivacaftor/tezacaftor, same convention as Fig. 1, Fig. 2. (G-H) Bubble responses on ivacaftor/tezacaftor. (G) CFTR-independent M-sweating induced by intradermal injection of methacholine. (H) CFTR-dependent C-sweating following intradermal injection of β-adrenergic cocktail. Sweat bubbles are numbered to indicate glands that produced them; several examples of merged bubbles can be seen. Scale bar = 1 mm.
Fig. 2, E-H show data for S2. Q-sweat tracings for S2 off-drug (2E) and on-drug (tezacaftor/ivacaftor, 2F). As for S1, no CFTR-dependent sweating was observed off-drug, but after starting tezacaftor/ivacaftor the β-adrenergic cocktail stimulated a secretory response of 201 nL/min/cm2 that corresponded to 50% of his cholinergic response. Fig. 2G shows cholinergically-induced sweat bubbles for S2, and 2H shows CFTR-dependent C-sweat bubbles, both while S2 was on tezacaftor/ivacaftor. The observed level of C-sweating is qualitatively similar to levels of C-sweating seen in carriers. It is quantified in the next section.
3.3 Gland-by-gland quantification of CFTR-dependent sweating on ivacaftor
Fig. 3 plots on-drug sweat bubble responses for the two S945L subjects. Fig. 3, A, B shows C-sweating vs. M-sweating. They were highly correlated as in controls, (S1: r = 0.86, n = 59, P < 0.00001; S2: r = 0.91, n = 91, P < 0.00001) with every gland producing measureable C-sweat. Note the large differences in sweat rates within and across subjects (reflecting gland size). Average M-sweat rates (nL·min−1·gl−1) were S1: 3.67 ± 0.22 and S2: 12.32 ± 0.69. Because of these large differences in sweat gland secretory capacities, meaningful comparisons of C-sweat rates among subjects require that they be expressed as a ratio of the M-sweat results [
Fig. 3Gland-by-gland quantification of sweat secretion on-drug for each subject. (A and B) Plots of C-sweat rate (y-axis) as a function of M-sweat rate (x-axis). Each point represents a single gland, and the slopes of the regression lines are significant (P < 0.00001). (C and D) Plots of the C-sweat/M-sweat ratios vs M-sweat rates for each of the identified glands. Each point represents a single gland. Regression line slopes have flattened. For S1 the slope is not significantly different from zero. For S2, the slope is modestly positive P = 0.01. (A, C) Subject 1, S945L/G542X on ivacaftor. (B, D) Subject 2, S945L/F508del on ivacaftor/tezacaftor. Healthy control mean (n = 40 subjects) is horizontal dashed line with error bar and shading showing +/−SD.
Fig. 3, C, D plot gland-by-gland C/M ratios vs. M-sweat rates for the two S945L subjects while on CFTR modulator treatment. The C/M ratio provides a near-linear readout of CFTR function [
]. Ratios were computed on a gland-by-gland basis, arrayed as a function of M-sweat rates for the glands, and then compared to results from controls to estimate the percentage of CFTR function in controls. The regression of C/M vs M approached a zero slope for S1 (r = 0.14, P = 0.29), and was weakly positive for S2 (r = 0.38, P = 0.01). The critical values are the mean C/M ratios while on CFTR modulator therapy, which were 0.09 ± 0.01 for S1 and 0.12 ± 0.004 for S2. These values were compared with the average results from an ongoing study of a cohort of 40 controls tested under identical conditions to those used here, which found an average C/M ratio of 0.19 ± 0.07 (Kim, Wine, and others, unpublished data). The C/M values for these two subjects are thus 47% and 63% of the average WT, which is the expected mid-range for carriers; [
] i.e. if the drug had acted on two S945L alleles the predicted values would have been 94–126% of control mean. This indicates that ivacaftor, with or without tezacaftor, restored essentially normal function to the S945L allele.
4. Discussion
We report for the first time in vivo CFTR function responses to modulators on two subjects carrying S945L. The responses observed are of a magnitude about half of those observed in controls and resemble the responses seen in carriers. We have never previously observed this magnitude of C-sweat in CF subjects not treated with CFTR modulators.
A limitation of this study is that no sweat bubble testing was done off-drug. However, the quantitative sweat secretion method did not detect C-sweating off-drug, and even though bubble testing is somewhat more sensitive than conventional evaporimetry [
], in vitro studies of S945L-CFTR suggest that only trace C-sweating would have been seen. S945L reduces both n and PO. When expressed in FRT cells, Van Goor and colleagues observed maturation of 42.4 ± 8.9% of WT [
]. In CHO cells, the PO of S945L-CFTR was approximately 1/3 that of WT CFTR. Together the two studies predict n·PO = 4–13% WT. Expression of S945L in FRT cells produced CFTR-mediated currents that were 6.0 ± 1.9% of WT [
], in good agreement with the average predicted n·PO. Given that our subjects have only a single copy of S945L, they would be expected to show approximately 3% WT response off-drug. This is below the detection level of evaporimetry and near the detection level of the bubble test. Although observations made in just two subjects could be seen as an additional limitation, the optical ratiometric assay provides the advantage of assessing multiple individual glands as the units of observation, which in our opinion makes it ideal for studies with small numbers of subjects or even at the individual level. It thus fulfills quite well the requirements of the emerging CF therapeutic paradigms of individualized precision medicine and ‘theraptyping’.
How is ivacaftor affecting S945L-CFTR? In FRT cells, ivacaftor increased Isc 12.5-fold, from 6 to 75% WT. If 3% WT C-sweat response is expected off-drug, then ivacaftor increased responding 13.7 fold, from 3 to 41% WT in S1. However, S945L PO was measured as ~0.12 [
]. If this measure is accurate, then even assuming ivacaftor could increase S945L-CFTR to 1.0, it would allow only an 8-fold increase in S945L function if the effect were entirely on PO. This raises the possibility that ivacaftor might also improve the maturation of S945L-CFTR. S2, who was taking the corrector tezacaftor in addition to ivacaftor, had on-drug function (in this single test) that was approximately 1/3 greater than S1. This could reflect the wide differences we see even among WT subjects, but it is also possible that it derives from restoration of some function to the F508del allele, or a corrective effect of tezacaftor on the maturation of the S945L allele.
In summary, two different methods that measured CFTR-dependent sweat secretion found that ivacaftor ± tezacaftor increased mutant CFTR function to a level roughly half that of controls. Because each subject had one S945L mutation, and because prior work showed a powerful increase of S945L function in vitro [
], we conclude that oral ivacaftor in vivo had a similar powerful increase on S945L. The results raise the possibility that ivacaftor, as well as tezacaftor, may improve the maturation of S945L-CFTR (i.e. increase n) as well as its PO.
Author contributions
CM and JJW designed the experiments. JK, ZD and CD performed the experiments. CM and JJW analyzed and interpreted the data and wrote the manuscript.
Acknowledgements
We are grateful to the subjects who participated in these trials. We thank members of the Cystic Fibrosis Foundation Therapeutics Sweat Consortium for discussions.
Funding to CM from CFFT (MILLA14Y0) and the Ross Mosier Laboratories gift fund, and to JJW from CFFT and CFRI.
References
Van Goor F.
Hadida S.
Grootenhuis P.D.
Burton B.
Cao D.
Neuberger T.
et al.
Rescue of CF airway epithelial cell function in vitro by a CFTR potentiator, VX-770.
Analysis of the 27 exons and flanking regions of the cystic fibrosis gene: 40 different mutations account for 91.2% of the mutant alleles in southern France.
30983-9&id=gr1.jpg){kind=link}
30983-9&id=gr2.jpg){kind=link}
30983-9&id=gr3.jpg){kind=link}