A little CFTR goes a long way: CFTR-dependent sweat secretion from G551D and R117H-5T cystic fibrosis subjects taking ivacaftor

Abstract

To determine if oral dosing with the CFTR-potentiator ivacaftor (VX-770, Kalydeco) improves CFTR-dependent sweating in CF subjects carrying G551D or R117H-5T mutations, we optically measured sweat secretion from 32-143 individually identified glands in each of 8 CF subjects; 6 F508del/G551D, one G551D/R117H-5T, and one I507del/R117H-5T. Two subjects were tested only (-) ivacaftor, 3 only (+) ivacaftor and 3 (+/-) ivacaftor (1-5 tests per condition). The total number of gland measurements was 852 (-) ivacaftor and 906 (+) ivacaftor. A healthy control was tested 4 times (51 glands). For each gland we measured both CFTR-independent (M-sweat) and CFTR-dependent (C-sweat); C-sweat was stimulated with a β-adrenergic cocktail that elevated [cAMP]i while blocking muscarinic receptors. Absent ivacaftor, almost all CF glands produced M-sweat on all tests, but only 1/593 glands produced C-sweat (10 tests, 5 subjects). By contrast, 6/6 subjects (113/342 glands) produced C-sweat in the (+) ivacaftor condition, but with large inter-subject differences; 3-74% of glands responded with C/M sweat ratios 0.04%-2.57% of the average WT ratio of 0.265. Sweat volume losses cause proportionally larger underestimates of CFTR function at lower sweat rates. The losses were reduced by measuring C/M ratios in 12 glands from each subject that had the highest M-sweat rates. Remaining losses were estimated from single channel data and used to correct the C/M ratios, giving estimates of CFTR function (+) ivacaftor = 1.6%-7.7% of the WT average. These estimates are in accord with single channel data and transcript analysis, and suggest that significant clinical benefit can be produced by low levels of CFTR function.

Figure 1. CFTR independent (M-sweat) and CFTR dependent (C-sweat) secretion in G551D CF subjects.

Each image shows a small region (∼1.5 mm²) of the stimulated field of sweat glands (full imaged field is 0.63 cm²) for each subject. Images for S1-3 and S5 were (+) ivacaftor and S6, 7 (−) ivacaftor. A portion of each field was selected to show glands that responded (or not) to the β-adrenergic cocktail (right column). (A-B). S1, male. Arrows in (A) point to air bubbles in the oil. (C-D) S2, female. (D) Shows the single sweat gland that responded to the β-adrenergic stimulus in this experiment. The dark stain is an ink spot on the freckle used for image registration. (E-F) S3, female. This subject had high background staining. (G-H) S5, male. (I-J) S6, female, (-) ivacaftor. This subject showed the most background staining of any subject in this series. (K-L) S7, female, (−) ivacaftor. (O) C-sweat collected in 2 µl constant bore capillary from S1 (+) ivacaftor. Calibration bars  = 0.5 mm.

Figure 2. M- and C-sweat secretion (±) ivacaftor for CF subject 1: male, F508del/G551D.

(A) Larger M-sweat responses (+) ivacaftor were seen (uniquely) in this subject. Scatter plot shows gland-by-gland volumes measured 7.5 min post MCh injection before any appreciable mergers for 46 glands. Each point is the average of 2–3 measures (2.93 measures per gland per condition), log transformed. Value (−) ivacaftor are on the x-axis and (+) ivacaftor on the y axis. Mean volume at the 7.5 min time point was 28.7±20 vs. 38.35±20 nl^−gland, p = 0.004, paired t-test on log transformed data. Diagonal line shows equivalence. (B) Bar graphs showing mean C-sweat volumes ± ivacaftor; 3 tests in each condition; zero C-sweat seen (−) ivacaftor. (C) Gland-by-gland sweat secretion (±) ivacaftor. Each point in these graphs is jointly determined by the mean of M-sweat responses (x axis) and C-sweat/M-sweat ratios expressed as a% of the mean WT C/M ratio (y axis). Top graph with truncated y axis shows measurements (−) ivacaftor. Each point is mean ratio measured in 1–3 tests (mean 2.1 tests) for each of 94 identified glands. C-sweating and hence ratio was zero for all glands. Bottom graph, (+) ivacaftor. 25/66 glands produced C-sweat on 1–3 trials (average 1.56 trials). The dark outlined red diamond and arrow show an outlier gland (G19) which secreted ∼10 times more than the mean of all responding glands on each of 3 (+) ivacaftor trials. Red line is linear fit.

Figure 3. M- and C-sweat secretion (±) ivacaftor for CF subject 2: female, F508del/G551D.

(A) M-sweat responses were unchanged by ivacaftor in this subject. Each point is log transformed values for a single pre-test and mean of 2 post-test measures for each of 87 glands. Mean volumes were 28.2±12.5 vs. 31.9±14.6 nl^−gland, pre and post, n = 87 glands p = 0.12, paired t-test on log transformed data, plotted as in Fig. 2A. (B) Bar graphs show C-sweat volumes (mean ± SEM) for (±) ivacaftor conditions; 2 tests in each condition. One (−) ivacaftor test was at a different site. (The means are based on glands 1–50). (C) Gland-by-gland C-sweat secretion ± ivacaftor. Each symbol shows correlations for each gland between mean M-sweat volumes (x axis) vs. C-sweat/M-sweat ratio expressed as a% of mean WT value (y axis). (Top graph, truncated y axis) shows measurements (−) ivacaftor. Each point is M-sweat volume for a single test for 195 identified glands at two sites; C-sweat was zero. (Bottom graph) shows measurements (+) ivacaftor (mean of 2 tests). Four of 87 glands produced C-sweat, only one gland secreted on both tests (red symbol). Red line is linear fit. This was the smallest response observed for any subject (+) ivacaftor.

Figure 4. C-sweat secretion (+) ivacaftor in CF subject 3: female, F508del/G551D.

S4 was tested only once, (+) ivacaftor. (A) Gland-by-gland correlations of M- and C-sweating using the same conventions as above. M-sweat was measured for 90 glands of which 21 also produced C-sweat with C/M ratios of 0.3% to 5% of the mean WT ratio. (B) Volume vs. time plots for C-sweating of each responding gland. β-adrenergic cocktail injected at time 0, and no observations were made during the first 3 min when atropine was taking effect and when the site was rinsed and dried before reapplying oil. Note slow time course of response.

Figure 5. C-sweat secretion (+) ivacaftor in CF subject 5: male, F508del/G551D.

S5 was tested only once, (+) ivacaftor. (A) Gland-by-gland correlations of M- and C-sweating using the same conventions as previously. M-sweating was measured in 58 glands. Of these, 29 (50%) also produced C-sweat, with C/M ratios of 0.6% to 5.9% of the mean WT ratio. (B) Volume vs. time plots for β-stimulated, C-sweating. β-adrenergic cocktail injected at time 0.

Figure 6. Absence of C-sweat secretion (−) ivacaftor in two G551D CF subjects.

(A) S6: female, F508del/G551D. Each symbol in the main graph shows M-sweat volumes produced by a single sweat gland (x-axis) and C-sweat (truncated y axis, all C-sweat values were zero). The distribution of M-sweat volumes is shown in the inset. (B) S7: female, F508del/G551D. Same format as for S6.

Figure 7. Images of C-sweat responses for CF subject 8: male, I507del/R117H-5T (±) ivacaftor (N of 1 trial).

S8 was tested 3 times (−) ivacaftor and 5 times (+) ivacaftor. Only the first 3 (+) ivacaftor tests are shown. (A-C) (−) ivacaftor; (D-F) (+) ivacaftor. Each image shows the field 30 min after the cocktail injection. A light freckle served as the landmark, with an ink spot on the freckle to improve registration and focusing. For image (F) the ink spot was too high causing the field to shift down. In this pancreatic sufficient subject, a few glands marked as A, B, or arrows) produced small amounts of C-sweat (−) ivacaftor.

Figure 8. Summary data for S8, R117H-5T, (±) ivacaftor.

(A) Mean ± SEM for the final (30 min) volume of C-sweat secretion per gland (50–53 glands measured for each test) (±) ivacaftor. (B) C-sweat/M-sweat ratios expressed as percent of the WT mean. Results are means ± SEM for 49–53 glands per test. (C-D) Gland-by-gland correlations of M-sweat and C-sweat responses in the absence (C), and presence (D) of ivacaftor. Each symbol represents a single gland and shows its average C/M ratio expressed as a% of the WT average (y-axis) vs. its mean M-sweat volume (x-axis). A single gland (G43) in the original ROI produced C-sweat (−) ivacaftor (red symbol); this gland had the second highest C/M ratio in the (+) ivacaftor condition (red symbol). In the (+) ivacaftor condition, 92% of the glands in the ROI produced C-sweat on at least one of the 3 trials, at C/M ratios 0.3%–8.7% of the mean WT C/M ratio. (E) C-sweat volume vs. time for 10 responding glands selected to represent the range of responses for S8 (+) ivacaftor. β-adrenergic cocktail injected at time 0. (F) Absence of ivacaftor effect on M-sweating in this subject. Each point represents a single identified sweat gland, and shows the average M-sweat volumes from 3 tests (−) ivacaftor (x-axis) and 5 tests (+) ivacaftor (y-axis). Dashed 45 degree line indicates equivalence; linear fit does not differ significantly from it.

Figure 9. Summary data for S4, a female CF subject having two potentially responsive mutations: G551D & R117H-5T who was tested 3 times (+) ivacaftor.

(A) Examples of M-sweat. (B) Examples of C-sweat. Calibration  = 0.5 mm. (C) Gland-by-gland correlations of M- and C-sweating using same conventions as previous figures. Symbols: Filled circles, C-sweat on 1/3 trials, blue triangles, 2/3 trials and red diamonds, 3/3 trials. (D) Volume vs. time plots for C-sweating from a single experiment (test 1). β-adrenergic cocktail injected at time 0.

Figure 10. M- and C-sweating in a female, healthy control subject.

(A) M-sweat and C-sweat responses for each of 4 tests (Mean ± SEM, n = 51 glands). Each open bar is the average final M-sweat volume in response to 15 min stimulation with MCh; each blue bar is the average final C-sweat volume in response to 30 min of β-adrenergic cocktail. (B) C-sweat/M-sweat ratios for each test. (C) Gland-by-gland correlation of mean C/M sweat ratios, expressed as a% of the WT average, vs. the mean M-sweat response. Each symbol represents mean values for a single gland based on ∼4 tests. Responses for this subject were at the low end of the control range. (D) Time course of C-sweating in 11 responsive glands selected to cover the range of C-sweat secretion rates.

Figure 11. Summary of CFTR-dependent sweat secretion for six CF subjects (+) ivacaftor.

(A) Columns (left axis) show C-sweat/M-sweat ratios (mean ± SEM) for each subject, expressed as% of mean WT value. Filled red gender signs (right axis) show M-sweat rates for each subject in the presence of ivacaftor (Mean ± SD). Error bars for S1, S4 and S8 are based on 3 tests, S2 on 2 tests. S1-S3 and S5 each has one G551D mutation, S8 has R117H-5T, and S4 has both. (B) Percentage of glands producing C-sweat in each subject (Mean ± S.D). For healthy control and CF heterozygote (carrier) subjects this value is 97–99±3% . (C) C/M ratios (+) ivacaftor at equivalent M-sweat rates. Filled red squares (right axis) show M-sweat rates (mean ± S.D) for a sample of 7–33 glands from each subject chosen to give a mean M-sweat volume of ∼55 nl^−gland ( = 3.67 nl^{−min−gland}). Columns (left axis) show C/M ratios (mean ± SEM), expressed as% of mean WT value for the same sample of glands. Numbers indicate sample size for each subject. S4 had the fewest glands and the lowest M-sweat rates, which prevented matching her sample to the others.

Figure 12. C-sweat values as a function of M-sweat rates.

(A) Summary data showing dependence of C/M ratios to M-sweat rates for all 6 subjects tested (+) ivacaftor. For clarity, each subject's glands, glands were rank ordered according to their M-sweat rates and then divided into 3–5 bins according to their M-sweat rates (mean of 16.6 glands per bin). Each point plots the average M-sweat response on the x-axis and the average C/M ratio as a% of the WT mean on the y-axis for each bin; lines are linear fits. Each gland was measured 1–3 times, and the average number of glands included in each bin was 16.6 (range 5–26 glands). C/M vs. M-sweat correlations were significant for all subjects except S3. (B) C-sweat response is coded as 1, if some C-sweat is observed on a test, and 0, otherwise; and P(C-sweat) denotes the probability that C-sweat is observed. Then P(C-sweat) is also the average of the recoded 0/1 C-sweat response. We predicted P(C-sweat) as a function of log M-sweat using the combined data from all 6 CF subjects tested (+) ivacaftor. A logistic mixed model analysis showed that CF patients differed in their overall P(C-sweat) levels, but not in the within-patient effect of M-sweat response on P(C-sweat) across glands (p>.57), and that the average across patients of the within-patient effect of M-sweat was significantly positive (b = 2.14, p<.0001). The predicted P(C-sweat) levels are plotted for each patient. See methods and Appendix S1.

Figure 13. Loss-reduced and loss-corrected estimates of CFTR function in CF subjects (+) ivacaftor.

Each blue column (left axis) shows mean C/M ratios (mean ± SEM) for the 12 glands with the highest M-sweat rates for each subject; higher sweat rates reduce C-sweat losses. Each transparent red column (right axis) shows the corrected C/M ratios (mean ± SEM) for the 12 fastest M-secreting glands, obtained by adding 0.029 nl^{−min−gland} to the observed C-sweat secretion rates (see text).

Figure 14. Two measures of CFTR function in the sweat gland.

The sweat chloride concentration (left y axis,) is plotted (solid line) against CFTR function as a% of average WT values, plotted on the x axis. Sweat chloride is an inverse logarithmic function of CFTR function (n P_O γ) and hence is most sensitive at the lowest levels of CFTR function. The C-sweat/M-sweat ratio as% WT (right axis), is plotted (dashed line) against CFTR function as a% of average WT values, plotted on the x axis. The C/M ratio provides a direct readout of CFTR function and is linear over most of the range; the loss-reduced and loss-corrected values extend the linear range close to zero CFTR function. Sweat chloride point 3 is from ref. , the rest are from ref. . C/M point 3 is from present paper, other points are from ref. . CFTR values 0, 50 and 100 are defined, 1.5% and 5% function assigned by patch clamp and mRNA measures.