Article Cite This: Anal. Chem. 2019, 91, 7929−7934
pubs.acs.org/ac
Salivary Thiocyanate as a Biomarker of Cystic Fibrosis Transmembrane Regulator Function Andrey V. Malkovskiy,*,† Alexander A. Yacob,‡ Colleen E. Dunn,‡ Jacquelyn M. Zirbes,‡ Sean P. Ryan,‡ Paul L. Bollyky,§ Jayakumar Rajadas,*,† and Carlos E. Milla*,‡ †
Biomaterials and Advanced Drug Delivery Laboratory, Stanford School of Medicine, Stanford, California 94304, United States Center for Excellence in Pulmonary Biology, Department of Pediatrics, Stanford University, Stanford, California 94304, United States § Department of Immunology, Stanford University, Stanford, California 94304, United States Downloaded via UNIV OF SOUTHERN INDIANA on July 22, 2019 at 05:57:38 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
‡
S Supporting Information *
ABSTRACT: Improved methods are needed to reliably assess Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) function in vivo in light of recent therapeutic developments targeting the CFTR protein. Oral fluid from patients with cystic fibrosis (CF) and healthy controls (HCs) were studied using colorimetry and nonresonant Raman spectroscopy. Colorimetry experiments showed only a 36% decrease in thiocyanate (SCN−) concentration, but a sharp Raman peak at 2068 cm−1, attributable to (SCN−) vibrations, normalized to C−H peak, was on average 18 times higher for HC samples. Samples from patients undergoing treatment with CFTR modulators including ivacaftor, lumacaftor, and tezacaftor showed a high normalized peak in response to therapy. The peak intensity was consistent in longitudinal samples from single donors and in stored samples. The Raman peak ratio is a more sensitive, convenient, noninvasive biomarker for assessments of the therapeutic efficacy of drugs targeting CFTR and provides a value that is in much better agreement with theoretical expectations of saliva SCN− concentrations compared to colorimetry. This insight may greatly facilitate assessments of CFTR modulator efficacy in individual patients. ystic fibrosis (CF) is a common life-shortening autosomal recessive disease, which occurs predominantly in Caucasians, with an estimated 30,000 affected patients in North America alone. The disease is caused by mutations in the gene encoding the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR), a protein with anion channeling properties that participates in electrolyte homeostasis and fluid movement across mucosal surfaces.1 In the airway surface, tight regulation of the hydration and ion homeostatic conditions of the mucosal are the key innate defense mechanisms against bacterial exposure.2 As a result, individuals with CF are prone to chronic airway bacterial infection and severe inflammation.3 Recent therapeutic advances targeting specific CFTR mutations have led to partial restoration of CFTR function in some affected patients with substantial clinical benefit.4−7 Clinical studies with the CFTR modulators ivacaftor, lumacaftor, and tezacaftor have demonstrated improvements in CFTR function in response to intervention.8 Well established CF diagnostic tools such as the sweat chloride test (SCT) have been of value in detecting responses to therapy and may predict eventual therapeutic success.9 However, SCT has some shortcomings, particularly in regards to quantifying low levels of CFTR function.10 Clinical parameters such as the FEV1 test11 are at best downstream
C
© 2019 American Chemical Society
surrogate markers of CFTR function. It has proven difficult to identify clear associations at the individual level between SCT values and other clinical parameters.12 Tests that are more sensitive to CFTR function such as the Nasal Potential Difference (NPD)13 or sweat secretion under beta-adrenergic stimulation14 have provided greater insight into the effectiveness of CFTR-targeted therapies,9,15 but the practical application of these methods may be limited to large clinical trials. Therefore, ongoing efforts in CF drug development create a need for improved biomarkers of CFTR function that can be interrogated in the clinical care setting. This is particularly important considering that over 2000 different disease-associated mutations have been reported in the CFTR gene. Most attention in CF research has focused on CFTRmediated defects in Cl− secretion.16 However, CFTR is an anion channel that also transports thiocyanate (SCN−) to the epithelial apical surface, and this transport is likewise deficient in CF.17 This deficiency is also observed in transgenic CF pigs, a robust model for the study of CF airway pathophysiology.18 Received: April 12, 2019 Accepted: May 21, 2019 Published: May 22, 2019 7929
DOI: 10.1021/acs.analchem.9b01800 Anal. Chem. 2019, 91, 7929−7934
Article
Analytical Chemistry In light of the role of CFTR in SCN− transport, we asked whether concentrations of SCN− could be used as a biomarker of CFTR function. Previously, no differences in SCN− levels were seen between CF patients and healthy controls (HCs) in nasal airway surface liquid (ASL) samples assessed using ion chromatography.18 Similarly, only a 25% difference was observed between CF and HC samples via colorimetry.19 Both these results have diminished the potential of SCN ion to be used as a biomarker. We sought to improve upon these methods by using Raman spectroscopy in conjunction with unprocessed oral fluid secretions (which we refer to here subsequently as “saliva”) that were collected from both CF and HC subjects. We selected saliva as it is easy to sample in subjects of any age; the gland has high levels of CFTR expression, and it is known to have abnormalities in its electrolyte composition in subjects with CF.
Colorimetry. Colorimetry analyses were based on monitoring the formation of Fe3+ and SCN− colored complexes20,21 and measuring solution absorbance at 450 nm. For this, a 50 μL sample of saliva was mixed with 50 μL of 20 mM FeCl3 solution in HCl in 96-well half-volume clear bottom plates. Air bubbles were removed by puncturing with a sharp needle, and the solutions were gently mixed with pipet tips. Measurements were conducted in a Molecular Devices Spectramax M2e with the absorbance recorded at 450 nm. For full UV−vis spectral recordings (300−600 nm) of saliva samples, cuvettes with a 1 cm path length and an Agilent Cary 6000i UV/vis/NIR spectrophotometer were used. Myeloperoxidase Assay. To test the potential effect on our measurements of inflammation due to infection in CF patients, myeloperoxidase (MPO) activity in saliva and sputum of HC and CF subjects was measured using the CBA024 Innozyme kit (EMD-Millipore) and following recommended manufacturer protocols. Calibration curves were built using standards provided with the kit. Allyl-isothiocyanate Challenge Assay. In an attempt to measure the effect of dietary consumption of SCN− containing food items on SCN− salivary production, healthy subjects consumed 5 and 20 μL of allyl-isothiocyanate (AITC). To avoid contaminating saliva with unswallowed AITC, the compound was placed in gel caps for oral intake. Unstimulated salivary secretion samples were taken 1 h before, directly before, and 1, 2, and 3 h after intake. Statistics. Statistical analysis was performed using a twosample unequal variance 2-tailed Student’s t test to obtain a significance level of the difference between groups of samples. Error measurements shown are expressed as standard deviation from the mean. Study Approval. All protocols were approved by the Stanford University Medical Center IRB, protocol #24185. All experimental methods were carried out in accordance with the relevant guidelines and regulations. Written informed consent was received from healthy controls and CF patients prior to inclusion in the study. The experimenters were blinded to participants’ information. Data Availability. Data from the study is available upon request.
■
EXPERIMENTAL SECTION Patients. We recruited HC volunteers and patients with CF from the Stanford Translational CF Research Center for saliva collection. Two sample types were collected, stimulated and unstimulated saliva. For the stimulated saliva sample set, samples of at least 0.1 mL volume were collected from 14 CF subjects (2 of whom were pancreatic sufficient), 11 HC subjects, and 2 CF patients carrying a CFTR G551D mutation undergoing treatment with ivacaftor. For the unstimulated saliva sample set, samples from 11 CF subjects and 12 HC subjects were collected, and 9 additional samples were obtained from CF subjects undergoing treatment with different CFTR modulators. For all subjects, data are included in Table S-1. Sample Collection. Saliva was collected into sterile sealed containers by applying gentle suction. For stimulated sample collection, a cotton tip soaked in 2% citric acid solution was gently applied to the lateral surface of the tongue on both sides to stimulate salivation. For unstimulated collection, no citric acid was used. Sampling procedures are described in detail in the Supporting Information. Raman Spectroscopy. Raman spectroscopy is a technique that relies on the detection of inelastic scattering of a laser source to profile vibrational modes in a sample and provide a structural fingerprint by which molecules can be identified. The Raman measurements were performed using backscattering geometry on 20 μL salivary samples dried under vacuum on aluminum or titanium foil. This routine was followed both for improved heat dissipation from the incident laser and for ease and reliability of its chemical precleaning and very low intrinsic Raman signal (zero for titanium foil). The illumination sources were 473 or 633 nm continuous wave diode-pumped solid-state lasers. The laser power at the sample was ∼1 mW. Raman spectra for the dried samples were collected with either NTEGRA Spectra or Horiba Xplora Raman spectrometers with a 600 grating. All Raman spectra were corrected for the fluorescent background, using an automated custom-made algorithm, as well as for CCD camera noise and dark counts. In the resulting spectra, the peak at 2068 cm−1 has been approximated with a single Lorentzian with a fixed width and position. The resulting peak areas have been normalized to the total area of the C−H peak, which has been fit with 4 Lorentzians. All of the measurement steps were performed by automated routines to be kept free from human bias.
■
RESULTS AND DISCUSSION Colorimetric Determination of SCN− Ion Content in Saliva. We hypothesized that the relative amount of SCN− in saliva would be dependent on CFTR function. We reasoned that, although both SCN− and Cl− transport are impacted by CFTR deficiency, Cl− reabsorption will be hindered more than SCN− secretion, resulting in 50% higher concentrations of Cl− in the saliva of CF patients.22,23 In addition, anion conductance selectivity by the CFTR channel becomes almost four times less favorable for SCN− for certain CFTR mutations.24 Moreover, SCN− is consumed upon reacting with the higher levels of OCl− that are expected for CF patients. Thus, we reasoned that in CF saliva the combination of the above factors should result in a substantial decrease of SCN− concentrations. The concentration of SCN− in saliva could therefore be expected to provide an indirect indication of CFTR function. A schematic of this chemistry is shown in Figure 1. We first replicated the results obtained by Minarowski et al., who observed only a 25% difference in SCN− levels between CF and HC samples via colorimetry.19 The individual results for different patients are plotted as squares in Figure 2a. The 7930
DOI: 10.1021/acs.analchem.9b01800 Anal. Chem. 2019, 91, 7929−7934
Article
Analytical Chemistry
Figure 1. Raman spectroscopy signature for SCN− provides an indicator of CFTR function. (a) A schematic showing the role of the CFTR in salivary SCN− secretion. The CFTR secretes both SCN− and Cl− into saliva. Once secreted, SCN− can be consumed into OSCN− either directly via salivary peroxidase (SPO) or by reacting with OCl−. Due to this and to the lower amount of SCN− transported through the channel, CF subjects have much lower SCN− saliva concentrations. (b) The SCN− signature as seen with Raman spectroscopy. In CF patients, a Raman peak corresponding to SCN− at 2070−2080 cm−1 is almost absent. (c) Comparison of SCN− Raman signals in dried saliva versus controls, including pure KSCN and KSCN dried from H2O, KOCN, and KCN. The thick black line is the Lorentzian fit for the saliva sample.
amount of SCN− in the saliva of HC and CF not on drug, as measured by colorimetry, was found to be only 36% (or 1.57 times) lower on average (nonsignificant, p = 0.12), in good agreement with the earlier findings19 (Figure 2a,c). There is, potentially, minor skewing of our results in favor of HC samples for both techniques. This could be the result of a diurnal decrease in SCN− secretion due to the circadian cycle, as initially observed by Schultz et al.,25 which is most drastic in the early hours of the day. For comparison, SCT results for all of the CF subjects are presented as well (Figure 2b,e), which are more sensitive to sample type (Figure 2a) than colorimetry. We then tested whether treatment with CFTR modulator drugs had any effect on the colorimetry result (Figure 3). We did not find a significant difference in the SCN− colorimetry reading between CF patients with or without treatment with modulator drugs (Figure 3a,c). Raman Signature of the SCN− Ion. We next assessed the overall spectral characteristics of salivary samples using Raman spectroscopy. For this, we examined a set of conserved peaks expected to be present in all saliva samples from 500 to 800 cm−1, attributed to gauche and trans C−S sulfonic bond vibrations of cysteine, respectively. We found that the ratio of S−S and C−S peaks is consistent over all samples (Figure S-1), CF or non-CF, and agrees well with published data.26,27 Bleaching of a saliva sample after prolonged exposure at high laser power partially decomposes it, resulting in S−O peaks from 1000 to 1100 cm−1 as a result of oxidation. However, we did not observe oxidative damage to the other spectral features of saliva, particularly the C−H bands, which we use to
Figure 2. Salivary SCN− Raman peak ratio discriminates CF from HC subjects. (a) SCN− peak ratio for HC (white bars) and CF (gray bars) subjects and colorimetry values (black and red squares, respectively). Arrows mark samples from the same patient 3 months apart; (b) sweat test results and patient mutation, (c) colorimetry values for HC and CF subjects now expressed as scatter plots and average values, (d) same as (c) for SCN− peak ratio, and (e) same as (c) for sweat test results. These data demonstrate that our SCN− score is consistent with sweat test results in these patients. **p = 0.0023. ***p = 0.000187. Data include 11 CF samples and 12 HC samples.
normalize the SCN− peak intensity, so even under these settings, useful measurements can be possible. We then sought to identify a Raman “fingerprint” for SCN−. When the CN triple bond is present in an organic compound, it gives rise to a characteristic strong and narrow Raman peak at 2100−2246 cm−1.28,29 Cyanide ions (CN−) produce a peak at around 2070 cm −1 30−32 as do isothiocyanates and diazo compounds.33,34 This signal is located in the area of the spectrum where very few other normal modes are found and thus can be clearly identified. To determine the identity of the peak in a saliva sample, we also examined the spectrum of dry potassium thiocyanate, KSCN (blue line in Figure 1c). This revealed a very sharp peak at 2060 cm−1, while the spectrum of KSCN dissolved in water showed a much broader peak at 2076 cm−1 due to hydrogen bonding.35 The peak remained at the same position (solid line) 7931
DOI: 10.1021/acs.analchem.9b01800 Anal. Chem. 2019, 91, 7929−7934
Article
Analytical Chemistry
patients (p = 0.000187, Figure 2a,d). Thus, the results for CF subjects were 17.7 times lower, on average. As for colorimetry, we tested whether treatment with CFTR modulator drugs changed the Raman result (Figure 3a,d). The SCN− peak ratio for subjects on treatment was significantly different (p = 0.0021). In particular, these results show Raman peak ratios closest to HC among patients undergoing treatment with ivacaftor and more modest differences for patients receiving treatment with lumacaftor + ivacaftor or tezacaftor + ivacaftor. This finding is in agreement with clinical observations of clinical responses to these drugs in large clinical trials.4−7 These differences are less pronounced in samples from stimulated salivation samples (Figure S-2), probably due to dilution from increased water content as a result of stimulation conditions. This is in agreement with findings of Tenovuo et al.36 The results are highly consistent across regions of the salivary sample itself (Figure S-3). These results were also independent of the saliva fraction assessed, which part of the samples had been probed by the Raman beam, or whether the sample was filtered or not (Figure S-4). No influence of patient age was observed on SCN−. There was perhaps a trend toward female subjects having lower values for SCN− (Figure 4), which would be consistent with Tenovuo
Figure 3. Salivary SCN− peak ratio discriminates CF patients receiving CFTR corrective therapy from CF patients not receiving this therapy. (a) SCN− peak ratio results for CF (gray bars) and CF on drug (hashed bars) subjects. Arrows mark samples from the same patient 3 months apart; (b) sweat test results, patient mutation, and type of ongoing drug treatment, (c) colorimetry values now expressed as scatter plots and average values, (d) same as (c) for SCN− peak ratio. *p = 0.035. **p = 0.0021. Data include 11 CF samples and 9 CF on drug samples.
Figure 4. Colorimetry and Raman results show dependence of SCN− concentrations on gender for unstimulated samples from healthy controls: (a) colorimetry results; (b) Raman peak ratios. Total of 6 M and 6 F healthy controls.
and Makinen,37 but this result was not statistically significant in our study. Note that, in our analysis, we did not use this gender correction factor when determining sensitivity of the technique. When this factor is accounted for, the difference between groups would be even more significant. Influence of Inflammation on Measurement Consistency. A concern was the potential influence of inflammation and oxidant activity due to the known chronic infections that affect CF patients. To evaluate for this, we assessed myeloperoxidase (MPO) peroxidation activity in saliva. If the peroxidation activity would be higher for CF patients, the measured SCN− levels would also be higher, resulting in values closer to those of HC and lowering the sensitivity of our tests to the subject condition. We found no significant differences between HC and CF (HC 35 ± 44 ng/mL, CF 53 ± 49 ng/ mL), despite very high levels in the sputum of CF patients (1644 ± 146 ng/mL), reflecting the known high inflammatory activity of CF.
when samples were dried up in low vacuum. Other possible compounds containing the CN group were also tested: KCN (2086 cm−1, red line) and KOCN (2173 cm−1, thick blue line). Their peak positions were clearly away from the SCN− peak. We then asked whether we could observe the same peak in dried saliva samples from HC subjects. We indeed observed a characteristic peak at 2076 cm−1 (Figure 1c) that matched the one observed for dried KSCN. This represented the level of SCN− present in saliva. Together, these data indicated that we were able to identify an SCN− signature in Raman spectroscopy of dried saliva samples. We then assessed samples from study subjects and found that, while the SCN− peak was uniformly present in HC samples, its amount was significantly lower in samples from CF 7932
DOI: 10.1021/acs.analchem.9b01800 Anal. Chem. 2019, 91, 7929−7934
Article
Analytical Chemistry We also ascertained whether individual subjects had consistent results over time using this approach. Longitudinal Raman and Raman/colorimetry results demonstrated (Figure 5) consistency of results for the same subject. The results for a
concentration drop in saliva of CF subjects. In contrast to colorimetry data, our simple Raman protocol of salivary secretions samples demonstrated highly significant differences between HC and CF, as well as among CF patients undergoing treatment with CFTR modulator drugs. Thus, Raman results provided for a much better measure of CFTR function and with much more clear separation between the study groups, compared to colorimetry or SCT results. We were unable to identify a dietary influence on intraday SCN− salivary secretions content after a challenge with AITC in our pilot experiments (n = 4, data not shown). We did notice a diurnal decrease in levels, and this is consistent with the FT-IR results of Schultz et al.,25 who observed a strong circadian dependence with the highest levels in the early morning and declining rapidly toward afternoon, after which the decline becomes much slower. Then, differences in timing of sampling and daily routine between the HC and CF subjects sampled may have influenced our results. However, this would affect colorimetry and Raman results in equal measure. Our data demonstrate that Raman spectrographic analysis of unprocessed salivary secretions clearly discriminates CF from healthy subjects and that this assay is sensitive to the effect of CFTR modulators on channel function. Our data indicate that the ratio of SCN− Raman peak normalized to CH peak (“SCN− peak ratio”) is a robust and practical biomarker of CFTR function. While there exists a large variability from patient to patient, the 17.7-fold difference between HC and CF samples will still allow for reliable identification of drug potency given enough subjects. However, as our longitudinal data shows, the Raman results are very consistent for each individual over time. Thus, for small size clinical trials, it might be beneficial to track the Raman peak for the same patient before and after onset of treatment. Our data suggest that Raman spectroscopy may distinguish between pancreatic-sufficient (PS) and nonsufficient (non-PS) CF patients (0.02 ± 0.00 vs 0.0078 ± 0.0055), while there is no difference in their sweat test results (108.5 ± 29.0 vs 94.3 ± 9.0) (Figure S-6). Further studies with larger patient groups will be required to explore this difference and its potential application to CF diagnostics, particularly for subjects with inconclusive sweat chloride results. We had only one subject (G551D mutation) for whom Raman data was collected before and after the onset of drug treatment with ivacaftor, but the result was very promising with a 7-fold increase in the Raman peak ratio (spectra not shown). This brought the Raman peak ratio to the level of HC controls, which was expected for this mutation. We hope to obtain results from more patients before and after drug treatment to further validate the sensitivity of the Raman method.
Figure 5. Longitudinal study of HC, CF, and CF on drug samples shows consistency of results for the same subject over time. Each experimental point was taken on a different day. For HC, CF + drug, and CF there were 7, 7, and 5 samples collected, respectively, over a 91 day time period. No two samples were collected on the same date. *****p = 0.000066.
subject homozygous for Phe508Δ mutation undergoing drug treatment with a combination of lumacaftor/ivacaftor demonstrated an improvement over those observed in untreated CF subjects, although not statistically significant Our data shows that SCN− results by colorimetry for CF patients are only 36% lower than for HC samples and with a high degree of variation. This result is in agreement with the work of Minarowski et al., who used colorimetry alone.19 Lorentzen et al.18 found that CFTR is not indispensable for the generation of a SCN− concentration gradient between nasal ASL and serum in adult human subjects. Their study found that the level of SCN− in nasal ASL was the same for CF and healthy subjects, thus also ruling out the possibility of using the SCN− concentration of ASL secretions to evaluate CFTR function in patients and monitor drug efficacy. Similarly, colorimetry did not appear as a sensitive enough technique. While the pooled results may show significance with a large enough sample number, the individual results show substantial variation (Figure 2a,c). Upon closer examination of the colorimetry assay, it appears that there is a strong influence of inherent haziness of salivary samples on the absorbance reading for a very wide range of wavelengths (Figure S-5b). This is even more pronounced for CF subject samples, due to the well-known dehydration of their saliva. Experiments of additions of known concentrations of SCN− to parotid, sublingual saliva, and 9% BSA solutions (Figure S-5c) show that the lines are clearly not linear when approaching zero concentration of added ion with a high nonzero absorbance value even for BSA solution without any SCN− concentration. Thus, single-wavelength colorimetry artificially measures a much higher SCN− concentration in samples. This results in much weaker sensitivity of the colorimetry assay and a significant underestimation of SCN−
■
CONCLUSION In summary, we have described a convenient, noninvasive, and accurate method to assess CFTR function in vivo in human subjects based on Raman measurements of SCN− in unprocessed salivary secretions. Such secretions, shipped in dry state, can be collected by mail and would not require a patient visit. More specifically, we have shown the deficiency of simple colorimetric determination of SCN− concentration so far used in the literature and have shown that the true concentration difference between HC and CF subjects is much greater, which agrees with the theory on CFTR permeability. Further, our method seems also well suited to detect changes in CFTR function in response to modulator therapy. Given the 7933
DOI: 10.1021/acs.analchem.9b01800 Anal. Chem. 2019, 91, 7929−7934
Article
Analytical Chemistry
(7) Rowe, S. M.; Daines, C.; Ringshausen, F. C.; Kerem, E.; Wilson, J.; Tullis, E.; Nair, N.; Simard, C.; Han, L.; Ingenito, E. P.; et al. N. Engl. J. Med. 2017, 377, 2024−2035. (8) Milla, C. Clin. Invest. 2014, 4 (4), 297−299. (9) Accurso, F. J.; Van Goor, F.; Zha, J.; Stone, A. J.; Dong, Q.; Ordonez, C. L.; Rowe, S. M.; Clancy, J. P.; Konstan, M. W.; Hoch, H. E.; et al. J. Cystic Fibrosis 2014, 13 (2), 139−147. (10) Kharrazi, M.; Milla, C.; Wine, J. Lancet Respir. Med. 2016, 4, 605−607. (11) Padoan, R.; Genoni, S.; Moretti, E.; Seia, M.; Giunta, A.; Corbetta, C. Acta Paediatr. 2002, 91, 82−87. (12) Fidler, M. C.; Beusmans, J.; Panorchan, P.; Van Goor, F. J. Cystic Fibrosis 2017, 16, 41−44. (13) Schüler, D.; Sermet-Gaudelus, I.; Wilschanski, M.; Ballmann, M.; Dechaux, M.; Edelman, A.; Hug, M.; Leal, T.; Lebacq, J.; Lebecque, P.; et al. J. Cystic Fibrosis 2004, 3 (2), 151−155. (14) Char, J. E.; Dunn, C.; Davies, Z.; Milla, C.; Moss, R. B.; Wine, J. J. PLoS One 2017, 12 (4), No. e0175486. (15) Kim, J.; Davies, Z.; Dunn, C.; Wine, J. J.; Milla, C. J. Cystic Fibrosis 2018, 17 (2), 179−185. (16) Hoo, A. F.; Thia, L. P.; Nguyen, T. T.; Bush, A.; Chudleigh, J.; Lum, S.; Ahmed, D.; Balfour Lynn, I.; Carr, S. B.; Chavasse, R. J.; et al. Thorax 2012, 67 (10), 874−881. (17) Moskwa, P.; Lorentzen, D.; Excoffon, K. J.; Zabner, J.; McCray, P. B., Jr; Nauseef, W. M.; Dupuy, C.; Bánfi, B. Am. J. Respir. Crit. Care Med. 2007, 175, 174−183. (18) Lorentzen, D.; Durairaj, L.; Pezzulo, A. A.; Nakano, Y.; Launspach, J.; Stoltz, D. A.; Zamba, G.; McCray, P. B., Jr; Zabner, J.; Welsh, M. J.; et al. Free Radical Biol. Med. 2011, 50, 1144−1150. (19) Minarowski, Ł.; Sands, D.; Minarowska, A.; Karwowska, A.; Sulewska, A.; Gacko, M.; Chyczewska, E. Folia Histochem. Cytobiol. 2008, 46 (2), 245−246. (20) Powell, W. N. J. Lab. Clin. Med. 1945, 30 (12), 1071−1075. (21) Betts, R. H.; Dainton, F. S. J. Am. Chem. Soc. 1953, 75, 5721− 5727. (22) di Sant-Agnese, P. A.; Grossman, H.; Darling, R. C.; Denning, C. R. Pediatrics 1958, 22, 507−514. (23) Gonçalves, A. C.; Marson, F. A.; Mendonça, R. M.; Ribeiro, J. D.; Ribeiro, A. F.; Paschoal, I. A.; Levy, C. E. Diagn. Pathol. 2013, 8, 46. (24) Linsdell, P.; Evagelidis, A.; Hanrahan, J. W. Biophys. J. 2000, 78, 2973−2982. (25) Schultz, C. P.; Ahmed, M. K.; Dawes, C.; Mantsch, H. H. Anal. Biochem. 1996, 240, 7−12. (26) Hayashi, M.; Shimanouchi, T.; Mizushima, S. J. Chem. Phys. 1957, 26, 608−612. (27) Lord, R. C.; Yu, N. J. Mol. Biol. 1970, 50, 509−524. (28) Shi, C.; Zhang, W.; Birke, R. L.; Lombardi, J. R. J. Electroanal. Chem. 1997, 423, 67−81. (29) Thygesen, L. G.; Jorgensen, K.; Moller, B. L.; Engelsen, S. G. Appl. Spectrosc. 2004, 58 (2), 212−217. (30) Volkov, S. V.; Evtushenko, N. P.; Yatsimirskii, K. B. Theor. Exp. Chem. 1975, 10 (2), 216−219. (31) Bron, M.; Holze, R. J. Electroanal. Chem. 1995, 385, 105−113. (32) Kettle, S. F. A.; Aschero, G. L.; Eliano, D.; Rosetti, R.; Stanghellini, P. L. Inorg. Chem. 2006, 45, 4928−4937. (33) Lin-Vien, D.; Colthup, N.; Fateley, W.; Grasselli, J. The Handbook of infrared and Raman characteristic frequencies of organic molecules; Academic Press, 1991; pp 478−486. (34) Davidovics, G.; Debu, F.; Marfisi, C.; Monnier, M.; Aycard, J. P.; Pourcin, J.; Bodot, H. J. Mol. Struct. 1986, 147, 29−45. (35) Pecile, C. Inorg. Chem. 1966, 5, 210−214. (36) Tenovuo, J.; Pruitt, K. M.; Thomas, E. L. J. Dent. Res. 1982, 61 (8), 982−985. (37) Tenovuo, J.; Makinen, K. K. J. Dent. Res. 1976, 55 (4), 661− 663.
multiple practical aspects of our assay, the Raman method presented here might offer advantages over other biomarkers and end points commonly used in CF. This insight may greatly facilitate the development of additional novel CFTR modulators.
■
ASSOCIATED CONTENT
* Supporting Information S
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.9b01800.
■
Additional experimental details, Raman spectra, SCN− peak ratios, Raman peak area ratios, illustration of imperfections in colorimetry measurements, and patient information tables (PDF)
AUTHOR INFORMATION
Corresponding Authors
*Phone: +1-330-208-7135. E-mail:
[email protected] (A.V.M.). *Phone: +1-650-724-6806. E-mail:
[email protected] (J.R.). *Phone: +1-650-736-9824. E-mail:
[email protected] (C.E.M.). ORCID
Andrey V. Malkovskiy: 0000-0002-5648-8602 Author Contributions
All authors have given approval to the final version of the manuscript. A.V.M. designed the study, designed and performed experiments, analyzed data, and prepared the manuscript. A.A.Y. performed the myeloperoxidase assay and collected samples. C.E.D., J.M.Z., and S.P.R. collected samples. P.L.B. and C.E.M. designed the study and prepared the manuscript. J.R. participated in experimental design. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank D. Mochly-Rosen (Department of Chemical Systems Biology, Stanford University) for useful suggestions. The work was supported by Cystic Fibrosis Research Inc. (CFRI), an Elizabeth Nash Memorial CF fellowship to A.V.M., a Ross Mosier Laboratories Fund to C.E.M., and R21 R21AI137432 to P.L.B. Part of this work was performed at the Stanford Nano Shared Facilities (SNSF) supported by the National Science Foundation under award ECCS-1542152.
■
REFERENCES
(1) Quinton, P. M. Physiol. Rev. 1999, 79, S3−S22. (2) Boucher, R. C. Adv. Drug Delivery Rev. 2002, 54, 1359−1371. (3) Gibson, R. L.; Burns, J. L.; Ramsey, B. W. Am. J. Respir. Crit. Care Med. 2003, 168 (8), 918−951. (4) Ramsey, B. W.; Davies, J.; McElvaney, N. G.; Tullis, E.; Bell, S. C.; D’evínek, P.; Griese, M.; McKone, E. F.; Wainwright, C. E.; Konstan, M. W.; et al. N. Engl. J. Med. 2011, 365, 1663−1672. (5) Wainwright, C. E.; Elborn, J. S.; Ramsey, B. W.; Marigowda, G.; Huang, X.; Cipolli, M.; Colombo, C.; Davies, J. C.; De Boeck, K.; Flume, P. A.; et al. N. Engl. J. Med. 2015, 373, 220−231. (6) Taylor-Cousar, J. L.; Munck, A.; McKone, E. F.; van der Ent, C. K.; Moeller, A.; Simard, C.; Wang, L. T.; Ingenito, E. P.; McKee, C.; Lu, Y.; et al. N. Engl. J. Med. 2017, 377, 2013−2023. 7934
DOI: 10.1021/acs.analchem.9b01800 Anal. Chem. 2019, 91, 7929−7934
