Biochemical Gas Sensors (Biosniffers) Using Forward and Reverse

Nov 9, 2017 - h Preventive Oral Health Care Science, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushi...
2 downloads 14 Views 1MB Size
Article Cite This: Anal. Chem. XXXX, XXX, XXX-XXX

pubs.acs.org/ac

Biochemical Gas Sensors (Biosniffers) Using Forward and Reverse Reactions of Secondary Alcohol Dehydrogenase for Breath Isopropanol and Acetone as Potential Volatile Biomarkers of Diabetes Mellitus Po-Jen Chien,† Takuma Suzuki,† Masato Tsujii,† Ming Ye,‡ Isao Minami,∥ Kanako Toda,g Hiromi Otsuka,h Koji Toma,‡ Takahiro Arakawa,‡ Kouji Araki,g Yasuhiko Iwasaki,§ Kayoko Shinada,h Yoshihiro Ogawa,∥,⊥,# and Kohji Mitsubayashi*,†,‡ †

Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, Japan ‡ Department of Biomedical Devices and Instrumentation, Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, 2-3-10 Kanda-Surugadai, Chiyoda-ku, Tokyo 101-0062, Japan § Faculty of Chemistry, Materials and Bioengineering, Kansai University, 3-3-35 Yamate-Cho, Suita-Shi, Osaka 564-0836, Japan ∥ Department of Molecular Endocrinology and Metabolism, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, Japan ⊥ Department of Medical and Bioregulatory Science, Graduate School of Medical Sciences, Kyushu University, 3-1-1, Maidashi, Higashi-ku, Fukuoka City, 812-8582, Japan # Department of Molecular and Cellular Metabolism, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, Japan g Educational System in Dentistry, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, Japan h Preventive Oral Health Care Science, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, Japan S Supporting Information *

ABSTRACT: This study describes two biosniffers to determine breath acetone and isopropanol (IPA) levels and applies them for breath measurement in healthy subjects and diabetic patients. Secondary alcohol dehydrogenase (S-ADH) can reduce acetone and oxidize nicotinamide adenine dinucleotide (NADH to NAD+) in a weak acid environment. NADH can be excited by 340 nm excitation lights and subsequently emit 490 nm fluorescence. Therefore, acetone can be measured by the decrease in NADH fluorescence intensity. S-ADH can also oxidize IPA and reduce NAD+ to NADH when it is in an alkaline environment. Thus, IPA can be detected by the increase of fluorescence. The developed biosniffers show rapid response, high sensitivity and high selectivity. The breath acetone and IPA analysis in healthy subjects shows that the mean values were 750.0 ± 434.4 ppb and 15.4 ± 11.3 ppb. Both acetone and IPA did not show a statistical difference among different genders and ages. The breath acetone analysis for diabetic patients shows a mean value of 1207.7 ± 689.5 ppb, which was higher than that of healthy subjects (p < 1 × 10−6). In particularly, type-1 diabetic (T1D) patients exhaled a much higher concentration of acetone than type-2 diabetic (T2D) patients (p < 0.01). The breath IPA also had a higher concentration in diabetic patients (23.1 ± 20.1 ppb, p < 0.01), but only T2D patients presented a statistical difference (23.9 ± 21.3 ppb, p < 0.01). These findings are worthwhile in the study of breath biomarkers for diabetes mellitus diagnosis. Additionally, the developed biosniffers provide a new technique for volatolomics research.

O

ver the last few decades, considerable interest has arisen over the volatolomics for medical application because the volatile organic compounds (VOCs) emitted from a human body contain abundant bioinformation, and the noninvasive measurement patterns can reduce the patient’s uncomfortable © XXXX American Chemical Society

Received: August 8, 2017 Accepted: October 26, 2017

A

DOI: 10.1021/acs.analchem.7b03191 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 1. (A) Schematic diagram of the acetone and IPA biosniffers. The inserted chart shows the structure of flow-cell. (B) Schematic diagram of the breath measurement using biosniffers. The sample breath was delivered to the biosniffer by an air pump, and the flow rate was 200 mL min−1 same as the standard gas measurement. (C) The selectivity of acetone and IPA biosniffers. N.D.: none detected. *: 2-butanol did not measure by the acetone biosniffer.

22.1 mg dL−1 in AKA patients.13 The detected IPA in DKA and AKA patients should be endogenous because the IPA-toacetone ratio was much higher than those in the IPA-ingested cases.13 It is speculated that endogenic IPA is generated from a reverse reaction of ADH in the liver, which metabolizes acetone to IPA when the human body is in an unusual condition, such as overstock of acetone or a high NADH-to-NAD+ ratio (reduced and oxidized form of nicotinamide adenine dinucleotide).15−17 NADH and NAD+ are commonly considered coenzymes in numerous biological reactions in the human body, such as energy metabolism and mitochondrial functions.18 The NADH-to-NAD+ ratio in the human body is regulated by several mechanisms,19 but it may change in particular body statuses, such as a high concentration of fatty acid, chronic alcohol abuse, and diabetes mellitus.15,16 Thus, the presence of high IPA concentrations in a human body may be a warning of the collapsing balance between NADH and NAD+. One of the breath analysis challenges is that a high-sensitivity and good-selectivity detection pattern commonly requires a trained operator and is coupled with bulky and expensive instruments such as a gas chromatography−mass spectrometer (GC-MS),20 which cause difficulty for medical applications. Several types of semiconducting metal oxides such as WO321−23 and SnO224 have been reported as chemoresistive-type acetone gas sensors. The advantages of these acetone sensors include fast responses, simple operation and ease as a portable device. However, their require working temperatures are approximately 200−400 °C and the selectivity remains a difficult challenge for

feelings. In particular, detection of acetone in breath is gaining attention because acetone is considered a potential biomarker of diabetes mellitus.1,2 The generation mechanism of acetone in the human body has been investigated in detail. Numerous studies reported higher concentrations of breath acetone in diabetic patients.1,3,4 For example, Artur Rydosz used a micropreconcentrator in low-temperature cofired ceramics (LTCC) technology with mass spectrometry to detect breath acetone. They found that the median concentrations in healthy subjects and type-1 diabetic (T1D) patients were 0.63 ± 0.12 ppm and 2.08 ± 0.47 ppm, respectively.3 Additionally, some studies reported good correlations between the breath acetone and blood glucose levels in diabetic patients.5,6 Thus, the determination of breath acetone in diabetic patients is extremely worthwhile. In addition to acetone, recently, isopropanol (IPA), which is an acetone-related metabolite, was proposed as a new potential biomarker in diabetic breathomics.7,8 A high concentration of IPA in a human body is usually considered to be deliberate or accidental IPA ingestion or inhalation.9,10 The exogenous IPA in a human body is mainly metabolized to acetone by alcohol dehydrogenase (ADH).11,12 However, an increasing number of studies reported that IPA was detected in body fluid in the diabetic ketoacidosis (DKA) or alcoholic ketoacidosis (AKA) patients who had no IPA exposure histories.13−16 For example, Petersen group measured IPA in post-mortem with ketoacidosis, including DKA and AKA, and found that the mean blood IPA was 15.1 ± 13.0 mg dL−1 in DKA cases and was 18.5 ± B

DOI: 10.1021/acs.analchem.7b03191 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

flow-cell equipped optical fiber probe (Ocean Optics Inc. USA). To eliminate the background noise, a band-pass filter (BPF, Asahi Spectra Co., Ltd., Japan) with a center wavelength of 340 ± 10 nm was embedded between the UV-LED and the bifurcated fiber; the other band-pass filter with a center wavelength of 490 ± 10 nm was set in front of the PMT. The sensing part of the biosniffer was constructed by inserting an optical fiber probe into the center of a flow-cell (Figure 1). An enzyme-immobilized membrane was attached to the flow-cell by a silicon O-ring. The buffer solution, which contained a coenzyme (NADH for acetone and NAD+ for the IPA biosniffer), was pumped into the flow channel, streamed through the space between the cell and the enzyme membrane, and withdrawn from the exit. The enzyme-immobilized membrane was fabricated by entrapment immobilization of SADH on a H-PTFE membrane using the PMEH polymer. The S-ADH membrane preparation method, characterization data of the enzyme membrane and other details of the biosniffer construction were described in our previous works.25−27 The biosniffers were validated by measuring a serial of standard acetone and IPA concentrations. For the acetone biosniffer, when it contacted the acetone vapor, the NADH in the buffer stream was immediately consumed by an enzymatic reaction of S-ADH, and the change (decrease) in fluorescence intensity was detected by the PMT. Simultaneously, the flowed buffer solution continuously removed the produced NAD+ and supplied fresh NADH. Thus, the working principle of the acetone biosniffer was established on a balance between the consumption and the delivery speed of NADH. Meanwhile, the IPA biosniffer used the NAD+-containing buffer solution. Similarly, when it contacted the IPA vapor, NADH was produced by S-ADH, and the change (increase) in fluorescence was measured by the PMT. A mass flow controller (Koflok Inc., Japan) maintained the air flow-rate of the standard vapor and breath sample at 200 mL min−1. Healthy Subjects and Diabetic Patients. This research was authorized by the Human Investigations Committee of Tokyo Medical and Dental University (authorization code 2015−06) and performed according to Declaration of Helsinki. In total, 25 diabetes mellitus patients were recruited from the Department of diabetes of the Tokyo Medical and Dental University Hospitals. Four of the patients were suffering from type-1 diabetes mellitus (T1DM), and the other 21 were diagnosed as type-2 diabetes mellitus (T2DM). No diabetic patients were hospitalized, and all were under a regular checkup every 2−3 months according to the medical order. Each participated patient was asked to supply a breath sample after the clinical examination. On average, every patient supplied 3− 4 exhaled breath samples during the entire experiment. The healthy subjects, who were not experiencing any treatment, chronic alcoholic abuse or diet, were recruited from the Tokyo Medical and Dental University. All participants were given a detail explanation of the experiment purpose and process and were asked to fill out a questionnaire, which included the sex, age, height, weight, last time of eating, smoking history, drinking status, and exercise habit. Breath Collection Method. To eliminate the interference of ambient air in the hospital, the breath collection place was performed in a clean room outside of the hospital area. Before the breath collection, all subjects rested in the room for at least 15 min until they felt comfortable and had easy breath. The subject was asked to take a deep inhalation, hold for 15 s, and gently blew the exhaled air through a T-shaped straw. One of

them. In our previous works, we have developed two gas-phase biochemical biosensors (biosniffers), which are highly sensitive, selective, and easy to operate, for the determination of breath acetone and IPA.25,26 These biosniffers measure the change in NADH fluorescent intensity, which results from the enzymatic reactions (eq 1) of the secondary alcohol dehydrogenase (SADH). S‐ADH pH ≤ 7

acetone + NADH + H+ XoooooooY IPA + NAD+ pH ≥ 8

(1)

S-ADH reduces acetone to IPA in a weak acid environment and oxidizes proportional NADH to NAD+. NADH has a unique optical property: when it is exposed to ultraviolet radiance with 340 nm wavelength, it releases the fluorescence with a center wavelength of 490 nm. Therefore, the acetone concentration can be measured by the decrease in NADH fluorescent intensity. S-ADH can also oxidize IPA to acetone and reduce NAD+ to NADH when it is in an alkaline environment. Thus, the IPA biosniffer determines the increase in NADH fluorescent intensity as a signal. Both developed biosniffers have been applied for a determination of human breath. The acetone biosniffer has successfully been used to evaluate the change of breath acetone in healthy subjects during aerobic exercise,25 and the IPA biosniffer has been applied to investigate the distribution of breath IPA in healthy people.26 On the basis of our results, the biosniffers provide benefits to the exhaled-air research field and may promote breath research in clinic applications. In this study, we used the biosniffers to investigate the exhaled acetone and IPA concentration in diabetic patients and compared the results with those in healthy subjects to enhance and promote the exploration of breath biomarkers in diabetes mellitus.



EXPERIMENTAL SECTION Materials. S-ADH was purchased from Daicel Chiral Technologies, Co., Japan. 2-Amino-2-hydroxymethyl-1,3-propanediol (Tris), hydrochloric acid (HCl), potassium dihydrogen phosphate (KH 2 PO 4 ), and dipotassium hydrogen phosphate (K2HPO4) were obtained from Wako Pure Chemical Industries, Japan. NAD+ and NADH were bought from Oriental Yeast, Co., Ltd., Japan. All prepared solutions used deionized distilled water from a Milli-Q purification system (Millipore, Co., USA). The H-PTFE membrane (pore size = 0.2 μm) was obtained from Millipore, USA. The gas sampling bag was purchased from As One Co., Ltd., Japan. The synthesis method of PMEH polymer, which was copolymerized by 2-methacryloyloxyethyl phosphorylcholine with 2-ethylhexyl methacrylate, was described in our previous work.27 Biosniffers for Acetone and IPA Measurements. A diagrammatic sketch of the biosniffer system is shown in Figure 1A. The system comprised an excitation unit, a bifurcated optical fiber, an optical fiber probe, which was equipped with a cylindrical flow cell, an enzyme-immobilized membrane, and a photon detection unit. The excitation unit to excite NADH was composed of a UV-LED (Sensor Electronic Technology, Inc., USA), whose central wavelength of the emission spectrum was 340 nm. A photomultiplier (PMT, Hamamatsu Corporation, Japan) was used as an NADH fluorescence detector. A Yshaped bifurcated optical fiber (Ocean Optics Inc. USA) was used to integrate these two units. The excitation unit and fluorescence detection unit were connected to two ports of the bifurcated fiber, and the common end was assembled with the C

DOI: 10.1021/acs.analchem.7b03191 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 2. (A) Calibration curves of the acetone biosniffer. The calibration range was 30−5468 ppb. (B) Typical responses of the acetone biosniffer to the breath samples; Healthy subject (523 ppb) and diabetic patient (1307 ppb). (C) Calibration curves of the IPA biosniffer. The calibration range was 1−9060 ppb. (D) Typical responses of the IPA biosniffer to the breath samples; Healthy subject (15.7 ppb) and diabetic patient (65.6 ppb).



RESULTS AND DISCUSSION Characteristics of the Biosniffers. The acetone biosniffer used a pH-7.0 phosphate buffer, which contained 50 μmol L−1 of NADH. As shown in Figure 1C (left), the acetone biosniffer rarely reacted with other substances with similar chemical structures. Only 2-butanone presented approximately 140% relative signals. We supposed that the interference was negligible because the exhaled acetone in healthy people was commonly hundreds of ppb,29,30 but 2-butanone was only approximately 0.38 ppb.31 The IPA biosniffer used a pH-8.5 Tris-HCl buffer, which contained 1 mmolL−1 of NAD+. Figure 1C (right) shows its selectivity; the IPA biosniffer almost did not respond to other commonly observed substances in human breath, such as ethanol and methanol. Only 2-butanol presented approximately 60% relative signals. It is considered that 2-butanol does not interfere with the breath IPA analysis because 2-butanol rarely exists in human breath.32 The calibration curve of the acetone biosniffer is shown in Figure 2A, which fits the sigmoidal shape and can be described by eq 2 with the correlation coefficient of 0.996:

the T-shaped straw ports was connected to a 3-litter sample bag, and the other end was an empty port. A three-way valve in the middle of the T-shaped straw was used to control the flow direction of the exhaled air from the subject. The first 3 s of the exhaled breath was abandoned through the empty end because it was supposed to be the dead space air without gas exchange with the alveolar.28 The remainder of the breath was guided to the other port and collected into a sample bag. The collected samples were stored in a heat-insulated box, which could maintain the temperature at approximately 37 °C, and immediately delivered to the laboratory. Figure 1B shows the scheme of the breath measurement method. The sample bag was connected to an air pump to deliver the exhaled air to the biosniffer, and the flow speed of the breath was set to be identical to a standard gas measurement. The breath samples from the subjects who smoked within an hour, had an alcoholic beverage drink within 24 h or used any alcohol-contained spray before the collection were excluded. Statistical Analyses. The breath data were presented as the mean ± standard deviations. A Student’s t-Test was used for the statistical analyses to detect a difference among the compared groups. The one-way analysis of variance (ANOVA) was used to compare the age-stratified groups. The selected threshold for statistical significance was the pvalue below 0.05. The Pearson Correlation analysis was used to examine a relationship between breath acetone and breath IPA. Statplus (mac LE, version 6.0.3, AnalystSoft Inc.) was used to calculate all analyses.

Δintensity = 1236.30 + 902361.59 /(1 + (0.16/acetone, ppb)0.48 )69.74

(2)

Δintensity is defined as an absolute value of the change in NADH fluorescence intensity. The dynamic range was confirmed to be 30−5468 ppb, which included breath acetone D

DOI: 10.1021/acs.analchem.7b03191 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 3. (A) Breath acetone in the healthy subjects and the age-stratified groups. No significant difference was observed (p = 0.46). (B) Breath acetone concentration in different gender. No significant was observed (p = 0.27). (C) Breath IPA in healthy subjects and the age-stratified groups. No significant difference was observed (p = 0.65). (D) Breath IPA concentration in different gender. No significant was observed (p = 0.55).

T1D patients and 75 samples from T2D patients. The IPA biosniffer measured 11 breath samples from T1D patients and 67 samples from T2D patients. The age of the healthy subjects was 30.8 ± 10.9, which showed significant differences compared to T1D (44.5 ± 17.2, p < 0.01) and T2D patients (64.5 ± 10.5, p < 0.01). The summary of subjects demographic was reported in Table S2. Breath Acetone and IPA in Healthy Subjects. Figure 3A shows the breath acetone concentrations in the healthy subjects and age-stratified groups. The mean concentration of breath acetone in all healthy subjects was 750.0 ppb. The exhaled acetone concentrations in the groups of under 30, 30−40, and over 40 years of age were 795.0, 661.4, and 675.5 ppb, respectively. There was no statistical difference for the age groups (p = 0.46). Figure 3B shows the mean breath acetone concentration in different sexes of healthy subjects: 719 ppb in male and 818 ppb in female. There was no statistical difference between males and females (p = 0.27). The breath IPA concentrations in the healthy subjects and age-stratified groups are shown in Figure 3C. The breath IPA concentrations in the under-30-, 30−40-, and over-40-year-old groups were 16.3, 13.4, and 14.0 ppb; the average for all healthy subjects was 15.4 ppb. There was no statistical difference for the age groups (p = 0.65). Figure 3D shows the mean concentrations of breath IPA in different sexes of healthy subjects: 14.9 ppb in male and 16.5 ppb in female, and no statistical difference was observed (p = 0.55). Thus, breath acetone and IPA may not correlate with age and sex. However, this conclusion must be treated with caution because it is not entirely consistent with previous research. The lack of correlation between breath acetone and age is consistent

in healthy people and diabetic patients. Figure 2B shows the typical responses of the acetone biosniffer to the breath samples of a healthy subject (523 ppb) and a diabetic patient (1307 ppb). The response time to reach 90% of the steady signal intensity was approximately 40 s. Figure 2C shows the calibration curve of the IPA biosniffer, which can be controlled by a power approximation of eq 3, and the correlation coefficient is 0.991: Δintensity = 2674.6 × (IPA,ppb)0.89

(3)

The calibration range was 1−9060 ppb, which was comparable to the ranges of other conventional IPA detection methods,33,34 and included the reported concentrations of breath IPA in healthy people and diabetic patients.7,35 Figure 2D shows the typical responses of two breath samples for the IPA analysis: the healthy subject (15.7 ppb) and the diabetic patient (65.6 ppb). The response time of the IPA biosniffer to reach 90% of the steady signal intensity was approximately 60 s. In addition to the calibration curves and response times, many other details of the characteristics and optimization conditions for the biosniffers were obtained and described in our previous work.25,26 A short comparison of our biosniffers with the reported acetone and IPA sensors is shown in Table S1.21−24,36−39 Subject Demographics. In total, 55 healthy volunteers, including 41 males and 14 females, supplied 108 breath samples for the acetone measurement, and 91 samples for the IPA analysis. Twenty-five diabetic patients joined the experiment, including 9 males and 16 females; 4 diabetic patients were diagnosed for T1DM, and the other 21 patients were T2DM. The acetone biosniffer measured 15 breath samples from the E

DOI: 10.1021/acs.analchem.7b03191 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 4. (A) Breath acetone in healthy groups and diabetic patients. (B) Breath acetone concentration in adifferent type of diabetes mellitus. (C) Breath IPA in healthy subjects and diabetic patients. (D) Breath IPA concentration in a different type of diabetes mellitus. Only T2D patients showed higher exhaled IPA concentrations. Other comparison did not observe a significant difference. (****: p < 1 × 10−6, ***: p < 0.001, **: p < 0.01, ns: no significant difference).

Figure 5. (A) Generation pathway of acetone in the liver. Acetone is metabolized from the decarboxylation of acetoacetate, which is one of the products during the fatty acid metabolism. (B) Speculated pathway of endogenic IPA generation in a human body. The polyol pathway produces NADH and consumes NAD+ that contributes to increasing the ratio of NADH/NAD+ and stimulates ADH reversely metabolize acetone to IPA.

Breath Acetone and IPA in Diabetic Patients. Figure 4A shows the comparison of breath acetone in healthy subjects and diabetic patients. The mean exhaled acetone in diabetic patients was 1207.7 ± 689.5 ppb, which was much higher than that in healthy subjects (750.0 ± 434.4 ppb) and showed a significant difference (p < 1 × 10−6). Figure 4B presents the breath acetone concentrations in different types of diabetic patients:

with Turner’s investigation, but the sex correlation is different.35 Turner et al. reported that the breath acetone levels were higher in men than in women with a significant difference (males = 558 ppb, females = 406 ppb, p = 0.02), whereas no significant difference were observed in our study. A more detailed investigation between breath acetone and sex in a wider population under dietary control is necessary. F

DOI: 10.1021/acs.analchem.7b03191 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry 1641.3 ± 845.5 ppb in T1D patients and 1121.0 ± 604.8 ppb in T2D patients. There are significant differences among the three groups: healthy subjects vs T1D patients (p < 0.001), healthy subjects vs T2D patients (p < 0.001), and T1D vs T2D patients (p < 0.01). The explanation of increasing acetone in diabetic patients and its generation pathway has been examined in many studies.40,41 Acetone is a ketone product in the metabolism of fatty acid and generated via decarboxylation of acetoacetate in the liver (Figure 5A).1,42 It is well-known that major diabetes mellitus results from the impairment of insulin secretion or insulin resistance,43 which makes the patients use fatty acid and proteins as the energy sources. Therefore, a higher concentration of breath acetone is easily found in diabetic patients. The much higher exhaled acetone concentration in T1D patients compared to that in healthy subjects is consistent with the previous study.1 T2D patients also exhaled more acetone than healthy subjects but significantly less than the T1D patients (p < 0.01). This observation may relate to the significantly more often occurrence of DKA in T1D patients.44 The results of breath IPA determination in the diabetic patients are shown in Figure 4C and D. For all patients, the breath IPA range was 1.6−87.6 ppb, and the mean was 23.1 ± 20.1 ppb, which showed a statistical difference compared to the healthy groups (p < 0.01). The separated analysis of diabetic types shows that breath IPA was higher in T2D patients (23.9 ± 21.3 ppb) and exhibited a significant difference to the healthy subjects (p < 0.01). However, no difference was observed in comparisons between healthy subjects versus T1D patients and T1D versus T2D patients. It is speculated that IPA in diabetic patients may result from a reverse metabolism of acetone and be caused by a high NADH/NAD+ ratio. The increase in NADH/NAD+ ratio in diabetes mellitus has been studied and proposed in some studies.45−47 A possible explanation is the metabolism of sorbitol. When the human body suffers from the hyperglycemic condition, blood glucose is metabolized by normal glycolysis and the polyol pathway (Figure 5B).15−17,45−48 In this pathway, blood glucose is catalyzed to sorbitol by aldose reductase; then, sorbitol dehydrogenase further metabolizes it to fructose. This process produces NADH and consumes NAD+, which increases the ratio of NADH/NAD+.45−48 Although the endogenous IPA is speculated from acetone, the correlation of IPA and acetone concentration remains unclear. Petersen et al. reported no relationship between blood acetone and IPA in post-mortem DKA patients,13 but another study showed a positive correlation between breath acetone and IPA in diabetic patients.7 Our results show a modestly positive correlation between exhaled IPA and acetone in the healthy subjects; the Pearson Correlation analysis gives r = 0.23 and p < 0.05. However, no correlation was observed among the diabetic patients (r = −0.05). This observation is consistent with previous studies on healthy subjects but not consistent with the diabetic group.7 A likely explanation of this observation is the effect of medication treatment. Li et al. show an investigation of breath acetone and IPA in the newly diagnosed (T2D patients with no surgical or medical treatment) and observed a positive correlation.7 Nonetheless, the recruited diabetic patients in our study were mostly under a regular medication treatment. Additionally, the blood examination of our T2D patients shows that the average concentration of glycated hemoglobin was approximately 7.0%, which indicates that their blood glucose are under a good control. Therefore, we suppose that the

metabolism of acetone to IPA may be affected by the blood glucose control. Another finding in our experiment is that the T1D patients exhaled much more acetone than the healthy subjects and T2D patients, but the breath IPA only observed a higher concentration in T2D patients. This observation may be explained by two reasons. First, the reverse metabolism of acetone to IPA may be more related to the NADH/NAD+ ratio than to the acetone concentration. Some researchers suggested that a high concentration of fatty acid, which is often found in T2DM and is considered one of the important factors of insulin resistance, might increase the NADH/NAD+ and acetyl-CoA/ CoA ratios.49,50 Second, this observation may simply relate to the present T1D patients because the participants appeared slightly insufficient. These results and discussion enhance the previous research of breath acetone as a biomarker for diabetes mellitus and supports the finding of the previous study on breath IPA in diabetic patients. Nevertheless, a more extensive comparison of breath acetone and IPA in T1D and T2D patients is suggested to clarify the contradictions with previous studies.



CONCLUSIONS In this study, two biosniffers, which used the forward and reverse enzymatic reactions of S-ADH and measured the change in NADH fluorescence intensity, were constructed and applied to determine the breath IPA and acetone concentrations in healthy subjects and patients with diabetes mellitus. Both acetone and IPA biosniffers showed wide calibration ranges, high sensitivities and rapid selectivity. The breath measurements indicate that both exhaled acetone and IPA may not correlate with age and sex. The breath analysis shows that diabetic patients exhaled more acetone than healthy subjects with a significant difference. In addition, both T1D and T2D patients exhaled higher acetone concentrations, and the T1D patients exhaled significantly more acetone than the others. The breath IPA detection found a higher concentration in T2D patients and showed a significant difference versus the healthy subjects, but no difference was observed between healthy subjects and T1D patients. Although the detected results in this study do not entirely support previous studies, we believe that our findings are worthwhile in the research of breath biomarkers for diabetes mellitus diagnosis. Furthermore, changing S-ADH to other enzymes enables the biosniffer to be applied to relevant VOC detection. We intend to continue pursuing and improving this gas measurement technique and further examine the disease diagnosis based on the humanemitted VOCs.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b03191. Comparison of reported acetone and IPA gas sensors to the bio-sniffers and summary of subjects’ demographics (PDF)



AUTHOR INFORMATION

Corresponding Author

*Tel: +81-3-5280-8091. Fax: +81-3-5280-8094. E-mail: m.bdi@ tmd.ac.jp. G

DOI: 10.1021/acs.analchem.7b03191 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry ORCID

(28) Paredi, P.; Loukides, S.; Ward, S.; Cramer, D.; Spicer, M.; Kharitonov, S. A.; Barnes, P. J. Thorax 1998, 53, 775−779. (29) Teshima, N.; Li, J. Z.; Toda, K.; Dasgupta, P. K. Anal. Chim. Acta 2005, 535, 189−199. (30) Kinoyama, M.; Nitta, H.; Watanabe, A.; Ueda, H. J. Health Sci. 2008, 54, 471−477. (31) Van den Velde, S.; Nevens, F.; Van Hee, P.; van Steenberghe, D.; Quirynen, M. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2008, 875, 344−348. (32) Libardoni, M.; Stevens, P. T.; Waite, J. H.; Sacks, R. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2006, 842, 13−21. (33) Rudnicka, J.; Kowalkowski, T.; Ligor, T.; Buszewski, B. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2011, 879, 3360−3366. (34) Kischkel, S.; Miekisch, W.; Sawacki, A.; Straker, E. M.; Trefz, P.; Amann, A.; Schubert, J. K. Clin. Chim. Acta 2010, 411, 1637−1644. (35) Turner, C.; Spanel, P.; Smith, D. Physiol Meas 2006, 27, 321− 337. (36) Nasution, T. I.; Nainggolan, I.; Hutagalung, S. D.; Ahmad, K. R.; Ahmad, Z. A. Sens. Actuators, B 2013, 177, 522−528. (37) Biswal, R. C. Sens. Actuators, B 2011, 157, 183−188. (38) Li, S.-H.; Chu, Z.; Meng, F.-F.; Luo, T.; Hu, X.-Y.; Huang, S.-Z.; Jin, Z. J. Alloys Compd. 2016, 688, 712−717. (39) Wu, Y.; Wen, F.; Liu, D.; Kong, H.; Zhang, C.; Zhang, S. Luminescence 2011, 26, 125−129. (40) Krebs, H. A. Adv. Enzyme Regul. 1966, 4, 339−354. (41) Reichard, G. A., Jr.; Haff, A. C.; Skutches, C. L.; Paul, P.; Holroyde, C. P.; Owen, O. E. J. Clin. Invest. 1979, 63, 619−626. (42) van Stekelenburg, G. J.; Koorevaar, G. Clin. Chim. Acta 1972, 39, 191−199. (43) Alberti, K. G. M. M.; Zimmet, P. Z.; Consultation, W. Diabetic Med. 1998, 15, 539−553. (44) Laffel, L. Diabetes/Metab. Res. Rev. 1999, 15, 412−426. (45) Rolo, A. P.; Palmeira, C. M. Toxicol. Appl. Pharmacol. 2006, 212, 167−178. (46) Nyengaard, J. R.; Ido, Y.; Kilo, C.; Williamson, J. R. Diabetes 2004, 53, 2931−2938. (47) Williamson, J. R.; Chang, K.; Frangos, M.; Hasan, K. S.; Ido, Y.; Kawamura, T.; Nyengaard, J. R.; van den Enden, M.; Kilo, C.; Tilton, R. G. Diabetes 1993, 42, 801−813. (48) Gabbay, K. H.; et al. N. Engl. J. Med. 1973, 288, 831−836. (49) Randle, P. J. Diabetes/Metab. Rev. 1998, 14, 263−283. (50) Le Marchand-Brustel, Y.; Gual, P.; Gremeaux, T.; Gonzalez, T.; Barres, R.; Tanti, J. F. Biochem. Soc. Trans. 2003, 31, 1152−1156.

Kohji Mitsubayashi: 0000-0002-1555-1281 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was partly supported by Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Number 26280053, by Japan Science and Technology Agency (JST) and by Ministry of Education, Culture, Sports, Science and Technology (MEXT). And a financial supporting for Po-Jen Chien by Tokyo Medical and Dental University Scholarship (Sony Corporation supported).



REFERENCES

(1) Wang, Z. N.; Wang, C. J. J. Breath Res. 2013, 7, 037109. (2) Yamada, Y.; Hiyama, S.; Toyooka, T.; Takeuchi, S.; Itabashi, K.; Okubo, T.; Tabata, H. Anal. Chem. 2015, 87, 7588−7594. (3) Rydosz, A. Metabolites 2014, 4, 921−931. (4) Deng, C. H.; Zhang, J.; Yu, X. F.; Zhang, W.; Zhang, X. M. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2004, 810, 269−275. (5) Turner, C.; Walton, C.; Hoashi, S.; Evans, M. J. Breath Res. 2009, 3, 046004. (6) Righettoni, M.; Schmid, A.; Amann, A.; Pratsinis, S. E. J. Breath Res. 2013, 7, 037110. (7) Li, W.; Liu, Y.; Liu, Y.; Cheng, S.; Duan, Y. RSC Adv. 2017, 7, 17480−17488. (8) Yan, Y.; Wang, Q.; Li, W.; Zhao, Z.; Yuan, X.; Huang, Y.; Duan, Y. RSC Adv. 2014, 4, 25430−25439. (9) Slaughter, R. J.; Mason, R. W.; Beasley, D. M. G.; Vale, J. A.; Schep, L. J. Clin. Toxicol. 2014, 52, 470−478. (10) Brugnone, F.; Perbellini, L.; Apostoli, P.; Bellomi, M.; Caretta, D. Occup. Environ. Med. 1983, 40, 160−168. (11) Slaughter, R. J.; Mason, R. W.; Beasley, D. M.; Vale, J. A.; Schep, L. J. Clin. Toxicol. 2014, 52, 470−478. (12) Dalziel, K.; Dickinson, F. M. Biochem. J. 1966, 100, 34−46. (13) Petersen, T. H.; Williams, T.; Nuwayhid, N.; Harruff, R. J. Forensic Sci. 2012, 57, 674−678. (14) Palmiere, C.; Sporkert, F.; Werner, D.; Bardy, D.; Augsburger, M.; Mangin, P. Leg. Med. 2012, 14, 17−20. (15) Davis, P. L.; Dalcortivo, L. A.; Maturo, J. J. Anal. Toxicol. 1984, 8, 209−212. (16) Dwyer, J. B.; Tamama, K. Clin. Chim. Acta 2013, 415, 245−249. (17) Jones, A. W.; Holmgren, A. Toxicologie Analytique et Clinique 2015, 27, 226−232. (18) Ying, W. Antioxid. Redox Signaling 2008, 10, 179−206. (19) McKenna, M. C.; Waagepetersen, H. S.; Schousboe, A.; Sonnewald, U. Biochem. Pharmacol. 2006, 71, 399−407. (20) Kim, K. H.; Jahan, S. A.; Kabir, E. TrAC, Trends Anal. Chem. 2012, 33, 1−8. (21) Choi, S.-J.; Lee, I.; Jang, B.-H.; Youn, D.-Y.; Ryu, W.-H.; Park, C. O.; Kim, I.-D. Anal. Chem. 2013, 85, 1792−1796. (22) Xiao, T.; Wang, X.-Y.; Zhao, Z.-H.; Li, L.; Zhang, L.; Yao, H.-C.; Wang, J.-S.; Li, Z.-J. Sens. Actuators, B 2014, 199, 210−219. (23) Righettoni, M.; Tricoli, A.; Gass, S.; Schmid, A.; Amann, A.; Pratsinis, S. E. Anal. Chim. Acta 2012, 738, 69−75. (24) Shin, J.; Choi, S.-J.; Lee, I.; Youn, D.-Y.; Park, C. O.; Lee, J.-H.; Tuller, H. L.; Kim, I.-D. Adv. Funct. Mater. 2013, 23, 2357−2367. (25) Ye, M.; Chien, P. J.; Toma, K.; Arakawa, T.; Mitsubayashi, K. Biosens. Bioelectron. 2015, 73, 208−213. (26) Chien, P. J.; Suzuki, T.; Tsujii, M.; Ye, M.; Toma, K.; Arakawa, T.; Iwasaki, Y.; Mitsubayashi, K. Biosens. Bioelectron. 2017, 91, 341− 346. (27) Kudo, H.; Sawai, M.; Wang, X.; Gessei, T.; Koshida, T.; Miyajima, K.; Saito, H.; Mitsubayashi, K. Sens. Actuators, B 2009, 141, 20−25. H

DOI: 10.1021/acs.analchem.7b03191 Anal. Chem. XXXX, XXX, XXX−XXX