Quantitative analysis of clonidine and ephedrine by a microfluidic system: On-chip electromembrane extraction followed by high performance liquid chromatography

Accepted Manuscript Title: Quantitative analysis of clonidine and ephedrine by a microfluidic system: On-chip electromembrane extraction followed by high performance liquid chromatography Authors: Mahroo Baharfar, Yadollah Yamini, Shahram Seidi, Monireh Karami PII: DOI: Reference:

S1570-0232(17)31157-1 https://doi.org/10.1016/j.jchromb.2017.10.062 CHROMB 20892

To appear in:

Journal of Chromatography B

Received date: Revised date: Accepted date:

4-7-2017 20-10-2017 31-10-2017

Please cite this article as: Mahroo Baharfar, Yadollah Yamini, Shahram Seidi, Monireh Karami, Quantitative analysis of clonidine and ephedrine by a microfluidic system: Onchip electromembrane extraction followed by high performance liquid chromatography, Journal of Chromatography B https://doi.org/10.1016/j.jchromb.2017.10.062 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Quantitative analysis of clonidine and ephedrine by a microfluidic system:

1

On-chip electromembrane extraction followed by high performance liquid

2

chromatography

3 4

Mahroo Baharfara, Yadollah Yaminia,*, Shahram Seidib, Monireh Karamia

5

Department of Chemistry, Faculty of Sciences, Tarbiat Modares University, P.O. Box: 14115-175, Tehran,

6

Iran

7

Department of Analytical Chemistry, Faculty of Chemistry, K.N. Toosi University of Technology, Tehran, Iran

8

a

b

9 10 11

Highlights     

A microfluidic device was developed for on-chip electromembrane extraction. It was applied for extraction ephedrine and clonidine from urine and plasma samples. The separation and determination of the analytes were performed by HPLC-UV. The limits of detection were less than 7.0 and 11 µg L-1 in urine and plasma samples. Smaller distance between electrodes makes it possible to apply low applied voltages.

12 13 14 15 16 17 18 19 20

Abbreviations:

21

CLO: clonidine; DEHP: di-(2-ethylhexyl) phosphate; EME: electromembrane extraction; EPH:

22

ephedrine; ER: extraction recovery; CCD: central composite design; HF-LPME: hollow fiber

23

liquid phase microextraction; HPLC: high performance liquid chromatography; LDR: linear

24

dynamic range; LLE: liquid-liquid extraction; LOD: limit of detection; NPOE: 2-nitrophenyl

25

octyl ether; PF: preconcentration factor; PPMA: polymethyl methacrylate; RR: relative

26

1

recovery; RSD: relative standard deviation; SLM: supported liquid membrane; TEHP: tris-(2-

27

ethylhexyl) phosphate.

28 29 30 31

Abstract In this work, a microfluidic device was developed for on-chip electromembrane extraction

32

of trace amounts of ephedrine (EPH) and clonidine (CLO) in human urine and plasma samples

33

followed by HPLC-UV analysis. Two polymethylmethacrylate (PMMA) plates were used as

34

substrates and a microchannel was carved in each plate. The microchannel channel on the

35

underneath plate provided the flow pass of the sample solution and the one on the upper plate

36

dedicated to a compartment for the stagnant acceptor phase. A piece of polypropylene sheet

37

was impregnated by an organic solvent and mounted between the two parts of the chip device.

38

An electrical field, across the porous sheet, was created by two embedded platinum electrodes

39

placed in the bottom of the channels which were connected to a power supply. The analytes

40

were converted to their ionized form, passed through the supported liquid membrane (SLM),

41

and then extracted into the acceptor phase by the applied voltage. All the effective parameters

42

including the type of the SLM, the SLM composition, pH of donor and acceptor phases, and

43

the quantity of the applied voltage were evaluated and optimized. Several organic solvents were

44

evaluated as the SLM to assess the effect of SLM composition. Other parameters were

45

optimized by a central composite design (CCD). Under the optimal conditions of voltage of 74

46

volts, flow rate of 28 μL min-1, 100 and 20 mM HCl as acceptor and donor phase composition,

47

respectively, the calibration curves were plotted for both analytes. The limits of detection

48

(LODs) were less than 7.0 and 11 µg L-1 in urine and plasma, respectively. The linear dynamic

49

ranges (LDR) were within the range of 10-450 and 25-500 µg L-1 (r2˃0.9969) for CLO, and

50

within the range of 20-450 and 30-500 µg L-1 (r2˃0.9907) for EPH in urine and plasma,

51

2

respectively. To examine the capability of the method, real biological samples were analyzed.

52

The results represented a high accuracy in the quantitative analysis of the analytes with relative

53

recoveries within the range of 94.6-105.2 % and acceptable repeatability with relative standard

54

deviations lower than 5.1%.

55

Keywords: Microfluidic device; On-chip electromembrane extraction; Ephedrine;

56 57

Clonidine; Biological samples.

3

58

1. Introduction Clonidine (CLO), chemically known as N-(2,6-Dichlorophenyl)-4,5-dihydro-1H-imidazol-

59

2-amine, is an imidazoline compound which has been prescribed as an antihypertensive

60

pharmaceutical for patients suffering from cardiovascular problems. This molecule exerts its

61

effect through binding to an α2-adrenergic receptor, a receptor which activates

62

neurotransmitters like norepinephrine to rise the blood pressure, with noticeable affinity to

63

control it. Moreover, this drug is usually used to treat attention-deficit and hyperactivity

64

disorders in children. The therapeutic dosage of this pharmaceutical is within the range of 75-

65

150 µg in urine and it would be valuable to determine its concentration in biological fluids of

66

patients being treated [1-5].

67

Ephedrine (EPH), 2-methylamino-1-phenylpropan-1-ol, is another medication extracted

68

from a plant called Ephedra sinica that acts as a sympathomimetic stimulant on central nervous

69

system to prevent low blood pressure in cardiovascular diseases and hypotension caused by

70

anesthesia [6]. This drug also affects adrenergic receptors so that it increases blood pressure

71

and heart rate. In addition, this stereoisomer is a natural alkaloid existing in green leaf tea and

72

some botanical supplements used along in combination with caffeine as an anti-obesity agent

73

by bodybuilders. However, adverse cardiovascular effects or death caused by misuse of this

74

drug have been reported [7]. Basically, the most amount of ephedrine remains unchanged in

75

biological samples such as urine, making it possible to determine this molecule in such fluids

76

[8].

77

Quantitative analysis of drugs in complicated matrices such as urine, saliva, plasma, and

78

blood samples is a formidable challenge. This fact is attributed to the presence of a vast variety

79

of contaminants in biological samples and low concentration of analytes of interest in these

80

samples, obscuring the analysis of target compounds. Therefore, selecting a suitable sample

81

preparation technique prior to the quantitative analysis of the target analytes is an essential step

82

4

in order to reduce the matrix effects, eliminate contaminants, preconcentrate the analytes, and

83

convert the sample into a compatible form with the analytical instrument [9,10].

84

Over the past years, various sample preparation techniques have been proposed and

85

developed to address the mentioned challenges. For example liquid-liquid extraction (LLE) is

86

one of the well-known conventional sample preparation methods which was widely utilized

87

prior to the quantitative analysis via analytical instruments [15,16]. Numerous innovative

88

techniques such as membrane-based liquid phase microextraction techniques, in which

89

analytes are extracted through a supported liquid membrane (SLM), have been derived from

90

this method [17]. In these techniques, a pH gradient termed hollow fiber liquid phase

91

microextraction (HF-LPME) [18] or an electrical filed called electromembrane extraction

92

(EME) [17] is applied as the driving force for the extraction of the target analytes.

93

Recently, the advent of microfluidic devices has made sample preparation techniques more

94

advantageous thanks to their remarkable features such as minimizing the cost, the amount of

95

required sample, and hazardous organic solvents. Besides, considering short diffusion

96

distances, microfluidic devices introduce rapid analysis and make the extraction process more

97

efficient on microfluidic platforms, which are attributed to the high surface-to-volume ratio

98

[11-13]. It is also noteworthy that automation, integration, and parallelization are more feasible

99

by these devices [14].

100

EME was introduced by Pedersen-Bjergaard et al. [19], comprising many advantages

101

compared with other similar techniques including rapid extractions, efficient sample cleanup,

102

and dispense with sample pretreatment [20]. Up to now, several developments have been

103

recommended to make EME more applicable [2-23]. Among them, performing EME on a chip-

104

based device is the most attractive one which goes back to 2010 [24]. Afterwards, numerous

105

interesting designs of this miniaturized method were published in the literature [25-27, 13].

106

5

In the present study, the advantages of electromembrane extraction and a microfluidic

107

device were combined and an on-chip EME was designated for the extraction and

108

preconcentration of ephedrine and clonidine from human urine and plasma samples. This

109

device, in which an electrical field is applied along the whole length of the dedicated extraction

110

channels, reduces power consumption, due to shorten the distances between the electrodes. It

111

also offers considerable extraction efficiency which can create a new way for designing

112

portable and analytically efficient microfluidic devices.

113

Also a central composite design was applied for the optimization of the effective parameters

114

on the extraction efficiency of the utilized on-chip EME procedure. Finally, applicability of the

115

method was successfully investigated in real urine and plasma samples.

116

2. Experimental

117

2.1. Chemical and reagents

118

Ephedrine (EPH) and clonidine (CLO) were kindly donated by Department of Pharmacy,

119

Tehran University (Tehran, Iran). The structure and corresponding physicochemical features

120

of the drugs can be seen in Table 1. HPLC grade methanol and acetonitrile were provided from

121

Daejung Chemicals and Metals (Siheung-city, South Korea). 2-Nitrophenyl octyl ether

122

(NPOE), tris-(2-ethylhexyl) phosphate (TEHP) and di-(2-ethylhexyl) phosphate (DEHP) were

123

obtained from Fluka (Buchs, Switzerland). 1-Octanol was purchased from Merck (Darmstadt,

124

Germany). All chemicals were of analytical reagent grade. The Accurel 2E HF (R/P)

125

polypropylene membrane sheet with a thickness of 150 mm, and a pore size of 0.2 mm was

126

purchased from Membrana (Wup- pertal, Germany). The water used in this work was purified

127

by a Younglin 370 series aqua MAX purification instrument (Kyounggi-do, Korea). Stock

128

solutions of EPH and CLO were prepared at the concentration of 1.0 mg mL-1 in ultra-pure

129

water. Standard solutions were prepared from the stock solutions by sufficient dilution. All of

130

the standard solutions were stored at 4 ºC and protected from light.

131

6

132

2.2. Real samples Plotting calibration curves and evaluating figures of merit were performed in plasma and

133

urine matrices. Urine and plasma samples were collected from volunteers. Sampling was

134

carried out based on the guidelines for research ethics and protocol was approved by an internal

135

review board. The urine samples were filtered by a 0.45 μm pore size cellulose acetate filter

136

provided from Milipore (Madrid, Spain). In order to prevent bacterial growth, the filtrate was

137

stored at 4 ºC in a clean glass vial. Two milliliter of each urine sample was spiked with a proper

138

amount of the mixed standard solution to obtain the desirable concentration and the pH of the

139

samples was adjusted. Plasma samples were obtained from Iranian Blood Transfusion

140

Organization (Tehran, Iran). The samples were stored in sterilized bottles at -4 ºC, thawed,

141

shaken, diluted 1:5, and their pH adjusted prior to use.

142

2.3. Chromatographic apparatus

143

The separation and detection of the two analytes were carried out using an Agilent 1260

144

HPLC system equipped with a quaternary pump, degasser, a 20 μL sample loop and UV-Vis

145

detector (Waldbornn, Germany). Results were recorded and analyzed by ChemStation for LC

146

system software (version B.04.03). The separations were accomplished on an ODS-3 column

147

(25 mm × 4.6 mm, with 5 μm particle size) provided from MZ-Analysenteknik (Maniz,

148

Germany). The separation of EPH and CLO was performed by an isocratic elution at the flow

149

rate of 1.0 mL min-1. The mobile phase constituents were acetonitrile and a 10 mmol L-1

150

phosphate buffer with a pH of 4.5 (80:20, v/v). The detection and quantification of both

151

analytes were carried out at the wavelength of 210 nm.

152

2.4. Preparation of chip for electromembrane extraction

153

Two polymethylmethacrylate plates (PMMA) were used as the substrates since its low price

154

and ease of fabrication by milling methods makes it the best choice. A long channel was carved

155

on each plate to provide the location of the donor and acceptor phases. The channels were 30

156

7

mm long, 500 μm deep and 1.0 mm wide. The structure of the chip is shown in Fig. 1. The

157

upper channel was dedicated for the stagnant acceptor phase and the lower channel was

158

exploited for the donor phase flow pass. Three holes were drilled for providing inlet and outlet

159

tubes and entering the electrodes. As is shown in the figure, holes a and b were connected to

160

the inlet and outlet tubes and hole c was used to mount the platinum electrodes (all holes had

161

I.D. of 0.5 mm). The platinum electrodes (provided from Pars Platin, Tehran, Iran), with a

162

diameter of 0.2 mm, were bent and located through the whole length of channels. All channels

163

and wholes were milled by the aid of a SMG-302 CNC micromilling machine from Sadrafan

164

Gostar Industries (Tehran, Iran).

165

A small piece of porous propylene sheet, impregnated by a proper organic solvent, was

166

located between the two parts of the chip to separate the donor and acceptor channels and the

167

whole device was fixed by bolts and nuts. The membrane sheet was replaced with a new one

168

for each extraction. Additionally, an external syringe pump from Fanavaran Nano-Meghyas

169

(Tehran, Iran) was exploited to flow the donor phase during the extraction procedure. The

170

acceptor phase, with a microliter volume, was introduced and withdrawn by a microsyringe in

171

each extraction.

172

2.5.On chip electromembrane extraction procedure

173

A piece of propylene sheet, with the dimension of 3 mm × 4 cm, was cut and dipped in

174

NPOE containing 10% (v/v) DEHP to impregnate the organic solvent into the pores of the

175

sheet. The excess amount of the organic solvent was wiped out by a piece of Kleenex. The

176

membrane sheet was mounted between the two parts of the chip device. Two milliliters of the

177

donor phase containing the target analytes was withdrawn into a syringe located on the syringe

178

pump and pumped through the related channel. Thirty five microliters of 100 mM HCl as the

179

acceptor phase were introduced into the upper channel of the chip device by a microsyringe.

180

After fulfillment of the extraction, the acceptor phase was collected by a microsyringe and

181

8

analyzed by HPLC-UV. After each extraction, the channels of the device were carefully

182

washed by ultrapure water and methanol.

183

2.6. Calculation of preconcentration factor, extraction recovery, and relative recovery

184

The preconcentration factor (PF) was defined as the ratio of the final analyte concentration

185

in the acceptor phase (Cf,a) to the initial concentration of analytes in the sample solution (Ci,s):

186

PF 

C f ,a

(1)

187

where Cf,a was determined according to a calibration graph obtained from the direct injection

188

of EPH and CLO standard solutions. The extraction recovery (ER%) was defined as the

189

percentage of the mole numbers of analyte extracted into the acceptor phase (nf,a) to that

190

originally present in the sample solution (ni,s).

191

Ci , s

ER % 

n f ,a

100 

C f ,a V f , a

100

(2)

192

where Vf,a and Vi,s indicate the volume of the acceptor phase and sample solution, respectively.

193

Relative recovery (RR) was calculated from the following equation:

194

RR% 

ni , s

C found  Creal

Ci , s Vi , s

100

(3)

195

where Cfound, Creal and Cadded represent the concentration of the analyte after adding a known

196

amount of the standard into the real sample, the concentration of analyte in the real sample,

197

and the concentration of a known amount of the standard spiked into the real sample,

198

respectively.

199

2.7. Data analysis and statistical methods

200

Cadded

In order to ascertain the optimal conditions for on-chip EME of EPH and CLO, central

201

composite design (CCD) was used. For this goal, Design-Expert software trial version 10.0

202

9

(Stat-Ease Inc., MN, USA) was utilized to generate an experimental matrix and evaluate the

203

results.

204

3. Results and discussion

205

3.1. Type of supported liquid membrane

206

The chemical nature and composition of the supported liquid membrane is a highly

207

influential factor affecting EME. There are several criteria which make an organic solvent a

208

suitable SLM for the EME procedure; these include certain electrical conductivity to provide a

209

low level of current, high permeability to make electrokinetic migration of the analytes feasible,

210

immiscibility in water, compatibility with propylene membrane sheet and less toxicity. 1-

211

Octanol and NPOE are the organic solvents which provide the mentioned requirements and are

212

frequently used in the EME. Moreover, it has been reported that the addition of ion-pairing

213

reagents to the SLM may present a beneficial effect on the EME performance [28]. To evaluate

214

this effect, 1-octanol, NPOE, NPOE containing 5, 10 and 15% (v/v) DEHP or TEHP and also

215

a mixture of NPOE with both DEHP and TEHP at various ratios were investigated. The results

216

are shown in Fig. 2A. As can be seen, 10% (v/v) DEHP in NPOE provided the highest

217

extraction efficiencies and then it was selected as the optimum SLM for further studies.

218

3.2. Results from central composite design (CCD)

219

In order to achieve optimal conditions to perform extractions, central composite design

220

(CCD) was utilized. Central composite design (CCD) covers factorial points, center points and

221

axial (star) points. The design included 20 experiments with four central points in random

222

order.

223

There are several factors affecting the on-chip EME efficiency that are as follows: the SLM

224

composition, flow rate, the applied voltage, and pH of the donor and acceptor phases. By a

225

given separate assessment, the best SLM composition was selected and the rest of the effective

226

10

parameters were optimized by taking advantage of experimental methods and reducing the

227

number of runs.

228

In the EME procedure, the applied electrical field is the driving force for migration of the

229

ionized form of the analytes. The extraction efficiencies increase by increasing the applied

230

voltage, but at high values of the applied voltage several problems which deteriorate system

231

efficiency are occurred. A few such problems include electrolysis and bubble formation, Joule

232

heating and also SLM punctuation [21]. These interfering processes are highly dependent on

233

the sample solution matrix.

234

For this reason, initial studies were carried out and the stability of the system was

235

investigated with several extraction voltages in urine and plasma matrices. The upper limit of

236

the voltage was determined before generating the experimental matrix and 90 V was chosen as

237

the highest voltage limit at which the system showed better stability. Notwithstanding, at higher

238

voltages, bubble formation and instability of the extraction system were apparent for the

239

biological matrices.

240

For further reduction in the number of runs, HCl concentration in the donor and the acceptor

241

phases were merged as a single parameter of ion balance (χ), which was defined as the total

242

ionic concentration in sample solution to that in the acceptor solution. For this purpose, in all

243

experiments, the concentration of HCl in the acceptor phase was maintained constant at 100

244

mM HCl and the corresponding concentration in the donor phase was varied between 0 to 100

245

mM HCl. The acceptor phase composition was selected based on initial experiments, in which,

246

as is shown in Fig. 3B, 100 mM HCl showed highest extraction efficiency as the acceptor phase

247

solution. Increasing the HCl concentration in the acceptor phase leads to enhancement of the

248

extractability due to facilitating the release of the analytes at the interface between the SLM

249

and the acceptor phase. On the other hand, more increasing of HCl concentration has a negative

250

effect on the recovery since it raises the electrolysis product, bubble formation and decreasing

251

11

the repeatability [17]. Therefore, a concentration of 100 mM HCl was a suitable choice as the

252

acceptor phase. Consequently, the ion balance parameter was kept in the range of 0-1.

253

Design matrix variable is presented in Table 2. An experimental response in each run

254

corresponds to the sum of peak areas. The acquired data were analyzed using analysis of

255

variance (ANOVA). A p-value less than 0.05 indicates the statistical significance of an effect

256

at 95% confidence level. The equation defining the final response for the effect of different

257

parameters on the extraction efficiencies is:

258

Sum of peak area = +361.45 + 17.71×A – 62.50×B – 23.46×C - 36.60×A×C + 66.80×B×C +

259

62.86×B2

260

(4)

Table 3 shows the ANOVA table in which the flow rate or the extraction time (B) is the most

261

influential factor in the on-chip EME extraction efficiency. According to Table 3, the model is

262

significant and there was not a significant lack of fit at the 95% confidence level.

263

Response surface methodologies (RSMs) have been widely used to assess the effect of

264

independent variables on the system performance. Graphical relationships between the

265

effective factors can be obtained via RSMs and it is a way to reach the exact optimum values.

266

Also, two-dimensional contour plots based on the model equations can be shown for each

267

response surface. These contour plots express the interactions between independent variables.

268

In Table 4, the coefficients of the studied model are displayed and, according to the given table,

269

the coefficient of determination of the model is 0.8320 indicating a good correlation between

270

the response and the model.

271

Total response surfaces are shown in Fig. 3. As a total result, the extraction efficiency of the

272

target analytes increased by increasing the experimental factors to the certain values and it then

273

gradually decreased. This observation can be interpreted by the fact that mass transfer and,

274

consequently, extraction efficiency increase by increasing the applied voltage and time in

275

EME. However, reduction in the response by further increases in the applied voltage and

276

12

extraction time can be ascribed to the instability problems and the non-exhaustive nature of

277

EME. In addition, formation of the mass transfer barriers, built-up boundary layers at both

278

sides of SLM which are mainly generated by hydrochloric acid ions, back extraction process

279

due to an increase in the pH of the acceptor phase by electrolysis reactions, and saturation of

280

the acceptor phase with the target analytes may also be responsible for the observed decline in

281

the extraction efficiency.

282

In EME, the analytes should exist as their ionized form to migrate under the electrical filed.

283

For basic analytes, an acidic medium improves the conversion of analytes into their ionized

284

form and thus increases the extraction efficiency. On the other hand, by increasing the acid

285

concentration in the donor phase, proton ions can compete with the analytes for migration into

286

the acceptor phase, which can cause a decrease in the extraction efficiency. Also, the

287

probability of Joule heating and electrolysis reactions in both the donor and the acceptor phases

288

increase. Considering the whole results, the optimized values were: a voltage of 74 V and a

289

flow rate of 28 µL min-1 and 20 mM HCl as the donor phase.

290

3.3. Method validation

291

To evaluate the applicability of the proposed method for the extraction of the target drugs

292

from real samples, drug-free human urine and plasma samples were spiked and analyzed under

293

the optimal conditions. Linear dynamic range (LDR), limit of detection (LOD), limit of

294

quantification (LOQ), preconcentration factor (PF) and extraction recoveries (ER%) were

295

calculated. Moreover, intra- and inter-assay RSDs% were calculated based on six replicate

296

measurements at the concentration level of 150 μg L-1 to evaluate the method precision. The

297

results are summarized in Table 5.

298

PF values were higher than 18 and 12 in urine and plasma samples, respectively. The method

299

showed good linearity with determination coefficient (R2) values higher than 0.9907 within the

300

concentration range of 25-500 μg L-1 and 30-500 μg L-1 for CLO and EPH in plasma samples,

301

13

respectively. The calibration curves in urine samples were linear with the R2 values higher than

302

0.9960 over the range of 10-450 μg L-1 and 20-450 μg L-1 for CLO and EPH, respectively.

303

LODs less than 11 μg L-1 and 7.0 μg L-1 were achieved in plasma and urine samples for both

304

analytes, respectively. The intra- and inter-day RSDs% were less than 6.1% and 8.2% for the

305

analytes in both matrices and indicated the acceptable precision of the method for the analysis

306

of CLO and EPH in plasma and urine samples which is owing to the fixed position of electrodes

307

and providing a homogeneous electrical field whole along the channel length of the device. In

308

addition, the application of the electrical field along the extraction channels caused acceptable

309

extraction efficiency and sensitivity, regarding the low volume of biological sample.

310

Table 7 provides a comparison between the proposed method and other works reported in

311

the literature for the quantitative analysis of EPH and CLO. As can be seen, the obtained results

312

by the chip device are completely comparable with conventional techniques. In comparison of

313

extraction efficiency of the proposed method with the conventional EME, the extraction

314

efficiency has been increased in plasma samples and it is somewhat the same in urine samples.

315

This issue shows that this method, from this standpoint, is more advantageous. In addition,

316

sample preparation methods are not the same. In the reported data for conventional

317

electromembrane extraction of EPH, protein precipitation and dilution were accomplished

318

before extraction, and urine samples were diluted 1:6. In the present study, however, the plasma

319

samples were just diluted 1:5 and their pH was adjusted. In addition, the urine samples were

320

used just by pH adjustment. Considering the low required volumes of sample for the proposed

321

microfluidic procedure, it is perfectly advantageous in comparison with conventional methods.

322

This efficient chip-based device can be introduced as a simple, portable, and useful method for

323

the analysis of biological samples even in a few microliter volumes. This technique can be a

324

prominent substitute for conventional sample preparation methods.

325 326

14

327

3.4. Analysis of real samples In order to assess the capability of the method for quantitative analysis of the target analytes

328

in real samples, the procedure was applied for the extraction and determination of CLO and

329

EPH in urine and plasma samples. The corresponding results are illustrated in Table 4. Fig. 4

330

shows typical chromatograms of an analyte-free plasma sample before and after spiking at the

331

concentration levels of 50 μg L-1 and 250 μg L-1. Corresponding relative recoveries in plasma

332

samples were within the range of 96.6% - 105.2%, indicating the applicability of the method

333

as an efficient sample clean-up technique for the determination of the drugs in plasma samples.

334

Fig. 5 shows the obtained chromatogram of a human urine sample taken from a patient

335

treated by CLO after 9 hours. The urine samples were spiked at the concentration levels of 50

336

and 150 µg L-1 of CLO and 200 and 240 µg L-1 of EPH. The values of errors% for the urinary

337

sample ranged between -5.4 to 2.7 which show the favorable accuracy of the proposed method.

338

Also, the low amounts of RSD% values for plasma and urine samples indicate the high

339

precision of the method in quantitative analysis of biological samples.

340

4. Conclusion

341

In this work, a chip-based electromembrane extraction was developed for the analysis of

342

trace amounts of EPH and CLO in biological fluids. The effective parameters of the extraction

343

procedure were optimized using CCD. Low required sample volume, good sample clean-up,

344

acceptable sensitivity, and low LODs are the advantages of the method. The results showed

345

that EME on a chip-based device is more suited than conventional EME methods for the

346

analysis of drugs in complicated biological matrices. On the basis of the obtained results, the

347

observed enhancement in the extraction efficiency can be ascribed to the increase of the surface

348

to volume ratio and exploitation of a homogeneous electrical field along the whole channel

349

length. More importantly, the smaller distance between the electrodes makes it possible to

350

provide larger electrical fields by applying low applied voltages [26]. This feature as well as

351

15

low required sample volume for on-chip EME makes the design of portable devices for analysis

352

feasible.

353

Acknowledgements

354

The authors gratefully acknowledge financial support from Tarbiat Modares University.

355

References

356

[1] K. Hall, J. Kossowsky, T. Oberlander, T. Kaptchuk, J. Saul, V. Wyller, E. Fagermoen, D.

357

Sulheim, J. Gjerstad, A. Winger, Genetic variation in catechol-O-methyltransferase

358

modifies effects of clonidine treatment in chronic fatigue syndrome, Pharmacogenomics

359

J. 16 (2016) 454–460.

360

[2] L. Hein, J.D. Altman, B.K. Kobilka, Two functionally distinct α2-adrenergic receptors

361 362

regulate sympathetic neurotransmission, Nature 402(6758) (1999) 181-184. [3] J.E. Piletz, G.A. Ordway, H. Zhu, B.J. Duncan, A. Halaris, Autoradiographic comparison

363

of [3H]-clonidine binding to non-adrenergic sites and α2-adrenergic receptors in human

364

brain, Neuropsychopharmacology 23 (2000) 697-708.

365

[4] K. Gamache, R.K. Pitman, K. Nader, Preclinical evaluation of reconsolidation blockade by clonidine as a potential

novel treatment

for

posttraumatic stress disorder,

366 367 368

Neuropsychopharmacology 37 (2012) 2789-2796. [5] D. Atlas, Y. Burstein, Isolation and partial purification of a clonidine‐displacing

369 370

endogenous brain substance, FEBS J. 144 (1984) 287-293. [6] Y. Zheng, Y. Yang, Y. Li, L. Xu, Y. Wang, Z. Guo, H. Song, M. Yang, B. Luo, A. Zheng,

371

Ephedrine hydrochloride inhibits PGN-induced inflammatory responses by promoting IL-

372

10

373

production

and

decreasing

proinflammatory

cytokine

secretion

via

the

PI3K/Akt/GSK3β pathway, Cell. Mol. Immunol. 10 (2013) 330-337.

374

[7] R.A. Niemann, M.L. Gay, Determination of ephedrine alkaloids and synephrine in dietary supplements

by

column-switching

cation 16

exchange

high-performance

liquid

375 376

chromatography with scanning-wavelength ultraviolet and fluorescence detection, J.

377

Agric. Food Chem. 51 (2003) 5630-5638.

378

[8] G. Aymard, B. Labarthe, D. Warot, I. Berlin, B. Diquet, Sensitive determination of

379

ephedrine and norephedrine in human plasma samples using derivatization with 9-

380

fluorenylmethyl chloroformate and liquid chromatography, J. Chromatogr. B 744 (2000)

381

25-31.

382

[9] K.D. Clark, C. Zhang, J.L. Anderson, Sample Preparation for Bioanalytical and Pharmaceutical Analysis, ACS Publications, 2016.

383 384

[10] S. Seidi, Y. Yamini, M. Rezazadeh, A. Esrafili, Low-voltage electrically-enhanced

385

microextraction as a novel technique for simultaneous extraction of acidic and basic drugs

386

from biological fluids, J. Chromatogr. A 1243 (2012) 6-13.

387

[11] X. Bian, Y. Lan, B. Wang, Y.S. Zhang, B. Liu, P. Yang, W. Zhang, L. Qiao, Microfluidic

388

Air Sampler for Highly Efficient Bacterial Aerosol Collection and Identification, Anal.

389

Chem. 88 (2016) 11504-11512.

390

[12] E.K. Sackmann, A.L. Fulton, D.J. Beebe, The present and future role of microfluidics in biomedical research, Nature 507 (7491) (2014) 181-189.

391 392

[13] S. Seidi, M. Rezazadeh, Y. Yamini, N. Zamani, S. Esmaili, Low voltage electrically

393

stimulated lab-on-a-chip device followed by red-green-blue analysis: a simple and

394

efficient design for complicated matrices, Analyst 139 (2014) 5531-5537.

395

[14] S. Haeberle, R. Zengerle, Microfluidic platforms for lab-on-a-chip applications, Lab Chip

396 397

7 (2007) 1094-1110. [15] C. Grosse, I. Davis, R. Arrendale, J. Jersey, J. Amin, Determination of remifentanil in

398

human blood by liquid-liquid extraction and capillary GC-HRMS-SIM using a deuterated

399

internal standard, J. Pharm. Biomed. Anal. 12 (1994) 195-203.

400

17

[16] L. Tavakoli, Y. Yamini, H. Ebrahimzadeh, S. Shariati, Homogeneous liquid–liquid

401

extraction for preconcentration of polycyclic aromatic hydrocarbons using a

402

water/methanol/chloroform ternary component system, J. Chromatogr. A 1196 (2008)

403

133-138.

404

[17] Y. Yamini, S. Seidi, M. Rezazadeh, Electrical field-induced extraction and separation

405

techniques: promising trends in analytical chemistry–a review, Anal. Chim. Acta 814

406

(2014) 1-22.

407

[18] J. Lee, H.K. Lee, K.E. Rasmussen, S. Pedersen-Bjergaard, Environmental and

408

bioanalytical applications of hollow fiber membrane liquid-phase microextraction: a

409

review, Anal. Chim. Acta 624 (2008) 253-268.

410

[19] S. Pedersen-Bjergaard, K.E. Rasmussen, Electrokinetic migration across artificial liquid

411

membranes: new concept for rapid sample preparation of biological fluids, J. Chromatogr.

412

A 1109 (2006) 183-190.

413

[20] A. Gjelstad, S. Pedersen-Bjergaard, Electromembrane extraction: a new technique for accelerating bioanalytical sample preparation, Bioanalysis 3 (2011) 787-797.

414 415

[21] M. Rezazadeh, Y. Yamini, S. Seidi, A. Esrafili, Pulsed electromembrane extraction: a new

416

concept of electrically enhanced extraction, J. Chromatogr. A 1262 (2012) 214-218.

417

[22] M. Rezazadeh, Y. Yamini, S. Seidi, B. Ebrahimpour, Electromembrane surrounded solid

418

phase microextraction: a novel approach for efficient extraction from complicated

419

matrices, J. Chromatogr. A 1280 (2013) 16-22.

420

[23] N.J. Petersen, S.T. Foss, H. Jensen, S.H. Hansen, C. Skonberg, D. Snakenborg, J.r.P.

421

Kutter, S. Pedersen-Bjergaard, On-chip electro membrane extraction with online

422

ultraviolet and mass spectrometric detection, Anal. Chem. 83 (2010) 44-51.

423

[24] N.J. Petersen, H. Jensen, S.H. Hansen, S.T. Foss, D. Snakenborg, S. Pedersen-Bjergaard, On-chip electro membrane extraction, Microfluid. Nanofluidics 9 (2010) 881-888.

18

424 425

[25] N.J. Petersen, J.S. Pedersen, N.N. Poulsen, H. Jensen, C. Skonberg, S.H. Hansen, S.

426

Pedersen-Bjergaard, On-chip electromembrane extraction for monitoring drug metabolism

427

in real time by electrospray ionization mass spectrometry, Analyst 137 (2012) 3321-3327.

428

[26] Y.A. Asl, Y. Yamini, S. Seidi, B. Ebrahimpour, A new effective on chip electromembrane

429

extraction coupled with high performance liquid chromatography for enhancement of

430

extraction efficiency, Anal. Chim. Acta 898 (2015) 42-49.

431

[27] Y.A. Asl, Y. Yamini, S. Seidi, M. Rezazadeh, Simultaneous extraction of acidic and basic

432

drugs via on-chip electromembrane extraction, Anal. Chim. Acta 937 (2016) 61-68.

433

[28] S. Seidi, Y. Yamini, M. Rezazadeh, Combination of electromembrane extraction with

434

dispersive liquid–liquid microextraction followed by gas chromatographic analysis as a

435

fast and sensitive technique for determination of tricyclic antidepressants, J. Chromatogr.

436

B 913 (2013) 138-146.

437

[29] Z. Zhang, C. Zhang, X. Su, M. Ma, B. Chen, S. Yao, Carrier-mediated liquid phase

438

microextraction coupled with high performance liquid chromatography for determination

439

of illicit drugs in human urine, Anal. Chim. Acta 621 (2008) 185-192.

440

[30] L. Fotouhi, Y. Yamini, S. Molaei, S. Seidi, Comparison of conventional hollow fiber based

441

liquid phase microextraction and electromembrane extraction efficiencies for the

442

extraction of ephedrine from biological fluids, J. Chromatogr. A 1218 (2011) 8581-8586.

443

[31] J. Zhou, Z. Zeng, Novel fiber coated with β-cyclodextrin derivatives used for headspace

444

solid-phase microextraction of ephedrine and methamphetamine in human urine, Anal.

445

Chim. Acta 556 (2006) 400-406.

446

[32] Q. Wang, C.r. Yin, L. Xu, Optimization of hydrophilic interaction LC by univariate and

447

multivariate methods and its combination with salting‐out liquid–liquid extraction for the

448

determination of antihypertensive drugs in the environmental waters, J. Sep. Sci. 36 (2013)

449

1007-1014.

450

19

[33] C. Ghosh, R.P. Singh, S. Inamdar, M. Mote, B.S. Chakraborty, Sensitive, selective, precise

451

and accurate LC–MS method for determination of clonidine in human plasma,

452

Chromatographia. 69 (2009) 1227-1232.

453 454

[34] https://chemicalize.com

455 456 457 458 459 460 461 462 463 464 465 466 467 468

Figure captions

469

Fig. 1. A schematic of chip structure.

470

Fig. 2. Effect of A) SLM composition B) composition of the acceptor phase on the extraction

471

efficiencies. Analytes were extracted across SLM from 0 mM HCl sample solution by applied

472

voltage of 60 V and flow rate of 40 μL min-1.

473

Fig. 3. Response surfaces and corresponding contour plots of sum of peak area against different

474

influential variables.

475 20

Fig. 4. Chromatograms of non-spiked (a), 50 μg L-1 (b), and 250 μg L-1 (c) spiked plasma

476

sample.

477

Fig. 5. Chromatograms resulted from extraction of CLO from non-spiked urine sample of a

478

patient treated by CLO after 9 h (a), 50 μg L-1 of CLO and 200 μg L-1 of EPH (b), 150 μg L-1

479

of CLO and 240 μg L-1 of EPH (c) spiked urine sample.

480 481

21

Table 1 Chemical structure and corresponding pKa and Log KO/W values of the analytes [35] Name Chemical structure pKa Log KO/W

Ephedrine

9.52

1.32

Clonidine

8.16

2.49

482 483

484 485

22

Table 2 Design matrix of desired factors and related response (sum of peak areas) Run A: Voltage B: Flow rate C: Ion balance (χ) Sum of peak area 1 26 28 0.8 424.1 2 74 82 0.2 287.0 3 26 82 0.8 329.4 4 74 28 0.2 593.1 5 26 82 0.2 235.6 6 74 82 0.8 397.9 7 50 55 0.5 281.3 8 50 55 0.5 312.6 9 50 55 0.5 295.4 10 74 28 0.8 273.2 11 50 55 0.5 316.5 12 26 28 0.2 433.9 13 50 100 0.5 477.9 14 10 55 0.5 356.7 15 50 55 0.5 394.8 16 50 10 0.5 703.4 17 90 55 0.5 424.3 18 50 55 0.5 450.1 19 50 55 0 497.8 20 50 55 1 381.6

23

486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511

Table 3 512 Analysis of variance (ANOVA) of quadratic model to predict the increase in extraction 513 efficiencies 514 Factor SS df MSS F P Model 1.695E+005 6 28252.30 15.86 ˂0.0001 (Significant) A 4284.76 1 4274.76 2.41 0.1469 B 53352.71 1 53352.71 29.95 0.0001 C 7518.02 1 7518.02 4.22 0.0624 AC 10718.02 1 10718.02 6.02 0.0304 BC 35694.85 1 35694.85 20.04 0.0008 B2 57945.48 1 57945.48 32.53 ˂0.0001 Residual 21374.12 12 1781.18 Lack of fit 19052.86 8 2381.61 4.10 0.0944 (Not significant) Pure error 2321.26 4 580.32 Corrected total 2.516E+005 19 SS: sum of square; df: degree of freedom; MSS: mean sum of square; F: Fisher value; p values <0.05 were considered to be significant, where A: voltage, B: flow rate, C: ion balance.

24

515 516 517

Table 4 Regression coefficients and standard errors (SE) of model elements Codded term Coefficient of regression (a) SE Intercept (a0) 361.45 12.25 A 17.71 11.42 B -62.50 11.42 B2. 62.86 11.02 C -23.46 11.42 AC -36.60 14.92 BC 66.80 14.92 2 R -adjusted: 0.8320

518 519

SE: standard error; A: voltage, B: flow rate, C: ion balance

520

25

521

Table 5 Analytical performance of on-chip EME-HPLC/UV for determination of EPH and CLO from urine and plasma samples LOD (μg L-1) LOQ (μg L-1) Linearity (μg L-1) R2 PFa ER% Plasma Urine a b

EPH CLO EPH CLO

11.0 8.0 7.0 3.0

30.0 25.0 20.0 10.0

30.0-500 25.0-500 20.0-450 10.0-450

0.9907 0.9969 0.9960 0.9996

12 13 18 19

21 23 32 34

522 523

RSD%b Inter-assay Intra-assay 6.1 8.2 5.0 6.1 5.7 7.2 4.5 5.3 524 525

Preconcentration factor at 150 μg L-1 Based on six replicate measurements at 150 μg L-1

26

526 527 528 529

Table 6 Determination of EPH and CLO in real samples using on-chip EME method Sample Plasma

Urine

a b

Analyte EPH

Creal (μg L-1) nda

CLO

nda

EPH

nda

CLO

159.4

Cadded (μg L-1) 250.0 50.0 250.0 50.0

Cfound (μg L-1) 259.1 52.6 260.1 48.3

RSD%b 5.1 5.4 4.3 4.8

Error% 3.6 5.2 4.0 -3.4

200.0 240.0 50.0 150.0

205.4 245.1 206.7 313.2

4.6 4.3 4.1 3.9

2.7 -2.1 -5.4 2.5 530 531

Not detected Based on six replicate measurements

27

532

Table 7 A comparison of extraction method with other proposed techniques for the extraction and determination of desired drugs. Method Analyte Matrix LOD LDR R2 RSD% ER% -1 -1 (µg L ) (µg L ) LPME/HPLC-UVa EPH Urine 50 100-10000 0.999 5.0 b HF-LPME/HPLC-UV EPH Urine 60 100-3000 0.991 7.5 10 Plasma 200 250-4000 0.988 8.6 2 c EME/HPLC-UV EPH Urine 5 15-750 0.994 5.2 34 Plasma 10 30-1000 0.993 7.3 14 d LLE/LC-UV EPH Plasma 2-300 0.998 3.0 HS-SPME/GC-FIDe EPH Urine 0. 33 20-20000 0.999 3.9 8 f SO-LLE/LC-UV CLO Water 1.9 10-1000 0.999 <7.0 g LLE/LC-MS CLO Plasma 0.25 0.47-73.98 >0.99 <9.3 OC-EME/HPLC-UVh CLO Urine 3 10-450 0.9996 4.5 34 Plasma 8 25-500 0.9969 5.0 23 EPH Urine 7 20-450 0.9960 5.7 32 Plasma 11 30-500 0.9907 6.1 21

533 534

Extraction time (min) 15 25

Ref.

15

[30]

40 35 71 71 71 71

[8] [31] [32] [33] This work

[29] [30]

535 536 537 538 539 540 541 542

a

Liquid-phase microextraction-liquid chromatography ultraviolet detection. Hollow fiber liquid-phase microextraction. c Electromembrane extraction. d Liquid-liquid extraction. e Headspace solid-phase microextraction–gas chromatography flame ionization. f Salting-out liquid-liquid extraction. g Liquid-liquid extraction-liquid chromatography mass spectrometry detection. h On-chip electromembrane extraction. b

28

543 544

Fig. 1

545 546

29

547

Fig. 2

548 549

30

550

Fig. 3

31

551

Fig. 4

552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 32

568

Fig. 5

569 570

33