Microfluidic immunoassay for rapid detection of cotininein saliva

Kaiping Cheng & Wang Zhao & Sixiu Liu & Guodong Sui

# Springer Science+Business Media New York 2013

Abstract A microfluidic immunoassay is successfully de-veloped for rapid analysis of cotinine saliva samples, whichis a metabolite of nicotine and is widely used as a biomarkerto evaluate the smoking status and exposure to tobaccosmoke. The core microfluidic chip is fabricated by polydi-methylsiloxane (PDMS) with standard soft lithography.Each chip is capable of eight parallel analyses of cotininesamples. The analyses can be completed within 40 min with12 μl sample consumption. The linear detection range is 1-~250 ng/ml and the minimum detectable concentration is1 ng/ml respectively. The correlation coefficient of the cali-bration curve established from standard samples is 0.9989.The immunoassay was also validated by real saliva samples,and the results showed good reproducibility and accuracy.All the results were confirmed with traditional ELISA mea-surements. The result from microfluidic chip device andELISA kits showed good correspondence, and the correla-tion coefficients are higher than 0.99. Compared with tradi-tional technique, this microfluidic immunoassay is moreeconomic, rapid, simple and sensitive, perfect for on-sitecotinine measurements as well as for the evaluation of theexposure to tobacco smoking. Moreover, this immunoassayhas potential to be applied in the analysis of other biomarkersin human saliva samples.

Keywords Cotinine . Microfluidic device . Saliva . Passivesmoking . Active smoking . ELISA

1 Introduction

It is well known that smoking is harmful to human health andenvironment, more than 4000 chemical components have beenidentified in the smoke, and about 60 of them are suspiciouscarcinogens (Adams and Hamilton 2008; Adhikari, et al. 2009;Thielen, et al. 2008). These chemicals may cause cancer(Bennett, et al. 1999; Chen, et al. 2003), apogeny (Pasinska,et al. 2009; Stekhun et al. 1979), cardiovascular (Ezzati, et al.2005; Thun, et al. 1999) and cerebrovascular diseases (Ezzati,et al. 2005; Lederle, et al. 2003;Mazzone, et al. 2010; Silvestrini,et al. 1996), and many other diseases (Braithwaite, et al. 2012;Austin, et al. 2002; Bisanovic, et al. 2011; Ueshima, et al.2004; Heitzer andMeinertz 2005). During smoking, two kindsof smoke with different composition are produced: mainstreamsmoke which is inhaled by the active smoker and sidestreamsmoke released into the environment through puffs from the litend of the cigarette (Thielen, et al. 2008). Compared withmainstream smoke, sidestream smoke was produced in largerquantities and contained 2 times more nicotine and 12 timesmore ammonia per cigarette, and higher concentrations ofcarcinogenic polycyclic aromatic hydrocarbons (Schick andGlantz 2005). So as for passive smoke (the inhalation of smokefrom other people’s cigarettes by a nonsmoker), also referred toas secondhand smoking, affects the health in the same way asregular smoking, increasing the chances of emphysema, cancerand respiratory distress (Kato, et al. 1999; Roddam, et al.2007), especially for children (Ino, et al. 2006; Bolat, et al.2012) and pregnant women (Bolat, et al. 2012; Titova, et al.2012). The personal and social problems caused by smokingand passive exposing to tobacco smoking have become aserious public issue.

Among all the harmful chemicals in cigarette, the majoraddictive substance in cigarette is nicotine (NIC), which alsois a major component in tobacco. Nicotine can be absorbedthrough mouth, nose skin and mucosal lining, or inhaled intothe lungs by both active and passive smokers (da Fonseca,

K. Cheng :W. Zhao : S. Liu :G. Sui (*)Department of Environmental Science and Engineering, Institute ofBiomedical Science, Fudan University, Shanghai 200433, People’sRepublic of Chinae-mail: gsui@fudan.edu.cn

G. SuiShanghai Tobacco Group, NO. 717, Changyang Road, Shanghai200082, People’s Republic of China

Biomed MicrodevicesDOI 10.1007/s10544-013-9786-4

et al. 2012; Cop-Blazic and Zavoreo 2009). After that, 70–80 % of nicotine is intensively metabolized in the liver andoxidized into cotinine (da Fonseca, et al. 2012; Dwoskin, et al.1999), and both of them transfer to all over human bodythrough blood circulation. So the concentration of nicotineand cotinine in human bodies can be used to evaluate expo-sure to tobacco smoke and smoking status. Identification ofcotinine levels in body fluid may provide useful informationfor nicotine replacement therapy, and smoking cessation plan-ning (Jaakkola, et al. 2003). As for passive smokers, theresearch can provide evidence for their exposure to smoke,as for them to ask for protection through legislation or rules.

The level of cotinine in the saliva, plasma and urine isproportionate to the amount of exposure to tobacco smoke(Xu et al. 2012), cotinine concentrations in saliva and plasmaare quite similar (Benowitz, et al. 2009). In addition, thein vivo half-life of cotinine is approximately 20 h, muchlonger than 2 h of nicotine, so cotinine is a widely used asreliable biomarker for evaluating tobacco smoke exposure(da Fonseca, et al. 2012). Generally, the saliva and urinesample is much more convenient to obtain, and the samplingprocess is non-invasive. Furthermore, saliva sample is easyto collect and the concentrations in it are highly correlatedwith blood and urine (da Fonseca, et al. 2012). So measuringcotinine concentration in saliva samples is one of the mostimportant methods in monitoring tobacco exposure. Thecotinine concentration less than 1 ng/ml are considered tobe associated with no active smoking. Heavy secondhandsmoke exposure can produce cotinine levels greater than10 ng/ml in nonsmokers (Benowitz, et al. 2009). Valuesbetween 10 ng/ml and 100 ng/ml are consistent with lightsmoking, which has overlap with heavy secondhand smokeexposure. For heavy smokers who smoke more than 20cigarettes a day, the values are generally over 300 ng/ml.

The accurate and sensitive measurement of cotinine can notonly be used to distinguish passive and active smokers, butalso be used to evaluate the severity of exposure to tobaccosmoking. There are several methods to detect cotinine insaliva, including immunoassays (Park, et al. 2010), colorimet-ric method (Dhar 2004), gas chromatography coupled tomass spectrometric (GC-MS) (Heinrich-Ramm, et al. 2002;Man, et al. 2006), high-performance liquid chromatographycoupled to mass spectrometric (LC-MS) (Baumann, et al.2010; Zhou, et al. 2011). The sensitivity of colorimetry meth-od is limited, not capable of analyzing samples from thepassive smokers since the cotinine concentration may belower than 10 ng/ml. The apparatus like GC, LC are expensiveand complex, requiring long analysis time and pretreatmentof the collected samples. All these methods needs complexinstruments and experienced operators, not suitable for on-site detection. There is a great need to develop a portable,high-throughput and sensitive method for rapid detection ofcotinine.

Microfluidics is the new technique developed in the pasttwo decades that can precise control and manipulate fluids inmicron size, typically in microliter, nanoliter or even picolitervolumes. Chemical and biological reactions which are tradi-tionally carried out in laboratories now can be miniaturized onthe microfluidic chip. It has advantages such as high through-put, high reaction efficiency, enhanced mobility as well asreduced sample and reagent consumption. Microfluidics havebeen utilized in many fields, particularly in the area of bio-chemical analysis (Bilitewski, et al. 2003). Many bioanalysisabout nucleic acids (Dimalanta, et al. 2004; Llopis, et al. 2007),amine acid (Guo, et al. 2005; Sandlin, et al. 2005), proteins(Choi, et al. 2002), cells (El-Ali, et al. 2006; Hao, et al. 2012)and toxin (Chiriaco, et al. 2011; Wu, et al. 2011; Zhang, et al.2011) by microfluidics technology have been reported.

Herein we propose a novel microfluidic immunoassay forrapid analysis of cotinine. The fabricated microfluidic deviceis capable of high-throughput quantification of cotinine fromsaliva samples, it can be used to evaluate the severity ofexposure to tobacco smoking and to distinguish passive andactive smokers.

2 Experimental

2.1 Chemicals and regents

PDMS (polydimethylsiloxane) (RTV 615) was purchased fromMomentive PerformanceMaterials (NY, USA). Protein A coatedmicrospheres were purchased from Bang’s Laboratory (Fishers,IN, USA). These microspheres (deameter about 9 μm) are madefrom polystyrene. Cotinine plate kit was purchased fromInnumalysis Corp. (Pomona, CA, USA). Amplex red reagentwas purchased from Invitrogen (Shanghai, China). Positive pho-toresist AZ-50XT was purchased from AZ Electronic Materials(Branchburg, NJ, USA). Negative photoresist SU8-2025 waspurchased from Microchem (Newton, MA, USA). Rabbit anticotinine was purchased from Genway Biotech (San Diego, CA,USA). Water was purified by Direct-Q system (Millipore,Bedford, MA, USA) and resistance is about 18.2 MΩ · cm. Allother reagents used were purchased from Sinopharm ChemicalReagent Co.Ltd. (Shanghai, China) at analytical grade.

2.2 Instruments

All the visible and fluorescence measurements were performedon an inverted fluorescence microscope (Nikon Ti-s) equippedwith monochrome CCD (Nikon DS-Ri1).

2.3 Microfluidic chips fabrication

The microfluidic chips were fabricated according to standardsoft lithography (Unger, et al. 2000. Zhang, et al. 2011). As

Biomed Microdevices

shown in Fig. 1, the microfluidic analytical chip includes twolayers (Fig. 1(a)): the patterned fluidic layer on the top andthe pneumatic control layer on the bottom. To create thefluidic layer and the pneumatic control layer, two differentmolds with different patterns have fabricated by photolitho-graphic processes. The mold to create the fluidic channelswas made by both positive photoresist (AZ-50 XT) andnegative photoresist (SU8 2025), while the control pneumat-ic mold was made by negative photoresist (SU8 2025). Forthe fabrication, the fluidic layer is made from PDMS (RTV615 A: B in ratio 5:1), and the pattern was transferred fromthe respective mold. The pneumatic control layer is madefrom PDMS (RTV 615 A: B in ratio 20:1). The two layerswere aligned and bonded together precisely, followed byheating at 80 °C overnight. Metal pins (23 gauge, NewEngland Small Tubing Corp.) connected with Tygon tubeswere employed in to introduce agents into the chip.

2.4 Preparation of the microspheres

The microspheres combined with secondary antibody IgGwere prepared according to the protocol. The brief proce-dures were described as following. The protein A coatedmicrospheres was washed with sodium borate, followed byaddition of IgG antibody suspension. The reaction mixturewas gently shaken for about 45 min. The microspheres werethen re-suspended in a solution of sodium borate andtriethanolamine in 9:1 ratio. A solution of 0.227 ul glutaral-dehyde dissolved in 10 ul triethanolamine was immediatelyadded to the microspheres. After 1 h, 10 ul ethanolamine wasadded into the mixture at room temperature for 10 min. Themicrospheres were then washed sequentially by 1 M NaCl,0.1 M glycine and water. The microspheres were stored at4 °C at storage buffer until being used.

Fig. 1 Schematic diagram of microfluidic chips fabricated for the rapiddetection of cotinine. Left: Two layers of the chip; Right: The pattern inthe mold and the functions of different parts

Fig. 2 a The photo of the microfluidic chip; a coin is placed nearby forsize comparison. b Illustration of the structure and functions of themicrofluidic chip. c Illustration of the structure of micro valves

b

Biomed Microdevices

2.5 Collection of human saliva samples

To confirm the reliability and practicability, actual sampleswere measured from negative and active smokers using thefabricated immunoassay, and compared with results by tradi-tional ELISA kit. For the saliva sample collection, volunteerswere instructed to accumulate in their months for about 5 min,and then expel the saliva sample into a plastic tube. Samplesare collected from 20 volunteers, in which 10 of them are non-smokers and the others are active smokers. 0.5 ml salivasamples were collected before and after smoking (for activesmokers)/exposing to smoking (for passive smokers). Salivasamples were analyzed within 4 h after collection.

3 Results and discussion

3.1 Microfluidic chip design and operation

Figure 2(a) shows the picture of the actual microfluidic chipfabricated and the size of the chip is 27×27 mm, which ismuch smaller than conventional instruments, make it possibleto significantly accelerate the detection process and de-crease the consumption of samples and reagents required

in the process. There are eight independent immunoassaycolumns in each microfluidic chip, each column can becontrolled independently by micromechanical valves, andeight samples can be analyzed simultaneously. Figure 2(b)illustrated the operation principle of the microfluidic chip.Inlet A is for reagents and buffers, and inlet B is only formicrospheres input. The inlet B can recharge two immuno-assay columns at the same time. Valve 3 is semipermeablevalve (Zhang, et al. 2011), so water can go through thevalve while the solid microspheres will be blocked insidethe column. Valve1, 2 and 4 are standard intercept valves,no fluidic or solid can go through it. The role of valve 1 and2 is to avoid water or microspheres flow back to the inlet,and valve 4 is to block the entire agent in the reactioncolumn during the reactions.

As shown in Fig. 2(c), two types of mechanical valveswere utilized in the experiment. A thin elastic PDMS mem-brane between the fluidic layer and the control layer wasfabricated. Upon applying sufficient pressure in the bottomcontrol layer, the elastic thin membrane can deform to blockthe fluid in the upper fluidic channel. Due to the profiles ofthe fluidic channel (rectangular or round shape), the valvescan be half-closed (rectangular shape) or totally closed(round shaped). The half-closed valve is semipermeable for

Fig. 3 The schematic diagram of operation procedures of themicrofluidic chip. Valve in red color represents the close state, whilevalve in gray color represents open state. Different colors of columns

represent different solutions pumped in (black: microspheres, green:wash buffer, pink: mixture solution, brown: reagent)

Fig. 4 Schematic diagram ofthe competitiveimmunoreactions in thecolumns

Biomed Microdevices

the fluid in the fluidic channel, which could be used to trap themicrospheres in the channel, while let the aqueous reagentpass though. The totally closed valve (intercept valve) couldbe used to control the reagent thoroughly.

The reaction procedures in microfluidic chip are as illus-trated in Fig. 3. The immune columns were prepared bysequentially loading microspheres (step 2), washing withPBS (step 3) and locking the column (step4). The device isready for analyzing samples after step 4. During the reaction(step 6) all valves should keep closed to make sure no fluidicleaking or interference. The last step (step 9) is fluorescencesignal detection under microscope. After the detection pro-cess is completed, the microspheres can be washed out andthe microfluidic chip can be reused.

3.2 Immunoassay scheme and fluorescence detection

The mechanism of the competitive immunoassay for cotinineanalysis in Step5 and 6 (Fig. 3) was illustrated in Fig. 4. Thiscompetitive immunoassay is based on the competitive bindingto antibody of enzyme labeled antigen and unlabeled antigen,in proportion to their concentration in the reaction mixture. Inthe detection process, the cotinine sample solution, HRP-cotinine and antibody were mixed for 5 min, and the ratio ofcotinine sample and HRP-cotinine is 1:1. For the amount of

the antibody, gradient experiments have been performed todetermine the proper amount to add. 1 μl of the mixture waspumped into the immunoassay columns to complete the com-petitive immunoreaction for 5 min. PBS was loaded to washthe unreacted solution for three times (step 7).

3.3 Detection of cotinine

When the competitive immunoassay was completed, 1 ulamplex red reagent and hydrogen peroxide (H2O2) (mixed by1:1) was pumped into the reaction channels (step 8, 9). Amplexred reagent is a highly stable and sensitive fluorogentic sub-strate for horseradish peroxidase (HRP). With the existing ofHRP, Amplex red reagent reacts with H2O2 with a 1:1 stoichi-ometry to generate the highly fluorescent product, resorufin.The less amount of antigen existed in sample solution, themore HRP-conjugated-cotinine will combine with the antibod-ies on the microspheres, so the stronger fluorescence intensitysignal will be obtained. Meanwhile, the relationship betweenthe concentration of cotinine samples and the intensity offluorescence is negative correlation.

Following step 2 to 10, the total detection time was less than40 min. Red fluorescence photos of the columns were takenusing an inverted microscope. All photos were taken under thesame objective (10X). Nikon NIS software was utilized foranalysis of the signal intensities.

3.4 Reaction time vs. the signal intensity

Generally, the incubate time for the reaction of Amplex redreagent and hydrogen peroxide with the existing of horseradishperoxidase is recommended to be 30 min or longer, but thereaction system ofmicrofluidic chip ismuch smaller and reactionspeed is generally faster than in conventional system. Cotininesample at concentration 25 ng/ml was used as a standard toinvestigate the resulting fluorescence intensity change trend vstime during step 9.

The relationship of the intensity of signal and the reactiontime was presented in Fig. 5 shows. It is clear that the intensity

Fig. 6 Photos of differentconcentration in the microscope

Fig. 5 The relationship between reaction time and average signal intensity

Biomed Microdevices

increased with time and reached maximum at about 30 min,compiled with protocol in conventional method. In order to getthe best results, 30 min was selected as the detection time instep 9 for all the experiments.

3.5 Measurement

Series of cotinine solutions were prepared to validate themicrofluidic device, and the concentrations are 1, 5, 10, 25,50, 100, 250, and 500 ng/ml respectively. The samples withthe concentration of 1 ng/ml and 250 ng/ml were dilutedfrom 5 ng/ml and 500 ng/ml respectively. Images of themicro columns for cotinine at different concentrations areshown in Fig. 6. The signal intensity decreases with theincrease of cotinine concentration. The detection and analysisprocess were repeated three times for each of the concentra-tions and the average intensity was used to create the calibra-tion curve. The calibration curve constructed from detecting ofgradient concentrations of standard samples was illustrated inFig. 7.

By taking the natural logarithm of each concentration andremove the sample of 500 ng/ml, the curve show excellent linerrange from 1 ng/ml to 250 ng/ml. The correlation coefficient ofthis method is 0.9989, much higher than ELISA kit which has

correlation coefficient at 0.984 (provided by cotinine ELISAkits instruction). The calibration curve is utilized to determinethe concentration of real samples, such as saliva from humanbodies, which will be presented in the following part.

In the analysis, the total time including reaction, washing,color development and detection is about 40 min. The size ofeach immunoreaction column is about 160 nanoliter. Sincethe column was filled with microspheres, the actual con-sumption is even less. The total volume of reaction agentprepared for this device to finish the analysis process is lessthan 12 ul. Compared with the conventional ELISA kit, thismethod can save at least 2/3 time and consumes 4 orders lessreagent.

Series of samples with concentration ranging from 0 ng/mlto 5 ng/ml were tested to determine the minimum detectableconcentration, when the concentration is lower than 1 ng/ml,the intensity of signals randomly show up, so the minimumdetectable concentration for this device is 1 ng/ml.

Fig. 8 Concentration changes of cotinine in human saliva (detected bymicrofluidic chips). a The concentration of cotinine in non-smokers’saliva samples before and after exposing to tobacco smoking. bTheconcentration of cotinine in active smokers’ saliva samples salivabefore and after smoking

Fig. 7 Standard curve of microfluidic detection

Biomed Microdevices

3.6 Detection of human saliva samples

Real saliva samples were collected to test the device, 10samples from active smokers and 10 from passive smokers.Each volunteers provided their saliva samples before and aftersmoking or exposed to tobacco smoking. Figure 8 shows thequantified concentration of cotinine in human saliva by usingmicrofluidic chips. Each sample analysis was repeated twice,and average value was used in the result. From the dataobtained, it is clear that in non-smokers saliva, nearly all theconcentrations were between 1 ng/ml to 15 ng/ml beforeexposing to smoking. After exposing to environmental tobac-co smoking (ETS), the concentration is between 7 ng/ml to22 ng/ml, the value increased for each individual, and theincreasing rate differs from 12.04 % to 206.59 % for differentindividuals. Heavy secondhand smoke exposure can producecotinine levels greater than 10 ng/ml in nonsmokers(Benowitz, et al. 2009). The data also suggested that evenpeople don’t expose to tobacco smoking intentional, the con-centration of cotinine in their bodies can demonstrate that theyare passive smokers, which means in public places, the phe-nomenon of exposing to ETS is inevitable. While as for theactive smokers, the concentration is much higher than non-smokers. Before smoking, the concentration of cotinine insaliva is between 50 ng/ml to 250 ng/ml for different individ-uals, and the value changed dramatically after they smoked.And the highest value is 355.3 ng/ml. Compared with passivesmokers, the concentration of cotinine in active smokers is atleast higher than 50 ng/ml, which can definitely prove that theyare active smokers.

To verify the results obtained from the microfluidic chipdevice, ELISA kit was used to measure the same samples. Theresults from two different detection methods were shown inFig. 9 in parallel manner. Figure 9(a) showed the data com-parison between two different methods for non-smokers be-fore exposing to ETS. For the results from microfluidic chip,the concentration is from 4.62 ng/ml to 16.99 ng/ml for

Fig. 10 Concentration of cotinine in active and passive smokers’ saliva

Fig. 9 Result comparison of cotinine concentration in human salivabetween ELISA kits and microfluidic chips a For non-smokers beforeexposing to ETS. b For non-smokers after exposing to ETS. c For activesmokers before smoking. d For active smokers after smoking

R

Biomed Microdevices

different individuals, while for the results from ELISA kit, theconcentration is from 4.4 ng/ml to 17.2 ng/ml. Under the sameexperiment conditions, results from microfluidic chip deviceare very close to those from ELISA kit. There is no significantdifference between two data groups and the correlation coef-ficient between two assays is 0.9915. Similar conclusions]can be made from data presented in Fig. 9(b), (c) and (d).Figure 9(b) showed the data comparison between two differ-ent methods for non-smokers after exposing to ETS, Fig. 9(c)showed the data comparison for active smokers beforesmoking, while Fig. 9(d) showed the data comparison foractive smokers after smoking. It is clear that the data frommicrofluidic chip device and ELISA kits show good corre-spondence, and the correlation coefficients are higher than0.99 for all the experimental setup. These results demonstratethat the microfluidic immunoassay has good reproducibility,calibration curve and accuracy.

Figure 10 clearly shows differences of cotinine concen-tration in passive and active smokers’ saliva. Generally forpassive smokers, the concentration of cotinine is lower than50 ng/ml; while for active smokers, the valve is higher than50 ng/ml. Active and passive smokers can be easily discrim-inated using the microfluidic device. The definition of pas-sive smokers is that the concentration of cotinine in theirbody is higher than 1 ng/ml, and the minimum detectableconcentration of the microfluidic chip is 1 ng/ml as well.

4 Conclusions

A simple and sensitive method employing microfluidic chipsto determine the concentration of cotinine in saliva sampleswas successfully developed. The device was fully validatedby lab prepared samples and real samples. Compared totraditional detection method, this microfluidic immunoassayis economic and the cost to perform one saliva sampleanalysis is less than 5 dollar. The time to perform parallel8 analyses is about 40 min, and consumption of reagents isonly about 12 ul. The minimum detection concentration is1 ng/ml and detection is accurate. The correlation coefficientof the calibration curve established from standard samples is0.9989, which is better than Elisa kit (0.984). The realsample analysis results showed great correspondence be-tween microfluidic chips and Elisa kits. It is believed to besuitable for on-site and rapid cotinine detection. By utilizingthis device, passive smokers and active smokers can beeasily distinguished for its accurate cotinine analysis. Theanalysis results of cotinine levels in body fluid will offersome evidence for medical diagnose or help to judge thehealth issues of populations, such as nicotine replacementtherapy, and smoking cessation planning.

Above all, this new method may be used for the evaluationof environmental tobacco smoking exposure, and will provide

support information for nicotine cessation programs, help toidentify the risks of smoking related diseases. Furthermore,this microfluidic immunoassay has the potential be used in theanalysis of other biomarkers in human saliva samples.

Ackonwledgment This study was supported by Shanghai TobaccoGroup SZBCW2010, NSFC (20975025), Cultivation Fund of the KeyScientific and Technical Innovation Project, Ministry of Education ofChina (708031), and Specialized Research Fund for the Doctoral Pro-gram of Higher Education (SRFDP 20090071110006).

References

E.R. Adams, P.B. Hamilton, Future Microbiol. 3, 167 (2008)B.R. Adhikari et al., Jpn. J. Infect. Dis. 62, 212 (2009)A.S. Austin, R.F.A. Logan, K. Thomason, G.K.T. Holmes, Scand. J.

Gastroenterol. 37, 978 (2002)F. Baumann, R. Regenthal, I.L. Burgos-Guerrero, U. Hegerl, R. Preiss,

J. Chromatogr. B-Anal. Technol. Biomed. Life Sci. 878, 107(2010)

W.P. Bennett et al., J. Nat. Cancer Inst. 91, 2009 (1999)N.L. Benowitz, J.T. Bernert, R.S. Caraballo, D.B. Holiday, J.T. Wang,

American J. Epidemiol. 169, 236 (2009)U. Bilitewski, M. Genrich, S. Kadow, G. Mersal, Anal. Bioanal. Chem.

377, 556 (2003)S. Bisanovic, B. Mehic, S. Sivic, Bosnian J. Basic Med. Sci. 11, 46

(2011)F. Bolat et al., Turkish J. Med. Sci. 42, 999 (2012)D. Braithwaite et al., Breast Cancer Res. Treat. 136, 521 (2012)Z.M. Chen et al., Int. J. Cancer 107, 106 (2003)M.S. Chiriaco et al., Lab on a Chip 11, 658 (2011)J.W. Choi et al., Lab on a Chip 2, 27 (2002)N. Cop-Blazic, I. Zavoreo, Acta Clinica Croatica 48, 371 (2009)B.M. da Fonseca et al., J. Chromatogr. B-Anal. Technol. Biomed. Life

Sci. 889, 116 (2012)P. Dhar, J. Pharm. Biomed. Anal. 35, 155 (2004)E.T. Dimalanta et al., Anal. Chem. 76, 5293 (2004)L.P. Dwoskin, L.H. Teng, S.T. Buxton, P.A. Crooks, J. Pharmacol. Exp.

Ther. 288, 905 (1999)J. El-Ali, P.K. Sorger, K.F. Jensen, Nature 442, 403 (2006)M. Ezzati, S.J. Henley, M.J. Thun, A.D. Lopez, Circulation 112, 489

(2005)Y.L. Guo, K. Uchiyama, T. Nakagama, T. Shimosaka, T. Hobo, Elec-

trophoresis 26, 1843 (2005)L. Hao, C.X. Xu, H.Y. Cheng, J.H. Liu, X.F. Yin, Prog. Chem. 24, 1544

(2012)R. Heinrich-Ramm, R. Wegner, A.H. Garde, X. Baur, Int. J. Hyg.

Environ. Health 205, 493 (2002)T. Heitzer, T. Meinertz, Z. Kardiol. 94, 30 (2005)T. Ino, T. Shibuya, K. Saito, J. Ohshima, R. Okada, Prev. Med. 42, 427

(2006)M.S. Jaakkola et al., Prev. Med. 36, 282 (2003)M. Kato et al., J. Hypertens. 17, 1395 (1999)F.A. Lederle, D.B. Nelson, A.M. Joseph, J. Vasc. Surg. 38, 329 (2003)S.L. Llopis, J. Osiri, S.A. Soper, Electrophoresis 28, 984 (2007)C.N. Man, L.H. Gam, S. Ismail, R. Lajis, R. Awang, J. Chromatogr. B-

Anal. Technol. Biomed. Life Sci. 844, 322 (2006)P. Mazzone et al., Int. J. Environ. Res. Public Health 7, 4111 (2010)S. Park, D.H. Lee, J.G. Park, Y.T. Lee, J. Chung, Clin. Chim. Acta 411,

1238 (2010)M. Pasinska, G. Przybylski, O. Haus, G. Ludwikowski, Med. Pr. 60,

117 (2009)

Biomed Microdevices

A.W. Roddam et al., Br. J. Cancer 97, 434 (2007)Z.D. Sandlin, M.S. Shou, J.G. Shackman, R.T. Kennedy, Anal. Chem.

77, 7702 (2005)S. Schick, S. Glantz, Nicotine Tob. Res. 7, 591 (2005)M. Silvestrini, E. Troisi, M. Matteis, L.M. Cupini, G. Bernardi, J.

Cereb. Blood Flow Metab. 16, 746 (1996)F. I. Stekhun, Vestnik Dermatologii I Venerologii, 61 (1979)A. Thielen, H. Klus, L. Muller, Exp. Toxicol. Pathol. 60, 141 (2008)M. Thun, J. Henley, L. Apicella, Environ. Heal. Perspect. 107, 841 (1999)O.E. Titova et al., Br. J. Nutr. 108, 1341 (2012)

H. Ueshima et al., Stroke 35, 1836 (2004)M.A. Unger, H.P. Chou, T. Thorsen, A. Scherer, S.R. Quake, Science

288(5463), 113–116 (2000)C.S. Wu, P.B. Lillehoj, L. Sabet, P. Wang, C.M. Ho, Biotechnol. J. 6,

150 (2011)X. Xu & Z. H. Fan, Electrophoresis 33, (2012)J.L. Zhang, S.X. Liu, P.Y. Yang, G.D. Sui, Lab on a Chip 11, 3516

(2011)W.H. Zhou, Y. Wang, L. Yu, Y.M. Jin, S.J. Zheng, Acta Chim. Sin. 69,

803 (2011)

Biomed Microdevices