- Email: [email protected]
Bianalyte multicommutated flow analysis system for microproteinuria diagnostics
Talanta 148 (2016) 707–711
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Bianalyte multicommutated flow analysis system for microproteinuria diagnostics$, $$ Kamil Strzelak a,b,n, Jagoda Misztal a, Łukasz Tymecki a, Robert Koncki a a b
University of Warsaw, Department of Chemistry, Pasteura 1, 02-093 Warsaw, Poland University of Warsaw, MISMaP College, Al. Żwirki i Wigury 93, 02-089 Warsaw, Poland
art ic l e i nf o
a b s t r a c t
Article history: Received 10 February 2015 Received in revised form 24 March 2015 Accepted 4 April 2015 Available online 17 April 2015
In this work a bianalyte multicommutated flow analysis (MCFA) system for determination of microproteinuria is presented. The developed MCFA system is based on two dedicated optoelectronic flowthrough detectors which allow estimation of urinary protein creatinine ratio. For total protein determination, turbidimetric Exton's method was used, whereas creatinine was determined by the photometric Jaffe reaction. The developed analytical system is fully-mechanized, easy to operate, economic in reagent consumption and characterized by satisfactory analytical parameters. It allows protein determination in the range 36–300 mg L 1 with 33 mg L 1 detection limit and simultaneous determination of creatinine in the range 0.045–2.50 mmol L 1 with 0.025 mmol L 1 detection limit. The measurement procedure for the presented MCFA system offers performing 30 peaks per hour for both analytes. To prove the analytical usefulness of the system, real human urine samples have been analyzed. The correlation and agreement between results offered by the developed system and clinical analyzers are fully acceptable. & 2015 Elsevier B.V. All rights reserved.
Keywords: Microproteinuria Protein creatinine ratio Multicommutation Paired emitter detector diode
1. Introduction Various kidneys' disorders could be diagnosed by quantitative determination of selected analytes in urine [1]. Because of urine composition fluctuations during the day, in urinalysis the 24 h urine collection is the most accurate way of receiving representative sample for analysis. However, such approach is time-consuming, inconvenient and often imprecise due to collection errors. It has been shown that incorrect urine collection concerns more than 25% of analyzed samples [2]. Thus, the alternative approach is to combine the analyte concentration measurement with simultaneous determination of renal biomarker such as creatinine. The expression of target analyte level as a relative to the creatinine concentration is used to correct the fluctuations in urine volume and the degree of dilution. Such approach has recently been carried out using several modern analytical techniques: sequential injection analysis [3] and lab-on-a-chip [4] for albumin determination, multicommutated flow analysis for ☆ A part of this work was presented and awarded in the course of XIX ICFIA (Fukuoka, 2014). ☆☆ Selected papers presented at The 19th International Conference on Flow Injection Analysis and Related Techniques and Related Techniques, Fukuoka, Japan, November 30–December 5, 2014. n Corresponding author at: University of Warsaw, Department of Chemistry, Pasteura 1, 02-093 Warsaw, Poland. E-mail address: [email protected] (K. Strzelak).
http://dx.doi.org/10.1016/j.talanta.2015.04.021 0039-9140/& 2015 Elsevier B.V. All rights reserved.
bilirubin determination [5], cross injection analysis for iron determination [6], liquid chromatography for determination of uric acid [7], as well as tryptophan and its metabolites [8]. The mentioned methodology refers also to the very important parameter of kidney disorder, which is used in diagnostics of the proteinuria: the spot urine protein creatinine ratio (PCR index). In this case, measurements of a random sample are well correlated with the 24 h urinary protein excretion and can therefore be used to quantify proteinuria without the need of time-consuming urine collection [9,10]. Under physiological conditions, up to 150 mg/L of proteins is excreted in urine. The concentrations that exceed this value are assigned to the pathological state of proteinuria [11]. It has been noticed that in many renal diseases the specific proteins can appear in low concentrations. Moreover, greater amounts of some proteins, which in physiological conditions are in trace levels, can be excreted. Such pathology cannot be revealed by the preliminary routine strip tests, because the concentration of total proteins do not exceed the critical value of 150 mg L 1 (in this case standard proteinuria measurements give negative results). Such situation refers to the state of subclinical proteinuria (microproteinuria) which is an early marker not only of progressive kidney damage [12] and diabetic nephropathy [13] but also circulatory system disorder [14], inflammatory bowel disease [15] and even cadmium accumulation symptom [16]. It has to be mentioned that albuminuria is a part of proteinuria, which is a broader concept. For early stage of renal damages, the concentration of albumins (in physiological conditions up to 20 mg/L)
708
K. Strzelak et al. / Talanta 148 (2016) 707–711
in urine samples is in good correlation with the total protein concentration and can be used alternately [12]. It is suitable especially in regions with limited resources because of the economy of total protein's determination in comparison to expensive immunochemical methods of albumins' determination [17]. In this communication, a bianalyte multicommutated flow analysis (MCFA) system dedicated for microproteinuria diagnostics is presented. This MCFA system is based on two miniaturized flowthrough detectors operating according to paired emitter detector diode (PEDD) principle described by Shockley equation and the Lambert–Beer law [18]. PEDD detectors, consisting of LED light emitter and the second LED playing the role of light detector, generate potentiometric signal which is proportional to the absorbance of sample and thereby to the concentration of analyte. It has been experimentally confirmed [19] that the reading of signal generated by PEDD using low input impedance voltmeter results in the significant enhancement of sensitivity caused by partial discharging of illuminated LED detector. Such approach causes a narrower range of linearity but the appropriate choice of power supplying LED-emitter allows to adjust the range of maximal sensitivity of device to the required range of detection. Obviously, the measurement setup based on low-budget voltmeter is more economic than this one based on pH-meter. For total protein determination according to Exton method a newly developed turbidimetric PEDD detector [20,21] is applied. Reference creatinine levels are determined with Jaffe method using photometric PEDD detector [22]. In this work the design, analytical performance and biomedical utility of the developed MCFA system based on two mentioned optoelectronic detectors allowing simultaneous determination of both analytes are presented.
2. Experimental Creatinine (as hydrochloride) has been obtained from SigmaAldrich (USA). Lyophilized bovine serum albumin (BSA), sulfosalicylic acid (SSA), picric acid (PA) and other reagents of analytical grade were obtained from POCh (Poland). For all experiments doubly distilled water was used throughout. For presented measurements the following solutions were used: 60 mmol/L PA, 20% (w/v) SSA and 0.8 mol/L NaOH. Analyzed samples of human urine with known concentrations of protein and creatinine were obtained from Central Clinical Laboratory of Medical University of Warsaw. These reference urine analyses were performed with clinical analyzer Cobas Integra 6000 (Roche Diagnostics). Light emitting diodes (LEDs) used to construct PEDDs (505 nm and 525 nm LEDs for creatinine detector and 565 nm and 600 nm LEDs for protein detector) were obtained from Optosupply (Hong Kong). All electronic components and solderless board were obtained from TME (Poland). LEDs operating as emitters of radiation (505 nm and 565 nm LEDs) were powered by stable over time currents using single, homemade low-voltage circuit based on L272 chip. Electromotive force generated by PEDDs, treated as analytical signal [19], was measured and recorded using two UNI-T multimeters (model UT70B, China) connected with data storage PC via RS232 interface. The optical flow-through cell of 60 mL internal volume integrated with respective LEDs was made of PEEK (PolyEther Ether Ketone) material that is resistant on acidic solutions. The cross-section of flow-through cell and the effect of its dimensions on analytical signal have been shown and discussed elsewhere [23]. The developed MCFA manifold has been arranged using microsolenoid pumps (stroke volume of 10 ml and 20 ml, product no. 120SP1210–4TE and 120SP1220–5TV, respectively), three-way microsolenoid valves (product no. 100T3MP12–62–5) purchased from Bio-Chem Fluidics (Boonton, USA) and PTFE Microbore tubing (ID 0.8 mm) obtained from Cole-Palmer (USA). The microsolenoid
devices were controlled by PC-programmed KSP Measuring System (Poland).
3. Results and discussion 3.1. Construction and operation of MCFA system The MCFA system developed for microproteinuria diagnostics is based on two optoelectronic detectors allowing simultaneous determination of proteins and creatinine in urine samples. PEDD constructed of 565 nm LED emitter and 600 nm LED detector is useful for turbidimetric detection of suspension formed by proteins in the SSA presence (Exton method). Photometric PEDD made of 505 nm LED emitter and 525 nm LED detector is sensitive on orange product formed in the course of reaction of creatinine with picrates in alkaline solution (Jaffe reaction). Both optoelectronic detectors are incorporated into MCFA manifold consisting of three flow modules shown in Fig. 1. The role of sample delivery module (SM) is to inject specified volume of a sample into flow system. As in conventional injection valve, the sample volume is defined by injection loop between two solenoid microvalves (V1 and V2). The sample's segment is aspirated by the operation of solenoid micropump P2. Such design enables dosage of constant injection volume which affects repeatability of analytical signals. Then, the segment of sample is pushed by water stream using P1 pump. The symmetric arrangement of a manifold from valve V2 to PEDDs (the same dimensions of a tubing) provides a split of a sample segment into two equal parts which are introduced into other parts of the manifold: creatinine determination module (CM) and protein determination module (PM). The basis of a operation of two remaining modules is to transport the reagent segments (picrate solution for CM and SSA solution for PM) and mix them with sample segments before they reaches the detection system, where PEDD signals are recorded simultaneously. To maintain the same pressure at the outlets of manifold, the stroke volume of P1 (20 ml per impulse) is twice greater than pumps P3 and P4 (10 ml per impulse). The program controlling the operation of developed MCFA system is given in Fig. 2. The first step is filling manifold with water by pump P1 and reagents by pumps: P4 (SSA) and P3 (alkaline picrate solution as the effect of continuous switching of valve V3 that mixes PA with sodium hydroxide. The result of this step are recordings of baseline signals. After this stage, which appears only before measurement cycle, the injection of sample into the manifold takes place. By operation of pump P2 the known volume of sample is aspirated through valve V2. The number of impulses responds to the volume of injection loop between valves V1 and V2 (dead volume of microsolenoid valves were taken into account during micropump propulsion to avoid systematic errors). After filling the injection loop with sample segment, the appropriate step of analytes determination begins. As an effect of transport of reagents and sample segments caused by pumps P1, P3 and P4, the reaction are pushed towards detectors. As a result, two peaks from both PEDD detectors appear simultaneously. The sequences of injection and obtaining an analytical signals are repeated throughout the all samples analysis. The last step finishing the procedure is flushing of system with water (P1) and NaOH solution (P3). In Fig. 2, the cycle construction based on actuation and de-actuation of microsolenoid pumps are presented. In step of recording analytical signals, impulses time is longer than in other steps and in effect the rate of fluid transport is reduced. It contributes to improving the mixing between segments of sample and reagents. Moreover, such impulse parameters were found as optimal for obtaining analytical signals (more details in Section 3.2).
K. Strzelak et al. / Talanta 148 (2016) 707–711
709
Fig. 1. The MCFA manifold (SM – sample delivery module, CM – creatinine determination module, and PM – proteins determination module).
Fig. 2. The program controlling the operation of MCFA system. Details are given in the text.
3.2. Analytical performance of MCFA system The optimization of presented MCFA system was performed using four standard solutions (BSA calibrants for total protein detection: 40, 60, 80 and 100 mg L 1 and 0.4, 0.6, 0.8, and 1 mmol L 1 calibrants for creatinine detection. The four relevant parameters were checked during studies, contribute to finding optimal conditions for bianalyte measurements: current powering LED emitters, length of reaction coils, volume of injected sample and flow rate. For both detection methods, an increase of supplying current of LED emitter causes increase of sensitivity of the method but also an increase of a background signal as an opposite effect for obtaining analytical signals [20] – above some current's value the limit of detection increases. Such effect is crucial in case of protein determination because reference value for microproteinuria is much closer to the detection limit offered by applied PEDD, then in case of creatinine measurements. 7.4 mA powering current for both detectors was found as optimal. Next parameters are directly related to MCFA manifold setup. The length's values of reaction coils, which are located before PEDD detectors, have to provide the compromise between kinetic character of applied reactions and dispersion of sample segments. The optimal values for both modules were 20 cm (the dimensions of reaction coils in CM and PM are the same to keep the equality of dynamic pressure throughout the system). Because the presented manifold is dedicated for analysis of easily available
sample (urine), the volume of a sample doesn't have to be reduced. The volume of sample segment, causing almost the saturation of PEDD signals, was found to be 0.7 mL. Also the influence of microsolenoid pump actuation, referred to the liquid transportation through the detection cells, was studied. It determines the height of a signals, thereby affecting the sensitivity of measurements. From the other hand, the dispersion effect can be observed as limitation factor. The found optimal value was 40 impulses per minute for each pump (P1 and P3 for CM and P1 and P4 for PM). Under optimal conditions, the calibrations using nine standards were performed. In Fig. 3, two recordings and corresponding calibration curves are shown. The linear response offered by the MCFA system is in the range of 36–300 mg L 1 for total protein measurements and 0.045–2.50 mmol L 1 for creatinine measurements. In both cases coefficients of determination are higher than 0.99. The sensitivities in linear ranges of determination for protein and creatinine are 0.7770.01 mV mg 1 L and 201.6070.27 mV mmol 1 L, respectively. The detection limits for total protein and creatinine determinations are 33 mg L 1 and 25 mmol L 1, respectively. The corresponding values of quantitation limits are 36 mg L 1 and 45 mmol L 1, respectively. For photometric measurements the relative standard deviations for 0.4 mmol L 1 and 1 mmol L 1 creatinine standards are 0.7% and 1.2%, respectively. In case of turbidimetric determination for 80 mg L 1 and 200 mg L 1 total protein standards, the relative standard deviations are 1.6% and 1.4%, respectively. The
710
K. Strzelak et al. / Talanta 148 (2016) 707–711
Fig. 3. The recordings from MCFA system calibration (left) and corresponding calibration graphs (right) for simultaneous determination of proteins (black) and creatinine (gray).
Fig. 4. The results of human urine analysis. The correlation graphs for proteins (A), creatinine (B) and PCR index (C).
Fig. 5. The results of human urine analysis. The Bland–Altman plots for proteins (A), creatinine (B) and PCR index (C).
K. Strzelak et al. / Talanta 148 (2016) 707–711
measurement procedure allows performing 15 measurements per hour (30 peaks per hour obtained for both analytes). The recording for creatinine detection shows that the baseline is stable over whole time of measurement. In case of protein determination, only for higher concentrations the baseline drift is observed. Most likely, it is caused by the adsorption of precipitated proteins on the surfaces of acrylic windows of PEDD detector. Such drawback is eliminated by periodical flushing of detector cell with SSA. The drift of a baseline does not affect the obtaining analytical signals (peak height). 3.3. Analysis of human urine samples The final stage of developing presented bianalytical MCFA system was the analysis of urine sample useful for microproteinuria/proteinuria diagnostics. For validation procedure, 13 samples of human urine with known concentrations of protein and creatinine were taken. The samples were analyzed in clinical laboratory. Firstly, undiluted urine samples were injected into the system. If the protein or creatinine level exceeded the upper limit of determination, then the measurements with 10–fold diluted sample were repeated. Each sample was injected in triplicate. The results of analysis obtained using investigated system and those obtained in clinical laboratory were compared. The regressions between two data sets for protein (A), creatinine (B) and PCR (C) dependences were y¼(0.89 70.01)x þ(44.74 70.29), y¼ (0.87 70.22)x þ(0.05 72.43) and y¼(1.017 0.07)x þ(4.16 73.63), respectively, with determination coefficients better than 0.94. To compare an accuracy of PCR measurements with reference methodology, the two-tail paired Student's t-tests with 12 degrees of freedom at a 95% confidence level were used. The calculated t-value was 1.946, which is lower than the tabulated one (2.179). The graphs of correlation are characterized by satisfactory coefficients of determination (Fig. 4). Moreover, to assess the agreement between presented method and reference one, the Bland–Altman plot was prepared by determining variable means on the horizontal axis and the differences plotted on the vertical axis (Fig. 5). All the data points are located within area defined by 72 SD of the mean of methods' differences. It indicates that these two methods can be used interchangeably. The data used for Bland–Altman plot corresponds with data shown in Fig. 4. Amidst the results presented in Figs. 4 and 5, five labeled points require a comment. In protein correlation plot (A), point nos. 3, 4 and 5 represent the pathological values of protein concentration in spot urine sample, greatly exceeding reference value. Other points are within the physiological reference values. In a case of creatinine dependence between developed method and reference one (B), only two samples correspond to pathological values: sample no. 1 with creatinine concentration below reference value and point “5” which refers to pathologically elevated creatinine level. All numbered points (nos. 1–5) show pathological values of PCR index (plot C). It is worth highlighting that even sample no. 2 which in both previous cases represent physiological values, has pathological value of PCR index (about 60 mg mmol 1). This point illustrates a diurnal fluctuation effect mentioned in Introduction section.
711
4. Conclusion The MCFA system for microproteinuria diagnostics has been developed. For the estimation of urinary protein creatinine ratio, the total protein determination based on Exton method and Jaffe method for determination of creatinine were used simultaneously. The developed MCFA system is fully-mechanized, easy to operate and economic in reagent consumption. The developed PEDD detectors provide satisfactory analytical parameters. To prove the analytical usefulness of the system, analysis of real urine samples has been performed. The correlation and agreement between results obtained by developed system and clinical laboratory methods were satisfactory. Therefore, it can be concluded that presented MCFA system is useful for routine clinical analysis and research investigations.
Acknowledgments Kamil Strzelak as a MISDoMP student acknowledges partial support from the EU through the European Social Fund, Contract number UDA-POKL.04.01.01-00-072/09-00. These investigations were granted by the Polish National Science Center in the frame of Opus project (Grant no. 2011/01/B/NZ5/00934).
References [1] C.C. Lin, C.C. Steng, T.K. Chuang, S. Lee, G.B. Lee, Analyst 136 (2011) 2669–2688. [2] A. Kumar, S. Kapoor, R.C. Gupta, J. Clin. Diagn. Res. 7 (2013) 622–626. [3] W. Siangproh, N. Teshima, T. Sakai, S. Katoh, O. Chailapakul, Talanta 79 (2009) 1111–1117. [4] C.C. Lin, J.L. Hsu, C.C. Steng, G.B. Lee, Microfluid. Nanofluid. 10 (2011) 1055–1067. [5] K. Ponhong, N. Teshima, K. Grudpan, J. Vichapong, S. Motomizu, T. Sakai, Talanta 133 (2015) 71–76. [6] N. Choengchan, T. Mantim, P. Inpota, D. Nacapricha, P. Wilairat, P. Jittangprasert, W. Waiyawat, S. Fucharoen, P. Sirankpracha, N. Phumala Morales, Talanta 133 (2015) 52–58. [7] Y. Zuo, Y. Yang, Z. Zhu, W. He, Z. Aydin, Talanta 83 (2011) 1707–1710. [8] J. Zhao, H. Chen, P. Ni, B. Xu, X. Luo, Y. Zhan, P. Gao, D. Zhu, J. Chromatogr. B 879 (2011) 2720–2725. [9] P. Ruggenenti, F. Gaspari, A. Perna, G. Remuzzi, Brit. Med. J. 316 (1998) 504–509. [10] A. Leanos-Miranda, J. Marquez-Acosta, F. Romero-Arauz, G.M. CardenasMondragon, R. Rivera-Leanos, I. Isordia-Salas, A. Ulloa-Aguirre, Clin. Chem. 53 (2007) 1623–1628. [11] R. Burden, C. Tomson, Clin. Med. 5 (2005) 635–642. [12] K.K. Venkatt, South. Med. J. 97 (2004) 969–979. [13] P. Martin, K.K. Hampton, C. Walton, H. Tindall, J.A. Davies, Diabet. Med. 7 (1990) 315–318. [14] F.C.T. Smith, P. Gosling, K. Sanghera, M.A. Green, I.S. Paterson, C.P. Shearman, Ann. Vasc. Surg. 8 (1994) 1–5. [15] A. Poulou, K. Goumas, D. Dandakis, I. Tyrmpas, M. Panagiotaki, A. Georgouli, D. Soutos, A. Archimandritis, World J. Gastroenterol. 12 (2006) 739–746. [16] M. Trzcinka-Ochocka, M. Jakubowski, T. Halatek, G. Razniewska, Int. Arch. Occup. Environ. Health 75 (2002) 101–106. [17] Z. Khatami, D.W. McIlveen, S.G. Nesbitt, I.S. Young, East. Mediterr. Health J. 11 (2005) 358–365. [18] Ł. Tymecki, M. Pokrzywnicka, R. Koncki, Analyst 133 (2008) 1501–1504. [19] Ł. Tymecki, R. Koncki, Anal. Chim. Acta 639 (2009) 73–77. [20] K. Strzelak, R. Koncki, Anal. Chim. Acta 788 (2013) 68–73. [21] K. Strzelak, A. Wiśniewska, D. Bobilewicz, R. Koncki, Talanta 128 (2014) 38–43. [22] Ł. Tymecki, J. Korszun, K. Strzelak, R. Koncki, Anal. Chim. Acta 787 (2013) 118–125. [23] Ł. Tymecki, K. Strzelak, R. Koncki, Anal. Chim. Acta 797 (2013) 57–63.
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Bianalyte multicommutated flow analysis system for microproteinuria diagnostics$, $$ Kamil Strzelak a,b,n, Jagoda Misztal a, Łukasz Tymecki a, Robert Koncki a a b
University of Warsaw, Department of Chemistry, Pasteura 1, 02-093 Warsaw, Poland University of Warsaw, MISMaP College, Al. Żwirki i Wigury 93, 02-089 Warsaw, Poland
art ic l e i nf o
a b s t r a c t
Article history: Received 10 February 2015 Received in revised form 24 March 2015 Accepted 4 April 2015 Available online 17 April 2015
In this work a bianalyte multicommutated flow analysis (MCFA) system for determination of microproteinuria is presented. The developed MCFA system is based on two dedicated optoelectronic flowthrough detectors which allow estimation of urinary protein creatinine ratio. For total protein determination, turbidimetric Exton's method was used, whereas creatinine was determined by the photometric Jaffe reaction. The developed analytical system is fully-mechanized, easy to operate, economic in reagent consumption and characterized by satisfactory analytical parameters. It allows protein determination in the range 36–300 mg L 1 with 33 mg L 1 detection limit and simultaneous determination of creatinine in the range 0.045–2.50 mmol L 1 with 0.025 mmol L 1 detection limit. The measurement procedure for the presented MCFA system offers performing 30 peaks per hour for both analytes. To prove the analytical usefulness of the system, real human urine samples have been analyzed. The correlation and agreement between results offered by the developed system and clinical analyzers are fully acceptable. & 2015 Elsevier B.V. All rights reserved.
Keywords: Microproteinuria Protein creatinine ratio Multicommutation Paired emitter detector diode
1. Introduction Various kidneys' disorders could be diagnosed by quantitative determination of selected analytes in urine [1]. Because of urine composition fluctuations during the day, in urinalysis the 24 h urine collection is the most accurate way of receiving representative sample for analysis. However, such approach is time-consuming, inconvenient and often imprecise due to collection errors. It has been shown that incorrect urine collection concerns more than 25% of analyzed samples [2]. Thus, the alternative approach is to combine the analyte concentration measurement with simultaneous determination of renal biomarker such as creatinine. The expression of target analyte level as a relative to the creatinine concentration is used to correct the fluctuations in urine volume and the degree of dilution. Such approach has recently been carried out using several modern analytical techniques: sequential injection analysis [3] and lab-on-a-chip [4] for albumin determination, multicommutated flow analysis for ☆ A part of this work was presented and awarded in the course of XIX ICFIA (Fukuoka, 2014). ☆☆ Selected papers presented at The 19th International Conference on Flow Injection Analysis and Related Techniques and Related Techniques, Fukuoka, Japan, November 30–December 5, 2014. n Corresponding author at: University of Warsaw, Department of Chemistry, Pasteura 1, 02-093 Warsaw, Poland. E-mail address: [email protected] (K. Strzelak).
http://dx.doi.org/10.1016/j.talanta.2015.04.021 0039-9140/& 2015 Elsevier B.V. All rights reserved.
bilirubin determination [5], cross injection analysis for iron determination [6], liquid chromatography for determination of uric acid [7], as well as tryptophan and its metabolites [8]. The mentioned methodology refers also to the very important parameter of kidney disorder, which is used in diagnostics of the proteinuria: the spot urine protein creatinine ratio (PCR index). In this case, measurements of a random sample are well correlated with the 24 h urinary protein excretion and can therefore be used to quantify proteinuria without the need of time-consuming urine collection [9,10]. Under physiological conditions, up to 150 mg/L of proteins is excreted in urine. The concentrations that exceed this value are assigned to the pathological state of proteinuria [11]. It has been noticed that in many renal diseases the specific proteins can appear in low concentrations. Moreover, greater amounts of some proteins, which in physiological conditions are in trace levels, can be excreted. Such pathology cannot be revealed by the preliminary routine strip tests, because the concentration of total proteins do not exceed the critical value of 150 mg L 1 (in this case standard proteinuria measurements give negative results). Such situation refers to the state of subclinical proteinuria (microproteinuria) which is an early marker not only of progressive kidney damage [12] and diabetic nephropathy [13] but also circulatory system disorder [14], inflammatory bowel disease [15] and even cadmium accumulation symptom [16]. It has to be mentioned that albuminuria is a part of proteinuria, which is a broader concept. For early stage of renal damages, the concentration of albumins (in physiological conditions up to 20 mg/L)
708
K. Strzelak et al. / Talanta 148 (2016) 707–711
in urine samples is in good correlation with the total protein concentration and can be used alternately [12]. It is suitable especially in regions with limited resources because of the economy of total protein's determination in comparison to expensive immunochemical methods of albumins' determination [17]. In this communication, a bianalyte multicommutated flow analysis (MCFA) system dedicated for microproteinuria diagnostics is presented. This MCFA system is based on two miniaturized flowthrough detectors operating according to paired emitter detector diode (PEDD) principle described by Shockley equation and the Lambert–Beer law [18]. PEDD detectors, consisting of LED light emitter and the second LED playing the role of light detector, generate potentiometric signal which is proportional to the absorbance of sample and thereby to the concentration of analyte. It has been experimentally confirmed [19] that the reading of signal generated by PEDD using low input impedance voltmeter results in the significant enhancement of sensitivity caused by partial discharging of illuminated LED detector. Such approach causes a narrower range of linearity but the appropriate choice of power supplying LED-emitter allows to adjust the range of maximal sensitivity of device to the required range of detection. Obviously, the measurement setup based on low-budget voltmeter is more economic than this one based on pH-meter. For total protein determination according to Exton method a newly developed turbidimetric PEDD detector [20,21] is applied. Reference creatinine levels are determined with Jaffe method using photometric PEDD detector [22]. In this work the design, analytical performance and biomedical utility of the developed MCFA system based on two mentioned optoelectronic detectors allowing simultaneous determination of both analytes are presented.
2. Experimental Creatinine (as hydrochloride) has been obtained from SigmaAldrich (USA). Lyophilized bovine serum albumin (BSA), sulfosalicylic acid (SSA), picric acid (PA) and other reagents of analytical grade were obtained from POCh (Poland). For all experiments doubly distilled water was used throughout. For presented measurements the following solutions were used: 60 mmol/L PA, 20% (w/v) SSA and 0.8 mol/L NaOH. Analyzed samples of human urine with known concentrations of protein and creatinine were obtained from Central Clinical Laboratory of Medical University of Warsaw. These reference urine analyses were performed with clinical analyzer Cobas Integra 6000 (Roche Diagnostics). Light emitting diodes (LEDs) used to construct PEDDs (505 nm and 525 nm LEDs for creatinine detector and 565 nm and 600 nm LEDs for protein detector) were obtained from Optosupply (Hong Kong). All electronic components and solderless board were obtained from TME (Poland). LEDs operating as emitters of radiation (505 nm and 565 nm LEDs) were powered by stable over time currents using single, homemade low-voltage circuit based on L272 chip. Electromotive force generated by PEDDs, treated as analytical signal [19], was measured and recorded using two UNI-T multimeters (model UT70B, China) connected with data storage PC via RS232 interface. The optical flow-through cell of 60 mL internal volume integrated with respective LEDs was made of PEEK (PolyEther Ether Ketone) material that is resistant on acidic solutions. The cross-section of flow-through cell and the effect of its dimensions on analytical signal have been shown and discussed elsewhere [23]. The developed MCFA manifold has been arranged using microsolenoid pumps (stroke volume of 10 ml and 20 ml, product no. 120SP1210–4TE and 120SP1220–5TV, respectively), three-way microsolenoid valves (product no. 100T3MP12–62–5) purchased from Bio-Chem Fluidics (Boonton, USA) and PTFE Microbore tubing (ID 0.8 mm) obtained from Cole-Palmer (USA). The microsolenoid
devices were controlled by PC-programmed KSP Measuring System (Poland).
3. Results and discussion 3.1. Construction and operation of MCFA system The MCFA system developed for microproteinuria diagnostics is based on two optoelectronic detectors allowing simultaneous determination of proteins and creatinine in urine samples. PEDD constructed of 565 nm LED emitter and 600 nm LED detector is useful for turbidimetric detection of suspension formed by proteins in the SSA presence (Exton method). Photometric PEDD made of 505 nm LED emitter and 525 nm LED detector is sensitive on orange product formed in the course of reaction of creatinine with picrates in alkaline solution (Jaffe reaction). Both optoelectronic detectors are incorporated into MCFA manifold consisting of three flow modules shown in Fig. 1. The role of sample delivery module (SM) is to inject specified volume of a sample into flow system. As in conventional injection valve, the sample volume is defined by injection loop between two solenoid microvalves (V1 and V2). The sample's segment is aspirated by the operation of solenoid micropump P2. Such design enables dosage of constant injection volume which affects repeatability of analytical signals. Then, the segment of sample is pushed by water stream using P1 pump. The symmetric arrangement of a manifold from valve V2 to PEDDs (the same dimensions of a tubing) provides a split of a sample segment into two equal parts which are introduced into other parts of the manifold: creatinine determination module (CM) and protein determination module (PM). The basis of a operation of two remaining modules is to transport the reagent segments (picrate solution for CM and SSA solution for PM) and mix them with sample segments before they reaches the detection system, where PEDD signals are recorded simultaneously. To maintain the same pressure at the outlets of manifold, the stroke volume of P1 (20 ml per impulse) is twice greater than pumps P3 and P4 (10 ml per impulse). The program controlling the operation of developed MCFA system is given in Fig. 2. The first step is filling manifold with water by pump P1 and reagents by pumps: P4 (SSA) and P3 (alkaline picrate solution as the effect of continuous switching of valve V3 that mixes PA with sodium hydroxide. The result of this step are recordings of baseline signals. After this stage, which appears only before measurement cycle, the injection of sample into the manifold takes place. By operation of pump P2 the known volume of sample is aspirated through valve V2. The number of impulses responds to the volume of injection loop between valves V1 and V2 (dead volume of microsolenoid valves were taken into account during micropump propulsion to avoid systematic errors). After filling the injection loop with sample segment, the appropriate step of analytes determination begins. As an effect of transport of reagents and sample segments caused by pumps P1, P3 and P4, the reaction are pushed towards detectors. As a result, two peaks from both PEDD detectors appear simultaneously. The sequences of injection and obtaining an analytical signals are repeated throughout the all samples analysis. The last step finishing the procedure is flushing of system with water (P1) and NaOH solution (P3). In Fig. 2, the cycle construction based on actuation and de-actuation of microsolenoid pumps are presented. In step of recording analytical signals, impulses time is longer than in other steps and in effect the rate of fluid transport is reduced. It contributes to improving the mixing between segments of sample and reagents. Moreover, such impulse parameters were found as optimal for obtaining analytical signals (more details in Section 3.2).
K. Strzelak et al. / Talanta 148 (2016) 707–711
709
Fig. 1. The MCFA manifold (SM – sample delivery module, CM – creatinine determination module, and PM – proteins determination module).
Fig. 2. The program controlling the operation of MCFA system. Details are given in the text.
3.2. Analytical performance of MCFA system The optimization of presented MCFA system was performed using four standard solutions (BSA calibrants for total protein detection: 40, 60, 80 and 100 mg L 1 and 0.4, 0.6, 0.8, and 1 mmol L 1 calibrants for creatinine detection. The four relevant parameters were checked during studies, contribute to finding optimal conditions for bianalyte measurements: current powering LED emitters, length of reaction coils, volume of injected sample and flow rate. For both detection methods, an increase of supplying current of LED emitter causes increase of sensitivity of the method but also an increase of a background signal as an opposite effect for obtaining analytical signals [20] – above some current's value the limit of detection increases. Such effect is crucial in case of protein determination because reference value for microproteinuria is much closer to the detection limit offered by applied PEDD, then in case of creatinine measurements. 7.4 mA powering current for both detectors was found as optimal. Next parameters are directly related to MCFA manifold setup. The length's values of reaction coils, which are located before PEDD detectors, have to provide the compromise between kinetic character of applied reactions and dispersion of sample segments. The optimal values for both modules were 20 cm (the dimensions of reaction coils in CM and PM are the same to keep the equality of dynamic pressure throughout the system). Because the presented manifold is dedicated for analysis of easily available
sample (urine), the volume of a sample doesn't have to be reduced. The volume of sample segment, causing almost the saturation of PEDD signals, was found to be 0.7 mL. Also the influence of microsolenoid pump actuation, referred to the liquid transportation through the detection cells, was studied. It determines the height of a signals, thereby affecting the sensitivity of measurements. From the other hand, the dispersion effect can be observed as limitation factor. The found optimal value was 40 impulses per minute for each pump (P1 and P3 for CM and P1 and P4 for PM). Under optimal conditions, the calibrations using nine standards were performed. In Fig. 3, two recordings and corresponding calibration curves are shown. The linear response offered by the MCFA system is in the range of 36–300 mg L 1 for total protein measurements and 0.045–2.50 mmol L 1 for creatinine measurements. In both cases coefficients of determination are higher than 0.99. The sensitivities in linear ranges of determination for protein and creatinine are 0.7770.01 mV mg 1 L and 201.6070.27 mV mmol 1 L, respectively. The detection limits for total protein and creatinine determinations are 33 mg L 1 and 25 mmol L 1, respectively. The corresponding values of quantitation limits are 36 mg L 1 and 45 mmol L 1, respectively. For photometric measurements the relative standard deviations for 0.4 mmol L 1 and 1 mmol L 1 creatinine standards are 0.7% and 1.2%, respectively. In case of turbidimetric determination for 80 mg L 1 and 200 mg L 1 total protein standards, the relative standard deviations are 1.6% and 1.4%, respectively. The
710
K. Strzelak et al. / Talanta 148 (2016) 707–711
Fig. 3. The recordings from MCFA system calibration (left) and corresponding calibration graphs (right) for simultaneous determination of proteins (black) and creatinine (gray).
Fig. 4. The results of human urine analysis. The correlation graphs for proteins (A), creatinine (B) and PCR index (C).
Fig. 5. The results of human urine analysis. The Bland–Altman plots for proteins (A), creatinine (B) and PCR index (C).
K. Strzelak et al. / Talanta 148 (2016) 707–711
measurement procedure allows performing 15 measurements per hour (30 peaks per hour obtained for both analytes). The recording for creatinine detection shows that the baseline is stable over whole time of measurement. In case of protein determination, only for higher concentrations the baseline drift is observed. Most likely, it is caused by the adsorption of precipitated proteins on the surfaces of acrylic windows of PEDD detector. Such drawback is eliminated by periodical flushing of detector cell with SSA. The drift of a baseline does not affect the obtaining analytical signals (peak height). 3.3. Analysis of human urine samples The final stage of developing presented bianalytical MCFA system was the analysis of urine sample useful for microproteinuria/proteinuria diagnostics. For validation procedure, 13 samples of human urine with known concentrations of protein and creatinine were taken. The samples were analyzed in clinical laboratory. Firstly, undiluted urine samples were injected into the system. If the protein or creatinine level exceeded the upper limit of determination, then the measurements with 10–fold diluted sample were repeated. Each sample was injected in triplicate. The results of analysis obtained using investigated system and those obtained in clinical laboratory were compared. The regressions between two data sets for protein (A), creatinine (B) and PCR (C) dependences were y¼(0.89 70.01)x þ(44.74 70.29), y¼ (0.87 70.22)x þ(0.05 72.43) and y¼(1.017 0.07)x þ(4.16 73.63), respectively, with determination coefficients better than 0.94. To compare an accuracy of PCR measurements with reference methodology, the two-tail paired Student's t-tests with 12 degrees of freedom at a 95% confidence level were used. The calculated t-value was 1.946, which is lower than the tabulated one (2.179). The graphs of correlation are characterized by satisfactory coefficients of determination (Fig. 4). Moreover, to assess the agreement between presented method and reference one, the Bland–Altman plot was prepared by determining variable means on the horizontal axis and the differences plotted on the vertical axis (Fig. 5). All the data points are located within area defined by 72 SD of the mean of methods' differences. It indicates that these two methods can be used interchangeably. The data used for Bland–Altman plot corresponds with data shown in Fig. 4. Amidst the results presented in Figs. 4 and 5, five labeled points require a comment. In protein correlation plot (A), point nos. 3, 4 and 5 represent the pathological values of protein concentration in spot urine sample, greatly exceeding reference value. Other points are within the physiological reference values. In a case of creatinine dependence between developed method and reference one (B), only two samples correspond to pathological values: sample no. 1 with creatinine concentration below reference value and point “5” which refers to pathologically elevated creatinine level. All numbered points (nos. 1–5) show pathological values of PCR index (plot C). It is worth highlighting that even sample no. 2 which in both previous cases represent physiological values, has pathological value of PCR index (about 60 mg mmol 1). This point illustrates a diurnal fluctuation effect mentioned in Introduction section.
711
4. Conclusion The MCFA system for microproteinuria diagnostics has been developed. For the estimation of urinary protein creatinine ratio, the total protein determination based on Exton method and Jaffe method for determination of creatinine were used simultaneously. The developed MCFA system is fully-mechanized, easy to operate and economic in reagent consumption. The developed PEDD detectors provide satisfactory analytical parameters. To prove the analytical usefulness of the system, analysis of real urine samples has been performed. The correlation and agreement between results obtained by developed system and clinical laboratory methods were satisfactory. Therefore, it can be concluded that presented MCFA system is useful for routine clinical analysis and research investigations.
Acknowledgments Kamil Strzelak as a MISDoMP student acknowledges partial support from the EU through the European Social Fund, Contract number UDA-POKL.04.01.01-00-072/09-00. These investigations were granted by the Polish National Science Center in the frame of Opus project (Grant no. 2011/01/B/NZ5/00934).
References [1] C.C. Lin, C.C. Steng, T.K. Chuang, S. Lee, G.B. Lee, Analyst 136 (2011) 2669–2688. [2] A. Kumar, S. Kapoor, R.C. Gupta, J. Clin. Diagn. Res. 7 (2013) 622–626. [3] W. Siangproh, N. Teshima, T. Sakai, S. Katoh, O. Chailapakul, Talanta 79 (2009) 1111–1117. [4] C.C. Lin, J.L. Hsu, C.C. Steng, G.B. Lee, Microfluid. Nanofluid. 10 (2011) 1055–1067. [5] K. Ponhong, N. Teshima, K. Grudpan, J. Vichapong, S. Motomizu, T. Sakai, Talanta 133 (2015) 71–76. [6] N. Choengchan, T. Mantim, P. Inpota, D. Nacapricha, P. Wilairat, P. Jittangprasert, W. Waiyawat, S. Fucharoen, P. Sirankpracha, N. Phumala Morales, Talanta 133 (2015) 52–58. [7] Y. Zuo, Y. Yang, Z. Zhu, W. He, Z. Aydin, Talanta 83 (2011) 1707–1710. [8] J. Zhao, H. Chen, P. Ni, B. Xu, X. Luo, Y. Zhan, P. Gao, D. Zhu, J. Chromatogr. B 879 (2011) 2720–2725. [9] P. Ruggenenti, F. Gaspari, A. Perna, G. Remuzzi, Brit. Med. J. 316 (1998) 504–509. [10] A. Leanos-Miranda, J. Marquez-Acosta, F. Romero-Arauz, G.M. CardenasMondragon, R. Rivera-Leanos, I. Isordia-Salas, A. Ulloa-Aguirre, Clin. Chem. 53 (2007) 1623–1628. [11] R. Burden, C. Tomson, Clin. Med. 5 (2005) 635–642. [12] K.K. Venkatt, South. Med. J. 97 (2004) 969–979. [13] P. Martin, K.K. Hampton, C. Walton, H. Tindall, J.A. Davies, Diabet. Med. 7 (1990) 315–318. [14] F.C.T. Smith, P. Gosling, K. Sanghera, M.A. Green, I.S. Paterson, C.P. Shearman, Ann. Vasc. Surg. 8 (1994) 1–5. [15] A. Poulou, K. Goumas, D. Dandakis, I. Tyrmpas, M. Panagiotaki, A. Georgouli, D. Soutos, A. Archimandritis, World J. Gastroenterol. 12 (2006) 739–746. [16] M. Trzcinka-Ochocka, M. Jakubowski, T. Halatek, G. Razniewska, Int. Arch. Occup. Environ. Health 75 (2002) 101–106. [17] Z. Khatami, D.W. McIlveen, S.G. Nesbitt, I.S. Young, East. Mediterr. Health J. 11 (2005) 358–365. [18] Ł. Tymecki, M. Pokrzywnicka, R. Koncki, Analyst 133 (2008) 1501–1504. [19] Ł. Tymecki, R. Koncki, Anal. Chim. Acta 639 (2009) 73–77. [20] K. Strzelak, R. Koncki, Anal. Chim. Acta 788 (2013) 68–73. [21] K. Strzelak, A. Wiśniewska, D. Bobilewicz, R. Koncki, Talanta 128 (2014) 38–43. [22] Ł. Tymecki, J. Korszun, K. Strzelak, R. Koncki, Anal. Chim. Acta 787 (2013) 118–125. [23] Ł. Tymecki, K. Strzelak, R. Koncki, Anal. Chim. Acta 797 (2013) 57–63.