- Email: [email protected]
An animal cell culture monitoring system using a smartphone-mountable paper-based analytical device
Accepted Manuscript Title: An animal cell culture monitoring system using a smartphone-mountable paper-based analytical device Author: Seong Hyun Im Ka Ram Kim Yoo Min Park Jae Ho Yoon Jung Woo Hong Hyun C. Yoon PII: DOI: Reference:
S0925-4005(16)30121-6 http://dx.doi.org/doi:10.1016/j.snb.2016.01.121 SNB 19632
To appear in:
Sensors and Actuators B
Received date: Revised date: Accepted date:
15-10-2015 25-1-2016 25-1-2016
Please cite this article as: Seong Hyun Im, Ka Ram Kim, Yoo Min Park, Jae Ho Yoon, Jung Woo Hong, Hyun C.Yoon, An animal cell culture monitoring system using a smartphone-mountable paper-based analytical device, Sensors and Actuators B: Chemical http://dx.doi.org/10.1016/j.snb.2016.01.121 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.
An animal cell culture monitoring system using a smartphone-mountable paper-based analytical device Seong Hyun Im1, Ka Ram Kim1, Yoo Min Park1, Jae Ho Yoon2, Jung Woo Hong2, Hyun C. Yoon1* 1
Department of Molecular Science and Technology, Ajou University, Suwon 443749, South
Korea 2
Research & Development Center, SPL Life Sciences Co. Ltd., Pocheon 487835, South Korea
* Author to whom correspondence should be addressed E-mail: [email protected]
Graphical abstract:
Highlights
A smartphone-based cell culture monitoring system was developed.
Paper-based analytical device was used for enzyme-mediated chromogenic assay.
Embedded flash light source and detector for smartphone were directly utilized.
Both glucose and lactate assays were accomplished simultaneously within a minute.
Real samples from animal cell culture were monitored easily and cost-effectively.
Abstract We developed a simple and low-cost cell culture monitoring system utilizing a paper-based analytical device (PAD) and a smartphone. The PAD simultaneously analyses glucose and lactate concentrations in the cell culture medium. Focusing on the fact that animal cells consume glucose and produce lactate under anaerobic conditions, oxidase- and horseradish peroxidase (HRP) enzyme-mediated colorimetric assays were integrated into the PAD. The PAD was designed to have three laminated layers. By using a double-sided adhesive tape as the middle layer and wax coating, a bifurcated fluidic channel was prepared to manipulate sample flow. At the inlet and the outlets of the channel, a sample drop zone and two detection zones for glucose and lactate, respectively, were positioned. When sample solution is loaded onto the drop zone, it flows to the detection zone through the hydrophilic fluidic channel via capillary force. Upon reaching the detection zone, the sample reacts with glucose and lactate oxidases (GOx and LOx) and HRP, immobilized on the detection zone along with colorless chromophores. By the Trinder’s reaction, the colorless chromophore is converted to a bluecolored product, generating concentration-dependent signal. With a gadget designed to aid the image acquisition, the PAD was positioned to the smartphone-embedded camera. Images of the detection zones were acquired using a mobile application and the color intensities were quantified as sensor signals. For the glucose assay using GOx/HRP format, we obtained the limit of detection (LOD ~ 0.3 mM) and the limit of quantification (LOQ ~ 0.9 mM) values in the dynamic detection range from 0.3 to 8.0 mM of glucose. For lactate assay using LOx/HRP, the LOD (0.02 mM) and the LOQ (0.06 mM) values were registered in the dynamic detection range from 0.02 to 0.50 mM of lactate. With the device, simultaneous analyses of glucose and lactate in cell culture media were conducted, exhibiting highly accurate and reproducible results. Based on the results, we propose that the optical sensing system developed is feasible for practical monitoring of animal cell culture.
Keywords: Paper-based analytical device, Animal cell culture, Microfluidics, Real-time analysis
1. Introduction
The efficiency of cellular production processes is directly related to the cell growth conditions. Therefore, developing methods and systems for the precise evaluation of growth conditions is becoming increasingly crucial to bioprocess engineering [1, 2]. During growth, cells located furthest from the medium-air interface are often exposed to hypoxic conditions. This is because the rate of cells digesting oxygen is faster than that of oxygen dissolving and diffusing into the medium to reach the cells [3, 4]. Oxygen-deficient conditions significantly inhibit cell growth resulting in either reduced productivity or apoptotic death. In general, a well-evaluated approach to analyzing cell growth conditions involves measuring the pH, CO2 and dissolved oxygen concentrations, and cell size, utilizing techniques such as flowinjection, near infrared (NIR) spectroscopy, and high-performance liquid chromatography (HPLC) [5–7]. Although these approaches offer high sensitivity and accuracy, they are less user-friendly owing to their expensive apparatus as well as complexity of their operations. To overcome these limitations, paper-based analytical devices (PAD) using biomarkers such as glucose and lactate have been developed, with focus on real-time analysis of the changes in biomarker concentrations during cell culture [8, 9]. Glucose is the most frequently used energy source for cell growth in culture; cells produce lactate under anaerobic conditions [10–12]. Therefore, we focused on the evaluation of glucose and lactate concentrations in cell culture media. Studies involving PADs have witnessed a recent growth owing to their cheap, simple, and easy technology that can be directly applied to commercial use [13]. Using PAD, we recently established an optical sensing system for monitoring glucose in serum to diagnose diabetes [14]. A single assay was conducted since the PAD was developed as a single channel system. However, for simultaneous assay of two different molecules in a single PAD with a single injection, here we employed a multilayered, bifurcated channel design in the PAD (Figure 1(A)) [15–17]. This PAD consists of three layers (top, middle, and bottom), where the top
layer provides sites for sample loading and detection zones for glucose and lactate. The bifurcated fluidic channel to transfer sample solution from the sample drop zone to the detection zone is positioned at the bottom layer, while these layers are attached by doubleside adhesive tape- the middle layer.
The channel was patterned on paper by employing the wax printing method [18], and later incubated in a vacuum oven at 150°C for 30 s to melt the printed wax permanently onto the paper (Figure 1(B)). The unprinted hydrophilic regions function as microfluidic channels that spontaneously transport the sample solution along the channel via capillary force [19]. For sample analysis, glucose and lactate oxidases (GOx and LOx, respectively), incorporated into HRP-mediated colorimetric assays using chromogenic substrates, were employed in the PAD for glucose and lactate measurements, respectively [20]. As shown in Figure 1(C), when a sample solution is introduced to the drop zone at the top layer, the solution is transferred to both the glucose and lactate detection zones along the fluidic channel on the bottom layer. Next, the solution participates in the HRP-mediated color-developing reaction. In the detection zone of the PAD, GOx and LOx digest glucose and lactate, respectively, generating H2O2. In the presence of H2O2, the HRP converts 4-aminoantipyrine (4-AAP) and N-ethyl-N(2-hydroxy-3-sulfopropyl)-3,5-dimethylaniline sodium salt monohydrate (MAOS) to a bluecolored product, which can be visualized in the detection zone. Its intensity is directly proportional to the concentrations of glucose and lactate in the applied sample solution (Figure 2(A) and (B)) [21-23].
For practical utilization of this cell monitoring system, a simple, reliable, and most importantly, user-friendly optical sensing device was necessary. Toward this, we developed an optical instrument employing an everyday information technology (IT) device- a smartphone. Among the high-performance sensors installed in smartphones (such as gyroscope, accelerometer, magnetometer, and proximity and light sensor), embedded chargecouple device (CCD) cameras, owing to their high resolution and definition, have emerged as promising alternatives to commercial high-end ones. Being a high-performance, user-friendly
device, the smartphone was the optimal platform to fabricate a simple optical sensing system as a signal transducer. In this study, we describe successful evaluation of glucose and lactate concentrations in cell culture media using the PAD in combination with a smartphone. By conjugating the PAD with a smartphone using a lightproof gadget, optical signal interference with the external light was optimized, making the assay suitable for real-time monitoring of glucose and lactate in the industry.
2. Experimental procedures
2.1 Materials
GOx (from Aspergillus niger), LOx (from microorganisms), and HRP were purchased from Toyobo Co., Ltd. β-D-glucose, lactate, cellulose, and 4-AAP were purchased from SigmaAldrich; while MAOS was acquired from Dojindo Molecular Technologies, Inc. RPMI-1640 (with HEPES), fetal bovine serum, and penicillin-streptomycin were purchased from Life Technologies. Filter paper was purchased from Whatman, while wax-coated channels were printed by wax print (Xerox, ColorQube 8570). Phosphate-buffered saline (PBS, pH 7.2) solution, containing 0.1 M phosphate and 0.15 M NaCl, was prepared in double-distilled and deionized water. PBST was prepared in the laboratory (50 mM PBS, 0.9% NaCl, 0.1% Triton X-100, pH 7.4). Samples of animal cell culture medium (mouse L-929 cells cultured in RPMI-1640, HEPES) were provided by SPL Life Science Co. Ltd.
2.2 Fabrication of PAD for glucose and lactate analyses
The PAD used in this study comprises three layers- top, middle, and bottom. The top and bottom layers have hydrophobic channels, with double-sided adhesive tape as the middle layer attaching the other two. For hydrophobic channels in each layer, a designed pattern was printed on filter paper by using wax printer. The printed paper was then heated in an oven at 150°C for 1 min. During this process, the wax solution was melted, which infiltrated into the porous paper, thereby forming the hydrophobic channel. The top layer contains one sampleloading zone and two detection zones, and the bottom includes the bifurcated flow channel,
which transfers the sample solution from the drop zone to each detection zone. For the colorimetric assay, 10 μL of solution containing 1 mg/mL HRP, 1 mM MAOS, 10 mM 4AAP, and oxidase was loaded onto the detection zones. GOx and LOx (10 mg/mL and 5 mg/mL, respectively) were immobilized on their corresponding detection zones. To transport the sample solution effectively, 8 μL PBST was used to pretreat the fluidic channel on the bottom layer. The surfactant, Triton X-100 in PBST, enhances hydrophilicity of the filter paper. Therefore, the sample solution rapidly moves from the loading to detection zones. The solution-treated top and bottom layers were dried for 2 h at room temperature. After attaching the prepared top layer to the middle one, 10 μL cellulose powder (2 mg/mL) was applied to the hole in the middle layer in order to fill the flow path (in our previous study, the optical signal in the detection zone was consistently developed through cellulose powder in the flow path [14]). Finally, the bottom layer was combined with the other side of the middle layer. The prepared PAD was stored at 4°C until further use.
2.3 Manipulation of smartphone gadget
The smartphone-adaptable lightproof case was constructed for complete protection from external light during the image acquisition process. A lightproof acryl case was bonded on a conventional smartphone case that covers the rear of the phone, along with a holder for the PAD and a reflection film, dispersion film, and trigonal prism. A combination of reflection and dispersion films was employed to spread light uniformly within the lightproof case during image acquisition using the smartphone-embedded light-emitting diode (LED) flash as a single light source. The optical signal from the PAD was registered on the smartphone CCD camera through the trigonal prism.
2.4 Glucose and lactate analyses employing enzymatic colorimetric assay
We prepared glucose and lactate samples in PBS (concentrations ranging from 0–16 mM and 0–1 mM, respectively), and applied 15 μL of each onto the sample-loading zone of the PAD. The resultant blue color in the detection zone of the PAD was registered every 5 s using the smartphone gadget developed. Similarly, samples of RPMI-1640, in which L-929 cells were cultured for 1, 24, 48, and 96 h, were loaded and the images recorded every 10 s. The color intensities were then quantified using the image analysis software, ImageJ.
3. Results and discussion
3.1 Principle of the smartphone-based optical sensing system
In this study, we developed an animal cell culture monitoring system using a novel PAD in combination with smartphone-embedded CCD camera as an optical transducer. Toward optical analyses of glucose and lactate, an HRP-mediated colorimetric biochemical assay was employed. It is widely known that HRP induces the transformation of colorless substrates into colored products in the presence of H2O2. GOx and LOx were employed along with HRP and the chromogen for glucose and lactate analyses, respectively, where the color intensities changed in proportion to the analytes concentration applied. These changes in the PAD were quantified using the smartphone-based optical sensing system as shown in Figure 3.
The prepared glucose and lactate solutions were applied onto the sample drop zone on the top layer of the PAD. As the dropped solutions reached the detection zones via capillary force along the fluidic channel on the bottom layer, glucose and lactate reacted with the enzymes, producing the colored product. The assayed PAD was positioned on the middle of the smartphone gadget, designed as described in Figure 3(A), and the image was registered through the trigonal prism on the embedded CCD camera in a light-protected environment. Constant and reproducible images were acquired using the LED flash as a single light source, whose homogeneous spreading was achieved with the reflection film. The dispersion film controlled the light intensity within the gadget in order to reduce sparkling effect of the water-drop in the PAD detection zone (Figure 3(B)).
3.2 Optimization of wax-printing size of the PAD detection zone (optimization of sample flow on the PAD)
In our previous study, we demonstrated a PAD-based glucose analysis principle using the
camera function of a commercial smartphone [14]. In that study, we precisely quantified the concentration of glucose in human serum. However, the principle enabled detection of only a single target at a time. An improved PAD design with bifurcated channel for simultaneous detection of glucose and lactate, therefore, has been developed in the current study. To effectively immobilize GOx and LOx in a fixed area allowing simultaneous detection of glucose and lactate on a single PAD, we designed multi-layered PAD containing one sample loading zone and two detection zones. For the sample movement to the desired area, a bifurcated fluidic channel was wax-printed in the middle layer. During the wax-printing and curing procedure, the printed size of wax pattern was slightly diminished (< 0.5 mm). To overcome this problem, we have conducted the optimization test for the wax-printing size and added margin by considering the wax diffusion and curing [14].
As shown in Figure 4, we optimized the size of the detection zone by comparing the assay results using two different sizes (3 and 4 mm diameter). We applied 15 μL of the sample solution, containing 5 mM glucose and 1 mM lactate, to the PAD prepared, and acquired images with the smartphone gadget. According to the assay results, the PAD with the 4 mmlarge detection zone displayed irregular color development for both glucose and lactate (Figure 4(A)). The chromophores appeared to be pushed ahead along the direction of flow, possibly misplaced by the excess solution that had continuously followed the reactants. For accurate analysis of glucose and lactate, an improved assay condition for uniform color development
was
required.
To
achieve
this,
we
could
either
increase
the
concentration/volume of chromophores or decrease the channel size. The latter was inducted into the design by adjusting the detection zone size from 4 mm to 3 mm. This maximized sample distribution in the detection zone, exhibiting uniform color intensities in both the glucose and lactate assays as shown in Figure 4(B). Moreover, the regular color development was stably maintained for several hours. Therefore, the 3 mm-large PAD detection zone was standardized for precise detection of glucose and lactate in the subsequent tests.
3.3 Optimization of enzymatic reaction time
In this study, glucose and lactate were analyzed based on the HRP/oxidase-associated
colorimetric principle. Generally, in enzymatic colorimetric assays, color development saturates over time. In order to detect the analyte and enhance the reliability of the developed assay principle, we optimized the time of the enzymatic reaction. Toward this, we assayed 0– 10 mM glucose and 0–0.75 mM lactate in PBS using the system developed, and registered images every 10 s using the time-lapse function of the smartphone (Figure 5).
In the glucose assay, the color intensity gradually increased for 60 s, followed by signal saturation. Although the color intensity increased over time, no significant differences were observed after 60 s (Figure 5(A)). Additionally, it was difficult to distinguish optical signals between each glucose assay within the 60 s. Based on the results, it was clear that 60 s were sufficient to effectively distinguish each glucose assay. The overall observation for the lactate assay was similar to that of glucose, with a sharp increase in color intensity for 60 s, followed by saturation. Therefore, we confirmed that lactate could also be analyzed at 60 s depending on the sensing principle (Figure 5(B)). Based on these findings, the detection time was determined to be 60 s, ideal for practical analyses of glucose and lactate.
3.4 Glucose and lactate assays using standard glucose and lactate in buffer
To examine the correlation between color intensities at the PAD detection zones and glucose and lactate concentrations. In the conventional animal cell culture, about 10 mM of glucose was used as an energy source, and only a small amount of lactate was produced in anaerobic condition [24, 25]. Based on this, glucose and lactate solutions (0–16 mM and 0–1 mM, respectively) in PBS were prepared and applied to the smartphone-based optical sensing system. To detect glucose and lactate simultaneously, the glucose solution was premixed with that of lactate in a 1:1 ratio. Here, 0 mM glucose with 1 mM lactate, and 0 mM lactate with 16 mM glucose were prepared to test the cross-reactivity of each oxidase. The color intensities for both glucose and lactate were proportional to the concentration of the analytes applied as shown in Figure 6(A) and (B).
Further, evident color differences between each assay result were observed. In the image acquired from 1 mM lactate assay, the color change was observed only in the lactate assay zone on the PAD, but not glucose detection zone. Similarly, optical signal from the 16 mM glucose assay was solely detected in the glucose detection zone. Based on these results, we verified that none of the oxidases was cross-reactive. This implied that we could selectively analyze the specific target analyte from the sample solution. The color intensities were quantified from the images acquired using ImageJ. The quantified data revealed an increase for both detection zones, proportional to each analyte concentration. The calibration curve of glucose assay using GOx/HRP catalysis format was plotted in Fig. 6(C). The curve showed that the optical signal was proportional to the glucose concentration and leveled off above 8 mM. Based on the calibration curve, the LOD and LOQ were calculated as 0.3 mM and 0.9 mM in the dynamic detection range from 0.3 to 8.0 mM of glucose. The LOD and LOQ of the calibration results were calculated as 3.3-times and 10-times the standard deviation of the background signal divided by the slope of the calibration curve, based on the ICH guidelines [26]. Meanwhile, the calibration result from the lactate assay using LOx/HRP catalysis was different from the glucose assay. In the developed enzyme-based biosensing system, the employed LOx has an affinity value to lactate (Km = 0.2 mM), which is significantly lower than that of GOx with glucose (Km = 7.0 mM) [27, 28]. Thus, the calibration signal was saturated and leveled off at the relatively low concentration of lactate, presenting a narrow dynamic range in comparison to the glucose assay. We found that the registered signal from LOx/HRP catalysis on PAD was increased in accordance with the lactate concentration, while the signal was leveled off above 0.5 mM (Fig. 6(D)). With the result, the LOD (0.02 mM) and LOQ (0.06 mM) for the lactate assay by employing the LOx/HRP bienzymatic catalysis in the dynamic detection range from 0.02 to 0.50 mM were obtained. By integrating the two bienzymatic reaction-mediated sensing including the GOx/HRP and LOx/HRP on a single PAD, the consumption of glucose as a carbon source and the production of lactate during an animal cell culture would be monitored simultaneously. To verify the accuracy of the smartphone-based assay, the same glucose and lactate samples were assayed at least thrice under the same reaction conditions (i.e., duration, temperature, and pH). Both the glucose and lactate assays exhibited coefficients of variation (COV) approximately within 5%, demonstrating high accuracy of the system. We thought that the smartphone-based optical sensing system, in combination with enzymatic colorimetric assay, would be useful for monitoring glucose and lactate levels in animal cell culture.
3.5 Glucose and lactate assays using cell culture media
To verify the applicability of the sensing system for monitoring changes in glucose and lactate concentrations in practical cell culture, L-929 cells were cultured in conventional RPMI-1640 media for 1, 24, 48, and 96 h. They were cultured without additional energy source under anaerobic conditions. For direct comparison, the media samples were also analyzed by high-performance liquid chromatography (HPLC) to determine the precise concentrations of glucose and lactate in the sample. As shown in Table 1, over time, glucose concentrations gradually reduced from 10.4 mM to 4.7 mM, while that of lactate increased from 3.3 mM to 9.1 mM.
S0925-4005(16)30121-6 http://dx.doi.org/doi:10.1016/j.snb.2016.01.121 SNB 19632
To appear in:
Sensors and Actuators B
Received date: Revised date: Accepted date:
15-10-2015 25-1-2016 25-1-2016
Please cite this article as: Seong Hyun Im, Ka Ram Kim, Yoo Min Park, Jae Ho Yoon, Jung Woo Hong, Hyun C.Yoon, An animal cell culture monitoring system using a smartphone-mountable paper-based analytical device, Sensors and Actuators B: Chemical http://dx.doi.org/10.1016/j.snb.2016.01.121 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.
An animal cell culture monitoring system using a smartphone-mountable paper-based analytical device Seong Hyun Im1, Ka Ram Kim1, Yoo Min Park1, Jae Ho Yoon2, Jung Woo Hong2, Hyun C. Yoon1* 1
Department of Molecular Science and Technology, Ajou University, Suwon 443749, South
Korea 2
Research & Development Center, SPL Life Sciences Co. Ltd., Pocheon 487835, South Korea
* Author to whom correspondence should be addressed E-mail: [email protected]
Graphical abstract:
Highlights
A smartphone-based cell culture monitoring system was developed.
Paper-based analytical device was used for enzyme-mediated chromogenic assay.
Embedded flash light source and detector for smartphone were directly utilized.
Both glucose and lactate assays were accomplished simultaneously within a minute.
Real samples from animal cell culture were monitored easily and cost-effectively.
Abstract We developed a simple and low-cost cell culture monitoring system utilizing a paper-based analytical device (PAD) and a smartphone. The PAD simultaneously analyses glucose and lactate concentrations in the cell culture medium. Focusing on the fact that animal cells consume glucose and produce lactate under anaerobic conditions, oxidase- and horseradish peroxidase (HRP) enzyme-mediated colorimetric assays were integrated into the PAD. The PAD was designed to have three laminated layers. By using a double-sided adhesive tape as the middle layer and wax coating, a bifurcated fluidic channel was prepared to manipulate sample flow. At the inlet and the outlets of the channel, a sample drop zone and two detection zones for glucose and lactate, respectively, were positioned. When sample solution is loaded onto the drop zone, it flows to the detection zone through the hydrophilic fluidic channel via capillary force. Upon reaching the detection zone, the sample reacts with glucose and lactate oxidases (GOx and LOx) and HRP, immobilized on the detection zone along with colorless chromophores. By the Trinder’s reaction, the colorless chromophore is converted to a bluecolored product, generating concentration-dependent signal. With a gadget designed to aid the image acquisition, the PAD was positioned to the smartphone-embedded camera. Images of the detection zones were acquired using a mobile application and the color intensities were quantified as sensor signals. For the glucose assay using GOx/HRP format, we obtained the limit of detection (LOD ~ 0.3 mM) and the limit of quantification (LOQ ~ 0.9 mM) values in the dynamic detection range from 0.3 to 8.0 mM of glucose. For lactate assay using LOx/HRP, the LOD (0.02 mM) and the LOQ (0.06 mM) values were registered in the dynamic detection range from 0.02 to 0.50 mM of lactate. With the device, simultaneous analyses of glucose and lactate in cell culture media were conducted, exhibiting highly accurate and reproducible results. Based on the results, we propose that the optical sensing system developed is feasible for practical monitoring of animal cell culture.
Keywords: Paper-based analytical device, Animal cell culture, Microfluidics, Real-time analysis
1. Introduction
The efficiency of cellular production processes is directly related to the cell growth conditions. Therefore, developing methods and systems for the precise evaluation of growth conditions is becoming increasingly crucial to bioprocess engineering [1, 2]. During growth, cells located furthest from the medium-air interface are often exposed to hypoxic conditions. This is because the rate of cells digesting oxygen is faster than that of oxygen dissolving and diffusing into the medium to reach the cells [3, 4]. Oxygen-deficient conditions significantly inhibit cell growth resulting in either reduced productivity or apoptotic death. In general, a well-evaluated approach to analyzing cell growth conditions involves measuring the pH, CO2 and dissolved oxygen concentrations, and cell size, utilizing techniques such as flowinjection, near infrared (NIR) spectroscopy, and high-performance liquid chromatography (HPLC) [5–7]. Although these approaches offer high sensitivity and accuracy, they are less user-friendly owing to their expensive apparatus as well as complexity of their operations. To overcome these limitations, paper-based analytical devices (PAD) using biomarkers such as glucose and lactate have been developed, with focus on real-time analysis of the changes in biomarker concentrations during cell culture [8, 9]. Glucose is the most frequently used energy source for cell growth in culture; cells produce lactate under anaerobic conditions [10–12]. Therefore, we focused on the evaluation of glucose and lactate concentrations in cell culture media. Studies involving PADs have witnessed a recent growth owing to their cheap, simple, and easy technology that can be directly applied to commercial use [13]. Using PAD, we recently established an optical sensing system for monitoring glucose in serum to diagnose diabetes [14]. A single assay was conducted since the PAD was developed as a single channel system. However, for simultaneous assay of two different molecules in a single PAD with a single injection, here we employed a multilayered, bifurcated channel design in the PAD (Figure 1(A)) [15–17]. This PAD consists of three layers (top, middle, and bottom), where the top
layer provides sites for sample loading and detection zones for glucose and lactate. The bifurcated fluidic channel to transfer sample solution from the sample drop zone to the detection zone is positioned at the bottom layer, while these layers are attached by doubleside adhesive tape- the middle layer.
The channel was patterned on paper by employing the wax printing method [18], and later incubated in a vacuum oven at 150°C for 30 s to melt the printed wax permanently onto the paper (Figure 1(B)). The unprinted hydrophilic regions function as microfluidic channels that spontaneously transport the sample solution along the channel via capillary force [19]. For sample analysis, glucose and lactate oxidases (GOx and LOx, respectively), incorporated into HRP-mediated colorimetric assays using chromogenic substrates, were employed in the PAD for glucose and lactate measurements, respectively [20]. As shown in Figure 1(C), when a sample solution is introduced to the drop zone at the top layer, the solution is transferred to both the glucose and lactate detection zones along the fluidic channel on the bottom layer. Next, the solution participates in the HRP-mediated color-developing reaction. In the detection zone of the PAD, GOx and LOx digest glucose and lactate, respectively, generating H2O2. In the presence of H2O2, the HRP converts 4-aminoantipyrine (4-AAP) and N-ethyl-N(2-hydroxy-3-sulfopropyl)-3,5-dimethylaniline sodium salt monohydrate (MAOS) to a bluecolored product, which can be visualized in the detection zone. Its intensity is directly proportional to the concentrations of glucose and lactate in the applied sample solution (Figure 2(A) and (B)) [21-23].
For practical utilization of this cell monitoring system, a simple, reliable, and most importantly, user-friendly optical sensing device was necessary. Toward this, we developed an optical instrument employing an everyday information technology (IT) device- a smartphone. Among the high-performance sensors installed in smartphones (such as gyroscope, accelerometer, magnetometer, and proximity and light sensor), embedded chargecouple device (CCD) cameras, owing to their high resolution and definition, have emerged as promising alternatives to commercial high-end ones. Being a high-performance, user-friendly
device, the smartphone was the optimal platform to fabricate a simple optical sensing system as a signal transducer. In this study, we describe successful evaluation of glucose and lactate concentrations in cell culture media using the PAD in combination with a smartphone. By conjugating the PAD with a smartphone using a lightproof gadget, optical signal interference with the external light was optimized, making the assay suitable for real-time monitoring of glucose and lactate in the industry.
2. Experimental procedures
2.1 Materials
GOx (from Aspergillus niger), LOx (from microorganisms), and HRP were purchased from Toyobo Co., Ltd. β-D-glucose, lactate, cellulose, and 4-AAP were purchased from SigmaAldrich; while MAOS was acquired from Dojindo Molecular Technologies, Inc. RPMI-1640 (with HEPES), fetal bovine serum, and penicillin-streptomycin were purchased from Life Technologies. Filter paper was purchased from Whatman, while wax-coated channels were printed by wax print (Xerox, ColorQube 8570). Phosphate-buffered saline (PBS, pH 7.2) solution, containing 0.1 M phosphate and 0.15 M NaCl, was prepared in double-distilled and deionized water. PBST was prepared in the laboratory (50 mM PBS, 0.9% NaCl, 0.1% Triton X-100, pH 7.4). Samples of animal cell culture medium (mouse L-929 cells cultured in RPMI-1640, HEPES) were provided by SPL Life Science Co. Ltd.
2.2 Fabrication of PAD for glucose and lactate analyses
The PAD used in this study comprises three layers- top, middle, and bottom. The top and bottom layers have hydrophobic channels, with double-sided adhesive tape as the middle layer attaching the other two. For hydrophobic channels in each layer, a designed pattern was printed on filter paper by using wax printer. The printed paper was then heated in an oven at 150°C for 1 min. During this process, the wax solution was melted, which infiltrated into the porous paper, thereby forming the hydrophobic channel. The top layer contains one sampleloading zone and two detection zones, and the bottom includes the bifurcated flow channel,
which transfers the sample solution from the drop zone to each detection zone. For the colorimetric assay, 10 μL of solution containing 1 mg/mL HRP, 1 mM MAOS, 10 mM 4AAP, and oxidase was loaded onto the detection zones. GOx and LOx (10 mg/mL and 5 mg/mL, respectively) were immobilized on their corresponding detection zones. To transport the sample solution effectively, 8 μL PBST was used to pretreat the fluidic channel on the bottom layer. The surfactant, Triton X-100 in PBST, enhances hydrophilicity of the filter paper. Therefore, the sample solution rapidly moves from the loading to detection zones. The solution-treated top and bottom layers were dried for 2 h at room temperature. After attaching the prepared top layer to the middle one, 10 μL cellulose powder (2 mg/mL) was applied to the hole in the middle layer in order to fill the flow path (in our previous study, the optical signal in the detection zone was consistently developed through cellulose powder in the flow path [14]). Finally, the bottom layer was combined with the other side of the middle layer. The prepared PAD was stored at 4°C until further use.
2.3 Manipulation of smartphone gadget
The smartphone-adaptable lightproof case was constructed for complete protection from external light during the image acquisition process. A lightproof acryl case was bonded on a conventional smartphone case that covers the rear of the phone, along with a holder for the PAD and a reflection film, dispersion film, and trigonal prism. A combination of reflection and dispersion films was employed to spread light uniformly within the lightproof case during image acquisition using the smartphone-embedded light-emitting diode (LED) flash as a single light source. The optical signal from the PAD was registered on the smartphone CCD camera through the trigonal prism.
2.4 Glucose and lactate analyses employing enzymatic colorimetric assay
We prepared glucose and lactate samples in PBS (concentrations ranging from 0–16 mM and 0–1 mM, respectively), and applied 15 μL of each onto the sample-loading zone of the PAD. The resultant blue color in the detection zone of the PAD was registered every 5 s using the smartphone gadget developed. Similarly, samples of RPMI-1640, in which L-929 cells were cultured for 1, 24, 48, and 96 h, were loaded and the images recorded every 10 s. The color intensities were then quantified using the image analysis software, ImageJ.
3. Results and discussion
3.1 Principle of the smartphone-based optical sensing system
In this study, we developed an animal cell culture monitoring system using a novel PAD in combination with smartphone-embedded CCD camera as an optical transducer. Toward optical analyses of glucose and lactate, an HRP-mediated colorimetric biochemical assay was employed. It is widely known that HRP induces the transformation of colorless substrates into colored products in the presence of H2O2. GOx and LOx were employed along with HRP and the chromogen for glucose and lactate analyses, respectively, where the color intensities changed in proportion to the analytes concentration applied. These changes in the PAD were quantified using the smartphone-based optical sensing system as shown in Figure 3.
The prepared glucose and lactate solutions were applied onto the sample drop zone on the top layer of the PAD. As the dropped solutions reached the detection zones via capillary force along the fluidic channel on the bottom layer, glucose and lactate reacted with the enzymes, producing the colored product. The assayed PAD was positioned on the middle of the smartphone gadget, designed as described in Figure 3(A), and the image was registered through the trigonal prism on the embedded CCD camera in a light-protected environment. Constant and reproducible images were acquired using the LED flash as a single light source, whose homogeneous spreading was achieved with the reflection film. The dispersion film controlled the light intensity within the gadget in order to reduce sparkling effect of the water-drop in the PAD detection zone (Figure 3(B)).
3.2 Optimization of wax-printing size of the PAD detection zone (optimization of sample flow on the PAD)
In our previous study, we demonstrated a PAD-based glucose analysis principle using the
camera function of a commercial smartphone [14]. In that study, we precisely quantified the concentration of glucose in human serum. However, the principle enabled detection of only a single target at a time. An improved PAD design with bifurcated channel for simultaneous detection of glucose and lactate, therefore, has been developed in the current study. To effectively immobilize GOx and LOx in a fixed area allowing simultaneous detection of glucose and lactate on a single PAD, we designed multi-layered PAD containing one sample loading zone and two detection zones. For the sample movement to the desired area, a bifurcated fluidic channel was wax-printed in the middle layer. During the wax-printing and curing procedure, the printed size of wax pattern was slightly diminished (< 0.5 mm). To overcome this problem, we have conducted the optimization test for the wax-printing size and added margin by considering the wax diffusion and curing [14].
As shown in Figure 4, we optimized the size of the detection zone by comparing the assay results using two different sizes (3 and 4 mm diameter). We applied 15 μL of the sample solution, containing 5 mM glucose and 1 mM lactate, to the PAD prepared, and acquired images with the smartphone gadget. According to the assay results, the PAD with the 4 mmlarge detection zone displayed irregular color development for both glucose and lactate (Figure 4(A)). The chromophores appeared to be pushed ahead along the direction of flow, possibly misplaced by the excess solution that had continuously followed the reactants. For accurate analysis of glucose and lactate, an improved assay condition for uniform color development
was
required.
To
achieve
this,
we
could
either
increase
the
concentration/volume of chromophores or decrease the channel size. The latter was inducted into the design by adjusting the detection zone size from 4 mm to 3 mm. This maximized sample distribution in the detection zone, exhibiting uniform color intensities in both the glucose and lactate assays as shown in Figure 4(B). Moreover, the regular color development was stably maintained for several hours. Therefore, the 3 mm-large PAD detection zone was standardized for precise detection of glucose and lactate in the subsequent tests.
3.3 Optimization of enzymatic reaction time
In this study, glucose and lactate were analyzed based on the HRP/oxidase-associated
colorimetric principle. Generally, in enzymatic colorimetric assays, color development saturates over time. In order to detect the analyte and enhance the reliability of the developed assay principle, we optimized the time of the enzymatic reaction. Toward this, we assayed 0– 10 mM glucose and 0–0.75 mM lactate in PBS using the system developed, and registered images every 10 s using the time-lapse function of the smartphone (Figure 5).
In the glucose assay, the color intensity gradually increased for 60 s, followed by signal saturation. Although the color intensity increased over time, no significant differences were observed after 60 s (Figure 5(A)). Additionally, it was difficult to distinguish optical signals between each glucose assay within the 60 s. Based on the results, it was clear that 60 s were sufficient to effectively distinguish each glucose assay. The overall observation for the lactate assay was similar to that of glucose, with a sharp increase in color intensity for 60 s, followed by saturation. Therefore, we confirmed that lactate could also be analyzed at 60 s depending on the sensing principle (Figure 5(B)). Based on these findings, the detection time was determined to be 60 s, ideal for practical analyses of glucose and lactate.
3.4 Glucose and lactate assays using standard glucose and lactate in buffer
To examine the correlation between color intensities at the PAD detection zones and glucose and lactate concentrations. In the conventional animal cell culture, about 10 mM of glucose was used as an energy source, and only a small amount of lactate was produced in anaerobic condition [24, 25]. Based on this, glucose and lactate solutions (0–16 mM and 0–1 mM, respectively) in PBS were prepared and applied to the smartphone-based optical sensing system. To detect glucose and lactate simultaneously, the glucose solution was premixed with that of lactate in a 1:1 ratio. Here, 0 mM glucose with 1 mM lactate, and 0 mM lactate with 16 mM glucose were prepared to test the cross-reactivity of each oxidase. The color intensities for both glucose and lactate were proportional to the concentration of the analytes applied as shown in Figure 6(A) and (B).
Further, evident color differences between each assay result were observed. In the image acquired from 1 mM lactate assay, the color change was observed only in the lactate assay zone on the PAD, but not glucose detection zone. Similarly, optical signal from the 16 mM glucose assay was solely detected in the glucose detection zone. Based on these results, we verified that none of the oxidases was cross-reactive. This implied that we could selectively analyze the specific target analyte from the sample solution. The color intensities were quantified from the images acquired using ImageJ. The quantified data revealed an increase for both detection zones, proportional to each analyte concentration. The calibration curve of glucose assay using GOx/HRP catalysis format was plotted in Fig. 6(C). The curve showed that the optical signal was proportional to the glucose concentration and leveled off above 8 mM. Based on the calibration curve, the LOD and LOQ were calculated as 0.3 mM and 0.9 mM in the dynamic detection range from 0.3 to 8.0 mM of glucose. The LOD and LOQ of the calibration results were calculated as 3.3-times and 10-times the standard deviation of the background signal divided by the slope of the calibration curve, based on the ICH guidelines [26]. Meanwhile, the calibration result from the lactate assay using LOx/HRP catalysis was different from the glucose assay. In the developed enzyme-based biosensing system, the employed LOx has an affinity value to lactate (Km = 0.2 mM), which is significantly lower than that of GOx with glucose (Km = 7.0 mM) [27, 28]. Thus, the calibration signal was saturated and leveled off at the relatively low concentration of lactate, presenting a narrow dynamic range in comparison to the glucose assay. We found that the registered signal from LOx/HRP catalysis on PAD was increased in accordance with the lactate concentration, while the signal was leveled off above 0.5 mM (Fig. 6(D)). With the result, the LOD (0.02 mM) and LOQ (0.06 mM) for the lactate assay by employing the LOx/HRP bienzymatic catalysis in the dynamic detection range from 0.02 to 0.50 mM were obtained. By integrating the two bienzymatic reaction-mediated sensing including the GOx/HRP and LOx/HRP on a single PAD, the consumption of glucose as a carbon source and the production of lactate during an animal cell culture would be monitored simultaneously. To verify the accuracy of the smartphone-based assay, the same glucose and lactate samples were assayed at least thrice under the same reaction conditions (i.e., duration, temperature, and pH). Both the glucose and lactate assays exhibited coefficients of variation (COV) approximately within 5%, demonstrating high accuracy of the system. We thought that the smartphone-based optical sensing system, in combination with enzymatic colorimetric assay, would be useful for monitoring glucose and lactate levels in animal cell culture.
3.5 Glucose and lactate assays using cell culture media
To verify the applicability of the sensing system for monitoring changes in glucose and lactate concentrations in practical cell culture, L-929 cells were cultured in conventional RPMI-1640 media for 1, 24, 48, and 96 h. They were cultured without additional energy source under anaerobic conditions. For direct comparison, the media samples were also analyzed by high-performance liquid chromatography (HPLC) to determine the precise concentrations of glucose and lactate in the sample. As shown in Table 1, over time, glucose concentrations gradually reduced from 10.4 mM to 4.7 mM, while that of lactate increased from 3.3 mM to 9.1 mM.