Technical Note pubs.acs.org/ac
Cite This: Anal. Chem. XXXX, XXX, XXX−XXX
Microfluidic Module for Real-Time Generation of Complex Multimolecule Temporal Concentration Profiles Kristina Woodruff and Sebastian J. Maerkl* Institute of Bioengineering, School of Engineering and School of Life Science, École Polytechnique Fédérale de Lausanne, Lausanne, Switzerland S Supporting Information *
ABSTRACT: We designed a microfluidic module that generates complex and dynamic concentration profiles of multiple molecules over a large concentration range using pulse-width modulation (PWM). Our PWM module can combine up to six different inputs and select among three downstream mixing channels, as required by the application. The module can produce concentrations with a dynamic range of three decades. We created complex, temporal concentration profiles of two molecules, with each concentration independently controllable, and show that the PWM module can execute rapid concentration changes as well as long-time scale pharmacokinetic profiles. Concentration profiles were generated for molecules with molecular weights ranging from 560 Da to 150 kDa. Our PWM module produces robust and precise concentration profiles under a variety of operating conditions, making it ideal for integration with existing microfluidic devices for advanced cell and pharmacokinetic studies.
P
Our chip successfully produces a large dynamic range of concentrations over both short and long time scales and can be
hysiologically relevant concentration changes such as pharmacokinetic drug concentration profiles occur over minutes to hours and follow a complex and continuous rise and fall pattern.1−3 However, in vitro experiments typically involve one or more step function changes, switching rapidly from one molecule to another or from one concentration to another. Pulse-width modulation (PWM) is one strategy to accurately and dynamically generate molecular concentrations that more accurately reflect in vivo conditions. PWM can be applied to control the flow of buffer and substrate solutions on microfluidic chips, generating alternating pulses that ultimately diffuse to homogeneity. Different concentrations are created by varying the duty cycle, which refers to the fraction of time occupied by the substrate pulse in comparison to the total cycle time. PWM has been incorporated into microfluidic diluters in the past to manipulate small molecules and generate nondynamic step functions for multiple molecules in parallel.4−7 The dynamic range that can be achieved on these devices has been limited, only small molecules (1.35−12.15 μM), and high (>12.15−109.35 μM) ranges. The difference between the maximum (109.35 μM) and minimum (0.15 μM) concentration values is 3 orders of magnitude. The generated concentration profile closely matched the programmed values for the three inlet pairs (Figure 3). The 729-fold concentration range presented in Figure 3 is physiologically relevant and covers the typical difference between minimum and maximum serum concentrations of antibiotics, therapeutic antibodies, and anticancer agents measured in clinical studies.20−22 Complex, Long-Term, and PK/PD Concentration Profiles. The PWM module can be easily integrated with chips fabricated by multilayer soft lithography.8 To simplify and broaden the use of the PWM module it can also be directly
mammalian cells.13−15 The faster flow rates (>100 μL/h) necessary for high-throughput culturing devices16−19 could be obtained by scaling up the features of the PWM chip. For flow rates of 30−36 μL/h, the chip achieves concentration changes with response times as low as 5 s (time to switch from 10 to 90% of the maximum concentration; Figures S-7 and S-8). For all experiments we used minimum and maximum duty cycles of 0.1 and 0.9, respectively, to ensure that the valve switching times were not too large compared to the pulse lengths. To broaden the dynamic range of output concentrations while keeping within these limits, we connected the chip to multiple preparations of the same substance. Our platform automatically selects which stock solution to use based on the desired output concentration. We mixed a series of 9-fold dilutions of 10 kDa FITC-dextran with buffer to create D
DOI: 10.1021/acs.analchem.7b04099 Anal. Chem. XXXX, XXX, XXX−XXX
Technical Note
Analytical Chemistry connected to a second chip, making it compatible with a wide range of existing devices. We tested this configuration using flexible PEEK tubing to connect the bypass channel outlet (Figure 1a) of the PWM module to the inlet of a second chip (Figure S-6). We used the two-chip setup to simultaneously program the concentrations of two substances as sine curves of changing periods (Figure 4a). Our setup enables the user to run custom concentration profiles and is not limited to simple mathematical functions. The device can also manipulate and sustain concentrations on the long time scales required for biological and cell culture studies. We performed a 48 h experiment in which sulforhodamine and 10 kDa FITC-dextran were pulsed in parallel, with each substance either stably maintained at 0.1 (10% of maximum) or gradually ramped up to, or down from, 0.9 (90% of maximum) (Figure 4b). Each peak spanned 2, 4, 6, or 8 h, which are time ranges relevant for drug dosage and pharmacokinetic studies.23,24 For the entire 48 h, the desired profile was generated without chip failure, demonstrating the robustness of the setup. We last generated dynamic profiles mimicking the plasma pharmacokinetic profile of orally administered drugs25 (Figure 4c). Sulforhodamine and FITC were programmed simultaneously; both molecules started at 0 concentration and quickly increased to 0.9 (90% intensity), corresponding to release of the drug from the formulation and absorption into the bloodstream.24,25 After reaching their respective maximum concentrations both substances decreased gradually (to final values of 0.1 for sulforhodamine and 0.29 for FITC), representing metabolism and elimination of the drug. The generated profiles closely match the programmed values in both shape and concentration (Figure 4c). These profiles can be customized to reflect different drugs, delivery methods, and formulations, facilitating studies for investigating the interactions of multiple drugs to identify potential synergy or antagonism.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: sebastian.maerkl@epfl.ch. Phone: +41 (0)21 693 7835. ORCID
Kristina Woodruff: 0000-0003-0554-2020 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This project is part of a Ph.D. dissertation30 and has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant Agreement No. 723106). We thank Nadanai Laohakunakorn and Mathieu Quinodoz for help with LabVIEW programming. We thank Gauthier Croizat for helpful discussions and assistance with fluidic modeling.
■
REFERENCES
(1) Meibohm, B.; Derendorf, H. J. Pharm. Sci. 2002, 91, 18−31. (2) Tuntland, T.; Ethell, B.; Kosaka, T.; Blasco, F.; Zang, R. X.; Jain, M.; Gould, T.; Hoffmaster, K. Front. Pharmacol. 2014, 5, 174. (3) Velkov, T.; Bergen, P.; Lora-Tamayo, J.; Landersdorfer, C. B.; Li, J. Curr. Opin. Microbiol. 2013, 16, 573−579. (4) Azizi, F.; Mastrangelo, C. H. Lab Chip 2008, 8, 907−912. (5) Ainla, A.; Gözen, I.; Orwar, O.; Jesorka, A. Anal. Chem. 2009, 81, 5549−5556. (6) Cao, L.; Zhang, X.; Grimley, A.; Lomasney, A. R.; Roper, M. G. Anal. Bioanal. Chem. 2010, 398, 1985−1991. (7) Zhang, X.; Grimley, A.; Bertram, R.; Roper, M. G. Anal. Chem. 2010, 82, 6704−6711. (8) Unger, M. A.; Chou, H.-P.; Thorsen, T.; Scherer, A.; Quake, S. R. Science 2000, 288, 113−116. (9) Thorsen, T.; Maerkl, S. J.; Quake, S. R. Science 2002, 298, 580− 584. (10) Woodruff, K.; Maerkl, S. J. Sci. Rep. 2016, 6, 23937. (11) Probst, C.; Grünberger, A.; Wiechert, W.; Kohlheyer, D. Micromachines 2013, 4, 357−369. (12) Mao, H.; Cremer, P. S.; Manson, M. D. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 5449−5454. (13) Liu, W.; Li, L.; Wang, X.; Ren, L.; Wang, X.; Wang, J.; Tu, Q.; Huang, X.; Wang, J. Lab Chip 2010, 10, 1717−1724. (14) Kobel, S.; Valero, A.; Latt, J.; Renaud, P.; Lutolf, M. Lab Chip 2010, 10, 857−863. (15) Kolnik, M.; Tsimring, L. S.; Hasty, J. Lab Chip 2012, 12, 4732− 4737. (16) Long, Z.; Nugent, E.; Javer, A.; Cicuta, P.; Scalvi, B.; Cosentino Lagomarsino, M.; Dorfman, K. D. Lab Chip 2013, 13, 947−954. (17) Dénervaud, N.; Becker, J.; Delgado-Gonzalo, R.; Damay, P.; Rajkumar, A. S.; Unser, M.; Shore, D.; Naef, F.; Maerkl, S. M. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 15842−15847. (18) Rajkumar, A. S.; Denervaud, N.; Maerkl, S. J. Nat. Genet. 2013, 45, 1207−1215. (19) Nobs, J.-B.; Maerkl, S. J. PLoS One 2014, 9, e93466. (20) Taccone, F. S.; Laterre, P. F.; Dugernier, T.; Spapen, H.; Delattre, I.; Wittebole, X.; De Backer, D.; Layeux, B.; Wallemacq, P.; Vincent, J. L.; Jacobs, F. Crit. Care 2010, 14, R126.
■
CONCLUSIONS Microfluidic chips are increasingly being used for biological studies due to reduced sample consumption, better regulation of the microenvironment, and compatibility with single-cell analysis. However, typical chips lack the ability to accurately produce physiologically relevant changes in concentrations. Our PWM module improves upon existing PWM devices by enabling long-term experiments and the dynamic manipulation of multiple substances in parallel for a variety of molecular sizes and flow rates (Table S-1). We thoroughly characterized the platform to identify the optimal conditions required for a given experiment (Figure S-9). The PWM module itself is easy to fabricate using standard multilayer soft lithography and the design files and software are available for download (lbnc.epfl.ch). The device can be programmed to generate pharmacokinetic profiles, a feature that is especially relevant in light of recent developments in organ-on-chip and microfluidic stem cell technologies.26−28 Complex temporal concentrations can furthermore be generated in real-time by the PWM module, which is a useful feature if closed loop, feedback experiments are performed.29 Overall, our PWM platform complements existing biomedical microfluidic devices and will facilitate studies that better reflect the in vivo environment.
■
Supporting figures and tables: timing of valve switching, data for 0.5 s cycle, data for fast and slow flow rates, data for response times, data for PWM used upstream of a second chip, table comparing other PWM devices, a guide to selecting operating conditions. Supporting experimental information: microfluidic device fabrication and operation (PDF).
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b04099. E
DOI: 10.1021/acs.analchem.7b04099 Anal. Chem. XXXX, XXX, XXX−XXX
Technical Note
Analytical Chemistry (21) Ng, C. M.; Stefanich, E.; Anand, B. S.; Fielder, P. J.; Vaickus, L. Pharm. Res. 2006, 23, 95−103. (22) Blair, E. Y. L.; Rivory, L. P.; Clarke, S. J.; McLachlan, A. J. Br. J. Clin. Pharmacol. 2004, 57, 416−426. (23) Levison, M. E.; Levison, J. H. Infect. Dis. Clin. North Am. 2009, 23, 791−797. (24) Nielsen, E. I.; Friberg, L. E. Pharmacol. Rev. 2013, 65, 1053− 1090. (25) Dhillon, S.; Gill, K. Clinical Pharmacokinetics; Pharmaceutical Press, 2006; pp 1−44. (26) Bhatia, S. N.; Ingber, D. E. Nat. Biotechnol. 2014, 32, 760−772. (27) Ertl, P.; Sticker, D.; Charwat, V.; Kasper, C.; Lepperdinger, G. Trends Biotechnol. 2014, 32, 245−253. (28) Luni, C.; Giulitti, S.; Serena, E.; Ferrari, L.; Zambon, A.; Gagliano, O.; Giobbe, G. G.; Michielin, F.; Knobel, S.; Bosio, A.; Elvassore, N. Nat. Methods 2016, 13, 446−452. (29) Milias-Argeitis, A.; Summers, S.; Stewart-Ornstein, J.; Zuleta, I.; Pincus, D.; El-Samad, H.; Khammash, M.; Lygeros, J. Nat. Biotechnol. 2011, 29, 1114. (30) Woodruff, K. Microfluidic platforms for high-throughput mammalian cell printing, transfection, and dosage-dependent studies. Ph.D. Dissertation, École Polythechnique Fédérale de Lausanne, Lausanne, CH, 2017.
F
DOI: 10.1021/acs.analchem.7b04099 Anal. Chem. XXXX, XXX, XXX−XXX
