Elasto-inertial microfluidics for bacteria separation from whole blood for sepsis diagnostics
© The Author(s) 2017
Received: 22 July 2016
Accepted: 3 December 2016
Published: 4 January 2017
Bloodstream infections (BSI) remain a major challenge with high mortality rate, with an incidence that is increasing worldwide. Early treatment with appropriate therapy can reduce BSI-related morbidity and mortality. However, despite recent progress in molecular based assays, complex sample preparation steps have become critical roadblock for a greater expansion of molecular assays. Here, we report a size based, label-free, bacteria separation from whole blood using elasto-inertial microfluidics.
In elasto-inertial microfluidics, the viscoelastic flow enables size based migration of blood cells into a non-Newtonian solution, while smaller bacteria remain in the streamline of the blood sample entrance and can be separated. We first optimized the flow conditions using particles, and show continuous separation of 5 μm particles from 2 μm at a yield of 95% for 5 µm particle and 93% for 2 µm particles at respective outlets. Next, bacteria were continuously separated at an efficiency of 76% from undiluted whole blood sample.
We demonstrate separation of bacteria from undiluted while blood using elasto-inertial microfluidics. The label-free, passive bacteria preparation method has a great potential for downstream phenotypic and molecular analysis of bacteria.
Despite progress in medical science, including the development of effective therapies, infectious diseases continue to cause millions of deaths worldwide, and pathogens in food, animals, water, and plants cause damage and production losses running into billions of dollars. To date, pathogens are typically detected only after they have already caused massive damage. Improved diagnostic methods for infectious pathogens are, therefore, urgently needed. For example, sepsis—an acute inflammatory response of immune-compromised patients to certain pathogens—is the third most common cause of death in Germany . In the case of septic shock, studies have shown that patient mortality will increase by 7.6% for each hour the antibiotic therapy is delayed , and, if the initial antibiotic therapy is inappropriate, the survival rate decreases from 52 to 10.3% . Therefore, in infectious diseases and sepsis diagnosis the foremost important step is to find the suitable treatment and identification of the bacteria to prevent evolution of resistant bacteria. Currently, blood-culturing method is the gold standard for identification of microorganism. However, the automated method still requires 24–72 h to get the results . This long turnaround time, especially for the identification of antimicrobial resistance, is driving the development of molecular diagnostics, often based on polymerase chain reaction (PCR) and are used to detect pathogens either from blood culture bottles [5–7] or directly from blood [8–13]. While the total time for diagnosis has been shortened significantly, the implementation of these molecular methods in clinics has been severely hampered by their lack of sensitivity in comparison with, for instance, blood culturing and the need for complex, multi-step sample preparation. Some of the factors affecting the quality in nucleic acid-based methods are PCR inhibitors, abundant interfering human DNA, the risk of carryover when processing several samples, inadequate lyses, and pathogens enclosed within or adhering to human cells. Despite improvements, sample preparation remains the bottleneck for the further development and implementation of molecular diagnostics in clinical settings. Hence, molecular diagnostics would benefit from a rapid, integrated sample-preparation assay method to isolate and enrich bacteria from complex sample matrices such as blood.
Microfluidics has the potential of eliminating the shortcomings associated with complex sample preparation. Microfluidics provide a higher surface to volume ratio, a faster rate of mass and heat transfer, and the ability to handle very small volumes of reagents in microchannels very precisely. Moreover, microfluidics open up the possibility for automated platforms with integrated microfluidic cartridges thereby reducing the risk of contamination . Therefore, a recent interest of microfluidic techniques has been towards the separation of microorganisms from blood. Different approaches to separate pathogens from blood using affinity separation [13–15], size [16, 17] or electrokinetic properties  have been demonstrated. These methods typically exploit the difference between cell properties, such as the size, shape, density, deformability, electric/magnetic susceptibility, and hydrodynamic properties. Among these parameters, size is an excellent label-free biomarker for bacteria separation from blood.
Very recent, inertial microfluidics has been described as a high-throughput, simple method for precise manipulation particles based on size . Recently, Wu et al. , separated bacteria from diluted red blood cells using ‘‘soft” inertial microfluidics that utilized deflection of larger cells in an asymmetrical sheath flow around a curvature while the smaller cells are kept on or near the original flow streamline. While the yield was about 62%, they obtained an impressive high purity of 99.7%. Similarly, Mach et al.  used a straight channel to separate bacteria from red blood sample using massively parallel channels. Here, size-dependent inertial lift forces were used to focus larger red blood cells as a method of cell separation and the authors achieved 80% removal of bacteria from diluted red blood cells after two passes of the single channel system.
While promising, the narrow size difference between microorganisms (typically 1–3 µm) and blood cells (3–15 µm) has shown to be very difficult for bacteria separation using inertial microfluidics. In addition, the fact that bacteria are smaller than blood cells and will end up in the “unfocused” stream, and in essence everywhere in the channel cross-section makes it difficult to achieve proper separation. Very recent, Hou et al.  used dean flow fractionation to address this by introducing a sheath flow at the inlet to pinch blood sample and size based migration of the blood cells towards the inner wall while the bacteria are lagging behind and could be extracted at an efficiency of 70% . In this work, we address this by employing elasto-inertial microfluidics instead to differentially migrate larger blood cells away from smaller bacteria in flow through straight channels.
Using elasto-inertial microfluidics, it is possible to migrate particles across streamlines and focus into a single stream in three-dimensional channel depending upon their size [24–30]. A number of investigations have recently focused on optimizing different conditions like concentration of non-Newtonian fluids (elastic forces) and flow rates (inertial forces) . The inertial and elastic forces have been used in combination to separate smaller and larger particles from each other  and for separation of blood cell components . Elasto-inertial microfluidics was utilized by Liu et al.  to separate bacteria (E. coli) from red blood cells without the use of sheath flow. However, the channel dimension used is not applicable for other blood cells as the smallest dimension of the cross-section was 10 μm, which would easily clog for applications using whole blood.
In this paper, using elasto-inertial microfluidics, we separate bacteria from undiluted whole blood by selectively migrating blood cells away from the walls towards the centerline of the channel while bacteria are remained in the streamline they enter and separated. We first investigate the elastic and inertial forces theoretically using simulations and experimentally using different sized particles and viscoelastic solutions. We optimized the flow conditions to continuously separate large particles (5 μm) from small particles (2 μm). Following, we applied the optimal flow conditions to continuously separate bacteria from undiluted whole blood.
Results and discussion
Elasto-inertial based particle focusing and separation
To quantitatively evaluate the effect of lift forces (FL) and elastic forces (Fe), it is possible to employ two dimensionless numbers: Reynolds number (Re) and Weissenberg number (Wi). The channel Reynolds number (Re) is defined to describe the magnitude ratio of inertial force to viscous force, while Wi describes the magnitude ratio of elastic force to viscous force (Additional file 1: Figure S1). Particle’s Reynolds number (Rp) is another important dimensionless number that accounts for particle size [Rp = Re (a/Dh)2].
Next, to mimic bacteria separation, we used the optimized channel dimension, PEO concentration and Re to separate 2 µm from 5 µm particles. The channel dimensions 50 µm × 65 µm (width × height) and PEO concentration of 500 ppm was used to achieve size-based separation of the particles. At the sample flow rate of 30 µl/h and PEO flow rate of 360 µl/h (corresponding to Rp 0.008), the minimum channel length to fully focus 5 µm particles was found around 25 mm (Fig. 3c). Using this channel length and flow conditions, we could successfully separated 5 µm particles from 2 µm particles (Fig. 3d). The yield of the 5 μm particles, calculated as fraction of 5 μm particles recovered through the middle outlet to the total count, was 95%, and the yield was 93% for the 2 μm particles in the side outlet.
Bacteria separation from whole blood
Culture-independent, PCR based, detection of pathogens directly from the patient’s blood is attractive to accelerate the diagnostic process. However, the use of whole blood in assays designed to detect pathogen nucleic acid is challenging. An excess of human DNA may hamper the detection of pathogen genomic material or inhibit the PCR reaction [36, 37]. Furthermore, hemoglobin traces may also inhibit PCR-based amplification . Therefore, molecular methods are often forced to use a relatively small volume of blood, which affects the sensitivity. As shown above (see Fig. 4c), our continuous flow sample preparation strategy significantly reduces the complexity by getting rid of majority RBCs and WBSs. However, for sepsis diagnostics, the amount of bacteria present in the blood is very low—in the order of 10–100 cfu/ml. Hence, to be clinically relevant it is imperative to further improve the method in order to recover all bacteria cells as well as improve the sample throughput.
In elasto-inertial microfluidics, the synergetic effect of visco-elastic forces and inertial forces are harnessed such that particles can migrate and occupy a single focusing point. This focusing phenomenon has been used for various applications, including sheathless cell ordering [39, 40] size based cell separation [33, 41, 42] and cell stretching measurements . Nam et al.  used similar channel geometry as ours to separate platelets from diluted blood components with extremely high purity (close to 99.9%). However, the relatively slow flow rate combined with the use of diluted blood sample (less than 1% solid content) makes it un-applicable for bacteria separation applications. Here, we have demonstrated bacteria separation from undiluted whole blood. While the method is immediately applicable as sample preparation for MALDI-ToF MS based identification of microorganisms from positive blood cultures, for blood culture independent molecular diagnostics the method needs to process ml volumes blood sample. The relative low volumetric flow rate is an inherent limitation of elasto-inertial microfluidics since the synergetic effect of the elastic forces and inertial forces are ideal at moderate flow rates. For instance, even at volumetric flow rate of 60 µl/h tested in this work, it would take about 17 h to process 1 ml blood. One way to improve throughput is therefore through parallelization of the channels. Towards this, we have recently reported on a highly scalable, parallel-channel, microfabrication method for passive size-based particle separation . Using the 16-channel parallel device , it would take only 1 h to process 1 ml. While outside the scope of this paper, we are currently working on combining the robust microfluidic fabrication process of parallel channels (64 channels), with bacteria separation based on elasto-inertial microfluidics. The method has potential value in clinical sample preparation applications for both molecular diagnostics as well as analysis by plating for antibiotic susceptibility.
We demonstrated bacteria separation from whole blood based on elasto-inertial microfluidics. By harnessing the synergetic effect of elastic and inertial forces, we first demonstrate efficient particle separation where we could separate 5 µm particles from 2 µm at a yield of 95% for 5 µm and 93% for 2 µm particles at the respective outlet fractions. Furthermore, we successfully demonstrated bacteria isolation from undiluted whole blood by selectively migrating the larger blood cell components from the sidewalls towards the centerline for separation. 76% of the bacteria were recovered at the side outlet while 92% of the WBCs could be separated into the middle outlet. The passive, label-free bacteria separation method is very promising and has great potential as stand-alone sample preparation method or integrated into lab-on-chip system for molecular and phenotypical based sepsis diagnostics.
Microfluidic device fabrication
The polydimethylsiloxane (PDMS) microfluidic chips with two inlets and two outlets were fabricated on a master mould that was produced through photolithography on a silicon wafer using SU-8 negative resist. PDMS Sylgard 184 was then poured onto the SU-8 master in a 10:1 ratio, degassed, and cured at 65 °C overnight. The PDMS slab was cut, the holes for inlets and outlet were punched, and covalently bonded to glass slides using oxygen plasma (CUTE Femto Science Co. Korea) treatment. The following cross section dimension (width × height) were fabricated: 50 µm × 65 µm; 100 µm × 65 µm and 150 µm × 65 µm.
Suspension of different fluorescent polystyrene particles (2, 5 and 10 µm) were prepared in Phosphate Buffered Saline (1× PBS). Poly (ethylene oxide) (Sigma Aldrich, St Louis), Mw = 2,000,000 was prepared at different concentrations of 250, 500, 750 and 1000 ppm in 1× PBS. To mimic the sepsis blood samples, blood samples obtained from healthy blood donors were spiked with bacteria (~106 cfu/ml). As a model strain, gram-negative E. coli (strains BL21-A1) cultured in liquid medium, collected at the mid-log phase, washed with PBS, was used to spike the blood samples.
For the particle based experiments, the solution containing the particles were introduced into side inlet using a syringe pump (Harvard apparatus PHD 2000, Harvard Apparatus, USA) and Non-Newtonian fluid of PEO was introduced into middle inlet by syringe pump (NEMESYS, Cetoni Gmbh, Germany). The particles collected from the side and center outlet were analyzed using Coulter counter for quantification. For the bacteria related work, initially PBS (1×) solution was spiked with bacteria. This was followed by whole blood experiments spiked with bacteria. The experimental procedure was similar as for the particle suspension. Coulter counter was used to quantify the white blood cells (WBCs) while plating was used to quantify the bacteria.
MAF performed the experiments including the chip fabrication, preparation, characterization, blood and bacteria separation experiment and analysis. HR contributed in chip design and provide assistance in experiments and analysis.IB performed the Comsol simulations while SA and SZ contributed to bacteria culturing. AR conceptualized, designed the chip and fabricated the master mould. All authors provided scientific input during these studies and assisted in preparation of the manuscript. All authors read and approved the final manuscript.
This work was financially supported by the Innovative Medicines Initiative, a public–private partnership between the European Union, and the European Federation of Pharmaceutical Industries and Associations (RAPP-ID project, Grant agreement, No. 115153).
The authors declare that they have no competing interests.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Engel C, Brunkhorst FM, Bone H-G, Brunkhorst R, Gerlach H, Grond S, et al. Epidemiology of sepsis in Germany: results from a national prospective multicenter study. Intensive Care Med. 2007;33:606–18.View ArticleGoogle Scholar
- Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, Sharma S, et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med. 2006;34:1589–96.View ArticleGoogle Scholar
- Kumar A, Ellis P, Arabi Y, Roberts D, Light B, Parrillo JE, et al. Initiation of inappropriate antimicrobial therapy results in a fivefold reduction of survival in human septic shock. Chest. 2009;136:1237–48.View ArticleGoogle Scholar
- Mancini N, Carletti S, Ghidoli N, Cichero P, Burioni R, Clementi M. The era of molecular and other non-culture-based methods in diagnosis of sepsis. Clin Microbiol Rev. 2010;23:235–51.View ArticleGoogle Scholar
- Lebovitz EE, Burbelo PD. Commercial multiplex technologies for the microbiological diagnosis of sepsis. Mol Diagn Ther. 2013;17:221–31.View ArticleGoogle Scholar
- Mancini N, Carletti S, Ghidoli N, Cichero P, Ossi CM, Ieri R, et al. Molecular diagnosis of polymicrobial sepsis. J Clin Microbiol. 2009;47:1274–5.View ArticleGoogle Scholar
- Towns ML, Jarvis WR, Hsueh P-R. Guidelines on blood cultures. J Microbiol Immunol Infect. 2010;43:347–9.View ArticleGoogle Scholar
- Afshari A, Schrenzel J, Ieven M, Harbarth S. Bench-to-bedside review: rapid molecular diagnostics for bloodstream infection—a new frontier? Crit Care. 2012;16:222.View ArticleGoogle Scholar
- Lin C-C, Wang J-H, Wu H-W, Lee G-B. Microfluidic immunoassays. J Assoc Lab Autom. 2010;15:253–74.View ArticleGoogle Scholar
- Mayeux PR. Pathobiology of lipopolysaccharide. J Toxicol Environ Health. 1997;51:415–35.View ArticleGoogle Scholar
- Ng AHC, Uddayasankar U, Wheeler AR. Immunoassays in microfluidic systems. Anal Bioanal Chem. 2010;397:991–1007.View ArticleGoogle Scholar
- Nguyen HB, Corbett SW, Steele R, Banta J, Clark RT, Hayes SR, et al. Implementation of a bundle of quality indicators for the early management of severe sepsis and septic shock is associated with decreased mortality. Crit Care Med. 2007;35:1105–12.View ArticleGoogle Scholar
- Yung CW, Fiering J, Mueller AJ, Ingber DE. Micromagnetic–microfluidic blood cleansing device. Lab Chip. 2009;9:1171.View ArticleGoogle Scholar
- Demirci U, Wang S, Inci F, Chaunzwa TL, Ramanujam A, Vasudevan A, et al. Portable microfluidic chip for detection of Escherichia coli in produce and blood. Int J Nanomedicine. 2012;7:2591.View ArticleGoogle Scholar
- Lee J-J, Jeong KJ, Hashimoto M, Kwon AH, Rwei A, Shankarappa SA, et al. Synthetic ligand-coated magnetic nanoparticles for microfluidic bacterial separation from blood. Nano Lett. 2014;14:1–5.View ArticleGoogle Scholar
- Leevy WM, Johnson JR, Lakshmi C, Morris J, Marquez M, Smith BD. Selective recognition of bacterial membranes by zinc(ii)-coordination complexes. Chem Commun. 2006;1595–7. doi:10.1039/b517519d.
- Leevy WM, Lambert TN, Johnson JR, Morris J, Smith BD. Quantum dot probes for bacteria distinguish Escherichia coli mutants and permit in vivo imaging. Chem Commun. 2008;2331–3. doi:10.1039/B803590C.
- White AG, Fu N, Leevy WM, Lee J-J, Blasco MA, Smith BD. Optical imaging of bacterial infection in living mice using deep-red fluorescent squaraine rotaxane probes. Bioconjug Chem. 2010;21:1297–304.View ArticleGoogle Scholar
- Di Carlo D. Inertial microfluidics. Lab Chip. 2009;9:3038.View ArticleGoogle Scholar
- Wu Z, Willing B, Bjerketorp J, Jansson JK, Hjort K. Soft inertial microfluidics for high throughput separation of bacteria from human blood cells. Lab Chip. 2009;9:1193.View ArticleGoogle Scholar
- Mach AJ, Di Carlo D. Continuous scalable blood filtration device using inertial microfluidics. Biotechnol Bioeng. 2010;107:302–11.View ArticleGoogle Scholar
- Hou HW, Warkiani ME, Khoo BL, Li ZR, Soo RA, Tan DS-W, et al. Isolation and retrieval of circulating tumor cells using centrifugal forces. Sci Rep. 2013;3:1259.Google Scholar
- Hou HW, Bhattacharyya RP, Hung DT, Han J. Direct detection and drug-resistance profiling of bacteremias using inertial microfluidics. Lab Chip. 2015;15:2297–307.View ArticleGoogle Scholar
- Caserta S, D’Avino G, Greco F, Guido S, Maffettone PL. Migration of a sphere in a viscoelastic fluid under planar shear flow: experiments and numerical predictions. Soft Matter. 2011;7:1100–6.View ArticleGoogle Scholar
- D’Avino G, Romeo G, Villone MM, Greco F, Netti PA, Maffettone PL. Single line particle focusing induced by viscoelasticity of the suspending liquid: theory, experiments and simulations to design a micropipe flow-focuser. Lab Chip. 2012;12:1638.View ArticleGoogle Scholar
- Ho BP, Leal LG. Migration of rigid spheres in a two-dimensional unidirectional shear flow of a second-order fluid. J Fluid Mech. 1976;76:783.View ArticleGoogle Scholar
- Leshansky AM, Bransky A, Korin N, Dinnar U. Tunable nonlinear viscoelastic “focusing” in a microfluidic device. Phys Rev Lett. 2007;98:234501.View ArticleGoogle Scholar
- Romeo G, D’Avino G, Greco F, Netti PA, Maffettone PL. Viscoelastic flow-focusing in microchannels: scaling properties of the particle radial distributions. Lab Chip. 2013;13:2802–7.View ArticleGoogle Scholar
- Villone MM, D’Avino G, Hulsen MA, Greco F, Maffettone PL. Particle motion in square channel flow of a viscoelastic liquid: migration vs. secondary flows. J Non Newton Fluid Mech. 2013;195:1–8.View ArticleGoogle Scholar
- Young Kim J, Won Ahn S, Sik Lee S, Min Kim J. Lateral migration and focusing of colloidal particles and DNA molecules under viscoelastic flow. Lab Chip. 2012;12:2807.View ArticleGoogle Scholar
- Ahn SW, Lee SS, Lee SJ, Kim JM. Microfluidic particle separator utilizing sheathless elasto-inertial focusing. Chem Eng Sci. 2015;126:237–43.View ArticleGoogle Scholar
- Yang S, Kim JY, Lee SJ, Lee SS, Kim JM. Sheathless elasto-inertial particle focusing and continuous separation in a straight rectangular microchannel. Lab Chip. 2011;11:266–73.View ArticleGoogle Scholar
- Nam J, Lim H, Kim D, Jung H, Shin S. Continuous separation of microparticles in a microfluidic channel via the elasto-inertial effect of non-Newtonian fluid. Lab Chip. 2012;12:1347.View ArticleGoogle Scholar
- Liu C, Xue C, Chen X, Shan L, Tian Y, Hu G. Size-based separation of particles and cells utilizing viscoelastic effects in straight microchannels. Anal Chem. 2015;87:6041–8.View ArticleGoogle Scholar
- Li G, McKinley GH, Ardekani AM. Dynamics of particle migration in channel flow of viscoelastic fluids. J Fluid Mech. 2015;785:486–505.View ArticleGoogle Scholar
- Rådström P, Knutsson R, Wolffs P, Lövenklev M, Löfström C. Pre-PCR processing. Mol Biotechnol. 2004;26:133–46.View ArticleGoogle Scholar
- Wiesinger-Mayr H, Jordana-Lluch E, Martró E, Schoenthaler S, Noehammer C. Establishment of a semi-automated pathogen DNA isolation from whole blood and comparison with commercially available kits. J Microbiol Methods. 2011;85:206–13.View ArticleGoogle Scholar
- Al-Soud WA, Radstrom P. Purification and characterization of PCR-inhibitory components in blood cells. J Clin Microbiol. 2001;39:485–93.View ArticleGoogle Scholar
- Nam J, Namgung B, Lim CT, Bae J-E, Leo HL, Cho KS, et al. Microfluidic device for sheathless particle focusing and separation using a viscoelastic fluid. J Chromatogr A. 2015;1406:244–50.View ArticleGoogle Scholar
- Won Seo K, Ran Ha Y, Joon Lee S. Vertical focusing and cell ordering in a microchannel via viscoelasticity: applications for cell monitoring using a digital holographic microscopy. Appl Phys Lett. 2014;104:213702.View ArticleGoogle Scholar
- Lim EJ, Ober TJ, Edd JF, Desai SP, Neal D, Bong KW, et al. Inertio-elastic focusing of bioparticles in microchannels at high throughput. Nat Commun. 2014;5:4120.Google Scholar
- Yang S, Lee SS, Ahn SW, Kang K, Shim W, Lee G, et al. Deformability-selective particle entrainment and separation in a rectangular microchannel using medium viscoelasticity. Soft Matter. 2012;8:5011.View ArticleGoogle Scholar
- Cha S, Shin T, Lee SS, Shim W, Lee G, Lee SJ, et al. Cell stretching measurement utilizing viscoelastic particle focusing. Anal Chem. 2012;84:10471–7.View ArticleGoogle Scholar
- Hansson J, Karlsson JM, Haraldsson T, Brismar H, van der Wijngaart W, Russom A. Inertial microfluidics in parallel channels for high-throughput applications. Lab Chip. 2012;12:4644.View ArticleGoogle Scholar