Elsevier

Biosensors and Bioelectronics

Volume 49, 15 November 2013, Pages 118-125
Biosensors and Bioelectronics

A multiplexed microfluidic platform for rapid antibiotic susceptibility testing

https://doi.org/10.1016/j.bios.2013.04.046Get rights and content

Highlights

  • Developed a microfluidic based biosensor platform for rapid antibiotic susceptibility testing.

  • Demonstrated advantages: low volume requirement, enhanced detection sensitivity and portability.

  • Enabled direct enumeration of phenotypes such as growth & morphology by detecting fluorescence.

  • Highlighted importance of thorough evaluation of antibiotic combinations prior to clinical use.

  • Showcased promise for utility of the platform for diagnosis and management of infections.

Abstract

Effective treatment of clinical infections is critically dependent on the ability to rapidly screen patient samples to identify antibiograms of infecting pathogens. Existing methods for antibiotic susceptibility testing suffer from several disadvantages, including long turnaround times, excess sample and reagent consumption, poor detection sensitivity, and limited combinatorial capabilities. Unfortunately, these factors preclude the timely administration of appropriate antibiotics, complicating management of infections and exacerbating the development of antibiotic resistance. Here, we seek to address these issues by developing a microfluidic platform that relies on fluorescence detection of bacteria that express green fluorescent protein for highly sensitive and rapid antibiotic susceptibility testing. This platform possesses several advantages compared to conventional methods: (1) analysis of antibiotic action in two to four hours, (2) enhanced detection sensitivity (≈1 cell), (3) minimal consumption of cell samples and antibiotic reagents (<6 µL), and (4) improved portability through the implementation of normally closed valves. We employed this platform to quantify the effects of four antibiotics (ampicillin, cefalexin, chloramphenicol, tetracycline) and their combinations on Escherichia coli. Within four hours, the susceptibility of bacteria to antibiotics can be determined by detecting variations in maxima of local fluorescence intensity over time. As expected, cell density is a major determinant of antibiotic efficacy. Our results also revealed that combinations of three or more antibiotics are not necessarily better for eradicating pathogens compared to pairs of antibiotics. Overall, this microfluidic based biosensor technology has the potential to provide rapid and precise guidance in clinical therapies by identifying the antibiograms of pathogens.

Introduction

In recent years, antibiotic resistance traits among microbial pathogens have escalated at alarming rates, which has spurred the development of technologies for rapid and accurate detection of antibiotic susceptibility profiles of pathogens (Rice, 2010). Established techniques for antibiotic susceptibility testing (AST), such as broth dilution and disc diffusion, involve multiple time-consuming steps (Lazcka et al., 2007, White et al., 1996) including: (1) isolation of pathogens from patient samples (24–48 h.), (2) pre-culturing of isolated bacteria to enrich cell density to detectable levels (24–48 h.), (3) incubation of cells with antibiotics in 96-well plates or petri dishes (24–48 h.), and (4) determination of bacterial growth using absorption spectroscopy or by visual assessment. Broth dilution and disc diffusion assays typically require significant quantities (10–30 mL) of patient samples such as blood, sputum, or urine for analysis (Mancini et al., 2010). In addition, the limited sensitivity of macroscale techniques for AST makes them unsuitable for detecting the presence of “persister” microbes. Although persister cells represent only a small fraction (≈10−5) of microbial cells, they tend to evade antibiotic mediated killing by switching to a metabolically dormant or “persistent” state (Balaban et al., 2004, Lewis, 2010). Persister cells constitute a significant threat due to their ability to re-initiate infection upon discontinuation of antibiotic therapy (Dawson et al., 2011). Finally, inconsistencies in results obtained from different AST techniques further complicate diagnosis and treatment (Gales et al., 2001, Goldstein et al., 2007, Lo-Ten-Foe et al., 2007, Nicodemo et al., 2004, Tan and Ng, 2007, Traub, 1970). Hence, in the absence of precise information about the antibiogram of particular pathogen, physicians often resort to empirical therapies that utilize broad-spectrum antibiotics. Indiscreet use of antibiotics in this manner is known to intensify the problem of antibiotic resistance (Alanis, 2005, Ang, 2001).

To address the aforementioned issues, biosensor platforms with improved sensitivity and fast analysis time have been developed for antimicrobial susceptibility testing (Chiang et al., 2009, Karasinski et al., 2007, Kinnunen et al., 2011, Koydemir et al., 2011, Nakamura et al., 1991, Tsou et al., 2010). For example, electrochemical sensors have been utilized to determine susceptibility by measuring small changes in growth of cells (Karasinski et al., 2007). Chiang et al., have developed a surface plasmon resonance-based biosensor platform to categorize strains as susceptible or resistant by detecting variations in optical properties of bacteria when treated with antibiotics (Chiang et al., 2009). Another interesting approach for antibiotic susceptibility testing utilizes an asynchronous magnetic bead rotation biosensor to monitor single cells or cell populations after treatment with antibiotics (Kinnunen et al., 2011). In addition, filter chip and optical detection biosensing system have been developed that can provide susceptibility results in one hour (Tsou et al., 2010). These microfluidic-based biosensor technologies are sensitive and rapid, however, most of these platforms lack multiplexing capabilities (Chiang et al., 2009, Kinnunen et al., 2011, Tsou et al., 2010). Hence, integrated microfluidics represents an attractive technology for the multiplexed implementation of biological assays with rapid turnaround times and minimal sample consumption (Sia and Whitesides, 2003). Several successful microfluidic platforms for AST have been reported (Boedicker et al., 2008, Chen et al., 2010, Churski et al., 2012, Cira et al., 2012, Ho et al., 2012, Kalashnikov et al., 2012, Sun et al., 2011). For example, droplet-based microfluidics has been utilized to compartmentalize bacterial cells, nutrients, antibiotics, and fluorescent viability indicators in water-in-oil emulsions (Boedicker et al., 2008, Churski et al., 2012). Sun et al. have reported on the development of a microfluidic platform for the confinement of bacterial cells in square microwells connected to a central flow channel that continuously delivers nutrients and antibiotics to cells (Sun et al., 2011). Choi et al. have reported a microfluidic agarose channel system for rapid antibiotic susceptibility testing by tracking single cell growth (Choi et al., 2013). Weibel and colleagues have developed a portable microfluidic chip for AST for point-of-care use (Cira et al., 2012). The key advantage of the portable chip is the automatic loading of bacterial cells into microfluidic chambers that had been preloaded with dehydrated antibiotics using a ‘degas driven flow’.

The existing approaches for microfluidic-based antibiotic susceptibility testing offer promising routes toward the development of a rapid and portable screening tool. However, many of these methods suffer from one or more of the following limitations: (1) complicated platform fabrication and/or operation procedures (Kalashnikov et al., 2012), (2) poor portability due to the requirement for syringe pumps, pneumatic actuators, and other ancillary equipment (Choi et al., 2013, Churski et al., 2012, Kalashnikov et al., 2012), and (3) unstable droplet formation (Theberge et al., 2010). In this work, we report on the design and fabrication of a microfluidic platform with biosensing capabilities featuring a spatially addressable 4×6-array of wells to simultaneously monitor the effects of multiple antibiotics at different concentrations, as well as their combinations, on bacterial cells for AST. This technology integrates ease-of-fabrication and use with enhanced combinatorial capabilities, and further provides improved portability and usability by circumventing the requirement for expensive syringe pumps and pneumatic actuators by implementing normally closed valves. In addition, the platform is amenable to automated analysis by using time-lapse fluorescence microscopy (TLFM). We employed the microfluidic platform to interrogate the antibiotic sensitivity profile of Escherichia coli to four commonly used bactericidal and bacteriostatic antibiotics. Furthermore, we explored synergistic and antagonistic effects of different antibiotic cocktails, as well as the effects of E. coli cell densities on dictating the efficiency of antibiotic action. Overall, this platform capitalizes on several key advantages of biosensor based integrated microfluidics technology including miniaturization of assays, expedited analysis, multiplexing, and improved detection sensitivity along with ease-of-use and portability.

Section snippets

Microfluidic chip fabrication

The microfluidic chip for AST was fabricated using standard soft lithographic techniques (Xia and Whitesides, 1998). Briefly, molds for casting the fluidic and control layers were made by patterning negative photoresist on silicon wafers using photolithography. A thin layer of 20:1 PDMS (weight ratio of polymer to cross-linker) was spin coated on to the fluidic layer master and 5:1 PDMS was poured on to the control layer master. The two layers were partially cured at 65 °C for 30 minutes. Next,

Design and validation of the microfluidic platform for biological studies

The microfluidic platform used in this study (Fig. 1) consists of a two-layer poly (dimethyl-siloxane) chip comprising: (1) a control layer for actuating the mixing and filling valves, and (2) a fluidic layer that houses the flow channels and a 4×6-array of wells. Each well consists of two half wells that house the cells and antibiotic solutions, respectively. Each half well is 400 μm wide, 15 μm tall and 500 μm long. In this way, the 4×6 array design facilitates the treatment of one microbial

Conclusions

In summary, we have developed a multiplexed microfluidic platform comprised of arrays of small volume chambers (3 nL) that enables monitoring the effects of several antibiotics and their combinations on E. coli cells. Conventional methods for antibiotic susceptibility testing (e.g., broth dilution and disc diffusion) are time consuming and tedious, generally requiring large sample and reagent volumes. From this perspective, our microfluidic biosensing platform provides substantial improvements

Acknowledgments

We acknowledge financial support from the National Science Foundation under awards CMMI 03–28162 and CMMI 07–49028 to Nano-CEMMS; Nano Science & Engineering Center (NSEC) on Nanomanufacturing for PJAK and a Packard Fellowship from the David and Lucile Packard Foundation for CMS. AM was supported in part by FMC Technologies Graduate Fellowship. We thank Dr. Amit Desai and Dr. Ashtamurthy Pawate for helpful discussions and for proof-reading the manuscript. In addition, we acknowledge Dr. Desai's

References (55)

  • A.J. Alanis

    Archives of Medical Research

    (2005)
  • M. Allewelt et al.

    Clinical Microbiology and Infection

    (2004)
  • Y.-L. Chiang et al.

    Biosensors and Bioelectronics

    (2009)
  • G. Cottarel et al.

    Trends in Biotechnology

    (2007)
  • P. Dejace et al.

    American Journal of Medicine

    (1986)
  • J. Karasinski et al.

    Biosensors and Bioelectronics

    (2007)
  • P. Kinnunen et al.

    Biosensors and Bioelectronics

    (2011)
  • O. Lazcka et al.

    Biosensors and Bioelectronics

    (2007)
  • R. Mohan et al.

    Sensors and Actuators B Chemical

    (2011)
  • N. Nakamura et al.

    Biosensors and Bioelectronics

    (1991)
  • P. Sun et al.

    Biosensors and Bioelectronics

    (2011)
  • T.Y. Tan et al.

    Clinical Microbiology and Infection

    (2007)
  • D.J. Tipper

    Pharmacology and Therapeutics

    (1985)
  • P.-H. Tsou et al.

    Biosensors and Bioelectronics

    (2010)
  • B.S.P. Ang

    Annals of the Academy of Medicine Singapore

    (2001)
  • N.Q. Balaban et al.

    Science

    (2004)
  • J.Q. Boedicker et al.

    Lab on a Chip

    (2008)
  • M.T. Butler et al.

    Proceedings of the National Academy of Sciences of the United States of America

    (2010)
  • R. Chait et al.

    Nature

    (2007)
  • C.H. Chen et al.

    Analytical Chemistry

    (2010)
  • J. Choi et al.

    Lab on a Chip

    (2013)
  • K. Churski et al.

    Lab on a Chip

    (2012)
  • N.J. Cira et al.

    Lab on a Chip

    (2012)
  • C.C. Dawson et al.

    PLoS Pathogens

    (2011)
  • T. Dorr et al.

    PLoS Genetics

    (2009)
  • A.C. Gales et al.

    Journal of Clinical Microbiology

    (2001)
  • F.W. Goldstein et al.

    Journal of Antimicrobial Chemotherapy

    (2007)
  • Cited by (0)

    1

    These authors contributed equally to the work.

    View full text