Article pubs.acs.org/ac
Rapid Identification and Susceptibility Testing of Uropathogenic Microbes via Immunosorbent ATP-Bioluminescence Assay on a Microfluidic Simulator for Antibiotic Therapy Tao Dong*,†,‡,§ and Xinyan Zhao‡,§ †
Institute of Applied Micro-Nano Science and Technology, Chongqing Engineering Laboratory for Detection, Control and Integrated System, Chongqing Technology and Business University, Nan’an District, Chongqing 400067, China ‡ Department of Micro and Nano Systems Technology (IMST), Faculty of Technology and Maritime Sciences (TekMar), Buskerud and Vestfold University College (HBV), Borre N3184, Norway S Supporting Information *
ABSTRACT: The incorporation of pathogen identification with antimicrobial susceptibility testing (AST) was implemented on a concept microfluidic simulator, which is well suited for personalizing antibiotic treatment of urinary tract infections (UTIs). The microfluidic device employs a fiberglass membrane sandwiched between two polypropylene components, with capture antibodies immobilized on the membrane. The chambers in the microfluidic device share the same geometric distribution as the wells in a standard 384-well microplate, resulting in compatibility with common microplate readers. Thirteen types of common uropathogenic microbes were selected as the analytes in this study. The microbes can be specifically captured by various capture antibodies and then quantified via an ATP bioluminescence assay (ATP-BLA) either directly or after a variety of follow-up tests, including urine culture, antibiotic treatment, and personalized antibiotic therapy simulation. Owing to the design of the microfluidic device, as well as the antibody specificity and the ATP-BLA sensitivity, the simulator was proven to be able to identify UTI pathogen species in artificial urine samples within 20 min and to reliably and simultaneously verify the antiseptic effects of eight antibiotic drugs within 3−6 h. The measurement range of the device spreads from 1 × 103 to 1 × 105 cells/mL in urine samples. We envision that the medical simulator might be broadly employed in UTI treatment and could serve as a model for the diagnosis and treatment of other diseases.
U
rinary tract infection (UTI) is among the most common bacterial infection and poses a significant healthcare burden.1−3 Almost 50% of the global population will experience a UTI at some point in their lives.1,4 UTIs are mainly caused by bacteria. The most common UTI-causing organism is Escherichia coli, with 80%−85% of the cases originating from these bacteria.3 Staphylococcus saprophyticus are responsible for 5%−10% of UTI cases,3,5 and in some rare cases, UTIs can also be caused by viral or fungal infections.3,6 Other groups of bacteria such as Klebsiella, Proteus, Pseudomonas, and Enterobacter can also cause UTIs.3,4 UTIs often present as a clinical conundrum; UTIs are rarely fatal, yet are associated high morbidity and affect all patient demographics.1 UTIs are often treated with antibiotics and by the removal of any urinary tract catheters and obstructions.4 The standard diagnosis of UTIs in hospitals relies on culturebased identification and antimicrobial susceptibility testing (AST), which have a typical delay of 2−3 days.7,8 In the absence of a definitive microbiological diagnosis, doctors frequently initiate an imprecise empirical broad-spectrum antibiotic treatment to meet the urgent needs of patients.1,9 © XXXX American Chemical Society
A wide variety of antimicrobial agents are used in the prophylaxis and treatment of UTI, but the widespread misuse and overuse of antibiotics has accelerated the selection of resistant pathogens and decreased the lifespan of antibiotics.1,9 A growing problem of worldwide concern is the increasing resistance of pathogens to conventional antibiotics.10 Thus, there is significant interest in developing rapid diagnostics of UTI, including pathogen identification and AST. In the past decade, the available diagnostic methods for UTI have marvelously developed,1,11 including immunoassays,12 urine dipsticks,13 urinary flow cytometry,14,15 bacteriophage technology,16 microfluidic sensors,17,18 matrix-assisted laser desorption/ionization time-of-flight mass spectrometry,19 rapid molecular pathogen identification by polymerase chain reaction,20 next-generation sequencing,21 and microarray analysis.22 Although the diagnostic technology has advanced, Received: November 20, 2014 Accepted: January 13, 2015
A
DOI: 10.1021/ac504428t Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry these novel methods remain relatively slow to realize timely and judicious antibiotic treatment in practice, and it remains difficult to provide intact and definitive microbiological diagnostic and susceptibility information, especially when some unknown genetic mutation mechanisms appear in the new resistant uropathogenic strains.11 Compared with these emerging methods, the culture-based method remains robust and more reliable. In fact, many methods and platforms for culture-based AST have been developed.23−26 The current technology bottleneck is at the level of sample preparation and system integration.1,27,28 In this study, a cell-based lab-on-a-chip (LOC) system employs the immunosorbent ATP-bioluminescence assay (IATP-BLA) as a fundamental measuring method to identify and evaluate the properties of uropathogenic bacteria. The ATP bioluminescence assay (ATP-BLA) is a multistep process involving the luciferase enzyme, luciferin substrate, oxygen, magnesium cation, and adenosine triphosphate (ATP), which will induce the efficient emission of photons (550−570 nm).29 Target microbes in the urine sample will be captured by specific antibodies on the fiberglass membrane inside of the device and then encapsulated by the calcium alginate gel in situ. Given enough growth medium, the immobilized microbes can survive there for on-chip AST prior to the IATP-BLA quantification. Basically, the magnitude of the ATP bioluminescence signal is proportional to the amount of corresponding living microbes. The entire process is smoothly integrated in this microfluidic simulator, so the specificity of immunoassay and the ultrasensitivity of the ATP-bioluminescence assay can be perfectly integrated. Compared with traditional microbial culture, the test cycle is reduced from a few days down to hours or even minutes. The 384-chamber microfluidic simulator can simultaneously perform an AST of 8 selected antibiotics on 13 types of microbe strains.
Figure 1. Architecture of the microfluidic simulator for the diagnosis and treatment of UTI. (A) Downward view of the chip layout. Except for the top and bottom covers, the chip comprises three layer components: the sample layer (SL), the membrane layer (ML), and the culture layer (CL). Air veins are engraved on the top of the SL; 16 × 24 vertical holes in both SL and CL are drilled through according to the dimensions of a standard 384-well microplate. (B) Upward view of the chip layout. Both SL and CL are engraved on double faces. Sample veins are on the lower face of the SL, and 11 groups of culture medium veins (CMVs) are engraved on the lower face of the CL. (C) The inlets and outlets of the device. Inlet A and outlet A are for air veins; inlet S and outlet S are for sample veins. The 11 pairs of inlets and outlets on the side face are for 11 groups of CMVs. (D) The meticulous structure of vertical reaction chambers is shown in the cross-sectional view of the device. (E) Thirteen groups of detection channels. Each group of CMVs is an independent detection channel in the simulator. Every CMV covers 16 reaction chambers, including 13 IgY-coated detection units for corresponding uropathogen, an ovalbumin control, a normal mouse serum control, and a blank control. Groups 1−2 have three replicates, respectively; groups 3−11 have only two replicates.
■
EXPERIMENTAL SECTION Chip Design and Fabrication. The microfluidic simulator is designed with 384 vertical reaction chambers, in which parallel reactions can be operated by a few processes on the valve and pump system. The position of each chamber in the simulator corresponds to that of each well on a standard 384well microplate, as shown in Figure 1. This design ensures that the microfluidic simulator can be directly read by a microplate reader. Accordingly, the existing instrument can be a reliable platform for further development of a relevant automatic simulation system.30 As a disposable biochip, the microfluidic device was made of two white polystyrene (PS) sheets, which were processed by laser ablation.31 A piece of Waterman grade GF-D fiberglass membrane filter was tightly clipped between two ablated PS layers. Through 24 “Culture Medium Veins” (CMVs), varied reagents can be lead into each reaction chamber (Figure 1, inset B and D). In total, there are 11 groups of CMVs on the culture layer. Each group shares the same inlet and outlet pair (Figure 1, inset C). More details of chip design and fabrication are illustrated in the Supporting Information. Each CMV is designed to pass through 16 vertical reaction chambers. Each spot on the fiberglass membrane inside of the 13 chambers is immobilized with 13 types of specific chicken egg yolk antibodies (IgY) against their respective corresponding uropathogenic microbes, including Escherichia coli, Staphylococcus saprophyticus, Staphylococcus epidermidis, Staphylococcus aureus, Klebsiella pneumonia, Proteus mirabilis, Pseudomonas aeruginosa, Enterobacter cloacae, Enterococcus faecium, Enter-
ococcus faecalis, Streptococcus viridans, Streptococcus pyogenes, and Candida albicans. The other three chambers in each CMV are for controls, including a blank control, a negative control (coated with normal chicken egg ovalbumin), and a normal mouse serum control. The 13 types of IgY were extracted from relevant immunized chicken eggs by the EGGstract IgY purification system (Promega, Madison, WI). The inactivated vaccines for immunizing the chickens were made from clinically isolated clones from our cooperative hospitals via traditional methods.32,33 Surface treatment of the fiberglass membrane was prepared in the same manner as poly-L-lysine-coated glass slides.34 Fluidic Manipulation. The schematic of on-chip IATPBLA is shown in Figure 2. Approximately 5 mL of urine sample is introduced every time. Owing to the hydrophilicity of glass fibers, the aqueous phase laterally penetrates toward the nearby B
DOI: 10.1021/ac504428t Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
the sample veins and penetrates into the interspace of the glass fibers that lies beneath. Accordingly, each hydrophilic chamber is isolated. All of the aforementioned steps are accurately operated by a manual valve and pump system in this study, as previously described,36 but the automation of fluidic manipulation should be realized in the future. The 11 groups of CMVs in the microfluidic simulator represent 11 detection channels (Figure 1, inset B). Channel 1 is often used to identify the class of uropathogenic microbe by the rapid IATP-BLA test, which could be completed within 20 min, including sample processing. ATP monitoring buffer is prepared by mixing the BacTiter-Glo buffer and its substrate (Promega, Madison, WI) before use. The buffer is accurately injected into inlets within 30 s; after that, the microfluidic simulator is quickly transferred into a microplate reader. The reading starts at the fifth minute after injection of the ATP monitoring buffer. If the ATP-BLA test in Channel 1 returns negative results, Channel 2 will be filled with BD BBL Mueller Hinton II Broth medium (Becton, Dickinson and Company, New Jersey) and incubated at 310 K for 3 h prior to the second round of the ATP-BLA test (Figure 3). When any positive result indicates that some uropathogenic bacteria exist, channels 3−11 will be employed to study their susceptibility to antibiotics. Up to eight types of antibiotic candidates can be arranged for the subsequent AST according to the pharmacopoeia and allergy conditions of the patient. Except for the blank channel, these channels are respectively filled with growth medium or artificial urine containing specific antibiotics and incubated for 3 h before ATP-BLA tests. The resistance of pathogens can be deduced from the bioluminescence signals. Rapid Identification of Pathogens. The 13 types of microbe clones were provided by our cooperative hospitals; these clones were same that were used as the antigens for chicken immunization. After 8−12 h proliferation at 310 K in BD BBL Mueller Hinton II Broth medium, the 13 microorganism cultures were rapidly counted under a microscope. Flinn Scientific artificial urine was spiked with the 13 types of quantified microbes to a respective concentration of 1 × 106 cells/mL. Then, the multimicrobial artificial urine was further diluted into a series of gradient mixture in Flinn Scientific artificial urine (1 × 105, 1 × 104, 1000, 100, 10, and 1 cell/mL). The series of multimicrobial artificial urine samples were tested by the microfluidic simulators. Channels 1−2 in each simulator device were tested by IATP-BLA. The relative luminescence unit (R.L.U.) data were collected by a microplate reader and analyzed by using Microsoft Excel and OriginPro. Moreover, the 13 microbe cultures were individually diluted by Flinn Scientific© artificial urine into 13 respective solutions of 1000 cells/mL. These artificial bacteriuria samples with single species were also tested by the microfluidic simulators to verify the onchip IATP-BLA specificity. Optimization of the Protocols for the Microfluidic Simulator. Nonspecific adsorption in the immunocapture step should be reduced by sufficient rinsing. The effect of the washing process was studied, in which experiment the multimicrobial artificial urine of 1000 cells/mL was used as the sample. After immunocapture, parallel microfluidic simulators were rinsed with calcium buffer at the same velocity of 50 μL/s, but the rinse durations on different simulators varied from 30 to 150 s. The R.L.U. data of the IATP-BLA tests were measured.
Figure 2. Typical protocol of the IATP-BLA test in the microfluidic simulator. The IATP-BLA test is processed in the vertical reaction chamber unit. Specific IgY antibodies were immobilized on the fiberglass membrane during fabrication. The standard operating procedure is demonstrated as follows: (1) the urine sample is loaded into the sample veins, and the specific microbes are captured by IgY antibodies; (2) unbound microbes are removed; (3) the captured cells are encapsulated in situ by calcium alginate gel; (4) the sodium alginate solution is washed away; and (5) paraffin oil is injected to isolate each reaction chamber. After that, different options will be selected for different channels. (6.0) The captured microbes can be reproduced before ATP-BLA; (6.2) the captured microbes can be inhibited by series of antibiotic regiments before ATP-BLA; or (6.1) the captured microbes can be inhibited by a single antibiotic in AST. Either way, (7) the cells will finally be quantified by ATP-BLA in a microplate reader.
reaction chambers with small particles such as bacteria, but large clots or mucilage are obstructed due to the limited size (
