A Microfluidic System for Saliva-BasedDetection of Infectious Diseases

ZONGYUAN CHEN,a MICHAEL G. MAUK,a JING WANG,a

WILLIAM R. ABRAMS,b PAUL L. A. M. CORSTJENS,c

R. SAM NIEDBALA,d DANIEL MALAMUD,b

AND HAIM H. BAUa

aUniversity of Pennsylvania School of Engineering and Applied Science,Philadelphia, Pennsylvania, USAbNew York University College of Dentistry, New York, New York, USAcLeiden University Medical Center, Leiden, the NetherlandsdDepartment of Chemistry, Lehigh University, Bethlehem, Pennsylvania, USA

ABSTRACT: A “lab-on-a-chip” system for detecting bacterial pathogensin oral fluid samples is described. The system comprises : (1) an oral fluidsample collector; (2) a disposable, plastic microfluidic cassette (“chip”)for sample processing including immunochromatographic assay with anitrocellulose lateral flow strip; (3) a platform that controls the cassetteoperation by providing metered quantities of reagents, temperature reg-ulation, valve actuation; and (4) a laser scanner to interrogate the lateralflow strip. The microfluidic chip hosts a fluidic network for cell lysis, nu-cleic acid extraction and isolation, PCR, and labeling of the PCR productwith bioconjugated, upconverting phosphor particles for detection on thelateral flow strip.

KEYWORDS: microfluidics; lab-on-a-chip; infectious diseases; diagno-stics; immunochromatography

INTRODUCTION AND BACKGROUND

Currently, point-of-care (POC) testing for infectious diseases is routinelyperformed using immunochromatographic assays of serum, urine, or saliva.These assays are typically implemented as easy-to-use nitrocellulose lateralflow test strips. Similar test strips are widely employed for a variety of applica-tions, such as home pregnancy testing and detecting drugs of abuse. Serologiclateral flow strips that detect HIV antibodies in oral fluid are completely self-contained, requiring no addition of reagents or accessory instrumentation.1

Address for correspondence: Haim H. Bau, 216 Towne Building, Mechanical Engineering andApplied Mechanics, University of Pennsylvania, Philadelphia, PA 19104-6315. Voice: 215-898-8363;fax: 215-573-6114.

bau@seas.upenn.edu

Ann. N.Y. Acad. Sci. 1098: 429–436 (2007). C© 2007 New York Academy of Sciences.doi: 10.1196/annals.1384.024

429

430 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES

The strips include a control test function for quality assurance, and can pro-vide test results in less than 10 min. The lateral flow test strip device providesconvenient, fast, and low-cost POC and home testing. However, there is con-siderable incentive for improving sensitivity and increasing functionality. Forexample, a more quantitative assay could be used to determine viral loads.Additionally, parallel testing of nucleic acid (DNA and RNA) and antigentargets, from both viral and bacterial pathogens, as well as detection of an-tibodies, would enable simultaneous robust, multipurpose, and confirmatorytesting from the same sample with the same test device. However, multiplexednucleic acid testing and immunoassays require elaborate sample processing,which complicates implementation as a simple-to-use POC device. Microflu-idic “lab-on-a-chip” devices in the form of miniaturized networks of conduits,chambers, and valves formed in a plastic substrate “chip” facilitate samplemetering, cell and virus lysis, isolation of nucleic acids and proteins, ampli-fication of nucleic acids using PCR or analogous techniques, and labeling ofanalytes for detection. Thus, a lateral flow strip combined with a microflu-idic cassette for sample processing would significantly expand the range ofapplications and tasks that could be performed by immunochromatographicmethods.

Most lateral flow test strips use colloidal gold, dye, or latex beads as re-porters to allow visual inspection of test results.1 Lateral flow strip tests thatincorporate upconverting phosphors as reporters have also been developed.2–5

Upconverting phosphor technology (UPT) is based on sub-micron-sized ce-ramic particles coated with lanthanides that absorb infrared light (excitation)and emit visible light (response signal). The particles can be functionalizedwith antigens and antibodies for use as labels on lateral flow strips. UPT par-ticles avoid interference from background fluorescence and do not exhibitphotobleaching effects. UPT reporters have been shown to improve the limitof detection in various bioassays by 10-fold or more.3–6 In one particular ap-plication, a POC diagnostics system (OraSure UplinkTM) includes an FDA-approved oral fluid sample collector, a molded plastic cartridge that housesthe lateral flow strip and features a loading port for the sample collector, and atabletop reader instrument with laser scanner and photomultiplier tube detectorto interrogate the lateral flow strip.7 Also, UPT-based immunochromatogra-phy flow strips have been adapted for detecting nucleic acid targets,3–5 suchas UPT-labeled PCR products, thus opening the way for applying this tech-nology for nucleic acid testing. This UPT lateral flow system thus providesan excellent platform for developing a microfluidics system for multiplexingnucleic acid–based diagnostics and immunoassays. The approach is to developa disposable credit card–sized plastic cassette that hosts a microfluidic circuitintegrating components for (1) sample metering; (2) enzymatic and chemicallysis using lysozyme, proteinase K, chaotropic salts, and detergents; (3) solid-phase extraction using a porous silica membrane embedded in the cassette; (4)PCR amplification; (5) labeling of the PCR product with UPT phosphor; and

CHEN et al. 431

(6) immunochromatography assay of the PCR amplicon with a nitrocelluloselateral flow strip mounted in the cassette.

Efforts to add microfluidic capabilities to lateral flow strip assays have al-ready met with success. Chen et al.8 developed a consecutive lateral flow assaymicrofluidic chip with UPT reporters for use with the UPlinkTM laser scanner.The chip includes functions for sample introduction, buffer distribution, me-tering, mixing, and thermopneumatic pumping. After a serum sample is loadedon the chip, the chip is connected to an instrumentation platform that providesfluidic power, temperature control, and valve actuation. Automated process-ing on the chip comprises a sequence of steps including sample aliquoting andmetering, dilution with buffer, and blotting on a nitrocellulose strip followedby a separate wash step of the strip with buffer, followed by addition of bufferwith UPT label. The UPT labels are dry-stored and preloaded on the chip in alyosphere. The chip is then inserted into the UPlinkTM reader for interrogationof the strip by IR (980 nm) and signal readout. Along similar lines, Wanget al.9 reported a disposable microfluidic cassette for DNA amplification anddetection combining a microfluidic PCR chamber and an incubation chamberto label the PCR product with UPT particles, integrated with a lateral flow stripto capture UPT-labeled PCR amplicon for detection in the laser scanner. Here,we report a microfluidic cassette for PCR-based detection of Gram-positivebacteria using a lateral flow strip with UPT reporters. The cassette providescomplete processing including sample introduction and metering, enzymaticcell lysis, nucleic acid isolation, PCR amplification with pathogen-specificprimers conjugated for labeling and capture on a lateral flow strip, and blot-ting of labeled PCR product on the lateral flow strip affixed to the cassette.The microfluidic system is tested with diluted B. cereus, a safe surrogate foranthrax, but the processing steps are sufficiently generic and the chip designis sufficiently flexible so that it can be adapted for detection of different viraland bacterial pathogens.

CHIP DESIGN, FABRICATION, AND OPERATION

FIGURE 1A is a photograph of a microfluidic chip for PCR-based detectionof bacteria targets. The chip is made as a thermally bonded three-layer lam-inate of polycarbonate sheets with channels and chambers defined in two ofthe layers by CNC machining. Cross-section dimensions of the conduits rangefrom 250 to 500 �m. The microfluidic cassette is mounted in a molded plasticcartridge compatible with insertion into the UPlinkTM laser scanner/reader unit(FIG. 1B). Phase-change “ice” valves actuated by small, electrically controlledthermoelectric elements on the processing platform are used for flow controland sealing the 10-�L PCR chamber.10 Flow is halted when the fluid in a chan-nel freezes at the section of conduit in contact with a cooled thermoelectricelement, and flow resumes when the thermoelectric element is heated to thaw

432 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES

FIGURE 1. (A) Polycarbonate microfluidic cassette for bacteria detection. The cassettefeatures conduits, chambers, and phase-change valve sites for sample metering, lysis, solid-phase extraction, and isolation of nucleic acids with a porous silica membrane-bindingphase, PCR, labeling, and nitrocellulose lateral flow test strip. The cassette measures 85 ×35 mm and is 3 mm thick. (B) Microfluidic cassettes with lateral flow strips are mountedin a molded plastic cartridge for the UPlinkTM laser scanner/reader unit.

the fluid. The PCR chamber is also heated and cooled by a thermoelectric ele-ment. The chip is interfaced with programmable syringe pumps, reservoirs forbuffers and reagents, and electronic controllers for the thermoelectric elements,controlled by a PC with LabVIEWTM software (FIG. 2).

CHEN et al. 433

FIGURE 2. Schematic of supporting instrumentation for microfluidic cassette includ-ing buffer and reagent reservoirs, programmable syringe pumps, thermoelectric elements foractuating phase-change “ice” valves and heating/cooling PCR chamber, controlled powersupplies, and PC.

The sample is introduced through a loading port and 50 �L of samplealiquoted in a metering chamber. The sample is then subjected to a two-step lysis process. The sample is first enzymatically digested for 30 min at37◦C with lysozyme (20 mg/mL lysozyme, 1.2% Triton X-100, 2mM EDTA,20 mM Tris-Cl), and then incubated with SDS detergent (0.1%), proteinaseK (10 mg/mL) and 6M guanidinium chloride at 60◦C for 15 min. The lysateis forced through a porous silica membrane embedded in the cassette. Thechaotropic salts induce nucleic acids binding to the silica. The silica-boundnucleic acids are washed with ethanol-based solutions to remove debris. Next,the nucleic acids are desorbed from the silica by elution with low-salt, pH-neutral buffer, and a 5-�L elution fraction is mixed with PCR reagents (0.3 UTaq polymerase, 200 �M dNTP, 0.1 �g/uL BSA, 50 mM Tris-Cl, 1.5–3.5 mMMgCl2) and primers (forward: 5′-TCT CGT TTC ACT ATT CCC AAG T-3′conjugated to dioxygenin; reverse: 5′ AAG GTT CAA AAG ATG GTA TTCAGG-3′ conjugated to biotin) and sealed in a 10-�L PCR chamber, where it

434 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES

is thermally cycled 25 times (95◦C, 15 sec; 55◦C, 25 sec; 72◦C, 20 sec). The305-bp amplicon is labeled by incubation at room temperature for 15 min with150 ng of avidin-conjugated UPT particles. The UPT-labeled PCR product isapplied to the sample pad of a lateral flow strip and captured by immobilizedanti-DIG antibodies at the test line stripe. The strip is then scanned with theUPlinkTM reader.

RESULTS

To assess microfluidic lysis and nucleic acid isolation for producing a DNAtemplate for PCR from cell culture samples, nucleic acid eluted from the chipsilica membrane was PCR-amplified and compared to PCR-amplified nucleicacid isolated by a standard benchtop protocol (QIAGEN DNeasyTM Kit). Inthe benchtop protocol, a 50-�L sample of B. cereus (∼106 cells/mL) was sub-jected to a two-step enzymatic lysis process with lysozyme, chaotropic salts,and detergents. The nucleic acid was then isolated from the lysate in successive

FIGURE 3. Comparison of on-chip and benchtop lysis and nucleic acid isolation. Todetermine the efficiency of on-chip lysis and nucleic acid isolation relative to benchtop lysisand nucleic acid isolation, elutions from the silica membrane on the chip and from a spincolumn were amplified by PCR and analyzed by gel electrophoresis. Control: PCR of pureB. cereus genomic DNA. Benchtop: B. cereus (50 �L, ∼106 cells/mL) was subjected totwo-step lysis (lysozyme digestion followed by treatment with chaotropic salt and proteinaseK according to QIAGEN DNeasyTM protocol). Nucleic acid was isolated from lysate usingQIAGEN DNeasyTM spin column. Aliquots of 5 �L from two successive 50-�L elutions(E1 and E2) from the spin column were mixed with 5 �L of PCR mix (Taq, dNTP, buffer,primers, BSA), amplified for 25 cycles, and run on an agarose gel (1.5%, ethidium bromidestaining). Chip runs (two replicates, I and II): Similar B. cereus samples (50 �L, ∼106

cells/mL) were loaded into the chip and subjected to two-step lysis and nucleic acid isolationusing the silica-membrane component of the chip. Aliquots (5 �L) from successive 50-�Lelutions (E1, E2, E3. . .) were mixed with 5 �L of PCR mix (Taq, dNTP, buffer, primers,BSA), amplified for 25 cycles on a benchtop thermal cycler, and run on an agarose gel(1.5%, ethidium bromide staining). Weak or absent PCR band for first elution (E1) for chipruns suggests a PCR inhibitor is contained in the initial elution. Subsequent elutions fromthe chip membrane, however, showed strong PCR amplification.

CHEN et al. 435

FIGURE 4. A representative UPlinkTM laser scan of the lateral flow strip for measure-ment of UPT-labeled PCR amplicon from bacteria (50 �L of ∼106 cells/mL) processed ona microfluidic cassette.

50-�L elutions from a QIAGEN DNeasyTM spin column (chaotrope–silicamethod), and amplified by benchtop PCR. Similarly, a 50-�L sample of B.cereus was loaded into the microfluidic cassette for on-chip lysis and nucleicacid isolation. Successive 50-�L elutions from the silica membrane embed-ded in the chip were collected and amplified by benchtop PCR. In each case,5-�L aliquots from successive 50-�L elution of the spin column and from thechip silica membrane were mixed with 5 �L of PCR mix (Taq, dNTP, buffer,primers, BSA), amplified for 25 cycles, loaded and run on a 1.5% agarosegel with ethidium bromide stain (FIG. 3). The weak or absent PCR productfrom the initial elution of the chip indicates the presence of an inhibitor, prob-ably residual ethanol from the silica membrane wash step. This initial elutionfraction should be avoided for providing a PCR template, but otherwise thesubsequent elutions from the chip yielded easily detected PCR product. Thereappears to be some benefit to removing inhibitors from the silica membraneby a vacuum drying step prior to elution (data not shown). To test the chip forcomplete sample processing (i.e., on-chip lysis, nucleic acid isolation, PCRamplification, labeling, and lateral flow assay), the chip was loaded with a50-�L aliquot of B. cereus (∼106 cells/mL), and operated according to thesequence detailed above. A typical UPlinkTM scan of the chip flow strip isshown in FIGURE 4, indicating a detectable B. cereus–specific PCR productproduced–from the sample.

DISCUSSION AND CONCLUSIONS

We have demonstrated a microfluidic cassette that can prepare UPT-labeledmicrofluidic PCR product for detection on a lateral flow strip assay interfacedwith the cassette, thus showing the feasibility of microfluidic PCR-based nu-cleic acid testing of raw samples. The current development effort is focused on

436 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES

Gram-positive bacterial targets (B. cereus, a safe surrogate for anthrax), but thesystem can be readily adapted for detection of viral and bacterial pathogens,multiplexed analysis of several pathogens, and serologic testing for antibod-ies. The device will serve as a module in a comprehensive, POC saliva-baseddiagnostic system that will also include modules for antigen and antibodyassays.

ACKNOWLEDGMENTS

The work was supported by NIH Grants UO1DE0114964 andUO1DE017855, and in collaboration with OraSure Technologies, Inc.(Bethlehem, PA, USA).

REFERENCES

1. KETEMA, F., H.L. ZINK, K.M. KREISEL, et al. 2005. A 10-minute, US Food andDrug Administration-approved HIV test. Expert Rev. Mol. Diagn. 5: 135–143.

2. NIEDBALA, R.S., H. FEINDT, K. KARDOS, et al. 2001. Detection of analytes byimmunoassay using up-converting phosphor technology. Anal. Biochem. 293:22–30.

3. CORSTJENS, P.L.A.M., M. ZUIDERWIJK, M. NILSSON, et al. 2003. Lateral-flow andup-converting phosphor reporters to detect single-stranded nucleic acids in asandwich-hybridization assay. Anal. Biochem. 312: 191–200.

4. CORSTJENS, P.L.A.M., M. ZUIDERWIJK, A. BRINK, et al. 2001. Use of up-convertingphosphor reporters in lateral flow assays to detect specific nucleic acid sequences:a rapid, sensitive DNA test to identify human papilloma type 16 infection. Clin.Chem. 47: 1885–1893.

5. HAMPL, J., M. HALL, N.A. MUFTI, et al. 2001. Upconverting phosphor reporters inimmunochromatographic assays. Anal. Biochem. 288: 176–187.

6. CORSTJENS, P.L.A.M., S. LI, M. ZUIDERWIJK, et al. 2005. Infrared up-convertingphosphors for bioassays. IEEE Proc. Nanobiotechnol. 152: 64–72.

7. MALAMUD, D., H. BAU, S. NIEDBALA & P. CORSTJENS. 2005. Point detection ofpathogens in oral samples. Adv. Dent. Res. 18: 12–16.

8. CHEN, Z., P.L.A.M. CORSTEJENS, M. ZUIDERWIJK, et al. 2005. A disposable mi-crofluidic point-of-care device for detection of HIV: a new upconverting phos-phor technology application. Proceedings of the 9th Int. Conf. Miniaturized Sys.Chem. Life Sci. 791–793.

9. WANG, J., Z. CHEN, P.L. CORSTJENS, et al. 2006. A disposable microfluidic cassettefor DNA amplification and detection. Lab. Chip 6: 46–53.

10. WANG, J., Z. CHEN, M. MAUK, et al. 2005. Self-actuated, thermo-responsive hy-drogel valves for lab on a chip. Biomed. Microdevices 7: 313–322.