Developing a multiplexed genotyping system for clinical diagnostic and pharmacogenetic testing.
 |
|
Figure 1. The eSensor XT-8
system by Osmetech Molecular Diagnostics (Pasadena, CA). |
Until recently, multiplex technology, such as real-time polymerase chain reaction (PCR), could analyze only a small number of polymorphisms in a single test. The multiplexing capabilities of DNA microarrays have improved, thereby increasing the number of independent genotypes that can be determined simultaneously. Early microarrays focused on high-density applications that were well suited for gene discovery. Recent developments of new detection methods and simplified methodologies have facilitated the transition from expensive high-density arrays to cost-effective, low-to-medium-density systems for clinical diagnostic and pharmacogenetic applications.1,2
Most microarray methods use optical detection of fluorescent, luminescent, or gold-nanoparticle-labeled probes that subsequently bind to surface immobilized probes.1–5 In some cases, multicolor labeling approaches are applied to allow for multiplex recognition.6 For example, the Verigene System by Nanosphere Inc. (Northbrook, IL) analyzes single-nucleotide variations using a gold-nanoparticle–based microarray platform.1 AutoGenomics Inc. (Carlsbad, CA) developed the Infiniti BioFilmChip that utilizes fluorescence detection on film-based microarrays for genomic and proteomic analyses.2 In general, most optical-detection-based approaches require large and expensive scanners, or charge-coupled-device–based imaging systems.
To develop cost-effective instrumentation and miniaturized point-of-care DNA diagnostic systems, IVD companies have taken a number of different approaches to electronic detection of DNA hybridization.7–9 Microarrays for such systems can be produced inexpensively with silicon-, ceramic-, or printed circuit board– based fabrication technologies.
Electrochemical test methods have been used in clinical chemistry, especially in blood glucose monitoring for diabetes care. However, although numerous studies have described molecular diagnostic technologies based on electrochemical detection, none have achieved practical implementation until recently.10,11
The eSensor cystic fibrosis carrier detection system by Osmetech Molecular Diagnostics (Pasadena, CA) is a DNA microarray product based on electrochemical detection that received FDA clearance and commercial application in the clinical laboratory.11 This system consists of the eSensor 4800 instrument, and the reagents and cartridges for genotyping 23 mutations and one polymorphism in the cystic fibrosis transmembrane regulator gene. This process is recommended for carrier screening by the American College of Obstetricians and Gynecologists, and the American College of Medical Genetics.12
This article discusses a new eSensor system that consists of the eSensor XT-8 instrument and microfluidic cartridges (see Figure 1). The new features in this system include higher array density, reduced hybridization time, and a more compact and modular instrument design offering true random-access operation and a user-friendly touch screen monitor. The modular design allows variation in the instrument capacity and throughput to address the testing needs of small hospital laboratories and larger reference laboratories. The first clinical application of the eSensor XT-8 system, the eSensor Warfarin Sensitivity Test, has been developed, and its performance has been evaluated in internal and clinical studies for FDA submission. However, performance characteristics have not yet been fully established.
DNA Detection Technology
 [5] |
|
Figure 2. (click to enlarge [5]) Schematic showing the principle of electrochemical detection of DNA using eSensor technology. Target (blue) is bound to the electrode via sequence-specific hybridization to the capture probe. Signal probe (red) hybridizes to the target sequence adjacent to the base of the capture probe, and the associated ferrocene labels are detected at the electrode surface by ACV.
|
The eSensor microarray is composed of a printed circuit board (PCB) consisting of an array of gold electrodes. Each electrode is modified with a multicomponent, self-assembled monolayer that includes presynthesized oligonucleotide capture probes (see Figure 2 [5]). Nucleic acid detection is based on a sandwich assay principle. Signal and capture probes are designed with sequences complementary to immediately adjacent regions on the corresponding target DNA sequence. A three-member complex is formed between capture probe, target, and signal probe based on sequence-specific hybridization. This process brings the 5'-end of the signal probe containing electrochemically active ferrocene labels into close proximity to the electrode surface.
The ferrous ion in each ferrocene group undergoes cyclic oxidation and reduction, leading to loss or gain of an electron, which is measured as current at the electrode surface using alternating-current voltammetry (ACV). Higher-order harmonic signal analysis also facilitates discrimination of ferrocene-dependent faradaic current from background capacitive current.13 Using this approach, the detection sensitivity of the eSensor technology has been measured at approximately 10 pM. The current sensitivity is sufficient for genotyping applications when combined with PCR amplification. Other approaches have demonstrated an increased sensitivity of electrochemical detection.9,14
The combination of electrochemical detection with ACV, harmonic signal analysis, and self-assembled monolayers is highly resistant to interference from sample constituents. Blood constituents that would normally interfere with fluorescence detection (e.g., hemoglobin, bilirubin) have no effect on signal detection using the eSensor technology. In addition, oxidation of the ferrocene labels and transmission of current through the monolayer depend on proximity of the label to the monolayer surface.
 [6] |
|
Figure 3. (click to enlarge [6]) Genotyping Assay Principle. Capture probes covalently bound to
the electrode hybridize equally to target DNA with both wild-type (WT) and mutant (MUT) sequences. Signal probes complementary to the WT and MUT target sequences are present in the hybridization buffer, and contain ferrocene labels with different redox potentials. The signal probes compete for binding to the region of the target DNA containing the mutation site, and binding of the perfect-match signal probe (i.e., WT signal probe to WT target) predominates. The genotype is determined by measuring the ratio of electrochemical signals from the WT and MUT signal probes. |
As a result, an unbound signal probe is not detected, and washing steps are not required to remove unbound reagents prior to the ACV measurement, even when a large number of signal probes representing multiple target sequences are present. The assay process is simplified, which allows hybridization and detection to be done in a small-footprint instrument without fluid handling or waste containers. In contrast, conventional microarrays require robotic instrumentation to automate multistage fluidic handling processes. Such instruments are bulky, complicated, and expensive, and limit microarrays to high-cost applications.
Genotyping Assay Principle
Genotyping of mutations or polymorphisms uses allele-specific signal probes containing ferrocene labels with distinguishable redox potentials. The signal probe matching the wild-type sequence contains a ferrocene label of one electrochemical potential, and a second signal probe matching the mutant sequence contains a second, distinguishable ferrocene label (see Figure 3 [6]). Both the wild-type and mutant targets bind to the capture probe at a site adjacent to the mutation. The wild-type and mutant signal probes compete for binding to their complementary sequences.
 [7] |
|
Figure 4. (click to enlarge) Figure 4. Genotyping scatter plot. Samples of known genotypes were analyzed by the eSensor method, and the magnitudes of the wild-type and mutant label signals determined. The vertical lines represent the established genotyping boundaries, and the horizontal line is the signal threshold. Data points between the boundary pairs are considered as indeterminate genotype, and signal values below the threshold are not evaluated for signal ratio.
|
The probe with the perfect match to the target is bound with a high degree of preference. The genotype is determined by the ratio of signals generated by the bound wild-type and mutant signal probes (see Figure 4 [8]). Genotyping boundaries are established based on statistical analysis of data from a large number of samples. Subsequent identification of unknown samples requires no further calibration of the instrument or cartridge lot. This approach can discriminate single- or multiple-base changes, insertions, and deletions. A mutation site with multiple alleles, or two adjacent mutation sites, can be genotyped using additional ferrocene labels.
System Overview
The eSensor XT-8 cartridge device consists of a PCB chip, a cover, and a microfluidic component (see Figure 5 [9]). The PCB chip contains 72 gold-plated working electrodes (as compared with 36 electrodes for the cystic fibrosis carrier detection chip), a silver/silver chloride reference electrode, and two gold-plated auxiliary electrodes. Each working electrode has a connector contact pad on the opposite side of the chip for electrical connection to the eSensor XT-8 instrument (see Figure 5d [9]).
 [9] |
|
Figure 5. (click to enlarge [9]) (a) Photograph of an eSensor cartridge. (b) Exploded schematic view of the cartridge components. Photographs of the PCB chip that measures 56 x 39 mm: (c) front (electrode pad) view; (d) back (connector pad) view.
|
The cartridge also contains an electrically erasable programmable read-only memory component, a memory device that stores information related to the cartridge (e.g., assay identifier, cartridge lot number, and expiration date). The microfluidic component is composed of a plate and a multilayer laminate, which includes a diaphragm pump and check valves in line with a serpentine channel that forms the hybridization chamber above the array of electrodes.
The PCB chip is prepared for an eSensor assay by depositing DNA capture probes and insulator molecules on the working electrodes. Each specific deposition solution is dispensed on the appropriate electrode using a robotic pipetting system. The capture probe and insulator react with the gold surface to form an insulating self-assembled monolayer.15 After capture-probe dispensing, the PCB chips are washed, dried, and assembled with the laminate, plate, and plastic cover into a cartridge to form a microfluidic circulating system that can hold approximately 140 µl.
 [10] |
|
Figure 6. (click to enlarge [10]) Schematic diagram showing the cross section of the cartridge.
|
The diaphragm pump in the cartridge is connected to a pneumatic source from the eSensor XT-8 instrument and provides unidirectional pumping of the hybridization mixture through the microfluidic channel during hybridization (see Figure 6 [10]). Using microfluidic technology to circulate the hybridization solution minimizes the unstirred boundary layer at the electrode surface and continuously replenishes the volume above the electrode that has been depleted of complementary targets and signal probes.16 This process reduces hybridization time from two hours for the cystic fibrosis carrier detection test to 30 minutes or less.
 |
|
Figure 7. A three-tower eSensor XT-8 Instrument with 24 independently operated cartridge slots.
|
The eSensor XT-8 instrument has a modular design consisting of a base module and one, two, or three cartridge-processing towers containing 8, 16, or 24 cartridge slots, respectively (see Figures 1 and 7). The number of towers can vary depending on the throughput needed, and the instrument can be upgraded to add towers as a lab's throughput needs increase. Assuming it takes 30 minutes for hybridization and detection, the throughput of the three-tower system can reach 300 tests per eight-hour shift.
The system's small footprint (46 × 40 × 41 cm for the one-tower system, 46 × 65 × 41 cm for the three-tower system) conserves bench space in a typical clinical laboratory. The cartridge slots operate independently of each other. Any number of cartridges can be loaded at one time, and the remaining slots are available for use while the instrument is running. When a single test is completed, that slot becomes available for another cartridge.
The base module controls each processing tower, provides power, and stores and analyzes data. The base module includes the user interface, and a 15-in. portrait-orientation display and touch panel. The instrument is designed to be operated solely with the touch screen interface. Entering patient accession numbers and reagent lot codes can be performed by the bar code scanner, the touch screen, or uploading a text file from a USB memory stick.
The processing tower consists of eight cartridge modules, each containing a cartridge connector, a precision-controlled heater, an air pump, and electronics. The cartridge connector has been designed to achieve high reliability and long life. The air pumps drive the diaphragm pump and valve system in the cartridge, eliminating fluid contact between the instrument and the cartridge, and minimizing the cartridge's complexity and cost.
Clinical Application
Warfarin is the most commonly prescribed anticoagulant in the United States.17 However, it exhibits a narrow therapeutic range, a wide interindividual variation in dosage required to reach optimal therapeutic effect, and severe adverse effects from overdosage, primarily due to excessive bleeding.18 Osmetech has developed the eSensor Warfarin Sensitivity Test, the first application for the eSensor XT-8 system. The test genotypes three polymorphisms in two genes that correlate with warfarin dose and allow individualization of therapy based on genotype.19,20
- The more active S-enantiomer of warfarin is metabolized to inactive forms by the liver enzyme cytochrome P450 (CYP450) 2C9. Two polymorphisms in the gene for CYP450 2C9 (*2, *3) reduce enzyme activity and are correlated with reduced warfarin dosage and an increased incidence of adverse side effects.
- The vitamin K epoxide reductase (VKOR) enzyme participates in the pathway of reactions leading to activation of clotting factors and is the target of warfarin action. A warfarin-sensitive haplotype has been identified, and the promoter polymorphism –1639G>A can identify patients with the warfarin-sensitive genotype (–1639AA).21
The warfarin sensitivity assay kit consists of a PCR buffer containing primers and dNTPs, thermostable DNA polymerase, exonuclease, genotyping reagent containing signal probes, cartridges, and ancillary buffer ingredients. The test is initiated by adding the PCR master mix to the genomic DNA sample. This mixture is subjected to thermal cycling in a standard instrument. After PCR, the entire product is treated briefly with bacteriophage l exonuclease that specifically recognizes and digests the nontarget amplicon strand.
The entire product of PCR and exonuclease digestion is used for genotyping with no further purification required. Genotyping is performed by adding the reagents containing the allele-specific, ferrocene-labeled signal probes to the digested PCR product, then placing the mixture into the sample reservoir of the cartridge. The entire Warfarin Sensitivity Test procedure can be completed in four hours, starting with isolated DNA. Assays can be performed on the eSensor XT-8 instrument in random-access mode, which is useful for this application.
Assay Performance
The performance of the Warfarin Sensitivity Test on the eSensor XT-8 instrument has been evaluated for accuracy, reproducibility, assay range, and the effects of interfering substances. Testing 40 replicates of a genomic DNA sample gave 100% call rates at 10, 100, and 1000 ng DNA/PCR, indicating a broad tolerance for input sample. This range is sufficient to span the lower and upper limits of expected DNA yield for most commercial whole-blood gDNA isolation kits.
Accuracy was evaluated by comparing genotyping of 101 genomic DNA samples by the eSensor method and bidirectional DNA sequencing. Agreement was 100%, with 97% of the samples giving results after the first round of testing, and a 100% calling rate after retesting of no-call results. Reproducibility was evaluated by testing a panel of 20 genomic DNA samples and three plasmid controls representing all possible panel genotypes. The panel was tested on five separate days with three lots of Warfarin Sensitivity Test kits. This protocol was performed using three instruments for a total of 345 tests. Agreement of the eSensor results with DNA sequencing was 100%, with 100% of the samples giving results after the first round of testing.
The effects of potential interfering substances on the Warfarin test were evaluated by adding to whole blood the following: human serum albumin (3 g/dl), human IgG (3 g/dl), bilirubin (0.3 mg/ml), triglycerides (500 mg/dl), hemoglobin (20 g/dl), and ethylene diamine tetraacetic acid (a concentration four times higher than what is used to prevent coagulation in order to simulate a low-volume blood draw). Such experiments were followed by extraction of genomic DNA and genotyping. None of these substances affected PCR efficiency, genotyping call rate, or accuracy.
Product Menu
Product development efforts are currently focused on assays for pharmacogenetics and genetic diseases. Osmetech is developing an extended eSensor Warfarin Sensitivity Test panel that determines the genotypes of eight single nucleotide polymorphisms (SNPs) in the CYP450 2C9 gene. These genotypes decrease enzyme activity, as well as the VKOR -1639G>A and a polymorphism in the CYP450 4F2 gene (rs2108622) 1347G>A.19 The genotype of the CYP450 4F2 SNP accounts for approximately 15% of the interindividual variability in warfarin dose.22
Pending FDA clearance, this test can identify patients at increased risk of adverse effects of warfarin and aid in determining the optimal initial warfarin dosage based on published algorithms. The 2C9 SNPs can also identify patients at risk of overmedication with other drugs metabolized by the CYP450 2C9 enzyme, such as phenytoin (an antiepileptic), glipizide and tolbutamide (antidiabetic agents), celecoxib (a nonsteroidal antiinflammatory agent), and losartan (an angiotensin I receptor antagonist).19
Other tests at varying development stages include genotyping tests for additional cytochrome P450 markers related to general drug metabolism, tests for infectious diseases, and combinations of markers related to the metabolism and efficacy of specific drug therapies.23
The recent efforts by FDA to encourage identification and validation of new pharmacogenetic biomarkers are leading to new diagnostic opportunities in genetics, cancer, infectious diseases, and pharmacogenetics.24 Such new markers can be adapted to a platform suitable for the clinical laboratory and submitted for FDA clearance, thus providing easy-to-use, cost-effective tests that can be run in hospitals of all sizes.
Conclusion
Electrochemical detection will allow further miniaturization of electronic components and integration of upstream sample extraction and amplification processes. Previous studies have demonstrated the feasibility of a single-use, sample-to-answer eSensor device, and recent developments in microfluidics provide additional tools to perform the required functions.25,26 Developing a point-of-care microarray system using electrochemical detection of nucleic acids will meet critical healthcare needs, including rapid genotyping to support accurate warfarin dosage in the cardiac care unit and detection of infectious diseases in the hospital and near-patient settings.
 |
 |
 |
 |
|
Robin H. Liu, PhD, is director of device technology at Osmetech Molecular Diagnostics (Pasadena, CA). He can be reached at robin.liu |
William A. Coty, PhD, |
Michael Reed, PhD, |
Gary Gust, PhD, is vice president of system development. He can be reached at gary.gust
@osmetech.com [14]. |
References
1. M Huber et al., “Automated, Enzyme-Free Genotyping Directly from Human DNA,” IVD Technology 13, no. 6 (2007): 37–45.
2. R Vairavan, “An Automated Multiplexing Platform for Genomic and Proteomic Analyses,” IVD Technology 14, no. 1 (2008): 29–35.
3. KL Gunderson, S Kruglyak, and MS Graige, “Decoding Randomly Ordered DNA Arrays,” Genome Research 14 (2004): 870–877.
4. J de Leon, “AmpliChip CYP450 Test: Personalized Medicine Has Arrived in Psychiatry,” Expert Review of Molecular Diagnostics 6, no. 3 (2006): 277–286.
5. M Schena, Microarray Biochip Technology (Natick, MA: Eaton Publishing, 2000).
6. P Fortina et al., “Simple Two-Color Array-Based Approach for Mutation Detection,” European Journal of Human Genetics 8 (2000): 884–894.
7. J Wang, “From DNA Biosensors to Gene Chips,” Nucleic Acids Research 28 (2000): 3011–3016.
8. ND Popovich, “Mediated Electrochemical Detection of Nucleic Acids for Drug Discovery and Clinical Diagnostics,” IVD Technology 7, no. 3 (2001): 36–42.
9. KM Roth et al., “Electrochemical Detection of Short DNA Oligomer Hybridization Using the CombiMatrix ElectraSense Microarray Reader,” Electroanalysis 18 (2006): 1982–1988.
10. K Kerman, M Kobayashi, and E Tamiya, “Recent Trends in Electrochemical DNA Biosensor Technology,” Measurement Science & Technology 15 (2004): R1–R11.
11. M Reed and W Coty, “eSensor: A Microarray Technology Based on Electrochemical Detection of Nucleic Acids and Its Application to Cystic Fibrosis Carrier Screening,” in Microarrays: New Development Towards Recognition of Nucleic Acid and Protein Signatures, ed. K Dill, R Liu, and P Grodzinski (New York: Springer, 2008).
12. MS Watson et al., “Cystic Fibrosis Population Carrier Screening: 2004 Revision of American College of Medical Genetics Mutation Panel,” Genetics in Medicine 6 (2004): 387–391.
13. TM Nahir and EFJ Bowden, “The Distribution of Standard Rate Constants for Electron Transfer between Thiol-Modified Gold Electrodes and Adsorbed Cytochrome C,” Journal of Electroanalytical Chemistry 410 (1996): 9–13.
14. R Gasparac et al., “Ultrasensitive Electrocatalytic DNA Detection at Two- and Three-Dimensional Nanoelectrodes,” Journal of the American Chemical Society 126, no. 39 (2004): 12270–12271.
15. RM Umek et al., “Electronic Detection of Nucleic Acids: A Versatile Platform for Molecular Diagnostics,” Journal of Molecular Diagnostics 3 (2001): 74–84.
16. R Lenigk et al., “Plastic Biochannel Hybridization Devices: A New Concept for Microfluidic DNA Arrays,” Analytical Biochemistry 311 (2002): 40–49.
17. JD Horton and BM Bushwick, “Warfarin Therapy: Evolving Strategies in Anticoagulation,” American Family Physician 59 (1999): 635–646.
18. L Brunton, J Lazo, and K Parker, The Pharmacological Basis of Therapeutics (New York: McGraw Hill, 2006).
19. M. Ingelman-Sundberg et al., “Influence of Cytochrome P450 Polymorphisms on Drug Therapies: Pharmacogenetic, Pharmacoepigenetic, and Clinical Aspects,” Pharmacology & Therapeutics 116 (2007): 496–526.
20. M Wadelius et al., “Association of Warfarin Dose with Genes Involved in Its Action and Metabolism,” Human Genetics 121 (2007): 23–34.
21. EA Sconce et al., “The Impact of CYP2C9 and VKORC1 Genetic Polymorphism and Patient Characteristics upon Warfarin Dose Requirements: Proposal for a New Dosing Regimen,” Blood 106 (2005): 2329–2333.
22. MD Caldwell et al., “CYP4F2 Genetic Variant Alters Required Warfarin Dose,” Blood 111 (2008): 4106–4112.
23. SD Vernon et al., “Bioelectronic DNA Detection of Human Papillomaviruses Using eSensor: A Model System for Detection of Multiple Pathogens,” BMC Infectious Diseases 3 (2003): 12.
24. “Table of Valid Genomic Biomarkers in the Context of Approved Drug Labels,” Food and Drug Administration (Rockville, MD: 2008 [cited 13 March 2008]); available from Internet: www.fda.gov/cder/genomics/genomic_ biomarkers_table.htm [15].
25. RH Liu et al., “Self-Contained, Fully Integrated Biochip for Sample Preparation, Polymerase Chain Reaction Amplification, and DNA Microarray Detection,” Analytical Chemistry 76 (2004): 1824–1832.
26. RH Liu and AP Lee, Integrated Biochips for DNA Analysis (New York: Springer, 2008).