Integrated amplification microarray system in a lateral flow cell for warfarin genotyping from saliva

Integrated amplification microarray system in a lateral flow cell for warfarin genotyping from saliva

Clinica Chimica Acta 429 (2014) 198–205 Contents lists available at ScienceDirect Clinica Chimica Acta journal homepage: www.elsevier.com/locate/cli...

283KB Sizes 1 Downloads 14 Views

Clinica Chimica Acta 429 (2014) 198–205

Contents lists available at ScienceDirect

Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim

Integrated amplification microarray system in a lateral flow cell for warfarin genotyping from saliva Thomas Sebastian a,⁎, Christopher G. Cooney a, Jennifer Parker a, Peter Qu a, Alexander Perov a, Julia B. Golova a, Lindsay Pozza b, Rafal M. Iwasiow b, Rebecca Holmberg a a b

Akonni Biosystems, Inc., 400 Sagner Avenue, Suite 300, Frederick, MD 21701, United States DNA Genotek, Inc., 2 Beaverbrook Road, Kanata, Ontario K2K 1L1, Canada

a r t i c l e

i n f o

Article history: Received 25 October 2013 Received in revised form 30 November 2013 Accepted 9 December 2013 Available online 17 December 2013 Keywords: Warfarin Microarray SNP Lateral flow cell Saliva

a b s t r a c t Background: Genetic polymorphisms in the CYP2C9 and VKORC1 genes have been linked to sensitivity of the anticoagulant drug warfarin. The aim of this study is to demonstrate a method for warfarin sensitivity genotyping using gel element microarray technology in a simplified workflow from sample collection to analysis and detection. Methods: We developed an integrated amplification microarray system combining PCR amplification, target labeling, and microarray hybridization within a single, closed-amplicon “lateral flow cell” for genotyping three single nucleotide polymorphisms (SNPs) that influence warfarin response. We combined nucleic acid extraction of saliva using the TruTip technology together with the lateral flow cell assay and with fully automated array detection and analysis. Results: The analytical performance of the assay was tested using 20 genotyped human genomic DNA samples and found to be sensitive down to 330 input genomic copies (1 ng). A follow-up pre-clinical evaluation was performed with 65 blinded saliva samples and the genotyping results were in agreement with those determined by bidirectional sequencing. Conclusions: Combined with the use of non-invasive saliva samples, rapid DNA extraction, the lateral flow cell, automatic imaging and data analysis provides a simple and fast sample-to-answer microarray test for warfarin sensitivity genotyping. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Warfarin is widely used as an anti-coagulant for the treatment and prevention of arterial and venous thromboembolism [1,2]. However, the drug is associated with the greatest number of serious adverse drug events over the past two decades, due to its narrow therapeutic index and the substantial inter-individual variability in dosing requirements [3,4]. There is increasing evidence indicating that a number of factors affect warfarin dosing, including non-genetic and genetic factors [2,4–6]. Genetic factors affecting warfarin pharmacokinetics and pharmacodynamics account for roughly 40% of warfarin dosing variability and stem primarily from two genes: warfarin's target gene, VKORC1 [7,8] and the gene of its main metabolizing enzyme, CYP2C9 [9]. Identifying individuals with the polymorphisms that produce various responses in

Abbreviations: CYP2C9, cytochrome P-450 variant 2C9; VKORC1, vitamin K epoxide reductase complex-1; SNP, single-nucleotide polymorphism; PCR, polymerase chain reaction. ⁎ Corresponding author. Tel.: +1 301 698 0101; fax: +1 301 698 0202. E-mail address: [email protected] (T. Sebastian). 0009-8981/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cca.2013.12.009

warfarin therapy and the concomitant adjustment of their warfarin dose is expected to confer substantial benefit to patients and in cost savings to the healthcare system. Clinical studies have shown a correlation between steady-state warfarin dose and allelic variants of CYP2C9 and VKORC1 [4,10]. To encourage the use of pharmacogenetic testing for patients starting warfarin therapy, the US Food and Drug Administration (FDA) mandated the inclusion of a warning label on warfarin packaging regarding the relationship of safe and effective dosage to individual patients with mutations in these specific genetic regions [11,12]. The anticipated need for warfarin genotyping prompted the development of numerous laboratory-developed clinical assays and commercial platforms focused on CYP2C9 430C N T (CYP2C9*2), CYP2C9 1075A N C (CYP2C9*3) and VKORC1 -1639 (or 3673) G N A genotyping [13–15]. Comparative studies of the several warfarin genotyping commercial platforms reported that the assays were found to differ in turnaround time and hands-on time, requirements for amount of input genomic DNA, accuracy and cost, specialized equipment usage and other factors that might makes each assay more favorable in different settings [13,14,16,17]. A wider adoption of the warfarin assay in clinical laboratories requires the development of testing platforms that can

T. Sebastian et al. / Clinica Chimica Acta 429 (2014) 198–205

reduce the cost of reagents and consumables, expensive supporting instrumentation, and the amount of handling required. Recent publications reported on the development of a few warfarin genotyping platforms offering a lower cost solution, but require multiple processing steps and create the potential for amplicon cross contamination [18,19]. DNA microarrays are a powerful genomic tool that can be applied to biomedical and clinical research [20,21]; however, a simplified workflow is required for a wider adoption of the microarray platform in diagnostic applications. Typical protocols incorporating microarray detection generally involve several steps, including nucleic acid purification, PCR amplification, amplified target purification and/or modification (e.g. fragmentation, labeling and/or single-stranded target generation), target hybridization, array washing, drying, imaging and complex manual data analysis. Chandler et al. reviewed the recent methods for condensing this workflow by combining target amplification, labeling, and microarray hybridization into a single, closedamplicon reaction chamber, and provided examples of this approach in infectious disease diagnostics [22]. A key element of this approach is the three dimensional gel element microarrays which enable an increase in hybridization efficiency and detection sensitivity compared to planar arrays [23,24] and is well suited to a lower cost imaging system. We recently reported on a valve-less lateral flow cell with a simple consumable architecture that supports target amplification and microarray hybridization in the same chamber [25]. This lateral flow cell also incorporates post hybridization wash steps while retaining an entirely closed-amplicon system, thus minimizing the potential for sample or amplicon cross-contamination. These advancements simplify the microarray workflow without custom instrumentation and increase its applicability for diagnostic applications. The objectives of this study were therefore to demonstrate the application of the integrated amplification microarray in a lateral flow cell for genotyping the *2 or *3 alleles of CYP2C9 and the VKORC1 allele to determine warfarin sensitivity, with an emphasis on clinical ease-of-use and reduced complexity. In order to develop a simplified workflow for routine warfarin genotyping, we combined the target amplification, labeling, and microarray hybridization within a single, closed-amplicon microfluidic chamber with fully automated array imaging and analysis. The assay performance of this system was optimized and evaluated using a series of genomic DNA (gDNA) samples. The final workflow incorporated saliva samples extracted using the rapid DNA extraction capability of the TruTip technology for an end-to-end solution. This approach was finally applied to the genotyping of a set of 65 blinded saliva samples for pre-clinical evaluation.

199

2. Materials & methods 2.1. Source DNA 2.1.1. Purified gDNA samples A set of 20 genotyped human genomic DNA samples (NA17075, NA17119, NA17204, NA17207, NA17210, NA17214, NA17215, NA17216, NA17219, NA17220, NA17221, NA17222, NA17229, NA17243, NA17247, NA17252, NA17254, NA17259, NA17285, NA17290) were obtained from the Coriell Institute Cell Repository (Camden, NJ). Each purified DNA sample is either wild type (WT), mutant (MUT), or heterozygous (HET) for each SNP target, CYP2C9*2, CYP2C9*3 and VKORC1. DNA concentration was estimated on a NanoDrop 1000 (Thermo Scientific, Wilmington, DE) before use. 2.2. Saliva samples Sixty-five de-identified saliva samples collected and preserved using the Oragene®•Dx self-collection kit were obtained from DNA Genotek, Inc. (Ontario, Canada). The warfarin genotype information was blinded in these samples. DNA was extracted and purified from the saliva samples using Akonni's TruTip gDNA Saliva Kit (Cat No. 300-20431) according to manufacturer's recommendations. Purified DNA was eluted from the TruTip matrix with 100 μl of elution buffer. The quantity and quality of the DNA were determined using a NanoDrop 1000 and quantification was verified by qPCR using the Quantifiler® Human DNA Quantification Kit (Life Technologies, Carlsbad, CA). DNA was stored at −20 °C until use. 2.3. Microarray design and manufacture Microarray probes were designed for each SNP target, two hybridization probes (one probe per allele) (Table 1). Probe sequences with artificial mismatch positioned relative to the SNP target were included for enhanced discrimination of single nucleotide polymorphisms by DNA hybridization [26]. All probes for microarray manufacturing were synthesized by Akonni with a custom 3′-methacryamido-linker and purified to N90% purity by HPLC. Control probes include Cy3-labeled 8-mer oligonucleotides for positional reference beacons of the array and manufacturing quality control and a plant sequence with no known homology to any human DNA sequence as a negative control probe. These probes provided an internal control of reagents, array integrity, and non-specific hybridization. Microarrays were manufactured

Table 1 DNA sequence of all PCR primers and microarray probes used in the assay. Primer ID

Sequence (5′–3′ direction)

Tm (°C)

Amplicon size, nt

CYP2C9*2-F CYP2C9*2-R CYP2C9*3-F CYP2C9*3-R VKORC1-F VKORC1-R

[Cy3] ATGGAAGGAGATCCGGCGTTT AGGTCAGTGATATGGAGTAGGG CCAGGAAGAGATTGAACGTGTG [Cy3} ACTTACCTTGGGAATGAGATA TGG GAA GTC AAG CAA GAG AAG ACC [Cy3]TGCTAGGATTATAGGCGTGAGCCA

62.6 62.7 62.7 56.7 64.6 64.6

155

Probe ID

Sequence (5′–3′ direction)

Tm (°C)

Probe Length

CYP2C9*2-3L CYP2C9*2-4L CYP2C9*3-5L CYP2C9*3-6L VKORC1-7S_A VKORC1-8S_A Negative control probe

TCTTGAACACGGTCCTCAATG TCTTGAACACAGTCCTCAATGC TCCAGAGATACATTGACCTTCTC TCCAGAGATACCTTGACCTTCT TTGaCCGGGTGCG ATTGaCCAGGTGCG TCT TCT TCC TCC TCC TCG TC

59.5 60.1 60.9 60.1 44 43.7 60.5

187 75

21 22 23 22 13 14 20

Bold letters in the probe sequences indicate SNP and lowercase letters in the probe sequences indicate intentional mismatches versus the genomic template. The melting temperature (Tm) was calculated with the Oligo Calc.

T. Sebastian et al. / Clinica Chimica Acta 429 (2014) 198–205

The valve-less integrated amplification microarray utilizes a singlestep protocol consisting of solution-phase PCR and hybridization in the same buffer system in the lateral flow cell. Construction and materials for the lateral flow cell were described previously [25]. Briefly, the components consisted of a double-sided adhesive spacer tape with cutouts for the inlet channel, reaction chamber, connecting channel, and waste chamber. A top film covers the array chamber and the waste chamber. The top film has an inlet hole and there is a small vent hole in the waste chamber. Assays were performed in triplicate with Coriell purified genomic DNA and saliva samples using 20–25 ng DNA. The Limit of Detection study used serial dilutions ranging from 1000 ng to 0.01 ng of purified gDNA. A negative PCR control containing no genomic DNA was included with each batch of samples to test for possible cross contamination. Primers listed in Table 1 were synthesized by Eurofins MWG Operon (Huntsville, AL) with HPLC purification. For asymmetric PCR, forward and reverse primers were used in ratios ranging from 1:10 to 1:50 (unlabeled to Cy3 labeled) with a final concentration of 0.025 to 1 μM of each primer. The asymmetric amplification master mix included 2 μl of template DNA in a 30 μl total reaction volume consisting of 15 μl 2 × platinum multiplex PCR mix (Life Technologies), 5 μl asymmetric PCR primer mix, non-acetylated BSA (1 μg/μl, Life Technologies), ET SSB (4 ng/μl, BioHelix, Beverly, MA), 5% formamide (Fisher Scientific, Pittsburgh, PA) and nucleasefree water. The reaction mix was introduced to the array chamber via pipette through the sample inlet port and the port was covered with a foil adhesive. The lateral flow cells were placed on the flat- block of the Quanta thermal cycler (Quanta Biotech, Surrey, UK) and the samples were amplified and hybridized using the following thermocycling protocol: 93 °C for 2 min, followed by 40 cycles of 90 °C for 30 s, 57.5 °C for 90 s, 68 °C for 60 s, and a final step of 50 °C for 60 min. Resulting Cy3-labeled, asymmetric amplicons ranged from 75 to 187 nt in length. After the amplification and hybridization, the array was washed by piercing the inlet port with a pipettor tip and sequentially dispensing 50 μl each of 1 × SSPE containing 0.01% Triton X-100, Milli-Q water, and acetone through the chamber. The lateral flow cell was inverted for 5 min at room temperature to allow the acetone to drain and the array to dry completely prior to imaging. 2.5. Imaging and data analysis The arrays were imaged using Akonni's prototype microarray analyzer (TruDx 2100), consisting of a high-intensity green light emitting diode (LED), custom optics, and CCD camera. During imaging, the lateral flow cells remained intact and arrays were imaged through the glass side of the chamber. The automated microarray analysis (AMA) software, integrated with the imager, locates and segments each spot and reports signal-to-noise ratios (SNR). A fixed circle algorithm was used for data extraction of probe signal from background pixels, and local background was subtracted from each gel element. Summary statistics and quality scores for every gel element on the array were automatically generated and imported to a Microsoft Excel workbook. Integrated background-corrected signal intensities were averaged across quadruplicate probes. SNRs for each probe were calculated from the median signal from quadruplicate gel elements, and normalized to three times the average standard deviation of the local background. The image and analysis performance were verified with a commercial imager, Sensovation FLAIR (Radolfzell, Germany) with a similar semi-automatic image analysis system for microarray image quantification for low density microarrays. Data is presented herein

A) 0.75 0.5 0.25

SNP Ratio

2.4. Amplification microarray in lateral flow cell

for the Akonni imager to support the full sample-to-answer system components. The genotype decision rule was constructed as follows: each SNP target has 2 probes immobilized on the array, one probe representing the wild-type (WT) sequence and one probe representing the mutant (MUT) sequence. A SNP target was declared “Detected” if the SNR of either WT or MUT probe was ≥ 3. A test was reported as “Invalid” if one or more of the three target variants resulted in a SNR b 3 (i.e. Not Detected) for both wild type and mutant probes, because every target should detect signal for at least one of the probes in the presence of

0 -0.25 -0.5 -0.75

B)

HET

MUT CYP2C9*2

HET

MUT

WT

All Pairs Tukey-Kramer 0.05

1 0.75 0.5

SNP Ratio

essentially as described by Golova et al. [27]. Each probe was printed in quadruplicate per array and Cy3-labeled 8-mer oligonucleotides served as positional reference beacons. Arrays were stored at room temperature and protected from light until used.

0.25 0 -0.25 -0.5 WT

CYP2C9*3

C)

All Pairs Tukey-Kramer 0.05

1

0.5

SNP Ratio

200

0

-0.5

-1

HET

MUT VKORC1

WT

All Pairs Tukey-Kramer 0.05

Fig. 1. Genotype determination of 20 gDNA samples by amplification microarray in the lateral flow cell. Shown is the SNP Ratio of different genotypes for CYP2C9*2 (A), CYP2C9*3 (B) and VKORC1 (C) with Tukey–Kramer statistical analysis depicted using circles on the right of each figure. The size of the circle is inversely proportional to the number of samples in the group. Confidence intervals of 95% are represented by the diamond; the middle bar corresponds to the mean values, with standard deviations above and below. Assay was performed in triplicate and all 60 replicates were plotted. NTCs were negative for all probes and so excluded from the analysis.

T. Sebastian et al. / Clinica Chimica Acta 429 (2014) 198–205

a human genomic DNA sample. For detected samples, the SNP ratio of different genotypes were calculated from SNR values of WT and MUT probes with the formula, (WT − MUT) / (WT + MUT) described earlier [28]. This formula determined whether the samples were WT, heterozygous (HET) or homozygous mutant (MUT) for each target, yielding a value between +1 and −1 (inclusive). For the blinded saliva sample study, samples were processed in triplicate. Invalid and Indeterminate samples were repeated and the repeated valid results were used in the average SNP ratio calculations. Replicate results were combined to determine the genotype call and final results were compared to bidirectional sequencing results obtained from DNA Genotek. SNP ratio data was analyzed and presented quantitatively as mean ± standard error of the mean (SEM). Differences between genotype calls for each SNP target data set were evaluated by one-way analysis of variance (ANOVA) and Tukey–Kramer statistical analysis (p-value). JMP (SAS, Cary, NC) and Excel (Microsoft, Redmond, WA) software were used for SNP Ratio calculations and statistical analysis. 3. Results 3.1. Assay development in lateral flow cell We have designed a multiplexed amplification assay in a valve-less lateral flow cell allowing for the detection and genotyping of 3 SNPs for warfarin sensitivity in a single sample using a low density microarray. We tested twenty gDNA samples that cover all possible genotypes in all three SNP targets, CYP2C9*2, CYP2C9*3 and VKORC1. Tests were performed in triplicate for each sample to determine the assay performance in the lateral flow cell before evaluating the test on blinded saliva samples. Average SNP allele ratio results for each separate target, CYP2C9*2, CYP2C9*3 and VKORC1, are shown in Figs. 1A–C, respectively. All wild type samples determined by the SNP ratios fell in a cluster on the intensity plot distributed mostly towards the +1 axis due to the predominant signal from the WT probe. Conversely, the mutant samples were in a similar cluster distributed mostly towards the −1 axis on the intensity plot due to the predominant signal from the MUT probe. The heterozygous samples with relatively equal probe signals fell between the centers of the intensity plot indicating an equal signal from both the WT and MUT probes and resulted in a call of HET. ANOVA analysis resulted in R squared (R2) values of 0.983, 0.954 and 0.955 for SNP targets CYP2C9*2, CYP2C9*3 and VKORC1 respectively. A Tukey–Kramer analysis of these distinct data sets shown as circles on the right side of the Fig. 1 plots verifies that each call type within each SNP target is statistically significant from the other two calls with p values b 0.0001.

201

These results indicate that specificity is maintained without interference when combining PCR and hybridization in a single reaction. The resulting calls from these clustered samples were used in combination with data obtained using separate amplification and hybridization (data not shown) to determine the SNP ratio ranges for each target: CYP2C9*2 (WT: 1.00 to 0.20, HET: 0.00 to − 0.20,MUT: -0.40 to −1.00); CYP2C9*3 (WT: 1.00 to 0.45, HET: 0.35 to 0.00, MUT: -0.20 to − 1.00); and VKORC1 (WT: 1.00 to 0.30, HET: 0.10 to − 0.10, MUT: -0.30 to − 1.00). Any samples resulting in a SNP ratio in between these ranges were reported as “Indeterminate”. These data indicate that the combination of asymmetric multiplex PCR with the microarray hybridization method in a lateral flow cell results in 100% accurate genotype calls for the three targets when compared to the predetermined genotypes. 3.2. Analytical sensitivity A limit of detection (LoD) study was performed to assess the genotyping performance of the lateral flow cell across a six-log range for genomic DNA input amounts. Twelve genomic DNA input concentrations were evaluated, and each serial dilution was assayed three times independently. Assay performance was evaluated with respect to the following criteria: probe signal intensity and SNP genotyping ratio. The result of this study is summarized in Table 2 and established the minimum input DNA amount for the lateral flow cell to be 1 ng DNA (330 gene copies). The analytical sensitivity obtained in this assay is comparable to other commercial assays already in clinical use [13,14]. In addition, the same study demonstrated that upper levels of DNA input amounts, 750 ng and 1000 ng did not interfere with the assay fidelity. The flexible range of input DNA for the assay is concordant with the DNA yield expected from typical saliva samples [29], so that the assay can be run without intermediate determination of concentration and subsequent dilution (data not shown). All subsequent assays were run with 20–25 ng input DNA which is comfortably above the limit of detection and well below the amount purified from the panel of saliva samples. 3.3. Analysis of blinded saliva samples Pre-clinical evaluation of the assay was performed with the genotyping of 65 blinded saliva samples which had been collected and stabilized with Oragene•Dx. Genomic DNA was extracted from each Oragene•Dx/ saliva sample in 7 min using the TruTip isolation technology. Concentrations ranged from 10 to 165 ng/μl with an average of 59 ng/μl.

Table 2 Detection sensitivity of the warfarin assay. Sample

Experiment-1

ng DNA

Genotype CYP2C9

1000 750 500 200 100 50 25 10 5 1 0.1 0.01 NTC

Experiment-2 Assay VKORC1 3673 (−1639)

*2

*3

MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT HET No call No call

WT WT WT WT WT WT WT WT WT WT WT No call No call

HET HET HET HET HET HET HET HET HET HET MUT No call No call

Assay

CYP2C9

Pass Pass Pass Pass Pass Pass Pass Pass pass Pass Fail Fail Pass

Experiment-3

Genotype VKORC1 3673 (−1639)

*2

*3

Mut Mut Mut Mut Mut Mut Mut Mut Mut Mut No call No call No call

WT WT WT WT WT WT WT WT WT WT No call No call No call

HET HET HET HET HET HET HET HET HET HET WT No call No call

Genotype

Assay

CYP2C9

Pass Pass Pass Pass Pass Pass Pass Pass Pass Pass Fail Fail Pass

VKORC1 3673 (−1639)

*2

*3

MUT MUT MUT MUT MUT MUT MUT MUT MUT MUT HET No call No call

WT WT WT WT WT WT WT WT WT WT HET No call No call

HET HET HET HET HET HET HET HET HET HET MUT No call No call

Pass Pass Pass Pass Pass Pass Pass Pass Pass Pass Fail Fail Pass

A range of gDNA (NA 17075) input concentrations (1000 ng to 0.01 ng) were assayed three times independently with amplification microarray in lateral flow cell. The assay result was represented as ‘pass’ or ‘fail’ based on the genotype decision rule described in the Materials & methods.

202

Sample No

SNP Ratio & Genotype

Bi-directional sequencing

CYP2C9

AK01 AK02 AK03 AK04 AK05 AK06 AK07 AK08 AK09 AK10 AK11 AK12 AK13 AK14 AK15* AK16 AK17 AK18 AK19 AK20 AK21 AK22 AK23* AK24 AK25 AK26* AK27 AK28 AK29 AK30 AK31 AK32

VKORC1 3673 (−1639)

*2-SNP Ratio (Mean ± SEM)

Genotype

*3-SNP Ratio (Mean ± SEM) WT

Genotype

SNP Ratio (Mean ± SEM)

Genotype

−0.13 0.42 −0.48 −0.07 −0.07 0.58 0.54 0.55 −0.08 0.54 0.45 −0.13 0.45 0.60 0.57 0.51 0.52 0.48 0.48 0.44 −0.17 0.50 0.52 −0.06 0.50 −0.07 0.51 0.51 0.43 0.47 0.46 0.44

HET WT MUT HET HET WT WT WT HET WT WT HET WT WT WT WT WT WT WT WT HET WT WT HET WT HET WT WT WT WT WT WT

0.70 0.63 0.60 0.78 0.70 0.75 0.74 0.74 0.71 0.22 0.22 0.62 0.53 0.28 0.72 0.25 0.92 0.84 −0.40 0.73 0.81 0.26 0.76 0.71 0.74 0.23 0.86 0.74 0.58 0.64 0.81 0.25

WT

−0.02 −0.04 −0.04 0.80 −0.68 −0.65 −0.84 0.76 0.00 −0.88 0.03 −0.65 0.04 0.95 0.83 −0.88 0.04 −0.02 −0.00 0.91 −0.91 −0.02 0.00 0.80 0.83 0.01 0.89 0.91 −0.03 −0.87 0.93 0.94

HET HET HET WT MUT MUT MUT WT HET MUT HET MUT HET WT WT MUT HET HET HET WT MUT HET HET WT WT HET WT WT HET MUT WT WT

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.02 0.02 0.04 0.00 0.01 0.02 0.02 0.02 0.02 0.01 0.02 0.03 0.05 0.02 0.01 0.01 0.01 0.00 0.06 0.01 0.02 0.11 0.03 0.01 0.02 0.01 0.01 0.02 0.02 0.03 0.00 0.02

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.04 0.06 0.03 0.04 0.06 0.06 0.06 0.03 0.09 0.03 0.04 0.02 0.03 0.02 0.00 0.01 0.01 0.03 0.01 0.03 0.02 0.01 0.04 0.06 0.02 0.03 0.04 0.05 0.07 0.05 0.03 0.01

WT WT WT WT WT WT WT HET HET WT WT HET WT HET WT WT MUT WT WT HET WT WT WT HET WT WT WT WT WT HET

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.01 0.03 0.02 .07 0.08 0.11 0.09 0.06 0.01 0.02 0.04 0.12 0.01 0.00 0.05 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.08 0.07 0.02 0.06 0.01 0.03 0.02 0.01 0.00

CYP2C9 (*2 and *3)

VKORC1 3673 (−1639)

*1/*2 *1/*1 *2/*2 *1/*2 *1/*2 *1/*1 *1/*1 *1/*1 *1/*2 *1/*3 *1/*3 *1/*2 *1/*1 *1/*3 *1/*1 *1/*3 *1/*1 *1/*1 *3/*3 *1/*1 *1/*2 *1/*3 *1/*1 *1/*2 *1/*1 *2/*3 *1/*1 *1/*1 *1/*1 *1/*1 *1/*1 *1/*3

G/A G/A G/A G/G A/A A/A A/A G/G G/A A/A G/A A/A G/A G/G G/G A/A G/A G/A G/A G/G A/A G/A G/A G/G G/G G/A G/G G/G G/A A/A G/G G/G

Genotype Agreement

Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

T. Sebastian et al. / Clinica Chimica Acta 429 (2014) 198–205

Table 3 Genotype determination of the 65 clinical samples.

0.43 0.43 0.39 0.63 0.50 −0.13 −0.15 0.45 0.53 0.51 0.51 0.54 0.49 0.43 0.48 0.44 −0.11 −0.11 −0.11 0.53 0.53 0.49 0.49 0.57 0.50 −0.11 0.48 0.49 0.40 0.37 0.41 0.36 −0.50

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.01 0.02 0.00 0.01 0.03 0.01 0.01 0.04 0.04 0.02 0.03 0.02 0.01 0.11 0.07 0.01 0.01 0.02 0.02 0.02 0.02 0.01 0.02 0.00 0.02 0.01 0.01 0.02 0.01 0.01 0.00 0.01 0.01

WT WT WT WT WT HET HET WT WT WT WT WT WT WT WT WT HET HET HET WT WT WT WT WT WT HET WT WT WT WT WT WT MUT

0.69 0.80 0.18 0.79 0.21 0.75 0.69 0.75 0.72 0.72 0.84 0.26 0.65 0.70 0.67 0.64 0.66 0.58 0.81 0.73 0.61 0.66 0.67 0.75 0.76 0.76 0.70 0.62 0.69 0.77 0.75 0.73 0.71

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.01 0.03 0.06 0.09 0.05 0.04 0.05 0.06 0.08 0.06 0.01 0.04 0.06 0.04 0.04 0.02 0.07 0.01 0.04 0.05 0.05 0.07 0.07 0.02 0.04 0.04 0.05 0.04 0.02 0.06 0.05 0.03 0.06

WT WT HET WT HET WT WT WT WT WT WT HET WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT

0.00 0.01 −0.90 0.05 −0.03 0.01 0.01 0.90 0.76 −0.00 −0.00 0.76 −0.04 0.73 −0.85 0.68 0.70 −0.68 0.02 0.03 −0.86 0.82 0.64 −0.65 0.80 0.90 0.87 0.03 −0.86 0.00 −0.79 0.64 0.78

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.02 0.02 0.03 0.01 0.05 0.03 0.03 0.03 0.10 0.01 0.01 0.18 0.02 0.09 0.03 0.09 0.10 0.13 0.02 0.01 0.05 0.07 0.10 0.14 0.07 0.04 0.07 0.04 0.02 0.02 0.09 0.04 0.04

HET HET MUT HET HET HET HET WT WT HET HET WT HET WT MUT WT WT MUT HET HET MUT WT WT MUT WT WT WT HET MUT HET MUT WT WT

*1/*1 *1/*1 *1/*3 *1/*1 *1/*3 *1/*2 *1/*2 *1/*1 *1/*1 *1/*1 *1/*1 *1/*3 *1/*1 *1/*1 *1/*1 *1/*1 *1/*2 *1/*2 *1/*2 *1/*1 *1/*1 *1/*1 *1/*1 *1/*1 *1/*1 *1/*2 *1/*1 *1/*1 *1/*1 *1/*1 *1/*1 *1/*1 *2/*2

G/A G/A A/A G/A G/A G/A G/A G/G G/G G/A G/A G/G G/A G/G A/A G/G G/G A/A G/A G/A A/A G/G G/G A/A G/G G/G G/G G/A A/A G/A A/A G/G G/G

Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

Average SNP ratio results from saliva samples processed by TruTip extraction with lateral flow cell amplification and detection as compared to bi-directional sequencing method. N = 3 per saliva sample. *Values are the mean of n = 2 replicates only if the third replicate was discordant. The genotype nomenclature used for wild-type and two variants of CYP2C9 allele was as follows: CYP2C9*1/*1 = WT; CYP2C9*1/*2 = HET for *2 and WT for *3; CYP2C9*1/*3 = WT for *2 and HET for *3; CYP2C9*2/*2 = MUT for *2 and WT for *3; CYP2C9*2/*3 = HET for *2 and *3; CYP2C9*3/*3 = WT for *2 and MUT for *3. For VKORC1, G/G = WT; G/A = HET, and A/A = MUT.

T. Sebastian et al. / Clinica Chimica Acta 429 (2014) 198–205

AK33* AK34 AK35 AK36 AK37 AK38 AK39 AK40 AK41 AK42 AK43 AK44 AK45 AK46 AK47 AK48 AK49 AK50 AK51 AK52 AK53 AK54 AK55 AK56 AK57 AK58 AK59 AK60 AK61 AK62 AK63 AK64 AK65

203

204

T. Sebastian et al. / Clinica Chimica Acta 429 (2014) 198–205

The average 260/280 nm absorbance ratio was 1.78 and average 260/230 nm absorbance ratio was 1.62 indicating quality genomic DNA. Isolated DNA was diluted to 10 ng/μl and used 2 μl (20 ng input DNA) per assay in the lateral flow cell. The resulting SNP genotypes were obtained from the microarray image analysis output in a total processing time of 4 h per sample. Using the decision logic detailed in the Materials & methods section for determining the genotype for each SNP target and previously determined ratio ranges, the result was determined for each of the 65 blinded samples. Table 3 lists the average SNP ratio, standard error, genotype call, comparison call determined by bidirectional sequencing and correlation between genotyping methods for the blinded sample set. Analysis of the SNP ratio for the CYP2C9*2 assay identified 49 samples as WT, 2 as MUT and 14 as HET for the CYP2C9*2 SNP. Review of the SNP ratio of the CYP2C9*3 assay determined 54 samples as WT, 10 as HET, and 1 as MUT. Assay results for the VKORC1 SNP showed that 24 samples were wild type, 15 as MUT, and 26 as HET. All genotyping calls were concordant with those determined using bidirectional sequencing. The p values resulting from Tukey–Kramer analysis for all comparisons indicated statistical significance of each genotype data set within each SNP target (p b 0.0001). In routine diagnostic use, a single test should be employed per patient. In this study the 65 clinical samples were processed in triplicate to evaluate repeatability of the assay, for a total of 195 lateral flow cells processed. Out of this total, four tests (2.05%) were invalid due to low signals in one or more of the SNP targets, resulting in an overall invalid test outcome, and two tests (1.03%) were considered to be indeterminant with a SNP ratio that fell in between the ranges set for each genotype call (data not shown). In all six cases, an additional replicate sample was processed with a valid outcome and resulted in the correct genotyping, suggesting that the problem is not related to the sample quality. Four additional assays (2.05%) resulted in SNR values above the set threshold yet the genotyping calls for at least one of the three SNP targets were discordant compared to the other two sample replicates and were therefore not included in the genotype determination. ANOVA analysis of all valid tests resulted in R2 values of 0.924, 0.673 and 0.966 for SNP targets CYP2C9*2, CYP2C9*3 and VKORC1 respectively. The lower R2 value for CYP2C9*3 is due to three incorrect genotypes in this set which may be explained by fluorescent artifacts that mask probe signals or amplification inhibition due to inconsistencies in the manual lateral flow cell assembly method. Further investigation into the source of the discordance is underway, with particular attention to the flow assembly process.

4. Discussion In this study, we have demonstrated an end-to-end platform for warfarin sensitivity genotyping from saliva samples using a novel extraction method and a single-step, closed-amplicon lateral flow cell for a simplified microarray detection approach. This system combines multiple steps into a simple, compact workflow without the need for complex instrumentation, and is less susceptible to contamination than conventional methods. The three dimensional gel element arrays offer increased signal relative to planar arrays, thus allowing for less sophisticated and less expensive optics. The prototype microarray imager used in this study represents a low-cost, robust imager such as this. However, a truly “easy-to-use” microarray diagnostic will further require fully automated image and data analysis which is also demonstrated herein with automatic grid placement, data extraction and analysis steps that do not require subjective input from the user. Saliva is a convenient, non-invasive, and cost-effective alternative to blood for DNA genotyping applications [29]. The fast and efficient isolation of genomic DNA from Oragene•Dx stabilized-saliva samples by the TruTip sample preparation method is a simple process that only requires a pipette for instrumentation [30] and is easy to automate

[31]. A sufficient amount of amplifiable human DNA was isolated from all tested Oragene•Dx/saliva samples in high quality. Overall, the TruTip extraction, lateral flow cell design and automated imaging and analysis allow for an expedient turnaround time of less than 4 h for a sample-to-answer result. This is comparable to other PCR-based commercial analytical platforms available for warfarin genotyping [13,16]. The simplicity and the closed amplicon design of the lateral flow cell reduced the handling steps as indicated by the number of pipetting steps per single sample run to 6, as compared to 18–71 for other commercial systems [13,19]. The analytical sensitivity experiments for determining the LoD to assess the genotyping performance showed that the lateral flow cell assay is robust over a wide range of input DNA amounts, as low as 330 copies of DNA. The end-to-end genotype testing using saliva samples demonstrated performance comparable to that using purified gDNA samples from Coriell Institute, and results directly correlated to those obtained from bidirectional sequencing. Performance of the assay with larger cohorts of patient samples will further determine its use as a potential clinical test to assist in patient care. In conclusion, this study describes a new technology for streamlining complex molecular diagnostic work flows. The use of non-invasive saliva samples stabilized with Oragene•Dx, rapid DNA extraction capabilities of TruTip technology, single-step lateral flow cell and the automated system for microarray image analysis, offers significant improvements in process efficiency compared to other test platforms. This system represents a considerable advance for an easy-to-use microarray-based diagnostics for warfarin sensitivity genotyping and is an important development towards the goal of a totally integrated system. Acknowledgments We thank Dane Brady for the excellent help in microarray manufacture and printing, Alissa Gindlesperger and Tinsley Stokes for DNA extraction from saliva samples by TruTip, and George Rudy for technical assistance in data analysis. We are grateful to Sensovation AG for semi-automated image and data analysis workflows on the FLAIR. References [1] Daly AK, King BP. Pharmacogenetics of oral anticoagulants. Pharmacogenetics 2003;13:247–52. [2] Wadelius M, Pirmohamed M. Pharmacogenetics of warfarin: current status and future challenges. Pharmacogenomics J 2007;7:99–111. [3] Budnitz DS, Pollock DA, Weidenbach KN, Mendelsohn AB, Schroeder TJ, Annest JL. National surveillance of emergency department visits for outpatient adverse drug events. JAMA 2006;296:1858–66. [4] Moyer TP, O'Kane DJ, Baudhuin LM, Wiley CL, Fortini A, Fisher PK, et al. Warfarin sensitivity genotyping: a review of the literature and summary of patient experience. Mayo Clin Proc 2009;84:1079–94. [5] Higashi MK, Veenstra DL, Kondo LM, Wittkowsky AK, Srinouanprachanh SL, Farin FM, et al. Association between CYP2C9 genetic variants and anticoagulationrelated outcomes during warfarin therapy. JAMA 2002;287:1690–8. [6] Veenstra DL, Blough DK, Higashi MK, Farin FM, Srinouanprachan S, Rieder MJ, et al. CYP2C9 haplotype structure in European American warfarin patients and association with clinical outcomes. Clin Pharmacol Ther 2005;77:353–64. [7] Li T, Chang CY, Jin DY, Lin PJ, Khvorova A, Stafford DW. Identification of the gene for vitamin K epoxide reductase. Nature 2004;427:541–4. [8] Rost S, Fregin A, Ivaskevicius V, Conzelmann E, Hörtnagel K, Pelz HJ, et al. Mutations in VKORC1 cause warfarin resistance and multiple coagulation factor deficiency type 2. Nature 2004;427:537–41. [9] Rettie AE, Wienkers LC, Gonzalez FJ, Trager WF, Korzekwa KR. Impaired (S)-warfarin metabolism catalysed by the R144C allelic variant of CYP2C9. Pharmacogenetics 1994;4:39–42. [10] Aithal GP, Day CP, Kesteven PJ, Daly AK. Association of polymorphisms in the cytochrome P450 CYP2C9 with warfarin dose requirement and risk of bleeding complications. Lancet 1999;353:717–9. [11] US Food and Drug Administration. New labeling information for warfarin (marketed as Coumadin). http://www.fda.gov/cder/drug/infopage/warfarin/ default.htm (Accessed August, 2007). [12] Gage BF, Lesko LJ. Pharmacogenetics of warfarin: regulatory, scientific, and clinical issues. J Thromb Thrombolysis 2008;25:45–51. [13] Maurice CB, Barua PK, Simses D, Smith P, Howe JG, Stack G. Comparison of assay systems for warfarin-related CYP2C9 and VKORC1 genotyping. Clin Chim Acta 2010;411:947–54. [14] Lefferts JA, Schwab MC, Dandamudi UB, Lee HK, Lewis LD, Tsongalis GJ. Warfarin genotyping using three different platforms. Am J Transl Res 2010;2:441–6.

T. Sebastian et al. / Clinica Chimica Acta 429 (2014) 198–205 [15] Spizz G, Young L, Yasmin R, Chen Z, Lee T, Mahoney D, et al. Rheonix CARD(®) technology: an innovative and fully automated molecular diagnostic device. Point Care 2012;11:42–51. [16] King CR, Porche-Sorbet RM, Gage BF, Ridker PM, Renaud Y, Phillips MS, et al. Performance of commercial platforms for rapid genotyping of polymorphisms affecting warfarin dose. Am J Clin Pathol 2008;129:876–83. [17] Langley MR, Booker JK, Evans JP, McLeod HL, Weck KE. Validation of clinical testing for warfarin sensitivity: comparison of CYP2C9–VKORC1 genotyping assays and warfarin-dosing algorithms. J Mol Diagn 2009;11:216–25. [18] Li Y, Jortani SA, Ramey-Hartung B, Hudson E, Lemieux B, Kong H. Genotyping three SNPs affecting warfarin drug response by isothermal real-time HDA assays. Clin Chim Acta 2011;412:79–85. [19] Poe BL, Haverstick DM, Landers JP. Warfarin genotyping in a single PCR reaction for microchip electrophoresis. Clin Chem 2012;58:725–31. [20] Stoughton RB. Application of DNA microarrays in biology. Annu Rev Biochem 2005;74:53–82. [21] Goldsmith ZG, Dhanasekaran N. The microrevolution: applications and impacts of microarray technology on molecular biology and medicine. Int J Mol Med 2004;13:483–95. [22] Chandler DP, Bryant L, Griesemer SB, Gu R, Knickerbocker C, Kukhtin A, et al. Integrated amplification microarrays for infectious disease diagnostics. Microarrays 2012;1:107–24. [23] Yershov G, Barsky V, Belgovskiy A, Kirillov E, Kreindlin E, Ivanov I, et al. DNA analysis and diagnostics on oligonucleotide microchips. Proc Natl Acad Sci U S A 1996;93:4913–8.

205

[24] Guschin D, Yershov G, Zaslavsky A, Gemmel A, Shick V, Prudnikov D, et al. Manual manufacturing of oligonucleotide microchips to enhance sequencing by hybridization. Anal Biochem 1997;250:203–11. [25] Cooney CG, Sipes D, Thakore N, Holmberg R, Belgrader P. Plastic disposable microfluidic flow cell for coupled on-chip PCR and microarray detection of infectious agents. Biomed Microdevices 2012;14:45–53. [26] Guo Z, Liu Q, Smith LM. Enhanced discrimination of single nucleotide polymorphisms by artificial mismatch hybridization. Nat Biotechnol 1997;15:331–5. [27] Golova JB, Chernov BK, Perov AN, Reynolds J, Linger YL, Kukhtin A, et al. Nonvolatile copolymer compositions for fabricating gel element microarrays. Anal Biochem 2012;421:526–33. [28] Bao YP, Huber M, Wei TF, Marla SS, Storhoff JJ, Müller UR. SNP identification in unamplified human genomic DNA with gold nanoparticle probes. Nucleic Acids Res 2005;33:2–7. [29] Hu Y, Ehli EA, Nelson K, Bohlen K, Lynch C, Huizenga P, et al. Genotyping performance between saliva and blood-derived genomic DNAs on the DMET array: a comparison. PLoS One 2012;7:e33968. [30] Chandler DP, Griesemer SB, Cooney CG, Holmberg R, Thakore N, Mokhiber B, et al. Rapid, simple influenza RNA extraction from nasopharyngeal samples. J Virol Methods 2012;183:8–13. [31] Holmberg RC, Gindlesperger A, Stokes T, Brady D, Thakore N, Belgrader P, et al. Highthroughput, automated extraction of DNA and RNA from clinical samples using TruTip technology on common liquid handling robots. J Vis Exp 2013:e50356. http://dx.doi.org/10.3791/50356.