Implementation of a Cost-Effective Unlabeled Probe High-Resolution Melt Assay for Genotyping of Factor V Leiden

Abstract

The Factor V Leiden mutation (FVL; c.1601G>A, p.Arg534Gln), the most common aberration underlying activated Protein C resistance, results in disruption of a major anticoagulation pathway and is a leading cause of inherited thrombophilia. A high-throughput assay for FVL mutation detection was developed using a single unlabeled probe on a high-resolution platform, the 96-well Roche 480 LightCycler (LC480) instrument. This method replaced the U.S. Food and Drug Administration-approved Roche Factor V Leiden kit assay on the LightCycler PCR instrument, decreasing total cost by 48%. The analytical sensitivity and specificity of the LC480 high-resolution assay approached 100% for the FVL mutation. Factor V mutations in proximity to the FVL locus may influence probe binding efficiency and melt characteristics. One out of three very rare variants tested in a separate study, 1600delC, was not distinguishable from FVL using the described high-resolution assay. However, a c.1598G>A variant, which changes the amino acid sequence from arginine to lysine at position 533, was detected by this high-resolution assay and confirmed by bidirectional sequencing. In the labeled probe LightCycler assay, the c.1598G>A variant was indistinguishable from the heterozygous FVL control. The c.1598G>A variant has not been described previously and its clinical significance is uncertain. In conclusion, the LC480 FVL assay is cost effective in a high-throughput setting, with capability to detect both previously described and novel FV variants.

Introduction

The Factor V Leiden mutation (FVL; c.1601G>A, p. Arg534Gln) (Dahlbäck et al., 1993; Bertina et al., 1994) renders the activated Factor V protein partially resistant to cleavage by activated Protein C (APC), thereby reducing the efficiency of a major negative feedback loop in the coagulation cascade and significantly increasing the risk for deep venous thrombosis. FVL is a founder mutation in the Caucasian population with highest prevalence in individuals of Northern European extraction. In this population, FVL is the most commonly detected risk factor for inherited thrombophilia. Heterozygosity for FVL confers a relative risk for a venous thrombotic event of about 4-8, while the risk for homozygous individuals is up to 80× that of normal. Currently used detection methods for FVL in clinical laboratories include restriction fragment length polymorphism (RFLP) methodology, which takes advantage of the fact that the FVL mutation abolishes a recognition site for the Mnl I endonuclease, allele-specific probe hybridization assays such as the nonamplification Invader assay (Third Wave Technology), and various polymerase chain reaction (PCR)-based techniques that are well suited for high-throughput environments. The U.S. Food and Drug Administration-approved Factor V Leiden kit manufactured by Roche Diagnostics is used on the LightCycler instrument. This kit features a pair of fluorescently labeled hybridization probes that bind to specific sites on an amplified fragment of the Factor V gene during the annealing phase of the PCR reaction. This allows the donor fluorophore (fluorescein) at the 3′-end of one probe to be in proximity to the acceptor fluorophore (LightCycler^® Red 640) on the other probe resulting in fluorescence resonance energy transfer (FRET) between the two fluorophores, as the donor fluorophore is excited by light emitted from the LightCycler instrument. The intensity of fluorescence emitted from the acceptor fluorophore is then measured by the instrument. After completion of the amplification reaction, a melting curve analysis is performed. This involves a gradual increase in temperature beyond the point where the shorter fluorescein-labeled probe, which spans the mutation site, dissociates (melts) from the template. If the mutation is present, the binding between the probe and the template will be less stable and melting will occur at a lower temperature. The melting temperature (Tm) of the probe corresponds to the temperature at which 50% of the probes have dissociated from the template. The Tm depends on the presence of the target mutation as well as other variants located under the probe. Melting analysis with determination of the Tm for a probe/template combination is used to confirm the presence of a specific mutation or variant.

Increased sensitivity for detection of unexpected sequence variants under the probe in melting analysis can be achieved by application of the delta Tm method (Lyon, 2005). This calculation method reduces the effects of variables such as initial target concentration, presence of amplification inhibitors, and differences in salt concentration. Delta Tm is calculated as the difference between the Tm of the wild-type and mutant peaks within a heterozygous sample. In a clinical run, the delta Tm values for each heterozygote are averaged, and the difference between the average delta Tm and the delta Tm for individual samples can then be calculated and compared with a threshold value. Within a run, the variation in delta Tm between different samples is generally <0.2. On the other hand, the variation in Tm of a specific allele (for instance, the wild-type peak) within a run may range from 0.2 to 0.5 (Lyon, 2005). In the clinical laboratory, samples with delta Tm that exceed the threshold are re-extracted and repeated. If the repeat confirms the delta Tm discrepancy, the sample undergoes bidirectional sequencing.

High-resolution melting analysis, that is, analysis of melting performed using specifically designed real-time PCR instruments with high precision and ability to acquire data at very high density, has emerged as a rapid and sensitive method for mutation scanning as well as targeted single-nucleotide polymorphism genotyping (for a review, see Erali et al., 2008). In high-resolution melting, samples can be genotyped by Tm as well as curve shape (Margraf et al., 2006a). The increased sensitivity makes it possible to distinguish variants with smaller differences in Tm compared with standard resolution techniques. With high-resolution melting, unlabeled probes can be used. For some applications involving small amplicons, it is also possible to perform the analysis without probes, by monitoring the melt of the amplicons themselves (Liew et al., 2004). During the initial cooling phase, the amplification products form homo- and heteroduplexes depending on whether a mutation is present in the sample. These variants will differ in their Tm and can be resolved by the melt analysis. However, amplicon melt analysis cannot always differentiate homozygous mutations from wild type, particularly in cases where the mutation results in insertion of a complementary base and the bases closest to the mutation are identical on both strands (nearest-neighbor thermodynamics symmetry). The ability to distinguish homozygous mutations from wild type can be increased by mixing the unknown sample with a sample of known genotype, thereby creating artificial heteroduplexes that can be more easily detected (Palais et al., 2005). The use of an unlabeled probe confers an added degree of sequence specificity desirable for clinical applications. The Tm of a probe is mainly determined by size and total GC content. The Tm shift induced by a mutation depends on the binding strength between the mutated base and its counterpart on the opposite strand, as well as nearest-neighbor effects. For defined variants, assay conditions such as temperature and ionic strength can be modified to optimize detection. Incorporation of locked nucleic acids (modified bases; LNA) into the probe may increase the ability to discriminate different sequence variants (Chou et al., 2005). The probe can also be modified using various masking techniques to remove variability in Tm induced by common benign variants under the probe (Margraf et al., 2006b). The development of plate-format instruments with high degree of temperature homogeneity within each well and across the plate, as well as improved intercalating dyes such as LCGreen, which are able to detect heteroduplexes, acting at saturating concentrations without inhibiting PCR (Wittwer et al., 2003) have been crucial for the advance of the high-resolution melting technology.

A new melting-based FVL assay was designed to further increase efficiency in a high-throughput setting. This laboratory-developed assay uses a short single unlabeled probe together with a saturating heteroduplex detecting dsDNA dye (Wittwer et al., 2003) on an integrated real-time PCR/high-resolution closed tube platform, the 96-well Roche 480 LightCycler instrument (LC480; Roche Diagnostics Corporation). To reduce hands-on time and increase speed for the entire analytic process, DNA extraction is performed using the Radius robotic system (Protedyne Corporation).

Several variants in proximity to the FVL locus have previously been identified using PCR/RFLP or PCR with fluorescent hybridization probes and melting analysis. Using the novel nonlabeled probe melting analysis assay on the real-time PCR/high-resolution melting LC480 platform, we detected a variant of uncertain significance, c.1598G>A, leading to an amino acid change (Arg533Lys) immediately adjacent to the Arg534 APC cleavage site in the activated Factor V protein. To our knowledge, this variant has not previously been described in the literature. When the same sample was tested using the fluorescent dual probe-based LightCycler assay, the peaks could not be discriminated from those produced by the heterozygous FVL mutation control.

Materials and Methods

DNA extraction

DNA was isolated using the MagNA Pure LC DNA Isolation Kit I (Roche Applied Science) on the Radius Instrument according to the manufacturer's recommended protocol (Protedyne Corporation). For repeats requiring manual DNA extraction, the Gentra Puregene DNA extraction kit (Qiagen Inc.) was used as detailed in the manufacturer's instructions.

PCR primers

Primers for amplification of a 151 bp fragment of the Factor V gene including the FVL locus were designed using Primer3 software as previously described (Rozen and Skaletsky, 2000) and synthesized at Integrated DNA Technologies, Inc. Primer sequences were 5′-CCC ATT ATT TAG CCA GGA GA-3′ for the forward primer and 5′-GCC TCT GGG CTA ATA GGA CT-3′ for the reverse.

Unlabeled probe

An unlabeled 31 nucleotide probe specific to the c.1601G (wild type, previously reported as 1691) with ∼40% GC content was synthesized at Integrated DNA Technologies, Inc. A 3′ amino modifier (AmM) was incorporated to prevent extension during the PCR reaction (Zhou et al., 2004; Dames et al., 2007). The probe sequence was 5′-TTC AAG GAC AAA ATA CCT GTA TTC CTC GCC T/3AmM/-3′. The Tm of the probe-amplicon pair was ∼71°C for the perfect match (wild type, c.1601G), and ∼66°C for the mismatched peak (c.1601G>A).

Genotyping assay on the LC480 instrument

PCR was set up with a final reaction volume of 10 μL on a 96-well block using the Radius instrument. The reaction contained 1 μL of DNA template with a concentration of 20-50 ng/μL, 1 μL Roche LC DNA Hybridization buffer 10×, 0.8 μL MgCl₂ (1 mM; final concentration 1.8 mM), 0.1 μL UNG (AmpErase™ Uracil N-Glycosylase, 1U/μL; final concentration 0.01U; Applied Biosystems), and 1 μL LCGreen Plus (final concentration 1×; Idaho Technology). The PCR was performed on a LightCycler 480 instrument (LC480; Roche Diagnostics). To optimize the probe binding reaction, an asymmetric prolonged PCR was performed allowing formation of excess single-stranded reverse DNA. The final concentration of the forward primer, the reverse primer, and the unlabeled probe was 0.1, 1, and 1 μM, respectively. Thermocycling was performed according to the following protocol: UNG denaturation (50°C for 3 min), polymerase activation (95°C for 5 min), followed by 60 amplification cycles (denaturation at 95°C for 15 s, annealing at 56°C for 15 s, and extension at 72°C for 15 s). After amplification, the PCR products were denatured at 95°C for 60 s and rapidly cooled to 40°C to generate probe/ssDNA amplicon heteroduplexes. Melting analysis was performed between 50°C and 88°C at 30 acquisitions per 1°C, using the melting curve analysis mode. Samples were then cooled to 40°C for 10 s. The rate for heating was 4.4°C/s, and the rate for cooling was 1.8°C/s.

Data analysis

Postamplification melting curves are analyzed using LightCycler 480 software version 1.2 (Roche Diagnostics Corporation) and MeltWizard software version 3.1, which determines Tm, clusters samples, and identifies outliers. MeltWizard 3.1 is designed for compatibility with automatic extraction and melting instrumentation through automated analysis and export of clinical results. In its first stage, the no template controls are automatically excluded, using an absolute amplitude criterion as well as a threshold for the ratio of signal to exponential background. The latter is obtained from a running estimate of a derivative-to-amplitude ratio, which would be expected to be constant for a pure exponential. The exponential background model is also used for normalization, as previously described (Palais and Wittwer, 2009). Briefly, an exponential is uniquely determined by its slope at two points. These slopes may be obtained from the slope of the raw data in any region where the melting curve is horizontal and the raw fluorescence vs. temperature is locally exponential, as identified by the constant ratio condition above.

To scan for heterozygous mutations, curves may be overlaid to emphasize shape differences over absolute Tm differences, but for genotyping, this is unnecessary and sacrifices the ability to distinguish homozygous variants with significant Tm shifts. For the Factor V assay, it is sufficient to determine a clustering hierarchy using unbiased single-linkage hierarchical clustering based on discrete L^p distance between curves. The genotype level of the hierarchy is automatically characterized by the unique maximum jump between intraclass and interclass distances, and the genotypes are redundantly identified as WT, HET, and MUT by the number and ranges of their Tm values. Delta Tm values are calculated for all specimens classified as heterozygotes, and an average delta Tm is determined.

Postamplification melting curves were analyzed using LightCycler 480 software version 1.2 (Roche Diagnostics Corporation) and MeltWizard software version 3.1, which assigns samples into subsets based on Tm (clustering), and identifies outliers. In MeltWizard, normalization was done by selecting linear regions before and after the melt (100% and 0% fluorescence, respectively). To facilitate clustering, temperature shifting of individual samples was performed by moving the melt curves along the X-axis. The Autogroup function was then used to induce clustering. Subtraction of each curve from a reference curve generated a difference plot, where fluorescence of all curves was plotted against temperature. Delta Tm values were calculated for all specimens classified as heterozygotes, and an average delta Tm was determined.

Quality control and assignment of genotypes

The following controls were included in each clinical run: one negative sample, one heterozygous sample, one homozygous sample, and two no template controls (no DNA template added). All control samples were previously genotyped using the FVL LightCycler fluorescent labeled hybridization probe assay. A tested sample that shows a single melting peak at 66°C is considered homozygous positive for the FVL mutation. A tested sample that shows a single peak at 71°C (WT) is considered to be FVL mutation negative (wild type). A sample that shows two peaks, at 66°C and 71°C is heterozygous for FVL.

Labeled probe LightCycler assay

The assay was performed using the Factor V Leiden Kit (Roche Diagnostics) according to the manufacturer's instructions. Briefly, a 222 bp fragment of the Factor V gene was amplified from isolated DNA using specific primers (LightCycler Factor V Leiden, Primer/Hybridization Probes; Roche Diagnostics), using a LightCycler instrument. Melting analysis was then performed using FRET hybridization probes (LightCycler DNA Master Hybridization probes). Genotyping was performed by analysis of Tm. In the presence of the FVL mutation, the resulting mismatch destabilized the hybrid, leading to a lower Tm compared with the wild type. To increase sensitivity for detection of unknown variants under the probes, delta Tm was calculated for all samples displaying two peaks, and the delta Tm values for individual samples were then compared with the average delta Tm for all heterozygotes. The threshold for calling a suspicious variant, prompting further investigation, was delta Tm >0.2.

Results

Validation of the unlabeled probe LC480 assay

In the initial validation experiments, 29 wild-type samples, 30 samples heterozygous for FVL mutation, and 30 samples homozygous for FVL were tested under the conditions described in the Materials and Methods section. The averaged Tm for the wild-type samples was 71.3°C ± 0.5°C (2 standard deviation [SD]). One of the heterozygous samples failed the first amplification, but amplified and showed the expected genotype when repeated. This sample was not included in the concordance study. The remaining samples showed an averaged FVL mutation peak (Tm1) of 66.1°C ± 0.2°C (2SD) and an averaged wild-type peak (Tm2) of 71.2°C ± 0.4°C (2SD). The average delta Tm (calculated as the difference between Tm1 and Tm2) was 5.1°C ± 0.3°C (2SD). The homozygous samples had an averaged Tm of 66.1°C ± 0.2°C (2SD). The analytical sensitivity and specificity, calculated from concordance with previous genotyping with the LightCycler labeled probe assay, was 100%.

Typical detection data from the probe-amplicon melt of a wild-type sample, a sample with the FVL mutation in heterozygous form, a sample with the FVL mutation in homozygous form, and a no template control are shown in Figure 1A.

FIG. 1.

(A) Detection of the FVL mutation using the unlabeled probe LC480 assay. Wild-type, homozygous mutant (c.1601G>A), heterozygous mutant (c.1601G>A), and no template control samples are indicated by arrows. (B) Detection of previously described Factor V variants (c.1600C>T, 1600delC, and c.1599G>A), as well as the FVL mutation, with the unlabeled probe LC480 assay. While the c.1600C>T and c.1599 G>A variants can be resolved due to delta Tm discrepancy, the c.1600delC variant cannot be distinguished from the FVL mutation. The wild-type, homozygous mutant (c.G1601A), heterozygous mutant (c.G1601A), heterozygous variant (c.G1599A), heterozygous variant (c.C1600T), heterozygous variant (c.1600delC); and no template control samples are indicated by arrows. (C) Detection of a novel variant (c1598G>A) on the LC480 instrument. The 1598G>A variant is readily distinguished from the FVL heterozygous sample and the wild type. FVL, Factor V Leiden.

To determine within-run reproducibility, one wild-type, one heterozygous, and one homozygous sample were tested in triplicate in one run. The averaged Tm for the wild-type sample was 70.8°C ± 0.3°C (2SD), and the averaged Tm for the homozygous sample was 66.1°C ± 0.1°C (2SD). For the heterozygous sample, the averaged Tm1 was 65.8°C ± 0.8°C (2SD), the averaged Tm2 was 70.9°C ± 0.7°C (2SD), and the averaged delta Tm was 5.0°C ± 0.3°C (2SD).

The between-run reproducibility was tested using the same samples as in the within-run reproducibility test. The samples were assessed in five consecutive runs and the values from each run were averaged. In this study, the averaged Tm for the wild-type sample was 70.8°C ± 0.4°C (2SD), and the averaged Tm for the homozygous sample was 66.0°C ± 0.4°C (2SD). For the heterozygous sample, the averaged Tm1 was 66.0°C ± 0.4°C (2SD), the averaged Tm2 was 70.9°C ± 0.6°C (2SD), and the average delta Tm was 4.9°C ± 0.3°C (2SD). The between-run reproducibility was also evaluated for a heterozygous FVL control sample in ten consecutive clinical runs. The averaged Tm1 was 65.245°C ± 0.0.86°C (2SD), the averaged Tm2 was 70.272°C ± 0.58°C (2SD), and the average delta Tm was 5.03°C ± 0.4°C (2SD).

In the validation, the no template control was noted to show a primer-dimer peak (confirmed by gel electrophoresis analysis) at ∼74°C-76°C; however, this peak did not interfere with the interpretation of the unlabeled probe/template DNA melting analysis performed between ∼62°C and 74°C. A noninterfering amplicon melting peak was noted at ≥80°C.

Detection of previously characterized FV variants using the unlabeled probe LC480 assay

Samples containing three FV variants located in proximity to the FVL mutation were genotyped using the novel unlabeled probe assay. Figure 1B demonstrates the detection of the FVL mutation and these sequence variants in a melt difference plot (c.1600C>T, c.1600delC, and c.1599G>A). In the accuracy study, the sample containing the heterozygous c.1599G>A variant showed a Tm1 of 67.9°C and a Tm2 of 71.2°C, with a delta Tm of 3.3°C; the sample with the c.1600C>T variant showed a Tm1 of 67.3C°C and a Tm2 of 71.2°C, with a delta Tm of 3.9°C; the sample with the c.1600delC variant had a Tm1 of 66.2°C and a Tm2 of 71.0°C, with a delta Tm of 4.8°C. The 90% confidence interval (CI) for the heterozygous FVL mutation was (5.02, 5.11), and the 99% CI was (5.00, 5.14). Using a threshold of 0.2 for a difference in delta Tm to suggest that a sample contains a variant, both the 1599G>A sample (delta Tm 3.3°C) and the 1600C>T sample (delta Tm 3.9°C) would have been referred for bidirectional sequencing if they had been detected in the clinical laboratory. However, the delta Tm of the 1600delC variant (4.8°C) could not be used to distinguish it from the FVL mutation (99% CI 5.00, 5.14).

In the within-run reproducibility test where the same sample set was tested in triplicate in the same run, the averaged delta Tm was 3.5°C ± 0.6°C (2SD; 90% CI [3.2, 3.5], 99% CI [3.1, 3.9]) for the sample with the heterozygous c.1599G>A variant, and 3.6°C ± 0.0°C (2SD; 90% CI [3.6, 3.6], 99% CI [3.6, 3.6]) for the sample with the heterozygous c.1600C>T variant. Thus, the delta Tm values of both of these variants were clearly distinguishable from that of the heterozygous FVL (90% CI [4.9, 5.0]; 99% CI [4.8, 5.2]). However, the delta Tm of the heterozygous 1600delC variant (90% and 99% CI [4.6, 4.7]); (4.6, 4.8) could not be used to distinguish this variant from the heterozygous FVL mutation.

Similarly, in a between-run reproducibility study where the same sample set was tested in five consecutive runs, the heterozygous c.1599G>A variant showed an averaged delta Tm of 3.7°C ± 0.1°C (2SD; 90% CI [3.60, 3.70], 99% CI [3.56, 3.74]); the sample with the c.1600C>T variant showed an averaged delta Tm of 3.9°C ± 0.3°C (2SD; 90% CI [3.74, 3.97], 99% CI [3.68, 4.03]). Thus, both the heterozygous c.1599G>A variant and the heterozygous c.1600C>T variant could be differentiated from the FVL variant (delta Tm 4.9°C ± 0.3°C [2SD], 90% CI [4.79, 5.02], 99% CI [4.73, 5.09]), on the basis of delta Tm. However, the delta Tm of the 1600delC variant 4.6°C ± 0.1°C (2SD; 90% CI [4.58, 4.66]; 99% CI [4.55, 4.68]) did not allow for this variant to be separated from the heterozygous FVL mutant.

Detection of a novel FV variant

In a clinical run using the described unlabeled probe assay, one sample was noted to have a single melting peak with a shifted Tm (69.6°C) when compared with the single melting peak of the wild-type samples (70.6°C) and the single melting peak of the homozygous FVL mutation controls (65.6°C) (Fig. 1C). Bidirectional sequencing revealed that the sample was negative for the FVL mutation, but heterozygous for a G to A nucleotide substitution at position 1598 (c.1598G>A). This variant changes the amino acid sequence from arginine to lysine at position 533 (p.Arg533Lys). To our knowledge, this variant has not been described previously, and the clinical significance is therefore uncertain. The sample containing the identified variant was also tested with the labeled probe FRET assay on the LightCycler instrument. Two peaks were seen, similar to the pattern for the heterozygous FVL control. The difference between the delta Tm of this sample and that of the heterozygous FVL control sample was less than the threshold for significant delta Tm differences of 0.20. Thus, with this assay, the sample would have been genotyped as heterozygous FVL mutation.

Discussion

A high-throughput unlabeled probe assay for the FVL mutation was developed on the Roche 480 LightCycler (LC480) 96-well plate platform. The LC480 instrument has previously been shown to compare well with other similar platforms (Herrmann et al., 2007). Compared with previous high-resolution platforms, the LC480 reduces workflow complexity, since it allows for more reactions to be performed simultaneously and for both the PCR and high-resolution melting analysis to be performed consecutively in a closed system. In the nonlabeled probe LC480 assay, MeltWizard3.1 software utilizes algorithms previously described (Palais and Wittwer, 2009) to classify samples as wild type, heterozygous, or homozygous by unsupervised clustering based on curve distance, and supervised learning algorithms based on Tm. Samples that are not classified with any of the three biallelic single-nucleotide polymorphism genotypes based on both intra- to intercluster metric, and Tm ranges are flagged as outliers. Automated analysis of 96-well plates can be performed in a matter of seconds from import to export using this system.

The LC480 assay replaced the fluorescently labeled probe Roche LightCycler assay in our laboratory. The cost of the two assays was monitored and compared during the transition. Cost calculations were performed as previously described (Farnsworth et al., 2007). With the new LC480 assay, labor cost was decreased by 18%. This was accomplished by reducing the complexity of performance and by using high-throughput instrumentation for DNA extraction, as well as for the amplification and melt reactions. The cost for direct reagent/supplies decreased by 56%, due to the use of a laboratory-developed procedure and the use of an unlabeled probe for melting analysis. The total cost per analyzed sample decreased by 48%.

As previously described, high-resolution assays are relatively sensitive to sample quality issues, such as low template concentration, which can lead to increased spread of melting curves within the wild-type and mutated sample groups (Krypuy et al., 2006; Seipp et al., 2007; De Leeneer et al., 2008). Tm is dependent on ionic strength, and this underscores the importance of using a DNA isolation protocol that allows for optimized and uniform sample extraction (Seipp et al., 2007). In cases where genotyping results are suboptimal, samples are re-extracted using a manual extraction method (Puregene). Evaporation during cycling can also affect PCR and melting quality in the LC480 system. This can be avoided by overlaying the samples with mineral oil before cycling.

Other sequence variants in proximity to the FVL mutation have been shown to interfere with detection of FVL in both RFLP and PCR-based assays, by affecting restriction enzyme cut sites and primer or probe binding sites, respectively. Several Factor V variants have previously been detected in our laboratory using the LightCycler assay with fluorescent hybridization probes and melting analysis, followed by confirmatory sequencing assay, as well as amplicon melting technology (Erali et al., 2008). Three such rare Factor V variants were tested using the LC480 assay. For the heterozygous c.1599G>A and c.1600C>T variants, the delta Tm values obtained would suggest that a variant may be present. This would prompt further exploration by sequencing. However, the heterozygous c.1600delC variant could not be differentiated from the heterozygous FVL mutation by the unlabeled probe LC480 assay. The one-base deletion leads to a shift in reading frame, and would be predicted to result in FV deficiency rather than increased risk for thrombosis. The prevalence for the c.1600delC, c.1599G>A, and c.1600C>T variants in the population tested for FVL is believed to be very low; likely <1/10,000 (Lyon, 2005). The unlabeled probe LC480 assay was, however, able to distinguish and prompt investigation by sequencing of another variant (c.1598G>A; p.Arg533Lys) that would have been interpreted as heterozygous FVL by the labeled probe LightCycler assay. This variant, which substitutes lysine for arginine at codon 533 adjacent to the FVL cut site, has not been previously described in the literature. This demonstrates that the unlabeled probe LC480 assay is capable of detecting several unexpected FV variants. Most variants identified in FV other than FVL are either silent or lead to factor deficiency. The p.Arg533Lys variant is conservative in terms of charge. Similar to the FVL locus, the c.1598G>A base is only moderately conserved. However, due to the location of this variant, the possibility that it could interfere with cleavage by APC, and thus contribute to thrombophilia, cannot be entirely ruled out. The patient with the c.1598G>A variant was tested for FVL after being hospitalized with pulmonary embolism. Given the dissimilarities between the unlabeled probe high-resolution assay and the labeled probe LightCycler assay in terms of probes, as well as amplification and melt analysis protocols, it is not surprising that the ability to distinguish certain unexpected FV variants differs between the two platforms.

In conclusion, the newly designed unlabeled probe high-resolution LC480 assay for detection of FVL was found to be cost effective in a high-throughput setting with capability to detect both some previously described Factor V variants and a novel Factor V variant, as compared with the previously used FRET probe LightCycler assay.

Footnotes

Acknowledgment

This work was supported by the ARUP Institute for Clinical and Experimental Pathology.

Disclosure Statement

No competing financial interests exist.

References

Bertina

, Koeleman

BPC

, Koster

et al. 1994. Mutation in blood coagulation factor V associated with resistance to activated protein C. Nature, 369:64-67.

Chou

, Meadows

, Wittwer

et al. 2005. Unlabeled oligonucleotide probes modified with locked nucleic acids for improved mismatch discrimination in genotyping by melting analysis. BioTechniques, 39:644-650.

Dahlbäck

, Carlsson

, Svensson

. 1993. Familial thrombophilia due to a previously unrecognized mechanism characterized by poor anticoagulant response to activated protein C: prediction of a cofactor to activated protein C. Proc Natl Acad Sci USA, 90:1004-1008.

Dames

, Margraf

, Pattison

et al. 2007. Characterization of aberrant melting peaks in unlabeled probe assays. J Mol Diagn, 9:290-296.

De Leeneer

, Coene

, Poppe

et al. 2008. Rapid and sensitive detection of BRCA1/2 mutations in a diagnostic setting: comparison of two high-resolution melting platforms. Clin Chem, 54:982-989.

Erali

, Voelkerding

, Wittwer

. 2008. High resolution melting applications for clinical laboratory medicine. Exp Mol Pathol, 85:50-58.

Farnsworth

, Jennison

, Theurer

et al. 2007. A systematic approach to direct cost analysis: the experience of one laboratory with labor and supply cost allocation. Clin Leader Manag Rev Jan/Feb, 4-11.

Herrmann

, Durtschi

, Wittwer

et al. 2007. Expanded instrument comparison of amplicon DNA melting analysis for mutation scanning and genotyping. Clin Chem, 53:1544-1548.

Krypuy

, Newnham

, Thomas

et al. 2006. High resolution melting analysis for the rapid and sensitive detection of mutations in clinical samples: KRAS codon 12 and 13 mutations in non-small cell lung cancer. BMC Cancer, 6:295.

10.

Liew

, Pryor

, Palais

et al. 2004. Genotyping of single-nucleotide polymorphisms by high-resolution melting of small amplicons. Clin Chem, 50:1156-1164.

11.

Lyon

. 2005. Discovering rare variants by use of melting temperature shifts seen in melting curve analysis. Clin Chem, 51:1331-1332.

12.

Margraf

, Mao

, Highsmith

et al. 2006a. Mutation scanning of the RET protooncogene using high-resolution melting analysis. Clin Chem, 52:138-141.

13.

Margraf

, Mao

, Wittwer

. 2006b. Masking selected sequence variation by incorporating mismatches into melting analysis probes. Hum Mutat, 27:269-278.

14.

Palais

, Liew

, Wittwer

. 2005. Quantitative heteroduplex analysis for single nucleotide polymorphism genotyping. Anal Biochem, 346:167-175.

15.

Palais

, Wittwer

. 2009. Mathematical algorithms for high-resolution DNA melting analysis. Methods Enzymol, 454:323-343.

16.

Rozen

, Skaletsky

. 2000. Primer3 on the WWW for general users and for biologist programmers. Krawetz

, Misener

. Bioinformatics Methods and Protocols: Methods in Molecular Biology. Humana: Press, Totowa, NJ, 365-386.

17.

Seipp

, Durtschi

, Liew

et al. 2007. Unlabeled oligonucleotides as internal temperature controls for genotyping by amplicon melting. J Mol Diagn, 9:284-289.

18.

Wittwer

, Reed

, Gundry

et al. 2003. High-resolution genotyping by amplicon melting analysis using LCGreen. Clin Chem, 49:853-860.

19.

Zhou

, Myers

, Vandersteen

et al. 2004. Closed-tube genotyping with unlabeled probes and a saturating DNA dye. Clin Chem, 50:1328-1335.