| Introduction
In collaboration with Dr. Brian Reid at the Fred Hutchinson Cancer Center (Seattle, WA), our laboratory is using LATE-PCR technologies to develop a clinically compatible platform for the convenient and reproducible detection of chromosomal DNA deletions involving tumor suppressor genes in human cancers. Chromosomal deletions leading to loss of heterozygosity (LOH) are among the earliest and most common genomic changes in many cancers. In particular, LOH events involving tumor suppressor genes serve as important indicators (biomarkers) of future cancer development in particular pre-malignant conditions. For esophageal cancers, a deadly malignancy with only 0.9% five-year survival rate, Dr. Reid and coworkers have established that deletion of tumor suppressor genes p53 (TP53) in chromosome 17p and p16 (CDKN2A) in chromosome 9p together with tetraploidy/aneuploidy are highly predictive of cancer risk in pre-malignant Barrett’s esophagus (BE). Translation of these findings to the clinic, however, has been severely hindered by the complexity and technical expertise required by current laboratory methods of LOH detection.
To overcome these shortcomings, we have devised a new, easy-to-use strategy based on Linear-After-The-Exponential (LATE) PCR endpoint assays. Currently, we are developing and validating LATE-PCR assays for reliable detection of 9p{p16} LOH and 17p{p53} LOH biomarkers in human cancers. These assays integrate genomic DNA preparation, efficient amplification of single-stranded amplicons, and a closed-tube, endpoint ratio-analysis of fluorescent probe signals at three temperatures. These ratios reveal with 99.7% accuracy whether the interrogated gene locus is diploid or has undergone LOH. As published in 2004, 2005, and 2006, LATE-PCR endpoint assays are remarkably convenient, robust, and sensitive: the assays are independent of the amount of input DNA or the extent of LATE-PCR amplification past the initial point of product detection and are sensitive down to the single genome level. Currently, we are rigorously evaluating the utility of these LATE-PCR assays by testing fresh/frozen BE epithelial cells of known LOH status provided to us by the Reid laboratory. Quantitative assay validation in our laboratory will facilitate subsequent clinical testing for early identification of pre-malignant epithelia in patients with the highest risk of developing cancer.
Background and significance
One of the major goals in cancer detection and diagnosis is to identify and evaluate molecular or cellular differences between normal, pre-malignant, and tumor tissues that are useful for the identification of individuals with increased risk of developing cancer or for clinical management of the disease. According to the multi-step model of neoplastic progression, cancers arise as a result of acquired genomic instability and evolution of clones with accumulated somatic genetic abnormalities. These somatic genetic changes can be used as potential biomarkers to distinguish patients at various risks for progression to malignancy with appropriate sensitivity and specificity. Such early detection is crucial to reduce cancer morbidity and mortality.
Validated cancer biomarkers may significantly improve cancer diagnosis and management by enabling cancer identification at the pre-symptomatic stage, by enhancing prognosis and disease monitoring in cases where clinical and histological features are not sufficient, and by providing a better assessment of future relapse and individual cancer risk. However, the large number of target genes in cancer and the mutiple potential alterations of these genes (including LOH, mutation, methylation) has made it difficult to identify key biomarkers that accurately predict future cancer development. In addition, effective use of cancer biomarkers requires detection technologies that are sensitive and accurate enough to identify the small initial number of biomarker positive cells in asymptomatic patients. Biomarker-based diagnostic assays must also be easily applied and of low cost to facilitate their ultimate implementation into routine clinical practice.
We are overcoming the above difficulties by (1) focusing on Barrett’s esophagus, a pre-malignant condition for which single-center, phase III and IV studies have identified and validated a specific biomarker panel capable of predicting the risk of developing esophageal adenocarcinoma better than existing diagnostic protocols based on morphological assessment, and (2) by employing a LATE-PCR platform that permits robust end-point detection of these biomarkers in a single-tube format with minimal operator expertise. Development of a clinical compatible platform for biomarker assessment in Barrett’s esophagus will accelerate validation of these and other LOH biomarkers increasingly being found in many other pre-malignant conditions.
Strategy for early cancer detection
Epithelial cancers represent the majority of human malignancies, and many studies have demonstrated that neoplastic progression in human epithelium is associated with accumulation of somatic genetic abnormalities. Loss of tumor suppressor genes due to LOH are among the earliest and most common genomic changes reported in many cancers. In particular, LOH involving the tumor suppressor genes p16 in chromosome 9 and p53 in chromosome 17 directly disrupt major cell cycle/instability pathways and are important contributors to cancer development. In accord with their fundamental role in carcinogenesis, these abnormalities occur in pre-malignant cells associated with many types of cancers, including esophageal cancer. Thus, there is great interest in using these LOH events as clinical biomarkers for the identification and early management of individuals at high risk for developing cancer.
Barrett’s Esophagus (BE), the only known precursor of Esophageal Adenocarcinoma (EA)
Esophageal cancer represents about 1 percent of the cancers diagnosed in the U.S. but its incidence has been increasing at the alarming rate of 300% and 350% in Caucasian males and females, respectively, in the Western world during the last three decades. These cancers have a 5 year survival rate of only 0.9% when detected late.
In BE, an intestinal metaplasia displaces the normal squamous epithelium lining the esophagus in response to damage caused by chronic acid and bile reflux into the esophagus. As a result, the esophagus lining turns from its normal whitish pink color to a diagnostic red color. BE patients (which includes as many as 5%-10% of chronic acid reflux individuals) have a 30- to 40-fold increased risk of developing EA and progress towards this condition at a rate of approximately 0.5%–1.0% per year.
The only proven cure for EA is surgical removal of the esophagus, a procedure with a mortality of 16%-23% in low-volume centers and 3%-8% in high-volume centers. Early identification of BE patients with the greatest risk of progressing to EA is, therefore, critical for improving survival: BE patients whose cancer was detected early by surveillance had a five year survival of 73% as compared to 0% among those whose cancers were detected by clinical symptoms. As a result, management of BE typically involves placing all patients in a surveillance program of periodic endoscopic biopsies. This is an expensive endeavor because the numbers of BE patients that will develop EA is small and regular endoscopic biopsies are costly. There is, therefore, a very urgent need to develop tests for risk stratification of BE patients in order to better focus surveillance efforts.
Biomarker studies in Barrett’s Esophagus
Current diagnostic criteria for EA risk is based on a five-tiered classification of dysplasia (negative, indefinite, low-grade dysplasia (LGD), high-grade dysplasia (HGD) and cancer) and is not particularly accurate due to observer variation and inadequacy of dysplasia as an indicator of EA development. Some BE patients with HGD progress to cancer, others regress, and yet others remain stable. As a result, the five-year accumulative cancer risk with HGD ranges from 15% to 57% in several studies. Two recent studies reported complete failure of endoscopic surveillance to detect early, curable cancers in BE patients. These findings have led to renewed calls to develop accurate diagnostic strategies based on objective molecular biomarkers.
Although more than 350 biomarkers have been proposed for BE, a recent review of the literature concluded that the principal genetic abnormalities in EA were p16 inactivation, p53 inactivation, aneuploidy, 5q LOH, 13q LOH, 18q LOH, and cyclin D1 overexpression. 5q LOH, 13q LOH, and 18q LOH could occur before or after cancer, making them less suitable as biomarkers for early detection. Although there was initial enthusiasm for cyclin D1 overexpression, it proved unreliable as a biomarker in subsequent studies. In contrast, LOH involving 9p{p16} and 17p{p53} together with DNA content abnormalities (tetraploidy, aneuploidy) consistently developed before EA. Alterations in the p16 (CDKN2A/INK4A) tumor suppressor gene (LOH, mutation, promoter hypermethylation) are among the earliest known genetic/epigenetic lesions in BE, occurring in 90% of BE segments that eventually progress to EA. Loss of the tumor suppressor gene p53 (TP53) due to LOH occurs later in disease progression than alterations in the p16 gene and may promote increased risk of progression towards EA formation by favoring the development of chromosomal unstable aneuploid populations. In a study of 269 patients, 17p (p53) LOH was associated with an increased rate of progression to EA formation (RR=16; {95% CI = 6.2,39; p<0.001}), to tetraploidy (RR = 6.1 {95% CI = 3.0,12; p<0.001}), and aneuploidy (RR = 7.5 {95% CI = 3.5,16; p<0.001}). These findings led the Reid laboratory to propose that a combined biomarker panel composed of p16 LOH, p53 LOH and abnormal DNA content would exhibit the greatest specificity and sensitivity for predicting EA. As shown below this hypothesis has been confirmed in a single center, phase IV study.
Single Center, Phase IV Evaluation of the Combined Barrett’s Esophagus Biomarker Panel (9p{p16} LOH, 17p{p53}LOH, tetraploidy, aneuploidy)
In a longitudinal study involving 274 BE patients followed for a median of 58 months, BE patients whose biopsies at the first endoscopy did not exhibit 9p{p16} LOH, 17p{p53}LOH, or tetraploidy/aneuploidy did not progress to EA during the course of the study. However, BE patients who had one, two, or three of these biomarker abnormalities at the first endoscopy had statistically significant increases in progression to EA. The five-year cumulative incidence of EA in BE patients with these three biomarker abnormalities was 79%, compared to less than 30% and 10% when only two or one abnormal biomarker occurred. This is a marked improvement over reports for high-grade dysplasia for which 5 to 8 per year cumulative incidences of EA have ranged from 9% to 59%. Other mechanisms of p16 and p53 gene inactivation (mutation, promoter hypermethylation) provided no independent predictive significance once 9p{p16} LOH and 17p{p53} LOH were incorporated into the three biomarker panel. The three biomarker panel also improved cancer risk prediction when stratified on dysplasia risk classification. Thus, among low-risk patients (negative, indefinite and LGD), the biomarker panel identified a small but significant population whose risk of progressing to EA was as high as if they had been diagnosed with HGD. Conversely, among BE patients with HGD, the biomarker panel identified a low-risk subset that followed a benign course throughout the study. Thus, based on these findings, it can be anticipated that a validated clinical assay for LOH detection will significant improve clinical management of BE patients predisposed to adenocarcinomas. Unfortunately, current research laboratory methods for LOH detection are too costly and technically complex for routine clinical use. As a result, there is a crucial need to develop alternative LOH biomarker assays that are affordable, simple, and easy to perform in a clinical setting.
The problem with current laboratory methods for LOH detection
Chromosomal deletions involving tumor suppressor genes may encompass as little as a single polymorphic genetic marker or as much as an entire chromosomal arm. Such deletions may even involve loss of the entire chromosome containing the normal allele, and may be accompanied by duplication of the remaining loci with no net change in copy number. As a result, identification of such chromosomal deletions requires genotyping methods capable of detecting loss of heterozygosity (LOH), rather that mere changes in sequence copy number. LOH describes the loss of an allele at one or more polymorphic loci for any given genetic marker (microsatellites markers, single nucleotide polymorphism or restriction fragment length polymorphisms). Since most tumor samples exhibiting LOH are contaminated with stromal and/or infiltrated inflammatory (normal) cells, LOH is often identified not as the absolute loss of an allele at a polymorphic locus but as an allele imbalance and often requires enrichment of epithelial tumor cells away from stromal and other cell types. As a result, LOH detection generally involves quantification of allele ratios at polymorphic genetic markers. The ideal LOH assay should provide both locus-specific genotype information and accurate quantification of relative allele copy number. Historically, enrichment of BE epithelial cells has been achieved by flow cytometry selection for Ki-67 expression, a proliferation marker in BE esophageal mucosa.
LOH was initially identified using polymorphic single tandem repeat (STR or microsatellites) as genetic markers for individual homologue chromosomes. STRs are highly informative because they typically consist of multiple alleles with different numbers of repeats. However, STRs are not very abundant and their detection requires sizing via capillary gel electrophoresis. More recently, STR markers have been replaced by single nucleotide polymorphisms (SNP) markers. SNPs are less informative than STRs because, when they are allelic, they generally only consist of two alleles. However, the number of identified SNP sites is 50-fold greater than the number of STRs in the human genome, with an average of one SNP site per every kilobase. Thus, LOH in virtually any region of the genome can be detected once a heterozygous SNP site in that region is identified. In addition, SNP alleles consist of unique base pair sequences (i.e. A, T, G, C), rather than size variants, making them far more convenient to detect the presence of heterozygosity among homologues.
LOH detection via SNP genotyping using RFLP analysis, real-time PCR, ligation-mediated PCR with mobility modifiers, DNA arrays, MALDI-TOF mass spectrometry, or Pyrosequencing is well established and is already being used in single center phase I, III and IV studies. SNP genotyping via Pyrosequencing, in particular, uses a sequencing-by-synthesis method that yields quantitative information on the fraction of the genomes within a sample corresponding to a particular variant. Similar quantitative information can also be obtained by MALDI-TOF (matrix-assisted laser de-absorption/ionization time-of-flight) mass spectrometry. In this method, individual SNP alleles are identified by the unique mass of small primer extension oligonucleotides that include the SNP site. Quantification of relative allele numbers in a Taqman-based homogeneous PCR assay requires monitoring of the amplification reaction in real-time (real-time PCR), use of allele-specific specific hybridization probes of different colors, and use of internal reference standard target of known copy number and similar amplification efficiency.
Despite these improvements, several technical considerations have limited the use of these methods for detection of LOH biomarkers in routine clinical practice. First, current research methods include multiple steps (i.e., DNA extraction, locus-specific PCR amplification, quantitation of relative allele copy number after PCR) which must be performed serially in separate tubes. Such experimental strategy risks introducing errors, and opens the possibility for sample loss, or even laboratory contamination with amplified products at any point during analysis. Second, these assays are not automatable and require significant technical expertise. These technical considerations complicate LOH assay design and increase assay costs. Thus, there is a need for robust, clinically compatible genotyping platform for LOH assessment that is quantitative, amenable to end-point analysis in a single tube format, and easy to design. As discussed in the next section, our experimental strategy satisfies all of these criteria without requiring an additional reference standard target.
Logic of LATE-PCR endpoint assays
Over the past five years, we have developed diagnostic assays that efficiently integrate the steps of sample preparation, PCR amplification, and end-point product copy number analysis into a single-tube format (see Figure 1). We refer to these assays as LATE-PCR endpoint assays. We have already demonstrated the utility of LATE-PCR endpoint assays for genotyping of SNP markers and gene mutations. We have also generated proof-of-principle data for detecting loss of tumor suppressor genes in cancer cells using the same method. These results demonstrate our expertise to develop robust, clinically compatible LATE-PCR endpoint assays for LOH detection in pre-malignant Barrett’s esophagus epithelia.
LATE-PCR is an advanced form of asymmetric PCR that uses unequal concentrations of specially designed primers to efficiently generate single-stranded DNA products under controlled conditions. LATE-PCR begins with an exponential phase of double-stranded DNA synthesis that is just as efficient as in symmetric PCR. When the limiting primer present at the lowest concentration is depleted, the reaction abruptly switches to a second phase of amplification during which large amounts of single-stranded DNA products accumulate with linear kinetics over many additional thermal cycles. Accumulated single-stranded DNA do not participate in the amplification process but are fully available to hybridize to specially designed mismatch-tolerant probes over a broad range of detection temperatures that are below the annealing temperature of the amplification reaction itself.
LATE-PCR endpoint assays are based on the ratio of fluorescence signals from a mismatch-tolerant probe collected at three specific temperatures. These temperatures are derived from the underlying thermal profiles of the mismatch tolerant probe binding to two target sequences homozygous for different SNP alleles, one perfectly complementary to the probe and the other partially complementary to the probe (see Figure 2).
Figure 2: Normalized thermal profiles of a mismatch-tolerant probe perfectly complementary to a SNP allele hybridized to targets that are either homozygous or heterozygous for said SNP allele (red and blue lines, respectively) or homozygous for the other SNP allele (green lines). The ratio of probe signals at middle and low temperatures normalized to the signals at the upper temperature reveals the percentage of targets with the interrogated SNP allele in the sample.
The upper temperature is too high for the probe to hybridize to either target and therefore provides a measure of background fluorescence. The middle temperature is derived from analysis of the temperature window of allele discrimination by the probe. Mismatch-tolerant probes hybridize to their perfectly complementary sequence over a higher range of temperatures than to partially-complementary sequences. The two melt curves together define a temperature window for allele discrimination by the probe (see red and green curves in Figure 2). Within that window, the desired middle temperature is the temperature at which there is the greatest difference between the amount of probe hybridized to the perfectly complementary and to the partially-complementary homozygous SNP allele target sequences. Heterozygous diploid samples are comprised of one perfectly-complementary and one partially-complementary target sequence and, therefore, fit a melt curve that, in most cases, lies virtually midway between the homozygous diploid alternatives (blue curves in Figure 2). Finally, the low temperature used for LATE-PCR endpoint assays corresponds to a temperature at which the mismatch-tolerant probe hybridizes to all SNP target sequences, including those partially-complementary to the probe (see Figure 2). The signals at the low and upper temperature are used to normalize all of the values between these two extremes. The normalized signal at the middle temperature can then be used to reveal whether SNPs that are known to be heterozygous near tumor suppressor genes in normal diploid genomes of a person are still heterozygous in pre-tumor genomes of that individual or have become hemizygous due to LOH.
The LATE-PCR LOH assays works by amplifying a particular heterozygous SNP sequence from normal and pre-malignant cells of the same individual using a single pair of LATE-PCR primers and then comparing the percentage of total amplified products that contain one of the two SNP alleles in each case based on the normalized fluorescent signal ratios described above (see Figure 2). If the pre-tumor genomes are still heterozygous at the interrogated SNP site, the mismatch-tolerant probe hybridizes to 50% of total amplified single-strands at the mid-temperature and to 100% of total single-strands at a lower temperature, just as it does in normal diploid cells. However, if the SNP site in question has become hemizygous due LOH, there are two possibilities. If the perfectly-complementary allele has been retained and the partially-complementary allele lost, the probe detects the same number of single-stranded molecules at both the mid-temperature and the lower temperature. In contrast, if the perfectly-complementary allele is lost and the partially-complementary allele retained, the probe detects little or none of the single-stranded molecules at the mid-temperature but all amplified single-stranded molecules at the lower temperature. The ratio for the normalized signals at the mid-temperature shifts up or down from the heterozygous value in accord with which allele has been lost due to LOH (see Figure 3).
Figure 3: The fluorescent signal ratios shift up or down from the heterozygous ratio value in accord with which allele has been lost due to LOH.
Experimental demonstration of identification of heterozygous SNP sites in normal diploid cells using LATE-PCR endpoint assays
Detection of LOH involving tumor suppressor genes in BE epithelial cells using LATE-PCR endpoint assays requires identification of heterozygous SNP sites in normal diploid cells of the BE patient, followed by comparison of the fluorescent ratios generated by LATE-PCR and mismatch-tolerant probes at these heterozygous SNP sites in normal and BE tissue. The example shown here illustrates the use of LATE-PCR endpoint genotyping to identify specific heterozygous SNP sites in the vicinity of the tumor suppressor p53 gene locus on human chromosome 17 in normal diploid cells from BE patients. The particular SNP sites examined (rs2270517 and rs858521) consist of either a C allele and a T allele or a C allele and a G allele, respectively, and were chosen for their high heterozygosity index (>40%) in Caucasian populations. For this pilot study, the Reid group provided us with replicate groups of one hundred normal diploid gastric epithelial cells and replicate groups of one hundred BE epithelial cells from six different individuals known to exhibit p53{17p} LOH in their Barrett’s epithelium. This material was derived from the biorepository at the Seattle Barrett’s Esophagus Study Group (CA91955, Brian Reid, MD PhD, PI). The genotype of the examined SNP sites and the p53{17p} LOH status had previously been established by the Seattle group but was blinded to us. Accordingly, we constructed sets of LATE-PCR primers that generated separate amplicons containing each of these SNP sites as well as mismatched-tolerant probes perfectly complementary to the C allele of each of these SNPs. Once the cells were placed in PCR tubes, the samples were lysed with QuantiLyse (a lysis buffer compatible with LATE-PCR, see below), and amplified with LATE-PCR primer/mismatched tolerant probe sets. All these steps were performed by serial dilution of reagents in the same PCR tube. Following LATE-PCR amplification, the tubes were cooled down and fluorescent readings from the hybridization probe were collected at 40°C, 52°C and 60°C (for the rs858521 SNP) or at 45°C, 57°C, and 71°C (for the rs2270517 SNP) as the mismatch- tolerant probes melted off their target. The resulting data were then used to calculate normalized fluorescent signal ratios. These specific temperatures were derived from previous analysis of the binding profile of mismatch-tolerant probes against synthetic matched and mismatched SNP allele targets at all temperatures. Replicate control samples (n=26) of known homozygous CC, heterozygous CG, and homozygous GG genotypes for the rs858521 SNP, or of known homozygous CC, heterozygous CT, and homozygous TT genotypes for the rs2270517 SNP were also analyzed to determine the range of normalized fluorescence ratios corresponding to each genotype. Figure 4 shows that the fluorescent signal ratios obtained after three-temperature normalization for each of the patient samples fell specifically within one of the fluorescent ratios ranges defined by each set of genotype controls (only rs2270517 SNP data is shown). The boxes in Figure 4 correspond to three-standard deviations of the range of fluorescence ratios that define the 99.7% confidence range for each genotype. There was 100% concordance in the genotype assignment when the genotype of the coded samples was revealed. The LATE-PCR endpoint assay was completed in less than 2 hours. The assay is robust because the affinity of mismatch-tolerant probes for their targets at different temperatures is an intrinsic thermodynamic property of the probe that is independent of the amount of target under conditions of probe excess. As a result, the normalized fluorescent ratios are independent of how many genomes were actually present in each sample at the start of the reaction and of the extent of LATE-PCR amplification past the point where amplification products are first detected (not shown).
Figure 4: Genotyping of specific SNPs in BE patients using LATE-PCR
Experimental demonstration of LOH detection using endpoint LATE-PCR assays
Based on the above confirmation of our assay, we moved to analysis of fluorescent ratios generated by LATE-PCR endpoint assays of pre-malignant BE genomes of the same patients. Replicates of 100 purified gastric cells and 100 purified BE epithelial cells were processed in QuantiLyse and amplified using the same LATE-PCR primers, as described above. At endpoint, fluorescent signals were read at the same three temperatures as above and then used to calculate fluorescence signal ratios. Figure 5 shows that the normalized fluorescent ratios from gastric cells indicated heterozygosity, as expected, and that the ratio of fluorescent signals from BE cells shift up from the heterozygous ratio value consistent with the loss of the non-interrogated SNP allele due to LOH, in accord with the LOH status of the samples provided to us by the Reid laboratory (only rs858521 SNP results shown). Once again, the boxes in Figure 5 correspond to three-standard deviations of the distribution of normalized fluorescence ratios that define the 99.7% confidence range for each genotype. These results also demonstrate that Quantilyse effectively lyses BE epithelial cells directly in the tube where PCR is done without disrupting amplification.
Figure 5: LOH detection using LATE-PCR
All of the BE epithelial samples in the experiment above were prepared by the traditional flow cytometry method, which is considered less suitable for clinically application. Figure 6 shows a LOH detection analysis in previously screened LOH BE cells for the rs3900787 SNP near the p16 gene in chromosome 9. In this case, the BE epithelial samples were prepared by the Reid laboratory different clinically-compatible protocols for epithelial sample preparation. Comparison of LATE-PCR fluorescent ratios derived from these samples and from normal diploid cells of the same individual clearly identified LOH regardless of which method was used to prepare the BE epithelial cells. These results demonstrate the validity of the more clinically compatible cytologic touch preparation and EDTA methods of sample preparation.
Figure 6: LATE-PCR LOH analysis of BE epithelial cells prepared by various methods (see text for details)
Optimization of LATE-PCR endpoint assays for LOH detection in samples containing mixtures of neoplastic and normal cells
Detection of LOH genomes can be confounded by the presence of normal diploid genomes from surrounding stromal tissue. To quantify the resolving power of LATE-PCR endpoint LOH assays, we analyzed artificially constructed mixtures of DNA homozygous for the A and C alleles of the rs3900787 SNP which resulted in different proportions of these two SNP alleles. Genomes heterozygous for the rs3900787 SNP (50% C, 50% A alleles) served as a point of reference. The mixtures consisted of 53% C and 47% A alleles (3% allele imbalance), 55.5% C and 44.4% A alleles (5.5% allele imbalance), and 60% C and 40% A alleles (10% allele imbalance). These allele imbalances corresponded to the presence of 88.8%, 80%, and 66.6% contaminating normal cells, or 8-fold, 4-fold, and 2-fold excess normal cell contamination of neoplastic cells, respectively. To improve the chances of detecting the smallest allele imbalances, we further optimized the design of the mismatched-tolerant probe. The improved probes readily detect allele imbalances as little as 3% (the equivalent to 88.8% normal cells to 11.2% LOH-positive cells, or 8-fold contamination of LOH-positive cells with normal diploid cells, see Figure 7). These data provide evidence that LATE-PCR assays can be employed for quantitative LOH detection.
Figure 7: Optimized design of mis-match tolerant probes permits LOH detection in the presence of an 8-fold excess normal diploid cells (see text for details). Boxes represent three standard deviations of each distribution.
Advantages of LATE-PCR endpoint assays
LATE-PCR endpoint assays are remarkably robust and powerful because fluorescence ratios indicative of genotype are independent of the starting amount of DNA present in the sample or the extent of product accumulation past the Ct value. As a result, there is no need to adjust the starting amounts of materials used in the assay or to use a fixed endpoint for analysis.
Fluorescence ratios generated by Two-Temperature LATE-PCR genotyping are independent of starting genome copy number. Fluorescence ratios generated by Two-Temperature LATE-PCR genotyping reflect solely the intrinsic thermodynamic properties governing the binding of the fluorescent probe to its matched and mismatched targets at two different temperatures as well as the quenching effect of temperature on fluorescence intensity. As a result, fluorescence ratios indicative of genotype should be independent of the number of starting genome copies in the amplification reaction. To test this hypothesis, Figure 8 compares the fluorescence signal ratios obtained in the course of LATE-PCR amplification for two replicate sets of samples homozygous for the C allele of the rs858521, a SNP site located in the vicinity of the human tumor suppressor gene p53. These samples consisted of either 2000 genome equivalents or 100 genome equivalents of genomic DNA at the start of the reaction. Figure 4 shows that these two sets of samples begin amplification and reach the same fluorescence ratio shortly after each set of samples reached its own Ct value. A single endpoint reading taken at any point after cycle 28 in Figure 4 identifies these samples as having identical genotypes (i.e., identical two-temperature fluorescence ratios) despite their differences in starting copy number. We conclude that LATE-PCR endpoint genotyping is robust because it allows for endpoint analysis independently of the starting number of genomes present in the sample.
Figure 8: Fluorescent ratios are independent of the number of starting DNA targets. Plot of fluorescent ratio signals at each cycle of a LATE-PCR reaction that amplifies a homozygous rs858521 [C/G] SNP site near the human p53 gene. The reaction was monitored with a mismatch-tolerant probe perfectly complementary to the rs858521 C allele. Red lines, homozygous CC replicates containing 2000 starting genomes; red dotted lines, homozygous CC replicates containing 100 starting genomes. Both sets of replicate homozygous samples generate the same two-temperature fluorescence ratio even though these samples differ in their amount of starting material.
Fluorescence ratios generated by LATE-PCR endpoint genotyping are independent of the extent of the amplification reaction. If fluorescence ratios indicative of genotype are independent of the number of starting genome copies in the amplification reaction, then these ratios should also be independent of the amplification cycle chosen for endpoint fluorescent signal ration analysis provided that the amplification reaction has progressed past the Ct value. To test this hypothesis, Figure 9, plots the fluorescent ratio signals during the course of a LATE-PCR amplification of replicate homozygous and heterozygous samples for the rs2270517 [C/T] SNP site, also near the human p53 gene. Once each sample begins amplification and reaches its own Ct value, the fluorescence values characteristic for each sample remain constant for the remaining of the amplification reaction. We conclude that LATE-PCR endpoint genotyping is robust because endpoint analysis can be performed at any cycle during LATE-PCR amplification past the thermal cycle where products are first detected.
Figure 9: Fluorescence ratios generated by LATE-PCR endpoint genotyping are independent of the extent of the amplification reaction (see text for details).
Conclusion
The LATE-PCR endpoint strategy outlined here is a new approach to endpoint genotyping and LOH analysis. The rapid detection, single-tube format, and compatibility with other potential biomarker assays (mutations, methylation changes) make LATE-PCR endpoint assays a clinically viable screen for LOH. Thanks to improved methods of epithelial cell preparation, Quantilyse, LATE-PCR, and our innovative strategy for normalization of end-point fluorescent signals, epithelial cells can be conveniently prepared from BE biopsies, lysed, amplified by LATE-PCR, and interrogated for LOH events all within the same PCR tube, without any additional amplicon purification steps or operator assistance. These features should greatly facilitate the implementation of future clinical assays for a panel of biomarker that is highly predictive of the development of adenocarcinoma in Barrett’s esophagus and promote studies on other cancers where LOH may be a component of a predictive biomarker panel.
|
