Review
`
`For reprint orders, please contact reprints@future-drugs.com
`
`Principle and applications of
`digital PCR
`Gudrun Pohl and Ie-Ming Shih†
`
`Digital PCR represents an example of the power of PCR and provides unprecedented
`opportunities for molecular genetic analysis in cancer. The technique is to amplify a
`single DNA template from minimally diluted samples, therefore generating amplicons
`that are exclusively derived from one template and can be detected with different
`fluorophores or sequencing to discriminate different alleles (e.g., wild type vs. mutant or
`paternal vs. maternal alleles). Thus, digital PCR transforms the exponential, analog signals
`obtained from conventional PCR to linear, digital signals, allowing statistical analysis of
`the PCR product. Digital PCR has been applied in quantification of mutant alleles and
`detection of allelic imbalance in clinical specimens, providing a promising molecular
`diagnostic tool for cancer detection. The scope of this article is to review the principles of
`digital PCR and its practical applications in cancer research and in the molecular
`diagnosis of cancer.
`
`Expert Rev. Mol. Diagn. 4(1), 41–47 (2004)
`
`Digital PCR was first developed by Vogelstein
`and Kinzler to extend the applications of con-
`ventional PCR [1]. This technology is based on
`applying optimal PCR conditions to amplify a
`single template and is followed by detection of
`sequence-specific PCR products (alleles) for
`allelic counting. Digital PCR has proven useful
`in detecting rare mutations in a bulk of wild
`type sequences and in assessing allelic imbal-
`ance in tumor tissue and in plasma DNA sam-
`ples. Therefore, this article will primarily focus
`on reviewing the principles of digital PCR and
`its applications in mutational analysis and
`assessment of allelic imbalance (TABLE 1).
`
`Principles of digital PCR
`The principle of digital PCR is illustrated in
`FIGURES 1 & 2. This new experimental approach
`involves two components [1]. First, the DNA
`to be analyzed is diluted into multi-well plates
`with one template molecule per two wells (on
`average) and PCR is performed in optimal
`conditions designed to amplify a single copy of
`PCR template. The amplicons are hybridized
`with fluorescence probes, such as molecular bea-
`cons, that allow detection of sequence-specific
`products using different fluorophores. Thus,
`
`digital PCR is employed to directly count, one
`by one, the number of each of the two (paternal
`vs. maternal or wild type vs. mutant) alleles in
`the samples. Second, several statistical analyses,
`including Bayesian-type likelihood methods, can
`be applied to measure the strength of the evi-
`dence for the allele distribution being different
`from normal [1]. This approach imparts a rigo-
`rous statistical basis to analyze allelic status and
`is expected to provide more reliable information
`than heretofore possible in allelic studies of tissue
`or body fluid samples. Therefore, digital PCR
`transforms the exponential and analog signals of
`conventional PCR to linear and digital signals.
`To perform digital PCR, genomic DNA
`samples from tissue or body fluid are diluted in
`384-well PCR plates so that there will be, on
`average, approximately 0.5 template mole-
`cules (genomic equivalent) per well. The opti-
`mal dilution of DNA samples can be achieved
`by DNA quantification kits to determine the
`amount of genomic equivalents in the original
`samples. As the PCR products from the ampli-
`fication of single template molecules are homo-
`geneous in sequence, a variety of conventional
`techniques could be used to assess their presence.
`Fluorescent probe-based reagents, which can be
`
`CONTENTS
`Principles of digital PCR
`Applications of digital PCR
`Expert opinion
`Five-year view
`Key issues
`References
`Affiliations
`
`†Author for correspondence
`Department of Pathology,,
`418 North Bond Street, B-315,
`Baltimore, MD 21231, USA
`Tel.: +1 410 502 7774
`Fax: +1 410 502 7943
`ishih@jhmi.edu
`
`KEYWORDS:
`cancer, detection, diagnosis,
`digital PCR
`
`www.future-drugs.com
`
`© Future Drugs Ltd. All rights reserved. ISSN 1473-7159
`
`41
`
`The Johns Hopkins University Exhibit JHU2003 - Page 1 of 7
`
`

`

`Pohl & Shih
`
`Table 1. Examples using digital PCR for molecular analysis in clinical samples.
`Application
`Findings
`Detection of KRAS mutations in stool
`KRAS mutations can be detected and quantified in stool DNA samples
`from colorectal cancer patients
`
`Detection of KRAS mutational status
`
`Low- and high-grade ovarian serous carcinoma develop through
`independent pathways
`
`Analysis of KRAS and AI in APC genes
`
`Mutations in KRAS and AI of APC occur in appendiceal adenomas
`
`Detection of AI and KRAS mutations
`
`Detection of LOH in APC locus
`
`High-grade ovarian serous carcinoma contain wild type KRAS and a high
`frequency of AI, even in small primary tumors
`
`Development of adenomatous polyps may proceed through a
`top-down mechanism
`
`Detection of AI of chromosomes 1p, 8p, 15q and 18q
`
`Evidence of AI occurs in early colorectal tumors
`
`Detection of AI of chromosome 18q
`
`AI of 18q is associated with vascular invasion in colorectal carcinomas
`
`Detection of AI of chromosomes 8p and 18q
`
`AI of 8p and 18q is a better predictor of prognosis than histopathological
`stage in colorectal cancer patients without metastasis
`
`Detection of AI using 8 SNP markers with high
`frequency of allelic loss in ovarian cancer
`
`AI can be detected with high specificity and sensitivity in plasma DNA
`samples from patients with ovarian cancer
`
`Detection of AI using 7 SNP markers in ovarian,
`colorectal and pancreatic cancers
`
`Detection of AI can be a useful adjunct for the detection of cancer in
`ascitic fluid
`
`Detection of BAT26 alterations in fecal DNA
`
`Quantitative detection of APC gene expression
`
`Presence of BAT26 mutations in fecal DNA provides a promising marker for
`colorectal cancer screening
`
`Small changes in expression of APC gene affect predisposition to familial
`polyposis coli
`
`AI: Allelic imbalance; APC: Adenomatous polyposis coli; LOH: Loss of heterozygosity; SNP: Single nucleotide polymorphism.
`
`Ref.
`[1]
`
`[18]
`
`[12]
`
`[17]
`
`[37]
`
`[6]
`
`[5]
`
`[19]
`
`[31]
`
`[32]
`
`[34]
`
`[35]
`
`used directly on the PCR products in the same wells, are particu-
`larly well suited for this purpose [2]. Currently, molecular
`beacons are extensively used to detect the PCR products in
`digital PCR assays [3]. For mutational analysis, a pair of molecu-
`lar beacons is designed with one hybridizing to the wild type
`sequence that harbors the mutation and the other hybridizing to
`the neighboring sequence (FIGURE 1). Therefore, the mutational
`status of a specific allele in a well is determined by the ratio of
`fluorescence intensity of the two beacons in that particular well.
`As multiple wells are counted, digital PCR can be used to detect
`mutations present at relatively low levels in the samples to be
`analyzed. The sensitivity of mutation detection depends on the
`number of wells that are included for analysis and the intrinsic
`mutation rate of the polymerase used for amplification. For
`assessing allelic imbalance, single nucleotide polymorphisms
`(SNPs) are used to represent the paternal or maternal alleles. A
`pair of PCR primers and a pair of molecular beacons are
`designed for each SNP (FIGURE 2). Digital PCR is performed
`using a SNP marker for which the patient is heterozygous. The
`resultant PCR products are then analyzed using molecular
`beacon probes to determine allelic representation. The mecha-
`nism of how molecular beacons discriminate between maternal
`and paternal alleles is briefly summarized. Molecular beacons are
`single-stranded oligonucleotides which contain a fluorescent dye
`and a quencher on their 5´ and 3´ ends, respectively (FIGURE 1).
`
`Both beacons are identical except for the nucleotide corre-
`sponding to the SNP and the fluorescent label (green or red).
`Molecular beacons include a hairpin structure, which brings the
`fluorophore closer to the quencher, and do not emit fluorescence
`when not hybridized to a PCR product [4]. Upon hybridization
`to their complimentary nucleotide sequences, the quencher is
`distanced from the fluorophore, resulting in increased fluores-
`cence. Therefore, the ratio of fluorescence intensity of two allele-
`specific beacons with either green or red fluorescence is calcu-
`lated to determine the allele type in one PCR reaction (well).
`With hundreds or thousands of wells (reactions) counted, the
`percentage of mutant alleles or the ratio of maternal and paternal
`alleles can be determined. For allelic status, a rigorous statistical
`method is then used to conclude whether allelic imbalance is
`present in the background of normal DNA [5,6].
`
`Applications of digital PCR
`Mutational analysis
`For a variety of basic research and clinical applications, the identi-
`fication of rare mutations is very important. Analysis of the early
`effects in tumorigenesis often depends on the ability to detect
`small populations of mutant cells [7,8]. Reliable technology to
`demonstrate the presence of mutations in clinical specimens
`holds great promise for cancer detection, as mutations represent
`a molecular genetic hallmark of neoplastic diseases.
`
`42
`
`Expert Rev. Mol. Diagn. 4(1), (2004)
`
`The Johns Hopkins University Exhibit JHU2003 - Page 2 of 7
`
`

`

`Digital PCR
`
`In another study, digital PCR has been used to identify KRAS
`mutations in paraffin tissues of appendiceal mucinous adeno-
`mas in identical twins [12]. One of the twins suffered from a rare
`disease called pseudomyxoma peritonei (PMP), which produces
`an overwhelming amount of mucin in the intra-abdominal
`cavity as a result of the rupture of the appendiceal mucinous
`tumor. As the mucinous adenoma is a single layer of neoplastic
`cells embedded in abundant stromal cells and mucin, tradi-
`tional methods, such as direct nucleotide sequencing, may not
`be sensitive enough to detect KRAS mutations, even when laser
`capture microdissection is employed to enrich the tumor cell
`population. In this study, the tumor tissue on paraffin sections
`was dissected under an inverted microscope and genomic DNA
`was purified and subjected to digital PCR. The study demon-
`strated that identical KRAS mutations were detected in the
`appendiceal adenoma and peritoneal tumor from the twin with
`PMP, whereas the adenoma from the other twin harbored a dif-
`ferent mutation. The KRAS mutational analysis supported the
`view of the authors that PMP is clonally derived from the asso-
`ciated appendiceal mucinous adenoma. The different types of
`mutations in KRAS in the tumors from both siblings suggested
`that mutation in KRAS occurs somatically in adenomas and is
`independent of the identical genetic background of the twins.
`
`Quencher
`
`R1
`
`R1
`
`To address whether digital PCR is useful for mutation detec-
`tion in cancer, Vogelstein and Kinzler have analyzed the DNA
`from stool specimens in patients with colorectal cancer [1]. Their
`study focused on the KRAS gene mutation, which is a frequent
`molecular genetic event in colorectal cancer [9,10]. As the stool
`DNA is pool DNA released from a mixed-cell population
`including both normal and tumor cells, approximately 1–10%
`of the KRAS genes purified from stool contained mutant alleles
`[11]. Therefore, digital PCR appears a well-suited technique to
`assess the presence of mutated KRAS gene in stool. A 384-well
`digital PCR experiment was established to include positive con-
`trols (48 of the wells contained 25-genome equivalents of DNA
`from normal cells) and negative controls (48 wells without
`DNA template). The other 288 wells contained an appropriate
`dilution of stool DNA. In this study, molecular beacon red fluo-
`rescence indicated that 102 of these 288 experimental wells con-
`tained PCR products, whereas the other 186 wells did not. The
`red/green ratios of the 102 positive wells suggested that five con-
`tained mutant KRAS alleles. To determine the nature of the
`mutant KRAS genes from stool in the five positive wells, the
`PCR products were sequenced directly to reveal Gly12Ala muta-
`tions (GGT to GCT at codon 12) in four of them, whereas the
`sequence of the other indicated a silent C>T transition at the
`third position of codon 13. This transition presumably resulted
`from a PCR error during the first productive cycle of amplifica-
`tion from a wild type template. Thus, approximately 4%
`(4/102) of the KRAS alleles present in this stool sample con-
`tained a Gly12Ala mutation. The mutant alleles in the stool pre-
`sumably arose from the colorectal cancer of the patient, as direct
`sequencing of PCR products generated from DNA of the cancer
`identified the identical Gly12Ala mutation [1].
`
`A
`
`F1
`
`B
`
`F1
`
`Wild type KRAS
`
`Mutant KRAS
`
`Codon 12/13
`
`X
`Codon 12/13
`
`Assessing allelic imbalance in tissue
`Genetic instability is a molecular signature of most human
`cancers [13] and at the molecular level is characterized by
`allelic imbalance (AI), representing losses or gains of defined
`chromosomal regions. Analysis of AI is important in elucidating
`the molecular basis in the development of cancer. There are,
`however, at least two major problems associated with the current
`methods for assessing AI in tissue sections
`using microsatellite markers. First, DNA
`purified from microdissected tissues is
`always a mixture of neoplastic and non-
`neoplastic DNA and the latter, released
`from non-neoplastic cells, can mask AI
`because it is difficult to quantify the allelic
`ratio using microsatellite markers. Second,
`such DNA is often degraded to a variable
`extent, producing artifactual enrichment
`of smaller alleles when microsatellite
`markers are used for analysis [14]. Thus,
`digital PCR promises to overcome these
`technical difficulties associated with the
`molecular genetic analysis of AI in which
`the paternal or maternal alleles within a
`plasma DNA sample are
`individually
`counted, thereby allowing a quantitative
`measure of such imbalance in the presence
`of normal DNA. Statistical methods are
`used to evaluate the strength of the evi-
`dence for loss of heterozygosity in each
`tumor sample. Currently, the sequential
`probability ratio test (SPRT) is used to
`
`Figure 1. Design of molecular beacons and PCR primers for digital PCR analysis to detect KRAS
`mutations. Forward (F1) and reverse (R1) primers amplify exon 1 of the KRAS gene containing the
`relevant codons 12 and 13. Asymmetrical amplification generates single-stranded DNA complementary
`to the molecular beacons by using excessive R1 primer. The green beacon recognizes the common
`sequence in both wild type and mutant, while the red beacon only recognizes wild type sequence
`containing codons 12 and 13. Therefore, both red and green fluorescence is detected in wild type DNA
`but predominant green fluorescence is detected in PCR products with mutations at and around codons
`12 and 13 as a result of mismatched hybridization.
`
`www.future-drugs.com
`
`43
`
`The Johns Hopkins University Exhibit JHU2003 - Page 3 of 7
`
`

`

`Pohl & Shih
`
`conclude the presence of AI in tumor tissues [15]. SPRT allows
`two probabilistic hypotheses to be compared as data accumu-
`late and, on average, guarantees a smaller amount of testing for
`a given level of confidence than any other method. Hypothesis
`one is that a sample has no loss of heterozygosity (LOH), that
`is, the tumor cells have the same proportion of alleles as normal
`cells. This corresponds to p = 50%, where p is the proportion
`of either allele in the overall sample. Hypothesis two is that the
`same one of the two alleles is absent in every tumor cell. This
`does not correspond to an allelic proportion of 100% in the
`
`A
`
`F1
`
`Digital PCR
`
`Hybridization with molecular beacon
`
`R1
`
`Complementary target
`
`T A
`
`C G
`
`Tumor tissue
`
`Noncomplementary or no template
`
`SNP
`
`Fluorophore
`
`Quencher
`
`B
`
`Normal tissue
`
`TA
`
`CG
`
`TA
`
`CG
`
`Allelic balance
`green:red = 32:34
`
`tested sample because isolation of pure tumor cell populations
`from human tumors is almost impossible for routine samples.
`Given the conservative assumption that at least 50% of the
`DNA from the microdissected samples originated from neo-
`plastic rather than non-neoplastic cells, the hypothesis of LOH
`corresponds to the probabilistic hypothesis that the observed
`proportion would be at least 66.7%. A SPRT is therefore
`constructed to choose between the hypotheses p = 50% and
`p = 66.7%, with a threshold likelihood ratio of 8. Generally, for
`each case, the number of alleles studied in each sample is plotted
`on the abscissa and the ratio of wells con-
`taining the allele with the higher counts to
`total number of wells containing either
`allele is plotted on the ordinate. Samples
`represented by points above curve one are
`interpreted to have allelic loss, meaning
`that the likelihood ratio for p = 66.7%
`versus p = 50% exceeds 8. The samples
`below the bottom curve are categorized as
`having no LOH. Indeed, it has been
`shown that AI can be demonstrated in a
`much higher percentage of colorectal car-
`cinomas using digital SNP analysis than
`the traditional method using microsatellite
`markers [5,16].
`Digital PCR has been used to charac-
`terize AI in small colorectal adenomas [6].
`The investigators analyzed the allelic status
`in a total of 32 adenomas with an average
`size of 2 mm (range 1 to 3 mm). AI of
`chromosome 5q markers occurred in
`55% of tumors analyzed, consistent with
`a gatekeeping role of the adenomatous
`polyposis coli (APC) tumor suppressor
`gene located at chromosomal position
`5q21. AI was also detected in each of the
`other four chromosomes tested. The frac-
`tion of adenomas with AI of chromosomes
`1p, 8p, 15q and 18q was 10, 19, 28 and
`28%, respectively. Over 90% of the
`tumors exhibited AI of at least one chro-
`mosome and 67% had AI of a chromo-
`some other than 5q. These findings dem-
`onstrate that AI is a common event, even
`in very small tumors, and led the authors
`to conclude that chromosomal insta-
`bility occurs very early during colorectal
`neoplasia [6].
`In another study, Singer and coworkers
`applied digital PCR to assess the pattern
`of AI during tumor progression in ovar-
`ian cancer [17]. This study demonstrated
`that a progressive increase in the degree of
`AI of chromosomes 1p, 5q, 8p, 18q, 22q
`and Xp was observed comparing serous
`
`Allelic imbalance
`green:red = 62:14
`
`Figure 2. Digital PCR analysis to assess allelic imbalance. (A) Molecular beacon design. A pair of
`primers are designed to amplify approximately 100 bp of PCR product that contains a single nucleotide
`polymorphism (SNP) marker in the center region. The two beacons used for analysis of a specific SNP
`are identical except for the base pair corresponding to the SNP and the fluorescence label. Green and
`red represent fluorescein and hex labels, respectively. The molecular beacons do not emit fluorescence
`when not hybridized to a PCR product, as the 3´-dabcyl group (open circle) quenches the signals. Upon
`hybridization to their complimentary sequences, the quencher is distanced from the fluorophore,
`resulting in increased fluorescence. (B) Schematic illustration of a digital PCR analysis of the previous
`format. DNA from samples is distributed to a 384-well PCR plate. After completion of the PCR,
`molecular beacons are added to each reaction to determine allele status. Modified protocol has been
`used by combining digital PCR and allelic determination of molecular beacons in a single step [17].
`
`44
`
`Expert Rev. Mol. Diagn. 4(1), (2004)
`
`The Johns Hopkins University Exhibit JHU2003 - Page 4 of 7
`
`

`

`Digital PCR
`
`borderline tumors to noninvasive and invasive micropapillary
`serous carcinomas (low-grade serous carcinomas). In contrast,
`high-grade (conventional serous carcinoma) tumors had a high
`frequency of AI, even in small (early) primary tumors, similar to
`that found in advanced-stage tumors. Based on these findings,
`together with mutational analysis of BRAF and KRAS genes
`[17,18], Singer and coworkers proposed a dualistic model for
`ovarian serous carcinogenesis. One pathway involves a stepwise
`progression from serous borderline tumor to noninvasive and
`then invasive micropapillary (low-grade) serous carcinoma. The
`other pathway is characterized by rapid progression from the
`ovarian surface epithelium or inclusion cysts to a conventional
`(high-grade) serous carcinoma.
`Zhou and coworkers have applied digital PCR to count
`alleles and use the presence of AI-specific chromosomal
`regions to predict recurrence of early-stage colorectal cancer
`[19]. They studied 180 colorectal cancer patients with no evi-
`dence of lymph node metastases or distant metastases at the
`time of surgery, and looked for the presence of AI on chromo-
`some 8p and 18q in these tissue samples. They divided tumors
`into three groups: L tumors (n = 93) had AIs of chromosomes
`8p and 18q, L/R tumors (n = 60) had AIs of either chromo-
`some 8p or 18q but not both, and R tumors (n = 27) retained
`allelic balance for both chromosomes. Five-year disease-free
`survival was 100% for patients with R tumors, 74% for
`patients with L/R tumors and 58% for those with L tumors.
`These differences were significant and independent of other
`variables. The authors concluded that in patients without
`metastasis, AI was found to be a better predictor of prognosis
`than histopathologic staging.
`
`Cancer detection in body fluid DNA
`It is well recognized that tumors release a significant amount
`of genomic DNA into the systemic circulation, probably
`through cellular necrosis and apoptosis [20–22]. This tumor-
`derived DNA can be detected as a result of specific genetic
`and epigenetic alterations in the tumors, such as microsatellite
`alterations, translocations, mutations and aberrant methy-
`lation. As previously described, genetic instability is a defining
`molecular signature of most human cancers [13,23] and at the
`molecular level, it is characterized by AI, representing losses
`or gains of defined chromosomal regions. Thus, analysis of AI
`may also provide a molecular basis for cancer detection. Using
`microsatellite markers, AI has been demonstrated in the
`serum or plasma obtained from patients with lung [24], breast
`[25,26], renal [27] and ovarian cancers [28] and melanoma [29].
`Some of these were small, early-stage neoplasms at the time of
`diagnosis, suggesting that detection of AI in plasma is a pro-
`mising method for population-based screening [30]. Although
`these studies provide encouraging results, as with assessing AI
`in tissues, there are the two major problems of plasma DNA
`being a mixture of neoplastic and non-neoplastic DNA (so AI
`may be masked as it is difficult to quantify the allelic ratio
`using microsatellite markers) and DNA is degraded to a variable
`extent (producing artifactual enrichment of smaller alleles
`
`when microsatellite markers are used for analysis [14]). There-
`fore, detection of AI using digital PCR analysis may overcome
`these technical difficulties.
`Chang and coworkers have performed digital PCR analysis
`to determine allelic status in plasma DNA and to evaluate the
`potential of this new technology for cancer detection using
`plasma samples [31]. This study first analyzed DNA concentra-
`tion in plasma samples from 330 patients, including 122
`patients with various cancers, 164 control patients with non-
`neoplastic disease and 44 individuals without apparent dis-
`eases. The area under the receiver-operating characteristic
`(ROC) curve for plasma DNA concentration was 0.90 for
`neoplastic versus healthy patients and 0.74 for neoplastic versus
`non-neoplastic patients. Given 100% specificity, the highest
`sensitivity achieved was 57%. Of the 330 patients, digital PCR
`analysis was performed on 54 ovarian cancer patients and 31
`non-neoplastic disease controls. AI of at least one SNP in
`plasma DNA was found in 87% (95% confidence interval
`[CI]: 60–98%) of Stage I/II and 95% (95% CI: 83–99%) of
`Stage III/IV patients and none of 31 patients without neoplas-
`tic disease (specificity 100%, CI: 89–100%). For the 63
`patients with serum CA125 data, DNA plasma concentration
`added information to serum CA125 levels by increasing the
`area under the ROC curve from 0.78 to 0.84. CA125 is the
`most commonly used tumor marker in ovarian cancer patients.
`Thus, measurement of plasma DNA levels may not be sensi-
`tive or specific enough for use as a cancer screening or diagnostic
`tool, even in conjunction with CA125. However, detection of
`AI in plasma DNA using the digital SNP analysis holds great
`promise for the detection of cancer.
`Besides plasma DNA, Chang and coworkers also applied
`digital PCR analysis to AI in ascites fluid to assess the feasibility
`of this new technology in detecting malignant ascites [32]. Cyto-
`logical examination of ascitic fluid is critical for clinical man-
`agement in patients with peritoneal or pelvic diseases. Such
`morphologic examination can only achieve a sensitivity less
`than 62% and thus a molecular test that is able to distinguish
`benign versus malignant ascites could be clinically useful [33].
`With digital PCR analysis, AI in at least one SNP marker was
`found in 19 of 20 (95%) ascitic fluid DNA samples obtained
`from patients with cytologically proven carcinomas in ascitic
`fluid. In contrast, AI was detected in only one of 20 patients
`with negative cytology. This latter patient with AI in her ascites
`had known Stage III ovarian carcinoma at the time of cytology
`sampling. The ascitic specimen of this patient demonstrated
`presence of carcinoma cells in culture with an identical AI
`pattern found in the ascitic supernatant and surgical specimen.
`These findings suggest that detection of AI using digital SNP
`analysis can be a useful adjunct for the detection of ovarian and
`other types of cancer in ascitic fluid.
`
`Cancer detection in stool DNA
`In addition to KRAS mutations, Traverso and coworkers have
`applied digital PCR to examine the alteration of a microsatellite
`marker, BAT26, in stool DNA from patients with proximal
`
`www.future-drugs.com
`
`45
`
`The Johns Hopkins University Exhibit JHU2003 - Page 5 of 7
`
`

`

`Pohl & Shih
`
`cancers (located at the right colon) to determine the feasibility,
`sensitivity and specificity of this new approach [34]. Their study
`focused on patients with proximal cancers, reasoning that such
`cancers are difficult to detect since they are located most dis-
`tally from the anus when colonoscopy is performed. Stool
`DNA was purified and subjected to digital PCR. The PCR
`products were sequenced to determine the status of BAT26
`(alterations/mutations vs. wild type). They found that 18 of 46
`cancers had microsatellite alterations in BAT26 and that identi-
`cal mutations could be identified in the fecal DNA of 17 of
`these 18 cases. Among the cancer patients with proximal
`lesions, the clinical sensitivity of the BAT26 fecal DNA test was
`37%. In contrast, there were no positives among 69 individuals
`with normal colonoscopy findings or among 19 patients with
`adenomas. The specificity was therefore 100%. This study
`provides a promising new molecular diagnostic technique for
`colorectal cancer screening.
`
`Quantification of gene expression of specific alleles
`Yan and coworkers have recently applied digital PCR analysis
`to quantitatively measure gene expression of specific alleles
`(based on SNP) using cDNA as templates [35]. The principle is
`similar to that for genomic DNA. The study showed that con-
`stitutional 50% decreases in expression of one APC gene allele
`can lead to the development of familial adenomatous polyposis.
`Similarly, Pohl and coworkers have used the digital PCR assay
`to quantify the ratio of wild type and mutant BRAF gene
`expression in ovarian cancer [POHL G, UNPUBLISHED RESULTS]. This
`approach can provide a powerful and useful technique for
`assessing the success of (epi)genetic knockout of specific alleles
`(wild type or mutant), such as somatic knockout and siRNA.
`
`Expert opinion
`Digital PCR represents an example of the power of PCR and
`provides new opportunities for genetic analysis. This technique
`is especially powerful in experiments requiring the quantitative
`investigation of individual alleles in DNA samples isolated
`from a mixed-cell population.
`There are several advantages of digital PCR compared with
`other types of PCR-based molecular genetic analyses. First, as
`compared with microsatellite markers, the PCR products
`derived from the two SNP alleles at every locus are the same
`size and therefore their analysis is not biased by the preferen-
`tial DNA degradation of larger alleles. Second, the digital
`PCR approach, which amplifies single-allele templates in the
`PCR reaction, can precisely determine the number of alleles
`examined in each experiment. Accordingly, SNP genotyping
`is digital, involving the detection of the presence or absence of
`
`a specific allele, rather than analog, as microsatellite genotyping
`is, which measures the length of microsatellites [1]. Third, a
`statistical method such as SPRT can be employed to conclude
`whether AI is present in the background DNA.
`
`Five-year view
`Although the sensitivity of digital PCR analysis in the current
`384-well format is usually satisfactory, higher sensitivity would
`be desirable. Sensitivity could be improved by analyzing more
`wells in the assay, although such an approach may not turn out
`to be cost effective. New technologies are being developed to
`perform digital PCR without using multiwell formats. For
`example, an innovative technology called the BEAMing
`(beads, emulsion, amplification and magnetics) method has
`been introduced [36]. In this method, each DNA molecule in a
`sample is converted into a single magnetic particle to which
`thousands of copies of DNA with the same sequence to the
`original are bound. This population of beads then corre-
`sponds to a one-to-one representation of the original DNA
`molecules. Variation within the primary population of DNA
`molecules can then be simply assessed by counting fluores-
`cently labeled particles using flow cytometry. Therefore, mil-
`lions of individual DNA molecules can be assessed at the same
`time. BEAMing can be used for the identification and quanti-
`fication of rare mutations, as well as to study variations in
`gene sequences or transcripts in specific populations or tissues.
`Such innovative techniques for digital PCR are expected to
`emerge in the next few years, and they may provide another
`wave of new technologies to facilitate researchers in both basic
`and clinical science.
`
`Key issues
`
`(cid:127) The study of DNA sequence variation is important for many
`areas of research. Current PCR format does not allow the
`identification and quantification of rare molecular genetic
`changes because conventional PCR amplifies a pool of DNA
`templates from the starting material.
`(cid:127) Digital PCR is used to amplify a single DNA template from
`minimally diluted samples, therefore transforming the
`exponential, analog signals from conventional PCR to linear,
`digital signals, allowing statistical analysis of the PCR products.
`(cid:127) Digital PCR has been applied in the quantification of mutant
`alleles and detection of allelic imbalance in clinical
`specimens, providing a promising molecular diagnostic tool
`for cancer detection.
`
`References
`Papers of special note have been highlighted as:
`(cid:127) of interest
`(cid:127)(cid:127) of considerable interest
`Vogelstein B, Kinzler KW. Digital PCR.
`1
`Proc. Natl Acad. Sci. USA 96(16),
`9236–9241 (1999).
`
`• Original paper describing the principle of
`digital PCR.
`2 Whitcombe D, Newton CR,
`Little S. Advances in approaches to
`DNA-based diagnostics. Curr. Opin.
`Biotechnol. 9(6), 602–608
`(1998).
`
`3
`
`4
`
`Tyagi S, Kramer FR. Molecular beacons:
`probes that fluoresce upon hybridization.
`Nature Biotechnol. 14(3), 303–308 (1996).
`Tyagi S, Bratu DP, Kramer FR.
`Multicolor molecular beacons for allele
`discrimination. Nat. Biotechnol. 16(1),
`49–53 (1998).
`
`46
`
`Expert Rev. Mol. Diagn. 4(1), (2004)
`
`The Johns Hopkins University Exhibit JHU2003 - Page 6 of 7
`
`

`

`Zhou W, Galizia G, Goodman SN et al.
`Counting alleles reveals a connection
`between chromosome 18q loss and vascular
`invasion. Nature Biotechnol. 19(1), 78–81
`(2001).
`Shih IM, Zhou W, Goodman SN et al.
`Evidence that genetic instability occurs at
`an early stage of colorectal tumorigenesis.
`Cancer Res. 61(3), 818–822 (2001).
`(cid:127) Demonstration of the usefulness of digital
`PCR in basic cancer research.
`Kumar R, Sukumar S, Barbacid M.
`Activation of ras oncogenes preceding the
`onset of neoplasia. Science 248(4959),
`1101–1104 (1990).
`Jonason AS, Kunala S, Price GJ et al.
`Frequent clones of p53-mutated
`keratinocytes in normal human skin.
`Proc. Natl Acad. Sci. USA 93(24),
`14025–14029 (1996).
`Forrester K, Almoguera C, Han K,
`Grizzle WE, Perucho M. Detection of high
`incidence of K-ras oncogenes during human
`colon tumorigenesis. Nature 327(6120),
`298–303 (1987).
`10 Bos JL, Fearon ER, Hamilton SR et al.
`Prevalence of ras