| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Division of Molecular Pathology of the Department of Pathology,
*
and the Department of Oncology,
Johns Hopkins Hospital, Baltimore; and the Department of Biostatistics,
Johns Hopkins School of Hygiene and Public Health, Baltimore, Maryland
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In this paper, we present an alternate method of engraftment analysis using single nucleotide polymorphisms (SNPs). SNPs are pairs of alleles which vary at a single DNA basepair. They constitute 90% of all polymorphisms in the human genome, occurring at an average rate of 1/1000 bp.11 Unlike the repeating sequences of VNTRs and STRs, SNPs are non-repetitive sequences, and can be found in both coding and non-coding regions of the genome. Interest in SNPs has surged as a result of applications in the mapping of cancer and other disease genes, population genetics, and pharmacogenomics.11, 12, 13, 14 Databases, both public (National Center for Biotechnology Information) and private (pharmaceutical), have been established to catalog the growing number of identified SNPs.11
The purpose of our experiments was to develop a rapid, highly sensitive method of quantifying bone marrow engraftment in patients without the need for subjective interpretation of assay results or post-PCR analysis. We focused on SNPs because of their prevalence and suitability to quantification by fluorogenic oligonucleotide probes. We describe a novel technique using multiplex real-time PCR amplification of a well-characterized, common, biallelic SNP located within the cytochrome P450 CYP2C9 gene.15 We used an assay containing two oligonucleotide reporter probes (allele 1-specific and allele 2-specific) and a detection system that is linear over at least five orders of magnitude.16
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CYP2C9 Genotype Assessment by
PCR-AvaII-Polyacrylamide Gel Electrophoresis
To identify samples of known genotype, we performed CYP2C9
genotyping on 50 ng of patient DNA using previously reported PCR
primers and conditions,19
in a final reaction volume of 50
µl. Twenty-five microliters of this product were digested with 12
units of AvaII (New England BioLabs, Beverly, MA) for 8
hours at 37°C, followed by electrophoresis on an 8% acrylamide gel.
The primers identified by Stubbins et al,19
which we have
also used, generate a PCR product that does not contain a second,
nonpolymorphic AvaII restriction site. The diagnosis of an
allele 2 homozygote would, therefore, be susceptible to a false
positive in the event of restriction enzyme failure or inhibition.
Accordingly, if one were to employ the RFLP assay for clinical use, the
primers would have to be redesigned to incorporate such an internal
control cut site. To confirm the genotype determined by RFLP analysis,
undigested amplicons were also sequenced using a Big Dye sequencing kit
(PE Biosystems, Foster City, CA) and products analyzed on a PE
Biosystems ABI 310 Gene Sequencer.
Real-Time PCR SNP Analysis at the CYP2C9 Locus
SNP analysis was also performed using a real-time PCR CYP2C9
allelic discrimination TaqMan assay (PE Biosystems), with minor
modifications. The assay includes proprietary nonlabeled forward and
reverse primers along with two proprietary fluorescent TaqMan
oligonucleotide probes (allele 1-specific probe labeled with VIC
fluorophore, allele 2-specific probe labeled with FAM
(6-carboxy-fluorescein fluorophore). The VIC and FAM reporter dyes are
covalently attached to the 5' terminal base of the two probes and the
nonfluorescent quencher dye is attached near the 3' ends. The probes
are capable of differentially binding to the amplicons generated during
PCR and selectively reporting their respective alleles.20
All PCR reactions were run in triplicate, and contained 100 ng of
patient DNA, 12.5 µl of TaqMan Universal PCR Master Mix, and 2.5 µl
of Allelic Discrimination Mix. Appropriate negative controls were also
run. Real-time PCR was performed on an ABI Prism 7700 Sequence
Detection System (SDS, PE Biosystems) using the following conditions:
50°C for 2 minutes, 95°C for 10 minutes, and then 40 cycles of
amplification (92°C denaturation for 15 seconds, 62°C
annealing/extension for 60 seconds). The annealing temperature was
empirically determined (data not shown) to promote high binding
specificity of the probes without loss of assay sensitivity. For each
cycle, the SDS software determines the 
Rn, which is the
normalized (ie, compared to a passive reference fluorophore)
fluorescent signal from the VIC- or FAM-labeled probe. For our analysis
we used the Rn value after the final cycle, because it proved more
reliable than the cycle at which the threshold was crossed
(CT value, data not shown). The
RnFAM/RnVIC
fluorophore signal ratio (hereafter abbreviated FAM/VIC) was calculated
for each individual real-time PCR reaction, and means with standard
deviations were determined for the triplicate runs.
RFLP-Southern Analysis
Three micrograms of DNA were digested with HinfI (New
England BioLabs), run on an agarose gel, transferred to a Zeta-Probe
nylon membrane (Biorad, Hercules, CA) and hybridized with a
32P-labeled D1S7 (MS-1) minisatellite
probe21
using standard techniques.22
The
percentage of each patient’s DNA was determined using a
Molecular Dynamics (Sunnyvale, CA) 445SI PhosphorImager.
Statistical Calculation of Loci Informativity
Two methods for calculating informativity are described (Table 1)
. In method A, only homozygotes are considered informative, whereas
with method B, both homozygotes and heterozygotes are considered
informative. For initial calculations, we assumed a theoretical optimal
SNP allele frequency of 50% in a population (p = q = 0.5).
For method A, the probability that a given donor is homozygous for the
p allele is 0.25 (p2 =
0.502) from the Hardy-Weinberg relationship
(p2 + 2pq + q2 = 1). The
probability that the recipient is homozygous for the q allele is
0.25, yielding a combined probability of 0.0625. Informativity is also
obtained if the donor is qq and the recipient is pp
(probability also equal to 0.0625). Therefore when only
homozygosity is considered informative, the total informativity of this
SNP is twice that value, 0.125 or 12.5%.
|
. The relative frequencies of each informative
genotype are calculated with the same optimal SNP. Together they
provide an overall informativity of 62.5%, when the allele frequency
of p is 0.5. Similar calculations using methods A and B were repeated
for lower p allele (and correspondingly higher q allele) frequencies.
Assuming an informativity of i for each SNP and independent assortment
of SNPs, the binomial theorem indicates the probability that a given
donor-recipient pair will yield at least one informative locus when n
loci are evaluated is:
 |
Assuming an i informativity and assembly of a panel of n SNPs, the
probability of having y number of loci be informative is given by the
general formula:
 |
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. This method entailed PCR amplification of DNA, AvaII
digestion, 8% polyacrylamide gel electrophoresis, and identification
of characteristic band patterns. In our study population of 20 patients
we found 4 (20%) heterozygotes, consistent with the 19% heterozygous
frequency of this SNP reported in general population
studies.19, 23
All other patients in our study were allele
1 homozygotes; no allele 2 homozygotes were found in our limited
population. Results of three representative samples are shown in Figure 1B
. Because no internal AvaII restriction site control is
present within the amplified product, the undigested PCR products shown
in Figure 1B
were sequenced to confirm their identity (data not
shown). These patient samples with defined genotypes were then used
for subsequent real-time PCR analysis.
|
. DNA mixes of two allele 2/allele 1 heterozygous samples were diluted
into an allele1 homozygous sample at proportions of 100, 90, 80, 70,
60, 50, 40, 30, 20, 10, 5, 3, 1, and 0% heterozygote (equivalent to
50, 45, 40, 35, 30, 25, 20, 15, 10, 5, 2.5, 1.5, 0.5, and 0%
allele 2). Real-time PCR was then performed using the conditions
detailed above, and two standard curves of % allele 2/% allele 1
versus FAM/VIC ratio were generated (Figure 3)
. These two parameters showed a linear relationship in both standard
curves (r2 = 0.987 and
r2 = 0.977), except in the mixes with
the lowest percentage of allele 2 (<5%). Thus, changes in the ratio
of allele 2 (FAM fluorophore) to allele 1 (VIC fluorophore) in a DNA
mix results in a predictable change in the ratio of intensities emitted
by their corresponding fluorescent probes after PCR amplification. The
plots from both mock chimeras were almost superimposable (within the
limits of pipetting and spectrophotometry error), indicating that the
standard curve is reproducible and perhaps applicable to other patient
mixes as well. If such is the case, generating standard curves for each
individual patient may not be necessary.
|
|
. In
the RFLP-Southern blot assay, % allele 2/% allele 1 was determined by
phosphoimager analysis. Overall, the results of the two assays were
highly correlated (r2 = 0.971),
as shown in Figure 4
.
|
In method A, only homozygotes are considered informative such that only
two of the nine possible allele combinations are informative (Table 1)
.
A given pair of individuals is informative if one person is homozygous
for one of the alleles and the other is homozygous for the other
allele. Under these conditions, we found the total informativity of an
optimal theoretical SNP to be 0.125, or 12.5% (see Materials and
Methods). If we assume that multiple SNPs would be analyzed before
transplantation, how many would be required to find an informative
locus in the majority of donor-recipient pairs? Assuming independent
inheritance (no linkage disequilibrium) and applying the binomial
theorem, we calculated that if only five optimal SNPs are analyzed,
only 49% of random donor-recipient pairs will be informative for one
of them. For 10 SNPs, 74% will have at least one informative locus,
and for a panel of 25 SNPs, 96.5% of pairs will have at least one
informative SNP. With a panel of 25 SNPs, the probability of finding no
informative alleles is only 3.5%, and the average number of
informative SNPs is expected to be about 3 (25 x 0.125).
In method B, both homozygotes and heterozygotes are considered
informative. In this setting, six of the nine possible combinations are
informative (Table 1)
, resulting in an overall informativity of the
single optimal SNP of 62.5%. Accordingly, far fewer SNPs would be
required to produce a functional panel for monitoring BMT engraftment.
Only three SNPs would provide an overall informativity of almost 95%;
with four SNPs, 98% of donor-recipient pairs would be informative for
at least one SNP. Transplantation between family members would require
a larger number of SNPs to be screened to achieve these same levels of
informativity. Thus, including informative heterozygotes in these
analyses has a profound effect on the number of loci required to
achieve a reasonable overall informativity.
Effect of Allele Frequency on SNP Panel Size
The number of SNPs required in a panel is determined in part by
the allele frequency of p and q in the population. Using various allele
frequencies of 2 to 50%, we have calculated the number of SNPs needed
to compose a panel so that at least one of the SNPs would be
informative between the donor and recipient 95% of the time (Figure 5)
. Using method B to define informativity, altering the p
allele frequency from 0.20 to 0.50 has very little effect on the number
of SNPs required. Method A is more dependent on variations in allele
frequency. Allele frequencies below 30% using method A, or below 2%
using method B, would require analysis of too many SNPs to be
practical.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In this report we provide evidence that SNPs can be used to monitor BMT engraftment. Nearly any SNP can be used in this assay; our choice of the CYP2C9 SNP was based solely on the frequency of the polymorphism and the availability of the primers and probes in a commercially available system. For clinical use, multiple SNPs would have to be evaluated as a panel in the donor and recipient by the real-time PCR assay, before transplantation, to identify an informative locus. How one defines informativity (homozygotes only versus both homozygotes and heterozygotes) has a profound effect on the number of SNP loci required for analysis. In method A, a panel of 23 highly informative SNPs would need to be run to be >95% certain that at least one SNP was informative, whereas in method B, only about 4 such loci are required to achieve a similar level of confidence. Method B is also much less sensitive to allele frequencies lower than the 50% theoretical optimum. Once an adequate SNP locus is identified which distinguishes the donor from the recipient, all posttransplant assays would then be run using only one of the informative SNPs identified.
The linear relationship between the percentage of allele 2 present in a
chimeric mix and the FAM/VIC emission ratio becomes curvilinear at the
lowest levels of percentage of allele 2 in Figure 3
. This may represent
nonspecific binding of some FAM probe to allele 1 DNA due to a relative
excess of allele 1 DNA. Additional experiments with mixes of known
allele 2 homozygotes will need to be performed with optimal
probes to further define the actual limit of detection of the SNP
real-time PCR approach. With further optimization, the SNP based
real-time PCR assay may attain the limit of detection of microsatellite
and minisatellite based assays currently in use.
BMT engraftment assessment using the real-time PCR SNP assay described herein may prove to be an improvement over current engraftment assays. Because the ABI 7700 can discriminate probe signals and measure them over a range of several orders of magnitude,16 relatively small amounts of recipient DNA should be detectable in a posttransplant marrow specimen. Fortunately, our data suggest that standard curves may extrapolate to other patients (ie, will not be patient-specific), though this will need to be confirmed with additional patient samples. Furthermore, the assay requires only minute amounts of DNA (100 ng, as typically obtained in cytopenic bone marrow specimens), does not require donor-recipient sex mismatch or radioactivity, and circumvents differential amplification of unequal length alleles. Finally, a real-time PCR-based assay has the potential to reduce technologist time, improve turnaround time, and consequently reduce medical cost.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Supported in part by National Institutes of Health grants KO8 CA66628 and RO1 CA81439 (to J. R. E.).
Accepted for publication August 30, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  E. Gineikiene, M. Stoskus, and L. Griskevicius Single Nucleotide Polymorphism-Based System Improves the Applicability of Quantitative PCR for Chimerism Monitoring J. Mol. Diagn., January 1, 2009; 11(1): 66 - 74. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-L. Liang, M.-T. Lin, M. J. Hafez, C. D. Gocke, K. M. Murphy, L. J. Sokoll, and J. R. Eshleman Application of Traditional Clinical Pathology Quality Control Techniques to Molecular Pathology J. Mol. Diagn., March 1, 2008; 10(2): 142 - 146. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Hurtle, E. Bode, D. A. Kulesh, R. S. Kaplan, J. Garrison, D. Bridge, M. House, M. S. Frye, B. Loveless, and D. Norwood Detection of the Bacillus anthracis gyrA Gene by Using a Minor Groove Binder Probe J. Clin. Microbiol., January 1, 2004; 42(1): 179 - 185. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. K. Cashion, C. J. Driscoll, and O. Sabek Emerging Genetic Technologies in Clinical and Research Settings Biol Res Nurs, January 1, 2004; 5(3): 159 - 167. [Abstract] [PDF]
|
|
|
|
 |
|
 E. P. Hochberg, D. B. Miklos, D. Neuberg, D. A. Eichner, S. F. McLaughlin, A. Mattes-Ritz, E. P. Alyea, J. H. Antin, R. J. Soiffer, and J. Ritz A novel rapid single nucleotide polymorphism (SNP)-based method for assessment of hematopoietic chimerism after allogeneic stem cell transplantation Blood, January 1, 2003; 101(1): 363 - 369. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Borlak and T. Thum Identification of Major CYP2C9 and CYP2C19 Polymorphisms by Fluorescence Resonance Energy Transfer Analysis Clin. Chem., September 1, 2002; 48(9): 1592 - 1594. [Full Text] [PDF]
|
|
|
|
|
|
B. Zimmermann, W. Holzgreve, F. Wenzel, and S. Hahn Novel Real-Time Quantitative PCR Test for Trisomy 21 Clin. Chem., February 1, 2002; 48(2): 362 - 363. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
T alteration destroys an AvaII
(GGACC) site.
B: A 190-bp fragment containing the SNP was PCR-amplified,
digested by AvaII, and run on a polyacrylamide
electrophoresis gel. The allele 1 sequence contains an AvaII
site and is digested into 75-bp and 115-bp fragments; the site is lost
in allele 2, so this amplicon remains 190-bp after digestion.
Heterozygous DNA produces 190-, 115-, and 75-bp fragments. Examples
from three different individuals are shown before and after cutting
with AvaII.
