Quantitative determination of the ratio of mutated to normal ras genes in the blood of leukemia patients by allele-specific PCR

~Pergamon

VoL 18, No. 9, pp. 693-702, 1994 Copyright© 1994Elsevier Science Ltd Printed in Great Britain. All rights reserved 0145 2126/94$7.00 + 0.00

Leukemia Research

0145-2126(94)00074-3

QUANTITATIVE DETERMINATION OF THE RATIO OF MUTATED TO NORMAL r a s GENES IN THE BLOOD OF LEUKEMIA PATIENTS BY ALLELE-SPECIFIC PCR TETSURO HORIKOSHI,* HEINZ-JOSEF LENZ,* KATHLEEN DANENBERG,* OLAF JOSEPH R . B E R T I N O t a n d P E T E R V . D A N E N B E R G *

M.

KOCH,t

*Kenneth Norris Jr Comprehensive Cancer Center, University of Southern California School of Medicine, Los Angeles, U.S.A. ; and tMemorial Sloan-Kettering Cancer Center, New York, U.S.A. (Received 5 January 1994. Revision accepted 7 June 1994) Abstract--By combining allele-specific PCR amplification and a PCR-based quantitation approach, a method has been developed to estimate the mutated K-ras gene content in the blood of AML patients as a percentage of total K-ras. One PCR primer set was designed not to discriminate between mutant K-ras and wild-type K-ras and thus amplified the total K-ras gene. The other PCR primer set was designed to be allele-specific for K-ras genes containing a G to C mutation at codon 12. This primer set could discriminate the mutant and wild-type genes when the proportion of the mutated sequence was 0.2% of the total K-ras gene. To test the method on biological specimens, genomic DNA samples were analyzed from the peripheral blood of a patient who had secondary AML with the same codon 12 K-ras mutation. Two samples taken from this patient 2 months apart during follow-up had myeloblest cell contents of 67 and 80%. However, the percentage of mutated K-raswas 50% in both samples, suggesting that this patient may be inherently heterozygotic in this particular mutation. This ratio of mutated to normal K-ras in the patient's cells was confirmed by RNA-SSCP analysis and RNA sequencing. This quantitation method can provide a sensitive and specific estimation of the content of mutated K-ras alleles in patient samples. Key words: Quantitation, allele-specific PCR, ras, leukemia, mutation.

Introduction

which in some cases approach a detection sensitivity of 1 malignant cell per 106 cells [2]. In the case of hematological malignancies, not only detection, but also an accurate quantitative determination of the percentage of malignant cells may be an important prognostic factor for predicting the course of disease progression. Quantitative analysis of the fraction of malignant cells in a mixed cell population could be used to measure the rate of increase of the malignant cells or to evaluate the efficacy of a given treatment in terms of reduction of the leukemic cell load. However, with one exception, all of the methods available to date for detection of malignant cells are difficult to use in a quantitative manner. Mutations of the ras gene occur with a high frequency in MDS and A M L patients [4] and thus could be conveniently used as markers for malignant and pre-malignant cells in many cases. In this paper, we describe an internally controlled approach for quantitative determination of the fraction of a mutated gene in a mixture of mutated and normal genes and its application for specifically quantitating

THE IMPORTANCE of developing sensitive methods for the detection of malignant cells in leukemia patients has been generally recognized because of the possibility that the persistence of low numbers of leukemic cells after treatment, or minimal residual disease (MRD), will eventually result in partial remission or relapse of the leukemia [1]. Standard approaches such as morphology, cytogenetics, immunological methods and Southern blotting are effective only when the amount of the malignant clone appreciably exceeds 1% [1]. More recently, very sensitive molecular methods have been developed that recognize specific genetic differences between the normal and malignant cells. For example, once a unique D N A or R N A sequence is identified in the malignant cells, allele-specific PCR techniques can be used Correspondence to: Peter V. Danenberg, Kenneth Norris Jr Comprehensive Cancer Center, University of Southern California School of Medicine, 1303 N. Mission Road, CRL 204, Los Angeles, CA 90033, U.S.A. 693

T. HORIKOSHIet al.

694

a K-ras gene containing a G to C transversion in c o d o n 12. T h e m e t h o d c o m b i n e s allele-specific P C R with a P C R - b a s e d m e t h o d that we designed earlier for accurate d e t e r m i n a t i o n of the ratios between different D N A segments in the same solution [3]. The discrimination sensitivity of the m e t h o d for this particular m u t a t e d K-ras gene was 0.2% of the total ras, while the detection sensitivity was 10 8 ng of template, or a b o u t 55 molecules. The m e t h o d was used to quantitate the fraction of m u t a t e d K-ras genes in m o n o n u c l e a r cells of a patient with secondary acute myeloid leukemia.

PCR (Fig. 1). The PCR reactions contained 50 pmol of the primers (KR7 and KR153, Table 1) 10 ~tl of 10 × Taq buffer (500 mM KCI, 100 mM Tris-HC1, pH 8.3), 200 ~tM deoxyribonucleoside triphosphates, 1.87 mM MgCI2, 0.2 pmol of fragment A (wild-type) or fragment B (mutant) and common overlapping fragment C. The reaction mixture was overlaid with mineral oil. After the reaction was heated for 5 min at 98°C, 1 min at 70°C, and i min 60°C in the thermocycler (TwinBlock, Ericomp, San Diego, Ca), 2.52 units of Taq polymerase in 8 ~tl of i × Taq buffer were added. The PCR conditions were 30 cycles of l m i n at 96°C, i min at 55°C, and 1 min at 72°C with a final extension step of 7 min at 72°C. The PCR products were purified by electrophoresis on a 6% polyacrylamide/8 M urea gel. The concentration was determined by UV spectrophotometry ( 1 0 . D . = 50 ~g/ml dsDNA).

Materials and M e t h o d s

Clinical specimens Blood mononuclear cells from healthy volunteers and bone marrow aspirates from a patient with secondary acute myeloid leukemia following MDS were separated by centrifugation on a Ficoll gradient, from which DNA was isolated according to published protocols [5]. Two genomic DNA samples were obtained from the same patient with secondary AML 2 months apart. Sample MDS 106 contained 67% myeloblasts and sample MDS 110 80% myeloblasts. Chemicals and enzymes Recombinant Taq D N A polymerase (AmpliTaq) was obtained from Perkin Elmer Cetus (Norwalk, CT); RNAguard RNase inhibitor, deoxynucleoside-5'-triphosphates, ribonucleoside-5'-triphosphates, and T7 RNA polymerase from Pharmacia; [a'-32p]CTP and [a~-32p]ATP were from Amersham (Arlington Heights, IL). The remaining chemicals and buffers obtained from various sources were molecular biology grade. Oligonucleotide primers and fragments for PCR The primers and fragments listed in Table 1 were synthesized by the phosphoramidite method using an Applied Biosystems (Foster City, CA) model 391 PCR-MATE DNA synthesizer. The T7 RNA polymerase promoter sequences are indicated by the prefix T7. The tetranucleotides in quotation marks are not part of the target sequence, but represent a T7 RNA polymerase promoter 'transcription initiation sequence' that increases the yield of the transcription product [6]. The human c-Ki-ras 2 (K-ras) proto-oncogene sequence for designing primers and synthesizing artificial gene templates was taken from the sequence data provided by McGrath et al. [7]. The nucleotide positions of each primer and fragment are schematically represented in Fig. 1. R N A sequencing R N A fragments from the transcription of PCR products were sequenced with AMV reverse transcriptase (Life Sciences, St. Petersburg, FL) according to Stoflet et al. [6]. PCR synthesis of wild-type and mutant K-ras fragments Segments of the wild-type K-ras gene and the corresponding mutant K-ras gene containing the base substitution found in the patient (codon 12, G G T - G C T ) of 166 base pairs were synthesized by overlapping fragment

Allele-specific PCR quantitation In order to determine the ratio of mutant fragment D to the total K-ras (wild-type + mutant) fragment E (Fig. 2), genomic D N A was amplified with primer pairs KR141 and KR7 to give PCR products corresponding to total K-ras, and in a separate reaction, the allele-specific primer KR161 was used along with KR141 to amplify the mutant Kras. PCR reactions contained 0-10 ~tl of genomic D N A solution, 12.5 pmol of the primer set (KR141 and KR7 or KR141 and KR161), 2.5 ~tl of 10 × Taq buffer, 200 ~tM dNTPs, 1.87 mM MgCI2 in a total volume of 25 ~tl. Mineral oil (ca. 40 ~tl) was placed on top of the aqueous layer. After pre-heating (5 min at 98°C, I min at 70°C and 1 rain at 60°C) in the thermocycler, 0.63 units of Taq D N A polymerase were added. The PCR cycles were 1 rain at 96°C, 30 s at 65°C and 30 s at 72°C for 27 cycles with a final extension step of 7 min at 72°C. PCR reactions were transcribed with T7 RNA polymerase. Transcription reactions contained 2.5 ~tl of 10 x transcription buffer (400 mM TrisHC1, pH 7.5, 120 mM MgCI2, 100 mM dithiothreitol and 10mM spermidine), 2.5 ~tl of 10raM solution of ribonucleotides (ATP, CTP, GTP and UTP), 0.25 ~tl of RNAguard, 15.8~tl of diethyl-pyrocarbonate-treated water, 0.25 ~tl of [ol-32p]CTP (3000 Ci/mmol), 3 ~tl of the PCR reaction and 1.62 ~tl of T7 RNA polymerase (50 units/~d, Epicentre). After incubation of the mixture for 1 h at 37°C, 0.75 ~tl of 0.5 M E D T A were added to stop the reaction. RNA products were electrophoresed on 6% polyacrylamide/8M urea gel. The gel was dried and exposed to a X A R film (Kodak, Rochester, NY). The bands were excised and counted in a liquid scintillation counter. R N A - S S C P analysis RNA-single strand conformational polymorphism (SSCP) analysis was performed as previously described [8]. Primers KR162 and KR155 were used to amplify a region of the K-ras gene from nucleotide positions 79 to 168. This fragment contains the G---,C mutation at position 130. The PCR amplifications were carried out for 35 cycles at 96°C for 1 min, 60°C for I min, and 72°C for I min with a final extension time of 7 min at 72°C. A portion (3 txl) of the PCR reaction was transcribed as described above except that 0.68 ~tl of T7 RNA polymerase (Pharmacia, 0.86 units/ ~tl) were used. Non-denaturing gel electrophoresis was carried out on HydroLink-MDEE gels. The gel was prepared with the MDEE 2 x concentrate (AT Biochemicals,

695

Ratios of ras genes determined by PCR TABLE 1. SEQUENCES OF PRIMERS AND SYNTHETIC FRAGMENTS USED*

Primers used for quantitation: 5'-T7t-'GGGA' :~GTCCTGCACCAGTAATATGC-3' KR141: 5'-AGGCCTGCTGAAAATGACTG-3' KR7: 5'-CTrGTGGTAGTTGGAGCTGC-3' KR161: Primers used for RNA-SSCP: 5'-T7-'GGGA'TATAAGGCCTGCTGAAAATG-3' KR162: 5'-'GGGA'GAATTAGCTGTATCGTCAAG-3' KR155: Primers used for RNA sequencing: KR141 and KR7 for PCR amplification and 32p-end labeled KR7 as sequencing primer Primers and template used to generate the synthetic WT and MU fragments: 5'-AGGCCTGCTGAAAATGACTG-3' KR7: 5'-'GGGA'GTCCTGCACCAGTAATATGC-3' KR153: 5'-TATAAGGCCTGCTGAAAATGACTGAATATAAACTFGTGGTAGTTGGAGCTGGTGGCGA: TAGGCAAGAGTGCCTrGACGATACAGCTAATTC-3' 5'-TATAAGGCCTGCTGAAAATGACTGAATATAAACTrGTGGTAGTTGGAGCTGCTGGCGB: TAGGCAAGAGTGCCTrGACGATACAGCTAATrC-3' 5'-GTCCTGCACCAGTAATATGCATATTAAAACAAGATTTACCTCTATTGTTGGATCATATTC: CGTCCACAAAATGATrCTGAATrAGCTGTATCGTC-3' * Nucleotide position: see Fig. 1. t T7 = AATTTAATACGACTCACTATA. $ 'GGGA' = translation initiation sequences. The G ~ C mutation indicated in bold. KR7

Results A

B

x

IlL C

"I

KR153 K-ras

100 ATG I

150 [

200 l

~

250 I

FIG. 1. Map of synthesized fragments for human c-K-ras 2 (K-ras) proto-oncogene. Sequences of primers and synthesized fragments are listed in Table 1. The X in fragment B indicates the G to C mutation in codon 12. KR161 D x

KR7

KR141 E KR141

ATGIO01

150

200

250

K-ras

FIG. 2. Map of primers for PCR amplification of total Kras and mutant K-ras. Primers KR7 and KR141 define total K-ras (fragment E). The allele-specific primers KR161 and KR141 selectively amplify the mutant K-ras (fragment D). The X in fragment D indicates the G to C mutation in codon 12. Malvern, PA) according to the manufacturer's recommendations for SSCP analysis. The transcription reaction mixture (10 ~tl) was mixed with 2 ~tl of 40% sucrose containing 0.05% bromophenol blue dye, loaded onto the MDEE gels and electrophoresed in 1 × TBE buffer at 24 W constant power per gel at 20-22°C by circulating tap water through the cooling tubes of the gel apparatus (model SE600, Hoefer, San Francisco, CA). The dried gel was exposed to XAR film (Kodak, Rochester, NY).

Identification of a K-ras mutation in an M D S patient Using R N A sequencing (Fig. 3) and R N A - S S C P (8) analysis (Fig. 4), we established the presence of a point mutation in the K-ras proto-oncogene at codon 12 (GGT---~GCT) in two samples from a patient with secondary A M L , taken 2 months apart (samples MDS106 and MDSl10) (Fig. 3). This single base change causing a Gly to Ala substitution is one of the common activated K-ras mutations [9]. The sequencing pattern showed C and G residues at the same position, indicating that these samples are a mixture of mutant and wt alleles. This is confirmed by the R N A - S S C P analysis (Fig. 4) of MDS106 and M D S l l 0 , which showed almost identical patterns that corresponded to the combined superimposed patterns of synthetic wild-type and synthetic mutant segments. The intensity of wild-type and mutant bands in the R N A - S S C P patterns were similar, indicating that the content of mutated and normal K-ras genes in the patient's tissues were also similar. The lack of additional R N A bands in the R N A - S S C P patterns other than those of the wild-type and codon 12 mutant segments indicates that during progression of AML, the patient did not develop any other mutation in codon 12 or 13 regions of the K-ras gene. Strategy for quantitation by allele-specific PCR The strategy that we devised for PCR-based quantitation of mutated K-ras gene content was similar in principle to our previously published general method

696

T. HORIKOSHIet al.

MDSI06

MDS 110

A

A

G C

T

G C T

codon

c

codon

-I

c

G 13

°t

T

T

G/C 12 G J : i ~

1

G 13 G J

1

G/C 12 G J

~

e

FIG. 3. Human K-ras gene sequences of MDS patient samples, MDS106 and MDS110. DNA fragments amplified by the PCR using primer pair of KR7 and KR141 were transcribed to RNA by T7 RNA polymerase. The resulting RNA fragments were sequenced using the 32p-labeled primer KR7 and AMV-reverse transcriptase in the presence of dideoxy-NTP. Both genomic DNA samples show a wild-type codon 12 sequence and a codon 12 GGT---~GCT mutation sequence, resulting in a substitution of Gly 12 to Ala 12.

for quantitation of gene expression [3]. This method involves PCR amplification of the gene (or its cDNA) of interest plus amplification in a separate reaction of an invariantly expressed internal standard gene (such as beta-actin) that is used as a denominator to normalize the amount of the target gene per cell. In order to quantitate mutated K-ras as a marker for malignant cells, the total ras gene itself is used as the denominator instead of an unrelated gene. Amplification of 'total ras' is accomplished with a 5' primer binding upstream from the site of the mutation (KR7 in Fig. 1) so that primers KR7 and KR141 would give a PCR product that is a mixture of the wild-type genes and mutant genes from the neoplastic cells. For the allele-specific amplification of only the mutant allele, the same 3' primer (KR141) but a new allelespecific 5' primer (KR161, Fig. 2) is used. KR161 has a C instead of a G on its 3' end that will, in theory, exclusively extend the mutant K-ras of the MDS patient (fragment E) and not the gene segment

with the wild-type codon 12. PCR of both total ras and mutant ras is carried out with serial dilutions of the D N A solution in order to establish the linear range of each PCR amplification reaction. An empirical ratio (Rapp) is then determined between the amounts of the K-ras allele-specific PCR product (fragment D, Fig. 2) and the total K-ras PCR product (fragment E, Fig. 2) in the linear ranges of amplification of each fragment: slope of the linear portion of the mutant-specific PCR product curve (cpm/vol of DNA solution) Rap p -

slope of the linear portion of the total K-ras PCR product curve (cpm/vol of DNA solution)

Rapp is the apparent fraction of mutated K-ras to total K-ras genes without taking into account either: (a) the efficiencies of the two different primer sets in the PCR; or (b) the difference in the amount of incorporation of the radiolabeled nucleotide into the

Ratios of ras genes determinedby PCR

1

2

3

4

FlG. 4. RNA-SSCP analysis of the K-ras gene mutation in the secondary AML patients. Genomic DNA samples from the patient, MDS106 and MDSll0, the synthesized fragment A containing wild-type sequence and the synthesized fragment B containing a point mutation at codon 12 (GGT--~GCT) were PCR-amplified using the primer pair KR162 and KR155. After transcription of the PCR-amplifled DNAs, the RNAs were applied to a native 0.5 × MDEE gel in 1 x TBE buffer and then electrophoresed at 20-22°C at 25 W constant power. Lane 1: synthetic wild-type (A); lane 2: synthetic mutant (B); lane 3: patient sample MDS106; lane 4: patient sample MDSll0.

two fragments. Determining the true ratio is described below. Selectivity of the allele-specific amplification To test the primer sets for allele-specific selectivity, and to demonstrate the above strategy for quantitation, segments of the K-ras gene from the secondary AML patient and a healthy volunteer with a wild-type K-ras sequence were analyzed. The genomic DNA solutions were serially diluted, PCRamplified, and transcribed to with T7 polymerase (Fig. 5). The main band of each PCR corresponded to the expected size of the RNA. Figure 5(A) shows that when DNA from the healthy volunteer was amplified with the allele-specific primer set KR161KR141, no band was visible on the autoradiogram and counting the region where the band would be expected in a liquid scintillation counter disclosed only a small amount of radioactivity (ca. 350 cpm)

697

close to the background level, which, if entirely due to the PCR fragment would correspond to Rapp = 0.0021. In contrast, amplification of MDS106 DNA with the allele-specific primer set gave a large amount of radiolabeled PCR-T7 transcription product corresponding to about one-third that of the total K-ras PCR product, for an Rapp value of 0.35 (Fig. 5(B)). The Rapp value for MDS110 was identical (0.35) (data not shown), demonstrating that the percentage of the mutated K-ras gene in this patient had not changed over the 2 month period between tissue sampling. To determine the precise selectivity (percent of false extension) of the allele-specific primer for this mutation (codon 12, GGT----~GCT) a fragment of the wild-type sequence was synthesized by Taq polymerase extension of overlapping fragments 169 and 171 (Fig. 1). This fragment was then PCR-amplified with the primer sets KR141/KR7 for total K-ras and the mutant allele-specific primer sets KR141-KR161. Figure 6(A) shows that at concentrations >10 -5 ng of the wild-type synthetic template per reaction, nonspecific extension of wild-type template by the mutant-specific primers occurred to the extent of about 0.2% of that of the K-ras reference amplification. At lower concentrations, however, the amount of false extension was not detectable above the background radioactivity. Thus, the discrimination ability of the allele-specific PCR for this particular mutation under our PCR conditions will be at least 0.2%.. Determining the absolute percentage of mutant K-ras to total K-ras Rapp values, the empirical ratios of the PCR products from amplification with mutant K-ras primers and total K-ras primers, probably would not be the same as the real ratios between these fragments due to different fragment lengths and possible differential efficiency of PCR amplification by the primer sets. To allow the conversion of Rapp values to the true ratios, the segment of K-ras containing the codon 12 GGT---,GCT mutation prepared as shown in Fig. 1 and the corresponding synthetic wild-type segment (Fig. 2) were mixed in various ratios ranging from 0 to 100% of the total. The mutant allele was quantitated in the mixtures with primer sets KR141KR161 as described above. The increase in Rapp values was linear with increasing amounts of the mutated fragment (Fig. 7), Figure 7 is a standard curve which can be used to find the true ratios between mutant K-ras and total K-ras once Rapp values are determined for samples of unknown composition. However, since the curve is linear, the correction factor is constant regardless of composition and is equal to the slope of the line, which

698

T. HORIKOSHIet al.

A :~

7o 2 o

B

7

DNA (hi)

Reference

K.qi4~Jq7

20

7

2

0.7

50

20

7

2

0.7

•

KR141/KR7

50

V_q1:41~

50

DNA

(hi)

Allele-specific KR141/KR161

el

100000

50000

25000

i

i

i

i

10

20

30

40

80000

DO

0 •..,. 2 0 0 0 0

60000 o

15000 0 k_

o L

D-

40000

t3_

10000

tY (.] 13_

5000 ( "~lm,

0

A

!

200

400

~

I

;

A

600

800

1000

Genomic DNA (hi)

20000

0

v

50

Cenomic DNA (nl)

FIG. 5. Relationship between starting amount of genomic DNA and the amount of the PCR products formed. Aliquots of a genomic DNA solution from healthy volunteer (A) and MDS106 (B) were PCR-amplified using either the primer set of KR141 and KR7 for K-ras reference amplification (©) or KR141 and KR161 for allele-specific amplification (0). The PCR product was transcribed to RNA and electrophoresed in 6% polyacrylamide/8 M urea gel. The bands were visualized by autoradiography (upper autoradiograms), excised, and quantitated by liquid scintillation counting. The DNA concentration of MDS106 was 630 ng/~tl, that of the healthy volunteer was too low to estimate the exact concentration. The slopes of first linear region were used for calculation of Rap p.

is 0.67. This n u m b e r is also obtained simply by amplifying only the mutated fragment, which gives an actual ratio of 1.0 between mutant K-ras and total K-ras. Dividing the Rap p value of 0.35 found for the patient samples by 0.67 gives an actual ratio of 0.52 between the mutant K-ras and total K-ras (Table 2). This curve also shows that, with the template concentrations used in this particular experiment, the

lower limit for m e a s u r e m e n t of the mutant K-ras/ total K-ras ratio was about 0.01. Discussion In this study, we developed a method for quantitating the amount of a mutated gene as a percentage of the total gene (the sum of both the wild-type and

Ratios of

ras

699

genes determined by PCR 0.8

A

/

0 i

i

~

i

i

73 0

i

10q000

/

0.6

Q-

//

/

/

¢v

.//

© o_

[3_ 0

0.4 "-

,J

/

/

/ //

o

i%ooo

*6 0.2 --,

./~//

p,,

0

,

/

0

0.0 tY o

t

0.0

,,,I

........

10 -a

i

10

........

i

........

10 . 6

-7

i

........

10 - 5

q

........

10 - ¢

i

10 -3

S y n t h e t i c WT ( n g / t u b e )

8 lO0,O0O

13.. (.3

10,000

o zl '13 o l_

CL r}," 0 13..

0.4

r 0.6

I 0.8

1.0

MU/(WT+MU)

1000

m

0.2

1000

i [,.,~

0 -a

........

i

10 -7

........

i

10-6

i i,uml

10 - 5

Synthetic MU (ng/tube) FIG. 6. Discrimination ability of allele-specific PCR amplification. A synthetic template of the wild-type K-ras gene was PCR-amplified: (A) using primer sets KR141/KR7 (©) and KR141/KR161 (Q); and (B) using primer sets KRI41/KR7 (©) and KR141/KR161 (O). The amount of initial DNA for amplification was determined by spectrophotometry.

mutated genes) in tissue specimens that are expected to contain a mixture of cells having both the mutated and the normal gene. The method combines allelespecific PCR amplification with the principle developed in our previously published PCR-based quantitation method of using an endogenous gene as a normalization denominator [3]. In this particular approach to quantitative PCR, the data are expressed as a ratio between two quantities which are measured by the same technique. Thus, it is not necessary to

FIG. 7. Linearity of quantitation of the mutant allele by allele-specific PCR. Solutions of synthesized templates of wild-type K-ras gene and that of mutant K-ras gene were mixed at specific ratios so that new solutions with different contents of mutant K-ras gene would be obtained. In each mixture, the total K-ras fragment was amplified using primer sets KR141 and KR7, and the K-ras gene mutant allele-specific fragment was amplified using primer sets KR141 and KR161. The ratios of allele-specific amplification and total K-ras (Rapp) amplification are plotted vs the true ratios of mutant allele to total K-ras.

measure the amount of RNA, D N A or protein in the specimens in order to obtain denominators. Because of the sensitivity of the PCR, one does not need to be concerned with obtaining enough tissue to be able to measure these quantities. Since the total gene amplification and the allele-specific mutant amplification are carried out in different tubes, it is possible to establish linear ranges of PCR amplification for each D N A segment separately simply by performing the PCR on serially diluted volumes of the D N A solution. We have shown that the PCR reactions are usually linear over about a 2 log concentration range of the D N A segments to be amplified [3]. To convert the ratio between the slopes of the linear PCR amplification ranges (Rapp) to the actual ratio between the mutated and the total gene requires correction for different segment length and PCR amplification efficiencies of the D N A segments by preparing a standard curve as shown in Fig. 7. However, for most purposes this will probably not be necessary. For example, to study time-dependent changes in the amount of the mutant allele in the same patient, it would be sufficient just to determine the Rap p value at various times. We chose to develop the initial application of this mutation quantitation method for a K - r a s mutation, because ras mutations may be among the most useful markers for malignant or pre-malignant cells in leukemia. Mutated ras occurs with a frequency of about 40-50% of MDS patients [10-12] and 30% of A M L

700

T. HORIROSHI et al. TABLE 2. FRACTION OF MUTANT K - r a s GENE DETERMINED BY ALLELE-SPECIFIC

PCR QUANTITATION Sample

Rapp

Corrected value

Normal blood DNA MDS106 MDSll0 Synthetic mutant DNA strand

2.1 × 10.3* 0.35 0.35 0.67 +- 0.12t

0.0032 0.52 0.52 1.00

* Maximal value. t Mean -+ S.D.

patients [12]. Exploiting ras mutations for quantitation of malignant cell fraction would require prior identification of the particular ras mutation in patients' cells but performing such analyses on a routine basis is now facilitated by recent developments in mutation detection technology such as an DNA-SSCP [13] or RNA-SSCP [8]. Apart from MRD measurement, establishing the presence or absence of ras mutations may itself be of prognostic value. Patients with MDS containing a ras mutation appear to have a greater chance to progress into AML and thus may have a poorer prognosis [10]. Patients with a ras mutation in a multipotent stem cell appear to be more resistant to chemotherapy [9]. Thus, quantitative analysis of the amount of mutated m s may identify patients who are at high risk and thereby qualify for an intensification of chemotherapy or for an early bone marrow transplantation. To demonstrate and validate the method for analysis of clinical material, we applied it to quantitate the mutated K - r a s gene content in a patient with secondary AML. The ras mutation in the malignant cells of the patient was first identified by RNA sequencing, and then PCR primers were designed for allele-specific PCR of this mutation. The PCR quantitation assay showed that the amount of the mutated K - r a s allele remained unchanged at the level of about 50% in two samples taken from the patient several months apart. These results were independently validated by RNA-SSCP and RNA sequence analysis. The RNA sequencing gels (Fig. 3) show that the ratios of the double bands corresponding to G and C at codon 12 in both tissue specimens from the MDS patient were similar, indicating no appreciable difference in the mutant to wild-type gene ratio between MDS106 and MDSll0. RNA-SSCP is based on the principle that metastable conformational states of RNA strands can be separated and observed upon gel electrophoresis of the RNA under non-denaturing conditions and that the resulting conformational band patterns often change upon even a single-base substitution [8].

Figure 4 (lanes 1 and 2) shows that the RNA conformation patterns of the wild-type K - r a s segment are different from those of the mutant fragment, and that the RNA patterns from the patient clearly consist of superimposed wild-type and mutant patterns at about the same intensity, thus confirming that wildtype and mutated alleles are present at similar levels in both of the tissue specimens from the patient. It is interesting that the mutated K - r a s content of the second clinical specimen remained unchanged at about 50% even though the percentage of myeloblasts increased from 67 to 80%. This observation suggests that this patient is heterozygotic for K - r a s because in other studies, the fraction of cells containing the mutated ras oncogenes were shown to increase concurrently with an increase in the fraction of blast cells [14]. It has been previously shown that r a s gene mutations can be incurred either in a multipotent stem cell or in a newly evolved leukemic cell clone later in the course of the disease [10]. The general applicability of this method for the quantitation of mutant genes will depend on the ability of allele-specific PCR to discriminate between the mutant and the corresponding wild-type genes. The principle behind allele-specific PCR is that primer extension by Taq polymerase is either absent or much less effective if there is no complementary base-pairing between the 3' terminus of the primer and the template. PCR primers with different 3' terminal bases complementary to either the altered base on the mutant or the original base on the wildtype allele should selectively amplify one allele in the presence of the other. In practice, some mismatched base pairs at the 3' end of the primer do allow significant primer extension [15], which would suggest that such mutations will not be quantifiable by allele-specific PCR. However, several studies have shown the amount of false extension of mismatched 3' termini is greatly dependent on the PCR conditions [16--19]. For example, A : A and T : T mismatches are extended at annealing temperatures of 50°C, but not at 55°C [16], Thus, if PCR conditions are optimized

Ratios of ras genes determined by PCR as to dNTP concentration, annealing temperature, and presence of PCR specificity reagents (for example, Perfect Match, Stratagene, La Jolla, CA) allele-specific quantitation may be possible for most mutations. In the present study, extension of the C : C mismatch between the primer and the template could not be detected in D N A from the healthy volunteer but was seen using high levels of the synthetic templates to the extent of about 0.2% that of the matched base pair in the wild-type K-ras template. Among the previously existing methods for detecting allelic differences in DNA, the single-nucleotide primer extension (SNuPE) method [20] is the only one that can be adapted for use in a quantitative manner without the need for measuring external normalizing factors. In the case of C : T or A : G mismatches the SNuPE assay can quantitate the mutant allele when it is as low as 0.1% of the total [21]. We did not test those particular base pair mismatches in the present study, but the 0.2% discrimination level obtained for the C : C mismatch without special efforts to optimize PCR conditions is comparable to that of the SNuPE assay. Because the present allele-specific quantitation method is carried within linear ranges of PCR amplification, additional information can be obtained about the gene of interest. For example, the copy number of the gene can be determined if another suitable gene is amplified for use as an internal reference. If cDNA is quantitated, the relative m R N A expression level of the gene of interest as well as that of the specific mutant allele can be compared with that of another specimen. By using the total gene itself as the quantitation denominator rather than some unrelated gene, the data are controlled for possible differences in expression of the denominator gene among tissue specimens. Such a capability may be useful after transfection of cells with a mutated ras gene to determine relative expression of mutated versus normal ras and to correlate these data with phenotypic changes in the cells. References 1. Van Dongen J. J. M., Breit T. M., Adriaansen H. J., Beishuizen A. & Hooijkaas H. (1992) Detection of minimal residual disease in acute leukemia by immunological marker analysis and polymerase chain reaction. Leukemia 6, 47. 2. Campana D., Yokota S., Coustan-Smith E., HansenHagge T. E., Janossy G. & Bartram C. R. (1990) The detection of residual acute lymphoblastic leukemia cells with immunological methods and polymerase chain reaction: a comparative study. Leukemia 4, 609. 3. Horikoshi T., Danenberg K. D., Stadlbauer T. H. W., Volkenandt M.~ Shea L. L. C., Aigner K.~ Gustavsson

701

B., Leichman L., Frosing R., Ray M., Gibson N. W., Spears C. P. & Danenberg P. V. (1992) Quantitation of thymidylate synthase, dihydrofolate reductase, and DT-diaphorase gene expression in human tumors using the polymerase chain reaction. Cancer Res. 52, 108. 4. Padua R. A., Carter G., Hughes D., Gow J., Farr C., Oscier D., McCormick F. & Jacobs A. (1988) Ras mutations in myelodysplasia detected by amplification, oligonucleotide hybridization and transformation. Leukemia 2, 503. 5. Maniatis T., Fritsch E. F. & Sambrook J. (1982) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Press, New York. 6. Stoflet E. S., Koeberl D. D., Sarkar G. & Sommer S. S. (1988) Genomic amplification with transcript sequencing. Science 239, 491. 7. McGrath J. P., Capon D. J., Smith D. H., Chen E. Y., Seeburg P. H., Goeddel D. V. & Levinson A. D. (1983) Structure and organization of the human Ki-ras proto-oncogene and a related processed pseudogene. Nature 304, 501. 8. Danenberg P. V., Horikoshi T., Volkenandt M., Danenberg K. D., Lenz H. J., Shea L. L. C., Dicker A. P., Simoneau A., Jones P. A. & Bertino J. R. (1992) Detection of point mutations in human DNA by analysis of RNA conformation polymorphism(s). Nucl. Acids Res. 20, 573. 9. Bos J. k. (1989) Ras oncogenes in human cancer: a review. Cancer Res. 49, 4682. 10. Jacobs A. (1992) Gene mutations in myelodysplasia. Leukemia Res. 16, 47. 11. Yunis J. J., Boot A. J. M., Mayer M. G. & Bos J. (1989) Mechanisms of ras mutation in myelodysplastic syndrome. Oncogene 4, 609. 12. Bos J. L., Verlaan-de Vries M., Van der Eb A. J., Janssen J. W. G., Delwel R., Lowenberg B. & Colly L. P. (1987) Mutations in N-ras predominate in acute myeloid leukemia. Blood 69, 1237. 13. Orita M., Suzuki Y., Sekiya T. & Hayashi K. (1989) Rapid and sensitive detection of point mutations and DNA polymorphisms using the polymerase chain reaction. Genomics 5, 874. 14. Hirai E., Okada M., Mizoguchi H., Mano H., Kobayashi Y., Nishida J. & Takaku F. (1988) Relationship between an activated N-ras oncogene and chromosomal abnormality during leukemic progression from myelodysplastic syndrome. Blood 71, 256. 15. Kwok S., Kellogg D. E., McKinney N., Spasic D., Goda L., Levenson C. & Sninsky J. J. (1990) Effects of primer-template mismatches in the polymerase chain reaction: human immunodeficiency virus type 1 model studies. Nucl. Acids Res. 18, 999. 16. Wu D. Y., Ugozzoli L., Pal B. K. & Wallace R. B. (1989) Allele-specific enzymatic amplification of/3globin genomic DNA for diagnosis of sickle cell anemia. Proc. natn. Acad. Sci. U.S.A. 86, 2757. 17. Ehlen T & Dubeau L. (1989) Detection of ras point mutations by polymerase chain reaction using mutation-specific inosine-containing oligonucleotide primers. Biochem. biophys. Res. Commun. 160, 441. 18. Okayama H., Curiel D. T., Brantly M. L., Holmes M. D. & Crystal R. G. (1989) Rapid, nonradioactive detection of mutations in thc human genome by allelespecific amplification. J. Lab. Clin. Med. 114, 105.

702

T. HORIKOSH~et al.

19. Major J. G. (1992) A rapid PCR method of screening for small mutations. BioTechniques 12, 40. 20. Kuppuswamy M. H., Hoffmann J. W., Kasper C. K., Spitzer S. G., Groce S. L. & Bajaj S. P. (1991) Single nucleotide primer extension to detect genetic diseases: experimental application to hemophilia B (factor IX)

and cystic fibrosis genes. Proc. natn. Acad. Sci. U.S.A.

88, 1143. 21. Singer-Sam J., LeBon J. M., Dai A. & Riggs A. D. (1992) A sensitive, quantitative assay for measurement of allele-specific transcripts differing by a single nucleotide. PCR Meth. Applic. 1, 160.