Inr. J . Cancer: 48, 51-56 (1991) 0 1991 Wiley-Liss, Inc.
Publication of the International Union Against Cancer Publication de I'Union Internationale Contre le Cancer
ALLELE-SPECIFIC DETECTION OF K-rus ONCOGENE EXPRESSION IN HUMAN NON-SMALL-CELL LUNG CARCINOMAS Robert J.C. SLEBOS',Gaston G.M. HABETS,Siegina G. EVERS,Wolter J. Moo1 and Sjoerd RODENHUIS Departments of Experimental Therapy, Medical Oncology and Cell Biology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands. Point mutations in codon I2 of the K-ras oncogene are frequent in human lung adenourcinomas. To study the expression of the K-ras gene in these tumors we have developed a mRNA detection technique based on the polymerase chain reaction (PCR). By this technique, K-ras expression can be detected semi-quantitatively in samples of less than 100 ng total RNA. Hybridization of the amplified cDNA sequences with mutation-specificoligonucleotides allows separate quantification of the expression of normal and point-mutated alleles in a single sample. RNA samples from 24 human nonsmall-cell lung carcinomas (NSCLC), from 2 lung metastases of colonic adenocarcinomas, from 3 human lung adenocarcinoma cell lines, and from normal lung tissue were analyzed. In most tumors, expression of K-ras was detected at levels equal to or several times higher than those found in normal lung tissue. A lung metastasis from a colon adenocarcinoma. known to contain an amplified K-ras gene, highly overexpressed the K-ras gene. In those tumors in which the K-ras oncogene was activated by a point mutation, both alleles of the gene were expressed. Our results show that a high overexpression of K-ras is a rare event in human lung carcinomas, but that a certain degree of over-expression of the mutated allele can be demonstrated in tumors with an activated K-ras gene. With the technique we describe here, adequate estimation of the expressionof specific genes in minimal amounts of tumor cells becomes possible.
Oncogenes of the ras family have been implicated in the pathogenesis of a wide variety of human and animal tumors (for reviews see Barbacid, 1987; Bos, 1988). Three of these genes have been extensively characterized: H-ras, K-ras and N-ras. All 3 code for closely related 21-kDa proteins (p21) with GTP-binding activity and a degree of homology with G proteins. Consequently, the normal ras proteins are believed to have a role in the transduction of growth signals. The ras proteins acquire transforming potential when point mutations in the genes lead to amino acid substitutions in one of the positions 12, 13 or 61, presumably because alterations at these sites are associated with loss of the intrinsic GTPase activity (Trahey and McCormick, 1987). The frequencies at which ras mutations can be demonstrated in uncultured human solid tumors depend on the tumor type of origin. Mutations in ras oncogenes are rare or do not occur in squamous-cell carcinomas of most organs or in adenocarcinomas of breast, stomach and ovary (Bos, 1988; Kraus et al., 1984; Sakato et al., 1984; Van de Vijver et al., 1989). K-ras activations occur in 30-50% of adenocarcinomas of the colon and lung (Rodenhuis er al., 1987; Forrester et al., 1987), and are almost invariably present in pancreatic cancers (Almoguera et al., 1988; Smit et al., 1988). The causes of this unequal distribution are unclear. We have previously shown that mutational activation of ras genes in NSCLC invariably occurs in codon 12 of K-ras (Rodenhuis et al., 1987). Furthermore, these mutations are restricted to the adenocarcinoma subtype, in which they occur in about one-third of all cases (Rodenhuis et al., 1988). Adenocarcinomas with mutations in K-ras have a different biological behavior from those without: in patients with radically resected tumors K-ras mutations predict early relapse and poor survival (Slebos et al., 1990). Because only one-third of adenocarcinomas of the lung harbor activating mutations in K-ras, the
possibility remains that ras genes are activated by other mechanisms in non-mutated cells. We investigated possible activation of ras by DNA amplification, but found no evidence for such a mechanism in NSCLCs (Slebos et al., 1989). In NIH3T3 cells, over-expression of K-ras can lead to malignant transformation (Pulciani et al., 1985), which raises the question whether K-ras could also be activated by increased expression in the absence of gene amplification in in vivo lung tumors. Due to usually limited sample size, and the requirement for special measures to prevent RNA degradation, assessing RNA expression in human tumors is not always possible. With the development of the polymerase chain reaction (PCR) (Mullis et al., 1987), these limitations have largely been overcome. To date, several PCR-based approaches to detect RNA expression in small samples have been described, all based on the synthesis of cDNA from cellular RNA preparations, followed by simultaneous amplification of an unrelated gene as an internal control together with the sequences of interest (Wang et al., 1989; Becker-Andr6 and. Hahlbrock, 1989). Under these conditions, the ratio between 2 different DNA fragments remains equal during the amplification reaction (Frye et al., 1989; Rappolee er al., 1989). Using this type of approach, we describe the determination of K-ras expression in NSCLC which, together with a mutation-specific oligonucleotide hybridization assay, allows simultaneous detection of the normal and the mutated allele of K-ras. Our findings show that K-ras is being actively transcribed in adenocarcinomas of the lung, and are consistent with a dominant role for the mutationally activated K-ras oncogene. MATERIAL AND METHODS
Tumor material and cell lines Tumor tissue was snap-frozen after surgery and stored at - 70°C. We investigated: 19 adenocarcinomas, 3 large-cell carcinomas, 2 squamous carcinomas, 2 lung metastases of colonic origin, normal lung tissue, and 3 lung adenocarcinoma cell lines. All tissues were carefully trimmed to remove as much non-neoplastic tissue as possible. From both sides of the tissue blocks 5-pm slices were cut and stained with eosin and hematoxylin. From these slides the percentage of tumor cells in each tissue sample was estimated. In most cases more than 70% of the sample consisted of tumor cells, while none of the samples had less than 50% tumor cells. Three human lung adenocarcinoma cell lines were investigated in this study: GLC-A1 (kindly donated by Dr. L. de Lxij, University of Groningen), HCLH23 (Carney et al., 1985) and A549 (Lieber et al., 1976); the latter only harbors the K-ras oncogene mutated at codon 12 (Valenzuela and Groffen, 1986). RNA isolation and cDNA synthesis Isolation of RNA was performed according to the guanidi-
'To whom correspondence should be addressed.
Received: November 1 1, 1990 and in revised form December 21, 1990.
52
SLEBOS ET AL.
nium isothiocyanate protocol (Chirgwin et al., 1979). Briefly, between 50 and 300 mg of tumor tissue were homogenized in a buffer containing 4 M guanidinium isothiocyanate, 5 mM sodium citrate, 100 mM P-mercaptoethanol and 0.5% sodium laurolylsarcoside, layered on a 5.7 M CsCl, cushion in 100 m~ EDTA and centrifuged for 18 hr at 60,000 g. The RNA pellet was dissolved in 10 m~ TRIS-HCI, 5 m~ EDTA and 1% SDS, extracted with chloroform/butanol (24: 1) and precipitated with ethanol. RNA yield was between 0.5 and 2 pg of total RNA per mg tissue. cDNA was primed using specific oligonucleotides for exon 2 of the K-rus gene (5’-ACTTGCTTCCTGTAGGAATC-3‘) (McGrath et al., 1983) and exon 3 of the (Persico G6PD gene (5’-ATGGTGGGGTAGATCTTCTT-3’) et al., 1986; Martini et al., 1986) the latter serving as an internal control. A typical cDNA reaction contained 5 pg of total RNA in 50 m~ TRIS-HCl PH 8.3, 40 m~ KCl, 8 m~ MgCI,, 0.25 mM sodium pyrophosphate, 100 p~ of each dNTP, 1 p~ of each cDNA primer, 10,OOO U/ml RNasin and 1000 U/ml AMV reverse transcriptase (both from BoehringerMannheim, Germany) in a 50-pl volume. Incubation was carried out for 1 hr at 37°C. Amplijication with the polymerase chain reaction From every cDNA sample, 5 10-fold dilutions were made starting with the amount of cDNA synthesized on 100 ng of mRNA. These dilutions were amplified after denaturation for 5 min at 100°C in a 50-pI reaction in 10 mM TRIS-HCl PH 8.8, 50 m~ KCl, 3 m~ MgCI,, 0.1% w/v gelatin, 100 p~ of each dNTP, 1 p~ of each oligonucleotide, and 20 U/ml Taq polymerase (Perkin Elmer Cetus, Vaterstetten, Germany). Amplification was performed at 92°C for 30 sec, at 55°C for 30 sec and at 70°C for 1 min, in a Dri-block device (Techne, Cambridge, UK). Sequences of the K-ras amplimers were: 5’-GACTGAGTATAAACTTGTGG-3’ and 5’-CCTGTAGGAATCCTCTATTG-3’, and for G6PD: 5’-CGGGATCCTGCGGGAAGAGC-3‘ and 5’-GGCCAGGTCACCCGATGCAC-3’. After 15 amplification cycles, new enzyme was added and another 15 cycles were performed. Detection of PCR reaction products The amplified sequences were slot-blotted on a nylon membrane (Hybond, Amersham, Little Chalfont, UK) with a slotblot apparatus (Schleicher and Schuell, Dassel, Germany). Detection of amplified sequences on slot-blots was done with 32P-labelled oligonucleotides specific for K-ras (wild-type): 5’-CCTACGCCACCAGCTCCAAC-3’and G 6 P D 5’-GATGGAAGGCATCGCCCTGG-3’ as described previously (Verlaan-de Vries et al., 1986). Hybridization of the slot-blots was done at 56°C in 3 M tetramethylammonium chloride (TMACl) (Aldrich, Beerse, Belgium), which makes the stability of the formed hybrids independent of their GC content (Wood et ul., 1985). Selective washing was done at 6042°C in 3 M TMACl for 30 min. Under these conditions, only perfectly matched oligonucleotides will remain hybridized to the
filter, which proves the identity of the amplified sequences. The housekeeping gene glucose-6-phosphate dehydrogenase was chosen as an internrfl standard, because it is expressed at levels comparable to those of K-rus (Hirono and Beutler, 1989). Detection was done by autoradiography using Kodak XAR-5 films. RESULTS
Amplification of cDNA To determine the conditions that allow a near-linear amplification of cDNA, we amplified different amounts of a Kras-containing plasmid pCD-Ki-ras2-76 (McCoy et ul., 1984). The reaction was started with 4 different amounts of plasmid: 0.016 ng, 0.125 ng, I ng and 8 ng, which were amplified for 3N (N = 0, I , . . . . , 7) PCR cycles. Theoretically, each series of 3 reactions should yield an increase of DNA by a factor of 8. As shown in Figure I , initially the signal increases as expected, but at higher cycle numbers a plateau is reached. Under the conditions used, near-linear amplification is achievable up to 15 cycles. When the same series of amplifications was performed, starting with picogram quantities of DNA, adding new enzyme after an initial 15 cycles, linearity was preserved (not shown). Thus, linear amplification conditions could be maintained with 2 series of 15 PCR cycles each. Using this approach, the detection of picogram amounts of cDNA, and thus picogram amounts of messenger RNA, is possible. It is reasonable to assume that a moderately expressed messenger RNA such as K-ras may account for about of total cellular RNAs, equivalent to picogram quantities of specific mRNA in 1 pg of total RNA. In serial dilutions, PCR products can be detected at levels of initial cDNA at which the plateau of the reaction has not yet been reached. To be able to correct for RNA degradation, differences in RNA yield or PCR efficiency, the housekeeping gene encoding glucose-6-phosphate dehydrogenase (G6PD) was reversetranscribed and amplified simultaneously with the K-ras sequences. Using specific oligonucleotides, 90-and 119-nt-long stretches of DNA from the G6PD and K-rus genes, respectively, will be amplified. Both the cDNA primers and the 3‘-amplimers were chosen in such a way that the amplified 90 and 119 nt fragments are interrupted by intron sequences more than 10 kb in length in genomic DNA. Thus, DNA sequences which might contaminate the RNA preparations cannot coamplify in the reaction. Because the amplified sequences are relatively small, limited degradation of RNA is acceptable for this technique. Separation of the PCR products on a 10% polyacrylamide gel confirmed that the expected 90 nt G6PD and 119 nt K-rus fragments were amplified (Fig. 2). Under the conditions used, G6PD amplifies linearly, which becomes apparent from Figures 4, 5 and 6, where 10-fold dilutions of the cDNAs result in equivalent changes in the signal obtained with the G6PD hybridization.
pCDKi-mS-76 0.018
ng
0.125
ng
1.0
ng
8.0
ng
cydes: 0 3 6 9 12 15 18 21 FIGURE1 - Hybridization of probe K12-gly with amplified sequences from plasmid pCD-Ki-ras2-76, showing the initial amounts of plasmid used for amplification (0.016-8 ng) and the number of PCR cycles that were performed (0 to 21). Signal intensity increased linearly up to 15 cycles whereafter a plateau was reached.
53
K-ras EXPRESSION IN LUNG TUMORS
nt
pBR322 9
xHpall DNA
1
8
16
18 20
21
51
-
- pCD-76 K-ras
Plasmid:
pCD-Ki-ras-76
PCD-SW-11 .1
100%
0%
75%
25%
50%
50%
25%
75%
0%
100%
Probe: K12-gly K12-val FIGURE3 - Simultaneous amplification of different admixtures of
147.
122. 110.
90-
76 -
FIGURE2 - Amplified cDNAs were end-labelled and separated on a PAA gel. Lane 1; marker pBR322 cut with HpaII: lane 2; one p,g of genomic DNA was reverse transcribed and amplified as described for RNA: lanes 3 to 9; lung tumor samples (the sample in lane 6 is from the same tumor as no. 2 in Figure 5): lane 12; amplification of 1 ng of plasmid pCD-Ki-ras2-76 (McCoy er al., 1984) with the sequence 5’-AC’ITGClTCCTGTAGGAATC-3‘ as a 3’-amplimer, amplifying a 127-nt-long PCR fragment.
Allele-specijicdetection of K-ras expression To investigate the separate detection of both alleles of K-rus, a total of 1 ng of 2 plasmids, pCD-K-rus2-76 and pCDSW-11.1, encoding the wild-type “GGT” and the mutated “GIT” at codon 12, respectively (McCoy era!., 1984), were mixed in different ratios and amplified for 15 cycles (Fig. 3). The result of this experiment demonstrated that the relative initial amounts of both plasmids are accurately reflected in the differential oligonucleotide hybridization on the amplified PCR products. To test this approach on total cellular RNA, 3 adenocarcinoma cell lines were investigated: A549, NCLH23, and GLC-A1 . The cell line A549 is known to harbor a “GGT” to “GAT” mutation at codon 12 which results in a gly to ser mutation in the encoded K-rus protein, while it has lost the wild-type allele of K-rus (Valenzuela and Groffen, 1986). The NCl-H23 cell line was heterozygous for a “GGT” to “TGT” mutation in codon 12 of K-rus (gly to cys), and expressed both
plasmids pCD-Krus-76 and pCD-SW- 1 1 . 1 , containing wild-type and mutated K-ras cDNAs, respectively. Starting with a total amount of 1 ng of plasmid DNA, 15 PCR cycles were performed, after which each PCR reaction was slot-blotted onto 2 separate membranes. Hybridizations with probe K12-gly and K12-val indicate that the initial ratio of both plasmids is retained during the PCR.
alleles of K-rus (see below). No mutations were found in any of the 3 rus genes in the cell line GLC-A1 (not shown). From these cell lines, 100 ng of total RNA were reverse transcribed and serially diluted in 10-fold steps. The cDNA was amplified in twice 15 cycles and slot-blotted on separate membranes. The blots were then hybridized to oligonucleotides specific for Krus wild-type, K-rus mutated and G6PD (Fig. 4). The signal obtained for G6PD is comparable in all 3 samples, indicating the integrity of the initial RNAs. The cell line GLC-A1 has K-rus levels about 10 times lower than those of G6PD. NCIH23 shows higher levels of K-rus, the mutated allele expressed preferentially at levels about 10 times higher than the wild-type allele. The cell line A549 has no wild-type allele, and thus only expression of the K-rus mutated allele was found. Detection of K-ras expression in NSCLCs RNA from 10 adenocarcinomas, 3 large-cell carcinomas, 2 squamous-cell carcinomas, 2 lung metastases of colonic primaries, and from normal lung tissue was analyzed (Table I). None of these samples had mutations in the K-rus oncogene (Rodenhuis et ul., 1987, 1988). Figure 5 shows hybridizations of amplified cDNAs from 6 representative lung tumors and from normal lung tissue, with the wild-type K-rus and G6PD detecting oligonucleotides. In all preparations sufficient amplification of the G6PD fragment was achieved, indicating the integrity of the starting material. Expression of G6PD does not vary significantly between the samples. The signal obtained with the K-rus probe in samples 33, 106 and normal lung tissue is about 10 times lower than that with the G6PD probe, while sample 58 has about 100-fold lower levels than G6PD. Samples 8 and 21 show equal levels with both probes, while in a lung metastasis from an adenocarcinoma of the colon (sample 2), in which we previously detected a 15- to 20-fold amplification of the K-rus oncogene (Rodenhuis er ul., 1987), the K-rus signal is about 10 times higher than the G6PD signal. Table II summarizes the results with tumors without mutations in codon 12 of K-rus. In 5 of 10 adenocarcinomas without a mutation in K-rus, the levels of the oncogene were higher than those found in normal lung tissue (Table II). All 3 large-cell lung carcinomas, and 1 of the 2 investigated squamous-cell carcinomas had expression levels at least 10 times higher than those found in normal lung tissue.
54
SLEBOS ET AL. GLC-A1
NCI-HZ3
A549
19
84
4
21
106
58
K12-W
K12-Sp
G6PD
K12-wl
K12-CyS
G6PD
FIGURE6 - Slot-blots of amplified serial dilutions of cDNA from 4 human lung adenocarcinomas that harbor a mutation in K-ras. The top row on each filter reflects the signal obtained with cDNA from 100 ng of total RNA; each lower row is a 10-fold dilution of the previous one. In samples 19, 87 and 90 the signals obtained with the mutationdetecting probe (K12-va1, KI2-asp, and K12-cys respectively) are higher than those of the wild-type probe (K12-gly). Sample 84 has levels of K-ras comparable to those obtained for G6PD.
K12-W
GGPD
FIGURE4 - Slot-blots of amplified serial dilutions of cDNA from 3 lung adenocarcinoma cell lines. The top row on each filter reflects the signal obtained with cDNA from 100 ng of total RNA; each lower row is a 10-fold dilution of the previous one. Cell line GLC-A1 has no mutation in codon 12 of K-ras, but expresses the wild-type sequence at levels about 10-times lower than G6PD. NCLH23 preferentially expresses the mutated K-ras allele to at least a 10-fold higher degree than the non-mutated allele. The cell line A549 harbors only a mutated allele, which is expressed at levels comparable to those of G6PD. FIGURE5 - Slot-blots of amplified serial dilutions of cDNA from 6 human lung adenocarcinomas and normal lung tissue. The top row on each filter reflects the signal obtained with cDNA from 100 ng of total RNA; each lower row is a 10-fold dilution of the previous one. Sample 2 is from a lung metastasis of a colonic adenocarcinoma with a 15- to 20-fold amplification of the K-ras oncogene. The K-ras signal in this sample is about 10 times higher than the G6PD signal. Sample 58 gave a K-ras signal which is at least 100 times lower than G6PD. samples 33 and 106 had about 10 times lower, while samples 8 and 21 show comparable levels of K-ras and G6PD. For comparison, the signals obtained with normal lung tissue are shown at the bottom of the Figure.
Using mutation-specific oligonucleotides, we were able to discriminate between the expression of the 2 alleles of K-rus in 9 adenocarcinomas (Table I) that had previously been shown to harbor mutations in codon 12 of K-rus. Detection of expression of both the wild-type and the mutated allele in 4 adenocarcinomas with a point mutation in K-rus is shown in Figure 6. In sample 84 the signal obtained with the mutation-detecting
probe is at least equal to that of the wild-type probe, while in the other samples the signals of the mutated allele are about 10 times higher than that obtained with the wild-type probe. In contrast to the results obtained with the cell lines, these results are influenced by nonmalignant cells present in the original tumor sample that may contribute to the wild-type K-rus signal. In summary, one adenocarcinoma had at least 10-fold lower levels of the mutated allele compared to the wild-type allele (not shown), 5 had equal levels of both alleles, while 3 samples showed at least 10-fold higher expression of the mutated allele than of the wild-type allele. DISCUSSION
Mutational activation of K-rus occurs in about one-thud of adenocarcinomas of the lung and is only very infrequently encountered in other subtypes of NSCLC (Rodenhuis et ul., 1987, 1988). The explanation for two-thirds of lung adenocarcinomas not having a mutationally-activated rus gene may well be that other genetic or epigenetic alterations may be substituted for the mutation. One likely cause of such an alternative activation was over-expression of the normal K-rus gene. Apart from oncogenic activation by point mutation, rus genes can become transforming in vitro when expressed at sufficiently high levels, either because of the action of a strong promoter (Chang et ul., 1982), or as a result of the incorporation of multiple copies in the genome of the recipient cell (Pulciani et ul., 1985). Furthermore, amplified rus genes have been detected in several human tumors (Barbacid, 1987). Amplification of the K-rus oncogene has been reported in a single metastatic lung carcinoma, but was not found in 25 primary lung tumors (Heighway and Hasleton, 1986), or in the tumors reported here (Slebos et ul., 1989). Detection methods based on cDNA synthesis followed by PCR have opened new possibilities for the detection of specific messenger RNAs in very small numbers of cells (Rappolee et ul., 1989; Noonan and Roninson, 1990). Several approaches have been developed, all of which are based on quantitative
55
K-ras EXPRESSION IN LUNG TUMORS TABLE I - HISTOLOGICAL CLASSIFICATION OF INVESTIGATED TISSUES AND CELL LINES Cancer
Number
Adenocarcinoma Large-cell carcinoma Squamous-cell carcinoma Solitary metastasis of colonic origin Adenocarcinoma cell line Normal lung tissue Total
19 3 2 2 3 I 30
TABLE 11 - EXPRESSION OF THE K-ras ONCOGENE IN LUNG TUMORS WITHOUT MUTATIONS IN C O W N 12 OF K-ras DETECTED IN SERIAL DILUTIONS OF cDNA. THE TABLE SHOWS THE NUMBER OF SAMPLES WITH SIGNAL LEVELS OF K-ras COMPARED TO THOSE OBTAINED FOR G6PD Histological type
Adenocarcinomas Large-cell carcinomas Squamous-cell carcinomas Lung metastases of colonic adenocarcinoma Cell line GLC-A1
K-ras expression levels relative to G6PD expression
lo-* 1 -
lo-’ 4
1 5
-
-
1
-
-
3 1 1
-
-
1
-
10’ 1
Total 10 3 2 2 1
amplification of cDNA. One strategy uses the addition of known amounts of synthetic RNAs to the RNA preparations to be investigated, thus providing an absolute assessment of expression levels (Becker-Andr6 and Hahlbrock, 1989; Wang et ul., 1989). However, this method does not correct for the degradation of RNA, a common feature of RNA preparations from human tumor tissue. When ubiquitously expressed genes (“housekeeping genes”) are used as an internal standard, and co-amplified together with the genes of interest in a single PCR, correction for limited degradation of RNA, or failure of the PCR, is possible (Rappolee e t a / . , 1989; Noonan and Roninson, 1990). To employ an internal standard that is expressed at levels comparable to those of K-rus, we used the housekeeping gene glucose-6-phosphate dehydrogenase, a key enzyme in the pentose phosphate pathway (Luzzatto and Battistuzzi, 1984). The G6PD gene is expressed at moderate levels in all human tissues thus far analyzed (Hirono and Beutler, 1989; Persico e t a / . , 1986). It is important to select an internal standard defining which mRNA levels are comparable to those of the gene of interest, since it is otherwise difficult to assure linearity of the amplification reaction for both substrates. Thus, for every individual gene that is to be investigated by quantitative PCR, a suitable control should be selected. We compared the expression levels to those in normal lung tissue and to those in a tumor containing a 15- to 20-fold K-rus amplification expected to have high levels of K-rus. Our results indicate that activation of K-rus by high overexpression is rare in human NSCLC, and is not likely to be a pathogenetic alternative for activation by codon 12 point mutations in adenocarcinomas of the lung. This is further supported by the finding that expression levels in tumors with and without a mutation in K-rus codon 12 were similar, and significantly lower than those in a tumor with a K-rus gene am-
plification. Thus, we believe that the moderately increased K-rus expression, as compared to expression in normal lung tissue, may be a consequence of proliferative activity rather than representing one of the transforming events in NSCLC. The specificity of K-rus codon 12 point mutations for adenocarcinomas is not reflected in the expression pattern of the gene: K-rus is expressed at equal levels in all histological subtypes of NSCLC. This would suggest that mutations in K-rus would result in an equally strong transforming signal in all types of NSCLC, and leaves the reason for the specificity of these mutations for the adenocarcinoma subtype unexplained. The detection of normal and mutant K-rus transcripts, using the RNaseA mismatch cleavage method, has been described for several human cell lines. In the cell lines Calu-1 and PR371 mutant K-rus transcripts were 5-10 times more abundant than their normal counterparts (Winter et ul., 1985), a difference that was also reflected in the relative abundance of the normal and mutated alleles in a cDNA library from Calu-1 (Capon et a/.,1983). In the case of PR371, the mutant K-rus allele also showed genomic DNA amplification (Winter et ul., 1985). Moreover, even loss of the wild-type allele in tumor cell lines with K-rus mutations has been described in several cases (Valenzuela and Groffen, 1986; Santos et ul., 1984). We investigated 2 K-rus-mutation-positive lung adenocarcinoma cell lines, and observed expression of the mutated allele in both. The NCl-H23 cell line had higher levels of the mutated allele than of the wild-type allele, while the A549 cell line did not contain a normal K-rus allele (Valenzuela and Groffen, 1986), and hence only expressed the mutated allele. The assessment of (onco)gene expression in fresh human tumors is often complicated by the presence of non-malignant cells (e.g., inflammatory cells) in the sample. When K-rus expression is determined by an allele-specific method, as described here, these normal cells could contribute to the K-rus wild-type signal. Consequently, the mutated allele was expressed preferentially only in the 3 samples in which the signal obtained with the mutationdetecting probe was higher than that found with the wild-type probe. It is likely that in these samples a selection for tumor cells with a preferential expression of the mutated allele of K-rus must have occurred. In the 5 cases in which equal levels of the wild-type and mutated allele were found, a preferential expression of the mutated allele could not be excluded. In colonic adenocarcinomas also, the relative amounts of normal and mutated K-rus mRNAs, detected by the RNaseA mismatch cleavage method, were comparable (Forrester et ul., 1987). These results provide further evidence for the hypothesis that the K-rus oncogene acts in a dominant manner in these tumors. The assay we describe here can be used to determine expression levels of other oncogenes on the basis of minimal amounts of tumor tissue, such as bronchial biopsies or cytological aspiration samples. ACKNOWLEDGEMENTS
We thank Drs. S.M. Bellot, P. Blok, J.A. Van Der Haar, N.A. Den Hartog, Th.M. Van Leeuwen, T.M. Vroom, and Sj.S. Wagenaar for selecting and providing the tumor samples. This research was supported by grants NKI 87-15 and NKI 88-7 from the Dutch Cancer Society KWF.
REFERENCES ALMOGUERA, C . , SHIBATA, D . , FORRESTER, K., MARTIN, J . , ARNHEIM, Bos, J.L., The rus gene family and human carcinogenesis.Mururion Res., N. and PERUCHO, M., Most human carcinomas of the exocrine pancreas 195, 255-271 (1988). contain mutant c-K-rus genes. Cell. 53, 549-554 (1988). CAPON,D.J., SEEBURG,P.H., MCGRATH, J.P., HAYFLICK, J.S., EDMAN, U . , LEVINSON, A.D. and GOEDDEL, D.V., Activation of Ki-rus 2 gene in BARBACID,M., rus genes. Ann. Rev. Biochern., 56, 779-827 (1987). human colon and lung carcinomas by two different point mutations. Nurure BECKER-ANDRB, M . and HAHLBROCK, K., Absolute mRNA quantification I L ~ ~ 304, , ) , 507-513 (1983). using the polymerase chain reaction (PCR). A novel approach by a PCR CARNEY, D.N., GAZDAR, A.F., BEPLER, G . , GUCCION,J.G., MARANGOS, aided transcript titration assay (PA’lTY). Nucl. Acids Res.. 127, 9437P.J., MOODY,T.W . , ZWEIG,M . H . and MINNA, J.D., Establishment and 9746 (1989).
56
SLEBOS ET AL.
identification of small cell lung cancer cell lines having classic and variant features. Cancer Res., 45, 2913-2923 (1985). CHANG,E.H., FURTH,M.E., SCOLNICK,E. and LOWY,D.R., Tumorigenic transformation of mammalian cells induced by a normal human gene homologous to the oncogene of Harvey murine sarcoma virus. Nature (Lond.J, 297,479484 (1982). CHIRGWIN,J.M., PRZYBYLA, A.E., MACDONALD,R.J. and RUTTER, W.J., Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry, 18, 52965299 (1979). FORRESTER,K., ALMOGUERA, C., HAN, K., GRIZZLE,W.E. and PERUCHO,M., Detection of high incidence of K-ras oncogenes during human colon tumorigenesis. Nature (Lond.). 327, 298-303 (1987). FRYE,R.A., BENZ, C.C. and LIU, E., Detection of amplified oncogenes by differential polymerase chain reaction. Oncogene, 4, 1153-1 157 (1989). HEIGHWAY, J. and HASLETON, P.S., c-Ki-ras amplification in human lung cancer. Erir. J . Cancer, 53, 285-287 (1986). WRONO,A. and BEUTLER,E., Alternative splicing of human glucose6-phosphate dehydrogenase messenger RNA in different tissues. J. clin. Invest.. 83, 343-346 (1989). KRAUS,M.H., YUASA,Y. and AARONSON, S.A., A position 12-activated H-ras oncogene in all HS578T mammary carcinosarcoma cells but not in normal mammary cells of the same patient. Proc. nat. Acad. Sci. (Wash.J, 81, 5384-5388 (1984). LIEBER,M., SMITH,B., SZAKAL,A., NELSON-REES,W. and TODARO, G . , A continuous tumor-cell line from a human lung carcinoma with properties of type I1 alveolar epithelial cells. Int. J. Cancer, 17, 62-70 (1976). LUZZATTO,L. and BATTISTUZZI,G . , Glucose-6-phosphate dehydrogenase. Advanc. hum. Gener., 14, 217-229 (1984). MARTINI,G., TONIOLO, D., VULLIAMY, T., LUZZATTO,L., DONO, R., G., D'URSO, M. and ~ R S I C O M.G., , StrucVIGLIETTO,G., PAONESSA, tural analysis of the X-linked gene encoding human glucose-6-phosphate dehydrogenase. EMBO J., 5, 1849-1855 (1986). McCoy, M.S., BARGMANN, C.I. and WEINBERG,R.A., Human colon carcinoma Ki-rus2 oncogene and its corresponding proto-oncogene. Mol. cell. Biol.. 4, 1577-1582 (1984). MCGRATH,J.P., CAPON, D.J., SMITH, D.H., CHEN, E.Y., SEEBURG, P.H.,GOEDDEL,D.V. and LEVINSON, A.D., Structure and organization of the human Ki-ras proto-oncogene and a related processed pseudogene. Nature (Lond.), 304, 501-506 (1983). MULLIS,K.B. and FALOONA, F.A., Specific synthesis of DNA in vitro via a polymerase catalyzed chain reaction. Merh. Enzymol., 155, 335-350 (1987). NOONAN,K.E. and RONINSON,I.B., Quantitative estimation of mdrl mRNA level by polymerase chain reaction. In: I.B. Roninson (ed.), Molecular and cellular biology of mulridrug-resistance in tumor cells. Plenum, New York (1990). (In press). ~ R S I C O M.G., , VIGLIETTO,G., MARTINI,G., TONIOLO, D., PAONESSA, G., MOSCATELLI, C., DONO, R., VULLIAMY,T., LUZZATTO,L. and D'URSO, M., Isolation of human glucose-6-phosphate dehydrogenast (G6PD) cDNA clones: primary structure of the protein and unusual 5 non-coding region. Nucl. Acids Res., 14, 2511-2522 (1986). PULCIANI, S., SANTOS,E., LONG,L.K., VICENZO,S. and BARBACID, M., ras gene amplification and malignant transformation. Mol. cell. Eiol., 5, 2836-2841 (1985).
RAPPOLEE,D.A., WANG,A,, MARK,D. and WERB,Z., Novel method for studying mRNA phenotypes in single or small number of cells. J . cell. Biochem.. 39, 1-11 (1989). RODENHUIS, S., SLEBOS,R.J.C., BOOT, A.J.M., EVERS,S.E., MOO], S J . ~ VAN . , BODEGOM, P.CH. and Bos, J.L., Incidence W.J., WAGENAAR, and possible clinical significance of K-ras oncogene activation in adenocarcinoma of the human lung. Cancer Res., 48, 5738-5741 (1988). RODENHUIS, S., VANDE WETERING,M.L., MOOI,W.J., EVERS,S.G., VAN ZANDWIJK,N. and Bos, J.L.,Mutational activation of the K-ras oncogene: a possible pathogenic factor in adenocarcinoma of the lung. N. Engl. J. Med., 317, 929-935 (1987). SAKATO, H., REDDY,E.P., PLJLCIANI, S., FELDMAN, R.J. and BARBACID, M., Transforming gene from human stomach cancers and a noncancerous portion of stomach mucosa. Proc. nat. Acad. Sci. (Wash.), 83,3997-4001 ( 1984). SANTOS,E., MARTIN-ZANCA, D., REDDY,E.P., PIEROTTI,M.A., DELLA M., Malignant activation of a K-ras oncogene PORTA,G. and BARBACID, in lung carcinoma but not in normal tissue of the same patient. Science, 223, 661464 (1984). SLEBOS,R.J.C., EVERS,S.E., WAGENAAR, SJ.S. and RODENHUIS,S., Cellular proto-oncogenes are infrequently amplified in untreated nonsmall-cell lung cancer. Brit. J . Cancer, 59, 76-80 (1989). SLEBOS,R.J.C., KIBBELAAR, R.E., DALESIO,O., KOOISTRA, A,, STAM, J . , MEIJER, C.J.L.M., WAGENAAR, S J . ~ . ,VANDERSCHUEREN, R.G.J.R.A., VAN ZANDWIJK,N., MOOI, W.J., Bos, J.L. and RODENHUIS, S., Kirsten-ras oncogene activation as a prognostic marker in adenocarcinoma of the lung. N. Engl. J . Med., 323, 561-565 (1990). SMIT, V.T.H.B.M., BOOT, A.J.M., SMITS, A.M.M., FLEUREN,G.J., CORNELISSE, C.J. and Bos, J.L., Kras codon 12 mutations occur very frequently in pancreatic adenocarcinomas. Nucl. Acids Res., 16, 77737782 (1988). TRAHEY,M. and MCCORMICK,F., A cytoplasmic protein stimulates normal N-ras p21 GTPase, but does not affect oncogenic mutants. Science, 238,542-545 (1987). VALENZUELA, D.M. and GROFFEN,J., Four human carcinoma cell lines with novel mutations in position 12 of c-K-ras oncogene. Nucl. Acids Res.. 14, 843-852 (1986). VAN DE VIJVER,M.J., PETERSE,J.L., MOOI, W.J., LOMANS,J., VERBRUGGEN, M., VAN DE BERSSELAAR, R., DEVILEE,P., CORNELISSE, c., J. and NUSSE,R., Oncogene activations in human BOS, J.L., YARNOLD, breast cancer. Cancer Cells, 7, 385-391 (1989). VERLAAN-DE VRIES,M., BOGAARD,M.E., ELST, H., VANBOOM,J.H., VAN DER EB, A.J. and Bos, J.L., A dot-blot screening procedure for mutated ras oncogenes using synthetic oligodeoxynucleotides. Gene, 50, 313-320 (1986). WANG,A.M., DOYLE,M.V. and MARK,D.F., Quantitation of mRNA by the polymerase chain reaction. Proc. nar. Acad. Sci. (Wash.), 86, 97179721 (1989). WINTER,E., YAMAMOTO, F., ALMOGUERA, C. and PF~RUCHO, M., A method to detect and characterize point mutations in transcribed genes: amplification and overexpression of the mutated c-Ki-ras allele in human tumor cells. Proc. nut. Acad. Sci. (Wash.), 82, 7575-7579 (1985). WOOD,W.I., GITSCHIER,J . , LASKY,L.A. and LAWN,R.M., Base composition-independent hybridization in tetramethylammonium chloride: a method for oligonucleotide screening of highly complex gene libraries. Proc. nat. Acad. Sci. (Wash.), 82, 1585-1588 (1985).
Publication of the International Union Against Cancer Publication de I'Union Internationale Contre le Cancer
ALLELE-SPECIFIC DETECTION OF K-rus ONCOGENE EXPRESSION IN HUMAN NON-SMALL-CELL LUNG CARCINOMAS Robert J.C. SLEBOS',Gaston G.M. HABETS,Siegina G. EVERS,Wolter J. Moo1 and Sjoerd RODENHUIS Departments of Experimental Therapy, Medical Oncology and Cell Biology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands. Point mutations in codon I2 of the K-ras oncogene are frequent in human lung adenourcinomas. To study the expression of the K-ras gene in these tumors we have developed a mRNA detection technique based on the polymerase chain reaction (PCR). By this technique, K-ras expression can be detected semi-quantitatively in samples of less than 100 ng total RNA. Hybridization of the amplified cDNA sequences with mutation-specificoligonucleotides allows separate quantification of the expression of normal and point-mutated alleles in a single sample. RNA samples from 24 human nonsmall-cell lung carcinomas (NSCLC), from 2 lung metastases of colonic adenocarcinomas, from 3 human lung adenocarcinoma cell lines, and from normal lung tissue were analyzed. In most tumors, expression of K-ras was detected at levels equal to or several times higher than those found in normal lung tissue. A lung metastasis from a colon adenocarcinoma. known to contain an amplified K-ras gene, highly overexpressed the K-ras gene. In those tumors in which the K-ras oncogene was activated by a point mutation, both alleles of the gene were expressed. Our results show that a high overexpression of K-ras is a rare event in human lung carcinomas, but that a certain degree of over-expression of the mutated allele can be demonstrated in tumors with an activated K-ras gene. With the technique we describe here, adequate estimation of the expressionof specific genes in minimal amounts of tumor cells becomes possible.
Oncogenes of the ras family have been implicated in the pathogenesis of a wide variety of human and animal tumors (for reviews see Barbacid, 1987; Bos, 1988). Three of these genes have been extensively characterized: H-ras, K-ras and N-ras. All 3 code for closely related 21-kDa proteins (p21) with GTP-binding activity and a degree of homology with G proteins. Consequently, the normal ras proteins are believed to have a role in the transduction of growth signals. The ras proteins acquire transforming potential when point mutations in the genes lead to amino acid substitutions in one of the positions 12, 13 or 61, presumably because alterations at these sites are associated with loss of the intrinsic GTPase activity (Trahey and McCormick, 1987). The frequencies at which ras mutations can be demonstrated in uncultured human solid tumors depend on the tumor type of origin. Mutations in ras oncogenes are rare or do not occur in squamous-cell carcinomas of most organs or in adenocarcinomas of breast, stomach and ovary (Bos, 1988; Kraus et al., 1984; Sakato et al., 1984; Van de Vijver et al., 1989). K-ras activations occur in 30-50% of adenocarcinomas of the colon and lung (Rodenhuis er al., 1987; Forrester et al., 1987), and are almost invariably present in pancreatic cancers (Almoguera et al., 1988; Smit et al., 1988). The causes of this unequal distribution are unclear. We have previously shown that mutational activation of ras genes in NSCLC invariably occurs in codon 12 of K-ras (Rodenhuis et al., 1987). Furthermore, these mutations are restricted to the adenocarcinoma subtype, in which they occur in about one-third of all cases (Rodenhuis et al., 1988). Adenocarcinomas with mutations in K-ras have a different biological behavior from those without: in patients with radically resected tumors K-ras mutations predict early relapse and poor survival (Slebos et al., 1990). Because only one-third of adenocarcinomas of the lung harbor activating mutations in K-ras, the
possibility remains that ras genes are activated by other mechanisms in non-mutated cells. We investigated possible activation of ras by DNA amplification, but found no evidence for such a mechanism in NSCLCs (Slebos et al., 1989). In NIH3T3 cells, over-expression of K-ras can lead to malignant transformation (Pulciani et al., 1985), which raises the question whether K-ras could also be activated by increased expression in the absence of gene amplification in in vivo lung tumors. Due to usually limited sample size, and the requirement for special measures to prevent RNA degradation, assessing RNA expression in human tumors is not always possible. With the development of the polymerase chain reaction (PCR) (Mullis et al., 1987), these limitations have largely been overcome. To date, several PCR-based approaches to detect RNA expression in small samples have been described, all based on the synthesis of cDNA from cellular RNA preparations, followed by simultaneous amplification of an unrelated gene as an internal control together with the sequences of interest (Wang et al., 1989; Becker-Andr6 and. Hahlbrock, 1989). Under these conditions, the ratio between 2 different DNA fragments remains equal during the amplification reaction (Frye et al., 1989; Rappolee er al., 1989). Using this type of approach, we describe the determination of K-ras expression in NSCLC which, together with a mutation-specific oligonucleotide hybridization assay, allows simultaneous detection of the normal and the mutated allele of K-ras. Our findings show that K-ras is being actively transcribed in adenocarcinomas of the lung, and are consistent with a dominant role for the mutationally activated K-ras oncogene. MATERIAL AND METHODS
Tumor material and cell lines Tumor tissue was snap-frozen after surgery and stored at - 70°C. We investigated: 19 adenocarcinomas, 3 large-cell carcinomas, 2 squamous carcinomas, 2 lung metastases of colonic origin, normal lung tissue, and 3 lung adenocarcinoma cell lines. All tissues were carefully trimmed to remove as much non-neoplastic tissue as possible. From both sides of the tissue blocks 5-pm slices were cut and stained with eosin and hematoxylin. From these slides the percentage of tumor cells in each tissue sample was estimated. In most cases more than 70% of the sample consisted of tumor cells, while none of the samples had less than 50% tumor cells. Three human lung adenocarcinoma cell lines were investigated in this study: GLC-A1 (kindly donated by Dr. L. de Lxij, University of Groningen), HCLH23 (Carney et al., 1985) and A549 (Lieber et al., 1976); the latter only harbors the K-ras oncogene mutated at codon 12 (Valenzuela and Groffen, 1986). RNA isolation and cDNA synthesis Isolation of RNA was performed according to the guanidi-
'To whom correspondence should be addressed.
Received: November 1 1, 1990 and in revised form December 21, 1990.
52
SLEBOS ET AL.
nium isothiocyanate protocol (Chirgwin et al., 1979). Briefly, between 50 and 300 mg of tumor tissue were homogenized in a buffer containing 4 M guanidinium isothiocyanate, 5 mM sodium citrate, 100 mM P-mercaptoethanol and 0.5% sodium laurolylsarcoside, layered on a 5.7 M CsCl, cushion in 100 m~ EDTA and centrifuged for 18 hr at 60,000 g. The RNA pellet was dissolved in 10 m~ TRIS-HCI, 5 m~ EDTA and 1% SDS, extracted with chloroform/butanol (24: 1) and precipitated with ethanol. RNA yield was between 0.5 and 2 pg of total RNA per mg tissue. cDNA was primed using specific oligonucleotides for exon 2 of the K-rus gene (5’-ACTTGCTTCCTGTAGGAATC-3‘) (McGrath et al., 1983) and exon 3 of the (Persico G6PD gene (5’-ATGGTGGGGTAGATCTTCTT-3’) et al., 1986; Martini et al., 1986) the latter serving as an internal control. A typical cDNA reaction contained 5 pg of total RNA in 50 m~ TRIS-HCl PH 8.3, 40 m~ KCl, 8 m~ MgCI,, 0.25 mM sodium pyrophosphate, 100 p~ of each dNTP, 1 p~ of each cDNA primer, 10,OOO U/ml RNasin and 1000 U/ml AMV reverse transcriptase (both from BoehringerMannheim, Germany) in a 50-pl volume. Incubation was carried out for 1 hr at 37°C. Amplijication with the polymerase chain reaction From every cDNA sample, 5 10-fold dilutions were made starting with the amount of cDNA synthesized on 100 ng of mRNA. These dilutions were amplified after denaturation for 5 min at 100°C in a 50-pI reaction in 10 mM TRIS-HCl PH 8.8, 50 m~ KCl, 3 m~ MgCI,, 0.1% w/v gelatin, 100 p~ of each dNTP, 1 p~ of each oligonucleotide, and 20 U/ml Taq polymerase (Perkin Elmer Cetus, Vaterstetten, Germany). Amplification was performed at 92°C for 30 sec, at 55°C for 30 sec and at 70°C for 1 min, in a Dri-block device (Techne, Cambridge, UK). Sequences of the K-ras amplimers were: 5’-GACTGAGTATAAACTTGTGG-3’ and 5’-CCTGTAGGAATCCTCTATTG-3’, and for G6PD: 5’-CGGGATCCTGCGGGAAGAGC-3‘ and 5’-GGCCAGGTCACCCGATGCAC-3’. After 15 amplification cycles, new enzyme was added and another 15 cycles were performed. Detection of PCR reaction products The amplified sequences were slot-blotted on a nylon membrane (Hybond, Amersham, Little Chalfont, UK) with a slotblot apparatus (Schleicher and Schuell, Dassel, Germany). Detection of amplified sequences on slot-blots was done with 32P-labelled oligonucleotides specific for K-ras (wild-type): 5’-CCTACGCCACCAGCTCCAAC-3’and G 6 P D 5’-GATGGAAGGCATCGCCCTGG-3’ as described previously (Verlaan-de Vries et al., 1986). Hybridization of the slot-blots was done at 56°C in 3 M tetramethylammonium chloride (TMACl) (Aldrich, Beerse, Belgium), which makes the stability of the formed hybrids independent of their GC content (Wood et ul., 1985). Selective washing was done at 6042°C in 3 M TMACl for 30 min. Under these conditions, only perfectly matched oligonucleotides will remain hybridized to the
filter, which proves the identity of the amplified sequences. The housekeeping gene glucose-6-phosphate dehydrogenase was chosen as an internrfl standard, because it is expressed at levels comparable to those of K-rus (Hirono and Beutler, 1989). Detection was done by autoradiography using Kodak XAR-5 films. RESULTS
Amplification of cDNA To determine the conditions that allow a near-linear amplification of cDNA, we amplified different amounts of a Kras-containing plasmid pCD-Ki-ras2-76 (McCoy et ul., 1984). The reaction was started with 4 different amounts of plasmid: 0.016 ng, 0.125 ng, I ng and 8 ng, which were amplified for 3N (N = 0, I , . . . . , 7) PCR cycles. Theoretically, each series of 3 reactions should yield an increase of DNA by a factor of 8. As shown in Figure I , initially the signal increases as expected, but at higher cycle numbers a plateau is reached. Under the conditions used, near-linear amplification is achievable up to 15 cycles. When the same series of amplifications was performed, starting with picogram quantities of DNA, adding new enzyme after an initial 15 cycles, linearity was preserved (not shown). Thus, linear amplification conditions could be maintained with 2 series of 15 PCR cycles each. Using this approach, the detection of picogram amounts of cDNA, and thus picogram amounts of messenger RNA, is possible. It is reasonable to assume that a moderately expressed messenger RNA such as K-ras may account for about of total cellular RNAs, equivalent to picogram quantities of specific mRNA in 1 pg of total RNA. In serial dilutions, PCR products can be detected at levels of initial cDNA at which the plateau of the reaction has not yet been reached. To be able to correct for RNA degradation, differences in RNA yield or PCR efficiency, the housekeeping gene encoding glucose-6-phosphate dehydrogenase (G6PD) was reversetranscribed and amplified simultaneously with the K-ras sequences. Using specific oligonucleotides, 90-and 119-nt-long stretches of DNA from the G6PD and K-rus genes, respectively, will be amplified. Both the cDNA primers and the 3‘-amplimers were chosen in such a way that the amplified 90 and 119 nt fragments are interrupted by intron sequences more than 10 kb in length in genomic DNA. Thus, DNA sequences which might contaminate the RNA preparations cannot coamplify in the reaction. Because the amplified sequences are relatively small, limited degradation of RNA is acceptable for this technique. Separation of the PCR products on a 10% polyacrylamide gel confirmed that the expected 90 nt G6PD and 119 nt K-rus fragments were amplified (Fig. 2). Under the conditions used, G6PD amplifies linearly, which becomes apparent from Figures 4, 5 and 6, where 10-fold dilutions of the cDNAs result in equivalent changes in the signal obtained with the G6PD hybridization.
pCDKi-mS-76 0.018
ng
0.125
ng
1.0
ng
8.0
ng
cydes: 0 3 6 9 12 15 18 21 FIGURE1 - Hybridization of probe K12-gly with amplified sequences from plasmid pCD-Ki-ras2-76, showing the initial amounts of plasmid used for amplification (0.016-8 ng) and the number of PCR cycles that were performed (0 to 21). Signal intensity increased linearly up to 15 cycles whereafter a plateau was reached.
53
K-ras EXPRESSION IN LUNG TUMORS
nt
pBR322 9
xHpall DNA
1
8
16
18 20
21
51
-
- pCD-76 K-ras
Plasmid:
pCD-Ki-ras-76
PCD-SW-11 .1
100%
0%
75%
25%
50%
50%
25%
75%
0%
100%
Probe: K12-gly K12-val FIGURE3 - Simultaneous amplification of different admixtures of
147.
122. 110.
90-
76 -
FIGURE2 - Amplified cDNAs were end-labelled and separated on a PAA gel. Lane 1; marker pBR322 cut with HpaII: lane 2; one p,g of genomic DNA was reverse transcribed and amplified as described for RNA: lanes 3 to 9; lung tumor samples (the sample in lane 6 is from the same tumor as no. 2 in Figure 5): lane 12; amplification of 1 ng of plasmid pCD-Ki-ras2-76 (McCoy er al., 1984) with the sequence 5’-AC’ITGClTCCTGTAGGAATC-3‘ as a 3’-amplimer, amplifying a 127-nt-long PCR fragment.
Allele-specijicdetection of K-ras expression To investigate the separate detection of both alleles of K-rus, a total of 1 ng of 2 plasmids, pCD-K-rus2-76 and pCDSW-11.1, encoding the wild-type “GGT” and the mutated “GIT” at codon 12, respectively (McCoy era!., 1984), were mixed in different ratios and amplified for 15 cycles (Fig. 3). The result of this experiment demonstrated that the relative initial amounts of both plasmids are accurately reflected in the differential oligonucleotide hybridization on the amplified PCR products. To test this approach on total cellular RNA, 3 adenocarcinoma cell lines were investigated: A549, NCLH23, and GLC-A1 . The cell line A549 is known to harbor a “GGT” to “GAT” mutation at codon 12 which results in a gly to ser mutation in the encoded K-rus protein, while it has lost the wild-type allele of K-rus (Valenzuela and Groffen, 1986). The NCl-H23 cell line was heterozygous for a “GGT” to “TGT” mutation in codon 12 of K-rus (gly to cys), and expressed both
plasmids pCD-Krus-76 and pCD-SW- 1 1 . 1 , containing wild-type and mutated K-ras cDNAs, respectively. Starting with a total amount of 1 ng of plasmid DNA, 15 PCR cycles were performed, after which each PCR reaction was slot-blotted onto 2 separate membranes. Hybridizations with probe K12-gly and K12-val indicate that the initial ratio of both plasmids is retained during the PCR.
alleles of K-rus (see below). No mutations were found in any of the 3 rus genes in the cell line GLC-A1 (not shown). From these cell lines, 100 ng of total RNA were reverse transcribed and serially diluted in 10-fold steps. The cDNA was amplified in twice 15 cycles and slot-blotted on separate membranes. The blots were then hybridized to oligonucleotides specific for Krus wild-type, K-rus mutated and G6PD (Fig. 4). The signal obtained for G6PD is comparable in all 3 samples, indicating the integrity of the initial RNAs. The cell line GLC-A1 has K-rus levels about 10 times lower than those of G6PD. NCIH23 shows higher levels of K-rus, the mutated allele expressed preferentially at levels about 10 times higher than the wild-type allele. The cell line A549 has no wild-type allele, and thus only expression of the K-rus mutated allele was found. Detection of K-ras expression in NSCLCs RNA from 10 adenocarcinomas, 3 large-cell carcinomas, 2 squamous-cell carcinomas, 2 lung metastases of colonic primaries, and from normal lung tissue was analyzed (Table I). None of these samples had mutations in the K-rus oncogene (Rodenhuis et ul., 1987, 1988). Figure 5 shows hybridizations of amplified cDNAs from 6 representative lung tumors and from normal lung tissue, with the wild-type K-rus and G6PD detecting oligonucleotides. In all preparations sufficient amplification of the G6PD fragment was achieved, indicating the integrity of the starting material. Expression of G6PD does not vary significantly between the samples. The signal obtained with the K-rus probe in samples 33, 106 and normal lung tissue is about 10 times lower than that with the G6PD probe, while sample 58 has about 100-fold lower levels than G6PD. Samples 8 and 21 show equal levels with both probes, while in a lung metastasis from an adenocarcinoma of the colon (sample 2), in which we previously detected a 15- to 20-fold amplification of the K-rus oncogene (Rodenhuis er ul., 1987), the K-rus signal is about 10 times higher than the G6PD signal. Table II summarizes the results with tumors without mutations in codon 12 of K-rus. In 5 of 10 adenocarcinomas without a mutation in K-rus, the levels of the oncogene were higher than those found in normal lung tissue (Table II). All 3 large-cell lung carcinomas, and 1 of the 2 investigated squamous-cell carcinomas had expression levels at least 10 times higher than those found in normal lung tissue.
54
SLEBOS ET AL. GLC-A1
NCI-HZ3
A549
19
84
4
21
106
58
K12-W
K12-Sp
G6PD
K12-wl
K12-CyS
G6PD
FIGURE6 - Slot-blots of amplified serial dilutions of cDNA from 4 human lung adenocarcinomas that harbor a mutation in K-ras. The top row on each filter reflects the signal obtained with cDNA from 100 ng of total RNA; each lower row is a 10-fold dilution of the previous one. In samples 19, 87 and 90 the signals obtained with the mutationdetecting probe (K12-va1, KI2-asp, and K12-cys respectively) are higher than those of the wild-type probe (K12-gly). Sample 84 has levels of K-ras comparable to those obtained for G6PD.
K12-W
GGPD
FIGURE4 - Slot-blots of amplified serial dilutions of cDNA from 3 lung adenocarcinoma cell lines. The top row on each filter reflects the signal obtained with cDNA from 100 ng of total RNA; each lower row is a 10-fold dilution of the previous one. Cell line GLC-A1 has no mutation in codon 12 of K-ras, but expresses the wild-type sequence at levels about 10-times lower than G6PD. NCLH23 preferentially expresses the mutated K-ras allele to at least a 10-fold higher degree than the non-mutated allele. The cell line A549 harbors only a mutated allele, which is expressed at levels comparable to those of G6PD. FIGURE5 - Slot-blots of amplified serial dilutions of cDNA from 6 human lung adenocarcinomas and normal lung tissue. The top row on each filter reflects the signal obtained with cDNA from 100 ng of total RNA; each lower row is a 10-fold dilution of the previous one. Sample 2 is from a lung metastasis of a colonic adenocarcinoma with a 15- to 20-fold amplification of the K-ras oncogene. The K-ras signal in this sample is about 10 times higher than the G6PD signal. Sample 58 gave a K-ras signal which is at least 100 times lower than G6PD. samples 33 and 106 had about 10 times lower, while samples 8 and 21 show comparable levels of K-ras and G6PD. For comparison, the signals obtained with normal lung tissue are shown at the bottom of the Figure.
Using mutation-specific oligonucleotides, we were able to discriminate between the expression of the 2 alleles of K-rus in 9 adenocarcinomas (Table I) that had previously been shown to harbor mutations in codon 12 of K-rus. Detection of expression of both the wild-type and the mutated allele in 4 adenocarcinomas with a point mutation in K-rus is shown in Figure 6. In sample 84 the signal obtained with the mutation-detecting
probe is at least equal to that of the wild-type probe, while in the other samples the signals of the mutated allele are about 10 times higher than that obtained with the wild-type probe. In contrast to the results obtained with the cell lines, these results are influenced by nonmalignant cells present in the original tumor sample that may contribute to the wild-type K-rus signal. In summary, one adenocarcinoma had at least 10-fold lower levels of the mutated allele compared to the wild-type allele (not shown), 5 had equal levels of both alleles, while 3 samples showed at least 10-fold higher expression of the mutated allele than of the wild-type allele. DISCUSSION
Mutational activation of K-rus occurs in about one-thud of adenocarcinomas of the lung and is only very infrequently encountered in other subtypes of NSCLC (Rodenhuis et ul., 1987, 1988). The explanation for two-thirds of lung adenocarcinomas not having a mutationally-activated rus gene may well be that other genetic or epigenetic alterations may be substituted for the mutation. One likely cause of such an alternative activation was over-expression of the normal K-rus gene. Apart from oncogenic activation by point mutation, rus genes can become transforming in vitro when expressed at sufficiently high levels, either because of the action of a strong promoter (Chang et ul., 1982), or as a result of the incorporation of multiple copies in the genome of the recipient cell (Pulciani et ul., 1985). Furthermore, amplified rus genes have been detected in several human tumors (Barbacid, 1987). Amplification of the K-rus oncogene has been reported in a single metastatic lung carcinoma, but was not found in 25 primary lung tumors (Heighway and Hasleton, 1986), or in the tumors reported here (Slebos et ul., 1989). Detection methods based on cDNA synthesis followed by PCR have opened new possibilities for the detection of specific messenger RNAs in very small numbers of cells (Rappolee et ul., 1989; Noonan and Roninson, 1990). Several approaches have been developed, all of which are based on quantitative
55
K-ras EXPRESSION IN LUNG TUMORS TABLE I - HISTOLOGICAL CLASSIFICATION OF INVESTIGATED TISSUES AND CELL LINES Cancer
Number
Adenocarcinoma Large-cell carcinoma Squamous-cell carcinoma Solitary metastasis of colonic origin Adenocarcinoma cell line Normal lung tissue Total
19 3 2 2 3 I 30
TABLE 11 - EXPRESSION OF THE K-ras ONCOGENE IN LUNG TUMORS WITHOUT MUTATIONS IN C O W N 12 OF K-ras DETECTED IN SERIAL DILUTIONS OF cDNA. THE TABLE SHOWS THE NUMBER OF SAMPLES WITH SIGNAL LEVELS OF K-ras COMPARED TO THOSE OBTAINED FOR G6PD Histological type
Adenocarcinomas Large-cell carcinomas Squamous-cell carcinomas Lung metastases of colonic adenocarcinoma Cell line GLC-A1
K-ras expression levels relative to G6PD expression
lo-* 1 -
lo-’ 4
1 5
-
-
1
-
-
3 1 1
-
-
1
-
10’ 1
Total 10 3 2 2 1
amplification of cDNA. One strategy uses the addition of known amounts of synthetic RNAs to the RNA preparations to be investigated, thus providing an absolute assessment of expression levels (Becker-Andr6 and Hahlbrock, 1989; Wang et ul., 1989). However, this method does not correct for the degradation of RNA, a common feature of RNA preparations from human tumor tissue. When ubiquitously expressed genes (“housekeeping genes”) are used as an internal standard, and co-amplified together with the genes of interest in a single PCR, correction for limited degradation of RNA, or failure of the PCR, is possible (Rappolee e t a / . , 1989; Noonan and Roninson, 1990). To employ an internal standard that is expressed at levels comparable to those of K-rus, we used the housekeeping gene glucose-6-phosphate dehydrogenase, a key enzyme in the pentose phosphate pathway (Luzzatto and Battistuzzi, 1984). The G6PD gene is expressed at moderate levels in all human tissues thus far analyzed (Hirono and Beutler, 1989; Persico e t a / . , 1986). It is important to select an internal standard defining which mRNA levels are comparable to those of the gene of interest, since it is otherwise difficult to assure linearity of the amplification reaction for both substrates. Thus, for every individual gene that is to be investigated by quantitative PCR, a suitable control should be selected. We compared the expression levels to those in normal lung tissue and to those in a tumor containing a 15- to 20-fold K-rus amplification expected to have high levels of K-rus. Our results indicate that activation of K-rus by high overexpression is rare in human NSCLC, and is not likely to be a pathogenetic alternative for activation by codon 12 point mutations in adenocarcinomas of the lung. This is further supported by the finding that expression levels in tumors with and without a mutation in K-rus codon 12 were similar, and significantly lower than those in a tumor with a K-rus gene am-
plification. Thus, we believe that the moderately increased K-rus expression, as compared to expression in normal lung tissue, may be a consequence of proliferative activity rather than representing one of the transforming events in NSCLC. The specificity of K-rus codon 12 point mutations for adenocarcinomas is not reflected in the expression pattern of the gene: K-rus is expressed at equal levels in all histological subtypes of NSCLC. This would suggest that mutations in K-rus would result in an equally strong transforming signal in all types of NSCLC, and leaves the reason for the specificity of these mutations for the adenocarcinoma subtype unexplained. The detection of normal and mutant K-rus transcripts, using the RNaseA mismatch cleavage method, has been described for several human cell lines. In the cell lines Calu-1 and PR371 mutant K-rus transcripts were 5-10 times more abundant than their normal counterparts (Winter et ul., 1985), a difference that was also reflected in the relative abundance of the normal and mutated alleles in a cDNA library from Calu-1 (Capon et a/.,1983). In the case of PR371, the mutant K-rus allele also showed genomic DNA amplification (Winter et ul., 1985). Moreover, even loss of the wild-type allele in tumor cell lines with K-rus mutations has been described in several cases (Valenzuela and Groffen, 1986; Santos et ul., 1984). We investigated 2 K-rus-mutation-positive lung adenocarcinoma cell lines, and observed expression of the mutated allele in both. The NCl-H23 cell line had higher levels of the mutated allele than of the wild-type allele, while the A549 cell line did not contain a normal K-rus allele (Valenzuela and Groffen, 1986), and hence only expressed the mutated allele. The assessment of (onco)gene expression in fresh human tumors is often complicated by the presence of non-malignant cells (e.g., inflammatory cells) in the sample. When K-rus expression is determined by an allele-specific method, as described here, these normal cells could contribute to the K-rus wild-type signal. Consequently, the mutated allele was expressed preferentially only in the 3 samples in which the signal obtained with the mutationdetecting probe was higher than that found with the wild-type probe. It is likely that in these samples a selection for tumor cells with a preferential expression of the mutated allele of K-rus must have occurred. In the 5 cases in which equal levels of the wild-type and mutated allele were found, a preferential expression of the mutated allele could not be excluded. In colonic adenocarcinomas also, the relative amounts of normal and mutated K-rus mRNAs, detected by the RNaseA mismatch cleavage method, were comparable (Forrester et ul., 1987). These results provide further evidence for the hypothesis that the K-rus oncogene acts in a dominant manner in these tumors. The assay we describe here can be used to determine expression levels of other oncogenes on the basis of minimal amounts of tumor tissue, such as bronchial biopsies or cytological aspiration samples. ACKNOWLEDGEMENTS
We thank Drs. S.M. Bellot, P. Blok, J.A. Van Der Haar, N.A. Den Hartog, Th.M. Van Leeuwen, T.M. Vroom, and Sj.S. Wagenaar for selecting and providing the tumor samples. This research was supported by grants NKI 87-15 and NKI 88-7 from the Dutch Cancer Society KWF.
REFERENCES ALMOGUERA, C . , SHIBATA, D . , FORRESTER, K., MARTIN, J . , ARNHEIM, Bos, J.L., The rus gene family and human carcinogenesis.Mururion Res., N. and PERUCHO, M., Most human carcinomas of the exocrine pancreas 195, 255-271 (1988). contain mutant c-K-rus genes. Cell. 53, 549-554 (1988). CAPON,D.J., SEEBURG,P.H., MCGRATH, J.P., HAYFLICK, J.S., EDMAN, U . , LEVINSON, A.D. and GOEDDEL, D.V., Activation of Ki-rus 2 gene in BARBACID,M., rus genes. Ann. Rev. Biochern., 56, 779-827 (1987). human colon and lung carcinomas by two different point mutations. Nurure BECKER-ANDRB, M . and HAHLBROCK, K., Absolute mRNA quantification I L ~ ~ 304, , ) , 507-513 (1983). using the polymerase chain reaction (PCR). A novel approach by a PCR CARNEY, D.N., GAZDAR, A.F., BEPLER, G . , GUCCION,J.G., MARANGOS, aided transcript titration assay (PA’lTY). Nucl. Acids Res.. 127, 9437P.J., MOODY,T.W . , ZWEIG,M . H . and MINNA, J.D., Establishment and 9746 (1989).
56
SLEBOS ET AL.
identification of small cell lung cancer cell lines having classic and variant features. Cancer Res., 45, 2913-2923 (1985). CHANG,E.H., FURTH,M.E., SCOLNICK,E. and LOWY,D.R., Tumorigenic transformation of mammalian cells induced by a normal human gene homologous to the oncogene of Harvey murine sarcoma virus. Nature (Lond.J, 297,479484 (1982). CHIRGWIN,J.M., PRZYBYLA, A.E., MACDONALD,R.J. and RUTTER, W.J., Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry, 18, 52965299 (1979). FORRESTER,K., ALMOGUERA, C., HAN, K., GRIZZLE,W.E. and PERUCHO,M., Detection of high incidence of K-ras oncogenes during human colon tumorigenesis. Nature (Lond.). 327, 298-303 (1987). FRYE,R.A., BENZ, C.C. and LIU, E., Detection of amplified oncogenes by differential polymerase chain reaction. Oncogene, 4, 1153-1 157 (1989). HEIGHWAY, J. and HASLETON, P.S., c-Ki-ras amplification in human lung cancer. Erir. J . Cancer, 53, 285-287 (1986). WRONO,A. and BEUTLER,E., Alternative splicing of human glucose6-phosphate dehydrogenase messenger RNA in different tissues. J. clin. Invest.. 83, 343-346 (1989). KRAUS,M.H., YUASA,Y. and AARONSON, S.A., A position 12-activated H-ras oncogene in all HS578T mammary carcinosarcoma cells but not in normal mammary cells of the same patient. Proc. nat. Acad. Sci. (Wash.J, 81, 5384-5388 (1984). LIEBER,M., SMITH,B., SZAKAL,A., NELSON-REES,W. and TODARO, G . , A continuous tumor-cell line from a human lung carcinoma with properties of type I1 alveolar epithelial cells. Int. J. Cancer, 17, 62-70 (1976). LUZZATTO,L. and BATTISTUZZI,G . , Glucose-6-phosphate dehydrogenase. Advanc. hum. Gener., 14, 217-229 (1984). MARTINI,G., TONIOLO, D., VULLIAMY, T., LUZZATTO,L., DONO, R., G., D'URSO, M. and ~ R S I C O M.G., , StrucVIGLIETTO,G., PAONESSA, tural analysis of the X-linked gene encoding human glucose-6-phosphate dehydrogenase. EMBO J., 5, 1849-1855 (1986). McCoy, M.S., BARGMANN, C.I. and WEINBERG,R.A., Human colon carcinoma Ki-rus2 oncogene and its corresponding proto-oncogene. Mol. cell. Biol.. 4, 1577-1582 (1984). MCGRATH,J.P., CAPON, D.J., SMITH, D.H., CHEN, E.Y., SEEBURG, P.H.,GOEDDEL,D.V. and LEVINSON, A.D., Structure and organization of the human Ki-ras proto-oncogene and a related processed pseudogene. Nature (Lond.), 304, 501-506 (1983). MULLIS,K.B. and FALOONA, F.A., Specific synthesis of DNA in vitro via a polymerase catalyzed chain reaction. Merh. Enzymol., 155, 335-350 (1987). NOONAN,K.E. and RONINSON,I.B., Quantitative estimation of mdrl mRNA level by polymerase chain reaction. In: I.B. Roninson (ed.), Molecular and cellular biology of mulridrug-resistance in tumor cells. Plenum, New York (1990). (In press). ~ R S I C O M.G., , VIGLIETTO,G., MARTINI,G., TONIOLO, D., PAONESSA, G., MOSCATELLI, C., DONO, R., VULLIAMY,T., LUZZATTO,L. and D'URSO, M., Isolation of human glucose-6-phosphate dehydrogenast (G6PD) cDNA clones: primary structure of the protein and unusual 5 non-coding region. Nucl. Acids Res., 14, 2511-2522 (1986). PULCIANI, S., SANTOS,E., LONG,L.K., VICENZO,S. and BARBACID, M., ras gene amplification and malignant transformation. Mol. cell. Eiol., 5, 2836-2841 (1985).
RAPPOLEE,D.A., WANG,A,, MARK,D. and WERB,Z., Novel method for studying mRNA phenotypes in single or small number of cells. J . cell. Biochem.. 39, 1-11 (1989). RODENHUIS, S., SLEBOS,R.J.C., BOOT, A.J.M., EVERS,S.E., MOO], S J . ~ VAN . , BODEGOM, P.CH. and Bos, J.L., Incidence W.J., WAGENAAR, and possible clinical significance of K-ras oncogene activation in adenocarcinoma of the human lung. Cancer Res., 48, 5738-5741 (1988). RODENHUIS, S., VANDE WETERING,M.L., MOOI,W.J., EVERS,S.G., VAN ZANDWIJK,N. and Bos, J.L.,Mutational activation of the K-ras oncogene: a possible pathogenic factor in adenocarcinoma of the lung. N. Engl. J. Med., 317, 929-935 (1987). SAKATO, H., REDDY,E.P., PLJLCIANI, S., FELDMAN, R.J. and BARBACID, M., Transforming gene from human stomach cancers and a noncancerous portion of stomach mucosa. Proc. nat. Acad. Sci. (Wash.), 83,3997-4001 ( 1984). SANTOS,E., MARTIN-ZANCA, D., REDDY,E.P., PIEROTTI,M.A., DELLA M., Malignant activation of a K-ras oncogene PORTA,G. and BARBACID, in lung carcinoma but not in normal tissue of the same patient. Science, 223, 661464 (1984). SLEBOS,R.J.C., EVERS,S.E., WAGENAAR, SJ.S. and RODENHUIS,S., Cellular proto-oncogenes are infrequently amplified in untreated nonsmall-cell lung cancer. Brit. J . Cancer, 59, 76-80 (1989). SLEBOS,R.J.C., KIBBELAAR, R.E., DALESIO,O., KOOISTRA, A,, STAM, J . , MEIJER, C.J.L.M., WAGENAAR, S J . ~ . ,VANDERSCHUEREN, R.G.J.R.A., VAN ZANDWIJK,N., MOOI, W.J., Bos, J.L. and RODENHUIS, S., Kirsten-ras oncogene activation as a prognostic marker in adenocarcinoma of the lung. N. Engl. J . Med., 323, 561-565 (1990). SMIT, V.T.H.B.M., BOOT, A.J.M., SMITS, A.M.M., FLEUREN,G.J., CORNELISSE, C.J. and Bos, J.L., Kras codon 12 mutations occur very frequently in pancreatic adenocarcinomas. Nucl. Acids Res., 16, 77737782 (1988). TRAHEY,M. and MCCORMICK,F., A cytoplasmic protein stimulates normal N-ras p21 GTPase, but does not affect oncogenic mutants. Science, 238,542-545 (1987). VALENZUELA, D.M. and GROFFEN,J., Four human carcinoma cell lines with novel mutations in position 12 of c-K-ras oncogene. Nucl. Acids Res.. 14, 843-852 (1986). VAN DE VIJVER,M.J., PETERSE,J.L., MOOI, W.J., LOMANS,J., VERBRUGGEN, M., VAN DE BERSSELAAR, R., DEVILEE,P., CORNELISSE, c., J. and NUSSE,R., Oncogene activations in human BOS, J.L., YARNOLD, breast cancer. Cancer Cells, 7, 385-391 (1989). VERLAAN-DE VRIES,M., BOGAARD,M.E., ELST, H., VANBOOM,J.H., VAN DER EB, A.J. and Bos, J.L., A dot-blot screening procedure for mutated ras oncogenes using synthetic oligodeoxynucleotides. Gene, 50, 313-320 (1986). WANG,A.M., DOYLE,M.V. and MARK,D.F., Quantitation of mRNA by the polymerase chain reaction. Proc. nar. Acad. Sci. (Wash.), 86, 97179721 (1989). WINTER,E., YAMAMOTO, F., ALMOGUERA, C. and PF~RUCHO, M., A method to detect and characterize point mutations in transcribed genes: amplification and overexpression of the mutated c-Ki-ras allele in human tumor cells. Proc. nut. Acad. Sci. (Wash.), 82, 7575-7579 (1985). WOOD,W.I., GITSCHIER,J . , LASKY,L.A. and LAWN,R.M., Base composition-independent hybridization in tetramethylammonium chloride: a method for oligonucleotide screening of highly complex gene libraries. Proc. nat. Acad. Sci. (Wash.), 82, 1585-1588 (1985).
