Predicting lung cancer by detecting aberrant promoter methylation in sputum.

Despite the promise of using DNA markers for the early detection of cancer, none has proven universally applicable to the most common and lethal forms of human malignancy. Lung carcinoma, the leading cause of tumor-related death, is a key example of a cancer for which mortality could be greatly reduced through the development of sensitive molecular markers detectable at the earliest stages of disease. By increasing the sensitivity of a PCR approach to detect methylated DNA sequences, we now demonstrate that aberrant methylation of the p16 and/or O6-methyl-guanine-DNA methyltransferase promoters can be detected in DNA from sputum in 100% of patients with squamous cell lung carcinoma up to 3 years before clinical diagnosis. Moreover, the prevalence of these markers in sputum from cancer-free, high-risk subjects approximates lifetime risk for lung cancer. The use of aberrant gene methylation as a molecular marker system seems to offer a potentially powerful approach to population-based screening for the detection of lung cancer, and possibly the other common forms of human cancer.

Aberrant methylation of p16(INK4a) is an early event in lung cancer and a potential biomarker for early diagnosis.

The p16(INK4a) (p16) tumor suppressor gene can be inactivated by promoter region hypermethylation in many tumor types including lung cancer, the leading cause of cancer-related deaths in the U.S. We have determined the timing of this event in an animal model of lung carcinogenesis and in human squamous cell carcinomas (SCCs). In the rat, 94% of adenocarcinomas induced by the tobacco specific carcinogen 4-methylnitrosamino-1-(3-pyridyl)-1-butanone were hypermethylated at the p16 gene promoter; most important, this methylation change was frequently detected in precursor lesions to the tumors: adenomas, and hyperplastic lesions. The timing for p16 methylation was recapitulated in human SCCs where the p16 gene was coordinately methylated in 75% of carcinoma in situ lesions adjacent to SCCs harboring this change. Moreover, the frequency of this event increased during disease progression from basal cell hyperplasia (17%) to squamous metaplasia (24%) to carcinoma in situ (50%) lesions. Methylation of p16 was associated with loss of expression in both tumors and precursor lesions indicating that both alleles were functionally inactivated. The potential of using assays for aberrant p16 methylation to identify disease and/or risk was validated by detection of this change in sputum from three of seven patients with cancer and 5 of 26 cancer-free individuals at high risk. These studies show for the first time that an epigenetic alteration, aberrant methylation of the p16 gene, can be an early event in lung cancer and may constitute a new biomarker for early detection and monitoring of prevention trials.

Effect of promoter and intron 2 polymorphisms on murine lung K-ras gene expression.

Differences in tumor formation among inbred mouse strains with high (A/J) and low (C3H) susceptibility for lung cancer have been linked to a repetitive element within the second intron of the K-ras gene. The purpose of this investigation was to determine whether differences within the K-ras gene promoter region or the intron 2 polymorphism affect K-ras gene expression in lung tumors and target alveolar type II cells isolated from A/J and C3H mice. Ribonuclease protection assays were performed using RNA isolated from 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK)-induced lung tumors from each mouse strain and alveolar type II cells isolated from A/J and C3H mice. An 838 bp fragment of the murine K-ras gene promoter region was amplified by PCR, cloned and sequenced from both mouse strains. Promoter regions from both mouse strains were inserted into a luciferase reporter gene vector, with and without the second intron polymorphism, and transfected into sensitive, intermediate and resistant lung tumor cell lines. No significant differences in K-ras gene promoter activity was found between the two strains using these specific reporter gene constructs. Consistent with these results, levels of K-ras expression did not differ between alveolar type II cells, whole lung or tumors induced by NNK in A/J or C3H mice. Moreover, in lung tumor cell lines derived from mice with differing susceptibility for lung cancer, K-ras expression did not correlate with the growth rate of tumors induced in nude mice from these cell lines. These results indicate that factors involved in modulating the rapid clonal expansion of the mutated K-ras allele from A/J mice are not directly linked to expression of this gene. Other genetic changes or losses in conjunction with hypothesized modifier loci, such as the Par1 locus, must play a significant role in establishing the phenotypic strain differences for lung tumor formation.

Frequent aberrant methylation of p16INK4a in primary rat lung tumors.

The p16INK4a (p16) tumor suppressor gene is frequently inactivated by homozygous deletion or methylation of the 5' CpG island in cell lines derived from human non-small-cell lung cancers. However, the frequency of dysfunction in primary tumors appears to be significantly lower than that in cell lines. This discordance could result from the occurrence or selection of p16 dysfunction during cell culture. Alternatively, techniques commonly used to examine tumors for genetic and epigenetic alterations may not be sensitive enough to detect all dysfunctions within the heterogeneous cell population present in primary tumors. If p16 inactivation plays a central role in development of non-small-cell lung cancer, then the frequency of gene inactivation in primary tumors should parallel that observed in cell lines. The present investigation addressed this issue in primary rat lung tumors and corresponding derived cell lines. A further goal was to determine whether the aberrant p16 gene methylation seen in human tumors is a conserved event in this animal model. The rat p16 gene was cloned and sequenced, and the predicted amino acid sequence of its product found to be 62% homologous to the amino acid sequence of the human analog. Homozygous deletion accounted for loss of p16 expression in 8 of 20 cell lines, while methylation of the CpG island extending throughout exon 1 was observed in 9 of 20 cell lines. 2-Deoxy-5-azacytidine treatment of cell lines with aberrant methylation restored gene expression. The methylated phenotype seen in cell lines showed an absolute correlation with detection of methylation in primary tumors. Aberrant methylation was also detected in four of eight primary tumors in which the derived cell line contained a deletion in p16. These results substantiate the primary tumor as the origin for dysfunction of the p16 gene and implicate CpG island methylation as the major mechanism for inactivating this gene in the rat lung tumors examined. Furthermore, rat lung cancer appears to be an excellent model in which to investigate the mechanisms of de novo gene methylation and the role of p16 dysfunction in the progression of neoplasia.

Avian thymic hormone and chicken (muscle) parvalbumin are distinct proteins: isolation of a muscle parvalbumin cDNA fragment by PCR.

Access to the nucleotide sequence of parvalbumin from chicken muscle was gained via the polymerase chain reaction. In the absence of specific amino acid sequence data, the PCR primers were based on consensus data for the two parvalbumin Ca2(+)-binding sites. The 137 bp fragment obtained by amplification clearly codes for a parvalbumin, as judged by the presence of 10 invariant codons within the sequence flanked by the primers. When used to probe poly(A)+ RNA from chicken muscle, the fragment recognizes an 800 nucleotide transcript. The translated nucleotide sequence of the muscle protein is unmistakably distinct from that of the thymus-specific parvalbumin known as avian thymic hormone. Of the 31 amplified residues, the two proteins differ at 14. The presence of a distinct parvalbumin in chicken thymus is consistent with the potent effector role proposed for the protein.

Molecular cloning of the thymus-specific parvalbumin known as avian thymic hormone: isolation of a full length cDNA and expression of the recombinant protein in Escherichia coli.

The complete coding sequence of the thymus-specific parvalbumin called avian thymic hormone (ATH) has been cloned into Escherichia coli. The translated amino acid sequence was found to be identical to the sequence of map turtle parvalbumin at 90 of 108 positions. Northern blot analysis of thymic RNA indicated a transcript length of approximately 1050 bp. However, the ATH cDNA probe failed to hybridize to poly(A)+ RNA from chicken leg muscle, a further indication that avian thymic hormone is distinct from the muscle-associated parvalbumin previously isolated from chicken. Southern analysis of chicken genomic DNA suggests the presence of a single copy of the ATH gene, and the absence of hybridization between an ATH cDNA fragment and genomic DNA from rat and rabbit is confirmatory evidence that ATH expression is restricted to avian species. One of the full length ATH cDNA clones harbored an insert that lacked all 5' noncoding sequences. This cDNA was inserted without further alteration into the prokaryotic expression vector, pKK223-3. The resulting construction, which contains eleven base pairs between the Shine-Dalgarno sequence and the initiation codon, affords reasonably high levels of expression in E. coli. In most respects, recombinant ATH mimics the tissue-derived protein, retaining a similarly high affinity for Ca2+ ion (KCa = 14 +/- 5 nM). However, in contrast to ATH isolated from chicken thymus tissue, the N-terminal alanine of recombinant ATH is unacetylated. As a result, the isoelectric point is shifted upward from 4.3 to approximately 4.8.

Site-specific replacement of amino acid residues within the CD binding loop of rat oncomodulin.

Relative to the same site in oncomodulin, the CD ion-binding domain of rat parvalbumin exhibits much greater affinity for Ca2+ and Mg2+. As part of an effort to understand the structural basis for these differences, site-specific variants of oncomodulin have been prepared in which the amino acid residues at positions 52, 54, 57, 59, and 60 have been replaced with the residues present at the corresponding positions in rat parvalbumin. The proteins resulting from the single-site substitutions at residues 52, 54, and 57 are indistinguishable from the wild-type protein on the basis Eu3+ luminescence spectroscopy, and none of the three variants displays increased affinity for Ca2+. By contrast, the substitutions at residues 59 and 60 perturb both the Eu3+ luminescence parameters and the Ca2+ and Mg2+ affinities, and these differences are amplified when both replacements are simultaneously incorporated into the protein. The Eu3+ 7F0----5D0 spectrum of the double variant (D59E/G60E) at pH 5.0, with a maximum at 5796 A and pronounced shoulder at 5791 A, strongly resembles that obtained with pike parvalbumin. Consistent with this increased parvalbumin-like character, KCa is decreased from 0.78 microM (for the wild-type protein) to 0.41 microM, and KMg is decreased from 3.5 to 0.74 mM. Nevertheless, the affinity of the CD ion-binding domain in D59E/G60E for Ca2+ remains almost 2 orders of magnitude lower than the corresponding site in rat parvalbumin, strongly suggesting that residues besides those present in the binding loop are involved in dictating the metal ion-binding properties of the oncomodulin CD site.

Eu3+ luminescence studies of oncomodulin. The origin of the pH-dependent behavior.

The Eu3+ 7F0----5D0 excitation spectra of parvalbumin and oncomodulin are pH-dependent. Until now, it had been assumed that both the CD and EF ion-binding sites shared this property and that deprotonation of water molecules coordinated to the bound Eu3+ ions might be responsible for the pH dependence. Results obtained with the site-specific variant of oncomodulin known as D59E, in which glutamate replaces the aspartate naturally present at position 59, have necessitated substantial revision of these ideas. It now appears that the pH-dependent behavior is confined to the CD site. Moreover, we observe no corresponding change in the number of O-H oscillators coordinated to the bound Eu3+ ions in the pH range over which we observe the spectroscopic alteration. It is likely that the behavior results from deprotonation of one or more carboxyl groups clustered at the COOH-terminal end of the CD domain.

Partial nucleotide sequence of the parvalbumin from chicken thymus designated "avian thymic hormone".

The incomplete amino acid sequence of the protein identified as avian thymic hormone was recently reported [Brewer et al. (1989), Biochem. Biophys. Res. Commun. 160, 11555-1161], and a very high degree of homology to the parvalbumins was apparent. Using mixed oligonucleotide primers based on the reported protein sequence, we have succeeded in amplifying and cloning a 188 bp fragment of the coding region for this protein, beginning with double-stranded cDNA prepared from chicken thymus mRNA. The translated nucleotide sequence of this fragment and the reported amino acid sequence display substantial disagreement. Most notably, the nucleotide sequence indicates that the CD site of avian thymic hormone is a typical parvalbumin CD site.

Site-specific substitution of glutamate for aspartate at position 59 of rat oncomodulin.

Replacement of the aspartate residue at position 59 of rat oncomodulin by glutamate by oligonucleotide-directed mutagenesis has afforded a protein which more closely resembles rat parvalbumin, at least judged by its interaction with the luminescent lanthanide ion Eu3+. The single-peak 7F0----5D0 spectrum observed at pH 5.0 with the fully bound wild-type protein is replaced by one which clearly shows two features at 5791 and 5796 A, arising from Eu3+ ions bound at the CD and EF sites, respectively. Furthermore, the pH dependence of the spectrum is substantially altered; the pKa observed for the CD domain, in which aspartate 59 residues, is shifted upward from pH 6.0 for the wild-type recombinant protein to pH 6.8 in the D59E mutant. Moreover, the maximum in the high-pH spectrum is shifted from 5781 to 5784 A. All three changes are indicative of a CD binding domain having increased parvalbumin-like character. Interestingly, however, the D59E substitution has only a modest effect on the Ca2+- and Mg2+-binding properties of the CD domain. For the wild-type protein, KCa = 7.8 x 10(-7) M and KMg = 3 x 10(-3) M. These affinities are more than an order of magnitude weaker than those seen for various parvalbumins and substantiate previous claims for calcium specificity made for the oncomodulin CD domain. Replacement of aspartate 59 by glutamate resulted in minor increases in affinity of the CD domain for Ca2+ (KCa = 5.5 x 10(-7) M) and Mg2+ (KMg = 1 x 10(-3) M). These findings strongly suggest that residues in oncomodulin besides aspartate 59 are important determinants of the observed calcium specificity of the CD calcium-binding domain. The consequences of the substitution at residue 59 appear to be confined to the CD domain. For the EF site in wild-type recombinant oncomodulin, KCa = 4.2 x 10(-8) M and KMg = 1.6 x 10(-4) M. The corresponding values for the D59E site-specific variant are identical within experimental error (KCa = 4.2 x 10(-8) M and KMg = 1.8 x 10(-4) M).