Detection of DNA methylation in the calcitonin gene in human leukemias using differential polymerase chain reaction.

One of the earliest events in the multistep process of malignant transformation is a change in the methylation pattern of certain genes. DNA methylation is usually detected by Southern blotting after restriction digest with methylation-sensitive endonucleases. Calcitonin gene hypermethylation has been described in a variety of human malignancies including lymphomas and leukemias. Here we report a technique based on the semi-quantitative differential polymerase chain reaction (PCR) which is capable of detecting subtle changes in the methylation pattern of the human calcitonin gene. This technique is based on two principles: (i) simultaneous coamplification of the target gene (5'-region of the calcitonin gene) and a reference gene for quantitative purposes; and (ii) simultaneous coamplification of a competitor with identical primer-binding sites as the target gene to control for proper restriction digest. Using this technique, we investigated calcitonin gene methylation in a variety of human cell lines, primary leukemias and normal human blood donors. The data revealed good correlation with standard Southern blotting. Weak calcitonin gene methylation was found in all normal blood donors tested (n = 14). In contrast, strong calcitonin gene methylation was detected in most acute leukemias (five of 10 acute myeloid leukemias (AML); six of seven acute lymphoblastic leukemias (ALL)). These data show that this technique can reliably be used to quantitate gene methylation and indicate that there exists heterogeneity with regard to methylation status in different leukemias, suggesting that hypermethylation of the calcitonin gene may play a role in the transformation process of some, but not all, human leukemias. Furthermore, differential PCR may facilitate determination of calcitonin gene methylation in clinical or archival tumor samples.

Mutations in the ras protooncogenes are rare events in renal cell cancer.

Mutations in codon 12, 13 or 61 of one of the three ras genes, Ha-ras, Ki-ras, and N-ras, convert these genes into active oncogenes. To determine the role mutated ras genes play in the carcinogenesis of renal cell carcinoma, we analysed tumour DNA and unaffected renal tissue derived from 55 patients. The polymerase chain reaction technique was used to amplify DNA fragments containing Ki-, Ha-, and N-ras codons 12, 13, and 61. The amplified fragments were then probed on slot-blots with labeled mutation-specific oligomers. A single Ki-ras mutation (codon 12, gly- greater than val) was detected in a patient with a pT2N2M1 tumour. We concluded that ras oncogene mutations do not play an important role in the initiation of renal cell carcinoma.

The status of p53 in the metastatic progression of colorectal cancer.

In order to investigate the role of TP53 in tumour progression and metastasis, we analysed 33 liver metastases of colorectal carcinomas and 19 primary colon carcinomas from the same hospital with respect to mutational changes, loss of heterozygosity and expression of the TP53 tumour suppressor gene. Direct sequencing of PCR products corresponding to the coding region of TP53 revealed that 13 of 19 primary tumours (68%) and 23 of 33 liver metastases (70%) had mutations in the TP53 gene. The distribution of mutations along the coding region of TP53 was similar in liver metastases compared to primary tumours. Thus, codon specificity did not seem to be a relevant factor and cells carrying specific TP53 mutations seem to have no selective advantage in the metastasising process. Comparing our data with the mutational spectra found in other countries did not reveal differences in the distribution of mutations along the coding region. Most of the metastases analysed showed loss of heterozygosity (LOH, 9 of 12 cases, 75%) and strong nuclear staining in immunohistochemistry (10 of 17 cases, 59%). Furthermore, with respect to mRNA expression levels, tumours carrying TP53 mutations showed significantly higher p53 mRNA levels compared to those without TP53 mutations. Thus, regulation of p53 mRNA levels seems to be subject to selection processes in tumourigenesis.

bcl-2 expression is reciprocal to p53 and c-myc expression in metastatic human colorectal cancer.

Apoptosis (programmed cell death) inhibition may be an important mechanism by which gastrointestinal mucosal cells containing damaged DNA evade normal clearance mechanisms and grow to become invasive tumours. Since bcl-2 is an apoptosis inhibitor, bcl-2 mRNA expression was measured in 21 metastases of colorectal cancer using reverse transcription-polymerase chain reaction analysis. The mean bcl-2 mRNA expression (0.45 U, P < 0.0001) was lower than that of normal mucosal controls (= 1 U). p53 expression was inversely correlated with bcl-2 expression (P = 0.021) in 19 evaluable samples, and in tumours where p53 expression was over twice that of normal colonic mucosal values, bcl-2 mRNA was significantly decreased (mean 0.30, P = 0.0052). c-myc was also inversely correlated with bcl-2 expression (P = 0.025). Decreased bcl-2 expression in metastatic colorectal cancer may be partly due to allelic loss, given the proximity of bcl-2 to the frequently deleted DCC gene on chromosome 18q. However, the inverse correlation to p53/c-myc suggests an active downregulation of bcl-2, possibly following delegation of its apoptosis inhibiting role to other genes.

Gene expression analysis by a competitive and differential PCR with antisense competitors.

We report a sensitive method for the reproducible and accurate measurement of gene expression from small samples of RNA. This method is based on a combination of two PCR techniques: First, an endogenous reporter gene and the gene of interest are simultaneously amplified in one tube after random-primed reverse transcription (RT) of RNA (differential RT-PCR). Second, exogenous homologous fragments of both genes with artificially introduced mutations are added and coamplified in the same reaction (competitive PCR). The first-strand cDNA, and the mutated antisense homologues of the reporter as well as the target gene compete for their respective primers and are therefore amplified with equal efficiencies. After PCR, restriction enzyme digestion allows visualization of the quantitative differences between the four resulting reaction products. The ratios of products that competed during PCR provide the quantitative information. The initial amount of a specific cDNA can be calculated from any competitor/cDNA ratio of reliably measurable PCR product amounts. Extensive competitor titration to experimentally approach the equilibrium is therefore unnecessary. The differential counterpart of competitive and differential RT-PCR (CD-RT-PCR) allows expression of the levels in reference to a reporter gene. MDR1 expression was determined in tumor cells by CD-RT-PCR.

The relation between inhibition of cell growth and of dihydroorotic acid dehydrogenase by brequinar sodium.

The growth inhibitory effects of Brequinar Sodium (DUP-785; NSC 368390) in 7 different cell lines were related to growth rates and to the inhibition of dihydroorotic acid dehydrogenase (DHO-DH) activity. IC50 values were between 0.2 and 5.8 microM; the fastest growing cell line was least sensitive. Despite a large variation in sensitivity, basal activity of DHO-DH showed little variation (only 2-fold) between the different cell lines. Residual activity of DHO-DH in the presence of Brequinar Sodium varied 30-fold. Drug sensitivity correlated with this residual DHO-DH activity; DHO-DH activity was only slightly inhibited by Brequinar Sodium in the most resistant lines, and almost completely in the most sensitive.

Retention of in vivo antipyrimidine effects of Brequinar sodium (DUP-785; NSC 368390) in murine liver, bone marrow and colon cancer.

Brequinar sodium (DUP-785) is a potent inhibitor of the pyrimidine de novo enzyme, dihydroorotic acid dehydrogenase (DHO-DH). In order to determine whether in vitro data could be extrapolated to the in vivo situation we investigated antipyrimidine effects of DUP-785 in mice bearing colon cancer. Two tumor models were used, Colon 26 and Colon 38, resistant and moderately sensitive to DUP-785, respectively. DUP-785 at 50 mg/kg caused a depletion of plasma uridine in mice, and depleted tissue uridine levels in Colon 38 down to 10%, which was retained for several days; in Colon 26 the decrease was less and tissue uridine levels recovered rapidly. In livers of these mice no significant effect on uridine was observed. DUP-785 depleted UTP in bone marrow cells within 2 hr to 25% of control levels, after 4 days normal levels were found. In livers of both Balb-c mice (bearing Colon 26) and C57Bl/6 mice (bearing Colon 38) a small decrease of uridine nucleotide pools was found. In Colon 26 DUP-785 increased uridine nucleotide pools to 170% after 2 hr, at 1 day normal levels were observed, but after 2 days again an increase was found. In Colon 38 DUP-785 decreased the uridine nucleotide pool by 50% after 1 and 2 days. DUP-785 did not affect cytidine nucleotide pools of livers and of Colon 26 and Colon 38. The ratio between uridine nucleotides and cytidine nucleotides decreased from 2.2 to 0.90 in Colon 38, in the other tissues the decrease was less. DHO-DH was measured in bone marrow cells and Colon 26 and 38 before and after treatment. Basal levels of DHO-DH were 3 times higher in Colon 26 than in Colon 38. In treated tumors DHO-DH was initially inhibited by more than 90%, after 7 days enzyme activity in Colon 26 was 50% and in Colon 38 about 200% of basal levels. In bone marrow cells DHO-DH was also rapidly inhibited but recovered within 4 days. It is concluded that the retention of antipyrimidine effects of DUP-785 in Colon 38 were more pronounced than in Colon 26, which is in agreement with the better antitumor effect of DUP-785 in Colon 38.

MDR1 expression correlates with mutant p53 expression in colorectal cancer metastases.

Overexpression of the multidrug resistance MDR1 gene is thought to contribute to drug resistance in non-responsive cancers like colorectal carcinoma. Little is known about the mechanisms by which expression of MDR1 is regulated in human tumours. However, there is growing evidence that regulation primarily takes place at the transcriptional level and that the process of tumour progression is related to activation of the MDR1 gene. Mutations in the p53 tumour-suppression gene occur in approximately 70% of colorectal cancers. As a transcriptional regulator, p53 might be involved in regulation of MDR1 expression in these tumours. We therefore determined MDR1 expression using the differential polymerase chain reaction technique in 30 colorectal tumours (4 primaries and 26 metastates) and correlated our results with previously reported data on p53 in the same group of patients. We found a significant positive correlation between p53 and MDR1 expression in p53-mutated tumours (P = 0.005; r = 0.596), but not in tumours without a p53 mutation. In addition, we observed a tendency towards higher MDR1 expression levels in tumours carrying p53 mutations (P = 0.14) compound to wild-type p53 tumours. These data indicate that mutant p53 may play a role in the regulation of MDR1 expression in human cholorectal cancer.

Competitive and Differential RT-PCR (CD-RT-PCR) for Measurement of Normalized Gene Expression Using Antisense Competitors.

Quantitative mRNA characterization by reverse transcription (RT) of RNA and subsequent polymerase chain reaction (PCR) (RT-PCR) is, compared to qualitative RT-PCR detection of RNA, more complicated because of two features inherent in in vitro amplification. First, during the exponential phase, minute differences in a number of variables can greatly influence reaction rates, with substantial effect on the yield of PCR products. Second, as a consequence of reaction components consumption and generation of inhibitors, the amplification enters a plateau phase. At this point, the reaction rate declines to an unknown level. Another source of errors in quantitative RT-PCR analysis lies in the determination of the amount of RNA to be analyzed for each sample. In small samples, the total amount of RNA may even be beyond the limit of detection. The sample loading problem can be solved by presenting the level of expression of the gene of interest in reference to a constitutively expressed gene. In a PCR, this can be done by the simultaneous amplification of two different genes in one reaction vessel, which was called differential PCR (1). However, in many cases, a quantitative PCR assay is desired that is internally controlled both for errors in comparison between samples and for the efficiency of the amplification reaction. To that end, a technique was devised that combines competitive PCR and differential RT-PCR by coamplification of two genes and their corresponding competitive templates (2). This chapter describes the working procedures for the complete assay called competitive and differential RT-PCR (CD-RT-PCR) and concomitant techniques.

Overexpression and amplification of c-myc during progression of human colorectal cancer.

Overexpression and amplification of the c-myc oncogene occur in approximately 70 and 10% of human primary colorectal carcinomas, respectively, indicating the importance of this gene in colorectal tumorigenesis. Little, however, is known about the involvement of c-myc in the progression of colorectal cancer. We therefore determined c-myc gene expression and amplification in a group of primary tumors and metastases from patients with colorectal cancer using quantitative PCR-based tests. While the percentage of metastases overexpressing c-myc (13/26 = 50%) was in the same range as reported for primary tumors by others, gene amplification of c-myc was significantly (p = 0.001) more frequent in metastases (16/27 = 59%) compared to primary tumors (1/23 = 4%) in our series. Interestingly, in 23 metastases where both expression and amplification of c-myc could be determined, there was no correlation between gene copy number and expression level (p = 0.18; r = 0.19). We conclude that amplification but not overexpression of c-myc is related to metastatic progression of colorectal cancer and that overexpression of c-myc is driven by mechanisms other than the number of c-myc copies in the tumors studied.

Expression and mutational analysis of Nm23-H1 in liver metastases of colorectal cancer.

It has been proposed that nm23-H1, a candidate suppressor gene for metastasis, plays an important role in metastasis formation of human tumours. In order to investigate its role in the progression of colorectal cancer, we analysed 22 liver metastases of this malignancy with respect to mutational changes, loss of heterozygosity and expression levels of nm23-H1. Although genetic alterations in nm23-H1 have recently been described in those colorectal adenocarcinomas which give rise to distant metastases, we were unable to detect any mutation in the coding sequence of nm23-H1 in the metastatic tissue itself. We further analysed the metastases with respect to allelic deletions at the chromosomal locus of nm23. However, no loss of heterozygosity could be detected in ten informative cases. Moreover, the mRNA expression levels of nm23-H1 in the metastatic tissues were not significantly different from those in normal colon mucosa. Thus, although nm23-H1 might be involved in metastasis suppression of certain tumour types, in colorectal tumour progression its role remains to be determined.

Evidence for a mutual regulation of p53 and c-myc expression in human colorectal cancer metastases.

BACKGROUND: Alterations of the c-myc and the p53 genes occur in a majority of human colorectal cancers, and functional interaction between these two genes has recently been suggested. PATIENTS AND METHODS: We analyzed p53 sequence and c-myc and p53 mRNA expression in 26 metastases and 4 advanced primaries of human colorectal cancer. RESULTS: Twenty-one of 30 tumors (=70%) carried mutations of the p53 gene. In these samples, c-myc and p53 were overexpressed in 70% (15/21) and 71% (14/20) of evaluable cases, respectively, while in tumors carrying only wild-type p53, overexpression of c-myc and p53 was observed in only 33% (3/9; p < 0.05) and 22% (2/9; p < 0.01), respectively. Expression of p53 and c-myc were positively correlated (p = 0.014; r = 0.563) in tumors carrying a p53 mutation, but not in those with only wild-type p53. CONCLUSION: We conclude that c-myc might induce p53 expression in human colorectal cancer and that wild-type but not mutant p53 might be involved in a negative feedback regulation of c-myc expression. The abrogation of this normal control mechanism seems to be an essential step during colorectal tumorigenesis and metastatic progression.

Multiple drug resistance gene expression in human renal cell cancer is associated with the histologic subtype.

Overexpression of p170 glycoprotein, the product of the multiple drug resistance (mdr) gene, has been associated with resistance to various cytotoxic drugs used in the treatment of human neoplasms. Normal renal epithelial cells express p170 as a function of their secretory capacity. Because renal cell carcinomas (RCC) respond poorly to chemotherapeutic regimens, p170 expression was studied in primary RCC. Such expression was measured in 40 human RCC and normal kidney tissues using immunohistochemical staining with the monoclonal antibody C-219. Staining intensities of the whole tumor and of different areas of the cryostat sections were transformed into digital numbers using an algorithm designed for this purpose. In most tumors, an inhomogeneous staining pattern and a correlation between grade of differentiation and C-219 immunoreactivity was observed. A comparison of the tumors according to their histopathologic subtypes showed clear differences. The means (range) of the staining intensities of the different types of RCC: clear cell carcinoma Grade 1 (n = 3), 2.0 (2.0 to 2.0); clear cell carcinoma Grade 2 (n = 19), 0.8 (0.0 to 2.9); clear cell carcinoma Grade 3 (n = 5), 0.1 (0.0 to 0.2); tubular carcinoma (n = 4), 2.0 (2.0 to 3.0); anaplastic carcinoma (n = 8), 0.05 (0.0 to 0.2); oncocytoma (n = 1), 0.0 (0.0 to 0.0); and normal kidney (n = 40), 0.5 (0.0 to 2.0). The differences between anaplastic, clear cell, and tubular carcinoma were significant (P less than 0.001 by Kruskal-Wallis test). In addition, the difference between the three subgroups of clear cell carcinoma was significant (P less than 0.01). It was concluded that the histopathologic subtypes of RCC correlate with the degree of mdr gene expression, as determined by staining with the C-219 monoclonal antibody.

Genetic heterogeneity in the apolipoprotein C-III promoter and effects of insulin.

In the present study, we have investigated the in vivo and in vitro role of two newly identified variants (G(-)944A and A(-)1180C) located in the upstream promoter region of the apolipoprotein C-III (apoC-III) gene. These variants were studied in 30 probands diagnosed with FCHL, their relatives, and spouses. The allele frequencies of both variants were not different between the groups. No significant associations between plasma lipid traits and DNA variants were observed. We further analyzed the effect of the presence of these variants in the upstream enhancing region of the apoC-III gene, as five different in vivo occurring haplotypes, on the transcriptional activity of apoC-III in both HepG2 and Caco-2 cells. All five promoter constructs, including the wild type, showed similar enhancing activity of the apoC-III gene. The average transcription efficiency was enhanced 19-fold in HepG2 cells and 11-fold in Caco-2 cells. Previous studies have shown in vitro insulin-dependent down-regulation of the apoC-III gene transcription in HepG2 cells by DNA variation in an insulin response element (IRE) in the apoC-III promoter. We observed a 30% insulin-dependent down-regulation of apoC-III expression that was, however, independent of the presence of the two IRE variants. In contrast, in Caco-2 cells, a more variable insulin-dependent up-regulation was found that was also independent of the presence of the IRE variants.In conclusion, our data suggested that the apoC-III gene transcription in vitro is regulated by insulin but independent of the presence of the two IRE variants at -455 and -482. We were unable to detect associations between these apoC-III variants and plasma lipids and insulin in our FCHL population. This means that in vivo apoC-III transcription not only depends upon insulin but appears to be mediated by other mechanisms.

PCR-determined expression of the MDR1 gene in chronic lymphocytic leukemia.

To determine the role the multiple drug-resistance (MDR 1) gene plays in chronic lymphocytic leukemia (CLL), we measured the expression of the MDR 1 gene in 30 patients with this disease. A rapid, highly sensitive, and nonradioactive technique based on the polymerase chain reaction (PCR) was used for that purpose. In this technique, called differential PCR, the target (MDR 1) and a reference gene (beta 2-microglobulin) are co-amplified by PCR from random hexamer-primed cDNA in the same reaction vessel. The level of target gene expression is reflected in the ratio between the intensities of the two resulting PCR product bands, as measured by high-performance liquid chromatography (HPLC). MDR 1 gene expression was detectable in 29/30 (97%) patients with CLL, with a median expression level of 0.36 U (human placenta = 1 U). There was no correlation between expression of the MDR 1 gene and clinical stage, time from diagnosis, absolute lymphocyte count, several lymphocyte surface markers, or prior treatment in the patients analyzed. Immunocytochemical studies of the same material using the monoclonal antibody C219 showed a very low or undetectable expression of the P-glycoprotein in the lymphocytes of all patients studied, whereas granulocytes were significantly more immunoreactive. We conclude that the level of expression of the MDR 1 gene in CLL is generally low, that the removal of granulocytes is important in studies of expression of MDR 1 mRNA in CLL, and that differential PCR provides a rapid and reliable method for quantifying the amount of a specific mRNA, even in very small samples of total RNA.

Molecular alterations in a patient with Turcot's syndrome.

Cells of a patient with Turcot's syndrome and of her parents were evaluated for the presence of molecular alterations in the p53 and the Ki-ras gene. Deletions on chromosome 17p, overexpression and point mutations of the p53 gene as well as mutations of the Ki-ras gene were detected in primary and metastatic tumour but not in the germline of the patient nor in her parents.

Altered expression of the retinoblastoma susceptibility gene in chronic lymphocytic leukaemia.

The pathogenesis of chronic lymphocytic leukaemia (CLL) is unknown. One of the most frequent cytogenetic abnormalities in CLL is a deletion within the long arm of chromosome 13, the region to which the retinoblastoma (Rb) gene has been mapped. Lack of Rb expression has been linked to the carcinogenic process in many human tumours. We therefore sought to investigate the role of Rb gene inactivation in CLL using differential polymerase chain reaction on reverse transcribed RNA. The result of the PCR was quantitated using HPLC. 5/39 patients revealed a lack or significantly impaired expression of the Rb gene upon differential PCR analysis. In addition, immunocytochemical studies were performed using the Rb-specific monoclonal antibody PMG245. 10/56 patients showed a weak or absent expression upon immunocytochemical analysis compared to monocytes or granulocytes. The samples lacking Rb were from both early and late stage CLL. Our results indicate that inactivation of the Rb protein occurs in a fraction of CLL cases and can be found in early and late stages of the disease.

Position specificity of Ki-ras oncogene mutations during the progression of colorectal carcinoma.

The Ki-ras proto-oncogene is converted into an active oncogene by mutations in codon 12, 13, or 61. The incidence of mutations in the Ki-ras oncogene in colorectal adenomas and primary colorectal carcinomas has been shown to be 50-75 and 40-65%, respectively. To determine the role activation of the Ki-ras oncogene plays in the progression of colorectal carcinoma, we analyzed DNA from 11 nude-mouse xenografts and from 24 metastases of 22 patients with colorectal carcinoma, using the polymerase chain reaction technique and hybridization with labeled mutation-specific oligomers. Eleven of the 24 metastases (46%) carried mutations, 7 in codon 12 and 4 in codon 13, whereas only 1 nude-mouse tumor (9%) harbored a Ki-ras codon-12 mutation. Eleven of these 12 mutations in advanced stages of colorectal cancer were localized to the second position of either codon 12 or codon 13, whereas a majority of published ras mutations in earlier stages are in the first position of codon 12 of the Ki-ras oncogene. We conclude that there is a position specificity of Ki-ras oncogene mutations in advanced stages of colorectal carcinoma. In general, however, these mutations do not seem to play an important role in the progression of this cancer.