Ha-ras-1 restriction fragment length polymorphism and susceptibility to colon adenocarcinoma.

It is not yet clear whether some polymorphic variants of the Ha-ras-1 gene confer genetic predisposition to cancer. However, recent data on myelodysplasia and lung cancer are controversial. To clarify this point, 62 colorectal adenocarcinoma patients were examined for the Ha-ras-1 gene restriction fragment length polymorphism and results were compared with those of 108 healthy blood donors. No Ha-ras-1 polymorphic variants specifically associated with the cancer patients were detected. However, a specific genotype was significantly more frequent in the healthy donors than in the cancer patients (16% versus 5%), suggesting an interaction between the two alleles of the gene. ImagesFigure 1Figure 2

Tumours of different histological types, both in human and blood donors (HBD) by restriction fragment length polyin animal systems, have been found to be associated with ras morphism (RFLP). proto oncogenes activated by mutation and/or overexpression. Activation by mutation has been demonstrated by the DNA Materials and methods transfection technique and ascribed to somatic single base mutations in the region of the 12th and/or 61st codons Subjects (McGrath et al., 1984;Sweet et al., 1984;Gibbs et al., 1984).
These mutations change the primary structure of the p21 Colorectal adenocarcinoma tissue samples and peripheral codified proteins, conferring on them transforming blood leucocytes (PBL) were obtained during surgery from properties. However, mutationally activated ras genes appear 62 CCP. Normal PBL were obtained from 108 HBD at the to be present in only a minority of naturally occurring Transfusion Center of the Pordenone Hospital and used as human tumours (Slamon et al., 1984;Fujita et al., 1984).
controls to determine frequency of the Ha-ras-l alleles in a Activation by overexpression was demonstrated by Chang et normal population. al. (1982), who transfected the human Ha-ras-I protooncogene linked to viral LTR in NIH/3T3 cells. In this configura-DNA extraction tion, the viral LTR enhances many fold the production of the normal human p21 protein in the recipient cells and DNA was extracted from carcinoma samples, from PBL of confers transforming properties on them.
CCP, and from PBL of HBD according to the method of Although it is difficult to demonstrate that ras proto- Wong-Staal et al. (1979).
oncogene overexpression confers transforming properties in natural systems, p21 overexpressions have been detected in a Probe DNAs large number of premalignant and malignant human tumours (Spandidos & Kerr, 1984). Plasmid p344 carrying the normal Ha-ras-1 human gene It is not currently known how ras gene expression is (Pulciani et al., 1982) was grown and purified by standard controlled. However, at least for the Ha-ras-1 proto-methods. Two fragments were used to study the polyoncogene, experimental data from Krontiris et al. (1985) morphism of the VTR region: (a) the BamHI 6.6 Kilobase seem to indicate that a repetitive genomic region called the pair (Kb) fragment encoding the complete Ha-ras-1 sequence variable tandem repetition region (VTR) could have an plus the VTR region, and (b) the 1 Kb MspI and HpaII important function in Ha-ras-I gene expression. fragment encoding the VTR region. The p344 fragments Golfarb et al. (1982) stated that the Ha-ras-I gene is were purified by preparative agarose electrophoresis and highly polymorphic in a human population. This poly-the low melting agarose procedure. Both probes were morphism, first detected by the BamHI restriction enzyme, 32P labelled by nick translation at specific activity was ascribed by Capon et al. (1983) to changes in the >108 cpm g-1. number of repeat units that form the VTR region. This was recently confirmed by Pierotti et al. (1986) by the use of Southern analysis TaqI restriction enzyme mapping. Ha-ras-1 alleles are inherited in a Mendelian fashion and do not arise de novo in Genomic DNAs (1Ougg) were digested to completion with the tumours (Krontiris et at., 1985;Pierotti et at., 1986). They appropriate restriction enzymes, as specified by the therefore supply a potentially useful tool for genetic analysis commercial supplier. Digested DNAs were subjected to of cancer susceptibility conferred by specific alleles of the electrophoresis on horizontal 0.7% or 1.0% (w/v) agarose Ha-ras-l gene (Krontiris et at., 1985). gels in 40 mM Tris acetate, 20mM Na acetate and 2mM To ascertain whether specific Ha-ras-l VTR conforma-EDTA buffer pH 7.6. Denatured DNA fragments were tions confer genetic predisposition to the development of blotted onto a Gene Screen Plus (New England Nuclear, colorectal cancer, we analyzed the distribution of Ha-ras-l1 Firenze, Italy) following standard procedures, as described alleles in 62 colorectal cancer patients (CCP) and 108 healthy by Southern (1975). Hybridization, washing and autoradiography were carried out as described by Ceccherini-Nelli et at. (1982)   Presumably, three fragments were produced, two of which functional Ha-ras-1 gene. These fragments range in size from coirae inorgl.Rsut'hrfr.idctdta 6.6 to 8.2 Kb (Figure la) owing to the variation in length of humian aa alleles could berdiide intocthe mai the VTR region of the gene. This is clearly demonstrated by .
Figure 1 Southern analysis of representative genomic DNAs variable fragments: 2.3 (plus 0.8 and 0.65 Kb), 2.5, 2.7, 2.9, from colonic adenocarcinoma patients and healthy blood donors 3.0, 3.7, 4.0 and 4.2 Kb, with some degree of microheterodigested with (a) BamHI and (b) TaqI restriction enzymes, geneity within each class. TaqI restriction maps were much probed with the p344 Ha-ras-l gene. more resolvable than the ones obtained with BamHI. We therefore classified Ha-ras-I alleles in accordance with TaqI-relevant restriction sites could generate the observed disgenerated fragments.
crepancies. Table I shows that all the Ha-ras-I genotypes, given by Table II gives Taql generated allele frequencies in CCP BamHI and TaqI restriction enzymes both in normal and and HBD populations. Statistical analysis performed by the patient populations, were in accordance with the previously chi square test on allele frequencies shown in Table II did depicted model based on the amplification of the VTR not indicate that any allele was significantly more frequent in region as the origin of the different polymorphic variants of CCP than in HBD. In contrast the genotype type IV (Table the gene (Figure 3). Only 5 out of 340 alleles analyzed I) was significantly more frequent in HBD than in CCP; it generated restriction fragments that did not follow the appeared in 17/108 HBD versus 3/62 CCP x2 =4.48; proposed scheme. We supposed that mutations at the P = 0.034. Repl.
-7500 --BamH  Figure 3 Schematic representation of the Ha-ras-l cloned T24-C3 gene (Pulcianai et al., 1982) is shown in type A: Black boxes represent exons and white boxes the VTR region. Type B: Alleles generated by amplification of the VTR region. Type C: Alleles generated by amplification of the VTR region plus reiteration of TaqI restriction sites within the VTR region. Finally, analysis of Ha-ras-1 RFLP in matched DNA samples derived from tumour and PBL of single patients showed identical restriction patterns by several restriction enzyme digestions.
In no case was evidence found of an allele lost in colon carcinoma tumoral DNA.

Discussion
The Ha-ras-I protooncogene is highly polymorphic in a human population, mainly due to a hypervariability in the length of the VTR region localized at about 1.5 Kb from the the 3' terminus of the gene (Capon et al., 1983). No defined biological properties have been associated with the VTR regions. However, the characteristic of this region (that is 28 bp consensus sequence reiterated 29 times in the p344 gene) suggests that the VTR region could have an important function in the regulation of the expression of this gene. This hypothesis is supported by the report of Krontiris et al. (1985), who found that EJ-ras subclones lacking the VTR region are expressed 5-to 10-fold less than the original clone. Moreover, Ishii et al. (1986) reported that VTR acts as an enhancer element of the Ha-ras-I gene and that specific conformations of this region have stronger enhancer activity.
Ha-ras-I alleles are inherited in a Mendelian fashion. This circumstance has been utilized by different authors to ascertain genetic susceptibility to cancer diseases conferred by this gene. Krontiris et al. (1985) described the association of rare Ha-ras-I alleles with tumours of different histological types. Thein et al. (1986) reported the lack of an association between Ha-ras-l alleles and myelodysplasia. Heighway et al. (1986) found a significant association between a specific Haras-l allele and non-small cell carcinomas of the lung when compared to unaffected controls and small cell carcinoma patients.
In the present report we analyzed Ha-ras-1 gene RFLP in 62 CCP and 108 HBD. Since the median age of the control group was lower than the age at which colon cancer becomes clinically evident, we concluded that allele distribution in our control group represented that present in a general population. The data obtained from the present study indicated that no significant association exists between any Ha-ras-I allele and predisposition to colorectal cancer. In fact, the observed frequencies of the four most abundant alleles were almost identical in patient and control groups. As regards the rare alleles, it is impossible, by this study, to define their influence in this pathology due to their very low frequency observed in the patient group. In contrast, our data seem to suggest that the genotypic asset at the Ha-ras-1 locus could have some influence in determining resistance to the development of colon cancer. In fact, the type IV genotype appears to be more frequent in the HBD than in the CCP group. The meaning of this finding is not yet clear. However, it could indicate that some interaction between the two alleles exists, with a consequent reduction in the frequency of colon carcinoma development in people carrying this genotype. If this situation is confirmed in a larger number of patients and in other tumour types, it could represent an important model to study at the molecular level.
More information will result from these studies when it is possible to link the gene structure analyzed here with defined biological properties; for the present, it is very important that investigators working on this topic decide on a standardized approach. In fact, the presently published studies are very heterogenous: Ha-ras-1 alleles range from 5 or 6 to more than 20 according to different authors, and this situation makes it very difficult to compare data obtained in different laboratories.