The New England Journal of Medicine
e-mail icon  FREE NEJM E-TOC    HOME   |   SUBSCRIBE   |   CURRENT ISSUE   |   PAST ISSUES   |   COLLECTIONS   |    Advanced Search
Sign in | Get NEJM's E-Mail Table of Contents — Free | Subscribe
 
Original Article
Volume 329:517-523 August 19, 1993 Number 8
NextNext

An Association between the Risk of Cancer and Mutations in the HRAS1 Minisatellite Locus
Theodore G. Krontiris, B. Devlin, Daniel D. Karp, Nicholas J. Robert, and Neil Risch

 

This Article
-Abstract

Tools and Services
-Add to Personal Archive
-Add to Citation Manager
-Notify a Friend
-E-mail When Cited

More Information
-PubMed Citation
ABSTRACT

Background The role of mutations in proto-oncogenes and their regulatory sequences in the pathogenesis of cancer is under close scrutiny. Minisatellites are unstable repetitive sequences of DNA that are present throughout the human genome. The highly polymorphic HRAS1 minisatellite locus just downstream from the proto-oncogene H-ras-1 consists of four common progenitor alleles and several dozen rare alleles, which apparently derive from mutations of the progenitors. We previously observed an association of the rare mutant alleles with many forms of cancer, and we undertook the present study to pursue this observation further.

Methods We conducted a case-control study, typing 736 HRAS1 alleles from patients with cancer and 652 from controls by Southern blotting of leukocyte DNA. We also carried out a meta-analysis of this study and 22 other published studies, estimating the relative risk of cancer (such as bladder, breast, or colorectal cancer) when one of the rare HRAS1 alleles was present.

Results Both the present case-control study (odds ratio, 1.83; 95 percent confidence interval, 1.28 to 2.67; P = 0.002) and the present study combined with our previous study (odds ratio, 2.07; 95 percent confidence interval, 1.47 to 2.92; P<0.001), as well as the meta-analysis of all 23 studies (odds ratio, 1.93; 95 percent confidence interval, 1.63 to 2.30; chi-square = 57.58; P<0.001), replicated our original finding and demonstrated a significant association of rare HRAS1 alleles with cancer. We found significant associations for four types of cancer: carcinomas of the breast, colorectum, and urinary bladder and acute leukemia. We also identified suggestive but not statistically significant associations for cancers of the lung and prostate and for non-Hodgkin's lymphoma.

Conclusions Mutant alleles of the HRAS1 minisatellite locus represent a major risk factor for common types of cancer. Although the relative risk associated with the presence of one rare allele is moderate, the aggregate prevalence of this class of mutant alleles implies an extremely important attributable risk: 1 in 11 cancers of the breast, colorectum, and bladder.

An intriguing feature of the human genome is the accumulation of distinct classes of repetitive DNA. These classes of DNA consist of several hundred to several thousand base pairs (bp), are dispersed throughout the genome, and are conserved in primates and other mammals. Except for the alpha satellite DNA that organizes the centromere, no repeat element possesses a clearly defined function.

Despite the mystery about both the origins and the functions of repetitive DNA, these elements have been implicated in the pathogenesis of human genetic disease. As passive participants in genetic damage, repeat sequences serve as sites of homologous recombination, resulting in the interstitial deletion and translocation of chromosomes. As active participants, certain repeat elements can transpose themselves to (become inserted in) new sites in the genome. Such movement has caused new mutations leading to factor VIII deficiency1. The expansion of trinucleotide repeats is the basis of Huntington's disease2 and myotonic dystrophy,3 and variation of dinucleotide repeats is observed in hereditary nonpolyposis colon cancer4.

One family of DNA repeat elements, designated "hypervariable minisatellites," "variable number of tandem repeats," or "variable tandem repeats," arises from the head-to-tail concatenation of short-sequence motifs 10 to 100 bp long5. These arrays, through a mutational process that increases or decreases the number of repeat motifs in a given allele, are frequently unstable and produce highly polymorphic loci that often possess dozens of alleles. Consequently, minisatellites have proved extremely useful both as markers in linkage analysis and as "DNA fingerprint" probes in forensic analysis.

Although minisatellites occur more frequently in telomeric regions (i.e., the natural ends of chromosomes), many are intimately associated with genes and gene clusters within chromosomes6. Both the alpha-globin and immunoglobulin heavy-chain loci contain several distinct minisatellites in both intergenic and intronic locations. Minisatellites appear within introns of the retinoblastoma, interleukin-6, von Willebrand factor, and myoglobin genes. Insulin, apolipoprotein B, and collagen type II genes bear minisatellites immediately upstream or downstream from coding sequences. Some genes, such as epithelial mucin, involucrin, and proline-rich proteins, contain transcribed and translated minisatellite arrays.

The proto-oncogene locus HRAS1, which encodes a protein involved in mitogenic signal transduction and differentiation,7 is also tightly linked to a minisatellite. The minisatellite is located approximately 1000 bp downstream from the gene's coding sequences and is composed of 30 to 100 units of a 28-bp consensus sequence8. Thirty alleles of 1000 to 3000 bp have been described9. The four most common alleles -- a1, a2, a3, and a4 -- represent 94 percent of all alleles in whites10 and have apparently served as progenitors for the remaining rare alleles. Molecular genetic analysis of the HRAS1 minisatellite has revealed an orderly locus structure in which the lineage of each rare allele may be traced, through the mutational insertion or deletion of 28-bp repeat units, to the common allele nearest in size (Figure 1)11,12.


View larger version (24K):
[in this window]
[in a new window]
  		Figure 1. Mutational Relation of Rare HRAS1 Alleles to Common Alleles.

In the human HRAS1 gene a BamHI (B) fragment contains the entire locus; the open rectangles represent the four coding exons, and the arrow denotes the initiation site for synthesis of messenger RNA, as well as the region in which transcriptional regulatory signals in the DNA are known to occur. The site of the polyadenylation (Poly A) is also indicated. Just downstream from this polyadenylation signal is the HRAS1 minisatellite. It is closely flanked by recognition sites for MspI (M), the restriction enzyme used to digest DNA for Southern blotting.

An expanded view of the minisatellite shows the common progenitor allele a4. The line of rectangles represents tandem repetition of the 28-bp consensus motif; only a few of the more than 85 units of this allele are shown. The solid, open, and hatched rectangles represent differences in DNA sequence that occur from unit to unit,8 imparting a "secondary structure" to the primary nucleotide sequence. The deletion or addition of repeat units produces mutant alleles, such as a3.5, which can be detected by Southern blotting of leukocyte DNA.

Southern blotting of two samples of leukocyte DNA reveals parental DNA bearing an a4 allele (left panel), which serves as the progenitor of the a3.5 mutation. This mutation is transmitted to the child. In this example, the child's a1 allele is derived from the other parent. Each of the four common alleles (a1, a2, a3, and a4) has given rise to such mutations, which in aggregate are the rare alleles associated with cancer.

 
Some years ago we reported that rare alleles of the HRAS1 minisatellite appear in the genomes of patients with cancer about three times as often as in controls without cancer13; many such alleles have been observed to date only in patients with cancer. Since that study was published, another 28, predominantly smaller, studies of HRAS1 alleles in patients with cancer and controls have appeared in the literature; the majority of these studies have not detected a statistically significant correlation of HRAS1 allelotypes with cancer. In this paper, we report a new case-control analysis that replicated our original findings. We then present a meta-analysis, using our data and data from other published studies, that further strengthens the association of cancer with rare HRAS1 alleles and defines specific cancers for which the association holds. Our conclusions, together with recent investigations demonstrating the interaction of certain minisatellites with transcriptional regulatory factors,14,15,16 lead us to propose that HRAS1 minisatellite mutations interfere with regulatory mechanisms governing the control of gene expression.

Methods

Present Case-Control Study

For our study, we recruited consecutive patients with cancer and control subjects from among unrelated white patients seen at the Division of Hematology/Oncology (case patients) and the Divison of General Medicine (controls) of the New England Medical Center Hospitals; recruitment was conducted at irregular intervals over a period of five years. The types of cancer represented in the study included carcinomas of the head and neck, gastrointestinal tract, breast, lung, kidney, bladder, prostate, ovary, uterus, and cervix, as well as lymphoma, acute and chronic leukemia, melanoma, and sarcoma. The most common tumors were bladder cancer (proportion of cases of cancer, 0.19), non-Hodgkin's lymphoma (0.11), breast cancer (0.10), prostate cancer (0.08), acute leukemia (0.08), lung cancer (0.04), melanoma (0.04), and colorectal cancer (0.03). The mean ages (±SD) of the case patients and the controls were 63.9 ±14.0 and 50.5 ±16.0 years, respectively.

The study protocol was approved by our institutional review board; we received informed consent from all study subjects.

HRAS1 Allele Typing

DNA purification, digestion, electrophoresis, and Southern blotting were performed as previously described9. In brief, leukocyte DNA was digested with MspI/HpaII (New England Biolabs) according to the manufacturer's instructions and subjected to electrophoresis through 1 percent agarose gels until a 1-kb marker migrated 130 mm. Southern blots were probed with nick-translated H-ras-1 DNA. DNA samples containing the four common HRAS1 alleles were included on each gel; other HRAS1 alleles were identified according to their migration relative to these four fragments. All autoradiograms were interpreted by one investigator who did not know whether the DNA was that of a case patient or a control.

Published Case-Control Studies

All studies included in our meta-analysis involved only white patients and controls. When analyzing these data, we classified the subjects' alleles into two groups -- common alleles and rare alleles -- and the subjects in two groups -- case patients and controls. We examined the sizes and frequencies of the alleles to identify the four common alleles and then other alleles by inference. Our classification of alleles according to study is presented elsewhere.* Of the studies investigating the hypothesis of an association between rare alleles and cancer, 22 were included in our analyses. Six other studies were excluded because they evaluated nonwhite subjects (three studies), used patients with cancer as controls (one study), or used patients without cancer as case patients (one study); for one study, we could not determine common alleles from the data presented.

Statistical Analysis

For each case-control study, the odds ratio and its associated statistics were estimated by unconditional maximum likelihood from the standard logit model. For the meta-analysis, the information from all the studies was combined, first with the use of unconditional maximum likelihood17 and the Mantel-Haenszel method18,19 to estimate the common odds ratio, and then with a random-effects model20 to estimate the mean difference between the case patients and the controls in the probability of rare alleles. Because the odds ratios produced with either unconditional maximum likelihood or the Mantel-Haenszel method were similar, we present only the results of the former method unless otherwise noted.

The heterogeneity of the odds ratios was evaluated by means of the statistic recommended by Breslow and Day21. We also performed a goodness-of-fit test that compared each specified logit model against the saturated model. All reported P values are two-tailed.

Results

Association of Rare HRAS1 Alleles with Cancer

Table 1 shows that in typing an additional 1388 alleles from case patients and controls, we replicated our original observations about the frequencies of common and rare alleles in these two groups. In the previous study,13 we found that rare alleles were 3 times more frequent in case patients than in controls (odds ratio, 3.22; 95 percent confidence interval, 1.42 to 7.31; P = 0.005); in the present study, we found that they were 1.8 times more frequent (odds ratio, 1.83; 95 percent confidence interval, 1.28 to 2.67; P = 0.002). When we combined our two studies, producing a total of 900 alleles from case patients and 882 from controls (the allele frequencies are available elsewhere*), the excess of rare alleles among the case patients corresponded to an odds ratio of 2.07 (95 percent confidence interval, 1.47 to 2.92; P<0.001).

View this table:
[in this window]
[in a new window]
  		Table 1. Frequency of HRAS1 Alleles in Two Studies of Patients with Cancer and Controls.

 
Association of HRAS1 Alleles with Cancer in Other Studies

The results of published studies,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43 including those of the present study, are summarized in Table 2. The four common alleles were put into a single class (common alleles), and the remaining alleles into another class (rare alleles). The odds ratio for 22 of the 23 studies was greater than 1 (Table 2), suggesting that the relative risk of cancer was increased among carriers of rare alleles.

View this table:
[in this window]
[in a new window]
  		Table 2. Results of 23 Published Studies of the Association between Rare HRAS1 Alleles and Cancer.

 
The estimate of the common odds ratio for this set of studies, 1.93, was significantly greater than 1 (95 percent confidence interval, 1.63 to 2.30; chi-square = 57.58; P<0.001). The fit of the unconditional-maximum-likelihood model was adequate (chi-square = 30.66, 22 df; P = 0.10). The estimate of the common odds ratio by the Mantel-Haenszel method, 1.92, was quite similar to that obtained by unconditional maximum likelihood, as were the estimates of the standard error of the log odds, 0.086 and 0.087. The odds ratios of the individual studies showed no significant heterogeneity (chi-square = 27.03, 22 df; P = 0.21).

We also examined the hypothesis of an association between rare alleles and cancer by determining the differences between the probabilities of rare alleles among the case patients and the probabilities among the controls. We estimated the mean difference across the studies with a random-effects model,20 which assumed heterogeneity among studies. Thus, with this model, variation of the differences resulted from both experimental error and real differences among studies. The estimate of the mean difference over the 23 studies, 0.040, differed significantly from 0 (t = 2.75, P = 0.006).

We analyzed subgroups of the 23 studies, specifically leaving out our present study, that of Lidereau et al.,33 or both. These two studies could potentially have had a large influence on the meta-analysis: ours because it contained the most subjects, and that of Lidereau et al. because it found an unusually high frequency of rare alleles.

The results of the analysis of these subgroups of studies did not differ substantially from the results of analysis of the original 23 studies. When we excluded our study, the odds ratio was 1.90 (95 percent confidence interval, 1.58 to 2.33; chi-square = 41.80; P<0.001); when we excluded the study by Lidereau et al., the odds ratio was 1.82 (95 percent confidence interval, 1.54 to 2.19; chi-square = 44.11; P<0.001). When we excluded both studies, the odds ratio was 1.74 (95 percent confidence interval, 1.44 to 2.15; chi-square = 28.73; P<0.001). In the subgroup of studies that had the greatest ability to identify alleles -- the studies using the enzyme combination MspI/HpaII (listed elsewhere*) -- the common odds ratio was 1.89. Again, this odds ratio was significantly greater than 1 (log-odds ratio [±SE], 0.64 ±0.11; chi-square = 36.56; P<0.001).

Possibility of Bias in the Meta-Analysis

A factor complicating meta-analysis is publication bias: favorable or unfavorable results remain unpublished because they run counter to prevailing views44. Publication bias was unlikely to have influenced our meta-analysis, for two reasons. First, the association between HRAS1 alleles and cancer has been controversial from the beginning35. Second, the number of published studies claiming a significant association between cancer and rare alleles was only 8 of 22, or 36 percent.

To emphasize that publication bias was an unlikely explanation for the association, we reevaluated the association by analyzing only the 14 studies that did not claim significance for it. In this subgroup, there was a weaker but still significant association between rare HRAS1 alleles and cancer (odds ratio, 1.48; log-odds ratio [±SE], 0.39 ±0.13; chi-square = 9.18; P = 0.0024).

We also examined potential publication bias with use of the visual diagnostic described by Light and Pillemer45; this examination revealed no evidence of bias. Finally, we observed no effect of population heterogeneity (unpublished data) or age.

Risk and Number of Rare Alleles in the Genotype

Ten studies22,24,25,28,30,31,32,38,39 (including our present study) presented genotypic data on case patients and controls, allowing us to distinguish among subjects with no rare alleles, those with one such allele, and those with two.*

Using this set of studies, we estimated the common odds ratio to be 1.85 (chi-square = 23.74; P<0.001) for the presence of one rare allele in the genotype and 4.62 (chi-square = 7.67; P = 0.0056) for the presence of two rare alleles. Therefore, the data suggested that the risk of cancer among carriers of two rare alleles was at least twice that among carriers of only one rare allele. As would be expected because of the small number of studies and genotypes involving two rare alleles, however, these two odds ratios were not significantly different from each other (chi-square = 2.64, 1 df; P = 0.10).

Association of HRAS1 Alleles with Specific Cancers

To assess the consistency of the association across cancers, we analyzed the association of rare alleles with six types of cancer (Table 3 and Table 4) that were represented in at least three studies besides our own. Table 4 presents the analysis of our present study, the previously published studies, and all studies combined. Some of the unpublished studies reported their results in sufficient detail to allow us to use portions of their data when appropriate.*

View this table:
[in this window]
[in a new window]
  		Table 3. Frequency of Rare HRAS1 Alleles among Case Patients in the Present Study, According to Type of Cancer.

 
View this table:
[in this window]
[in a new window]
  		Table 4. Odds Ratios for the Association between Rare HRAS1 Alleles and Cancer.

 
The association between rare HRAS1 alleles and cancer showed some variability across types of cancer, whether only data from our study (Table 3 and Table 4) or those from the published studies (Table 4) were analyzed. When all the studies were included in the analysis, the odds ratio ranged from 1.55 for lung cancer to 2.30 for bladder cancer. At least one of the log-odds ratios for each type of cancer was significantly greater than 0 when the level of significance was 0.05, except for melanoma, which had an associated P value of 0.093 (when all studies were analyzed). In addition, the diseases were not significantly heterogeneous in their association with HRAS1 rare alleles (chi-square = 3.23, 5 df; P = 0.67) when the log odds for each disease were compared with the estimated common log-odds ratio (0.61).

The odds ratio for each disease was consistent with the common odds ratio derived from all the published studies, and the fit of the unconditional-maximum-likelihood model was generally satisfactory. There was a significant lack of fit only for the studies of leukemia, which included a subgroup of data from one study30 that had an odds ratio of 0.91 and two studies23,24 that had inestimable odds ratios (Table 2).

Data from our study of two other tumors, non-Hodgkin's lymphoma and prostate cancer, are shown in Table 3. The raw totals of alleles for these two cancers are presented because both types were well represented among our case patients (11.1 and 8.0 percent, respectively). The estimated odds ratios for these two types of cancer were 1.79 and 1.68 -- values quite close to the common odds ratio derived from the meta-analysis, 1.93. However, neither value was statistically significant.

Discussion

We have shown, both in a single large case-control study and in a meta-analysis including 22 smaller studies, a significant association between a subgroup of HRAS1 alleles and cancer. In aggregate, these 23 studies represent the typing of 8500 HRAS1 alleles, one of the largest molecular genetic analyses of any human proto-oncogene locus. Of the 23 studies, 9 claimed a significant positive association between the mutant alleles and cancer, and 14 did not. However, our meta-analysis of these 14 studies alone also revealed a significant positive association. Therefore, the likely explanation for these "negative" studies was the small size of their study populations and consequent loss of power, rather than an absence of an actual effect. Meta-analysis identified bladder, breast, and colorectal cancer, as well as acute leukemia, as individual types of neoplasms demonstrating the risk associated with rare HRAS1 alleles. Our study suggested an association with melanoma, lung and prostate cancer, and non-Hodgkin's lymphoma, but the data were not sufficient for a definitive conclusion.

From our results, we calculated the risk of cancer attributable to a rare HRAS1 allele. Assuming a relative risk of 1.85 for a heterozygote and 4.62 for a homozygote and a total frequency of 0.058 for rare alleles, we obtained an attributable risk of 0.090. Hence, despite the moderate relative risks computed from the aggregate data, the prevalence of rare HRAS1 alleles implied that 1 in 11 cancers of the breast, colon, and bladder may be attributed to this genetic factor. By contrast, using data derived by Claus et al.,46 we obtained an attributable risk of 0.046 for the breast-cancer-susceptibility gene, BRCA1, recently mapped to chromosome 17q47. This attributable risk was half our estimate for the HRAS1 locus because of the low frequency of BRCA1 in the general population (0.003). If, as now seems likely, lung and prostate cancer are also among the cancers associated with HRAS1, more than 50,000 cases of cancer a year may be attributed to the possession of these alleles.

Two principal issues arise from our observations. The first, of course, is the nature of the phenomenon underlying the association. Rare alleles of the HRAS1 minisatellite may demonstrate linkage disequilibrium with inherited pathogenetic lesions of the HRAS1 locus or other potential disease loci in the vicinity of 11p15.5. This would mean that the rare alleles are simply markers for the risk of cancer but are not involved in its pathogenesis. We cannot yet exclude this explanation, although it seems unlikely11. Several recent findings concerning the functional properties of the HRAS1 minisatellite lead us to propose an alternative hypothesis: new mutations of the HRAS1 minisatellite disrupt the controlled expression of nearby genes, including HRAS1, by interacting directly with transcriptional regulatory mechanisms. We have recently shown that the HRAS1 minisatellite binds at least four members of the rel/NF-kappaB family of transcriptional regulatory factors14. Furthermore, the minisatellite is capable of activating and repressing transcription; intriguingly, allele-specific effects have been observed16. Given these findings, we suggest that most mutations of the four common HRAS1 alleles, whether or not the minisatellite has a physiologic role in the regulation of HRAS1, disrupt nonpathogenetic interactions with rel proteins and render these interactions pathogenetic (Figure 2).


View larger version (25K):
[in this window]
[in a new window]
  		Figure 2. Model Depicting the Influence of the HRAS1 Minisatellite on Transcriptional Regulation.

Proteins encoded by rel/NF-kappaB family members form dimers to produce a wide variety of DNA-binding regulatory factors. These interactions with DNA result in both stimulatory and inhibitory effects on transcription. We predict that mutations of common alleles may disrupt these regulatory networks in two ways, which may affect the regulation of HRAS1 expression. (Shaded circles denote inhibitory factors, and hatched rectangles their preferred recognition sites. Only a few of more than 85 potential binding sites on a4 are represented.) First, altering the number of subunits with preferences for particular classes of factors (some inhibitory subunits are deleted in a3.5 in the figure) may alter the equilibrium concentrations of factors available for binding regulatory sites elsewhere in the genome -- for example, in the transcriptional regulatory region upstream from HRAS1. Second, mutations may lead to the increased binding of classes of rel proteins with more uniform functional consequences that can directly affect HRAS1 transcription. Each mutation can produce a wide range of effects, depending on the combination of inhibitory and stimulatory rel interactions affected.

 
The second issue is whether our results apply to other diseases. Since there are many gene-associated minisatellites, might some exert genetic influence on diseases other than cancer? We have recently shown that the minisatellite within the intron separating the diversity and joining segments of the human immunoglobulin heavy-chain gene binds a transcriptional regulatory protein closely related to members of the mycHLH family of proteins; this binding apparently sequesters the factor in a form that can no longer activate transcription15. The minisatellite just upstream from the insulin gene can suppress transcription but, interestingly, not in pancreatic cells48. A subgroup of alleles at the minisatellite locus for the insulin gene has been associated with type I diabetes49. Like the early studies of HRAS1 in various populations, however, these analyses have produced conflicting results. The intronic minisatellite of the interleukin-1alpha gene possesses binding sites for the transcription factor, Sp-1,50 although no binding of this protein has been directly demonstrated. Collectively, such observations suggest the possibility of a broader contribution of minisatellites to the genetic risk of disease.

Supported by a grant (CA-45052) from the National Institutes of Health.

We are indebted to the physicians, nurses, and staff of the Divisions of Hematology/Oncology and General Medicine of the New England Medical Center Hospitals for their cooperation with this study.

* See NAPS document no. 05045 for 12 pages of supplementary material. To order, contact NAPS c/o Microfiche Publications, 248 Hempstead Tpk., West Hempstead, NY 11552.


Source Information

From the Division of Hematology/Oncology, Department of Medicine, Tufts University School of Medicine and New England Medical Center Hospitals, Boston (T.G.K., D.D.K., N.J.R.); and the Departments of Epidemiology and Public Health (B.D., N.R.) and Genetics (N.R.), Yale University School of Medicine, New Haven, Conn.

Address reprint requests to Dr. Krontiris at Box 245, New England Medical Center Hospitals, 750 Washington St., Boston, MA 02111.

References

  1. Dombroski BA, Mathias SL, Nanthakumar E, Scott AF, Kazazian HH Jr. Isolation of an active human transposable element. Science 1991;254:1805-1808. [Free Full Text]
  2. Macdonald ME, Ambrose CM, Duyao MP, et al. A novel gene containing a trinucleotide repeat that is expanded and unstable on Huntington's disease chromosomes. Cell 1993;72:971-983. [CrossRef][Medline]
  3. Buxton J, Shelbourne P, Davies J, et al. Detection of an unstable fragment of DNA specific to individuals with myotonic dystrophy. Nature 1992;355:547-548. [CrossRef][Medline]
  4. Aaltonen LA, Peltomaki P, Leach FS, et al. Clues to the pathogenesis of familial colorectal cancer. Science 1993;260:812-816. [Free Full Text]
  5. Jeffreys AJ, Wilson V, Thein SL. Hypervariable `minisatellite' regions in human DNA. Nature 1985;314:67-73. [CrossRef][Medline]
  6. Dover GA. DNA fingerprints: victims or perpetrators of DNA turnover? Nature 1989;342:347-348. [Medline]
  7. Barbacid M. ras Genes. Am Rev Biochem 1987;56:779-827.
  8. Capon DJ, Chen EY, Levinson AD, Seeburg PH. Goeddel DV. Complete nucleotide sequences of the T24 human bladder carcinoma oncogene and its normal homologue. Nature 1983;302:33-37. [CrossRef][Medline]
  9. Krontiris TG, DiMartino NA, Colb M, Mitcheson HD, Parkinson DR. Human restriction fragment length polymorphisms and cancer risk assessment. J Cell Biochem 1986;30:319-329. [CrossRef][Medline]
  10. Krontiris TG. Detection of cancer predisposition by hypervariable region analysis. In: Spatz L, Bloom AD, Paul NW, eds. Detection of cancer predisposition: laboratory approaches. White Plains, N.Y.: March of Dimes Birth Defects Foundation, 1990:129-40.
  11. Kasperczyk A, Mermer BA, Parkinson DR, Lonergan JA, Krontiris TG. Allele-specific deletion in exon 1 of the HRAS1 gene. Am J Hum Genet 1989;45:689-696. [Medline]
  12. Kasperczyk A, DiMartino NA, Krontiris TG. Minisatellite allele diversification: the origin of rare alleles at the HRAS1 locus. Am J Hum Genet 1990;47:854-859. [Medline]
  13. Krontiris TG, DiMartino NA, Colb M, Parkinson DR. Unique allelic restriction fragments of the human Ha-ras locus in leukocyte and tumour DNAs of cancer patients. Nature 1985;313:369-374. [CrossRef][Medline]
  14. Trepicchio WL, Krontiris TG. Members of the rel/NF-kappaB family of transcriptional regulatory factors bind the HRASI minisatellite DNA sequence. Nucleic Acids Res 1992;20:2427-2434. [Free Full Text]
  15. Trepicchio WL, Krontiris TG. IGH minisatellite suppression of USF-binding-site- and Emicro-mediated transcriptional activation of the adenovirus major late promoter. Nucleic Acids Res 1993;21:977-985. [Free Full Text]
  16. Green M, Krontiris TG. Allelic variation of reporter gene activation by the HRAS1 minisatellite. Genomics 1993;17:429-434. [CrossRef][Medline]
  17. Breslow N. Odds ratio estimators when data are sparse. Biometrika 1981;68:73-84. [Free Full Text]
  18. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22:719-748.
  19. Robins J, Breslow N, Greenland S. Estimates of the Mantel-Haenszel variance consistent in both sparse data and large-strata limiting models. Biometrics 1986;42:311-323. [CrossRef][Medline]
  20. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177-188. [CrossRef][Medline]
  21. Breslow NE, Day NE. Statistical methods in cancer research. Vol. 1. The analysis of case-control studies. Lyon, France: International Agency for Research on Cancer, 1980. (IARC scientific publications no. 32.)
  22. Barkardottir RB, Johannsson OT, Arason A, Gudnason V, Egilsson V. Polymorphism of the c-Ha-ras-1 proto-oncogene in sporadic and familial breast cancer. Int J Cancer 1989;44:251-255. [Medline]
  23. Boehm TLJ, Hirth H-P, Kornhuber B, Drahovsky D. Oncogene amplifications, rearrangements, and restriction fragment length polymorphisms in human leukaemia. Eur J Cancer Clin Oncol 1987;23:623-629. [Medline]
  24. Carter G, Worwood M, Jacobs A. The Ha-ras polymorphism in myelodysplasia and acute myeloid leukaemia. Leuk Res 1988;12:385-391. [Medline]
  25. Ceccherini-Nelli L, De Re V, Viel A, et al. Ha-ras-1 restriction fragment length polymorphism and susceptibility to colon adenocarcinoma. Br J Cancer 1987;56:1-5. [Medline]
  26. Corell B, Zoll B. Comparison between the allelic frequency distribution of the Ha-ras 1 locus in normal individuals and patients with lymphoma, breast, and ovarian cancer. Hum Genet 1988;79:255-259. [CrossRef][Medline]
  27. Diedrich U, Eckermann O, Schmidtke J. Rare Ha-ras and c-mos alleles in patients with intracranial tumors. Neurology 1988;38:587-589. [Free Full Text]
  28. Gerhard DS, Dracopoli NC, Bale SJ, et al. Evidence against Ha-ras-1 involvement in sporadic and familial melanoma. Nature 1987;325:73-75. [CrossRef][Medline]
  29. Hall JM, Huey B, Morrow J, et al. Rare HRAS alleles and susceptibility to human breast cancer. Genomics 1990;6:188-191. [CrossRef][Medline]
  30. Hayward NK, Keegan R, Nancarrow DJ, et al. c-Ha-ras-1 alleles in bladder cancer, Wilms' tumour and malignant melanoma. Hum Genet 1988;78:115-120. [Medline]
  31. Heighway J, Thatcher N, Cerny T, Hasleton PS. Genetic predisposition to human lung cancer. Br J Cancer 1986;53:453-457. [Medline]
  32. Klingel R, Mittelstaedt P, Dippold WG, Meyer zum Buschenfelde K-H. Distribution of Ha-ras alleles in patients with colorectal cancer and Crohn's disease. Gut 1991;32:1508-1513. [Free Full Text]
  33. Lidereau R, Escot C, Theillet C, et al. High frequency of rare alleles of the human c-Ha-ras-1 proto-oncogene in breast cancer patients. J Natl Cancer Inst 1986;77:697-701.
  34. Mackay J, Elder PA, Porteous DJ, et al. Partial deletion of chromosome 11p in breast cancer correlates with size of primary tumour and oestrogen receptor level. Br J Cancer 1988;58:710-714. [Medline]
  35. Peto TEA, Thein SL, Wainscoat JS. Statistical methodology in the analysis of relationships between DNA polymorphisms and disease: putative association of Ha-ras-1 hypervariable alleles and cancer. Am J Hum Genet 1988;42:615-617. [Medline]
  36. Radice P, Pierotti MA, Borrello MG, Illeni MT, Rovini D, Della Porta G. HRAS1 proto-oncogene polymorphisms in human malignant melanoma: TaqI defined alleles significantly associated with the disease. Oncogene 1987;2:91-95. [Medline]
  37. Riou G, Barrois M, Sheng ZM, Duvillard P, Lhomme C. Somatic deletions and mutations of c-Ha-ras gene in human cervical cancers. Oncogene 1988;3:329-333. [Medline]
  38. Ryberg D, Tefre T, Ovrebo S, et al. Ha-ras-1 alleles in Norwegian lung cancer patients. Hum Genet 1990;86:40-44. [Medline]
  39. Saglio G, Camaschella C, Giai M, et al. Distribution of Ha-ras-1 proto-oncogene alleles in breast cancer patients and in a control population. Breast Cancer Res Treat 1988;11:147-153. [CrossRef][Medline]
  40. Sheng ZM, Guerin M, Gabillot M, Spielmann M, Riou G. c-Ha-ras-1 polymorphism in human breast carcinomas: evidence for a normal distribution of alleles. Oncogene Res 1988;2:245-250. [Medline]
  41. Weston A, Vinseis P, Caporaso NE, Krontiris TG, Lonergan JA, Sugimura H. Racial variation in the distribution of Ha-ras-1 alleles. Mol Carcinog 1991;4:265-268. [Medline]
  42. White GR, Santibanez-Koref M, Heighway J, Thatcher N. Constitutional frequencies of c-Ha-ras alleles in patients with different types of lung cancer. Br J Cancer 1990;61:186-186. [Medline]
  43. Wyllie FS, Wynford-Thomas V, Lemoine NR, Williams GT, Willams ED, Wynford-Thomas D. Ha-ras restriction fragment length polymorphisms in colorectal cancer. Br J Cancer 1988;57:135-138. [Medline]
  44. Dear KBG, Begg CB. An approach for assessing publication bias prior to performing a meta-analysis. Stat Sci 1992;7:237-45.
  45. Light RJ, Pillemer DB. Summing up: the science of reviewing research. Cambridge, Mass.: Harvard University Press, 1984.
  46. Claus EB, Risch N, Thompson WD. Genetic analysis of breast cancer in the Cancer and Steroid Hormone Study. Am J Hum Genet 1991;48:232-242. [Medline]
  47. Hall JM, Lee MK, Newman B, et al. Linkage of early-onset familial breast cancer to chromosome 17q21. Science 1990;250:1684-1689. [Free Full Text]
  48. Takeda J, Ishii S, Seino Y, Imamoto F, Imura H. Negative regulation of human insulin gene expression by the 5'-flanking region in non-pancreatic cells. FEBS Lett 1989;247:41-45. [CrossRef][Medline]
  49. Raffel LJ, Vadheim CM, Roth M-P, Klein R, Moss SE, Rotter JI. The 5' insulin gene polymorphism and the genetics of vascular complications in type 1 (insulin-dependent) diabetes mellitus. Diabetologia 1991;34:680-683. [Medline]
  50. Haugen A, Mann D, Murray C, Weston A, Willey JC. Interleukin-1 alpha gene intron containing variable repeat region coding for SP1 transcription factor recognition sequence is polymorphic. Mol Carcinog 1989;2:68-71. [Medline]

 

This Article
-Abstract

Tools and Services
-Add to Personal Archive
-Add to Citation Manager
-Notify a Friend
-E-mail When Cited

More Information
-PubMed Citation

This article has been cited by other articles:


HOME  |  SUBSCRIBE  |  SEARCH  |  CURRENT ISSUE  |  PAST ISSUES  |  COLLECTIONS  |  PRIVACY  |  TERMS OF USE  |  HELP  |  beta.nejm.org

Comments and questions? Please contact us.

The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved.