Genomics and Cancer Prevention Answers
This web page is archived for historical purposes and is no longer being maintained or updated.
On the partner strand of DNA, the “antisense” strand, what is the corresponding base change?
C to T
The authors present background material describing how intron 1, and specifically the region where the G to A change occurs, is involved in regulation of transcription of the ODC gene. How does this regulation occur?
The protein c-myc binds to the DNA in intron 1 and activates, or increases, transcription of ODC. C-myc is a proto-oncogene, meaning that it can function as an oncogene or cancer-promoting gene when it is highly active in cells. The action of the c-myc protein is opposed, or antagonized, by the MadI protein. MadI is in turn activated by the protein product of the APC gene.
APC is a tumor suppressor gene . What happens when APC is functioning normally?
When APC is functioning normally, MadI acts to oppose c-myc, and this shuts off the ODC gene. The ODC enzyme is involved in production of polyamines, including putrescine, spermidine, and spermine. The ODC enzyme converts ornithine to putrescine, the key first step in polyamine synthesis. Polyamines are found in all bacteria and most animal cells. They perform a variety of functions, mainly serving as growth factors, and help to stabilize membranes in growing cells. The important point to remember is that polyamines are growth promoting. That is why in Scheme 1 on page 7859, when polyamines are reduced, tumorigenesis is reduced: when there are less growth promoting influences, there is less stimulus for carcinogenesis.
What happens when APC is not functioning normally?
When APC is mutated or missing (due to deletion), there is extra c-myc protein around and no MadI to oppose it. Thus, there will be increased ODC activity, increased polyamines, and increased tumorigenesis. The authors reference papers showing that this occurs in mice, and it is likely that a similar situation occurs in humans.
What happens with the ODC intron 1 G allele is changed to A?
The ODC A allele is more responsive to MadI than the G allele. Thus, there will be less polyamines produced in cells with the A allele than in cells with the G allele.
What do the authors hypothesize about the ODC intron 1A allele and risk of polyps?
The authors hypothesize that having one or more copies of the ODC intron 1 A allele will reduce the risk of colon polyps, since there will be less polyamines present and less stimulus for cell growth in the colon.
How does aspirin fit into the scheme?
Aspirin stimulates catabolism, or the breakdown, of polyamines. If polyamine levels drop, there is less stimulus for cell growth in the colon.
What do the authors hypothesize about interactions between the ODC polymorphism and aspirin?
The authors hypothesize that the association between ODC genotype and risk of colon polyps will be “modified by aspirin use” (p. 7860).
How was the reliability of the assay determined? Do you think these measures were adequate?
The authors genotyped 27 (4%) of the 688 samples twice. A more standard protocol is to repeat genotyping on a random 10% of all study samples.
How was the association between ODC genotype and polyp recurrence measured?
Using odds ratios calculated from logistic regression models. The largest group (G/G) is used as the reference group, and odds ratios are calculated for A/A versus G/G and G/A versus G/G.
How was interaction between aspirin use and ODC genotype assessed?
The joint effects of aspirin use and ODC genotype were estimated using individuals with no aspirin use and homozygous for the G allele (G/G) as a common referent group. In this manner, one can estimate the effects of aspirin use alone, A/A genotype alone, and the combination of aspirin use and A/A genotype.
Are the ODC genotypes in Hardy Weinberg equilibrium in the study population?
Using the information in Table 1, the distributions of ODC genotypes in the total cohort are as follows:
To calculate departures from Hardy-Weinberg equilibrium, we first calculate allele frequencies:
G allele frequency = 2 (382) + 1 (264) / 2 (688) = 0.75
A allele frequency = 2 (42) + 1 (264) / 2 (688) = 0.25
Allele frequencies are calculated as the number of chromosomes with a specific allele / total number chromosomes.
Next, determine the estimated distribution of genotypes based upon allele frequencies:
The expected values are based upon the Hardy Weinberg equilibrium (HWE) proportions. Assuming “random mating,” no “genetic drift,” no “new mutations,”no “selection” and no “admixture” within a population, two alleles at a locus should occur in combinations (genotypes) that roughly approximate Hardy-Weinberg equilibrium (HWE).
p = frequency of one allele
q = frequency of the other allele
where p and q are two alternative alleles at a single locus.
If HWE is present, then:
p 2 = frequency of p/p homozygotes
2pq = frequency of p/q heterozygotes
q 2 = frequency of q/q homozygotes
These calculations assume that NO ERRORS have occurred in the genotyping assay. In our laboratory, we use HWE calculations as a check on the validity of our laboratory assays. This is because we know that for most loci, HWE conditions hold among the controls (unaffected persons) within our study population.
Expected ODC genotypes
G/G (0.75) (0.75) x 688 = 387
G/A 2 (0.75)(0.25) x 688 = 258
A/A (0.25) (0.25) x 688 = 43
Calculate a chi-square test statistic, which is the sum of (Obs – Expected) 2 :
(382-387)2 + (264-258)2 + (42-43)2
387 258 43 = 0.23
The degrees of freedom is the number of alleles minus one = (2-1) = 1.
On a chi – square table, the p-value corresponding to a chi – square statistic of 0.23 with df = 1: P= 0.75.
Interpret the p-value for the Hardy Weinberg equilibrium test.
A high p-value is what you want for the Hardy Weinberg test in this situation. If the p value is high, this means that if you randomly selected study participants from a population in Hardy Weinberg equilibrium and genotyped them, and did this an infinite number of times, there is high (in this case 75%) probability that the difference you found (or a greater difference) between the observed and expected genotype distributions would occur simply by chance. Thus, if we assume a threshold of 0.05 for our type one error rate, there are no statistically significant departures from Hardy Weinberg equilibrium for the intron 1 ODC genotypes in this study population. The study population is in HWE for the ODC intron 1 locus. The result is reassuring, given our concern that only 47% of participants gave blood for genotyping: it is unlikely, given the HWE test results, that ODC genotype was related to obtaining blood samples.
How does a Hardy Weinberg equilibrium test serve as a quality control check on genotyping procedures as well as on the underlying study design?
When we are conducting a genotyping assay in the laboratory, the observed genotype frequencies should correspond to the expected genotype frequencies, as calculated on the basis of observed allele frequencies. If this sounds strange, as if you are using the data to check on itself, you are right! We are!
Think of it this way. Your assay took a bunch of DNA and detected alleles on a bunch of chromosomes. If you took these results, and put all the alleles in a big pile, disregarding whom they came from, and then sorted the alleles into piles based upon an equilibrium distribution, how close would the equilibrium distribution approximate what you found in your original population?
If the observed and expected genotype distributions differ, this could be due to:
(a) Population dynamics (genetic drift, selection, new mutations, etc.) leading to a population that is not in Hardy Weinberg equilibrium.
(b) Failure to observe HWE could also be due to failure to draw a random (unbiased) sample of individuals from a population that is in HWE (in other words, due to selection bias). Note: a population sample could exhibit HWE for one or more genetic markers and still be non-representative for one or more environmental factors.
(c) A problem in your laboratory assay. The most common sources of laboratory error are mixing up samples, sample contamination, or a faulty genotyping procedure (e.g., mistakenly genotyping a pseudo-gene, improper design of primers, probes or other reagents). There are several published examples where genotyping errors have lead to failure to achieve HWE.
Note: Not every population, or every population sample, will be in Hardy-Weinberg equilibrium for every genetic marker. Human populations are in flux, so one or more of the conditions in (a) may apply. However, one should always try to minimize the chances of the other two possibilities, (b) and (c). Achieving good response rates for participant enrollment and blood collection are helpful in avoiding (b). Using positive controls (DNA samples of known genotypes) in your assay and repeating some portion of samples using a more rigorous “gold standard” laboratory method are helpful in avoiding (c). For example, the Taqman based methods used in this paper could be checked against a PCR-based restriction-fragment length polymorphism analysis using the PstI restriction enzyme.
Figure 1 presents joint effects where both aspirin use and ODC genotype are taken into account. How would you characterize the joint effects of aspirin use and ODC genotype?
The effects of ODC genotype and aspirin use are greater than multiplicative. The odds ratio (OR) for aspirin alone (among participants with G/G genotype) was 0.83 (95% CI 0.51-1.34). The OR for ODC A/A genotype alone (in non-aspirin users) was 0.68 (95% CI 0.30-1.51). The combined effect of aspirin use and ODC A/A genotype was strongly inverse, OR = 0.10 (0.02-0.66). Note that this estimate is fairly imprecise, due the small number of participants with A/A genotype (n = 42). The joint OR (0.10) indicates a stronger inverse relation than one would have expected based upon simply multiplying the two independent effects together (0.83 * 0.68 = 0.56). The observation that 0.10 is much lower than 0.56 is evidence of gene-environment interaction.
When two factors operate at separate steps in a common multi-step pathway, the joint effects are expected to be greater than multiplicative. Since we know that ODC is involved in polyamine synthesis, observing interactions between the ODC polymorphism and aspirin use provides evidence that the protective effect of aspirin use may also be via polyamines.
How do the authors interpret the findings?
The authors point to the biologic plausibility of the findings. ODC is involved in the first step of polyamine synthesis, and aspirin is involved in catabolism, the breakdown of polyamines: both exposures are affecting a common metabolic pathway. Thus, it makes sense that ODC genotype and aspirin use would act together to lower polyamine levels and thereby lower the risk of polyp recurrence.
What are the implications of the findings for public health?
The main importance of these findings lies in the insight they provide into the biology of polyp formation: ODC genotype helps implicate specific biologic processes, namely the synthesis of polyamines, in the generation of colon polyps. This is in turn helps us to better understand the preventive effects of aspirin and may help find other drugs or dietary interventions that interfere with polyamine synthesis.
Would you recommend that aspirin use for the prevention of polyps only be targeted to persons with ODC A/A genotype?
No, there was a weak inverse association of aspirin use in persons with the ODC G/G genotype.
Should persons with the ODC G/G or G/A genotype be targetted for increased screening for polyps and colorectal cancer, for example, through fecal occult blood testing or colonoscopy?
In this study, 98% of persons were G/G or G/A genotype, so not much would be gained in screening programs by focusing on carriers of the G allele. Furthermore, the risk of polyp recurrence in those with A/A genotype in this study of persons with previous polyps was high enough (36%) that one would not want to deprive A/A homozygotes of screening.
What other studies should be done?
The authors describe a variety of additional laboratory studies that could be done to better understand the effects of the ODC genotype and interactions with aspirin and other non-steroidal anti-inflammatory drugs. It would be interesting to genotype participants in a randomized trial of aspirin use for the ODC polymorphism. In such a trial, aspirin use would be controlled and dose/duration carefully monitored, thereby increasing the strength of evidence. It would be interesting to examine the association of ODC genotype, and joint effects with aspirin use, using colorectal cancer as the outcome, as well as other cancers (e.g. breast cancer) for which aspirin has shown protective effects. ODC genotype might be helpful in sorting out why some studies and some study populations show protective effects of aspirin and cancer while others do not.