Gene Polymorphisms MTHFRC677T and MTRA2756G as Predictive Factors in Adjuvant Chemotherapy for Stage III Colorectal Cancer

Discussion

Chemotherapy is used in treatment of colorectal cancer in several ways related to the indication (3). The treatment itself carries risks of side-effects, which, in milder cases can affect the patients' well-being and quality of life. However, the treatment can also lead to life-threatening conditions such as severe infections and toxicity of the bone marrow.

Adjuvant treatment is used for patients who has confirmed lymph node metastasis, but are without radiological evidence of distant metastasis and thus having stage III disease. The risk assessment can be difficult and differs from the palliative situation. There are some prognostic measures, such as lymph node ratios, which can assess the risk at group level, but the risk for the individual patient is unknown (13). Thus, it can be hard to ascertain the benefit of the treatment for the individual patient. Development of predictive factors is important in order to increase treatment accuracy and minimize patient toxicity. An individualized chemotherapy regimen, by the use of predictive factors, should have several important benefits, such as less patient suffering and a better chance of treatment completion by reducing the risk of side-effects.

Patients with the MTHFR 677 CC genotype had a significantly lower risk of chemotherapy related toxicity and a better survival than those with CT/TT genotypes. These findings concur with those that have been previously reported, including from our own group (14, 15). The main target of 5-FU is folate metabolism, by inhibiting TS (4). The associated enzyme MTHFR converts 5,10MTHF to 5THF. The product provides a methyl group to convert homocysteine to methionine which is supported by MTR. Thus, a less active enzyme system, with MTHFR 667 CC being more effective than MTHFR 667 CT/TT, could theoretically result in increased 5,10MTHF-levels. This could lead to a higher efficacy of 5-FU in binding and inhibiting TS and could thus increase the toxicity. Moreover, it could also lead to a lower level of substrate for the conversion of homocysteine to methionine by MTR. This could affect the availability of methyl groups needed in DNA gene regulation through the methylation process and might even have a prognostic impact.

Previous findings on TS and MTHFR gene polymorphisms as predictive factors looked promising as described in several publications (7, 8, 16). However, the data have been hard to validate and have even occasionally been contradicted (17, 18). Thus, no real clinical utilization has yet been found, as noted by Tejpar et al. (19). There are many difficulties in the search for predictive factors, including the increasing molecular heterogeneity of the colorectal cancer itself (20, 21). The absence of validated conclusive findings could also suggest that the molecular heterogeneity makes the use of single markers more difficult.

The MTR A2756G polymorphism has been suggested to affect the risk of developing cancer but has otherwise, to our knowledge, not been clinically studied in cancer disease (22, 23). In the present paper, there was a link between MTR genotype and the risk for a need for dose reduction due to toxicity. The lack of findings for toxicity in association with MTR itself suggested a possible functional interdependency. Thus, the following exploration of a connection between the different genotypes of MTHFR and MTR gave findings that challenge and modifiy previous results. A patient who has an optimal enzymatic activity of MTHFR, which genotype 677 CC provides, in combination with MTR 2756 AA has the least risk of side-effects. The risk of side-effects increases for patients who instead have the MTR AG or GG genotypes. The pattern is similar but more pronounced for patients with MTHFR 677 CT and TT genotypes. The risk of nausea could serve as an example, where a patient with the MTHFR 667 CT or TT and MTR 2756 AG or GG genotype combination had 4.5 times higher risk (p<0.014) than patients with the MTHFR CC and MTR AA genotype combination. The functional interdependency did also, in this material, affect the survival, where again the difference by MTHFR function only is seen for the patients with MTR 2756 AA genotype.

Distinction can be made between prognostic and predictive factors where the latter relates to the prognosis for a specific treatment rather than being an overall indicator of survival. The MTHFR C677T polymorphism could then be suggested as a predictive factor for 5-FU treatment in colorectal cancer. However, that for MTHFR would be valid only for 95 out of 150 patients and the findings in the present study suggest that the outcome can be altered by MTR A2756G, which then should be added as a potential factor. Thus, it could be difficult to advocate the use of a single marker for prediction considering these findings. It is also highly plausible that the same connection and interdependence could exist with other enzymes such as TS. It is not unreasonable that the findings could, at least partly, explain some of the difficulties in reaching coherent and conclusive results with single predictive markers for use in tailoring and individualising chemotherapy. Identifying and testing markers and combinatory risk scores for the overall function of the metabolism could lead to further progress.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Survival after adjuvant chemotherapy in stage III colorectal cancer where the MTHFR 677 CC genotype has a better outcome the MTHFR CT/TT genotype. The difference was only seen and significant for the patients who also had a MTR 2756 AA genotype.

The study was performed at a single centre treating patients from a specific geographic region. Thus, the treatments and assessments were carried out coherently and along the same guidelines for all patients and also allowed for a good follow-up of the patients. The cancer staging was adequately performed, with adequate node assessment data, thus minimizing the risk of stage migration effects (24). The dietary folate content has been suggested to possibly influence the effect of the polymorphisms. The actual folate levels in blood or colorectal mucosa was not measured in this material, which is a possible weakness of the study. There is always an inter-individual variation in dietary intake and it has been suggested that low levels could affect the genotype-related risks. However, since the present patient cohort consists of an unselected material from a region without special dietary limitations, it is reasonably safe to assume that the dietary difference at a group level is not significant. There could also be geographical variations in genotype prevalence between different regions of the world, which could affect the findings from other research groups. The results here could support a hypothesis of MTHFR C677T beeing a predictive factor for 5-FU treatment in stage III colorectal cancer. However, the MTR A2756G gene polymorphism can affect and modify the risk of toxicity and the interdependency should not be neglected. It would be of interest to confirm these findings in further studies. A study with actual measurement of tissue levels of the folate metabolites would also be of interest and could bring new information on the complex function of folate metabolism as a whole.