Known prognostic indicators such as pathological criteria including tumor size, lymph-node status and hormone receptor status are pathologically used in clinical practice [2]. Although numerous studies have shown that disease-free survival after breast cancer treatment may be partially predicted by pathological factors [3,4], they are not yet predicted prognostic factors for each patient in clinical practice.

Epigenetic aberrations occur in tumor cells; especially, CpG-island hypermethylation and global genomic hypomethylation are common features of breast cancer cells [5]. A lack of methyl supply can induce DNA global hypomethylation as well as incomplete conversion of dUMP to dTMP leading to uracil misincorporation into DNA [6]. One-carbon metabolism is a network of interrelated biochemical reactions that include the transfer of one-carbon groups [7], which have two major functions: DNA methylation and DNA synthesis [8].

The principal element of one-carbon metabolism is folate, since the one-carbon transfer reactions involve interconversion between several forms of this nutrient. Other important nutrients in one-carbon metabolism include vitamins B₂, B₆ and B₁₂, which act as essential cofactors for one or more enzymes that catalyze one-carbon transfer reactions. Vitamin B₂ is a cofactor for methylenetetrahydrofolate reductase, the critical folate-dependent enzyme. Vitamin B₆ has a role in the conversion of tetrahydrofolate to 5, 10-methylenetetrahydrofolte, which is required for the synthesis of thymidylate and a precursor for purine synthesis [9,10].

Previous studies have suggested that a high status of one-carbon metabolism factors has reduced the risk of several cancers, including colorectal, pancreatic, esophageal, renal cell, and breast cancer [11-15]. In studies related to breast cancer risk, folate has been investigated many times. Although a number of cohort and case–control studies have suggested a protective effect for a high folate status on breast cancer risk, these results are far from conclusive. And other B vitamins did not showed any associations with breast cancer risk [15]. For studies on one-carbon metabolism related nutrients and breast cancer survival, a few investigated the association between folate intake and breast cancer survival with inconsistent results. In a Swedish mammography study, high folate intake before breast cancer diagnosis improved the prognosis of the breast cancer and overall survival [16]. However, in the Long Island Breast Cancer Study Projects and Nurses’ Health study, the association between vitamin B₂, B₆ and folate and breast cancer prognosis had null results [17,18].

Table ​Table22 presents the association between one-carbon metabolism related nutrients intake and disease progression in breast cancer patients. The mean intakes±SD of vitamin B₂, B₆ and folate were 1.0±0.4mg, 1.5±0.5mg and 213.8±99.3mg respectively. Patients with a high (range, > median) intake of vitamin B₂, B₆, and folate had an increased HR for the recurrence compared to the patients with a low (range, < median) intake; however, the association was not statistically significant. When examining the association between the combined effects of vitamin B₂, B₆ and folate intake on the progression of breast cancer, the HR was 1.46 (95% CI, 0.87-2.43) for the high combined score group (range, 3) compared to the low combined score group (range, 0).

Table ​Table33 presentes the association between the median of one-carbon metabolism related nutrients intake and disease progression in breast cancer patients stratified by clinicopathological characteristics. No differential association was found between the intake levels of vitamin B₂, B₆, and folate and disease progression stratified by TNM stage, lymph-node status, or nuclear grade. However, in ER–/PR– breast cancers, a high intake of vitamin B₂ and folate statistically elevated the HR of breast cancer progression compared to a low intake (HR=2.28; 95% CI, 1.20-4.35, HR=1.84; 95% CI, 1.02-3.32, respectively). Low versus high (reference=low) intake of vitamin B₆ showed an increased HR for breast cancer progression; however, the association was not significant (P_interaction=0.04, HR=1.86; 95% CI, 0.97-3.56).

Furthermore, we evaluated the association between the combined scores of one-carbon metabolism related nutrients intake and disease progression in breast cancer patients stratified by hormone status (Table ​(Table4).4). The positive association between ER/PR status was profound when the nutrient intakes were categorized by combined score (P_interaction=0.018). In ER–/PR– cancers, the high combined score (range, 3) group was associated with a significantly poor DFS compared to those belonging to the low score (range, 0) group (HR=3.84; 95% CI, 1.70-8.71).

In our study, the range for the intake of vitamin B₂, B₆ and folate were 0.1-3.4mg, 0.1-4.8mg and 9–970mg, respectively. These levels of vitamin B₆ and folate were less than the tolerable upper intake level of Dietary Reference Intakes for Koreans (KDRIs). The tolerable upper intake of vitamin B₂ level was not determinable by the KDRIs. In the KDRIs, the tolerable upper intake level of vitamin B₆ and folate were 100mg and 1000mg. The levels of dietary intake of vitamin B₂, B₆, and folate in breast cancer patients were hardly reported. Even though there were many previous studies on the association between vitamin B₂, B₆ and folate and breast cancer risk, they reported the tertiles or quartile levels in both the cases and the controls. Thus, it is hard to compare the nutrients intake level of patients’ with other studies that used different designs.

In observational studies, the results for the relationship between micronutrient intakes and all-cause mortality were inconsistent among breast cancer patients [21]. Only one study investigated one-carbon metabolism related nutrients intake and all-cause of mortality in women with breast cancer, and showed null results for vitamin B₂, B₆ and folate [17]. Few studies have investigated the association between folate intake and breast cancer survival. McEligot et al. studied 516 postmenopausal women diagnosed with breast cancer for 6.6years (median follow-up) and showed that women in the highest tertile for dietary folate intake had an HR of 0.34 (95% CI, 0.18-0.67) regarding all-cause mortality [22]. In addition, in the Swedish mammography cohort, dietary folate intake was inversely associated with overall mortality (HR=0.79; 95% CI, 0.66-0.96) [16]. However, in the Iowa Women’s Health Study, Sellers et al. reported that among 177 breast cancer patients, folate intake had no association with breast cancer prognosis [23]. Moreover, in the Nurses’ Health Study, Holmes et al. provided additional support for no association of breast cancer prognosis with vitamin B₂, B₆ and folate [18]. In addition, Rossi et al. measured the folate levels in plasma from 1024 breast cancer patients in Australia and reported that plasma folate was not significantly associated with breast cancer survival [24]. No other prior studies have reported the relationship between one-carbon metabolism related factors and their combined intake effects on the prognosis of breast cancer. The lack of consensus in the results of previous studies could be explained by the variation in study methodologies, micronutrient source, and total micronutrient level. Thus, it is difficult to determine how the micronutrient intake distribution reported in one study compares to other studies [25].

We observed that the combined intake effects of vitamin B₂, B₆, and folate were associated with breast cancer progression in patients depending on their ER/PR status. No previous studies examining the combined intake effects of one-carbon metabolism related nutrients intake and hormone specific breast cancer have been identified. In the Swedish Mammography Cohort (SMC), only folate intake, which is one of the one-carbon metabolism related nutrients studies, was assessed for association with breast cancer progression [20]. Although Harris et al. have reported that dietary folate intake has shown protective effects on breast cancer-specific mortality in ER-negative tumors, our results do not support this effect. It is difficult directly compare our study’s results to theirs since the distribution of the hormone receptors in the study subjects was different (ER-negative breast cancer, less than 20% in SMC; 39.2% in the present study). For breast cancer risk, few studies have evaluated the relationship between hormone status and folate intake. A higher folate intake was associated with a lower risk of ER-negative breast cancer in the Nurses’ Health Study [26], and in the VITamins And Lifestyle (VITAL) cohort [27]. In SMC, a high folate intake was related to a decreased risk of ER+/PR– breast cancer [28]. Otherwise, few cohort studies have reported largely null results, with no findings for associations between folate intake and ER+, ER–, PR+or PR– breast cancers [26,27,29,30]. From various laboratory studies, their results were inconsistent with our study, which examined whether DNA methylation of the ER CpG island may play a role in suppressing ER gene expression in ER-negative breast cancer cells [31,32]. A greater supply of methyl through a high intake of one-carbon metabolism related nutrients may induce the suppression of gene expression in patients with ER–/PR– breast cancers. In addition, a high intake of one-carbon metabolism related nutrients may have a stronger effect on ER–/PR– breast cancer progression than the other types of breast cancers since ER–/PR– breast cancers are less responsive to hormone therapies. Further studies are needed to elucidate the possible association between one-carbon metabolism related nutrients and the hormone receptors in breast cancer patients.

The results of our study were contrary to the hypothesis based on previous studies that a higher intake of B vitamins would have a protective effect on breast cancer survival in population based studies. One prospective cohort study suggested the association of major energy sources with breast cancer survival may be U-shaped rather than linear [33]. As far as this idea, the association has not yet been proven; however, a midrange intake of vitamins is associated with the most favorable outcomes, and extremes are associated with less favorable outcomes.

This study has some limitations. First, there were likely errors in our estimate of dietary habits. Patients were asked about their dietary intake for the year preceding the diagnosis using the FFQ. Due to this, measurement errors likely occurred because of poor recall despite the validity and reproducibility evidence of the questionnaire. The FFQ correlations were lower than that reported in western countries which were between 0.5-0.7 [34]. In Asia, the median of the correlation coefficients for the FFQ has ranged from 0.3-0.5 in Japan [35,36] and Korea [37,38], and a lower FFQ correlation may have been caused by the dining etiquette and cultural foods of Korea [20]. Though the correlation coefficients were low, to date, FFQ is the only method in which long-term usual dietary intake of an individual can be easily obtained with a single measurement [39]. Second, intake of supplements was not available for the calculations. However, the intake of supplements use can improve the dietary quality for certain micronutrients [25]. Lastly, we could not evaluate the association between one-carbon metabolism related nutrients and breast cancer specific mortality because the data for the cause of death were not available. Thus, the results must be interpreted cautiously and need to be confirmed by a study that investigates the association of one-carbon metabolism related nutrients with breast cancer specific survival.