Introduction

Over the last few decades, there has been increased interest in how an array of dietary components may affect cancer progression and whether any of these components have potential for use as chemopreventative agents. Folate, also known as vitamin B9, has been the subject of numerous studies seeking to understand its role in cancer progression [1–6]. This is due, in part, to the 1996 mandate by the United States Food and Drug Administration (FDA) requiring the supplementation of wheat, rice, bread, cereals, pasta, and other grain products with folic acid in an effort to reduce the incidence of neural tube defects in newborns [7]. Following this mandate, several studies were published that put forth conflicting findings regarding the influence of folate supplementation on colorectal cancer (CRC) progression; nearly two decades later, the debate over whether folate promotes or prevents CRC continues. The controversy surrounding the influence of folate on tumorigenesis is likely due to its position at the center of a complex metabolic network that influences a variety of biological processes including nucleotide synthesis, DNA integrity, and gene expression through epigenetic regulation [2]. The most thoroughly characterized function of folate is its immediate role in one-carbon metabolism. Through a series of metabolic reactions, dietary folate can be converted into several forms, each of which plays a role in distinct metabolic pathways. The three most thoroughly investigated folate metabolites are 5-methyl tetrahydrofolate (THF), 10-formyl THF, and 5,10-methylene THF. These three metabolites all provide one-carbon groups for use in a variety of biosynthesis reactions (Fig. 1) [8]. 10-Formyl THF and 5,10-methylene THF are key intermediates required for thymidylate and pyrimidine biosynthesis, respectively, while 5-methyl THF feeds one-carbon groups into the homocysteine metabolic cycle, the pathway which serves as the primary source of methyl groups used in DNA methylation reactions [9]. It is because of folate’s involvement in many biological processes relevant to cancer development, and its potential to modulate these processes, that it has been widely considered for its potential chemopreventive benefit. However, despite significant potential and a large body of research into its multifaceted functions, understanding of the intricate effects of folate intake and supplementation on cancer risk remains elusive.

Fig. 1
figure 1

a Overview of folate metabolism. Dietary folate is converted into several key metabolites by a series of enzymatic conversions. b 5,10-Methylene tetahydrofolate is used as the substrate for thymidylate synthesis, where a methyl group is donated to deoxyuridine monophosphate (dUMP) to form deoxythymidine monophosphate (dTMP). c 10-Formyl tetrahydrofolate performs a similar function in the synthesis of the purine nucleotide precursor, inosine monophosphate (IMP). Initially, 10-FTHF donates a formyl group to convert glycinamide ribonucleotide (GAR) to N-formylglycinamide ribonucleotide (FGAR). In a later step, 10-FTHF donates a formyl group to convert 5-Aminoimidazole-4-carboxamide ribonucleotide (AICAR) into 5-formamidoimidazole-4-carboxamide ribotide (FAICAR). d Methyl groups can be shuttled from 5-methyl tetrahydofolate (5-MTHF) to the homocysteine cycle, where they are used to regenerate the “universal methyl donor,” S-adenosyl methionine (SAM). SAM in turn serves as the substrate for the DNA methyltransferase enzymes, which transfer a methyl group from SAM to cytosine, resulting in the formation of 5-methylcytosine

Full size image

Folic Acid and Colorectal Cancer Risk

A large number of studies, both preclinical and clinical, have sought to understand how intake of folic acid influences tumor progression; a collection of key studies can be found in Table 1. Unfortunately, this body of work has not yet been able to provide a clear picture of how folate intake influences colorectal cancer risk. The earliest investigations into the relationship between folate and CRC risk were typically epidemiological case–control studies examining associations between CRC incidence and folate intake determined from dietary questionnaires. Three such studies that were among the first to report on folate intake and CRC risk demonstrated an inverse relationship between folate consumption and colorectal cancer incidence [2, 10–12]. Importantly, all three of these studies were conducted in countries that, at the time, had not implemented folic acid fortification programs and were therefore more likely to have larger portions of the population with inadequate folate intake. For the most part, studies of folic acid intake and the risk of colorectal neoplasia conducted in the United States after the implementation of the national fortification program have also supported an inverse relationship between intake and risk in the general population [13–15]. However, a key randomized clinical trial conducted by Cole and Baron [16] challenged this long-standing paradigm by demonstrating that very high doses of folic acid (1 mg/day) in individuals with a history of colorectal adenoma do not decrease the risk of future adenoma and may actually result in an increased risk. Additionally, a more recent meta-analysis of six large folic acid supplementation trials has concluded that folic acid supplementation is associated with an increased risk of cancer [17••].

Table 1 Key preclinical and clinical studies of folate intake and colorectal cancer risk
Full size table

Studies in rodent models have suggested that once neoplastic lesions have been established folate deficiency slows tumor progression [18, 19, 20•]. As a result, it has been suggested that folic acid has a “dual modulatory effect” on carcinogenesis [21]; prior to the establishment of neoplastic lesions, high folate intake decreases the risk of cancer, while high folate intake after the establishment of neoplasia drives progression. Thus, within the context of folate-based chemoprevention strategies, careful consideration of dosage and timing with respect to carcinogenic initiation are critical determinants of the safety and efficacy of an intervention.

Folic Acid as a Carbon Source in De Novo Nucleotide Synthesis

Folic acid metabolites are used as substrates in the biosynthesis of both thymidylate and purine nucleotides. Once folate has entered the cell, it is reduced to tetrahydrofolate (THF), which is the derivative that acts as the central nexus of the folic acid metabolic cycle. THF goes through a series of enzymatic conversions, with each step in the pathway giving rise to an intermediate that serves a unique role. The most important folate derivatives for nucleotide metabolism are 5,10-methylene THF and N10-formyl THF. 5,10-methylene THF is the substrate for the enzyme thymidylate synthase which transfers a methyl group from 5,10-methylene THF to deoxyuridine monophosphate (dUMP) to generate thymidine monophosphate (dTMP) [22]. N10-Formyl THF, on the other hand, is the substrate for two enzymes involved in separate steps in the biosynthesis of purine nucleotides: Phosphoribosylglycinamide formyltransferase (GART) and phosphoribosylaminoimidazolecarboxamide formyl transferase (AICARFT) [23, 24]. GART transfers a methyl group from N10-formyl THF to glycinamide ribonucleotide (GAR) to generate the purine intermediate N-formylglycinamide ribonucleotide. Later in the purine biosynthetic pathway, AICARFT transfers another carbon from N10-formyl THF to AICAR to generate the purine intermediate 5-formamidoimidazole-4-carboxamide ribonucleotide (Fig. 1) [25]. Thus, the supply of folate is incredibly important for nucleotide biosynthesis, as it is the primary source of crucial structural carbons in purine precursors and of the carbon that is necessary to convert dUMP to dTMP. Imbalances in these key folate metabolites have previously been detected in colorectal tumors, underscoring the relevance of this metabolic pathway to colon carcinogenesis [26].

It is folic acid’s role in nucleotide metabolism that has made it an interesting chemopreventive candidate from a mechanistic standpoint. It has been hypothesized that folate deficiency causes DNA damage by reducing the availability of nucleotide synthesis precursors and thus causing an imbalance in nucleotide pools [27]. Such imbalances are known to lead to uracil misincorporation into DNA, DNA mismatches, and double-stranded breaks [27, 28]. Inhibition of one-carbon metabolism, such as that caused by dietary folate deficiency, has been shown to cause uracil misincorporation and DNA double-stranded breaks in lymphocytes and colonic epithelial cells in rodent models [29–31]. It has also previously been shown that dietary folate deficiency is capable of inducing both uracil misincorporation and chromosome breaks in human DNA [32, 33]. Thus, when folate status is low in a normal cell, that cell will accumulate DNA damage and have a higher probability of undergoing transformation; on the one hand, folic acid supplementation has been proposed to reduce CRC risk by ensuring the ample supply of nucleotide precursors. On the other hand, this same mechanism might help to explain the inhibitory effect of folate deficiency on tumor growth. Cells which have undergone transformation tend to have a higher rate of replication and a higher demand for nucleotides than their comparatively slow-replicating normal counterparts. In these transformed cells, low folate status places a severe restriction on nucleotide biosynthesis, inhibiting DNA replication and slowing tumor growth [34]. A recent study by Witherspoon [35•] revealed that the chemopreventive agent α-difluoromethylornithine (DFMO) exerts its protective effects through a similar mechanism; DFMO alters the metabolic availability of folate derivatives and inhibits thymine biosynthesis.

It is possible that transformed cells are unable to cope with the rapid accumulation of DNA damage caused by folate deficiency and eventually activate apoptotic pathways, resulting in the death of the initiated cell before it can develop into a more advanced lesion. In fact, this mechanism has been used for many years as the basis for antimetabolite chemotherapy; 5-fluorouracil (5-FU) acts by inhibiting thymidylate synthase which leads to the reduced availability of thymidylate, inhibition of DNA replication, and ultimately, the induction of apoptosis [36]. Furthermore, antifolate drugs such as methotrexate and pemetrexed function by inhibiting steps in folate one-carbon metabolism and have consequences for DNA integrity that are similar to those of 5-FU [37].

Confusion regarding the effects of folate intake on CRC risk may arise from the fact that these two effects of deficiency are not necessarily mutually exclusive. An individual with low folate intake may have two distinct subpopulations of colonic epithelial cells: normal cells which are sustainin DNA damage and are pushed towards transformation and a population of initiated cells the progression of which is inhibited by folate restriction. For the purpose of chemoprevention, then, a folate-based intervention may take one of two forms depending on an individual’s personal risk profile. Endoscopic detection of an adenoma is a known risk factor for future neoplasia, suggesting that individuals with a history of adenoma may possess multiple synchronous neoplastic foci capable of progressing to a more advanced lesion. In these cases, it may be of greater benefit to the patient to restrict folate intake so as to inhibit the progression of existing neoplasia; in a young individual with a colon free of neoplastic foci, it may be of greater benefit to supplement folate intake in order to prevent DNA damage and reduce the likelihood of neoplastic initiation. A key task which must be completed in order to maximize the utility of folate chemoprevention is the characterization of the effects of various levels of folate intake on cellular viability at different stages of carcinogenesis. From these data, it may be possible to identify individuals who belong to specific at-risk subpopulations that stand to benefit most from folate supplementation or restriction. The overall effect on carcinogenesis of a given folate intake level may be quite different for certain high-risk individuals, such as patients with a personal history of colorectal cancer or individuals with familial adenomatous polyposis, than it is for the general population.

Folic Acid as a Source of One-Carbon Groups for DNA Methylation

A second mechanism through which folate intake has been proposed to offer chemopreventive effects is the modulation of DNA methylation patterns. In a landmark paper, Feinberg and Vogelstein established the connection between genomic hypomethylation and colorectal carcinogenesis by demonstrating that, compared with matched normal tissue, primary tumor samples were significantly hypomethylated at a specific subset of genes [38]. This finding was expanded upon when Feinberg later demonstrated that, compared with normal mucosa, adenomas and adenocarcinomas have reductions in total genomic 5-methylcytosine content of 8 % and 10 %, respectively [39]. A link between hypomethylation and carcinogenesis was later demonstrated in an animal model when Gaudet found that mice expressing hypomorphic Dnmt1 developed aggressive lymphomas between 4 and 8 months of age [40]. It was later revealed that, while tumors do undergo a global reduction in DNA methylation, they also become hypermethylated at specific CpG islands in a tissue-specific manner, suggesting that there exists a more general relationship between the disruption of normal methylation patterns and colorectal carcinogenesis.

The supply of one-carbon groups for use in methylation reactions is dependent on metabolic flux through the homocysteine cycle [41], which generates S-adenosyl methionine (SAM), the “universal methyl donor.” Dietary folic acid is eventually converted to the metabolite 5-methyltetrahydrofolate (5-MTHF); a methyl group is then transferred to homocysteine to generate methionine, which reacts with ATP to generate SAM. SAM is the substrate for DNA methyltransferase enzymes (DNMTs) which transfer a methyl group to DNA CpGs, converting SAM to S-adenosyl homocysteine (SAH) and the target cytosine to 5-methylcytosine (Fig. 1) [42]. The availability of SAM and SAH are important regulators of DNMT activity and the maintenance of patterns of DNA methylation. Considering the importance of folate for homocysteine metabolism and DNA methylation, it has been suggested that folate-based chemoprevention may help to maintain normal patterns of methylation [43, 44].

Studies of the relationship between folic acid intake and DNA methylation status have produced inconsistent results. While some studies have reported that folate supplementation has no effect on global DNA methylation [33, 45, 46], there is also ample evidence to support a relationship between folate intake and methylation. Several studies have shown that increased folic acid intake is associated with an increase in global DNA methylation in both leukocytes and colonic epithelial cells [47, 48]. It has also been shown that folate deficiency causes both a global reduction in DNA methylation and specific hypomethylation at the p53 gene promoter in human colon carcinoma cells [43]; it was further demonstrated that both types of hypomethylation could be reversed when folic acid was added to the culture medium. This finding was partially recapitulated in an animal model when Apc Min/+ mice were fed a diet deficient in folate: folate deficiency was associated with both a decrease in DNA hypomethylation and SAM [49]. Interestingly, in a 2007 study by Keyes [48], this positive correlation between folate intake and global DNA methylation was found to be present in 18-month-old mice but not in 4-month-old mice. A more recent study showed that, in old mice (18 months), administration of a diet deficient in folate and supplemented with 18 % ethanol caused a significant reduction in global DNA methylation compared with control [50]. These studies suggest an age-dependent susceptibility of DNA methylation patterns to modulation by folate intake.

There have been only a few clinical studies investigating the relationship between folate intake and DNA methylation in humans, but, in general, they support a positive correlation between folic acid intake and DNA methylation. Early studies indicated that supplementation of folic acid intake caused an increase in DNA methylation in cells of the colonic mucosa [47, 51, 52]. However, one drawback of these early studies is that they all relied on the methyl acceptance assay, which has been reported to be subject to a large degree of variation [53]. Two more recent studies, which measured DNA methylation with more sophisticated techniques, have also produced inconsistent findings; one study confirmed the positive correlation between folate intake and DNA methylation in the colon while the other did not [33, 54].

Importantly, the findings of two key studies may warrant a reevaluation of our understanding of the effects of folate intake on methylation during colorectal carcinogenesis. These studies showed that, when the activity of either Dnmt1 or Dnmt3a is reduced in Apc Min/+ mice, there is a dramatic reduction in intestinal tumor multiplicity; in the case of Dnmt1, tumorigenesis is suppressed completely. These studies suggest that, in the intestine and colon, carcinogenesis is dependent on DNMT activity. It is well known that DNA methylation is critical for regulating chromatin conformation [55], and it is thought that hypomethylation of a particular region of DNA may cause it to be more susceptible to accumulating damage. In the past, it has been proposed that folic acid supplementation might provide chemopreventive protection by preventing hypomethylation and reducing susceptibility to genomic damage [56]. If, however, colorectal carcinogenesis is dependent on methylation, it may actually be beneficial to restrict the generation of methyl groups destined for use in DNA methylation in certain situations. This view of folate-dependent methylation in the context of colorectal cancer may help to explain the anti-tumor effects of folate deficiency that have been recently described in mouse models [19–21]. Just as the effects of folate intake on tumorigenesis appear to vary depending on timing and dosage, they may also vary depending on an individual’s age and epigenetic background.

Taken together, the variability in these studies suggests that the effects of folate intake on DNA methylation may be modulated by confounding factors, including age and intake of alcohol. If modulation of methylation patterns is a key mechanism through which folic acid chemoprevention may exert its effects, it will be important to tease out which members of the DNA methylation pathway (SAM, SAH, DNMTs, etc.) are the dominant regulators. In addition, it is possible that, based on their unique epigenetic profile, certain individuals may benefit from folate-based chemoprevention more than others. It will be important to understand how outside factors, such as other dietary factors or advanced age, interact with folate-dependent methylation.

Folic Acid as a Requirement for Immune Cell Proliferation

Although it is not discussed as frequently as biosynthesis or DNA methylation within the context of folate chemoprevention, one-carbon metabolism is important for the maintenance of the immune system and the inflammatory response. There is considerable evidence that inflammation is a significant risk factor for the development of colorectal cancer [57–59]. Much like cancer cells, immune cells divide rapidly following activation and are therefore highly dependent on an adequate supply of nucleotides [60–62]. Thus, it is possible that lowered folate intake and the subsequent restriction of nucleotide availability will inhibit immune cell proliferation and lead to an attenuated inflammatory response. While inhibition of the immune system comes with a host of potential complications, it may serve as the basis for a useful chemopreventive strategy for particular subpopulations, such as individuals with inflammatory bowel disease (IBD). The practice of disrupting one-carbon metabolism as an anti-inflammatory therapy has been in use for some time; the antifolate methotrexate (MTX) has been administered at low doses to treat rheumatoid arthritis for several decades [63]. Investigations into the anti-inflammatory mechanism of MTX showed that it acts by inhibiting replication of the immune cells responsible for mediating the inflammatory response [64]. Mild to moderate deficiency in folic acid intake, in certain subpopulations, may function through the same mechanism to reduce inflammation and cancer risk.

With the exception of a recent study which demonstrated that mouse monocytes cultured in a folate-depleted medium significantly upregulate expression of inflammatory cytokines, most evidence points towards an immunosuppressive effect of folate deficiency. Depletion of folic acid has been shown to inhibit the proliferation of T cells in vitro [65]; this finding is particularly interesting because ulcerative colitis , a form of IBD which carries a significant increase in CRC risk, is characterized by T cell infiltration of the colonic mucosa [66]. Studies in animal models and in humans also generally support an anti-inflammatory effect of folic acid deficiency. In order to better understand the role of folate metabolism in the process of tumor development, our laboratory [20•] recently restricted the flux of one-carbon metabolism by administering a diet deficient in the methyl donors folate, choline, methionine, and B12 to Apc Min/+ mice. In this study, it was demonstrated that methyl donor deficiency dramatically reduces tumor burden in the small intestine; it was further shown that this reduction in tumor burden was associated with a decrease in immune cell populations and in the expression of inflammatory cytokines. A prior clinical trial in which adult men were placed on a folate-deficient diet for either 4 or 8 weeks showed a similar reduction in inflammatory markers [33]. A study of folic acid intake in the AOM/DSS mouse model of colitis-associated CRC demonstrated that tumor multiplicity was positively correlated with folic acid intake; folate-deficient mice had the lowest number of tumors [67]. Together, these studies demonstrate the potential for a folic acid-based anti-inflammatory intervention, which has important implications for colorectal cancer chemoprevention.

There is considerable evidence to support the use of non-steroidal anti-inflammatory drugs (NSAIDs) for the chemoprevention of colorectal cancer [68]. Interestingly, it has been shown that a large number of NSAIDs commonly considered for use as chemopreventive agents act as competitive inhibitors of folate-dependent enzymes in addition to their primary mechanisms of action [69]. This observation may suggest that these NSAIDs have chemopreventive efficacy due, in part, to the inhibition of folate one-carbon metabolism and presents the possibility of using antifolate drugs for chemoprevention. It may be worthwhile to reevaluate existing antifolates for use as chemopreventive agents, particularly in high-risk populations such as individuals who have recently undergone colorectal tumor resection and are therefore at a high risk of relapse [70]. Drugs such as methotrexate, which is already used at low dosages for long-term management of rheumatoid arthritis, may be the best candidates to be repurposed for use in chemoprevention.

Conclusion

Folic acid may still prove to be a useful chemopreventive agent, but it will likely find the most utility when interventions are tailored for an individual based on his personal medical history and unique genetic background. Due to the multiple mechanisms through which the modulation of folic acid intake may exert its chemoprotective effects, and the differences among at-risk populations, it will be necessary to carefully tailor interventions to an individual’s unique risk profile. If, for example, folate supplementation promotes the progression of established neoplasia [16], it may be prudent for an individual to reduce folate consumption following endoscopic detection of an adenoma. It is especially important to understand the optimal folate intake for specific high-risk subpopulations considering the FDA-mandated national fortification program and the effects it may have on non-target populations. Though this program achieved its goal of reducing the incidence of neural tube defects by raising folate intake in women of childbearing age, it has concurrently raised folate intake nationwide, with men and women over the age of 50 years being the most significantly affected [71, 72]. Collectively, the data suggest that dietary supplementation with folic acid is not protective against CRC in all instances and that successful folate-based chemoprevention will require better understanding of the optimal intake level for particular groups of people as well as carefully targeted interventions.

While nucleotide biosynthesis is the best-characterized folate-dependent molecular mechanism [73], steps still need to be taken to understand how its role in cancer risk changes throughout the carcinogenic process. As a cell acquires more genomic changes and progresses towards cancer, its metabolic demands change. This should be taken into account when considering a biosynthesis-based mechanism of folate chemoprevention. Similarly, DNA methylation profiles are known to be influenced by a number of outside factors including age and the consumption of alcohol [50, 74, 75]. Thus, not only may different subpopulations require unique chemopreventive strategies, but the best strategy for a particular individual may change over time in order to adapt to changes in diet, age, and personal history of neoplasia. Finally, effects of modulated folate intake which are not commonly considered for chemopreventive use may find relevance in specific high-risk subpopulations, such as the potential use of the anti-inflammatory effects of folate deficiency in individuals with IBD. While folate chemoprevention has not turned out to be the “magic bullet” it was once hoped to be, it still has potential to be an effective strategy when interventions are founded in a mechanistic understanding and tailored to an individual’s unique requirements.