J Slavin, Department of Food Science and Nutrition, University of Minnesota, 225 FScN, 1334 Eckles Avenue, Saint Paul, MN 55108, USA. E-mail: email@example.com. Phone: +1-612-624-7234. Fax: +1-612-625-5272.
Prebiotics may prevent colorectal cancer (CRC) development in humans by modifying the composition or activity of the colorectal microflora. Epidemiologic and animal studies have shown a reduction in CRC or CRC biomarkers after the administration of prebiotics. Studies using indirect chemical biomarkers of CRC in humans, however, gave mixed results. Recently, human studies measuring direct physical indices of CRC risk after prebiotic consumption have been published. The purpose of this review is to summarize those studies to provide recommendations for the use of prebiotics in CRC risk reduction. A PubMed search was conducted, revealing nine studies. One tested lactulose, two evaluated a blend of oligofructose and inulin, and six measured resistant starch. Lactulose reduced adenoma recurrence, while resistant starch had no effect on adenoma or CRC development. Crypt mitotic location, gene expression, and DNA methylation were somewhat improved after resistant starch consumption. No changes in cell proliferation and apoptosis, crypt morphology, or aberrant crypt foci were found. More human studies measuring physical changes to the gut are needed.
Colorectal (CRC) cancer was the third most common cancer in the United States in 2010. In addition, the mortality rate of CRC was the second highest of any cancer.1 While cancer treatments have made large strides in recent decades, chemoprevention by diet and other lifestyle factors offers a more desirable alternative. Various dietary interventions have been examined for their effectiveness in preventing cancer, but there is still much work to be done.2 Diet-cancer interactions are made even more complex in the colon and rectum by the microflora, which has unique effects on the colorectal environment.
The human gut contains hundreds of bacterial species, of which the genera Bacteroides and Bifidobacterium comprise approximately 50%. The relationship between the composition of the intestinal microflora and CRC in humans is just beginning to be explored. The microflora may contribute to either prevention or promotion of colorectal carcinogenesis by way of mutagenic degradation products, colonocyte proliferation, stool frequency, inflammation and immunity, toxification of procarcinogens by bacterial enzymes, or prevention of pathogenic organism overgrowth.3,4 The differences in microflora composition between subjects with colon or rectal cancer or adenomas and subjects with normal colonoscopies have recently been investigated in cross-sectional studies, which showed conflicting results.5–7 The abundance of Bacteroides spp. was significantly greater in cases versus controls in one study5 and significantly lower in another,6 while a third found different effects from various Bacteroides spp.7 The third study also found a reduction in butyrate-producing bacteria and an increase in opportunistic pathogens in CRC subjects.7 In other studies, Bifidobacterium and Lactobacillus, thought to be beneficial bacterial genera, were found to have anticancer properties,8–10 but research on their role in preventing human CRC when given as probiotic supplements is limited.11 Research on dysbiosis and the properties of intestinal microflora could lead to interventions to prevent the development of adenomas or cancer, or could lead to the development of microbial biomarkers for CRC, but the conflicting results indicate that more investigation of the relationship between intestinal microflora composition and CRC is needed.
Some species have a preference for carbohydrate or protein substrates. As such, these dietary components stimulate certain bacterial species and may compound or attenuate the risk for CRC.12,13 Carbohydrate-consuming bacteria, such as Bifidobacterium and Lactobacillus, may provide protection against CRC,14 while protein-consuming bacteria may increase CRC risk through the production of toxic protein breakdown products.12 In general, only carbohydrates that are not broken down by the enzymes in the human small intestine reach the colon for use by the microflora. Those that stimulate the microflora in a way that is beneficial to the host are called prebiotics.
The most recent definition states that a prebiotic is “a selectively fermented ingredient that allows specific changes, both in the composition and/or activity in the gastrointestinal microflora that confers benefits upon host well-being and health.”15 To date, all known and suspected prebiotics are carbohydrate compounds, primarily oligosaccharides, known to resist digestion in the human small intestine and reach the colon, where they are fermented by the gut microflora. The degree to which their fermentation in the large intestine results in beneficial changes to the host has been difficult to assess. Studies have provided evidence that inulin and oligofructose (OF), lactulose, and resistant starch (RS) meet all aspects of the definition, including the stimulation of Bifidobacterium, a beneficial bacterial genus. Other compounds, such as galactooligosaccharides, transgalactooligosaccharides, stachyose, and raffinose, may also have prebiotic effects, but additional work is needed.15
Inulin and OF are found in a variety of foods, including chicory, onions, leeks, garlic, asparagus, bananas, and artichokes. They are fructans composed of a linear fructose chain with or without a glucose subunit at one end. Inulin has a chain length of 2–60 fructose or glucose subunits, while OF is the enzymatic hydrolysis product of inulin, with a chain length of 2–20 subunits. The bonds that connect the fructose subunits cannot be cleaved by the enzymes in the human small intestine.16,17 As a result, inulin and OF reach the large intestine, where they are known to have a bifidogenic effect.13,18 Lactulose is a nondigestible chemically manufactured disaccharide with a galactose subunit attached by a β(1–4) bond to fructose.19 Its consumption in human subjects has led to increased Bifidobacterium concentrations.20–22 Resistant starch is the component of food or chemically altered starches that is not digested and absorbed in the small intestine of humans. There are four RS types, with different properties that lead to their nondigestibility. RS1 is physically unavailable due to limited cell wall degradation of the grain or seed. RS2 is unavailable in its raw form due to starch granules that are resistant to hydrolysis. RS3 becomes unavailable after cooking and storage due to recrystallization of the starch chains. Finally, RS4 is chemically modified to resist hydrolysis by intestinal enzymes.23 All four RS types have been shown to increase Bifidobacterium in humans and animal models.24–26
Additionally, simultaneous doses of probiotics and prebiotics, known as synbiotics, may have benefits beyond probiotics or prebiotics alone. Probiotics and synbiotics may be more difficult to incorporate into the diet and may be more expensive than prebiotics27; however, they carry less potential for adverse gastrointestinal symptoms, such as flatulence and diarrhea, which may result from prebiotics due to their ready fermentability.28
Though epidemiologic studies of CRC risk typically do not evaluate prebiotic consumption, one study calculated daily RS intake as 5% of the daily starch intake and found it had an inverse association with CRC incidence.29 CRC risk has also been reduced after prebiotic treatment in animal models, as evidenced by a reduction in aberrant crypt foci (ACF) and other markers of CRC risk30,31 There are a number of studies in human subjects measuring indirect biomarkers of CRC risk after prebiotic treatment. Some evidence suggests that prebiotics contribute to the reduction of bacterial enzymes, protein degradation products, and secondary bile acids.21,32–34 Additionally, a synbiotic treatment was found to reduce genotoxicity and cell proliferation in the colon.35 Other studies, however, suggest that prebiotics have little to no effect on CRC biomarkers.36–39
A large number of potential biomarkers for CRC in humans have been proposed, but not all have been validated. Precursors to CRC, such as ACF and adenomas, are indicative of CRC.40 Proteins associated with cell proliferation and apoptosis, certain genes, and epigenetic factors like DNA methylation have been associated with CRC risk.41–46 Genotoxic or cytotoxic compounds in fecal water, such as bile acids, N-nitroso compounds, and heterocyclic amines, have been proposed as biomarkers for CRC. Additionally, polyphenols and short-chain fatty acids (SCFAs), e.g., butyrate, in fecal water may be appropriate biomarkers because of their chemoprotective properties.47 Fecal water compounds have been proposed as biomarkers because they can be collected in a noninvasive fashion, but data on the strength of the relationship between fecal water biomarkers and the development of CRC in vivo in humans is limited.48 Therefore, this review will focus on the development of colorectal carcinoma, precursors to CRC, colonic mucosa biomarkers, gene expression, and epigenetic factors.
A larger literature base exists for dietary fiber and the risk of CRC than for prebiotics and the risk of CRC. Systematic reviews of dietary fiber in human and animal studies have reported mixed results49–51; however, the data on dietary fiber and CRC is of limited value to the study of prebiotics because even though inulin, OF, and RS are typically considered dietary fibers, the majority of fibers are not thought to have a prebiotic effect. Among studies of prebiotics and CRC risk, most have been in animal models or have measured fecal water biomarkers of CRC risk. Recently, a number of studies have tested CRC development, the development and recurrence of CRC precursors, colonic mucosa biomarkers, gene expression, and epigenetic measures of CRC risk in humans in response to prebiotics, but no systematic review has been conducted to summarize these articles. Therefore, a systematic review was performed to determine the effect of prebiotics on measures of CRC risk in adult humans. This review will include established prebiotic types as well as those that have only preliminary, but promising, research. It will not include in vitro studies, animal studies, or studies with fecal water biomarkers as endpoints.
LITERATURE SEARCH METHODS
A systematic review was conducted using an adaptation of the American Dietetic Association (ADA) evidence analysis process.52 A research question was developed using the PICO (patient/problem, intervention, comparison, outcome) method53 to assess the effect of prebiotics on CRC in adult humans. Two PubMed searches were carried out in February 2011 to find primary research articles in which prebiotics were given as an intervention, and CRC or biomarkers for CRC were measured. The first search used keywords related to prebiotics and CRC. When this search failed to retrieve an adequate number of articles meeting the criteria, a second search was performed using medical subject headings (MeSH) terms related to prebiotics and CRC. A full list of the terms is given in the legend to Figure 1. Additional articles were identified from the reference lists of articles uncovered in the PubMed search. Randomized and nonrandomized controlled and crossover trials were included for review. Studies using dietary fiber or other low- or non-digestible carbohydrates with unknown prebiotic effect were excluded, as were studies measuring indirect biomarkers of colorectal carcinogenesis, such as butyrate concentrations, bacterial enzyme activity, fecal water cytotoxicity, fecal bile acids, urinary protein degradation products, and the presence of other compounds thought to be related to CRC risk.
The study design, execution, and reporting of data were evaluated for each study using the ADA Quality Criteria Checklist.52 The Quality Criteria Checklist includes questions on subject inclusion and exclusion, potential sources of bias, generalizability of the results, and data collection and analysis. Studies that addressed all or almost all of the applicable quality criteria questions were given a positive rating. Those that did not address two or three key questions but were otherwise positive were given a neutral rating. A negative rating was reserved for studies that did not address the majority of questions on the Quality Criteria Checklist. Quality ratings, as well as prebiotic type, dose, study duration, and CRC biomarker, were taken into account when drafting conclusions.
A summary of the search strategy and results is given in Figure 1. Nine articles were identified for inclusion in the review. Three prebiotic treatments, as well as a number of CRC markers, were represented in the articles. A summary of the study characteristics and the results of each of the nine studies is given in Table 1.
Table 1. Human intervention studies measuring the effects of prebiotics on colorectal cancer risk.
Abbreviations and symbols: ACF, aberrant crypt foci; CRC, colorectal cancer; GI, gastrointestinal; OF, oligofructose; RS2, RS3, resistant starch types 2 and 3; +, positive quality rating (all or almost all quality questions addressed); 0, neutral quality rating (two or three key questions not addressed).
No effect on cell proliferation or crypt height, minimal effect on DNA methylation
Inulin and oligofructose
Two controlled trials tested the effects of inulin and OF on markers of CRC risk. Langlands et al.54 conducted a 2-week trial in 29 adults with no history of gastrointestinal disease who underwent colonoscopy. The subjects were split into experimental and control groups, but information on randomization or blinding was not provided. A supplement of 7.5 g of OF plus 7.5 g of inulin was consumed daily. The control group did not receive a supplement. Cell proliferation was measured by differences in MCM2, Ki67, and proliferating cell nuclear antigen levels between the experimental and control groups; however, no differences were found. Due to the limited information on randomization, blinding, and the lack of a placebo, this study was given a neutral rating. The second inulin and OF study (Limburg et al.55) was a randomized controlled trial in which 85 adults ≤40 years of age with a history of CRC resection and at least five ACF at the time of the study were given 12 g/day OF-enriched inulin (experimental group) or 12 g/day maltodextrin (control group). Percent changes in ACF, proliferation (via Ki67), and apoptosis (via caspase-3) were measured. No change was seen in any of the measurements.
One study, by Roncucci et al.,56 tested the effect of 20 g/day lactulose for up to 3 years on adenoma recurrence in 209 adults who had previously had polyps removed. The study found a significant reduction in adenoma recurrence after lactulose supplementation, but limited information on subject characteristics was given, and there was no placebo for the control group. Therefore, this study was given a neutral rating.
Five studies measured the effect of RS consumption on CRC risk. Burn et al.57 found no effect on neoplasia occurrence or severity from a 30 g/day 1:1 mix of RS2 and RS3 in 727 adults with Lynch syndrome (a hereditary nonpolyposis colon cancer). Dronamraju et al.58 also used a 30 g/day 1:1 RS2 and RS3 mixture and found no effect on cell proliferation (Ki67), crypt number, or crypt size. However, RS produced a trend toward a reduction in percent mitotic cells in the top half of the crypt and an increase in the expression, in tumor tissue, of genes CDK4 and GADD45A, which are associated with reduced cell proliferation and genomic stability, respectively. Grubben et al.59 used a large (45 g/day) dose of RS2 but found no effect on cell proliferation. Van Gorkom et al.60 administered 30 g/day RS2 but found no changes in cell proliferation (BrdU assay) or crypt length. The largest RS dose was provided by Wacker et al.,61 who administered 50–60 g/day RS2 but found no effect on cell proliferation (BrdU assay). This study was given a neutral quality rating because it was not randomized and it lacked generalizability, as it tested a small sample with a narrow age range. Finally, in a crossover trial by Worthley et al.,62 treatment with 25 g/day RS2 produced no effect on cell proliferation (Ki67) or crypt height but resulted in an increase in MINT2, one of 17 gene markers in which DNA methylation was measured.
The purpose of this review was to compile human studies to find the effect of prebiotics on CRC occurrence or risk factors. Three of the nine studies showed a significant reduction in any marker of CRC risk. Only one study demonstrated a significant decrease in a direct CRC endpoint: Roncucci et al.56 reported a reduction in adenoma recurrence. Overall, the studies revealed a weak association between prebiotic consumption and CRC risk. The individual biomarkers will be discussed here, along with potential reasons for the differences in outcome between the studies.
Adenoma occurrence or recurrence, CRC development, and aberrant crypt foci
Three studies measured changes in the development of ACF, adenomas, and CRC. One,56 in which 20 g/day lactulose was given for up to 3 years, reported significantly fewer recurrences of adenoma in postpolypectomy subjects consuming lactulose than in those with no treatment. In contrast, a 6-month study55 of 12 g/day OF-enriched inulin found no effect on change in ACF in adults with a history of CRC, and a 2-year study57 of 30 g/day RS found no effect on the development of adenoma or CRC, or the percentage of advanced cases, in subjects with Lynch syndrome. The study with the longest treatment period was the only study with significant results, but given that all three studies were conducted with different types of prebiotics and included subjects with different characteristics, it is not possible to draw conclusions about the value of a longer treatment duration.
There was no effect of prebiotics on cell proliferation in any of the seven studies in which it was measured. However, when crypt mitotic location was measured by Dronamraju et al.,58 there was a trend toward a reduction in the percentage of cells undergoing mitosis in the top half of the crypt after RS supplementation compared with corn starch (control) treatment. Normal cell proliferation occurs at the base of the crypt, so the change that occurred with RS indicates a possible protective effect. Increased cell proliferation has been found in patients with established adenomatous polyps and CRC,63 but whether it occurs in response to short-term changes without the development of a polyp or tumor is unknown. Because cell proliferation occurs in response to cell death, which can occur from cytotoxic or genotoxic compounds.,64 perhaps the treatment duration of the seven studies in which cell proliferation was measured (2 weeks to 6 months) did not allow sufficient changes in the colorectal environment to prevent or induce cell death.
Crypt morphology, gene expression, and DNA methylation
No changes were seen in crypt number or size in three studies58,60,62 ranging from 2 weeks to 2 months. As such, this review revealed no differences in colorectal location, given that crypt morphology was measured in various regions of the colon and rectum. Expression of CDK4 and GADD45A genes in tumor tissue was significantly increased after 2–4 weeks of RS treatment.58 Because subjects in this study already had symptoms or a diagnosis of CRC, these findings may only describe the effect of RS among cancer survivors, rather than the effect on primary cancer prevention. Worthley et al.62 reported a minor increase in the methylation of MINT2, one of 17 methylation markers measured. Since MINT2 represents such a small proportion of the total assay, they reported that the increase was likely due to chance.
Differences in prebiotic types could not be assessed here due to the low number of studies for each type. Prebiotic dose did not appear to have an effect on the results, given that the highest doses administered resulted in no effect on CRC risk. The effects of treatment duration and sample size were unclear, given that, of the two studies with the longest treatment periods and largest sample sizes, one found a significant effect of lactulose on adenoma recurrence,56 and one found no effect of RS on neoplasia.57
Hypotheses regarding changes in the microflora and CRC risk seem well founded, yet little effect was observed in these nine studies. Perhaps the beneficial effect of prebiotics and nondigestible starches in epidemiologic and animal studies is due to nutrient replacement. Diets high in fiber and fermentable carbohydrates are low in fat, making it difficult to determine whether differences in disease incidence or health status are linked to high-fiber diets or low-fat diets. Animal studies of experimentally induced colon cancer have evaluated the interaction between dietary fiber and fat intake.65 Different lipids (corn oil or fish oil) were fed to rats. The authors found a significant interaction between fat, fiber, and p21 expression, with one combination being protective (fish oil) and the other promotive (corn oil) of colon carcinogenesis, supporting the need to control fat intake in fiber intervention studies.
Fermentation of fiber or prebiotics yields SCFAs in the colon that are absorbed as an energy source. SCFAs produced in the colon from fermentation of nondigestible carbohydrates by the resident microbiota contribute about 10% of daily energy requirements in humans.66 One SCFA, butyrate, is thought to be particularly protective against colon cancer. Not all studies support a relationship between butyrate and colon cancer, a finding termed the “butyrate paradox.” The chemopreventive benefits of butyrate depend on the amount administered, the time of exposure with respect to the tumorigenic process, and the type of fat in the diet. Butyrate targeted miRNA-dependent p21 gene expression in human colon cancer cells, supporting that butyrate regulates host gene expression involved in intestinal homeostasis as well as carcinogenesis through modulation of miRNAs.67 The role of microbial-generated products, including butyrate, in disease prevention is an exciting field, but difficult to study.
In controlled feeding studies, the fat content of the diet can be held constant and a prebiotic just added to usual intake. Thus, protein and fat are controlled and prebiotics supplement the normal diet, making it possible to determine the role of prebiotics alone. Only one study61 in this review controlled the diet of the subjects, but no effect of RS on cell proliferation was seen.
Six of the nine studies were given a high quality rating according to the ADA Quality Criteria Checklist, which strengthens the results. The measurement of neoplasia occurrence, cell proliferation, gene expression, and DNA methylation are likely more indicative of CRC risk than secondary markers such as toxic compounds in the colon or rectum, protein degradation, and bacterial enzyme activity because the secondary markers require more intermediate steps before CRC develops. However, the links between ACF and adenoma occurrence, cell proliferation, gene expression, DNA methylation, and CRC have not been fully established as risk factors, and until then, only the measurement of CRC occurrence itself can fully represent the effect of treatments on CRC.45 Even then, the complexity of the relationship between diet and the colorectal environment and the length of time needed for the development of CRC requires more control over the diets of subjects to prevent influencing factors, such as protein or fat consumption, than would a short-term study using a secondary biomarker. In addition, CRC susceptibility may vary greatly among patients, and larger sample sizes may be needed for adequate study power. Future studies should also examine changes in microflora composition to verify the sufficiency of the delivery method and the dose of the prebiotic. Additionally, individual biomarkers may not represent the entire gut environment, and it may be better to use a combination of biomarkers to evaluate CRC risk. Along the same lines, a metagenomic approach may be preferable to a study of changes to individual bacterial genera because prebiotics, probiotics, and other potential chemopreventers or chemopromoters may have wide-reaching effects on the gut microbiome. Finally, epigenetics, such as DNA methylation and chromatin remodeling, is an interesting new approach that has the potential to identify biomarkers for CRC risk and therefore would be a valuable addition to future studies.
Little effect was seen in CRC risk reduction as a result of prebiotic administration in humans. Only one study found a significant change in a direct CRC endpoint, but that study was given a neutral quality rating. Thus, the evidence is weak that prebiotic-supplemented diets, in comparison with diets without prebiotics, lower CRC risk in adult humans. Nevertheless, more human studies with improved study designs are needed to build upon the evidence obtained to date.
Funding. No funding was received for this project.
Declaration of interest. The authors have no relevant interests to declare.