Epidemiological studies support the involvement of short-chain fatty acids (SCFA) in colon physiology and the protective role of butyrate on colon carcinogenesis. Among the possible mechanisms by which butyrate may exert its anti-carcinogenicity an antioxidant activity has been recently suggested. We investigated the effects of butyrate and mixtures of SCFA (butyrate, propionate and acetate) on DNA damage induced by H2O2 in isolated human colonocytes and in two human colon tumour cell lines (HT29 and HT29 19A). Human colonocytes were isolated from endoscopically obtained samples and the DNA damage was assessed by the comet assay. H2O2 induced DNA damage in normal colonocytes in a dose-dependent manner which was statistically significant at concentrations over 10 μM. At 15 μM H2O2 DNA damage in HT29 and HT29 19A cells was significantly lower than that observed in normal colonocytes (P < 0.01). Pre-incubation of the cells with physiological concentrations of butyrate (6.25 and 12.5 mM) reduced H2O2 (15 μM) induced damage by 33 and 51% in human colonocytes, 45 and 75% in HT29 and 30 and 80% in HT29 19A, respectively. Treatment of cells with a mixture of 25 mM acetate + 10.4 mM propionate + 6.25 mM butyrate did not induce DNA damage, while a mixture of 50 mM acetate + 20.8 mM propionate + 12.5 mM butyrate was weakly genotoxic only towards normal colonocytes. However, both mixtures were able to reduce the H2O2-induced DNA damage by about 50% in all cell types. The reported protective effect of butyrate might be important in pathogenetic mechanisms mediated by reactive oxygen species, and aids understanding of the apparent protection toward colorectal cancer exerted by dietary fibres, which enhance the butyrate bioavailability in the colonic mucosa.
Several epidemiological and experimental studies have been devoted to the definition of the effect of dietary factors on the pathogenesis and incidence of colorectal cancer (1). There is general agreement, also based on meta-analysis data (2), about the protective effect of dietary fibre which, by increasing stool volume, reduces both faecal concentrations of carcinogens and intestinal transit time (3). In addition, by-products of the bacterial fermentation of dietary fibre, the short-chain fatty acids (SCFA), play an important role in colon physiology and a central role in colonocyte metabolism (4–6). Indeed, these compounds reduce the faecal pH and consequently decrease the activity of 7α-dehydroxylase, a bacterial enzyme responsible for the conversion of bile acids from primary to secondary (7), the latter are considered to be cancer promoters. In particular, among the SCFA, sodium butyrate is an important energy source for colonocytes, a potent controller of cell growth and differentiation and an inducer of apoptosis (8). Butyrate is thought to act at the gene level by increasing the accessibility of DNA to a variety of transcription factors (9–11); however, the molecular mechanisms by which butyrate prevents colon cancer are still unclear. Recently, Abrahamse et al. (12) showed that butyrate is able to reduce DNA damage induced by hydrogen peroxide in freshly isolated rat colon cells, suggesting that this compound may exert its anticarcinogenic effect through an antioxidant activity.
The involvement of reactive oxygen species (ROS) in some degenerative diseases, including cancer, is now widely accepted (13,14). In fact, although ROS production is involved in many physiological processes, including defence against pathogenic microorganisms, cell proliferation control and apoptosis induction (15,16), an overproduction or a reduction in the removal of ROS from cells, a condition referred to as `oxidative stress', may cause damage through mutations which may lead to cancer (17,18). Some studies, however, have evidenced a `pro-oxidant activity' of butyrate. In colorectal cancer cell lines butyrate increases peroxide production (19). In K562 cells, for instance, the differentiation induced by butyrate is mediated by ROS (20) and, moreover, the butyrate trigger activity in apoptosis is mediated by ROS overproduction (21). It should be noted that all these effects involving ROS overproduction are directed against undifferentiated tumour cell lines; therefore, it is not in conflict with the protective role of butyrate.
In the present study, we have focused our attention on the protective role of butyrate against oxidative DNA damage induced in human colonocytes, since it has been reported that rat colonocytes differ in their susceptibility to various genotoxic compounds with respect to human colonocytes (22,23). Freshly isolated human colonocytes and two human colon tumour cell lines (HT29 and HT29 19A) have been studied to evaluate the possible different response of normal and tumour cells to oxidative stress and the role of butyrate as a protective agent. Furthermore, the effect of a physiological mixture of SCFAs in this system has been evaluated.
Materials and methods
Multiple biopsies were taken from normal sigmoid mucosa of 26 subjects of both sexes undergoing routine colonoscopic examinations for screening purposes. After careful explanation of the aims of the investigation, each subject gave informed consent. From a series of preliminary experiments, in which various bioptic samples were tested to determine colonocyte yield, it was established that at least four biopsies were necessary to obtain a sufficient number of cells.
Colonocytes were prepared as previously described (24). Briefly, biopsy samples from each donor were pooled, minced and digested with a protease/collagenase mixture (6 mg protease + 3 mg collagenase in 3 ml Hank's balanced salt solution) for 35 min at 37°C. The digestion was stopped by adding Dulbecco's modified Eagle medium (DMEM), supplemented with 10% fetal calf serum (FCS), 2 mM l-glutamine, 1 mM sodium pyruvate, 1 mM non-essential amino acids, 100 U/ml penicillin and 100 μg/ml strepto- mycin. The resulting cells were washed (1400 r.p.m. for 7 min) and suspended in the same medium. Cell numbers and viability were determined by trypan blue exclusion and the concentration was adjusted to 0.5×106 cells/ml for subsequent experiments. On average, 2×106 cells (85% viability) were recovered from each sample. Fibroblasts eventually present were removed by adherence for 30 min at 37°C in a 6-well plastic plate.
Cell type characterization
A mixture of monoclonal antibodies (mAbs)—phycoerythrin-conjugated anti-CD3, fluorescein (FITC)-conjugated anti-CD56 and anti-My4—were employed to detect the percentage of contaminating lymphocytes/monocytes. For staining, 3×105 cells were suspended in 50 μl phosphate-buffered saline (PBS; 10 mM sodium phosphate, pH 7.4, containing 120 mM NaCl and 2.7 mM KCl), incubated at 4°C for 30 min, washed and analysed by flow cytometry (FACScan; Becton Dickinson, Mountain View, CA). Contaminant lymphocytes/monocytes were removed according to a previously described method (25), by negative immunomagnetic bead selection using a mixture of the mAbs anti-CD3 and anti-My4 coated with magnetic beads (Immunotech, Marseille, France). For each mAb, 100 μl washed beads were added to 1×106 cells suspended in 1 ml PBS + 30% FCS, gently mixed and incubated for 10 min at room temperature. After 5 min, the suspension was mixed again to allow the immunobeads to bind to the antigenic target and form rosettes with cells. The tube tests were then placed in contact with the magnet for 10 min and the negative cells were carefully collected from the supernatant.
The cells obtained after these treatments were cytocentrifuged, dried and stained with Giemsa to morphologically assess the percentage of colonocytes.
Cancer cell lines
Human carcinoma cell lines (HT29 and HT29 19A) were seeded at 3×105/ml and grown for 48 h in DMEM. The cells were then detached by a trypsin–EDTA solution (Gibco BRL, Paisley, UK), washed, counted and adjusted to 0.5×106/ml with DMEM and used for the subsequent experiments.
Cell treatment and comet assay
Freshly isolated colonocytes and human colon tumour cells (HT29 and HT29 19A) were incubated with the different substances (H2O2, sodium butyrate, sodium proprionate and sodium acetate; Sigma–Aldrich, Milan, Italy) in DMEM at 37°C and 5% CO2 under various experimental conditions (see Results). After each treatment, cell viability was determined by trypan blue exclusion and DNA damage by the comet assay.
The comet assay or `single cell gel electrophoresis assay' (SCGE) was carried out as previously described (26). Cell suspensions (200 μl, 0.1×105 cells) were transferred into 1.5 ml Eppendorf tubes and centrifuged at 1200 r.p.m. for 5 min. The supernatant was discarded and the pellet was mixed with 75 μl 0.7% low melting agarose (Gibco-BRL) in PBS and distributed onto conventional microscopic slides pre-coated with 0.5% normal melting agarose (Gibco-BRL) in PBS and dried at 50°C. After the agarose was solidified (4°C for 10 min), a second layer of low melting agarose was applied. Successively, the slides were immersed for 1 h at 4°C in lysis solution (2.5 M NaCl, 100 mM Na2EDTA, 10 mM Tris–HCl pH 10, containing freshly added 1% Triton X100 and 10% DMSO) and then placed into a horizontal electrophoresis apparatus filled with freshly made solution (1 mM Na2EDTA, 300 mM NaOH). Following 20 min of pre-incubation (unwinding of DNA), electrophoresis was run for 20 min at a fixed voltage of 25 V and 300 mA, adjusted by raising or lowering the electrophoresis buffer in the tank. Then the slides were washed three times with neutralization buffer (0.4 M Tris–HCl pH 7.5), stained with 50 μl 20 μg/ml ethidium bromide and kept in a moisture chamber in the dark at 4°C until analysis. All the above steps were carried out in a red light environment to prevent any additional DNA damage.
Cells were analysed 24 h after staining at 400× magnification, using a fluorescence microscope (Axiolab; Carl Zeiss, Jena, Germany) equipped with a 50 W mercury lamp. Microscopic images revealed circular shapes (undamaged DNA) and `comet' like shapes, in which the DNA had migrated from the head to form a tail (damaged DNA). For each slide, 100 comets were scored by a computerized image analysis system (Comet Assay II; Perceptive Instruments, Haverhill, UK) able to calculate the `tail moment', the parameter considered to be more directly related to DNA damage (27). The tail moment is defined as the product of DNA content in the tail and the mean distance of migration in the tail. DNA damage was expressed as arbitrary units (AU), obtained by dividing the comets into five classes, each class representing an increase in the extent of DNA damage according to the tail moment values: class 0, tail moment <1; class 1, tail moment 1–5; class 2, tail moment 5–10; class 3, tail moment 10–20; class 4, tail moment >20, and then by multiplying the number of cells belonging to each class by the class number. Thus, the total score of 100 comets could range from 0 (all cells undamaged) to 400 (all cells heavily damaged) (28). The percentage of protection exerted by butyrate or SCFA mixtures on hydrogen peroxide-induced DNA damage was evaluated as above
The two-tailed Student's paired t-test was adopted for statistical analysis. Values of P < 0.05 were chosen for rejection of the null hypothesis. Data are expressed as means ± SEM.
Characterization of cell types isolated from biopsies
Fibroblasts were consistently absent from each sample. Staining with mAbs directed toward lymphocyte and monocyte antigens showed that the percentage of contaminating cells was on average less than 6%. These results were also confirmed by direct observation of the Giemsa-stained cytospin. Therefore, we were able to carry out experiments in a setting yielding 95% or more of freshly isolated colonocytes.
DNA damage exerted by H2O2 on normal colonocytes
First, the susceptibility of human colonocytes to treatment with H2O2 was assessed. Freshly isolated cells were incubated with increasing concentrations of H2O2 in DMEM for 15 min at 37°C and 5% CO2, followed by the measurement of cell viability and the evaluation of DNA damage by the trypan blue and comet assays, respectively. After treatment, cell viability always remained over 85% while DNA damage showed a dose-dependent effect that became statistically significant at H2O2 concentrations over 10 μM (Figure 1A). An increase in the concentration of H2O2 to 80 μM resulted in nearly all cells being highly damaged (data not shown). An increase in the exposure time of cells to 15 μM H2O2 showed no further increase in DNA damage. Instead, genotoxic activity was reduced with incubation times greater than 30 min (Figure 1B).
The response of colonocytes isolated from different subjects to H2O2 was, as expected, somewhat variable, in that we were dealing with fresh biological samples and not with cell clones. Figure 2 shows such a variability in basal values and after 15 μM H2O2 treatment in the 26 subjects we studied.
Protective effect of butyrate and SCFA mixtures on H2O2-induced DNA damage in normal colonocytes
For these experiments, devoted to the investigation of the protective effect of butyrate on DNA damage, the oxidative stress was induced by exposing cells to 15 μM H2O2 for 15 min. Colonocytes were first pre-incubated with butyrate (6.25 or 12.5 mM) for 15 min, washed with DMEM and then treated with H2O2. The basal level of DNA damage did not change in cells exposed to butyrate alone at both concentrations, while H2O2 exposure induced significant damage (Figure 3). In cells pre-incubated with butyrate, there was a statistically significant reduction in DNA damage (33 and 51% at butyrate concentrations of 6.25 and 12.5 mM, respectively). A prolonging of the pre-incubation time with butyrate of up to 2 h showed no further increase in the protective activity at both concentrations (data not shown). Further experiments were performed using two different mixtures of SCFA, whose composition was derived from the concentrations of butyrate used previously (6.25 and 12.5 mM) and from the relative amounts of SCFAs found in the lumen of the human colon (acetate/propionate/butyrate at a relative concentration of 60:25:15) (29). The results of these experiments, reported in Figure 4, show that incubating normal colonocytes with a mixture of 25 mM acetate + 10.4 mM propionate + 6.25 mM butyrate (mixture 1) did not induce DNA damage, while a mixture of 50 mM acetate + 20.8 mM propionate + 12.5 mM butyrate (mixture 2) was slightly but significantly genotoxic (P = 0.048). However, both mixtures were able to protect cells from H2O2-induced DNA damage in a similar manner (47%).
Protective effect of butyrate and SCFA mixtures on H2O2-induced DNA damage in colon tumour cell lines
Two human colon tumour cell lines (HT29 and HT29 19A), which differ in their level of differentiation, were employed. HT29 contains almost no differentiated cells, whereas HT29 19A is a clone cell differentiated by butyrate and possesses several characteristics typical of normal colonocytes. As shown in Figure 5, the basal DNA damage was significantly lower in HT29 cells (16 ± 4.5 AU) than in both HT29 19A cells (33 ± 7 AU, P < 0.05) and in normal colonocytes (47 ± 3.8 AU, P < 0.001; Figure 3). H2O2-induced DNA damage in HT29 (56 ± 12 AU) and HT29 19A (63 ± 11 AU) was significantly lower than that observed in normal colonocytes (131 ± 5.7 AU, P < 0.01). Butyrate alone was not genotoxic towards tumour cell lines at both concentrations; pre-treatment with the highest butyrate concentration reduced H2O2-induced DNA damage in both cell lines (Figure 5). Exposure of HT29 and HT29 19A cells to SCFA mixtures did not induce genotoxic effects (Figure 6). Pre-incubation of tumour cell lines with the two SCFA mixtures showed a marked protective effect against H2O2 oxidative DNA damage; the percentages of protection were 58 and 52% in HT29 cells and 47 and 55% in HT29 19A cells treated with mixture 1 and 2, respectively (values similar to that registered in normal colonocytes, 47%).
Most of the literature on this topic refers to experimental animal models, and those human studies available have not usually specified the total colonocytes yield; therefore, there is a possibility that other contaminant cells (in particular, fibroblasts and immunologically competent ones) might `contaminate' the bulk of the obtained cells. The isolation method we employed allowed us to obtain a high percentage of isolated single cells composed of an almost homogeneous colonocyte population (94%).
Furthermore, employing a less time consuming (45 min) and less stringent method in this study for the isolation of colonic cells avoided the spontaneous apoptosis of colonic epithelial cells, due to the loss of anchorage to the basal membrane and loss of cell to cell contact (30,31). This phenomenon is strictly dependent on the single cell isolation method and on the time employed for detachment of the cells. In fact, Grossmann et al. (30) demonstrated that an appreciable apoptotic (18%) index occurs after 2 h from the cell isolation, and more recently it has been shown that the sub-G1 cell population, indicative of apoptosis, became evident after 1 h (31). Moreover, cell activation, as shown by an increase in both urokinase production and interleukin-8 secretion, may also be avoided by using a less stringent isolation method (32).
Several points are worth discussing further. First, treatment of human colonocytes with a concentration of H2O2 as low as 15 μM for 15 min caused evident DNA damage whereas cell viability was not affected. Comparison of these results with those reported by Abrahamse et al. (12) in rat colonocytes, in which DNA damage was observed at concentrations of H2O2 higher than 100 μM (12), suggests that human cells are much more sensitive to oxidative stress than those of the rat. This conclusion is supported by previous studies in which the effect of 200 μM H2O2 was compared in rat and human colonocytes (33). In addition, human cells respond differently from rat cells when exposed to other genotoxic agents such as N-methyl-N-nitro-N-nitrosoguanidine (34).
Secondly, H2O2 treatment showed that DNA damage values were lower in tumour cell lines than in normal colonocytes. These results suggest that, when cells are adapted to in vitro culture, they react more promptly against oxidative stress caused by H2O2 (via the involvement of DNA repair systems) than the normal colonocytes. Differences among cell types and cell growth status in response to the oxidative damage could be ascribed to the involvement of various detoxifying activities and/or of the DNA repair systems. Several studies suggest that these systems may represent the most effective defense against oxidative stress by H2O2 in actively proliferating cells, but not in differentiated or undividing cells (28,35,36). Even though the mechanisms by which ROS damage cells have not been fully elucidated, there are several studies on their ability to cause DNA strand breakage (12,28,37). In particular, it has been proposed that H2O2 DNA damage takes place via the iron-induced Fenton reaction that leads to the reduction of H2O2 to hydroxyl radicals, which react with DNA and cause strand breaks (36).
The butyrate and SCFA mixtures showed antioxidant properties, protecting both the freshly isolated colonocytes and the colon tumour cell lines from H2O2-induced DNA damage. It should be stressed that butyrate and SCFA mixtures were used at concentrations and relative ratios thought to be physiological in the colon lumen (29). Incubation of colonocytes and tumour cell lines with butyrate at both concentrations tested, and with the two SCFA mixtures, did not induce any change in the basal level of DNA damage, with the only exception, which cannot easily be explained, of a statistically significant (P = 0.048) increase registered in colonocytes when incubated with SCFA mixture 2. That the same protective effect was observed as a result of treatment with either butyrate alone or the SCFA mixtures suggests that the presence of propionate and acetate does not interfere with the butyrate effect. This finding agrees with the general opinion that, among SCFA, butyrate is the most active in biological effects, such as apoptosis and inhibition of proliferation in tumour cells (38) and in being utilized as a primary energy source in normal colonocytes (39). Any possible relationship between these effects and DNA damage protective activity remains to be elucidated.
The results on the protective effect of the SCFA mixture obtained in this study differed from those obtained by Abrahamse et al. (12), that showed no protection in rat colonocytes. However, the data are not directly comparable because the SCFA mixtures used by these authors employed a buty rate concentration of 6.25 mM and relative ratios of the SCFA (75:15:10 and 41:21:38) different from those used in the present investigation.
The mechanism by which butyrate reduces H2O2-induced DNA damage in colon cells is not known. However, the simple scavenger activity of butyrate against oxygen free radicals is probably not the main mechanism, since this compound was removed from the medium during the treatment of cells with H2O2. In addition, the chemical structure of butyrate makes the simple action as scavenger improbable. In this content, it is important to consider that DNA damage represents a steady-state between the initiation of DNA damage and its repair by cellular processes (40). Thus, DNA damage depends on protective endogenous factors, such as DNA repair and levels of antioxidant systems, which could be modified by the treatment with butyrate. The effects of butyrate on the chromatin structure and on the DNA repair mechanisms are known (41). Butyrate is able to increase both chromatin accessibility to DNA repair enzymes in HT29 cells (42) and DNA excision repair in human fibroblasts (43). These activities could result in a reduction of DNA damage, as shown in the present investigation. Alternatively, butyrate could affect the intracellular antioxidant enzymes (i.e. catalase and glutathione peroxidase) responsible for the reduction of H2O2 (44). According to previous studies in which treatment with 5 mM butyrate increases catalase activity in artery smooth muscle cells (45), a similar effect could take place in colonocytes, thus resulting in an enhanced disproportionation of H2O2 and a reduced level of damage. However, it is likely that butyrate may act at many different levels (46) and most effects seem to result from direct action on intracellular targets (47).
In conclusion, this study on freshly isolated human colonocytes shows that butyrate (alone or in SCFAs mixtures) displays a protective effect on oxidative DNA damage induced by H2O2. This effect might be of importance in pathogenetic mechanisms mediated by ROS, and it may help in a better understanding of the apparent protection toward colorectal cancer exerted by ingestion of less fermentable dietary fibre, as shown by epidemiological studies (48–51).
We are warmly indebted to Professor G.Rechkemmer for the kind gift of cell lines HT29 and HT29 19A.