This study demonstrates that treatment with 2.5 g/kg, 5 g/kg and 10 g/kg of GBR (weight of GBR (g)/weight of rat (kg)) significantly suppressed the total number of ACF, AC and multicrypt of ACF per colon, yet was insignificant among the GBR-treated groups. It is postulated that the chemopreventive effects of GBR was dose-independent. The ACF was found to consist of 1, 2, 3, 4 or more crypts. It is believed that the number of crypts increases with time due to crypt multiplication or branching . The crypts reproduce themselves by a fission mechanism . However, the number of ACFs correlates poorly with cancer risk. It is the crypt multiplicity that is more predictive of malignant transformation . It has been hypothesized that many ACFs regress and only larger foci or lesions progress to cancer . Therefore, GBR seems to be a potential agent to inhibit the progression of colon cancer since the number of larger foci (4 or more crypts) was found to be significantly less in all groups treated with it compared to the control group (Figure 1).
All tissue samples from G5, which represent the normal colon tissue did not show β-catenin immunoreactivity (Figure 2). There was no localization of stain at cell membrane, cytoplasm and nucleus of epithelial cells and goblet cells. This may indicate that the level of β-catenin in the normal colon mucosa is too low to be detected by this method of immunohistochemistry (IHC). In contrast, previous studies by Hong  reported that all the normal adjacent tissue of human colorectal sample showed β-catenin immunoreactivity in the plasma membrane at cell to cell border, and cytoplasm of the epithelial cells and goblet cells, but no nuclear staining was found. Takahashi et al.  also showed localization of β-catenin at membranes of cell-to-cell borders in normal colon epithelial cells of male F344 rats. The divergence in our staining result may be caused by inter-species and inter-strain variation between the Sprague Dawley rats and the other samples (human and other breeds of rats).
In the tissues of positive control group (G1), where colon carcinogenesis was initiated by injection of AOM but no treatment with GBR was given, intense β-catenin stain (immunoreactivity in the cytoplasm of epithelial cells is prominent) was observed, representing much higher level of β-catenin (Figure 2). The intensity of the brown stain is variable, where some cells exhibit greater intensity at the cell membrane compared to the cytoplasm, while other cells lost or have reduced brown stain at the cell membrane, thus appearing that the cytoplasm compartment has more localization of β-catenin. Overall scoring showed that the epithelial cells of G1, which was shown to have the highest number of aberrant crypt foci (ACF) among all groups (Table 1), has the highest score in the evaluation of immunohistochemical staining of β-catenin compared to other groups.
According to Goss and Groden , β-catenin binds to the cytoplasmic tail of α-catenin and E-cadherin and indirectly to the cytoskeleton. It may dissociate from them and enters cytoplasm as free unbound β-catenin. Since β-catenin is an important Wnt signaling pathway activator, its level in the normal tissue is constantly regulated by APC, axin and glycogen synthase kinase-3β (GSK-3β) complex that targets its rapid degradation in the cytoplasm to prevent translocation of β-catenin into the nucleus to promote transcription of various target genes .
Mutations in β -catenin, APC, axin and GSK-3β may result in the inability of the complex to phosphorylate β-catenin, resulting in failure of ubiquitination; whereas E-cadherin mutation can cause dissociation of β-catenin from the complex, releasing free β-catenin into the cytoplasm. This explains loss or reduced β-catenin protein at the cell-to-cell borders but an over-expression in the cytoplasm of the G1 tissues.
In the treatment groups (G2, G3, and G4), heterogenous expression of β-catenin was observed in the cell membrane and cytoplasm compartment. Nevertheless, all the three treatment groups showed significant decrease in β-catenin expression compared to G1. This indicates that GBR has the potential of reducing the expression of β-catenin. It is postulated that GBR is a natural chemopreventive agent which acts through pathway involving β-catenin.
Tissue samples from G5, which represent the normal colon tissue showed some degree of COX-2 immunoreactivity, where the expression of COX-2 was found to be weak, diffused and heterogenous (Figure 3). Lack of expression of COX-2 in tissue samples from G5 indicates that most normal colonic cells were not undergoing inflammation. However there was still a score of 3 as to indicate that inflammation does exist in normal condition but at a very slow rate. COX-2 is known to be an inducible enzyme that plays an important role in induction of pain and inflammation . Localization of COX-2 could be seen mainly at the cytoplasm of colon epithelial cells, and some cells also showed clear nuclear membrane stain. No nuclear staining was observed. Similarly, Tomozawa et al.  and Takahashi et al.  have also reported COX-2 expression in normal rat colon mucosa using IHC method.
In the tissues of G1, the colon epithelia showed much greater COX-2 immunoreactivity compared to all other groups (Figure 3). The staining was also heterogenous throughout the tissues, mostly in the cytoplasm. No nuclear staining was found. Many studies reported well-documented observation of COX-2 expression in colonic tumour. For example, Tomozawa et al.  and Takahashi et al.  have reported positive immunoreactivity in cytoplasm and nuclear membranes of tumour cells. The staining pattern in the treatment groups is similar, but with apparent reduction in both intensity and percentage of positive cell, compared to G1. This shows that GBR prevents colon carcinogenesis by down-regulating the COX-2 expression.
As for the comparison among the treatment groups, there were significant differences (p = 0.0001) between G2 and G3, and between G2 and G4. This indicates that the expression of COX-2 decreased with increase in the concentration of GBR. APC, β-catenin, axin, and GSK-3β mutations can cause increased expression of β-catenin in the cytoplasm and nucleus of colorectal carcinoma, which leads to Wnt pathway activation. Since COX-2 may be a downstream target of Wnt signaling pathway, COX-2 gene is upregulated and COX-2 expression is increased in tumour cells [37, 38].
Spearman rank correlation test showed a statistically positive linear relationship between β-catenin and COX-2 scores. This is in accordance with findings by Kawasaki et al.  that showed positive correlation of COX-2 over expression with cytoplasmic β-catenin expression. Thus, indicating a possible link between COX-2 and the Wnt signaling pathway . COX-2 has been shown to activate β-catenin through prostaglandin E2 and G protein-coupled receptor EP2 (E prostanoid 2) . A study by Lee and Jeong has shown that β-catenin stabilizes COX-2 mRNA by interacting with AU-rich elements in a 3' untranslated region . These findings suggest the existence of a positive feedback loop between β-catenin and COX-2. Therefore, GBR might have suppressed β-catenin expression through regulation of COX-2 or vice versa. It is also possible that GBR regulates expression of both proteins simultaneously.
This study indicates that GBR possesses good chemopreventive effects, more likely due to the nutritional components that are increasing in content following the process of germination such as phytic acid (IP6), ferulic acid, inositol and dietary fiber . It has been reported that inositol hexaphosphate (IP6) reduced the carcinogen-induced large bowel cancer and inhibited growth of transplanted tumors . However, the mechanism by which IP6 exerts chemopreventive is not completely understood. In general, IP6 is rapidly absorbed by cells and metabolized to lower phosphates and inositol . Both IP6 and its lower phosphates have metal chelating activity and may interfere with tumor formation by suppressing metal catalyzed oxidation of fats. It regulates the cell cycle to block uncontrolled cell division and force malignant cells either to differentiate or to undergo apoptosis . Alternatively, IP6 may block the activity of key enzyme(s) affecting cell proliferation. One enzyme candidate is PI-3kinase which plays a central role in signal transduction and cell transformation triggered by growth factor or tumor promoter. IP6 has been reported to inhibit PI-3 kinase activity in vitro . In addition, the antioxidant properties of IP6 may also contribute to its antineoplastic activity . Ferulic acid inhibits the growth of colonic ACF and suppresses the progression of preneoplastic to malignant neoplasia. The suppression of metabolic activation and enhancement of detoxification may become one of the mechanisms for chemopreventive action of ferulic acid .
As functional food component, Shamsuddin  has reported that inositol plays a role in suppressing tumor formation. Meanwhile, Alabaster  has found that the dietary fiber is linked to the prevention of colon carcinogenesis. A high fiber intake result in a high stool bulk, which reduces stool transmit time and thus lowers exposure to potential carcinogens .
This findings show that germinated brown rice has chemopreventive properties in rats induced with colon cancer. The administration of 2.5 g/kg, 5 g/kg and 10 g/kg of GBR (weight of GBR (g)/weight of rat (kg)) significantly reduced the total number of ACF, AC, and multicrypt of ACF, and the expression of β-catenin and COX-2 compared to the positive control (without GBR). It is suggested that GBR can be a promising dietary intake for prevention of human colon cancer.