Red meat and colon cancer : how dietary heme initiates hyperproliferation

Colorectal cancer is a leading cause of cancer deaths in Western countries. The risk to develop colorectal cancer is associated with the intake of red meat. Red meat contains the porphyrin pigment heme. Heme is an irritant for the colonic wall and it is previously shown that the addition of heme to the diet of rats induces hyperproliferation. Hyperproliferation increases the risk of endogenous mutations, which subsequently increases the risk to develop colon cancer. The aim of this thesis was to elucidate the diet-modulated signaling from an injured surface epithelium to the proliferative stem cells in the crypt to initiate compensatory hyperproliferation. In chapter 2 we showed that when heme is added to the diet of mice, there is an increased cytotoxicity of the colonic contents. Heme-fed mice showed decreased apoptosis and increased compensatory epithelial hyperproliferation resulting in hyperplasia. Gene expression levels of mouse colon after heme feeding were analyzed by microarrays and showed 3,710 differentially expressed genes (q<0.01) of which many were involved in proliferation and stress response. Stainings for the enzyme Heme oxygenase-1 and expression levels of heme- and stress-related genes showed that heme affected the epithelial surface cells, but that heme did not reach the crypt cells. Heme caused injury of the surface epithelial cells, and as proliferation originates from the stem cells in the crypts this implied that there must be a signaling mechanism from the injured surface to the stem cells in the crypts to start the hyperproliferation. In chapter 2 several surface to crypt signaling molecules were identified. Heme downregulated inhibitors of proliferation, such as Wnt inhibitory factor 1, Indian hedgehog and Bone morphogenetic protein 2. Furthermore, heme downregulated the cytokine Interleukin-15. Heme upregulated the expression of the growth factors Amphiregulin, Epiregulin and of Cyclooxygenase-2 mRNA in the surface. However, their protein/metabolite levels were not increased as heme induced surface-specific inhibition of translation by increasing the levels of the translation inhibitor 4E-BP1. We concluded that heme induced colonic hyperproliferation and hyperplasia by downregulating the surface to crypt signaling of feedback inhibitors of proliferation. Besides many proliferation and stress-related genes, many PPARα target genes were upregulated upon heme feeding. As PPARα is proposed to protect against oxidative stress and lipid peroxidation, we hypothesized in chapter 3 that absence of PPARα leads to more colonic surface injury, which subsequently leads to increased compensatory hyperproliferation in colonic crypts upon heme-feeding. This hypothesis was tested using wild-type and PPARα knockout mice receiving a heme diet. Proliferation levels and gene expression profiles were determined. Heme induced luminal cytotoxicity and lipid peroxidation to the same extent in wild-type and PPARα knockout mice. We showed that PPARα does not play a role in the heme-induced hyperproliferation, as heme induced hyperproliferation both in wild-type as well as in PPARα knockout mice. Stainings for alkaline phosphatase activity and expression levels of Vanin-1 and Nrf2-targets indicated a compromised antioxidant defense in the heme-fed PPARα knockout mice. We concluded that PPARα plays a protective role in colon against oxidative stress, but PPARα does not mediate heme-induced hyperproliferation. This implied that oxidative stress of surface cells is not the main determinant of heme-induced hyperproliferation and hyperplasia. Heme was shown to increase both reactive oxygen species as well as cytotoxicity of the colonic contents of mice. So far, the time dependency of the heme-induced oxidative stress and cytotoxic stress on the initiation of hyperproliferation was not studied. Therefore, in chapter 4 the effects of dietary heme on the colonic mucosa after 2, 4, 7 and 14 days of heme feeding were determined. This study showed that the effects of dietary heme on the colonic mucosa can be separated in acute and delayed effects. Acutely, heme increased oxidative stress which caused an increase in lipid peroxidation products. Besides, there was an acute activation of PPARα target genes, most probable induced by the generated oxidized lipids. Nrf2 target genes were activated acutely which played a role in the protection against oxidative stress. Delayed effects which occurred after day 4 of heme feeding, were increased luminal cytotoxicity and the induction of hyperproliferation. This suggested that the cytotoxicity, rather than oxidative stress, induced hyperproliferation. Remarkably, the surface epithelial cells sensed heme after day 4, although heme was present in the colon several hours after ingestion of the heme diet. This suggested that the mucus barrier played a role in the protection of the surface epithelium the first days of heme feeding. As the colon is densely populated by bacteria, the microbiota might play a role in modulating the surface to crypt signaling inducing hyperproliferation. To explore the role of the colonic microbiota we simultaneously investigated the effects of dietary heme on colonic microbiota and on the host mucosa of mice (chapter 5). Using 16S rRNA phylogenetic microarrays, it was determined that heme increased Bacteroidetes and decreased Firmicutes in colonic contents. This shift in the microbiota was most likely caused by a selective susceptibility of Gram-positive bacteria to heme cytotoxic fecal water. This susceptibility was not observed for Gram-negative bacteria and allowed the expansion of the Gram-negative community. The increased amount of Gram-negative bacteria, which likely caused an increased mucosal exposure to lipopolysaccharide (LPS), did not elicit a detectable immune reaction in the host mucosa. The absence of an immune reaction might be influenced by the strong upregulation of Secretory leukocyte peptidase inhibitor (Slpi) at gene and protein level, which is known to suppress excessive immune reactions. We showed that there was no functional change in the sensing of the bacteria by the mucosa, as changes in inflammation pathways and Toll- like receptor signaling were not detected. In conclusion, the change in microbiota did not cause the observed hyperproliferation and hyperplasia via inflammation pathways. In the study described in chapter 6 we investigated whether microbiota play a causal role in the heme-induced hyperproliferation. In this study mice received a control or a heme diet with or without broad spectrum antibiotics (Abx). Similar to previous experiments, heme induced epithelial hyperproliferation. Interestingly, when heme was administered together with Abx there was no induction of hyperproliferation. Heme induced oxidative stress in the heme group as well as in the heme plus Abx group. Cytotoxicity was also induced in both heme groups. As bacteria were decreased by 100 to 1000 fold in abundance upon Abx treatment it is unlikely that bacteria play a major role in the formation of the cytotoxic factor. Whole genome transcriptomics showed that Abx blocked the heme-induced differential expression of oncogenes, tumor suppressors and cell turnover genes. Moreover, Abx blocked the mucosal sensing of luminal cytotoxicity indicating that Abx increased the mucus barrier. Abx eliminated mucin-degrading bacteria, such as Akkermansia, and sulfate-reducing bacteria (SRBs) that produce sulfide. In-vitro studies showed that sulfide is more potent than N-acetylcysteine and cysteine in splitting disulfide bonds, indicating that SRB generated sulfide can denature mucins and thus open the mucus barrier. This study showed that the microbiota plays an important facilitating role in the heme-induced hyperproliferation and hyperplasia by breaking the mucus barrier and thereby decreasing the protection against luminal irritants such as the toxic heme metabolite