Diarrhea or colorectal cancer: Can bacterial toxins serve as a treatment for colon cancer?

Colorectal cancer exhibits a low incidence in under-developed countries even though it is the third most common neoplasm worldwide and the second most common in the U.S. (1, 2). This geographical imbalance suggests an environmental contribution to the resistance of endemic populations to intestinal neoplasia. Although the epidemiology of colon cancer remains poorly understood, there is clearly an unexplained inverse relationship between the incidence of colorectal cancer and enterotoxigenic Escherichia coli (ETEC) infections (3, 4). Drawing from these observations, Pitari et al. (5) introduced an interesting hypothesis that specific peptides (STa) elaborated from ETEC may prevent the hyperproliferative and neoplastic development of intestinal epithelial cells that are associated with initiation and progression of colorectal cancer. Although a direct causal relationship between STa-mediated infectious diarrhea and low cancer rates in under-developed countries has not been proven, the authors provide convincing evidence of the presence of a novel intracellular signaling pathway initiated by STa that prevents proliferation of colon cancer cells.

Mechanistically, STa binds to guanylyl cyclase-C (GC-C) receptors specifically expressed in intestinal cells (6). Ligand binding to GC-C activates the intracellular synthesis of the second messenger, cGMP. STa hyperactivates this signaling receptor, causing large increases of intracellular cGMP ([cGMP]_int) levels (7). Despite this strong negative selective pressure of STa-mediated infection and diarrhea, GC-C and its signaling pathway have been evolutionarily conserved in a wide variety of animal species, suggesting that it must play a role in an important aspect of intestinal physiology. In this way, STa peptides represent molecular mimicry wherein enterotoxigenic bacteria have evolved a strategy for its transmission that exploits normal intestinal physiology (8, 9). Indeed, STa peptides are structurally and functionally homologous to the endogenous peptides guanylin and uroguanylin, which mediate autocrine/paracrine control of intestinal fluid and electrolyte homeostasis (9).

Recently, a new role for GC-C has been implied: a tumor suppressor. In fact, GC-C and its ligands have been implicated in the regulation of the balance of proliferation and differentiation along the crypt-to-villus axis in the intestine (10, 11). Interestingly, expression of guanylin and uroguanylin is lost during colon cancer tumorigenesis (12–14). In support of these studies, targeted inactivation of the mouse guanylin gene results in increased colonic epithelial proliferation (15). As a result, subsequent loss of the initiation of GC-C signaling may represent one key mutational event underlying neoplastic transformation in the colon. However, intestinal GC-C and its downstream intracellular effector molecules are conserved in colorectal tumors (16, 17), thus providing a means of restoring the signaling cascade associated with the tumor suppressor phenotype. In this manner, Forte and coworkers demonstrated that oral administration of uroguanylin suppresses the formation and progression of adenomatous polyps in the Min/+ mouse animal model of colorectal cancer (13, 18).

Mechanisms by which the GC-C ligands repress proliferation and colorectal tumorigenesis are unknown. In a recent issue of PNAS, Pitari et al. (5) demonstrated the presence of a previously unrecognized STa/GC-C-induced cGMP-dependent signaling pathway, through cyclic nucleotide-gated (CNG) channels and calcium, responsible for the antiproliferative action of bacterial enterotoxins on human colon carcinoma cells (5). Similar to results shown from previous studies (10), STa-mediated inhibition of DNA synthesis and cellular proliferation in colon cancer cells (17, 19) was GC-C-dependent because controls using colon cancer cells devoid of GC-C (i.e., SW480 cells; ref. 19) were without effect. Commensurate with the toxin-induced effect was a concomitant increase in [cGMP]_int. In fact, 8-Br-cGMP, a cell-permeable, nonhydrolyzable cGMP analog mimicked the antiproliferative effects of STa in colon cancer cells. In addition, inhibitors of colonic cell-specific cGMP-dependent phosphodiesterases, which catabolize cGMP to its respective biologically inactive noncyclic nucleotide 5′-GMP, potentiate antiproliferation by STa in human colon carcinoma cells, presumably by increasing the accumulation of [cGMP]_int. Pitari et al. (10) further demonstrated that this GC-C mediated inhibition of proliferation was associated with a reduced rate of DNA synthesis. Uroguanylin mimicked the antiproliferative and cytostatic effects of STa in colon cancer cells whereas an inactive-analog of STa was without effect, suggesting the specificity and requirement for the STa/uroguanylin receptor, GC-C.

How does cGMP suppress proliferation of colon cancer cells? To address this question, Pitari et al. examined the antiproliferative effects of intracellular effector molecules of cGMP by STa in T84 cells by using a pharmacologic approach. Selective inhibitors of protein kinase G (PKG) and protein kinase A (PKA), which are previously characterized downstream effectors of cGMP and STa-mediated intestinal secretion (20, 21), did not influence the inhibition of proliferation by STa. Inhibition of phosphodiesterase 3, a target for negative cross-talk with cGMP pathways and an activator of the PKA pathway (22), also had no effect on STa-induced inhibition of DNA synthesis and tumor cell growth. Thus, conventional downstream effectors of cGMP did not mediate the antiproliferative effects of STa on human colon carcinoma cells.

cGMP can also exert its actions through direct activation of cyclic nucleotide-gated (CNG) channels and/or inhibition of Na⁺/Ca²⁺ exchange, leading to alterations in intracellular calcium ([Ca²⁺]_int) (21). In fact, apart from cGMP, STa induces intracellular signaling through other second messengers (23, 24). In this respect, calcium does play a pivotal role. It should be noted that most, if not all, of the studies showing rise of [Ca²⁺]_int by STa have been restricted to the enterocytes of animals (25, 26). On the other hand, reports in T84 colon cancer cells and colonic epithelia are inconsistent. Although the chloride secretory responses in colonic cells have been shown to be influenced by both STa and calcium, a synergistic elevation of [Ca²⁺]_int by STa has yet to be observed (27, 28). These observations may be best explained by Brayden et al. (29) who found that increases in cytosolic free calcium in nonconfluent colon cancer cells activate conductances that differ from those in confluent monolayers. Thus, STa-dependent calcium influx is more than likely dependent on cell cycle and growth. In colon carcinoma cells, STa activation of GC-C mobilizes [Ca²⁺]_int by a cGMP-dependent mechanism (30) and induces Ca²⁺ influx through CNG channels (31). Pitari et al., in turn, used a specific inhibitor, l-DLT, to show the involvement of CNG channels in the antiproliferative properties of STa. Based on an elaborate series of electrophysiological studies and Ca²⁺ transport experiments, the authors clearly established that STa (and cGMP) targets the CNG channels in colon cancer cells which leads to a direct influx of Ca²⁺ into the cells. By using specific channel protein inhibitors, Pitari et al. further found that the STa- and 8-Br-cGMP-induced changes in conductance were due mostly to a Ca²⁺-dependent K⁺ current (K_Ca), indicating that STa-dependent activation of K_Ca reflected cGMP-induced Ca²⁺ influx through specific l-DLT-sensitive CNG channels. The essential role of Ca²⁺ influx is underscored by chelation of free cytosolic Ca²⁺, which reversed the antiproliferative action of STa. Also, depletion of [Ca²⁺]_ext abolished the ability of STa to inhibit cancer cell proliferation, whereas increases in [Ca²⁺]_ext restored the antiproliferative effect of STa, without perturbing [cGMP]_int accumulation. A proposed model describing the intracellular signaling and antiproliferation initiated by STa, including the CNG channel-dependent pathways introduced by Pitari et al. is depicted in Fig. 1.

Figure 1.

Proposed model for the regulation of ion transport by STa and GC-C and mechanism for cell proliferation in the colon. ETEC elaborate STa in the intestinal lumen. Activation of the GC-C receptor by STa (or guanylin and uroguanylin, which are secreted into the lumen by specific intestinal cells) increases cGMP, which in turn, directly activate cGMP-dependent protein kinase II (PKG-II), cAMP-dependent protein kinase (PKA), and/or cGMP-regulated cAMP phosphodiesterase (PDE) (9). PKG-II activation by cGMP stimulates both Cl⁻ and HCO secretion by regulating the CFTR chloride channel, and reduces Na⁺ absorption, presumably by inhibiting the Na⁺/H⁺ exchanger NHE-3. cGMP may activate PKA directly or indirectly (by inhibiting cAMP hydrolysis via PDE, thereby increasing intracellular cAMP levels). PKA, in turn, regulate CFTR activity. Lastly, cGMP can also directly activate the cyclic nucleotide gated (CNG) channels in colon cells, allowing Ca²⁺ to enter the cell. Basolateral entry of Cl⁻, Na⁺, and K⁺ occur, in part, through the Na⁺-K⁺-2Cl⁻ cotransporter and Na⁺/K⁺-ATPase. Mechanisms that lead to cellular proliferation or apoptosis presumably are affected downstream of PKG-II/PKA activation and/or Ca²⁺ entry.

The studies presented by Pitari et al. provide convincing evidence that STa inhibits DNA synthesis in colon carcinoma cells by a signaling mechanism initiated by activation of GC-C, accumulation of [cGMP]_int, and Ca²⁺ influx through CNG channels. The impact of this manuscript is heightened by the fact that it is first to report the regulation of cell proliferation by a cGMP-dependent mechanism mediated by CNG channels and Ca²⁺ influx. However, two questions arise from this study: (i) what is the precise mode of Ca²⁺ delivery to its intracellular site of action to induce intestinal cell cytostasis and (ii) by what mechanism(s) is the STa/GC-C/cGMP-induced Ca²⁺ influx mediating DNA synthesis? In regards to Ca²⁺ delivery, intracellular mobilization, and resultant downstream activation, reports have suggested that inositol triphosphate activation, stimulation of diacylglycerol formation, and translocation and activation of protein kinase C (PKC) may play a role (30, 32–34). In intestinal and other cells types, each has been shown to be involved in proximal signals that lead to mitogenesis and/or apoptosis (35, 36). However, the authors found no evidence of apoptosis by STa (5, 10). In this manner, STa-induced inhibition of proliferation must be tightly regulated (37) because Ca²⁺ can promote apoptosis in colon cancer (38, 39).

In their investigation into the underlying mechanism for the antiproliferative properties of STa and cGMP in colon carcinoma cells, Pitari and coworkers showed that the inhibition of proliferation was independent of apoptosis and necrosis (5, 10). These results are in direct contrast with those reported previously using GC-C agonists (i.e., uroguanylin, STa) and/or specific phosphodiesterase inhibitors in colon cancer cells (13, 40–43). Also, in other cell types, such as vascular smooth muscle cells (VSMC) and hepatocytes, nitric oxide and certain natriuretic peptides have been shown to activate PKG by concurrent increases in [cGMP]_int, and induce apoptosis (44, 45). The mechanisms by which these apoptotic events occur have not yet been resolved, though activation of specific members of the mitogen-activated protein kinase (MAPK) cascades have been suggested (44). Bennett (46) and others (47, 48) have also demonstrated in the same cell type that cGMP and PKG lead directly to the down-regulation of the c-myc protooncogene, resulting in growth arrest and apoptosis. Alternatively, in kidney cells, it was demonstrated through transfection of dominant-negative and constitutively active forms of PKG that this kinase inhibits cell growth by decreasing MAPK activation (49, 50). cGMP and cGMP analogs were strictly dependent on PKG, and PKG activation inhibited the MAPK pathway (i) by inhibiting an upstream activator, c-Raf, and (ii) by inducing MAPK phosphatase-1 expression (one of many enzymes that may inactivate MAPK family members; ref. 50). In reference to the intestine, many reports discuss apoptosis along the rostrocaudal and crypt-villus axes. However, only few researchers have attempted to study the mechanism of action and the intracellular signaling mechanisms (i.e., cGMP, MAPK, etc.) that are associated with programmed cell death in these cells. Soh et al. have recently found that cGMP mediates apoptosis in human colon cancer cells through the activation of PKG and a member of the MAPK cascades (43, 51). Recently, it was found that STa and uroguanylin activate MAPK specific cascades in colon carcinoma cells in a cGMP-dependent fashion and play contributory roles in the peptide-induced apoptosis.† Thus, future investigations into MAPK signaling and the mechanisms underlying GC-C/cGMP-induced cellular proliferation, cell cycle control, and apoptosis in colorectal cancer are warranted (53) and may elucidate conflicting reports such as those presented by Pitari and colleagues (5, 10) and Shailubhai et al. (13).

In conclusion, Pitari et al. present very interesting findings that greatly extend our understanding of the role that GC-C and its ligands play in the control of epithelial tumor cell growth. The signaling cascade suggested by the authors, ETEC → STa → GC-C → cGMP → CNG channel → Ca²⁺ entry → growth inhibition, has not previously been recognized in these cells. This information should be of interest to investigators working in a variety of areas including tumorigenesis, functional roles of CNG channels, factors that control epithelial growth, differentiation, and cell cycle control, the physiological properties of the guanylin/uroguanylin peptide family, and mechanisms of ETEC-induced diarrhea. Lastly, these observations offer a possible mechanistic insight into the resistance to colorectal cancer seen in geographic areas in which ETEC is endemic. However, a causal relationship has not been proven; a person may have to be chronically (or subclinically) infected with toxin-producing ETEC for many years in order for the toxin to have significant impact on colorectal tumor development. Indeed, Forte and coworkers proposed that STa-mediated diarrhea, which occurs repetitively throughout life in people that live in developing countries, is a factor that contributes to the low incidence of colon cancer in third world countries (9, 13). The significance of the antiproliferative pathway demonstrated by Pitari et al. is highlighted by the neoplastic transformation of epithelial cells that follows loss of expression of guanylin and uroguanylin in the colon (13–15). In turn, the conservation of GC-C and its downstream intracellular effector molecules in colorectal tumors (16, 17, 19, 54) provide a potential endogenous therapeutic target for restoration of this signaling cascade and preservation of the tumor suppressor phenotype. Thus, oral administration of specific agents that play a role in intestinal salt and water transport processes, for example, GC-C peptides (i.e., guanylin, uroguanylin, STa), cell-permeable agonists of cGMP (i.e., cGMP analogs, intestinal-specific phosphodiesterase inhibitors), and calcium, offer a novel approach to the prevention and/or therapy of colorectal cancer.