Robert E. Carroll

Phase IIA Clinical Trial of Curcumin for the Prevention of Colorectal Neoplasia

Curcumin is derived from the spice tumeric and has antiinflammatory and antineoplastic effects in vitro and in animal models, including preventing aberrant crypt foci (ACF) and adenomas in murine models of colorectal carcinogenesis. Inhibiting the production of the procarcinogenic eicosanoids prostaglandin E2 (PGE2) and 5-hydroxyeicosatetraenoic acid (5-HETE) can suppress carcinogenesis in rodents. Curcumin reduces mucosal concentrations of PGE2 (via inhibition of cyclooxygenases 1 and 2) and 5-HETE (via inhibition of 5-lipoxygenase) in rats. Although preclinical data support curcumin activity in many sites, the poor bioavailability reported for this agent supports its use in the colorectum. We assessed the effects of oral curcumin (2 g or 4 g per day for 30 days) on PGE2 within ACF (primary endpoint), 5-HETE, ACF number, and proliferation in a nonrandomized, open-label clinical trial in 44 eligible smokers with eight or more ACF on screening colonoscopy. We assessed pre- and posttreatment concentrations of PGE2 and 5-HETE by liquid chromatography tandem mass spectroscopy in ACF and normal-tissue biopsies; ACF number via rectal endoscopy; proliferation by Ki-67 immunohistochemistry; and curcumin concentrations by high-performance liquid chromatography in serum and rectal mucosal samples. Forty-one subjects completed the study. Neither dose of curcumin reduced PGE2 or 5-HETE within ACF or normal mucosa or reduced Ki-67 in normal mucosa. A significant 40% reduction in ACF number occurred with the 4-g dose (P < 0.005), whereas ACF were not reduced in the 2-g group. The ACF reduction in the 4-g group was associated with a significant, five-fold increase in posttreatment plasma curcumin/conjugate levels (versus pretreatment; P = 0.009). Curcumin was well tolerated at both 2 g and 4 g. Our data suggest that curcumin can decrease ACF number, and this is potentially mediated by curcumin conjugates delivered systemically. Cancer Prev Res; 4(3); 354–64. ©2011 AACR.

Aberrant expression of gastrin-releasing peptide and its receptor by well-differentiated colon cancers in humans

Epithelial cells lining the adult human colon do not normally express gastrin-releasing peptide (GRP) or its receptor (GRPR). In contrast, approximately one-third of human colon cancers and cancer cell lines have been shown to express GRP-binding sites. Because GRPR activation causes the proliferation of many cancer cell lines, GRP has been presumed to act as a clinically significant growth factor. Yet GRP has not been shown to be expressed by colon cancers in humans nor has the effect of GRP and/or GRPR coexpression on tumor behavior been investigated. We therefore determined GRP and GRPR expression by immunohistochemistry in 50 randomly selected colon cancers resected between 1980 and 1997, all 37 associated lymph node and liver metastases, and 20 polyps. Tumor sections studied were those that contained the margin and adjacent nonmalignant epithelium. Overall, 84% of cancers aberrantly expressed GRP or GRPR, with 62% expressing both ligand and receptor, whereas expression was not observed in adjacent normal epithelium. Consistent with the previously established mitogenic capabilities of GRP, tissues coexpressing GRP and GRPR were more likely to express proliferating cell nuclear antigen than tissues not expressing both ligand and receptor. Yet GRP/GRPR coexpression was seen with equal frequency in stage A as in stage D cancers and was only detected in 1 in 37 metastases. Furthermore, Kaplan-Meier analysis did not reveal any difference in patient survival between those whose tumors did or did not express GRP/GRPR. In contrast, GRP/GRPR coexpression was found in all well-differentiated tumor regions, whereas poorly differentiated tissues never coexpressed GRP/GRPR. Overall, these data indicate that, although GRP is a mitogen, it is not a clinically significant growth factor in human colon cancers. Rather, the strong association of GRP/GRPR coexpression with tumor differentiation raises the possibility that these proteins primarily act in vivo as morphogens.Epithelial cells lining the adult human colon do not normally express gastrin-releasing peptide (GRP) or its receptor (GRPR). In contrast, approximately one-third of human colon cancers and cancer cell lines have been shown to express GRP-binding sites. Because GRPR activation causes the proliferation of many cancer cell lines, GRP has been presumed to act as a clinically significant growth factor. Yet GRP has not been shown to be expressed by colon cancers in humans nor has the effect of GRP and/or GRPR coexpression on tumor behavior been investigated. We therefore determined GRP and GRPR expression by immunohistochemistry in 50 randomly selected colon cancers resected between 1980 and 1997, all 37 associated lymph node and liver metastases, and 20 polyps. Tumor sections studied were those that contained the margin and adjacent nonmalignant epithelium. Overall, 84% of cancers aberrantly expressed GRP or GRPR, with 62% expressing both ligand and receptor, whereas expression was not observed in adjacent normal epithelium. Consistent with the previously established mitogenic capabilities of GRP, tissues coexpressing GRP and GRPR were more likely to express proliferating cell nuclear antigen than tissues not expressing both ligand and receptor. Yet GRP/GRPR coexpression was seen with equal frequency in stage A as in stage D cancers and was only detected in 1 in 37 metastases. Furthermore, Kaplan-Meier analysis did not reveal any difference in patient survival between those whose tumors did or did not express GRP/GRPR. In contrast, GRP/GRPR coexpression was found in all well-differentiated tumor regions, whereas poorly differentiated tissues never coexpressed GRP/GRPR. Overall, these data indicate that, although GRP is a mitogen, it is not a clinically significant growth factor in human colon cancers. Rather, the strong association of GRP/GRPR coexpression with tumor differentiation raises the possibility that these proteins primarily act in vivo as morphogens.

Expression of Vitamin D Receptor and 25-Hydroxyvitamin D3-1α-Hydroxylase in Normal and Malignant Human Colon

Considerable evidence exists to support the use of vitamin D to prevent and/or treat colorectal cancer. However, the routine use of bioactive vitamin D, 1,25-dihydroxyvitamin D3, is limited by the side effect of toxic hypercalcemia. Recent studies, however, suggest that colonic epithelial cells express 25-hydroxyvitamin D3-1α-hydroxylase, an enzyme that converts nontoxic pro-vitamin D, 25-hydroxycholecalciferol [25(OH)D3], to its bioactive form. Yet, nothing is known as to the cellular expression of 1α-hydroxylase and the vitamin D receptor (VDR) in the earliest histopathologic structures associated with malignant transformation such as aberrant crypt foci (ACF) and polyps [addressing the possibility of using nontoxic 25(OH)D3 for chemoprevention], nor is anything known as to the expression of these proteins in colorectal cancer as a function of tumor cell differentiation or metastasis [relevant to using 25(OH)D3 for chemotherapy]. In this study, we show that 1α-hydroxylase is present at equal high levels in normal colonic epithelium as in ACFs, polyps, and colorectal cancer irrespective of tumor cell differentiation. In contrast, VDR levels were low in normal colonic epithelial cells; were increased in ACFs, polyps, and well-differentiated tumor cells; and then declined as a function of tumor cell de-differentiation. Both 1α-hydroxylase and VDR levels were negligible in tumor cells metastasizing to regional lymph nodes. Overall, these data support using 25(OH)D3 for colorectal cancer chemoprevention but suggest that pro-vitamin D is less likely to be useful for colorectal cancer chemotherapy.

The case for gastrin-releasing peptide acting as a morphogen when it and its receptor are aberrantly expressed in cancer

Gastrin-releasing peptide (GRP) and its receptor (GRP-R) are frequently expressed by cancers of the gastrointestinal tract, breast, lung, and prostate. Most studies have found that GRP and its amphibian homologue bombesin act to increase tumor cell proliferation, leading to the hypothesis that this peptide hormone is a mitogen important for the growth of various cancers. Yet GRP/GRP-R co-expression in cancer promotes the development of a well-differentiated phenotype; while multiple studies suggest that the presence of these 2 proteins confer a survival advantage. Along with recent reports showing that GRP and its receptor critically regulate aspects of colon and lung organogenesis, we argue that these proteins do not function primarily as mitogens when aberrantly expressed in cancer. Rather, we postulate that GRP/GRP-R are onco-fetal antigens that function as morphogens, with their effect on tumor cell proliferation being a component property of their ability to regulate differentiation. Thus aberrant GRP/GRP-R expression in cancer recapitulates, albeit in a dysfunctional manner, their normal role in development.

Azoxymethane-induced fulminant hepatic failure in C57BL/6J mice: characterization of a new animal model

Without transplantation, ∼50-90% of all patients with fulminant hepatic failure (FHF) die. This poor outcome is due in part to the absence of an appropriate animal model, which would allow for a greater understanding of the pathophysiology of this syndrome. Given the reports of liver injury in humans and livestock fed cycad palm nuts on the island of Guam, we hypothesized that the active ingredient azoxymethane (AOM) could cause FHF. We therefore evaluated AOM in C57BL/6J mice. Histologically, we observed microvesicular steatosis 2 h, sinusoidal dilatation 4 h, and centrilobular necrosis 20 h after AOM administration, and transmission electron microscopy showed that this agent causes mitochondrial injury. FHF was associated with all four stages of encephalopathy, as well as by a prodromal period of decreased eating and drinking lasting ∼15 h before the development of stage I encephalopathy (i.e., loss of scatter reflex). Late encephalopathy was associated with increased arterial ammonia, decreased serum glucose, and evidence of brain edema (astrocyte swelling). We show that AOM-induced FHF is highly reproducible, without evidence of lot-to-lot variability, and is dose dependent. These findings therefore suggest that AOM is an excellent agent for the study of FHF, as well as indicate that Guamanian FHF may be due to AOM found in unwashed cycad palm nuts.Without transplantation, approximately 50-90% of all patients with fulminant hepatic failure (FHF) die. This poor outcome is due in part to the absence of an appropriate animal model, which would allow for a greater understanding of the pathophysiology of this syndrome. Given the reports of liver injury in humans and livestock fed cycad palm nuts on the island of Guam, we hypothesized that the active ingredient azoxymethane (AOM) could cause FHF. We therefore evaluated AOM in C57BL/6J mice. Histologically, we observed microvesicular steatosis 2 h, sinusoidal dilatation 4 h, and centrilobular necrosis 20 h after AOM administration, and transmission electron microscopy showed that this agent causes mitochondrial injury. FHF was associated with all four stages of encephalopathy, as well as by a prodromal period of decreased eating and drinking lasting approximately 15 h before the development of stage I encephalopathy (i.e., loss of scatter reflex). Late encephalopathy was associated with increased arterial ammonia, decreased serum glucose, and evidence of brain edema (astrocyte swelling). We show that AOM-induced FHF is highly reproducible, without evidence of lot-to-lot variability, and is dose dependent. These findings therefore suggest that AOM is an excellent agent for the study of FHF, as well as indicate that Guamanian FHF may be due to AOM found in unwashed cycad palm nuts.

Tumor suppressor functions for the Cdk inhibitor p21 in the mouse colon

The Cdk inhibitor p21 regulates p53-mediated growth arrest following DNA damage. It is expressed during epithelial differentiation in a variety of organs including colon. We investigated susceptibility of p21-deficient mice to the colon carcinogen azoxymethane (AOM). After AOM injections, rodents develop putative premalignant lesions called aberrant crypt foci (ACF) that are localized to the distal three centimeters of the colon. p21-deficient mice developed significantly higher numbers of ACF than wild-type mice in response to AOM, and these were not restricted to distal colon. After AOM treatment, increased numbers of lymphoid aggregates were detected in p21-deficient colon. Proliferation was similar in wild type and p21-deficient colon before and after AOM injection, but AOM-induced apoptosis was detected only in wild-type crypt epithelial cells, and not in the p21-deficient colon. The proapoptotic function uncovered for p21 was unexpected, because p21 acts as an inhibitor of apoptosis in many systems, and is not required for p53-dependent apoptosis. Enhanced formation of ACF in p21-deficient mice supports a tumor suppressor function for p21 in the colon. Reduced apoptosis of colon epithelial cells with deleterious mutations may be an initiating event in the formation of ACF, with inflammatory cell cytokine expression contributing to their further expansion.

LOCATION AND CHARACTERIZATION OF THE HUMAN GRP RECEPTOR EXPRESSED BY GASTROINTESTINAL EPITHELIAL CELLS

The exact location of normal gastrin-releasing peptide (GRP) receptor expression by epithelial cells lining the human gastrointestinal (GI) tract is not known; yet this receptor is found on upwards of 50% of GI cancers. Furthermore, the pharmacology reported for GRP receptors expressed by GI cancers varies considerably. Therefore, the purpose of this study was to determine the normal distribution of GRP receptor expression by cells lining the human GI tract, and then determine the normal pharmacology of the human receptor when ectopically expressed by the nonmalignant human colon epithelial cell line NCM460. We obtained endoscopic pinch biopsies of, and extracted the RNA from, epithelial cells lining the esophagus, stomach, jejunum, ileum, and proximal and descending colon, RT-PCR demonstrated that GRP-R expression is limited to cells lining the gastric antrum, indicating that this receptor is aberrantly expressed by GI cancers. To determine the normal pharmacology of this receptor when expressed by nonmalignant human tissues for the first time, we transfected NCM460 cells with the cDNA for the human GRP receptor. By studying three stable NCM460 cell lines expressing varying numbers of receptors, we demonstrate that agonist and antagonist binding affinity, binding kinetics, and G-protein coupling are all independent of receptor number. Finally, by comparing GRP receptors expressed by GI cancers with those on NCM460-transfected cells, we show that the pharmacology of the aberrantly expressed receptors is significantly altered. Thus, these data demonstrate that GI cancers aberrantly express GRP receptors that then behave abnormally.

Allelic Loss of the Gene for the GPX1 Selenium-Containing Protein Is a Common Event in Cancer

Selenium has been shown to reduce cancer incidence in animal models and more recent data indicate that it may be protective in humans as well. However, little is known about the mechanism by which selenium prevents cancer. Cytosolic glutathione peroxidase (GPX1), a selenium-containing antioxidant enzyme, has been implicated in the development of cancer of the head and neck, lung, and breast, in part because of allelic loss at the GPX1 locus. The study of allelic loss at the GPX1 locus in colon cancer was investigated by examining loss of heterozygosity (LOH) in DNA extracted from both tumor and adjacent histopathologically normal tissue obtained by laser capture microdissection. Tissue samples were obtained from 53 colon cancer patients. Two highly polymorphic markers, alanine codon repeats and a proline-leucine polymorphism (198P/L) present in the GPX1 gene, were used to examine LOH at this locus. Analysis of both polymorphisms identified LOH at GPX1 in a significant percentage of colorectal cancer (42%). These results indicated that LOH at the GPX1 locus is a common event in cancer development and that GPX1 or other tightly linked genes may be involved in the etiology of this disease.

Colon cancer chemoprevention by a novel NO chimera that shows anti-inflammatory and antiproliferative activity in vitro and in vivo

Chemopreventive agents in colorectal cancer possess either antiproliferative or anti-inflammatory actions. Nonsteroidal anti-inflammatory drugs (NSAID) and cyclooxygenase-2 inhibitors have shown promise, but are compromised by side effects. Nitric oxide donor NSAIDs are organic nitrates conjugated via a labile linker to an NSAID, originally designed for use in pain relief, that have shown efficacy in colorectal cancer chemoprevention. The NO chimera, GT-094, is a novel nitrate containing an NSAID and disulfide pharmacophores, a lead compound for the design of agents specifically for colorectal cancer. GT-094 is the first nitrate reported to reduce aberrant crypt foci (by 45%) when administered after carcinogen in the standard azoxymethane rat model of colorectal cancer. Analysis of proximal and distal colon tissue from 8- and 28-week rat/azoxymethane studies showed that GT-094 treatment reduced colon crypt proliferation by 30% to 69%, reduced inducible NO synthase (iNOS) levels by 33% to 67%, reduced poly(ADP-ribose)polymerase-1 expression and cleavage 2- to 4-fold, and elevated levels of p27 in the distal colon 3-fold. Studies in cancer cell cultures recapitulated actions of GT-094: antiproliferative activity and transient G2-M phase cell cycle block were measured in Caco-2 cells; apoptotic activity was examined but not observed; anti-inflammatory activity was seen in the inhibition of up-regulation of iNOS and endogenous NO production in lipopolysaccharide (LPS)-induced RAW 264.7 cells. In summary, antiproliferative, anti-inflammatory, and cytoprotective activity observed in vivo and in vitro support GT-094 as a lead compound for the design of NO chimeras for colorectal cancer chemoprevention. [Mol Cancer Ther 2007;6(8):2230–9]

Constitutive activation of the gastrin-releasing peptide receptor expressed by the nonmalignant human colon epithelial cell line NCM460.

Gastrin-releasing peptide (GRP) causes multiple effects in humans by activating a specific heptaspanning receptor. Within the gastrointestinal tract, GRP receptors (GRP-R) are not normally expressed by mucosal epithelial cells except for those lining the gastric antrum. In contrast, recent studies have shown that up to 40% of resected colon cancers aberrantly express this receptor. This is important because the GRP-R can cause the proliferation of many, but not all, tissues in which it is expressed. Since GRP and other agonists are not known to exist in the colonic lumen, it has not been clear how or even if GRP-R expression in colon cancer contributes to cell proliferation. To evaluate the functional consequence of GRP-R expression on colonic epithelium, we transfected the recently isolated nonmalignant human colon epithelial cell line NCM460 with the cDNA for this receptor. All NCM460 cell lines expressing varying numbers of GRP-R bound selected agonists and antagonists indistinguishably from receptors expressed by other human tissues. Furthermore GRP-R-expressing transfected cell lines, but not wild-type NCM460 cells, proliferated independently of serum or other growth factors. Further evaluation revealed that GRP-R in these cells tonically stimulated G alpha q/11, resulting in increased phospholipase C activation. Since transfected cells do not secrete GRP, nor is their growth influenced by exposure to receptor-specific antagonists, these data indicate that GRP-R ectopically expressed by NCM460 cells are constitutively active. This report provides the first evidence of mutation-independent heptaspanning receptor constitutive activation resulting in cell proliferation, and identifies a potential mechanism whereby the GRP-R may act as an oncogene in human colon cancer.