The side population (SP) phenotype, defined as the reserpine-blockable ability to efflux the nucleic acid dye Hoechst 33342, has been claimed to be enriched for stem cells in several human normal tissues, cancers and cell lines, and thus may be useful for the identification and isolation of cancer stem cells. We demonstrated the presence of SP fractions in all of seven tested gastrointestinal cancer cell lines. Four cell lines were selected (HT29, HGT101, Caco2 and HRA19a1.1) for detailed phenotypic and behavioural analysis with respect to stem cell characteristics. Cell surface marker analysis showed that, contrary to non-SP cells, the SPs entirely lack the expression of CD34. This difference, however, disappeared when the cells were cultured, rendering both populations CD34-positive. Expression of other putative stem cell markers (CD133, CD44, Hes-1, beta-catenin, Musashi-1, Oct-4 and CD117) was identical on SP and non-SPs before and after culturing. Sorted SP and non-SP cells were similarly clonogenic in vitro, tumourigenic in vivo, and displayed similar multipotential differentiation potential in vitro and in vivo. Additionally, culturing cytometrically-sorted SP and non-SP cells showed that the populations are interconvertible, each giving rise to the other. Expression of ABCG2 and Mdr-1, two membrane transporter proteins that have been suggested to be responsible for the drug-effluxing capacities of SP cells, including Hoechst 33342, was identical in non-SP and SP cells, indicating that there may be additional factors responsible for the Hoechst effluxing property in gastrointestinal cancer SP cells. Here, we show that the SP and non-SP fractions, albeit phenotypically distinct populations, do not differ with respect to stem cell-like cell number or behaviour. We thus conclude that the concept of the SP phenotype as a universal marker for stem cells does not apply to gastrointestinal cancer cells. These findings stand in contrast to the observations made in many other tissues and harbour important implications for the future search for intestinal cancer stem cell markers.

The authors have previously reported the derivation of colonic subepithelial myofibroblasts (SEMFs) in both humans and mice from bone marrow (BM). In the pathogenesis of inflammatory bowel disease (IBD), such as Crohn's disease and ulcerative colitis, colonic SEMFs mediate several types of inflammatory response. In the present study, interleukin (IL)-10-/- mice were used as a model of IBD to investigate the involvement of BM-derived cells in the inflamed mucosa. Male whole BM [either C57/BL10 (wild type: WT) or IL-10-/- donor mice] was used to perform bone marrow transplantation (BMT) into both WT and IL-10-/- female mice. Tissue samples were evaluated by immunohistochemistry for alpha-smooth muscle actin expression and by in situ hybridization using a Y-chromosome-specific probe to track the donor-derived colonic SEMFs. The mucosal expression of mRNA for pro-inflammatory cytokines was analysed by reverse transcriptase-polymerase chain reaction (RT-PCR). In addition, mRNA expression of matrix metalloproteinase (MMP)-7 and osteopontin in the inflamed mucosa was assessed using in situ hybridization. Body weights and histological scores showed that IL-10-/- mice that received WT BM had an improved course of colitis, decreased mucosal pro-inflammatory mRNA expression, and up to 30% of their SEMFs were of BM origin. Conversely, IL-10-/- mice receiving IL-10-/- BM progressed to extensive colitis, and Y probe analysis revealed that up to 45% of colonic SEMFs were of BM origin. WT mice receiving IL-10-/- or WT BM had no signs of colonic inflammation. The expression of MMP-7 and osteopontin was up-regulated in the inflamed mucosa. In conclusion, IL-10-/- mice displayed ameliorated disease activity after WT BMT, whilst colitis was not induced in WT mice by IL-10-/- BMT. The contribution of BM-derived cells to colonic SEMFs was significantly increased in the inflamed mucosa compared with non-inflamed mucosa.

Trefoil factor family (TFF) peptides have many in vivo and in vitro effects on restitution, wound healing, apoptosis, cell motility, adhesion and vectorial ion pumping, amongst others. (125)I-TFF peptides bind to cell membranes with classical saturable ability. It would be surprising if there were not TFF-protein interactions that would explain these actions, but to date no convincing TFF-binding partner has been shown which unambiguously takes part in any of these functions. Nevertheless, several TFF-binding proteins exist, including the small intestinal CRP-ductin (muclin), which binds TFF2, and the recently described gastric foveolar proteins TFIZ1 (TFF1-binding) and blottin (TFF2-binding), any of which may yet interact in novel ways to elicit TFF-mediated events. This review describes the expression and, where known, the functions of such proteins.