|Home | About | Journals | Submit | Contact Us | Français|
Notch downstream targets are fundamental to intestinal cell lineage commitment and are suggested as therapeutic targets for colon cancer cells. However, the role of endogenous Notch signaling through receptor-ligand interaction, and effects of its longer-term down-regulation on intestinal homeostasis, are unclear.
To address these issues, the gene encoding protein O-fucosyltransferase 1, an enzyme required for the Notch ligand binding and thus activation of all Notch receptors, was deleted in the mouse intestinal and colonic epithelium, through Villin-Cre-mediated recombination.
Pofut1 deletion inactivated Notch signaling giving rise to smaller but viable mice. These mice exhibited a large increase in all intestinal secretory cell lineages which accumulated in the crypts, resulting in crypt hyperplasia. Although proliferating cells were largely reduced in the colon, the transit amplifying compartment was maintained in the upper crypts of the intestinal mucosa. By 9 months, these perturbations in cell maturation altered mucus-associated gut microbiota and caused chronic intestinal inflammation, with evidence for bacterial translocation to the mesenteric lymph nodes, macrophage and T lymphocyte infiltration, and Th1/Th17 immune response. Dysplastic foci were also observed in Pofut1 deficient small intestine with occasional progression to tumor formation.
Mucus hypersecretion upon Pofut1 inactivation is accompanied by alteration of the mucus-associated flora, which likely contributes to the development of entero-colitis. Therefore, these data identify important potential complications in strategies to target Notch signaling in therapeutic approaches to colon cancer.
Notch signaling is a highly evolutionarily conserved network that orchestrates cell fate determination, involving regulation of proliferation, migration, differentiation and cell death in organisms ranging from insects to humans. In mammals, Notch signaling is mediated by four Notch genes (Notch1 through Notch4) and five cell-membrane-associated ligands (Jag1, Jag2, Dll1, Dll3 and Dll4). Upon ligand binding on neighboring cells, Notch receptors undergo γ-secretase mediated proteolytic cleavage, which releases the Notch intracellular domain (NICD) from the membrane 1. NICD subsequently translocates to the nucleus and forms a complex with the DNA-binding factor RBP-Jκ which upregulates the transcription of target genes2, including members of the hairy and enhancer of split (HES) and hairy related transcription factors (HRT). In particular, NICD stimulates expression of HES1 3, which represses the activity of other bHLH transcription factors, including MATH1 (or human homolog Hath1) 4, 5.
Gain-of-function approaches have shown that intestinal crypt progenitors cease to mature towards secretory cell types and continue to proliferate upon constitutive Notch1 activation 6, 7. Consistent with this, conditional inactivation of RBP-Jκ signaling drives cellular differentiation towards goblet and Paneth cell lineages in the mouse intestine 8. As Notch signaling appeared to be a key, albeit redundant, pathway in crypt homeostasis 5, 6, 8–11, its activation was reported indirectly by investigating expression of Notch target genes in human and mouse colorectal and intestinal tumors respectively 8, 12. In this context, inactivation of the Notch pathway by use of a γ-secretase inhibitor (GSI), can drive a small percentage of intestinal tumor cells in the ApcMin mice to differentiate and undergo growth arrest, providing proof of principal that this pathway is a potential target for therapeutic approaches to colon cancer 8.
However, important questions remain unanswered. First, what is the requirement for endogenous Notch/ligand interaction in regulation of intestinal homeostasis? Although the function of HES1, MATH1 and RBP-Jκ, downstream components of the Notch pathway, have been addressed by loss of function studies 5, 8, 9, 13, these models may not reflect the complex signaling triggered by the different combinations of Notch ligand/receptor interaction 14 along the gut. Second, with regard to the potential use of GSI as a therapeutic approach in colon cancer, what are the longer-term consequences of disruption of intestinal homeostasis by targeting Notch signaling?
To address these issues, we adopted the strategy of blocking all Notch/ligand interactions by inactivating the expression of protein O-fucosyltransferase 1 (Pofut1) specifically in intestinal and colonic epithelial cells in the mouse. The rationale for this approach is that Notch receptors require the addition of O-fucose by POFUT1 to conserved Ser or Thr residues in epidermal growth factor-like repeats of their extracellular domain in order to signal through Delta-like and Jagged ligands 15–19. We report that intestinal Pofut1 deficiency dramatically inhibits Notch signaling, and promotes differentiation of intestinal proliferating progenitors to secretory cell fates. By 9 months, the disruption in intestinal homeostasis gives rise to extensive inflammation throughout the small and large intestine linked to alterations in the gut microbial population, with evidence for development of dysplasia and progression to tumor formation.
Pofut1F/F mice18 were on a mixed 129/C57BL/6 background. Transgenic Tg(Vil-cre)997Gum mice expressing Cre recombinase under control of the mouse villin 1 promoter 20 were purchased from the Jackson Laboratory. Breeding generated mice that were homozygous for the Pofut1 floxed allele and that harbored a Villin-Cre transgene. Mice were genotyped by PCR analysis as previously described 18, 20. All protocols were approved by the Animal Care and Use Committee at Montefiore Medical Center and the Albert Einstein College of Medicine. Small and large intestine were removed immediately after sacrifice of 4 or 36 week old mice, opened longitudinally and inspected under a dissecting microscope.
The small and large intestine were dissected, everted, filled to distension with PBS, and incubated with shaking at 37°C in 1.5 mM EDTA buffer. The sequential isolation of intestinal and colonic epithelial cells along the crypt-villus axis (CVA) was performed as previously described and validated 21, 22. Resulting fractions of dissociated epithelial cells were harvested by centrifugation at 1500 rpm at 4°C for 5 min, cell pellets were snap frozen in liquid nitrogen, and stored at −80°C.
Intestinal and colonic tissues were fixed overnight in 10% neutral buffered formalin, paraffin-embedded and sectioned at 5 μm. After dewaxing and hydration, sections were pretreated with 3% H2O2 for 10 min at room temperature and incubation of antibodies was performed overnight in BSA/PBS at 4°C. Antigen retrieval was achieved by boiling in 10mM citrate buffer pH 6.0 (30 min) for all antibodies, except anti-BrdU, which used 0.1% trypsin (20 min at 37°C). HRP-conjugated secondary antibodies were detected with the diaminobenzidine peroxidase substrate kit (Vector Laboratories). For assay of alkaline phosphatase activity, the substrate (Vector red; Vector Laboratories) was applied to the sections for 10 min.
Histological scores were assigned as follows: 0 = no significant change; 1 = minimal to mild change; 2 = moderate change, and 3 = severe change. This scoring system was applied to: crypt hyperplasia (elongation of the crypts, resulting in mucosa thickening), crypt dilation (crypt lumen widening associated with luminal mucous), inflammatory infiltrates into the lamina propria of the mucosa (evaluated by cell type and number). Moderate increases of infiltrates resulted in the separation of glands by inflammatory cells, more notable in marked inflammation, and associated with some crypt loss.
Protein extracts were prepared by homogenization of cell pellets in RIPA lysis buffer at 4°C, centrifugation at 15,000g for 15 min at 4°C, and supernatants were collected for Western blot analysis. Protein concentration was quantified using the Bradford assay (Bio-Rad Laboratories). Solubilized proteins (50–200 μg) were separated by electrophoresis on SDS-PAGE gels, transferred onto activated PVDF membranes and subjected to immunoblot analysis. Antibodies were then used according to the manufacturer’s protocols. Horseradish peroxidase (HRP)-conjugated secondary antibodies were detected by chemiluminescence. Equal protein loading was verified by reprobing membranes for β-actin.
Total RNA was extracted from frozen tissue cell pellets with Trizol reagent (Invitrogen). First-strand cDNA was synthesized from 1 to 5μg total RNA with 200 units of reverse transcriptase using the Superscript II kit (Invitrogen) according to the manufacturer’s recommendations.
cDNA was amplified using SYBR Green PCR Master Mix, and the ABI PRISM 7900HT Sequence Detection System Real Time PCR system (Applied Biosystems). Primer sequences (available upon request) were supplied by Sigma-Genosys. Murine actin was used as an internal reference. For each PCR product, there was a single dissociation curve, confirming specificity of the amplification. All reactions were carried out in duplicate. A standard curve was generated by similar analysis of 5-fold serial dilutions of a standard template DNA. Relative values for each PCR product were expressed in arbitrary units (AU) as a ratio of the target transcript normalized to mouse actin.
1.5cm of colonic flat mucosa was washed in cold PBS supplemented with penicillin and streptomycin, then cut into small pieces and maintained in culture for 24h at 37°C in 2ml of serum-free RPMI medium containing BSA (0.01%) and antibiotics (200g/ml penicillin, 200U/ml streptomycin). Before subsequent cytokine profile analysis, samples were centrifuged at 5000g for 5 minutes at 4°C, supernatants were collected and stored at −80°C
Mesenteric lymph nodes (MLN) and colonic mucosa were removed under sterile conditions, prepared with a sterile grinder and then plated onto either Trypticase soy agar with 5% sheep’s blood agar, MacConkey agar, or CDC anaerobes blood agar. Cultures were incubated at 35°C and examined at 24-h intervals for 2 days. Bacterial growth on the agar plates was quantified as the number of colony forming units per μg of tissue.
We first investigated the distribution of Pofut1 in the normal adult mouse intestinal and colonic epithelium by real time PCR. Fractionation of intestinal epithelial cells along the CVA revealed that fraction 4 was enriched in Sox9-expressing progenitor cells from the crypts whereas fraction 1 mainly contained Fabp2-expressing differentiated cells from the top of the villi (Fig 1A). Pofut1 expression was upregulated by more than 4-fold in the crypt compartment and thus displayed a similar expression pattern as Notch receptors and ligands (Fig S1). In the colon, Sox9 expression was enhanced in fraction 4, although there was a modest, not significant increase in Pofut1 in the crypt (Fig 1B). Whereas the complex distribution of Notch receptors and ligands suggested specific roles for different Notch-receptor complexes along the gut cephalo-caudal axis (Fig S2), the expression of Pofut1 was constant along this same axis (data not shown), consistent with its universal role in providing the necessary fucosylation of all Notch receptors.
To achieve inactivation of Pofut1 in the intestinal and colonic epithelium, Pofut1F/F mice 18 were crossed with a constitutive Tg(Vil-cre)997Gum in which Cre recombinase expression was driven by the villin promoter 20. The resulting Pofut1F/F:Villin-Cre progeny were obtained in the expected Mendelian ratio. Efficient recombination was demonstrated by the presence of only the deleted form of the floxed allele in isolated epithelial cells from Pofut1 mutant intestine or colon (Fig S3A). As predicted, recombination was specific for small and large intestine; it was not detected in the stomach or other peripheral tissues (Fig S3B). Pofut1-deficient mice appeared normal at birth, but smaller than mice carrying either no or one recombined Pofut1 floxed allele (Fig 1C). The mean reduction in weight was 42% in 4 week old Pofut1F/F:villin-Cre mice compared to Pofut1F/F littermates (n = 15) and this was associated with a modest decrease in survival (Fig S4).
Pofut1 has been shown in both Drosophila 16, 19 and mouse 18, to act upstream in the canonical Notch signaling pathway as a necessary component for the binding activity of Notch receptors to their ligands. In our model, functional blockage of Notch receptors signaling was achieved through inactivation of the Pofut1 gene, as illustrated by dramatically reduced NICD1, NICD2 and HES1 expression (Fig 1D, Table S1) in intestinal and colonic epithelial cells of Pofut1F/F:villin-Cre mice. Hes5 transcripts were repressed specifically in the small intestine (in contrast to the colon) upon Pofut1 gene excision. Concomitantly, Math1 mRNA expression was enhanced (Table S1) and MATH1-positive cells accumulated in the crypts of the small intestine (Fig 1E), providing direct evidence that Pofut1 expression is necessary for intestinal Notch signaling.
The substantial elevation of all intestinal secretory cell lineages following Pofut1 loss (Fig 2, Table S2) was reminiscent of Hes1 or Rbp-Jκ deletion 8, 9, providing strong evidence that POFUT1 targets Notch signaling in early crypt progenitors to control their lineage determination. Phenotypic alterations occurred mainly in the crypts and were homogeneous throughout the intestinal and colonic mucosa, again demonstrating highly efficient excision of the Pofut1-floxed alleles by the Cre transgene and the absence of a mosaic phenotype.
H&E staining (Fig 2A–B) and enhanced expression of lysozyme, MMP7 and Pla2g2a (Fig S5A–B, S6A) showed that the Paneth cell population was greatly expanded in the small intestine of Pofut1-deficient animals (Fig S5E). The expression of SOX9, necessary for proper differentiation of this cell type 23, 24, was also upregulated (Fig S6D). Moreover, Pofut1F/F:Villin-Cre mice exhibited goblet cell hypertrophy and hyperplasia in both the small and large intestine, as identified by Alcian blue and TFF3 staining (Fig 2C–D, S7, S5C–D), and by enhanced levels of Tff3, Muc2, Fizz2, and Spdef transcripts (Fig S6B). This expended goblet cell population secreted both neutral and acidic sialo- and sulfo-mucins (Fig S8). Most of these post-mitotic goblet cells accumulated in the crypt compartment, above the Paneth cells, where progenitors normally undergo rapid division (Fig 2D). Finally, the Pofut1-deficient intestine had an increased number of enteroendocrine cells all along the crypt-villus axis, although these cells resided more frequently in the vicinity of the crypt (Fig 2E–F, S5F). Moreover, the expression level of genes encoding gut hormones (CCK, GIP, Glucagon, Ghrelin, Somatostatin) revealed that all endocrine subtypes examined were expanded across the small intestine (Fig S6C, Table S2). Analysis of several transcription factors important for enteroendocrine specification and differentiation (Insm1, NeuroD, Ngn3, Gfi1) showed significantly increased expression in Pofut1F/F:Villin-Cre small intestine (Fig S6D).
Pofut1 inactivation also enhanced secretory cell lineage differentiation at the modest expense of absorptive cell differentiation in the small intestine. In fact, attenuation in villus length (8%) (Fig S9A), and decreased transcript levels of several enterocyte specific genes (Fabp2, Dpp4, lactase, Cdx2; Fig S9B) suggested that the number of enterocytes was reduced in Pofut1-deficient intestine, although alkaline phosphatase staining and ultrastructural examination showed maintenance of the enterocyte brush border in Pofut1F/F:Villin-Cre mice (Fig S9C–F). These data demonstrated that Pofut1 loss triggers the binary choice of proliferating progenitors between differentiation toward secretory or absorptive cell lineages.
Loss of function experiments have shown that neither NOTCH1 nor HES1 are essential for crypt progenitor cell proliferation 9, 10, suggesting redundancy in Notch receptors 11 or Notch targets in stem cell maintenance, in the crypt epithelium. This is consistent with the altered proliferation observed in crypt progenitors from Pofut1-deficient intestine, where the entire Notch signaling pathway is presumably turned off. Importantly, reduction in Ki67 staining was more pronounced in the colon of Pofut1-deficient mice compared to the small intestinal mucosa (Fig 3A–E). In parallel, the expression of the Cdk inhibitor p21 was upregulated in Pofut1-deficient intestine or colon whereas cyclin D1 protein level was decreased (Fig 3F). In the small intestinal epithelium, the transient amplifying compartment, which normally occupies the entire crypt (Fig 3A), was relocated to the upper part of the crypt, above the expanded secretory cell population (Fig 3B). Cell proliferation counts revealed a 25% reduction in proliferation, but paradoxically, there was an increase in total number of cells per crypt (Fig 3E) as well as a 1.6 fold increase in crypt depth (Fig S9A), likely reflecting the accumulation of secretory cells in the crypt of Pofut1F/F:Villin-Cre mice. The slight reduction in migration of BrdU-labeled cells in the small intestine (Fig S10) suggested that secretory progeny migrated out of the crypts more slowly in Pofut1-deficient intestine. Alternatively, fewer cells may divide, but with a shortened cell cycle, producing a larger number of secretory cells. In the distal colon, proliferation of crypt progenitors was dramatically decreased (estimated as 65% of Ki67 positive cells per crypt) (Fig 3E). We did not observe intestinal or colonic crypt cells stained for cleaved caspase 3 or TUNEL in Pofut1-deficient mice (data not shown). Confirming a marked reduction in apoptosis in the Pofut1F/F:Villin-Cre mice, both cleaved caspase 3 and PARP were undetectable in the colonic mucosa of these mice, in contrast to their prominence in the control Pofut1F/F mice (Fig 3F).
The essential role of POFUT1 in maintenance of adult intestinal homeostasis was demonstrated by the chronic and diffuse entero-colitis that developed in 100% (18/18) of 36 week old Pofut1F/F: Villin-cre mice. The inflammation was evident at 4 weeks, as shown by the thickened intestinal wall (Fig S7B, S7D) and the increased paracellular mucosal permeability (Fig S11), but severity of inflammation substantially increased in 36 week old Pofut1-deficient mice, in which changes were more pronounced in the colon (Fig 4A–D). The goblet cell hypertrophy and hyperplasia described above persisted in these older mice (Table I). The entero-colitis was characterized primarily by crypt hyperplasia, edema in the lamina propria with mixed inflammatory infiltrates, and frequent crypt abscesses and transmural inflammation. The colonic mucosa was thickened predominantly by T lymphocytes and macrophages, as demonstrated by CD3 and Mac-387 (Fig 4E–F) staining. This was consistent with enhanced production of chemoattractants for activated T cells and macrophages in total colonic mucosa from Pofut1F/F: Villin-cre mice (Fig 5, Table S3). Increased expression of the α1, α4, β1, and β2 integrins and their ligands (MadCAM1, VCAM1, ICAM1 and fibronectin) likely contributed to immune cell adhesion to endothelium, mucosal homing, and activation of lymphocytes (Fig 5, Table S3). In addition, colonic secretions of the Th1 (IL2 and IFNγ), and the Th17 cytokines (IL17, IL23) were elevated after Pofut1 deletion, whereas Th2 cytokines (IL4, IL10) production was not changed (Fig 5, Table S4). Macroscopically, the entero-colitis was accompanied by lymph vessel dilation (Fig 4G), a more prominent gut-associated lymphoid tissue (GALT) and occasional spleen hypertrophy (Fig 4H), suggesting a chronic immune stimulation.
Histopathologic examination of the small intestine of Pofut1 mutant mice also demonstrated the presence of dysplastic foci (Fig 6A–B), morphologically distinct from aberrant crypt foci. These small nodules of dysplastic glandular epithelium had few to no goblet cells, no villi (i.e. were sessile lesions), and atypical epithelial cells, usually smaller than their adjacent non-dysplastic counterparts.
Although the cause of the inflammation was uncertain, we hypothesized that the enhanced production of mucus in Pofut1-deficient intestine and colon (Fig 6C, 6E) could result in changes in the mucus-associated flora, thereby allowing surface associated bacteria with increased inflammatory potential to become permanently established and contribute to the development of chronic inflammation. In fact, frequent crypt necrosis was observed in the intestinal and colonic mucosa of Pofut1F/F: Villin-cre mice at 9 months (Fig 6D), suggesting bacterial infection.
To investigate this, we performed bacterial culture and histological scoring of Warthin-Starry stained (silver stain for spirilliform bacteria) sections from colonic mucosa of Pofut1F/F and Pofut1F/F:Villin-cre mice. Despite no significant overgrowth of aerobic or anaerobic bacteria associated with the colonic mucosa of the mutant mice, Gram-negative bacteria colonies were more frequently associated with the large intestinal mucosa in Pofut1F/F:Villin-cre mice (Table II) and spiral-shaped organisms (interpreted to be Helicobacter spp) accumulated within the dilated and mucous-filled glands of the large intestine of Pofut1F/F:Villin-cre mice (Fig 6F–G). Interestingly, Gram-negative bacteria translocation across the intestinal epithelium to mesenteric lymph nodes was observed exclusively in Pofut1F/F:Villin-cre mice (Table II). In sum, these data suggest that disruption in the luminal environment of Pofut1-deficient mice contributes to an alteration in the composition of bacterial flora associated with the colonic mucosa which likely initiates the pronounced inflammation.
Here, we report that blocking of all Notch/ligand interaction by targeted inactivation of the Pofut1 gene in mouse small intestinal and colonic epithelia enhanced the commitment of crypt progenitor cells towards secretory cell lineages, and leads to hypersecretion of mucus, modification of the gut microbiota and the development of intestinal inflammation.
Consistent with previous studies 25, 26, we report that activation of Notch signaling occurs through concentration of Notch ligands and receptors predominantly in epithelial cells of the crypts, where Pofut1 expression is the highest. The efficient down-regulation of NICD and HES1 expression, and the substantial elevation of all intestinal secretory cell lineages following Pofut1 loss, which was reminiscent of Hes1 or Rbp-Jκ deletion 8, 9, provided strong evidence that POFUT1 targets Notch signaling in early crypt progenitors to control their lineage determination. Importantly, we show that POFUT1 is necessary for the transcriptional program of early secretory precursors, since expression of Gfi1, that selects Paneth/goblet versus enteroendocrine cell fates 27, is induced in Pofut1-deficient intestine. We also demonstrate that Pofut1 loss controls an earlier event in proliferating progenitors where it triggers the binary choice between differentiation toward secretory or absorptive cell lineages, since reduced enterocytic differentiation was accompanied by an increased number of secretory cells. This is in accordance with previous in vivo experiments, in which inactivation of Rbp-Jκ or Hes1 led to a reduced number of differentiated enterocytes 8, 9, whereas only enterocytes emerged from Math1-mutant crypts13.
The extent of phenotypic changes, such as proliferation index, observed in small compared to large intestine lacking Pofut1 cannot be accounted for by differential expression of the O-fucosyltransferase. Rather, the differences likely arise from the complex expression pattern of Notch receptors and their ligands in the intestinal epithelium, which we have documented along the cephalo-caudal axis (Fig S1, S2). In fact, the type and the combination of the Notch ligand/receptor interactions in a given environment may have different effects on target gene expression14, transcriptional activity of RBP-Jκ and thus biological effect. Interestingly, compared to the total absence of proliferating progenitors in Rbp-Jκ-deficient intestine8, the more modest reduction in cell proliferation in Pofut1F/F:Villin-cre mice was consistent with their viability (Fig S4). This discrepancy could underlie transcriptional activity of RBP-Jκ independent of NICD, as reported in the adult exocrine pancreas 28 and in 293T kidney cells 29. However, intestinal phenotypic changes in mice treated with a GSI were indistinguishable from those observed after Rbp-Jκ deletion 8, suggesting that Notch receptors might partially transduce signals independently of POFUT1 in the intestinal mucosa 30. An additional explanation for the differences between Pofut1- and Rbp-Jκ-deficient mice might be a different expression pattern of POFUT1 and RBP-Jκ in crypt progenitors or stem cells. RBP-Jκ might target earlier intestinal progenitors than Pofut1.
The ability of Pofut1 inactivation in the intestinal epithelium to induce both mucus hypersecretion and colitis was unexpected since colonic inflammation is usually accompanied by a compromised mucus barrier and goblet cell hypoplasia31. However, subsequent to the enhanced production of goblet cells in Pofut1-deficient mice, increased number of Gram negative bacteria associated with the mucus layer may contribute to the pronounced chronic inflammatory reaction, similar to abnormalities observed in the small intestine of the mouse model for cystic fibrosis32. Interestingly, the enhanced exposure of the colonic mucosa to spiral-shaped bacteria (such as Helicobacter species), which naturally colonize the mucus layer and accumulate in the crypts of Pofut1F/F:Villin-cre, may trigger the strong Th1/Th17-associated adaptative immune response and the GALT hypertrophy. Infection with Helicobacter spp. has been associated with the development of T-cell mediated inflammatory bowel disease in various mouse models33. Additional mechanisms independent of Notch pathway may be important. Glycosylation changes in gastro-intestinal POFUT1 target, such as Muc13 34, may affect bacterial adherence to the mucosa and then lead to increased inflammation. Moreover, the inflammation may be a predisposing factor for dysplastic lesions to progress to tumors in the Pofut1F/F: Villin-cre mice, since 100% of the mice displayed mucosal inflammation and frequent dysplastic foci, though only one of ten Pofut1-deficient mice examined displayed an early stage adenoma in the small intestine (Fig S12). Thus, the Pofut1-deficient mouse may be a new model in which to dissect the complex relationship between colitis and elevated risk for colon cancer development and its modulation.
In conclusion, the reduced epithelial proliferation, and enhanced secretory cell differentiation in response to Pofut1 inactivation demonstrate that the interactions between Notch receptors and ligands provide growth and differentiation signals in the intestinal proliferating cell compartment. Moreover, activation of Notch signaling, reflected by increased expression of HES1 and reduced HATH1 levels, has been described in human colorectal tumors and mouse models of intestinal tumorigenesis 8, 12. In this regard, down-regulation of Notch signaling through the use of γ-secretase inhibitors decreased cell proliferation and drove cells of ApcMin mouse intestinal tumors to differentiate, and is thus a potential new approach for therapy of colon tumors8. However, we demonstrate that the overproduction of mucus following intestinal Pofut1 deletion is paralleled by modifications of the gut microbiota that may contribute to the development of intestinal inflammation. Since alterations in lineage specific differentiation and mucus secretion are seen with loss of Notch signaling 8, 11, there is the potential that development of therapeutic approaches for colon tumors that target this pathway may have to address effects similar to those that develop longer term through Pofut1 inactivation, including cytotoxicity of normal crypts, chronic intestinal inflammation, and dysplasia.
We thank T. Sudo (Toray Inductries, Inc., Kamakura, Japan) for the gift of Hes1 antibody. We are grateful to A. Velcich, L. Klampfer, and JM. Mariadason for helpful comments. This work was supported by NCI grants RO1 CA114265 and U54 CA100926 to LH Augenlicht and RO1 CA95022 to Pamela Stanley, by the Philippe Foundation, and by the Albert Einstein Cancer Center grant PO1 13330.
No conflict of interest to disclose
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.