MET Signaling Mediates Intestinal Crypt-Villus Development, Regeneration, and Adenoma Formation and Is Promoted by Stem Cell CD44 Isoforms.

BACKGROUND & AIMS: Resistance of metastatic human colorectal cancer cells to drugs that block epidermal growth factor (EGF) receptor signaling could be caused by aberrant activity of other receptor tyrosine kinases, activating overlapping signaling pathways. One of these receptor tyrosine kinases could be MET, the receptor for hepatocyte growth factor (HGF). We investigated how MET signaling, and its interaction with CD44 (a putative MET coreceptor regulated by Wnt signaling and highly expressed by intestinal stem cells [ISCs] and adenomas) affects intestinal homeostasis, regeneration, and adenoma formation in mini-gut organoids and mice. METHODS: We established organoid cultures from ISCs stimulated with HGF or EGF and assessed intestinal differentiation by immunohistochemistry. Mice with total epithelial disruption of MET (AhCre/Metfl/fl/LacZ) or ISC-specific disruption of MET (Lgr5Creert2/Metfl/fl/LacZ) and control mice (AhCre/Met+/+/LacZ, Lgr5Creert2/Met+/+/LacZ) were exposed to 10 Gy total body irradiation; intestinal tissues were collected, and homeostasis and regeneration were assessed by immunohistochemistry. We investigated adenoma organoid expansion stimulated by HGF or EGF using adenomas derived from Lgr5Creert2/Metfl/fl/Apcfl/fl and Lgr5Creert2/Met+/+/Apcfl/fl mice. The same mice were evaluated for adenoma prevalence and size. We also quantified adenomas in AhCre/Metfl/fl/Apcfl/+ mice compared with AhCre/Met+/+/Apcfl/+ control mice. We studied expansion of organoids generated from crypts and adenomas, stimulated by HGF or EGF, that were derived from mice expressing different CD44 splice variants (Cd44+/+, Cd44-/-, Cd44s/s, or Cd44v4-10/v4-10 mice). RESULTS: Crypts incubated with EGF or HGF expanded into self-organizing mini-guts with similar levels of efficacy and contained all differentiated cell lineages. MET-deficient mice did not have defects in intestinal homeostasis. Total body irradiation reduced numbers of proliferating crypts in AhCre/Metfl/fl/LacZ mice. Lgr5Creert2/Metfl/fl/LacZ mice had impaired regeneration of MET-deficient ISCs. Adenoma organoids stimulated with EGF or HGF expanded to almost twice the size of nonstimulated organoids. MET-deficient adenoma organoids did not respond to HGF stimulation, but did respond to EGF. ISC-specific disruption of Met (Lgr5Creert2/Metfl/fl/Apcfl/fl mice) caused a twofold increase in apoptosis in microadenomas, resulting in an approximately 50% reduction of microadenoma numbers and significantly reduced average adenoma size. Total epithelial disruption of Met (AhCre/Metfl/fl/Apcfl/+ mice) resulted in an approximate 50% reduction in (micro)adenoma numbers. Intestinal crypts from Cd44-/- mice did not expand to the same extent as crypts from Cd44+/+ mice on stimulation with HGF, but had the same response to EGF. The negative effect on HGF-mediated growth was overcome by expression of CD44v4-10, but not by CD44s. Similarly, HGF-mediated expansion of adenoma organoids required CD44v4-10. CONCLUSIONS: In studies of intestinal organoid cultures and mice with inducible deletion of MET, we found HGF receptor signaling to regulate intestinal homeostasis and regeneration, as well as adenoma formation. These activities of MET are promoted by the stem cell CD44 isoform CD44v4-10. Our findings provide rationale for targeting signaling via MET and CD44 during anti-EGF receptor therapy of patients with colorectal cancer or in patients resistant to EGF receptor inhibitors.

CD44 expression in intestinal epithelium and colorectal cancer is independent of p53 status.

CD44 marks stem cell-like cells in a number of tumour types, including colorectal cancer (CRC), while aberrant CD44 expression conveys increased tumourigenic, invasive, and metastatic potential. Previous data indicate that CD44 is a direct target of p53-mediated transcriptional repression in breast cancer. Since inactivating p53 mutations are frequent genetic events in CRC these could unleash expression of CD44. In the present study, we therefore explored the relation between p53 mutational status and CD44 expression in a cohort of 90 localized primary CRCs and studied the effect of radiation-induced p53 activation on CD44 expression. Interestingly, we observed that, in contrast to breast cancer, loss of function p53 mutations were not associated with elevated CD44 expression in colon cancer. Moreover, DNA-damage induced p53 activation did not result in repression of CD44 expression, neither in colon cancer cells nor in normal intestinal epithelial cells. Our data demonstrate that CD44 expression in normal and malignant intestinal epithelial cells is not regulated by p53, implying that regulation of this potentially important therapeutic target is tissue and cancer-type specific.

WNT signaling controls expression of pro-apoptotic BOK and BAX in intestinal cancer.

In a majority of cases, colorectal cancer is initiated by aberrant activation of the WNT signaling pathway. Mutation of the genes encoding the WNT signaling components adenomatous polyposis coli or ß-catenin causes constitutively active ß-catenin/TCF-mediated transcription, driving the transformation of intestinal crypts to cancer precursor lesions, called dysplastic aberrant crypt foci. Deregulated apoptosis is a hallmark of adenomatous colon tissue. However, the contribution of WNT signaling to this process is not fully understood. We addressed this role by analyzing the rate of epithelial apoptosis in aberrant crypts and adenomas of the Apc(Min/+) mouse model. In comparison with normal crypts and adenomas, aberrant crypts displayed a dramatically increased rate of apoptotic cell death. Expression profiling of apoptosis-related genes along the crypt-villus axis and in Apc mutant adenomas revealed increased expression of two pro-apoptotic Bcl-2 family members in intestinal adenomas, Bok and Bax. Analysis of the colon of familial adenomatous polyposis (FAP) patients along the crypt-to-surface axis, and of dysplastic crypts, corroborated this expression pattern. Disruption of ß-catenin/TCF-4-mediated signaling in the colorectal cancer cell line Ls174T significantly decreased BOK and BAX expression, confirming WNT-dependent regulation in intestinal epithelial cells. Our results suggest a feedback mechanism by which uncontrolled epithelial cell proliferation in the stem cell compartment can be counterbalanced by an increased propensity to undergo cell death.

The AC133 epitope, but not the CD133 protein, is lost upon cancer stem cell differentiation.

Colon cancer stem cells (CSC) can be identified with AC133, an antibody that detects an epitope on CD133. However, recent evidence suggests that expression of CD133 is not restricted to CSCs, but is also expressed on differentiated tumor cells. Intriguingly, we observed that detection of the AC133 epitope on the cell surface decreased upon differentiation of CSC in a manner that correlated with loss of clonogenicity. However, this event did not coincide with a change in CD133 promoter activity, mRNA, splice variant, protein expression, or even cell surface expression of CD133. In contrast, we noted that with CSC differentiation, a change occured in CD133 glycosylation. Thus, AC133 may detect a glycosylated epitope, or differential glycosylation may cause CD133 to be retained inside the cell. We found that AC133 could effectively detect CD133 glycosylation mutants or bacterially expressed unglycosylated CD133. Moreover, cell surface biotinylation experiments revealed that differentially glycosylated CD133 could be detected on the membrane of differentiated tumor cells. Taken together, our results argue that CD133 is a cell surface molecule that is expressed on both CSC and differentiated tumor cells, but is probably differentially folded as a result of differential glycosylation to mask specific epitopes. In summary, we conclude that AC133 can be used to detect cancer stem cells, but that results from the use of this antibody should be interpreted with caution.

Deletion of the WNT target and cancer stem cell marker CD44 in Apc(Min/+) mice attenuates intestinal tumorigenesis.

Mutation of the genes encoding the WNT signaling components adenomatous polyposis coli or beta-catenin plays a critical role in the initiation of colorectal cancer. These mutations cause constitutively active beta-catenin/TCF-mediated transcription, driving the transformation of intestinal crypts to colorectal cancer precursor lesions, called dysplastic aberrant crypt foci. CD44 is a prominent WNT signaling target in the intestine and is selectively expressed on the renewing epithelial cells lining the crypts. The expression of CD44 is dramatically increased in aberrant crypt foci in both humans and tumor-susceptible Apc(Min/+) mice, suggesting a role for CD44 in intestinal tumorigenesis. To study this role, we crossed C57BL/6J-Cd44(-/-) mice with C57BL/6J-Apc(Min/+) mice. Compared with C57BL/6J-Cd44(+/+)/Apc(Min/+) mice, C57BL/6J-Cd44(-/-)/Apc(Min/+) mice showed an almost 50% reduction in the number of intestinal adenomas. This reduction was primarily caused by a decrease in the formation of aberrant crypts, implying the involvement of CD44 in tumor initiation. The absence of CD44 in the normal (nonneoplastic) crypts of Cd44(-/-)/Apc(Min/+) mice did not alter the proliferative capacity and size of the intestinal stem cell and transit-amplifying compartments. However, compared with Cd44(+/+)/Apc(Min/+) mice, Cd44(-/-)/Apc(Min/+) showed an increase in the number of apoptotic epithelial cells at the base of the crypt which correlated with an increased expression of the proapoptotic genes Bok and Dr6. Our results show an important role for CD44 in intestinal tumorigenesis and suggest that CD44 does not affect proliferation but is involved in the control of the balance between survival and apoptosis in the intestinal crypt.