A two-step model for colon adenoma initiation and progression caused by APC loss

Abstract

Aberrant Wnt/beta-catenin signaling following loss of the tumor suppressor adenomatous polyposis coli (APC) is thought to initiate colon adenoma formation. Using zebrafish and human cells, we show that homozygous loss of APC causes failed intestinal cell differentiation but that this occurs in the absence of nuclear beta-catenin and increased intestinal cell proliferation. Therefore, loss of APC is insufficient for causing beta-catenin nuclear localization. APC mutation-induced intestinal differentiation defects instead depend on the transcriptional corepressor C-terminal binding protein-1 (CtBP1), whereas proliferation defects and nuclear accumulation of beta-catenin require the additional activation of KRAS. These findings suggest that, following APC loss, CtBP1 contributes to adenoma initiation as a first step, whereas KRAS activation and beta-catenin nuclear localization promote adenoma progression to carcinomas as a second step. Consistent with this model, human FAP adenomas showed robust upregulation of CtBP1 in the absence of detectable nuclear beta-catenin, whereas nuclear beta-catenin was detected in carcinomas.

Figure 1. KRAS promotes intestinal cell proliferation and nuclear localization of β-catenin following loss of APC

(A) apc^mcr zebrafish embryos were injected at the one cell stage with KRAS mRNA. 72 hpf embryos were fixed, sectioned and stained for hematoxylin and eosin (H&E) (Row 1, sagital section; row 2, cross section), DNA (magenta- middle and bottom) and either PCNA (middle-green) or β-catenin (bottom- green). Overlapping expression is shown in white. (B) Protein lysates from 72 hpf embryos were subjected to western blot analysis for either V5-tag (top-left) and total RAS (bottom-left) or phospho-Erk (top-right) and vinculin (bottom-right). (C) Zebrafish harboring an integrated β-catenin TOPGFP reporter were injected with KRAS mRNA, APC morpholino or both. 72 hpf embryos were subjected to whole mount in situ hybridization for GFP. Boxes indicate the intestine (top) and arrows indicate the hindbrain (bottom). All images were captured using the same exposure and represent at least three independent experiments. (Scale Bar: 10μm)

Figure 2. KRAS and RAF1 are necessary for nuclear localization of β-catenin and intestinal cell proliferation following loss of APC

(A) apc^mcr zebrafish embryos injected with constitutively active RAF1 or MEK1. 72 hpf embryos were stained for hematoxylin and eosin (H&E) staining (top), DNA (magenta-middle and bottom) and either PCNA (middle-green) or β-catenin (bottom-green). (B) SW-480 cells transfected with control, KRAS^G12V, RAF1 directed siRNA or DN-MEK1 constructs were stained for DNA (magenta) and β-catenin (green). (C) RNA was harvested and subjected to rt-PCR for Axin2. (D) SW-480 cells transfected with KRAS^G12V-specific siRNA were cotransfected with constitutively active KRAS, RAF1 or MEK1 and stained for DNA (magenta) and β-catenin (green). Overlapping expression is shown in white. All images were captured using the same exposure and represent three independent experiments. (Scale Bar: 10μm)

Figure 3. KRAS and RAF1 direct stabilized β-catenin to the nucleus

(A) WT and KRAS-injected zebrafish embryos treated with PGE₂ were stained for DNA (magenta), PCNA (green) and β-catenin (green). (B) Human 293 cells were transfected with constitutively active KRAS, RAF1 or MEK1 and either DMSO (top) or PGE₂ (bottom). Cells were stained for DNA (magenta) and β-catenin (green). (C) 293 cells treated with DMSO, PGE₂ or PGE₂ and leptomycin B were stained for DNA (magenta) and β-catenin (green). (D) 293 cells were transfected with control or KRAS- or RAF1-directed siRNA and treated with EGF and PGE₂ or vehicle then stained for DNA (magenta) and β-catenin (green). (E) 293 cells were subjected to an MTT assay (*p<0.05 vs DMSO). Overlapping expression is shown in white. All images were captured using the same exposure and represent three independent experiments. (Scale Bar: 10μm)

Figure 4. KRAS/RAF1 Regulation of β-catenin requires the Activity of RAC1

(A, B) SW-480 cells transfected with vehicle or myc-tagged (A, S191A or S605A) or flag-tagged (B, S552A or S675A) β-catenin mutants were stained for DNA (magenta) and α-myc (A, green) or α-flag (B, green). (C) SW-480 cells were transfected with RAC1-directed siRNA, DN-RAC1, DN-Cdc42 or treated with the RAC1-specific inhibitor NSC23766 were stained for DNA (magenta) and β-catenin (green). (D) SW-480 cells were subjected to an MTT assay (*p<0.05 vs DMSO). (E) SW-480 cells were transfected with vehicle, KRAS^G12V-targeted siRNA, RAF1 siRNA or DN-MEK1 and subjected to a RAC1 activity assay. The western blot was probed for RAC1 (top). Control lysates were probed for total RAC1 (bottom). (F) SW-480 cells treated as above were stained for phospho-cJun. (G) Human 293 cells transfected with constitutively active RAC1 and treated with PGE₂ or DMSO were stained for DNA (magenta) and β-catenin (green). All images were captured using the same exposure and are representative of at least three independent experiments. (Scale Bar: 10μm)

Figure 5. KRAS-mediated intestinal cell proliferation following loss of APC requires β-catenin

(A) Zebrafish embryos were injected with β-catenin^S45A mRNA (left panel) along with KRAS mRNA (middle panel). Also shown is a representative image of the APC-KRAS embryo (right panel). At 72hpf, the embryos were fixed and photographed. (B) The percent cyclops was analyzed (*p<0.01 WT vs β-catenin-KRAS, **p<0.01 WT vs. APC-KRAS). (C) Protein was harvested from apc^mcr embryos treated with DMSO or NS-398 and subjected to western blot analysis for β-catenin (top) or vinculin (bottom). (D) Wildtype uninjected or apc^mcr embryos injected with KRAS mRNA treated with VEH (top) or NS-398 (bottom) were stained by H&E (WT-left, APC-KRAS-right) and for DNA (magenta), PCNA (green) and β-catenin (green). TOPGFP-APC^mo-KRAS embryos were stained for GFP expression (purple). Boxes indicate the intestine. (E) SW-480 cells treated with DMSO or NS-398 were stained for DNA (magenta) and β-catenin (green). Protein lysates were subjected to western blot analysis for β-catenin (top) and vinculin (bottom). (F) Cells were subjected to MTT analysis (*p<0.05 vs DMSO). Overlapping expression is shown in white. All images were captured using the same exposure and represent at least three independent experiments. (Scale Bar: 10μm)

Figure 6. apc control of cellular fate and differentiation are mediated by ctbp1 and are β-catenin-independent

(A) Protein from 48 hpf apc^mcr zebrafish embryos injected with either 6xHis-APC^955-2075 or 6xHis-APC^955-2075-AALP were subjected to western blot analysis for β-catenin (top), ctbp1 (second), 6xHis (third) or vinculin (bottom). (B) Embryos injected as above were fixed at 72 hpf and subjected to in situ hybridization for i-fabp. (C) apc^mcr embryos were treated with DMSO or NS-398 or injected with ctbp1-directed morpholino in the presence or absence of NS-398. At 72 hpf, the embryos were fixed and subjected to in situ hybridization for i-fabp, NaPi, hoxa13a or evx1. All images are representative of at least three independent experiments.

Figure 7. JNK1 activation is coincident with nuclear accumulation of β-catenin in human tumor samples

FFPE human matched grossly uninvolved and adenoma samples obtained from FAP patients and unmatched sporadic carcinomas were stained to indicate DNA (magenta) and (A) CtBP (green), (B) β-catenin* (green) or (C) phospho-cJun (green) as an indicator of JNK activity. Overall, 20 patient matched grossly uninvolved and adenoma or unmatched carcinoma tissue samples were stained. Shown are two representative samples. All images were captured using the same exposure and overlapping expression is shown in white. *Note: a section from each of the β-catenin-stained adenomas was enlarged to the right. Each antibody was evaluated using serial sections. (Scale Bar: 5μm)