Activation of β-Catenin-Tcf Signaling in Colon Cancer by Mutations in β-Catenin or APC

See allHide authors and affiliations

Science  21 Mar 1997:
Vol. 275, Issue 5307, pp. 1787-1790
DOI: 10.1126/science.275.5307.1787


Inactivation of the adenomatous polyposis coli (APC) tumor suppressor gene initiates colorectal neoplasia. One of the biochemical activities associated with the APC protein is down-regulation of transcriptional activation mediated by β-catenin and T cell transcription factor 4 (Tcf-4). The protein products of mutant APC genes present in colorectal tumors were found to be defective in this activity. Furthermore, colorectal tumors with intact APC genes were found to contain activating mutations of β-catenin that altered functionally significant phosphorylation sites. These results indicate that regulation of β-catenin is critical to APC's tumor suppressive effect and that this regulation can be circumvented by mutations in either APC or β-catenin.

Mutations of the APC gene are the most common disease-causing genetic events in humans; about 50% of the population will develop colorectal polyps initiated by such mutations during a normal life-span (1). Individuals who inherit APC mutations develop thousands of colorectal tumors, consistent with the tumor suppressor or “gatekeeping” role of APC protein in colorectal tumorigenesis (2, 3). APC homodimerizes through its NH2-terminus (4) and interacts with at least six other proteins: β-catenin (5), γ-catenin (plakoglobin) (6), tubulin (7), EB1 (8), hDLG, a homolog of the Drosophila Discs Large tumor suppressor protein (9), and glycogen synthase kinase-3β (GSK-3β) (10), a mammalian homolog of ZW3 kinase. Whether any of these interacting proteins communicate APC growth-controlling signals is unknown. Here, we used a genetic approach to investigate the role of β-catenin in APC's tumor suppressor function.

Although β-catenin was originally discovered as a cadherin-binding protein, it has recently been shown to function as a transcriptional activator when complexed with members of the Tcf family of DNA binding proteins (11). One family member, hTcf-4, is expressed in normal and neoplastic colorectal epithelium, and wild-type (WT) APC can suppress signaling by the β-catenin-Tcf complex (12). If this inhibitory activity is critical for APC's tumor suppressor function, then mutant APC proteins should be defective in this activity.

To evaluate this hypothesis, we tested four APC mutants (Fig. 1A) for their ability to inhibit β-catenin-Tcf-regulated transcription (CRT) in transfection assays. The first mutant, APC331Δ, represents a type of mutation commonly found in the germ line of familial adenomatous polyposis patients as well as in sporadic tumors (2). The APC331Δ protein is truncated at codon 331, NH2-terminal to the three 15-amino acid (aa) β-catenin-binding repeats between codons 1020 and 1169. The second mutant, APC1309Δ, is the most common germline APC mutation (2), a 5-base pair (bp) deletion that produces a frameshift at codon 1309 and truncation of the protein. The APC1309Δ protein retains the 15-aa β-catenin-binding repeats but lacks the seven 20-aa repeats between codons 1323 and 2075 that have been implicated in binding and degradation of β-catenin (5). The third mutant, APC1941Δ, represents one of the most distal somatic mutations observed in colorectal tumors (13). The APC1941Δ protein is truncated at codon 1941 and therefore contains the 15-aa repeats and all but the last two 20-aa repeats. Finally, APC2644Δ represents a germline mutation resulting from a 4-bp deletion in codon 2644. Patients with this type of unusual COOH-terminal mutation develop few polyps (attenuated polyposis) but have pronounced extracolonic disease, particularly desmoid tumors (14).

Fig. 1.

Effects of APC mutations on CRT. (A) Schematic of wild-type (WT) and mutant APC. APC is a 2843-aa protein (23) that contains Armadillo (ARM) repeats in the NH2-terminus (24), 15- and 20-aa β-catenin-binding repeats in the central region (5, 6), and a basic region in the COOH-terminus (23). The COOH-terminus also contains a TXV sequence, which mediates DLG binding (9). (B) Effects of WT and mutant APC on CRT. SW480 cells containing endogenous mutant APC were transfected with the APC expression vectors shown in (A) and CRT was measured (25). Cells were transfected with increasing amounts of WT APC (0, 0.15, and 0.5 μg, indicated by the black wedge) or 0.5 μg of mutant APC. CRT reporter activities are expressed relative to assays containing no WT APC and are the means of three replicates. Error bars represent standard deviations.

Each of the APC mutants was cotransfected with a CRT reporter into the SW480 colorectal cancer cell line. SW480 cells have truncated APC and constitutively active CRT, which can be suppressed by exogenous WT APC (12). Although all four mutants produced comparable amounts of APC protein after transfection (15), they varied in their CRT inhibitory activity. The three mutants found in patients with typical polyposis or cancer were markedly deficient in inhibition of CRT (Fig. 1B). The reduced activity of APC1309Δ and APC1941Δ suggests that β-catenin binding is not sufficient for APC-mediated inhibition of CRT and that the complete set of 20-aa repeats is required. The inhibitory activity of the APC2644Δ mutant associated with attenuated polyposis was comparable with that of WT APC (Fig. 1B), suggesting that the DLG-binding domain at the COOH-terminus of APC is not required for down-regulation of CRT.

If APC's inhibition of CRT is critical to suppression of colorectal tumorigenesis, cancers with WT APC must escape this inhibition through alternative mechanisms. To investigate this possibility, we evaluated CRT in two colorectal tumor cell lines (HCT116 and SW48) that express full-length APC (Fig. 2A). Both HCT116 and SW48 displayed constitutively active CRT and, in contrast to cell lines with truncated APC (DLD1 and SW480), this activity was not inhibited by exogenous WT APC (Figs. 1B and 2B). Other (noncolorectal cancer) cell lines expressing WT APC do not display constitutive CRT activity (12).

Fig. 2.

Evaluation of CRT in colorectal cancer cell lines with WT APC. (A) Immunoblot of endogenous APC in the SW480, DLD1, HCT116, SW48, and 293 cell lines, developed with APC monoclonal antibody FE9 (26). (B) Effects of exogenous WT APC on CRT in cell lines with endogenous mutated or WT APC. Cells were transfected with increasing amounts (0, 0.15, and 0.5 μg for DLD1 and SW48; 0, 0.5, and 5 μg for HCT116) of WT APC or APC1309Δ mutant (0.5 μg for DLD1 and SW48; 5 μg for HCT116), and CRT was assessed as in Fig. 1. CRT reporter activities are expressed relative to activity in extracts without exogenous APC and are the means of three replicates. Error bars represent standard deviations.

These transfection results suggested that the constitutive CRT in HCT116 and SW48 might be due to an altered downstream component of the APC tumor suppressor pathway. We therefore evaluated the status of a likely candidate, β-catenin, in the same four lines. All four lines expressed similar amounts of apparently intact β-catenin, as assessed by immunoblots (Fig. 3A). However, sequence analysis revealed that both HCT116 and SW48 harbored mutations in the β-catenin gene (CTNNB1) (Fig. 3B). HCT116 had a 3-bp deletion that removed one amino acid (Ser45) and SW48 had a C-to-A missense mutation that changed Ser33 to Tyr. Analysis of paraffin-embedded archival tissue from the HCT116 patient confirmed the somatic nature of this mutation and its presence in the primary tumor before culture (16). Both mutations affect serines that have been implicated in the down-regulation of β-catenin through phosphorylation by the GSK-3β kinase in Xenopus embryos (Fig. 3C) (17, 18).

Fig. 3.

Evaluation of β-catenin in colorectal cancer cell lines with WT APC. (A) Immunoblot of the cell lines used in this study, developed with β-catenin monoclonal antibody C19220 (Transduction Laboratories, Lexington, Kentucky) (22). (B) Sequence of CTNNB1 in HCT116 and SW48 (27). The left panel (nucleotides 121 to 143 from HCT116) reveals the presence of a deletion in addition to the WT sequence. The middle panel (antisense strand of nucleotides 156 to 113 of the WT and deleted alleles of HCT116) reveals the 3-bp deletion (ΔTCT) that removed codon 45 in half the clones. The right panel (nucleotides 80 to 113 from SW48) reveals a C-to-A transition affecting codon 33 (TCT to TAT). (C) Schematic of β-catenin illustrating the Armadillo repeats (24) and negative regulatory domain. The residues in larger type fit the consensus sequence for GSK-3β phosphorylation (20) and those in bold have been demonstrated to affect down-regulation of β-catenin through GSK-3β phosphorylation in Xenopus embryos (17). The five mutations found in human colon cancers are indicated at the top. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

To test the generality of this mutational mechanism, we evaluated five primary colorectal cancers in which sequencing of the entire coding region of APC had revealed no mutations (13). Three of these five tumors were found to contain CTNNB1 mutations (S45F, S45F, and T44A) that altered potential GSK-3β phosphorylation sites (19) (Fig. 3C). Each mutation was somatic and appeared to affect only one of the two CTNNB1 alleles.

Because the β-catenin mutations were heterozygous, we hypothesized that the mutations might exert a dominant effect, rendering a fraction of cellular β-catenin insensitive to APC-mediated down-regulation. To test this notion, we performed gel shift analyses with nuclear extracts from untransfected HCT116 cells. In contrast to noncolorectal cancer cell lines with intact APC, HCT116 cells contained a β-catenin-Tcf complex that gel-shifted an optimized Tcf-binding oligonucleotide, and this complex supershifted with antibody to β-catenin (Fig. 4A). We also constructed β-catenin expression vectors and compared the biologic activity of the mutant β-catenin from HCT116 (β-Cat Δ45) and SW48 (β-Cat S33Y) with that of their WT counterpart. For these experiments, we used the 293 kidney epithelial cell line because it is highly transfectable, exhibits low endogenous CRT, and contains a large amount of endogenous APC (Fig. 2A). In the presence of endogenous APC, both mutant β-catenins were at least six times as active as the WT protein, and this activity was inhibited by dominant-negative hTcf-4 (Fig. 4B).

Fig. 4.

Functional evaluation of β-catenin mutants. (A) Constitutive nuclear complex of β-catenin and Tcf in HCT116 cells. The presence of nuclear β-catenin-Tcf complexes was assessed by gel shift assays (12). Lanes 1 to 3, the optimal Tcf retardation probe shifted with nuclear extract from HCT116 cells when no antibody was added (lane 1) or with addition of antibody to β-catenin (0.25 μg, lane 2) or an irrelevant antibody (0.25 μg, lane 3); lane 4, mutant Tcf retardation probe shifted with nuclear extract from HCT116 cells; n.s., nonspecific shifting seen with the mutant probe. (B) Effects of the β-catenin mutations on CRT. Human 293 cells were transfected with WT or mutant (Δ45, S33Y) β-catenin and CRT was assessed (28). CRT reporter activities are expressed relative to WT β-catenin and are the means of three replicates. Error bars represent standard deviations. DN Tcf refers to dominant-negative hTcf-4.

Together, these results indicate that disruption of APC-mediated regulation of CRT is critical for colorectal tumorigenesis. This is most commonly achieved by recessive inactivating mutations of both APC alleles but, as shown here, can also be achieved by dominant mutations of CTNNB1 that render CRT insensitive to the effects of WT APC. Our results suggest that APC inhibition of CRT requires phosphorylation of β-catenin at multiple sites. These potential phosphorylation sites are consistent with the known specificity of GSK-3β (20), a serine/threonine kinase that negatively regulates β-catenin in Xenopus and Drosophila cells (17) and that interacts with APC and β-catenin in mammalian cells (10). Our results also suggest a functional basis for the occasional CTNNB1 mutations observed in other tumor types (21) and illustrate how a critical pathway in human disease can be illuminated by the discovery of mutations in different components of the pathway. The next step in understanding APC function will be the identification of the genes that are activated by hTcf-4-β-catenin complexes and inhibited by WT APC. These genes are likely to be related to APC's ability to induce apoptosis in colorectal cancer cells (22).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
View Abstract

Navigate This Article