Molecular basis for the tissue-specificity of β-catenin oncogenesis

Abstract

Wnt-β-catenin-T Cell Factor signaling is causally linked to c-myc dependent tumorigenesis in mouse and human colon epithelial cells. By contrast, β-catenin is not similarly associated with oncogenic transformation of other tissues, including T cells. The molecular basis for tissue specificity of β-catenin dependent oncogenesis is unknown. Here we demonstrate that Adenomatous Polyposis Coli mutant APC^Min/+ mice, which have increased expression of β-catenin in all tissues, develop severe intestinal neoplasia, but fail to develop thymic lymphoma. Whereas β-catenin elicits a proliferative response from intestinal cells, thymocytes experience oncogene-induced-senescence (OIS), growth arrest and apoptosis. We demonstrate that the differential cellular response of thymocytes and intestinal epithelial cells is a direct consequence of the gene expression elicited by β-catenin expression in each tissue. We find that whereas intestinal cells induce genes that promote proliferation thymocytes induce expression of genes associated with OIS, growth arrest and p53-dependent apoptosis. We correlate gene expression pattern with the role β-catenin plays in the development of each tissue and suggest that susceptibility of transformation by β-catenin is intimately related to its function during development. We propose that when oncogenes are used as signaling molecules, safety nets in the form of OIS, growth arrest and apoptosis are in place to prevent accidental transformation.

Introduction

The canonical Wnt-β-catenin-T Cell Factor (TCF) signaling pathway plays an essential role in maintaining the intestine by providing proliferative signals to precursor stem cells¹. Impairment of this signaling pathway results in a severe loss of the proliferative progenitors of the prospective crypts in fetal and adult intestine^2–4. By contrast, aberrant activation of this pathway due to mutations in adenomatous polyposis coli (APC), a tumor suppressor gene that is inactivated in most colorectal cancers, induces hyperproliferation of the intestinal epithelium and results in cancer^5–7. Specifically, TCF4 and β-catenin target gene, c-Myc, is causative in promoting colon cancer^8,9, whereas target gene TCF1 behaves as tumor suppressor in this model¹⁰. As Wnt-signaling pathway is required for T cell development and function it might be expected to induce lymphomagenesis but few examples of this are currently reported. As a matter of fact, it has been previously noted that β-catenin-TCF1 signaling pathway may not be not causally linked with human T cell lymphomas has been noted^11–13. Furthermore, analysis of acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL) showed that these cancers were distinguishable based on the expression of β-catenin, with AML cells expressing higher levels compared to ALL cells¹⁴. In other examples of transformed human T cells noticeable β-catenin accumulation was correlated with reduced expression of TCF1 indicating that TCF1 and β-catenin dependent transcription may not be a major contributor to T cell cancer^15,16. This contrast in the ability of β-catenin to transform intestinal cells, but not T cells, is of particular interest because both cell types utilize Wnt-β-catenin-TCF signaling pathway for tissue renewal throughout the life of the organism¹⁷.

Whereas intestinal cells utilize β-catenin signals for proliferation in response to a Wnt gradient, β-catenin expression regulates T cell development^18–21. Specifically, we have shown that developmental signals transiently induce β-catenin expression in immature thymocytes, which is required to be turned down for proper T cell development^22–24. In a mouse model with transgenic expression of oncogenic β-catenin in thymocytes from the proximal Lck promoter, failure to turn down β-catenin expression results in developmental block²³. Developmentally blocked thymocytes induce genes that regulate OIS, growth arrest and p53-dependent apoptosis resulting in a hypocellular thymus. This cellular response also protects thymocytes from transformation, which is permitted when p53 function is removed as thymocytes expressing oncogenic β-catenin fail to undergo p53-dependent apoptosis²⁴. These observations show that developing intestinal cells and immature thymocytes exhibit distinct response to expression of β-catenin. However, the molecular basis for transformation of intestinal cells, but not thymocytes, in response to β-catenin expression remains unknown.

In this report we provide a molecular basis for the tissue-specificity of β-catenin-dependent oncogenesis. We demonstrate that the differential cellular response of thymocytes and intestinal epithelial cells is a direct consequence of the gene expression elicited by β-catenin expression in each tissue. We find that whereas Min intestinal cells induce genes that promote proliferation thymocytes induce expression of genes associated with OIS, growth arrest and p53-dependent apoptosis. We correlate gene expression pattern with the role β-catenin plays in the development of each tissue. Our data suggest that susceptibility of transformation by β-catenin is intimately related to its function during development. In light of these observations we propose that oncogenesis is regulated by the response of the cell type to expression of the oncogene, which in turn is congruent with the role of the oncogene as a signaling molecule in the development of the tissue.

Results

Min intestinal cells and Min thymocytes respond distinctly to enhanced β-catenin expression

To address directly the molecular basis for distinct response of the intestinal tissue and thymocytes we have utilized APC^Min/+ (Min) mouse model. Min mice bear a heterozygous germ line mutation at codon 850 of the APC gene and consequently over-express endogenous β-catenin in all tissues^25–27. Min mice develop adenomatous polyps²⁸ and Fig. 1a). By contrast, Min mice have small dramatically hypocellular thymuses (Fig. 1b, c). We found that relative abundance of β-catenin protein levels was increased in Min thymocytes and gut compared to control (Fig. 1d). However, the relative abundance of β-catenin protein levels was much higher in the gut compared to thymocytes from both control and Min mice (Fig. 1d). Accordingly, we found that expression levels of Wnt target genes in the control gut were approximately two-fold higher than in the thymus (Fig. 1e). We also found that both Min thymocytes and Min gut polyps showed increased expression of target genes such as Axin2, c-jun and c-fos (Fig. 1e) compared to control thymocytes and gut. These data demonstrate increased β-catenin dependent signaling in Min thymocytes but a failure to develop thymic lymphoma. Min mice are known to undergo loss of the wild-type Apc allele in intestinal epithelium²⁹. Therefore we assayed for the status of the Apc alleles in Min thymocytes and found that in the Min thymocytes one allele was wild type and one allele was mutant (Fig. 1f). To determine if the lack of thymic lymphoma correlated with changes in cell death or proliferative status we stained ex-vivo thymocytes with AnnexinV and studied BrdU uptake in-vivo. We found that Min thymocytes had increased apoptosis (Fig. 1g) and reduced proliferation (Fig. 1h). Thus, increased expression of β-catenin induces apoptosis and growth arrest in developing thymocytes, which results in small hypocellular thymuses in Min mice.

Figure 1. Min thymuses are hypocellular with enhanced apoptosis and impaired proliferation.

(a) Representative intestine sections from 4 to 5 months old wild-type (WT) control and Min mice photographed to show macroscopically visible polyps in the Min intestines compared with WT control intestines.

(b) Representative thymuses from 4 to 5 months old wild-type (WT) control and Min mice photographed to show size differences in Min thymuses compared with WT control thymus.

(c) Total numbers of thymocytes are shown for WT control and Min thymuses. Numbers are shown as mean ± SEM. *** P ≤ 0.0005; Student’s t-test. n = 24 for control and Min thymus.

(d) Purified DN thymocytes, normal intestine and intestinal adenomas from wild-type (WT) control and Min mice were analyzed for the abundance of β-catenin protein by Western Blot. Representative blots (top) and densitometric analysis (bottom) of three independent experiments is shown.

(e) Real time RT PCR analysis of the abundance of mRNA of Wnt-signaling target genes presented relative to the abundance of GAPDH in DN4 thymocytes and intestinal tissues from WT control and Min. Data are pooled from two independent experiments with a total of four to six independent samples (error bars, s.e.m.).

(f) Allele-specific APC genomic PCR analysis of the normal wild-type and mutant Min APC alleles in tail DNA and thymocyte DNA from WT control and Min mice.

(g) Flow-cytometric analysis of Annexin V staining in control and Min thymocytes. Numbers indicate percentage of Annexin V⁺ cells. Data are representative of three independent experiments with a total of six mice from each category (error bars, s.e.m.). ** P ≤ 0.005; Student’s t-test.

(h) In vivo BrdU incorporation analysis of control and Min thymocytes. Numbers indicate percentage of BrdU⁺ cells. Data are representative of two independent experiments with a total of four mice from each category (error bars, s.e.m.). * P ≤ 0.05; Student’s t-test.

Thymocyte development is stalled in Min mice

Thymocyte development is defined based on the expression of cell surface markers CD4 and CD8, with the most immature thymocytes lacking these markers. The immature CD4- and CD8-double negative (DN) thymocyte population can be further dissected into developing subsets using cell surface expression of c-kit and CD25 (ETP:c-kit+CD25−; DN2:c-kit+CD25+; DN3:c-kit-CD25+; DN4/pre-DP: c-kit-CD25−)³⁰. We found that Min thymocytes were blocked at the DN4/pre-DP stage of development (Fig 2a–c). DN4/pre-DP Min thymocytes expressed intracellular T cell receptor β chain (TCRβ_IC), which identifies developmentally competent DN4 thymocytes (Fig. 2d). TCRα chain is expressed in DP thymocytes following TCRβ chain expression. We found that Min-pre-DP thymocytes did not express cell-surface TCRα (Fig. 2e). The lack of TCRα chain expression confirmed that these stalled thymocytes were attenuated in development at the pre-DP stage. Molecular features of developmentally stalled Min pre-DP thymocytes showed a higher abundance of Egr 1, 2, and 3, which indicates that these thymocytes have received pre-T cell receptor signals (Fig. 2f). However, further development is blocked at the DN4/pre-DP stage as the stalled thymocytes express low levels of RORγt compared with control DN4 cells (Fig. 2f). These data demonstrate that stalled thymocytes had failed to extinguish pre-TCR signals (represented by higher expression of Egr proteins) and to induce transcription factor RORγt, which is required for transition to the DP stage, which is consistent with previously described block at the DN4/pre-DP stage upon thymocyte-specific loss of the APC gene³¹. Thus, failure to down-regulate expression of β-catenin past the pre-DP/DN4 stage blocks thymocyte development to the DP stage.

Figure 2. Min thymocytes are blocked in T cell development at pre-DP stage and impaired in maturation to the DP stage.

(a) Schematic of cell-surface marker expression during T-cell developmental stages in thymus.

(b) Flow cytometric analysis of WT control and Min thymocytes stained with antibodies to CD4 and CD8 (left panel) and numbers (mean ± SEM) of thymocytes for each subset. Data are representative of ten experiments with at least two mice from each category per experiment.

(c) WT control and Min thymocytes were stained with lineage specific antibody cocktail (CD4, CD8, TCRβ, TCRγδ, B220, Gr1, Mac1 and NK1.1), antibodies to CD25 and c-kit and analyzed by flow cytometry. Distribution (left panel) and numbers (right panel) of electronically gated lineage negative thymocyte subsets: ETP (CD25⁻c-Kit⁺), DN2 (CD25⁺c-Kit+), DN3 (CD25⁺c-Kit⁻) and DN4 (CD25⁻c-Kit⁻) are shown. Data are representative of ten experiments with at least two mice from each category per experiment.

(d) Flow cytometry of intracellular TCRβ expression in electronically gated control and Min DN4 thymocytes representated as histogram overlays with isotype control antibody shown in black. Data are representative of six independent experiments.

(e) Flow cytometric analysis of surface TCRα expression in electronically gated DN (CD4−CD8−NK1.1−) (upper panel) and mature (CD4+CD8+NK1.1+TCRβ+) (lower panel) thymocytes of control and Min total thymocytes stained with a combination of antibodies to a collection of TCR Vα proteins along with antibodies to CD4, CD8, NK1.1 and TCRβ. Numbers in the plots indicate percentage of TCRα+ cells. Data are representative of six independent experiments.

(f) Real-time RT PCR analysis of the abundance of Egr1, Egr2, Egr3 and Rorγt mRNA in sorted DN4 thymocytes from control and Min thymuses; results are presented relative to the abundance of GAPDH. * P ≤ 0.05; *** P ≤ 0.0005; Student’s t-test. Data are pooled from two to three independent experiments with a total of four to eight independent samples (error bars, s.e.m.).

To confirm that the developmental block resulted from cell-intrinsic over-expression of β-catenin we studied T cell development in mice that express transgenic β-catenin from the Lck proximal promoter in developing thymocytes (CAT-Tg mice)³². CAT-Tg mice show an increase in DN cells, which are blocked at DN4/pre-DP stage (Supplementary Fig. 1), and have higher expression of Egr 1, 2, and 3, and lower expression of RORγt (Supplementary Fig. 2). These data shows that Min thymocytes share several features with CAT-Tg thymocytes and thereby demonstrate that thymic hypocellularity in Min mice was not a result of the intestinal tumors as intestinal cells do not over-express β-catenin in CAT-Tg mice and the intestine is healthy. Finally, we note that even in Min mice adenoma count was comparable in mice with normal or low thymocyte count (60 ±3 and 68 ±4 respectively). These data confirm a lack of any connection between adenoma and thymic cellularity (Supplementary Fig. 3). Together this data demonstrates that cell intrinsic expression of β-catenin leads to a developmental block in T cell development in Min mice.

Differential gene expression in Min hypocellular thymus and intestinal cells is the basis for tissue-specificity of β-catenin oncogenesis

Next we addressed the molecular basis for tissue-specific oncogenic function of β-catenin by studying gene expression in thymocytes and intestinal epithelial cells obtained from control and Min mice. We have previously reported that cell intrinsic expression of β-catenin in DN4-pre-DP thymocytes induces expression of gene associated with growth arrest, OIS and apoptosis²⁴. Stalled Min pre-DP/DN4 thymocytes induced expression of these genes (Fig. 3a), including genes associated with p53-dependent apoptosis (Fig. 3b). To confirm that induction of these genes was in response to increased expression of β-catenin we analyzed thymocytes from CAT-Tg mice. We found that CAT-Tg DN4 thymocytes also showed increased expression of genes linked with growth arrest, OIS and apoptosis (Supplementary Fig. 4). These results demonstrate that β-catenin over-expression in thymocytes (driven by mutation in the APC gene or as β-catenin transgene) induces genes associated with growth arrest, OIS and apoptosis. In striking contrast, genes associated with growth arrest, OIS and apoptosis were not induced, instead were sometimes suppressed, in the intestinal adenomatous polyps (Fig. 3c and 3d), which is consistent with the proliferative response elicited by increased expression of β-catenin in these cells. These data demonstrate that the molecular response of thymocytes and intestinal epithelial cells to over-expression of oncogenic β-catenin is very disparate and matches the distinct cellular responses.

Figure 3. Differential gene expression in Min hypo-cellular thymus and intestinal adenomas.

(a) Real time RT PCR analysis of the abundance of mRNA of growth arrest and cellular senescence associated genes in sorted WT control and Min DN4 thymocytes.

(b) Real time RT PCR analysis of the abundance of mRNA of p53-mediated apoptosis-associated genes in sorted WT control and Min DN4 thymocytes.

(c) Real time RT PCR analysis of the abundance of mRNA of growth arrest and cellular senescence associated genes in normal intestine and intestinal adenomas.

(d) Real time RT PCR analysis of the abundance of mRNA of p53-mediated apoptosis-associated genes in normal intestine and intestinal adenomas.

All results are presented relative to the abundance of GAPDH. * P ≤ 0.05; ** P ≤ 0.005; *** P ≤ 0.0005; ns = not significant; Student’s t-test. Data are pooled from two to three independent experiments with a total of four to eight independent samples (error bars, s.e.m.).

p53-dependent apoptosis protects Min mice from thymic lymphoma

The induction of growth arrest, OIS and apoptosis genes in response to β-catenin expression in thymocytes is very likely the reason for the absence of β-catenin dependent thymic lymphomas. Growth arrest, OIS and apoptosis have all been shown to be p53-dependent under various conditions³³. To directly test the role of p53 in Min mice we generated p53-deficient Min mice. We found (data not shown), as has been previously reported³⁴, that rate of adenoma formation and progression to malignancy in p53^−/−Min mice was similar to Min mice. In sharp contrast, absence of p53 permitted the generation of thymic lymphoma with immature pre-DP phenotype in p53^−/−Min mice (Fig. 4). Thymocytes from control, p53^−/−Min and p53^−/−CAT-Tg mice were compared for cell size and expression of developmental markers, CD4 and CD8. Control thymocytes showed the typical butterfly pattern of CD4 and/or CD8 expression (Fig. 4a) and were small in size (Fig. 4b), as thymocytes have previously been shown to be in G1/G0 stage of cell cycle. By contrast, Min thymocytes and p53^−/−Min thymocytes were mostly CD4–CD8- (Fig. 4a) and were large in size (Fig. 4b), which indicates a change in cell cycle. The phenotype of this lymphoma was similar to thymic lymphoma in p53^−/−CAT-Tg mice (Fig. 4a and b, right panel; ²⁴. p53^−/−Min mice displayed enlarged thymus, spleen, and lymph nodes (data not shown). Enlarged peripheral lymphoid tissues in p53^−/−Min mice had largely DN lymphoma cells (Fig. 4c), which also demonstrates the metastatic potential of the transformed thymocytes. p53^−/−Min DN thymocytes showed increased expression of OIS associated genes like p15 and Dec1 and other cell cycle inhibitors (Fig. 4d). However, we note that expression of the cell cycle inhibitor p21 was decreased in p53^−/−Min DN lymphoma cells, which correlates with increased proliferation (Fig. 4d). The number of thymocytes was also significantly increased in p53^−/−Min compared to extremely hypocellular Min thymus (Fig. 4e), which indicates thymocyte proliferation. The increased cell death observed in Min DN4 thymocytes compared to control DN4 thymocytes was abrogated in p53^−/−Min DN4 thymocytes (Fig. 4f). These data are congruent with our previous report showing p53 independentβ-catenin-induced OIS and growth arrest and p53 dependent apoptosis in p53^−/−CAT-Tg thymic lymphoma²⁴. These data show that increased expression of β-catenin (endogenous or enforced) promotes thymic lymphoma when p53 function is removed. Further analysis showed that similar to Min thymocytes thymic lymphoma cells in p53^−/−Min mice have increased β-catenin protein expression (Fig. 4g) and increased expression of target genes (Fig. 4h). Analysis of the status of the Apc alleles in p53^−/−Min thymic lymphoma showed the presence of both wild type and mutant the Apc alleles (Fig. 4i). Together these data show that p53-dependent apoptosis plays a critical role in protecting thymocytes, but not intestinal cells, from β-catenin dependent transformation.

Figure 4. β-Catenin-triggered DN lymphomas develop in p53^−/−Min mice.

(a–b) Representative dot plots showing (a) cell surface staining with antibodies to CD4 and CD8 and (b) forward scatter (FSC) and side scatter (SSC) in total thymocytes of WT control (n=6), Min (n=6), p53^−/−Min (n=6), and p53^−/−CAT-Tg (n=4). Numbers indicate percentage of cells in the quadrants. The strains of mice are listed above each dot plot.

(c) Lymph node cells from WT control, Min and thymic lymphoma bearing p53^−/−Min mice were stained with antibodies to CD4, CD8 and TCRβ and analyzed by flow cytometry. Representative dot plots showing CD4 and CD8 expression on gated TCR⁺ lymph node cells are shown.

(d) Real time RT PCR analysis of the abundance of mRNA of growth arrest and cellular senescence associated genes in sorted DN4 thymocytes from WT control, Min, pre-lymphoma p53^−/−Min and p53^−/−Min mice with thymic lymphoma.

(e) Total numbers of thymocytes are shown for Min and p53^−/−Min thymuses. Numbers are shown as mean ± SEM. *** P ≤ 0.0005; Student’s t-test. n = 8 for Min and p53^−/−Min thymus.

(f) Flow-cytometric analysis of Annexin V staining in control, Min and p53^−/−Min DN4 thymocytes. Numbers indicate percentage of Annexin V⁺ cells. Data are representative of two independent experiments with a total of four mice from each category (error bars, s.e.m.). *** P ≤ 0.0005; Student’s t-test.

(g) A Western blot showing abundance of β-catenin protein in purified DN thymocytes from wild-type (WT) control and p53^−/−Min thymic lymphoma (representative of three independent experiments).

(h) Real time RT PCR analysis of the abundance of mRNA of Wnt-signaling target genes presented relative to the abundance of GAPDH in p53^−/− and p53^−/−Min thymic lymphoma. Data are pooled from two independent experiments with a total of four independent samples (error bars, s.e.m.).

(i) Allele-specific APC genomic PCR analysis of the normal wild-type and mutant Min APC alleles in tail DNA and thymocyte DNA from p53^−/− and p53^−/−Min thymic lymphoma bearing mice.

Molecular basis for the formation of intestinal polyp and thymic lymphoma are distinct

Gene expression analysis shows that intestinal polyps express high levels of c-myc (Fig. 5a), which along with other signals that promote proliferation and polyp formation. By contrast immature Min thymocytes do not induce expression of c-myc in response to β-catenin expression even in the absence of p53 when they develop into lymphoma (Fig. 5b). Because Notch signaling has been implicated in thymic lymphomagenesis we assayed for the expression of Notch target genes, Dtx1 and Hes1. Neither gene was up regulated in p53-deficient Min lymphoma cells (Fig. 5b). This data suggests that β-catenin-dependent lymphoma in p53-deficient immature thymocytes was not due to c-myc or Notch-dependent signals. We conclude that in addition to initial response to β-catenin expression being tissue specific, the molecular basis for the oncogenic function of β-catenin is distinct in each cell type.

Figure 5. DN Tumors in p53^−/−Min mice are not c-Myc or Notch dependent.

(a) Real-time RT PCR analysis of the abundance of c-Myc mRNA in normal intestine and Min intestinal adenomas.

(b) Real-time RT PCR analysis of the abundance of c-Myc, Dtx1 and Hes1 Id2 mRNA in DN4 thymocytes purified from WT control and p53−/− Min DN tumor thymus; results are presented relative to the abundance of GAPDH. ** P ≤ 0.005; *** P ≤ 0.0005; Student’s t-test. Data are pooled from two independent experiments with a total of four to eight independent samples (error bars, s.e.m.).

Discussion

In this paper we demonstrate that enforced expression of oncogene β-catenin elicits distinct cellular response from intestinal epithelial cells and thymocytes, which is the consequence of differential gene expression pattern induced by β-catenin expression. In intestinal cells oncogenic β-catenin induces expression of c-myc among other targets and promotes proliferation and transformation. By contrast in thymocytes β-catenin expression induces expression of genes associated with OIS, growth arrest and p53-dependent apoptosis. We note that these responses are congruent with the role played by β-catenin in the development of each tissue. In light of these data we propose that oncogenesis is regulated by the molecular and cellular response of the tissue in which it is expressed, which in turn is congruent with the signaling role played by the oncogene in the development and maintenance of that tissue.

The distinct tissue-specific response to a ubiquitously expressed oncogene provides a plausible explanation for the tissue-specificity of oncogenesis. In this report we provide evidence that β-catenin expression in intestinal tissue induces gene expression congruent with proliferative response whereas in thymocytes genes associated with OIS, growth arrest and apoptosis are induced. A previous report showed that loss of Apc in renal epithelium led to accumulation of β-catenin and rapid development of renal carcinoma which is further accelerated by p53 deficiency³⁵. Interestingly, a subsequent study from the same group showed that Apc loss from the renal and intestinal epithelium resulted in OIS in renal cells but not in intestinal cells, with a double loss of p21 and Apc resulting in rapid renal carcinoma³⁶. Tissue-specific regulation of p19^Arf has been shown to dictate response to oncogenic K-ras³⁷. These studies are congruent with the notion that protective mechanisms including OIS and p21-dependent growth arrest provide built-in protective mechanisms in tissues that use oncogenes as signaling agents. In addition to differential gene expression to the initial expression of β-catenin, molecular basis for transformation is also distinct in different tissues. In contrast to intestinal cells, p53-deficient immature thymocytes did not respond to β-catenin expression by inducing expression of c-myc. As a matter of fact, expression of Dtx1 and Hes1, which are associated with thymic lymphoma, was also not induced. This data suggests that neither c-myc nor Notch signaling was induced in stalled p53-deficient Min thymocytes. Differential function of c-myc in β-catenin dependent transformation of different tissues has been reported previously³⁸. In light of the above observations we propose that in addition to initiation of transformation, which results from tissue-specific response to oncogene expression, molecular basis of transformation by an oncogene is also tissue-specific.

We note that lack of a causal connection between oncogene β-catenin and T cell lymphoma is not due to paucity of research in this area. The search for a role for the Wnt-β-catenin-TCF1 pathway in T cell lymphoma began with the original observation from the Clevers laboratory that TCF1 was expressed in T cells but not in B cells³⁹. This report was followed by observations that β-catenin was highly expressed in some human T-ALL cell lines such as Jurkat^40,41, but TCF1-dependent transcription activation could not be demonstrated in Jurkat cells^42,43. More recent reports have provided a mixed picture of β-catenin expression in human ALL cells, with a fraction of cells also showing Exon3 mutations that might result in β-catenin accumulation⁴⁴. However, it remains to be demonstrated that accumulated β-catenin leads to increased function of TCF1-dependent gene expression, which causally result in thymic lymphoma. This is not surprising in light of the observation that several molecular manipulations that lead increased β-catenin accumulation in developing thymocytes result in a developmental block and thymic atrophy^{20,23,24,31,45,46}. We propose that it is the induction of OIS, growth arrest and apoptosis in developing thymocytes in response to β-catenin that results in the absence of causal connection between T cell transformation and β-catenin over-expression. In addition we suggest that constitutive and high level of TCF1 expression in thymocytes might function as a built in tumor suppressor and protect from lymphomagenesis.

These data support a model in which, during normal intestinal renewal, intestinal precursor cells induce β-catenin in response to a Wnt gradient to promote proliferation^8,47. As the developing cells migrate away from the crypt Wnt gradient diminishes leading to decrease in proliferation, which allows for controlled expression of β-catenin to regulate cellularity. However, in the intestine abnormal induction of β-catenin due to mutations in the genes of the Wnt-β-catenin-TCF4 pathway results in uncontrolled proliferation ⁸ and (Supplementary Fig. 5a). By contrast, in thymocytes β-catenin is induced transiently and provides signals for maturation and then thymocytes must turn the protein expression down for further development (Supplementary Fig. 5b). Failure to turn down expression of β-catenin during maturation leads to a developmental block in thymocytes at the pre-DP stage^22,23. The developmentally stalled thymocytes experience additional events, become pre-cancerous and experience OIS. We have also found that stalled cells are growth arrested and are then removed from the developmental pool by p53-dependent apoptosis thereby protecting developing T cells from β-catenin dependent oncogenesis²⁴. Importantly, this protective response is strictly restricted to the immature thymocytes that utilize β-catenin for maturational signals. Over-expression of β-catenin using CD4-Cre dependent deletion of exon 3 that induces β-catenin stabilization in more mature CD4+ CD8+ double positive (DP) thymocytes supported c-myc dependent oncogenesis⁴⁸. Because DP thymocytes normally do not express β-catenin, we propose that ectopic expression of β-catenin induces c-myc and potentially other targets genes in DP thymocytes, which then supports transformation as there are no inherent mechanisms of protection. Alternately, β-catenin protein level may regulate extent of Wnt-related signaling and c-myc expression as dose of Wnt-signaling has been correlated with gene expression⁴⁹. In light of these observations we suggest that oncogenic function of β-catenin is orchestrated by the tissue-specific response of the cells and intimately related to the role it plays in the development of the tissue. We propose that when oncogenes are used to provide developmental signals ‘safety nets’ comprising of OIS, growth arrest and apoptosis are put in place to prevent oncogenesis.

In conclusion, tissue specificity of oncogenesis is defined by tumors being confined to a few tissues despite expression of oncogenes in multiple tissues. Data provided in this paper provides a novel perspective on the molecular basis for the difference in the response of intestinal epithelial cells and thymocytes to β-catenin expression and the disparate outcomes. Based on the data presented in this paper we propose that the oncogenic function of β-catenin is intimately related to the role it plays in the development of the particular tissue. Finally, as inhibitors of β-catenin pathway^50–52, and inducers of p53 function^53,54 have been defined and are available for use in clinical situations these observations provide new opportunities for designing tissue specific therapeutic strategies for treating the human cancers.

Materials and methods

Mice

APC^Min/+ (Min) mice and p53^−/− mice are on a C57BL/6 genetic background and were purchased from The Jackson Laboratory. The generation of CAT-Tg and p53^−/−CAT-Tg mice has been described previously^24,32. p53^−/−Min were generated by crossing p53^−/− mice with Min mice. Age-matched littermate control mice or C57BL/6 mice were used in all experiments. All mice were bred and maintained in the animal facility of the National Institute on Aging according to regulations of the National Institutes of Health and were in compliance with the guidelines of the animal resources facility of the National Institute on Aging, which operates under the regulatory requirements of the US Department of Agriculture and Association for Assessment and Accreditation of Laboratory Animal Care.

Western blot analysis

Immunoblotting was performed with whole-cell lysates resolved on 4–12% Bis-Tris SDS-PAGE, transferred to nitrocellulose membrane and detected using Abs against mouse β-catenin (14/Beta-Catenin; 610154; BD Pharmingen) or rabbit-anti-PKC-μ (C-20; sc-639; Santa Cruz) followed by horseradish peroxidase-conjugated anti–mouse IgG (sc-2005) or anti–rabbit IgG (sc-2004; both from Santa Cruz). Reactivity was visualized by enhanced chemiluminescence.

Allele-specific APC genomic PCR analysis

The Min mutation was identified by allele-specific PCR done as per Jackson laboratory protocol. Briefly, the mutant Min primer (5′-TTCTGAGAAAGACAGAAGTTA -3′), whose 3′-most nucleotide is complementary to the Min nonsense mutation at nucleotide 2549 of Apc, and the primer MAPC 15 (5′-TTCCACTTTGGCATAAGGC-3′) is used in PCR. A PCR product of ~340 bp is generated in the presence of the Min allele, but no product is generated in its absence. As an internal positive control for the PCR, a third primer, MAPC 9 (5′-GCCATCCCTTCACGTTAG-3′) was included in the reaction. MAPC 9 and MAPC 15 together generate a 618 bp product surrounding the Min nonsense mutation detected in wildtype and APC^Min/+.

Antibodies and flow cytometry

Cells were collected, stained and analyzed on a FACSCalibur (Becton Dickinson). Dead cells were excluded by assessing forward light scatter data. All the data were acquired and are presented on a log scale. The following antibodies conjugated to FITC, PE, peridinin chlorophyll protein-cyanine 5.5, allophycocyanin or biotin (all from BD Pharmingen) were used for staining: anti-CD4 (GK1.5), anti-CD8α (53-6.7), anti-CD25 (PC61), anti-c-Kit (2B8), anti-TCRβ (H57-597), anti-TCRγδ (GL3), anti-B220 (RA3-6B2), anti-Gr-1 (RB6-8C5), anti-Mac-1 (M1/70), anti-NK1.1 (PK136), anti-Ter119 (Ter119), anti-TCRVα₂ (B20.1), TCRVα_3.2 (RR3–16)_, TCRVα_8.3 (B21.14) and TCRVα_11.2 (RR8-1). Lineage specific cocktail contained antibodies to CD4, CD8, TCRβ, TCRγδ, B220, NK1.1, Gr1, and Mac-1. For DN purification, cell suspensions of thymocytes were stained with biotinylated anti-CD4 and anti-CD8, followed by incubation with anti-biotin microbeads (Miltenyi Biotec) and magnetic depletion of DP and SP thymocytes. DN enriched cell suspensions were surface-stained with lineage Ab mixture, anti-CD25 and anti-c-kit, followed by sorting using MoFlo. To assess apoptosis, freshly isolated thymocytes were first stained with surface markers, followed by staining with FITC-conjugated annexin V (Roche) according to the manufacturer’s instructions.

In vivo BrdU Incorporation Assay

BrdU flow kit (BD Pharmingen) was used for proliferation assay. Briefly, 100 μl of 10 mg/ml of the BrdU solution was injected intraperitoneally. BrdU incorporation was assayed 2 h post-injection. Thymocytes were first stained with surface markers and followed by intracellular staining with anti-BrdU antibody conjugate.

Isolation of intestinal adenomas

WT control and Min mice were sacrificed at the age of 4 months and whole intestinal rolls were removed; the lumenal contents were flushed out with PBS, the tissue was slit open longitudinally and laid flat mucosal surface up. Adenomatous polps were counted and scored visually, then intestine was chopped into ~5 cm or smaller pieces and the fragments were washed three times in cold PBS. Polyps were carefully dissected to exclude non-adenomatous tissue and homogenized in lysis buffer (Qiagen).

Real-Time RT-PCR

mRNA was extracted from sorted DN4 thymocytes and intestinal tissue using RNeasy Micro Kit (Qiagen). Total mRNA was reversed transcribed into cDNA with poly(dT) and Superscript III reverse transcriptase (Invitrogen). SYBR green quantitative real-time RT-PCR was performed, using PCR Master Mix (Applied Biosystems), for 40 cycles with annealing and extension at 60 °C (primer sequences, Supplementary Table 1). The expression of target gene was determined relative to GAPDH and calculated as 2^{−(CtTarget gene - CtGAPDH)}.