Proliferation, but not apoptosis, is associated with distinct beta-catenin expression patterns in non-small-cell lung carcinomas: relationship with adenomatous polyposis coli and G(1)-to S-phase cell-cycle regulators

Abstract

beta-catenin (beta-cat) is a versatile component of homotypic cell adhesion and signaling. Its subcellular localization and cytoplasmic levels are tightly regulated by the adenomatous polyposis coli (APC) protein. Mutations in beta-cat (exon 3) or APC (MCR) result in beta-cat aberrant overexpression that is associated with its nuclear accumulation and improper gene activation. Data from experimental models have shown that beta-cat overexpression has a multitude of effects on cell-cycle behavior. In many of these aspects its function depends on major G(1) phase regulators. To the best of our knowledge, most of these issues have never been addressed concurrently in tumors. For this reason we investigated in a panel of 92 non-small-cell lung carcinomas, beta-cat and APC expression, and their relationship with cell-cycle kinetics (PI and AI) and ploidy status. Moreover, the above correlations were examined in relation to the main G(1)/S-phase checkpoint regulators. Four beta-cat immunohistochemical expression patterns [membranous (11.1%), membranous-cytoplasmic (54.3%), cytoplasmic (9.9%), cytoplasmic-nuclear (24.7%)] and three APC immunohistochemical expression patterns [cytoplasmic (37.7%), cytoplasmic-nuclear (58%), nuclear (4.3%)] were observed, which were further confirmed by Western blot analysis on subcellular fractions in representative samples. The frequent presence of beta-cat in the cytoplasm is an indication of aberrant expression, whereas membranous and nuclear localization were inversely related. Absence of mutations in beta-cat (exon 3) and APC (MCR) suggest that beta-cat destruction mechanisms may be functional. However, expression analysis revealed attenuated levels for APC, indicating a residual ability to degrade beta-cat. Decreased levels were associated with loss of heterozygosity at the APC region in 24% of the cases suggesting that additional silencing mechanisms may be involved. Interestingly, the 90-kd APC isoform associated with apoptosis, was found to be the predominant isoform in normal and cancerous lung tissues. The most important finding in our study, was the correlation of nuclear beta-cat immunohistochemical localization with increased proliferation, overexpression of E2F1 and MDM2, aberrant p53, and low expression of p27(KIP), providing for the first time in vivo evidence that beta-cat-associated proliferation correlates with release of E2F1 activity and loss of p53- and p27(KIP)-dependent cell-cycle checkpoints. Loss of these checkpoints is accompanied by low levels of APC, which possibly reflects a diminished ability to degrade beta-cat. Taken together our data indicate that cases with nuclear beta-cat immunohistochemical expression represent a subset of non-small-cell lung carcinomas that have gained an increased proliferation advantage in contrast to the other beta-cat immunohistochemical expression profiles.

Figure 1.

β-cat analysis in NSCLC. A: Representative IHC results. Streptavidin-biotin-peroxidase technique with E-5 anti-β-cat antibody (DAB as chromogen) and hematoxylin counterstain (see Materials and Methods). a: Linear membranous (arrows) immunodetection (M) of β-cat-positive tumor cells in squamous carcinoma (case 20). b: Cytoplasmic and membranous immunoreactivity (MC) of β-cat-positive tumor in squamous carcinoma (case 18). c: β-cat-positive tumor cells exhibiting homogenous cytoplasmic staining in squamous carcinoma (C) (case 52. d: Cytoplasmic and nuclear localization (CN) of β-cat in positive tumor cells in adenocarcinoma (case 60). e: β-cat-positive tumor cells exhibiting cytoplasmic (arrowhead) and membranous (arrow) immunoreactivity (MC) in adenocarcinoma (case 25). f: Normal bronchial epithelium with membranous and cytoplasmic β-cat immunopositivity is shown. B: Western blot analysis of β-cat subcellular distribution. Nuclear (Nuc) and membranous/cytoplasmic (M/C) fraction analysis for β-cat presence in representative samples exhibiting M (case 1), MC (case 56), and CN (case 22) β-cat IHC patterns. β-cat presence was more abundant in the corresponding fractions for each case, respectively, confirming the initial IHC expression results. The increased sensitivity of Western blot combined with an enhanced chemiluminescent system (see Materials and Methods) also showed the presence of residual levels of β-cat in the opposite fractions, that could be because of the contaminating presence of normal tissue and/or the CN shuttling of β-cat. Original magnifications in A: ×400 (a, c, e, f); ×100 (b, d).

Figure 2.

APC analysis in NSCLC. A: Representative IHC results. Streptavidin-biotin-peroxidase technique with C-20 anti-APC antibody (DAB as chromogen) and hematoxylin counterstain (see Materials and Methods). a: Cytoplasmic immunodetection (C) of APC-positive tumor cells (arrows) (case 58). b: Cytoplasmic (arrowheads) and nuclear (arrows) immunoreactivity (CN) of APC-positive tumor cells (case 56). c: Nuclear staining (N) of APC-positive tumor cells (arrows) (case 71). B: Protein analysis. Results of the Western blotting analysis showing decreased APC protein levels in tumor areas (lanes 2 and 4) in comparison to corresponding normal tissue (lanes 1 and 3). Faint bands for APC levels were detected in tumor lanes, but because of image editing they were not accurately reproduced. Equal loading of total protein extracts was verified by actin blotting (second lane). C: Western blot analysis of APC subcellular distribution. Nuclear (Nuc) and membranous/cytoplasmic (M/C) fractions from representative samples exhibiting C (case 10), CN (case 22), and N (case 49) APC IHC patterns, respectively, were examined for its expression. APC was more abundant in the corresponding fractions of each case confirming the initial IHC results. The increased sensitivity of Western blot combined with an enhanced chemiluminescent system (see Materials and Methods) revealed the presence of residual APC levels in the opposite fractions, probably because of contamination from normal tissue and/or the nucleocytoplasmic shuttling ability of APC. Because of image editing APC residual levels were not accurately reproduced. D: LOH analysis. Loss of one of the two constitutive alleles in tumor areas (arrows) in comparison to their normal counterparts. Original magnifications in A, ×400.

Figure 3.

Diagrams (box plots) demonstrating the correlation between β-cat IHC patterns and PI (a) and AI (b). a: The main difference in proliferation is between cases with nuclear localization versus cases with membranous presence of β-cat (M and MC patterns, P = 0.005 and P = 0.002, respectively, Bonferoni analysis; Table 4 ▶ ). b: There were no significant differences in the apoptotic rate between all β-cat profiles. Similar results were obtained for GI (data not shown).

Figure 4.

A hypothetical model summarizing β-cat association with proliferation in relation with G₁ phase cell-cycle regulators in NSCLCs, based on the literature and our data (present and previous work): A: Data from experimental systems have shown that β-cat cellular pools under normal conditions are involved in homotypic cell adhesion (M). Activation of Wnt signaling induces a nuclear accumulation (N), in which β-cat functions as a transcription factor inducing among others, cellular proliferation. This β-cat-associated increase in proliferation is tightly controlled by the following mechanisms: 1) nuclear sequestration and cytoplasmic degradation of β-cat by APC; 2) activation of p53, that accumulates in response to β-cat signaling and triggers a cell growth arrest; and 3) p27^KIP-dependent cell-cycle arrest that results from the interplay between activation of nuclear β-cat transcriptional targets (eg, myc 76 ) and E-cad. Note: Depending on the histological context the exact relationships between β-cat expression status, cell-cycle kinetics, and expression of G₁ cell-cycle regulators may vary. B: In the present analysis, the group of cases with β-cat membranous localization (Mem) was related with intact p53/MDM2 loop, high levels of p27^KIP expression, and decreased APC levels. This group of cases demonstrated a lower PI compared to the β-cat nuclear-associated group (Nu). Taking into consideration our findings and the data in the literature we suggest that the following mechanisms may apply in this group. Functional p53 and p27^KIP checkpoints should prevent β-cat oncogenic effects by eliciting an anti-proliferative response, as previously predicted. Despite decreased APC levels, the low cell-cycle turnover may prevent saturation of degradation mechanisms thus providing sufficient β-cat down-regulation response. Also membranous association may reflect intact E-cad sequestration of β-cat, that further adds to proliferation control. C: In the present analysis, the group of cases with β-cat nuclear localization (Nu), were associated with defected p53 (either via p53 mutations or MDM2 overexpression 53,54 ), decreased p27^KIP expression levels, increased E2F1 expression levels, and decreased APC levels. In this group of cases the above correlations are related with higher PI. Taking into consideration our findings and the data in the literature we suggest that the following mechanisms may apply in this group. Loss of p53 and p27^KIP checkpoints should release β-cat from cellular growth restrains. Its nuclear accumulation leads to aberrant transcriptional activation (eg, activation of c-myc, cyclin D) and consequently an increase in proliferation. Low levels of APC and increased cell-cycle proliferation probably result in a saturation, and therefore render inefficient, the β-cat down-regulating mechanisms. Furthermore, E-cad might also be inactivated, as previously suggested. It is either not expressed or other factors (eg, γ-cat or activated growth-related tyrosine kinase receptors) compete for binding.