| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
From the Department of Pathology, University of ErlangenNürnberg, Erlangen, Germany
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
80% of the sporadic tumors.1
They
lead to an increase of the cytoplasmatic free pool of ß-catenin,
which is a component of the WNT/wingless pathway.2,3
In
an E-cadherin-bound, cell
membrane-associated fraction, ß-catenin participates in cell-cell
adhesion. The oncogenic potential of ß-catenin is derived from its
nuclear pool that associates with members of the TCF (T cell factor)
family of transcription factors. The resulting transcription complex
was shown to activate the genes of c-myc and cyclin
D1 associated with proliferation.4,5 Later, the matrix metalloproteinase-7, an important molecule for tissue remodeling,6,7 and the mesenchymal marker protein fibronectin8 were found to be target genes of the ß-catenin transcriptional activity. Moreover, a spatial predominance of nuclear ß-catenin expression was found at the invasion front.9 Most recent data indicate that heterogeneous intensities and spatial distributions of nuclear ß-catenin expression do not correlate with the rate of proliferation, but with tumor size10 and the grade of dysplasia11 in adenomas as well as with the invasiveness in adenocarcinomas.9
These observations prompted us to evaluate the spatial patterning and nuclear expression of ß-catenin in colonic carcinogenesis. Patterning or pattern formation is defined as a spatiotemporal process by which ordered arrangements of cells and tissue structure are attained. The term is mostly applied to embryonic cells and the compartimentation and morphogenesis in developmental biology, but it is useful for neoplastic cells and the neomorphogenesis in tumor biology, too. Thus growth and invasiveness in colonic neoplasms can be considered as a patterning of neoplastic tubules. Growth will be achieved by a new branching and extension of tubules in addition to cellular proliferation. Invasion requires an ingression of tumor cells combined with a reconstruction of tubules, which is also necessary in the satellite foci and metastases. Many studies of colon neoplasms dealt with the proliferation and differentiation,12,13 whereas the patterning was rarely investigated and is poorly understood.14,15
ß-catenin is decisively involved in the first patterning of a tubule in embryogenesis, which creates the primitive gut (archenteron) and is called the gastrulation.16,17 During this process the WNT signal pathway is activated, causing nuclear expressions of ß-catenin similar to colon carcinoma.16-22 The program of gastrulation is highly conserved from echinodermata to mammals and its simplest appearance is well documented for the sea urchin. Thus, nuclear ß-catenin could link a basic phylogenetic process of patterning from embryogenesis to carcinogenesis. To test this hypothesis we analyzed the patterning in colonic adenomas and adenocarcinomas in relation to the nuclear ß-catenin expression and in comparison with the embryonic gastrulation in the sea urchin.
| Materials and Methods |
|---|
|
|
|---|
The study was comprised of 88 colon adenomas and 45 colon adenocarcinomas. Of five adenocarcinomas, lymph node metastases were also evaluated. Only tubular and tubulovillous adenomas were studied. They were included irrespective of an immunohistochemically detectable nuclear ß-catenin expression. Adenocarcinomas had to exhibit clear tubular structures and an immunohistochemically detectable nuclear ß-catenin expression for the inclusion into the study. Both findings were present in the majority of colonic adenocarcinomas from the files of our institute.
Immunohistochemistry
Expression of ß-catenin was investigated by a mouse anti-ß-catenin monoclonal antibody (1:100, clone 14; Transduction Labs., Lexington, KY) and a rabbit anti-ß-catenin antiserum (1:750; Sigma, Deisenhofen, Germany). In addition, all tumors were immunostained by mouse anti-cytokeratin 18 (1:200; DAKO, Hamburg, Germany) and mouse anti-E-cadherin monoclonal antibodies (1:100, clone 36; Transduction Labs.). All carcinomas were stained with a rabbit anti-fibronectin antiserum (1:500; DAKO). Single staining on paraffin sections was performed as previously described.6 For double-immunofluorescence both first antibodies (mouse anti-E-cadherin and rabbit anti-fibronectin antiserum) were incubated together overnight at room temperature. After washing, the slides were incubated with the corresponding secondary antibodies labeled with the fluorescence dyes anti-mouse Alexa 488 (1:500, green) and anti-rabbit Alexa 546 (1:500, red; Molecular Probes, Leiden, Netherlands).
Scoring of Immunostaining and Branching
Normal colonic epithelia adjacent to the adenomas were used as internal control of staining efficiency. The number of neoplastic cells with nuclear expression of ß-catenin detected with the monoclonal antibody was never homogeneous, but varied in the adenomas. The following scoring system was used to evaluate the percentage of cells with nuclear staining: no staining, score 0; low-grade staining, score 1 (<20%); medium-grade staining, score 2 (20 to 60%); and high-grade staining, score 3 (>60%). In addition, the branching of tubules was evaluated by a scoring system in the adenomas. Branching tubules were defined as tubules with an obvious tubular ramification or an intensely folded outline because of multiple branches in an immunostaining for cytokeratin 18. The percentage of branching tubules was scored, as follows: no branching, score 0; low-grade branching, score 1 (<20%); medium-grade branching, score 2 (20 to 60%); high-grade branching, score 3 (>60%). For the scoring of nuclear ß-catenin immunostaining the adenomas were screened by x200 and x400 magnification (ocular, x10; objective, x20 and x40). For the scoring of branching a x100 and x200 magnification (ocular, x10; objective, x10 and x20) were used. The areas of highest branching intensity were taken for the score of branching and nuclear ß-catenin immunostaining in adenomas with heterogeneous patterns.
Statistics
Statistical analysis was performed using the SPSS software (SPSS
Standard version 9.0.0; SPSS Inc., Chicago, IL). The correlation
between absolute size (in mm) and branching was evaluated by the
Mann-Whitney U-Test. Correlation between branching activity
and nuclear ß-catenin staining was evaluated by
-square test
(linear-by-linear association).
Comparative Analysis of Patterning
Patterning was analyzed by a comparative analysis of colonic
adenomas and adenocarcinomas with the gastrulation of the sea urchin,
which exhibits the simplest and best documented patterning of
gastrulation. The following steps can be distinguished (see Figure 7
):23
1) epithelio-mesenchymal transition and ingression
of the primary mesenchyme: at the center of the vegetal plate of the
blastula, some epithelial cells undergo a transition to mesenchymal
cells that dissociate from the epithelial layer and migrate as primary
mesenchyme into the blastocoel. Together with the secondary mesenchyme,
which dissociates during the extension of the primitive gut
(archenteron), these cells develop the mesoderm. 2) Endodermal
invagination and formation of the primitive gut (archenteron): the
epithelial cells of the vegetal plate become the endoderm. They
rearrange after the ingression of the mesenchyme and push into the
blastocoel as a coherent epithelial layer. After the initial budding of
the endodermal cells a primitive tubule (archenteron) is formed by
convergent extension. Epithelio-mesenchymal transition during
gastrulation is characterized by a loss of membranous E-cadherin and an
expression of fibronectin. The cells undergoing this transition
(micromeres) exhibit a stronger nuclear ß-catenin expression than the
cells undergoing endodermal differentiation and invagination
(macromeres).16,17,22
|
| Results |
|---|
|
|
|---|
Two types of patterning were distinguished in colonic adenomas and
adenocarcinomas: tubular branching and invasion with tubular
reconstruction. Tubular branching was characterized by a budding and
ramification of epithelial cells from pre-existing tubules (Figures 1 and 3)
. Tubules with a high degree of
branching showed intensely folded outlines (see Figure 3
). Branching
into the tubules caused cribriform patterns in some tumors. The
patterning of tubular branching was found in adenomas and
adenocarcinomas. Only 30 very small adenomas (median size, 2.6 mm)
exhibited no tubular branching (Figure 2)
. The intensity of branching measured
by a score was significantly correlated with the size of adenomas
(Figure 2)
.
|
|
|
|
|
The patterning of tubular branching was correlated with nuclear
ß-catenin expressions in epithelial tumor cells in the adenomas
(Figure 3)
. Most adenomas showed spatial
expressions of nuclear ß-catenin. Only 10 of 88 adenomas had strong
expression of nuclear ß-catenin in >60% of the neoplastic cells
(Table 1)
. Spatial expressions
predominated in tubules with epithelial budding and branching (Figures 1 and 3)
. Adenomas without tubular branches were negative for nuclear
ß-catenin. Morphometric analyses demonstrated a significant
correlation between the branching score and the score for nuclear
ß-catenin expression in the adenomas (Table 1)
. The adenocarcinomas
usually showed intense tubular branching in the inner zone that was
correlated with nuclear ß-catenin expression in neoplastic cells
(Figure 4, a and b)
. The nuclear staining
for ß-catenin in branching tubules of adenomas and adenocarcinomas
was weaker by the polyclonal than by the monoclonal antibody used in
this study.
|
The patterning of invasion with reconstruction was associated with
nuclear ß-catenin expressions. The invasion front showed the
strongest spatial nuclear ß-catenin immunostaining in most
adenocarcinomas (Figure 4, a and c)
. The stronger nuclear ß-catenin
staining in areas of invasion than in areas of branching was better
distinguishable by the polyclonal than by the monoclonal antibody. The
strong nuclear ß-catenin expressions were found in all patterning
steps of invasion, dissociation, and reconstruction of tubules (Figure 4, a and c)
. The neoplastic patterning in lymph node metastases
resembled the patterning of invasion with reconstruction and was
associated with strong nuclear ß-catenin expression, too (not shown).
Patterning of Invasion with Reconstruction and Epithelio-Mesenchymal Transitions
A membranous E-cadherin expression characterizes a
phenotype of coherent stationary epithelial cells. Fibronectin
production is typical for dissociated migrating cells of mesenchymal
phenotype.8
E-cadherin was detected in the normal cryptal
and the coherent neoplastic epithelial cells. Fibronectin was found in
the extracellular matrix of the normal colonic mucosa and in the
stroma of colonic tumors by immunohistochemistry (Figures 5a and 6)
.
|
Analogies of the Tumor Patternings with Embryonic Gastrulation
The patterning of invasion with reconstruction showed analogies
with the epithelio-mesenchymal transition, the ingression of the
primary mesenchyme and the rearrangement of endodermal cells in the
first phase of gastrulation in the sea urchin (Figure 7)
. The analogies were: 1) the transition
of epithelial tumor cells to a mesenchymal phenotype; 2) the
dissociation and ingression of single tumor cells with the transient
mesenchymal phenotype into the stroma; and 3) the rearrangement of
tumor cells as new tubules at the invasion front (Figures 4 and 7)
.
Similar to the zone of epithelio-mesenchymal transition and mesenchymal
ingression at the vegetal plate of the sea urchin, the strongest
nuclear ß-expressions were associated with the patternings of
invasion with reconstruction in the adenocarcinomas (Figure 4)
. The
patterning of tubular branching exhibited analogies with the endodermal
invagination during the second phase of the gastrulation in the sea
urchin (Figure 7)
. The budding of epithelial cells from neoplastic
tubules resembled the initial invagination of endodermal cells in
gastrulation. The extension of tubular branches corresponded to the
convergent extension of the primitive gut (Figure 1, ac)
. The
observation of a weaker nuclear ß-catenin expression at branching
than at invasion (Figure 4)
is equivalent to a weaker nuclear
ß-catenin expression in the zone of invaginating endodermal cells
than the zone of epithelio-mesenchymal transition and ingression at the
vegetal plate of the sea urchin.22
| Discussion |
|---|
|
|
|---|
In contrast, the adenomas grow by tubular branching without invasion
and repeat only the second phase without the first phase of
gastrulation. Nuclear ß-catenin expression and tubular branching were
absent in very small adenomas. This lack of nuclear ß-catenin goes
along with an absence of tubular branching. Wasan et al24
previously described that microadenomas in human familial adenomatosis
polyposis and in the multiple intestinal neoplasia mouse enlarge by
elevated rates of crypt fission, which is the process for new crypt
formation during the extension of the colon in development. In larger
adenomas they described a high percentage of abnormal and heterogeneous
crypt fissions, which morphologically corresponds to the tubular
branching described here. Studies in both epithelial cell
culture25
and intestinal epithelial cells in
mice26
demonstrated that forced overexpression of
ß-catenin leads to polyp formation by increased abnormal branching.
These findings fit well to our observations, that the presence and
increasing number of cells with nuclear ß-catenin are significantly
related to the presence and increasing rate of tubular branching in the
adenomas. The opposite direction of tubular invagination in adenomas
and in gastrulation (see Figure 7
) might be because of the localization
of the basal membrane, which is outside in neoplastic tubules, but
inside in the blastocoel.
Taken together the colonic adenoma-carcinoma sequence seems to reflect a step-wise return to the patternings of the second and first phase of the gastrulation, which goes along with an increasing extension and intensity of nuclear ß-catenin expression in the neoplastic epithelial cells. The distribution and amount of nuclear ß-catenin expression in colonic neoplasms has consequently to be regulated and this regulation or misregulation could be decisive for the neoplastic progression. The question therefore is, what regulates and forces the nuclear translocation of ß-catenin? The APC gene mutation is probably the gatekeeper, but itself not a sufficient cause for the accumulation of nuclear ß-catenin. This is supported by observations that small adenomas lack detectable nuclear ß-catenin, despite APC gene mutations.10 Additional dominant mutations in signal pathways, eg, in the ras or src oncogene, that might influence the ß-catenin function, would cause genetically fixed nuclear expressions, but cannot explain the regulation and spatial nuclear translocation in carcinomas. We speculate that signals from the tumor microenvironment, eg, by soluble factors or extracellular matrix molecules, are important regulators. For instance stimulation by epidermal growth factor27,28 or trefoil factors29 was shown do release E-cadherin-bound ß-catenin by a tyrosine phosphorylation. Also an activation of integrins by extracellular matrix molecules can influence intracellular ß-catenin distribution, as it was demonstrated for the integrin ß1 and the associated integrin-linked kinase.30,31
The increase of nuclear ß-catenin in neoplastic cells could finally
reach a critical level that is necessary for epithelio-mesenchymal
transitions and ingressions of cells with mesenchymal phenotype at the
invasion front. The critical level is probably powerfully
contraregulated, because it initiates a decisive alteration and new
rearrangement of the cellular context. However, when that level is
crossed the patterning of invasion could start and perpetuate itself by
ongoing epithelio-mesenchymal and mesenchymo-epithelial transitions in
the primary tumors and in metastases. Particularly the second
retransition step, enabling a tubular reconstruction, can explain the
metastatic growth with an identical differentiated morphology as the
primary tumor after the dissemination of isolated tumor cells. Based on
the present observations we suggest that nuclear ß-catenin could be a
prerequisite for the transition processes to drive tumor cells toward
either a stationary epithelial or a migratory mesenchymal phenotype.
Tumors could thus be considered as dynamic epithelial-mesenchymal
equilibriums that are balanced by contrarotating signal gradients, as
in developmental biology (Figure 8)
.
|
How can the additional signaling from the microenvironment be defined? During gastrulation the transforming growth factor-ß-like molecule BMP4 counteracts the functions of nuclear ß-catenin in the endo-mesoderm formation.21 Interestingly mutations inactivating the transforming growth factor-ß pathway, eg, in the DPC4 gene37 or transforming growth factor-ß type II receptor,38 are found in late dysplastic adenomas, where a strong increase in nuclear ß-catenin is observed.10 Moreover, some factors are known that can drive epithelio-mesenchymal transitions. For instance the hepatocyte growth factor can initiate a mesenchymal transition of epithelial cells by activating the Ets-1 transcription factor.39 In contrast the leukemia inhibitory factor was shown to initiate a mesenchymo-epithelial conversion.40 Because extracellular matrix proteins can also have different effects via integrin binding on tumor cell phenotypes,41 the spatial heterogeneities of the extracellular matrix composition in tumors and metastases could have important impacts. For example experimental mouse mammary tumors demonstrate, that metalloproteases produced by tumor cells can induce epithelio-mesenchymal transitions by a remodeling of the extracellular matrix.42,43 Intratumorous signaling gradients could consequently be built up by combinations of soluble morphogens and modified extracellular matrix components. In conclusion the present study provides some intriguing implications for colonic carcinogenesis: 1) the adenoma-carcinoma sequence can be considered as a sequential return of neoplastic cells to a transcriptional susceptibility of embryonic gastrulation. 2) The return could be because of a step-wise deregulation of the ß-catenin turnover and translocation to the nucleus that primarily starts by a decrease of ß-catenin degradation because of APC loss of function mutations. 3) The intratumorous nuclear ß-catenin expression might enable a step-wise transcriptional activation and initiation of patternings repeating basic morphogenic processes of embryonic gastrulation. 4) The microenvironmental regulation of such processes could be the driving force of tumor growth and progression.
The present data for these conclusions are observational and their limitations have to be overcome by additional molecular and functional studies. Moreover, the study is representative just for the common type of colon carcinoma, exhibiting clear tubular structures as well as nuclear ß-catenin expression. However, the novel approach to link tumor patternings with developmental biology already offers remarkable new insights into the complexity of colon cancer. It might be fruitful for other cancers, too.
| Acknowledgements |
|---|
| Footnotes |
|---|
Accepted for publication July 1, 2000.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
M. Misikangas, A.-M. Pajari, E. Paivarinta, S. I. Oikarinen, J. Rajakangas, M. Marttinen, H. Tanayama, R. Torronen, and M. Mutanen Three Nordic Berries Inhibit Intestinal Tumorigenesis in Multiple Intestinal Neoplasia/+ Mice by Modulating {beta}-Catenin Signaling in the Tumor and Transcription in the Mucosa J. Nutr., October 1, 2007; 137(10): 2285 - 2290. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Mazzone and P. M. Comoglio The Met pathway: master switch and drug target in cancer progression FASEB J, August 1, 2006; 20(10): 1611 - 1621. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. E. Witek, K. Nielsen, R. Walters, T. Hyslop, J. Palazzo, S. Schulz, and S. A. Waldman The Putative Tumor Suppressor Cdx2 Is Overexpressed by Human Colorectal Adenocarcinomas Clin. Cancer Res., December 15, 2005; 11(24): 8549 - 8556. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Pena, J. M. Garcia, J. Silva, V. Garcia, R. Rodriguez, I. Alonso, I. Millan, C. Salas, A. G. de Herreros, A. Munoz, et al. E-cadherin and vitamin D receptor regulation by SNAIL and ZEB1 in colon cancer: clinicopathological correlations Hum. Mol. Genet., November 15, 2005; 14(22): 3361 - 3370. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. Brabletz, S. Spaderna, J. Kolb, F. Hlubek, G. Faller, C. J. Bruns, A. Jung, J. Nentwich, I. Duluc, C. Domon-Dell, et al. Down-Regulation of the Homeodomain Factor Cdx2 in Colorectal Cancer by Collagen Type I: An Active Role for the Tumor Environment in Malignant Tumor Progression Cancer Res., October 1, 2004; 64(19): 6973 - 6977. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Saegusa, M. Hashimura, T. Kuwata, M. Hamano, and I. Okayasu {beta}-Catenin Simultaneously Induces Activation of the p53-p21WAF1 Pathway and Overexpression of Cyclin D1 during Squamous Differentiation of Endometrial Carcinoma Cells Am. J. Pathol., May 1, 2004; 164(5): 1739 - 1749. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. E. Baldus, S. P. Monig, S. Huxel, S. Landsberg, F.-G. Hanisch, K. Engelmann, P. M. Schneider, J. Thiele, A. H. Holscher, and H. P. Dienes MUC1 and Nuclear {beta}-Catenin Are Coexpressed at the Invasion Front of Colorectal Carcinomas and Are Both Correlated with Tumor Prognosis Clin. Cancer Res., April 15, 2004; 10(8): 2790 - 2796. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Uraguchi, M. Morikawa, M. Shirakawa, K. Sanada, and K. Imai Activation of WNT Family Expression and Signaling in Squamous Cell Carcinomas of the Oral Cavity J. Dent. Res., April 1, 2004; 83(4): 327 - 332. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Chilosi, V. Poletti, A. Zamo, M. Lestani, L. Montagna, P. Piccoli, S. Pedron, M. Bertaso, A. Scarpa, B. Murer, et al. Aberrant Wnt/{beta}-Catenin Pathway Activation in Idiopathic Pulmonary Fibrosis Am. J. Pathol., May 1, 2003; 162(5): 1495 - 1502. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. P. S. Monga, W. M. Mars, P. Pediaditakis, A. Bell, K. Mule, W. C. Bowen, X. Wang, R. Zarnegar, and G. K. Michalopoulos Hepatocyte Growth Factor Induces Wnt-independent Nuclear Translocation of {beta}-Catenin after Met-{beta}-Catenin Dissociation in Hepatocytes Cancer Res., April 1, 2002; 62(7): 2064 - 2071. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. Lustig, B. Jerchow, M. Sachs, S. Weiler, T. Pietsch, U. Karsten, M. van de Wetering, H. Clevers, P. M. Schlag, W. Birchmeier, et al. Negative Feedback Loop of Wnt Signaling through Upregulation of Conductin/Axin2 in Colorectal and Liver Tumors Mol. Cell. Biol., February 15, 2002; 22(4): 1184 - 1193. [Abstract] [Full Text] [PDF] |
||||
![]() |
F. Hlubek, A. Jung, N. Kotzor, T. Kirchner, and T. Brabletz Expression of the Invasion Factor Laminin {gamma}2 in Colorectal Carcinomas Is Regulated by {beta}-Catenin Cancer Res., November 1, 2001; 61(22): 8089 - 8093. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Jung, M. Schrauder, U. Oswald, C. Knoll, P. Sellberg, R. Palmqvist, G. Niedobitek, T. Brabletz, and T. Kirchner The Invasion Front of Human Colorectal Adenocarcinomas Shows Co-Localization of Nuclear {beta}-Catenin, Cyclin D1, and p16INK4A and Is a Region of Low Proliferation Am. J. Pathol., November 1, 2001; 159(5): 1613 - 1617. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. Lifschitz-Mercer, R. Amitai, B. B.-S. Maymon, L. Shechtman, B. Czernobilsky, L. Leider-Trejo, A. Ben-Ze'ev, and B. Geiger Nuclear Localization of {beta}-Catenin and Plakoglobin in Primary and Metastatic Human Colonic Carcinomas, Colonic Adenomas, and Normal Colon International Journal of Surgical Pathology, October 1, 2001; 9(4): 273 - 279. [Abstract] [PDF] |
||||
![]() |
T. Brabletz, A. Jung, S. Reu, M. Porzner, F. Hlubek, L. A. Kunz-Schughart, R. Knuechel, and T. Kirchner Variable beta -catenin expression in colorectal cancers indicates tumor progression driven by the tumor environment PNAS, August 28, 2001; 98(18): 10356 - 10361. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||