HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 18, 15843-15850, May 3, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/18/15843    most recent M200184200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wong, M. H. Articles by Gordon, J. I. Search for Related Content PubMed PubMed Citation Articles by Wong, M. H. Articles by Gordon, J. I. Social Bookmarking                      What's this?

Selection of Multipotent Stem Cells during Morphogenesis of Small Intestinal Crypts of Lieberkühn Is Perturbed by Stimulation of Lef-1/-Catenin Signaling^*

Melissa H. Wong^§, Joerg Huelsken^¶, Walter Birchmeier^¶, and Jeffrey I. Gordon

From the  Department of Molecular Biology and Pharmacology, Washington University School of Medicine, St. Louis, Missouri 63110 and the ^¶ Max-Delbrueck-Center for Molecular Medicine, 13125 Berlin, Germany

Received for publication, January 8, 2002, and in revised form, February 8, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Studies of chimeric mice have disclosed that the stem cell hierarchy in the small intestinal epithelium is established during formation of its proliferative units (crypts of Lieberkühn). This process involves a selection among several multipotential progenitors so that ultimately only one survives to supply descendants to the fully formed crypt. In this report, we examine the hypothesis that the level of -catenin (-cat)-mediated signaling is an important factor regulating this stem cell selection. In the canonical Wnt signaling pathway, -catenin can partner with Lef-1/Tcf high mobility group (HMG) box transcription factors to control gene expression. Both Lef-1 and Tcf-4 mRNAs are produced in the fetal mouse small intestine. Tcf-4 expression is sustained, whereas Lef-1 levels fall as crypt formation is completed during the first two postnatal weeks. A Tcf-4 gene knockout is known to block intestinal epithelial proliferation in late fetal life. Therefore, to test the hypothesis, we enhanced -catenin signaling in a chimeric mouse model in which the stem cell selection could be monitored. A fusion protein containing the HMG box domain of Lef-1 linked to the trans-activation domain of -catenin (Lef-1/-cat) was constructed to promote direct stimulation of signaling without being retained in the cytoplasm through interactions with E-cadherin and Apc/Axin. Lef-1/-cat was expressed in 129/Sv embryonic stem cell-derived small intestinal epithelial progenitors present in developing B6-ROSA26129/Sv chimeras. Lef-1/-cat stimulated expression of a known -catenin target (E-cadherin), suppressed expression of Apc and Axin, and induced apoptosis in 129/Sv but not in neighboring B6-ROSA26 epithelial cells. This apoptotic response was not associated with any detectable changes in cell division within the Lef-1/-cat-expressing epithelium. By the time crypt development was completed, all 129/Sv epithelial cells were lost. These results indicate that developmental changes in -catenin-mediated signaling can play an important role in establishing a stem cell hierarchy during crypt morphogenesis.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In the canonical Wnt/Wingless pathway, extracellular Wnt signals are transduced through two receptors, Frizzled and the low density lipoprotein-related receptor, to inhibit cytosolic glycogen synthase kinase-3 (Gsk-3)-mediated phosphorylation of the N-terminal region of -catenin (1, 2). Wnt inhibition of -catenin phosphorylation prevents axin/adenomatous polyposis coli (Apc)-directed targeting of -catenin to an ubiquitin-proteasomal degradation pathway (3-5). This allows unphosphorylated cytoplasmic -catenin to enter the nucleus where it interacts with Lef-1/Tcf (lymphocyte enhancer factor-1/T cell factor) HMG¹ box transcription factors to regulate expression of a number of target genes (6-10). The DNA binding domain of Lef-1/Tcf and the transactivation domain of -catenin function together as a bipartite transcriptional activator. Targets of the Wnt/-catenin pathway include genes involved in regulating a variety of developmental processes, including definition of cell polarity, regulation of cell proliferation, specification of cell fates, and pattern formation (8, 9, 11).

Recent reports have emphasized the role of -catenin, Tcf-3, and Lef-1 in regulating the ability of multipotent mouse skin stem cells to give rise to descendant lineages. A skin-specific knockout of -catenin attenuates placode formation during embryogenesis and dramatically restricts specification of cell fates by the multipotent bulge stem cell after completion of the initial hair cycle. As the second anagen commences, -catenin-deficient stem cells cannot give rise to any follicular epithelial cells and are restricted to producing epidermal keratinocytes (12). Merrill et al. (13) generated several pedigrees of transgenic mice that overexpress wild-type Lef-1, Tcf-3, and dominant negative derivatives. They showed that the balance of signaling through these HMG box transcription factors has profound effects on cell fate selection by descendants of the pilosebaceous unit's multipotent bulge stem cells.

The role of the canonical Wnt/-catenin pathway in regulating the activity and survival of the multipotent intestinal epithelial stem cell is less well understood. Studies of chimeric mice have disclosed that the stem cell hierarchy of the intestine is established coincident with formation of its crypts of Lieberkühn. These flask-shaped mucosal invaginations are the proliferative units of the adult intestine. Crypts develop from a flat intervillus epithelium located between finger-like villi. Villi first arise from the pseudostratified gut epithelium at E14 (14). During this stage of development, multipotent stem cells in the intervillus epithelium are able to specify three of the four epithelial lineages that are ultimately represented in the mature intestine: absorptive enterocytes, goblet, and enteroendocrine cells (15, 16). The intervillus epithelium undergoes reshaping to yield fully formed crypts by the end of the second postnatal week (17). In late fetal life, the intervillus epithelium of a chimeric mouse contains several active multipotent stem cells representing both genetic backgrounds used to generate the chimera (i.e. it is polyclonal). As the intervillus epithelium develops into crypts, a poorly understood process of stem cell "purification" occurs that converts a polyclonal nascent crypt to a monoclonal mature crypt where all epithelial cells share a common genotype (18). This phenomenon has been interpreted to mean that all active stem cells in each purified monoclonal crypt are descended from a single crypt progenitor cell that occupies the highest position in the established stem cell hierarchy.

Korinek et al. (19) showed that mice homozygous for a Tcf-4 null allele die shortly after birth with an absence of proliferative activity in the intervillus epithelium. Proliferative activity appears normal at E14.5 in Tcf4^/ animals but is lost by E16.5 when only differentiated epithelial lineages are evident. The loss of proliferative activity indicates that multipotent stem cells in the developing intestine require some level of Wnt/-catenin signaling for their maintenance. Further understanding of the role of Wnt/-catenin signaling in establishing the stem cell hierarchy of the crypt has been impeded by several factors. First, Tcf-4^/ mice die before the hierarchy is established. Second, conventional knockout mice are not suited for an assessment of the evolution of the hierarchy; chimeric mice currently provide the only direct way of monitoring the stem cell selection process. Third, information is lacking about the impact of augmented Wnt/-catenin signaling on gut development. Prior genetic studies that provided an opportunity to examine the consequences of enhanced signaling have been limited by embryonic lethality (e.g. Apc^/ mice die prior to gastrulation (20)), have not described the effects on the intestine (as in a recently described Apc truncation mutant that produces viable mice (21)), or have not focused on the developing gut (e.g. mice expressing a mutant -catenin with an N-terminal truncation that removes Gsk-3 phosphorylation sites and thereby increases protein stability (22,23)).

In this report, we test the hypothesis that the level of -catenin-mediated signaling is an important factor in stem cell selection during crypt morphogenesis. The experimental approach was based on the following. We had previously identified transcriptional regulatory elements from a fatty acid-binding protein gene (Fabpl) that function in the multipotent intestinal stem cell prior to, during, and after completion of crypt purification (24). We knew that 129/Sv ES cells transfected with Fabpl/reporter constructs could be introduced into C57Bl/6-ROSA26 (B6-ROSA26) blastocysts (25), and the effect(s) of transgene expression monitored in the resulting developing or adult chimeric mouse by comparing the phenotype of its 129/Sv intestinal epithelial cells with the phenotype of readily identifiable neighboring B6-ROSA26 epithelium (22, 26). Finally, we reasoned that a fusion protein containing an HMG box domain linked to the trans-activation domain of -catenin would allow direct stimulation of signaling without being retained in the cytoplasm through interactions with Apc/Axin. As described below, expression of this fusion protein had a dramatic effect on the survival of multipotent 129/Sv intestinal stem cells during crypt formation/purification.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Expression of a Lef-1/-Catenin Fusion Protein in Cultured Cells-- p3021 encodes a protein consisting of an N-terminal 15-residue c-Myc epitope tag, amino acids 33-368 of human Lef-1, and the C-terminal 124 residues of human -catenin (abbreviated Lef-1/-cat; see Fig. 1A). 2 × 10⁶ IIAI.6 B cells were transfected with a CAT reporter (pTK(56)₇; see Ref. 27) and (i) p3021 or (ii) expression plasmids encoding wild-type mouse Lef-1 (p2983) plus wild-type human -catenin (p306) or (iii) plasmids encoding a mutant Lef-1 lacking its -catenin binding domain (Lef-1CatBD, deleted residues from Lef-1 = 1-32) plus wild-type -catenin or (iv) plasmids specifying wild-type Lef-1 plus a mutant -catenin lacking its trans-activation domain (CatTransAct; deleted residues from -catenin = 696-781). 48 h after transfection, cellular CAT activities were determined (27).

Generation of Chimeric Mice-- A 1540-bp NcoI-EcoRI fragment from p3021 containing Lef-1/-cat coding sequences was subcloned at engineered NcoI-MfeI sites in pLPNDon (28), yielding pLF:Lef-1/-cat. This subcloning placed Lef-1/-cat downstream of nucleotides 596 to +21 of rat Fabpl and upstream of nucleotides +3 to +2152 of the human growth hormone (hGH) gene and a pgk-neomyocin-resistance selection cassette. Multiple stop codons separated the open reading frame encoding Lef-1/-cat and the initiator Met of hGH, thereby ensuring that hGH would not be produced from transcripts derived from Fabpl-Lef-1/-cat-hGH. The 6.24-kb Fabpl-Lef-1/-cat-hGH-pgk-neo insert in pLF:Lef-1/-cat was excised with SacII, purified by gel electrophoresis, and electroporated into D3 129/Sv ES cells (22). Stably transfected ES cells were identified using PCR. Twelve ES cell lines were established and then injected (separately) into B6-ROSA26 blastocysts (22) to produce B6-ROSA26129/Sv(Lef-1/-cat) chimeras. Non-transfected ES cells were used to generate control B6-ROSA26129/Sv chimeras. The 129/Sv contribution in 6-week-old chimeras derived from all ES cell lines ranged from 20-60% based on coat color. Chimerism in E13.5-P7 mice was assessed by glucose phosphate isomerase assay of limb tissue (29).

Assaying Lef-1/-Cat Expression-- Chimeras generated from each of the 12 different cell lines were analyzed for transgene expression. Total cellular RNA was isolated (RNeasy kit, Qiagen, Valencia, CA) from the entire small intestine of E13.5-E18.5 and postnatal day 1 (P1) B6-ROSA26129/Sv(Lef-1/-cat) chimeras and from the middle third of the small intestines of P42 adult mice.

RT-PCR was used to identify the Fabpl-Lef-1/-cat-hGH mRNA transcript. Two µg of DNase-treated total cellular RNA was used in each oligo(dT)-primed cDNA synthesis reaction (final volume of 200 µl; described in Ref. 24). Two-µl aliquots from the cDNA synthesis reaction mixture were added to RT-PCR assays, each containing one of two primer pairs. One primer pair produced a 150-bp amplicon from hGH sequences (forward primer, 5'-AGGTGGCCTTTGACACCTACCAGG-3'; reverse primer, 5'-TCTGTTGTGTTTCCTCCCTGTTGG-3'; note that the amplicon spanned an intron/exon junction). The other primer pair generated a 722-bp amplicon from the junction between Lef-1/-catenin and hGH (-cat forward primer, 5'-GGACTTGATATTGGTGCCCAG-3'; hGH reverse primer, 5'-TCTGTTGTGTTTCCTCCCTGTTGG-3'). The annealing temperature for PCR was 67 °C (total of 30 cycles). Primers designed to produce a 300-bp amplicon from -actin mRNA were used to confirm the integrity of intestinal RNAs (forward primer, 5'-CACCACACCTTCTACAATGAGCTG-3'; reverse primer, 5'-TCATCAGGTAGTCAGTGAGGTCGC-3'; annealing at 65 °C). Control reactions contained cDNA prepared from intestinal RNAs that had been isolated from age-matched normal (B6-ROSA26129/Sv) chimeras.

To detect Lef-1/-cat, proteins were extracted from the small intestines of B6-ROSA26129/Sv(Lef-1/-cat) and control B6-ROSA26129/Sv chimeras using previously published procedures (22). Western blots of extracted small intestinal proteins were probed with rabbit antibodies raised against a C-terminal peptide from -catenin (22) or the N-terminal c-Myc tag (Upstate Biotechnology, Lake Placid, NY, 1:100) of Lef-1/-cat. Antigen-antibody complexes were detected using the Western Light kit from Tropix (Foster City, CA).

X-Gal Genotyping of Small Intestinal Epithelium from Chimeric Mice-- Small intestines were removed from E13.5-E18.5, P1, P28, and P42 mice. For E13.5-P1 animals, the intestines were fixed immediately in periodate-lysine-paraformaldehyde (PLP) for 1 h at 24 °C, washed 3 times with PBS, and incubated overnight at 4 °C in X-gal solution (22). Intestines dissected from chimeric mice were embedded in 2% agar and subdivided into 2-mm segments with a razor blade. Each segment was placed cut side down on a flat surface (carefully preserving their cephalocaudal orientation) and lined up into parallel rows of six segments each. The arrayed clusters were first embedded in 2% agar, then embedded in paraffin. Serial 5-µm sections of the tissue block were cut and counterstained with nuclear fast red. Surveys of these serial sections allowed the distribution of 129/Sv and B6-ROSA26 epithelial cells to be defined throughout the small intestine.

For P28 and P42 mice, the small intestine was removed en bloc immediately after sacrifice and flushed with ice-cold PBS followed by PLP. The intestine was opened with a longitudinal incision along its mesenteric side, pinned onto dissecting wax, and fixed in PLP for 1 h at 24 °C. Whole mounts were washed three times with PBS (5 min each) followed by a 45-min incubation in 20 mM dithiothreitol, 20% EtOH, 150 mM Tris-HCl (pH 8.0) to remove surface mucus. Following three more PBS washes, the whole mount preparations were placed in X-gal solution for 12 h at 4 °C and photographed.

Quantitation of Apoptosis and Cell Division-- Serial sections were prepared from the middle third of X-gal-stained small intestines obtained from E17.5, E18.5, and P1 B6-ROSA26129/Sv(Lef-1/-cat) mice and age-matched normal chimeric controls (n = 2-3 animals/group/time point). Sections were counterstained with hematoxylin and eosin, and the number of apoptotic and M-phase cells in B6-ROSA26 and 129/Sv intervillus epithelium was scored (n = 1767-2865 intervillus regions surveyed/time point). Mean values ±S.E. per genotype were computed. Observed differences between 129/Sv and B6 epithelium were analyzed using Student's t test.

Multilabel Immunohistochemistry-- E18.5 small intestines were fixed in PLP for 1 h at 24 °C, washed three times with PBS, cryo-protected in 15% sucrose/PBS (12-16 h at 4 °C), frozen in OCT (VWR, Batavia, IL), and 5-8 µm sections were cut. The protocol used for multilabel immunohistochemical studies has been described in an earlier publication (22). Sections were stained with affinity-purified rabbit anti-Escherichia coli -galactosidase (Ref. 22; diluted 1:500 in blocking buffer (1% bovine serum albumin, 0.3% Triton X-100, 1 mM CaCl₂ in PBS)) and with rabbit anti-Apc (raised against residues 1034-2130 of the human protein (30, 31); a gift from Paul Polakis, Genentech; final dilution = 1:500)). Antigen-antibody complexes were detected with indocarbocyanine- or fluorescein isothiocyante-conjugated donkey anti-rabbit Ig (Jackson ImmunoResearch Laboratories; West Grove, PA; 1:500).

Real-time Quantitative RT-PCR (qRT-PCR)-- SYBR Green-based qRT-PCR (32) was used to quantitate levels of gene expression in age-matched B6-ROSA26129/Sv(Lef-1/-cat) and normal B6-ROSA26129/Sv chimeras. Primer pairs were designed to the 3' end of the targeted transcript to produce 100-200-bp amplicons and, whenever possible, spanned an intron-exon boundary. A melting curve (T_m) was used to identify a temperature at which only the amplicon, and not primer dimer, accounted for the SYBR Green-bound fluorescence (24, 33). Each reaction contained 1× SYBR Green PCR Master Mix buffer (Applied Biosystems, Foster City, CA), 0.25 UDP-N-glycosidase (Invitrogen), 900 nM forward and reverse primers, and a 2-µl aliquot from the 200-µl cDNA synthesis reaction mixture. The primers and melting temperatures used for qRT-PCR were as follows: glyceraldehyde 3-phosphate dehydrogenase (GAPDH; 5'-TGGCAAAGTGGAGATTGTTGCC, 5'-AAGATGGTGATGGGCTTCCCG, 80 °C); Lef-1 (5'-AGACACCCTCCAGCTCCTGA, 5'-CCTGAATCCACCCGTGATG, 78.6 °C); Tcf-4 (5'-GGCGTTGGACAGATCACC, 5'-GGTGAAGTGTTCATTGCTGTACTG, 82.4 °C); -catenin (5'-AGCCGAGATGGCCCAGAAT, 5'-AAGGGCAAGGTTCGAATCAA, 79.2 °C); E-cadherin (5'-GTCAACACCTACAACGCTGCC, 5'-GTTGTGCTCAAGCCTTCGC, 80.2 °C); Apc (5'-TGACAAGACGGCAGCTGGAG, 5'-TCTTCGCTGTGCACGCTTC, 79.2 °C); Axin (5'-CCCCCATACAGGATCCGTAA, 5'-GGTACCCGCCCATTGACTT, 76.8 °C); and cyclooxygenase-2 (5'-TGAGTACCGCAAACGCTTCTC, 5'-TGGACGAGGTTTTTCCACCAG, 80 °C). All assays were performed in triplicate, on three separate occasions, using an Applied Biosystems Model 7700 Sequence Detector.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Real-time Quantitative RT-PCR Analysis of Lef-1, Tcf-4, and -Catenin Expression in the Developing Small Intestine-- Members of the Lef-1/Tcf HMG box family of transcription factors have overlapping patterns of expression during development (34). In situ hybridization studies have shown that Tcf-4 and Tcf-3 are expressed during and following development of the mouse small intestine and colon (35). Lef-1 shares a high degree of sequence homology with Tcf-4/Tcf-3 and recognizes a similar DNA binding site ((A/T)(A/T)CAA(A/T)GG); see Ref. 36). Although Lef-1 is not expressed in the adult intestinal epithelium, it is expressed in colorectal tumors (37, 38). Since gene expression during tumorigenesis can recapitulate patterns observed during normal organogenesis, we examined whether Lef-1 mRNA is present in the developing mouse intestine.

Fig. 1 presents the results of a real-time quantitative RT-PCR study of the relative levels of Lef-1, Tcf-4, and -catenin mRNAs in the small intestines of E13.5-P14 B6 mice. The time points encompass the period prior to the initiation of villus formation (E13.5), the period when nascent villi and the proliferating intervillus epithelium appear (E14.5-E18.5), and the period when crypts form from the intervillus epithelium with concomitant establishment of a final stem cell hierarchy (P1-P14). The figure shows that Tcf-4 mRNA levels are relatively constant during these periods. In contrast, Lef-1 expression is significantly higher than Tcf-4 expression from E13.5-17.5 and then drops precipitously from E17.5 to P1. -Catenin mRNA levels vary <3-fold between E13.5 and P14 and are 2000-5000-fold higher than the levels of either Lef-1 or Tcf-4 mRNA.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Real-time quantitative RT-PCR study of Lef-1, Tcf-4, and -catenin expression during normal mouse intestinal development. RNA was prepared from pooled intestines from B6 mice (n = 3-5 animals/time point), and SYBR Green-based qRT-PCR was performed. All data were normalized to an internal GAPDH mRNA control, and the levels of all mRNA species were expressed relative to the level of Tcf-4 mRNA at the postnatal day 1 (P1) time point. Mean values ±S.E. from three different determinations, each performed in triplicate, are plotted.

Design and in Vitro Assay of a Lef-1/-Catenin Fusion Protein-- Because Lef-1 expression changes during the critical period when stem cell selection is occurring in the developing crypt, we proceeded to design a fusion protein that would allow us to enhance Lef-1/-catenin trans-activation of target genes as crypts form. Fig. 2A outlines the salient features of the fusion protein. Residues 33-368 of human Lef-1, encompassing a functional HMG box and nuclear localization signal, were linked to the 13th armadillo repeat and C-terminal trans-activation domain of human -catenin. Physical linking of these two domains obviated the need for the -catenin binding domain of Lef-1. Moreover, exclusion of armadillo repeats 1-12 minimized potential interactions with cytoplasmic E-cadherin, Apc/Axin, and Gsk-3 that might compromise the ability of the Lef-1/-cat fusion protein to activate gene expression in intestinal epithelial cells.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Lef-1/-cat transactivates a target reporter. A, elements present in the fusion protein. B, co-transfection studies of cultured mouse IIAI.6 B cells comparing the trans-activation potential of Lef-1/-cat relative with wild-type -catenin and Lef-1 or to mutant -catenin and Lef-1 proteins that lack domains required for their interaction or transactivation. Mean values ±S.D. are plotted for replicate assays of CAT activity expressed from a plasmid containing seven tandem repeats of a Lef-1 binding site (CTTTGTT). NLS, nuclear localization signal.

The ability of Lef-1/-cat to activate transcription of a target reporter gene was first tested in a cultured mouse B cell line (IIAI.6) that lacks endogenous Lef-1 mRNA and contains very low levels of -catenin (39). IIAI.6 cells were transfected with a reporter plasmid containing seven tandem repeats of the Lef-1 DNA consensus binding site (CTTTGTT) linked to the thymidine kinase gene promoter and a bacterial chloramphenicol transferase (CAT) gene (pTK(56)₇; see Ref. 40). Co-transfection with plasmids encoding wild-type Lef-1 and -catenin produced a 10-fold increase in CAT activity as compared with cells containing the Lef-1 plasmid plus the reporter plasmid (Fig. 2B). Expression of Lef-1/-cat resulted in even more robust induction of CAT activity (15-fold as compared with Lef-1 alone; Fig. 2B). Control co-transfections with (i) plasmids encoding wild-type -catenin plus a mutant Lef-1 lacking a -catenin binding domain or (ii) plasmids encoding wild-type Lef-1 plus a mutant -catenin lacking its trans-activation domain resulted in significantly lower cellular CAT activities than those obtained by co-expression of wild-type Lef-1 and wild-type -catenin or by expression of Lef-1/-cat (Fig. 2B). These results support the requirement for close physical proximity between the DNA binding domain of Lef-1 and the transactivation domain of -catenin.

Generation of B6-ROSA26129/Sv(Lef-1/-Cat) Chimeric Mice-- Based on these cell culture results, we used Lef-1/-cat to test our hypothesis that the overall levels of -catenin-mediated signaling play an important role in regulating stem cell selection/survival during crypt morphogenesis. The experiment required that Lef-1/-cat be targeted to the stem cells, that expression of Lef-1/-cat be sustained throughout the period when endogenous Lef-1 gene expression declines, and that a reference control population of stem cells would be available to determine the effects of Lef-1/-cat on stem cell survival.

As noted in the Introduction, chimeric mice, generated by introducing 129/Sv ES cells into B6-ROSA26 blastocysts, can be used to monitor changes in multipotent stem cell populations as developing crypts are converted from polyclonality to monoclonality (Fig. 3A). Moreover, nucleotides 596 to +21 of Fabpl are known to reliably direct expression of various gene products to the multipotent small intestinal stem cell and all of its descendants, beginning as early as E13.5 and lasting through adulthood (24). Expression occurs throughout the proximal two-thirds of the small intestine and declines abruptly in the distal-most portion (terminal ileum (41)). Therefore, 129/Sv ES cells were stably transfected with a recombinant DNA containing Lef-1/-cat under the control of these Fabpl regulatory elements. Twelve different stably transfected, cloned ES cell lines or control non-transfected ES cells were each injected (separately) into B6-ROSA26 blastocysts to generate B6-ROSA26129/Sv(Lef-1/-cat) and control B6-ROSA26129/Sv mice, respectively.

View larger version (60K): [in this window] [in a new window]   Fig. 3.   Analysis of Lef-1/-cat expression in the small intestine of chimeric mice. A, sections from the mid-portion of the small intestine of a normal E18.5 and a normal P42 chimeric mouse. Sections were stained with X-gal and hematoxylin and eosin (left) or nuclear fast red (right). At E18.5, the intervillus epithelium is polyclonal, containing a mixture of B6-ROSA26 (blue) and 129/Sv cells. By P42, crypt morphogenesis has been completed. All crypts are monoclonal: i.e. composed exclusively of B6-ROSA26 or 129/Sv epithelial cells but not a mixture of both. Note the orderly migration of epithelial cells from each crypt up a neighboring villus. Villi supplied by both monoclonal B6-ROSA26 and monoclonal 129/Sv crypts have a striped appearance. Bars = 25 µm. B, RT-PCR study of RNA prepared from the small intestines of B6-ROSA26129/Sv(Lef-1/-cat) and normal chimeric mice (3-6 intestines pooled/time point). The arrow points to a 150-bp amplicon derived from Lef-1/-cat mRNA generated from the transgene. Note the age-associated reduction in levels of Lef-1/-cat mRNA. The integrity of each RNA preparation was verified using primers that generate a 300-bp amplicon from -actin mRNA. C, immunoblot analysis of E17.5 total small intestinal proteins (100 µg/lane). The blot was probed with antibodies to -catenin. D and E, X-gal-stained whole mount preparation of the small intestine from an adult (P42) B6-ROSA26129/Sv(Lef-1/-cat) mouse with 60% 129/Sv contribution to coat color. Panel D shows the mid-portion of the small intestine (jejunum). Only LacZ-positive B6-ROSA26 epithelium is retained (blue). Bar = 0.66 mm. Panel E, the distal 10% of the small intestine (ileum) showing a mixture of 129/Sv(Lef-1/-cat) epithelium (white) and B6-ROSA26 epithelium. F and G, X-gal-stained whole mount of jejunum and terminal ileum, respectively, from a P42 normal chimera whose 129/Sv contribution to coat chimerism was similar to the animal shown in panels D and E. Note the presence of 129/Sv and B6-ROSA26 epithelium in jejunum and ileum. Bars in F and G = 0.82 mm.

E15.5, E17.5, E18.5, and P1 chimeric mice were studied from each ES cell line. Total cellular RNA was isolated from their intact small intestines and assayed for transgene expression by RT-PCR. The expected size amplicon from the Fabpl-Lef-1/-cat transcript was identified in all 12 B6-ROSA26129/Sv(Lef-1/-cat) chimeric lines (a line of chimeric mice is defined as animals derived from a given stably transfected cloned ES cell line; n = 2-4 mice surveyed/line/time point). The amplicon was absent in intestinal RNA prepared from age-matched normal chimeras (n = 3/time point; Fig. 3B).

Lef-1/-cat expression was independently confirmed by Western blot analysis of total small intestinal proteins extracted from E17.5 B6-ROSA26129/Sv(Lef-1/-cat) and normal chimeras. When the blots were probed with antibodies to the C terminus of -catenin, endogenous wild-type -catenin (91 kDa) and Lef-1/-cat (70 kDa) were both detected (Fig. 3C). The presence of the 70-kDa Lef-1/-cat fusion protein was independently confirmed using antibodies to its N-terminal Myc tag (data not shown). As expected from the RT-PCR study, Lef-1/-cat was not detectable in the small intestines of normal chimeras (Fig. 3C). Steady state levels of the 91-kDa product of the endogenous -catenin gene (Catnb) were similar in chimeric-transgenic and normal chimeric mice (Fig. 3C).

Surprisingly, RT-PCR assays of small intestinal RNA prepared from adult (P42) B6-ROSA26129/Sv(Lef-1/-cat) mice representing each of the 12 ES cell lines failed to detect Lef-1/-cat mRNA (see Fig. 3B for results from one line; each of the other 11 lines exhibited a similar, progressive age-associated loss of this mRNA species). X-Gal staining of whole mounts of the proximal 90% of the small intestines of adult chimeras generated using each ES cell line revealed only B6-ROSA26 epithelium (see Fig. 3D for results representative of chimeric-transgenic mice generated from all 12 ES cell lines). Fabpl/reporter transgenes are generally not expressed in the distal 10% of the small intestine or in the colonic epithelium. These regions in B6-ROSA26129/Sv(Lef-1/-cat) chimeras retained scattered populations of 129/Sv cells (Fig. 3E), as did their other organs (e.g. liver and kidney; data not shown). In contrast, small intestines from age-matched normal chimeras contained a mixture of 129/Sv and B6-ROSA26 crypt-villus units throughout their length (Fig. 3, F and G). There were no significant differences in the size of the small intestines of adult B6-ROSA26129/Sv(Lef-1/-cat) mice as compared with age-matched normal chimeras.

129/Sv(Lef-1/-Cat) Cells Undergo Apoptosis-- Serial sections representing all regions of X-gal-stained small intestines from E15.5-E18.5 and P1 B6-ROSA26129/Sv(Lef-1/-cat) mice (n = 10/time point) were stained with hematoxylin and eosin. There were no apparent perturbations in villus formation, nor were there any histological changes indicative of perturbed cellular differentiation or cellular census when compared with age-matched normal chimeras (n = 5-15/time point). Quantitative histochemical analyses of the middle third of the small intestines of E17.5, E18.5, and P1 mice disclosed statistically significant 10-15-fold increases in the number of apoptotic cells in 129/Sv(Lef-1/-cat) as compared with neighboring B6-ROSA26 intervillus epithelium (p < 0.05; Fig. 4A-C). In contrast, there were no statistically significant differences in the frequency of apoptotic cells between the 129/Sv and B6-ROSA26 intervillus epithelium of normal age-matched chimeras (Fig. 4C). This apoptotic response was independently confirmed by terminal deoxynucleotidyltransferase-mediated dUTP nick end-labeling (TUNEL) assay (data not shown) and could not be attributed to enhanced proliferation. Quantitative studies failed to reveal any statistically significant difference in the number of M-phase cells in the 129/Sv versus B6-ROSA26 intervillus epithelium of chimeric-transgenic or normal chimeric mice (Fig. 4D).

View larger version (47K): [in this window] [in a new window]   Fig. 4.   Apoptotic response of 129/Sv(Lef-1/-cat) cells. A and B, serial sections from the middle of the small intestine of an E18.5 B6-ROSA26129/Sv(Lef-1/-cat) mouse. Panel A shows a section genotyped with X-gal and counterstained with nuclear fast red. Panel B shows a higher power view of the adjacent section stained with hematoxylin and eosin. Arrows point to an apoptotic cell. Bar in A = 25 µm. C, quantitation of apoptosis in the 129/Sv and B6-ROSA26 epithelium of age-matched B6-ROSA26129/Sv(Lef-1/-cat) chimeric-transgenic and normal chimeric mice. Sections prepared from the middle third of the small intestine were stained with X-gal and counterstained with hematoxylin and eosin. The number of apoptotic cells was scored in 129/Sv and B6-ROSA26 intervillus epithelium. Mean values ±S.E. are plotted (n = 2-3 mice/group/time point). D, the same sections used for the study in panel C scored for M-phase cells. There are no significant differences in cell division between B6-ROSA26 and 129/Sv epithelium at any time point in either group of age-matched mice.

Since the complete loss of 129/Sv epithelium occurs by the time crypts have completed their morphogenesis (P14), we concluded that all 129/Sv-derived intestinal stem cells failed to survive during crypt purification (see "Discussion"). The absence of detectable morphologic or histological abnormalities in the developing and adult intestine indicates that the effect is cell-autonomous: i.e. B6-ROSA26 stem cells survive and are able to support a normal level of production of epithelial descendants.

Assaying -Catenin Signaling in Chimeric-Transgenic Mouse Intestine-- To obtain direct evidence for enhanced -catenin signaling in 129/Sv(Lef-1/-cat) small intestinal epithelial cells, SYBR Green-based qRT-PCR was used to examine expression of a known downstream target of -catenin, E-cadherin (Cdh1, see Refs. 22, 42, 43). RNA was prepared from the small intestines of E18.5 B6-ROSA26129/Sv(Lef-1/-cat) and normal chimeras (n = 5/group). Levels of E-cadherin mRNA were first normalized to an internal reference mRNA (GAPDH) and the normalized value from each B6-ROSA26129/Sv(Lef-1/-cat) intestine as compared with the normalized values from age-matched normal chimeric intestinal RNA. The results revealed that E-cadherin mRNA levels are elevated an average of 4-fold in E18.5 B6-ROSA26129/Sv(Lef-1/-cat) intestines (Fig. 5A).

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Effects of Lef-1/-cat on E-cadherin, Apc, Axin, and Cox-2 expression. A, qRT-PCR study of E18.5 small intestine from chimeric-transgenic mice. Values (means ±S.E.) are expressed relative to age-matched normal chimeras (n = 5/group). B, qRT-PCR study of Apc and Axin expression in the developing intestine of normal B6 mice. Data have been normalized to an internal GAPDH mRNA control, and levels of Apc and Axin mRNA were expressed relative to the E13.5 time point. C, multilabel immunohistochemical study of a jejunal villus from an E18.5 chimeric-transgenic mouse. The villus has been sectioned perpendicular to the intervillus epithelium-villus axis. Apc levels are diminished in LacZ-negative 129/Sv(Lef-1/-cat) epithelium relative to the adjacent B6-ROSA26 epithelium. D, immunohistochemical study of a jejunal villus from an E18.5 normal chimeric mouse. Levels of immunoreactive Apc are similar in adjacent 129/Sv and B6-ROSA26 epithelium. Bars = 25 µm.

Expression of Apc and Axin Is Repressed by Lef-1/-Cat-- Apc regulates cellular -catenin levels by targeting it for degradation (44). A qRT-PCR study of RNA prepared from E13.5, E14.5, E15.5, E16.5, E17.5, E18.5, P1, and P14 normal B6 mouse small intestines (n = 3-5 animals/time point) established that Apc mRNA levels rise modestly as Lef-1 expression diminishes (compare Fig. 5B with Fig. 1). At P1, the level of Apc mRNA is 51-fold greater than that of Lef-1 mRNA.

Axin binds to LRP-5 in the Wnt receptor complex (45) and acts as an intracellular scaffold for Gsk-3, Apc, and -catenin assembly. This has led to the proposal that Axin is instrumental in coordinating the intracellular response to the extracellular Wnt signal (46). Unlike Apc mRNA, small intestinal levels of Axin mRNA remain constant from E13.5-P14 (Fig. 5B).

Because both Apc and Axin act to regulate -catenin-mediated signaling, we wondered how they responded to conditions in which there was an engineered forced enhancement of signaling. qRT-PCR analysis of RNAs prepared from E18.5 chimeric-transgenic mice representing four different ES cell lines revealed that steady state Apc mRNA concentrations were on average 9-fold lower as compared with normal chimeric small intestines (p < 0.05; Fig. 5A). These results were confirmed by immunohistochemistry. Frozen sections from E18.5 chimeric-transgenic small intestine were incubated with polyclonal antibodies that recognize epitopes in the C-terminal domain of Apc (31) and with antibodies raised against E. coli -galactosidase. Levels of immunoreactive Apc were markedly lower in LacZ-negative 129/Sv(Lef-1/-cat) epithelium as compared with adjacent LacZ-positive B6-ROSA26 epithelium (n = 5 mice/ES cell line; e.g. Fig. 5C). In contrast, there were no detectable differences in the level of immunoreactive Apc in juxtaposed 129/Sv and B6-ROSA26 epithelium of age-matched normal chimeras (Fig. 5D). Axin mRNA levels were suppressed 2-fold in the intestines of E18.5 chimeric-transgenic mice representing four different ES cell lines (p < 0.05; Fig. 5A).

The finding that both Apc and Axin are repressed under conditions of augmented -catenin signaling is novel. However, the mechanism remains obscure. Neither Axin nor Apc contains demonstrable Lef-1/Tcf binding sequences. At first glance, it would appear that suppression of Apc would further exacerbate enhanced signaling by limiting degradation of endogenous -catenin. Because Axin binds to the Wnt receptor complex and is thought to participate in transducing the extracellular signal, it is possible that a heretofore unappreciated negative feedback loop acts to dampen -catenin-mediated signaling by reducing the availability of the Axin.

Analysis of Cyclooxygenase-2 Expression-- At present, the precise relationship between cyclooxygenase-2 (Cox-2) expression and -catenin signaling is unclear. Forced expression of Cox-2 in rat intestinal epithelial cells has been shown to induce apoptosis (47). However, genetic and pharmacologic experiments in mice indicate a relationship between the loss of Cox-2 function and reduced intestinal tumorigenesis associated with Apc mutations (48, 49). Other studies have reported that the loss of Apc and/or elevated levels of -catenin produce increased Cox-2 expression in the intestine (50, 51).

We used qRT-PCR to determine whether Cox-2 gene (Ptgs2) expression was affected by the enhanced signaling generated by Lef-1/-cat and/or by the concomitant reduction in Apc. Fig. 5A shows that Cox-2 mRNA levels in E18.5 B6-ROSA26129/Sv(Lef-1/-cat) small intestine are indistinguishable from those in age-matched B6-ROSA26129/Sv controls. These results are compatible with recent cell culture studies indicating that Cox-2 is not a downstream target of Wnt-3 signaling (52). Moreover, the promoter region of Cox-2 contains a single imperfect cAMP-response element-binding protein site (GACCTCA) that makes it a candidate for regulation by a -catenin-responsive, Lef-1-independent mechanism (53).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Studies of normal chimeric mice without genetically engineered perturbations in Wnt/-catenin signaling indicate that establishment of the stem cell hierarchy in developing crypts is associated with the selection of one population of stem cells belonging to one genotype over stem cells of the other genotype. This selection is highly efficient, proceeding to complete the elimination of one cellular genotype, and appears to occur in a stochastic, crypt-autonomous fashion. The simplest explanation for this purification process is that all active stem cells in a given mature crypt are ultimately derived from a single selected active progenitor.

A role for -catenin-mediated signaling through Tcf-4 in maintaining multipotent stem cells in the developing small intestine was first suggested when Tcf-4^/ knockout mice were found to abruptly lose proliferative activity in their intervillus epithelium (19). Our analysis of chimeric-transgenic mice provides evidence that stimulating -catenin-mediated signaling through forced expression of a Lef-1/-cat fusion protein during a time in development when endogenous Lef-1 gene expression is normally falling greatly influences stem cell selection in nascent polyclonal crypts of Lieberkühn. In B6-ROSA26129/Sv(Lef-1/-cat) chimeras, where Lef-1/-cat is expressed in 129/Sv intervillus epithelium and nascent crypts under the control of transcriptional regulatory elements known to function in the multipotent stem cell, selection of stem cells capable of giving rise to descendant lineages is dramatically skewed toward the B6-ROSA26 progenitor: i.e. the progenitor whose normal developmentally programmed level of -catenin-mediated signaling has not been genetically manipulated. We postulate that forestalling the late fetal decline in Lef-1 by the engineered expression of Lef-1/-cat results in augmented cellular -catenin signaling and "removal" of the Lef-1/-cat-producing progenitor. As a result, surviving crypts in the adult small intestine of these chimeras are all populated exclusively by B6-ROSA26 epithelium. The death response only needs to be targeted to the small subset of 129/Sv intervillus epithelial progenitors to result in subsequent elimination of all vestiges of this genotype from the developing polyclonal crypt epithelium.

Interestingly, the apoptotic response induced by Lef-1/-cat is not associated with enhanced cell proliferation. There is precedence for enhanced -catenin signaling stimulating apoptosis during development. For example, Ahmed et al. (54) have shown that neuronal apoptosis is a consequence of augmented Armadillo (-catenin homolog) signaling in the Drosophila eye. Precise regulation is required; reduced Armadillo signaling results in aberrant neuronal differentiation (54). In addition, a conditional knockout of Apc in the neural crest of mice resulted in massive apoptosis of cephalic and cardiac neural crest cells during development (55). Apoptosis occurred in regions where intracellular -catenin accumulated. Moreover, 5-bromo-2'-deoxyuridine labeling studies indicated that proliferation was unaffected.

If an absolute level of -catenin signaling is required for stem cell purification and if a nascent crypt contains three or more mitotically active progenitors, a mechanism is needed to explain how the selection is coordinated to yield one progenitor. The Wnt molecule is a morphogen (56-58). Lickert et al. (59) have profiled Wnt expression in the mouse intestine and found that six family members (Wnt2a, 4, 5a, 5b, 6, and 11) are expressed at E14.5. Expression subsequently varies as a function of location along the proximal-distal axis and as a function of developmental stage (59). We postulate that during development, cells exposed to concentrations of the Wnt morphogen above a critical threshold should have abnormally high levels of -catenin signaling and phenocopy the apoptotic response of stem cells to forced expression of Lef-1/-cat. In this proposed scheme, cells exposed to levels of Wnt morphogen below some critical threshold should have low levels of -catenin signaling and phenocopy the differentiation response of progenitors lacking Tcf-4. Cells exposed to an "adequate" threshold dose of morphogen are able to maintain levels of -catenin-mediated signaling sufficient for sustained proliferation, leading to their predominance over other progenitors in the nascent crypt. Thereafter, once the stem cell hierarchy is established, all active multipotent stem cells derived from the selected progenitor are able to survive having adapted to a level of -catenin signaling.

Finally, it is important to consider the general paradigm that expression of stem cell features is highly dependent upon cellular and molecular elements present in its niche (60). As noted above, the apoptotic response was specific for 129/Sv(Lef-1/-cat) cells in developing polyclonal crypts. Moreover, the small intestines of adult B6-ROSA26129/Sv(Lef-1/-cat) mice lack 129/Sv epithelium but are similar in size to normal chimeras. This suggests that the ultimate size of the small intestine's active stem cell population was not greatly affected and that Lef-1/-cat expression during crypt morphogenesis does not irrevocably perturb the crypt stem cell niche.

	ACKNOWLEDGEMENTS

We thank Sabrina Wagoner, David O'Donnell, and Maria Karlsson for invaluable technical assistance.

	FOOTNOTES

^* This work was supported by Grant DK30292 from the National Institutes of Health.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Present address: Dept. of Dermatology, Cell and Developmental Biology, Oregon Health and Sciences University, Portland, OR 97201.

To whom correspondence should be addressed: Dept. of Molecular Biology and Pharmacology, Washington University School of Medicine, 660 S. Euclid, Campus Box 8103, St. Louis, MO 63110. Tel.: 314-362-7243; Fax: 314-362-7047; E-mail: jgordon@molecool.wustl.edu.

Published, JBC Papers in Press, February 19, 2002, DOI 10.1074/jbc.M200184200

	ABBREVIATIONS

The abbreviations used are: HMG, high mobility group; -cat, -catenin; RT, reverse transcription; qRT-PCR, real time quantitative reverse transcriptase-PCR; X-gal, 5-bromo-4-chloro-3-indolyl--D-galactopyranoside; ES, embryonic stem; PLP, periodate-lysine-paraformaldehyde; PBS, phosphate-buffered saline; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; E, embryonic day; P, postnatal day.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES