Anteroposterior Gradient of Epithelial Transformation during Amphibian Intestinal Remodeling:...

DEVELOPMENTAL BIOLOGY 192, 149–161 (1997)ARTICLE NO. DB978749

Anteroposterior Gradient of EpithelialTransformation during Amphibian IntestinalRemodeling: Immunohistochemical Detectionof Intestinal Fatty Acid-Binding Protein

Atsuko Ishizuya-Oka,*,1 Shuichi Ueda,* Sashko Damjanovski,†Qing Li,† Vivia C.-T. Liang,† and Yun-Bo Shi†*Department of Histology and Neurobiology, Dokkyo University School of Medicine,Mibu, Tochigi 321-02, Japan; and †Laboratory of Molecular Embryology,National Institute of Child Health and Human Development,National Institutes of Health, Bethesda, Maryland 20892-5431

To determine whether the remodeling of the well-organized intestinal epithelium during amphibian metamorphosis isregionally regulated along the anteroposterior axis of the intestine, we raised a polyclonal antibody against the Xenopuslaevis intestinal fatty acid-binding protein (IFABP), which is known to be specifically expressed in intestinal absorptive cells,and examined immunohistochemically the differentiation, proliferation, and apoptosis of the epithelial cells throughout X.laevis small intestine. During pre- and prometamorphosis, IFABP-immunoreactive (ir) epithelial cells were localized onlyin the anterior half of the larval intestine. At the beginning of metamorphic climax, apoptotic cells detected by nick end-labeling (TUNEL) suddenly increased in number in the entire larval epithelium, concurrently with the appearance of adultepithelial primordia. Subsequently, the adult primordia in the anterior part of the intestine developed more rapidly byactive cell proliferation than those in the posterior part, and replaced the larval epithelial cells earlier than those in theposterior part. IFABP-ir cells in the adult epithelium were first detectable at the tips of newly formed folds in the proximalpart of the intestine. Thereafter, IFABP expression gradually progressed both in the anteroposterior direction and in thecrest–trough direction of the folds. These results suggest that developmental processes of the adult epithelium in the X.laevis intestine are regionally regulated along the anteroposterior axis of the intestine, which is maintained throughoutmetamorphosis, and along the trough–crest axis of the epithelial folds, which is newly established during metamorphosis.Furthermore, the regional differences in IFABP expression along the anteroposterior axis of the intestine were reproducedin organ cultures in vitro. In addition, IFABP expression was first down-regulated and then reactivated in vitro when theanterior part, but not the posterior part, of the larval intestine was treated with thyroid hormone (TH) for extended periods.Therefore, it seems that, in addition to TH, an endogenous factor(s) localized in the intestine itself with an anteroposteriorgradient participates in the development of the adult epithelium during amphibian metamorphosis. q 1997 Academic Press

INTRODUCTION such as the clustered homeobox genes (Wright et al., 1989;Reuter and Scott, 1990; Reuter et al., 1990; Pollock et al.,1992; Yokouchi et al., 1995). In mammals, even within theRecently, there has been a growing body of evidence thatsmall intestine, it has been reported that differentiation oforgan-specific development of the digestive system is re-epithelial cells is regionally regulated along the anteropos-gionally regulated along the anteroposterior axis of the bodyterior axis of the intestine (Cohn et al., 1992; Crossmannwhich is established during early development by geneset al., 1994; Simon et al., 1995), although the molecularmechanisms of this axial patterning remain unknown.

During a short period of amphibian metamorphosis, the1 To whom correspondence should be addressed. Fax: /81-282-86-1463. E-mail: [email protected]. small intestine is rapidly remodeled from the larval to adult

149

0012-1606/97 $25.00Copyright q 1997 by Academic PressAll rights of reproduction in any form reserved.

AID DB 8749 / 6x31$$$$41 11-06-97 22:25:59 dbas

150 Ishizuya-Oka et al.

form. It is interesting, from a developmental biology stand- sion within the small intestine, observed during spontane-ous metamorphosis, can be reproduced in isolated intestinalpoint, to determine whether the remodeling of a well-orga-

nized intestine during amphibian metamorphosis is also fragments in vitro. To answer these questions, we organo-typically cultured tissue fragments isolated from the proxi-regionally regulated along this axis. However, little atten-

tion has been directed so far to regional differences during mal and distal parts of the small intestine of X. laevis tad-poles or frogs according to the published method (Ishizuya-intestinal remodeling at the cellular level, although the long

larval intestine is generally known to shorten during meta- Oka and Shimozawa, 1991), and chronologically examinedIFABP expression in vitro in the presence or absence of TH.morphosis (Marshall and Dixon, 1978; Ishizuya-Oka and

Shimozawa, 1987). At the cellular level, the epithelial lar- The results provide evidence for an organ autonomous re-sponse to TH, depending on the anteroposterior location ofval-to-adult transformation of the amphibian intestine is

mainly caused by (1) programmed cell death (apoptosis) of the intestinal fragments.the primary larval epithelium (Ishizuya-Oka and Shimo-zawa, 1992a), and (2) proliferation and differentiation of thesecondary adult epithelial cells (Hourdry and Dauca, 1977; MATERIALS AND METHODSYoshizato, 1989; Shi and Ishizuya-Oka, 1996). Recently, wehave shown that apoptosis of larval epithelial cells prog-

Tissue Preparation for Immunohistochemistryresses concurrently with active proliferation of adult cellsin the proximal part of Xenopus laevis small intestine dur-

Tadpoles of the South African clawed frog (X. laevis) from stageing the early period of metamorphic climax (Ishizuya-Oka 50 (about 2 weeks after fertilization) to 66 (about 2 months afterand Ueda, 1996), but the other parts of the intestine have fertilization and just after completion of metamorphosis) were pur-not yet been examined. chased from a commercial source in Hamamatsu, Japan. Tissue

In the present study, to determine whether epithelial lar- fragments were isolated from the proximal part just behind the bileval-to-adult transformation is regionally regulated during duct junction of the small intestine, the middle part between outer

and inner spirals, and the distal part just before the large intestineamphibian metamorphosis, we chronologically examinedof the tadpoles. Developmental stages were determined accordingapoptosis of larval epithelial cells and proliferation of adultto the external criteria of Nieuwkoop and Faber (1967) except forepithelial cells throughout X. laevis small intestine, by us-stages 59–61, when the external criteria do not always agree withing the TdT-mediated dUTP–biotin nick end-labeling (TU-the development of the small intestine itself. The stages 59–61NEL) method (Gavrieli et al., 1992) and labeling with bro-were determined according to the histological criteria in the proxi-

modeoxyuridine (BrdU), an analogue of thymidine (Gratz- mal part of the intestine as follows: the primordia of adult epithe-ner, 1982), respectively. In addition, we studied epithelial lium cannot be detected at stage 59, the primordia can be firstdifferentiation by analyzing the expression of the intestinal recognized as small islets at stage 60, and the growing primordiafatty acid-binding protein (IFABP) in X. laevis intestine. The protrude into the connective tissue at stage 61 (Ishizuya-Oka andexpression of IFABP mRNA can be detected in X. laevis Ueda, 1996).intestine from early developmental stages (the mid to latetailbud period) (Shi and Hayes, 1994; Henry et al., 1996) andis strictly localized in the absorptive epithelial cells in the Nick End-Labeling of Apoptotic DNAanterior half of the intestine, except for the early period of

DNA fragmentation associated with apoptosis was detected ac-metamorphic climax when its mRNA cannot be detectedcording to the TUNEL method of Gavrieli et al. (1992). Briefly,(Ishizuya-Oka et al., 1994). Thus, IFABP is a useful tooltissue fragments were fixed in 4% paraformaldehyde at 47C for 4for studying the origin and differentiation of the absorptiveh and embedded in paraffin. Sections cut at 5 mm were treated withepithelial cells in amphibian small intestine. In this study,20 mg/ml proteinase K (Boehringer Mannheim) at room temperature

we generated a polyclonal anti-Xenopus IFABP antibody for 15 min, and immersed in 1% H2O2 at room temperature for 15and, by immunohistochemistry with this antibody, pre- min to inactivate endogenous peroxidase. They were then incu-cisely examined the expression of IFABP in both the larval bated with TdT buffer (100 mM potassium cacodylate, pH 7.2, 2and adult epithelia of X. laevis small intestine throughout mM cobalt chloride, 0.2 mM dithiothreitol; GIBCO-BRL) con-

taining 0.3 equivalent U/ml terminal deoxynucleotidyl transferasemetamorphosis, paying special attention to regional differ-(TdT; GIBCO-BRL) and 0.04 mM biotinylated dUTP (Boehringerences along the anteroposterior axis of the intestine.Mannheim) at 377C for 90 min. The reaction was terminated byIn the amphibian intestine, the IFABP gene is known asplacing the slides in a washing solution (300 mM sodium chloride,a late thyroid hormone (TH) response gene (Shi and Brown,30 mM sodium citrate) for 15 min at room temperature. After the1993), and its mRNA has been shown to be temporarytreatment with 10% normal rabbit serum for 10 min, the sectionsdown-regulated by treating X. laevis tadpoles with TH inwere incubated with peroxidase-labeled streptavidin (Nichirei) for

vivo (Shi and Hayes, 1994). However, the expression of 30 min and then with 0.02% 3,3*-diaminobenzidine–4HCl (DAB)IFABP has not yet been analyzed in the isolated intestine and 0.006% H2O2 for 5 min. As a positive control, some sectionsin vitro. The questions arise as to whether IFABP expression were pretreated with 2 U/ml DNase I (Sigma) in 30 mM Tris buffer,is organ-autonomously regulated by the inductive action of pH 7.4, for 10 min before the treatment with TdT. The omission

of either TdT or dUTP gave completely negative results (notTH and whether the distribution pattern of IFABP expres-

Copyright q 1997 by Academic Press. All rights of reproduction in any form reserved.

AID DB 8749 / 6x31$$$$41 11-06-97 22:25:59 dbas

151Axial Gradient of Xenopus Intestinal Remodeling

shown). At least five tadpoles were examined for each develop- Total RNA (5 mg) was electrophoresed through a 1% agaroseformaldehyde gel and blotted onto a GeneScreen membrane (NEN)mental stage.as described (Ranjan et al., 1994). The membrane was UV irradiated(120 mJ, Stratalinker 1800, Stratagene), baked for 1 h at 807C undervacuum, and then stained with methylene blue to verify RNA qual-Staining with Methyl Green–Pyronin Yity and equal RNA loading (Herrin and Schmidt, 1988). Hybridiza-

To identify adult epithelial primordia, some of the sections pre- tion was carried out overnight at 427C in 10% dextran sulfate,pared for the nick end-labeling were stained with methyl green– 50% formamide, 51 Denhardt’s, 51 SSPE, 0.1% SDS, 100 mg/mlpyronin Y (Muto) as described (Ishizuya-Oka and Ueda, 1996). The denatured salmon sperm DNA and [32P]-dCTP–IFABP–cDNAadult primordia were intensely stained red because of their RNA- probe labeled using a random-priming method (Amersham). Therich cytoplasm, whereas the larval epithelial cells at and after stage filter was washed three times at room temperature with 21 SSC60 were stained very weakly. and 0.2% SDS, and twice for 20 min each at 657C with 0.251 SSC

and 0.2% SDS.

BrdU ImmunohistochemistryOverproduction of X. laevis IFABP in EscherichiaTo examine regional differences in cell proliferation when thecoli and Antibody Preparationadult epithelial cells are replacing the larval ones, tissue fragments

were isolated from the proximal and distal parts of the intestine The cDNA encoding the full-length IFABP (Shi and Hayes, 1994)from the same tadpoles at stages 60–63. For labeling of S-phase was PCR-amplified by using two primers, each containing a BamHIcells, they were immersed in a solution of 20 mM 5-bromo-2*-deoxy- site at the 5*-end. The PCR fragment was cloned in-frame intouridine (BrdU; Amersham) in 60% Leibovitz 15 medium (GIBCO- the BamHI site of the pET-15b vector (Novagen) to produce anBRL) at 267C for 2 h. They were then fixed in 70% ethanol at 47C overexpression vector for the IFABP with an N-terminal histidinefor 4 h, embedded in paraffin, and sectioned at 5 mm. The sections tag. The clone was verified by sequencing.were treated first with 1% H2O2 and 10% normal rabbit serum as To overproduce the fusion protein, the expression vector abovedescribed above, and then with anti-BrdU monoclonal antibody was transformed into E. coli strain BL21 (DE3) according to the(SANBIO) at room temperature for 1 h. They were then visualized manufacturer’s instruction (Novagen). Five milliliters of LB me-by sequential incubations with streptavidin–biotin–peroxidase dium was inoculated with the transformed cells and allowed tocomplex and DAB/H2O2 as described above. As a negative control, grow to saturation overnight. The next morning this culture wassome sections were processed according to the standard procedure added to 1 liter of fresh LB. The cells were grown to an OD600 ofexcept that the treatment with anti-BrdU antibody was omitted. 0.6–0.8 and then induced with 0.4 mM isopropyl thiogalactosideNo control sections showed any signals (not shown). The labeling (IPTG) for 2 h. The cells were subsequently pelleted and storedindices were calculated as the ratio of labeled cells to total epithe- frozen at 0707C. After thawing, the cells were resuspended into 10lial cells in more than 10 sections randomly selected for each speci- ml of binding buffer containing 20 mM Tris–HCl, pH 7.9, 500 mMmen. More than 3 tadpoles were examined for each experimental NaCl, 5 mM imidazole, and 1 mg/ml lysozyme. After incubationpoint. on ice for 15 min, the cells were lysed by sonication.

Overexpression of IFABP in E. coli led to the production of insol-uble IFABP. Therefore, IFABP was purified from the insoluble frac-

Organ Culture tion, i.e., the inclusion bodies, by using the His-Bind metal chela-tion resin column (Novagen) (Stolow et al., 1996). The purifiedTissue fragments were isolated from the proximal and distalIFABP was electrophoresed on an SDS–protein gel to separate itparts of the small intestine of tadpoles at stage 57 (before metamor-from the remaining contaminants and the gel slice containing thephic climax) and at stage 66, and cultured as described (Ishizuya-purified protein was injected into a rabbit to generate a polyclonalOka and Shimozawa, 1991). Briefly, they were placed on membraneantibody.filters (Type HAWP; Millipore Corp.), and the filters were put on

stainless-steel grids in culture dishes containing 60% Leibovitz 15medium supplemented with 100 IU/ml of penicillin, 100 mg/ml of Western Blot Analysisstreptomycin, and 10% fetal bovine serum which had been treatedwith activated charcoal (Sigma) according to the method of Yoshi- E. coli or tadpole intestinal protein extracts were electrophoresedzato et al. (1980) to remove endogenous thyroid hormones (CTS on a 15% SDS–protein gel and transferred to a nitrocellulose filtermedium). To induce metamorphosis, 3,3*,5-triiodo-L-thyronine, in- (Ausubel et al., 1988). To verify the protein levels, the filter wassulin, and hydrocortisone (Sigma) were added to CTS medium at stained with ponceau S (Sigma) prior to Western blotting using the1008 M, 5 mg/ml, and 0.5 mg/ml, respectively. The culture medium IFABP antibody and the ECL kit (Amersham).was changed every other day for the 7-day treatment at 267C.

Immunohistochemistry for IFABPRNA Isolation and Analysis

Tissue fragments and explants were fixed with 4% paraformalde-hyde in phosphate-buffered saline (pH 7.4), processed for paraffinTadpoles were reared in dechlorinated tapwater and staged indi-

vidually, according to Nieuwkoop and Faber (1967). Total RNA was embedding, and cut at 5 mm. The sections were incubated for 1 hwith the polyclonal antibody (diluted 1:1000) and then visualizedisolated from whole embryos or tadpoles using the guanidinium

thiocyanate method (Ranjan et al., 1994). The synchronized em- by sequential incubation with streptavidin–biotin–peroxidasecomplex and DAB/H2O2 as described above. There was no positivebryos were generated by in vitro fertilization (Dmitrov et al., 1993).


AID DB 8749 / 6x31$$$$41 11-06-97 22:25:59 dbas


larval epithelial cells underwent apoptotic degeneration,and finally reactivated toward the end of metamorphosis(stages 64–66) when adult epithelial cells become differenti-ated (Fig. 1).

To determine whether the IFABP protein levels were sim-ilarly regulated during development, we first generated apolyclonal antibody against an overexpressed IFABP fusionprotein. The antibody recognized the fusion protein pro-duced in E. coli with high specificity (Fig. 2A). Western blotanalysis with the polyclonal antibody revealed that like themRNA, IFABP protein was readily detectable in premeta-morphic intestines from stages 54 to 57 (Fig. 2B). It was alsopresent in the intestines of postmetamorphic (stage 66) andadult frogs, although at lower levels. At the climax of meta-morphosis (at stage 62), the protein was at a minimal level.Subsequently, with the progress of the adult epithelial dif-ferentiation, the protein became detectable again (stage 64).Thus, IFABP mRNA and protein levels were regulated simi-larly during intestinal development. We then used immuno-

FIG. 1. Northern blot showing that IFABP gene is first activated histochemistry to determine the regulation of the IFABParound stages 33/34, immediately prior to hatching at stages 35/ protein levels in relation to tissue transformations in the36, during embryonic stages. The mRNA is then maintained at

intestine during metamorphosis.high levels throughout tadpole development and subsequentlydownregulated during metamorphic climax (stages 60–62). ThemRNA levels are again upregulated toward the end of metamorpho-

IFABP Is Present in Larval Intestinal Epitheliumsis (from stages 64–66). Equal loading was confirmed by stainingat Stage 59 or Earlier When Little Apoptosisthe membrane for total RNA (not shown). Developmental stages

9, 10/11, 16/17, and 23/24 correspond to midblastula transition, Takes Placegastrula, neurula and early tailbud stages, respectively. The posi-

Throughout the larval period up to stage 59, intestinaltions of 18 and 28S RNA are indicated.epithelium consisted of a simple layer of columnar epithe-

staining when preimmune rabbit serum was applied (diluted1:1000) as the specificity control (not shown). More than threefragments were examined for each experimental point.

RESULTS

Both IFABP mRNA and Protein AreDownregulated during Intestinal Metamorphosis

We have previously isolated the X. laevis IFABP gene asa gene regulated by TH in the metamorphosing intestine(Shi and Hayes, 1994). Our earlier studies have shown that FIG. 2. Western blot analysis shows that IFABP protein levelsIFABP mRNA is expressed only in differentiated epithelial are also downregulated during metamorphic climax. (A) The anti-

IFABP antibody recognizes specifically X. laevis IFABP overex-cells of the larval and adult intestines (Shi and Hayes, 1994;pressed in E. coli. Equally amounts of the uninduced (0) and IPTG-Ishizuya-Oka et al., 1994). To determine the earliest stageinduced (/) protein extracts from E. coli transformed with thewhen epithelial cell differentiation occurs, we isolated totalIFABP overexpression vector were loaded onto an SDS–protein gelRNA from embryos at various stages and analyzed theand analyzed by Western blotting. (B) Regulation of IFABP proteinmRNA levels for IFABP. The results in Fig. 1 show that atlevels in the intestine during development. Twenty micrograms ofno stage prior to stages 33/34 was IFABP mRNA detectable,total proteins from the intestine of tadpoles at different stages and

indicating that intestinal epithelial differentiation begins adult frogs was analyzed using the IFABP antibody by Westernaround stages 33/34, just prior to tadpole hatching at stages blotting. Membrane staining with ponceau S indicated that similar35/36. Subsequently, the mRNA reached the highest levels amounts of the proteins were present in each lane (not shown).by stages 44/45 when tadpole feeding begins. Its expression The IFABP proteins (arrows) are of the expected sizes with the

fusion protein (A) slightly larger than the native protein (B).was then repressed to a minimum by stages 60–62 as the


AID DB 8749 / 6x31$$$$41 11-06-97 22:25:59 dbas


lial cells along its entire length. The majority of epithelial ward stage 63 the adult epithelium in the proximal partbecame differentiated into a simple columnar epitheliumcells were absorptive cells possessing a brush border on their

luminal surface. The connective tissue was very thin except (Fig. 5E) and cell proliferation rapidly decreased. On theother hand, in the distal part, morphogenesis of the intesti-in the single longitudinal fold, the typhlosole (Ishizuya-Oka

and Shimozawa, 1987), which appeared just behind the gas- nal folds began after stage 62, and larval epithelial cellsremained until stage 63 (Fig. 5F), when cell proliferation intric region on the side of pancreas around stage 50 (Fig. 3A)

and was present in approximately the first third or fourth the distal part was more active than that in the proximalpart (Fig. 6). By stage 64, the adult epithelium completelyof the larval small intestine before metamorphic climax.

Both apoptotic nuclei labeled by TUNEL and proliferating replaced the larval one in the distal part and began to differ-entiate into the simple columnar epithelium.cells labeled by BrdU were small in number at stage 59 or

earlier. Changes in IFABP expression. During stages 60–63 thelarval epithelial cells in the intestinal anterior region re-The expression of IFABP was detected only in the absorp-

tive epithelial cells, but not in the other types of cells. The mained positive for IFABP as long as they were morphologi-cally intact (Figs. 3E and 3F). In contrast, the adult epithelialstaining was intense in the proximal part of the intestine

including the typhlosole (Fig. 3B), and gradually decreased primordia showed no IFABP-positive cells when they firstappeared at stage 60 (Fig. 3E), and remained negative untilin intensity toward the middle part of the intestine (Fig.

3C). In the distal part close to the large intestine, no positive the completion of replacement of the larval epithelial cellsby the adult ones (Fig. 3F). Around stage 63, the adult epithe-cells were observed (Fig. 3D). In contrast to this regional

difference along the anteroposterior axis of the intestine, lial cells immunoreactive to IFABP became detectable atthe tips of the newly formed intestinal folds in the proximalthe intensity of the staining was independent of the position

of absorptive cells along the trough–crest axis of the typhlo- part of the small intestine (Figs. 3G and 3H).sole.

Reexpression of IFABP as Adult Epithelial CellsCorrelation of IFABP Level with IntestinalDifferentiate during the Second Half ofEpithelial Apoptosis and Cell ProliferationMetamorphic Climax (Stages 64–66)during the First Half of Metamorphic

Climax (Stages 60–63) After the completion of replacement of the larval epithe-lial cells by the adult ones, both apoptotic cells detected byApoptosis of the larval epithelium. At stage 60, apo-

ptotic cells labeled by TUNEL suddenly increased in num- TUNEL and proliferating cells detected by BrdU-labelingbecame small in number in every part of the intestine. To-ber throughout the intestine, but only in the larval epithe-

lium (Figs. 4A and 4B). During subsequent stages, a dramatic ward the end of metamorphosis (stage 66), a new distribu-tion pattern of epithelial cell proliferation and apoptosisloss of larval epithelial cells occurred in the entire intestinal

epithelium. The labeling by TUNEL remained high in the was established in the entire small intestine; BrdU-labeledproliferating cells became localized in the troughs of thelarval epithelium until it finally disappeared around stage

62 in the proximal part of the intestine, and around stage newly formed intestinal folds, while apoptotic cells labeledby TUNEL became localized at their tips as we previously63 in its distal part (Figs. 4C and 4D).

Growth of the adult epithelial primordia. Concomi- reported in the proximal part of the intestine (Ishizuya-Okaand Ueda 1996).tantly with the sudden increase in the number of apoptotic

cells in the larval epithelium around stage 60, the adult Adult epithelial cells immunoreactive to IFABP, whichwere first localized at the tips of the developing intestinalprimordia were first detected as round small islets con-

sisting of a few undifferentiated cells intensely stained with folds in the proximal part of the intestine, gradually in-creased in number in two directions: toward the middle partmethyl green–pyronin Y in every part of the intestine exam-

ined (Figs. 5A and 5B). However, at this stage, the primordia of the intestine and toward the troughs of the intestinalfolds. By stage 66, the regional differences of IFABP expres-were detectable in the proximal part but not the distal part

of only some tadpoles. Then, the adult primordia rapidly sion along the two axes became evident; the signals de-creased in intensity from the proximal to the middle partsincreased in size and number, and protruded into the devel-

oping connective tissue. At stage 61 the sizes of the adult of the intestine just like those in the intestine during thelarval period, and from the crests to the troughs of the intes-primordia were always larger in the proximal part (Fig. 5C)

than those in the distal part (Fig. 5D) because of more active tinal folds (Figs. 7A and 7B). In the distal part of the intes-tine, the adult epithelium remained negative for IFABPcell proliferation during stages 60–61 (Fig. 6). In contrast,

the ratio of larval cells to adult cells was higher in the distal throughout the metamorphosis (Fig. 7C).The correlation of the larval-to-adult epithelial replace-part than that in the proximal part. At stage 62 morphogene-

sis of intestinal folds began in the proximal part. The degen- ment with the expression of IFABP in X. laevis small intes-tine during spontaneous metamorphosis described above iserating larval epithelium was almost completely replaced

by the adult epithelium except at the tips of the folds. To- schematically summarized in Fig. 8.


AID DB 8749 / 6x31$$$$41 11-06-97 22:25:59 dbas


FIG. 3. Immunohistochemical detection of IFABP in X. laevis small intestine during metamorphosis. (A) Longitudinal section, at stage51, between the proximal part of the intestine including the typhlosole (Ty), which runs longitudinally on the side of pancreas (Pa), andthe gastric region (Ga). Only the larval intestinal epithelium (LE) facing the lumen (L) is strongly positive. Before metamorphic climax(at stage 57) the staining of the epithelium is intense in the proximal part of the intestine (B), very weak (arrows) in the middle part (C),and not detectable in the distal part (D). At stage 60, in the proximal part of the intestine (E), the larval epithelial cells remain positive,but the adult epithelial primordia (arrowheads) in the epithelial basal region facing the connective tissue (CT) are negative. Then, in theproximal part at stage 61 (F), only a small number of the larval epithelial cells localized at the tip of the typhlosole remain positive, butthe adult epithelium (AE) is negative. (G) Longitudinal section between the proximal (P) and distal (D) parts of the intestine at stage 63.Positive cells are detectable only in the adult epithelium of the proximal part (arrows). In the proximal part at this stage (H), the stainingis the strongest at the tips of developing intestinal folds (arrows). Bars, 20 mm.


AID DB 8749 / 6x31$$8749 11-06-97 22:25:59 dbas


FIG. 4. Apoptosis detected by TUNEL in the small intestine during metamorphosis. At stage 60, in both the proximal (A) and distalparts (B) of the small intestine, labeled nuclei and nuclear fragments with various sizes (arrows) are numerous in the larval epithelium(LE) facing the lumen (L). At stage 62, in the proximal part of the intestine (C), almost all of the epithelial cells are adult-type (AE), andthe labeling becomes very rare. On the other hand, in the distal part at this stage (D), the labeling is detected in the larval epithelial cells(arrows) which are localized in the upper region of developing folds (Fo). Bars, 20 mm.

Regulation of IFABP Expression in Organ Cultures epithelial cells remained positive until they finally disappeared,whereas the adult epithelial cells were negative when theyReveals the Organ-Autonomous Nature ofwere replacing the larval cells (Fig. 9E). Then, by the seventhIntestinal Transformationday of cultivation, the adult epithelial cells completely replacedThe expression of IFABP in the proximal part of X. laevisthe larval cells and differentiated into the simple columnar

small intestine before and after metamorphosis was main-epithelial cells expressing IFABP (Fig. 9F). However, in explants

tained in vitro when explants isolated from both the larval isolated from the distal part of the larval intestine, no expres-intestine at stage 57 (Fig. 9A) and the adult intestine at stage sion of IFABP was detectable throughout the cultivation (Fig.66 were cultured in the control medium deprived of TH 9G). Thus, the intestinal metamorphosis is organ-autonomous.(Fig. 9B). In contrast, in explants isolated from the distalpart of both the larval and adult intestines, no IFABP-posi- DISCUSSIONtive cells were detected in vitro (Figs. 9C and 9D).

Regional Regulation of the Larval-to-AdultIn the medium including TH, chronological changes inTransformation of the Intestinal EpitheliumIFABP expression in explants isolated from the proximal part

of the larval intestine were similar to those in the same part The present study reports for the first time the antero-posterior progress of development of the adult epitheliumduring spontaneous metamorphosis described above. The larval


AID DB 8749 / 6x31$$$$41 11-06-97 22:25:59 dbas


FIG. 5. Development of the adult intestinal epithelium during metamorphosis, stained with methyl green–pyronin Y. In both theproximal (A) and distal parts (B) of the small intestine at stage 60, small primordia of adult epithelium (arrows) appear as round isletsbetween the larval epithelium (LE) facing the lumen (L) and the connective tissue (CT). Then, at stage 61, the adult epithelium (AE) inthe proximal part of the intestine (C) grows more rapidly and is larger than that (arrows) in the distal part (D). At stage 63 the adultepithelium completely replaces the larval epithelium and becomes simple columnar in the proximal part of the intestine (E), whereas thelarval epithelium remains in the crests of developing folds (Fo) in the distal part (F). Bars, 20 mm.


AID DB 8749 / 6x31$$8749 11-06-97 22:25:59 dbas


in the small intestine during amphibian metamorphosis.We showed that the adult epithelial primordia in the ante-rior part of the intestine grow more rapidly than those inthe posterior part during the early period of metamorphicclimax, even if the adult primordia appear almost simulta-neously (around stage 60) throughout the intestine. In addi-tion, we found by using the BrdU immunohistochemistrythat the differences in adult epithelial growth are caused byhigher levels of DNA synthesis in the anterior part com-pared to the posterior part during stages 60–61. In contrast,the TUNEL analysis indicates that massive apoptosis of thelarval epithelial cells starts almost simultaneously (aroundstage 60) throughout the intestine and continues in the en-tire larval epithelium until the final disappearance of thelarval cells. In other amphibian organs, there have been fewreports on regional differences in the larval-to-adult conver-sion during metamorphosis except for the skin. In X. laevis,FIG. 6. BrdU-labeling indices of the epithelium in the proximalit has been reported that cell growth and differentiation(h) and distal (…) parts of X. laevis small intestine during stagesof the adult dorsal muscle beneath the skin proceed in an60–63, when the adult epithelium is rapidly replacing the larvalanterior-to-posterior direction, but the direction of cellepithelium. Cell proliferation during stages 60–61 is more active

in the proximal part than that in the distal part. Then, following death gradient is opposite to this (posterior-to-anterior di-the completion of the epithelial cell replacement, the labeling indi- rection) (Nishikawa and Hayashi, 1994, 1995). In the presentces in the proximal part decrease earlier than those in the distal study, although the ratio of apoptotic larval cells to totalpart. Each value represents the mean { SE of more than three epithelial cells was higher in the distal part of the intestinetadpoles. than that in the proximal part just like that during the dorsal

muscle remodeling (Nishikawa and Hayashi, 1994), no pos-

FIG. 7. Distribution pattern of IFABP in X. laevis small intestine just after the completion of metamorphosis (stage 66). (A) The proximalpart of the intestine. Only the epithelial absorptive cells (AE) facing the lumen (L) are strongly positive except for the troughs (Tr) ofnewly formed intestinal folds (Fo). The other types of cells such as goblet cells (Go) and connective tissue cells (CT) are negative. Thestaining of the epithelial absorptive cells is much weaker in the middle part (B), and not detectable in the distal part (C). Bars, 20 mm.


AID DB 8749 / 6x31$$$$41 11-06-97 22:25:59 dbas


FIG. 8. Correlation between the epithelial replacement and expression of IFABP in X. laevis small intestine during spontaneous metamor-phosis. Around stage 60, the adult epithelial primordia appear as small islets throughout the intestine. During stages 61–63, the epithelialreplacement progresses in the anterior-to-posterior direction and in the crest-to-trough direction of developing intestinal folds. In thelarval epithelium, IFABP is distributed with an anteroposterior gradient. On the other hand, IFABP in the adult epithelial cells becomesdetectable at stage 63 in the crests of the developing folds in the proximal part of the intestine. Then, IFABP expression propagates towardthe middle part of the intestine and toward the lower region of the folds.

terior-to-anterior progression of the larval cell death could agree well with the localization of IFABP mRNA shown inbe detected in the intestine by the TUNEL analysis. These our previous in situ hybridization (ISH) analysis (Ishizuya-results agree well with our previous in vitro study where the Oka et al., 1994). In mammals, many immunohistochemi-adult epithelial primordia developed better in the fragments cal studies have reported that expression of IFABP is strictlyisolated from the more anterior part of the intestine, while regulated along both the anteroposterior and the crypt–vil-the larval epithelium degenerated in every fragment with- lus axes of the adult small intestine (Schields et al., 1986;out any regional specificity (Ishizuya-Oka and Shimozawa, Gordon et al., 1992; Green et al., 1992) and that these two1992b). Therefore, it seems likely that there are some endog- axes are first detected in late fetal period (Rubin et al., 1989;enous factor(s) that are distributed with an anteroposterior Green et al., 1992; Cohn et al., 1992). In contrast, in thegradient and are involved in the development of the adult amphibian small intestine (the present study), regional dif-intestinal epithelium. ferences in the amount of IFABP before metamorphosis

In addition, the present study indicates that the adult were observed only along the anteroposterior axis of theepithelium develops not only along the anteroposterior axis intestine, indicating that this axis is established earlier thanbut also along the trough–crest axis of the newly formed the trough–crest axis of the folds.intestinal folds; the adult epithelial cells first appear in the Although IFABP mRNA in the larval epithelium cannottroughs of the developing folds and then gradually extend be detected during stages 60–62 by ISH (Ishizuya-Oka ettoward their tips by active proliferation. al., 1994), the present study indicates that the protein re-

mains in the larval epithelial cells as long as they are mor-phologically intact, whereas the adult cells have no IFABP

Expression of IFABP during Metamorphosis during this period. It remains uncertain whether the adultcells originate from a very small number of IFABP-negative,The present immunohistochemical results on the local-undifferentiated cells or from the IFABP-positive cellsization of IFABP in the absorptive epithelial cells, with an

anteroposterior gradient before and after metamorphosis, whose IFABP has been rapidly degraded just before stage 60.


AID DB 8749 / 6x31$$$$41 11-06-97 22:25:59 dbas


FIG. 9. In vitro detection of IFABP in intestinal explants organotypically cultured in the absence (A–D) and the presence of TH (E–G).On the fifth day of cultivation, both the larval (LE) and adult epithelia (AE) isolated from the proximal part of the larval intestine at stage57 (A) and the adult intestine at stage 66 (B), respectively, remain positive. In contrast, both the larval and adult epithelia isolated fromthe distal part of the intestines at stage 57 (C) and at stage 66 (D), respectively, remain negative. In the TH-treated explant isolated fromthe proximal part of the larval intestine at stage 57 (E), the epithelial cell replacement is taking place on the fifth day; the larval cells arepositive, but the adult epithelial cells are negative. Thereafter, on the seventh day (F), the adult epithelial cells completely replace thelarval cells and begin to express IFABP. In contrast, in the TH-treated explant isolated from the distal part of the intestine at stage 57(G), no positive cells are detected in the epithelium (E) throughout 7 days of cultivation. CT, connective tissue. Bars, 20 mm.

In any case, at least by stage 60, the epithelial cells are In the adult epithelium, our work indicates that IFABPexpression begins from the proximal part of the intestinedistinctly divided into two types of cells according to the

presence or absence of IFABP. Therefore, IFABP immuno- just after the completion of the epithelial cell replacementand propagates toward the middle part. Thus, not only cellhistochemistry is very useful for distinguishing the larval

cells from the adult cells during the epithelial replacement. proliferation but also differentiation of the adult epitheliumproceeds along the anteroposterior axis. In addition, in thisFurthermore, the present study has shown that the IFABP-

positive larval cells remain until much later stage at the study, the regional differences in IFABP expression alongthe trough–crest axis of intestinal folds, equivalent to thosetips of the intestinal folds and/or the typhlosole than in the

troughs. along the crypt–villus axis in the mammals (Rubin et al.,


AID DB 8749 / 6x31$$$$41 11-06-97 22:25:59 dbas


Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman,1989; Green et al., 1992; Cohn et al., 1992), become detect-J. G., Smith, J. A., and Struhl, K. (1988). ‘‘Current Protocols inable in the adult epithelium during metamorphic climax.Molecular Biology.’’ Wiley, New York.Taken together with the localization of epithelial cell prolif-

Calvert, R., and Pothier, P. (1990). Migration of fetal intestinaleration in the troughs of the folds (Ishizuya-Oka and Ueda,intervillous cells in neonatal mice. Anat. Rec. 227, 199–206.1996), the amphibian intestinal remodeling during meta-

Cheng, H., and Bjerknes, M. (1985). Whole population cell kineticsmorphosis may serve as a good model system for the study and postnatal development of the mouse intestinal epithelium.of the establishment of the crypt–villus axis in the mamma- Anat. Rec. 211, 420–426.lian intestine at the time of birth (Hermos et al., 1971; Cohn, S. M., Simon, T. C., Roth, K. A., Birkenmeier, E. H., and Gor-Cheng and Bjerknes, 1985; Schmidt et al., 1988; Calvert and don, J. I. (1992). Use of transgenic mice to map cis-acting ele-Pothier, 1990), which is believed to be regulated by the ments in the intestinal fatty acid binding protein gene (Fabpi)

that control its cell lineage-specific and regional patterns of ex-surrounding mesenchyme (Dauca et al., 1990; Adams andpression along the duodenal–colonic and crypt–villus axes of theWatt, 1993). The coincidence of the establishment of thegut epithelium. J. Cell Biol. 119, 27–44.trough–crest axis of the intestinal folds with the develop-

Crossmann, M. W., Hauft, S. M., and Gordon, J. I. (1994). Thement of the connective tissue in the amphibian intestinemouse ileal lipid-binding protein gene: A model for studying axialalso suggests the importance of the connective tissue inpatterning during gut morphogenesis. J. Cell Biol. 126, 1547–establishing this axis.1564.

Dauca, M., Bouzigers, F., Colin, S., Kedinger, M., Kekker, J. M.,Schilt, J., Simon-Assmann, P., and Haffen, K. (1990). Develop-Hormone- and Region-Dependent Expressionment of the vertebrate small intestine and mechanisms of cellof IFABP in Vitrodifferentiation. Int. J. Dev. Biol. 34, 205–218.

Dmitrov, S., Almouzni, G., Dasso, M., and Wolffe, A. P. (1993).The present study reports, for the first time, IFABP ex-Chromatin transitions during early Xenopus embryogenesis:pression in isolated intestine in vitro. We have shown thatChanges in histone H4 acetylation and in linker histone type.the anteroposterior gradient of IFABP expression can beDev. Biol. 160, 214–227.maintained in both the larval and adult intestines even if

Gavrieli, Y., Sherman, Y., and Ben-Sasson, S. A. (1992). Identifica-they are separated from the other organs. In addition, wetion of programmed cell death in situ via specific labeling ofhave not only confirmed previous Northern blot analysisnuclear DNA fragmentation. J. Cell Biol. 119, 493–501.

that IFABP gene is a TH-response gene in vivo (Shi and Gordon, J. I., Schmidt, G. H., and Roth, K. A. (1992). Studies ofHays, 1994), but also directly demonstrated TH-regulated intestinal stem cells using normal, chimeric, and transgenicexpression of IFABP at the cellular level. By the action of mice. FASEB J. 6, 3039–3050.TH, IFABP in the larval epithelial cells rapidly decreases Gratzner, H. G. (1982). Monoclonal antibody to 5-bromo- and 5-in amount, while the adult cells begin to express IFABP. iododeoxyuridine: A new reagent for detection of DNA replica-

tion. Science 218, 474–475.Furthermore, the present work indicates that the antero-Green, R. P., Cohn, S. M., Sacchettini, J. C., Jackson, K. E., and Gor-posterior gradient in IFABP expression can be reproduced

don, J. I. (1992). The mouse intestinal fatty acid binding proteinin vitro even if the larval epithelial cells are completelygene: Nucleotide sequence, pattern of developmental and re-replaced by the adult cells. This means that some endoge-gional expression, and proposal product. DNA Cell Biol. 11, 31–nous factor(s) distributed in precursor cells of the adult epi-41.thelium itself, and/or the other tissues surrounding the

Henry, G. L., Brivanlou, I. M., Kessler, D. S., Hemmati-Brivanlou,adult epithelium with an anteroposterior gradient, is in- A., and Melton, D. A. (1996). TGF-b signals and a prepatternvolved in the differentiation of adult epithelial cells. in Xenopus laevis endodermal development. Development 122,

In conclusion, although the larval-to-adult transforma- 1007–1015.tion in the amphibian intestine is triggered by TH, the de- Hermos, J. A., Mathan, M., and Trier, J. S. (1971). DNA synthesisvelopment of the adult epithelium is regionally regulated and proliferation by villous epithelial cells in fetal rats. J. Cellalong both the anteroposterior axis, which is established Biol. 50, 255–258.

Herrin, D. L., and Schmidt, G. W. (1988). Rapid reversible stainingduring early development and is maintained by some endog-of Northern blots prior to hybridization. BioTechniques 6, 196–enous factor(s) in the organ itself, and along the trough–200.crest axis of the intestinal folds, which is newly established

Hourdry, J., and Dauca, M. (1977). Cytological and cytochemicalduring metamorphosis possibly through interactions be-changes in the intestinal epithelium during anuran metamorpho-tween the epithelium and the surrounding mesenchyme.sis. Int. Rev. Cytol. (Suppl.) 5, 337–385.Biochemical analyses of regional differences within the

Ishizuya-Oka, A., and Shimozawa, A. (1987). Development of thesmall intestine will be an interesting future subject. connective tissue in the digestive tract of the larval and metamor-

phosing Xenopus laevis. Anat. Anz. 164, 81–93.Ishizuya-Oka, A., and Shimozawa, A. (1991). Induction of metamor-

REFERENCES phosis by thyroid hormone in anuran small intestine culturedorganotypically in vitro. In Vitro Cell. Dev. Biol. 27A, 853–857.

Ishizuya-Oka, A., and Shimozawa, A. (1992a). Programmed cellAdams, J. C., and Watt, F. M. (1993). Regulation and differentiationby extracellular matrix. Development 117, 1183–1198. death and heterolysis of larval epithelial cells by macrophage-


AID DB 8749 / 6x31$$$$41 11-06-97 22:25:59 dbas


like cells in the anuran small intestine in vivo and in vitro. J. saic patterns of gene expression. Proc. Natl. Acad. Sci. USA 86,1278–1282.Morphol. 213, 185–195.

Schields, H. M., Bates, M. L., Bass, N. M., Best, C. J., Alpers, D. H.,Ishizuya-Oka, A., and Shimozawa, A. (1992b). Connective tissue isand Ockner, R. K. (1986). Light microscopic immunocytochemi-involved in adult epithelial development of the small intestinecal localization of hepatic and intestinal types of fatty acid-bind-during anuran metamorphosis in vitro. Roux’s Arch. Dev. Biol.ing proteins in rat small intestine. J. Lipid Res. 27, 549–557.201, 322–329.

Schmidt, G. H., Winton, D. J., and Ponder, B. A. J. (1988). Develop-Ishizuya-Oka, A., Shimozawa, A., Takeda, H., and Shi, Y.-B. (1994).ment of the pattern of cell renewal in the crypt villus unit ofCell-specific and spatio-temporal expression of intestinal fattychimeric mouse intestine. Development 103, 785–790.acid-binding protein gene during amphibian metamorphosis.

Shi, Y.-B., and Brown, D. D. (1993). The earliest changes in geneRoux’s Arch. Dev. Biol. 204, 150–155.expression in tadpole intestine induced by thyroid hormone. J.Ishizuya-Oka, A., and Ueda, S. (1996). Apoptosis and cell prolifera-Biol. Chem. 268, 20312–20317.tion in the Xenopus small intestine during metamorphosis. Cell

Shi, Y.-B., and Hayes, W. P. (1994). Thyroid hormone-dependentTissue Res. 286, 467–476.regulation of the intestinal fatty acid-binding protein gene duringMarshall, J. A., and Dixon, K. E. (1978). Cell specialization in theamphibian metamorphosis. Dev. Biol. 161, 48–58.epithelium of the small intestine of feeding Xenopus laevis. J.

Shi, Y.-B., and Ishizuya-Oka, A. (1996). Biphasic intestinal develop-Anat. 126, 133–144.ment in amphibians: Embryogenesis and remodeling duringNieuwkoop, P. D., and Faber, J. (1967). ‘‘Normal Table of Xenopusmetamorphosis. Curr. Top. Dev. Biol. 32, 205–235.laevis (Daudin).’’ North-Holland, Amsterdam.

Simon, T. C., Roberts, L. J. J., and Gordon, J. I. (1995). A 20-nucleo-Nishikawa, A., and Hayashi, H. (1994). Isoform transition of con-tide element in the intestinal fatty acid binding protein genetractile proteins related to muscle remodeling with an axial gradi-modulates its cell lineage-specific, differentiation-dependent,ent during metamorphosis in Xenopus laevis. Dev. Biol. 165, 86–and cephalocaudal patterns of expression in transgenic mice.94.Proc. Natl. Acad. Sci. USA 92, 8685–8689.Nishikawa, A., and Hayashi, H. (1995). Spatial, temporal and hor-

Stolow, M. A., Bauzon, D. D., Li, J., Segwick, T., Liang, V. C.-T.,monal regulation of programmed muscle cell death during meta-Sang, Q. A., and Shi, Y.-B. (1996). Identification and characteriza-morphosis of the frog Xenopus laevis. Differentiation 59, 207–tion of a novel collagenase in Xenopus laevis: Possible roles dur-214.ing frog development. Mol. Biol. Cell 7, 14771–1483.Pollock, R. A., Jay, G., and Bieherich, C. J. (1992). Altering the

Wright, C. V. E., Schnegelsberg, P., and De Robertis, E. M. (1989).boundaries of Hox3.1 expression: Evidence for antipodal geneXIHbox8: A novel Xenopus homeo protein restricted to a narrowregulation. Cell 71, 911–923.band of endoderm. Development 106, 787–794.Ranjan, M., Wong, J., and Shi, Y.-B. (1994). Transcriptional repres-

Yokouchi, Y., Sakiyama, J., and Kuroiwa, A. (1995). Coordinatesion of Xenopus TRb gene is mediated by a thyroid hormoneexpression of Abd-B subfamily genes of the HoxA cluster in theresponse element located near the start site. J. Biol. Chem. 269,developing digestive tract of chick embryo. Dev. Biol. 169, 76–24699–24705.89.Reuter, R., Panganiban, G. E. F., Hoffman, F. M., and Scott, M. P.

Yoshizato, K. (1989). Biochemistry and cell biology of amphibian(1990). Homeotic genes regulate the spatial expression of putativemetamorphosis with a special emphasis on the mechanism ofgrowth factors in the visceral mesoderm of Drosophila embryos.removal of larval organs. Int. Rev. Cytol. 119, 97–149.Development 110, 1031–1040.

Yoshizato, K., Kikuyama, S., Shioya, N. (1980). Stimulation of glu-Reuter, R., and Scott, M. P. (1990). Expression and function of thecose utilization and lactate production in cultured human fibro-homeotic genes Antennapedia and Sex combs reduced in theblasts by thyroid hormone. Biochim. Biophys. Acta 627, 23–29.embryonic midgut of Drosophila. Development 109, 289–303.

Rubin, D. C., Ong, D. E., and Gordon, J. I. (1989). Cellular differenti- Received for publication June 10, 1997Accepted September 17, 1997ation in the emerging fetal rat small intestinal epithelium: mo-


AID DB 8749 / 6x31$$$$41 11-06-97 22:25:59 dbas

Anteroposterior Gradient of Epithelial Transformation during Amphibian Intestinal Remodeling:...

Documents

Transcript of Anteroposterior Gradient of Epithelial Transformation during Amphibian Intestinal Remodeling:...