The role of transforming growth factor beta-2, beta-3 in mediating apoptosis in the murine intestinal mucosa

GASTROENTEROLOGY 2002;122;1364-1375

The Role of Transforming Growth Factor Beta-2, Beta-3 in Mediating Apoptosis in the Murine Intestinal Mucosa NICOLE DONKER, KAI SCHMITT, NORBERT SCHUSTER, and KERSTIN KRIEGLSTEIN Anatomy and Cell Biology, University of Saarland, Homburg/Saar, Germany

Background & Aims: Apoptosis is especially relevant in the gastrointestinal tract because the mammalian intestinal mucosa undergoes continual epithelial regeneration. Most recently, we confirmed the proapoptotic role of endogenous transforming growth factor (TGF)-~ in the developing chick retina as well as in chick ciliary, dorsal root, and spinal motor neurons. In the present study, we determined to establish the role of TGF-~2 and TGF-~3 in mediating apoptosis in non-neuronal tissue by analyzing the intestinal mucosa of Tgf~2 +/- and Tgf{33+/heterozygous mice. Methods: Intestinal localization of

TGF-~2 and TGF-~3 isoforms and antiapoptotic molecules Bcl-xL and Bcl-2 was examined immunocytochemically and by Western blot analysis. Apoptosis was detected by enzyme-linked immunosorbent assay and terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling, and proliferation was detected by proliferating cell nuclear antigen stains. Results: TGF-~2 was detected in endocrine cells, whereas TGF-~3 was predominantly found in goblet cells. Programmed cell death was significantly reduced in the intestinal mucosa of Tgf[32+/- and Tgf~3 +/- heterozygous mice. This decrease in apoptosis was accompanied by an increase in villus length; proliferation, however, seemed to remain unchanged. The level of Bcl-xL and Bcl-2 was significantly up-regulated in Tgf(32+/- and Tgf~3 +/- mice. Conclusions: Our data show that TGF-~2 and TGF-~3 play an important role in mediating apoptosis in the intestinal mucosa and regulating apoptosisassociated proteins Bcl-xL and Bcl-2 in vivo.

Poptosis is now recognized as an important process responsible for maintenance of the cellular balance between proliferation and death. Apoptosis is especially relevant in the gastrointestinal tract because the mammalian intestinal mucosa undergoes continual cell turnover.1 Cell proliferation is confined to the crypts, differentiation occurs during migration to the villus, and differentiated enterocytes undergo a process of programmed cell death (PCD). The crypt-to-villus and crypt-to-surface epithelial cuff axes of the mouse gut represent an anatomically well-organized, continuous developmental system. 2 The adult mouse small intestine

A

contains approximately 1.1 million crypts of Lieberktihn. These anatomically well-demarcated, flask-shaped structures represent the proliferative unit of the intestinal epithelium. The crypt stem cells proliferate to produce the epithelial cells (enterocytes, goblet cells, and endocrine cells), which migrate upward and proliferate several times before reaching villi as differentiated epithelial cells. Given the dynamic of this system, where cell proliferation has to be counterbalanced by cell death, the gastrointestinal tract offers an ideal system to study the apoptotic process in vivo. In the gastrointestinal tract, most studies on the renewal of the intestinal epithelium have traditionally focused on the control of cell proliferation, its control by growth factors, and changes that may result in hyperproliferation and carcinogenesis. However, in recent years it became increasingly apparent that the control of cell death is an equally, if not more, important regulator of cell number. Apoptosis is an important mechanism for eliminating excess cells. Thus, both decreased proliferation and/or increased cell death may reduce cell number, whereas increased proliferation and/or decreased death may increase cell number. The intestinal epithelium is an ideal model for investigating the relationship between these 2 processes and their effects within a tissue environment. For example, the small intestine has very high levels of proliferation and cell loss, which occur in a well-defined and polarized topographical organization in which the hierarchy or cellular "age" can be assessed by the position of a cell in the tissue) In normal adult small intestine, a persistent low frequency of apoptosis can be seen in the crypts. This Abbreviations used in this paper: DAB, 3,3'-diaminobenzidine;dUTP, deoxyuridine triphosphate; ELISA, enzyme-linked immunosorbent assay; M.O.M., mouse on mouse; PBA, periodic acid-Schiff reaction; PCD, programmed cell death; PCNA, proliferating cell nuclear antigen; T~R, transforming growth factor ~ receptor; TGF, transforming growth factor; TUNEL, terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling. © 2002 by the American Gastroenterological Association 0016-5085/02/$35.00 doi:10.1053/gast.2002.32991

May 2002

spontaneous apoptosis occurs predominantly in the lower regions of the crypt, the stem cell proliferation zone, and is increased in the context of genotoxic insult) ,4 In addition, it is widely held that in the gastrointestinal tract, cell loss occurs by passive shedding into the gut lumen) Evidence for this, however, is not compelling. Deletion of excess cells by apoptosis may also be responsible for removing effete ceils from the villus tip. The maturation, differentiation, and senescence of enterocytes on the villus is a genetically programmed event) It seems logical that differentiated villus ceils are likely to die by apoptosis and be extruded into the lumen or taken up by macrophages. However, there has been some debate as to whether this is true apoptosis. Recent evidence supports the idea that apoptosis is responsible for controlling the majority of intestinal epithelial cell loss. 1,6 Benedetti et al. 6 conducted a quantitative evaluation on the occurrence of cell death in normal gastrointestinal mucosa of young rats and showed that apoptosis is a rare but constant phenomenon: a mean of 3 apoptotic bodies were observed in duodenal villi. Although studies have shown apoptosis in crypt stem cells, 4 work from different laboratories ~,6-8 indicates that apoptosis occurs in nonproliferative compartments, predominantly in the villus tip cells. Apoptotic cells with intense DNA fragmentation were seen at the villus tip, suggesting that these ceils were exfoliated into the lumen after induction of apoptosis. Engulfment by adjacent epithelial cells, subepithelial macrophages, and lymphocytes has been implicated in the rapid removal of apoptotic cells from the villus tip. v,9 Although a large amount of data on the mechanisms and inducers of apoptosis are available, most of the work has been on cell lines and little information is available on the mechanisms and the inducers involved in apoptosis of the enterocytes in vivo. The regulation of proliferation and the balance between cell proliferation, differentiation, and cellular senescence are maintained through complex interactions of many different factors) ° Most observations made in past years underscored the central importance of regulatory peptides produced by cellular constituents of the gastrointestinal tract in these processes. Growth factors like transforming growth factor (TGF)-[3 are remarkably pleiotropic and seem to play key regulatory functions in a diverse spectrum of biological processes. H These include modulation (either up- or down-regulation) of proliferative activity, cellular differentiation, development of many tissues, formation of extracellular matrix, and, as discovered most recently, apoptosis) 2,13 In mammalian tissues, TGF-~s appear in

TGF-~ MEDIATED APOPTOSIS IN THE MURINE GUT 1365

3 major isoforms (TGF-[31-3); 2 additional forms were found in avian and amphibian tissues (TGF-[34 and 5). All 3 major TGF-[3 isoforms have been found in both the adult and the embryonic intestine. 14-16 TGF-[31 and lesser amounts of TGF-[33 are present in embryonic intestinal epithelium. In mature animals, TGF-[31, TGF-[32, and TGF-[33 may be detected in all gastrointestinal tract tissues, whereby the expression has been demonstrated in both lamina propria cells and the epithelium. All 3 TGF-[3 isoforms were found broadly distributed in the colonic epithelium as well. 15,17 Two TGF-[3 transcripts (2.2 and 1.8 kilobases) have been detected in nontransformed rat intestinal epithelial cell lines (IEC-6). TGF-[3 is a potent inhibitor of proliferation in this crypt cell-like cell line) 8 Little is known about functional effects of TGF-[3s in other cellular elements of the intestinal mucosa. Morphologic changes associated with apoptosis are closely correlated with the expression of specific proteins. TGF-[3s play a well-established role in regulating gene transcription of proteins associated with apoptosis, such as Bcl-2 family members Bcl-2 and Bcl-xL. An in vitro study by Hague et al. ~9 demonstrated that TGF-[3induced apoptosis is preceded by a reduction in p26Bcl-2 protein levels in colorectal adenoma cells. We have shown previously that TGF-132 and TGF-133 not only regulate proliferation but are also actively involved in mediating apoptosis in vivo. 12a3 PCD of chick ciliary, dorsal root, and spinal motor neurons is largely prevented following application of a neutralizing antibody that recognizes all 3 TGF-[3 isoforms. ~2 The proapoptotic role of endogenous TGF-[32 and TGF-[33 was confirmed by showing the capacity of TGF-[3 to prevent cell death in the developing chick retina33 Neutralization of endogenous TGF-[3 during the early period of PCD in the chick retina resulted in a significant reduction of PCD revealed by a marked decrease of terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling (TUNEL)-positive cells. ~3 Analyzing a knockout mouse model, the work presented here addresses the question whether TGF-[32 and TGF-~3 might play an active role in intestinal epithelium regeneration by regulating PCD of enterocytes in vivo.

Materials and Methods Animals

Tgf[32 +/- and Tgf[:33 +/- heterozygous mice were offspring from breeding pairs kindly provided by T. Doetschman, University of Cincinnati (Cincinnati, OH). The generation of

1366 DONKERETAL.

these strains is described elsewhere. 2°,21 TGF-[3 double-heterozygous mice were generated by intercrossing the 2 heterozygous TGF-[3 strains resulting in TglCfl2 +/- Tg~fl3 +/- offspring according to the Mendelian ratio.

Tissue Preparation Three-week-old mice were killed by cervical dislocation. Defined pieces of small intestine (downwards from duodenal bulb) and colon (excluding cecum and rectum) were dissected, fixed in Bouin's fixative (picrid acid, formaldehyde, and glacial acetic acid) for several hours, dehydrated in a graded series of ethanol, and embedded in paraffin wax. Some paraffin cross sections (10-12 btm) of wild-type, T j f l 2 +/-, TgiCfl3 +/-, and TglCfl2 +/- TgiCfl3 +/- intestinal tubes were counterstained with H&E for morphologic studies. Digital images of stained sections were taken to calculate length of villi (base of villus to top of villus) and cell diameters of the small intestinal mucosa using the Axiovision morphometric software (Zeiss).

Irnmunohistochemistry Paraffin sections (10 txm) were deparaffinized and heated in citrate buffer in a microwave oven to improve antigen retrieval. Sections were preincubated with 10% normal goat serum in PBS containing 0.3% Triton-X 100 for 1 hour. Immunostaining was performed using isoform-specific anti-TGF-[3 and anti-TGF-[3 receptor (T[3R) antibodies = (TGF-[32, sc-90; TGF-[33, sc-82; T[3R-II, sc-400; Santa Cruz Biotechnology, Heidelberg, Germany) as well as a chromogranin A-specific antibody (DAKO, Hamburg, Germany) at a dilution of 1:200. The reaction was visualized with the following: (1) 3,3'-diaminobenzidine (DAB; Kem-En-Tec, Copenhagen, Denmark) using the rabbit Vectastain Elite ABC avidin-biotin-kit (PK-6101; Alexis, formerly Vector Laboratories, Griinberg, Germany) or using fluorescein isothiocyanate (green fluorescence; used for TGF-[32, -[33 labeling); or (2) Cy3 (red fluorescence; used for chromogranin A stain) goat antirabbit immunoglobulin (Ig) G-conjugated secondary antibodies (Alexa) at a dilution of 1:1000 in 10% normal goat serum/phosphate-buffered saline (PBS)/Triton-X 100. As controls, PBS was substituted for the primary antisera to test for nonspecific labeling. No specific cellular staining was observed when the primary antiserum was omitted. Stained sections were analyzed under light and fluorescence microscope, as well as by confocal imaging (EZ 2000; Nikon, Diisseldorf, Germany).

Periodic Acid-Schiff Reaction Stain Goblet cells of the small intestinal mucosa were stained specifically with a modified periodic acid-Schiff reaction (PBA). In brief, deparaffinized sections were treated with 1% periodic acid (Roth, Karlsruhe, Germany) for 5 minutes, rinsed in deionized H20, and left in 10% sodium disulfide (Merck, Darmstadt, Germany) for 18 hours. After 3 washes in H20 and ethanol, sections were stained with aldehydthionin

GASTROENTEROLOGYVo1.122, No. 5

(Merck) for 1 hour, subsequently rinsed in ethanol and H20 again, and counterstained with nuclear fast red.

TUNEL Staining Cell death was detected by TUNEL staining. Paraffin sections (10 ~m) were deparaffinized and stained with in situ cell-death detection kit (Roche Laboratories, Mannheim, Germany; cat. no. 1684795). In brief, sections were treated with the TUNEL reaction mixture (including fluorescein-deoxyuridine triphosphate [dUTP]) and incubated in the dark for 1.5 hours at 37°C, followed by washes in PBS. Incorporated fluorescein was detected by a sheep antifluorescein antibody conjugated with horseradish peroxidase. The substrate reaction was visualized with nickel-intensified DAB and the sections were counterstained with nuclear fast red. Apoptosis was quantified by counting TUNEL-positive cells in 50 intestinal villi per animal. Because the TUNEL technique is believed also to detect necrosis, only those cells exhibiting condensed chromatin in addition to being TUNEL-positive have been scored.

Quantification of Cell Death Cell death was quantified in defined intestinal pieces (small intestine, 2 cm from duodenal bulb onwards; colon, 2 cm excluding cecum and rectum) applying an enzyme-linked immunosorbent assay (ELISA) (Roche; cat. no. 1544675). This assay allows the quantification of soluble nucleosomes in cell lysates using a combination of antibodies recognizing histones and DNA. Scraped intestinal villi of wild-type, Tgffl2 +/-, and Tg/fl3 +l- heterozygous mice were homogenized i n 200 IxL PBS buffer and centrifuged at 13,000 rpm for 30 minutes at 4°C. A portion of the supernatant was used to quantify proteins by standard methods, and the rest was diluted in the buffer provided by the supplier and processed according to the manufacturer's manual. Absorbance values were normalized with respect to the values obtained with wild-type animals.

Proliferating Cell Nuclear Antigen Stain Paraffin sections (10 gtm) were deparaffinized and heated in citrate buffer in a microwave oven to improve antigen retrieval as described above. Sections were treated with 3% H202 in methanol to block endogenous peroxidase activity and preincubated with avidin-biotin blocking solution for 15 minutes, followed by an additional blocking step with mouse on mouse (M.O.M.) IgG blocking reagent (see below). Immunostaining was performed using a M.O.M. staining kit (Alexis, formerly Vector) following the manufacturer's instructions. In brief, the mouse-anti-PCNA antibody (Novocastra Laboratories Ltd., Dossenheim, Germany) was diluted 1:100 in M.O.M. diluent, incubated 1 hour at room temperature, and subsequently detected with Vectastain Elite ABC-reagent (Alexis, formerly Vector) included in the M.O.M. kit. As for the TUNEL staining, the reaction was visualized with DAB and the sections were counterstained with nuclear fast red. Cell proliferation was quantified counting all proliferating cell nuclear antigen (PCNA)-positively stained cells in 10 crypts per animal and calculating the ratio of proliferative cells per crypt.

May 2002

Western Blot Analysis As for the ELISA assays, the small intestinal mucosa (including the lamina propria) and parts of the submucosa were scraped from defined gut pieces opened longitudinally, homogenized, and protein levels quantified by standard methods. For Western blot analysis, proteins were transferred to a polyvinylidene difluoride membrane by tank blotting with 20 mmol/L Tris/HCl, pH 8.7, and 150 mmol/L glycine as a transfer buffer. Membranes were blocked in PBS with 0.1% Tween 20 and 5% dry milk for 1 hour at room temperature and incubated overnight at 4°C with primary antibodies (TGF-[32, sc-90; TGF-[33, sc-82; T[3R-I, sc-402; T[3R-II, sc-400; bax, sc-526; Bcl-xL, sc-8392; Bcl-2, sc-7382; Santa Cruz Biotechnology) diluted 1:100 in PBS-Tween 20 with 1% dry milk. After 3 washes with PBS-Tween 20, the membranes were incubated for 1 hour with the respective secondary antibody (peroxidase-conjugated goat anti-rabbit and goat anti-mouse IgG; Dianova, Hamburg, Germany) at a dilution of 1:10.000 in PBS-Tween 20 with 1% dry milk. Signals were developed by the Renaissanceenhanced luminol reagent (NEN Life Science Products, Packard BioScience,Dreieich, Germany) and visualized in a chemiluminescence imager (Raytest, SprockhiSvel, Germany). Densitometric measurements were performed by quantifying the density of the bands vs. the local background using AIDA-biopackage (Raytest).

Results Expression of Endogenous TGF-1~2, TGF-~3, and TI3R-II in the Murine Intestinal Mucosa In the present study, we localized TGF-[32 and TGF-[33 immunocytochemicallyin the intestinal mucosa of 3-week-old mice using isoform-specific anti-TGF-[3 antibodies. For both isoforms, prominent immunostaining was found in scattered ceils along the villus axis as well as in the small intestine crypts (Figure 1A and B). Cells staining positively for TGF-[32 and TGF-[33 were also found in the colon (Figure 1C and D). TGF-~3 immunoreactivity was even more abundant in the colon, where intense staining was localized in the crypts. Labeling for TI3R-II was distributed evenly over the entire intestinal mucosa, being somewhat more prominent at the villus tip (Figure 1E). Expression of TGF-[3 ligands and TGF-~ receptors was also analyzed by Western blot analysis using 100 b~g homogenized wild-type mouse gut extracts (Figure 1F and G). TGF-~ ligands (Figure 1F) and TGF-[3 receptor I and II (Figure 1F and G) could be detected in the small intestine and the colon of 3-week-old mice. The level of TGF-~2 expression seems to be higher in the small intestine (than in the colon), whereas TGF-[33 levels are significantly increased in the colon (than in the small intestine; Figure 1F).

TGF-13 MEDIATED APOPTOSIS IN THE MURINE GUT 1367

Isoform-Specific, Nonoverlapping Expression of Endogenous TGF-132 and TGF-1~3 in the Small Intestine In the small intestine, TGF-~2 and TGF-[33 exhibit an isoform-specific, nonoverlapping expression pattern: endocrine cells stain intensely positive for TGF-[32 (Figure 2A), whereas TGF-[33 immunoreactivity was found predominantly in goblet ceils (Figure 2B). Cellspecific fluorescein isothiocyanate labeling of the 2 antibodies was confirmed by the fact that goblet cells, intensely labeled with TGF-~3-specific antibody, remained unstained for TGF-[32 immunoreactivity (Figure 2A). The cell type identity and specificity of labeling pattern was analyzed in more detail by the following: (1) confocal microscopy of a TGF-~2 positive cell, (2) confocal microscopy of a chromogranin A staining, and (3) PBA stain, specifically labeling mucus-secreting goblet ceils. Confocal imaging visualized basal concentrations of secretory granules in cells staining positively with the TGF-[32-specific antibody (Figure 2C; green fluorescein isothiocyanate fluorescence). Besides, the enteroendocrine nature of TGF-~2-positive cells was confirmed by immunohistochemical labeling for chromogranin A, an acidic glycoprotein specifically expressed in secretory granules of endocrine ceils. Confocal imaging of chromogranin A labeling (Figure 2C; red Cy3 fluorescence) likewise visualized basal concentration of hormone-accumulating secretory granules typical for enteroendocrine ceils. PBA-labeled goblet cells strongly resembled TGF[33-positive cells and vice versa (Figure 2E).

Reduction of PCD in the Intestinal Mucosa of Tgf[32 +/- and Tgf[33 +/- Heterozygous Mice The preceding immunohistochemical and Western blot analyses revealed the presence and localization of TGF-[32 and TGF-[33 isoforms in the intestinal tract. To address the question whether TGF-[32 and TGF-~3 play a role in intestinal epithelium regeneration by regulating PCD of enterocytes, we analyzed the guts of 2 knockout mouse strains, deficient for either TGF432 or TGF-]33. Because homozygous Tgf[32 - / - and TgoC133- / - knockout mice die already during or shortly after birth when the intestinal tract is not yet fully differentiated, we analyzed TglC[32+/- and TglC[33+/- heterozygous as well as Tglc[32+/- TgtC~3 +/- double heterozygous mice. Analyses were performed at an age of 3 weeks, the earliest stage at which morphology of crypts and villi resemble the adult stage. Western blot analyses were performed, confirming that half gene dosages of the 2 Tg/Cj8genes in Tglc~2 +/-

@

:~g: --

am~ :1

i

~ :~

~F~2.

-,;

:

~__

D

TGF-B2

m

TGF-&3

4r

TGF-B2

@ chromogranine A

'qvrp,p,

Figure 2. Immunofluorescence showing the differential localization of (A) TGF-~2 and (/3) TGF-133in small intestinal epithelial cells. TGF-132and TGF-~3 exhibit an isoform-specific, nonoverlapping expression pattern: TGF-132 was mainly localized in endocrine cells (bold arrows in A), whereas TGF-~3 immunoreactivity was predominantly confined to goblet cells (arrowheads in B). Arrowheads in A demarcate goblet cells, which remained unstained for TGF-~2 immunoreactivity. Cell type identity and specificity of the labeling patterning was confirmed by (C) confocal imaging of a TGF-!82-immunoreactive cell and (D) confocal imaging of chromogranin A labeling, both visualizing basal concentration of hormone-accumulating secretory granules typical for enteroendocrine cells (small arrows in C and D). Enterocytes labeled by PBA, specifically staining mucus-secreting goblet cells (arrowheads in E), morphologically strongly resemble (B) TGF-133-positive cells. Scale ban A, B, and D, 201~m.

T6RII

4 F

TGF-62 and TGF-B3 14 kDa

TGF-82

TGF-B3

G 97.4 66.2

wmmqlmm

45

gel

31

~k,

21 14.4 ,~,

""

Figure I . (A-D) Expression of TGF62 and TGF-133 in the (A and B) wild-type murine small intestine and (Cand D) colon. TGF-132(arrows in A) and TGF!83 (arrowheads in B) immunoreactive cells were found along the small intestine crypt-to-villus axis. Cells staining positively for TGF-#2 were also found in the colon (arrows in C). TGF~3 immunoreactivity was even more abundant in the colon, where intense staining was localized in the crypts. Scale bar:A and B, 20Fm; Cand D, 100 t~m. (E) Localization of the TI~R-II in the small intestine. T6R-II is expressed throughout the intestinal mucosa, but expression seems to be more concentrated at the villus tip (white arrowheads). Scale bar: E, 25 i~m. (Fand G) Western blot analysis of (F) TGF-~ ligand and (G) TI3R expression. Homogenized wildtype mouse gut extracts (100 ~g) were separated on a 12.5% sodium dodecyl sulfate-polyacrylamide gel and blotted onto a polyvinylidene difluoride membrane. The membrane was incubated with specific antibodies for TGF#2, TGF~3, T~R-I, and TI3R-II. Both TGF-13 ligands and receptors could be detected in the gut. The level of TGF-132 expression seems to be higher in the small intestine, whereas TGF-#3 levels are significantly higher in the colon. Expression of T6R-I and T#R-II was confirmed for the small intestine (G) and colon (data not shown).

May 2002

TGF-~ MEDIATED APOPTOSIS IN THE MURINE GUT 1369

and Tg[~3 +/- heterozygous mice do in fact lead to a significant reduction of 50% in T G F - ~ ligand levels in the small intestine (Figure 3). To reveal apoptotic enterocytes, cross section of intestinal tubes of Tgf~2 +/- and TgOc~3+/- heterozygous mice were labeled by T U N E L staining, partly converted with peroxidase. Apoptosis was mainly confined to the top of intestinal villi (Figure 4). Detection of D N A fragmentation by fluorescein-dUTP incorporation clearly confirmed the apoptotic nature of enterocyte cell death at the villus tip (Figure 4B). Many more apoptotic nuclei were labeled in wild-type mice compared with Tg/~2 +/- and TgoC~3+/- heterozygous littermates (Figure 4C and D). Apoptotic levels were quantified in homogenates of small intestine and colon using the nucleosome assay (ELISA) and in intestinal tube sections using the T U N E L assay. The ELISA assays showed that compared with wild-type littermates, cell death was significantly reduced in the small intestine of Tglr~32+/- and TgoC~3+/heterozygous as well as in Tgf~2 +/- Tg,#~3 +/- double heterozygous mice (Figure 5A). In the colon, however, apoptotic levels seem to be unaffected in Tg/~2 +/-, TgoC~3+/-, and Tg/'j~2+/- TgiC~3+/- mice. Cell counts of TUNEL-positive enterocytes confirmed these results and revealed that there was a reduction of about 50% in the number of apoptotic nuclei in Tg/~2 +/- and TglC~3+/- heterozygous as well as in

TGF-g2

TGF-g3

100

o~ 80 c~ ¢--

T

60

intestinal lumen

+1-

~

~

• 40 Q.

20 n=3 I

11=3 n:

Figure 3. Western blot analysis of TGF-~2 and TGF-~3 expression in the small intestine of Tgf[~2+/- and Tgf~3+/- heterozygous mice. Homogenized wild-type mouse gut extracts (200 i~g) were separated on a 12.5% sodium dodecyl sulfate-polyacrylamide gel and blotted onto a polyvinylidene difluoride membrane. The membrane was incubated with specific antibodies for TGF-~2 and TGF-[33. The level of TGF-~2 expression was found to be lower in the small intestine of Tgf~2+/- heterozygous mice, whereas TGF-~3 levels are significantly lower in the gut of Tgf~3÷/- heterozygous mice. Densitometric values were normalized with respect to values obtained with wild-type extracts.

Figure 4. TUNE[- labeling of apoptotic cells in the small intestine. Apoptetic enterecytes, labeled by peroxidase-converted TUNE[- staining, were mainly confined to the top of intestinal villi (arrowheads in A and arrows in C and D). Detection of DNA fragmentation (arrowheads in N by fluorescein-dUTP incorporation (see Materials and Methods) clearly confirmed the apoptotic nature of entemcyte cell death. Many mere apoptotic nuclei (arrows in C and D) were labeled in (C) wild-type

mice compared with (D) rgf~3+/ heterozygous littermates. Scale bar'. A and B, 20 ixm; C and D, 100 ~m.

1370

DONKER ET AL.

GASTROENTEROLOGY Vol. 122, No. 5

Tg[~2 +/- TgiC~3+/- double heterozygous mice compared with wild-type littermates (Figure 5B).

A 1,2

Unaffected Proliferation Rate but Increase in Villus Length in Tgf[32 +1- and Tgf~3 +/Heterozygous Mice

ELISA J ~

small intestine

•"~ 1,0 >,.~, O 0,8

On g =o,, f" 8

0,4

v

0.2 •

r,i,.

0,0

B 120

TUNEL

.~_ lO0 U) 0 ,t.*

0Q. 80 60

O~ e~

40

o

D.

2O

Figure 5. Quantification of cell death. (A) Apoptotic levels were quantified in homogenates of small intestine (11) and colon ([3) using the nucleosome assay (ELiSA; see Materials and Methods). Cell death is significantly reduced in the small intestine of Tgfi82+/ and Tgf#3+/heterozygous as well as in Tgf182+/- Tgf#3+/- double heterozygous mice compared with wild-type littermates. In the colon, however, apoptotic levels seem to be unaffected in Tgf#2+/-, Tgf#3+/-, and Tgf#2+/- Tgf#3 +/- mice. (B) Apoptotic levels were quantified in small intestinal tube sections using the TUNEL assay. The number of apoptotic cells is clearly reduced in Tgf#2÷/ and Tgf~83+/- heterozygous as well as in Tgf#2+/- Tgfj83+/- double heterozygous mice compared with wild-type littermates. Values are means _+ SEM. *P < 0.01; * * P < 0.005; unpaired Student t test.

To investigate whether the reduction in apoptosis detected in the intestinal tract of TgiCj82+/-, Tg/~3 +land TglC[S2+/- Tg/cj83+/- mice might be counterbalanced by a reduction in cell proliferation or accompanied by either changes in cell diameter or villus length, all 3 parameters were subjected to quantitative measurements. Neither cell proliferation rate detected by PCNA stain (Figure 6A) nor endothelial cell diameter (Figure 6B) seemed to be altered in the small intestine of TglC~2+land TgoC~3+/- heterozygous as well as Tg/'~2 +/TglC~3 +/- double heterozygous mice compared with wild-type littermates. However, small intestinal villi of Tg{~2 +/- and Tg,#~3 +/- heterozygous mice exhibit a significant increase in length compared with wild-type villi (Figure 6C). Up-regulation of Bcl-2 and Bcl-xL Expression in the Intestine of Tgf[32 +1- and Tgf[33 +/Heterozygous Mice To address the question of whether TGF-[32 and TGF-~3 also influence the levels of pro- and antiapoptotic proteins in vivo, the expression of Bax, Bcl-2, and Bcl-xL in the murine intestine was assayed by Western blotting. Bax, Bcl-2, and Bcl-xL were expressed in the murine small intestine at the age of 3 weeks, Bax and Bcl-xL at moderate, and Bcl-2 at somewhat higher levels (data not shown). Because TGF-[32 and TGF-[33 seemed to have no effect on the expression level of the proapoptotic protein Bax (data not shown) in vitro, we focused our in vivo analyses on the TGF-[3-regulated expression of the antiapoptotic proteins Bcl-xL and Bcl-2. Compared with wild-type littermates, protein levels of both Bcl-xL and Bcl-2 are significantly increased in Tg/~2 +/- and TgoC~3+/- heterozygous mice(Figure 7). These results were reproduced for wild-type vs. heterozygous littermates in 3 independent experiments (Figure 7A and B) and confirmed by densitometric measurements quantifying the density of the bands (Figure 7C and D).

Conclusions In the present study we show that PCD is significantly reduced in the intestinal mucosa of TglC~2+/- and Tgf~3 +/- mice as confirmed by ELISA and TUNEL labeling. Our data show for the first time that TGF-[32

May 2 0 0 2

TGF-13 MEDIATED APOPTOSIS IN THE MURINE GUT 1371

and TGF-[33 play an important role in mediating apoptosis in the murine intestinal mucosa in vivo.

A ~,

16

Discussion

14

Differential Endogenous Expression of TGFJ32 and TGF-I~3 in the Intestinal Mucosa

O

""

12

a) O O

10

._>

8

O

6

<~ Z O n

4

B

2

6 S

E =1.

'--' Q)

E "O

--

Q O

4

3 2 1

C

700 6OO

..j E

soo

•~ 400

5 300 ol 200

TGF-[3s signaling via specific TGF-~ receptors belong to a family of polypeptides with diverse biological functions that include the stimulation of inhibition of cell proliferation, enhancement of cell differentiation, and stimulation of matrix formation.23 TGF-~s thus play a central role in the control of cell growth in adult tissue, in embryonic development, and in carcinogenesisY In the present study we showed that the localization of TGF-~2 is confined to enteroendocrine cells, whereas TGF-[33 is mostly localized in goblet cells. Barnard et al. z5 found TGF-~31, TGF-~2, and TGF-[33 mRNA in homogenates from the intact mouse jejunum and colon. Unlike the present study, in which we clearly demonstrated an isoform-specific localization of TGF-~, the authors state that the 3 isoforms colocalize in these tissues and found that the expression is most prominent on the villus tip. 15 In contrast to our results, they detected gradually decreasing staining along the crypt-tovillus axis and found that staining in the crypt itself was negligible. Besides, no labeling was detected in goblet cell cytoplasm, which was intensely stained for TGF-~3 immunoreactivity in our investigations. These differences in expression patterns may either result from the fact that they used adult mice or that they used a different set of antibodies and the respective antibodies might cross-react. In the study presented here, Western blot analyses revealed that the level of TGF-~2 expression seems to be higher in the small intestine, whereas TGF-~3 levels are significantly increased in the colon. These findings are in good accordance with the cell-specific localization of the TGF-~ isoforms because enteroendocrine cells exhibiting TGF-[32 immunoreactivity are more abundant in the small intestine, whereas goblet cells staining positively for TGF-[33 are the predominant cell type of the colon.

100

Figure 6. Quantification of (A) cell proliferation, (B) cell diameter, and (C) villus length. Neither cell proliferation rate as detected by (A) PCNA stain (see Materials and Methods) nor (B) endothelial cell diameter is altered in the small intestine of Tgf~82+/- and Tgf#3 +/heterozygous as well as Tgfi82+/- Tgf~3 +/- double heterozygous mice compared with wild-type littermates. (C) Small intestinal villi of Tgf#2 +/- and Tgf#3 +/- heterozygous mice exhibit a significant increase in length compared with wild-type villi. Values are means + SEM. * * P < 0.005; unpaired Student t test.

1372

DONKER El" AL.

A

~

B

GASTROENTEROLOGY Vol. 122, No. 5

~

~

~

lib

am Ill

bcl-2

bcl-xl ¢:

e:3

.o

o

2000

o~ 4O00

J¢

m

bcl-2 28 kDa

D

-o5OOO

_

bcl-xl 29 kDa

~

-~

C

,,Q

~

(.l

1500 ¸

J~

3000

g ill o

o

2000

7. (A and B) Analysis and (C and D) densitometric quantification of Bcl-xL and Bcl-2 expression levels in the small intestine of wild-type and Tgf#2 +/- and Tgf,83+/- heterozygous mice by Western blotting analysis using (A) BclxL- and (B) Bcl-2-specific antibodies. Protein levels of both Bcl-xL and Bcl-2 are significantly increased in Tgf#2 +~and Tgf#3 +/ heterozygous compared with wild-type mice. * P < 0.001; * * P < 0.005; unpaired Student t test. Figure

1000

500 ¸ O1

.=_ xV

Avery et al. iv detected all 3 TGF-13 isoforms in normal colon mucosa within epithelial cells, whereby those in the upper parts of the crypts show enhanced immunoreactivity compared with cells in the proliferative compartment. In the rat jejunal epithelium, TGF-13s are likewise expressed predominantly in the villus tip, the region of the crypt villus unit that is characterized by the terminally differentiated phenotype. 25 This pattern of differential staining suggests that TGF-13s have an important role in the control of growth and differentiation in the intestinal mucosa and may function in coordination of the rapid cell turnover typical for the intestinal epithelium)V, 25 TGF-13s signal via ligand binding TGF-13 receptors, which are not necessarily located on the surface of those cells expressing TGF-132 or TGF-133. Thus, it is not TGF-~3s expressing ceils, which undergo apoptosis, but ceils that (besides other cues) express the T13R-II. A paracrine mode of action might be postulated for the induction of apoptosis at the intestinal villus tips. This hypothesis is strengthened by the finding that T13R-II, although evenly distributed throughout the intestinal

mucosa, is expressed to a higher degree at the villus tip, the region where apoptotic nuclei were found. Secreted TGF-132, produced by enteroendocrine cells, and TGF133, expressed by goblet cells, might influence apoptosis of enterocytic cells at the villus tip, expressing the T13RII, thereby inducing apoptosis of these cells in a paracrine fashion.

Reduced Apoptosis in the Small Intestine of Tgf[32 +/- and Tgf[33 +/- Heterozygous Mice We demonstrated an isoform-specific localization of TGF-132 and TGF-133 in the small intestine. One might speculate that the different TGF-13 isoforms have differing and potentially opposing effects with regard to cell differentiation, proliferation, or epithelial regeneration. As far as apoptosis is concerned, however, TGF-132 and TGF-133 seem to have the same proapoptotic effect on enterocytes because PCD was reduced to the same extent in both TgoC~32+/- and Tgf~33 +/- heterozygous mice.

May 2002

Epithelial cells in the gastrointestinal tract have a high turnover, and because it is essential to maintain a normal cell balance, cell death, especially apoptosis, is crucial for maintenance of normal morphology and function, m The mechanism of cell loss was widely held to be some form of exfoliation from the superficial portion of the small intestine villus and the top of the colonic crypt. Cell loss by exfoliation is, however, not a compelling explanation for the massive physiological loss of epithelial cells from the superficial aspect of the gastrointestinal tract. Although studies have shown apoptosis in crypt stem cells,4 work from Ramachandran's laboratory and others 1,6,7 indicates that apoptosis occurs predominantly in the villus tip cells. Similar to the results of the present study, Hall et al. v identified apoptotic bodies within the epithelium of the majority of small intestinal villi and stated that they are more frequent towards the top of the villus. Apoptotic cells at the villus tip form a prominent "apoptotic cuff ''26 also seen in our TUNEL stains (Figure 4). That some loss of cells or fragments thereof occur by shedding in vivo can, however, not be denied) Detailed studies by Potten 2v of epithelial cell kinetics in the mouse gastrointestinal tract have shown that apoptosis is rare in normal epithelium, with one apoptotic body being seen every fifth crypt. The paucity of apoptotic cells on histologic sections reflect the rapid kinetics of apoptosis and the removal of apoptotic cells by phagocytosis. 28 Potten et al) term cell loss from the villus anoikis ("homelessness") as they hypothesize that this cell loss occurs via lost adhesion, which, according to the authors, may in turn induce apoptosis) Altered adhesion signals have been shown to initiate apoptosis, e.g., by altered extracellular matrix composition. These data additionally argue for a role of TGF-13 in intestinal apoptosis because TGF-[3 is known to mediate synthesis of extracellular matrix components. 29 The main importance of apoptosis in intestinal disease lies in its role in carcinogenesis and hence the promise of treatment for cancer. The TGF-[3 pathway is known to play an important role in both human and murine colon cancer. 24,3° Mutations in the human T[3R-II have been found in both sporadic and inherited colon cancers) 1 Like smad2, smad4 has been strongly implicated in human cancer as a potent tumor suppressor) 2 Mutations in the smad4 gene are involved in 50% of pancreatic cancers. However, smad4 heterozygous mice display no increase in tumorigenicity for a period of up to 11 months. 33 Deficiency of smad4 in the Apc mouse model of small intestinal cancer has been shown to increase adenoma size and promote progression to invasive carci-

TGF-I~ MEDIATED APOPTOSIS IN THE MURINE GUT 1373

noma. 34 smad3 ( - / - ) mice are viable and fertile; however, at the age of 4 - 6 months, smad3 mutant mice bred onto a 129 background develop metastatic colorectal cancer. 35 TGF-~I seems to suppress early events in colon cancer development, whereas TGF-~2 and TGF-~3 are obviously not involved in early events because TgiC~3I-~knockout mice develop only nonmetastatic colon cancer, 36 whereas smad3-deficient mice exhibit metastatic colon cancer) 5 One might hypothesize that TGF-~2 and/or TGF-[33 inhibits metastasis at late stages of tumorigenesis) 6 The fact that Tgf[32 - / - and Tg/:[33 - l knockout mice do not develop tumors may be explained by the fact that they die in utero or shortly after birth. 2°,21,3v,38 Thus, early lethality is likely to preclude tumor formation.

Up-regulation of Bcl-2 and Bcl-xL Levels in Tgf[32 +/- and Tgf[33 +/- Heterozygous Mice In the present study, we demonstrated the regulative influence of TGF-IB2 and TGF-133 on Bcl-2 and Bcl-xL expression levels in vivo demonstrating the upregulation of Bcl-2 and Bcl-xL levels in the small intestine of TgoC~32+/- and TgiC~33+/- heterozygous mice. Morphologic changes associated with apoptosis are closely correlated with the expression of specific proteins. In vitro data from cell culture experiments suggest that TGF-{3 regulates the expression of the Bcl-2 family members. 39 A study by Hague et al. 19 demonstrated that TGF-J31-induced apoptosis is preceded by a reduction in p26-Bcl-2 protein levels in colorectal adenoma cells. Our data confirmed the results of a study by Salzman et al) 9 showing a decrease in Bcl-xL protein level 24 hours after TGF-13 treatment. Pro- and antiapoptotic proteins Bax, Bcl-2, and Bcl-xL were found in the small intestine and colon; the localization of these proteins is, however, controversial. 8,4°,4~ High levels of proapoptotic protein Bax have been detected in terminally differentiated cells. Although there is minimal expression of antiapoptotic protein Bd-2 in the small intestine of the mouse, it has been detected at the base of colonic crypts, 4z where Bcl-2 is commonly overexpressed in colonic adenocarcinomas. The causeeffect relationships between the expression of these proteins and DNA degradation are barely known. For studying expression of apoptosis-related proteins in relation to apoptosis, the small intestine with its spatially organized continuum of proliferation, differentiation, and death is a very useful preparation. Enterocytes towards the apex of the villi become increasingly susceptible to apoptosis. A study by Aschoff et al. 8 showed that the gradient of Bcl-2 staining is similar to the gradient of DNA fragmentation. Immunoreactivity for Bcl-2 is most intense in ceils

1374

DONKER ET AL.

that are prone to become apoptotic next, not in cells in an advanced apoptotic state. 8 These findings suggest that sublethally damaged cells may up-regulate Bcl-2 as part of a protective mechanism. 8 In the present study, we confirmed TGF-~'s wellestablished role in regulating gene transcription of proteins associated with apoptosis showing the down-regulation of Bcl-xL and Bcl-2 levels upon TGF-~ treatment in vitro. The in vivo data are even more interesting because the significant reduction in enterocyte apoptosis found in Tgf[32 +/- and Tgf[33 +/- heterozygous mice is accompanied by a respective up-regulation of Bcl-2 and Bcl-xL levels in the small intestine of knockout mice deficient for TGF-[3.

Proliferation TGF-[3s have been recognized to alter proliferation in nearly all cell types. TGF-[3s stimulate proliferation in a number of mesenchymal cell populations, including osteoblasts and Schwann cells. 1° In contrast, TGF-13s inhibit proliferation of all epithelial cell populations, e.g., embryonic fibroblasts, apparently through prolongation of the G1 phase. 42 In the present study we could not detect a balancing decrease in proliferation rate in the intestines of Tgf~2 +/- and TgiC~3 +/- heterozygous mice, although a significant reduction in apoptotic nuclei at the villus tips was demonstrated. One might speculate that the PCNA labeling used to stain for proliferative cells is not sensitive enough (e.g., compared with BrdU incorporation) or that increased apoptosis is not counterbalanced by decreased proliferation, but cell cycle arrest plays a critical role. Besides, the cumulative effect of decreased apoptotic levels on tissue architecture is straightforward as far as a detectable increase in villus length is concerned.

Synopsis In the present study we have established the role of TGF-[32 and TGF-133 in mediating apoptosis in the intestinal mucosa, a tissue in which apoptosis is especially relevant as this epithelium undergoes continual regeneration. We have found that programmed cell death was significantly reduced in the intestinal mucosa of Tg/C~2 +/- and TgoC133+/- heterozygous mice. This decrease in apoptosis was accompanied by a significant up-regulation of levels of the antiapoptotic proteins Bcl-xL and Bcl-2. Although apoptosis is essential for the maintenance of normal gut epithelial function, dysregulated apoptosis is seen in a number of pathological conditions in the gastrointestinal tract. 1 A shift in proliferation/cell death balance might be a predisposition for tumorigenesis. The TGF-[3 pathway is known to play an

GASTROENTEROLOGY Vol. 122, No. 5

important role in colorectal carcinogenesis. 2<3° Mutations in the human TI3R-II have been found in both sporadic and inherited cancers. 31 Besides, many colorectal tumors have elevated levels of Bcl-2, suggesting that deregulated Bcl-2 expression may be important in tumor development. 1 Thus, at present, the significance of apoptosis for intestinal disease seems to be primarily linked to carcinogenesis. In consequence, therapeutic strategies for treatment of intestinal cancer may be developed based on drugs interfering with epithelial cell proliferation and apoptotic cell death. On the other hand, it is conceivable that application of neutralizing antibodies to TGF-[3s may be effective in diseases that therapeutically require stimulation of epithelial regeneration and suppression of apoptosis.

References 1. Ramachandran A, Madesh M, Balasubramanian KA. Apoptosis in the intestinal epithelium: its relevance in normal and pathophysiological conditions. J Gastroenterol Hepatol 2 0 0 0 ; 1 5 : 1 0 9 - 1 2 0 . 2. Hermiston ML, Simon TC, Crossman MW, Gordon Jl. Model systems for studying cell fate specification and differentiation in the gut epithelium. In: Johnson LR, ed, Physiology of the gastrointestinal tract. 3rd ed. New York: Raven, 1994:521-569. 3. PoRen CS, Wilson JW, Booth C. Regulation and significance of apoptosis in the stem cells of gastrointestinal epithelium. Stem Cells 1997;15:82-93. 4. Potten CS, Owen G, Roberts SA. The temporal and spatial changes in cell proliferation within the irradiated crypts of the murine small intestine. Int J Radiat Biol 1990;57:185-199. 5. Creamer B, Shorter RG, Bamforss J. The turnover and shedding of epithelial cells. Part I1. The shedding in the small intestine. Gut 1961:117-118. 6. BenedeRi A, Mancini R, Marucci L, Paolucci F, Jezquel AM, Orlandi F. Quantitative study of apoptosis in normal rat gastroduodenal mucosa. J Gastroenterol Hepatol 1 9 9 0 ; 5 : 3 6 9 - 3 7 4 . 7. Hall PA, Coates PJ, Ansari B, Hopwood D. Regulation of cell number in the mammalian gastrointestinal tract: the importance of apoptosis. J Cell Sci 1 9 9 4 ; 1 0 7 : 3 5 6 9 - 3 5 7 7 . 8. Aschoff AP, Ott U, FiJnfstOck R, Stein G, Jirikowski GF. Colocalization of Bax and BCL-2 in small intestine and kidney biopsies with different degrees of DNA fragmentation. Cell Tissue Res 1999;296:351-357. 9. Iwanaga T. The involvement of macrophages and lymphocytes in the apoptosis of enterocytes. Arch Histol Cytol 1 9 9 5 ; 5 8 : 1 5 1 159. 10. Podolsky DK. Peptide growth factors in the gastrointestinal tract. In: Johnson LR, ed. Physiology of the gastrointestinal tract. 3rd ed. New York: Raven, 1 9 9 4 : 1 2 9 - 1 6 7 . 11. Sporn MB, Roberts AB. Peptide growth factors are multifunctional. Nature 1988;322:217-219. 12. Krieglstein K, Richter S, Farkas L, Schuster N, DOnker N, Oppenhelm RW, Unsicker K. Reduction of endogenous TGF-13 prevents ontogenetic neuron death. Nature Neurosci 2 0 0 0 ; 3 : 1 0 8 5 - 1 0 9 0 . 13. DOnker N, Schuster N, Krieglstein K. TGF-13 modulates programmed cell death in the retina of the developing chick embryo. Development 2 0 0 1 ; 1 2 8 : 1 9 3 3 - 1 9 4 2 . 14. Thompson NL, Flanders KC, Smith JM, Ellingsworth LR, Roberts AB, Sporn MB. Expression of transforming growth factor-131 in specific cells and tissues of adult and neonatal mice. J Cell Biol 1989;108:661-669.

May 2002

15. Barnard JA, Ginger JW, Gold LI. Localization of transforming growth factor ~ isoforms in the normal murine small intestine and colon. Gastroenterology 1993;105:67-73. 16. Pelton RW, Saxena B, Jones M, Moses HL, Gold LI. Immunohistochemical localization of TGF~I, TGF#2 and TGF~3 in the mouse embryo: expression patterns suggest multiple roles during embryonic development. J Cell Biol 1991;115:1091-1105. 17. AveryA, Paraskeva C, Hall P, Flanders KC, Sporn M, Moorghen M. TGF-~ expression in the human colon: differential immunostaining along crypt epithelium. Br J Cancer 1993;63:137-139. 18. Koyama S, Podolsky DK. Differential expression of transforming growth factors alpha and beta in rat intestinal epithelial cells: mirror-image gradients from crypt to villus. J Clin Invest 1989; 83:1768-1773. 19. Hague A, Bracey TS, Hicks DJ, Reed JC, Paraskeva C. Decreased levels of p26-Bcl-2, but not p30 phosphoryiated Bcl-2, precede TGF-#l-induced apoptosis in colorectal adenoma cells. Carcinogen 1998;19:1691-1695. 20. Proetzel G, Pawlowski SA, Wiles WV, Yin M, Boivin GP, Howles PN, Ding J, Ferguson MWJ, Doetschman T. Transforming growth factor-1~3 is required for secondary palate fusion. Nat Gen 1995; 11:409-414. 21. Sanford LP, Ormsby I, Gittenberger-de Groot A, Sariola H, Friedman R, Boivin GP, Cardell EL, Doetschman T. TGF~2 knockout mice have multiple developmental defects that are non-overlapping with other TGF~8knockout phenotypes. Development 1997; 127:2659-2670. 22. Flanders KC, Thompson, NL, Cissl DS, Van Obberghen-Schilling E, Baker CC, Kass ME, Ellingsworth LR, Roberts AB, Sporn MB. Transforming growth factor-~l: histochemical localization with antibodies to different epitopes. J Cell Biol 1989;108:653-660. 23. Roberts AB, Sporn MB. The transforming growth factor-~s. In: Sporn MB, Roberts AB, eds. Handbook of experimental pharmacology. Volume 95. Heidelberg: Springer, 1990:419-472. 24. Roman C, Saha D, Beauchamp RD. TGF-# and colorectal carcinogenesis. Microsc Res Tech 2001;52:450-457. 25. Barnard JA, Beauchamp RD, Coftey RJ, Moses HL. Regulation of intestinal epithelial cell growth by transforming growth factor type I~. Proc Natl Acad Sci U S A 1989;86:1578-1582. 26. Westcarr S, Farshori P, Anderson WA. Apoptosis and differentiation in the crypt-villus unit of the rat small intestine. J Submicrosc Cytol Pathol 1999;31:15-30. 27. Potten CS. The significance of spontaneous and induced apoptosis in the gastrointestinal tract of mice. Cancer Metastasis Rev 1992;11:179-195. 28. Watson AIM. Necrosis and apoptosis in the gastrointestinal tract. Gut 1995;37:165-167. 29. Hocevar BA, Howe PH. Analysis of TGF-18-mediatedsynthesis of extracellular matrix components. Methods Mol Biol 2000;142: 55-65. 30. Markowitz S. TGF-~ receptor and DNA repair genes, coupled targets in a pathway of human colon carcinogenesis. Biochim Biophys Acta 2000;1470:M13-M20. 31. Markowitz S, Wang J, Myeroft L, Parsons R, Sun L, Lutterbaugh J, Fan R, Zborowska E, Kinzler K, Vogelstein B, Brattain M, Willson

TGF-I3 MEDIATED APOPTOSIS IN THE MURINE GUT 1375

32. 33.

34.

35. 36.

37.

38.

39.

40.

41.

42.

J. Inactivation of the type II TGF-~ receptor in colon cancer cells with microsatellite instability. Science 1995;268:1336-1338. Duff EK, Clarke AR. Smad4 (DPC4) - a potent tumour suppressor? Br J Cancer 1998;78:1615-1619. Sirard C, de la Pompa JL, Elia A, Itie A, Mirtsos C, Cheung A, Hahn S, Wakeham A, Schwartz L, Kern SE, Rossant J, Mak TW. The tumor suppressor gene smad4/Dpc4 is required for gastrulation and later for anterior development of the mouse embryo. Gen Dev 1998;12:107-119. Takaku K, Oshima M, Miyoshi H, Matsui M, Seldin MF, Taketo MM. Intestinal tumorigenesis in compound mutant mice Of both Dpc4(Smad4) and Apc genes. Cell 1998;92:645-656. Zhu Y, Richardson JA, Parada LF, Graft JM. Smad3 mutantmice develop metastatic colorectal cancer. Cell 1998;94:703-714. Engle SJ, Hoying JB, Boivin GP, Ormsby I, Gartside PS, Doetschman T. Transforming growth factor 1~1 suppresses nonmetastatic colon cancer at an early stage of tumorigenesis. Cancer Res 1999;59:3379-3386. Kaartinen V, Voncken JW, Shuler C, Warburton D, Bu D, Heisterkamp N, Groffen J. Abnormal lung development and cleft palate in mice lacking TGF-beta 3 indicates defects of epithelialmesenchymal interaction. Nat Genet 1995;11:415-421. DUnker N, Krieglstein K. Targeted mutations of transforming growth factor-beta genes reveal important roles in mouse development and adult homeostasis. Eur J Biochem 2000;267:69826988. Salzman A, Munro R, Searfoss G, Franks C, Jaye M, Ivaschenko Y. Transforming growth factor-beta-mediated apoptosis in the Ramos B-lymphoma cell line is accompanied by caspase activation and Bcl-xl downregulation. Exp Cell Res 1998;242:244254. Krajewski S, Krajewska M, Shabaik A, Miyashita T, Wang HG, Reed JC. Immunohistochemical determination of in vivo distribution of BAX, a dominant inhibitor of Bcl-2. Am J Pathol 1994;145: 1323-1336. Hockenbery DM, Zutter M, Hickey W, Nahm M, Krosmeyer SJ. BCL2 protein is topographically restricted in tissues characterized by apoptotic cell death. Proc Natl Acad Sci U S A 1991;88: 6961-6965. Merritt AI, Potten CS, Watson AJ, Loh DY, Nakayama K, Hickman JA. Differential expression of BCL-2 in intestinal epithelia: correlation with attenuation of apoptosis in colonic crypts and the incidence of colonic neoplasia. J Cell Sci 1995;108:2261-2271.

Received July 7, 2001. Accepted January 10, 2002. Address requests for reprints to: Nicole DLinker, Ph.D., Center of Anatomy, Georg-August-University G6ttingen, Kreuzbergring 36, 37075 G6ttingen, Germany. e-mail: nduenke@ gwdg.de; fax: (49) 55139-14016. The authors' current affiliation is Center of Anatomy, Neuroanatomy, University of G6ttingen, G6ttingen, Germany. Supported by grants from the Deutsche Forschungsgemeinschaft. The authors thank T. Doetschman for generously providing Tgf~2 +land Tgf133+/- breeding pairs, and S. Brundaler and G. KLihnreich for excellent technical assistance.