Functions of the Growth Arrest Specific 1 Gene in the ... - ScienceDirect

Developmental Biology 234, 188 –203 (2001) doi:10.1006/dbio.2001.0249, available online at http://www.idealibrary.com on

Functions of the Growth Arrest Specific 1 Gene in the Development of the Mouse Embryo K. K. H. Lee,* ,1 A. K. C. Leung,* M. K. Tang,* D. Q. Cai,† C. Schneider,‡ C. Brancolini,‡ and P. H. Chow* *Department of Anatomy, Faculty of Medicine, The Chinese University of Hong Kong, Shatin, Hong Kong, People’s Republic of China; †Department of Orthopedics and Traumatology, The Chinese University of Hong Kong, Shatin, Hong Kong, People’s Republic of China; ‡Laboratorio Nazionale Consorzio Interuniversitario Biotecnologie, Area Science Park Padriciano 99, Trieste, Italy and Dipartimento Scienze e Tecnologie Biomediche, Facolta’ di Medicina, University of Udine, p. Kolbe 4, Udine, Italy

The growth arrest specific 1 (gas1) gene is highly expressed in quiescent mammalian cells (Schneider et al., 1988, Cell 54, 787–793). Overexpression of gas1 in normal and some cancer cell lines could inhibit G 0/G 1 transition. Presently, we have examined the functions of this gene in the developing mouse embryo. The spatial–temporal expression patterns for gas1 were established in 8.5- to 14.5-day-old embryos by immunohistochemical staining and in situ hybridization. Gas1 was found heterogeneously expressed in most organ systems including the brain, heart, kidney, limb, lung, and gonad. The antiproliferative effects of gas1 on 10.5 and 12.5 day limb cells were investigated by flow cytometry. In 10.5 day limbs cells, gas1 overexpression could not prevent G 0/G 1 progression. It was determined that gas1 could only induce growth arrest if p53 was also coexpressed. In contrast, gas1 overexpression alone was able to induce growth arrest in 12.5 day limb cells. We also examined the cell cycle profile of gas1-expressing and nonexpressing cells by immunochemistry and flow cytometry. For 10.5 day Gas1-expressing heart and limb cells, we did not find these cells preferentially distributed at G 0/G 1, as compared with Gas1-negative cells. However, in the 12.5 day heart and limb, we did find significantly more Gas1-expressing cells distributed at G 0/G 1 phase than Gas1-negative cells. These results implied that Gas1 alone, during the early stages of development, could not inhibit cell growth. This inhibition was only established when the embryo grew older. We have overexpressed gas1 in subconfluent embryonic limb cells to determine the ability of gas1 to cross-talk with various response elements of important transduction pathways. Specifically, we have examined the interaction of gas1 with Ap-1, NF␬B, and c-myc responsive elements tagged with a SEAP reporter. In 10.5 day limb cells, gas1 overexpression had little effect on Ap-1, NF␬B, and c-myc activities. In contrast, gas1 overexpression in 12.5 day limb cells enhanced AP-1 response while it inhibited NF␬B and c-myc activities. These responses were directly associated with the ability of gas1 to induce growth arrest in embryonic limb cells. In the 12.5 day hindlimb, gas1 was found strongly expressed in the interdigital tissues. We overexpressed gas1 in these tissues and discovered that it promoted interdigital cell death. Our in situ hybridization studies of limb sections and micromass cultures revealed that, during the early stages of chondrogenesis, only cells surrounding the chondrogenic condensations expressed gas1. The gene was only expressed by chondrocytes after the cartilage started to differentiate. To understand the function of gas1 in chondrogenesis, we overexpressed the gene in limb micromass cultures. It was found that cells overexpressing gas1/GFP could not participate in cartilage formation, unlike cells that just express the GFP reporter. We speculated that the reason gas1 was expressed outside the chondrogenic nodules was to restrict cells from being recruited into the nodules and thereby defining the boundary between chondrogenic and nonchondrogenic forming regions. © 2001 Academic Press Key Words: growth arrest specific 1 gene; mouse embryo; cell cycle; signal transduction.

INTRODUCTION The normal development of the embryo involves an integration of cell growth and differentiation. Many growth 1

To whom correspondence should be addressed at Department of Anatomy, Basic Medical Science Building, The Chinese University of Hong Kong, Sha Tin, Hong Kong, People’s Republic of China. Fax: (852) 2603 5031. E-mail: [email protected].

188

factors have now been identified in the developing embryo and their ability to stimulate cell proliferation have been established (Louvi et al., 1997; Webb and Lee, 1997; Webb et al., 1997; Ortega et al., 1998; Vaccarino et al., 1999). However, cell proliferation is not governed solely by the presence or absence of growth factors but a balance between growth-promoting and growth-suppressing factors. The products of the growth arrest-specific (gas) genes are believed to be important growth inhibitors of normal and 0012-1606/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

189

gas1 in Embryo Development

FIG. 1. In situ hybridization showing the expression pattern for gas1 mRNA in E8, 5–10.5 day embryos. Cells stained blue/purple are gas1 ⫹. (A–C) In the 8.5-day-old embryo, gas1 is expressed in the connective tissues (ct), the epicardium and endocardium (ec) of heart, and cells in the somitocoele (sc) of the somites (sm). Gas1 is not expressed in the cephalic neural folds (cf) and neural tube (nt). (D–F) In the 9.5-day-old embryo, the connective tissues (ct) of the head region and trunk continue to express gas1. Gas1 is also detected in the myocardium of the heart (h) and the first branchial arch (1). (G–I) In the 10.5-day-old embryo, gas1 is expressed in the fore-, mid-, and hindbrain (hb), spinal cord (sp), and the first (1) and second branchial arch (2). e, eye; drg, dorsal root ganglion. Bars ⫽ 100 ␮m.

transformed cells, both in vitro (Schneider et al., 1988; Del Sal et al., 1992; Evdokiou and Cowled, 1998a) and in vivo (Evdokiou and Cowled, 1998b). The gas gene family is composed of six members (gas1–gas6). Gas2 is a component of the cytoskeleton (Brancolini et al., 1992). Gas2 overexpression can inhibit cell growth and act as a cell death substrate (Brancolini and Schneider, 1994; Brancolini et al., 1995; Lee et al., 1999). Gas6 is a ligand and has mitogenic and survival activities in chondrocytes (Loeser et al., 1997),

Schwann cells (Li et al., 1996), and vascular smooth muscle cells (Nakano et al., 1997). Presently, we are interested in the role gas1 plays during mouse embryo development. Gas1 is highly expressed in NIH 3T3 fibroblasts under growth arrest conditions, such as high-cell density when cells are contact inhibited and serum deprivation (Schneider et al., 1988). Under growth promoting conditions, such as the addition of growth factors to growth arrested cultures, gas1 expression is rapidly down regulated. Moreover, it has been reported that overexpression of gas1 in fibroblasts and some tumor cell lines can block these cells from proliferating in vitro (Del Sal et al., 1992, 1994; Evdokiou and Cowled, 1998a) and in vivo (Evdokiou and Cowled, 1998b). The gene can apparently prevent cells at G 0 from traversing into the S phase. However, this exit from the cell cycle is not permanent but reversible. Gas1 has now been mapped to mouse chromosome 13 (Webb et al., 1992) and human chromosome 9q21.3-22.1 (Del Sal et al., 1994; Blair et al., 1997). The product of gas1 is located on the cell membrane and it has a large extracellular domain (Del Sal et al., 1992). It has been reported that Gas1-induced growth arrest requires endogenous p53 and that the N-terminal domain of p53 is dispensable for Gas1 transactivation (Del Sal et al., 1995). Ruario et al. (1997) have performed a detailed site-specific mutagenic examination of p53 and identified a proline-rich region (mouse, amino acids 63– 85) that is required for gas1 transduction. Besides p53, c-myc transcriptional factor can also influence gas1 expression. The induction of c-myc can promote cell proliferation and also transcriptionally repress gas1 expression (Lee et al., 1997). Similarly, effects have also been reported for v-Src oncogene (Grossi et al., 1998). These workers proposed that down regulation of gas1 expression plays an important role in regulating the efficient entry of growth arrested cells into S phase. Although we now know some of the physiological functions of gas1 in vitro, little is known of its function in vivo. The mouse embryo is a good model for determining the function of gas1 gene because cell growth and differentiation are an integral part of embryogenesis. Cells in the embryo multiply rapidly during development, so there must be genes involved in arresting this proliferation at the appropriate time, otherwise the embryo will grow out of control. Moreover, growth arrest is closely related to differentiation, with terminally differentiated cells leaving the cell cycle permanently. We have already investigated the role gas2 gene plays in mouse embryo. Gas2 was found to be a multifunctional gene involved in the regulation of programmed cell death and chondrogenesis during development (Lee et al., 1999). In this study, we have (1) established the spatiotemporal expression pattern of gas1 in the developing mouse embryo, (2) determined whether cells expressing gas1 in the embryo were growth arrested, (3) determined whether overexpression of gas1 in embryonic limb cells could inhibit cell proliferation, (4) established that gas1 could induce interdigital cell death, and (5) examined the transduction profile for gas1.

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.

190

Lee et al.

MATERIALS AND METHODS Embryos Postimplantation mouse embryos were harvested from pregnant ICR mice (The Chinese University of Hong Kong). The presence of a vaginal plug was designated as embryonic day 0.5. The mice were killed by cervical dislocation and 9.5- to 14.5-day-old embryos were isolated from the decidua in prewarmed Dulbecco’s phosphatebuffered saline containing 0.4% bovine serum albumin (PB1; Sigma, St. Louis, MO). The embryos were directly fixed in 4% cold paraformaldehyde and used for in situ hybridization and immunohistological studies.

Histology All fixed embryos were dehydrated in a graded series of cold ethanol, cleared in xylene, and embedded in wax. The specimens were sectioned at 7 ␮m and mounted onto to TESPA-treated slides. All solutions used were RNase free.

In Situ Hybridization Digoxigenin-labeled sense and antisense riboprobes were synthesized from a 1.1-kb mouse gas1 cDNA fragment inserted into pBSKII (Stratagene) using T3 and T7 RNA polymerase (Boehringer Mannheim Biochemica), respectively. The size of the probes was reduced by hydrolysis in 40 mM sodium bicarbonate and 60 mM sodium carbonate at 60°C for 15 min. The probes were precipitated with 4 M LiCl and ethanol and then resuspended in in situ hybridization buffer. For in situ hybridization, paraffin sections were dewaxed, hydrated, and equilibrated in PBS. The sections were then treated with 10 ␮g/ml of proteinase K for 10 min and postfixed in 4% paraformaldehyde. The treated specimens were washed twice with PBS for 10 min and then incubated in prehybridization buffer for 3 h. Afterward, 1 ␮g/ml of digoxigenin-labeled sense or antisense riboprobes were diluted in hybridization buffer and added to the treated specimens. The sense probes were used as the control. The hybridization temperature for the sense and antisense riboprobes was 55°C, and the incubation time was 18 h. After hybridization, all unbound cDNA probes were stringently washed free in 2⫻ SSC at 42°C for 10 min (three changes); followed by 60% formamide in 0.2⫻ SSC buffer for 15 min and then 0.2⫻ SSC buffer for 10 min. The location of hybridized probes was determined using alkaline phosphatase-conjugated digoxigenin antibodies. The antibodies were added to the specimens for 1 h, washed in PBS, and developed in NBT/BICP.

Plasmids pRcCMV-gas1, containing the full-length human gas1 cDNA downstream to a CMV promoter (generously supplied by P.A. Cowled), was employed in the overexpression studies. The parental vector, pRcCMV, was used as a negative control. A HAT–Gas1 fusion plasmid was constructed using PCR for the production of recombinant Gas1 protein. Briefly, two sets of primers (forward as 5⬘CGG TCT CGG ATC CCT CGT CTC CTT TCC CCT CCT CTC C and reverse as 5⬘ CTG AGC GGA ATT CCT ATC TGT CCC AAG CCA CAG GTG C) were used to amplify the entire gas1 open reading frame (ORF) from the pRcCMV-gas1 plasmid. Two restriction sites (BamHI and EcoRI) were introduced into the flanking regions of the ORF. The amplified products were digested

with BamHI and EcoRI and subcloned into the corresponding sites of a pHAT10 expression vector (Clontech, Palo Alto, CA). To construct the GFP–Gas1 fusion plasmid, gas1 cDNA was excised from the pHAT-gas1 plasmid at the BamHI and EcoRI sites. The excised gas1 cDNA was then subcloned in frame between the BglII and EcoRI sites of pEGFP-C1 (Clontech) to create pEGFP-Gas1. All of the plasmids were isolated and purified using a Promega’s High Pure Plasmid Isolation Kit (Promega, Madison, WI), according to the manufacturer’s instructions.

Cell Culture The proximal regions of day 12.5 hindlimbs were harvested from embryos in PB1 medium. After rinsing the limb fragments with PBS, they were dissociated by incubation with 0.5% trypsin and 0.25% pancreatin in MEM–Hepes medium (Sigma) containing 2.2% sodium bicarbonate, for 30 min at 4°C, then 10 min at 37°C followed by trituration with a Pasteur pipette to mechanically disrupt the tissues. The enzymatic reaction was inhibited by the addition of 10% fetal bovine serum (FBS; Gibco BRL) and all undissociated clumps of cells were removed by filtration through a Nylon filter (21-␮m pore size) to produce a single-cell suspension. The dissociated cells were pelleted by centrifugation at 250g for 3 min and then resuspended in Dulbecco’s modified essential medium (DMEM; Sigma) containing 10% FBS. Cell concentration was established with an improved Neubauer haemocytometer and viability was assessed with Trypan blue. The cells were plated out at 2 ⫻ 10 5 cells/ml onto 0.1% gelatin-coated coverslips housed inside four-well tissue culture plates. Interdigital and NIH 3T3 cultures were produced in the same manner. The cultures were maintained at 37°C and 5% CO 2 for 48 h. At 24 h, the cells were used for gas1 overexpression studies.

Expression Vector and Cell Transfection To examine the function of gas1 gene, we overexpressed gas1 in embryonic cells and NIH 3T3 fibroblasts. The embryonic cells were derived from the proximal regions of 10.5 and 12.5 day hindlimbs. Cell were also obtained from the interdigital regions of 12.5 day hindlimbs. pRcCMV-gas1, which carries the full-length human gas1 cDNA driven by the cytomegalovirus promoter, was used to transiently express gas1 in cultured cells (Evdokiou and Cowled, 1998a). In all transfection assays, we cotransfected pRcCMV-gas1 with a plasmid containing the green fluorescent protein gene (pEGFP). The pEGFP reporter was used to identify which cells have been transfected. The plasmids were transfected into the cells using cationic liposomes (Lipofectamine Plus, Life Technologies, Inc.). The transfections were performed according to the manufacturer’s instructions.

SEAP Reporter Assay A transduction profile of NIH 3T3 fibroblasts and embryonic limb cells response to gas1 was determined using a Mercury Pathway Profiling System (Clontech). The enhancer elements Ap-1, NF␬B, and E-box (c-myc) were used. These responsive elements were linked with a TAL promoter and a secreted alkaline phosphatase (SEAP) reporter gene. NIH 3T3 fibroblasts and primary limb cells were plated at 1 ⫻ 10 4 cells/well (24-well plate) 24 h prior to transfection. The cells were transfected with pRcCMV-gas1 plasmid (0.5 ␮g/well) and one of the SEAP reporter plasmid (0.5 ␮g/well) using lipofactamine reagent. For the control experiments,


191


the total amount of plasmid DNA introduced was kept constant and the plasmid pRcCMV was used. After 48 h of culture, 25 ␮l of supernatant was harvested from each well of the 24-well plates. A Great EscAPe SEAP Fluorescence Detection Kit (Clontech) was used to determine the concentration of SEAP in the supernatant. All the assays were performed in a white 96-well flat-bottom microtiter plate. Dilution buffer (25 ␮l) was added to each of the 25 ␮l supernatant and mixed gently. The mixture was incubated for 30 min at 65°C. Three microliters of 1 mM MUP was then placed into the samples and incubated for 60 min in the dark at room temperature. The SEAP activity was measured at 360 nm using a Victor-2 Multilabel Counter (EG&G WALLAC). The average (⫾SD) relative fluorescent units (RFU) for each transfection was then determined.

Production of Gas1 Antibody Gas1 polyclonal antibody was produced by immunizing white New Zealand rabbits (approximately 3 kg) with purified recombinant Gas1 protein. The protein was produced by transforming pHAT-gas1 into JM105 bacterium and purified using a HAT Protein Expression and Purification System (Clontech). A 10-ml test bleed was taken from each rabbit prior to immunization. Each rabbit was inoculated with 1 ml of a 1:1 mixture of Gas-1 protein (1 mg/ml H 2O) and Complete Freund’s Aduvant (Calbiochem), subcutaneously. At 3 and 4 weeks, the inoculated rabbits were injected with 1:1 mixture of Gas1 and Incomplete Freund’s Adjuvant (Calbiochem). Blood was then collected from the ears and allowed to clot for 24 h at 4°C, separated by centrifugation (9000g, 15 min), and stored at 4°C after the addition of 0.02% sodium azide. Finally, Gas1 antibodies were affinity purified. Indirect ELISA was performed to confirm the specificity of the antibody.

Flow Analysis All culture cells were dissociated using a solution of 0.1% trypsin and 1 mM EDTA in PBS for 2–5 min at 37°C. Embryonic tissues were also similarly dissociated, but with 0.1% collagenase added. The embryonic tissues were further disaggregated by trituration through a fine-bore pipette. All dissociated cells were immediately fixed in 2% cold paraformaldehyde/PBS for 2 h. For proliferation studies, the cells (1 ⫻ 10 6/ml) were stained in a mixture of 20 ␮g propidium iodide (Sigma), 0.2 mg DNase-free RNase, and 0.1% Triton X-100 made up in 1 ml of PBS. The DNA in the nucleus was stained after 30 min at 37°C. For studying apoptosis, the cells were stained for DNA strand breaks using an In Situ Cell Death Detection Kit (Roche Molecular Biochemicals), according to instructions supplied by the manufacturer. Cells were also immunohistologically stained using our polyclonal Gas1 antibodies, at 20 ␮g/ml for 4 h without permeabilizing the membrane. The treated cells were developed with 1/200 dilution of FITCconjugated anti-rabbit IgG antibody (Zymed, San Francisco, CA). Flow analysis was performed on a Coulter EPIC ALTRA Flow Cytometer (Beckman Coulter). All data were analyzed using an EXPO V.2 Cytometer software and a MODFIT software (Beckman Coulter).

RT-PCR Analysis NIH 3T3 fibroblasts were harvested for PCR analysis 24 h after gas1 transfection. The transfected cells were detached from 35-mm culture dishes by trypsin-digest for 3 min and pooled by centrifu-

gation at 1200g. Total mRNAs were isolated from the cell pellet using a Total mRNA Micro Isolation Kit (Pharmacia). The mRNAs were immediately reversed transcribed into cDNA (T-Prime FirstStrand cDNA Synthesis Kit) according to the manufacturer’s instruction (Pharmacia). PCR was performed using 1 ␮l of the total cDNA mixed with 1⫻ PCR buffer, 1.5 mM MgCl2, 0.2 mM dNTP, 1 ␮M of each gene-specific oligonucleotide primers (mouse ␤-actin forward 5⬘AGGACTCCTA-TGTGGGTGACG and reverse 5⬘CTGGAAAAGAGCCTCAGGGCA; h-Gas1 forward 5⬘GCAAGAAGTGGGCAGGACTTGG and reverse 5⬘AAGAGCGGCCCAAGCAGCA-GC). PCR amplifications were performed 35 cycles, using the following conditions. After 5 min denaturation at 95°C, 35 cycles of PCR were performed with each cycle including denaturation at 95°C (15 s), 30 s annealing at 60°C (30 s), and primer extension at 72°C (45 s). A final extension was carried out for 5 min and stopped at 4°C. The PCR products were visualized following electrophoresis on a 1.5% agarose gel and ethidium bromide staining. The expected products size for mouse ␤-actin is 641 bp; the expected products size for human Gas1 is 1142 bp.

Immunohistochemistry For immunohistochemical analysis, embryo and limb sections were dewaxed in xylene and rehydrated. Endogenous peroxidase was blocked with 0.3% hydrogen peroxide. After inhibiting nonspecific protein interaction with horse serum, the sections were incubated in 1:50 Gas1 polyclonal antibody overnight inside a humid chamber. The antibody staining was viewed using an anti-rabbit ABC kit (Vectastain) with DAB as the substrate.

BrdU Incorporation Studies To determine the effects of gas1 overexpression on cell proliferation, BrdU was added to the cell cultures 5 h before harvest. A BrdU cell staining kit (Zymed) was used to identify cells that have incorporated BrdU. Cell proliferation was also examined in the embryo at different stages of development. BrdU labeling solution (0.2 ml; Amersham Life Science) was injected into the peritoneum of 11.5- to 14.5-day-old pregnant mice. The mice were sacrificed 6 h after injection. The embryos were then extracted from these mice and processed for BrdU analysis.

RESULTS Gas1 Expression in 8.5- to 14.5-day-old Embryos In situ hybridization and immunohistological staining techniques were used to establish the spatiotemporal expression pattern of gas1 in the developing mouse embryo. Gas1 sense riboprobe was used as a negative control. No hybridization signal was detected in all embryos hybridized with the sense probe. In the 8.5-day-old embryo, gas1 was mainly expressed by the soft connective tissues (Fig. 1A), the developing heart (Fig. 1B), and cells located in the somitocoele (Fig. 1C). No expression was found in the cranial neural folds and neural tube. In the 9.5-day-old embryo, gas1 was expressed in the soft connective tissues of the head and trunk (Figs. 1D and 1E) and the epicardium and endocardium of the heart (Fig. 1F). Strong expression was also detected in the first but not the second branchial


192

Lee et al.

FIG. 2. (A) Representative 13.5-day-old embryo stained with Gas1 antibodies. (B) Negative control. Gas1 is strongly expressed in the mantle layer of the hindbrain (C), the intervertebral tissues of the vertebral column (D), the endocardium and myocardium of the heart (E), bronchioles of the lung (F), and soft connective tissues of the kidney (G). hb, hindbrain; int, intervertebral tissues; h, heart; lu, lung; bron, bronchiole; kid, kidney; neph, metanephric tubules. Bars (A, B) ⫽ 500 ␮m; bars (C–G) ⫽ 100 ␮m.

FIG. 3. In situ hybridization showing the expression patterns of gas1 14.5-day-old embryos (A–H). Gas1 is expressed in the hindbrain (B), the vertebral cartilage, intervertebral tissues and spinal cord (C), dorsal root ganglion (D), endocardium and myocardium of the heart (E), bronchioles of the lung (F), kidney and adrenal capsule (G), and mesonephros (H). ad, adrenal gland; atr, atrium; drg, dorsal root ganglion; int, intervertebral tissues; gd, gonad; li, liver; mes, mesonesphros; nt, neural tube; pg, primordial glomerulus; t, tongue; vc, vertebral cartilage; ven, ventricle of heart. Bars (A) ⫽ 500 ␮m; bars (B–H) ⫽ 100 ␮m.


193


dorsal root ganglia, the liver, and the eyes. Both the first and the second branchial arches at this developmental stage expressed gas1. Using Gas1 antibody, the expression pattern of Gas1 protein was examined in 13.5-day-old embryos (Figs. 2A–2G). Gas1 was found strongly expressed in the midbrain (Fig. 2C), the intervertebral mesenchyme (Fig. 2D), and the visceral epicardium and myocardium of the heart (Fig. 2E). Strong expression was also detected in the segmental bronchioles of the lungs (Fig. 2F) and soft connective tissues surrounding the metanephric tubules in the kidney (Fig. 2G). We have also examined the 14.5-day-old embryo because most organ systems were present at this developmental stage. Gas1 was found strongly expressed in most tissues (Fig. 3A). It was strongly expressed in the hindbrain (Fig. 3B), spinal cord, and vertebral cartilage (Fig. 3C), the dorsal root ganglia (Fig. 3D), and the myocardium of the heart (Fig. 3E). In the lung, the epithelial cells lining the bronchioles also strongly expressed gas1 (Fig. 3F). In the kidney, gas1 was mainly expressed in the soft connective tissues and a small number of kidney tubules (Fig. 3G). Gas1 expression in the gonads was patchy but expression in the accompanying mesonephros expression was very strong (Fig. 3H). The overall expression patterns for gas1, at all developmental stages examined, are summarized in Table 1. To determine whether there was any correlation between cell proliferation and gas1 expression, we injected day 13.5 pregnant mice with BrdU. Tissues from the heart and vertebral column were then examined using double labeling with BrdU and Gas1 antibodies. In the heart, it was found that there were Gas1 ⫺/BrdU ⫹, Gas1 ⫹/BrdU ⫺, and Gas1 ⫹/ BrdU ⫹ cardiomyocytes (Fig. 4A). In the intervertebral tissues, the cells were strongly stained for Gas1 and only a few of the cells were Brdu ⫹ (Fig. 4B). In the vertebral cartilage, we found the presence of Gas1 ⫹/BrdU ⫺ and Gas1 ⫹/BrdU ⫹ chondrocytes. These results suggest that some of the cells expressing Gas1 are also capable of incorporating BrdU. However, the data does not imply that Gas1 is incapable of inhibiting cell growth and this is confirmed by quantitative studies using flow cytometry shown below.

Gas1 Expression in the Developing Hindlimb FIG. 4. Double immunohistological staining for the presence Gas1- (red) and BrdU- (green) positive cells. (A) In the myocardium of the heart, it is possible to detect the presence of Gas1 ⫹/BrdU ⫹ (a), Gas1 ⫺/BrdU ⫹ (b), and Gas1 ⫹/BrdU ⫺ (c) cells. (B) In the vertebral column, the intervertebral tissues are strongly stained for Gas1. Very few BrdU ⫹ cells are found in the intervertebral tissues. IV, intervertebral tissue; VC, vertebral cartilage. Bar ⫽ 50 ␮m.

arches. In contrast to 8.5 day, the midbrain, hindbrain, and spinal cord of 10.5-day-old embryos strongly transcribed gas1 (Figs. 1G, 1H, and 1I). No expression was found in the

Gas1 expression in day 12.5–14.5 hindlimbs was examined by in situ hybridization and immunohistochemistry. In the day 12.5 limb, gas1 transcripts (Fig. 5A) and Gas1 proteins (Fig. 5B) were found strongly expressed by the interdigital tissues but not the adjacent digits. The interdigital tissues normally undergo apoptosis at day 13.5 and the process is completed at day 14.5 (Lee et al., 1999). Immunohistological staining of limbs exposed to BrdU revealed that the interdigital regions were filled with BrdUpositive cells at day 12.5 (Fig. 5C). In contrast, few BrdUpositive cells were found in the digits. The results implied that there was no correlation between cell proliferation and gas1 expression. Gas1 expression in the interdigital regions was maintained in the day 13.5 limb (Fig. 5D). In the day


194

Lee et al.

TABLE 1 Localization of gas1 mRNA Expression in the Developing Embryos

Brain

Spinal cord

Dorsal root ganglion

8.5 9.5 10.5

⫺ ⫺ ⫹

⫺ ⫺ ⫹⫹

NA NA ⫺

11.5 12.5 14.5

⫹ ⫹ ⫹⫹

⫹⫹ ⫹⫹ ⫹⫹

⫺ ⫺ ⫹⫹

Stage (days)

Branchial arches NA ⫹ (1st) ⫹ (1st) ⫹ (2nd) NA NA NA

Soft connective tissue

Heart

Lung

Adrenal gland

Kidney

Liver

Lens

Vertebral cartilage

Mesonephros

⫹ ⫹ ⫹

⫹ ⫹ ⫹⫹

NA NA NA

NA NA NA

NA NA NA

NA NA ⫺

NA NA ⫺

⫹a NA ⫺

NA NA NA

⫹ ⫹ ⫹⫹

⫹⫹ ⫹ ⫹⫹

NA ⫹ ⫹⫹

NA ⫺ ⫹c

NA ⫹ ⫹⫹

⫺ ⫺ ⫺

⫺ ⫺ NA

⫺ ⫹ ⫹⫹

⫺b NA ⫹⫹

Note. ⫹, gas1 weakly expressed; ⫹⫹, gas1 strongly expressed; ⫺, no expression. NA, not available. a gas1 weakly expressed in somitocole cells. b gas1 weakly expressed in the connective tissue. c gas1 weakly expressed in the adrenal capsule.

13.5 limb, there were fewer BrdU-positive cells present in the interdigital tissues. BrdU-positive cells were found mainly distributed in the growing digital tips and the perichondrium surrounding the digits (Fig. 5E). By day 14.5, gas1 was expressed in the soft connective tissues, interdigital tissues, cartilage, and developing joints (Fig. 5F). Gas1 expression in the developing cartilage occurred in a proximodistal fashion. The femur, fibula, and tibia (which are developmentally more mature than the metatarsals and digits) strongly expressed gas1, while the less mature metatarsals and digits weakly expressed gas1. Gas1 was also expressed differently within the cartilage. In the femur, fibula, and tibia, gas1 expression was much stronger in the hypertrophic and prehypertrophic regions than the proliferative chondrogenic region (Fig. 5H). In the less mature digits, only chondrocytes in the central but not peripheral regions expressed gas1. Gas1 was also expressed in the in the skeletal muscle cells (Fig. 5G). In the day 15.5 limb, Gas1 protein was strongly expressed in the developing joints (Fig. 5I), which normally undergo apoptosis.

Effects of Gas1 Overexpression on Embryonic Limb Cell and NIH 3T3 Fibroblast Proliferation To determine the action of gas1, we first transfected NIH 3T3 fibroblasts with a human gas/GFP plasmid. The human gas1 plasmid was used in order to distinguish it from the endogenously expressed gas1. The fibroblast cultures were transfected for 24 h. RT-PCR analysis confirmed that the transfected cultures could express human gas1 mRNAs (Fig. 6A). The transfection rate for GFP control (Fig. 6B) and GFP plus gas1 (Fig. 6C) was approximately 25–30%. To test the ability of our gas1 construct to inhibit cell growth, we first transfected the fibroblasts and then pulse labeled them with BrdU. The results showed that fibroblasts transfected with the GFP plasmid did not inhibit BrdU incorporation, i.e., prevent G 0/G 1-S progression (Figs. 7A and 7B). In

contrast, fibroblasts overexpressing gas1/GFP generally did not incorporate BrdU during culture (Figs. 7C and 7D). The antiproliferative effects of gas1 on embryonic cells were also investigated by flow cytometry. The cells were obtained from the hindlimb buds of 10.5- and 12.5-day-old embryos. In 10.5 day limb cells, comparison between gas1/ GFP and GFP overexpression showed no difference in the cell cycle distribution (Table 2). This suggests that gas1 overexpression in 10.5 day limb cells could not prevent G 0/G 1-S progression. It has been reported that the inhibitory action of gas1 on cell growth arrest was dependent on the presence of p53 (Ruaro et al., 1997). Immunohistochemical staining revealed that the 10.5 day hindlimb did not expressed p53 (data not shown). Hence, we examined whether it was the absence of p53 that prevented gas1 from inducing growth arrest in 10.5 day limb cells. The limb cells were transfected with a p53 construct alone (control) or in combination with gas1. The transfected cells were then evaluated by flow cytometry after 24 h culture. No difference was found in the cell cycle profile in cells expressing p53 alone or gas1 alone. However when p53 and gas1 were coexpressed, there were significantly more transfected cells distributed at G 0/G 1 than cells expressing p53 or gas1 alone (Table 2). For 12.5 day limb cells, gas1 overexpression alone was sufficient to increase the number of cells held at G 0/G 1 as compared with GFP transfected cells.

Gas1 Overexpression Induces Interdigital Cell Death The in situ hybridization and immunohistological results demonstrated that gas1 was expressed in the interdigital tissues and interzonal tissues of the developing digits. These tissues normally undergo programmed cell death during development. However, when interdigital tissues are isolated from the limb and maintained in culture, they do not die but instead form cartilage and soft connective tissues (Tang et al., 2000). To determine whether gas1


195


TABLE 2 Effects of Gas1 and p53 Overexpression on Cell Cycle Progression Transfected DNA Cells NIH 3T3 10.5 day Limb cells

12.5 day Limb cells

Percentage of cells

gas1

p53

G 0/G 1

⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹

⫺ ⫺ ⫺ ⫺ ⫹ ⫹ ⫺ ⫺

55.2 ⫾ 4.5 73.1 ⫾ 2.4 a 58.1 ⫾ 1.9 d 59.4 ⫾ 4.2 b 54.4 ⫾ 3.0 c 78.2 ⫾ 6.1 d 56.7 ⫾ 4.7 e 69.9 ⫾ 3.1 e a

S

G 2/M

29.4 ⫾ 3.1 13.2 ⫾ 3.8 28.9 ⫾ 3.1 26.3 ⫾ 2.9 26.9 ⫾ 3.1 14.0 ⫾ 2.0 30.7 ⫾ 2.7 22.2 ⫾ 4.1

13.1 ⫾ 2.1 8.6 ⫾ 1.9 14.0 ⫾ 1.1 16.7 ⫾ 1.7 18.7 ⫾ 2.6 7.2 ⫾ 2.1 14.6 ⫾ 2.2 7.9 ⫾ 1.3

a

Compared with the control, Gas1 overexpression produced significantly more growth arrested cells. The difference in the percentage of cells held at G 0/G 1 was significant (P ⬍ 0.001). b,c Gas1 or p53 overexpression did not induce growth arrest, respectively. d Gas1 and p53 coexpression was capable of producing growth arrest. The difference in the percentage of cells held at G 0/G 1 was significant (P ⬍ 0.002). e At 12.5 day, Gas1 overexpression alone could prevent G 0/G 1 3 S progression (P ⬍ 0.004).

played a role in regulating interdigital cell death, we overexpressed gas1 in 12.5 day interdigital cells in vitro. The interdigital cells were transfected with a plasmid carrying the gas1 full-length cDNA or the empty vector of gas1. The cells were harvested 24 h after transfection. The presence of apoptotic cells was determined using TUNEL staining. In the control, flow cytometry revealed that 48.2 ⫾ 12.1% (N ⫽ 3) of interdigital cells were apoptotic (Fig. 8). For gas1, the apoptotic rate increased dramatically to 85.6 ⫾ 18.9%.

Effects of gas1 Overexpression in Limb Micromass Cultures In situ hybridization revealed that gas1 was not expressed during the early stages of cartilage development in the 11.5to 12.5-day-old limbs. Gas1 was only expressed as the cartilage matured. To examine the role that gas1 played in cartilage development, we examined its expression pattern during chondrogenesis in limb micromass culture. Under high-cell density conditions, the limb micromass cultures produced numerous aggregates after 1-day culture. Lee et al. (1999) reported that WGA (a lectin) labeled with TRITC was a good marker for chondrogenic cells. Presently, we have performed WGA staining to confirm that the aggregates present in the 1-day cultures were composed of chondrogenic cells (Figs. 9A and 9B). These cells did not transcribe gas1 (Fig. 9C). In more mature cultures (after 3 days culture), cells surrounding the chondrogenic nodules but not the chondrocytes expressed gas1 (Fig. 9D). In order to

understand the expression pattern, gas1/GFP and GFP (control) expression plasmids were transfected into the limb micromass cultures. In both the control and the experimental cultures, numerous chondrogenic aggregates were found. In the controls, GFP-expressing cells were found in the perichondral and central regions of these chondrogenic aggregates (Fig. 9E). In contrast, cells overexpressing gas1 were unable to participate in chondrogenesis (Fig. 9F). We found only a few gas1/GFP-positive cells inside the chondrogenic aggregates (Table 3).

Signal Transduction Profile of gas1 We have overexpressed gas1 in subconfluent NIH 3T3 fibroblasts and embryonic limb cells to determine the ability of gas1 to cross-talk with various response elements of important transduction pathways. Specifically, we have examined the interaction of gas1 with Ap-1, NF␬B, and c-myc responsive elements linked with a SEAP reporter. In NIH 3T3 cells, under normal growth conditions, gas1 overexpression increased the activity of Ap-1 responsive elements by approximately 250% (Fig. 10A). In contrast, it inhibited the activity of NF␬B and c-myc target reporters by 50%. In 10.5 day limb cells, gas1 overexpression barely affected Ap-1, NF␬B, and c-myc activity (Fig. 10B). The transduction response of 12.5 day limb cells to gas1 was found to be similar to NIH3T3 cells. Gas1 enhanced Ap-1 response by 24%, while it inhibited NF␬B and c-myc activity by 45 and 68%, respectively (Fig. 10C).

Cell Cycle Analysis of gas1 Expressing Cells Embryonic cells were isolated from the heart and limbs of 10.5-day-old embryos. These cells were fluorescently labeled with Gas1 antibodies and counterstained with propidium iodide. The aim was to determine the relationship between Gas1 expression and the cell cycle in vivo. Flow analysis of the 10.5 day heart revealed that overall the cardiac cells were rapidly dividing with an average of 78.5% cells in the S phase and 11.9% in the G 2/M phase (Table 4). No significant difference was discernable between the Gas1-expressing and nonexpressing cardiac cells. In contrast, there were significantly fewer dividing cells in the 12.5 day heart as compared with 10.5 day. There were significantly more Gas1-positive cells distributed at G 0/G 1 than Gas1-negative cells. (Table 4). For 10.5 forelimb cells, we found these cells proliferating very rapidly, with approximately 57.6% of cells distributed in the S phase and 11.3% in the G 2/M phase. Surprisingly, there were also comparatively less Gas1-positive cells held at G 0/G 1 than Gas1-negative cells (Table 4).

DISCUSSION The proper development of the embryo requires an integration of cell cycle exit and differentiation pathways.


196

Lee et al.

FIG. 5. Gas1 expression in day 12.5–14.5 hindlimbs was examined by in situ hybridization (A, D, F, G, H) and immunohistochemistry (B, C). At 12.5 day (A, B), gas1 was strongly expressed in the interdigital tissues (id) and the cells at the boundary of the developing digits (d). BrdU incorporation assay did not demonstrate a relationship between growth inhibition and gas1 expression. Gas1 was found in the developing joints (F, I), skeletal muscles (G), and hypertrophic chondrocytes (H). tib, tibia; fib, fibula; interzonal mesenchyme (arrows). Bars ⫽ 100 ␮m.

Embryonic growth is regulated by a balance between growth promoting and inhibitory influences. Enzymes such as cyclin-dependent kinases are positive regulators of cell proliferation while inhibitors of these enzymes block G 1/S transition (Harper and Elledge, 1996). Gas1 is also another growth inhibitory factor that is capable of inhibiting G 1/S

progression (Del Sal et al., 1992, 1994). Presently, we have examined the functions of this gene in the developing mouse embryo. We first established the spatial–temporal expression pattern for gas1 in 8.5- to 14.5-day-old mouse embryos. It was found that gas1 was heterogeneously expressed in most embryonic tissues including the brain,


197


FIG. 6. (A) RT-PCR showing pRcCMV-gas1/GFP is capable of inducing human gas1 mRNA expression in mouse 3T3 fibroblasts. (B) Representative fluorescent image of a fibroblast transfected with the GFP reporter gene. (C) Fibroblast transfected with the GFP-gas1 construct. The GFP reporter showed the product of gas1 is localized in the perinuclear region and the cell membrane. Bars ⫽ 15 ␮m.

tified a proline-rich region that is indispensable for gas1 transduction (Ruario et al., 1997). Presently, we have found that gas1 is widely expressed in the developing mouse embryo. To understand how gas1 affected cell proliferation in the embryo, we overexpressed gas1 in 10.5 embryonic limb cells. Gas1 is normally expressed in the mesenchymal cells of 10.5-day-old limbs. Prior to experimentation, we determined that our gas1 construct was capable of inhibiting NIH 3T3 fibroblast proliferation. Using the same construct, we found that gas1 overexpression did not prevent embryonic limb cells from dividing. We speculated the reason 10.5 day limbs cells were incapable of responding to gas1 may be attributed to an absence of p53. Gottlieb et al. (1997) have reported that, besides the tip of the limb bud, there were no p53 transcriptional activities in the limbs of 10.5 day transgenic embryos carrying the p53 responsive promoter. Indeed, we were able to elicit a growth-inhibitory response from 10.5-day-old limb cells when both gas1 and p53 were simultaneously overexpressed. The inhibitory response was greater than p53 overexpression alone and confirmed that gas1 could play an inhibitory role in embryonic cells when all the necessary components were in place. In contrast, overexpression of gas1 alone was sufficient to induce growth arrest in 12.5 day proximal limb cells. To investigate further the function of gas1 in the embryo, we examined the cell cycle profile of gas1-expressing cells by immunochemistry and flow analyses. For 10.5 day Gas1-

heart, kidney, limb, lung, and gonad. Gas1 was expressed especially early in the heart, from 8.5 to 14.5 days. Since it is expressed so early on in development, it argues against a growth inhibitory role for this gene in the embryo. Indeed, our BrdU incorporation assay revealed that overall the Gas1-positive tissues also contained BrdU-positive cells. However, this does not imply that Gas1 is incapable of preventing cell proliferation. Our flow cytometric analysis demonstrated that, quantitatively, Gas1 could inhibit cell growth at specific stages of development and this is discussed below. The gas1 expression pattern in the embryo is very similar, apart from the heart, to gas2 (Lee et al., 1999). This is not surprising since these genes are coexpressed in fibroblasts under growth arrest conditions (Schneider et al., 1988). Gas2 has already been shown to be an important regulator for chondrogenesis and apoptosis. Hence, similar to gas2, gas1 may also play similar roles in these morphogenetic processes.

Gas1 Gene and Growth Arrest The ability of gas1 to inhibit cell growth has been well documented in a number of normal and cancer cell lines (Del Sal et al., 1992, 1994; Evdokiou and Cowled, 1998a,b). Del Sal et al., (1995) have reported that gas1-induced growth arrest required the presence of endogenous p53. Moreover, site-specific mutagenic examination of p53 iden-

FIG. 7. Effects of GFP (A, B) and gas1/GFP (C, D) overexpression on BrdU incorporation (B, D) in NIH 3T3 fibroblasts. Cells overexpressing gas1 generally do not incorporate BrdU. Arrows, BrdUpositive cells. Bar ⫽ 20 ␮m.


198

Lee et al.

TABLE 3 Effect of Gas1 Overexpression on Chondrogenesis DNA transfected

No. of nodules examined

No. of nodules contained GFPpositive cell(s)

% of nodules contained GFPpositive cell(s)

GFP GFP/gas1

44 39

31 2

70.4 5.1

P p

expressing heart and limb cells, we did not find these cells preferentially distributed at G 0/G 1, as compared with Gas1negative cells. Surprisingly, Gas1-positive cells were found at high frequencies both in the S and in the G 2/M phases. In the 12.5 day heart, we did find significantly more Gas1expressing cells distributed at G 0/G 1 phase than Gas1negative cells. These results suggest that Gas1 alone, during the early stages of development, cannot inhibit cell growth. This inhibition is only established when the embryo grows older. Since growth is determined by an intricate balance between growth-promoting and -inhibiting forces, we speculate that the reasons gas1 could not inhibit the growth of early embryonic tissues is because gas1 could not offset the balance of the growth-promoting influences. Indeed, our flow analysis revealed that heart and limb cells normally proliferated very rapidly at 10.5 day, suggesting the presence of a strong growthstimulatory influence. Another explanation is that there may not be an adequate number of negative components present (for example, p53) for gas1 to be effective and that these components are only made available in older embryos.

Gas Genes and Interdigital Cell Death The interdigital tissues are overtly discernible in the 12.5 day hindlimbs following the formation of the digits. These tissues normally undergo apoptosis between 13.5 and 14.5

days so that the digits could separate from each other during limb development (Lee et al., 1994). This morphogenetic process is highly regulated, both spatially and temporally, by cytokines (Macias et al., 1997; Merino et al., 1998), growth factors (Macias et al., 1996), receptors (Yokouchi et al., 1996; Zou and Niswander, 1996), and the physical presence of the ectoderm and digits (Tang et al., 2000). Moreover, we have reported that gas2 is expressed in the interdigital tissues and that it acted as a death substrate during interdigital cell death (Lee et al., 1999). Although it is still not known exactly how these factors interplay to specify interdigital cell death, we do know that it is relatively easy to divert 12.5 day interdigital cells from the apoptotic pathway. Lee et al. (1994) have reported that 12.5 day interdigital cells maintained in culture generally do not die but instead differentiate into chondrogenic and connective tissue cells. Presently, we have demonstrated that gas1 is expressed in the interdigital tissues of 12.5 day hindlimb. To determine the function of gas1 in these tissues, we overexpressed the gene in cultured 12.5 day interdigital cells. We found that up to 86% of interdigital cells, compared with 48% in the control, were induced to undergo apoptosis. These results imply that gas1 is involved in specifying interdigital cell death. Besides the interdigits, the interzonal mesenchyme also strongly express gas1 and these tissues normally die to form the joints in the developing digits. The expression pattern for gas1 is almost identical to those that we have already reported for gas2 in the embryonic limb. We propose that, similar to gas2, gas1 could be added to the growing list of proapoptotic genes involved in the regulation of interdigital cell death. This is the first report that gas1 can induce apoptosis. Previously, gas1 has only been known for its ability to induce growth arrest in cultured cell lines. Moreover, this induction required the presence of p53 (Del Sal et al., 1995; Ruaro et al., 1997). p53 is not expressed in the interdigital tissues of 12.5 and 13.5 day hind limbs (unpublished observation). This

TABLE 4 Cell Cycle Distribution of Embryonic Cells in Vivo Percent Tissue

Stage (days)

Cell attribute

Percent overall

G 0/G 1

S

G 2/M

Heart

10.5

Heart

12.5

Forelimb

10.5

Overall Gas1 Ab ⫹ Gas1 Ab ⫺ Overall Gas1 Ab ⫹ Gas1 Ab ⫺ Overall Gas1 Ab ⫹ Gas1 Ab ⫺

100 60.8 ⫾ 0.7 39.2 ⫾ 0.7 100 64.2 ⫾ 2.1 35.8 ⫾ 2.1 100 85.0 ⫾ 3.7 15.0 ⫾ 3.7

9.6 ⫾ 2.6 14.1 ⫾ 1.3 10.0 ⫾ 0.9 60.7 ⫾ 4.3 77.4 ⫾ 2.0 a 43.0 ⫾ 5.7 a 31.1 ⫾ 3.2 28.7 ⫾ 0.5 c 68.6 ⫾ 0.1 c

78.5 ⫾ 3.4 74.8 ⫾ 6.0 68.9 ⫾ 4.1 28.8 ⫾ 1.4 18.9 ⫾ 1.9 b 40.6 ⫾ 3.7 b 57.6 ⫾ 0.5 62.4 ⫾ 0.8 d 26.7 ⫾ 2.5 d

11.9 ⫾ 6.0 11.1 ⫾ 7.3 21.0 ⫾ 3.2 6.7 ⫾ 2.2 4.0 ⫾ 1.7 11.1 ⫾ 2.3 11.3 ⫾ 2.7 9.0 ⫾ 1.3 4.7 ⫾ 2.4

a,b,c,d

Significant difference in distribution between Gas1-positive and -negative cells (P ⬍ 0.001).


199


implies that the presence of p53 is not required for gas1induced interdigital cell death. We have not attempted to establish which transduction pathway was involved in gas1-induced interdigital cell death because the cells died too fast before our profiling system could be established. Besides the interdigital tissues, we have also detected gas1 expression in other apoptotic regions such as the interzonal and intervertebral mesenchyme.

Signal Transduction In order to understand the functions of gas1, we have examined the ability of gas1 to cross-talk with various responsive elements of major transduction pathways. In 3T3 fibroblasts, we discovered that gas1 overexpression stimulated Ap-1 responsive elements, while it inhibited NF␬B and c-myc responsive elements. Ap-1, NF␬B, and c-myc are transcriptional factors. They all play an important regulatory role in numerous cellular activities. c-myc is a strong potentiator of cell proliferation. Its expression reduces the capacity of cells to differentiate and increases the potential for cells to enter the cell cycle (Bouchard et al., 1998). Since gas1 can prevent fibroblast proliferation, it is not surprising to find that c-myc expression is also correspondingly suppressed. Lee et al. (1997) have previously reported that c-myc can repress gas1 promoter activity. Our results suggest that it can also operate in the opposite direction by gas1-inhibiting c-myc transactivation. In 10.5 day embryonic limb cells, gas1 overexpression had little effects on the c-myc responsive elements. This was expected since gas1 was unable to prevent limb cells from proliferating. However, for 12.5 day limb cells, it was possible to repress the c-myc activity. Besides c-myc, gas1 can also inhibit the nuclear activation of NF␬B in NIH 3T3 fibroblasts. The translocation of cytoplasmic NF␬B into the nucleus is often associated with cell survival (Foo and Nolan, 1999). In fibroblasts that overexpressed gas1, we did not find an increase in the incidence of cell death due to a lack of “survival factors.” This agrees with previous findings that gas1 dose not induce apoptosis in fibroblasts (Del Sal et al., 1992; Evdokiou and Cowled, 1998a). However in our limb experiments, we did find an increase in interdigital cell death. We speculate that this induction may be attributed to the ability of gas1 to inhibit the nuclear translocation of NF␬B. We were unable to profile the interdigital cells to confirm our hypothesis because the cells died too rapidly after gas1 transfection. However, it has recently been reported that interdigital cell death is inhibited in embryos lacking IKK␣ (Takeda et al., 1999). A lack of IKK␣ disrupts NF␬B activation. Besides interdigital cell death, NF␬B is essential for maintaining cell proliferation in the progressive zone and the outgrowth of the limb (Kanegae et al., 1998). In our experiments on 10.5 day limb cells, we found that gas1 slightly elevated nuclear NF␬B. This would suggest that gas1 could weakly stimulate cell proliferation. Indeed in our gas1 flow analyses, we actually found there were

FIG. 8. Flow cytometric analysis of apoptotic cell death in interdigital cells overexpressing gas1. (A) Showing the extent of cell death in interdigital cells transfected with the control plasmid. (B) Showing gas1 overexpression in interdigital cells significantly increased the incidence of interdigital cell death. (C) Negative control, showing the intensity of background staining. M1 designates the region containing TUNEL-positive cells.

significantly more gas1-positive cells in S-phase than gas1 negative cells. The 12.5 day limb cells responded similarly to fibroblasts and this corresponded with a decrease in cell proliferation as observed by flow analyses. The Ap-1 transcriptional factor is a dimeric protein made up of heterodimers between Fos and Jun family gene products or homodimers of Jun family gene products. Ap-1 has been implicated in the regulation of programmed cell death (Marti et al., 1994). It is possible that, through gas1, Ap-1 can also regulate interdigital cell death. However, RofflerTarlov et al. (1996) reported that cell death was unaffected


200

Lee et al.

the apparent differences in the findings. One possible explanation may be that gas1 affected other members of the jun family.

Gas1 and Chondrogenesis

FIG. 9. Limb micromass cultures were used to demonstrate the function of gas1 during chondrogenesis. (A) Showing the presence of numerous newly formed cell aggregations in the limb micromass cultures. (B) These aggregates are composed of chondrogenic cells, as identified by WGA-TRITC staining. In situ hybridization revealed that gas1 is not expressed in the chondrogenic aggregates (pc) during the first 24 h of culture (C). However, after 72 h, gas1 is specifically expressed at the borders of the chondrogenic aggregates (D). To demonstrate the function of gas1, limb mesenchymal cells were transfected with a GFP or a GFP/gas1 expression vector. (E) Limb cells expressing GFP alone (white arrow) were found within the chondrogenic aggregates (white dotted line). In contrast, cells overexpressing GFP/ gas1 (white arrows) could not participate in chondrogenesis (F). Hence, GFP/gas1 expressing cells are found mainly outside the chondrogenic aggregates (white dotted line). pc, prechondrogenic aggregates; c, chondrogenic aggregates; black arrow, region expressing gas1 mRNA. Bar ⫽ 50 ␮m.

in c-fos and c-jun null mutant embryos. Our results appear to contradict previous findings that forced expression of gas1 did not interfere with c-fos and c-jun protein expression (Del Sal et al., 1992). At present we could not explain

The limb skeleton is initially formed as a result of loose limb mesenchymal cells condensing together to produce chondrogenic nodules. These nodules subsequently grow to form a cartilage model that anatomically resembles the skeleton found in the limb. This condensing process involves (1) the migration of putative chondrogenic cells (Ede and Wilby, 1981); (2) cell– cell interaction (Caroline and Kosher, 1991); and (3) cell–matrix interaction (Frenz et al., 1989). These processes can be followed in vitro using limb micromass cultures (Daniels et al., 1996). Our in situ hybridization studies of limb sections and micromass cultures revealed that, during the early stages of chondrogenesis, only cells surrounding the chondrogenic condensations expressed gas1. It was only after the chondrogenic nodules had maturated that gas1 was expressed in the cartilage. To understand the function of gas1 in chondrogenesis, we overexpressed the gene in limb micromass cultures. We found that cells expressing gas1/GFP could not participate in cartilage formation, unlike cells that just express the GFP marker. We speculate that the reason gas1 is expressed outside the chondrogenic nodules is to restrict cells from being recruited into the nodules. It is now established that the product of gas1 is located in the cell membrane and that it contains a large extracellular domain (Del Sal et al., 1992). This domain contains an arginine– glycine– aspartic acid sequence (RGD), which suggests that it could interact with integrin-type molecules (Ruoslahi, 1996). Therefore, gas1 could be involved in cell– cell and cell–matrix interactions. It is possible that cells expressing gas1 are more anchored to the extracellular matrix and thereby inhibited from participating in chondrogenesis. Indeed, Evdokiou and Cowled (1998a) reported that gas1-transfected fibroblasts are very large with a huge cytoplasm containing a single nucleus. In addition, cells transfected with gas1-antisense plasmids grow in a crisscross fashion without contact inhibition. Presently, we have not determined whether gas1 overexpressing limb cells are larger than normal because in micromass cultures it is difficult to judge cell size. In the future, more studies will be required to determine whether gas1 could induce embryonic limb cells to be more adhesive and whether Gas1 could bind with integrins. In sum, it appears that gas1 is a multifunctional gene. The gene is able to induce cell growth arrest at late, but not early, stages of embryo development. In addition, gas1 is involved in regulating limb chondrogenesis and interdigital cell death.


201


FIG. 10. Transduction profile of NIH 3T3 fibroblasts, 10.5 day limb cells and 12.5 limb cells. (A) gas1 overexpression (hatched bars) in NIH 3T3 cells stimulated Ap-1 responsive element while it inhibited ␬B and E-box responsive elements. (B) 10.5 day limb cells were generally unresponsive to gas1 overexpression. Only a slight stimulation was obtained for ␬B responsive elements. In contrast, 12.5 day limb cells responded to gas1 overexpression in the same manner as NIH 3T3 cells (C).

ACKNOWLEDGMENTS We are extremely grateful to Dr. P. A. Cowled for providing the gas1 probes. This study was supported by an RGC Earmarked Grant 4273/98M awarded to K. K. H. Lee.

REFERENCES Blair, I. P., Dawkins, J. L., and Nicholson, G. A. (1997). Fine mapping of the hereditary sensory neuropathy type I locus on

chromosome 8q22.1 3 q22.3: Exclusion of GAS1 and XPA. Cytogenet. Cell Genet. 78, 140 –144. Bouchard, C., Staller, P., and Eilers, M. (1998). Control of cell proliferation by Myc. Trends Cell Biol. 8, 202–206. Brancolini, C., Bottega, S., and Schneider, C. (1992). Gas2, a growth arrest-specific protein, is a component of microfilament network system. J. Cell Biol. 117, 1251–1261. Brancolini, C., and Schneider, C. (1994). Phosphorylation of the growth arrest-specific protein Gas2 is coupled to actin rearrangements during G0 3 G1 transition in NIH 3T3 cells. J. Cell Biol. 124, 743–756. Brancolini, C., Benedetti, M., and Schneider, C. (1995). Microfilament reorganization during apoptosis: the role of Gas2, a


202

Lee et al.

possible substrate for ICE-like proteases. EMBO J. 14, 5179 – 5190. Caroline, C. N. D., and Kosher, R. A. (1991). Gap junctional communication during limb cartilage differentiation. Dev. Biol. 144, 47–53. Cowled, P. A., Ciccarelli, C., Coccia, E., Philipson, L., and Sorrentino, V. (1994). Expression of growth arrest-specific (gas) genes in senescent murine cells. Exp. Cell Res. 211, 197–202. Daniels, K., Reiter, R., and Solursh, M. (1996). Micromass cultures of limb and other mesenchyme. [Review] Methods Cell Biol. 51, 237–247. Del Sal, G., Ruaro, M. E., Philipson, L., and Schneider, C. (1992). The growth arrest specific gene gas 1 is involved in growth suppression. Cell 70, 595– 607. Del Sal, G., Collavin, L., Ruaro, M. E., Edomi, P., Saccone, S., Della Valle, G., and Schneider, C. (1994). Structure, function, and chromsome mapping of the growth-suppressing human homologue of the murine gas1 gene. Proc. Natl. Acad. Sci. USA 91, 1848 –1852. Del Sal, G., Ruaro, M. E., Utrera, R., Cole, C. N., Levine, A. J., and Schneider, C. (1995). Gas1-induced growth suppression requires a transactivation-independent p53 function. Mol. Cell Biol. 15, 7152–7160. Ede, D. A., and Wilby, O. K. (1981). Golgi orientation and cell behaviour in the developing pattern of chondrogenic condensations in chick limb bud mesenchyme. Histochem. J. 13, 615– 630. Evdokiou, A., and Cowled, P. A. (1998a). Growth-regulatory activity of the growth arrest-specific gene gas1, in NIH3T3 fibroblasts. Exp. Cell Res. 240, 359 –367. Evdokiou, A., and Cowled, P. A. (1998b). Tumor-suppressive activity of the growth arrest-specific gene, gas1 in human tumor cell lines. Int. J. Cancer 75, 568 –577. Foo, S. Y., and Nolan, G. P. (1999). NF-kappaB to the rescue: RELs, apoptosis and cellular transformation. [Review] Trends Genet. 15, 229 –235. Frenz, D. A., Jaikaria, N. S., and Newman, S. A. (1989). The mechanism of precartilage mesenchymal condensation: A major role for interaction of the cell surface with the amino-terminal heparin-binding domain of fibronectin. Dev. Biol. 136, 97–103. Gottlieb, E., Haffner, R., King, A., Asher, G., Gruss, P., Lonai, P., and Oren, M. (1997). Transgenic mouse model for studying the transcriptional activity of the p53 protein: Age- and tissuedependent changes in radiation-induced activation during embryogenesis. EMBO J. 16, 1381–1390. Grossi, M., La Rocca, S. A., Pierluigi, G., Vannucchi, S., Ruaro, W. M., Schneider, C., and Tato, F. (1998). Role of gas1 downregulation in mitogenic stimulation of quiescent NIH3T3 cells by v-Src. Oncogene 17, 1629 –1638. Harper, J. W., and Elledge, S. J. (1996). Cdk inhibitors in development and cancer. Curr. Opin. Genet. Dev. 6, 56 – 64. Kanegae, Y., Tavares, A. T., Izpisua Belmonte, J. C., and Verma, I. M. (1998). Role of Rel/NF-kappaB transcription factors during the outgrowth of the vertebrate limb. Nature 392, 611– 614. Lee, K. K. H., Li, F. C. H., Yung, W. T., Kung, J. L. S., Ng, J. L., and Cheah, K. S. E. (1994). Influence of digits, ectoderm, and retinoic acid on chondrogenesis by mouse interdigital mesoderm in culture. Dev. Dyn. 201, 297–309. Lee, K. K. H., Tang, M. K., Yew, D. T. W., Chow, P. H., Yee, S. P., Schneider, C., and Brancolini, C. (1999). Gas2 is a multifunctional gene involved in the regulation of apoptosis and chondrogenesis in the developing mouse limb. Dev. Biol. 207, 14 –25.

Lee, T. C., Li, L., Philipson, L., and Ziff, E. B. (1997). Myc represses transcription of the growth arrest gene gas1. Proc. Natl. Acad. Sci. USA 94, 12886 –12891. Li. R.-H., Chen, J., Hammonds, G., Phillips, H., Armanini, M., Wood, P., Bunge, R., Godowski, P. J., Sliwkowski, M. X., and Mather, J. P. (1996). Identification of gas6 as a growth factor for human schwann cells. J. Neurosci. 16, 2012–2019. Loeser, R. F., Varnum, B. C., Carlson, C. S., Goldring, M. B., Liu, E. T., Sadiev, S., Kute, T. E., and Wallin, R. (1997). Human chondrocytes expression of growth-arrest-specific gene 6 and the tyrosine kinase receptor axl: Potential role in autocrine signaling in cartilage. Arthritis Rheum. 40, 1455–1465. Louvi, A., Accili, D., and Efstratiadis, A. (1997). Growth-promoting interaction of IGF-II with the insulin receptor during mouse embryonic development. Dev. Biol. 189, 33– 48. Macias, D., Ganan, Y., Ros, M. A., and Hurle, J. M. (1996). In vivo inhibition of programmed cell death by local administration of FGF-2 and FGF-4 in the interdigital areas of the embryonic chick leg bud. Anat. Embryol. 193, 533–541. Macias, D., Ganan, Y., Sampath, T. K., Piedra, M. E., Ros, M. A., and Hurle´, J. M. (1997). Role of BMP-2 and OP-1 (BMP-7) in programmed cell death and skeletogenesis during chick limb development. Development 124 1109 –1117. Marti, A., Jehn, B., Costello, E., Keon, N., Ke, G., Martin, F., and Jaggi, R. (1994). Protein kinase A and AP-1 (c-Fos/JunD) are induced during apoptosis of mouse mammary epithelial cells. Oncogene 9, 1213–1223. Merino, R., Ganan, Y., Macias, D., Economides, A. N., Sampath, K. T., and Hurle, J. M. (1998). Morphogenesis of digits in the avian limb is controlled by FGFs, TGF␤s, and noggin through BMP signaling. Dev. Biol. 200, 35– 45. Nakano,T., Kawamoto, K., Kishino, J., Nomura, K., Higashino, K., and Arita, H. (1997). Requirement of gamma-carboxyglutamic acid residues for the biological activity of Gas6: Contribution of endogenous Gas6 to the proliferation of vascular smooth muscle cells. Biochem. J. 323, 387–392. Ortega, S., Ittmann, M., Tsang, S. H., Ehrlich, M., and Basilico, C. (1998). Neuronal defects and delayed wound healing in mice lacking fibroblast growth factor 2. Proc. Natl. Acad. Sci. USA 95, 5672–5677. Roffler-Tarlov, S., Brown, J. J., Tarlov, E., Stolarov, J., Chapman, D. L., Alexiou, M., and Papaioannou, V. E. (1996). Programmed cell death in the absence of c-Fos and c-Jun. Development 122, 1–9. Ruaro, M. E., Collavin, L., Del Sal, G., Haffner, R., Oren, M., Levine, A. J., and Schneider, C. (1997). A proline-rich motif in p53 is required for transactivation-dependent growth arrest as induced by Gas1. Proc. Natl. Acad. Sci. USA 94, 4675– 4680. Ruoslahi, E. (1996). RDG and recognition sequences for integrins. Annu. Rev. Cell Dev. Biol. 12, 697–715. Schneider, C., King, R. M., and Philipson, L. (1988). Genes specifically expressed at growth arrest of mammalian cells. Cell 54, 787–793. Takeda, K., Takeuchi, O., Tsujimura, T., Itami, S., Adachi, O., Kawai, T., Sanjo, H., Yoshikawa, K., Terada, N., and Akira, S. (1999). Limb and skin abnormalities in mice lacking IKKalpha. Science 284, 313–316. Tang, M. K., Leung, A. K., Kwong, W. H., Chow, P. H., Chan, J. Y., Ngo-Muller, V., Li, M., and Lee, K. K. H. (2000). Bmp-4 requires the presence of the digits to initiate programmed cell death in limb interdigital tissues. Dev. Biol. 218, 89 –98.


203


Vaccarino, F. M., Schwartz, M. L., Raballo, R., Nilsen, J., Rhee, J., Zhou, M., Doetschman, T., Coffin, J. D., Wyland, J. J., and Hung, Y. T. (1999). Changes in cerebral cortex size are governed by fibroblast growth factor during embryogenesis. Nat. Neurosci. 2, 246 –253. Webb, G. C., Cowled, P. A., Evdokiou, A., Ford, J. H., and Forbes, I. J. (1992). Assignment, by in situ hybridization, of the growth arrest-specific gene, Gas-1, to mouse chromosome 13 and bands B3–C2. Genomics 14, 548 –549. Webb, S. E., and Lee, K. K. H. (1997). Effect of platelet-derived growth factor isoforms on the migration of mouse embryo limb myogenic cells. Int. J. Dev. Biol. 41, 597– 605. Webb, S. E., Lee, K. K. H., Tang, M. K., and Ede, D. A. (1997).

Fibroblast growth factor 2 and 4 stimulate migration of mouse embryonic limb myogenic cells. Dev. Dyn. 209, 206 –216. Yokouchi, Y., Sakiyama, J., Kameda, T., Iba, H., Suzuki, A., Ueno, N., and Kuroiwa, A. (1996). BMP-2/-4 mediate programmed cell death in chicken limb buds. Development 122, 3725–3734. Zou, H., and Niswander, L. (1996). Requirement for BMP signaling in interdigital apoptosis and scale formation. Science 272, 738 – 741. Received for publication September 11, 2000 Accepted February 26, 2001 Published online April 25, 2001
