Differentiation of Embryonal Stem Cells into ... - ScienceDirect.com

DEVELOPMENTAL BIOLOGY 179, 184–196 (1996) ARTICLE NO. 0250

Differentiation of Embryonal Stem Cells into Keratinocytes: Comparison of Wild-Type and b1 Integrin-Deficient Cells Claudia Bagutti, Anna M. Wobus,* Reinhard Fa¨ssler,† and Fiona M. Watt1 Imperial Cancer Research Fund, London WC2A 3PX, United Kingdom; *Institut fu¨r Pflanzengenetik und Kulturpflanzenforschung, 06466 Gatersleben, Germany; and †Max Planck Institut fu¨r Biochemie, 82152 Martinsried/Munich, Germany

b1 Integrins are known to regulate terminal differentiation and morphogenesis in the adult epidermis. We have investigated their role in the embryonic development of keratinocytes by comparing the differentiation of wild-type and b1-null mouse embryonal stem (ES) cells. By 12–15 days in culture, differentiation of embryonic or simple epithelial cells occurred in both ES cell populations, as detected by expression of keratins 8, 18, and 19. From 21 days, expression of keratins 10 and 14 and of the cornified envelope precursor involucrin indicated that some of the wild-type cells had differentiated into keratinocytes. In contrast, keratinocyte markers were not expressed in b1-null cultures. The b1-null cells failed to express the a2 and a3 integrin subunits on the cell surface, consistent with the association of these a subunits with b1 . Furthermore, a6 and b4 expression was reduced in the b1-null cultures. Although b1-null ES cells failed to undergo differentiation into keratinocytes in vitro, they did form keratinocyte cysts expressing a6b4 , keratins 1 and 14, and involucrin when allowed to form teratomas by subcutaneous injection in mice; furthermore, b1-null keratinocytes were found in the epidermis of a wild-type/b1-null chimeric mouse. As judged by immunofluorescence microscopy, extracellular matrix assembly was severely impaired in b1-null ES cell cultures, but not in the teratomas or chimeric mouse skin. We therefore speculate that the failure of b1-null cells to differentiate into keratinocytes in vitro may reflect an inability to assemble a basement membrane. q 1996 Academic Press, Inc.

INTRODUCTION In epidermis, b1 integrins not only mediate adhesion to extracellular matrix (ECM), but also regulate terminal differentiation and tissue assembly (Watt and Hertle, 1994). Furthermore, different subpopulations of proliferating keratinocytes can be distinguished by the levels of b1 integrins they express (Jones and Watt, 1993; Jones et al., 1995). In keratinocytes, the b1 subunit forms heterodimers with several a subunits, including a2 , a3 , and a5 , forming receptors for collagen, laminin, and fibronectin, respectively (reviewed by Watt and Hertle, 1994). Although the a6 subunit forms a6b1 heterodimers in other cell types, it is found exclusively as an a6b4 heterodimer in keratinocytes. a6b4 mediates binding of keratinocytes to laminin and is a component of hemidesmosomes (Sonnenberg et al., 1991). 1 To whom correspondence should be addressed. Fax: 44-171-2693078.

In view of the role of b1 integrins in regulating the behavior of keratinocytes in the adult organism, they might be expected to play a role in the differentiation of keratinocytes during embryogenesis. Targeted deletions of the b1 gene in mice have recently been shown to lead to embryonal death shortly after implantation (Fa¨ssler and Meyer, 1995; Stephens et al., 1995), before the epidermis develops from a single ectoderm layer. Alternative strategies to study keratinocyte differentiation are to examine in vitro differentiation of b1-null ES cells, teratoma formation by ES cells (e.g., Lee et al., 1995), and the epidermis of wild-type/b1-null chimeric mice. ES cells cultured in vitro retain their undifferentiated pluripotent state (Doetschmann et al., 1985). Differentiation is induced upon aggregation of ES cells to form embryoid bodies from which the cells differentiate spontaneously into a large variety of cell types. ES cell cultures are therefore a useful tool for analyzing differentiation of cells carrying a mutation lethal early in development. In recent years, several in vitro models of ES cell differentiation have been 0012-1606/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

184

AID

DB 8341

/

6x13$$$221

09-23-96 15:23:16

dbal

AP: Dev Bio

185

Epithelial Differentiation of b1-Null ES Cells

described which give rise to a variety of cell types, including hematopoietic (Wiles and Keller, 1991), neuronal (Stru¨bing et al., 1995), myogenic (Rohwedel et al., 1994), and cardiac muscle (Wobus et al., 1991; Maltsev et al., 1994) cells. Epithelial cells can be identified by the expression of keratin intermediate filaments, the composition of which, paired type I (acidic) and type II (basic/neutral) keratins, is specific for particular types of epithelial differentiation and development (Lane and Alexander, 1990). Keratin 8 (K8) and K18 form the keratin pair that is expressed earliest during embryonic development (Oshima et al., 1983) and serves as a marker for simple epithelium. K19 can associate with different partners and is expressed in embryonic, simple, and stratified epithelia (Stasiak and Lane, 1987). The keratin pair K5 and K14 is a marker for the basal layer of epithelia, such as the epidermis, that can undergo stratification. In the epidermis, keratinocytes that undergo terminal differentiation switch from expression of K5 and K14 to K1 and K10 (Fuchs and Green, 1980). During terminal differentiation keratinocytes also initiate expression of proteins that form the cornified envelope of the outermost cell layers: involucrin is an abundant envelope precursor in both mouse and human epidermis (Simon, 1994). Comparison of the capacities of wild-type and b1-null ES cells to differentiate in vitro has already proved useful for investigating the role of b1 integrins in the differentiation of cells into cardiac muscle (Fa¨ssler et al., submitted for publication). This approach has therefore been applied to keratinocytes. Here, we report on the expression of markers of epithelial differentiation in wild-type and b1-null ES cells in vitro and after subcutaneous injection in vivo. In addition, we have examined whether the epidermis of a wildtype/b1-null chimeric mouse contains b1 integrin-deficient keratinocytes.

MATERIALS AND METHODS Cell culture and in vitro differentiation. The wild-type embryonic stem cell line D3 (Doetschmann et al., 1985) was cultured on a feeder layer of mitomycin C-treated primary mouse embryonic fibroblasts in Dulbecco’s modified Eagle’s medium supplemented with 15% fetal calf serum (FCS), 2 mM L-glutamine, nonessential amino acids, 50 U/ml penicillin/streptomycin, and 50 mM b-mercaptoethanol (all reagents from Gibco BRL) (Wobus et al., 1984, 1991). The b1 integrin-deficient cell line D3/G-201 (Fa¨ssler et al., 1995) was grown without any feeder layer in the above medium containing 5000 U/ml of leukemia-inhibiting factor (LIF, Williams et al., 1988). The procedure of differentiation is schematically illustrated in Fig. 1. Embryoid body formation was initiated in hanging drops of 600 cells in 20 ml of medium containing 20% FCS (Rudnicki and McBurney, 1978). After 2 days the embryoid bodies were washed off and cultivated for another 3 (D3) or 7 (G-201) days in suspension in bacteriological petri dishes (Greiner) before they were transferred to gelatinized (D3) or uncoated (G-201) cell culture dishes (Falcon). Coverslips were added to culture dishes to be used for immunofluorescence analysis. ES cell tumors. Tumors were generated by subcutaneous injec-

tion of 1 1 107 D3 or G-201 cells on the back of male 129/Sv mice. When the D3 tumors reached a volume of approximately 1 g, both D3 and G-201 mice were sacrificed. The tumors were dissected, embedded in OCT (Miles Inc.), and frozen, and 6-mm cryosections were cut. The first and last of each series of sections were stained with hematoxylin and eosin. Antibodies. The following primary monoclonal antibodies were used: LL001, LE41, and LE61 (mouse-anti K14, K8, and K18, respectively; Lane, 1982; Purkis et al., 1990), LP2K and LL25 (mouse anti-K19 and K16, respectively, kind gift of E. B. Lane, University Dundee, UK), MB1.2 (rat-anti mouse b1 integrin subunit, kindly provided by B. M. C. Chan, University Western Ontario), 346-11A (rat anti-mouse b4 , kind gift of S. J. Kennel, Oak Ridge; Kennel et al., 1986), Hal/29 (hamster anti-rat a2 , kindly provided by D. Mendrick, Harvard Medical School; Mendrick and Kelly, 1993), GoH3 (rat anti-a6 , Serotec; Sonnenberg et al., 1986). We also used the following rabbit antisera: LH1 (anti-K1, kind gift of E. B. Lane, University Dundee, UK), anti-mouse involucrin (from BAbCo), DH4 (rabbit antiserum to cytoplasmic domain of the human a3 subunit; Bishop et al., 1995), anti-mouse type IV collagen and laminin 1 (both from Becton–Dickinson), and anti-mouse fibronectin (from Biogenesis). The secondary antibodies used were: FITC-conjugated rabbit anti-mouse, donkey anti-rabbit, goat anti-rat, goat anti-hamster IgG (all from Sigma Chem.), DTAF-conjugated goat anti-mouse and Texas Red-conjugated donkey anti-mouse, donkey anti-rabbit, goat anti-rat, and goat anti-hamster IgG (all from Jackson Lab.). Immunofluorescence. Embryoid bodies grown on coverslips were rinsed in PBS, fixed with either acetone for 8 min at 0207C or 4% paraformaldehyde for 10 min at room temperature, and then incubated with 10% FCS/PBS for 30 min at 377C. Incubation with both primary and secondary antibodies was performed for 30 min at 377C and the cells were washed in PBS after each incubation. Finally the specimens were mounted in Gelvatol and examined and photographed using a Zeiss Axiophot microscope (Carl Zeiss Ltd.) or, where indicated, a MRC-600 laser scanning confocal microscope (Bio-Rad Microscience). The cryosections of ES cell tumors and of wild-type/b1-null chimeric mouse skin (see below) were fixed with acetone for 8 min at 0207C and otherwise treated in the same way as described for the embryoid body cultures. Flow cytometry analysis. Embryoid body cultures were disaggregated using 0.05% trypsin/0.02% EDTA and incubated with anti-b1 , anti-b4 , or anti-a6 antibodies diluted in PBS for 15 min on ice. After washing with cold PBS containing 10% FCS, the cells were resuspended in the appropriate FITC-conjugated secondary antibody as before. Control samples were labeled with secondary antibodies alone. After washing, the cells were passed through a FACScan (Becton– Dickinson). Five thousand to 10,000 cells were analyzed per sample and dead cells (positive for propidium iodide) were gated out. RT–PCR analysis. Fifty embryoid bodies including outgrowths were harvested in 500 ml of 4 M guanidinium thiocyanate, 0.1 M Tris/HCl, pH 7.5, and 0.1% b-mercaptoethanol. Poly(A)/ RNA was isolated using HYBOND-messenger affinity paper (Amersham) according to the method of Sheardown (1992). DNase I-digested poly(A)/ RNA was reverse transcribed using random hexamer primers and MMLV reverse transcriptase (Gibco, Paisley, UK). The products of the reverse transcription reactions were denatured for 5 min at 957C and one-tenth was subjected to amplification by 30 cycles of PCR at 607 and 947C. The PCR products were then separated on 1.9% agarose gels. The oligonucleotide pairs used for PCR amplification were chosen from the published sequences of the individual genes (Steinert et al., 1983; Krieg et al., 1985; Singer et al., 1986; Tokunaga et al., 1986; Djian et al., 1993; Gandarillas and Watt, 1995) and were as follows:

Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

AID

DB 8341

/

6x13$$$222

09-23-96 15:23:16

dbal

AP: Dev Bio

186

Bagutti et al.

FIG. 1. Schematic illustration of the experimental protocol used for the in vitro differentiation studies.

K18, 5*-TTGTCACCACCAAGTCTGCC-3* and 5*-TTTGTGCCAGCTCTGACTCC-3*; K14, 5*-GTGTCCACTGGCGATGTGAACGTGG-3* and 5*-GCTGCCGCAGTAGCGACTCTACTGT-3*; K10, 5*-CGCAAGGATGCTGAAGAGTGGTTC-3* and 5*-TGGTAC-

TCGGCGTTCTGGCACTCGG-3*; involucrin, 5*-GGTGTACAGAAGCTTCCAAGATGTCC-3* and 5*-GGCATTGTGTAGGATGTGGAGTTGG-3*; and actin, 5*-GTTTGAGACCTTCAACACCCC-3* and 5*-GTGGCCATCTCCTGCTCGAAGTC-3*.

TABLE 1 Immunofluorescence Staining of Embryoid Body Cultures at Different Stages of Differentiation Time of differentiation (days in culture) 12 Protein K19 K8/18c K14 K1 Involucrin

15

21

23

27

33

Wild-type b1-Null Wild-type b1-Null Wild-type b1-Null Wild-type b1-Null Wild-type b1-Null Wild-type b1-Null /a 0 0 n.d.d n.d.

/ //0 0 n.d. n.d.

/ / 0 n.d. n.d.

/ / 0 n.d. n.d.

//0 / / 0 0

//0 / 0 0 0

//0 / / 0 0

//0 / 0 0 0

//0 / / 0 0

a

//0 / 0 0 0

Scoring was as follows: (/) many positive patches, (//0) a few positive patches, (0) no positive staining at all. Occasional positively stained cells were detected. c Staining with antibodies to K8 and K18 gave the same results. d n.d., not determined. b


AID

DB 8341

/

6x13$$$222

09-23-96 15:23:16

dbal

AP: Dev Bio

0b / / 0 0

0b / 0 0 0

187


FIG. 2. Expression of keratins in wild-type and b1-null embryoid body cultures. Immunofluorescent staining of (A, B) K8 and (C, D) K14 on the 22nd day of culture. E and F represent secondary antibody controls (ns). (A, C, E) Wild-type; (B, D, F) b1-null ES embryoid bodies. Bar in F, 33 mm (C, D, E, F); 82 mm (A, B).

Preparation and analysis of wild-type/b1-null chimeric mouse skin. Wild-type/b1-null chimeric mice were generated as described by Fa¨ssler and Meyer (1995). Skin samples from an adult mouse were embedded in OCT (Miles Inc.) and snap frozen in liquid nitrogen-cooled isopentane. Seven-micrometer sections were cut on a cryostat. The b-galactosidase gene product was identified by X-Gal staining as described by Carroll et al. (1993).

RESULTS Detection of Epithelial-Specific Gene Products by Immunofluorescence and RT–PCR The wild-type D3 mouse ES cells were grown on a feeder layer of mouse embryonal fibroblasts and had the typical


AID

DB 8341

/

6x13$$$222

09-23-96 15:23:16

dbal

AP: Dev Bio

188

Bagutti et al.

FIG. 3. Detection of epithelial-specific gene expression by RT–PCR at different stages of differentiation (A) (i.e., at 5 (wild-type only), 8, 12, 15, 20, 23 days of culture) and (C) in undifferentiated ES cells and mouse embryonal (ME) feeder cells. Sizes of the amplified products in (A) are as follows: K18, 213 bp; K14, 330 bp; K10, 278 bp; inv (involucrin), 150 bp; actin (control), 320 bp. (B) Control reactions were performed using actin primers and omitting the reverse transcriptase.

morphology in the undifferentiated state (Doetschmann et al., 1985). In contrast, their b1-deficient variants D3/G-201 (Fa¨ssler et al., 1995), which have been generated by targeted disruption of the b1 gene in D3 cells, creating homozygous b1 0/0 cells, would only grow without a feeder layer and displayed a significantly different morphology (Fa¨ssler et al., 1995). In order to study the influence of b1 on the differentiation of ES cells into epithelial cells, we used the in vitro differentiation protocol illustrated in Fig. 1.

The differentiation of both wild-type D3 and b1-deficient G-201 ES cells was initiated by growing the cells in hanging drops, where they formed aggregates within 2 days (Fig. 1). Thereafter, the growing embryoid bodies were cultured in suspension for another 3 (wild-type) or 7 (b1-null) days. As described recently (Fa¨ssler et al., submitted for publication), it was necessary to incubate wild-type and b1-null cells for different times in suspension, in order to achieve equivalent development. The 5 (wild-type)- and 9 (b1-null)-day-old em-


AID

DB 8341

/

6x13$$$222

09-23-96 15:23:16

dbal

AP: Dev Bio

FIG. 4. Expression of integrins by wild-type and b1-null embryoid body cultures at 21 days. Confocal microscope analysis of (A, B) b1 ; (C, D) a3 ; (E, F) a6 ; and (G, H) b4 subunits. I and J represent secondary antibody controls (ns). (A, B, E–H) Nonpermeabilized cells; (C, D, I, J) acetone-permeabilized. Bar in J, 40 mm (A, B, E, F, I, J) and 28 mm (C, D, G, H). 189

09-23-96 15:23:16

dbal

AP: Dev Bio

190

Bagutti et al.

FIG. 5. Flow cytometry analysis of integrin expression. Wild-type (solid lines) and b1-null cells (stippled lines) were cultured for 39 (A, B) and 25 (C) days. Antibodies were used to recognize the b1 (A), a6 (B), and b4 (C) integrin subunits. Vertical mark on each x axis indicates maximal fluorescence of cells incubated with the secondary antibody alone.

bryoid bodies were allowed to attach to culture dishes so that cells could grow out from the embryoid bodies. At various stages of differentiation, the embryoid bodies and

their outgrowths were examined for the presence of epithelial-specific proteins by immunofluorescent staining and for the corresponding mRNAs by RT–PCR. As indicated by the presence of keratin 19 filament networks in both wild-type and b1-null cells (Table 1), the first epithelial cells appeared after 12 days in culture. After 3 additional days, two further keratins, K8 and K18, were detected in both wild-type and b1-null embryoid bodies, a few positive cells being present as early as Day 12 in the b1-null cells (Table 1, Figs. 2A and 2B). From the 21st day of culture onward, patches of K14positive cells appeared in the wild-type cultures (Table 1, Figs. 2C and 2D). In contrast, K14 was not detected in b1deficient embryoid bodies at the corresponding stage. Both types of culture were still positive for K8 and K18 but the number of cells expressing K19 began to decline (Table 1). The presence of K14 in the wild-type embryoid bodies suggested that keratinocytes might have arisen in the cultures. However, when the cultures were examined for two markers of the epidermal keratinocyte terminal differentiation programme, K1 and involucrin, neither protein was detected (Table 1). Even at later stages (27th–33rd days of differentiation), when the wild-type and b1-null embryoid bodies had lost K19 almost completely but were still positive for K8 and K18, immunofluorescent staining for K1 and involucrin was negative (Table 1). In addition, antibodies to K16, a keratin marker of terminal differentiation in hyperproliferative epidermis (Lane and Alexander, 1990), never gave positive staining of the cultures (data not shown). To compare the expression of the epithelial and epidermal cell-specific proteins with the expression of the corresponding mRNAs and to use a more sensitive method of detecting gene expression, we performed RT–PCR. For this purpose, mRNA was isolated from the embryoid bodies of both wildtype and b1-null cells and reverse transcribed. Specific primers were used to amplify the transcripts of K18, K14, K10, and involucrin. As a control, PCR was performed with actin primers on all RNA lysates without prior reverse transcription reaction (shown in Fig. 3B). In addition, all epithelialspecific genes tested were found to be activated only upon differentiation since undifferentiated ES cells failed to produce the transcripts (Fig. 3C). As shown in Fig. 3A, K18 transcripts were already detectable in 5-day wild-type and 8-day b1-null embryoid bodies, several days earlier than the proteins were visualized by immunofluorescence. The RT–PCR results for the b1-null embryoid bodies confirmed those obtained by immunofluorescent staining, with respect to the absence of K14 in the b1-null cultures. The expression of K10 and involucrin in the wild-type cultures on the 20th and 23rd days confirmed that the ES cells were capable of developing into terminally differentiating keratinocytes, although neither of the two proteins were detectable by immunostaining. Whereas we could never detect any K10 transcripts in the b1-null embryoid bodies, involucrin amplification sometimes (e.g., Fig. 3A) produced a faint band on the 20th day only.


AID

DB 8341

/

6x13$$$222

09-23-96 15:23:16

dbal

AP: Dev Bio

191


FIG. 6. Identification of keratinocyte cysts in ES cell tumors. (A, C, E) Wild-type; (B, D, F) b1-null. Cysts were double labeled with antibodies to K14 (C, D) and involucrin (E, F). (A, B) phase-contrast images of cysts in C, E and D, F, respectively. Bar in E, 165 mm (A, C, E) and 350 mm (B, D, F).

Identification of Integrin Subunits in Embryoid Body Cells Since the lack of b1 should affect the surface expression of the a integrin subunits that form heterodimers with b1 , we analyzed the expression of the a2 , a3 , and a6 subunits. We also examined the surface expression of the b4 subunit which

forms a heterodimer with a6 in keratinocytes and other epithelial cells. Monoclonal antibodies to extracellular domain epitopes were used, except in the case of a3 , where a rabbit antibody to the cytoplasmic domain was applied. Confocal microscope analysis of immunofluorescent staining of embryoid body cultures showed b1 labeling of the membrane of the wild-type but not of b1-null embryoid


AID

DB 8341

/

6x13$$$222

09-23-96 15:23:16

dbal

AP: Dev Bio

192

Bagutti et al.

FIG. 7. Immunofluorescence staining of keratinocyte cysts in ES cell tumors, using antibodies to b1 (A, B), a3 (C, D), a6 (E, F), and b4 (G, H) integrin subunits. (A, C, E, G) Wild-type; (B, D, F, H) b1-null. All sections were double labeled with antibody to K14 (data not shown). Bar in G, 165 mm (A, B, E, F) and 82 mm (C, D, G, H).

bodies (Figs. 4A and 4B), confirming the absence of the b1 subunit in the b1-null cells (Fa¨ssler et al., 1995). In addition, the b1-knockout embryoid bodies expressed virtually no surface a2 , a3 , a6 , or b4 , whereas all these integrin subunits were detectable at high levels on the surface of the wildtype cells (Figs. 4C–4H and data not shown). The weak cytoplasmic fluorescence observed in b1-null cells labeled with the anti-a3 antiserum (Fig. 4D cf. 4J) probably reflects

a3 integrin that has been synthesized but not exported to the plasma membrane (see Hotchin et al., 1995). The immunofluorescent staining observations were supported by flow cytometry (Fig. 5). Disaggregated embryoid body cells were fluorescently labeled using the same integrin antibodies as for immunofluorescence and subjected to flow cytometry on Days 25, 30, and 39 of culture. Both the b1 and a6 integrin subunits were highly expressed on the surface of


AID

DB 8341

/

6x13$$$222

09-23-96 15:23:16

dbal

AP: Dev Bio

193


contained a peripheral ring of cells which stained positively for K14, corresponding to basal keratinocytes (Figs. 6C and 6D). The interior living layers of the cysts expressed K1 and involucrin and corresponded to suprabasal layers of keratinocytes undergoing terminal differentiation (Figs. 6E and 6F and data not shown). In contrast to our findings in the differentiation studies in vitro, no significant difference was detectable between the differentiation capacity of wild-type and b1-null ES cells in tumors. Staining of the b1-null tumors with an anti-b1 antibody was negative (Fig. 7B cf. 7A). Staining for both a2 (data not shown) and a3 subunits was positive, but diffuse, which might represent accumulation in the cytoplasm rather than on the cell surface (Fig. 7D cf. 7C). a6 and b4 integrin subunits were both concentrated at the basement membrane zone in wild-type- and b1-null-derived teratomas (Figs. 7E– 7H). Double-label immunofluorescence for the integrins and K14 demonstrated coexpression, confirming that integrin expression was confined to the basal layer of keratinocytes (data not shown; see Watt and Hertle, 1994).

Identification of b1-Null Keratinocytes in the Epidermis of a Wild-Type/b1-Null Chimeric Mouse

FIG. 8. Detection of b1-null and wild-type (b1-positive) cells in the skin of a chimeric wild-type/b1-null mouse. (A) X-Gal staining shows a patch of b1-null keratinocytes. (B, C) Immunofluorescence labeling of b1 integrins in a black (i.e., wild-type; B) and an agouti (i.e., b1-null; C) coat-colored region. Bar in A, 62 mm.

wild-type cells and were absent from the surface of b1-null cells (Figs. 5A and 5B, and data not shown). There was also a marked reduction in b4 expression in b1-null cells, although the flow cytometry profile was bimodal and indicated that some cells expressed the same level of b4 as the wild-type (Fig. 5C).

To study the ability of b1-null ES cells to differentiate into keratinocytes during embryonic development, we analyzed the skin of a wild-type/b1-null chimeric mouse. Epidermis from regions of the skin with agouti-colored hairs, indicating a high degree of chimerism, was stained with X-Gal to identify b-galactosidase-expressing b1-null keratinocytes. Figure 8A shows that there were patches of b1-null keratinocytes in the epidermis. The histology and expression of K14 and K1 were identical to those of nonchimeric epidermis from skin with black-colored hairs (data not shown). Confirming the results obtained with X-Gal staining, positive immunofluorescence staining of b1 integrins was demonstrated in epidermis of skin with black-colored hairs (Fig. 8B), whereas patches of b1-null keratinocytes were associated with agouti coat color (Fig. 8C). The sections containing b1-null keratinocytes in the epidermis were also stained with antibodies to a3 , a6 , and b4 ; a3 staining was positive but diffuse and a6b4 showed normal polarization at the basement membrane (data not shown).

Assembly of the Extracellular Matrix Differentiation and Integrin Expression of ES Cells in Experimental Tumors in Mice Subcutaneous transplantation of ES cells into syngeneic mice results in teratomas containing many different cell types including keratinocytes (e.g., Lee et al., 1995). We therefore compared the ability of wild-type and b1-null ES cells to differentiate into keratinocytes as tumors in vivo. Although the b1lacking tumors grew less well (R.F., unpublished), they formed various complex structures similar to the wild-type cells. Most markedly, cryosections of the tumors revealed several cysts resembling stratified squamous epithelia in both wildtype and b1-null tumors (Figs. 6A and 6B). These structures

To assess whether the b1-null cells might be impaired in assembling an ECM, we analyzed the embryoid body cultures, the ES cell tumors, and the epidermis of the wild-type/b1-null chimeric mouse for the presence of ECM components. As shown by immunofluorescent staining, wild-type cells from embryoid body outgrowths (22 days of culture) were capable of forming a dense network of fibronectin fibers, whereas b1-null cells produced very few fibronectin fibers, which were shorter and thicker than those in the wild-type (Figs. 9A and 9B). Similar results were obtained for laminin 1 (Figs. 9C and 9D) and type IV collagen (data not shown). In contrast, in the teratomas, there was no difference in the dis-


AID

DB 8341

/

6x13$$$222

09-23-96 15:23:16

dbal

AP: Dev Bio

194

Bagutti et al.

FIG. 9. Analysis of ECM assembly in 22-day-old embryoid body cultures (A–D), teratomas (E and F), and wild-type/b1-null chimeric mouse skin (G and H). Wild-type (A, C, E) and b1-null (B, D, F) cells or tumor sections were immunofluorescently stained with antibodies to fibronectin (A, B), laminin 1 (C, D), and type IV collagen (E, F). The chimeric skin was stained for type IV collagen (G) and laminin 1 (H). Bar in E, 165 mm (A–D), 82 mm (E, F), and 41 mm (G, H).

tribution of collagen (Figs. 9E and 9F), fibronectin and laminin (data not shown) between the wild-type- and b1-null-derived tumors (in agreement with recent data of R.F., unpublished). We also found collagen and laminin to be expressed normally at the basement membrane zone in the skin of the wild-type/ b1-null chimeric mouse (Figs. 9G and 9H).

DISCUSSION We have shown that wild-type ES cells are capable of differentiating in vitro into epithelial cells expressing mark-

ers of simple (K8, K18, K19) and stratified (K14, K19) epithelia and furthermore that markers of the keratinocyte terminal differentiation pathway (K10, involucrin) can be detected by RT–PCR. Interestingly, the temporal sequence of events, K8, K18, and K19 appearing first and K14, K10, and involucrin later in the differentiating cultures, fits with the known timing of appearance of these keratins in developing mouse embryos (Oshima et al., 1983; Blystone et al., 1994). The differentiating b1-null cells expressed the simple epithelial keratins, but not K14 or K10, and involucrin was detected only occasionally and transiently. We therefore


AID

DB 8341

/

6x13$$$223

09-23-96 15:23:16

dbal

AP: Dev Bio

195


conclude that the capacity of the b1-null cells to differentiate into keratinocytes in vitro is severely impaired. The finding that the b1-null cells were lacking surface expression of the a2 and a3 integrins was expected since these subunits form heterodimers exclusively with the b1 subunit. The surface expression of b4 , which only partners a6 , was markedly reduced and a6 , which forms heterodimers with b1 and b4 , was undetectable. The mechanism of down-regulation of b4 has still to be elucidated: it might be due to the requirement of b1 for the synthesis of b4 in vitro, although this seems unlikely since b3 and b5 expression is normal in b1-null F9 embryonal carcinoma cells (Stephens et al., 1993). Alternatively, the low level of b4 may reflect impaired differentiation of keratinocytes and other cell types that normally express a6b4 (Sonnenberg et al., 1991). In addition to changes in integrin expression, the assembly of the ECM was grossly reduced in the b1-null cultures, consistent with the observation of Wennerberg et al. (1996). Subcutaneous implantation of wild-type and b1-null ES cells into mice showed that in vivo the b1-null cells were able to differentiate into keratinocytes as previously reported for wild-type cells (Lee et al., 1995). The cysts formed by the tumors not only resembled stratified squamous epithelia by their histological appearance but were identified as such by the presence of K14 in the outer ring of basal cells and K1 and involucrin in the centres of the cysts. The basal cells were stained positively for the a2 , a3 , a6 , and b4 integrins. The diffuse distribution of a2 and a3 is probably due to the requirement of b1 for transport to the cell surface (e.g., Hotchin et al., 1995). The expression of a6 and b4 was normal in respect of localization at the basal surface of the basal cells (Sonnenberg et al., 1991). The staining for collagen, laminin, and fibronectin was indistinguishable in the wild-type and b1-null teratomas. b1-Null cells were also able to differentiate into keratinocytes in chimeric mouse skin. As in the teratomas, staining for ECM proteins was normal, a6b4 was polarized to the basement membrane zone and the b1-null keratinocytes underwent normal terminal differentiation. High levels of b1 integrins distinguish stem cells from keratinocytes of lower proliferative potential in human epidermis (Jones and Watt, 1993; Jones et al., 1995) and it will therefore be interesting to examine the proliferative potential of keratinocytes that differentiate from b1-null ES cells. In summary, b1-null cells in vitro have an impaired capacity to differentiate into keratinocytes, to assemble an ECM, and to express a6b4 . In vivo, in teratomas and in chimeric skin, b1-null cells can differentiate into keratinocytes and, at least as judged by immunofluorescence, they can assemble a basement membrane and express a6b4 in a polarized manner. We therefore speculate that in the presence of a basement membrane in vivo, the a6b4 integrin might be able to fulfil the role of b1 integrins in regulating terminal differentiation. Work in other experimental models has established a direct requirement for functional integrins for ECM assembly (Wu et al., 1995; Wennerberg et al., 1996; Fa¨ssler et al., submitted for publication); it is known that both keratinocytes and fibroblasts contribute to the forma-

tion of the epidermal basement membrane (Marinkovich et al., 1993) and that a6b4 polarization in keratinocytes is dependent on ECM (De Luca et al., 1990). We suggest that the basement membrane assembly associated with b1-null keratinocytes in vivo may thus be due to the contribution of wild-type cells from surrounding stromal tissue. If this hypothesis is correct, then failure of differentiation in vitro is due to absence of ECM rather than absence of b1 per se.

ACKNOWLEDGMENTS C.B. was funded by the Swiss National Science Foundation, the Huggenberger-Bischoff Foundation for Cancer Research, and the Ciba-Geigy-Jubilee-Foundation. A.M.W. was supported by DFG (SFB366). We thank everyone who provided us with reagents and J. M. Carroll for advice on the design of the PCR primers.

REFERENCES Bishop, L. A., Rahman, D., Pappin, D. J. C., and Watt, F. M. (1995). Identification of an 80kD protein associated with the a3b1 integrin as a proteolytic fragment of the a3 subunit: Studies with human keratinocytes. Cell Adhesion Commun. 3, 243–255. Blystone, S. D., Graham, I. L., Lindberg, F. P., and Brown, E. J. (1994). Integrin avb3 differentially regulates adhesive and phagocytic functions of the fibronectin receptor a5b1 . J. Cell Biol. 127, 1129–1137. Carroll, J. M., Albers, K. M., Garlick, J. A., Harrington, R., and Taichman, L. B. (1993). Tissue- and stratum-specific expression of the human involucrin promoter in transgenic mice. Proc. Natl. Acad. Sci. USA 90, 10270–10274. De Luca, M., Tamura, R. N., Kajiji, S., Bondanza, S., Rossino, P., Cancedda, R., Marchisio, P. C., and Quaranta, V. (1990). Polarized integrin mediates human keratinocytes adhesion to basal lamina. Proc. Natl. Acad. Sci. USA 87, 6888–6892. Djian, P., Phillips, M., Easley, K., Huang, E., Simon, M., Rice, R. H., and Green, H. (1993). The involucrin genes of the mouse and the rat: Study of their shared repeats. Mol. Biol. Evol. 10, 1136–1149. Doetschmann, T. C., Eistetter, H. R., Katz, M., Schmidt, W., and Kemler, R. (1985). The in vitro development of blastocyst-derived embryonic stem cell lines: Formation of visceral yolk sac, blood islands and myocardium. J. Embyol. Exp. Morphol. 87, 27–45. Fa¨ssler, R., and Meyer, M. (1995). Consequences of lack of b1 integrin gene expression in mice. Genes Dev. 9, 1896–1908. Fa¨ssler, R., Pfaff, M., Murphy, J., Noegel, A. A., Johansson, S., Timpl, R., and Albrecht, R. (1995). Lack of b1 integrin gene in embryonic stem cells affects morphology, adhesion and migration but not integration into the inner cell mass of blastocysts. J. Cell Biol. 128, 979–988. Fuchs, E., and Green, H. (1980). Changes in keratin gene expression during terminal differentiation of the keratinocyte. Cell 19, 1033–1042. Gandarillas, A., and Watt, F. M. (1995). The 5* noncoding region of the mouse involucrin gene: Comparison with the human gene and genes encoding other cornified envelope precursors. Mamm. Genome 6, 680–682. Hotchin, N. A., Gandarillas, A., and Watt, F. M. (1995). Regulation of cell surface b1 integrin levels during keratinocyte terminal differentiation. J. Cell Biol. 128, 1209–1219.


AID

DB 8341

/

6x13$$$223

09-23-96 15:23:16

dbal

AP: Dev Bio

196

Bagutti et al.

Jones, P. H., Harper, S., and Watt, F. M. (1995). Stem cell patterning and fate in human epidermis. Cell 80, 83–93. Jones, P. H., and Watt, F. M. (1993). Separation of human epidermal stem cells from transit amplifying cells on the basis of differences in integrin function and expression. Cell 73, 713–724. Kennel, S. J., Foote, L. J., and Flynn, K. M. (1986). Tumor antigen on benign adenomas and on murine lung carcinomas quantitated by a two-site monoclonal antibody assay. Cancer Res. 46, 707– 712. Krieg, T. M., Schafer, M. P., Cheng, C. K., Filpula, D., Flaherty, P., Steinert, P. M., and Roop, D. R. (1985). Organisation of a type I keratin gene. J. Biol. Chem. 260, 5867–5870. Lane, E. B. (1982). Monoclonal antibodies provide specific intramolecular markers for the study of epithelial tonofilament organization. J. Cell Biol. 92, 665–673. Lane, E. B., and Alexander, C. M. (1990). Use of keratin antibodies in tumour diagnosis. Semin. Cancer Biol. 1, 165–179. Lee, S. L., Tourtellotte, L. C., Wesselschmidt, R. L., and Milbrandt, J. (1995). Growth and differentiation proceeds normally in cells deficient in the immediate early gene NGFI-A. J. Biol. Chem. 270, 9971–9977. Mackenzie, I. C. (1994). Epithelial-mesenchymal interactions in the development and maintenance of epithelial tissues. In ‘‘The Keratinocyte Handbook’’ (I. M. Leigh, E. B. Lane, and F. M. Watt, Eds.), pp. 243–257. Cambridge Univ. Press, Cambridge. Marinkovich, M. P., Keene, D. R., Rimberg, C. S., and Burgeson, R. E. (1993). Cellular-origin of the dermal-epidermal basement membrane. Dev. Dyn. 197, 255–267. Maltsev, V. A., Wobus, A. M., Rohwedel, J., Bader, M., and Hescheler, J. (1994). Cardiomyocytes differentiated in vitro from embryonic stem cells developmentally express cardiac-specific genes and ionic currents. Circ. Res. 75, 233–244. Mendrick, D. L., and Kelly, D. M. (1993). Temporal expression of VLA-2 and modulation of its ligand specificity by rat glomerular epithelial cell in vitro. Lab. Invest. 69, 690–702. Oshima, R. G., Howe, W. E., Klier, F. G., Adamson, E. D., and Shevinsky, L. H. (1983). Intermediate filament protein synthesis in preimplantation murine embryos. Dev. Biol. 99, 447–455. Purkis, P. E., Steel, J. B., Mackenzie, I. C., Nathrath, W. B. J., Leigh, I. M., and Lane, E. B. (1990). Antibody markers of basal cells in complex epithelia. J. Cell Sci. 97, 39–50. Rohwedel, J., Maltsev, V., Bober, E., Arnold, H.-H., Hescheler, J., and Wobus, A. M. (1994). Muscle cell differentiation of embryonic stem cells reflects myogenesis in vivo: Developmentally regulated expression of myogenic determination genes and functional expression of ionic channels. Dev. Biol. 164, 87–101. Rudnicki, M. A., and McBurney, M. W. (1978). Cell culture methods and induction of differentiation of embryonal carcinoma cell lines. In ‘‘Teratocarcinomas and Embryonic Stem Cells: A Practical Approach’’ (E. J. Robertson, Ed.), pp. 19–49. IRL Press, Oxford. Sheardown, S. A. (1992). A simple method for affinity purification and PCR amplification of poly(A)/ mRNA. Trends Genet. 8, 121. Simon, M. (1994). The epidermal cornified envelope and its precursors. In ‘‘The Keratinocyte Handbook’’ (I. M. Leigh, E. B. Lane, and F. M. Watt, Eds.), pp. 275–292. Cambridge Univ. Press, Cambridge. Singer, P. A., Trevor, K., and Oshima, R. G. (1986). Molecular cloning and characterisation of the endo B cytokeratin expressed in preimplantation mouse embryos. J. Biol. Chem. 261, 538–547. Sonnenberg, A., Calafat, J., Janssen, H., Daams, H., van der Raaij-

Helmer, L. M. H., Falcioni, R., Kennel, S. J., Aplin, J. D., Baker, J., Loizidou, M., and Garrod, D. (1991). Integrin a6/b4 complex is located in hemidesmosomes, suggesting a major role in epidermal cell-basement membrane adhesion. J. Cell Biol. 113, 907– 917. Sonnenberg, A., Daams, H., van der Valk, M. A., Hilkens, J., and Hilgers, J. (1986). Development of mouse mammary gland: Identification of stages in differentiation of luminal and myoepithelial cells using monoclonal antibodies and polyvalent antiserum against keratin. J. Histochem. Cytochem. 34, 1037–1046. Stasiak, P. C., and Lane, E. B. (1987). Sequence of cDNA coding for human keratin 19. Nucleic Acids Res. 15, 10058. Steinert, P. M., Rice, R. H., Roop, D. R., Trus, B. L., and Steven, A. C. (1983). Complete amino acid sequence of a mouse epidermal keratin subunit and implications for the structure of intermediate filaments. Nature 302, 794–800. Stephens, L. E., Sonne, J. E., Fitzgerald, M. L., and Damsky, C. H. (1993). Targeted deletion of b1 integrins in F9 embryonal carcinoma cells affects morphological differentiation but not tissuespecific gene expression. J. Cell Biol. 123, 1607–1620. Stephens, L. E., Sutherland, A. E., Klimanskaya, I. V., Andrieux, A., Meneses, J., Pedersen, R. A., and Damsky, C. H. (1995). Deletion of b1 integrins in mice results in inner cell mass failure and peri-implantation lethality. Genes Dev. 9, 1883–1895. Stru¨bing, C., Ahnert-Hilger, G., Shan, J., Wiedenmann, B., Hescheler, J., and Wobus, A. M. (1995). Differentiation of pluripotent embryonic stem cells into the neuronal lineage in vitro gives rise to mature inhibitory and excitatory neurons. Mech. Dev. 53, 1– 13. Tokunaga, K., Taniguchi, H., Yoda, K., Shimizu, M., and Sakiyama, S. (1986). Nucleotide sequence of a full-length cDNA for mouse cytoskeletal beta-actin mRNA. Nucleic Acids Res. 14, 2829. Watt, F. M., and Hertle, M. D. (1994). Keratinocyte integrins. In ‘‘The Keratinocyte Handbook’’ (I. M. Leigh, E. B. Lane, and F. M. Watt, Eds.), pp. 153–164. Cambridge Univ. Press, Cambridge. Wennerberg, K., Lohikangas, L., Gullberg, D., Pfaff, M., Johansson, S., and Fa¨ssler, R. (1996). b1 integrin-dependent and -independent polymerization of fibronectin. J. Cell Biol. 132, 227–238. Wiles, M. V., and Keller, G. (1991). Multiple hematopoietic lineages develop from embryonic stem (ES) cells in culture. Development 111, 259–267. Williams, R. L., Hilton, D. J., Pease, S., Willson, T. A., Stewart, C. I., Gearing, D. P., Wagner, E. F., Metcalf, D., Nicola, N. A., and Gough, N. M. (1988). Myeloid leukaemia inhibiting factor maintains the developmental potential of embryonic stem cells. Nature 336, 684–687. Wobus, A. M., Holzhausen, H., Ja¨kel, P., and Scho¨neich, J. (1984). Characterisation of a pluripotent stem cell line derived from a mouse embryo. Exp. Cell Res. 152, 212–219. Wobus, A. M., Wallukat, G., and Hescheler, J. (1991). Pluripotent mouse embryonic stem cells are able to differentiate into cardiomyocytes expressing chronotropic responses to adrenergic and cholinergic agents and Ca2/ channel blockers. Differentiation 48, 173–182. Wu, C., Keivens, V. M., O’Toole, T. E., McDonald, J. A., and Ginsberg, M. H. (1995). Integrin activation and cytoskeletal interaction are essential for the assembly of a fibronectin matrix. Cell 83, 715–724. Received for publication April 22, 1996 Accepted July 29, 1996


AID

DB 8341

/

6x13$$$223

09-23-96 15:23:16

dbal

AP: Dev Bio