A comparison of normal epithelial cells with their transformed counterparts could lead to the definition of parameters related to growth and differentiation which are altered by viral transformation and which may be relevant to malignant changes in vivo. Using the SV40-transformed human keratinocyte line, SVK14, which exhibits characteristics of simple, nonkeratinizing epithelia, we have shown that IGF I stimulation of these cells results in extensive multilayering, increased cell size, accumulation of involucrin, modulation of keratin 18 and expression of keratins 14 and 10, whilst T-antigen expression is maintained in the multilayered cells. Since T-antigen expression is correlated directly with impairment of stratification and differentiation, it is interesting that treatment of SVK14 with a single growth factor, IGF I, results in molecular events characteristic of differentiating normal keratinocytes.

Normal human keratinocytes in culture stratify and form cornified envelopes reminiscent of the defined program of differentiation of natural epidermis. The onset of the terminal differentiation program is characterized by increased cell size, an increase in the total amount of keratin species and synthesis of involucrin, as the cells pass through the different layers. Involucrin, a soluble protein precursor of the cross-linked envelope, is a marker of an early stage in the pathway of terminal differentiation. It is believed to be synthesized as keratinocytes leave the basal layer and begin to enlarge, but some time before the envelope crosslinking, which occurs only in the outermost cell layers (Watt, 1983).

The differentiation process in cells is frequently impaired by malignant transformation. As the common lethal cancers in man develop from epithelial cells there is considerable interest in the biology of such cells and how they are altered by malignant transformation. Transformation with DNA tumour virus nuclear products, such as the large T-antigen of SV40, confer on cells in culture the ability to grow in the presence of low concentrations of serum, unlimited potential for growth and a block in the differentiation program of a variety of cells that originally show a capacity to differentiate (Chang, 1986). Transformation of human epidermal keratinocytes by the oncogenic virus SV40 leads to progressive inhibition of the normal differentiation process in vitro, accompanied by disruption of the stable pattern of keratin synthesis such that abundance of mRNA for keratins 5 and 14 (markers for tissues of stratified epithelial origin in vivo and in culture) and expression of keratins 6 and 16 (found in hyperproliferating keratinocytes in vivo and normal keratinocytes in culture) are reduced, while abundance of mRNA for keratins 8, 18 and 19 (characteristic of simple nonstratifying epithelia in vivo and in culture) is significantly increased (Morris et al. 1985; Hronis et al. 1984). The transformed keratinocyte line, SVK14 for example, has been reported to be almost completely unable to differentiate (Bernard et al. 1985), marked by lack of stratification, negligible expression of involucrin (Michel et al. 1988), inability to form cornified envelopes, and expression of keratins 7, 8 and 18, characteristic of simple non-differentiating epithelia (Bernard et al. 1985; Taylor-Papadimitriou et al. 1982). They thus more closely resemble the fetal periderm (Lane et al. 1985) suggesting a block at an early step in the program of terminal differentiation.

In investigating the conditions for optimal growth of SVK14 in growth-factor-defined medium, we have found that insulin-like growth factor, IGF I, causes increased cell size, multilayering, modulation of keratin 18, production of keratins 14 and 10 and of involucrin in SVK14, whilst still maintaining T-antigen expression.

Materials

Growth factors

Bovine pancreatic insulin was supplied by Sigma. Insulin-like growth factors I and II (IGF I & IGF II) protein preparations were gifts from Dr D. Morrell (Inst. Child Health). Recombinant IGF I was supplied by Amersham.

Primary antibodies

Mouse monoclonal antibody against keratins 10, LH3, (Leigh, personal communication) was a gift from Dr I. Leigh (The London Hospital). Mouse monoclonal antibodies to keratin 14, L1001, (Leigh et al. 1989) and keratin 18, LE61, (Lane, 1982) were gifts from Dr B. Lane (ICRF, Clare Hall). The above antibodies were used undiluted for immunofluorescence and diluted 1:5 in PBS containing 1% BSA and 0·1 % azide for immunoblotting. Mouse monoclonal antibody against involucrin, 3 Al (Rupniak et al. 1986), a gift from Dr T. Rupniak (Glaxo, Greenford), was used undiluted for immunofluorescence and diluted 1:3 in PBS containing 5% FCS, 0·5 % Tween 20 and 0·1 % azide for immunoblotting. A mouse monoclonal antibody specific for the SV40 T-antigen, PAb416 (Harlow et al. 1981), a gift from Dr L. Crawford (ICRF, Lincolns Inn Fields), was used undiluted for immunofluorescence. 14 and 18 respectively, gifts from Dr B. Lane (ICRF, Clare Hall), were used undiluted for immunofluorescence and diluted 1:5 in PBS containing 1 % BSA and 0T % azide for immunoblotting. Mouse monoclonal antibody against involucrin, 3A1, a gift from Dr T. Rupniak (Glaxo, Greenford), was used undiluted for immunofluorescence and diluted 1:3 in PBS containing 5 % FCS, 0·5 % Tween 20 and 0 ·1% azide. Mouse monoclonal antibody specific for the SV40 T-antigen, 416, a gift from Dr L. Crawford (ICRF, Lincolns Inn Field), was used undiluted for immunofluorescence.

Secondary antibodies

Fluorescein-conjugated goat anti-mouse immunoglobulin (Cappel Laboratories) was used for immunofluorescence at a dilution of 1:50 in PBS containing 1 % BSA and 0T % azide. Horseradish-peroxidase-conjugated rabbit anti-mouse immunoglobulin, (DAKOPATT), was used for immunoblotting at a dilution of 1:50 in PBS containing 1 % BSA.

Cell culture

Culture medium

SV40-transformed human epidermal keratinocytes, SVK14, were a gift from Dr J. Taylor-Papadimitriou, ICRF, Lincolns Inn Fields. Stock cultures of SVK.14 were grown in RPMI 1640 supplemented with 10% fetal calf serum (FCS), 50ngmP cholera toxin and 5μg ml-1 hydrocortisone. ‘Growth factor defined medium’ was composed of RPMI 1640 supplemented with 1 % dithiothreitol-treated new bom calf serum, (dtt-NCS), (Van Zoelen et al. 1985), 5;<g ml-1 transferrin, 5μg ml-1 hydrocortisone and 50 ng ml-1 cholera toxin. All cells were grown at 37 °C in a humidified 12% CO2 atmosphere.

Growth factor stimulation studies

Except where otherwise stated, exponentially growing SVK14 cells were seeded at 3×105 cells in 2 ml of medium in 3 cm dishes and left for 48 h. The cells were then fed with test medium, which was changed every day, for 4 days, when the cell system was assayed.

For autoradiography, SVK14, grown in test conditions for 4 days as above, were subjected to a 60 min pulse of [3H]thymidine (5μCiml-1) after which time the cells were washed with PBS, fixed and embedded in resin for sectioning as below. Autoradiography on sections was performed as described by Brooks (1975) with the following modifications; Ilford K5 emulsion, Ilford ID19 developer, diluted 1:1 with water and Kodak Unifix were used. Exposure time was approximately 7 days.

Involucrin expression in presence of2·5mm-hydroxyurea

Exponentially growing cells were seeded at 1×105 in 2 ml of growth-factor-defined medium in a 3 cm dish and left for 24 h after which time the dishes were changed over to test medium supplemented with 2·5 mm-hydroxyurea.

Immunofluorescence

Staining for involucrin and keratins 10, 14 and 18

SVK14 cells grown under test conditions were fixed in methanol:acetone (1:1) for 10min, air dried for 30min and rehydrated with PBS for 10 min. Antibody binding to fixed cells was performed at 37°C in a humidified chamber, using an incubation period of 60 min for both primary and secondary antibodies.

Staining for SV40 T-antigen

SVK14 cells grown under test conditions were washed in PBS, fixed in 4% formaldehyde for 20 min and then given three washes in PBS. Free aldehyde groups were quenched with 0 ·lM-glycine for 10min. The cells were then permeabilized using 0 ·5 % (w/v) Triton X-100 in PBS for 5 min, and washed extensively with PBS.

The preparations were mounted in glycerol containing 1,4-diazabicyclo[2,2,2]octane (DABCO) as an antifade agent (Johnson et al. 1982) and visualized on a fluorescence microscope.

FACS analysis

SVK14 grown in growth-factor-defined medium and growth-factor-defiried medium with insulin (5μg ml-1) for 4 days with daily medium changes, were stained in suspension (in a volume of 50 μl), with a monoclonal antibody against involucrin, followed by a fluorescein-labelled secondary antibody and subjected to analysis by the fluorescence-activated cell sorter, where fluorescence intensity of 5000 cells was measured and plotted against cell number. In order to minimize cell loss (caused by the methanol: acetone fixative reacting with the coating of the tubes used, rendering them ‘sticky’), the cells were fixed in 0 · 5 –1 ml of fixative and transferred to a new tube for further processing. After each step, the cells were spun down at 1300g.

Flow cytometry

For analysis of DNA content, cells were trypsinized, washed in PBS and fixed with 50% methanol. 5xl(r cells were then resuspended in 0·5 ml of PBS containing ethidium bromide (5 mg per 100 ml) and 0 · 1% Triton X-100, and subjected to analysis in the Cambridge MRC custom-built dual laser flow cytometer (Watson, 1980, 1981). The Innova 70–5 argon ion laser was tuned to the 488 nm line at light power of 200 mW.

Immunoblots

Involucrin

Whole cell extracts equivalent to 1×106 cells were subjected to electrophoresis on sodium dodecyl sulphate (SDS)-polyacrylamide gels (7%) (Laemmli, 1970), electrophoretically transferred to nitrocellulose (Towbin et al. 1979) and probed with a monoclonal antibody against involucrin. After electrophoretic transfer, the nitrocellulose filter was blocked with PBS containing 0·5% Tween 20, and incubated with a monoclonal antibody against involucrin. The filter was then washed in PBS containing 0·5 % Tween 20 for 30 min at room temperature and incubated with horseradish-peroxidase-conjugated rabbit anti-mouse immunoglobulin. The filter was washed with PBS containing 0·5% Tween 20 for 30min at room temperature and peroxidase activity visualized by using 4-chlor-l-naphthol (Andrews, 1986). The blocking and antibody incubation steps were performed by gentle shaking at room temperature, overnight.

Keratins

Keratins were extracted (Rosenberg et al. 1988) from multilayered regions of SVK14 produced as a result of insulin treatment, after dissection of these regions out of culture. 100 fig of keratin extract was separated by discontinuous SDS-PAGE (8·5 %), electrophoretically transferred to nitrocellulose filter and probed with monoclonal antibodies to keratin 14 and keratin 10 as described above.

Cornified envelope formation assay

The cells, once trypsinized, were washed and resuspended in 0·5 ml serum-free RPMI to which 50 μ d of 10% (w/v) SDS containing 200mm-dithiothreitol was added. The suspension was heated to 95°C for 10min and, after cooling, 10l of a DNase I solution (I mg ml-1) was added to prevent aggregation of the envelopes. Envelopes were counted with a haemocytometer, as intact structures (Rice & Green, 1979).

Histological examination of SV KI 4

SVK14 cells grown under test conditions were washed in PBS and incubated with Dispase I (1 unit ml-1) for 1 h at 37°C. The detached intact sheet of cells was then fixed with 4% formaldehyde, embedded in resin (LR White), sectioned at a thickness of 5 pm and stained with toluidine blue or haematoxylin for histological examination.

Estimation of cell number and cell size

Cell number and cell size were estimated using a Coulter counter and channelyser, after trypsinization with 0· 06% trypsin (Difco 1:250) and 0·015% EDTA.

Growth factor requirement of SVK14 cells

Treatment with dithiothreitol (dtt) destroys the majority of peptide growth factors in serum (since most contain essential disulphide bonds) yet other growth promoting components remain intact. By adding back purified growth factors to such dtt-treated serum, Van Zoelen et al. (1985) were able to obtain sustained proliferation of NRK cells, whereas the growth factors alone were inadequate. We have adopted a similar procedure in order to study the growth factor dependence of SVK14 cells. Using RPMI 1640 medium supplemented with 1 % dtt-serum, together with transferrin, hydrocortisone and cholera toxin (‘growth-factor-defined’ medium), we found that SVK14 cells could be grown from low density almost as well as in medium containing fetal calf serum (FCS). The initial plating efficiency tended to be slightly lower than in 10 % FCS, but thereafter, the growth rate was comparable (Fig. 1A). Moreover, the final saturation density was similar (Fig. 1B), though the average cell size tended to be less (Fig. 1C).

Fig. 1.

Growth curves and size histograms of SVK14 from low cell densities to show growth rate (A), from high cell densities to show the effect of insulin on cell saturation density (B), and cell size after 4 days growth in test conditions (C). (A) Exponentially growing SVK14 were seeded in 24-well dishes at 3×104 cells ml-1 per well in growth-factor-defined medium (▵), growth-factor-defined medium supplemented with 5μ -1 linsulin (▴) and 10% FCS (•). The medium was changed every 4 days. (B) Exponentially growing SVK14 were seeded in 3 cm dishes containing 2 ml of medium with 10% FCS, at 3×105. After 2 days the medium was changed to test conditions with daily medium changes. (C) Cell size after 4 days growth in test conditions; 1, growth-factor-defined medium; 2, 10% FCS; 3, insulin in growth-factor-defined medium. The arrowheads indicate the diameters of standard latex beads.

Fig. 1.

Growth curves and size histograms of SVK14 from low cell densities to show growth rate (A), from high cell densities to show the effect of insulin on cell saturation density (B), and cell size after 4 days growth in test conditions (C). (A) Exponentially growing SVK14 were seeded in 24-well dishes at 3×104 cells ml-1 per well in growth-factor-defined medium (▵), growth-factor-defined medium supplemented with 5μ -1 linsulin (▴) and 10% FCS (•). The medium was changed every 4 days. (B) Exponentially growing SVK14 were seeded in 3 cm dishes containing 2 ml of medium with 10% FCS, at 3×105. After 2 days the medium was changed to test conditions with daily medium changes. (C) Cell size after 4 days growth in test conditions; 1, growth-factor-defined medium; 2, 10% FCS; 3, insulin in growth-factor-defined medium. The arrowheads indicate the diameters of standard latex beads.

Morphological changes induced by insulin

During the course of these studies, we observed that the addition of insulin to the growth-factor-defined medium (5 μg ml-1) brought about dramatic changes in cell behaviour. Apart from general improvements in plating efficiency at low densities (Fig. 1A), the final saturation density was increased about threefold over that obtained with 10% FCS (Fig. 1B). In addition, the average cell volume at saturation density was increased by roughly 50% (Fig. 1C). Much more striking, however, was the change in culture morphology. When subjected to daily medium changes (as in Fig. 1B), the insulin-treated cultures developed a pattern of ridges and nodules, interspread with regions of monolayer (Fig. 2C). These ridges and nodules were many cells thick, as shown by sections cut perpendicular to the cell layer (Fig. 2D). In some cases, the nodules had a mushroom-like appearance (Fig. 3A). In contrast, control cultures in which confluent cells were subjected to daily changes of medium (growth-factor-defined or FCS-supplemented), remained essentially as monolayers (Figs 2A,B, 3B, 4E). Similar patterns developed when insulin was added to serum-containing medium (changed daily), and also, though less extensively, when fresh insulin was added daily, without medium changes (data not shown). The absence of multilayering under routine culture conditions (10% FCS) is not therefore due to the presence of an inhibitor in serum. Rather, the morphological changes are the direct consequence of insulin treatment.

Fig. 2.

Morphology of SVK14 cultures under different growth conditions. (A,C,E) Phase contrast; (B,D,F) histological sections. (A,B) Control cultures grown in medium supplemented with 10 % FCS. (C,D) Cells grown in defined medium supplemented with insulin (5μg ml-1). (E,F) Cells grown in defined medium supplemented with IGF I (100ng ml-1). Bar for A-D, 300gm and 200gm for E,F.

Fig. 2.

Morphology of SVK14 cultures under different growth conditions. (A,C,E) Phase contrast; (B,D,F) histological sections. (A,B) Control cultures grown in medium supplemented with 10 % FCS. (C,D) Cells grown in defined medium supplemented with insulin (5μg ml-1). (E,F) Cells grown in defined medium supplemented with IGF I (100ng ml-1). Bar for A-D, 300gm and 200gm for E,F.

Fig. 3.

Autoradiographs of SVK14 cells labelled with [3H]thymidine, without (B) and with (A) insulin treatment. The arrowheads indicate labelled cells in the monolayer. Though less obvious in the photograph, the presence of silver grains over these cells was confirmed by direct observation under the microscope, at high magnification. Bar, 30pm for A, and 50 μm for B.

Fig. 3.

Autoradiographs of SVK14 cells labelled with [3H]thymidine, without (B) and with (A) insulin treatment. The arrowheads indicate labelled cells in the monolayer. Though less obvious in the photograph, the presence of silver grains over these cells was confirmed by direct observation under the microscope, at high magnification. Bar, 30pm for A, and 50 μm for B.

A titration of the concentration of insulin required to produce multilayering showed that insulin used at concentrations ranging from physiological (0·4 ng ml -1) up to 1μg ml-l did not induce multilayering. Insulin in the range of 1 – 3μ g ml-1 caused morphological changes indicative of multilayering; however, use of insulin at 3 – 5μg ml-1 led to the multilayering of the type shown in Fig. 2C,D. Since the levels of insulin required to promote multilayering in SVK14 are approximately 12500 times higher than the physiological levels of insulin, it was apparent that the induction of multilayering is caused either by a contaminating activity in bovine insulin preparations or by insulin cross-reacting and stimulating receptors other than the insulin receptor. As all effects of small concentrations of IGF on cultured cells can be mimicked by unphysiologically high concentrations of insulin (Froesch et al. 1985) and since the IGF I receptor binds insulin with weak affinity (Rechler & Nissley, 1985), it is probable that insulin-promoted multilayering is due to insulin cross-reacting with the IGF I receptor. Also, as IGF I and IGF II have been shown to be important in cellular growth and differentiation (Froesch et al. 1985; Beebe et al. 1987), we have chosen to assess the ability of these growth factors to promote differentiation in SVK14. Use of IGF II at 100 ng ml-1 did not result in multilayering. However, use of purified or recombinant IGF I over a range of 50–200 ng ml-1 leads to the morphological pattern observed with insulin at 5μm-1 (Fig. 2E,F), as well as to similar increases in saturation density and cell volume (not shown). A closer examination of the cells grown with IGF I shows that IGF I at 50 ng ml-1 caused visible multilayering but not as extensive as 100 ng ml-1. At 100–200 ng ml-1 no increase in degree of multilayering was observed.

Origin of multilayered ridges and nodules induced by insulin/IGF 1

Since SVK14 cells were derived by transformation of normal keratinocytes with SV40, it might be argued that the ridges and nodules induced by insulin/l GF I treatment were little more than the ‘piling up’ typical of many transformed lines. However, the pattern of ridges and nodules, interspersed with regions of monolayer, persists however long the cultures are maintained. If the patterns were due solely to ‘piling up’ then one would expect the monolayer regions eventually to give way to uniform multilayering. This does not happen. The focal pattern of multilayering suggests a lack of homogeneity in the culture. However, similar morphological patterns were observed in several independent clones derived (by limiting dilution) from early and late passages of the original population. It is therefore most unlikely that the ‘foci’ are due to the presence of variant cells within the culture, unless these arise at extremely high frequency, and yet fail to have a selective advantage.

To determine whether the multilayered regions arose as a result of the stimulation of cell division, we used flow cytometry to examine the distribution of cells through the cell cycle, at a point equivalent to day 4 in Fig. 1B. This showed about 21 % of the cells to be in S phase both in the presence and absence of insulin/lGF I. The fractions of the cells in Gt (58 %) and G2 (21 %) were also identical under both conditions. There is thus no indication that the multilayering induced by insulin is caused by an increased rate of cell cycle initiation.

Autoradiography following pulse labelling of similar cultures with [3H]thymidine, on sections cut perpendicular to the cell layer, confirmed the high proportion of cells in S phase (Fig. 3). This is especially significant in the case of the cultures not treated with insulin (Fig. 3B) which had reached saturation density. From this, together with the flow cytometry data, it can be concluded that these cultures are not quiescent, but have reached an equilibrium between cell death and cell proliferation. Indeed, dead cells are seen floating in the medium once confluency is achieved. In the light of this, it is plausible that the increased cell density following insulin treatment is a consequence of improved cell survival, though this in itself does not account for the pattern of ridges and nodules.

In autoradiographs of the insulin-treated cultures (Fig. 3A), the most heavily labelled cells were located predominantly on the outer surface of the nodules. Labelled cells may also be found within the adjacent monolayer regions, but with a very much lower grain density (arrowheads in Fig. 3A). Such inequalities in labelling intensity were less obvious in control cultures not treated with insulin. It is possible that the differences in labelling intensity reflect differences in the uptake of [3H]thymidine unrelated to the rate of DNA replication. For instance, the cells in the outer surface of the nodules appear flatter than those within the crowded monolayer regions, and may thus have had better access to the medium. The labelling pattern is also consistent with a locally higher rate of DNA replication (and hence shorter S phase and overall cell cycle time) at the surface of the nodules.

Induction of differentiation by insulin/IGF I

The multilayering of the SVK14 cells induced by insulin/lGF I treatment is unlike the stratification of normal keratinocytes in that it is nonuniform and proliferating cells are not restricted to the basal layer. Despite this, we chose to investigate whether the multilayering was associated with any changes in the differentiation state of the cells.

The extent of involucrin expression in SVK14 when grown in either FCS-supplemented medium, or growthfactor-defined medium with or without insulin/lGF I was investigated by immunostaining. Cells grown in growth-factor-defined medium failed to express any detectable involucrin (Fig. 4A,B). Cells grown in the FCS-supplemented medium were seen to stain very weakly (Fig. 4D,F), while cells grown in presence of IGF I or insulin stained intensely in the multilayered regions (Fig. 4C,D). Involucrin expression by SVK14 as a result of insulin/IGF I treatment was further confirmed by FACS analysis (Fig. 5) and immunoblotting (Fig. 6, panel 1). These results clearly show SVK14 are capable of involucrin expression when stimulated by insulin/IGF I. The low levels of expression observed in the presence of FCS (Fig. 4F) may be due to small amounts of insulin-like growth factors present in the batch of FCS used.

Fig. 4.

Comparative expression of involucrin in SVK14 cells as judged by immunostaining using a monoclonal antibody against involucrin. (A,C,E) Phase contrast; (B,D,F) fluorescence. (A,B) Cells grown in growth-factor-defined medium; (C,D) cells grown in growthfactor-defined medium supplemented with insulin (5μg ml-1); (E,F) cells grown in medium supplemented with FCS. Bar, 100 μm.

Fig. 4.

Comparative expression of involucrin in SVK14 cells as judged by immunostaining using a monoclonal antibody against involucrin. (A,C,E) Phase contrast; (B,D,F) fluorescence. (A,B) Cells grown in growth-factor-defined medium; (C,D) cells grown in growthfactor-defined medium supplemented with insulin (5μg ml-1); (E,F) cells grown in medium supplemented with FCS. Bar, 100 μm.

Fig. 5.

FACS analysis of SVK14 cells after staining with a monoclonal antibody to involucrin, presenting fluorescence intensity of the cells with and without insulin treatment.

Fig. 5.

FACS analysis of SVK14 cells after staining with a monoclonal antibody to involucrin, presenting fluorescence intensity of the cells with and without insulin treatment.

Fig. 6.

Immunoblots of SVK14 cell extracts probed with a monoclonal antibody to involucrin (panel 1) and monoclonal antibodies to keratin 14 and keratin 10 (panel 2). Panel 1. Whole cell extract from cells grown in growth factor defined medium, without (B) and with (A) insulin treatment, probed with a monoclonal antibody to involucrin. Similar results were obtained after loading the gels with extracts containing equivalent protein concentration rather than equivalent cell number. Panel 2. Keratin extract from the cells in the multilayer regions only, probed with monoclonal antibodies to keratin 14 (A) and keratin 10 (B).

Fig. 6.

Immunoblots of SVK14 cell extracts probed with a monoclonal antibody to involucrin (panel 1) and monoclonal antibodies to keratin 14 and keratin 10 (panel 2). Panel 1. Whole cell extract from cells grown in growth factor defined medium, without (B) and with (A) insulin treatment, probed with a monoclonal antibody to involucrin. Similar results were obtained after loading the gels with extracts containing equivalent protein concentration rather than equivalent cell number. Panel 2. Keratin extract from the cells in the multilayer regions only, probed with monoclonal antibodies to keratin 14 (A) and keratin 10 (B).

Differentiation in keratinocytes has been shown to have a cell-density-related component (Chang, 1986). In order to distinguish the effect of increased cell density from direct IGF I stimulation on expression of involucrin in SVK14, sparsely seeded cells were grown in the presence of hydroxyurea, so that cell division was prevented without affecting cellular metabolic activity. Cells grown in defined medium failed to express detectable levels of involucrin, whilst cells treated with IGF I were clearly seen to accumulate involucrin (Fig. 7). As these cultures were maintained at low cell density, we have concluded that IGF I has a direct effect on involucrin expression.

Fig. 7.

Immunofluorescent staining of SVK14 cells grown in the presence of 2·5mm-hydroxyurea, using a monoclonal antibody against involucrin. (A,C) Phase contrast; (B,D) fluorescence. (A,B) Cells grown in growth-factor-defined medium; (C,D) cells grown in growth-factor-defined medium supplemented with insulin (5/μg ml-1). Bar, 100urn.

Fig. 7.

Immunofluorescent staining of SVK14 cells grown in the presence of 2·5mm-hydroxyurea, using a monoclonal antibody against involucrin. (A,C) Phase contrast; (B,D) fluorescence. (A,B) Cells grown in growth-factor-defined medium; (C,D) cells grown in growth-factor-defined medium supplemented with insulin (5/μg ml-1). Bar, 100urn.

Immunostaining has shown that all SVK14 cells express SV40 T-antigen. Cells in multilayered regions produced after insulin/lGF I treatment, also maintain T-antigen expression (Fig. 8) while expressing involucrin.

Fig. 8.

Immunostaining of SVK14 cells stratified as a result of insulin treatment, with a monoclonal antibody against SV40 T-antigen. (A) phase contrast; (B) fluorescence. Bar, 250 μ m.

Fig. 8.

Immunostaining of SVK14 cells stratified as a result of insulin treatment, with a monoclonal antibody against SV40 T-antigen. (A) phase contrast; (B) fluorescence. Bar, 250 μ m.

IGF-I-treated SVK14 fall short of terminal differentiation

The extent of terminal differentiation in normal keratinocytes is assessed by cornified envelope formation (Rice & Green, 1979) which is preceded by a characteristic change in the total amount of keratin species expressed. Keratin 18 is a marker of simple epithelia, which is absent in stratifying epithelia (Moll et al. 1982). Immunostaining showed that SVK14 cells grown in growth-factor-defined medium express keratin 18 (Fig. 9A,B), while those grown in the presence of IGF I lacked keratin 18 in the centre of multilayered regions (Fig. 9C,D). Further, following insulin/lGF I treatment, SVK14 cells showed positive immunostaining with antibodies to keratin 14 (Fig. 10A,B), characteristic of basal cells of stratifying epithelia, and keratin 10 (Fig. 10C,D), characteristic of suprabasal cells of the epidermis (Moll et al. 1982). By immunoblotting, the antibody to keratin 14 revealed a single band of relative molecular mass 50 × 103 compatible with keratin 14 (Fig. 6, panel 2A). The antibody to keratin 10 stained a number of bands, the strongest having a relative molecular mass of 56 ·6 ×103 compatible with keratin 10 (Fig. 6, panel 2B). There seems little doubt, therefore, that insulin/lGF I treatment induces the appearance of both keratin 14 and keratin 10, though two-dimensional gels would be required to confirm this unambiguously. It is interesting that lack of expression of keratin 18 and expression of keratins 14 and 10 is confined to the centre of the multilayered regions, where the cells appear not to be dividing (Fig. 3A). Insulin/lGF I treatment may thus favour a shift to a more normal pattern of keratin expression. Nevertheless, the incidence of cornified envelopes remains low. Without ionophore treatment (Rice & Green, 1979), cornified envelopes are undetectable in the absence of IGF I treatment (0 cornified envelopes observed per 2-8 x104 cells). In the presence of IGF I, the incidence rises to approximately 1% (250 cornified envelopes observed per 2-7×104 cells) but this falls short of the 5-10% reported for growing colonies of normal keratinocytes (Sun & Green, 1976).

Fig. 9.

Immunostaining of SVK14 cells with a monoclonal antibody against keratin 18, (A,C) phase contrast; (B,D) fluorescence. (A,B) SVK14 without insulin treatment; (C,D) SVK14 with insulin treatment. Bar, 250 μ m.

Fig. 9.

Immunostaining of SVK14 cells with a monoclonal antibody against keratin 18, (A,C) phase contrast; (B,D) fluorescence. (A,B) SVK14 without insulin treatment; (C,D) SVK14 with insulin treatment. Bar, 250 μ m.

Fig. 10.

Immunostaining of SVK14 cells without (E,F) and with insulin treatment (A,D). (A,C,E) Phase contrast; (B,D,F) fluorescence. (A,B) Stained with monoclonal antibody to keratin 14. (C,D) Stained with a monoclonal antibody to keratin 10. (E,F) Stained with a monoclonal antibody to keratin 14. (Cells not treated with insulin and stained with an antibody to keratin 10 were identical to E,F.) Bar, 100 μm.

Fig. 10.

Immunostaining of SVK14 cells without (E,F) and with insulin treatment (A,D). (A,C,E) Phase contrast; (B,D,F) fluorescence. (A,B) Stained with monoclonal antibody to keratin 14. (C,D) Stained with a monoclonal antibody to keratin 10. (E,F) Stained with a monoclonal antibody to keratin 14. (Cells not treated with insulin and stained with an antibody to keratin 10 were identical to E,F.) Bar, 100 μm.

Multilayering and differentiation in SVK14

The keratinocyte line SVK14 has been reported to lack almost completely the capacity to differentiate in culture, and instead exhibits characteristics of simple epithelia (Bernard et al. 1985; Taylor-Papadimitriou et al. 1982). We show here, using a growth-factor-defined medium that supports optimal growth of SVK14 cells, that the addition of a single growth factor, IGF I, brings about profound changes in behaviour, in part reminiscent of the differentiation of normal keratinocytes. Following IGF I treatment, the cells develop extensive, complex, patterns of multilayering, accompanied by a reduction in keratin 18 expression, and the induction of keratin 14, keratin 10 and involucrin expression. Changes during differentiation (Woodcock-Mitchell et al. 1982) and development (Dale et al. 1985; Moll et al. 1982) in epithelial cells coincide with alterations in keratin synthesis, resulting in a change in the pattern of keratins expressed as the cells undergo a commitment to terminal differentiation. In their expression of kera-tins 7, 8 and 18, and their lack of stratification, SVK14 cells resemble fetal periderm rather than normal keratinocytes (Bernard et al. 1985; Lane et al. 1985). It is therefore possible that the changes induced by IGF I represent a shift to a more mature phenotype.

The multilayering of SVK14 ceils following insulin/ IGF I treatment is unlike the stratification of normal keratinocytes. Not only is it nonuniform, with regions of monolayer interspersed among patches of multilayering, but proliferating cells are not restricted to the basal layer. Nevertheless, the multilayered regions are not simply random ‘piling up’ of cells, since the multilayering remains nonuniform however long the cultures are maintained. Nor do the ridges and nodules represent transformed foci arising from variant cells within the population, since the same patterns are observed in several independent clones. Further, transformed foci are not normally considered to be inducible, whereas the ridges and nodules appear only as a result of insulin/lGF I treatment. The significance of the pattern of multilayering remains unclear. However, during development, the fetal epidermis invaginates to generate hair follicles and sweat glands under inductive stimuli arising from the dermis. In view of the similarities between SVK14 cells and fetal periderm, we are intrigued by the possibility that the patterns represent some caricature of normal morphogenesis. To investigate this further, we are currently examining the behaviour of SVK14 cells on collagen gels.

Involucrin production in SVK14 appears to be directly stimulated by IGF I treatment and is clearly seen in multilayered regions by immunofluorescence. Expression of involucrin in SVK14 has been previously reported to be negligible (Bernard et al. 1985), undetectable by immunofluorescence and on SDS-polyacrylamide gels (Michel et al. 1987) or present only in trace amounts (Michel et al. 1988). Further, a study by Ponec et al. (1985) shows Ca2+-ionophore-treated SVK14 to be capable of producing 100% cornified envelopes without reporting any stratification. It is likely that these cornified envelopes result from the cross-linking of proteins other than involucrin (Michel et al. 1987, 1988). Ionophore-induced cornified envelope formation is therefore an unreliable marker of differentiation. For this reason, we have considered only the spontaneous incidence of cornified envelopes (i.e. without ionophore treatment). This clearly rises in IGF-I-treated cultures (from 0 to 1%), but remains lower than the 5–10% expected for normal growing keratinocytes (Sun & Green, 1976). Thus, even assuming the spontaneous cornified envelopes to be entirely normal (which we have not yet demonstrated), it is evident that the IGF-I-induced differentiation remains incomplete.

T-antigen expression in SVK14

Established cell lines offer distinct advantages in studies requiring prolonged passaging, as normal cells in culture undergo senescence after a limited number of subcultures. SV40 transformation of normal keratinocytes, has led to the establishment of several keratinocyte cell lines, with varying degrees of transformation-induced changes in their growth and differentiation ability (Chang, 1986; Steinberg & Defendi, 1983), similar to those observed in some squamous cell carcinomas, suggesting that SV40 virus transformation can somehow reproduce the molecular events occurring in naturally arising squamous cell carcinomas (Bernard et al. 1985). The keratinocyte cell line HE-209 has been transformed with a temperature-sensitive T-antigen gene (tsA209) and a study of this cell line shows that T-antigen expression is directly correlated with impairment of stratification and differentiation, which are only exhibited at the nonpermissive temperature (Banks-Schlegel & Howley, 1983). In the same study, the cell line HE-SV, which has been transformed with a wild-type T-antigen gene, shows a limited degree of stratification although it is not clear whether T-antigen expression in this line is maintained during differentiation (Banks-Schlegel & Howley, 1983). In both of these examples, keratins present in nontransformed keratinocytes were observed and in the case of HE-209 their amounts were found to be elevated at restrictive temperatures. SVK14 differs from these examples by exhibiting characteristics of simple epithelia in that they express keratins 18, 7 and 8, and show monolayer growth, even though they have been derived by SV40 transformation of cells from stratifying epithelia (Taylor-Papadimitriou et al. 1982). It is interesting that IGF-I-induced stratification in these cells results in readdressing keratin expression to that of stratifying epithelia, while T-antigen expression is maintained in the stratifying cells. These results do not distinguish between modulation of T-antigen expression as a result of IGF I treatment to levels that would allow cells to enter differentiation or that IGF I can somehow over-ride the de-differentiation state controlled by T-antigen or that IGF-I-induced differentiation is not affected by SV40 transformation and/or T-antigen.

Studies on the role of the IGF family of proteins in development have shown IGF I levels to be low in. neonatal rats and rise as the animals mature (Daugha-day et al. 1982). A similar increase in serum IGF I is seen during the rapid growth phase in young chickens (Huybrechts et al. 1985). These temporal changes in serum IGF I levels may be related to the observation that embryonic chicken serum does not support lens fibre differentiation, although adult chicken serum does (Beebe et al. 1987). Several recent studies have provided evidence for the importance of IGFs in other examples of differentiation. IGF I receptors were shown to be present at very early stages of chicken embryo development, at which time, they are more abundant than insulin receptors (Bassas et al. 1985). IGF I has been shown to stimulate erythroid differentiation from late erythroid progenitor cells (Kurtz et al. 1985) and to promote the differentiation of ovarian granulosa cells (Veldhuis & Demeres, 1985). A molecule having IGF I activity appears to be both necessary and sufficient to cause DNA synthesis and mitosis in postmetamorphic frog lens epithelial cells (Rothstein et al. 1980) and lenopterin, a protein related to the IGF family of proteins has been shown to stimulate lens fibre differentiation.

These observations show the importance of IGFs in selected examples of cell differentiation and, together with our data, highlight the importance of IGF I in the control of cellular growth and differentiation.

Our special thanks to Dr B. Lane and Dr I. Leigh for the very generous gift of anti-keratin 14 antibody, L1001, and anti-keratin 10 antibody, LH3. The authors would like to thank Dr T. Rupniak and Dr L. Crawford for gifts of antibodies, Dr J. Taylor-Papadimitriou for SV-K14 cells, Dr M. Green for advice, Dr D. Capellaro and collegues for FACS analysis, Dr J. Watson and Miss H. Cox for flow cytometry, Dr N. Holder for providing electrophoresis apparatus and Dr C. Colaco for much help and advice on gel work. T.K. is indebted to Dr L. Buluwela for much help, critical discussion and reading of this manuscript. We would also like to thank Cancer Research Campaign for supporting this work.

Andrews
,
A. T.
(
1986
).
In Electrophoresis: Theory, Techniques, and Biochemical and Clinical Applications
. Second edition (ed.
A. R.
Peacocke
and
W. F.
Harrington
).
Oxford
:
Clarendon Press
.
Banks-Schlegel
,
S. P.
&
Howley
,
P. M.
(
1983
).
Differentiation of human epidermal cells transformed by SV 40
.
J. Cell Biol
.
96
,
330
337
.
Bassas
,
L.
,
De Pablo
,
F.
,
Lensniak
,
M.
&
Roth
,
J.
(
1985
).
Ontogeny of receptors for insulin-like peptides in chick embryo tissues: early dominance of insulin-like growth factor over insulin receptors in brain
.
Endocrinology
117
,
2321
2329
.
Beebe
,
D. C.
,
Silver
,
M. H.
,
Belcher
,
K. S.
,
Van Wyk
,
J. J.
,
Svoboda
,
M. E.
&
Zelenka
,
P. S.
(
1987
).
Lenopterin, a protein that controls lens fibre formation is related functionally and immunologically to the insulin like growth factors
.
Proc. natn. Acad. Sci. U.S.A
.
84
,
2327
2330
.
Bernard
,
B. A.
,
Robinson
,
S. M.
,
Semât
,
A.
&
Darmon
,
M.
(
1985
).
Reexpression of fetal characteristics of simian virus 40 transformed human kératinocytes
.
Cancer Research
45
,
1707
1716
.
Brooks
,
R. F.
(
1975
).
The kinetics of serum-induced initiation of DNA synthesis in BHK21/C13 cells, and the influence of exogenous adenosine
.
J. cell. Physiol
.
86
,
369
377
.
Chang
,
S. E.
(
1986
).
In-vitro transformation of human epithelial cells
.
Biochem. Biophys. Acta
823
,
161
194
.
Dale
,
B. A.
,
Holbrook
,
K. A.
,
Kimball
,
J. R.
,
Hoff
,
M.
&
Sun
,
T. T.
(
1985
).
Expression of epidermal keratins and filagrin during human fetal skin development
.
J. Cell. Biochem
.
101
,
1257
1269
.
Daughaday
,
W. H.
,
Parker
,
K. A.
,
Borowsky
,
S.
,
Trivendi
,
B.
&
Kapadia
,
M.
(
1982
).
Measurement of somatomedin-related peptides in fetal, neonatal, and maternal rat serum by insulin like growth (IGF) I radioimmunoassay, IGF-II radioreceptor assay (RRA), and multiplication stimulating activity RRA after acid-ethanol extraction
.
Endocrinology
110
,
575
581
.
Froesch
,
E. R.
,
Schmid
,
C.
,
Schwander
,
J.
&
Zapf
,
J.
(
1985
).
Action of insulin-like growth factors
.
A. Rev. Physiol
.
47
,
443
467
.
Harlow
,
E.
,
Crawford
,
L. V.
,
Pim
,
D. C.
&
Williamson
,
N. M.
(
1981
).
Monoclonal antibodies specific for simian virus 40 tumor antigen
.
J. Virol
.
39
,
861
869
.
Hronis
,
T. S.
,
Steinberg
,
M. L.
,
Defendí
,
V.
&
Sun
,
T. T.
(
1984
).
Simple epithelium nature of some simian virus-40 transformed human epidermal kératinocytes
.
Cancer Research
44
,
5797
5804
.
Huybrechts
,
L. M.
,
King
,
D. B.
,
Lauterio
,
T. J.
&
Scanes
,
L. G.
(
1985
).
Plasma concentrations of somatomedin-C in hypophysectomized, dwarf and intact growing domestic fowl as determined by heterologous radioimmunoassay
.
J. Endocrinol
.
104
,
233
239
.
Johnson
,
G. D.
,
Davidson
,
R. S.
,
McNamee
,
K. C.
,
Russell
,
G.
,
Goodwin
,
D.
&
Holborow
,
E. J.
(
1982
).
Fading of immunofluorescence during microscopy: a study of phenomenon and its remedy
.
J. Immunological Methods
55
,
231
242
.
Kurtz
,
A.
,
Hartl
,
W.
,
Jelkmann
,
W.
,
Zapf
,
J.
&
Bauer
,
C.
(
1985
).
Activity in fetal bovine serum that stimulates erythroid colony formation in fetal mouse livers is insulin like growth factor I
.
J. clin. Invest
.
76
,
1643
1648
.
Laemmli
,
U. K.
(
1970
).
Cleavage of structural proteins during the assembly of the head of bacteriophage T4
.
Nature, Lend
.
227
,
680
685
.
Lane
,
E. B.
(
1982
).
Monoclonal antibodies provide specific intramolecular for the study of epithelial tonofilament organization
.
J. Cell Biol
.
92
,
665
673
.
Lane
,
E. B.
,
Barter
,
J.
,
Purkis
,
P. E.
&
Leigh
,
I. M.
(
1985
).
Keratin antigens in differentiating skin
.
A. N.Y. Acad. Sci
.
455
,
241
258
.
Leigh
,
L M.
,
Purku
,
P. E.
,
Steel
,
J. B.
&
Lane
,
E. B.
(
1989
).
Manuscript in prep
.
Michel
,
S.
,
Schmidt
,
R.
,
Robinson
,
S. M.
,
Shroot
,
B.
&
Reichert
,
U.
(
1987
).
Identification and subcellular distribution of cornified envelope precursor proteins in the transformed human kératinocyte line SV-K14
.
J. invest. Dermatol
.
88
,
301
305
.
Michel
,
S.
,
Schmidt
,
R.
,
Shroot
,
B.
&
Reichert
,
U.
(
1988
).
Morphological and biochemical characterization of the cornified envelopes from human epidermal kératinocytes of different origin
.
J. invest. Dermatol
.
91
,
11
15
.
Moll
,
R.
,
Moll
,
I.
&
Wiest
,
W.
(
1982
).
Changes in the pattern of cytokeratin polypeptides in epidermis and hair follicles during skin development in human foetus
.
Differentiation
23
,
170
178
.
Morris
,
A.
,
Steinberg
,
M. L.
&
Defendí
,
V.
(
1985
).
Keratin gene expression in simian virus 40 transformed human kératinocytes
.
Proc. natn. Acad. Sci. U.S.A
.
82
,
8498
8502
.
Ponec
,
M.
,
Lavrusen
,
S.
,
Kempenaar
,
J.
,
Havekes
,
L.
&
Boonstra
,
J.
(
1985
).
SV40 transformed (SV-K14) and normal kératinocytes: similarities in the expression of low density lipoproteins, epidermal growth factor, glucocorticoid receptor, and the regulation of lipid metabolism
.
J. invest. Dermatol
.
85
,
476
482
.
Rechler
,
M. M.
&
Nissley
,
S. P.
(
1985
).
The nature and regulation of the receptors for insulin like growth factors
.
A. Rev. Physiol
.
47
,
425
442
.
Rice
,
R. H.
&
Green
,
H.
(
1979
).
Presence in human epidermal cells of a soluble protein precursor of the cross-linked envelope: action of cross-linking by Ca2+ ions
.
Cell
18
,
681
694
.
Rosenberg
,
M.
,
Raychaudhury
,
A.
,
Shows
,
T. B.
,
Le Beau
,
M. M.
&
Fuchs
,
E.
(
1988
).
A group of type I keratin genes on human chromosome 17: Characterization and expression
.
Mol. cell. Biol
.
8
,
722
736
.
Rothstein
,
H.
,
Van Wyk
,
J. J.
,
Hayden
,
J. H.
,
Gordon
,
S. R.
&
Weinsieder
,
A.
(
1980
).
Somatomedin C: restoration in vivo of cycle traverse in GQ/G] blocked cells of hypophysectomized animals
.
Science
208
,
410
412
.
Rupniak
,
H. T.
,
Turner
,
D. M.
,
Wood
,
E. J.
&
Cunuffe
,
W. J.
(
1986
).
Characteristics of the cell envelope system in normal and psoriatic epidermis
.
J. invest. Dermatol
.
87
,
164
.
Steinberg
,
M. L.
&
Defendí
,
V.
(
1983
).
Transformation and immortilization of human kératinocytes by SV40
.
J. invest. Dermatol
.
81
,
131
136
.
Sun
,
T. T.
&
Green
,
H.
(
1976
).
Differentiation of epidermal kératinocytes in cell culture: formation of cornified envelopes
.
Cell
9
,
511
521
.
Taylor-Papadimitriou
,
J.
,
Purkis
,
P.
,
Lane
,
B. E.
,
McKay
,
J. A.
&
Chang
,
S. E.
(
1982
).
Effect of SV40 transformation on the cytoskeleton and behavioral properties of human kératinocytes
.
Cell Differentiation
11
,
169
180
.
Towbin
,
H.
,
Staehelin
,
T.
&
Goron
,
J.
(
1979
).
Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications
.
Proc. natn. Acad. Sci. U.S.A
.
76
,
4350
4354
.
Van Zoelen
,
E. J. J.
,
Van Oostwaard
,
T. M. J.
,
Van Der Sagg
,
P. T.
&
De Laat
,
S. W.
(
1985
).
Phenotypic transformation of normal rat kidney cells in a growth factor defined medium; introduction by a neuroblastoma derived transforming growth factor independently of the EGF receptor
.
J. cell. Physiol
.
123
,
151
160
.
Veldhuis
,
J. D.
&
Demeres
,
L. M.
(
1985
).
A role for somatomedin C as a differentiating hormone and amplifier of hormone action on ovarian cells: studies with synthetically pure human somatomedin C and swine granulosa cells
.
Biochem. Biophys. Res. Commun
.
130
,
234
240
.
Watson
,
J. V.
(
1980
).
Enzyme kinetic studies in cell populations using fluorogenic substrates and flow cytometric techniques
.
Cytometry
1
,
143
000
.
Watson
,
J. V.
(
1981
).
Dual laser beam focussing for flow cytometry through a single crossed cylindrical lens pair
.
Cytometry
2
,
14
00
.
Watt
,
F. M.
(
1983
).
Involucrin and other markers of kératinocyte terminal differentiation
.
J. invest. Dermatol
.
81
,
1005
1035
.
Watt
,
F. M.
&
Green
,
H.
(
1981
).
Involucrin synthesis is correlated with cell size in human epidermal cultures
.
J. Cell Biol
.
90
,
738
742
.
Woodcock-Mitchell
,
J.
,
Eichner
,
W. G.
,
Nelson
,
W. G.
&
Sun
,
T. T.
(
1982
).
Immunolocalization of keratin polypeptides in human epidermis using monoclonal antibodies
.
J. Cell Biol
.
95
,
580
588
.