ABSTRACT
Chick embryo neural retinal cells transdifferentiate extensively into lens cells when cultured in Eagle’s MEM containing horse and fetal calf sera (FHMEM). Such cultures express elevated levels of pp60c-src-associated tyrosine kinase activity relative to parallel cultures prevented from transdifferentiating by the addition of supplementary glucose (FHGMEM) or replacement of MEM by medium 199 (F199). Northern blotting and in vitro translation studies suggest that c-src mRNA levels are only slightly higher in late transdifferentiating (FHMEM) cultures as compared to parallel blocked (FHGMEM or F199) cultures. By immunocytochemical staining, we show that pp60c-src protein is largely localized in cell groups undergoing conversion into lens (i.e. expressing δ crystallin) in late FHMEM cultures. Initial studies of pp60c-sre in chick lens tissues during development indicate that higher kinase activity is found in the epithelial cells relative to mature lens fibres. Thus pp60c-src may be expressed both during the differentiation of lens cells in vivo and during the transdifferentiation of neural retina cells into lens in vitro.
Introduction
The appearance of lens-fibre-like cells in long-term monolayer cultures of embryonic chick neural retina (NR) cells is one example of a phenomenon termed transdifferentiation (Okada, Itoh, Watanabe & Eguchi, 1975; Okada, 1980), used here to describe the conversion of partially differentiated cells from one tissue type (NR) into fully differentiated cells of a foreign tissue type (lens). In this case, the initial retinal tissue (from 7- to 9-day embryos) is not yet fully differentiated, but clearly lens cells would never appear during normal neuroretinal development in vivo. The transdifferentiated lens cells appear mainly as groups of lens-fibre-like cells, termed lentoid bodies, after about 30 days of NR culture (Okada et al. 1975). In this system, the extent of NR transdifferentiation is dependent upon a variety of factors including embryonic age (de Pomerai & Clayton, 1978) and the ambient culture conditions. Thus, in NR cultures maintained in Eagle’s MEM supplemented with 5 % horse serum and 5 % fetal calf serum (FHMEM), δ-crystallin mRNA is detectable after about 25 days (Carr & de Pomerai, 1985), and by 40 days up to 20% of the total soluble protein consists of δcrystallin (de Pomerai & Gali, 1982), which is the major structural protein of embryonic chick lens fibres (Piatigorsky, 1984). By contrast, when NR cultures are maintained in FHGMEM medium (i.e. FHMEM supplemented with glucose to 18 mm final) or in medium 199 supplemented with 10 % fetal calf serum (F199), transdifferentiation and δ-crystallin accumulation are largely inhibited even after 40−50 days in vitro (de Pomerai & Gali, 1982; Gali & de Pomerai, 1984; Carr & de Pomerai, 1985).
Notably, the lens crystallin genes (including δ) are expressed at very low levels in the initial NR tissue in early embryos (Clayton, Thomson & de Pomerai, 1979; Agata, Yasuda & Okada, 1983); thus transdifferentiation in this system involves an increase in the expression of these genes from low to high levels during in vitro culture, rather than de novo activation of a set of genes never normally transcribed in retinal cells.
Cellular genes homologous to the oncogenes of transforming retroviruses have been detected in the genomes of all metazoan species so far examined. These proto-oncogenes have been implicated in the processes of cellular proliferation and differentiation (Bishop, 1983; Heldin & Westermark, 1984; Adamson, 1987). The normal cellular counterpart of the src oncogene product of Rous sarcoma virus (RSV) is a 60×103Mr tyrosine-specific phosphokinase (pp60c-src) and is a highly conserved protein found in species ranging from sponge to man (Schartl & Barnekow, 1982). The ability of the viral src gene product (pp60v-src) to induce cell proliferation in RSV-transformed cells (Hanafusa, 1977) at first sight suggests a metabolic role for pp60c-src in normal cells related to cell division or growth. The discovery that pp60c-src is found at high levels in postmitotic nervous tissues (Schartl & Barnekow, 1984; Cotton & Brugge, 1983; Fults, Towle, Lauder & Maness, 1985) suggests a possible additional role in differentiation. Sorge, Levy & Maness (1984) showed that pp60c-src is developmentally regulated in chick neural retina, being expressed in retinal neurones at the onset of differentiation and persisting in postmitotic, fully functional neurones. This evidence suggests that the c-src gene product may be more important in neuronal differentiation than in cell proliferation.
We have examined the activity of the pp60c-src kinase in embryonic chick neuroretinal cells cultured in media permissive (FHMEM) or nonpermissive (FHGMEM, F199) for transdifferentiation, by measuring in vitro phosphorylation of the 53×103Mr IgG heavy chain in immunoprecipitates using RSV-induced tumour-bearing-rabbit (TBR) serum (Collett & Erikson, 1978). The localization of pp60c-scr in transdifferentiating cultures was determined by immunocytochemistry using a rabbit antiserum against pp60v-src (kindly donated by Dr P. F. Maness, University of North Carolina). Double labelling of cultures was accomplished using the rabbit anti-pp60v-csr and a rat anti-chick-δ -crystallin antiserum, which were subsequently detected by PAP staining and rhodamine fluorescence, respectively. This allowed the simultaneous detection of both pp60c-src and δcrystallin (lens marker protein) in situ in transdifferentiating NR cultures.
Our results demonstrate an increase in pp60c-src tyrosine kinase activity in late-stage transdifferentiating (FHMEM) cultures, relative to a decrease in blocked (FHGMEM, F199) cultures. Furthermore, the bulk of the pp60c-scr protein appears to be localized in cell groups undergoing conversion into lentoids in these cultures. However, the levels of c-src transcripts are only slightly higher in late FHMEM as compared to FHGMEM or F199 cultures. A preliminary survey of pp60c-scr kinase activity in lens epithelial cultures and during embryonic lens development in vivo has shown that elevated c-src kinase activities may also characterize the differentiation of normal lens cells. The function of pp60c-src in developing lens cells remains obscure, but might be related to its role in neuronal differentiation.
Materials and methods
Fertile chicken eggs were obtained from G. W. Padley Ltd (Grantham, Lincs, UK); culture medium components and sera were from Gibco-Europe Ltd (Paisley, Scotland). Immunochemicals were from Miles Laboratories, radiochemicals and in vitro translation kits from Amersham International p.l.c.; all other chemicals were from Sigma Ltd (Poole, Dorset) unless otherwise stated.
The tumour-bearing rabbit (TBR) serum used in the protein kinase assay ‘was a generous gift from Dr A. Barnekow, Institut für Virologie, Giessen, FDR (Schartl & Barnekow, 1982), as was the Schmidt-Ruppin D strain of RSV used for cell transformation. The rabbit anti-pp60v-src serum was raised against pp60v-src purified from bacteria expressing a cloned v-src gene, but was able specifically to recognize the pp60s-src protein (Sorge et al. 1984); this antiserum was a generous gift from Dr P. F. Maness, University of North Carolina, USA. The anti-δ -crystallin serum was raised in rats and was monospecific as judged by immunoblotting criteria (de Pomerai, Ellis & Carr, 1984a; de Pomerai, Takagi, Kondoh & Okada, 1984a). For Northern blotting, we used an src-specific probe comprising a 612 bp PstI subfragment of the 800 bp PvuII fragment of cloned RSV-DNA V (De Lorbe et al. 1980) inserted in pBR322; this clone was a generous gift from Dr M. Hayman, ICRF, London.
Cell culture
Neural retina cell culture procedures were as described previously (de Pomerai & Gali, 1982). Cultures were maintained throughout in either medium FHMEM (Eagle’s MEM with Earle’s salts, 5% (v/v) fetal calf serum, 5% (v/v) horse serum, 26mM-NaHCO3, 2mM-L-glutamine, 100i.u.ml-1 penicillin and 100μgml-1 streptomycin), FHGMEM (as FHMEM but with supplementary glucose to 18 mm final concentration) or F199 (containing 10% (v/v) fetal calf serum and medium 199; NaHCO3, L-glutamine, penicillin and streptomycin as above) as described previously (de Pomerai & Gali, 1982; Gali & de Pomerai, 1984; Carr & de Pomerai, 1985). Neuroretinal or fibroblast cells from 7-day-old chick embryos were infected with the Schmidt-Ruppin D strain of RSV in medium FHMEM containing 1 % (v/v) dimethyl sulphoxide (Pessac & Calothy, 1974). Lens epithelial cultures were prepared from 19-day-old embryonic chick lenses as described by de Pomerai, Clayton & Pritchard (1978), and were maintained for up to 35 days in FHMEM medium.
Preparation of cell extracts, immunoprecipitation and protein kinase assays
Clarified cell extracts were prepared and 0−2 mg of total soluble protein per sample was used for immunoprecipitation with TBR serum as described by Schartl & Barnekow (1982), except that protein A-sepharose CL-4B (Pharmacia) was used to adsorb the immunoprecipitates. Determination of the protein concentrations of clarified cell extracts was carried out according to the method of Lowry, Rose-brough, Farr & Randall (1951).
The protein kinase assay was performed according to a modification of the method of Collett & Erikson (1978). Washed immunoprecipitates were suspended in kinase buffer (20 mm-bis-tris propane, 50mM-MgCl2) containing 5μCi [y-32P]ATP (3000 Ci mmol-1) per sample, and incubated at 4°C for 5 min. The labelled immunocomplexes were electrophoresed under reducing conditions on 11% SDS-polyacrylamide gels, and the labelled proteins detected (after Coomassie staining) by autoradiography at −70°C with an intensifying screen. For quantification of kinase activity, the heavy-chain IgG bands (53×10 Mr) were cut out from the gel, solubilized in Soluene (Packard), and their 32P-radioactivity determined by liquid scintillation counting.
To confirm that the phosphorylation site(s) in the IgG 53×103Mr heavy chain was tyrosine, phosphoamino acid analysis was carried out according to the method of Hunter & Sefton (1980).
Immunocytochemistry
To detect and localize pp60c-src in NR cultures in situ, a method based on that of Sternberger (1979) was used. Briefly, cultures were washed twice in ice-cold phosphate-buffered saline (PBS) and fixed in 1 % (v/v) glutaraldehyde, 1% (v/v) formaldehyde, 1% (v/v) DMSO. After two washes in PBS, cultures were incubated in PBS containing 20 % normal goat serum (NGS) and then agitated successively in anti-pp60v-src (1:1000 dilution for 24 h), in goat anti-rabbit IgG (1:100 dilution for 2h) and in horseradish peroxidase-anti-peroxidase complex (PAP) (1:100 dilution overnight). Washes between these incubations were in several changes of PBS; for control experiments, preimmune rabbit serum was used in place of the anti-pp60v-src serum for the initial incubation. PAP staining was developed with 0·05% (v/v) 3’3-diaminobenzidine (DAB) and 0·01 % (v/v) H2O2 in PBS.
For subsequent localization of ó crystallin, the immunofluorescence method described by de Pomerai et al. (1984b) was used. Cultures already stained as above were washed in PBS and incubated successively in rat anti-δ crystallin antiserum (1:1000 dilution overnight) and in TRITC-conjugated anti-rat IgG (1:1000 dilution for 2h); after each incubation, cultures were washed several times in PBS. Rhodamine-fluorescent cultures were photographed under appropriate u.v. illumination (de Pomerai et al. 1984b). Controls omitted the first antibody or used preimmune rat serum.
RNA preparation, Northern blotting and in vitro translation
Total cellular RNA was prepared from fresh NR tissue as described previously (Carr & de Pomerai, 1985) and enriched for poly (A)+ RNA by one cycle of oligo-(dT)-cellulose chromatography (Aviv & Leder, 1972). In the case of cultured NR material the oligo (dT)-cellulose selection step was omitted and total cytoplasmic RNA was used.
RNA samples were denatured and fractionated through 1·5% (v/v) agarose gels as described by Lehrach, Diamond, Wozney & Boedtker (1977), except that the gel buffer contained 1·1 m-formaldehyde. RNA was then transferred to nitrocellulose membranes (Schleicher and Schuell), and hybridized for 20 h with a 32P-labelled nick-translated src-specific probe (specific activity >108ctsmin-1 ftg-1) according to the method of Thomas (1980). After hybridization, filters were washed four times in 0·1 × SSPE, 0·1% SDS for 10 min at 50°C (1 × SSPE = 0·18M-NaCl, 1m?-EDTA,10mm-sodium phosphate, pH7-4). Dried filters were then used to expose Kodak X-Omat-S film with an intensifying screen for 3 days (poly(A)+ RNAs) or 2 weeks (total cytoplasmic RNAs) at −70°C. In vitro translation was performed as described previously (Carr & de Pomerai, 1985) using total cytoplasmic RNA preparations. Immunoprecipitation of translation products with anti-δ crystallin or with anti-pp60v-src antibodies (see above) was also as described by Carr & de Pomerai (1985).
Results
pp60c-src in fresh and cultured NR and lens
The pp60c-src kinase activity present in transdifferentiating (FHMEM) NR cultures and in parallel ‘blocked’ (FHGMEM or F199) cultures Was assessed by in vitro IgG phosphorylation assays (Figs 1, 2); parallel determinations of kinase activity were carried out with fresh NR and lens tissues from developing chick embryos. The smears of radioactivity apparent in all lanes of Fig. 1A may possibly be caused by lipid, as documented previously for chick brain tissue (Barnekow & Bauer, 1984). The kinase activity in fresh NR peaks at around the 12th day of embryonic development (Figs IB, 2B), in agreement with previous reports (Sorge et al. 1984). The additional higher Mr bands apparent in Fig. IB, lane 2, probably represent other proteins phosphorylated in the immune complexes prepared with TBR serum, since similar bands are found on prolonged exposure of all samples showing high kinase activity. Such bands are not observed when anti-pp60v-src is substituted for TBR serum in the kinase assays (Ellis, PhD thesis, University of Nottingham, 1986; data not shown). At 13 days of development, there is also a peak in the level of accumulated c-src RNA transcripts, as determined by Northern blotting analysis of poly(A)+ RNA isolated from fresh NR tissue (Fig. 3B). The predominant retinal poly(A)+c-src mRNA species is of approximate size 4 kb, close to the value of 3-9 kb reported by Gonda, Sheiness & Bishop (1982).
pp60c-src kinase activity in neural retina cultures, showing phosphorylation of the 53× 103M? heavy chain IgG in TBR-serum immunoprecipitates from 0·2 mg soluble protein extracts. Lanes 1-3, 8-day cultures; lanes 4−6, 20-day cultures; lanes 7−9, 40-day cultures. Lanes 1, 4 and 7, F199 cultures; lanes 2, 5 and 8, FHGMEM cultures; lanes 3, 6 and 9, FHMEM cultures. (B) pp60c-scr kinase activity in embryonic neural retina extracts. Lanes 1−3; 7, 12 and 17 days respectively.
pp60c-src kinase activity in neural retina cultures, showing phosphorylation of the 53× 103M? heavy chain IgG in TBR-serum immunoprecipitates from 0·2 mg soluble protein extracts. Lanes 1-3, 8-day cultures; lanes 4−6, 20-day cultures; lanes 7−9, 40-day cultures. Lanes 1, 4 and 7, F199 cultures; lanes 2, 5 and 8, FHGMEM cultures; lanes 3, 6 and 9, FHMEM cultures. (B) pp60c-scr kinase activity in embryonic neural retina extracts. Lanes 1−3; 7, 12 and 17 days respectively.
pp60c-src kinase activity in (A) neural retina and lens epithelium cultures in vitro; (B) embryonic neural retina and lens tissues in vivo. Phosphorylation of the 53×103Mr heavy chain IgG in TBR-serum immunoprecipitates was determined using 0·2 mg soluble protein extracts. (A) NR and LE cultures: •—•, FHMEM; ▪—▪, FHGMEM; ▴— ▴, F199; ▿—▿, LE cultures (FHMEM). (B) Fresh NR and lens tissues: •—•, neural retina; ▪—▪, whole lens; ▵—▵, lens epithelium; □—□, lens cortex; ○—○, lens nucleus. Each point shows the mean and standard error derived from four (part A) or two (part B) batches of samples (each in duplicate). For comparison, the average c-src kinase activity in 40-day cultures of SR-D RSV-transformed NR cells is 300 ± 20disintsmin -1pg−1 soluble protein.
pp60c-src kinase activity in (A) neural retina and lens epithelium cultures in vitro; (B) embryonic neural retina and lens tissues in vivo. Phosphorylation of the 53×103Mr heavy chain IgG in TBR-serum immunoprecipitates was determined using 0·2 mg soluble protein extracts. (A) NR and LE cultures: •—•, FHMEM; ▪—▪, FHGMEM; ▴— ▴, F199; ▿—▿, LE cultures (FHMEM). (B) Fresh NR and lens tissues: •—•, neural retina; ▪—▪, whole lens; ▵—▵, lens epithelium; □—□, lens cortex; ○—○, lens nucleus. Each point shows the mean and standard error derived from four (part A) or two (part B) batches of samples (each in duplicate). For comparison, the average c-src kinase activity in 40-day cultures of SR-D RSV-transformed NR cells is 300 ± 20disintsmin -1pg−1 soluble protein.
Northern blotting analysis of c-src mRNA levels in NR cells in vitro and in vivo. (A) Northern blot of total cytoplasmic RNAs (10μg per lane) isolated from 14day (lanes 1, 2, 3) and 40-day (lanes 4, 5, 6) cultures of NR cells maintained in FHMEM (lanes 1 and 4), F199 (lanes 2 and 5) or FHGMEM (lanes 3 and 6) media. Hybridization with 32P-labelled src-specific probe. (B) Northern blot of poly(A)+ RNA from embryonic neural retina (5 μg per lane) hybridized with the same probe. Lanes 1-5; 6, 9, 13, 16 and 20 days of embryonic development. Approximate size of transcripts (4 kb) estimated by comparison with methylene-blue-stained ribosomal RNA markers on same gel (not shown).
Northern blotting analysis of c-src mRNA levels in NR cells in vitro and in vivo. (A) Northern blot of total cytoplasmic RNAs (10μg per lane) isolated from 14day (lanes 1, 2, 3) and 40-day (lanes 4, 5, 6) cultures of NR cells maintained in FHMEM (lanes 1 and 4), F199 (lanes 2 and 5) or FHGMEM (lanes 3 and 6) media. Hybridization with 32P-labelled src-specific probe. (B) Northern blot of poly(A)+ RNA from embryonic neural retina (5 μg per lane) hybridized with the same probe. Lanes 1-5; 6, 9, 13, 16 and 20 days of embryonic development. Approximate size of transcripts (4 kb) estimated by comparison with methylene-blue-stained ribosomal RNA markers on same gel (not shown).
The level of pp60c-src kinase activity in early cultures of dissociated NR cells is approximately twofold lower than in the tissue from which they are derived (Figs 1A, IB and 2A). In the case of NR cultures grown in medium types that inhibit transdifferentiation (FHGMEM or F199), this level of activity is maintained until around 25 days, but by day 38 falls to low levels. By contrast, in transdifferentiating NR cultures (FHMEM; permissive for lens fibre production; Carr & de Pomerai, 1985), kinase activity increases sharply after 25 days, and by day 38 this activity is seven times higher than in parallel ‘blocked’ cultures (where transdifferentiation is inhibited by FHGMEM or F199 media; Figs 1A, 2A). Preliminary kinase assays on very late cultures in FHMEM (postlentoidogenesis, i.e. after 50 to 60 days in vitro) show decreased levels of kinase activity relative to a peak at around 40 days (results not shown). The contrasting levels of kinase activity apparent after 38 days of NR culture (Figs 1A, 2A) could be explained in several ways. These include: (i) increased levels of c-src mRNA and/or pp60c-src protein in transdifferentiating (FHMEM) as compared to blocked (FHGMEM, F199) cultures; or (ii) post-translational activation of pp60c-src kinase activity in FHMEM cultures, without concomitant c-src gene activation, as documented in other systems (e.g. during granulocytic and monocytic differentiation of HL60 cells; Bamekow & Gessler, 1986).
Northern blotting analysis (Fig. 3A) shows that c-src transcripts are present at similar levels in the total cytoplasmic RNA populations from 14 day NR cultures under all three medium conditions; but even after 40 days in vitro, c-src transcripts are only slightly more prevalent in FHMEM as compared to FHGMEM or F199 cultures (Fig. 3A). This pattern is in broad agreement with the corresponding kinase activities (Figs 1A, 2A), but the differences are much less marked. A similar conclusion emerges from in vitro translation studies using total RNA extracted from 40-day cultures maintained in the three medium types (Fig. 4A). A faint pp60c-src band can be immunoprecipitated from the FHMEM translation products (comigrating with the pp60v-src band precipitated from the translation products of RSV-infected chick embryo fibroblast RNA; see Fig. 4A, lanes 1 and 4), whereas the corresponding pp60c-scr bands from F199 or FHGMEM translation products are barely detectable (Fig. 4A, lanes 2 and 3). For comparison, a parallel study using our anti-δ -crystallin antibody immunoprecipitates a strong δ -crystallin band from among the translation products of late FHMEM culture RNA, but not from those using late FHGMEM or F199 culture RNAs (Fig. 4B).
In vitro translation analysis of total cytoplasmic RNAs from cultured NR cells. Cytoplasmic RNA was extracted from 40-day NR cultures maintained in FHMEM (lanes 1 and 5), FHGMEM (lanes 2 and 6) or F199 (lanes 3 and 7) media, and also from RSV-infected chick embryo fibroblasts (lane 4). In vitro translation products were immunoprecipitated with either anti-pp60v-src antibody (part A, lanes 1-4) or anti-δ-crystallin antibody (part B, lanes 5-7), and the precipitates subjected to SDS-PAGE and autoradiography. Arrowheads point to the 60×103;Mr pp60 products (pp60c-scr in lanes 1−3, pp60v-src in lane 4) in part A, and to the 50×103 Mrδ crystallin in part B. Size estimation was by comparison with total chick lens proteins run on the same gel and stained with Coomassie blue (not shown).
In vitro translation analysis of total cytoplasmic RNAs from cultured NR cells. Cytoplasmic RNA was extracted from 40-day NR cultures maintained in FHMEM (lanes 1 and 5), FHGMEM (lanes 2 and 6) or F199 (lanes 3 and 7) media, and also from RSV-infected chick embryo fibroblasts (lane 4). In vitro translation products were immunoprecipitated with either anti-pp60v-src antibody (part A, lanes 1-4) or anti-δ-crystallin antibody (part B, lanes 5-7), and the precipitates subjected to SDS-PAGE and autoradiography. Arrowheads point to the 60×103;Mr pp60 products (pp60c-scr in lanes 1−3, pp60v-src in lane 4) in part A, and to the 50×103 Mrδ crystallin in part B. Size estimation was by comparison with total chick lens proteins run on the same gel and stained with Coomassie blue (not shown).
In the developing chick lens, the pp60c-src tyrosine kinase activity of whole lens extracts increases generally between the 4th and 18th days of embryogenesis and falls slightly prior to hatching (21 days; Fig. 2B). In fractions prepared from the lenses of both newly hatched and adult chickens, kinase activity is highest in the epithelial cells and much lower in the innermost nuclear lens fibres (Fig. 2B). The pp60c-src kinase activity is also much lower in adult lenses than in the embryonic lens (Fig. 2B). Lens epithelial cultures prepared from 19-day embryonic lenses were found to have very similar levels of kinase activity at the two stages examined, both at 10 days (shortly before lentoids first appear) and at 35 days (cultures confluent, with many lentoids present), as shown in Fig. 2A.
Two-dimensional phosphoamino acid analysis was performed on the 32P-labelled IgG heavy chain bands (53X103) immunoprecipitated from cell extracts by the TBR serum for all samples (fresh tissues and cultures) where radioactive incorporation was sufficiently high (data not shown); this analysis confirmed that only tyrosine residues were phosphorylated in the immunoprecipitates.
Localization of pp60c-src in transdifferentiating NR cultures
In order to ascertain whether the observed increase in c-src kinase activity in transdifferentiating (FHMEM) NR cultures was confined to a certain type of cell or distributed throughout the cell population, immunocytochemical localization of the pp60c-src protein was carried out using an anti-pp60v-src serum in conjunction with PAP staining. In 15-day FHMEM cultures, staining is essentially confined to the overlying neuronal cells (Fig. 5A; Sorge et al. 1984). Controls using preimmune rabbit serum show no staining, even of lentoids in late FHMEM cultures (Fig. 5C). The majority of the detectable pp60c-src in late-stage FHMEM cultures is found to be localized in the lensfibre aggregates known as lentoids (Fig. 5B,E,G), with a minority subpopulation of single cells and groups of cells in monolayer regions also staining positively for pp60c-src (Fig. 5B). In the remainder of the monolayer cell sheet in FHMEM cultures, and also in all regions of cultures of similar age maintained in media FHGMEM and F199 (not shown), pp60c-src is barely detectable above background levels by this method. Interestingly, the pattern of PAP staining for pp60c-src within the lens-fibre aggregates (lentoids) in late FHMEM cultures is not uniform (Fig. 5E,G); certain regions stain very strongly while other regions stain weakly or not at all. This would not be predicted if PAP staining of lentoid bodies were merely artefactual due for example to their increased thickness (see also control in Fig. 5C).
Immunocytochemical localization of pp60c-src and δ crystallin in 15-day (A) and 40-day (B-H) NR cultures maintained in FHMEM. Bar 100μm. pp60c-src was detected by using anti-pp60v-src antiserum and PAP staining; δ crystallin by using a rat antiserum against chick δ crystallin followed by rhodamine-conjugated antirat IgG. A-C,E and G show PAP staining by transmitted light; D,F and H show rhodamine fluorescence under appropriate u.v. illumination. Bar 100μm (A) 15-day FHMEM culture showing pp60c-src in neurone-like cells. (B) 40-day FHMEM culture showing pp60c-src in lentoids and in some monolayer cells. (C) 40-day FHMEM culture; pp60 control with initial anti-μ-pp60v-src antiserum replaced by preimmune rabbit serum, showing unstained lentoid. (D) 40-day FHMEM culture; δ control with initial anti-δ-crystallin antiserum omitted, showing very faint rhodamine fluorescence in a lentoid, exposure 3min. (E-H) Colocalization of pp60c-src and δ crystallin within lentoid regions; (E,G) 40-day FHMEM cultures stained for pp60c-src (F,H) the same fields stained for δ crystallin and viewed under u.v., exposure 30s.
Immunocytochemical localization of pp60c-src and δ crystallin in 15-day (A) and 40-day (B-H) NR cultures maintained in FHMEM. Bar 100μm. pp60c-src was detected by using anti-pp60v-src antiserum and PAP staining; δ crystallin by using a rat antiserum against chick δ crystallin followed by rhodamine-conjugated antirat IgG. A-C,E and G show PAP staining by transmitted light; D,F and H show rhodamine fluorescence under appropriate u.v. illumination. Bar 100μm (A) 15-day FHMEM culture showing pp60c-src in neurone-like cells. (B) 40-day FHMEM culture showing pp60c-src in lentoids and in some monolayer cells. (C) 40-day FHMEM culture; pp60 control with initial anti-μ-pp60v-src antiserum replaced by preimmune rabbit serum, showing unstained lentoid. (D) 40-day FHMEM culture; δ control with initial anti-δ-crystallin antiserum omitted, showing very faint rhodamine fluorescence in a lentoid, exposure 3min. (E-H) Colocalization of pp60c-src and δ crystallin within lentoid regions; (E,G) 40-day FHMEM cultures stained for pp60c-src (F,H) the same fields stained for δ crystallin and viewed under u.v., exposure 30s.
The colocalization of pp60c-src and δ crystallin (characteristic lens protein) in the lentoid bodies of transdifferentiating FHMEM cultures was demonstrated by immunostaining of pp60c-src (Fig. 5E,G) followed by immunofluorescent detection of δ crystallin (Fig. 5F,H). Only faint staining of lentoid bodies is observed in controls where the initial anti-δ-crystallin antibody was omitted (Fig. 5D) or where preimmune rat serum was used (not shown). Those monolayer regions of late FHMEM cultures showing positive PAP staining for pp60c-src sometimes show faintly positive fluorescent staining for δ crystallin. In general, most regions staining positively for the c-src protein also stain positively for δ crystallin, although some lentoid regions showing strong immunofluorescence for δ crystallin stain weakly or not at all for the c-src protein.
Note. These photographs were derived from two separate experiments, in which the background and intensity of staining were somewhat variable. The c-src control (C) belonged to the same series as A,B and G, while E was from the other series in which all cultures (including controls) acquired a brownish background during PAP staining.
Discussion
We have shown that in late-stage (approximately 40 day) cultures of embryonic neuroretinal cells maintained in a medium (FHMEM) permissive for transdifferentiation, the level of pp60c-src-associated tyrosine kinase activity is about seven times higher than in parallel cultures maintained in media that inhibit the production of lentoids (FHGMEM or F199). The fact that the overall growth rates for NR cells in FHMEM, FHGMEM and F199 media are similar (de Pomerai & Gali, 1982; Gali & de Pomerai, 1984) argues against the possibility that the rise in kinase activity in late FHMEM cultures is merely a result of differential cell proliferation, particularly since the increase occurs at a stage when cell numbers plateau in all three media.
Cellular pp60c-src is expressed in the developing neurones of the neural retina at the onset of differentiation, persisting throughout development into adulthood when the neurones are postmitotic and fully functional (Sorge et al. 1984). Retinal pp60c-src activity peaks at around day 12 in vivo, and this is paralleled by an accumulation of poly(A)+c-src RNA transcripts (Fig. 3B), implying that the levels of pp60c-src are regulated at least in part at the level of transcription.
In vitro, the pp60c-src protein is localized mainly in neurone-like cells (Sorge et al. 1984) in early NR cultures under all medium conditions tested, but later on it is found at high levels in lentoid bodies under FHMEM conditions only (see Fig. 5). Control experiments with preimmune rabbit serum (Fig. 5C), together with the nonuniform staining patterns observed in lentoids (Fig. 5B,E,G), strongly suggest that the observed distribution of pp60c-src protein in late FHMEM cultures is not artefactual. This is supported by the high pp60c-src kinase activities observed in late FHMEM (but not FHGMEM or F199) cultures and also in lens epithelial cells and cortical lens fibres from newly hatched chicks in vivo (see Fig. 2). On this basis, we suggest that the high kinase activity of late FHMEM cultures (Figs 1A, 2A) is due to pp60c-src protein localized primarily in the developing lentoid bodies (Fig. 5E,G) which also express δ crystallin (Figs4B, 5F,H). If this is the case, one might have expected a greater disparity between the c-src transcript levels detected in total RNA from late FHMEM as compared to FHGMEM or F199 cultures (Figs 3A, 4A). In fact, total cytoplasmic RNA from late FHMEM cultures contains only slightly (approximately twofold) higher levels of c-src transcripts as compared to RNA from parallel FHGMEM or F199 cultures, whether analysed by Northern blotting with a c-src probe (Fig. 3A) or by immunoprecipitation of in vitro translation products (Fig. 4A). Several independent batches of cultures have yielded similar results when comparing total RNA populations (Ellis & Carr, unpublished data). Furthermore, poly(A)+-selected RNAs from these three types of culture show no detectable differences in c-src transcript levels, perhaps implying that there is some contribution of poly(A)-c-src mRNAs under transdifferentiation-permissive (FHMEM) conditions (Ellis & Carr, unpublished data; not shown). This point is under further investigation.
The apparent anomaly here may be explained in at least two ways, (i) Translational and/or post-translational controls may amplify a relatively small change in steady-state c-src mRNA levels, (ii) Because neurones are lost more rapidly from FHMEM cultures than from FHGMEM (de Pomerai & Gali, 1982) or F199 (Gali & de Pomerai, 1984) cultures, it is possible that low-level c-src expression by large numbers of surviving neuronal cells may occur in these ‘blocked’ cultures during the later stages in vitro when most such cells have already disappeared from FHMEM cultures and lentoids are beginning to appear.
Our present data do not exclude either of these explanations. However, if the increased kinase activity of late FHMEM cultures were due merely to post-translational activation (as occurs during HL60 myelocytic cell differentiation; Barnekow & Gessler, 1986), then the observed concentration of immunoreactive pp60c-src protein in developing lentoids (Fig. 5) would not be expected. Our results suggest rather that increased levels of pp60c-src protein become expressed in developing lens cells during transdifferentiation.
The changes in cell population that occur during NR culture under the three medium conditions studied here are complex and difficult to quantify. In FHMEM cultures, aggregates of neurone-like cells (expressing pp60c-src; Fig. 5A) overlie a monolayer of flattened glial-like cells during the first 2 to 3 weeks in vitro (Okada et al. 1975). These cultures express a variety of neuronal markers (de Pomerai et al. 1983), amongst which cell-surface toxin receptors persist for longest. However, all such markers decline to low levels by 20 to 25 days in vitro (Crisanti-Combes, Pessac & Calothy, 1978), and few morphologically recognizable neurones survive beyond 30 days, the stage at which lentoid bodies begin to develop (Okada et al. 1975; Okada, 1980). In FHGMEM cultures, neurones survive for considerably longer, and choline acetyltransferase (CAT) activity does not decline to low levels until about 30 days in vitro (de Pomerai & Gali, 1982). By contrast, CAT levels remain low throughout in F199 cultures, apparently because the rounded neurone-like cells remain dispersed on top of the cell sheet and do not form aggregates as in FHMEM or FHGMEM cultures (Gali & de Pomerài, 1984). Nevertheless, such rounded cells remain abundant until at least 35 days in F199 cultures.
These observations imply that low-level contributions of c-src transcripts from numerous neuronelike cells in late FHGMEM or F199 cultures could affect the results shown in Figs 3A and 4A, where c-src RNA levels are only slightly higher in transdifferentiating (FHMEM) as compared to blocked (FHGMEM, F199) cultures. However, the observed pp60c-src protein distribution in late FHMEM cultures (primarily in lentoids; Fig. 5) would imply that most c-src transcripts in these cultures should be confined to developing lentoids. Preliminary in situ hybridization experiments (using a double-stranded c-src probe within the pBR322 plasmid, allowing possible concatemer formation) seem to confirm this pattern, with dense silver grains over parts of many lentoids, whereas the plasmid alone does not hybridize to late FHMEM cultures above background levels (results not shown; Ellis, PhD thesis. University of Nottingham, 1986).
Taken together, our data show that pp60c-scr protein and kinase activity are expressed at elevated levels during the development of lens cells (lentoids) in transdifferentiating FHMEM cultures. In blocked FHGMEM or F199 cultures, there may be some persistence of c-src expression in surviving neuronelike cells. In order to resolve some of the remaining ambiguities, we plan to study pp60c-src expression (i) in neurone-stripped NR cultures, which are still able to transdifferentiate extensively into lens (de Pomerai & Gali, 1981; Shinde & Eguchi, 1982) and (ii) in transdifferentiating cultures of retinal pigmented epithelial cells (Eguchi & Okada, 1973).
Our immunocytochemical data also suggest that pp60c-src expression may occur in lens precursor cells as well as in newly formed lens fibres in transdifferentiating cultures. Fig. 5 shows that some individual cells or groups of cells in monolayer (nonlentoids) regions of late FHMEM cultures stain strongly for pp60c-src but (usually) weakly for δ crystallin; such cells may be embarking upon conversion into lens. Conversely, some groups of cells within well-developed lentoids are almost negative for pp60c-src but strongly positive for δ crystallin. Tentatively, we suggest that pp60c-src expression may not persist in established lens fibres, a view supported by the kinase activity data for fresh lens (Fig. 2B). The highest kinase activities observed in lens are found in the epithelial fraction of newly hatched chick lens (Fig. 2B). However, it should.be noted that lens-fibre cells contain high levels of crystallins, which would tend to reduce the apparent kinase activity of lens tissue when measured (as here) per milligram protein rather than per cell. Even so, cortical fibres show rather more pp60c-src activity than do nuclear fibres, again suggesting that mature lens fibres contain lower levels of the c-src kinase. In lens epithelial (LE) cultures, lentoid bodies (lens-fibre-like cells) differentiate after about 2 weeks in vitro (Okada, Eguchi & Takeichi, 1971). Our measurements of c-src kinase activity in cultures of 19-day LE cells after 10 days (prelentoid stage) or 35 days (numerous lentoids present) in vitro both give similar results to fresh lens epithelium tissue (Fig. 2A), consistent with our suggestion that elevated pp60?-s?? levels may characterize lens-fibre precursor cells as well as newly formed lens fibres. It is perhaps surprising that the 35day LE cultures (abundant lentoids) do not show higher c-src kinase activity, as found in transdifferentiating NR cultures (Fig. 2A). One possible explanation is that these late LE cultures are effectively postlentoidogenic (cf. 50- to 60-day NR cultures in FHMEM). Alternatively, it is possible that pp60c-irc levels may be exaggerated in transdifferentiated (NR-derived) as compared to normal lens cells. The highest activity detected in lens tissue (21-day lens epithelium) is not as high as that found in transdifferentiating NR cultures; however, it is possible that the pp60c-src activity in lens epithelium may be confined to the equatorial germinal zone (where new lens fibres are being formed). If so, there may be no significant difference in pp60c-src kinase activity between lens fibre differentiation in situ (lens) and in transdifferentiating NR cultures. We are currently using immunocytochemical staining of sectioned lenses from various stages in development in order to determine whether pp60c-src kinase is indeed located mainly in the germinal zone in vivo. Furthermore, the relative levels of pp60c-src protein and kinase activity will need to be quantified during lens cell development by both normal and foreign (transdifferentiation) pathways. Our present data leave open the possibility that transdifferentiation may involve an abnormal degree of cellular oncogene (c-src) expression relative to lens development in vivo.
ACKNOWLEDGEMENTS
The authors would like to thank Dr A. Barnekow, Institut für Virologie, Giessen, FDR, for valuable advice on the protein kinase assay, and Dr R. H. Scott for advice on immunocytochemistry. We would also like to thank Mr M. Hutchings for preparing the photographs and Miss W. M. Lister for typing the manuscript. D.K.E. and A.C. were in receipt of SERC studentships and this work was supported by a grant from the Cancer Research Campaign.