/ . Embryol. exp. Morph. Vol. 62, pp. 291-308, 1981
Printed in Great Britain © Company of Biologists Limited 1981
291
Influence of serum factors
on the prevalence of "normal" and "foreign"
differentiation pathways in cultures of chick
embryo neuroretinal cells
By D. I. DE POMERAI AND M. A. H. GALI
From the Department of Zoology, University of Nottingham
SUMMARY
Embryonic (9-day) chick neuroretinal cells transdifferentiate extensively into lens and
pigment cells during prolonged culture (4-5 weeks) in media containing foetal calf serum.
Medium conditions which promote the attachment and differentiation of neural cells in other
culture systems (e.g. horse serum, high glucose levels) both delay the onset and greatly
reduce the extent of transdifferentiation in retinal cultures. In the presence of high glucose,
horse serum (but not foetal calf serum) also favours cholinergic neuronal differentiation
during the early phases of culture, as shown by the levels of choline acetyltransferase activity
and accumulation of labelled choline. Substrate conditions have some effect on cholinergic
differentiation (promoted by polylysine-coated dishes) but do not affect later transdifferentiation. These effects may be due in part to selective survival or growth of particular retinal cell
types under the various medium conditions tested. Cultures stripped of neuronal cells contain
negligible choline acetyltransferase activity, but still transdifferentiate into both lens and pigment cells, although more slowly than control cultures. Cell size distributions reveal a
significant depletion of the larger cells in high glucose media with foetal calf serum, but not
in those with horse serum.
INTRODUCTION
The term metaplasia is used to describe the conversion of a cell from one
fully-differentiated phenotype to another, a process involving dedifferentiation
followed by redifferentiation (as exemplified by Wolffian lens regeneration from
the dorsal iris epithelium in newts; reviewed by Yamada, 1977). Transdifferentiation, by contrast, is a term of wider application, used to describe the conversion of partially-differentiated cells from one tissue into fully-differentiated
cells of a foreign tissue type (see Okada, 1976). The appearance of lens and
pigment cells during long-term culture of embryonic chick neural retina (Okada,
Itoh, Watanabe & Eguchi, 1975; Itoh, Okada, Ide & Eguchi, 1975) is of this
Author's address: Department of Zoology, University of Nottingham, University Park,
Nottingham NG7 2RD, U.K.
292
D. I. DE POMERAI AND M. A. H. GALI
type, since the initial tissue is not yet fully differentiated, although clearly neither
lens nor pigment cells would ever appear among normal neuroretinal cells in
vivo. This capacity for transdifferentiation during long-term culture in vitro may
reflect some relict potential of the optic cup tissues to regenerate both lens (as
in Wolffian lens regeneration) and each other (Lopashov, 1963; Coulombre &
Coulombre, 1965). In vivo however the neural retina develops into a multilayered neural tissue comprising supportive Miiller cells (glia) and several
morphologically distinctive neuronal cell types (photoreceptor cells, bipolar
and horizontal cells, amacrine cells and ganglion cells; reviewed by Cajal, 1893).
These layers become established by about the 15th day of embryonic development in chick, at which stage transdifferentiation into lens in culture is still
possible, though much less pronounced than at earlier stages (de Pomerai &
Clayton, 1978; Nomura & Okada, 1979). The various cell layers of the avian
retina utilize a number of different neurotransmitter systems, including acetylcholine (Vogel & Nirenberg, 1976; Cristani-Combes, Pessac & Calothy, 1978),
catecholamines (Ehinger, 1967), GABA (Marshall & Voaden, 1974) and possibly
also substance P (Karten & Brecha, 1979) and serotonin (Suzuki, Noguchi &
Yagi, 1978).
In cultures of 3^-day embryonic chick neuroretinal cells, there is an initial
increase in choline acetyltransferase (CAT; E.C.2.3.1.6) activity, followed by
a sharp decline when 8 crystallin begins to appear (Okada, Yasuda & Nomura,
1979). CAT is a marker of cholinergic neuronal differentiation in chick retina
(Crisanti-Combes et al. 1978), while 8 crystallin is confined to chick lens fibres
laid down during embryonic and early post hatching development (Clayton,
1974, 1979). The present paper extends this approach, using cultures of neuroretinal cells from later 9-day chick embryos, and examining two parameters of
cholinergic neuronal differentiation (CAT activity and choline accumulation)
under culture conditions used widely in other neural systems to encourage
neuronal differentiation. These include the use of horse serum (Fedoroff & Hall,
1979) in combination with high glucose levels (Adler, Manthorpe & Varon,
1979), and highly adhesive substrates such as polylysine-coated dishes (Pettmann, Louis & Sensenbrenner, 1979). No combination of these conditions yet
tested can block transdifferentiation completely, though media based on horse
serum both delay its onset and greatly reduce its extent, while favouring various
aspects of cholinergic neuronal differentiation during the earlier phases of
culture. The effects of these conditions on other aspects of retinal differentiation
(including markers for other neurotransmitter systems and for Miiller cells) in
cultures of 9-day and later embryonic neuroretinal cells, are under current
investigation.
MATERIALS AND METHODS
Fertile chick eggs were obtained from J. B. Eastwood Ltd (Nottingham).
Sources of chemicals and apparatus are given below where appropriate.
Pathways of retinal differentiation in culture
293
Cell culture
Neuroretinal cells from 9-day chick embryos were cultured as described
previously (Okada et al. 1975; de Pomerai, et al. 1977; de Pomerai & Clayton,
1978), except that retinas were dissociated for 60min. in 0-002% Trypsin
(Sigma Type 1) in calcium- and magnesium-free saline (CMF) to minimize
damage to neuronal cells (Varon & Raiborn, 1969). Standard culture medium F
was Eagle's MEM with Earle's salts and 2 mM L-glutamine, supplemented with
10% foetal calf serum (FCS), 100 units/ml Penicillin, lOO^g/ml Streptomycin
(all from GIBCO - Europe, Paisley, Scotland) and 26 mM NaHCO 3 . H medium
contained 10 % horse serum (GIBCO) in place of FCS. Hg and Fg media were
H and F media containing extra glucose to 28 mM (insteaed of 6 mM) final. Some
dishes were coated with polylysine (5/tg/ml; M.W.>70000; Sigma Type IB)
according to Pettmann et al. 1979. Certain cultures were largely (> 90%)
stripped of neuronal cells after 4 days in vitro by gentle swirling following a
15-min incubation in 0-001% Trypsin/0-01 % collagenase (type 1, Sigma) in
CMF (K. Yasuda, personal communication). All cultures were seeded with
3-7 x 106 neuroretinal cells per ml of culture medium, and were maintained
in vitro for up to 8 weeks as described previously (de Pomerai et al. 1977), the
medium being changed every 2-3 days. Cell counting after trypsinisation
(0-2 % Sigma Type 1 in CMF for 2-3 h) and photography of cultures were also
as described previously (de Pomerai et al. 1977). A model ZB1 Coulter counter
was used to obtain cell size distributions (cell suspension diluted 1-5 ml to
30 ml) at 1 /im diameter intervals across the size range 1-25 ju,m. The counter
was calibrated using 18-5 fim diameter Latex beads (Coulter Ltd).
Identification of crystallins
Analysis of the soluble proteins extracted from cultures was performed by
SDS-polyacrylamide gradient slab gel electrophoresis, as detailed previously
(de Pomerai & Clayton, 1978), and protein contents were determined by the
Lowry method (Lowry, Rosebrough, Farr & Randall, 1951). Crystallins were
identified by comparison with chick lens protein standards (Thomson et al.
1978).
Haemagglutination inhibition assays were performed as described previously
(de Pomerai & Clayton, 1978,1980), except that the marker sheep red blood cells
were coated with a mixture of equal parts of soluble proteins from newlyhatched chick lens (rich in 8 crystallin) and from adult chick lens cortex (rich
in a and /? crystallins) by the method of Evans, Steel and Arthur (1974). This
gave good agglutination with the anti-a, anti-total-/? and anti-# crystallinspecific antisera, which were the same as described previously (de Pomerai &
Clayton, 1978, 1980). Assays were always performed in duplicate, and several
sets of such assay results have been averaged to give a final mean.
294
D. I. DE POMERAI AND M. A. H. GAL1
Days in culture
Fig. 1. Growth of 9-day embryonic neuroretinal cultures in F, Fg, H and Hg media.
Cultures were set up and maintained in these four media as described in Methods.
At the indicated time points duplicate dishes were trypsinized and the cells counted
by haemocytometer. At least 2000 cells were counted per sample, and duplicate
dishes varied by less than 10% from each other. Mean cell counts from both dishes
have been plotted without ranges for the sake of clarity. •
• , F medium;
•
D,Fg medium; O
O, H medium; #
# , Hg medium.
Assessment ofneuronal differentiation
(a) Choline acetyltransferase (CAT) assays.
These assays were performed on culture extracts as detailed by CrisantiCombes et al. (1978), except that all assay components except [14C]Acetyl-Co A
(59/*Ci//*mole; Radiochemical Centre, Amersham) were present at one fifth
of the specified final concentrations. Blank assay values (using extraction buffer
in place of culture extract) have been subtracted from all experimental results.
All assays were performed in duplicate, and several sets of data from a number
of independent culture batches have been averaged to give the means and
standard errors plotted in Fig. 7.
(b) Choline accumulation
Choline incorporation into membrane components was assessed by the
method of Barald & Berg (1979). Cultures were washed and incubated in
Pathways of retinal differentiation in culture
295
B
i i i i i i i i i i i i i i i i i i i i i i i
r i i T i iT M I i i i i i 1 I I I I
10 12 14 16 18 20 22 24 2 4 6
Cell diameter Oim)
I I I I I I I I I I I I I I I I I
M
8 10 12 14 16 18 20.22 24
Fig. 2. Cell size distributions for retinal cultures in F, Fg and Hg media. These cell
size distributions were obtained using a Model ZBl Coulter counter with appropriate
threshold settings for 1 /*m diameter intervals in the range 1-25 /im. The four parts
of Fig. 2 were derived from a single experiment, the cells (1-5 ml suspension diluted
to 30 ml) from one dish of each type being counted (01 ml per count) at each stage.
Part A, freshly dissociated retina. Part B, retinal cultures at 4 days. Part C, retinal
cultures at 10 days. Part D, retinal cultures at 20 days. For parts B, C, and D, histo, F medium;
, Fg medium; . . ., Hg
gram bars are identified as follows:
medium. Similar assessments have been performed with three independent sets of
cultures, the overall size distributions in the three media being virtually
indistinguishable from that shown. N.B. The histograms plotted in this figure
give actual cell numbers within the indicated size ranges from 01 ml samples of the
diluted suspension; no attempt has been made to correct for the different
numbers of cells present under,the three medium conditions (see Fig. 1).
1 /tM-[3H]choline (55 Ci/mmole; Radiochemical Centre, Amersham) in phosphate-buffered saline (PBS) for one hour, followed by extensive washing for a
further hour in PBS containing 10/AM unlabelled choline. Hemicholinium III
(10/*M) was present throughout the procedure in some cases, so as to inhibit
specific choline uptake (Barald & Berg, 1979).
296
D. I. DE POMERAI AND M. A. H. GALI
RESULTS
A. Growth and morphology
Of the four culture media originally tested (see Fig. 1), that with horse serum
but no added glucose (H) failed to support cell growth and was not further
studied. During the first two weeks of culture, cell numbers fall progressively
under all conditions, but more markedly in media containing foetal calf serum
(F, Fg) as compared to horse serum (H, Hg) and independent of added glucose
(Fig. 1). When cell numbers start to recover, however, unsupplemented F
medium supports significantly faster culture growth, whereas cell numbers
increase more slowly in the glucose-supplemented media, independent of the
serum type (Hg and Fg; Fig. 1). This suggests that the neuronal and glial cell
populations present in retinal cultures (Okada et al. 1979) may respond differently to these various medium conditions in terms of survival and/or growth.
Further indications of this are given in Fig. 2, which shows a series of cell size
distributions obtained from freshly-dissociated 9-day embryo retinas and from
cultures of the same cell batch maintained for 4, 10 and 20 days in F, Fg and
Hg media. Of particular interest at 10 and 20 days is the tail-out towards larger
cell sizes (10-16 fim diameter) prominent in F and especially Hg cultures
(relative to the numbers of smaller cells) but severely depleted in Fg medium.
It should be noted that the largest (> 12 /tm) class of cells initially present in
fresh retina almost disappears by 4 days in culture.
Figure 3 compares the development of retina cultures in F, Fg and Hg media
on plastic dishes, while Fig. 4 compares development in F and Hg media on
polylysine-coated dishes. In morphological terms, neuronal differentiation is
more apparent in Hg and Fg media as compared to F medium at 15 days
(Fig. 3, i-iii); cell processes are also better developed on a polylysine substrate
in Hg as compared to F medium (Fig. 4, i-iv; at 6 and 15 days). By 40 days,
lentoids (aggregates of lens fibre cells wrapped around each other; Okada et al.
1975) and pigment cells are prominent in all cultures maintained in F medium,
(Fig. 3, iv), but few if any could be found in those kept in Hg medium (Fig. 3, vi).
This remains true even after 60 days of culture in Hg medium (results not shown).
Some lentoids and pigment cells also develop in Fg cultures after about 40 days
in vitro (Fig. 3, v), though less extensively than in F medium (Fig. 3, iv). Cells of
neuronal morphology, however, can survive in Hg and Fg cultures for periods
of 40 to 50 days (Fig. 3, v, and vi), whereas most such cells disappear from F
cultures after about 30 days (Fig. 3, iv; c.f. Okada et al. 1975; de Pomerai et al.
1977). The contrast between F and Hg media in terms of lentoid formation does
not seem to be affected by the culture substrate, as shown by parallel cultures
on polylysine-coated dishes (Fig. 4, v and vi; at 40 days).
Pathways of retinal differentiation in culture
297
Fg
Hg
15 days
40 days
100 Mm
i
,
Fig. 3. Development of retinal cultures in F, Fg and Hg media. This figure shows
typical fields from living cultures on plastic dishes photographed under phase
contrast. Top row, cultures at 15 days. Bottom row, cultures at 40 days. Superscripts F, Fg and Hg denote the media used. Bar represents 100 fim.
B. Crystallin appearance
Figures 5 and 6 demonstrate the appearance of chick lens crystallins in these
cultures, using haemagglutination inhibition assays with specific anticrystallin
antisera (Fig. 5) and SDS poly aery lamide gradient slab gel electrophoresis
(Fig. 6). The crystallins (especially 8) appear in much greater amounts in cultures
maintained in F medium as compared to Hg medium (a+/3 + S crystallin represents some 40 % of the total soluble protein extracted from 52-day-old cultures in
F medium, as compared to 3 % in Hg medium at the same stage; Fig. 5, A and
C). Cultures in Fg medium tend to degenerate after about 50 days in vitro, and
reliable data were not obtained beyond 46 days. Nevertheless, it is clear that
298
D. I. DE POMERAI AND M. A. H. GALI
6 days
15 days
*
•
•
* »
TL. * * »
40 days
100/im
Fig. 4. Effect of substrate on retinal culture differentiation in F and Hg media.
These fields were photographed from living cultures maintained in F (first column)
and Hg (second column) media on polylysine-coated plastic dishes. Top row, cultures
at 6 days; middle row, cultures at 15 days; bottom row, cultures at 40 days. Bar
represents 100 /«n.
this medium gives intermediate results throughout (Fig. 5B). fi crystallin is more
prominent than 8 for cultures in Hg medium, whereas the converse is true for
Fg and particularly F media. This is in accord with previous results showing
that ot+fi:8 crystallin ratio increases with declining extent of transdifferentia-
Pathways of retinal differentiation in culture
299
Z
K O
20
30
40
50 20
30
40
50 20
30
40
50
Days in culture
Fig. 5. Appearance of crystallins in retinal cultures under various conditions. This
figure shows the results of haemagglutination inhibition assays performed on the
soluble proteins extracted from 20- to 52-day-old cultures of 9-day embryonic chick
neuroretinal cells (see Methods). Parts A, B and C compare the appearance of
crystallins in these cultures under three medium conditions, namely Hg (part A),
Fg (part B) and F (part C). At least four sets of duplicate assay results were used in
compiling thesefigures,the mean and standard error (vertical bar) being plotted for
all points except that at 20 days (when standard error bars lie within the symbol
area). A
A, # crystallin; #
• , /? crystallins; •
• a
cry stall in.
tion, whether due to advancing embryonic age (de Pomerai & Clayton, 1978) or
reduced culture growth rate (de Pomerai & Clayton, 1980). Table IB compares
the amounts of crystallins after 34 and 50 days in vitro in controls and in retinal
cultures which had been stripped of at least 90 % of their neuronal cells (de
Pomerai and Gali, unpublished data). Both lentoids and pigment cells appear
in these 'epithelial' cultures, although more slowly than in controls. This is true
for all three classes of crystallin, the levels being four- to eightfold lower than
controls at 34 days, but at most twofold lower by 50 days. Again /? crystallin is
more prominent in epithelial cultures than in controls (/? levels at 50 days are
almost identical). This is compatible with the suggestion (de Pomerai & Clayton,
1980) that neuronal cells may contribute (directly or indirectly) a population of
#-rich lentoids during transdifferentiation, although other explanations (e.g. in
terms of cell density) have not been tested yet.
In Fig. 6, the prominence of protein bands migrating to the same positions
300
D. I. DE POMERAI AND M. A. H. GALI
2
3
4
5
6
7
5—
Fig. 6. Appearance of crystallins in retinal cultures maintained in F, Fg and Hg
media. Soluble proteins were extracted from F and Fg cultures at 25 and 35 days,
and 100 fig samples run on an SDS-polyacrylamide gradient slab gel as described in
Methods. Slot 1: F cultures after 25 days. Slot 2; Fg cultures after 25 days. Slot 3:
F cultuies after 35 days. Slot 4: Fg cultures after 35 days. Slot 5: 50/fg of newlyhatched chick lens soluble protein. 80 /*g samples of soluble protein from 36 day
cultures maintained in F and Hg media were run as above on a separate gel. Slot 6:
75 fig of newly-hatched chick lens soluble protein. Slot 7: Hg cultures after 36 days.
Slot 8: F cultures after 36 days. Crystallin bands identified according to Thomson
et ah (1978).
as the a and 8 crystallins correlates well with the results shown for F, Fg and
Hg cultures in Fig. 5 (A, B, and C). After 36 days in vitro, these proteins are
more prominent in F than in Fg cultures and are virtually absent in Hg cultures.
However, a minor retinal protein of the same apparent subunit molecular
Pathways of retinal differentiation in culture
7 -
301
Fg
6 -
I?
5 4 3 -
S
21-
T
12
T
16
~i—i—i
12
16
20
Days in culture
Fig. 7. Choline acetyltransferase activity in F, Fg and Hg cultures. CAT assays were
performed (as described in Methods) at various stages during cell culture in the three
media, and results are given as net dpm (above blank values) per /<g soluble protein.
Mean and standard error (vertical bar) are plotted for all points where three or
more independent cultures (each sample in duplicate) have been assayed. Some
additional points are included where only one or two cultures have been assayed
(in duplicate); these are plotted as single mean points and should be regarded as
merely indicative. The data plotted in this figure were obtained from several
different sets of cultures set up at initial cell densities ranging from 3 to 7 x 106 cell
per ml. Part A, F medium; part B, Fg medium; part C, Hg medium.
weight as 8 crystallin is present in Hg cultures and also at 25 days in F and Fg
cultures; in all these cases, 8 crystallin is barely detectable immunologically
(c.f. de Pomerai et al 1977; de Pomerai & Clayton, 1978). Soluble proteins
were also compared between F and Hg media for 36 day cultures on plastic,
collagen or polylysine substrates (results not shown); few if any differences
according to substrate were found, although the pattern of protein bands was
consistently different in Hg as compared to F cultures. In particular, the
prominent bands found at the a and 8 crystallin positions in all F culture samples
were much reduced or absent in all Hg culture samples (c.f. Fig. 6, Hg and F at
36 days). Overall, Fg medium reduces lentoid and crystallin production relative
to F medium, while Hg medium largely eliminates both. Substrate conditions
have little effect on this contrast.
C. Choline metabolism
Figure 7 shows the results of choline acetyltransferase (CAT) assays using
cultures grown under all three medium conditions. After an initial two- to
threefold increase above the levels found in fresh 9-day embryo retinal cells,
there is a decline in CAT activity in F and most markedly in Fg cultures, whereas in Hg cultures the CAT activity reaches higher levels and is maintained for
302
D. I. DE POMERAI AND M. A. H. GALI
Table 1 a. Choline acetyltransferase activity in 12-day cultures
of retinal cells
Medium
CAT activity (dpm
per fig protein)
Substrate
Plastic
Polylysine-coated
plastic
Plastic
Hg
Polylysine-coated
plastic
'Epithelial' cultures,, from which
> 90 % of neurones had been
removed on day 4 (F
i medium,
plastic dishes)
F
51 ± 13
384 ±46
362 ±23
607 ±71
15±4
This table compares CAT activity (expressed as dpm per //g soluble protein) at day 12 of
culture in both F and Hg media using untreated plastic and polylysine-coated plastic dishes.
In addition, 'epithelial' cultures (F medium, untreated plastic dishes) were prepared by
removing over 90% of the neuronal cells after 4 days in culture (see Methods), and CAT
activity was again assayed at 12 days. This table gives means plus standard errors obtained
from three cultures of each type, each sample being assayed in duplicate.
Table 1 b. Crystallin contents of control and epithelial cultures
Crystallin (as percentage of total soluble protein)
Crystallin
Class
ctj
and 8
a
fi
8
a
fi
8
Controls (F medium,
plastic dishes)
(O/\
\/o)
All « 1
51 ±0-7
40 ±0-5
6-8 ±0-9
8-4±21
15-2±3-2
26-5 ±4-1
Epithelial cultures
(stripped of neurones
on day 4; F medium,
plastic dishes).
(%)
Days in
culture
20
Barely detectable
0-9 ±0-2
1-2 ±0-2
0-8±01
6-2 ±1-3
14-5±3-4
12-4±2-6
1
J
1
J
34
50
The content of a, /? and 8 crystallin in' epithelial' (as above) and in control (unfractionated)
cultures was assessed at 20, 34 and 50 days by means of haemagglutination inhibition assays
as described in methods. Mean percentages (±S.E.) derived from three duplicate assays are
given for all three crystallin classes.
Pathways of retinal differentiation in culture
303
Table 2. Choline accumulation in membrane components of F, Fg
and Hg cultures
Net [3H]choline uptake in 60 min
(d.p.m. per/fg protein)
Medium
F
Fg
Hg
F
Fg
Hg
F
Fg
Hg
Mean % inhibition
by Hemicholinium
Days in
culture Control (no HCII1) WithlO/*mHCIlI
III
3
3
3
10
10
10
19
19
19
1060 ±50
1200 ±40
430±15
550 ±25
400 ±20
310±20
100±10
140±15
210+10
300 ±30
400 + 20
110±10
110+ 10
180±10
8O±8
120 ± 5
13O±5
120+10
73
66
75
66
55
74
17
6
42
Cultures maintained in F, Fg or Hg media were compared after 3, 10 or 19 days in vitro.
Each 1 hour incubation in [3H]choline (1 iiu) was followed by a 1 h chase in unlabelled
choline (10//M): see Methods and Barald & Berg (1979) for details. Duplicate samples from
each of three culture dishes were used to obtain each mean and standard error. Hemicholinium III, when present, was at 10/*M throughout.
considerably longer. After about 20 days, CAT activity is negligible in all types
of retinal culture studied to date (de Pomerai, unpublished data). It should be
noted that the data included in Fig. 7 are derived from several different sets of
cultures (not merely that used for Fig. 1) and that the differences in CAT
activity (per /ig protein) were maintained even between low-density Hg cultures
and high-density F cultures. A non-specific cell-density effect on CAT activity
is thus unlikely to explain these results (see also Discussion). Substrate conditions have little effect during the first 8 days of culture, but polylysine-coated
dishes promote CAT activity markedly between 8 and 16 days in vitro, as shown
in Table 1A (at 12 days). Conversely, cultures which have been largely (> 90 %)
stripped of neuronal cells after 4 days in vitro, contain negligible CAT activity
at 12 days (Table 1 A). Overall, CAT activity correlates well with the morphological development of neuronal cells under F and Hg (but not Fg) conditions
during the first 2 weeks of culture (c.f. Figs. 3 and 4). Thereafter, the surviving
cells of neuronal morphology lack significant CAT activity, even though they
can apparently still bind the neurone-specific marker tetanus toxin (Mirsky et
al. 1979), as revealed by indirect immunofluorescence (de Pomerai, unpublished
results).
The trend shown by Fig. 7 is supported by preliminary studies of [3H]choline
accumulation. For cultures at 8 days, the rate of choline uptake (over a 5 to
60 minute assay period) is consistently higher for Hg than for F medium across
the concentration range 10~8 to 10~6 M choline, and at these low concentrations,
304
D. I. DE POMERAI AND M. A. H. GALI
choline is apparently taken up against a concentration gradient (de Pomerai,
unpublished data). Further studies of the kinetics and sodium and temperature
dependence of this process are now under way.
For Table 2, retinal cultures were labelled for 1 h in 1 /AM [3H]choline and then
washed for a further hour in 10 pm unlabelled choline. According to Barald and
Berg (1979) this technique detects mainly the choline incorporated into membrane components. For Hg cultures at 8 days, the rate of choline uptake (per
unit protein) immediately after labelling is higher than for F cultures (see
above), yet choline conversion into membrane material is markedly lower at
3 and 10 days for Hg as compared to F or Fg cultures (Table 2). This might
suggest that a larger fraction of the choline taken up by Hg cultures may be
metabolised e.g. as acetylcholine, consistent with the higher CAT activity in
these cultures (Fig. 7). Hemicholinium III inhibits choline uptake into membrane material effectively (66 to 74%) under all medium conditions at 3 days
(Table 2). At 10 or 19 days, however, this inhibition is much more effective in
Hg medium (74 % and 42 % respectively) than in F or Fg media (55-66 % a n d
6-17% respectively). Taken together with the results in Fig. 7, this evidence
indicates that choline metabolism falls off more rapidly in F and Fg media than
in Hg medium.
Preliminary evidence from autoradiography of these choline-labelled cultures
(data not shown) also tends to support this inference. Silver grains occur more
frequently over neuronal aggregates than over flattened epithelial cells in the
case of Hg but not F or Fg cultures, nor in Hg cultures incubated with 10 JUM
hemicholinium III. Owing to geometrical considerations, the grain densities
over these different cell types cannot be compared directly, and the results
obtained to date are merely indicative of higher neuronal uptake in Hg as
compared to F or Fg media.
DISCUSSION
The results reported in this paper show that cholinergic neuronal differentiation in chick embryonic neuroretinal cultures is favoured by horse serum (in
conjunction with high glucose), which also practically blocks later transdifferentiation into lens. The factors responsible for these effects have not been
characterized as yet. Some inhibition of lens cell formation can be obtained with
high glucose alone (i.e. in combination with foetal calf serum), while polylysinecoated substrates may promote neuronal differentiation without greatly affecting
transdifferentiation. Several qualifications must therefore be made, if the main
results are to be interpreted in terms of a choice between developmental pathways in culture.
First and foremost, neuroretinal cell cultures are not homogenous, but
initially comprise a monolayer of flattened epithelial cells (glial Miiller cells;
Linser & Moscona, 1979), overlaid by aggregates of small rounded neuronal
cells which become interconnected by long processes between the first and third
Pathways of retinal differentiation in culture
305
weeks of culture. The neuronal identity of these aggreates has been demonstrated by staining with Merocyanine 540 (Okada et al. 1979), by tetanus toxin
immunofluorescence (de Pomerai, preliminary results; Mirsky et al. 1978), and
indirectly by the fact that cultures stripped of neuronal cells show negligible
CAT activity compared with controls (this paper).
Secondly, there is as yet only indirect evidence that lens differentiation can be
initiated within these neuronal aggregates as well as from the underlying epithelial cells (Okada et al. 1979). Intermediate structures apparently showing
partial conversion of neuronal aggregates into lentoids are frequently encountered in cultures around the third week in vitro (de Pomerai & Clayton,
1980), and time-lapse film studies have shown cells within a neuronal aggregate
swelling up and elongating into 'bottle cells' and lentoid structures during long
term culture (K. Yasuda, pers. comm.; Okada et al. 1979). However, these data
do not preclude the possibility that neuronal-glial cell interactions may promote
lentoid development in vitro. The neurone-stripped cultures examined here
retained a few neuronal aggregates (< 10 %), and of course large numbers were
present up to the time of separation (day 4 in culture). To date there are no
reported methods for obtaining cultures of completely purified epithelial or
neuronal cell fractions from retinal material, in order to determine whether
either alone can produce lentoids in vitro.
A third problem is that markers of cholinergic neuronal differentiation
decline to low levels well before lentoids begin to appear in neuroretinal cultures.
This is true whether the lentoids appear early and in large numbers (F medium)
or late and very sparsely (Hg medium). Numerous neuronal aggregates and
even some interconnecting processes can be found in 5- to 6-week-old neuroretinal cultures maintained in Hg or Fg media (Fig. 3), but these no longer
possess high levels of cholinergic markers (Fig. 7). However, preliminary
evidence suggests that they can still be stained by tetanus toxin immunofluorescence (de Pomerai, unpublished) and may thus remain 'neuronal' in terms of
some cell surface features. This might imply gradual dedifferentiation of retinal
neuronal cells under in vitro culture conditions, with cholinergic markers being
lost more rapidly in some media (F and especially Fg) than in others (Hg).
It has been shown previously that crystallin production in F cultures first
becomes detectable between 12 and 16 days in vitro (de Pomerai & Clayton,
1978; de Pomerai et al. 1977). During this phase of culture, cholinergic markers
are at low levels in F and Fg media, but remain relatively high (albeit declining)
in Hg medium. Further experiments are needed to establish the true importance
of this phase, but ideally they will require the development of medium conditions
able completely to abolish transdifferentiation into lens.
A final problem is the possibility that differential culture growth, cell density,
and (most crucially) cell selection by the medium., might substantially affect
these results. At least the first two of these parameters are known to affect the
extent of transdifferentiation in chick neuroretinal cultures (Clayton, de
306
D. I. DE POMERAI AND M. A. H. GALI
Pomerai & Pritchard, 1977; de Pomerai & Clayton, 1980). However, the differences in culture growth rate and final cell density are only about twofold
between Fg and Hg media and F medium (Fig. 1), yet the inhibition of crystallin
production by Hg medium (Fig. 5 A) is far greater than that produced by Fg
medium (Fig. 5 A) or by a 10-fold difference in the initial cell density under
F conditions (Clayton et al. 1977). It might be argued that the greater cell
density in Hg medium (as compared to F or Fg media) during the first 12 days
of culture (see Fig. 1) could favour CAT activity (Fig. 7) via a non-specific celldensity effect. In fact, the data in Fig. 7 were derived from a number of separate
culture experiments in which the initial cell densities varied from 3 to 7 x 106
cells per ml. Even in the low-density Hg cultures, CAT activity per fig protein
remained consistently higher than in high-density F or Fg cultures. Thus, a
non-specific cell density effect on CAT activity seems unlikely, just as the later
differences in growth rate seem inadequate to account for the differential effects
of F and Hg media on crystallin production.
The cell size distribution in Fig. 2 show that cultures in Fg medium are
depleted in cells belonging to the larger (10-16 /*m diameter) size-classes (Fig. 2,
10 and 20 days), whereas these cells are prominent in both F and Hg media.
While this result does support the possibility of cell selection occurring in
neuroretinal cultures, there is no obvious correlation with any of the differentiation parameters assessed here. There are suggestions that the largest class
of cells in embryonic chick neuroretina may be predominantly ganglion cells, as
shown by Percoll fractionation of 14-day chick embryo retina cells (Sheffield,
Pressman & Lynch, 1980). Further work is in progress to characterize the
various size classes of retinal cell by means of specific neuronal markers, and
also Miiller cell markers such as glutamine synthetase (Linser & Moscona,
1979) and glial fibrillary acidic protein (Bignami & Dahl, 1979).
The authors would like to thank Drs N. L. Robinson and I. R. Duce for advice on the
assessment of neuronal differentiation, Mr R. Searcy for preparing the photographs, and
Mrs R. Clayton (University of Edinburgh) for generous gifts of anticrystallin antisera.
M.A.H.G. is in receipt of a postgraduate training award from the Government of the Republic
of Iraq. This work was supported by a grant from the Medical Research Council.
REFERENCES
ADLER, R., MANTHORPE, M. & VARON, S. (1979). Separation of neuronal and non-neuronal
cells in monolayer cultures from chick embryo optic nerve. Devi Biol. 69, 424-435.
BARALD, K. & BERG, D. K. (1979). Autoradiographic labelling of Spinal cord neurons with
high affinity choline uptake in cell culture. Devi Biol. 72, 1-14.
BIGNAMI, A. & DAHL, D. (1979). The radial glia of Muller in the rat retina and their response
to injury. An immunofluorescence study with antibodies to the glial fibrillary acidic
(GFA) protein. Expl Eye Res. 28, 63-69.
Y CAJAL, R. (1893). 'The vertebrate retina', republished in translation in RODIECK, R. W.
(1973). The Vertebrate Retina: Principles of Structure and Function, pp. 775-904. San
Francisco: W. H. Freeman & Co.
Pathways of retinal differentiation in culture
307
R. M. (1974). Comparative aspects of lens proteins. In The Eye, vol. V (ed. H.
Davson & L. T. Graham, Jr.), pp. 399-494. New York and London: Academic Press.
CLAYTON, R. M. (1979). Divergence and convergence in lens cell differentiation: regulation
of the formation and specific content of lens fibre cells. In Stem Cells and Tissue Homeostasis (British Society for Cell Biology Symposium 2) (ed. B. Lord, C. Potten and R. Cole),
pp. 115-138. Cambridge, London, New York and Melbourne: Cambridge University
Press.
CLAYTON, R. M., DE POMERAI, D. I. & PRITCHARD, D. J. (1977). Experimental manipulation
of alternative pathways of differentiation in cultures of embryonic chick neural retina.
Devi, Growth and Differ. 19, 319-328.
COULOMBRE, J. L. & COULOMBRE, A. J. (1965). Regeneration of neural retina from the pigmented epithelium in the chick embryo. Devi Biol. 12, 79-92.
CRISANTI-COMBES, P., PESSAC, B. & CALOTHY, G. (1978). Choline acetyl transferase activity
in chick embryo neuroretinas during development in ovo and in monolayer culture. Devi
Biol. 65, 228-232.
EHINGER, B. (1967). Adrenergic nerves in the avian eye and ciliary ganglion. Z. Zellforsch.
Mikrosk. Anat. 82, 577-588.
EVANS, J., STEEL, M. & ARTHUR, E. (1974). A haemagglutination inhibition technique for
detection of immunoglobulins in supernatants of human lymphoblastoid cell lines. Cell 3,
153-158.
FEDOROFF, S. & HALL, C. (1979). Effect of horse serum on neural cell differentiation in tissue
culture. In Vitro 15, 641-648.
ITOH, T., OKADA, T. S., IDE, H. & EGUCHI, G. (1975). The differentiation of pigment cells of
chick embryos. Devi, Growth and Differ. 17, 39-50.
KARTEN, H. J. & BRECHA, N. (1980). Localisation of substance P immunoreactivity in amacrine cells of the retina. Nature, Lond. 283, 87-88.
LINSER, P. & MOSCONA, A. A. (1979). Induction of glutamine synthesase in embryonic neural
retaina: localisation in Miiller fibres and dependence on cell interactions. Proc. natn. Acad.
Sci., U.S.A. 76, 6476-6480.
LOPASHOV, G. V. (1963). Developmental Mechanisms of Vertebrate Eye Rudiments.Translated
by J. Medawar. Oxford: Pergamon Press.
LOWRY, O. H., ROSEBROUGH, N., FARR, A. & RANDALL, R. (1951). Protein measurements
with the Folin-phenol reagent. /. biol. Chem. 193, 265-275.
MARSHALL, J. & VOADEN, M. J. (1974). An autoradiographic study of the cells accumulating
3
H y-aminobutyric acid in isolated retinas of pigeons and chickens. Invest. Opthalmol. 13,
602-607.
MIRSKY, R., WENDON, L. M. B., BLACK, P., STOLKIN, C. & BRAY, D. (1978). Tetanus toxin:
a cell surface marker for neurones in culture. Brain Res. 148, 251-259.
NOMURA, K. & OKADA, T. S. (1979). Age-dependent change in the trans-differentiation
ability of chick neural retina in cell culture. Devi, Growth and Differ. 21, 161-168.
OKADA, T. S. (1976). 'Transdifferentiation' of cells in specialised eye tissues in cell culture.
In Tests of Teratogenicity In Vitro (ed. J. Ebert & M. Marois), pp. 91-105. Amsterdam:
N. Holland Press.
OKADA, T. S., ITOH, Y., WATANABE, K. & EGUCHI, G. (1975). Differentiation of lens in
cultures of neural retinal cells of chick embryos. Devi Biol. 45, 318-329.
OKADA, T. S., YASUDA, K. & NOMURA, K. (1979). The presence of multi-potential progenitor
cells in embryonic neural retina as revealed by clonal cell culture. In Cell Lineage, Stem
Cells and Cell Determination, (ed. N. le Douarin). I.N.S.E.R.M. Symp. 10, pp. 335-346.
Amsterdam, New York and Oxford: North Holland Publishing Co.
PETTMANN, B., LOUIS, J. C. & SENSENBRENNER, M. (1979). Morphological and biochemical
maturation of neurones cultured in the absence of glial cells. Nature Lond. 281, 378-380.
DE POMERAI, D. I., PRITCHARD, D. J. & CLAYTON, R. M. (1977). Biochemical and immunological studies of lentoid formation in cultures of embryonic chick neural retina and dayold chick lens epithelium. Devi Biol. 60, 416-427.
DE POMERAI, D. I. & CLAYTON, R. M. (1978). Influence of embryonic stage on the transdifferentiation of chick neural retina cells in culture. /. Embryol. exp. Morph. 47, 179-193.
CLAYTON,
308
D. I. DE POMERAI AND M. A. H. GALI
DE POMERAI, D.
I. & CLAYTON, R. M. (1980). The influence of growth-inhibiting and growthpromoting medium conditions on crystallin accumulation in transdifferentiating cultures of
embryonic chick neural retina. Devi Growth and Differ. 22, 49-60.
SHEFFIELD, J. B., PRESSMAN, D. & LYNCH, M. (1980). Cells isolated from the embryonic
neural retina differ in behavior in vitro and membrane structure. Science 209, 1043-5.
SUZUKI, O., NOGUCHI, E. & YAGI, K. (1978). Uptake of 5-hydroxytryptamine by chick retina.
/ . Neurochem. 30, 295-296.
THOMSON, I., WILKINSON, C. E., BURNS, A. T. H., TRUMAN, D. E. S. & CLAYTON, R. M.
(1978). Characterisation of chick lens soluble proteins and the control of their synthesis.
Expl Eye Res. 26, 351-362.
VARON, S. & RAIBORN, C. W. (1969). Dissociation, fractionation and culture of embryonic
brain cells. Brain Res. 12, 180-199.
VOGEL, Z. & NIRENBERG, M. (1976). Localisation of acetylcholine receptors during synaptogenesis in retina. Proc. natn. Acad. Sci., (U.S.A. 73, 1806-1810.
YAMADA, T. (1977). Control Mechanisms in Cell-type Conversions in Newt Lens Regeneration.
Monographs in Developmental Biology (ed. A. Wolsky), p. 13. Basel: S. Karger.
(Received 27 June 1980, revised 24 October 1980)