/ . Embryol. exp. Morph. Vol. 62, pp. 47-62, 1981 4 7Printed in Great Britain © Company of Biologists Limited 1981

Transdifferentiation of chicken embryoneural retina into pigment epithelium:

indications of its biochemical basis

ByD. J. PRITCHARDFrom the Department of Human Genetics

University of Newcastle upon Tyne

SUMMARY

Neural retina from 8- to 9-day embryo chickens was grown in long-term cell culture in anexperiment to test the hpothesis that one step during the in vitro transdifferentiation of neuralretina into pigment cells occurs in response to stimulation of tricarboxylic acid (TCA) cycleactivity. Time-lapse photography showed that pigment-cell formation occurs through theintermediate stages of 'undistinguished cells', 'pavement epithelium' and 'potential pigmentcells'. Mitosis of undistinguished cells to pavement epithelium was proportional to malonateover most of the tested range of concentrations and was inhibited by succinate, which res-pectively depress and stimulate the TCA cycle. Conversely mitosis of pavement epitheliumto potential pigment cells occurred in proportion to succinate concentration over most of thetested range and was inhibited by malonate, in support of the hypothesis under test.

Melanin synthesis begins in a minority of 'pigment leader cells' uniquely stimulated by thelowest concentration of malonate, although higher concentrations blocked pigment synthesisin all cell types. The pigment leader cells appear to act as centres of influence upon neigh-bouring potential pigment cells, which subsequently also beome pigmented. Lactate inhibitedmost or all of the steps in formation of pigment epithelium.

Between three and five mitoses occur in the production of pigment cells, whereas multi-layers and lentoid bodies seem to be formed by expansion of undistinguished cells, probablywithout mitosis.

The observations lead to a general theory that metaplastic conversion between cell typesin eye tissues may require the physical isolation of overtly differentiated, multipotent cellsfrom 'leader' cells which normally hold them in physiological subjugation.

INTRODUCTION

The in vitro ' transdifferentiation' of 8- to 9-day chicken embryo neuralretina (NR) into lens fibre-like cells and pigment epithelium, first reported byT. S. Okada's group at Kyoto (Okada, Itoh, Watanabe & Eguchi, 1975;Itoh, Okada, Ide & Eguchi, 1975), is now a well-accepted phenomenon (seereviews by Clayton, 1978, 1979 and Okada, 1976). The direction and degree ofchange can be controlled by modification or choice of culture medium (Clayton,

Author's address: Department of Human Genetics, The University, 19 Claremont Place,Newcastle upon Tyne, NE2 4AA, U.K.

48 D. J. PRITCHARD

de Pomerai & Pritchard, 1977; Itoh, 1976; Okada, 1976, 1977; Pritchard,Clayton & de Pomerai, 1978); by selection of embryos of particular ages (Araki& Okada, 1978; Nomura & Okada, 1979; de Pomerai & Clayton, 1978); byadjusting the cell inoculum density; and by formation of artificial multilayers(Clayton et al. 1977). Significantly, the area occupied by the immediate pigment-cell precursors, or 'potential pigment cells', is directly proportional to bicar-bonate over the range covered by standard media, whereas that occupied byother cell types is not (Pritchard et al. 1978).

Retinal tissues show significantly different physiological behaviour in bi-carbonate-buffered, as compared with phosphate-buffered medium (Graymore,1969) due to incorporation of bicarbonate ions into the tricarboxylic acid(TCA) cycle (Crane & Ball, 1951a, b). It was to test the hypothesis that TCA-cycle stimulation through CO2 fixation was the basis of the bicarbonate effectupon transdifferentiation, that the experiments described below were carried out.

Cultures of chicken embryo NR were treated with sodium succinate andsodium malonate respectively to stimulate and inhibit the TCA cycle (Lehninger,1972). One expected outcome of TCA-cycle inhibition would be the increasedconversion of pyruvate into lactate. To control for the effects of lactate, asdistinct from TCA-cycle inhibition per se, cultures were grown also in mediumsupplemented with sodium lactate. The results support the hypothesis thatcontrol of TCA-cycle activity is a primary feature of the transdifferentiation ofneural retina into pigment cells.

MATERIALS AND METHODS

Neural retina cultures were established from trypsin-dissociated tissues takenfrom 8- to 9-day chicken embryos, as described by Okada et al. (1975). Theywere grown in loosely stoppered polystyrene tissue-culture flasks, in modified orstandard versions of Minimal Essential Medium (Eagle, 1959) based on Earle'ssalts, 'EMEM' (Earle, 1943), supplemented with 6 % foetal calf serum which wastaken from a single batch, 100 i.u./ml penicillin and 100/tg/ml streptomycin(all materials from Gibco Europe). The cells were inoculated at 7 x 106 perflask of floor area 25 cm2 and grown at 37 °C in a humid atmosphere of 5 %CO2 in air,

In the main experiment 40 identical, serially numbered cultures were set up,each in 5 ml of standard EMEM, which was replaced with 8 ml on day 2. Onday 6 they were divided into sets of four, each set containing numbers distri-buted throughout the series. From this time onwards one control set received8 ml of standard EMEM, while the other sets all received 8 ml of EMEMsupplemented with 1-5,0-75 or 0-15 g/1 of the appropriate biochemical: disodiumsuccinate hexahydrate (Sigma) at 5-6 (SH), 2-8 (SI) or 0-56 mM (SL); disodiummalonate monohydrate (Sigma) at 8-9 (MH), 4-5 (MI) or 0-9 mM (ML) andsodium lactate (L(+) lactic acid (Sigma grade L-l) titrated to neutrality with

Transdifferenlialion of chick embryo neural retina 49sodium hydroxide) at 16-7 (LH), 8-4 (LI) or 1-7 mM (LL). This supplementationdid not affect the pH of the media.

The medium was replenished every two or three days and the cultures wereassessed stereologically with a phase-contrast microscope around day 60, at30-100 sites per flask.

Six further cultures set up from the same cell preparation were given standardEMEM up to day 56. They were then divided into two equally pigmented setsof three. One set was given medium supplemented with 16*7 mM sodium lactate,while the others continued receiving standard EMEM. These 'late lactate'cultures and their controls were assessed and harvested on day 76.

Four cultures grown under control conditions were examined by time-lapsephotography at two fields per culture. Relocation of the fields was achievedby the method of Pritchard & Ireland (1977).

The growth curve was based on haemocytometer counts of suspended cellsfrom three replicate cultures at each time point.

RESULTS

Aggregate bodies

During the earliest stages of culture development, the 'minute cells' of thefreshly dissociated NR settle down singly, or in 'aggregate bodies' (Pritchardet al. 1978) composed of many cells, before expanding and spreading over thevessel floor as a sheet of large 'undistinguished cells' with no major distinctivefeatures (see Fig. 1B,C). At day 21, cultures supplemented with succinate,LL and LI contained normal numbers of aggregate bodies, but LH and allconcentrations of malonate produced marked reductions.

Minute and large cells

When released by trypsinization, the undistinguished cells and their derivativesadopt a spherical shape of diameter 10-20/tm, whereas those which have notexpanded become spheres of 2-5 /im diameter (Fig. 1 A). Minute cells multiplyat a rate similar to that of the large cells (Fig. 2) although the origin of newminute cells is not known.

Large cell numbers showed a similar pattern of response to the supplementsto that for production of potential pigment cells (Fig. 6B), but minute cellnumbers were less variable.

Total area of cell sheets

The final total areas of the cell sheets are shown in Fig. 3. Lactate and malo-nate depressed colonization of the vessel, while succinate marginally promotedit. Malonate caused considerable death of undistinguished cells.

50 D. J. PRITCHARD

Fig. 1. (A) Large and minute cells (some arrowed) in a suspension prepared bytrypsinization of an established culture. (B-E) The same area of a control culturegrown in standard EMEM and photographed under phase-contrast illumination atdifferent stages of development. (B) Aggregate bodies at day 3. (C) Undistinguishedcells with two aggregate bodies at day 14. (D) Pavement epithelium at day 37. (E)Potential pigment cells at day 53. (F, G) The same area of a culture at day 53photographed by normal, direct light (F) and phase-contrast illumination (G),showing melanin in the pigment leader cells. All photographs are to the same scale,the bar represents 100 /im.

Transdifferentiation of chick embryo neural retina 51

20 30Time (days)

Fig. 2. Growth curves of large, minute and total neural retina cells in a controlset of cultures grown in standard EMEM. Each point represents the mean haemo-cytometer count of cells harvested from three flasks. # , Total; O, large; #, minute.

00 0-4 0-8 1-2Concentration of supplement (g/1)

Fig. 3. The relationship between percentage areas of vessels occupied by cells andthe concentrations of supplements in the medium. Cultures were assessed aroundday 60. The supplements were disodium succinate hexahydrate, disodium malonatemonohydrate and lactic acid titrated to neutrality with sodium hydroxide. Eachpoint represents the mean assessment of four cultures, the standard error of the meanis denoted by an error bar unless smaller than the symbol. , succinate;

, lactate; , malonate.

52 D. J. PRITCHARD

Transdifferentiation of chick embryo neural retina 53

Gross appearance of mature cultures

The gross appearance of typical terminal cultures is shown in Fig. 4. Culturesgrown in MI, MH and LH conditions developed no pigment colonies. UnderLI and succinate supplementation their numbers were greatly reduced, but MLand LL conditions produced no marked effect (see Table 1).

Sequence of differentiative events

Time-lapse photographic analysis of cultures raised in standard EMEMrevealed the sequence of morphological changes which occur as undistin-guished cells differentiate into pigment epithelium (Figs. 1B-E, 8). The samesequence was recorded in three fields and accords with my general observationsof several hundred similar cultures.

From days 12-16 some of the undistinguished cells, of in situ diameterabout 37 jiim, divide once or twice and their outlines become more discrete asthey form a pavement epithelium with cells of mean in situ diameter about26/*m. These cells then subdivide a further two or three times to producepotential pigment cells of about 12/mi diameter (Pritchard et al. 1978), whichaccumulate pigment from about day 40. In a minority of areas the intermediatepavement epithelium stage is not detectable. The process of subdivision ofpavement cells spreads outwards from scattered foci (Pritchard et al. 1978).

Proportion of cell types in mature cultures

The relative areas occupied by the different cell types in mature cultures areshown in Fig. 5. There were great variations in their proportions.

Transdifferentiation was blocked before the potential-pigment-cell stage inLH cultures and at the melanogenic stages by MI and MH conditions.

Formation of pavement epithelium

Figure 6 A shows the fractions of the cell sheets which became convertedinto pavement epithelium, these values being derived by combining the areasfinally occupied by pavement, potential pigment and pigment cells. Formationof pavement epithelium was proportional to malonate concentration, althoughonly MH caused an increase relative to the control. Both lactate and succinateproduced significant reductions.

Fig. 4. Typical neural retina cultures photographed at day 60, showing variationin pigmentation. (A) Control culture grown in standard EMEM. (B-K) Culturesgrown in medium supplemented with low, intermediate and high levels of succinate(B, C, D), malonate (E, F, G) and lactate (H, J, K).

54 D. J. PRITCHARD

Table 1. The effects of the supplements upon pigmentation

Mean areaNo. of pigment Pigmentation initia- (mm2) of pigmentcolonies (S.E.) tion frequency* (S.E.) colonies f (S.E.)

ControlSuccinate

Low

IntermediateHigh

MalonateLow

Intermediate

High

LactateLow

Intermediate

High

102 (19-2)

15 (4-1)P < 002

59 (20-2)50 (3-7)

P < 005

133 (25-3)

0P < 002

0P < 002

71 (5-2)

6(3-8)P < 002

0P < 002

12-2(1*15)

10-3 (1-70)

18-5(2-15)14-8 (2-20)

270 (2-60)P < 002

0P < 002

0P < 002

14-0(1-74)

11-8(8-75)

t

5-0(1-36)

2-9 (0-43)

2-9 (0-35)1-7(0-58)

P < 005

2-7(1-57)

0P < 002

0P < 002

1-2(0-18)P < 005

~ 1-2probably significant

0P < 002

* Pigment colonies per cm2 of combined area of pigment and potential pigment cells.t Area occupied by pigment cells divided by number of pigment colonies.t This figure could not be calculated as there were no potential pigment cells.

Formation of potential pigment cells

The proportions of pavement epithelium which became subdivided intopotential pigment cells are shown in Fig. 6B. Malonate and lactate stronglyinhibited this step, but potential-pigment-cell formation was proportional tosuccinate concentration, although no treatment caused an increase relative tothe control. Since this is the major mitotic step, this pattern was closely similarto that of large cell numbers (results not shown).

Development of pigmentation

Melanogenesis develops only in large spreads of potential pigment cells andbegins in the foci where potential pigment cells first appeared. Groups of abouttwenty 'pigment leader cells' begin to synthesize pigment with no change ofform (Pritchard et al. 1978), or three or four adjacent cells expand, forcing theirneighbours to adopt a concentric pattern about them, before they becomepigmented (Fig. 1F, G). Pigmentation then spreads outwards from these foci.

Transdifferentiation of chick embryo neural retina 55

Fig. 5. Diagram to show relative percentage areas of vessels occupied by undis-tinguished cells (D), pavement epithelium (EH) potential pigment (^) and pigmentcells ( • ) in terminal cultures of neural retina. (A-K) as in Fig. 4.

The relative frequencies of sites of initiation of pigmentation were estimatedby relating numbers of pigment colonies to areas of cells which reach thepotential-pigment-cell state (see Table 1). Initiation of pigmentation wasunaffected by succinate or lactate. It was completely inhibited in MI and MHcultures, but interestingly, ML conditions caused a significant in crease.

Efficiency of spread of melanogenesis through the potential pigment cells isindicated by the size of the pigment colonies (Table 1). Mean colony area wassignificantly reduced by lactate and SH, but not SL or SI. The effects of MI andMH were not tested, as no colonies were initiated, but significantly ML had noeffect, suggesting that spread of pigment synthetic activity is biochemicallydistinct from that of its initiation.

56 D. J. PRITCHARD

- 2 0 _

00 0-4 0-8 1-2Concentration of supplement (g/1)

1-6

"oo

<D

S,

tia

c

ote

o.

E-t

elii

ith

D.

E

pave

o

ca

c<L>H<uD,00

80

60

40

20

0

B~ #

-

-

-1

\

• \ .4

1

4l ~ — •

*""̂ -^» ^̂i i

00 0-4 0-8 1-2

Concentration of supplement (g/1)

1-6

Fig. 6. (A) The relationship between the percentage area of the cell sheet which be-comes converted into pavement epithelium, or remains as undistinguished cells,and concentrations of supplements in the medium. (B) The relationship betweenthe percentage area of pavement epithelium which becomes converted into potentialpigment cells, and concentrations of supplements in the medium. Details as inFig. 3.

The 'late lactate' experiment

The observations recorded in Table 2 suggest that lactate probably exerts aninhibiting effect at most stages in the production of pigment cells, but this isstatistically significant only at the conversion of pavement epithelium to poten-tial pigment cells.

Undistinguished cells and multilayers

The proportion of undistinguished cells was markedly increased in allexperimental cultures, except MI and MH, in which there was much cell death(Figs 5, 6 A). Comparison of Figs 6 A and 7 demonstrates that the areasoccupied by undistinguished cells and multilayers are affected by the supple-ments in similar ways, but to different degrees, in accordance with the theory

Transdifferentiation of chick embryo neural retina 57

Table 2. The effect oflactate on mature cultures

Cultures were exposed to medium containing 16-7 mM sodium lactate at day 56and assessed twenty days later.

MeasureControl

(S.E.)

Lactate-treated

(S.E.) Significance

Cell sheet to pavement*

Pavement to pot. pig.*

Pot. pig. to pigment*

No. of pigmentcolonies

Pigmentation initiationfrequency

Mean area ofpigment colonies (mm2)

48%(6-2)89%(13)56%(30)128(116)121(0-69)4-67(0-38)

* The mean percentage area of the first eeltiation.

43%(46)75%(53)50%(32)83(84)10-5(17)4-90(0-68)

1 type which reached

N.S.

P < 005

N.S.

P < 002

N.S.

N.S.

the next stage of differen-

0-4 0-8 1-2Concentration of supplement (g/1)

1-6

Fig. 7. The relationship between the percentage area of the cell sheet which be-comes multilayered, and concentrations of supplements in the medium. Detailsas in Fig. 3.

that undistinguished cells produce multilayers by upward extrusion of cells.There is no indication that undistinguished cells are produced by any meansother than expansion of minute cells, and detailed counts of the different celltypes suggest that in these mass cultures (cf. clonal cultures, Okada et al. 1979)undistinguished cells do not undergo mitosis without yielding pavementepithelium.

58 D. J. PRITCHARD

^Vl

VI

Transdifferentiation of chick embryo neural retina 59

DISCUSSION

Figure 8 summarizes a tentative interpretation of the sequence of differentia-tive changes in long-term NR cultures and the biochemical influences whichaffect them. One or more of the supplements exerted a significant effect on everyvariable examined, and it is concluded that the TCA cycle has a profound effecton the transdifferentiation of NR cells.

The production of potential pigment cells from pavement epithelium occurredin proportion to succinate concentration and was inhibited by malonate, insupport of the hypothesis under test, namely that the promotory effect ofbicarbonate upon production of potential pigment cells involves TCA-cyclestimulation through CO2 fixation (see Introduction). However, relative to theunsupplemented control, this involved depression of mitosis at SL and noincrease, even at SH (Fig. 6B). Possibly the TCA cycle cannot be stimulated togreater activity than that in EMEM, which contains a very high concentrationof bicarbonate, but the depression of mitosis by SL requires explanation. Therewas a similar depression in the mitosis of pavement epithelium to potentialpigment cells due to ML, despite its increase with malonate concentration(Fig. 6 A). Also initiation of pigmentation was significantly increased comparedwith controls, by ML, but MI and MH completely inhibited it. Three possibleexplanations are offered for these non-linear effects:

(i) Some differentiate events occur only, or optimally, at critical levels ofTCA-cycle activity.

(ii) Important subsidiary biochemical pathways involving the supplementsmay be more susceptible to low-level supplementation than the TCA cycleitself.

(iii) The supplements may evoke compensatory responses which over-com-pensate at low levels of supplementation.

Fig. 8. Tentative scheme of differentiation in long-term mass cultures of neural retinafrom 8- to 9-day chicken embryos. Minute cells (I), in young cultures expand to formundistinguished cells (II, III) which produce multilayers (IV) probably by extrusion,and then lentoid bodies (V). Undistinguished cells divide to form pavement epi-thelium (VI), which on further mitosis yields potential pigment cells (VII, VIII). Thelatter occasionally arise direct from undistinguished cells. Pigment synthesis occursfirst at pigmentation initiation foci (VIII) containing pigment leader cells, then spreadsoutwards to produce pigment epithelium (IX). Occasionally lentoid bodies arisewithin the pigment epithelium. Step I-II is inhibited by bicarbonate (Pritchardet al. 1978) and steps II-1V by malonate. Step II-VI occurs in the presence of malo-nate, but is inhibited by succinate and lactate. Step VI-VII is promoted by bicarbo-nate (Pritchard et al. 1978) and occurs in the presence of succinate, but isinhibited by lactate and malonate. Step VII-VIII is promoted by low concentrationsof malonate, but inhibited by lactate and intermediate and high concentrations ofmalonate. Step VIII-IX is inhibited by lactate. M, Observed mitotic events.

60 D. J. PRITCHARD

Low concentrations of malonate significantly boosted initiation of pigmentsynthesis in the pigment leader cells, but had no effect on its subsequent spreadthrough the potential pigment cells. This implies that the leaders are particularlyresponsive cells which, using a different biochemical means, influence theirless responsive neighbours to adopt, overtly at least, a similar phenotype.Pigment density tends to be uniform within any one colony, although variablebetween colonies in the same culture, suggesting physiological communicationbetween contiguous cells (see Loewenstein, 1968). The concept that multi-potent follower cells are only overtly subjugated by the physiology of neigh-bouring determined leaders might explain the capacity of pigment epithelialcells from mature NR cultures to transdifferentiate into lens when isolated insubculture (Okada, Yasuda, Araki & Eguchi, 1979). A variety of other pheno-mena of regeneration, metaplasia and transdifferentiation of eye tissues couldalso be explained if the same situation applies in vivo (see reviews by Clayton,1978, 1979; Coulombre, 1965; Eguchi, 1979; Lopashov, 1963; Okada, 1976;Yamada, 1976.

In Wolffian lens regeneration in lentectomized newt eyes, Yamada (1976)reported that six cell divisions are required for the irreversible metaplasia ofpigment epithelium into lens. In mass cultures of NR from 8- to 9-day chickenembryos, very little or no mitosis seems to occur in the differentiation of lenscells (although the evidence is indirect, being based on counts of other celltypes), but crowding does seem to be essential (see Pritchard et al. 1978). Incontrast, during pigment-cell formation, cell counts indicate that between threeand five mitoses occur (see Fig. 1C-E). Of these the first one or two require aninactive TCA cycle, whereas the remainder require TCA-cycle activity. Mitosiswould seen necessary for reduction of cell volume, but as yet there is no evidencethat it is a prerequisite for melanin synthesis. Cessation of mitosis probably isessential for melanin synthesis (Whittaker, 1974).

The most significant outcome of this work is the discovery that imposition ofa defined, but atypical physiological regime upon a vertebrate cell type cancause it to change to a recognizably different one. Other authors have shownthat such changes involve modifications in the patterns of gene expression(Itoh et al, 1975; Okada et al 1975; de Pomerai, Pritchard & Clayton, 1977;Thomson, de Pomerai, Jackson & Clayton, 1979), possibly by amplifying low-level syntheses of mRNAs characteristic of the new cell types (Clayton, 1978,1979; Jackson et al 1978). Examples of the control of gene expression by simplephysiological stimuli are rare in eukaryotes. The identification in this paper ofan area of biochemistry significant in this system, opens the way for a detailedanalysis of the intra- and extracellular biochemical conditions which initiateand stabilize new patterns of gene activity during differentiation of eukaryotecells. Work in progress in this laboratory is directed toward this goal.

Transdijferentiation of chick embryo neural retina 61It is with pleasure that I acknowledge the encouraging interest in this work shown by my

friends, relatives and colleagues. In particular I wish to thank Professor D. F. Roberts,Dr S. S. Papiha, and my wife Penny, for reading and criticizing the manuscript. Thanks arealso due to Mr M. Booth and Miss F. Behjati for technical assistance, and to Mrs P. Dun-woodie, Miss S. Mitchinson and Mrs D. Towell for the typing.

REFERENCES

ARAKI, M. & OKADA, T. S. (1978). Effects of culture media on the 'foreign' differentiationof lens and pigment cells from neural retina in vitro. Devi. Growth & Differ. 20, 71-78.

CLAYTON, R. M. (1978). Divergence and convergence in lens cell differentiation: regulationof the formation and specific content of lens fibre cells. In Stem Cells and Tissue Homeo-stasis (ed. B. Lord, C. Potten & R. Cole). Cambridge University Press.

CLAYTON, R. M. (1979). Genetic regulation in the vertebrate lens cell. In Mechanisms of CellChange (ed. J. Ebert & T. S. Okada). New York: Wiley.

CLAYTON, R. M., DE POMERAI, D. I. & PRITCHARD, D. J. (1977). Experimental manipulationof alternative pathways of differentiation in cultures of embryonic chick neural retina.Devi. Growth and Differ. 19, 319-328.

COULOMBRE, A. J. (1965). The eye. In Organogenesis (ed. R. L. de Haan & H. Ursprung),pp. 219-251. New York, Chicago, San Francisco, Toronto, London: Holt, Rinehart andWinston.

CRANE, R. K. & BALL, E. G. (1951 a). Factors affecting the fixation of C14O2 by animaltissues. J. biol. Chem. 188, 819-832.

CRANE, R.K. & BALL, E.G. (19516). Relationship of C14O2 fixation to carbohydratemetabolism in retina. / . biol. Chem. 189, 269-276.

EAGLE, H. (1959). Amino acid metabolism in mammalian cell cultures. Science, N.Y. 130,432-437.

EARLE, W. R. (1943). Production of malignancy in vitro. IV. The mouse fibroblast andchanges seen in the living cells. J. Natn. Cancer Inst. 4, 165-212.

EGUCHI, G. (1979). 'Transdifferentiation' in pigmented epithelial cells of vertebrate eyes invitro. In Mechanisms of Cell Change (ed. J. Ebert & T. S. Okada). New York: Wiley.

GRAYMORE, C. N. (1969). General aspects of the metabolism of the retina. In The Eye, vol. 1,Vegetative Physiology and Biochemistry (ed. H. Davson), second edition. New York andLondon: Academic Press.

ITOH, Y. (1976). Enhancement of differentiation of lens and pigment cells by ascorbic acidin cultures of neural retinal cells in chick embryos. Devi Biol. 54, 157-162.

ITOH, Y., OKADA, T. S., IDE, H. & EGUCHI, G. (1975). The differentiation of pigment cellsin cultures and chick embryonic neural retinae. Devi. Growth and Differ. 17, 39-50.

JACKSON, J. F., CLAYTON, R. M., WILLIAMSON, R., THOMSON., I., TRUMAN, D. E. S. & DEPOMERAI, D. I. (1978). Sequence complexity and tissue distribution of chick lens crystallinmRNAs. Devi. Biol. 65, 383-395.

LEHNINGER, A. L. (1972). Biochemistry. New York: Worth.LOEWENSTEIN, W. R. (1968). Emergence of order in tissues and organs. Communication

through cell junctions, implications in growth control and differentiation. Devi Biol.Suppl. 2, 151-185.

LOPASHOV, G. V. (1963). Developmental Mechanisms of Vertebrate Eye Rudiments. Trans-lated by J. Medawar. Oxford, etc.: Pergamon Press.

NOMURA, K. & OKADA, T. S. (1979). Age-dependent change in the transdifferentiationability of chick neural retina in cell culture. Devi. Growth and Differ. 21, 161-168.

OKADA, T. S. (1976). Transdifferentiation of cells of specialised eye tissues in cell culture. InTests of Teratogenicity In Vitro, pp. 91-105. Amsterdam: North-Holland.

OKADA, T. S. (1977). A demonstration of lens-forming cells in neural retina in clonal cellculture. Devi. Growth and Differ. 19, 47-55.

OKADA, T. S., ITOH, Y., WATANABE, K. & EGUCHI, G. (1975). Differentiation of lens in culturesof neural retinal cells of chick embryos. Devi Biol. 45, 315-329.

3 EMB 62

62 D. J. PRITCHARD

OKADA, T. S., YASUDA, K., ARAKI, M. & EGUCHI, G. (1979). Possible demonstration ofmultipotential nature of embryonic neural retina by clonal cell culture. Devi Biol. 68,600-617.

DE POMERAI, D. I. & CLAYTON, R. M. (1978). Influence of embryonic stage on the trans-differentiation of chick neural retina cells in culture. / . Embryol. exp. Morph. 47, 179-193.

DE POMERAI, D. I., PRITCHARD, D. J. & CLAYTON, R. M. (1977). Biochemical and immuno-logical studies of lentoid formation in cultures of embryonic chick neural retina and day-oldchick lens epithelium. Devi Biol. 60, 416-427.

PRITCHARD, D. J., CLAYTON, R. M. & DE POMERAI, D.I. (1978). 'Transdifferentiation' ofchicken neural retina into lens and pigment epithelium in culture: controlling influences./ . Embryol. exp. Morph. 48, 1-21.

PRITCHARD, D. J. & IRELAND, M. J. J. (1977). A method for relocation of specified regions intissue culture dishes. Experientia 33, 1120.

THOMSON, I., DE POMERAI, D. I., JACKSON, J. F. & CLAYTON, R. M. (1979). Lens-specificmRNA in cultures of embryonic chick neural retina and pigmented epithelium. Expl CellRes. 122, 73-81.

WHITTAKER, J. R. (1974). Aspects of differentiation and determination in pigment cells. InConcepts of Development (ed. J. Lash & J. R. Whittaker), pp. 163-178. Stamford, Connecti-cut: Sinauer Associates Inc.

YAMADA, T. (1976). Dedifferentiation associated with cell-type conversion in newt lensregenerating system: a review. In Progress in Differentiation Research (ed. N. Muller-Berat et al.), pp. 355-360. Amsterdam: North-Holland.

(Received 30 June 1980, revised 20 October 1980)