Cloned mouse melanocyte lines carrying the germline mutations albino and
brown: complementation in culture
DOROTHY C. BENNETT, PHILIP J. COOPER, TIM J. DEXTER, LINDA M. DEVLIN, JANET
HEASMAN and BARRY NESTER
St George's Hospital Medical School, Department of Anatomy, Cranmer Terrace, London SW17 ORE, UK
Summary
We have established two new immortal lines of mousemelanocytes, melan-b and melan-c, from mice homo-zygous for the brown (b) and albino (c) mutationsrespectively. Both lines were derived through differen-tiation in vitro of embryonic epidermal melanoblasts.The brown melanocytes are visibly brown by lightmicroscopy, and centrifuged cell suspensions formbrown pellets. The albino melanocytes form white pelletsand contain abundant unpigmented premelanosomes asshown by transmission electron microscopy. Like nor-mal, non-immortal melanocytes and like the immortalblack melanocyte line melan-a, both lines show little orno growth in a standard, serum-supplemented medium,but proliferate well in the presence of 12-o-tetradecanoylphorbol-13-acetate (TPA). Sustained growth of the al-bino cells also requires either keratinocyte feeder cells or
2-mercaptoethanol (2-ME). The modal chromosomenumbers are 39 for melan-b and 40 (diploid) for melan-c. Neither line is tumorigenic in nude mice. Hetero-karyons between the two lines can be constructed andform wild-type, black pigment. Melanocyte lines cannow be reproducibly generated from mice of differentstrains, and provide tools for molecular studies ofgermline coat-colour mutations. These two lines provideelegant means to study the developmentally controlledexpression of the two complementary genes, B and C,with black melanin pigment as a readily detectablenatural marker.
Coat-colour mutations account for over 60 of the knowngenetic loci of the mouse, and include mutations actingat various stages of development from the neural creststage onwards (Silvers, 1979). The mechanisms ofaction of many of these mutations are unknown, butprogress in understanding them has already been ac-celerated by the development of techniques for theselective cell culture of mouse neural crest (Ito &Takeuchi, 1984), mouse melanoblasts (Mayer & Oddis,1977; Mayer, 1982; Bennett et al. 1987) and postnatalmouse melanocytes (Sato et al. 1985; Tamura et al.1987). Melanoblasts or melanocytes can now beselected readily in primary cultures by the use of mediasupplemented with TPA and cholera toxin, as originallyused for human melanocytes (Eisinger & Marko, 1982).
A further technical development, which will greatlyfacilitate biochemical and molecular-genetic studies, isthe isolation of immortal lines of mouse melanocytes.Two clonal lines have been derived from black (a/a)mice (Sato et al. 1985; Bennett et al. 1987). Othermelanocyte lines have been established from the dermis
of mice carrying mutations including several at thealbino (c) locus, although these lines were not clonedand no evidence was presented that the lines containedno other dermal cells such as fibroblasts (Tamura et al.1987; Halaban et al. 1988). Albino melanocyte lineswere also reported in an abstract (Abe & Takeuchi,1985). The C locus is thought to code for the pigment-synthesizing enzyme tyrosinase (Silvers, 1979; Kwon etal. 1987; Halaban et al. 1988). Here we describe theestablishment of two new cloned melanocyte lineshomozygous respectively for albino (c/c) and for brownip/b), another locus at which the expression appears tobe developmentally controlled and melanocyte-specific. A tyrosinase-related protein, possibly anotherenzyme or a melanosomal structural protein, maps at ornear the B locus (Jackson, 1988).
We also describe somatic complementation betweenthe b and c mutations in heterokaryons.
Materials and methods
MaterialsTissue-culture media and plastics (Nunc) were obtained from
380 D. C. Bennett and others
Gibco Europe (Uxbridge, UK), foetal calf serum from TissueCulture Services (Slough, UK) and cholera toxin fromSchwarz-Mann (Orangeburg, NY, USA). Polystyrene latexbeads, phytohaemagglutinin (PHA/P), TPA, mitomycin C,2-ME and 4-norleucine, 7-D-phenylalanine-a--melanocytestimulating hormone (N-MSH) were from Sigma ChemicalCo. (Poole, UK). TPA, cholera toxin and N-MSH weredissolved and stored as previously described (Bennett et al.1985). Polyethylene glycol (PEG) 1540 was from Koch-Light(Haverhill, Suffolk, UK).
MediaThe basic culture medium was a supplemented Eagle'sminimal essential medium (SMEM) containing penicillin,streptomycin, sodium pyruvate and nonessential amino acids(Kreider et al. 1975); and prepared with only 25mM-sodiumbicarbonate as previously described (Bennett et al. 1987) togive a pH of 6-9 with 10% CO2. Ham's F10 medium wassupplemented with 18 mM bicarbonate to give the same pH.
AnimalsAlbino embryos were obtained from a female mouse of theoutbred LAC-MF1 strain (B/B,c/c), which was maintained atSt George's Hospital Medical School. Brown embryos werefrom a mating of a brown male and female from the partiallyinbred 'Q' strain, maintained at the MRC Mammalian Devel-opment Unit, London NW1, and the pregnant female waskindly donated by Dr Ian Jackson. Random-bred nude(nu/nu) mice were maintained at the Imperial CancerResearch Fund Animal Breeding Unit, South Mimms.
Keratinocyte feeder cellsThe XB2 immortal mouse keratinocyte line (Rheinwald &Green, 1975) was kindly provided by Dr J. G. Rheinwald andwas adapted in our laboratory to grow in the absence of 3T3feeder cells, in Dulbecco's modified Eagle's medium (DEM)and 10% fetal calf serum (FCS). To prepare keratinocytefeeder cells, nearly confluent cultures of XB2 cells on 85 mmplates were growth-inactivated by incubation for 2h in 5 mlDEM and 10% FCS containing 4^gml~1 mitomycin C, thenwashed in DEM and incubated for lOmin in fresh DEM+FCSwithout mitomycin. They were then subcultured and eitherreplated at 3xlO4 cells ml"1 or frozen in liquid nitrogen andreplated when needed at 5xl04ml~1 (allowing for reducedviability).
Primary cultures of melanoblastsThese were prepared essentially as described previously(Mayer & Oddis, 1977; Bennett et al. 1987). The albinoembryos were 16-5 days old and the brown embryos were 18-5days old. In brief, the epidermis was separated with trypsin,washed in medium (SMEM+10% FCS), minced in a drop ofmedium using two scalpels, and pipetted vigorously in about2ml medium with a siliconized Pasteur pipette. The suspen-sion was suitably diluted with the same medium, then sup-plemented with cholera toxin (lOnivi) and N-MSH (100 pM)and plated into 50 mm culture plates containing XB2 feedercells, at 4 ml medium per dish. The cultures were incubated at37°C with 10 % CO2. TPA (200 nin) was added after 3-4 days.The medium was initially changed twice weekly, but this wasadjusted according to the number of cells present.
The most usual medium during the early passages wasSMEM with 10% FCS, TPA (200 nM), N-MSH (100 pM) andcholera toxin (10 nM, reduced to lnM after the first 1-2passages). Preliminary work has indicated that MSH increasesthe yield of diploid mouse melanocytes (unpublished data).With immortal mouse melanocytes, MSH stimulates both
proliferation and pigment synthesis (Tamura et al. 1987 andour unpublished data), but this is not known for diploid cells.In our current method for culture of diploid melanocytes, theconcentration of FCS has been reduced to 5 % throughout.For subculture, cell suspensions were made as for the estab-lished lines (see below), but were replated on to fresh XB2feeder cells.
Culture of established linesThe final growth medium used for both lines was SMEM with5% FCS, TPA (200nM), and 2-ME (100^M). Both weresubcultured approximately every 7-10 days, by the proceduredescribed previously for melan-a cells (Bennett et al. 1987).Cells were replated at 2-3xl04ml"L, 10 ml per 85 mm dish, ingrowth medium without feeder cells.
Frozen cell stocks in liquid nitrogen were prepared bystandard procedures except that the freezing medium wasSMEM with 5% FCS, 100//M-2-ME and 7-5% dimethylsulphoxide (DMSO), and care was taken to avoid a rise in pHthrough loss of CO2. We now use this method for allmelanocyte cultures.
Cloning of melanocytesA cell suspension was prepared as for subculture. Cells wereplated, either by limiting dilution or by manual selection witha drawn-out Pasteur pipette and mouth tube, into 6-mmculture wells containing XB2 feeder cells. The cloning me-dium was 50 % SMEM and 50 % Ham's F10 medium, pH6-9(Materials and methods), with 5 % FCS, conditioned for 1 dayby the feeder cells and then supplemented with TPA andcholera toxin. This method gave consistently high cloningefficiencies of up to 80 %.
Light microscope photographyCultures were washed in Dulbecco's phosphate-bufferedsaline lacking calcium and magnesium chlorides (PBSA),fixed in 4% formalin in PBSA and washed with water. Theywere photographed with water in the dishes, using an Olym-pus inverted microscope with planachromat objectives.
Electron microscopyCell suspensions were prepared as above from two confluent85 mm plates of melan-c cells at passage 26. The cells werecentrifuged, washed with PBSA and resuspended in fixative:2% glutaraldehyde in 0-lM-cacodylate buffer, pH7-2. Theywere fixed at refrigerator temperature for 1 h, rinsed twice incacodylate buffer, centrifuged at 4000 revs min"1 and the cellpellet postfixed for lh in osmium tetroxide. The pellet wasdehydrated and embedded in Araldite. Thin sections werecut, and examined and photographed with a Philips EM301electron microscope.
Tumorigenicity testsMelan-b and melan-c cells were harvested as above, washed inPBSA and resuspended at the required concentrations inPBSA. They were injected in 0-2 ml volumes into female,thymus-deficient nude (nu/nu) mice aged 16 weeks. 10 micereceived 2xlO6 cells subcutaneously in the flank region and 5received 106 cells intravenously in the tail vein.
Preparation of heterokaryonsPolystyrene latex beads, diameter ljum, were suspended andstored at 4xl09ml~1 in 70% ethanol, and were then pre-sumed to be sterile. Melan-c cells were labelled by growth for10 days in the presence of beads (2xl07ml~1 medium).Suspensions of melan-b and bead-labelled melan-c cells in
•2
4 'r •
1A
B
Fig. 1. Diploid melanocytes and melanoblasts in a primaryculture. These are black (a/a) cells, as a generalillustration. (A) Phase-contrast; (B) bright-field optics. Themelanin pigment is seen clearly in B. The cells can bearranged along a morphological spectrum, as exemplified bythe series 1—4, so that the cells with appearance 1 arepresumed to be melanoblasts. These emerge and proliferatefrom explants before melanocytes are seen. Themelanoblasts are very small, bipolar and/or crescent-shapedwith a ruffled membrane on the convex side. Very fewother cells (primary or XB2 keratinocytes or fibroblasts) arepresent. Scale bar, 200 fim.
14 A \ ^ ^*37&i
Fig. 2. Cloned, immortal melanocytes by phase-contrastoptics. (A) melan-a, passage 17; (B) melan-b, passage 10;(C) melan-c, passage 13. At this magnification, melan-acells show abundant pigment granules, melan-b cells havefine, inconspicuous granules and melan-c do not appeargranular. Bar, 200 fim.
Fig. 3. Pelletted melanocytes. From left to right: melan-a,melan-b and melan-c (10 cells each). The cells were fixedin 4% formalin in suspension before centrifugation.
complete medium were prepared as above and plated on 33-mm dishes at 105mr', 2 ml per dish, both separately (con-trols) and in a 1:1 mixture. The cells were allowed to attachand spread, then fused in situ by the following series of washes(lml per dish): (1) Three washes in serum-free SMEM; (2)PHA/P, 100f<gml"1 in serum-free SMEM, 15 min; (3) 10%DMSO in SMEM, 10s; (4) PEG 1540, 45% w/v in SMEMwith 10% DMSO, final pH approximately 7-6, 1 min; (5)10 % DMSO in PBSA, 10s; (6) SMEM+1-8 mM-EGTA, a fewseconds, and (7) complete growth medium+1-8 mM-EGTA,20 min in incubator. See Shay (1982) for reviews on the use ofbead-labelling, PHA, DMSO and EGTA in PEG-mediatedcell fusion. PHA and DMSO increase efficiency of fusion,while EGTA increases viability.
Results
Derivation of melan-bPrimary cultures were plated in February, 1986. Many
Cloned albino and brown mouse melanocyte lines 381
20
4Fig. 4. Electron micrographs of melan-c cells. (A) Part of atypical cell showing many premelanosomes in the peripheralcytoplasm. They are the pale organelles containinglongitudinal filaments. (B) Premelanosome at highermagnification showing transverse striations. Bars, (A)500 nm and (B) 100 nm.
Fig. 5. Proliferation of melan-c cells in media with andwithout 2-ME or heat-inactivated serum. SMEM wassupplemented with 5 % FCS which was either untreated orheated for 1 h at 56°C. 2-ME was added to media the daybefore use to allow equilibration. All media contained TPA(200 nM). A melan-c cell suspension was prepared as forsubculture in medium with 1 % FCS, diluted into each testmedium at 2xlO4 cells ml"1, plated at 2ml per 33 mm dishand incubated. On the specified days, cells wereresuspended and counted by haemocytometer. The mediawere renewed after 4 days. Each point show the mean andstandard error of 6 counts, 2 from each of 3 dishes. Key:• , SMEM+untreated FCS; • , same+2-ME, 100^M;D, SMEM+FCS to which 100/iM-2-ME had been added,giving a final concentration of 5^M-2-ME, V, SMEM+heated FCS.
groups of unpigmented melanoblasts were seen within afew days; as usual these proliferated and matured topigmented melanocytes over the first few weeks, afterwhich no melanoblasts were detected. The epidermalkeratinocytes failed to proliferate in the selective me-dium, and these and the feeder cells died out over about2 weeks. Typical melanoblasts and melanocytes areshown for reference in Fig. 1; these are from black(a I a) mice and therefore resemble wild-type cells, asmutations at the A locus do not directly affect melano-cytes (Silvers, 1979). Living b/b melanocytes, wherelocally crowded, were visibly brown by bright-fieldmicroscopy; elsewhere the pigment was very faint.
The cells were subcultured after 3 weeks, and in total8 times over the first 7 months. Proliferating melano-cytes could be observed throughout this period, asjudged by the presence of colonies of small, healthy-looking, pigmented cells with some mitoses (see Ben-nett et al. 1987 for an illustration of pigmented melano-cytes in mitosis). However, growth was balanced bycellular senescence and death so that the number ofcells showed no net increase during this time. Fibro-blast-like cells also grew but did not outgrow themelanocytes so long as cholera toxin was present.
382 D. C. Bennett and others
It Rt,
* • '
• -*., \ B
\
V 4?
\ • \
From passage 8, melanocyte growth accelerated andXB2 feeder cells and N-MSH were now omitted. Atpassage 19 and 10-5 months of culture, single-cell cloneswere prepared, by limiting dilution (Materials andmethods). Numerous proliferating clones were ob-tained. Two pigmented melanocyte clones wereselected, from wells containing no unpigmented (fibro-
blast-like) cells, and subcultured. They now grew welland growth was not noticeably impaired by simplifi-cation of the medium to SMEM, 5 % FCS, 200 nM-TPAand 100^IM-2-ME. 2-ME was now added because ofresults obtained with the albino cell line (see below).Frozen stocks from both clones were prepared and onewas selected for characterization and named 'melan-b'.
Cloned albino and brown mouse melanocyte lines 383
Fig. 6. Fused melanocytes. Cultures of melan-b, melan-c ora 1:1 mixture of both were plated at 105 cells ml"1 andfused in situ as described (Materials and Methods). Thecultures were then incubated for 3 days in growth mediumwith 1 nM-N-MSH. MSH was included in case it wouldincrease the rate of pigment synthesis as it does inmelanoma cells (reviews: Pawelek, 1979; Bennett, 1988)and in melan-a melanocytes (unpublished data). In eachcase, the same field is shown by phase-contrast (left) andbright-field optics. Care was taken that photographicprocedures were identical in all cases. All cultures containdead cells killed by the fusion treatment. (A,B) melan-bonly. Only a faint granulation is visible in either mono- ormultinucleate living cells. Dead, rounded-up cells andfragments look more granular; this was a consistentproperty of melan-b cells. (C,D) melan-c only. Polystyrenebeads in living and dead cells appear prominent and darkby bright-field or refractile by phase-contrast (arrowheads);no pigment is present. (E-H): melan-b-l-melan-c, twofields. Many cells in these cultures contained heavy, darkpigment like that of black melanocytes. Such cells weretypically giant and tortuous in shape, and contained beads(arrowheads); they were thus presumably hybrid cells. Scalebar, 200 [im. Inset in G and H: part of cell on left at 3xhigher magnification, for clearer images of beads (bright byphase contrast) and melanin.
Since the cells had emerged from a phase of senescencethey were presumed to be established and immortal(Pollack, 1981).
Derivation of melan-cTo identify and purify albino melanocytes, as they lackpigment, we could use only morphological criteriatogether with the knowledge that an almost-pure mel-anocyte population is normally selected under theconditions used. Cells which were small, elongated,epithelioid to dendritic and having nuclei which weredark by phase-contrast optics (Bennett et al. 1987) wereprovisionally taken to be melanocytes.
The primary culture (May, 1986) was passaged after 3weeks and the resulting plates were kept for 7 weeks,after which large, healthy, melanocyte-like colonieswere observed. One of these colonies, assumed to be aclone, was subcultured by local application of trypsinand EDTA solution into two 16 mm wells containingXB2 feeder cells. This was counted as passage 2. Noflbroblast-like cells were observed in this culture sub-sequently, but net growth was very poor, largely be-cause many cells died after subculture or addition offresh medium, even though feeder cells were present.The total cell number increased only slightly during afurther 7 subcultures and 5 months. At this pointimproved growth was observed on reducing the serumconcentration from 10 % to 5 %, suggesting an inhibi-tory effect of FCS. Therefore some cultures weretransferred to medium with the reducing agent 2-ME(100/M), which might alter some serum components.2-ME is used in the culture of other cell types such ashybridomas (Langone & Van Vunakis, 1986) and earlymouse embryo cells and embryonal carcinoma cells,which it enables to grow in the absence of feeder cells
(Smith & Hooper, 1987). A marked effect was noticed:cell death ceased immediately and within a few daysthere were substantially more cells than in controlcultures. Quantitative studies at a later passage con-firmed this effect (see below).
From passage 9, 5% FCS and 2-ME were usedroutinely and the cells grew progressively. Feeder cellscould now be omitted. Frozen stocks were preparedfrom passage 12 onwards. The cells now behaved as anestablished line and were named 'melan-c'. Subclonesof the line have also been prepared (Materials andmethods).
General characteristics of melan-b and melan-c cellsAll information from this section onwards relates to theestablished lines grown in the absence of feeder cells.
Fig. 2 shows the appearance of the two lines by lightmicroscopy, with melan-a melanocytes for comparison.All three populations have a similar appearance, exceptthat melan-a cells are less aligned than the others andmelan-b cells are relatively flat. Melan-c cells alsobecame flatter at later passages. Melan-b cells have onlylight pigmentation, and melan-c cells are unpigmentedalthough granular at very high magnifications (notshown). The granules are probably premelanosomes(see below). The pigmentation of melan-b cells isvisibly brown, both by light microscopy of densecultures and as seen macroscopically in pelleted cells(Fig. 3). This is different from a low density of blackpigment, which gives a grey cell pellet.
Both lines grow slowly in the earlier passages, withpopulation doubling times of approximately 2-5-3 daysas estimated from cell yields at subculture. As withmelan-a (Bennett et al. 1987), the growth rate increasesgradually with passage number.
Ultrastructure of melan-c cellsAlbino melanocytes in vivo produce premelanosomes,the unpigmented precursors of melanosomes (pigmentorganelles) (Parakkal, 1967; Hearing et al. 1973). Theseorganelles are unique to pigment cells and are readilyidentifiable by electron microscopy. Ultrastructuralexamination of melan-c cells revealed abundant pre-melanosomes, especially in the peripheral cytoplasm asexpected (Fig. 4). Most of them were classifiable asstage II melanosomes, which contain longitudinal fila-ments with cross-striations (Quevedo et al. 1987).
Growth responses to 2-ME and TPAThe effects of 2-ME on the growth of melan-c cells wereretested at later passages. Fig. 5 shows a typical exper-iment, in which the effect of heat-inactivation of theFCS was also tested. Heated FCS inhibited prolifer-ation, but 2-ME (100 ;UM) produced a marked increasein cell yield over 8 days. Addition of 2-ME at 100 /AM tothe FCS before use was less stimulatory (Fig. 5). Fromthese observations, it is unlikely that the favourableaction of 2-ME was due to inactivation of complement,or any other effect on the serum. When the concen-tration of 2-ME was varied, the best growth wasobtained with concentrations between 30 and 300 JUM
384 D. C. Bennett and others
(further data not shown), so routine use at 100^M wascontinued. A small, but statistically significant, im-provement in the growth of melan-b cells (about 40 %more cells in 6 days) was obtained with the samesupplementation.
Both cell lines grew poorly in the absence of TPA. Ina typical experiment the numbers of melan-c cellsobtained after 7 days' growth in different media were asfollows (mean and standard error in units of104 cells ml"1; other details as for Fig. 5): completegrowth medium (GM), 16-19 ±0-86; GM minus ME,6-38 ± 0-37; GM minus TPA, 4-87 ± 0-20; GM minusME and TPA, 2-24 ± 0-38. Both melan-b and melan-ccells became flattened and polygonal instead ofelongated when TPA was not included, as previouslydescribed and illustrated for melan-a cells (Bennett etal. 1987). Diploid, non-immortal mouse and humanmelanocytes show a similar proliferative and morpho-logical response to TPA, which all three cell lines havethus retained.
KaryologyChromosome spreads (Rothfels & Siminovitch, 1958)were prepared from melan-b cells at passage 15 andmelan-c cells at passage 17. The modal chromosomenumbers were 39 for melan-b (30/50 cells) and 40(diploid number) (38/50 cells) for melan-c. A subpopu-lation of 10/50 melan-b cells (20%) had 40 chromo-somes.
Tumorigenicity testsAs neither line was from fully inbred mice, theirtumorigenicity was tested in immunodeficient (nude)mice (Materials and methods). Neither line formed anytumour or transient nodule in any of 15 injected miceper cell line. The mice were killed after 6 months, whenhistological examination of the injected sites, in thoseanimals that had received subcutaneous injections,revealed no evidence of tumour initiation.
Somatic complementation testThe b and c mutations are recessive and complementeach other in heterozygotic animals. Therefore it isreasonable to expect complementation in hetero-karyons between these b and c melanocytes in vitro,given that melan-c cells are from MF1 (B/B) mice.Expression of a gene believed to be the B gene islimited to melanocytes and melanoma (Jackson, 1988),and this specificity is also indicated by the fact thatmutations at the b locus do not appear to affectanything but pigmentation in mice (Silvers, 1979). Thuscomplementation would also be an indication of nor-mal, melanocytic differentiation in melan-c cells.
Melan-c cells were prelabelled with polystyrenebeads and fused to melan-b cells with PEG 1540,PHA/P and DMSO (Materials and methods). 3 dayslater the cultures were examined. The results are shownin Fig. 6. Fused (and other) cells in the control culturesof melan-c or melan-b alone had no pigment or theinconspicuous brown pigment, respectively, but numer-ous cells in the mixed cultures contained conspicuous
pigment similar to that of black melanocytes (seeBennett et al. 1987). All such dark cells containedbeads, providing evidence that they were hybrid cells.To assess the proportion of heterokaryons that ex-pressed black pigment, a count was made of theproportion of cells with prominent pigment out of 200cells with beads, pigment (faint or prominent) and morethan 1 nucleus. The ratio was 182/200 or 91 %. In fusedmelan-b cultures, the corresponding proportion of cellswith two or more nuclei that had prominent pigmentwas 0/200. The 9% of apparent heterokaryons thatcontained only faint pigment may alternatively havebeen brown homokaryons which had adsorbed orphagocytosed beads from dead melan-c cells, of whichmany were present. Thus somatic complementation wasobserved in at least 90% of heterokaryons.
Discussion
Proliferating mouse melanocyte cultures can readily bederived through differentiation in vitro of embryonicmouse melanoblasts. From such cultures, we haveestablished two new melanocyte lines carrying homo-zygous recessive mutations which produce clearlyobservable phenotypes in the cells. The methods de-scribed yield immortal melanocyte lines at a highfrequency (approximately one in two attempts). Theisolation of further mutant lines is in progress. Amethod for cloning mouse melanocytes is also de-scribed. The cloning efficiency is high and furthersubclones of both lines have been readily obtained.
Several observations identify the unpigmentedmelan-c cells firmly as melanocytes. The most conclus-ive is the presence of many premelanosomes in thecytoplasm. In addition, the cells contain the B gene inan expressible state - not, for example, repressed bymethylation - as shown by somatic complementation;they grow well in the presence of TPA and cholera toxinand very slowly in the absence of TPA; they originatefrom epidermis and they have the correct morphology.
The reducing agent 2-ME proved essential for theestablishment of the melan-c line, which still growspoorly without it; however, its stimulation of the growthof melan-b cells is slight and preliminary tests onprimary melanoblast/melanocyte cultures indicate aninhibition of growth (unpublished data). The reason isunknown, but 2-ME should evidently be used withcaution in pigment cell culture.
The melan-b and melan-c lines will be useful asfurther lines of non-tumorigenic melanocytes, for ex-perimental comparisons with melanoma cells. More-over, they provide excellent vehicles for the study ofmammalian gene expression. We have shown that eachof the two mutations can be directly complemented byfusion with a melanocyte carrying the wild-type gene.We have now obtained efficient transfection and inte-gration of exogenous DNA carrying drug-resistancemarkers, with both melan-b and melan-c cells. Theselines thus permit gene transfer experiments aimed atthe identification of the B and C genes, which appear to
Cloned albino and brown mouse melanocyte lines 385
be developmentally controlled. Putative sequences forboth these genes have recently been cloned (Shibaharaet al. 1986; Kwon et al. 1987; Yamamoto et al. 1987;Jackson, 1988; reviewed by Bennett, 1988), includingone sequence asserted to be either the tyrosinase gene(Shibahara et al. 1986), or a sequence at or near the Blocus (Jackson, 1988). Once genomic B and C se-quences are identified, it will be possible to study theexpression of manipulated constructs in melan-b andmelan-c cells, using the very prominent natural marker,black melanin.
We are most grateful to Dr Ian Hart of the Imperial CancerResearch Fund for carrying out the tumorigenicity tests, andto Drs Ian Jackson and Ruth Halaban for communication ofunpublished work. This research was supported by the CancerResearch Campaign and a Wellcome Vacation Scholarshipand Foulkes Foundation Fellowship to P.J.C.
References
ABE, H. & TAKEUCHI, T. (1985). Establishment of albinomelanocyte cell lines. Zool. Sci. 2, 900.
BENNETT, D. C. (1988). Mechanisms of differentiation in melanomacells and melanocytes. In Regulation of differentiation ineukaryotic cells (ed. A. M. Jetten). Environmental HealthPerspectives 80 (In press).
BENNETT, D. C , BRIDGES, K. & MCKAY, I. A. (1985). Clonalseparation of mature melanocytes from premelanocytes in adiploid human cell strain: spontaneous and induced pigmentationof premelanocytes. J. Cell Sci. 77, 167-183.
BENNETT, D. C , COOPER, P. J. & HART, I. R. (1987). A line ofnon-tumorigenic mouse melanocytes, syngeneic with the B16melanoma and requiring a tumour promoter for growth. Int. J.Cancer 39, 414-418.
EISINGER, M. & MARKO, O. (1982). Selective proliferation ofnormal human melanocytes in the presence of phorbol ester andcholera toxin. Proc. natn. Acad. Sci. U.S.A. 79, 2018-2022.
HALABAN, R., MOELLMAN, G., TAMURA, A., KWON, B. S.,
KUKLINSKA, E., POMERANTZ, S. H. & LERNER, A. B. (1988).Tyrosinases of murine melanocytes with mutations at the albinolocus. Proc. natn. Acad. Sci. U.S.A. 85, 7241-7245.
HEARING, V. J., PHILLIPS, P. & LUTZNER, M. A. (1973). The fine
structure of melanogenesis in coat color mutants of the mouse.J. Ultrastnict. Res. 43, 88-106.
ITO, K. & TAKEUCHI, T. (1984). The differentiation in vitro of theneural crest cells of the mouse embryo. J. Embryol. exp. Morph.84, 49-62.
JACKSON, I. J. (1988). A cDNA encoding tyrosinase-related proteinmaps to the brown locus in mouse. Proc. natn. Acad. Sci. U.S.A.85, 4392-4396.
KREIDER, J. W., WADE, D. R., ROSENTHAL, M. & DENSLEY, T.
(1975). Maturation and differentiation of B16 melanoma cellsinduced by theophylline treatment. J. natn. Cancer Inst. 54.1457-1467.
KWON, B. S., HAQ, A. K., POMERANTZ, S. H. & HALABAN, R.
(1987). Isolation and sequence of a cDNA clone for humantyrosinase that maps at the mouse c-albino locus. Proc. natn.Acad. Sci. U.S.A. 84, 7473-7477.
LANGONE, J. J. & VAN VUNAKIS, H., eds (1986). ImmunochemicalTechniques Part I: Hybridoma Technology and Monoclonalantibodies. Methods Eniymol. 121. New York: Academic Press.
MAYER, T. C. (1982). The control of embryonic cell proliferation inculture by cyclic AMP. Devi Biot. 94, 509-514.
MAYER, T. C. & ODDIS, L. (1977). Pigment cell differentiation inembryonic mouse skin and isolated epidermis: an in vitro study.J. exp. Zool. 202, 415-424.
PARAKKAL, P. F. (1967). Transfer of premelanosomes into thekeratinizing cells of albino hair follicles. J. Cell Biol. 35,473-477.
PAWELEK, J. (1979). Melanoma cells. Methods Enzymol. 48,564-570.
POLLACK, R. ed. (1981). Readings in Mammalian Cell Culture. NewYork: Cold Spring Harbor Laboratory.
QUEVEDO, W. C . , FlTZPATRICK, T. B . , SZABO, G. & JlMBOW, K.
(1987). Biology of melanocytes. In Dermatology in GeneralMedicine (ed. T. B. Fitzpatrick, A. Z. Eisen, K. Wolff, I. M.Freedberg & K. F. Austen), 3rd edition, pp. 225-251. NewYork: McGraw-Hill.
RHEINWALD, J. G. & GREEN, H. (1975). Formation of akeratinizing epithelium in culture by a cloned cell line derivedfrom a teratoma. Cell 6, 317-330.
ROTHFELS, K. H. & SIMINOVITCH, L. (1958). An air-dryingtechnique for flattening chromosomes in mammalian cells grownin vitro. Stain Technol. 33, 73-77.
SATO, C , ITO, S. & TAKEUCHI, T. (1985). Establishment of a mousemelanocyte clone which synthesizes both eumelanin andpheomelanin. Cell. Struct. Funct. 10, 421-425.
SHAY, J. W. (ed.) (1982). Techniques in Somatic Cell Genetics. NewYork: Plenum Press.
SHIBAHARA, S., TOMITA, Y., SAKAKURA, T., NAGER, C , CHAUDHURI,
B. & MULLER, R. (1986). Cloning and expression of cDNA formouse tyrosinase. Nucleic Acids Res. 14, 2413-2427.
SILVERS, W. K. (1979). The Coat Colors of Mice. New York:Springer-Verlag.
SMITH, A. G. & HOOPER, M. L. (1987). Buffalo rat liver cellsproduce a diffusible activity which inhibits the differentiation ofmurine embryonal carcinoma and embryonic stem cells. DeviBiol. 121, 1-9.
TAMURA, A., HALABAN, R., MOELLMAN, G., COWAN, J. M.,
LERNER, M. R. & LERNER, A. B. (1987). Normal murinemelanocytes in culture. In Vitro Cell. Devi Biol. 23, 519-522.
YAMAMOTO, H., TAKEUCHI, S., KUDO, T., MAKINO, K., NAKATA,
A., SHINODA, T. & TAKEUCHI, T. (1987). Cloning and sequencingof mouse tyrosinase cDNA. Jpn. J. Genet. 62, 271-274.
