AU600393B2 - Generation of neural precursor cell lines - Google Patents

Description

~,~YW~i~l OF A U L R A L A9

COMMONWEALTH

PATENTS ACT 1952 COMPLETE SPECIFICATION (Original) FOR OFFICE USE Class Int. .Class Application Number: Lodged: Complete Specification Lodged: Accepted: Published: Priority: T is docunment contains the a i; idmnuts made under SSec:ion 49 and is correct for printing.

Related Art: Name of Applicant: Address of Applicant: AMRAD CORPORATION LIMITED.

0000 Do 0 C, a oc00 01 0 0 0 to 6 04 Q00a0 0 a4 a 0 a0 O CC (0(

((CC

C Ct

C

6/663 Victoria Road, Abbotsford, 3067, Victoria, Australia.

Actual Inventor(s): ORA BERNARD and PERRY FRANCIS BARTLETT Address for Service: DAVIES COLLISON, Patent Attorneys, 1 Little Collins Street, Melbourne, 3000.

Complete specification for the invention entitled: "GENERATION OF NEURAL PRECURSOR CELL LINES" The following statement is a full description of this invention, including the best method of performing it known to US -1- 1 "GENERATION OF NEURAL PRECURSOR CELL LINES" This invention relates to the generation of neuroepithelial and neural crest cell lines that are capable of differentiating in vitro into cells expressing neuronal and/or glial cell markers. In particular, the oo°o, 5 invention relates to the immortalisation of neural O0 precursor cells contained in neuroepithelium and neural 0 04 aeo crest cells by the introduction of members of the mry family of oncogenes in a retrovirus vector.

Whilst the detailed description herein relates 10 specifically to the generation of mouse neuroepithelial and neural crest cell lines, it will be understood that 4 the invention is not restricted to murine cell lines and that in fact it extends to cell lines of other mammalian S" species including humans.

o, 15 The two major cell types in the mammalian central nervous system, neurons and glia, are developmentally derived from the neuroepithelium that forms the neural ,o tube. Neuroepithelial cells have been shown to give rise 0 0 Sto both types of neural cells in yitro and precursor 20 cells committed to a particular cell lineage have been 2 identified within mouse neuroepithelial cells as early as embryonic day 10 The fact-rs and genes that regulate precursor cell differentiation in the mammalian nervous system are largely unknown, and the establishment of cell lines representative of these neural precursor populations would greatly assist in their identification.

Work leading to the present invention has included the generation of cell lines derived from mouse neuroepithelial cells by the introduction of the c-myc oncogene with a murine retrovirus delivery system using the pDol vector These cell lines have the characteristics and surface phenotype of freshly isolated neuroepithelial cells. They express cytokeratin but do not express neuronal or glial cell markers, and can be induced to express class I histocompatibility antigens upon stimulation with interferon-y (yIFN). While they do not differentiate spontaneously in vitro, exposure to basic fibroblast growth factor (bFGF) induces their o""ao differentiation into neurofilaments positive neurons and oso0 '2p9 glial fibrillary acidic protein (GFAP) positive glial o cells.

*O It has also been found that different types of oo* mouse neuroepithelial and neural crest cell lines can be generated by the introduction of a different recombinant retroviral vector bearing the N-myc, the c-my or the L-myc oncogenes and that some of these cell lines have the I capacity to differentiate spontaneously in vitro into neurons and/or glial cells. Many of these lines are factor dependent and can be used as target populations to rapidly screen for the potential neurotrophic factors.

These cell lines can also be used for the production of factors important for the maintenance of replication and t differentiation of cells in the central and peripheral nervous systems. Finally, these immortalised cell lines may also be used as model systems to study the possibility of using cell lines to restore damaged brain after an accident, stroke or in diseases such as Parkinson, Huntington, Alzeheimer, etc.

The myc proto-oncogene family has at least four different members: c-my, N-my, L-my (see reference 4 for review), and the recently cloned B myc The first three share the ability of being able to compliment a mutated ras gene in the transformation of primary rat embryo fibroblasts They are expressed in the developing brain, in E10 neuroepithelium and during embryogenesis, but no expression can be detected two weeks after birth Furthermore, N-myc is also expressed in neuroblastomas and is frequently amplified in these tumours L-myc is expressed in small cell carcinoma of the lung, however brain tumours expressing L-my, have not been identified.

The effects of retroviruses bearing the different oa' my c gene on differentiation of E10 neuroepithelial cells o. 20 has been investigated by the generation of immortalised mouse neuroepithelium cell lines using new retroviral vectors pZen and pZenSVNeo (11) bearing the c-mr,, the jo «N-myg or the L-myc proto-oncogenes. These zen retroviruses express higher levels of the inserted genes than the vector (pDol) used previously. The majority of the new cell lines spontaneously differentiate into mature i neural cells without the addition of exogenous growth I factors.

*r "According to the present invention, there is provided a method for the in vitro production of immortalised neural precursor cells, which comprises the S, step of infecting neuroepithelium or neural crest cells with a retroviral vector carrying one of the myc oncogenes.

Preferably the myQ oncogene carried by the retroviral vector is the c-myc, the N-myc or the L-myc oncogene.

The present invention also extends to immortalised or continuous cell lines which are neuroepithelial or neural crest cells which have been infected with a retroviral vector carrying one of the myc oncogenes.

In addition to immortalisation of the neural precursor cells, in certain instances the infection of these cells with a retroviral vector carrying one of the my oncogenes has been found to produce cell lines having neuronal and/or glial cell characteristics.

In particular embodiments of this invention, the pDol vector (17) has been used to construct a Dol(c-myq.) retrovirus which gives rise to continuous lines resembling freshly isolated neuroepithelium cells, and the pZen and pZenSVNae vectors (11) have been used to give Zen(myc) 0o" viruses (including c-myq, N-myc and L-myc) that give rise 20 to cells capable of differentiating in yitro.

o, Members of the myc family of oncogenes are not *got capable of directly transforming primary cells such as It $l mouse embryo fibroblasts but are capable of immortalising these cells. Normally, growth of mouse embryo fibroblasts Jn vitro is limited to about 60 generations, however, I after introduction of the c-myq oncogene the cells become immortalised and grow forever These immortalised cells have growth characteristics of normal cells and they do not develop into tumors when injected into nude mice This capacity of the myc oncogenes to immortalise cells has been used in the present invention to generate cell lines of embryonic neural cells.

t-- All the cells of the central nervous system (CNS) are derived from the neuroepithelium which forms the neural tube The cells of the peripheral nervous system are derived from the neural crest cells which migrate from the neural tube The neuroepithelial cells can give rise, in culture, to glial cells and to neurons but these cells survive in culture only for a very limited period of time Studies on early brain development have been difficult to perform but the use of cell lines represents a major leap forward in the technology to study these processes. In addition, isolation of homogeneous populations of cells from normal mouse brain are also very difficult to obtain in large numbers. Therefore the myc oncogenes have been used in order to generate continuous cell lines representing brain cells during different stages of brain differentiation.

The cell lines so obtained: 1. are immortalised; O.°o 2. have the capacity to differentiate in vitro into oo a 20 neurons or glial cells; and 3. have the capacity to respond to growth factors 0 0.0* such as basic fibroblast growth factor (bFGF) 0oo0 (16) and also growth factors produced by S°themselves and other neuroepithelium cell lines.

Further details of the present invention are set out in the following specific Examples. It will be appreciated that while the specific examples relate to work done in the mouse cells, the techniques and 0: procedures are equally applicable to human cells.

In the accompanying figures: igure 1 shows So schematic representation of recombinant retroviral S" plasmid used to produce the Dol(c-myc) retrovirus. The 0 a expression of the c-myc gene is controlled by the promoter and enhancer in the Moloney long terminal repeats (LTRs).

R

The Ne R gene is expressed through the simian virus early promoter and enhancer pBRori is the pBR322 origin of replication. R, EcoRI; B, BamHI: X,XbaI.

Southern blot analysis of DNA from neuroepithelium (lanes NE) and neuroepithelial cell lines (lanes 2.3Q, 2.3D.12, and 2.3D)(25). Cellular DNA (2pg) was digested with EcoRI (RI) or Xbal size-fractionated on 0.7% agarose, and transferred onto a Zeta-Probe membrane (Bio-Rad) in 0.4M NaOH. The filter was hybridised to 32 SP-labelled c-myc cDNA clone Hybridisation and washing conditions were as described (19a). Solid arrowheads mark the position of the endogenous c-my -bearing fragments, and the open arrowheads indicate the proviral c-rmy DNA bands. Sizes of fragments are in kilobase pairs. The intensities of the bands in various lanes vary because of unequal amounts of cDNA loaded onto the gel.

O Figure 2 shows 20 Expression of c-myQ in neuroepithelial cell lines.

Poly(A)+ RNA from several cell lines (indicated above the lanes) was fractionated on a formaldehyde/agarose gel and transferred to nitrocellulose filters for wit 32 hybridisation with P-labelled probe (19a). The c-myc probe was as above and the Ne1 probe was a 1.4-kilobase 1 HindIII-Sal I fragment from pSV2Neo Lines 2.3D.17 and 2.3.1.9 are clones isolated from 2.3D and 2.3S cell lines, respectively. Astroglia were from cultured populations of 19 day embryo cerebellar cells.

32 The same filter was hybridised to a P-labelled N.e probe. The full size viral RNA corresponds to the 6-kilobase transcript hybridising to the c-myc probe; the smaller 3-kilobase RNA corresponds to the subgenomic simian virus 40-Neo mRNA. The amount of poly(A)' RNA L7C loaded onto the gel for the 2.3S.19 cells and the astroglia was at least 3 times lower than in other lanes.

Figure 3 is a schematic representation of retroviruses expressing c-myQ, N-myc and L-my. All the vectors have a pBR322 backbone for replication in E.coli. SD and SA indicate splice donor and slice acceptor respectively.

42 denotes the packaging signal. SV indicates origin of replication. The three transcripts expected from the provirus are shown. Restriction endonuclease sites are abbreviated as follows: Xh, XhoI; Xb, XbaI; C, ClaI B, BamHI.

Figure 4 shows Southern blot analysis of DNA isolated from immortalised neuroepithelial cell lines infected with MPZenSVNeQ(N-myQ) virus. Digests were with XhaI(X), EcRI(RI(RI) and HiindIII(H). The probe was the SacII/HindII fragment of the mouse genomic N-myc clone (36).

Arrowheads indicate the proviral N-my fragment. Note that not all the digested DNA samples were run on the same S gel.

20 Figure 5 shows expression of MPZenSVNeo(N-my) recombinant retroviruses in neuroepithelial cell lines.

Poly(A)+ RNA was prepared from E10 neuroepithelial cells (lane MP2enSVNeQ(N-myc) infected NE cell lines (lanes neuroblastoma C1300N2a (lane The filter was first hybridised to N-myc probe then stripped off and hybridised to c-myc and actin probes. Sizes of transcripts are in kilobases 2.3 and 2.9 transcripts represent cellular c-myc and N-myq respectively. The proviral transcripts are 6 and 5.6 kb.

The 5.6 kb transcripts are not resolved in the over exposed lanes (lanes 2,5,6) and are not detectable in the 0 under exposed lane (lane 3).

i -z 8 Eigure.6 shows expression of c-myQ and N-my. in embryonic brain. Poly(A) RNA was prepared from EJ0 neuroepithelium (E10), brain of 16-day old embryo (E16) and neuroepithelial cell lines infected with Dol(c-my.) virus Astroglia were from cultured populations of 19-day embryo cerebellar cells (E19 ast), glioma cell line (CG glioma) and neuroblastoma cl300N2a The filter was hybridised to c-mya probes, stripped off and hybridised to a N-my- probe.

Figure 7 shows expression of c-my_ in Zen(c-my-Q) based virus infected neuroepithelial cell lines. Poly(A)+ RNA was prepared from E10 neuroepithelial cells (NE), Dol(c-myc) infected neuroepithelial cell line (2.3D), ZenSVN&(c-myQ) (R1.3) and MPZenSVNeo(c-my.) (MM2.4) virus infected neuroepithelial cell lines, and Zen(c-my (R15.1, R15.6) and MPZen(c-my (M40.4, M40.5) infected neuroepithelial cell lines. Cellular c-my-Q is 2.3kb, the proviral transcripts in the ZenSVN (c-myc) lines are 5.1 and 4.7 kb while those from the Zen(c-myc) lines are 3.4 and Figure 8 shows Western blot analysis of myc protein expressed in the different cell lines. +2 NZen6- 2 line producing MPZenSVNQp(N-myc) virus; NZen17 and a NZen 40 neuroepithelial cell lines generated by infected with the MPZenSVNeQ(H-MyEq) virus. qY2.3 q2 line producing Dol(c-myg) virus; 2.3D neuroepithelial cell line generated by infection of neuroepithelial cells with Dol(c-myc) virus; HL-60 promyelocytic cell line.

EXAMPLE 1 MATERIALS AND METHODS Construction of Dol(c-myc) retrovirus.

The c-myg retrovirus used to infect neuroepithelial cells was constructed by using the shuttle vector pDol as described This vector carries the bacterial neomyocin resistance (Ne

R

gene that confers resistance to the antibiotic G418 The Xho I fragment of the murine c-my cDNA was inserted into the BamHI site of the pDol vector. This fragment contains the entire c-myc coding region A schematic diagram of the recombinant retroviral plasmid is shown in Fig.1A. To avoid possible effects induced by a helper virus in subsequent experiments, virus stocks were produced by transfection of y-2 fibroblasts (20) with the retroviral vector. Transfected -2 cells were selected for resistance to G418, and a cell line producing the highest 4 titer of the c-mny virus (10 viruses per ml) was used for all the experiments.

Infection of neuroepithelial cells.

Virus-producing y-2 cells were cultured in 24-well Linbro plates in 1 ml of Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (vol/vol) fetal calf serum at 5x103 cells per well and were irradiated with 2800 rads. After 24hr, single-cell oa. 20 suspensions of neuroepithelial cells isolated from mesencephalon of 10 day embryo CBA/CaH mice were cultured on the irradiated cells at various concentrations ranging 3 5 6 4 from 10 to 5x10 cells per well. After culturing the cells for 2-5 days, 400pg of Genetecin (G418) (GIBCO) was added to each well, and fresh medium containing G418 was subsequently added every 3-4 days. Cell lines were cloned by limiting dilution by using 3T3 fibroblasts as an underlayer.

Immunofluorescent staining.

Cell lines were cultured on glass coverslips contained in 24-well Linbro plates at a density of 2x10 3 cells per ml. After incubation for 48hr coverslips were removed and transferred to a humidified Terasaki plate.

To stain for surface antigenic determinants, 50pl of the primary antibody was added per coverslip and incubated for min at room temperature. Coverslips were washed by gentle immersion into beakers of isotonic phosphatebuffered saline (PBS), and 50pl of fluorescein-labelled second antibody was applied to each coverslip and incubated for a further 30 min. Coverslips were washed as before and fixed by the addition of acid alcohol (vol/vol) absolute ethanol and 5% (vol/vol) acetic acid] for 20 min at -20 0 C. The coverslips were inverted and mounted in glycerol containing 2.6% (wt/vol) 1,4-diazobicyclo- [2.2.2]octane (Dabco, Merck, Melbourne, Australia). To visualise the glial fibrillary acidic protein (GFAP) antigen and the neurofilament proteins, cells were fixed in acid alcohol prior to the addition of the primary antibody.

Biological reagents.

an anti-ganglioside antibody was used at a dilution of 1:100 of ascites fluid. Anti-GFAP antibody was a polyclonal antiserum from Dakopatts (Sydney, Australia) and was used at a dilution of 1:30.

Anti-neurofilament antibody (22) was obtained from Immuno Nuclear (Stilwater, MN) and was used at a dilution of 1:150. K2F2, an anti-cytokeratin antibody was obtained from D.Hewish (CSIRO, Melbourne, Australia) and was used at a dilution of 1:100. WEHY-NEP-6, an anti-glial cell precursor monoclonal antibody, was C prepared in our laboratory and used at a dilution of 1 1:10. Anti-H-2K monoclonal antibody was from hybridoma clone 11-4.1 and anti-IA monoclonal antibody was from 30 hybridoma clone 10-2.16 both hybridomas were obtained from the American Type Culture Collection.

Acidic FGF from bovine brain was obtained from R D Systems (Minneapolis, MN). Recombinant bovine basic FGF was obtained from Collaborative Research (Waltham, 11 MA). Recombinant interferon y was obtained from Genentech (South San Francisco, CA).

RESULTS

Infection of neuroepithelial cells.

To maximise the efficiency of infection, neuroepithelial cells were co-cultivated with the virus-producing y-2 cell lines. A y-2 line producing equivalent titers of the pDol virus carrying the Neo gene was used as a control. The majority of the neuroepithelial cells used for the infection expressed the cytokeratin intermediate filament marker as shown by immunofluorescent staining with the antibody K2F2, indicating that this population was not significantly contaminated with mesenchymal elements. The G418-resistant cells were evident 7-10 days after the addition of the drug, whereas, in control cultures containing only neuroepithelium, all the cells were dead oan within 48hr of G418 addition, After 10-14 days of culture, no q-2 cells survived, leaving only the 0 infected neuroepithelial cells to proliferate. Do (Neo) .o virus-infected cells grew slowly and only survived for 2-3 oo!o months. In marked contrast, cells infected with 8o Dol(c-myc) virus grew rapidly after an initial lag period of 1-2 weeks and have continued to proliferate for >12 months. The frequency of precursors capable of giving Srise to Ne cell lines was assessed by limiting a°a dilution and found to be 2x10.

Clonality of the cell lines.

30 To establish that the cell lines did indeed o V harbour c-my_ virus and to investigate their clonal o composition, DNA from selected lines was digested with EcoRI with XbaI restriction endonucleases and subjected to oa Southern blot analysis (25) with a radioactive c-myc 12 0 00 06 0 0004 OcO *0 0d* probe. Fic.lB shows the results for the lines 2.3Q and 2.3D as well as for 2.3D.12, which is a clonal derivative of 2.3D. Xba I cuts within each proviral long terminal repeat, releasing a 5.5-kilobase fragment. Since, for 2.3Q cells, the intensity of this band was about half that of the 8.5-kilobase fragment from the endogenous c-myc alleles, it is clear that this line carries only a single provirus and is, therefore, clonal. The line 2.3D harbours multiple inserts, as judged by the relative intensity of the Xba I fragments. Individual proviral inserts can be distinguished in EcQRI digests: there is only one EcoRI site within the provirus, and the size of the released fragment thus depends on the location of the nearest EcQRI site in the flanking DNA. By this criteron, 2.3D cells carry three inserts, two within E_cRI fragments of similar size. Since the number and size of the proviral fragments was identical in the subclone 2.3D.12 (Fig.lB), the 2.3D line is also clonal.

Expression of c-myc mRNA.

20 Analysis of mRNA from several neuroepithelial cell lines for c-myc expression showed that both the proviral and endogenous c-my. genes are expressed in all cell lines (Fig.2A). As expected from the number of integrated proviruses, the level of the 6-kilobase proviral c-myc transcript is higher in the 2.3D line than in the 2.3Q line. Significantly, the 2.3-kilobase endogenous c-myc transcript is expressed in all the lines examined. Thus, the introduction of the proviral c-myc gene has not silenced the endogenous c-my. gene. This result contrasts with those obtained for B lymphocytes in transgenic mice carrying a c-myc oncogene driven by the immunoglobulin heavy chain enhanced, where no expression of the endogenous c-my gene could be detected (26,27).

However, it is similar to that reported for some t fibroblast cell lines where expression of the endogenous c-my_ gene is not suppressed by the introduced c-my gene (28).

Tumoriaenicity of the neuroepithelial cell lines.

The c-myc virus-infected cells were not tumorigenic. Syngeneic animals injected with 5x10 6 cells from several lines were monitored for 3 months, but no tumours were detected. Thus, by itself, deregulated c-mya expression appears not to be sufficient to transform neuroepithelial cells. However, after continual passage in vitro for over a year the 2.3S and 2.3Q cell lines were capable of generating tumours when injected subcutaneously into mice. The transformation of these cell lines was not associated with any discernible alteraation in morphology or surface phenotype. Furthermore, RNA gel blot analysis has shown no significant changes in the expression of either the endogenous or the proviral c-myc genes (data not shown). It is presumed that transformation resulted from accumulation of additional genetic change(s) during S* 20 long term culture.

Phenotype of neuroepithelial cell lines.

t: The majority of the cell lines generated from the o neuroepithelium appear to represent cells at a very early stage of neural differentiation. They have the morphological appearance of normal neuroepithelium 00cultured in vitro and express similar antigenic markers (Table All the cell lines express the cytokeratin intermediate filament; this marker is also' found in °0wa neuroepithelium but is lost during neural differentiation 30 and is not present in mature glial or neuronal cells. A o o further indication that these lines are of an immature phenotype is that the majority of the cell lines do not express neurofilaments or the neuronal surface marker (21) (also expressed on some glial cell precursors or the astrocyte-associated intermediate filament GFAP. In addition, the lines do not express the surface marker NEP6 normally present on the glial precursor cells prior to expression of GFAP. In an attempt to induce these lines to differentiate or to shift antigenic phenotype, several agents known to induce cellular differentiation, such as phorbol esters, retinoic acid, and 8-azacytidine have been used, but no surface phenotypic changes have been observed in any of the lines examined. However, IFN-y and acidic and basic FGF have been effective in initiating phenotype changes.

Induction of surface antigens by IFN-y.

The neuroepithelial cell lines are similar to neuroepithelium in that they normally do not express surface class I and class II histocompatibility antigens.

This property is also shared by mature neurons and glia; however, it has been shown that some mature neural cells can be induced to express both classes of molecules after incubation with IFN-y It has been found in all 20 the 140 neuroepithelial cell lines tested that classes I and II histocompatibility molecules can be induced within 48 hr after IFN-y treatment (Table The expression of class I antigens is similarly induced on freshly Sisolated E10 neuroepithelium However, the expression of class II antigens is not seen on normal neuroepithelium and is only found on GFAP-positive astrocytes. The 8 "o significance of this finding is unclear, although it may indicate that these lines are biased toward the glial differentiation pathway. As shown in Fig.2A the S 30 pretreatment of the neuroepithelial cell lines with IFN-y has no effect upon endogenous or proviral c-mya eooa gene expression.

FGF stimulates differentiation.

Acidic and basic FGF are found in relatively high concentrations in the brain (31) and have been shown to enhance neuronal survival in vitro and in vivo (32).

Results from our laboratory indicate that FGF is a potent proliferative stimulus to freshly isolated neuroepithelium and, at concentrations >5 ng/ml, induced morpho-differentiation. Cells from the 2.3D cell line were incubated with FGF (5ng/ml) and showed obvious morpholical changes within 24hr of incubation, and some of the cells begin to round up and form aggregates that sometimes detached from the plate. Staining of coverslips of these lines after treatment with FGF for the presence of neuronal and glial markers has revealed cells containing GFAP as well as a process bearing population that is both A2B5-positive and also contains neurofilament protein. A large percentage of the cells (40-50%) express neurofilaments by 3 days, whereas GFAP is not detected until 7 days after the addition of FGF.

9 9 44449 9 44 94 4 4 9r 9 4~ 4 94 49 4 o 44 94 49 44 4 tI 6 1 91 4 4 L~'j 16 TABLE 1 Phenotype of neuroepithelial cell lines Antibody markers Cell(s) A2B5 GFAP NEP6 IAI* ElONE 2.3D,2.3S, and 2.3Q. of lines screened 2.3A 2.3R Cells cultured on glass coverslips were examined for expression of various markers by immunofluorescence with the following antibodies: A2B5, anti-ganglioside; anti-GFAP; K2F2, anti-cytokeratin; WEHY-NEP-6, anti-glial cell precursor; anti-H-2Kk monoclonal antibody clone 11-4.1; and anti-I-Ak monoclonal antibody clone 10-2-16. All cells listed were K2F2-positive, H-2-negative, H-2-positive in the presence of IFN-y, and I-A-negative. The antigenic phenotype of the E10 neuroepithelium (NE) was similar to that of 95% of the cell lines screened.

Absence of marker; presence of marker.

Cells were incubated with IFN-y (1 unit/mi) for 24 hr prior to staining; indicates the ability to express these antigens in the presence of IFN-y.

EXAMPLE 2 EXPERIMENTAL PROCEDURES Construction of Zen-mvc retroviruses.

The retroviral vectors used in these experiments, pZen, pMPZen, pZenSVejQ and pMPZenSVNeQ (11) were derived from pZipleoSV(33) and from a variant plasmid pMPZip~gg r containing the enhancer from the LTR of the myeloproliferative sarcoma virus The pZen(myv), pMPZen(c-myg) pZenSVNa(c-myc) and pMPZenSVNe(c-myc) were 10 constructed by insertion of the 1391 bp XhQI fragment of the murine c-my cDNA (19) into the XhI site of the pZen, 0 4a pMPZen, pZenSV~eg and pMPZenSViQ vectors. The I- pMPZenSVNe(N-myc) vector was constructed by insertion of the 4160 bp HindII/SacII fragment of the murine N-myc gene into the XhoI site of the pMPZenSVNeo vector DNA.

This fragment contains exons 2 and 3, 820bp of IVS-1, IVS-2 and 70bp of the 8' untranslated region. The pMPZenSVNeo(L-myc) vector was constructed by insertion of 1400 bp SaI/HindIII fragment containing the murine cDNA clone. The L-myc cDNA was constructed by ligating the SacI/StuI fragment isolated from exon 2 of the genomic L-myc clone and the StuI/HindII fragment of the L-Imy cDNA clone This full length cDNA construct contains 166 bp of the 5' untranslated region, the complete L-myg coding sequences and 42 bp of the 3' untranslated region (Fig.3).

Generation of y2 virus producing lines.

P2 cells (20) were transfected with retroviral plasmid DNA using electroporation apparatus (Biorad). 5 or 10pg of plasmid DNA were mixed with 1x10 l 2 cells in 0.8ml of DMEM supplemented with i °o 20 10% calf serum (when the retroviral plasmid DNA did not contain the SVNeo gene 30pg of retroviral plasmid were mixed with 3pg of pSV2N~e plasmid DNA The cells O 0 were incubated at 40 for 10 min prior to electroporation.

They were then pulsed with 500 volts and 25pF and allowed to stand on ice for a further 5 min prior to plating onto 8 dishes (10cm) in DMEM supplemented with calf serum. After 48 hrs fresh medium containing 400pg/ml of G418 was added and resistant colonies became o evident 10-14 days after transfection, they were ring 30 cloned and expanded for further analysis.

Infection of neuroepithelial cells.

Infection of neuroepithelial cells was performed in exactly the same way as in Example 1.

18 Immunofluorescent staining.

Immunofluorescent staining was performed by the method described in Example 1.

Growth factors.

Acidic FGF from bovine brain was obtained from R D systems, Minneapolis MN, USA; recombinant human EGF, recombinant bovine bFGF, recombinant human IGF-I(Thr-59) and recombinant PDGF(v-sis) were obtained from Amersham, Buck.,UK. IGF-II (CR-multiplicator stimulating activity) and NGF (2.5S) were from Collaborative Research, Mass., USA. Growth factors were added to the cell lines at the following concentrations: EGF, 50ng/ml; IGFI, 50ng/ml; PDGF, 50ng/ml; NGF, bFGF, 5ng/ml; aFGF, 5ng/ml; IGFII, 100ng/ml.

Cell proliferation assay.

Cells were harvested from high density cultures of each cell line, washed 3x with DMEM containing 1% FCS and 10pl of cells at lxlO4/ml were plated onto 60 well HLA plates (Terasaki plates). Growth factors or conditioned medium from the different cell lines were then added to the cultures and the plates were incubated for 48-72 hrs. After incubation the number of cells was determined by counting under an inverted phase microscope at 100X magnification. To exclude non-viable cells Eosin was added to each well (2 min) prior to counting.

Analysis of RNA and DNA.

6 DNA was prepared from 5x10 cells by lysis with guanidine hydrochloride and analysed by Southern blot hybridisation PolyA-RNA was prepared as described by Gonda et.al.(27). Samples of 2pg were fractionated on 1% formaldehyde/agarose gel as previously described The c-myq probe was the XhoI fragment of the c-mye cDNA clone pMc-myc54 (19) while the N-myc probe was the HindII/SacII fragment containing exons 2 and 3 of the 9- :r i; 19 N-myQ genomic clone The probes were labelled with 32 c( P) dATP using a random hexamer priming kit (Besatec Limited, South Australia).

Western blots.

Cell lysates were prepared from approximately 2x107 cells at log phase. Cells were scraped from dishes with a rubber policeman, washed twice with PBS pH 7.5 and lysed immediately in lml of sample buffer (0.1MTris pH 6.8, 10% glycerol, 0.1% triton X-100). The lysates were sonicated, centrifuged, and the supernatants aliquoted and stored at -70 0 C until used. All lysates were used within 2 weeks of preparation. 80 pg of cell lysate proteins were fractionated on 10% SDSpolyacrylamide gel and electroblotted onto nitrocellulose paper Following transfer the filter was washed in PBS and pre-incubated for 30 min in PBS containing nonfat dried milk. The filter was then incubated for 2 hrs with sheep anti-myc oncoprotein antibody (DCP801 Cambridge Research Biochemicals) diluted 1:100, washed and 1 20 further incubated for 1 hr with rabbit anti-sheep IgG.

After washing, the filter was incubated for 1 hr with 125I-protein A (1x105 cpm/ml, 40 pci/pg), washed and autoradiographed at -70 0 C. All incubations were at room temperature. After exposure, the nitrocellulose containing the myc protein was cut out and counted in X counter to estimate the amount of myc protein in the different cell lines.

o t

RESULTS

30 Generation of neural cell lines with new N-myc, c-myc and L-myc retroviruses.

It has been shown previously that the Dol(c-myc) retrovirus can immortalise neural precursors and that the lines so generated retain characteristics of early neural tr-m cells and do not spontaneously differentiate (see Example The effects of higher levels of constitutive N-myc, c-imyc and L-myq expression on neuroepithelial cells has been investigated utilising a family of new retroviral vectors, the Zen vectors (11) (Fig.3) which were shown to express high levels of the inserted genes. These vectors were derived from the previously described ZipNeoSV(X) (33) and pMPZipN e The Zen vectors all expressed the inserted gene from a spliced subgenomic mRNA and utilise either the Moloney virus enhancer sequence (pZen,pZenSVNeo) or that from the myeloproliferative sarcoma virus (pMPZen,pMPZenSVNeo). The N-myc, c-myc and L-myc retroviral vectors were constructed by insertion of the genomic N-myc HindII/sacII fragment, the Xhol fragment of the c-my cDNA, and a L-myc cDNA construct, respectively, into the XhoI site of the pZen vectors (see experimental procedures).

N-myc, c-myc and L-myc virus producing lines were a generated by electroporating the retroviral plasmid DNA S0* 20 into $2 cells together with the plasmid pSVNeo where necessary. Colonies were selected in G418 expanded and o 4a assayed directly for their capacity to immortalise oa neuroepithelial cells as previously described (Example 5 S1). In brief, 5x10 virus producing p 2 cells, plated onto 24-well Linbro plates, were irradiated with 2800rads prior to addition of single cell suspensions of neuroepithelial cells at various concentrations 3 5 ranging from 10 to 5x10 cells per well. Of the two 2 lines producing MPZenSVNeo(N-myc) virus, one had the capacity to immortalise E10 neuroepithelial cells. This cell line, 4 2 NZen6 was used to generate all the neuroepithelial lines immortalised with N-my. All the six 2 lines producing MPZenSVRNe(c-myc) viruses and the two 12 lines producing MPZen(c-myc) and Zen(c-my) n' II. 21 viruses were capable of immortalising neuroepithelium. However, only two of the six 42 producing ZenSVNeo(c-myc) viruses were capable of immortalising E10 neuroepithelial cells. The titers of the N-myc and c-myc viruses produced by the l2 lines capable of immortalising neuroepithelial cells were found to be between 1x10 5x10 colony forming units per ml (17).

Cells infected with c-myc or N-myc Zen viruses can differentiate spontaneously in vitro into neurnal and glial cells.

Mock-infected neuroepithelial cells formed small aggregates which disappeared within one week. In marked contrast, within one week of infection, neuroepithelial cells infected with the my- viruses formed large aggregates of cells that express the neural marker which is characteristic of the neuronal cells but is also expressed by some glial precursors In addition, a j o°9 *large number of these cells also expressed 20 neurofilaments. The cells did not however, express the l glial cell marker glial fibrillary acidic protein (GFAP).

4 4 The size of the aggregates increased with time and networks of neuronal like cells were observed both in the aggregates and in the surrounding flat cells attached to the dish. All 48 culture wells contained cells of identical morphology. After two weeks, some cells in each well expressed the glial cell marker, GFAP. However, these cells were distinct from those expressing the and neurofilaments. The morphology as well as the heterogeneity of these cell lines was maintained during the first 5-6 passages.

1 Each cell line was examined by fluorescence immunocytochemistry for the presence of the neural marker for the neurofilament marker identifying neurons and vector pDol as described This vector carries the 22 for GFAP, the marker of glial cells (Table All the lines contain cells that express both neuronal and glial cell markers, although the relative frequency of each cell type varied from line to line. Several of the cell lines have been cloned by limiting dilution, a procedure which proved difficult without plating the cells onto an irradiated feeder layer of 3T3 cells. The cloned lines differed from the founder cell lines, in that they mainly grew as an epithelial sheet resembling both normal neuroepithelium and the previously described 2.3D cell line derived by infection of neuroepithelial cells with the Dol(c-my) retrovirus (Example However, unlike the 2.3D cell line and the other cell lines derived by infection with the Dol(c-myg) virus, the newly-generated cell lines differentiated spontaneously into both neurons and glial cells, when the cells reached high density.

They also formed aggregates on the top of the epithelial-like monolayer similar to the aggregates formed during the first few days after the retroviral infection.

Some of the cloned cell lines have now been passaged in yitra for overe a year and still retain the ability to spontaneously differentiate at high density, although their ability to do this appears to decrease with increasing passages. The ability of the cloned cell lines to differentiate into both neurons and glial cells indicates that immortalised bipotential neuroepithelial stem cells have been immortalised.

Clonalitv of the cell lines.

Southern blot analysis revealed that all the cell lines harboured a full length provirus (Fig.4). The N-myc virus-infected cell lines (designated NZen) all generated a 5.5Kb N-myc fragment on digestion of the DNA with XbaI, which cuts once within each LTR. Thus, as expected, the a 41 t N-myc intron present within the transfected plasmid was l I t 23 removed from the viral RNA by splicing. The DNA from most of the lines yielded a unique c-myc or N-myc bearing fragment after cutting with EcQRI or HindIII which cut only once within the retroviral vector, indicating that they derived from an integration of a single provirus copy. The presence of unique N-myc o7 c-myc. EQcoRI and HindIII fragments at intensities equivalent to half of the endogenous myc alleles indicate that the cell lines are indeed clonal.

The expression of the cellular c-myc mRNA is suppressed in the neuroepithelial cell lines.

All the neuroepithelial cell lines generated by either the N-myc or the c-my- viruses expressed high levels of proviral transcripts as well as smaller processed transcripts from which the viral myc protein is translated. The N-myc virus-infected lines express a 6 Kb viral N-myZG mRNA and a smaller 5.6 Kb sub-genomic N-my_ mRNA. No expression of the endogenous Kb N-myc mRNA t6. was observed in these lines (Fig.5, lanes Whilst it 20 cannot be concluded that the lack of the endogenous N-myc S0 expression is due to suppression of the cellular allele, 0000 it is clear that in these lines cell lines c-myg expression is turned off. Like all dividing cells, neuroepithelial cells express high levels of c-myc (Figs.5,6), however cellular c-myc expression was not detected in the N-myc infected lines even on extended exposure of the autoradiograph (Fig.5, lanes 2-6).

S" Cells infected with the ZenSVNeQ(c-my.) viruses and the MPZenSVNeg (c-myc) viruses also express two viral S 30 c-myc transcripts of 5.1 and 4.7 Kb (Fig.7). While their Slevel of expression is very high when compared to that expressed by E10 neuroepithelial cells, it is similar to a that expressed by the neuroepithelial cell lines generated by infection with the Dol(c-myc) virus (Fig.7). However, -it- 24 in contrast to the 2.3D cell line which also expressed the cellular c-myc, only trace amounts of cellular c-myc are expressed by the cell lines immortalised with the Zen(c-myg) virus. The cell lines infected with Zen(c-myc) (R15.1 and R15.6) and with MPZen(c-myc) (M40.4 and M40.5) viruses also express two c-my viral transcripts of 3.4 and 3.0 Kb with similar intensities. As in the case of the ZenSVNeo(c-myxc) virus-infected lines the expression of the cellular c-myc is completely suppressed.

Expression of the myc protein in neuroepithelial cell lines.

The Dol(c-my.Q) and the Zen-myc virus-infected neuroepithelial cell lines express similar levels of myc mRNA (Fig.7). Since the levels of mRNA expressed may not always reflect the amount of the translated protein, the levels of myc protein synthesised by the different cell lines have been measured using Western blot analysis (Fig.8). The results in Figure 8 indicate that the q2 so cells producing the Dol(c-myc) virus as well as the S\ 20 neuroepithelium cell line 2.3D generated by infection with this virus synthesise very small amounts of the c-myc proteins while the 9 2 cells producing the Zen-myc viruses as well as the infected neuroepithelium cell lines produce high levels of myn protein. It was estimated by counting the different myq protein bands that the cells transfected or infected with the Zen-based viruses synthesise about 10 times more myc protein (N-myc or c-myc) than the Dol-myc cells.

Effect of fibroblast growth factor on the differentiation of the myc virus infected cell lines.

It has previously been reported that 0 neuroepithelial cells as well as the neuroepithelial cell line 2.3D can be induced to differentiate into neurons and glial cells by the addition of acidic or basic FGF I (Example In order to examine the effect of FGF and other growth factors on the cell lines generated with the Zen-based viruses some of the cell lines were tested with a panel of recombinant or purified growth factors known for their ability to induce cellular differentiation or to support cellular growth. All the factors used have been identified in normal brain tissue, namely, acidic and basic FGF, epidermal growth factor (EGF), nerve growth factor (NGF), platelet derived growth factor (PDGF) and insulin-like growth factor I and II (IGFI, IGFII).

Interestingly, it was found that acidic and basic FGF were the only known growth factor capable of significantly influencing cellular differentiation. The effect of FGF was rapid: cell aggregation and neurofilament expression occurring within 2 days of FGF addition. When both heparin and FGF were added the response was enhanced so that 5-6 days after FGF addition, the majority of the cells formed aggregates and detatched from the glass 'coverslip.

4 t I Si

I

S lWvzijLtuu iur some 3 t Table 2 Phenotype of Cell Lines Obtained from ElO Neuroepithelium Infected with RetrovirUSes Containing N-myc or c-myc Cell Retrovirus Clonall Response to Antigenic Markers Expressed 3 line construct FGF 2 GFA P A2B5 Neurofilaments MPZenSVNeo[N-mycl NZen6 it- NZen7 it+ NZen8 of+ NZen8.2 11++ NZen8.6 to4 NZen9 it NZenl3 it+ NZenl4 N.D. +4 4 NZenl7 +4 4 NZen36 NZen37 to+ Rl.2 ZenSVNeo[c-nyc] R1.3 it R1.4 ft R2.2 it+ R15.3 MPZen[c-mycj N.D. 4++ R15.6 11+ MM4.2 MPZenSVNeo~c-mycj N.D. +4 4+ MM4.4 itN.D. MM5.3 toN.D. MM5.6 f 4- 27 Table 2 [cont] 1 Cell lines were cloned by limiting dilutions and their clonality was determined by Southern blot analysis as previously described N.D. not done.

2 100 cells of each of the cell lines were plated in a Terasaki well with or without 5ng/ml bFGF in 10pl DMEM containing 1% FCS. The assays were read 48 hrs later. The response to FGF was scored positive when the number of cells in the presence of FGF were at least twice that of the control. Each cell line was assayed in six replicate cultures.

3 Cells were grown on coverslips at high cell density for 2 days before staining for the expression of the surface marker A2B5 and the two intermediate filament markers GFAP and neurofilaments [see experimental procedures for details]. They were examined for the expression of these markers using fluorescence microscopy. The stained cells were scored as follows: o a 0 a 4 0 0 4 i6 t 0 i O 28 Those skilled in the art will appreciate the significance and potential use of the production of immortalised neural precursor cells in accordance with the present invention. In particular, this invention will enable the study of neural cell lineage during brain development and provide a source of new neurotrophic factors that regulate proliferation and differentiation of neurons and glial cells for the purification and cloning of these factors. The cell lines will also provide a clonal population of target cells to assay new neurotrophic and differentiation factors, as well as providing a model system for the study of replacement of damaged nervous tissue. It will also be appreciated that the production of neurotrophic factors from these cell lines opens the way to the therapeutic application of these factors to damaged nervous tissue, in addition to the direct use of in vitro generated cells in transplantation into damaged nervous a tissue.

S*oo 0 9 n 4t 0 a 4 0 4 4* 29 1. Abney, Bartlett, P.F and Raff, M.C. (1981) Devy.Biol. 21, 301-310.

2. Bartlett, Bailey, K.A. and Wycherley, K.A., (1987) Neurosci. Let..Supyp. 21, 42-43.

3. Bailey, Wycherley, K.A. and Bartlett, P.F.

(1987) Neuroses Lett.Supp. 21, 53.

4. Alt, DePinho, Zimmerman, Legoug, Hatton, Ferrier, Tesfaye, A., Yancopoulos, Nisen, P. (1986). The human my- gene family. Cold Spring Harbor Symp.

Quant.Biol. U1: 931-941.

Ingravsson, Asker, Axelson, Klein, G.

and Sumagi, J. (1988). Mol.Cell.Biol.

_a:3168-3174.

6. Land, Parada, L.F. and Winberg, R.A. (1983).

Na~ture UA~, 596-602.

7. Yancopolous, Nisen, Tesfaye, A., I Kohl, Goldfarb, M.P. and Alt, F.W. (1985).

Proc.Natl.-Acad.Sci. .8d2:1195-1198.

a oi 8. DePinho, Hatton, Tesfaye, A., Yancopoulos, G.D. and Alt, F.W. (1987). Genes.

and Development 1:1311-1326.

449. Legouy, DePinho, Zimmerman, Collum, Yancopoulos, Kitsock, Kriz, R. and Alt, F.W. (1987) EI4BO1 Ji: 3359-3366.

Zimmerman, Yancopoulos, Collum, R., Smith, Kohl, Denis, Nau, Witte, Toran-Allerand, Gee, Minna, J. and Alt, F. (1986). Na~ture 3J-U: 780-783.

11. Hariharan, Adams, J.M. and Cory, S. (1988).

Oncogene Res. (In press).

12. Mougneau, Lemieux, Rassoulzadegan, M. and Cuzin, F. (1983). Proc.Natl.Acad. Sci. USA. R81, 5758-5762.

13. Zerlin, Julius, Cerni, C. and Marcu, K.B. (1987). Qncogen 1, 19-27.

14. Cowan, (1979). Sci. American. 2A1(3), 106-117.

Weston, J.A. (1970). Adv. Morphogen..a, 41-114.

16. Gospodarowicz, D. (1985). In Ford, R.J. and Maizel, A.L. (eds). "Mediators in Cell Growth and Differentiation." Raven Press, New York, NY ppl09-134.

17. Cory, Bernard, Bowtell, Schrader, S.

and Schrader, J.W. (1987), Oncooene Res. 1: 67-76.

018. Southern, P.J. and Berg, P. (1982). J.MolAppj, Genet.1 327-344.

A

19 a 00 0 0 4 4.

44 0 9 Stanton, Watt, R. and Marcu, K.B. (1983) Nature (London) M: 401-406.

Reed, K.C. and Mann, D.A. (1985) Nucleic Acids Rga. 7207-7221.

Mann, Mulligan, R.C. and Baltimore, D.

(1983). Cel 31 153-159.

Eisenbarth, Walsh, F.S. and Nirenberg, M.

(1979). PLQop.Natl.Acad.Sci. USA. 76 4913-4917.

Drager, Edward, D.L. and Kleinschmidt, J.

(1983). Proc.Natl.Acad.Sci USA U&:6408-6412.

French, P.W. and Hewish, D.R. (1986) JJll.

IUZ: 14 12 -1418.

Oi, Jones, Goding, J.W. and Herzenberg, L.A. (1978) Cur.Top.Microbiol.

Imnol. *BJ.:115-129.

Southern, E.M. (1975), J.Mol.iol. 91: 503-517.

Adams, Harris, Pinkert, C.A., Corcoran, Alexander, Gory, S., Palmiter, R.D. and Brinster, R.L. (1985). Naur (London)3 3J)a: 533-538.

Alexander, Schrader, J.W. and Adams, J.M.

(1982) Mol.Cell.Biol. 7: 1436-1444.

Keath, Caimi, P.G. and Cole, M.D. (1985) Cell a2j: 339-348.

24.

26.

4Cr 4444 4 SS .4 41 44 a 4 4 4.5 28.

32 29. Raff, Miller, R.H. and Noble, M. (1983) Nature (London) M 390-396.

Wong, Bartlett, Clark-Lewis, I., McKimrn-Breschkin, J.L. and Schrader, J.W. (1984) Nature (London) 31-Q:588-691.

31. Gospodarowicz, Cheng, Lui, Baird, A. and Bohlent, P. (1984). Proc.Natl.Acad.SCi.tJSA .81: 6963-6967.

32. Morrison, Sharmna, deVellis, J. and Bradshaw, R.A. (1986). Proc.Natl.Acad.Spi.USA 7537-7541.

33. Cepko, Roberts, B.E. and Mulligan, R.C.

34. Bowtell, Johnson, Kelso, A. and Cory, S. (1987). Mol.Biol.Med. 4:229-250.

DePinho, Legouy, Feldman, Kohl, N., Yancopoulos, G.D. and Alt, F.W. (1986).

Proc.Natl.Acad.Sci.USA R3.: 1827-1831.

36. Bowtell, D.D.L. (1987). Anal.Biochem, 1_62: 463-465.

37. Gonda, Sheivess, D. and Bishop, J.M.

(1982). Mol.Cell.Biol. Z: 617-624.

38. Thomas, P.S. (1980). Proc.Natl.Acad.Sci.USA 77: 5201-5203.

39. Towbin, Staehelin, and Gordon, J. (1979).

Proc.Natl.Acad.Spi.USA Jj: 4350.

4~044 4 4 4444 4 44 04 4 4 44 41 0444 044 4 4

Claims (7)

1. A method for the in vitr production of immortalised neural precursor cells, which comprises the step of infecting neuroepithelium or neural crest cells with a retroviral vector carrying a member of the myg family of oncogenes.

2. A method according to claim 1 wherein the myc oncogene carried by the retroviral vector is the c-myc, the N-myQ or the L-myc. oncogene. 0000 on 3. A method according to claim 1 or claim 2 wherein 0 0 the retroviral vector comprises the vector pDol carrying o the c-myc oncogene. o 4. A method according to claim 3 wherein the retroviral vector is used to transfect fibroblast cells, and virus-producing fibroblast cells so produced are then co-cultivated with neuroepithelial or neural crest cells a 4 to infect said neuroepithelial or neural crest cells with said retroviral vector. a A method according to claim 4, wherein the immortalised neural precursor cells are induced with acidic or basic fibroblast growth factor to differentiate to form neuronal and/or glial cells.

6. A method according to claim 1 or claim 2 wherein the retroviral vector comprises a vector selected from the group consisting of pZen, pZenSVife., pMPZen and pMPZenSVeiQ, carrying the c-myQc, the N-myc or the L-myQ oncogene. -o i

7. A method according to claim 6 wherein the retroviral vector is used to transfect fibroblast cells, and virus-producing fibroblast cells so produced are then co-cultivated with neuroepithelial or neural crest cells to infect said neuroepithelial or neural crest cells with said retroviral vector.

8. A method according to claim 7, wherein the immortalised neural precursor cells are induced with acidic or basic fibroblast growth fac-or to differentiate to form neuronal and/or glial cells. ,ao 9. An immortalised or continuous cell line 0°e comprising neuroepithelial or neural crest cells which S have been infected with a retroviral vector carrying a o. member of the myc family of oncogenes. 0 oO"o 10. A cell line according to claim 9, wherein the myc oncogene carried by the retroviral vector is the c-my_, the N-myc or the L-myc oncogene.

11. A cell line according to claim 9 or claim wherein the retroviral vector comprises the vector pDol carrying the c-myg oncogene.

12. A cell line according to claim 9 or claim wherein the retroviral vector comprises a vector selected from the group consisting of pZen, pZenSVNe_, pMPZen and pMPZenSVNaQ, carrying either the c-myc or the N-myc oncogene. Dated this 24th day of May, 1990 AMRAD CORPORATION LIMITED By Its Patent Attorneys DAVIES COLLISON I 0r