Adenoviral Vector as a Gene Delivery System into Cultured Rat Neuronal and Glial Cells

European Journal of Neuroscience, Vol. 5, pp. 1287-1291

0 1993 European Neuroscience Association

Adenoviral Vector as a Gene Delivery System into Cultured Rat Neuronal and Glial Cells Catherine Caillaudls, Sa‘id Aklils, Emmanuelle Vigne2, Annette Koulakoff3, Michel Perricaudet2, Livia Poenarul, Axel Kahnl and Yoheved B e ~ a l d - N e t t e r ~ ’U129 INSERM, lnstitut Cochin de Gknktique Molkculaire (ICGM), 24 rue du Faubourg Saint Jacques, 75014 Paris, France *URA 1301 CNRS, lnstitut Gustave Roussy, PR2, 39 rue Camille Desmoulins, 94805 Villejuif, France 3Laboratoire de Biochimie Cellulaire, College de France, 11 place Marcellin-Berthelot, 75231 Paris Cedex 05, France §These two authors have contributed equally to this investigation Key words: defective recombinant adenovirus, @-galactosidase,brain-specific gene regulation, brain transplantation

Abstract Previous studies have demonstrated that a defective recombinant adenovirus can infect a wide range of postmitotic and slowly proliferating cell types such as hepatocytes, myotubes, pneumocytes and intestinal cells (Stratford-Perricaudet etal., Hum. Gene Ther., 1, 241 -256, 1990; Quantin etal., Roc. Natl. Acad. Sci. USA, 89, 2581 - 2584, 1992; Jaffe et a/., Nature Genetics, 1, 372 - 378, 1992). We have used a defective recombinant adenovirus, Ad.RSV@gal,containing the Escherichia coli @-galactosidasegene targeted to the nucleus under the transcriptional control of the Rous sarcoma virus long terminal repeat promoter (StratfordPerricaudet etal., J. Clin. Invest., 90, 626-630, 1992) to infect non-dividing neural cells in primary culture. We show that 80- 1000/0 of neuronal and astroglial cells infected with a viral titre lower than lo9 p.f.u./ml express P-galactosidase for at least 1 month without cell damage. These results demonstrate the potential usefulness of recombinant adenovirus infection for the analysis of brain-specific gene regulation and for the transfer of genes into neural cells before their transplantation into the brain.

Introduction Retroviral vectors are very efficient in the transfer of foreign genes into various cell types manipulable ex vivo (Temin, 1986). The limitation of recombinant retroviral vectors is their requirement for exogenous DNA integration into replicating target cells or cells that are proliferation-inducible. Non-dividing cells are not permissive for infection by such vectors (Culver et al., 1992). Conversely, adenoviral vectors are not dependent on host cell replication and allow stable expression of the foreign gene in an episomal state. They may thus be useful for gene transfer into postmitotic neural cells. Other advantages of adenoviral vectors are: (i) the large amount of exogenous DNA that can be inserted (> 7 kb), (ii) their low pathogenicity in humans (Chanock et al., 1966; Straus, 1984), and (iii) the possibility of obtaining a high titre of virus. Previous studies have demonstrated the efficiency of adenoviral vectors in gene transfer into postmitotic cells such as hepatocytes and myotubes (Stratford-Perricaudet et al., 1990; Quantin et al., 1992; Jaffe et al., 1992) and their ability to partially correct ornithine transcarbamylase deficiency in spfash mutant mice (Stratford-Perricaudet et al., 1990). Because of the blood-brain barrier, the systemic route used in these studies did not allow us to determine whether CNS neural cells are permissive to adenoviral infection. The infection of cultured CNS-derived cells by recombinant

Correspondence to: Axel Kahn, as above Received 22 September 1992; revised 28 April 1993; accepted I June 1993

adenovirus could be very useful for both experimental and therapeutic purposes, e.g. for analysing the regulatory regions of brain-specific genes, modifying the phenotype or modulating the function of CNSderived cells, and transfer of a therapeutic gene into cells before their transplantation into the brain. In the present study, we optimized the conditions of infection of neuronal and glial cells by a defective adenovirus that directs the expression of an Escherichia coli lacZ gene. We show that the majority of cultured neurons and glial cells can be infected without detectable toxicity by the defective adenoviral vector that directs stable expression of the lacZ gene for at least 1 month.

Materials and methods Cell culture For culture of neurons, the cells were derived from brains of 15-day mouse or rat embryos. After removing the brain, dissociated cells were obtained by gentle mechanical disruption of cell-cell contacts. The dissociated cells were suspended in culture medium and seeded on polyornithine-treated 12-well culture plates at 1 x lo6 viable cells per well. Replication of glial cells was suppressed by addition of the

1288 Cultured CNS cells infected with an adenoviral vector antimitotic agents cytosine arabinoside (1 pg/ml) or fluorodeoxyuridine M) 48 h after plating (Berwald-Netter et al., 1981; Couraud et al., 1986). Cultures of glia without neurons were obtained by dissociation of cerebral hemispheres from newborn mice or rats by mild treatment with trypsin/collagenase followed by pipette trituration. Cells were used as secondary cultures seeded at 2.5 X 105 cells per well in 12-cell culture plates. Culture conditions were as described (Gautron et al., 1992).

Construction of recombinant plasmid pAd. RSVPgal and of recombinant adenovirus Ad.RSVpgal The procedures to obtain these two constructs have been described in detail by Stratford-Perricaudet et al. (1992). Briefly, the Ad.RSV0gal is a replicationdeficient recombinant adenovirus (El ,E3 region deleted) obtained by in vivo recombination between the Addl327 mutant and the plasmid pAd.RSVpgal depicted in Figure 1. After isolation, the recombinant adenoviral DNA was amplified in 293 cells, a transcomplementing cell line for E l function (Graham et al., 1977), to obtain a culture medium supernatant containing non-purified recombinant adenovirus at a titre of 1O'O p.f.u./ml. The adenoviral suspension was further purified by caesium chloride density centrifugation.

ltCoK'lTR enh

pLTR ()gal PIX

nls Lac2

-

Infection Rat or mouse neuronal (1 X lo6 viable cells per well) and glial (2.5 X 105 viable cells per well) cells in culture were infected with different titres of viral suspension diluted in 250 pl culture medium and incubated at 37°C for 75 min. 750 p1 of culture medium was then added per well, and the plates were incubated at 37°C with medium replaced every 3 -4 days. After infection, plates were maintained for between 30 h and 1 month at 37"C, fixed and assayed in situ for P-galactosidase activity using the chromogenic substrate 5-bromo-4chloro-3-indolyl-P-~-galactoside (X-Gal). ,8-galactosidase cytochemistry Cultures were rinsed with 150 mM NaCl, 15 mM Na phosphate, pH 7.3 (phosphate-buffered saline, PBS), and fixed for 5 min in 0.37% formaldehyde plus 0.2% glutaraldehyde in PBS. The cells were then washed twice with PBS and covered with a reaction mixture containing 0.4 mg/ml X-Gal, 4 mM potassium ferricyanide, 4 mM potassium ferrocyanide and 2 mM MgC12 in PBS. The X-Gal was dissolved in dimethyl sulphoxide at 40 mglml and then diluted into the reaction mixture. Incubation was for 2 h at 37°C.

Results The defective recombinant adenovirus Ad.RSVpgal used in this study has been described by Stratford-Perricaudet et al. (1992). The plasmid pAd.RSVBgal (Fig. 1) contains the 5' inverted terminal repeat and adjacent packaging sequences of adenovirus-5, the P-galactosidase gene of E. coli containing the amino acid sequence of SV40 that specifies nuclear expression (nlslacZ) under the control of the RSV LTR promoter, a poly(A) signal from SV40, and sequences coding for adenovirus-5 polypeptide IX with 3' flanking sequences to allow homologous recombination. To determine ifAd.RSVfigal can be expressed in CNS neurons, neuron cultures derived from fetal mouse or rat brain were infected with different concentrations of Ad.RSVOgal, non-purified or purified as described in Materials and methods, and tested for P-galactosidase activity. When using non-purified adenovirus at 4 x lo7 p.f.u./ml, < 1% of neuronal cells, in primary cultures infected at day 7, expressed 0-galactosidase. At 2 x 108 p.f.u./ml, 10% of neurons were permissive

FIG. 1 . Structure of recombinant plasmid pAd.RSVpgal. The plasmid contains the following genetic elements: ITR,inverted terminal repeat, necessary for replication; nlslacz, gene coding for E.coli p-galactosidase containing a nuclear localization signal (nls) under the transcriptional control of RSV LTR and pIXAd5, coding sequence for polypeptide IX of adenovirus-5. After homologous recombination with defective adenovirus-5, this construct allows the constitutive expression of P-galactosidase targeted to the nucleus of infected cells.

to adenovirus. Maximum infectability was obtained at a titre of 4 x lo8 p.f.u./ml, with nearly 30% of 0-galactosidase-positivecells. However, infective doses of > 4 x 10s p.f.u./ml of non-purified adenovirus caused severe cell damage leading to cell death within 3 -7 days after infection. Similar experiments were realized with a purified adenoviral suspension. At 4 x lo7 p.f.u./ml, 20% of neuronal cells were 0-galactosidase-positive and 65 % of neurons were permissive to adenoviral infection at 2 x lo8 p.f.u./ml. A maximum of 80% of neuronal cells expressed P-galactosidase when infected with 4 x los p.f.u./ml of Ad.RSV@gal (Fig. 2A). Labelled cells were not detected in parallel non-infected cultures. Infectability of CNS neurons was tested as a function of time in culture (Fig. 3) using purified Ad.RSVPgal suspension at 2X lo8 p.f.u./ml. Between 1 h and day 3, 25 % of cells were permissive for adenovirus infection. A maximal level of infection for this viral titre (65 % of cells expressing P-galactosidase) was attained on day 7 of culture, and the level of infection remained stable as long as the cultures lasted. Astroglial cells were even more infectable by recombinant adenovirus and able to express /3-galactosidase. Cultured glial cells containing essentially astrocytes ( > 95 % pure) were infected with purified Ad.RSV@gal and tested for P-galactosidase activity. At 4 X 10' p.f.u./ml, 20% of glial cells expressed P-galactosidase,but nearly 100% of cells were 6-galactosidase-positive in cultures infected with 2 x lo8 p.f.u./ml (Fig. 2B). The identity of the infected cells was examined by immunolabelling with specific markers: antibodies against y enolase-a predominantly neuronal isoenzyme (Secchi et al., 1980)-and antibodies to glial fibrillary acidic protein (GFAP)-an astrocyte-specific intermediate filament protein (Eng et al., 1971). As expected, the vast majority of cultured cells in neuronal cultures were positive for both y enolase

-

Cultured CNS cells infected with an adenoviral vector 1289

FIG. 2. Expression of nucleus-targeted 8-galactosidase in neurons and astroglial cells infected with 2 X lo8 p.f.u. Ad.RSV8gal. (A) Cultures of rat brain neurons infected 7 days after plating. (B) Cultured rat astroglial cells 40 h after infection. @-galactosidase expression is seen as nuclear blue reaction product. (C) Immunoperoxidase staining of neurons with anti-y-enolase antibody. (D) h u n o p e r o x i d a s e staining of astroglia with anti-GFAP antibody. I00

aJ ln m

z

VI 0

L1

m U

80 90

1

m

m CI

-mc ln 0) L

a X

aJ

-

ln

a3

U L

0

R

0

1

2

3

4

5

6

7

8

9

10

I1

12

13

14

15

16

17

18

19

20

days post-plating

FIG. 3. Infectability of rat neuronal cells during in v i m maturation. Cells were infected with 2 x lo8 p.f.u. Ad.RSV@gal at different times after plating. 0-galactosidase was visualized in the nuclei by a cytochemical staining. Values are mean + SD (pooled values from three independent experiments).

and 0-galactosidase (Fig. 2C). However, a small proportion of /3-galactosidase-positivecells were y enolase-negative, indicating that some non-neuronal cells were present in the culture. These cells were

identified as astrocytes by anti-GFAP immunostaining (Fig. 2D). The level of this ‘contamination’ was estimated at 5 - 10%of cells in primary neuron cultures maintained in the presence of an antimitotic agent. As

1290 Cultured CNS cells infected with an adenoviral vector previously observed, the morphology of astroglia grown in the absence of neurons remained ‘protoplasmic’ (Fig. 2B), whereas that of astroglia grown in the presence of neurons became ‘fibrous’ and clearly astrocytic (Fig. 2D). Both forms were susceptible to the Ad.RSV@galinfection and both bound antibodies to GFAP. Similar results were obtained for rat and mouse neurons and for rat and mouse astroglia (not shown). The level of expression of @-galactosidasein both neurons and glia infected with purified adenovirus was maintained for at least 1 month. For the viral titres ~ 4 x 1 p.f.u./ml, 0 ~ no signs of cytopathic effect were observed in cultures infected with the purified virus preparations compared to non-infected neurons or glia maintained under the same culture conditions. Cytopathic effects were only observed for viral titres > lo9 p.f.u./d.

Discussion Previous studies have demonstrated that adenovirus can infect a wide range of cell types including hepatocytes, pneumocytes, myotubes and intestinal cells (Stratford-Pemcaudet er al., 1990; Quantin er al., 1992; Jaffe er al., 1992). In these studies, adenovirus was used as a gene delivery system into postmitotic or slowly proliferating cells. Here, we show that cultured neurons and astroglial cells can also be infected with high efficiency by such vectors. Nearly 80% of neuronal cells in primary culture were infected and expressed @-galactosidasefor at least 1 month. The transfected gene expression was detectable within hours after cell plating and reached a maximum at day 7, for a viral titre of 4 x lo8 p.f.u./ml. This titre seems to be optimal for obtaining the maximal infection ratio without cytopathic effect. The identity of the infected cells was verified by immunostaining with specific markers: anti-y enolase for neurons and anti-GFAP antibodies for astroglia. Virtually all the glial cells showed expression of 6-galactosidase. No cytopathic effect was observed with purified adenovirus if viral titre was < lo9 p.f.u./ml, and the life-span in culture of infected cells was comparable to that of non-infected cells (at least 1 month) or of cells infected with a non-recombinant defective adenovirus (not shown). The cytopathic effect observed in neuronal and glial cells infected with non-purified Ad.RSV@gal was probably due to the presence of toxic products in the supernatant of viral suspension (such as empty viral capsids or other viral proteins, altered forms of viral DNA . . .) all of which may affect cell metabolism. In particular, the endosmolytic properties of adenovirus structural proteins, especially of penton (Horwitz er al., 1991), can explain the toxic effect due to massive cellular uptake of viral material. The same explanation can apply to the cytotoxicity observed after infection with a high-titre suspension of purified adenovirus . In conclusion, we have demonstrated that defective adenoviral vectors allow the transfer and expression of foreign genes into 80 - 100% of CNS-derived cells in cultures consisting of neurons and/or astroglial cells without any detectable alteration of the cell viability. Moreover, the transferred genes are expressed for prolonged periods in nondividing cells. Adenoviral vectors could thus be a valuable tool for the analysis of brain-specific gene regulation and to test the biochemical and physiological consequences of the expression of various genes in cultured brain cells. Consequently, adenovirus-mediated gene transfer could become an essential method in molecular neurobiology . The therapeutic use of brain transplantation of fetal tissues is being tested in Parkinson’s disease (Lindvall et al., 1990), and a variety of genetically modified cell lines have been transplanted into the CNS to serve as potential delivery systems for the transgene of interest (Wojcik er al., 1993). It could be interesting in the future to try to

improve the results of such trials by introducing the appropriate genes into cells before transplantation. These transgenes could encode neurotransmitter metabolism enzymes, neurotrophic factors or any other type of biologically active polypeptide. Obviously, adenovirus appears to be the vector of choice for genetic engineering of cell grafts before their transplantation into the CNS. While this manuscript was in preparation, results from our laboratory (Akli et al., 1993) and from three other groups (Le Gal La Salle et al., 1993; Davidson er al., 1993; Bajocchi er al., 1993) have demonstrated that adenoviral vectors are very efficient in transferring foreign genes into brain cells in vivo.In these experiments, recombinant adenovirus suspensions were applied by stereotaxy , resulting in stable expression of the transferred gene for >45 days. As in culture, cytotoxicity was observed with high viral titres only. The results in vivo corroborate the validity of the in vitro culture model system.

Acknowledgements We wish to thank Dr L. Eng and Dr L. Legault-Demare for generously providing antibodies. We are grateful to C. Gruszczynski for expert assistance in the preparation of cell cultures. This study was supported by the Institut National de la Santk et de la Recherche Maicale and the Centre National de la Recherche Scientifique, and by grants from the Association Franqaise contre. les Myopathies, the Direction des Recherches Etudes et Techniques, Vaincue les Maladies Lysosomales and the Caisse Nationale d’Assurance Maladie et Maternitk.

Abbreviations p .f. u . CNS GFAP PBS RSV LTR X-Gal

plaque- forming units central nervous system glial fibrillary acidic protein phosphate-buffered saline Rous sarcoma virus long terminal repeat 5-bromo-4-chloro-3-indolyl-~-~-galactos~de

References Akli, S., Caillaud, C., Vigne, E., Stratford-Perricaudet, L. D., Poenaru, L., Perricaudet, M., Kahn, A. and Peschanski, M. (1993) Transfer of foreign genes into the brain using adenovirus vectors. Nature Genetics, 3,224-228. Bajocchi, G., Feldman, S. H., Crystal, R. T. and Mastrangeli, A. (1993) Direct in vivo gene transfer to ependymal cells in the central nervous system using recombinant adenovirus vectors. Nature Genetics, 3, 229-234. Benvald-Netter, Y., Martin-Moutot, N., Koulakoff, A. and Couraud, F. (1981) Na+-channel-associated scorpion toxin receptor sites as probes for neuronal evolution in vivo and in virro. Proc. Natl. Acad. Sci. USA, 78, 1245- 1249. Chanock, R. M., Ludwig, W., Heubner, R. J., Cate, T. R. and Chu, L.-W. (1966) Immunization by selective infection with type 4 adenovirus grown in human diploid tissue culture. I. Safety and lack of oncogenicity and tests for potency in volunteers. J. Am. Med. Assoc., 195, 445 -452. Culver, W. C., Ram, Z., Wallbridge, S., Ishii, H.,Oldfield, E. H. and Blaese, M. (1992) In vivo gene transfer with retroviral vector-producer cells for treatment of experimental brain tumors. Science, 256, 1550- 1552. Couraud, F., Martin-Moutot, N., Koulakoff, A. and Berwald-Netter, Y.(1986) Neurotoxin-sensitive sodium channel in neurons developing in vivo and in vitro. J. Neurosci., 6, 192-198. Davidson, B. L., Allen, E. D., Kozarsky, K. F., Wilson, J. M. and Roessler, B. J. (1993) A model system for in vivo gene transfer into the central nervous system using an adenoviral vector. Nature Generics, 3, 219-223. Eng, L. F., Vanderhaeghen, J. J., Bignami, A. and Gerstl, B. (1971) An acidic protein isolated from fibrous astrocytes. Brain Res., 28, 351 -354. Gautron, S., Dos Santos, G., Pinto-Henrique, D., Koulakoff, A,, Gros, F. and Berwald-Netter, Y. (1992) The glial voltage-gated sodium channel: cell- and tissue-specific mRNA expression. Proc. Natl. Acad. Sci. USA, 89, 7272-7276. Graham, F. L., Smiley, J., Russel, W. C. and Nairu, R. (1977) Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J. Gen. Virol., 36, 59-72. Honvitz, M. S. (1991) Adenoviridae and their replication. In Fields, B. N.

Cultured CNS cells infected with an adenoviral vector and Knipe, D. M. (eds), Virology. Raven Press, New York, pp. 1679- 1721. Jaffe, H. A,, Danel, C., Longenecker, G., Metzger, M., Setoguchi, Y., Rosenfeld, M. A., Gant, T. W., Thorgeirsson, S. S., Stratford-Perricaudet, L. D., Perricaudet, M., Pavirani, A., Lecocq, J.-P. and Crystal, R. G. (1992) Adenovirus-mediated in vivo gene transfer and expression in normal rat liver. Nature Genetics, 1, 372-378. Le Gal La Salle, G., Robert, J. J., Berrard, S., Ridoux, V., StratfordPerricaudet, L. D., Perricaudet, M. and Mallet, J. (1993) An adenovirus vector for gene transfer into neurons and glia in the brain. Science, 259, 986 -988. Lindvall, O . , Brundin, P., Widner, H., Rehncrona, S., Gustavii, B., Frackowiak, R., Leenders, K. L., Sawle, G., Rothwell, J. C., Marsden, C. D. and BjorMund, A. (1990) Grafts of fetal dopamine neurons survive and improve motor function in Parkinson’s disease. Science, 247,574-577. Quantin, B., Perricaudet, L. D., Tajbakhsh, S. and Mandel, J. L. (1992) Adenovirus as an expression vector in muscle cells in vivo. Proc. Natl. Acad. Sci. USA,89, 2581-2584. Secchi, J., Lecaque, D., Cousin, M. A., Lando, D., Legault-Demare, L. and

1291

Raynaud, J. P. (1980) Detection and localization of 14-3-2 protein in primary cultures of embryonic rat brain. Brain Res., 184, 455 -466. Stratford-Perricaudet, L. D., Levrero, M., Chase, J. F., Perricaudet, M. and Briand, P. (1990) Evaluation of the transfer and expression in mice of an enzyme-encoding gene using a human adenovims vector. Hum. Gene &r., 1, 241-256. Stratford-Perricaudet,L. D., Makeh, I., Perricaudet, M. and Briand, P. (1992) Widespread long-term gene transfer to mouse skeletal muscles and heart. J. Clin. Invest., 90, 626-630. Straus, S. E. (1984) Adenovirus infections in humans. In Ginsberg, H. S. (ed.), The Adenovirus. Plenum Press, New York, pp. 45 1-496. Temin, H. (1986) Retrovirus vectors for gene transfer: efficient integration into and expression of exogenous DNA in vertebrate cell genomes. In Kucherlapati, R. (ed.), Gene Transfer. Plenum, New York, pp. 149-1237, Wojcik, B. E., Nothias, F., Lazar, M., Jouin, H., Nicolas, J. F. and Peschanski, M. (1993) Catecholaminergicneurons result from intracerebral implantationof embryonal carcinoma cells. Proc. Natl. Acad. Sci. US4,90, 1305 - 1309.