Characterization of Promoter Function and Cell ... - Journal of Virology

JOURNAL OF VIROLOGY, Dec. 2000, p. 11254–11261 0022-538X/00/$04.00⫹0 Copyright © 2000, American Society for Microbiology. All Rights Reserved.

Vol. 74, No. 23

Characterization of Promoter Function and Cell-Type-Specific Expression from Viral Vectors in the Nervous System R. L. SMITH,1 D. L. TRAUL,2 J. SCHAACK,1 G. H. CLAYTON,1 K. J. STALEY,1

AND

C. L. WILCOX2*

Department of Neurology and Pediatrics, University of Colorado Health Sciences Center, Denver, Colorado 80262,1 and Department of Microbiology, Colorado State University, Ft. Collins, Colorado 80523-16772 Received 9 June 2000/Accepted 30 August 2000

Viral vectors have become important tools to effectively transfer genes into terminally differentiated cells, including neurons. However, the rational for selection of the promoter for use in viral vectors remains poorly understood. Comparison of promoters has been complicated by the use of different viral backgrounds, transgenes, and target tissues. Adenoviral vectors were constructed in the same vector background to directly compare three viral promoters, the human cytomegalovirus (CMV) immediate-early promoter, the Rous sarcoma virus (RSV) long terminal repeat, and the adenoviral E1A promoter, driving expression of the Escherichia coli lacZ gene or the gene for the enhanced green fluorescent protein. The temporal patterns, levels of expression, and cytotoxicity from the vectors were analyzed. In sensory neuronal cultures, the CMV promoter produced the highest levels of expression, the RSV promoter produced lower levels, and the E1A promoter produced limited expression. There was no evidence of cytotoxicity produced by the viral vectors. In vivo analyses following stereotaxic injection of the vector into the rat hippocampus demonstrated differences in the cell-type-specific expression from the CMV promoter versus the RSV promoter. In acutely prepared hippocampal brain slices, marked differences in the cell type specificity of expression from the promoters were confirmed. The CMV promoter produced expression in hilar regions and pyramidal neurons, with minimal expression in the dentate gyrus. The RSV promoter produced expression in dentate gyrus neurons. These results demonstrate that the selection of the promoter is critical for the success of the viral vector to express a transgene in specific cell types. The ability to introduce foreign genes into nondividing cells, such as neurons, has benefited from the availability of viral vectors that include recombinant herpes simplex virus vectors (9, 17, 23, 26), herpes simplex virus-based amplicons (18, 19), adenovirus vectors (6, 7, 12, 29), vaccinia virus-based vectors (11, 24), adeno-associated virus vectors (2, 16, 21, 27), and lentivirus vectors (5, 37, 38). Virtually all of these vectors have provided the opportunity to modify gene function, although the efficiency of transduction, the cytotoxicity, and the patterns of gene expression vary considerably. Adenoviral vectors have been used extensively for gene transfer into neurons. The human cytomegalovirus (CMV) immediate-early promoter, the Rous sarcoma virus (RSV) long terminal repeat, and the adenovirus E1A promoter have been successfully used in adenovirus vectors to drive expression of foreign genes in neurons (6, 7, 12, 28, 29, 45). Direct comparison of the efficiency of the adenovirus-mediated gene transfer has been complicated by differences in the adenovirus vector background, the wide range of targeted cell types, and complications caused by immune responses in animal studies. Therefore, the extent to which the promoter determines the efficiency of transgene expression in the context of the viral vector remains largely unknown. The CMV and RSV promoters have been widely used to provide high levels of expression in many cell types (1, 3, 25, 30, 31, 47). It has generally been assumed that transgene expression obtained with either the CMV or the RSV promoter reflects the limitation of cell-typespecific infection. Based on published data regarding the use of these promoters in viral vectors, significant differences in ex* Corresponding author. Mailing address: Colorado State University, Department of Microbiology, Ft. Collins, CO 80523-1677. Phone: (970) 491-2552. Fax: (970) 491-1815. E-mail: [email protected] .edu.

pression would not be predicted but have not been compared directly in the same vector background. A third viral promoter, the E1A promoter from adenovirus, was also examined in these studies. The E1A promoter has been shown to have special utility for expression of toxic genes that cannot be incorporated into adenoviral vectors using either conventional or plasmid-based recombination in bacteria (28, 45). While these published studies suggest the utility of the E1A promoter, characterization of the functional properties of the E1A promoter has not been available to establish the general utility of this promoter. To facilitate comparison of vector-mediated gene expression, adenovirus vectors were constructed in an identical adenovirus vector background, Ad5dl327, which contains deletions in the E1 and E3 regions. Expression of two reporter genes from each of the promoters was also compared; vectors were constructed to express either ␤-galactosidase or the enhanced green fluorescence protein (EGFP). The use of EGFP allowed visualization of viable cells expressing the transgene over time. These adenovirus vectors were utilized to examine transgene expression in primary sensory neurons in culture, in vivo following stereotaxic injection of the dentate gyrus of the rat hippocampus, and in acutely prepared slices from the hippocampal region of adult rats. MATERIALS AND METHODS Construction of adenovirus vectors. Construction of the adenovirus vector encoding lacZ under the control of the CMV promoter was described previously (42). To construct viruses with lacZ or EGFP expression controlled by the E1A or RSV promoters and EGFP expression controlled by the CMV promoter, bacterial plasmid vectors containing the left end of the adenovirus chromosome through the XhoI site at bp 5788 were used. The E1A promoter plasmid (a generous gift of L.-J. Su and I. Maxwell, University of Colorado Health Sciences Center [UCHSC]) has the region between bp 502 and the BgllII site at bp 3328 replaced by a multiple cloning site. The HindIII-BamHI fragment of pON249 (44) containing the entire lacZ coding sequence and eukaryotic translational

11254

VOL. 74, 2000

CHARACTERIZATION OF VIRUS VECTORS IN THE NERVOUS SYSTEM

11255

FIG. 1. Histochemical detection of ␤-galactosidase expression from neuronal cultures infected with adenovirus vectors expressed under control of different viral promoters. DRG neuronal cultures were infected with each of the adenovirus vectors at an MOI of 50 PFU per cell. At the times indicated, the expression of ␤-galactosidase was examined using X-Gal histochemistry as described in Materials and Methods. Representative fields are shown. Similar results were obtained in three separate experiments.

initiation site was inserted in the sense orientation 3⬘ to the E1A promoter. To construct the plasmids containing the lacZ gene or EGFP gene (Clontech, Palo Alto, Calif.) driven by the RSV, the RSV promoter (a generous gift of R. Mahalingham, UCHSC) was cloned into a promoterless plasmid containing the left end of the adenovirus genome with a multiple cloning site replacing the adenovirus fragment from the SacII site at bp 357 to the BgllII site at bp 3328 (a generous gift of L.-J. Su and I. Maxwell, UCHSC). The pON249 fragment containing the lacZ gene was then inserted in the sense orientation 3⬘ to the RSV promoter. The E1A-lacZ, RSV-lacZ, E1A-EGFP, RSV-EGFP, and CMV-EGFP constructs were introduced into adenovirus by overlap recombination (10) in 293 cells (20) using the 2.5- to 100-map unit fragment of Ad5dl327BB (42) isolated by centrifugation on a sucrose gradient. Plaques were screened for the presence of ␤-galactosidase activity in the presence of the chromogenic substrate X-Gal (5-bromo-4-chloro-3-indolyl-␤-D-galactopyranoside) or for the presence of EGFP using epifluorescence. DNA from purified plaque preparations was characterized by its DNA endonuclease restriction pattern. Positive plaques, which contained the promoter-reporter gene construct replacing the adenovirus E1 region, were replication defective and, therefore, required growth and titer determination on the complementing 293 cells (20). The 293 cells were grown in Dulbecco modified Eagle medium containing high levels of glucose and supplemented with 10% bovine calf serum (Life Technologies). The viral vector stocks were prepared, the titers were determined, and the stocks were stored under essentially identical conditions. Preparation of neuronal cultures. Dorsal root ganglion (DRG) neuronal cultures were prepared and maintained as previously described (43). After two weeks in culture, DRG neurons were infected at a multiplicity of infection (MOI) of 50 PFU of adenoviral vector per cell for 1 h at 35°C. Mock-infected cultures were incubated with medium alone. At 1, 3, 7, 14, 28, and 35 days postinfection, the transgene expression was analyzed. ␤-Galactosidase activity was assessed qualitatively by histochemical staining with X-Gal or quantitatively in protein extracts using the ␤-galactosidase assay described below. Neuronal cultures were

examined qualitatively using epifluorescence to detect EGFP in living cultures or quantitatively for expression of EGFP in protein extracts using the fluorometeric assay described below. Neutral red assay for viability. Neuronal survival after adenovirus vector infection was assessed using a neutral red assay essentially as described by Greer and Shewen (22). The viability of cultures was examined as a function of neutral red uptake in infected cultures and compared to uninfected, age-matched control cultures. Briefly, medium was removed from the neuron cultures and replaced with 0.5 ml of a 1:10 dilution of a 3.33-mg/ml neutral red solution (Life Technologies) in phosphate-buffered saline (PBS, pH 7.4). After 1 h at 35°C, the neutral red solution was removed and the cultures were washed with 0.5 ml of PBS for 1 h at 35°C. Cells were lysed overnight before reading the absorbance at 540 nm. ␤-Galactosidase detection and quantification. For histochemical detection of ␤-galactosidase, neuronal cultures were washed once with PBS, fixed with 4% paraformaldehyde for 5 min, washed with PBS again, and reacted with X-Gal as described previously (41). To quantify ␤-galactosidase activity, neuronal cultures were harvested in 50 ␮l of lysis buffer and analyzed using the Galacton-Plus assay (Tropix, Bedford, Mass.). Briefly, 1 ␮l of each sample was diluted in 9 ␮l of lysis buffer containing 0.1 mM dithiothreitol and added to 100 ␮l of reaction buffer containing the Galacton substrate. After 1 h at 37°C, the accelerator buffer was added, and the luminescence was measured in a 20E luminometer (Turner Designs, Sunnyvale, Calif.). Luminescence was standardized to values obtained from reactions performed with known amounts of purified enzyme. The results are expressed as femtograms of ␤-galactosidase per microgram of total protein. Protein concentrations were determined using the BCA Protein Assay Kit (Pierce, Rockford, Ill.). In vivo dentate gyrus injections and analysis of lacZ expression. Stereotaxic injection of rats anesthetized with Metafane was performed using standard methods, targeting the dorsal limb of the dentate gyrus with weight-normalized coordinates (anterior-posterior [AP], ⫺3.3 mm; medial-lateral [ML], ⫺2.2 mm;

11256

SMITH ET AL.

FIG. 2. Quantitative measurement of ␤-galactosidase activity in neuronal cultures after infection with adenovirus vectors. ␤-Galactosidase activity was measured in extracts prepared from neuronal cultures infected with adenovirus vectors as described in Fig. 1. Using the assay for ␤-galactosidase as described in Materials and Methods, ␤-galactosidase activity was quantified using a standard curve. Samples were normalized to the total protein present in extracts, and the values are in femtograms of ␤-galactosidase per microgram of total protein at various times after infection. The data shown are the mean ⫾ the standard error of the mean (SEM) (n ⫽ 3).

J. VIROL. dorsal-ventral [DV], ⫺3.0 mm; average weight, 300 g) (34). Injection was performed using needles fabricated from glass pipettes filled with 1 to 1.5 ␮l of virus vector stock containing 105 PFU of virus. Virus was pressure injected over 10 min, and the pipette was left in place for another 5 min to facilitate the controlled delivery of virus. Acetaminophen (2 mg/ml) was provided ad libitum during the first 12 h of recovery from surgery. Three days after injection, animals were anesthetized with pentobarbitol and intracardic perfusion with 2% paraformaldehyde–PBS was performed. Brains were removed and coronally sectioned into 1-cm-thick slices. Thick slices were incubated in X-Gal buffer for 30 min at 37°C, followed by reaction in buffer containing X-Gal (1 mg/ml) (41). Tissue blocks were postfixed in 4% paraformaldehyde–PBS, cryoprotected in 20% sucrose, and cut as 25- to 40-␮m frozen sections to localize the X-Gal staining. In vitro brain slice preparations and inoculation with virus. Rat brain slices (400 ␮m thick) were prepared as described previously for acute experiments (46). The brain slices were prepared under aseptic conditions, and 2 ␮g of gentamicin per ␮l was added to the artificial cerebral spinal fluid to retard bacterial growth. Slices were maintained at the gas-medium interface in an atmosphere of 95% O2–5% CO2 at a temperature of 35°C. Infection with virus was performed by gently pipetting 5 ␮l of viral stock (108 PFU/ml) on the exposed surface of the dentate-hilar region. Photography and image preparation. Photographs were obtained using a Nikon FX-35DX camera on a Nikon Diaphot 200 microscope using 35-mm Ektachrome film (Eastman Kodak Co., Rochester, N.Y.). Slides were scanned with a Nikon LS-1000 Film Scanner, and the resulting images were prepared using Photoshop 5.0 software (Adobe Systems, Mountain View, Calif.).

RESULTS Many of the adenoviruses utilized for construction of recombinant viral vectors differ in the promoters used to express

FIG. 3. Fluorometric detection of EGFP expression from neuronal cultures infected with adenovirus vectors expressed under control of different viral promoters. DRG neuronal cultures were infected with each of the adenovirus vectors at an MOI of 50 PFU per cell. At the times indicated, the expression of EGFP was examined under fluorescence microscopy. Representative fields are shown. Similar results were obtained in three separate experiments.

VOL. 74, 2000


FIG. 4. Quantitative measurement of EGFP activity in neuronal cultures following infection with adenovirus vectors. EGFP expression was measured in extracts prepared from neuronal cultures at various times following infection with adenovirus vectors as described in Fig. 3. EGFP fluorescence was quantified using a standard curve. The data shown are the average relative fluorescence values (n ⫽ 2).

transgenes and in the viral vector background, especially in the E3 region. These differences have complicated the direct comparison of adenovirus vector functions, and the contribution of specific differences to vector function remains unknown. To facilitate the direct comparison of the functions of several promoters, a set of adenovirus vectors was constructed in the identical vector background, Ad5dl327, which has deletions in the E1 and E3 genes. The recombinant viral vectors were constructed to express either ␤-galactosidase or EGFP under control of the CMV, RSV, or E1A promoter. Comparison of the expression patterns of promoters using the recombinant adenovirus vectors in primary neurons in culture. Primary neuronal cultures prepared from embryonic rat DRG sensory neurons are efficiently infected with adenoviral vectors (45, 48). For the studies presented here, DRG neuronal cultures were inoculated at an MOI of 50 PFU per cell, which was previously shown to produce efficient infection of the neurons without apparent toxicity (45). Neurons expressing ␤-galactosidase were examined following infection using X-Gal histochemistry (Fig. 1). All of the vectors demonstrated transgene delivery to the neurons, with the highest expression from the CMV and RSV promoters. In DRG neuronal cultures, both the CMV and the RSV promoters appeared to produce similar levels of ␤-galactosidase expression and similar numbers of positively stained neurons, although with different time courses (Fig. 1). The histochemical detection of expression from the E1A promoter was significantly delayed and never attained the significant levels of expression observed from the CMV and RSV promoters. Quantitative measurement of ␤-galactosidase activities indicated that the levels of expression from the CMV, RSV, and E1A promoters varied by almost 2 orders of magnitude, a difference that was greater than was suggested by the histochemical methods (Fig. 2). The CMV promoter produced the highest level of function of the three promoters. While expression from the CMV peaked rapidly and remained high, the RSV promoter produced levels of ␤-galactosidase that continued to increase for several days in culture and subsequently decreased. These results were confirmed using EGFP as the reporter gene under the control of the CMV, RSV, or E1A promoter. The Ad-CMV-EGFP and Ad-RSV-EGFP produced fluorescence in virtually all the neurons, although with different time courses (Fig. 3). The E1A promoter produced little detectable

11257

evidence of EGFP expression at any of the times examined, a result consistent with reduced levels of expression from the lacZ gene. As shown in Fig. 4, direct measurements of EGFP fluorescence in cell lysates confirmed the pattern of expression determined using ␤-galactosidase as the reporter. Adenovirus vectors produce no detectable neuronal cell death. Previous studies utilized counts of cell profiles at early time points after infection as a measure of viral toxicity (45). To confirm and extend these analyses, vital staining using neutral red uptake was performed after viral infection of DRG neurons in culture to determine if the vectors produced measurable cytotoxicity or if expression of either reporter gene in the cultures resulted in neuronal cell death. No significant changes in neutral red uptake were observed following inoculation with the viral vectors compared to the uninfected control cultures (Fig. 5). These results indicate that no detectable neuronal cell death occurred as a result of the viral vector infection or reporter gene expression, even in neurons expressing very high levels of the transgene products over relatively long periods of time. Promoter-dependent ␤-galactosidase expression in the hippocampal region in vivo. Both the CMV and the RSV promoters have been used in vivo for gene transfer studies in the brain (12, 29). However, little information is available about the comparative functions of the two promoters, and analysis is complicated by the use of different viral vector backgrounds. To address these issues, the Ad-CMV-lacZ and Ad-RSV-lacZ vectors were used to transduce expression in brain structures in vivo. To allow direct comparison, small-volume (1 ␮l) inoculations of the dorsal limb of the dentate gyrus were performed using 105 PFU of adenoviral vector. The hippocampal region was selected for in vivo injection because of the presence of unique anatomic features that facilitate comparison of transduction. In addition, the hippocampal region has been extensively targeted for viral gene transfer because of its important role in epilepsy. Rat brain sections were examined histochemically for ␤-galactosidase after inoculation with the viral vector. The expression patterns were most distinct at 72 h postinjection (Fig. 6). These differences in the pattern of expression were clearly evident by 24 h postinjection and remained very

FIG. 5. Cell death is not detected following infection of neuronal cultures with the adenovirus vectors. Neuronal cultures were infected with the adenovirus vectors at an MOI of 50 PFU per cell. At the times indicated, the neuronal cultures were harvested and used in the neutral red viability assay as described in Materials and Methods. Neutral red uptake is shown as the percentage relative to the uninfected control neuronal cultures. The data shown are the means ⫾ the SEM (n ⫽ 3).

11258

SMITH ET AL.

J. VIROL.

FIG. 6. Adenovirus vector-mediated ␤-galactosidase expression in rat dentate gyrus in vivo. Adenovirus vectors constructed to express ␤-galactosidase under control of the RSV promoter (Ad-RSV-lacZ) or the CMV promoter (Ad-CMV-lacZ) were injected into the dorsal blade of the dentate gyrus. Representative views are shown of ␤-galactosidase histochemistry (blue), within the dentate gyrus, which was counterstained with cresyl violet (purple). (A) Ad-RSV-lacZ expressing ␤-galactosidase was detected within most classes of cells within the dentate region of the hippocampus, as indicated by the arrows. (B) Enlarged view of the boxed area in panel A. Note the cells throughout the upper blade of the dentate region, indicated by the bracket, expressing ␤-galactosidase. (C) Ad-CMV-lacZ expressing ␤-galactosidase was largely restricted to cells within the lower portions of the denate granule cell layer. (D) Enlarged view of the boxed area in panel C. Expression of ␤-galactosidase was evident primarily in the cells underlying the dentate gyrus, with a lack of labeled cells in the upper blade of the dentate gyrus, as indicated by the bracket. The bar shown in panel A is 500 ␮m and also applies to panel C. The bar shown in panel B is 100 ␮m and also applies to panel D; the asterisk indicates the hilar region of the hippocampus, and the arrows indicate the upper and lower blades of the dentate gyrus. Similar results were obtained in all three of the animals injected for each experimental group.

similar at 7 days postinjection (data not shown). With AdCMV-lacZ a small subpopulation of large cells localized to the subgranular zone of the dentate gyrus expressed ␤-galactosidase, although scattered cells within the granular cell layer were also labeled, confirming that the viral inoculum spread throughout dentate gyrus. Ad-RSV-lacZ produced greater expression in the granular cell populations, as well as in scattered large cells. Expression from the Ad-E1A-lacZ vector within the hippocampal region was not detected (data not shown). As reported in previous studies, inflammatory changes were evident in the hippocampus at early times after injection, as demonstrated by a diffuse infiltrate of small cells and damage to the cytoarchitecture of the granular cell layer. Although the analysis of the extent of the differences was complicated by the presence of inflammatory changes and cell loss triggered by the viral injection, these results suggest that differences exist in the neuronal cell type specificity of expression from the promoters. Hippocampal slice infections show that cell-type-specific expression is dependent upon the promoter used to drive expression of ␤-galactosidase. In vivo studies strongly suggested that marked differences in cell-type-specific expression were evident comparing the Ad-CMV-lacZ and Ad-RSV-lacZ. Ideally, analysis under conditions allowing control of inoculation and without the inflammatory response would greatly extend these

results. Hippocampal slice preparations have been used to examine viral vector function with several types of vectors (8, 32, 35, 36). Hippocampal slices acutely prepared from rat brain can be maintained for 24 h or longer in culture with maintenance of metabolic and electrophysiologic properties (13, 33, 40). The ability to maintain brain slices from mature animals for a period of greater than 24 h allowed the functions of the adenoviral vectors to be examined in mature neurons that maintain many of their in vivo connections and functions without an inflammatory response. Studies indicated that the neurons in the slice preparations were normal, as evidenced by the stability of the excitatory postsynaptic potentials and the population spikes recorded by stimulation of the lateral perforant path (data not shown). The hippocampal formation contains multiple neuronal cell types, which can be identified based on location and cell morphology. This provides a preparation for examining the properties of vector-mediated expression in different classes of adult neurons. Viral inoculations were performed after the slice preparations had stabilized in the incubation chamber for 1 h. Adenoviral vectors were introduced in a volume of 5 ␮l (109 PFU/ml) onto the surface of the individual slices, directly over the hippocampal formation, and allowed to diffuse over the slice passively. At 12 or 24 h after inoculation, the slices

VOL. 74, 2000


11259

FIG. 7. Adenovirus-vector-mediated ␤-galactosidase expression in rat hippocampal brain slices in vitro. Hippocampal slices were prepared from adult rat brain and infected with adenovirus vectors as described in Materials and Methods. At 24 h after infection with 5 ⫻ 107 PFU of each viral vector, hippocampal slices were processed for the histochemical detection of ␤-galactosidase. Regions of dark blue indicate ␤-galactosidase expression. Representative results are shown. (A and D) Ad-RSVlacZ-infected slice. (B and E) Ad-CMV-lacZ-infected slice. (C and F) Ad-E1A-lacZ-infected slice. (A to C) Low-power views of the hippocampal slices. The bar in panel A is 500 ␮m and applies to panels A to C. (D to F) High-power views of corresponding boxed regions. The bar in panel D is 100 ␮m and applies to panels D to F. The bracket indicates the upper blade of the dentate gyrus. Similar results were obtained in three separate experiments.

were processed for detection of ␤-galactosidase using X-Gal histochemistry. Only the upper, exposed cell layer of the slice showed ␤-galactosidase expression, suggesting that the virus was not able to penetrate beyond a single cell layer. CMV promoter-mediated expression was detected as early as 12 h after inoculation in a limited population of cells in the hilar region of hippocampus, whereas RSV promoter-mediated expression was not detected at this time (data not shown). By 24 h after infection with the Ad-CMV-lacZ vector, intense ␤-galactosidase expression in the hilus obscured identification of individual cells and included large regions of pyramidal cells extending to CA1 (Fig. 7B and E). However, expression in the granular cell layer was minimal. In contrast, the Ad-RSV-lacZ vector produced significant levels of ␤-galactosidase staining by 24 h after inoculation in the granular cell layer, which extended into the molecular layer. A more restricted staining pattern in the hilar region was observed and included many cell profiles that were readily identifiable as neuronal (Fig. 7A and D). Expression of ␤-galactosidase under the control of the E1A promoter was not detected over the time course of these experiments, a result consistent with the reduced levels of expression observed in the DRG neurons in culture (Fig. 7C and F). The Ad-RSV-lacZ demonstrated a pattern of infection in hippocampal slices in vitro that was similar to what was detected in vivo and confirmed published results (29). In contrast, the expression from the CMV promoter was markedly restricted both in vivo and in vitro in the granular cell layer. DISCUSSION Despite significant technical improvements, viral vector development presents technical challenges with the requirements for the appropriate expression of functional transgenes (28, 45). The choice of the promoter in a viral vector is one of the determinants of the overall utility of the vector. As demonstrated here, promoter selection can have important impli-

cations for transgene expression. The striking differences observed with the three promoters would not have been predicted based on the high activity of these promoters in several cell lines (unpublished observation). This strongly indicates the need for empiric examination of promoter function in the target cells. The increasing availability of promoter-reporter combinations in viral vectors will allow testing of the promoters in the target cells prior to construction of recombinant viral vectors for delivery of functional transgenes. The studies in the DRG neurons in tissue culture showed potentially important differences in the levels and duration of expression from each of the promoters that we compared. The CMV promoter clearly produced the highest levels of expression over the entire time course examined. The reason for the slight decrease in expression from the CMV and RSV promoters following the peak of expression is not clear, but it does not appear to be the result of the death of the neurons. The apparent heterogeneity in the levels of expression in the neuronal cultures may be the result of different copy numbers of the vectors in the cells; however, this most likely reflects the heterogeneous populations of neurons present in the DRG in vivo and in vitro (14, 15, 39). The differences in function of the CMV and RSV promoters were unexpected, especially with regard to the neuronal cell type specificity demonstrated by the in vivo injections and in the brain slices. In other studies, the CMV and RSV promoters have been shown to provide high levels of expression in many cell types (1, 3, 25, 30, 31, 47). The use of an identical adenovirus vector background allowed the demonstration that differences were at the level of the promoter and not related to differences in the viral vector background. The use of the identical adenovirus vectors expressing different reporter gene products provided a mechanism to determine if any observed differences reflected cell-type-specific differences in the mRNA levels, protein stability, or cytotoxicity of the reporter gene products. The function of the E1A promoter is an important issue,

11260

SMITH ET AL.

since intrinsic regulatory properties of this promoter have allowed construction of viral vectors containing toxic genes (28, 45). One of the utilities of the use of the E1A promoter in adenovirus vectors is based on the ability of the E1A gene product to repress the E1A promoter. This allows the packaging 293 cells line, which supplies the E1A gene product, to repress the E1A promoter and prevent expression of a toxic gene. This allows generation of the recombinant adenovirus vector in the 293 cells (28, 45). However, the E1A promoter demonstrated a nearly 100-fold-lower level of expression in the DRG neurons in culture compared to the CMV promoter, and expression from the E1A promoter was not detectable using X-Gal histochemistry in hippocampal brain slices. While the level of E1A expression was significantly reduced compared to the other promoters, it has been demonstrated to produce sufficient expression to produce functional gene products in neurons (28, 45). The E1A promoter provides low levels of expression that could have great utility for specific types of gene delivery. The hippocampal slice provides a useful preparation for comparison of differences in gene transfer to different cell types and classes of neurons. The hippocampal slice preparation is widely used as a model system for the examination of neuronal physiology. Rat hippocampal slices can be maintained for 24 h or longer to achieve viral-vector-mediated gene transfer and expression (13, 33, 40). The data presented here demonstrated that significant expression of transgenes was achieved within the time of viability of brain slices with use of the CMV or RSV promoters, although the cell type specificity of expression differed considerably. The restricted expression of the CMV promoter in dentate gyrus was also demonstrated by injection of the virus in vivo. Therefore, the cell-type-specific expression of the CMV promoter was not the result of conditions of the in vitro experiments. These promoter-specific effects are also likely to be relevant to other vector systems. Similar results were reported using an adeno-associated virus, showing that a CMV promoter driven construct did not express significant levels of ␤-galactosidase in granular neurons compared to hilar regions (4). Choice of the promoter will be an important consideration for the design of viral vectors to target expression in hippocampal cells and possibly other cell types.

J. VIROL.

8.

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

24.

ACKNOWLEDGMENTS This work was supported by National Institutes of Health grants NS 29046 (C.L.W.) and NS 01741 (R.L.S.) and by a USDA grant (C.L.W.).

25.

REFERENCES

26.

1. Acsadi, G., A. Jani, B. Massie, M. Simoneau, P. Holland, K. Blaschuk, and G. Karpati. 1994. A differential efficiency of adenovirus-mediated in vivo gene transfer into skeletal muscle cells of different maturity. Hum. Mol. Genet. 3:579–584. 2. Alexander, I. E., D. W. Russell, A. M. Spence, and A. D. Miller. 1996. Effects of gamma irradiation on the transduction of dividing and nondividing cells in brain and muscle of rats by adeno-associated virus vectors. Hum. Gene Ther. 7:841–850. 3. Baldwin, H. S., C. Mickanin, and C. Buck. 1997. Adenovirus-mediated gene transfer during initial organogenesis in the mammalian embryo is promoterdependent and tissue-specific. Gene Ther. 4:1142–1149. 4. Bartlett, J. S., R. J. Samulski, and T. J. McCown. 1998. Selective and rapid uptake of adeno-associated virus type 2 in brain. Hum. Gene Ther. 9:1181– 1186. 5. Blomer, U., L. Naldini, T. Kafri, D. Trono, I. M. Verma, and F. H. Gage. 1997. Highly efficient and sustained gene transfer in adult neurons with a lentivirus vector. J Virol. 71:6641–6649. 6. Boviatsis, E. J., M. Chase, M. X. Wei, T. Tamiya, R. K. Hurford, Jr., N. W. Kowall, R. I. Tepper, X. O. Breakefield, and E. A. Chiocca. 1994. Gene transfer into experimental brain tumors mediated by adenovirus, herpes simplex virus, and retrovirus vectors. Hum. Gene Ther. 5:183–191. 7. Caillaud, C., S. Akli, E. Vigne, A. Koulakoff, M. Perricaudet, L. Poenaru, A.

27. 28.

29. 30. 31. 32. 33.

Kahn, and Y. Berwald-Netter. 1993. Adenoviral vector as a gene delivery system into cultured rat neuronal and glial cells. Eur. J. Neurosci. 5:1287– 1291. Cassacia-Bonnefil, P., E. Bendikz, H. Shen, A. Stelzer, D. Edelstein, M. Geschwind, B. M., H. J. Federoff, and P. J. Bergold. 1993. Localized gene transfer into organotypic hippocampal slice cultures and acute hippocampal slices. J. Neurosci. Methods. 50:341–351. Chang, J. Y., E. M. Johnson, Jr., and P. D. Olivo. 1991. A gene delivery/recall system for neurons which utilizes ribonucleotide reductase-negative herpes simplex viruses. Virology 185:437–440. Chinnadurai, G., S. Chinnadurai, and J. Brusca. 1979. Physical mapping of a large plaque mutation of adenovirus type 2. J. Virol. 32:623–628. Cook, D. G., R. S. Turner, D. L. Kolson, V. M. Lee, and R. W. Doms. 1996. Vaccinia virus serves as an efficient vector for expressing heterologous proteins in human NTera 2 neurons. J. Comp. Neurol. 374:481–492. Davidson, B. L., E. D. Allen, K. F. Kozarsky, J. M. Wilson, and B. J. Roessler. 1993. A model system for in vivo gene transfer into the central nervous system using an adenoviral vector. Nat. Genet. 3:219–223. Djuric, B., R. Berger, and W. Paschen. 1994. Protein synthesis and energy metabolism in hippocampal slices during extended recovery following different periods of ischemia. Metab. Brain Dis. 9:377–389. Dodd, J., and T. M. Jessel. 1985. Lactoseries carbohydrates specify subsets of dorsal root ganglia neurons projecting to the superficial dorsal horn of rat spinal cord. J. Neurosci. 5:3278–3294. Dodd, J., D. Solter, and T. M. Jessel. 1984. Monoclonal antibodies directed against carbohydrate differentiation antigen identify subsets of primary sensory neurons. Nature 311:469–472. Du, B., P. Wu, D. M. Boldt-Houle, and E. F. Terwilliger. 1996. Efficient transduction of human neurons with an adeno-associated virus vector. Gene Ther. 3:254–261. Fink, D. J., L. R. Sternberg, P. C. Weber, M. Mata, W. F. Goins, and J. C. Glorioso. 1992. In vivo expression of beta-galactosidase in hippocampal neurons by HSV-mediated gene transfer. Hum. Gene Ther. 3:11–19. Freese, A., A. I. Geller, and R. Neve. 1990. HSV-1 vector-mediated neuronal gene delivery. Strategies for molecular neuroscience and neurology. Biochem. Pharmacol. 40:2189–2199. Geller, A. I., and X. O. Breakefield. 1988. A defective HSV-1 vector expresses Escherichia coli beta-galactosidase in cultured peripheral neurons. Science 241:1667–1669. Graham, F. L., J. Smiley, R. W. C., and R. Narin. 1977. Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J. Gen. Virol. 36:59–72. Grant, C. A., S. Ponnazhagan, X. S. Wang, A. Srivastava, and T. Li. 1997. Evaluation of recombinant adeno-associated virus as a gene transfer vector for the retina. Curr. Eye Res. 16:949–956. Greer, C. N., and P. E. Shewen. 1986. Automated colorimetric assay for the detection of Pasteurella haemolytica leucotoxin. Vet. Microbiol. 12:33–42. Ho, D. Y., E. S. Mocarski, and R. M. Sapolsky. 1993. Altering central nervous system physiology with a defective herpes simplex virus vector expressing the glucose transporter gene. Proc. Natl. Acad. Sci. USA 90:3655– 3659. Hsu, H., E. Huang, X. C. Yang, A. Karschin, C. Labarca, A. Figl, B. Ho, N. Davidson, and H. A. Lester. 1993. Slow and incomplete inactivations of voltage-gated channels dominate encoding in synthetic neurons. Biophys. J. 65:1196–1206. Johnson, L. G., R. J. Pickles, S. E. Boyles, J. C. Morris, H. Ye, Z. Zhou, J. C. Olsen, and R. C. Boucher. 1996. In vitro assessment of variables affecting the efficiency and efficacy of adenovirus-mediated gene transfer to cystic fibrosis airway epithelia. Hum. Gene Ther. 7:51–59. Johnson, P. A., K. Yoshida, F. H. Gage, and T. Friedmann. 1992. Effects of gene transfer into cultured CNS neurons with a replication-defective herpes simplex virus type 1 vector. Brain Res. Mol. Brain Res. 12:95–102. Jomary, C., K. A. Vincent, J. Grist, M. J. Neal, and S. E. Jones. 1997. Rescue of photoreceptor function by AAV-mediated gene transfer in a mouse model of inherited retinal degeneration. Gene Ther. 4:683–690. Kuhn, T. B., M. D. Brown, C. L. Wilcox, J. A. Raper, and J. R. Bamburg. 1999. Myelin and collapsin-1 induced motor neuron growth cone collapse through different pathways: inhibition of collapse by opposing mutants of rac1. J. Neurosci. 19:1965–1975. Le Gal La Salle, G., J. J. Robert, S. Berrard, V. Ridoux, L. D. StratfordPerricaudet, M. Perricaudet, and J. Mallet. 1993. An adenovirus vector for gene transfer into neurons and glia in the brain. Science 259:988–990. Lee, A. H., Y. S. Suh, J. H. Sung, S. H. Yang, and Y. C. Sung. 1997. Comparison of various expression plasmids for the induction of immune response by DNA immunization. Mol. Cell 7:495–501. Meyrelles, S. S., R. V. Sharma, C. A. Whiteis, B. L. Davidson, and M. W. Chapleau. 1997. Adenovirus-mediated gene transfer to cultured nodose sensory neurons. Brain Res. Mol. Brain Res. 51:33–41. Ozaki, M., K. Matsumura, S. Kaneko, M. Satoh, Y. Watanabe, and T. Aoyama. 1993. A vaccinia virus vector for efficiently introducing genes into hippocamal slices. Biochem. Biophys. Res. Commun. 193:653–660. Pakhoten, P. I., A. B. Belousov, and N. A. Otmakhov. 1990. Functional

VOL. 74, 2000

34. 35. 36. 37.

38. 39. 40. 41.


stability of the brain slices of ground squirrels kept in conditions of prolonged deep periodic hypothermia. Neuroscience 38:591–598. Paxinos, G., and C. Watson. 1986. The rat brain in stereotaxic coordinates. Academic Press, Inc., San Diego, Calif. Pettit, D. L., T. Koothan, D. Liao, and R. Malinow. 1995. Vaccinia virus transfection of hippocampal slice neurons. Neuron 14:685–688. Pettit, D. L., S. Perlman, and R. Malinow. 1994. Potentiated transmission and prevention of further LTP by increased CaMKII activity in postsynaptic hippocampal slice neurons. Science 266:1881–1885. Poeschla, E., J. Gilbert, X. Li, S. Huang, A. Ho, and F. Wong-Staal. 1998. Identification of a human immunodeficiency virus type 2 (HIV-2) encapsidation determinant and transduction of nondividing human cells by HIV-2based lentivirus vectors. J. Virol. 72:6527–6536. Poeschla, E. M., F. Wong-Staal, and D. J. Looney. 1998. Efficient transduction of nondividing human cells by feline immunodeficiency virus lentiviral vectors. Nat. Med. 4:354–357. Price, J. 1985. An immunohistochemical and quantitative examination of dorsal root ganglion neuronal subpopulations. J. Neurosci. 5:2051–2059. Reid, K. H., H. L. Edmonds, A. Schurr, M. T. Tseng, and C. A. West. 1988. Pitfalls in the use of brain slices. Prog. Neurobiol. 31:1–18. Sanes, J., J. Rubenstein, and J.-F. Nicolas. 1986. Use of a recombinant retrovirus to study post-implantation cell lineage in mouse embryos. EMBO J. 5:3133–3142.

11261

42. Schaack, J., S. Langer, and X. Guo. 1995. Efficient selection of recombinant adenoviruses by vectors that express beta-galactosidase. J. Virol. 69:3920– 3923. 43. Smith, R. L., and C. L. Wilcox. 1996. Studies of herpes simplex virus 1 latency using primary neuronal cultures of dorsal root ganglion neurons, p. 221–231. In P. Lowenstein and L. Enquist (ed.), Protocols for gene transfer in neuroscience: towards gene therapy of neurological disorders. John Wiley & Sons, Ltd., Sussex, England. 44. Spaete, R., and E. Mocarski. 1987. Insertion and deletion mutagenesis of the human cytomegalovirus genome. Proc. Natl. Acad. Sci. USA 84:7213–7217. 45. Staley, K., R. Smith, J. Schaack, C. Wilcox, and T. J. Jentsch. 1996. Alteration of GABAA receptor function following gene transfer of the CLC-2 chloride channel. Neuron 17:543–551. 46. Staley, K. J., T. Otis, and I. Mody. 1992. Membrane properties of dentate gyrus granule cells comparison of sharp microelectrode and whole-cell recordings. J. Neurophysiol. 67:1346–1358. 47. Teramoto, S., T. Matsuse, E. Ohga, T. Nagase, Y. Fukuchi, and Y. Ouchi. 1997. Kinetics of adenovirus-mediated gene transfer to human lung fibroblasts. Life Sci. 61:891–897. 48. Wilcox, C. L., R. L. Smith, R. D. Everett, and D. Mysofski. 1997. The herpes simplex virus type 1 immediate-early protein ICP0 is necessary for the efficient establishment of latent infection. J. Virol. 71:6777–6785.