Overcoming cellular immunity to prolong adenoviral-mediated gene expression in sciatic nerve.

In a previous report, we demonstrated that a first generation (E1- and E3-deleted) recombinant adenovirus can transduce expression of the E. coli lacZ gene into Schwann cells, both in vitro and in vivo, suggesting that this method might be useful for future therapy of peripheral neuropathy, including CMT1. Adenoviral-mediated gene transfer was limited, however, by demyelination and Wallerian degeneration at the site of virus injection, as well as by attenuation of viral gene expression over time. In our current work we have optimized adenoviral-mediated gene expression after intraneural injection into sciatic nerve. Using an improved injection protocol, peak expression of lacZ occurs between 10 and 14 days after injection of 2-week-old animals, decreases thereafter, and there is minimal associated tissue injury. In contrast, very few adenoviral-infected Schwann cells are found in nerves of adult animals 10 days after injection, probably due to immune clearance of viral-infected cells. Consistent with this notion, high levels of lacZ are found in sciatic nerve 30 days after injection of adult SCOD mice, which have a genetic defect in both cellular and humoral immunity, of adult beta 2 microglobulin-deficient mice (beta 2 M-/-), which have a genetic defect in cellular immunity, or of adult mice treated with the immunosuppressing agent FK506. In addition, adenoviral-infected Schwann cells co-cultured with axons in vitro, in the absence of a host immune response, ensheath axons and express lacZ for at least 8 weeks. These data thus demonstrate that expression of first generation recombinant adenovirus in sciatic nerve in adult mice, as in other tissues, is limited mainly by the host cellular immune response to the virus, which can be overcome by attenuation of host cell-mediated immunity. Adenoviral vectors might thus be used to modulate Schwann cell gene expression in patients with peripheral neuropathy after appropriate immunosuppression.

Modulation of cell-mediated immunity prolongs adenovirus-mediated transgene expression in sciatic nerve.

In a previous report, we demonstrated that a first-generation (E1- and E3-deleted) recombinant adenovirus can transduce expression of the E. coli lacZ gene into Schwann cells, both in vitro and in vivo, suggesting that this method might be useful for future therapy of peripheral neuropathy, including CMT1. Adenovirus-mediated gene transfer was limited, however, by demyelination and Wallerian degeneration at the site of virus injection, as well as by attenuation of viral transgene expression over time. In our current work we have optimized adenoviral vector-mediated transgene expression after intraneural injection into sciatic nerve. Using an improved injection protocol, peak expression of lacZ occurs between 10 and 14 days after injection of 2-week-old rats, decreases thereafter, and there is minimal associated tissue injury. In contrast, few lacZ-expressing Schwann cells are found in nerve of adult animals 10 days after injection, probably owing to immune clearance of virus-infected cells. Consistent with this notion, high levels of LacZ are found in sciatic nerve 30 days after injection of adult SCID mice, which have a genetic defect in both cellular and humoral immunity, of adult beta2-microglobulin-deficient mice (beta2M4-/-), which have a genetic defect in cellular immunity, or of adult mice treated with the immunosuppressing agent FK506. In addition, adenovirus-infected Schwann cells cocultured with axons in vitro, in the absence of a host immune response, ensheathe axons and express lacZ for at least 8 weeks. These data thus demonstrate that lacZ transgene expression of first-generation recombinant adenovirus in sciatic nerve in adult mice, as in other tissues, is limited mainly by the host cellular immune response to the virus, which can be overcome by attenuation of host cell-mediated immunity. Adenoviral vectors might thus be used to modulate Schwann cell gene expression in patients with peripheral neuropathy after appropriate immunosuppression.

Laminin alpha2 muscular dystrophy: genotype/phenotype studies of 22 patients.

OBJECTIVE: To determine the number of primary laminin alpha2 gene mutations and to conduct genotype/phenotype correlation in a cohort of laminin alpha2-deficient congenital muscular dystrophy patients. BACKGROUND: Congenital muscular dystrophies (CMD) are a heterogeneous group of muscle disorders characterized by early onset muscular dystrophy and a variable involvement of the CNS. Laminin alpha2 deficiency has been reported in about 40 to 50% of cases of the occidental, classic type of CMD. Laminin alpha2 is a muscle specific isoform of laminin localized to the basal lamina of muscle fibers, where it is thought to interact with myofiber membrane receptor, such as integrins, and possibly dystrophin-associated glycoproteins. METHODS: Seventy-five CMD patients were tested for laminin alpha2 expression by immunofluorescence and immunoblot. The entire 10 kb laminin alpha2 coding sequence of 22 completely laminin alpha2-deficient patients was screened for causative mutations by reverse transcription (RT)-PCR/single strand conformational polymorphisms (SSCP) analysis and protein truncation test (PTT) analysis followed by automatic sequencing of patient cDNA. Clinical data from the laminin alpha2-deficient patients were collected. RESULTS: Thirty laminin alpha2-negative patients were identified (40% of CMD patients tested) and 22 of them were screened for laminin alpha2 mutations. Clinical features of laminin alpha2-deficient patients were similar, with severe floppiness at birth, delay in achievement of motor milestones, and MRI findings of white matter changes with normal intelligence. Loss-of-function mutations were identified in 95% (21/22) of the patients studied. SSCP analysis detected laminin alpha2 gene mutations in about 50% of the mutant chromosomes; PTT successfully identified 75% of the mutations. A two base pair deletion mutation at position 2,096-2,097 bp was present in 23% of the patients analyzed. CONCLUSIONS: Our data suggest that the large majority of laminin alpha2-deficient patients show laminin alpha2 gene mutations.

Interferons impair early transgene expression by adenovirus-mediated gene transfer in muscle cells.

Recombinant adenovirus (AVR) promises to be an efficient vector in gene therapy for neuromuscular diseases, but in preclinical experiments the expression of therapeutic genes is shorter lived in immunocompetent animals than in immunocompromised hosts. Interferons (IFN), which are known to have a role both in early antiviral activity and in late cytotoxic immunoreaction against the virus or transduced cells, may influence the efficiency of gene transfer. In this study we investigated the role of IFNs in determining the efficiency of gene transfer by AVR. AVRs expressing beta-galactosidase (beta-gal) from either a cytomegalovirus (CMV) or a troponin-I promoter were used. Muscle cells were infected by AVR after exposure to various IFNs. The alphaIFN treatment significantly reduced (up to fivefold) the CMV promoter-driven gene expression in muscle cells in vitro and in immature muscles in vivo, while the least effective inhibitor was betaIFN. The decrease in gene expression by IFNs was more pronounced with the CMV-driven transgene than troponin-I promoter-driven one and was due to a decrease in transcript level. Intrinsic IFNs that are triggered by AVR administration can decrease the efficiency of gene transfer in muscle cells. Therefore the use of muscle specific promoters in AVR and/or IFN inhibitory agents will likely improve the prospects of effective gene therapy by AVR.

Generation, validation, and large scale production of adenoviral recombinants with large size inserts such as a 6.3 kb human dystrophin cDNA.

Human, serotype 5 (Ad 5), replication-defective recombinant adenoviruses (AdVs) expressing a 6.3 kb partial dystrophin cDNA (Becker) under the control of either the CMV early or the RSV LTR promoter/enhancer in combination with various polyadenylation sequences (polyA), were developed for gene transfer studies aimed at Duchenne muscular dystrophy. Based on previous experience, a strategy for generation, screening and validation of AdVs with relatively large size gene expression cassette inserts was established. Here we focus on some aspects of stability and safety of such AdVs as gene therapeutic tools based on relevant molecular biological methods. Furthermore, the quality of our best AdV-minidystrophin construct was validated following its large scale production and purification as well as its delivery in mdx mice. These results are of interest for establishing other AdVs, where the combined length of a tissue specific promoter, the gene of interest and the polyA sequences reach the upper limit of the packaging capacity of first generation AdVs.

Toxicity of replication-defective adenoviral recombinants in dissociated cultures of nervous tissue.

Replication-defective human type 5 adenoviral recombinants (AVR) are very efficient means of introducing foreign genes into neurons in vitro and in vivo; however, a significant reduction in the number of cells expressing reporter genes has been reported to occur over time. In vitro, this may be due to direct toxicity of the protein product of the transgene or adenoviral molecules. In vivo, in addition, an immune attack by the host could eliminate the transduced cells. To assess the direct toxicity of AVR or reporter gene products, a quantitative study of survival of transduced neurons over a period of 4 weeks was conducted in primary neural cultures. Cultures of dissociated murine spinal cord-dorsal root ganglia were exposed to AVR containing the Escherichia coli lacZ (E. coli lacZ) gene under control of either the very efficient cytomegalovirus enhancer/promoter or the fast muscle troponin I promoter, which is not active in these cells. Two factors contributed to loss of neuronal and nonneuronal cells: (i) direct toxicity of (E1 + E3)-deleted replication-incompetent AVR at high titers [> or = 5 x 10(8) viral particles/ml or multiplicity of infection (m.o.i.) 1000] and (ii) high levels of expression of the reporter gene product, beta-galactosidase, at titers that result in 55-75% transduction efficiency (5 x 10(7)-5 x 10(8) viral particles/ml or m.o.i. 100-1000). Despite the efficacy of adenoviral vectors in introducing foreign genes into primary, postmitotic cells, specific precautions must be taken in their use because of the narrow margin between concentrations of recombinants that transduce a sufficient percentage of cells and those that are cytotoxic.

Phenotypic and immunologic factors affecting plasmid-mediated in vivo gene transfer to rat diaphragm.

Little is known about the molecular mechanisms governing adaptive responses of the diaphragm in the setting of lung disease. By permitting the study of regulatory elements and the effects of overexpressing genes of interest, direct in vivo gene transfer to the diaphragm could be used as a tool to address such questions. Therefore, we evaluated parameters affecting transfection efficiency and duration of foreign gene expression in the diaphragm after plasmid-mediated gene transfer. Reporter gene constructs were injected into adult rat diaphragm and hindlimb muscles. Transfection efficiency at 8-10 days postinjection was decreased in large caliber ( > 1,000 microns2) and type II myosin heavy chain (MHC)-expressing fibers. There were also strong trends toward augmented transfection efficiency in type I MHC- and embryonic MHC-expressing fibers. All diaphragms demonstrated evidence of muscle injury and inflammatory cell infiltrates at this early time point. By 30 days postinjection, however, neither inflammation nor reporter gene expression was detectable in diaphragm or hindlimb muscles of immunocompetent animals. By contrast, immunosuppressed rats (given cyclosporine; 15 mg.kg-1. day-1) showed high levels of foreign gene expression at 30 days postinjection, which remained stable up to 60 days. Therefore, exploitation of plasmid-mediated in vivo gene transfer as a tool for studying regulated gene expression in the diaphragm may be facilitated by the use of immunodeficient animal models.

Dystrophin expression in muscles of mdx mice after adenovirus-mediated in vivo gene transfer.

We have generated high-titer adenoviral recombinants (AVR) expressing a 6.3-kb partial dystrophin cDNA insert under the control of either the Rous sarcoma virus (RSV) or cytomegalovirus (CMV) promoter. These AVR preparations were free of both E1-containing AVR and AVR with a nonfunctional dystrophin expression cassette. With these optimal AVR preparations, we have obtained a high degree of short-term (10 days) expression of a truncated (approximately 200 kD) dystrophin in dystrophin-deficient mdx muscles injected in the neonatal period; a lesser degree of expression of dystrophin was found in muscles injected in the young adult age and in old animals. Microscopic indices of muscle damage revealed that the truncated dystrophin provided a significant protection of the transduced muscle fibers. However, by 60 days post-injection, a substantial reduction of the number of dystrophin-positive fibers was noted, even in the neonatally injected muscles, and near-total elimination of dystrophin-positive fibers occurred in muscles injected in the adult age. These effects appeared to be brought about by the activity of CD8+ cytotoxic lymphocytes directed against the transduced cells, leading to their eventual elimination. In severe combined immunodeficiency (SCID) mice, lacking both humoral and cellular immune competence, muscles transduced (either in the neonatal or adult age) by AVR containing a CMV-LacZ expression cassette maintained the early (10 day) transduction level up to 30 days post-injection. Systemic administration of AVR (i.e., into the left ventricle of the heart) led in 5 days to a high number of dystrophin-positive fibers in heart, diaphragm, and intercostal muscles but not in limb muscles.

Efficiency and functional consequences of adenovirus-mediated in vivo gene transfer to normal and dystrophic (mdx) mouse diaphragm.

The protein dystrophin is absent in muscles of patients with Duchenne muscular dystrophy (DMD) as well as in mdx mice. The mdx mouse diaphragm closely resembles the human DMD phenotype and should serve as an appropriate model for future studies of dystrophin gene replacement. In this regard, recombinant adenovirus (AV) holds great promise as a vector for delivering a functional dystrophin gene to muscle. However, the use of AV is hampered by the development of an immune response against transduced cells, resulting in short-lived transgene expression as well as possible adverse effects on organ function. In the present study, sensitive reporter genes were employed to determine the efficiency and functional consequences of AV-mediated gene transfer to the diaphragm in both normal and mdx adult mice. One week after direct intramuscular injection of AV into the diaphragm, the level of transgene expression was significantly increased in mdx compared with normal diaphragms. In addition, small-caliber fibers (< 500 microns2) demonstrated preferential transduction in both groups of mice. Normal diaphragms receiving AV exhibited a substantial reduction in maximal twitch and tetanic force generation, whereas no significant effect on diaphragm contractility was found in the mdx group at 1 wk after injection. At 1 mo after AV administration, however, there was a significant decrease in force production by both normal and mdx diaphragms. Immunosuppression with cyclosporine A over 1 mo did not augment the level of transgene expression, but a beneficial effect on diaphragm force-generating capacity was observed in both groups of animals. We conclude the following: (1) short-term transduction of the diaphragm is more efficient in mdx than in normal mice; (2) AV leads to reduced force production by the diaphragm, with this effect being more pronounced in normal than in mdx in the early (but not the late) postinjection period; and (3) immunosuppressive therapy with cyclosporine has a partially protective effect on muscle function after AV administration, which is apparently unrelated to sparing of transduced fibers from elimination by the host immune system. These findings have important implications for the application of AV-mediated dystrophin gene transfer to the treatment of DMD.

Differential short-term transduction efficiency of adult versus newborn mouse tissues by adenoviral recombinants.

We demonstrated different transduction efficiency in several major organs of the immature (newborn) versus mature (adult) mice using adenoviral recombinants containing expression cassettes for either firefly luciferase or bacterial beta-galactosidase reporter genes. The studied tissues included skeletal muscle, heart, brain, lung, kidney, and liver. The transduction efficiency in all tissues, especially skeletal muscle, was significantly less in adults than in newborns, with two exceptions. In the heart, transduction efficiency was the same in newborns and adults, while in brain, it was greater in the adult than in the newborn. The cited differences in transduction efficiencies between newborn and adult tissues applied approximately equally to both reporter genes. The alpha v integrin level showed the same trend as the transduction efficiency in all tissues, except the heart. Polymerase chain reaction showed a specific adenoviral product in proportion to the reporter gene expression in muscle, heart, and brain. The results of this study should be considered in designing gene therapy strategies in genetic diseases.

The route of administration is a major determinant of the transduction efficiency of rat tissues by adenoviral recombinants.

One of the key factors that determines the efficacy of adenovirus-mediated gene therapy in genetic diseases, is the degree and extent of transduction of the target cells by adenovirus (AV)-recombinants carrying the therapeutic gene or cDNA. In this paper we provide experimental evidence which indicates that the route of administration of the AV-recombinants has a major influence on the transduction of various tissues in young rats. The heart, diaphragm, intercostal muscles and thymus show high transduction after intra-arterial (left cardiac ventricle) injection. By contrast, the liver shows a high transduction after intravenous injection. A substantial viremia develops within 2 h of gastric-rectal, intraperitoneal and intracardiac administration of AV recombinants. The number of adenoviral DNA copies per nucleus of transduced cells ranged from one to three in most tissues. These numbers correlated well with the overall transduction efficiency of the tissue determined by reporter gene expression. The various factors that determine which route of administration favors a high transduction rate in a particular tissue can be analyzed and this can lead to an improved efficiency of gene therapy in targeting a particular tissue in a disease.

Emergence of early region 1-containing replication-competent adenovirus in stocks of replication-defective adenovirus recombinants (delta E1 + delta E3) during multiple passages in 293 cells.

Early region 1 (E1)-deleted human adenovirus (AV) recombinants have been shown to be powerful tools of gene transfer in vivo and in vitro and are considered for application in human gene therapy. We could detect increasing titers of E1-containing adenovirus in two independent E1 + E3-deleted recombinant AV stocks during multiple passages in 293 cells, most likely due to a recombinant event with the host cell genome. We show the deleterious effects of this E1-containing, mostly replication-competent AV subpopulation in vivo and compare different screening methods of AV stocks for its detection. These considerations are important for the safety of human gene therapy trials.

The principles of gene therapy in Duchenne muscular dystrophy.

Somatic cell gene therapy (GT) for genetic disease such as Duchenne muscular dystrophy entails the introduction of normal, or at least functionally adequate, alleles of a gene into target cells for correction or mitigation of deleterious consequences of the disease's characteristic mutation. The following factors have a major impact upon the efficiency of GT: the artificial gene construct, the promoter, the delivery system, and the mode of dissemination of the therapeutic genes. For skeletal muscles, replication-defective adenovirus represents an efficient delivery system, but only if immature muscle cells are abundant in the muscle. The major drawback of adenoviruses is that the maximal insert capacity is only about 7.5 kb, which is only sufficient to accommodate a dystrophin minigene (6.3 kb) with a constitutive promoter. These and many other problems still require solution in experimental animals before single-muscle pilot studies of GT can be undertaken for such human muscle diseases as Duchenne muscular dystrophy.

Cultured human myoblasts and myotubes show markedly different transducibility by replication-defective adenovirus recombinants.

Human adenovirus (AV) is a favored vector for delivery of therapeutic genes into certain target cells, such as skeletal muscle cells for gene therapy. Here we show that replication-defective (E1 + E3 deleted) human type 5 adenovirus (AV) recombinants containing a reporter gene insert (RSV-luciferase or RSV-Lux) can very efficiently transduce cultured human myoblasts. However, transduction efficiency is about one order of magnitude less in cultured myotubes 16 days postfusion. The high transduction of myoblasts by AV-RSV-Lux could be effectively blocked by an arginine-glycine-asparagine (RGD) oligopeptide that serves as a ligand for the natural internalization receptor of AV. The normalized level of beta 3/beta 5-integrin, the main component of the internalization receptor for AV is about three times as abundant in myoblasts than in myotubes. This could contribute, among other things, to the relatively high susceptibility of myoblasts to AV infection and AV-mediated gene transduction.

Gene transfer into skeletal muscles by isogenic myoblasts.

The best way to overcome immunorejection in heterologous myoblast transfer (HMT) is by the use of immunodeficient and/or highly immunosuppressed mice as hosts. The same may be attained by autologous myoblast transfer (AMT). In this paper, we describe myoblast transfer in mdx and normal mice where the donor myogenic cells originated from highly inbred litter mates that are considered to be isogenic and thus the procedure is analogous to AMT. The myoblasts were marked in vitro with Rous Sarcoma Virus (RSV)-luciferase (Lux) or RSV-beta-galactosidase (LacZ) reporter genes through transduction mediated by an autonomously replication-defective recombinant human adenovirus. This permitted us to follow their fate after transplantation. mdx and normal mice were irradiated with 20 Gray gamma rays; necrosis and regeneration were induced by intramuscular notexin prior to myoblast injection. In both mdx and normal mice, the expression of luciferase rapidly declined after the injection implying that a large portion of the injected myoblasts were lost by 48 hr, due to undetermined cause(s). The surviving, injected myoblasts well-mosaicized large groups of host fibers but only in the immediate vicinity of the injection. Substantial expression of the reporter gene continued up to 1 month post-transplantation in normal mice, but there was a gradual decline and eventual disappearance of the reporter gene expression in mdx mice. This latter phenomenon was due to the ongoing intense necrosis of muscle fibers in mdx. There was no evidence of immunorejection. These experiments indicate that even in the absence of immunorejection, myoblast transfer suffers from important negative features: major loss of myoblasts within 48 hr after the injection and lack of significant spread of the injected cells from the injection site in the host muscle. These factors, plus the limited proliferative and fusion capacity of Duchenne muscular dystrophy (DMD) myoblasts, make them less than an ideal vector for the dystrophin cDNA for dystrophin gene replacement therapy in DMD.

A differential efficiency of adenovirus-mediated in vivo gene transfer into skeletal muscle cells of different maturity.

High titre (10(11)-10(12) pfu/ml) suspensions of autonomously replication-defective type 5 human adenovirus (AV) recombinants with different reporter gene inserts (CMV-Luciferase (Lux), CMV-beta-galactosidase (Lac Z), RSV-Lux and RSV-Lac Z) were injected into intact quadriceps muscles of 1-5 day old (Group 1) or 35-45 day old (Group 2) normal mice, as well as regenerating adult mouse muscles (Group 3) and 35 day old mdx muscles (Group 4). The expression of the reporter genes was quantitated 10 days and 2 months later. At 10 days postinjection all reporter gene expression was very high in the neonatally injected (Group 1) muscles. In Group 2 muscles the transduction was markedly less. In Group 3 muscles the gene expression was significantly better than in the Group 2 muscles. In adult mdx muscles (Group 4) where spontaneous regeneration is usually present, the results were similar to those in Group 3 animals. At 2 months post-injection in Group 1 animals, the RSV-Lux expression was even higher than at 10 days postinjection. The cell surface density of alpha v-integrin-containing molecules including the internalization receptor for AV in Groups 1, 2, 3 and 4 showed a positive correlation with AV transducibility. We conclude that adenovirus vector in high titre (10(10) pfu/ml or above) is capable of efficiently transducing only immature muscle cells but not mature muscle fibers in vivo and this appears to correlate with a higher surface density of the available AV internalization receptor in immature muscle cells and lower level in mature muscle fibers.(ABSTRACT TRUNCATED AT 250 WORDS)

The potential for gene therapy in Duchenne muscular dystrophy and other genetic muscle diseases.

Dystrophin cDNAs have been introduced into skeletal muscle fibers of dystrophin-deficient mice (mdx) through direct DNA injection in plasmid expression vectors and by replication-defective recombinant adenovirus vectors. The introduced genes appear to protect those muscle fibers from necrosis in which they become expressed. By direct injection of dystrophin cDNA in plasmid expression vector, only 1-2% of adult mdx muscle fibers of the injected muscle expressed dystrophin. On the other hand, by recombinant adenovirus injection into very young mdx muscle, a better efficiency has been reported. We have discussed several putative and proven factors that may contribute to the thus far demonstrated relatively low efficiency of dystrophin gene transfer. These include poor uptake of gene constructs by muscle fibers, degradation of the injected DNA, and poor access of gene constructs to the nuclear compartment. Neutralization or elimination of these factors could improve the efficiency of gene transfer so that it might, in the future, qualify as an effective therapy for DMD and some other genetic diseases of muscle.

Long-term persistence of plasmid DNA and foreign gene expression in mouse muscle.

Plasmid pRSVL persisted and expressed luciferase for at least 19 months in mouse skeletal muscle after intramuscular injection. Other injected plasmids also stably expressed long-term suggesting that any plasmid DNA could stably persist and express in muscle. Plasmid DNA was demonstrated by quantitative PCR in some of the muscle DNA samples for at least 19 months after injection. The methylation pattern of the plasmid DNA remained in its bacterial form indicating that the foreign DNA did not replicate in the muscle cells. The electroporation of total cellular DNA from injected muscles into bacteria indicated that the plasmid DNA was extrachromosomal. Chromosomal integration of plasmid DNA was searched for by electroporating the injected muscle DNA into bacteria after restriction enzyme digestion and ligation. No plasmids containing plasmid/chromosome junctions were observed in over 1800 colonies examined. Lack of integration increases the theoretical safety of this gene transfer technique. Long-term stability of plasmid DNA in muscle indicates that muscle is an attractive target tissue for the introduction of extrachromosomal plasmid or viral DNA for the purpose of gene therapy.

Persistence of plasmid DNA and expression in rat brain cells in vivo.

The purpose of this study was to determine whether plasmid DNA is able to persist in nondividing or slowly dividing brain cells in vivo. A new cationic lipid formulation which contains 70 mol% of DOTMA (N[1-(2,3-dioleyloxy)propyl]-N, N, N-trimethylammonium) and 30 mol% of cholesterol was used to transfect reporter genes into fetal brain cells in culture that were then transplanted into adult host brains. Gene expression was localized both to glial and neuronal cells after transfection of fetal brain cells with pRSVLac-Z, the gene coding for Escherichia coli beta-galactosidase protein. After the transfection of pRSVL plasmid which contains the firefly luciferase gene into fetal brain cells that were transplanted, substantial amounts of luciferase and pRSVL DNA were present in the host brains for 1 to 2 months. These results have implications for intracerebral viral infections and gene therapy of brain disorders.