Preexisting antiadenoviral immunity is not a barrier to efficient and stable transduction of the brain, mediated by novel high-capacity adenovirus vectors

The utility of first-generation adenovirus vectors for long-term gene transfer in humans is limited by preexisting antiadenoviral immunity. We demonstrate here that new-generation high-capacity adenovirus vectors (HC-Ads) can efficiently transduce the brain and mediate stable transgene expression for at least 2 months, even in the presence of a preexisting antiadenoviral immune response. First-generation vector-mediated transduction was almost completely abolished in preimmunized animals within 60 days of the vector injection. Levels of HC-Ad-mediated transduction by 3 days postinjection were not significantly affected by preimmunization, were reduced within 14 days to 56% of those levels seen in nonimmunized animals, and remained stable until day 60 postinjection. Acute brain inflammation elicited by the HC-Ad vector injection was more transient, and was reduced in intensity compared with brain inflammation elicited by the first-generation vector injection in immunized animals. Inflammation was significantly higher in all immunized animals than in nonimmunized animals. Our results show that preexisting antiadenoviral immunity does not significantly reduce initial HC-Ad-mediated infection of the brain and is not a barrier to stable HC-Ad vector-mediated transduction of the CNS. Although input HC-Ad capsid proteins injected into the brain may contain transient targets for a brain-infiltrating cellular adenovirus-specific immune response, this fails to eliminate transgene expression. Thus HC-Ads show promise for gene therapy of chronic brain disease.     

Systemic mannitol-induced hyperosmolality amplifies rAAV2-mediated striatal transduction to a greater extent than local co-infusion.

Recombinant viral vectors derived from adeno-associated virus serotype 2 (rAAV2) have been investigated as highly effective vehicles for gene transfer to the central nervous system (CNS). Transduction with rAAV2 vectors results in long-term transgene expression in CNS neurons. Optimal injection parameters leading to efficient targeting and spread of the transgene to large neuroanatomical regions are important in molecular gene therapy studies of the CNS. In the present study, we reexamined the effects of both local and systemic administration of mannitol-induced hyperosmolality on facilitation of transgene expression and vector distribution in the CNS. Systemic intraperitoneal administration of mannitol prior to vector administration improved gene transfer to striatal neurons, increasing the total number of transduced cells by 400% and vector distribution by 200%. On the other hand, local coadministration of mannitol in the striatum increased the number of transduced striatal neurons by 25% and had little effect on transduction volume. In conclusion, we have demonstrated that systemic coadministration of mannitol significantly enhances transgene spread of rAAV2 viral delivery in the brain to a much greater degree than local coadministration.     

Lentiviral vectors encoding HIV-1 polyepitopes induce broad CTL responses in vivo.

Lentiviral vectors have been tested as vaccination vectors in anti-tumoral and anti-viral models. They efficiently transduce dendritic cells and stimulate strong T-cell responses against the encoded antigen. However, their capacity to stimulate a cytotoxic T-lymphocyte (CTL) response against several antigens has not been evaluated. Broad anti-human immunodeficiency virus 1 (HIV-1) T-cell immune responses are important for the control of HIV replication. We evaluated the potential of polyepitope-encoding lentiviral vectors to induce broad anti-HIV CTL responses. We constructed two lentiviral vectors coding for an HLA-A2- or HLA-B7-restricted polyepitope and evaluated their immunogenicity by direct injection of vector particles in HLA-A2 or HLA-B7 transgenic mice. In vitro cytotoxicity assays showed that a single immunization induces a strong, diversified, and long-lasting CTL response in both mouse models. CTL responses were directed against all 13 epitopes in the HLA-A2 system and 8 out of 12 in the HLA-B7 system. A second immunization augmented the number of responding mice in the HLA-A2 system but not in the HLA-B7 system. HLA-B7-immunized mice mounted strong interferon-gamma (IFN-gamma)-secreting T-cell responses against a majority of the epitopes and lysed peptide-loaded target cells in vivo. CTL responses in HLA-B7 mice were only partially dependent on CD4 T-cell help. This work underlines the potential of lentiviral vectors as candidates for therapeutic vaccination against acquired immunodeficiency syndrome.     

Evaluation of Tet-on system to avoid transgene down-regulation in ex vivo gene transfer to the CNS

Ex vivo gene transfer to the CNS has so far been hampered by instability of transgene expression. To avoid the phenomenon of transgene down-regulation, we have employed strong, constitutive promoters and compared this expression system with the inducible Tet expression system incorporated in a single plasmid vector or in lentiviral vectors. Plasmid-based transgene expression directed by the constitutive, human ubiquitin promoter, UbC, was stable in transfected HiB5 cells in vitro and comparable in strength to the CMV promoter. However, after transplantation of UbC and CMV HiB5 clones to the rat striatum, silencing of the transgene occurred in most cells soon after implantation of transfected cells. The Tet-on elements were incorporated in a single plasmid vector and inducible HiB5 clones were generated. Inducible clones displayed varying basal expression activity, which could not be ascribed to an effect of cis-elements in the vector, but rather was due, at least in part, to intrinsic activity of the minimal promoter. Basal expression activity could be blocked in a majority of cells by stable expressing the transrepressor tTS. Fully induced expression levels were comparable to CMV and UbC promoters. Similar to the constitutive promoters transgene expression was down-regulated soon after grafting of inducible HiB5 clones to the rat striatum. Lentiviral vectors can direct long-term stable in vivo transgene expression. To take advantage of this quality of the lentiviral vector, the Tet-on elements were incorporated in two lentiviral transfer vectors followed by transduction of Hib5 cells. Interestingly, all HiB5 clones established by lentiviral transduction showed very similar expression patterns and tight regulatability that apparently was independent of transgene copy number and integration site. Nevertheless, transgene expression in all lentiviral HiB5 clones was down-regulated shortly after transplantation to the rat striatum. These results confirm the general phenomenon of transgene down-regulation. Moreover, the results suggest that the considerable advantages offered by lentiviral vectors for direct gene delivery cannot necessarily be transferred directly to ex vivo gene delivery. This emphasizes the need for alternative vector strategies for ex vivo gene transfer.     

Lentiviral vectors pseudotyped with baculovirus gp64 efficiently transduce mouse cells in vivo and show tropism restriction against hematopoietic cell types in vitro

The envelope glycoprotein from vesicular stomatitis virus (VSV-G) has been used extensively to pseudotype lentiviral vectors, but has several drawbacks including cytotoxicity, potential for priming of immune responses against transgene products through efficient transduction of antigen-presenting cells (APCs) and sensitivity to inactivation by human complement. As an alternative to VSV-G, we extensively characterized lentiviral vectors pseudotyped with the gp64 envelope glycoprotein from baculovirus both in vitro and in vivo. We demonstrated for the first time that gp64-pseudotyped vectors could be delivered efficiently in vivo in mice via portal vein injection. Following delivery, the efficiency of mouse cell transduction and the transgene expression is comparable to VSV-G-pseudotyped vectors. In addition, we found that gp64-pseudotyped lentiviral vectors could efficiently transduce a variety of cell lines in vitro, although gp64 showed a more restricted tropism than VSV-G, with especially poor ability to transduce hematopoietic cell types including dendritic cells (DCs). Although we found that gp64-pseudotyped vectors are also sensitive to inactivation by human complement, gp64 nevertheless has advantages over VSV-G, because of its lack of cytotoxicity and narrower tropism. Consequently, gp64 is an attractive alternative to VSV-G because it can efficiently transduce cells in vivo and may reduce immune responses against the transgene product or viral vector by avoiding transduction of APCs such as DCs.     

Efficient gene transfer into the CNS by lentiviral vectors purified by anion exchange chromatography

Lentiviral vectors have been shown to stably transduce dividing and non-dividing target cells in vitro and in vivo. However, in vivo gene transfer applications with viral vectors in the central nervous system require highly efficient vector preparations, because only very small volumes can be injected stereotactically without damage to the brain tissue. Since lentiviral vectors are generated in transient co-transfection systems, viral preparations need to be purified and efficiently concentrated before injection into the brain. We describe an alternative procedure to concentrate lentiviral preparations by binding viral particles to an anion exchange column. Viral particles are eluted with sodium chloride, desalted and further concentrated by ultrafiltration. These vector preparations allowed high levels of gene transfer into terminally differentiated neuronal and glial cells and long-term transgene expression without any signs of acute and long-term toxicity or inflammation. The purification of lentiviral vectors from large-scale preparations by anion exchange chromatography allowed us to concentrate the virus to small volumes and to use these preparations to genetically modified target cells in vivo without signs of acute inflammatory responses.     

Immune Responses to AAV in Canine Muscle Monitored by Cellular Assays and Noninvasive Imaging

We previously demonstrated that direct intramuscular injection of rAAV2 or rAAV6 in wild-type dogs resulted in robust T-cell responses to viral capsid proteins, and others have shown that cellular immunity to adeno-associated virus (AAV) capsid proteins coincided with liver toxicity and elimination of transgene expression in a human trial of hemophilia B. Here, we show that the heparin-binding ability of a given AAV serotype does not determine the induction of T-cell responses following intramuscular injection in dogs, and identify multiple epitopes in the AAV capsid protein that are recognized by T cells elicited by AAV injection. We also demonstrate that noninvasive magnetic resonance imaging (MRI) can accurately detect local inflammatory responses following intramuscular rAAV injection in dogs. These studies suggest that pseudotyping rAAV vectors to remove heparin-binding activity will not be sufficient to abrogate immunogenicity, and validate the utility of enzyme-linked immunosorbent spot (ELISpot) assay and MRI for monitoring immune and inflammatory responses following intramuscular injection of rAAV vectors in preclinical studies in dogs. These assays should be incorporated into future human clinical trials of AAV gene therapy to monitor immune responses.     

Adeno-associated virus-mediated gene transfer to the brain: duration and modulation of expression

Adeno-associated virus (AAV) is a promising vector for central nervous system (CNS) gene transfer, but a number of issues must be addressed if AAV is to be used for widespread delivery throughout the CNS. Our aim was to test the effect of dose, route of delivery, and hydroxyurea treatment on brain expression of beta-galactosidase activity after cerebral inoculation with an rAAV-lacZ vector (rAAV-beta-gal). We also wished to test whether an immune response appeared against the vector and the transgene product. We found in BALB/c mice that beta-Gal expression increased during the first 2 months after inoculation, then decreased slightly by 4 months, and continued out to 6, 12, and 15 months in single animals. Cerebral injection produced localized beta-Gal expression that did not diffuse to other regions despite a fivefold increase in injection volume. Intraventricular injection resulted in negligible transduction. Antibodies to AAV capsid protein and beta-Gal appeared at low levels at 2 and 4 months, but correlated poorly with beta-Gal expression and did not prevent readministration of rAAV-beta-gal. Hydroxyurea treatment did not result in increased transduction in vivo. We conclude that our study confirms rAAV vectors as having considerable potential for CNS gene transfer; however, several important problems must be addressed if this vector system is to be used for long-term transduction of the entire brain. Sustained, regulatable expression will be needed if rAAV is to be used in the treatment of chronic CNS disease. The difficulty in delivering AAV to diverse regions of the brain is an important problem that must be overcome if these vectors are to be used for anything beyond localized transduction.     

Absence of an intrathecal immune reaction to a helper-dependent adenoviral vector delivered into the cerebrospinal fluid of non-human primates

Inflammation and immune reaction, or pre-existing immunity towards commonly used viral vectors for gene therapy severely impair long-term gene expression in the central nervous system (CNS), impeding the possibility to repeat the therapeutic intervention. Here, we show that injection of a helper-dependent adenoviral (HD-Ad) vector by lumbar puncture into the cerebrospinal fluid (CSF) of non-human primates allows long-term (three months) infection of neuroepithelial cells, also in monkeys bearing a pre-existing anti-adenoviral immunity. Intrathecal injection of the HD-Ad vector was not associated with any sign of systemic or local toxicity, nor by signs of a CNS-specific immune reaction towards the HD-Ad vector. Injection of HD-Ad vectors into the CSF circulation may thus represent a valuable approach for CNS gene therapy allowing for long-term expression and re-administration.     

Striatal delivery of rAAV-hAADC to rats with preexisting immunity to AAV.

We tested the hypotheses that initial immunization of rats with rAAV might limit subsequent transduction by rAAV-hAADC when stereotaxically infused into the striatum and that the level of inhibition would correlate with AAV neutralizing antibody titers. Immunohistochemical detection of AADC and analysis by stereology revealed that the control group (no immunization) had the greatest volume of distribution of AADC (20.32 +/- 2.03 mm3) (+/-SD). There was a 58% decrease in spread (8.46 +/- 3.67 mm3, P < 0.008) in the high-dose immunization group (5 x 10(10) vg rAAV-null). Transduction weakly correlated with preexisting titer levels of neutralizing antibody at the time of intrastriatal rAAV-hAADC infusion. Only rats with neutralizing antibody titers of 1:1208 +/- 332 had significantly decreased AADC transgene expression compared to the unimmunized control group. Immunohistochemistry on serial sections for inflammatory markers including GFAP, CD11b, CD4, and CD8a revealed normal morphology and no cellular infiltration, suggesting little immune reaction in the CNS. We conclude that rAAV vectors can transduce brain tissue in the context of preexisting immunity, but that efficiency of transduction declines significantly in the presence of very high titers of neutralizing antibodies. These results have important implications for gene therapy for CNS disorders.     

Transient Pretreatment With Glucocorticoid Ablates Innate Toxicity of Systemically Delivered Adenoviral Vectors Without Reducing Efficacy

More than 300 human clinical trials utilize recombinant adenoviruses (rAds) as a gene transfer vector, confirming that rAds continue to be of high clinical interest. A primary weakness of rAds is their known propensity to trigger an innate, proinflammatory immune response rapidly after high-dose, systemic administration. In this study, we investigated what affects that pre-emptive treatment with anti-inflammatory glucocorticoids might have upon Ad vector-triggered inflammatory immune responses. We found that a simple pretreatment regimen with Dexamethasone (DEX) can significantly reduce most Ad-induced innate immune responses. DEX prevented rAd induction of systemic cytokine/chemokine releases in a dose-dependent fashion, with higher dosages preventing rAd induction of acute thrombocytopenia, endothelial cell activation, proinflammatory gene induction, and leukocyte infiltration into transduced organs. Transient glucocorticoid pretreatment also significantly reduced rAd-induced adaptive immune responses, including a decreased induction of Ad-neutralizing antibodies (NAbs). Importantly, use of DEX did not reduce the efficacy of rAd-mediated gene transduction nor rAd-derived transgene expression. Our results demonstrate that a simple, pre-emptive and transient glucocorticoid pretreatment is a viable approach to reduce rAd-associated acute toxicities that currently limit the use of Ad vectors in systemic clinical applications.     

Lentiviral gene transduction of kidney

Gene transfer into kidney holds great potential as a novel therapeutic approach. We have studied the transduction of kidney in vivo after delivery of lentiviral vectors by various routes of administration. A lentiviral vector expressing the bacterial lacZ gene from the cytomegalovirus early promoter was used. The lentiviral vector was delivered into the kidneys of BALB/c mice by retrograde infusion into the ureter, by injection into the renal vein or artery, or by direct injection into the renal parenchyma. Expression of the reporter gene was achieved independently of the route of administration, although it appeared more efficient after parenchymal or ureteral administration. After parenchymal or ureteral infusion, expression of the transgene was localized to the outer medulla and corticomedullary junction. In the case of parenchymal injection, expression of the reporter gene extended to the cortex. Detection of the transgene in the renal proximal tubules was confirmed by in situ polymerase chain reaction after parenchymal or ureteral infusion. On delivery of the lentiviral vector through the renal artery or vein, expression of the reporter gene was markedly lower than was observed with parenchymal or ureteral infusion and was limited to the inner medullary collecting ducts. No apparent histological abnormality was observed after virus administration and transgene expression was stable for at least 3 months. These results provide the first evidence that lentiviral vectors can stably transduce renal cells in vivo and may be effective vehicles for gene delivery to the kidney.     

PEGylation of E1-deleted adenovirus vectors allows significant gene expression on readministration to liver

Systemic administration of adenoviral vectors leads to activation of innate and antigen-specific immunity. In an attempt to diminish T and B cell-specific immune responses to E1-deleted adenoviral vectors, capsid proteins were modified with various activated monomethoxypolyethylene glycols (MPEGs). The impact of this modification was studied in a murine model of liver-directed gene transfer in which an E1-deleted adenovirus expressing the lacZ gene was given intravenously. The efficiency of vector transduction of hepatocytes in vivo was not compromised by any of the polymer chemistries. PEGylation of the virus, however, diminished the activation of cytotoxic T lymphocytes and helper T cells of the type 1 subset (Th1 cells) against native viral antigens; neutralizing antibodies to native virus were also diminished. PEGylation prolonged transgene expression and allowed partial readministration with native virus or with a virus PEGylated with a heterologous chemical moiety. Apparently, modification of the capsid leads to a shift in antigenic epitopes because vector readministration was not possible when the immunizing vector had been modified by the same PEGylation chemistry used to modify the second vector. In light of these results, the concept of improving the performance of adenoviral vectors through modification of the capsid with PEG shows promise.     

Readministration of helper-dependent adenoviral vectors to mouse airway mediated via transient immunosuppression

The efficacy of adenovirus-mediated gene therapy is attenuated by the host immune responses to both vector and transgene products. Even for helper-dependent adenoviral (HD-Ad) vectors, which have all viral-coding sequences deleted, the viral capsid proteins still cause immune reactions. In order to improve the efficiency in transgene expression during HD-Ad vector readministration, we administered cyclophosphamide to transiently modulate the mouse immune system. We delivered a high dose (5 × 10(10) vector particles (vp) per mouse) of empty HD-Ad to the mouse airway to induce an initial immune response. After 4 weeks, the mice were readministered with an HD-Ad vector containing either the reporter gene, LacZ, or the gene for the human cystic fibrosis transmembrane conductance regulator (CFTR) (1.5 × 10(10) vp per mouse). We found that the expression of both transgenes was greatly improved by the administration of cyclophosphamide when compared with the expression in mice without the immunosuppressing drug. We also found that the high dose of the empty HD-Ad vector administered intranasally does not induce an acute systemic immune response, but it does elicit an acute local response of proinflammatory cytokine production. Antibodies against Ad vector, including the neutralizing antibodies, were greatly reduced by the presence of cyclophosphamide in vector readministratiton. Moreover, cyclophosphamide reduced the infiltration of inflammatory cells, including total leukocytes, lymphocytes, CD4+ and CD8+T cells. These results indicate that transient administration of immunosuppressive agent can be used to extend transgene expression as well as attenuating immunogenicity to HD-Ad vectors in airway readministration.     

Progress and challenges in viral vector-mediated gene transfer to the brain

Gene transfer into the brain allows the manipulation of transgene expression in both time and space. Recently developed gene transfer technologies allow transgenes to be expressed in any anatomically, biochemically or functionally distinct group of brain cells. Gene transfer has been used to alter the expression of neurotransmitter receptors, ion channels, signaling proteins, neuronal growth, differentiation and survival factors, and thus to modify brain anatomy, neuron physiology, behavior and pathology. However, challenges remain in making gene therapy a more widespread tool for the treatment of neurological disease. We have identified the following as areas needing development: access and delivery of viral vectors to the brain; diffusion of viral vectors and transgenes throughout large areas of brain tissue; viral vector side effects and toxicity, inflammatory and immune responses to vectors; long-term stable transgene expression; cell type-specific expression of transgenes; and the ability of the experimenter or physician to switch transgene expression 'on' and 'off' at will. In the last year, neuro-gene therapy has shown that brain defects in experimental disease models can be prevented and corrected, and that viral vectors and encoded transgenes can be made to diffuse over larger brain areas. In addition, the cause of vector-induced inflammation and immune responses have begun to be elucidated, so that rational approaches can be developed to avoid these complications. Further improvements in viral vectors will facilitate clinical trials in the near future.     

Optimized lentiviral vector production and purification procedure prevents immune response after transduction of mouse brain

HIV-derived lentiviral vectors are efficient vehicula to deliver genes into the brain and hold great promise for future gene therapy of neurodegenerative disorders. However, administration of the current vector preparations in mouse brain was found to induce a systemic immune response to vector proteins and a modest inflammation in the brain. Moreover, serum antibodies from vector-treated animals were capable of partially neutralizing lentiviral vector-mediated transduction in cell culture. To avoid this unexpected immune reaction, we have optimized new vector production and purification protocols. Purification by sucrose gradient ultracentrifugation abolished the immune response, but vector titers also decreased substantially. Lentiviral vector production in the absence of serum in the cell culture medium equally reduced immunogenicity without affecting transduction efficiency. These results have important implications for future clinical use of lentiviral vectors, and for the use of lentiviral vectors to create animal models for neurodegenerative diseases that have an important neuroinflammatory component.     

IL10 Released by a New Inflammation-regulated Lentiviral System Efficiently Attenuates Zymosan-induced Arthritis

Administration of anti-inflammatory cytokines is a common therapeutic strategy in chronic inflammatory diseases. Gene therapy is an efficient method for delivering therapeutic molecules to target cells. Expression of the cell adhesion molecule E-selectin (ESEL), which is expressed in the early stages of inflammation, is controlled by proinflammatory cytokines, making its promoter a good candidate for the design of inflammation-regulated gene therapy vectors. This study describes an ESEL promoter (ESELp)-based lentiviral vector (LV) that drives localized transgene expression during inflammation. Mouse matrigel plug assays with ESELp-transduced endothelial cells showed that systemic lipopolysaccharide (LPS) administration selectively induces ESELp-controlled luciferase expression in vivo. Inflammation-specific induction was confirmed in a mouse model of arthritis, showing that this LV is repeatedly induced early in acute inflammation episodes and is downregulated during remission. Moreover, the local acute inflammatory response in this animal model was efficiently blocked by expression of the anti-inflammatory cytokine interleukin-10 (IL10) driven by our LV system. This inflammation-regulated expression system has potential application in the design of new strategies for the local treatment of chronic inflammatory diseases such as cardiovascular and autoimmune diseases.