Jaagsiekte sheep retrovirus pseudotyped lentiviral vector-mediated gene transfer to fetal ovine lung.

Viral vector-mediated gene transfer to the postnatal respiratory epithelium has, in general, been of low efficiency due to physical and immunological barriers, non-apical location of cellular receptors critical for viral uptake and limited transduction of resident stem/progenitor cells. These obstacles may be overcome using a prenatal strategy. In this study, HIV-1-based lentiviral vectors (LVs) pseudotyped with the envelope glycoproteins of Jaagsiekte sheep retrovirus (JSRV-LV), baculovirus GP64 (GP64-LV), Ebola Zaire-LV or vesicular stomatitis virus (VSVg-LV) and the adeno-associated virus-2/6.2 (AAV2/6.2) were compared for in utero transfer of a green fluorescent protein (GFP) reporter gene to ovine lung epithelium between days 65 and 78 of gestation. GFP expression was examined on day 85 or 136 of gestation (term is ∼145 days). The percentage of the respiratory epithelial cells expressing GFP in fetal sheep that received the JSRV-LV (3.18 × 10(8)-6.85 × 10(9) viral particles per fetus) was 24.6±0.9% at 3 weeks postinjection (day 85) and 29.9±4.8% at 10 weeks postinjection (day 136). Expression was limited to the surface epithelium lining fetal airways <100 µm internal diameter. Fetal airways were amenable to VSVg-LV transduction, although the percentage of epithelial expression was low (6.6±0.6%) at 1 week postinjection. GP64-LV, Ebola Zaire-LV and AAV2/6.2 failed to transduce the fetal ovine lung under these conditions. These data demonstrate that prenatal lung gene transfer with LV engineered to target apical surface receptors can provide sustained and high levels of transgene expression and support the therapeutic potential of prenatal gene transfer for the treatment of congenital lung diseases.

Early gestational gene transfer with targeted ATP7B expression in the liver improves phenotype in a murine model of Wilson's disease.

The ideal gene therapy for metabolical liver disorders would target hepatocytes before the onset of disease and be durable, non-toxic and non-immunogenic. Early gestational gene transfer can achieve such goals. Here, we demonstrate that prenatal gene transfer of human Atp7b reduces liver pathology and improves biochemical markers in Atp7b(-/-) mice, a murine model of Wilson's disease (WD). Following prenatal injection of lentivirus vector containing the human Atp7b gene under the transcriptional control of a liver-specific promoter, the full-length ATP7B was detectable in mouse livers for the entire duration of experiments (20 weeks after birth). In contrast to a marked pathology in non-injected animals, livers from age-matched treated mice consistently demonstrated normal gross and histological morphology. Hepatic copper content was decreased in the majority of treated mice, although remaining copper levels varied. Improvement of hepatic copper metabolism was further apparent from the presence of copper-bound ceruloplasmin in the sera and normalization of the mRNA levels for HMG CoA-reductase. With this approach, the complete loss of copper transport function can be ameliorated, as evident from phenotypical improvement in treated Atp7b(-/-) mice. This study provides proof of principle for in utero gene therapy in WD and other liver-based enzyme deficiencies.

Early intra-amniotic gene transfer using lentiviral vector improves skin blistering phenotype in a murine model of Herlitz junctional epidermolysis bullosa.

Mutations of the LAMB3 gene cause a lethal form of junctional epidermolysis bullosa (JEB). We hypothesized that early intra-amniotic gene transfer in a severe murine model of JEB would improve or correct the skin phenotype. Time-dated fetuses from heterozygous LAMB3(IAP) breeding pairs underwent ultrasound guided intra-amniotic injection of lentiviral vector encoding the murine LAMB3 gene at embryonic day 8 (E8). Gene expression was monitored by immunohistochemistry. The transgenic laminin-ß3 chain was shown to assemble with its endogenous partner chains, resulting in detectable amounts of laminin-332 in the basement membrane zone of skin and mucosa. Ultrastructually, the restoration of ∼60% of hemidesmosomal structures was also noted. Although we could correct the skin phenotype in 11.9% of homozygous LAMB3(IAP) mice, none survived beyond 48 h. However, skin transplants from treated E18 homozygous LAMB3(IAP) fetuses maintained normal appearance for 6 months with persistence of normal assembly of laminin-332. These results demonstrate for the first time long-term phenotypic correction of the skin pathology in a severe model of JEB by in vivo prenatal gene transfer. Although survival remained limited due to the limitations of this mouse model, this study supports the potential for treatment of JEB by prenatal gene transfer.

Early gestational gene transfer of IL-10 by systemic administration of lentiviral vector can prevent arthritis in a murine model.

Gene therapy has been applied to murine models of rheumatoid arthritis (RA) using a number of different strategies to downregulate inflammation in synovial joints. However, prolonged joint expression has been problematic. Our laboratory has found that early gestational intravascular injection of lentiviral vector leads to efficient transduction and sustained transgene expression in articular cartilage and synovium. In this study, we show that in utero gene transfer of IL-10 can prevent and decrease pathology in a murine model of RA. Following prenatal injection of lentiviral vector containing murine IL-10 gene, the cytokine was detectable in the serum, and the green fluorescent protein reporter gene was detectable in chondrocytes and synoviocytes of adult mice up to 21 weeks of age. Adult mice that had been treated prenatally were later immunized against type II collagen to induce an autoimmune arthritis. Compared with controls, prenatally treated mice demonstrated delayed onset of arthritis, decreased frequency of arthritis and markedly decreased severity of disease, by both clinical and histological criteria. This effect was directly related to levels of IL-10 expression, but no immunosuppressive effects of the therapy were observed. This study demonstrates proof of principle for the prenatal prevention and amelioration of RA by early gestational gene transfer of the anti-inflammatory cytokine, IL-10.

SLAM-enriched hematopoietic stem cells maintain long-term repopulating capacity after lentiviral transduction using an abbreviated protocol.

Gene transfer to long-term repopulating hematopoietic stem cells (HSCs) using integrating viral vectors is an important goal in gene therapy. The SLAM (signaling lymphocyte activation molecule)-family receptors have recently been used for the isolation of highly enriched murine HSCs. This HSC enrichment protocol is relatively simple, and results in an HSC population with comparable repopulating capacity to c-kit(+)lin(-)Sca-1(+) (KSL) HSCs. The capacity to withstand genetic manipulation and, most importantly, to maintain long-term repopulating capacity of SLAM-enriched HSC populations has not been reported. In this study, SLAM-enriched HSCs were assessed for transduction efficiency and in vivo long-term repopulating capacity after lentiviral transduction using an abbreviated transduction protocol and KSL-enriched HSCs as a reference population. SLAM- and KSL-enriched HSCs were efficiently transduced by lentiviral vector using a simple protocol that involves minimal in vitro manipulation and no pre-stimulation. SLAM-HSCs are at least equal to KSL-HSCs with respect to efficiency of transduction and maintenance of long-term repopulating capacity. Although there was a reduction in repopulating capacity related to enrichment and culture manipulations relative to freshly isolated bone marrow (BM) cells, no detrimental effects were identified on long-term competitive capacity related to transduction, as transduced cells maintained stable levels of chimerism in competition with non-transduced cells and freshly isolated BM cells. These results support the SLAM-HSC enrichment protocol as a simple and efficient method for HSC enrichment for gene transfer studies.

The developmental stage determines the distribution and duration of gene expression after early intra-amniotic gene transfer using lentiviral vectors.

Gene transfer after intra-amniotic injection has, in general, been of low efficiency and limited to epithelial cells in the skin, pulmonary and gastrointestinal system. We have recently shown that early gestational administration results in a more efficient gene transfer to developmentally accessible stem cell populations in the skin and eye. In this study we present a comprehensive analysis of patterns of tissue expression seen after early intra-amniotic gene transfer (IAGT) using lentiviral vectors. To assess the influence of developmental stage on tissue expression, injections were administered from the late head fold/early somite stage (E8) to E18. In early gestation (E8-10), green fluorescent protein (GFP) expression was observed in multiple organs, derived from all three germ layers. Remarkably, GFP expression was observed in tissues derived from mesoderm and neural ectoderm at E8, whereas expression was limited to only epithelial cells of ectoderm- and endoderm-derived organs after E11. The amount and duration of gene expression was much higher after IAGT at early gestational time points. The observed temporal patterns of gene expression correspond to the predicted developmental accessibility of organ-specific cell populations. This model may be useful for the analyses of mechanisms of genetic and/or developmental disease and for the development of prenatal gene therapy for specific disorders.

Frequent expression of a mutant epidermal growth factor receptor in multiple human tumors.

The epidermal growth factor receptor has received much interest as a target for various antineoplastic agents, but a complication is that many normal tissues also express this receptor. We have previously identified in human glial tumors an 801-bp in-frame deletion within the epidermal growth factor receptor gene that created a novel epitope at the junction. By using Western blot assays with a mutant-specific antibody as a rapid and sensitive means for detecting this alteration in primary human tumors, it was found that 57% (26 of 46) of high-grade and 86% (6 of 7) of low-grade glial tumors, but not normal brain, express this protein. This altered receptor was also present in 66% (4 of 6) of pediatric gliomas and 86% (6 of 7) of medulloblastomas, 78% (21 of 27) of breast carcinomas, and 73% (24 of 32) of ovarian carcinomas. The fact that this receptor is frequently found in tumors but not in normal tissue makes it an attractive candidate for various antitumor strategies.

Formation of undifferentiated mesenteric tumors in transgenic mice expressing human neurotropic polymavirus early protein.

The human polyomavirus, JCV, is the established etiologic agent of the human demyelinating disease, progressive multifocal leukoencephalopathy (PML) seen in immunosuppressed individuals. In PML patients, the viral early protein, which is produced exclusively in glial cells is responsible for initiation of the viral lytic cycle. The JCV early protein, T-antigen, has greater than 70% homology to the well characterized SV40 early protein which has established oncogenic properties. To investigate the role of JCV T-antigen in tumorigenesis, transgenic mice containing the viral early genome were produced. Of the four positive transgenic animals, one developed severe neurological abnormalities and succumbed to death at 3 weeks of age. Another animal died with no visible gross pathology and the cause of death was not determined. The remaining two founders developed massive, undifferentiated, solid mesenteric tumors with no obvious neurological symptoms. Results from histologic analysis demonstrated the presence of highly cellular, poorly differentiated neoplastic cells in the tumor tissue. Electron microscopic evaluation of the tumor revealed the presence of a small blue cell-like tumor of epithelial/neuroectodermal origin. Results from RNA analysis by non-quantitative and highly sensitive RT-PCR indicated the presence of the JCV early transcript in various tissues, including kidney, liver, spleen, heart, lung, and brain, as well as in the tumors. However, analysis of the viral early protein by Western blot and immunohistochemistry indicated high level production of JCV early protein in the tumor tissue, but not in any other tissues. These observations present the first evidence for the development of inheritable neuroectodermal tumors induced by the human polyomavirus, JCV, early protein in a whole animal system.

High-titer adeno-associated viral vectors from a Rep/Cap cell line and hybrid shuttle virus.

Adeno-associated virus (AAV) is a potential vector for in vivo gene therapy. A critical analysis of its utility has been hampered by methods of production that are inefficient, difficult to scale up, and that often generate substantial quantities of replication-competent AAV. We describe a novel method for producing AAV that addresses these problems. A cell line, called B50, was created by stably transfecting into HeLa cells a rep/cap-containing plasmid utilizing endogenous AAV promoters. Production of AAV occurs in a two-step process. B50 is infected with an adenovirus defective in E2b, to induce Rep and Cap expression and provide helper functions, followed by a hybrid virus in which the AAV vector is cloned in the E1 region of a replication-defective adenovirus. This results in a 100-fold amplification and rescue of the AAV genome, leading to a high yield of recombinant AAV that is free of replication-competent AAV. Intramuscular injection of vector encoding erythropoietin into skeletal muscle of mice resulted in supraphysiologic levels of hormone in serum that was sustained and caused polycythemia. This method of AAV production should be useful in scaling up for studies in large animals, including humans.

A general method for the induction and screening of antisera for cDNA-encoded polypeptides: antibodies specific for a coronavirus putative polymerase-encoding gene.

A prokaryotic vector, pGE374, containing the recA and lacZ genes, out-of-frame, was used for the expression of cDNA derived from the putative polymerase-encoding gene of the coronavirus mouse hepatitis virus strain A59 (MHV-A59). The pGE374/viral recombinant vector generates a tripartite bacterial/viral protein composed of a segment of the RecA protein at the N terminus, the coronaviral sequences in the middle, and an enzymatically active beta-galactosidase at the C terminus. Rabbits immunized with such recombinant proteins generated antibodies to the MHV-A59 portion of the tripartite protein. Because the MHV-A59 polymerase proteins have been difficult to identify during infection, we used a novel method to demonstrate the viral specificity of the antiserum. The viral cDNA was excised from the expression vector, and transferred to a pGem vector, downstream from and in-frame with a portion of the cat gene. This construct contained a bacteriophage RNA polymerase promoter that enabled the cell-free synthesis of a fusion protein that was used to verify that antibodies were generated to the expressed viral DNA. This strategy was shown to successfully result in the specific generation of antibodies to the encoded information of the viral cDNA. Furthermore, this method has general applicability in the generation and characterization of antibodies directed against proteins encoded in cDNAs.

Regulated gene expression in gene therapy.

The original model of gene therapy, that of efficient delivery, durable transfer, and stable expression of transgenes to correct a gene defect underlying an inherited disease, is limited in light of improved understanding of the processes involved. Techniques that enable regulated expression of transgenes may enhance safety and allow us to regulate the timing and level of expression with a goal of precisely targeting a therapeutic level between the extremes of suboptimal and supraoptimal thresholds. Using regulated systems to control protein expression has practical and possibly essential roles for the success of safe and effective gene therapy in a number of clinical situations. Pharmacologically regulated gene expression is an evolving tool, and no individual system may be effective in all clinical applications.

Detection of mouse hepatitis virus nonstructural proteins using antisera directed against bacterial viral fusion proteins.

Mouse hepatitis virus, strain A59 cDNAs were inserted into the procaryotic fusion vector pGE374. RecA/viral/LacZ tripartite fusion proteins were synthesized from these plasmids and purified from E. coli. Antisera were raised in rabbits against these fusion proteins. Viral nonstructural proteins were detected in infected murine fibroblasts and glial cells. The anti-gene B, ORF1 sera detect a 30K cytoplasmic protein while the anti-gene E, ORF2 sera detect a 9.6K protein. Sera raised against proteins encoded in cDNAs from 5' portions of gene A immunoprecipitate the 200-250K polypeptides synthesized in vitro from genome RNA. Antisera raised against proteins encoded in both 5' and 3' portions of gene A immunoprecipitate membrane associated polypeptides of 150K and greater than 600K from MHV infected cells.

A quantitative nonimmunogenic transgene product for evaluating vectors in nonhuman primates.

The success of gene therapy depends on safe, effective vectors to transfer genetic information. We have developed a means to quantitatively assess efficacy of gene transfer vectors by using a biologically inert, secreted reporter molecule, the beta chain of chorionic gonadotropin (beta-CG). Using an isogenic beta chain subunit of CG in a recombinant adeno-associated virus (rAAV) vector, overall gene transfer of rhesus macaque muscle is demonstrated over time by measuring the serum concentration of beta-CG. Endogenous levels of gonadotropins are not detectable in healthy, nonpregnant primates as confirmed by the inability to detect serum levels of beta-CG prior to vector administration. The serum concentration of beta-CG also provides a measure for the transfer efficiency in liver and/or muscle in immunodeficient mice using recombinant adenovirus and rAAV vectors. No biological effect was observed in animals tested. Assaying for the serum level of beta-CG serves as a surrogate quantitative marker of gene transfer without interfering with the biology of the host or transduction process.

The ns 4 gene of mouse hepatitis virus (MHV), strain A 59 contains two ORFs and thus differs from ns 4 of the JHM and S strains.

The sequence of the MHV-A 59 non-structural gene 4 (ns 4) reveals two open reading frames. The upstream ORF potentially encodes a 19 amino acid (2.2 kDa) polypeptide and the downstream ORF potentially encodes a 106 amino acid (11.7 kDa) polypeptide. This is in contrast to MHV-JHM gene 4 which expresses a 15 kDa protein. Cell free translation of a synthetic mRNA containing both ORFs of MHV-A 59 ns 4 results in the synthesis of a 2.2 kDa polypeptide; the predicted 11.7 kDa product of the MHV-A 59 downstream ORF is not detected during cell free translation nor in infected cells. These results add to the recent data suggesting that expression of some of the ns proteins of MHV is not necessary for efficient growth in tissue culture.

Mouse hepatitis virus ORF 2a is expressed in the cytosol of infected mouse fibroblasts.

A 266-bp fragment of cDNA from within gene B, ORF 2a, of MHV-A59 was used to construct a vector encoding a bacterial/viral fusion protein. Antiserum raised against this fusion protein specifically immunoprecipitates a 30K protein from infected 17Cl-1 mouse fibroblasts. The protein is localized primarily in the cytosol and not in the membranes. This is consistent with its predicted sequence and potential role as an RNA binding protein.

Molecular cloning of the gene encoding the putative polymerase of mouse hepatitis coronavirus, strain A59.

Complementary DNA (cDNA) libraries were constructed representing the genome RNA of the coronavirus mouse hepatitis virus, strain A59 (MHV-A59). From these libraries clones were selected to form a linear map across the entire gene A, the putative viral polymerase gene. This gene is approximately 23 kb in length, considerably larger than earlier estimates. Sequence analysis of the 5' terminal region of the genome indicates the presence of the 66-nucleotide leader that is found on all mRNAs. Secondary structure analysis of the 5' terminal region suggests that transcription of leader terminates in the region of nucleotide 66. The sequence of the first 2000 nucleotides is very similar to that reported for the closely related JHM strain of MHV and potentially encodes p28, a basic protein thought to be a component of the viral polymerase (L. Soe, C. K. Shieh, S. Baker, M. F. Chang, and M. M. C. Lai, 1987, J. Virol., 61, 3968-3976). Gene A contains two of the consensus sequences found in intergenic regions. One is adjacent to the 5' leader sequence and the other is upstream from the initiation codon for translation of gene B.

Biology of E1-deleted adenovirus vectors in nonhuman primate muscle.

Adenovirus vectors have been studied as vehicles for gene transfer to skeletal muscle, an attractive target for gene therapies for inherited and acquired diseases. In this setting, immune responses to viral proteins and/or transgene products cause inflammation and lead to loss of transgene expression. A few studies in murine models have suggested that the destructive cell-mediated immune response to virally encoded proteins of E1-deleted adenovirus may not contribute to the elimination of transgene-expressing cells. However, the impact of immune responses following intramuscular administration of adenovirus vectors on transgene stability has not been elucidated in larger animal models such as nonhuman primates. Here we demonstrate that intramuscular administration of E1-deleted adenovirus vector expressing rhesus monkey erythropoietin or growth hormone to rhesus monkeys results in generation of a Th1-dependent cytotoxic T-cell response to adenovirus proteins. Transgene expression dropped significantly over time but was still detectable in some animals after 6 months. Systemic levels of adenovirus-specific neutralizing antibodies were generated, which blocked vector readministration. These studies indicate that the cellular and humoral immune response generated to adenovirus proteins, in the context of transgenes encoding self-proteins, hinders long-term transgene expression and readministration with first-generation vectors.

Identification of polypeptides encoded in open reading frame 1b of the putative polymerase gene of the murine coronavirus mouse hepatitis virus A59.

The polypeptides encoded in open reading frame (ORF) 1b of the mouse hepatitis virus A59 putative polymerase gene of RNA 1 were identified in the products of in vitro translation of genome RNA. Two antisera directed against fusion proteins containing sequences encoded in portions of the 3'-terminal 2.0 kb of ORF 1b were used to immunoprecipitate p90, p74, p53, p44, and p32 polypeptides. These polypeptides were clearly different in electrophoretic mobility, antiserum reactivity, and partial protease digestion pattern from viral structural proteins and from polypeptides encoded in the 5' end of ORF 1a, previously identified by in vitro translation. The largest of these polypeptides had partial protease digestion patterns similar to those of polypeptides generated by in vitro translation of a synthetic mRNA derived from the 3' end of ORF 1b. The polypeptides encoded in ORF 1b accumulated more slowly during in vitro translation than polypeptides encoded in ORF 1a. This is consistent with the hypothesis that translation of gene A initiates at the 5' end of ORF 1a and that translation of ORF 1b occurs following a frameshift at the ORF 1a-ORF 1b junction. The use of in vitro translation of genome RNA and immunoprecipitation with antisera directed against various regions of the polypeptides encoded in gene A should make it possible to study synthesis and processing of the putative coronavirus polymerase.

Isolation and characterization of a type II JC virus from a brain biopsy of a patient with PML.

Brain tissue of a patient with multiple myeloma suffering from neurological disorders similar to those seen in progressive multifocal leukoencephalopathy (PML) patients was evaluated for the presence of the papovavirus, JCV. Results from polymerase chain reaction (PCR) revealed the presence of JCV with structural organization at the control region which is distinct from well-characterized isolates, ie Mad-1 and Mad-4. The control region of the new isolate, named JCVPhila-1' contains a 22 nucleotide insertion which separates the TATA box from the NF-1 regulatory motif. Only 18 nucleotides of the insert are duplicated in the second copy of the enhancer/promoter of the new isolate, which is 84 nucleotides in size. Results from a transcription assay indicate a modest elevated level of JCVPhila-1 early promoter activity compared to that of JCVMad-4 in glial cell lines. The basal and T-antigen-induced transcriptional activities of the JCVPhila-1 late promoter was lower with respect to Mad-4 late gene activity in glial cells. Of particular interest was the observation that in the cells producing the early protein, T-antigen, JCVPhila-1 DNA replicated more efficiently then the Mad-4 DNA. These results suggest that the alterations seen in the JCVPhila-1 control region may differentially influence early and late gene expression and facilitate amplification of the viral genome in cells derived from the CNS.