Intravenous administration of an AAV-2 vector for the expression of factor IX in mice and a dog model of hemophilia B.

Previous experiments have demonstrated the stable expression of factor IX (FIX) protein in mice and canine models of hemophilia B following portal vein gene transfer with a recombinant adeno-associated virus (rAAV) vector encoding FIX. Here, we present the results of studies that further optimized the rAAV vector transgene cassette used to express FIX and explored the use of the less-invasive intravenous (i.v.) route of vector administration for the treatment of hemophilia B. First, a liver-specific promoter was evaluated in conjunction with cis-acting regulatory elements in mice. Constructs that included both the beta-globin intron and the woodchuck hepatitis virus post-transcriptional regulatory element resulted in the highest level of FIX expression in vivo. Using this optimized vector, we demonstrate that i.v. injection was feasible for hepatic gene transfer in mice, achieving 70-80% of portal vein expression levels of FIX. In further studies using the Chapel Hill strain of hemophilia B dogs, we demonstrate for the first time FIX expression and partial correction of the bleeding disorder following i.v. administration of an AAV vector.

Dopamine production in the caudate putamen restores feeding in dopamine-deficient mice.

Dopamine-deficient (DD) mice cannot synthesize dopamine (DA) in dopaminergic neurons due to selective inactivation of the tyrosine hydroxylase gene in those neurons. These mice become hypoactive and hypophagic and die of starvation by 4 weeks of age. We used gene therapy to ascertain where DA replacement in the brain restores feeding and other behaviors in DD mice. Restoration of DA production within the caudate putamen restores feeding on regular chow and nest-building behavior, whereas restoration of DA production in the nucleus accumbens restores exploratory behavior. Replacement of DA to either region restores preference for sucrose or a palatable diet without fully rescuing coordination or initiation of movement. These data suggest that a fundamental difference exists between feeding for sustenance and the ability to prefer rewarding substances.

Epitope mapping of human anti-adeno-associated virus type 2 neutralizing antibodies: implications for gene therapy and virus structure.

Recombinant adeno-associated virus type 2 (AAV) is a common vector used in human gene therapy protocols. We characterized the humoral immune response to AAV and observed that 80% of normal human subjects have anti-AAV antibodies and that 18% have neutralizing antibodies. To analyze the effect of neutralizing antibodies on AAV readministration, we attempted to deliver recombinant AAV expressing human factor IX (AAV-hFIX) intraportally into the livers of mice which had been preexposed to AAV and shown to harbor a neutralizing antibody response. While all naive control mice expressed hFIX following administration of AAV-hFIX, none of the mice with preexisting immunity expressed hFIX, even after transient immunosuppression at the time of the second administration with anti-CD4 or anti-CD40L antibodies. This suggests that preexisting immunity to AAV, as measured by a neutralizing antibody response, may limit AAV-mediated gene delivery. Using human sera in an enzyme-linked immunosorbent assay for AAV and a capsid peptide scan library to block antibody binding, we mapped seven regions of the AAV capsid containing immunogenic epitopes. Using pools of these peptides to inhibit the binding of neutralizing antibodies, we have identified a subset of six peptides which potentially reconstitute a single neutralizing epitope. This information may allow the design of reverse genetic approaches to circumvent the preexisting immunity that can be encountered in some individuals.

Viral gene delivery selectively restores feeding and prevents lethality of dopamine-deficient mice.

Dopamine-deficient mice (DA-/- ), lacking tyrosine hydroxylase (TH) in dopaminergic neurons, become hypoactive and aphagic and die by 4 weeks of age. They are rescued by daily treatment with L-3,4-dihydroxyphenylalanine (L-DOPA); each dose restores dopamine (DA) and feeding for less than 24 hr. Recombinant adeno-associated viruses expressing human TH or GTP cyclohydrolase 1 (GTPCH1) were injected into the striatum of DA-/- mice. Bilateral coinjection of both viruses restored feeding behavior for several months. However, locomotor activity and coordination were partially improved. A virus expressing only TH was less effective, and one expressing GTPCH1 alone was ineffective. TH immunoreactivity and DA were detected in the ventral striatum and adjacent posterior regions of rescued mice, suggesting that these regions mediate a critical DA-dependent aspect of feeding behavior.

Correction of hemophilia B in canine and murine models using recombinant adeno-associated viral vectors.

Hemophilia B, or factor IX deficiency, is an X-linked recessive disorder occurring in about 1 in 25,000 males. Affected individuals are at risk for spontaneous bleeding into many organs; treatment mainly consists of the transfusion of clotting factor concentrates prepared from human blood or recombinant sources after bleeding has started. Small- and large-animal models have been developed and/or characterized that closely mimic the human disease state. As a preclinical model for gene therapy, recombinant adeno-associated viral vectors containing the human or canine factor IX cDNAs were infused into the livers of murine and canine models of hemophilia B, respectively. There was no associated toxicity with infusion in either animal model. Constitutive expression of factor IX was observed, which resulted in the correction of the bleeding disorder over a period of over 17 months in mice. Mice with a steady-state concentration of 25% of the normal human level of factor IX had normal coagulation. In hemophilic dogs, a dose of rAAV that was approximately 1/10 per body weight that given to mice resulted in 1% of normal canine factor IX levels, the absence of inhibitors, and a sustained partial correction of the coagulation defect for at least 8 months.

Selective uptake and sustained expression of AAV vectors following subcutaneous delivery.

BACKGROUND: Recombinant adeno-associated viral (rAAV) vectors are capable of long-term expression of secreted and intracellular proteins following delivery to muscle, liver, and the central nervous system. In this study, we have evaluated subcutaneous injection of rAAV encoding a variety of transgenes as an alternative route of administration for the systemic delivery of therapeutic proteins. METHODS: rAAV vectors encoding the human factor IX, human interferon-alpha 2a, murine erythropoietin (epo), and Escherichia coli lacZ genes were used for subcutaneous delivery into mature immunocompetent mice. Expression of factor IX and interferon in mouse serum was measured by ELISA. Expression of Epo was monitored by an increase in hemotocrit and by RIA. The tissue tropism of AAV transduction was determined by histochemistry following administration of the lacZ vector. RESULTS: Long-term protein expression (at least one year) is demonstrated in the serum of immunocompetent mice following subcutaneous delivery of AAV vectors encoding the human factor IX and interferon genes. The murine epo gene delivered via this route resulted in levels of Epo that correlate with increased hematocrits of up to 90% for a duration of nine months. rAAV encoding the lacZ gene revealed that the panniculus carnosus, a skeletal muscle layer of the skin, was transduced upon subcutaneous administration. CONCLUSIONS: This study shows that long-term expression of secreted proteins can be achieved using rAAV vectors injected subcutaneously as a single administration. The observation that the panniculus carnosus is the primary tissue transduced by rAAV illustrates the high tropism of rAAV for skeletal muscle.

Regulation of gene expression in vivo following transduction by two separate rAAV vectors.

Control of gene expression is important to gene therapy for purposes of both dosing and safety. In vivo regulation of gene expression was demonstrated following co-injection of two separate recombinant adeno-associated virus vectors, one encoding an inducible murine erythropoietin transgene and the other a transcriptional activator, directly into the skeletal muscle of adult immunocompetent mice. Transcription was controlled by systemic administration or withdrawal of tetracycline over an 18 week period, demonstrating that the two vectors were capable of transducing the same cell. Cellular or humoral immune responses against the transactivator protein were not detected.

Nucleotide sequence context effect of a cyclobutane pyrimidine dimer upon RNA polymerase II transcription.

We have studied the role of sequence context upon RNA polymerase II arrest by a cyclobutane pyrimidine dimer using an in vitro transcription system consisting of templates containing a specifically located cyclobutane pyrimidine dimer (CPD) and purified RNA polymerase II (RNAP II) and initiation factors. We selected a model sequence containing a well characterized site for RNAP II arrest in vitro, the human histone H3.3 gene arrest site. The 13-base pair core of the arrest sequence contains two runs of T in the nontranscribed strand that impose a bend in the DNA. We hypothesized that arrest of RNAP II might be affected by the presence of a CPD, based upon the observation that a CPD located at the center of a dA6.dT6 tract eliminates bending (Wang, C.-I., and Taylor, J.-S. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 9072-9076). We examined the normal H3.3 sequence and a mutant sequence containing a T --> G transversion, which reduces bending and efficiency of arrest. We show that a CPD in the transcribed strand at either of two locations in the arrest site is a potent block to transcription. However, a CPD in the nontranscribed strand only transiently pauses RNAP II. The CPD in concert with a mutation in the arrest site can reduce the extent of bending of the DNA and improve readthrough efficiency. These results demonstrate the potential importance of sequence context for the effect of CPDs within transcribed sequences.

Effects of aminofluorene and acetylaminofluorene DNA adducts on transcriptional elongation by RNA polymerase II.

A prominent model for the mechanism of transcription-coupled DNA repair proposes that an arrested RNA polymerase directs the nucleotide excision repair complex to the transcription-blocking lesion. The specific role for RNA polymerase II in this mechanism can be examined by comparing the extent of polymerase arrest with the extent of transcription-coupled repair for a specific DNA lesion. Previously we reported that a cyclobutane pyrimidine dimer that is repaired preferentially in transcribed genes is a strong block to transcript elongation by RNA pol II (Donahue, B.A., Yin, S., Taylor, J.-S., Reines, D., and Hanawalt, P. C. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 8502-8506). Here we report the extent of RNA polymerase II arrest by the C-8 guanine DNA adduct formed by N-2-aminofluorene, a lesion that does not appear to be preferentially repaired. Templates for an in vitro transcription assay were constructed with either an N-2-aminofluorene adduct or the helix-distorting N-2-acetylaminofluorene adduct situated at a specific site downstream from the major late promoter of adenovirus. Consistent with the model for transcription-coupled repair, an aminofluorene adduct located on the transcribed strand was a weak pause site for RNA polymerase II. An acetylaminofluorene adduct located on the transcribed strand was an absolute block to transcriptional elongation. Either adduct located on the nontranscribed strand enhanced polymerase arrest at a nearby sequence-specific pause site.

Transcript cleavage by RNA polymerase II arrested by a cyclobutane pyrimidine dimer in the DNA template.

A current model for transcription-coupled DNA repair is that RNA polymerase, arrested at a DNA lesion, directs the repair machinery to the transcribed strand of an active gene. To help elucidate this role of RNA polymerase, we constructed DNA templates containing the major late promoter of adenovirus and a cyclobutane pyrimidine dimer (CPD) at a specific site. CPDs, the predominant DNA lesions formed by ultraviolet radiation, are good substrates for transcription-coupled repair. A CPD located on the transcribed strand of the template was a strong block to polymerase movement, whereas a CPD located on the nontranscribed strand had no effect on transcription. Furthermore, the arrested polymerase shielded the CPD from recognition by photolyase, a bacterial DNA repair protein. Transcription elongation factor SII (also called TFIIS) facilitates read-through of a variety of transcriptional pause sites by a process in which RNA polymerase II cleaves the nascent transcript before elongation resumes. We show that SII induces nascent transcript cleavage by RNA polymerase II stalled at a CPD. However, this cleavage does not remove the arrested polymerase from the site of the DNA lesion, nor does it facilitate translesional bypass by the polymerase. The arrested ternary complex is stable and competent to resume elongation, demonstrating that neither the polymerase nor the RNA product dissociates from the DNA template.

Repair and transcription. Collision or collusion?

While some proteins have distinct responsibilities in both transcription and DNA repair, additional proteins are needed to couple these essential DNA transactions in expressed genes.

Characterization of a DNA damage-recognition protein from mammalian cells that binds specifically to intrastrand d(GpG) and d(ApG) DNA adducts of the anticancer drug cisplatin.

A factor has been identified in extracts from human HeLa and hamster V79 cells that retards the electrophoretic mobility of several DNA restriction fragments modified with the antitumor drug cis-diamminedichloroplatinum(II) (cisplatin). Binding of the factor to cisplatin-modified DNA was sensitive to pretreatment with proteinase K, establishing that the factor is a protein. Gel mobility shifts were observed with probes containing as few as seven Pt atoms per kilobase of duplex DNA. By competition experiments the dissociation constant, Kd, of the protein from cisplatin-modified DNA was estimated to be (1-20) X 10(-10) M. Protein binding is selective for DNA modified with cisplatin, [Pt(en)Cl2] (en, ethylenediamine), and [Pt(dach)Cl2] (dach, 1,2-diaminocyclohexane) but not with chemotherapeutically inactive trans-diamminedichloroplatinum(II) or monofunctionally coordinating [Pt(dien)Cl]Cl (dien, diethylenetriamine) complexes. The protein also does not bind to DNA containing UV-induced photoproducts. The protein binds specifically to 1,2-intrastrand d(GpG) and d(ApG) cross-links formed by cisplatin, as determined by gel mobility shifts with synthetic 110-bp duplex oligonucleotides; these modified oligomers contained five equally spaced adducts of either cis-[Pt(NH3)2d(GpG) or cis-[Pt(NH3)2d(ApG)]. Oligonucleotides containing the specific adducts cis-[Pt(NH3)2d(GpTpG)], trans-[Pt(NH3)2d(GpTpG)], or cis-[Pt(NH3)2(N3-cytosine)d(G)] were not recognized by the protein. The apparent molecular weight of the protein is 91,000, as determined by sucrose gradient centrifugation of a preparation partially purified by ammonium sulfate fractionation. Binding of the protein to platinum-modified DNA does not require cofactors but is sensitive to treatment with 5 mM MnCl2, CdCl2, CoCl2, or ZnCl2 and with 1 mM HgCl2.(ABSTRACT TRUNCATED AT 250 WORDS)

Isolation of cDNAs encoding a human protein that binds selectively to DNA modified by the anticancer drug cis-diamminedichloroplatinum(II)

DNA modified by the antitumor drug cis-diamminedichloroplatinum(II) (cis-DDP or cisplatin) was used to identify a factor in mammalian cells that binds to cis-DDP-damaged DNA and hence may play a role in repair. This factor selectively recognizes double-stranded DNA fragments modified by cis-DDP or [Pt(en)Cl2] (en, ethylenediamine). Little or no binding occurs to unmodified double-stranded DNA or to DNA modified with the clinically ineffective compounds trans-DDP and [Pt(dien)Cl]Cl (dien, diethylenetriamine). Low levels of binding to single-stranded DNA modified by cis-DDP are observed. The apparent molecular mass of the factor in a variety of mammalian cells is approximately 100 kDa, as determined by modified Western blotting. Two recombinant phage have been isolated from a human B-cell lambda gt11 library by using a cis-DDP-modified DNA restriction fragment as a probe. The two clones have insert sizes of 1.88 and 1.44 kilobases and are aligned at their 5' ends. The polypeptides encoded by the recombinant phage exhibit DNA binding properties similar to those of the cellular factor identified in crude extracts prepared from mammalian cells. Northern analysis with one of the clones revealed an mRNA of 2.8 kilobases that is conserved in humans and rodents. The methods used here should be applicable in studies of other damage-specific DNA binding proteins.

Separation of the SOS-dependent and SOS-independent components of alkylating-agent mutagenesis.

Escherichia coli plasmids containing the rpsL+ gene (Strs phenotype) as the target for mutation were treated in vitro with N-methyl-N-nitrosourea. Following fixation of mutations in E. coli MM294A cells (recA+ Strs), an unselected population of mutant and wild-type plasmids was isolated and transferred into a second host, E. coli 6451 (recA Strr). Strains carrying plasmid-encoded forward mutations were then selected as Strr isolates, while rpsL+ plasmids conferred the dominant Strs phenotype in the second host. Mutation induction and reduced survival of N-methyl-N-nitrosourea-treated plasmids were shown to be dose dependent. Because this system permitted analysis and manipulation of the levels of certain methylated bases produced in vitro by N-methyl-N-nitrosourea, it afforded the opportunity to assess directly the relative roles of these bases and of SOS functions in mutagenesis. The methylated plasmid DNA gave a mutation frequency of 6 X 10(-5) (a 40-fold increase over background) in physiologically normal cells. When the same methylated plasmid was repaired in vitro by using purified O6-methylguanine DNA methyltransferase (to correct O6-methylguanine and O4-methylthymine), no mutations were detected above background levels. In contrast, when the methylated plasmid DNA was introduced into host cells induced by UV light for the SOS functions, rpsL mutagenesis was enhanced eightfold over the level seen without SOS induction. This enhancement of mutagenesis by SOS was unaffected by prior treatment of the DNA with O6-methylguanine DNA methyltransferase. These results demonstrate a predominant mutagenic role for alkylation lesions other than O6-methylguanine or O4-methylthymine when SOS functions are induced. The mutation spectrum of N-methyl-N-nitrosourea under conditions of induced SOS functions revealed a majority of mutagenic events at A . T base pairs.

Induction of sister chromatid exchanges and gene mutations in Chinese hamster ovary cells by psoralens.

Three linear psoralen compounds, 8-methoxypsoralen (8-MOP), 5-methoxypsoralen (5-MOP), 3-carbethoxypsoralen (3-CP), and one angular psoralen, 5-methylangelicin (5-ANG), were tested for their ability to induce both sister chromatid exchanges (SCE) and gene mutations (hypoxanthine-guanine phosphoribosyltransferase locus) in vitro in Chinese hamster ovary cells (CHO line). All the compounds induced both SCE and mutations in the presence of UV irradiation (UVA; peak at 330-380 nm), but no increases were observed in its absence. The frequency of both responses increased with either 1) increasing concentration of compound with a fixed amount of UVA or 2) increasing amount of UVA with a fixed concentration of psoralen. Significant increases in SCE were seen for 8-MOP, 5-MOP, and 5-ANG at concentrations near 1 X 10(-6) M, whereas concentrations near 20 X 10(-6) M of 3-CP were needed before increases in SCE were observed. The induction of gene mutations followed a similar pattern; concentrations of 50-100 X 10(-6) M of 3-CP were needed to induce large increases in mutations, but much lower concentrations of 8-MOP, 5-MOP, and 5-ANG (5-10 X 10(-6) M) were sufficient to induce large increases in mutations. The ratio of induced mutations to induced SCE was similar for 8-MOP, 5-MOP, and 3-CP; that of 5-ANG was much higher, which indicated that the linear furocoumarins produce a different spectrum of DNA damage from that produced by the angular psoralen.