EP1084236A1

EP1084236A1 - Novel adenoviral vectors for gene therapy

Info

Publication number: EP1084236A1
Application number: EP99927250A
Authority: EP
Inventors: Manal A. Morsy; Volker Sandig
Original assignee: Merck and Co Inc
Current assignee: Merck and Co Inc
Priority date: 1998-06-09
Filing date: 1999-06-04
Publication date: 2001-03-21
Also published as: JP2002517234A; EP1084236A4; WO1999064577A1; CA2329947A1

Abstract

Helper dependent adenoviral vectors are suitable for use in gene therapy, as they are stable, contain only a very small amount less than about 0.5 % of helper virus contamination, and are less immunogenic than previous adenoviral vectors. The only viral DNA in the vectors are ITRs and packaging signals, and overall size is about 26-38 kb.

Description

TITLE OF THE INVENTION

NOVEL ADENOVIRAL VECTORS FOR GENE THERAPY

FIELD OF THE INVENTION This invention related to adenoviral vectors which are useful as vectors for gene delivery in gene therapy applications.

BACKGROUND OF THE INVENTION

Adenoviral (Ad) vectors are currently among the most efficient gene transfer vehicles for both in vitro and in vivo delivery, but the utilization of current Ad vectors for many gene therapy applications is limited by the transient nature of transgene expression obtained by these vectors. The development of Ad vectors that are deleted in all viral protein-coding sequences offers the prospect of a potentially safer, less immunogenic vector with an insert capacity of up to approximately 38 kb. These vectors require supplementation by viral regulatory and structural proteins supplied in trans for packaging and rescue, and are thus termed "helper-dependent" (HD).

Fisher et al, 1996 Virology 217:11-22 describes a spontaneous concatamerization of HD-adeno viral vectors. This resulted in a population containing three types of concatamerized viruses with different genome sizes: Form 1 with a genome size of the original transfected DNA, Form 2 with a size twice that of the monomer, and predominant Form 3 with a larger size. Individual genomes were fused in 3 different ways: "head-to-tail" configuration (5'-3'-5'-3', i.e. the ends contained a 5' region and 3' region, with a 3' region joined to a 5' region in the center); a "head-to- head" configuration (S'-S'-S'-S¹, i.e. both ends contained 5' regions, with two 3' regions joined in the center); and a "tail-to-tail" configuration (3'-5'-5'-3', i.e. both ends contained 3' regions, with two 5' regions joined in the center). These viruses were present as a mixture of multimers of a small genome, wherein the only adenoviral sequences were the ITR and packaging regions. Parks et al, 1997, J Virology 71(4):3293-3298 describes a population of HD-adeno viruses in which a 15.1 kb virus concatamerized spontaneously in culture, and formed two different species of 30.2 kb dimers. Parks, using terminology different from that used by Fisher, described the following configurations of concatamers: "head-to-tail" dimer ( 5'-3'-5'-3', i.e. one end contained a 5' region and the other end contained a 3' region, with a 3' region joined to a 5' region at the center); and another concatamer where the dimer is a "tail-to-tail" dimer (i.e. both ends contained 3¹ regions, with the 5' regions joined in the center). These vectors contained a large segment of adenoviral genome (+5,700 bp of the 5' region and +6,100 bp of the 3' region), including a disrupted El A gene.

Kochanek et al, 1996 Proc. Natl. Acad. Sci. USA 93:5731-5736 describes an adenoviral vector which has been extensively deleted so that the only viral sequences present are viral terminal repeats (ITRs) and full length packaging signals (ψ) required for packaging the DNA into viral capsids. These vectors were able to carry a large heterologous DNA insert. However, DNA from the plasmid used for replication prior to a rescue step was also present in the vector (including the ori and ampicillin resistance gene. These vectors, however suffered from a high level of contamination of a helper virus, approximately 1%, which is too high for human gene therapy applications.

One of the problems associated with HD-adenoviruses as vectors for gene therapy is contamination of the HD-virus population with the helper viruses needed for viral culture and replication. It would be desirable to have a HD-system wherein helper virus contamination was low.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is a diagram of vectors HD-leptin-monomer and HD-2ψ- leptin. Figure 1 A is the DNA composite fragments of ρΔSTK120-HCMV-mOb- BGHpA (approximately 19.6 kb total size). From left to right the fragments are: the left end terminus of Ad5, containing the inverted terminal repeat sequences and the packaging signal ψ (nucleotides (nt.) 1-440, solid arrow); the 5072 bp fragment of hypoxanthine guanine phosphoribosyltransferase (HPRT) (nt.12373-17853 in Genbank gb:humhprtb; shown as a bar) (inserted in the complementary orientation); the Hindlll 9063 bp fragment of C346 cosmid (nt.12421-21484 in gb:L31948, shaded bar); and the right end terminus of Ad5, composed of the ITR sequence (nt. 35818-35935). The ITRs are flanked by unique Pmel restriction sites used to liberate the vector fragment from the plas id backbone prior to the initial transfection into 293-cre4 cells for viral rescue and propagation (released fragment is 16.7 kb).

Figure IB shows a virus according to this invention, HD-2ψ-leptin, which is a tail-to-tail concatamerization of the virus of Figure 1 A.

Figure 2 is a series of gels demonstrating sizes of vectors made in accordance with this invention, and sizes of vectors cut with various restriction enzymes. Further details are given in the Examples.

Figure 3 is a diagram of vectors (A) stkl 20-1 ψ-CMV-hVEGF 145.

SV40pA and (B) stkl20-2ψ-CMV-hVEGFl45-SV40pA. From left to right the fragments in figure 3(A) are: the left end terminus of Ad5, containing the inverted terminal repeat sequences and the packaging signal ψ (nucleotides (nt.) 1-440, solid arrow); the 16054 bp fragment of hypoxanthine guanine phosphoribosyltransferase (HPRT) (nt.12373-17853 in Genbank gb:humhprtb; shown as a bar) (inserted in the complementary orientation); the Hindlll 9063 bp fragment of C346 cosmid (nt.12421- 21484 in gb:L31948, shaded bar); and the right end terminus of Ad5, composed of the ITR sequence (nt. 35818-35935). The fragments in (B) are identical to those in (A) exept for the right end terminus of Ad5 which is replaced by an inverted left end terminus, containing the inverted terminal repeat sequences and the packaging signal ψ (nucleotides (nt.) 1-440, solid arrow). The ITRs are flanked by unique Pmel restriction sites used to liberate the vector fragment from the plasmid backbone prior to the initial transfection into 293-cre4 cells for viral rescue and propagation (released fragment is 29 kb).

Figure 4 shows a comparison of the vectors HD-lψ and HD-2ψ after propagation. (A) Restriction analysis of DNA isolated from HD-vector preparations at passage 6. HD-vector and helper were not separated during purification. Restriction fragments were radioactivly labeled prior to electrophoretic separation and the gel was exposed to a film. Sizes of expected restriciton fragments for both HD-vectors are shown next to the autoradiograph. The band containing the second packaging signal (450 bp) in the 2ψ vector overlays the 475 bp band, the unique band of the right ITR (145 bp) is not visible in this autoradiograph. The restriction pattern of the helper virus is shown for comparison. (B) Characteristics of HD vector preparations at passage 6: productivity, HD-vector particles obtained per cell from preparations at passage 6 determined by optic density (OD) after purification and purity, % contamination of the HD- Vectors by helper virus determined by quantitative PCR using Taqman Technology.

As used throughout the specification and claims, the following definitions apply:

Stuffer DNA: non-expressed DNA which is inserted to an adenoviral vector to increase the size of virus to at least about 75% of wild type, more preferably from at least about 75-105% wild-type (about 26 -38 kb).

HD-virus: A "helper-dependent" virus: an adenovirus which has native genes deleted so that it cannot replicate unless the products of the deleted genes are provided by another source, such as a helper virus or by a cell line.

Expression Cassette: contains a gene whose product is to be expressed, along with any and all necessary control sequences.

Tail-to-tail Concamerization: two virus are joined so that their regions are 5'-3'-3*-5'.

Head-to-tail Concatamerization: two viruses are joined so that their regions are 5'-3'-5'-3'.

Head-to-head Concatamerization: two viruses are joined so that their regions are 3'-5'-5'-3'. Plasmid-based sequences: refers to DNA and genes present in plasmids which are used during construction of the adenoviral vector, but which are not adenoviral DNA, nor the heterologous gene which is to be delivered by the adenoviral vector. Examples of plasmid-based sequences include: the origin of replication (o ), antibiotic resitance genes used as markers (ampicillin resistance), and the like. Adenoviral vectors are generally useful for delivering genes in gene therapy. In the past, adenoviral vectors have been made which contain a deletion in the section of their genomes coding for replication-related functions, such as the El and E3 genes. These deletions have two different functions: aside from making a safe, non-replicating viral vector, the deletions allow room for insertion of transgenes. Such viruses with deleted replication functions are generally referred to as helper-dependent (HD)-viruses. They cannot replicate on their own, but the replication-related functions can be provided for in trans by helper viruses which are co-cultivated with the HD- viruses. While propagating HD-viruses using conventional cell lines and helper viruses containing genes for viral replication, a spontaneous concatamerization occurred. These new HD-2ψ-viruses were rescued and propagated, and were analyzed both for structure verification and for helper-load contamination. Althernatively, these viruses may be made by ligating together the various described segments, using conventional genetic engineering techniques.

The concatamerized HD-adenoviral vectors according to this invention contain the following elements:

(A) a first 5'-inverted terminal repeat sequence (ITR);

(B) a first packaging signal sequence element; (C) a first heterologous expression cassette;

(D) a second packaging signal sequence element; and

(E) a second 5TTR.

In addition, the vector may contain a second heterologous expression cassette (F) between the first heterologous expression cassette (element C) and the second packaging signal (element D). The first and second expression cassette may express the same gene, or they may express different gene products.

The overall size of the vector should be approximately 26 kb to approximately 38 kb, and preferably from about 26 kb to about 37 kb for efficient packaging. If the above-described elements occupy less than about 26 kb, then "stuffer" DNA should be added in an amount so that the total length of the concatamerized viral vector is approximately 26-38 kb, or preferably about 26-37 kb. Thus the absolute amount of stuffer DNA may vary from vector to vector depending on the size of other elements present, but will generally be up to about 26 kb in length. It is preferred that the stuffer DNA not be expressed by the host cell. Generally, the stuffer DNA will be inserted between the first packaging signal (element B) and the first heterologous expression cassette (element C) and/or between the first or second expression cassette (elements C or F) and the second packaging signal sequence element (element D). In preferred embodiments, the stuffer DNA is inserted prior to the concatamerization event, in an amount approximately one-half of the desired final size, or up to about 13 kb per non-concatamerized virus. This will result in a concatamerized virus containing up to about 26 kb stuffer DNA. Examples of suitable stuffer DNA sequences include: fragments of the hypoxanthine guanine phosphoribosyl-transferase gene, and other non-expressed human or mouse DNA fragments, preferably comprising minimal or no repeat regions. The first 5'ITR and packaging signal (ψ) will generally be a sequence of up to approximately 500 bp. The adenoviral ITRs are known in the art. One such ITR is encompassed in the left end of Ad5, base pairs 1-360, although additional base pairs may be added without affecting the ITR and packaging signal function. 5' ITRs are generally preferred to 3' ITRs. The expression cassette contains the gene of interest, along with control sequences (i.e. promoter, termination signals, enhancers, etc.). These may be virtually any known gene, known promoter, and known control sequences which are desired.

The viruses of this invention have many advantages over prior adenoviral vectors. Their structure is very stable during multiple propagations. While not wishing to be bound by theory, it appears that the combination of viral size and arrangement of packaging signals is an important feature of this invention, and leads to unexpectedly beneficial results. Repeated rescue results in an unexpectedly consistent single concatamerization species, the tail-to-tail dimer, so generation of these types of viruses are repeatable. Unlike previous literature reports describing heterogeneous populations of concatamerized adenoviruses, no other concatamerization species were identified in any of the three independent rescues (by restriction mapping and analysis of radio-labeled digestion fragments).

A further significant advantage of the viruses of this invention is that the helper virus contamination load is consistently very low, less than 0.1% per infectious HD unit and less than 1 pfu of helper virus/100,000 OD particles (minimum estimated HD infectious unit: OD particle is 1 : 100). Thus one aspect of this invention is a population of helper-dependent adenoviruses which comprises less than about 0.5% helper contamination or less than lpfu helper virus per 100,000 OD particles. In contrast, HD-leptin-monomer virus containing: a 5TTR and packaging signal sequence at the left arm only; stuffer DNA; the expression cassette; additional stuffer DNA; and the 3' ITR of Ad5; consistently results in at least 2-10 fold higher load of helper virus contamination (1 pfu of helper virus/103-104 OD particles/ml) in HD-leptin monomer stock.

A further advantage with the concatamerized virus according to this invention is the longevity of transgene expression achieved. The leptin model used in these examples illustrates this advantage. The transgene was expressed for a significantly longer time when introduced into the mouse using the 2ψ-vector as compared to a non-concatamerized control adenoviral vector.

The concatamerized viruses of this invention expressed a transgene, leptin, at levels comparable to its counterpart first generation Ad-leptin. Thus, when using a virus of this invention one can obtained sustained expression for a longer period of time, yet retain comparable levels of expression over this time period.

One virus according to this invention has been designated HD-2ψ- leptin. It was generated from a 16.7 kb vector fragment (Figure 1). This fragment when transfected and propagated in the presence of a helper virus resulted in a concatamerized virus with a full length of approximately 33 kb. The full length structure of the tail-to-tail concatamerized recombinant virus is detailed in Figure IB.

Mice, either control lean or obese (ob/ob) were treated with a single tail intravenous infusion of 1-2 x 10l 1 particles of either: • HD-2ψ-leptin ~ a concatamerized virus according to this invention; • Ad-leptin ~ a so-called first generation adenoviral vector, non-concatamerized, containing viral genes other than El, and a leptin transgene expression cassette;

• control virus Ad-β-gal ~ the same first generation adenoviral vector as Ad-leptin, except that the transgene is β-gal; • an equal volume of dialysis buffer control.

Liver toxicity. Liver toxicity, as reflected by the significant elevation in AST and ALT serum levels over basal control levels, was observed only in mice treated with Ad-β-gal and Ad-leptin, but not HD-2ψ-leptin. Ad-vector-associated toxicity observed in both the lean and ob/ob treated mice was most significant at one week, was present but to a less significant extent at two weeks, and was resolved by four weeks post-treatment. In contrast, HD-2ψ-treatment was not associated with liver toxicity as reflected by the AST and ALT serum levels which were essentially indistinguishable from controls.

Hepatic sections of HD-2ψ-leptin treated lean mice were histologically also indistinguishable from control liver sections at all time points tested post treatment (one, two and four weeks). Occasional perivascular clusters of 50:50 T- and B-cells and small foci of cellular infiltrates in HD-2ψ-leptin-treated as well as in untreated control mice was observed. In contrast, Ad-leptin and Ad-β-gal treated mice displayed hepatic pathology throughout the post-treatment intervals. At one week post- treatment, both Ad-β-gal and Ad-leptin treated mice displayed degenerative hepatic pathology characterized by foci of round cell infiltration composed almost entirely (>98%) of T-cells, individual liver cell necrosis, increased liver cell mitotic activity, and dissociation of hepatic cords. At two weeks post-treatment, Ad-leptin treated mice displayed a similar, but less pronounced hepatic pathology. The inflammation and cellular infiltration observed resolved by four weeks post-treatment, there was almost an absence of lesions in the Ad-leptin treated mice, with only a trace of individual cell death present, which is within normal ranges. Examination of liver sections obtained from ob/ob mice reflected similar Ad-vector associated histopathology. Similar to the observations in lean mice, evidence of toxicity associated with Ad-vectors was not observed with HD-2ψ-leptin treatment in ob/ob mice; however, a slight cellular infiltrate was detected, which may be attributed to the immunogenicity of leptin in these leptin-deficient mice. Nonetheless, the extent of inflammation and cellular infiltrates remained significantly less than that observed with Ad-leptin.

Duration of transgene expression. In the lean mice, treatment with Ad-leptin resulted in a transient increase in serum leptin levels and weight loss that lasted for only 7-10 days. In contrast, treatment with HD-2ψ-leptin resulted in high serum leptin levels (6- to 10-fold over background) and approximately 20% weight loss that persisted at least two months. Weight loss inHD-2ψ-leptin-treated mice was associated with satiety that persisted over a longer period (2-3 weeks) than in those treated with Ad-leptin (5- 7 days). Vector DNA in the livers of Ad-leptin treated mice was rapidly lost and fewer than 0.2 copies per cell were detected, compared to 1-2 copies per cell following HD- 2ψ-leptin treatment at 8 weeks post-injection. These effects can be correlated with the duration of gene expression obtained with these two vector types. Gene expression mediated by Ad-leptin was transient and almost undetectable as early as 1 week post- treatment as seen by northern blot analysis of total liver RNA, whereas that mediated by HD-2ψ-leptin persisted for at least eight weeks. No changes in serum glucose or insulin levels in the treated lean mice were detected throughout the study. Vector DNA levels were stable at 1-2 copies per cell at 1, 2, 4 and 8 weeks post-treatment.

Leptin serum levels. The ob/ob mice are naive to leptin and thus transgene immunogenicity is not an unexpected finding. In these animals, similar to what was observed in the lean mice, HD-2ψ-leptin was found to be less immunogenic than the first-generation Ad-leptin vector. In the ob/ob mice treated with Ad-leptin, serum levels of leptin increased only for a short period during the first four days of treatment, returning to baseline levels within ten days post-injection. Increased leptin levels was associated with transient body weight loss of approximately 25%, followed by weight gain two weeks after treatment. Similar to the results obtained in lean mice, the Ad- leptin vector DNA was rapidly lost (less than 0.2 copies per cell were detected by 2 weeks post treatment, and undetectable by 8 weeks). In contrast, the ob/ob HD-2ψ-leptin-treated mice had increased serum leptin levels up to approximately 15 days post-treatment, after which the levels gradually dropped to baseline over the subsequent 25 days. The initial rise in leptin levels correlated with rapid weight reduction resulting in greater than 60% weight loss (reaching normal lean weight) by one month. Weight loss was maintained for a period of 6-7 weeks post-treatment. As leptin levels dropped to baseline, a gradual increase in body weight was observed. Satiety was observed in association with increased leptin levels, and appetite suppression was sustained for a longer period (approximately 1 month) compared to the short transient effect induced by Ad-leptin (approximately 10 days) (data not shown).

Leptin-specific antibodies were detected in the sera of ob/ob Ad-leptin- and HD-2ψ-leptin -treated mice, therefore it was essential to determine whether the drop observed in serum leptin levels was due to interference of the antibodies with the ELISA assay utilized to measure leptin, or a loss of vector DNA and/or gene expression. Although by Southern blot analysis greater stability of HD-2ψ-leptin DNA was observed over Ad-vector DNA in livers of ob/ob treated mice compared at similar time points, the analysis revealed eventual loss of the HD-2ψ-leptin DNA over the 8 week time interval. Approximately 75% less vector DNA was detected in the livers of HD-2ψ-leptin-treated mice at 4 and 8 weeks post-treatment compared to the persistent levels found in the livers of HD-2ψ-leptin -treated lean littermates at similar time points. Gene expression in ob/ob Ad-leptin-treated mice correlated with the DNA findings, RNA levels were beyond the sensitivity level of detection at one week post- treatment, whereas in HD-2ψ-leptin-treated mice, gene expression was detected up to four weeks post-injection and was undetectable at eight weeks.

Serum glucose and insulin. Serum glucose and insulin levels dropped during the first 1-2 weeks post-treatment to normal lean values in both HD-2ψ-leptin- and Ad-leptin- treated mice, although the effects of HD-2ψ-leptin treatment were sustained for longer periods, which parallels what was seen with weight loss, satiety, DNA stability and leptin gene expression. The subsequent increase in glucose and insulin levels in both vector treatments correlated with the drop observed in serum leptin levels and eventual loss of vector DNA. The overall HD-2ψ-leptin-mediated prolonged effect was also reflected in the accompanying phenotypic correction, which lasted longer than that seen in litter mates treated with Ad-leptin (6-7 versus 2-3 weeks).

Our studies clearly illustrate that HD-2ψ-leptin achieved a significant improvement in the safety profile and longevity of gene expression over that achieved with the first-generation Ad-leptin vector. The significant differences observed in the extent of cellular infiltrate in the liver, together with the pronounced liver toxicity as measured by approximately 10- and 5-fold increases in AST and ALT serum values associated with Ad-leptin but not HD-2ψ-leptin treatment in lean mice, can be directly attributed to the elimination of the Adenoviral protein-coding DNA sequences, since the leptin expression cassette was identical in both vectors. The appearance of a leptin-specific antibody response, gradual loss of gene expression and vector DNA observed in the ob/ob (leptin-deficient) but not in the lean mice (leptin-wild type) treated with HD-2ψ-leptin reflects an independent immune response event related to leptin tolerance.

It is to be understood that the use of the leptin transgene cassette is just a single embodiment of this invention. Virtually any expression cassette containing any desired transgene may be substituted for the illustrated leptin cassette.

In another embodiment of this invention, a non-concatamerized vector comprises the following elements (in 5' to 3' order): a) a first 5'-inverted terminal repeat sequence (ITR); b) a first packaging signal; c) one or more heterologous expression cassettes; d) a second packaging signal; and e) a second 5' ITR wherein the vector has an overall size of about 26 to about 38 kb, and wherein the only adenoviral sequences present in the vector are ITRs and packaging signals, and wherein the vector contains no bacterial plasmid-based sequences. Like the concatamerized vectors described previously, the non- concatamerized vector may contain "stuffer" DNA to increase its size. In this vector, the stuffer DNA may be placed anywhere in the vector, but is preferably placed between the first packaging signal and the hererologous expression unit (items b) and c)), and/or between expression units (c) or between the last expression unit and the second packaging signal (c) and d)). The amount of stuffer DNA will depend on the size of the heterologous expression cassette(s), but will be enough to bring the total size of the vector to about 26 to about 38 kb.

Another important feature of these vectors is that they contain non- bacterial plasmid based sequences. This is not to say that the heterologous expression cassettes (element C) cannot be a bacterial gene—the heterologous expression cassette may contain any desired gene and control (promoter, enhancer, termination sequence), regardless of origin. The vector does not contain extraneous sequences which are from the plasmid, however, such as a bacterial origin of replication, or marker genes which are soley useful in the plasmid culture of the vector.

It has been found that these vectors can be purified so that a population of the rescued vectors contains a very low level of contaminating helper virus (less than about 0.1%), thus making them useful as vectors for human gene therapy.

The following non-limiting Examples are presented to better illustrate the invention.

EXAMPLE 1 Construction of Vectors Construction of the first generation vectors designated "Ad-leptin" and "Ad-^β-gal" was essentially as described by Morsy, et al, 1998 Gene Therapy 5(1): 8- 18, which is hereby incorporated by reference. Briefly, the expression cassettes contain the HCMV promoter (Invitrogen, Carlsbad, CA), the transgene (either leptin or β-galactosidase), and the bovine growth hormone poly-A sequence. These vectors were propagated and titered as previously described by Graham et al, 1977 J. Gen. Virol. 36:59-72; and Graham et al, 1991 in Gene Transfer and Expression Protocols, ed. Murray, E. J. (The Humana Press Inc., Clinton, NJ), pp. 109-128, both of which are hereby incorporated by reference.

HD-2ψ-leptin was prepared by releasing the linear backbone structure of HD-leptin from its plasmid pΔSTK120-HCMV-mOb-BGHpA (by Pmel digest) (described in co-pending U.S. application Serial No. 08/878,737 and WO 97/48806, published December 24, 1997, both of which are hereby incorporated by reference, and further described below) and transfecting the linear DNA into 293 -CRE cells followed by helper infection as described below.

Two different structures were used for rescuing HD viruses expressing leptin, HD-2ψ-leptin and HD-leptin-monomer. The structure of HD-2ψ-leptin plasmid is a pBluescriptlLKS based plasmid that contains (in the following order): a) the Ad5 inverted terminal repeat sequences (ITR) and the packaging signal ψ, 440 base pairs (bp), (nucleotides (nt.) 1-440); b) "stuffer DNA" which is a 5072 bp fragment of hypoxanthine guanine phosphoribosyl transferase (HPRT) (nt.12373 - 17781 in Genbank gb : humhprtb); c) the leptin gene expression cassette, 1835 bp including the HCMV promoter, the leptin cDNA and the bovine growth hormone polyadenylation signal sequence; d) a Hindlll 9063 bp fragment of: C346 cosmid (nt.12421-21484 in Genbank gb :L31948); and e) right end terminus of Ad5, composed of the 3' ITR sequence, 117bp (nt. 35818-35935); with the intervening multiple cloning sites between junctions of the different fragments.

The total size of this construct is 19.6 kb, which includes 2.9 kb of the pBluescriptllKS. The 2.9 kb of pBluescriptllKS is eliminated prior to HD-vector rescue by linearizing the plasmid with two Pmel flanking sites.

HD-leptin-monomer plasmid (pSTK120-HCMV-mOb-BGHpA) differs in that the HPRT "stuffer DNA" is a larger fragment of 16054 bp (nt. 1799-17853 in Genbank gb:humhprtb). Total size of the HD-leptin-monomer plasmid is approximately 30 kb, which includes 2.9 kb of the pBluescriptllKS, which as in the case of HD-2ψ-leptin plasmid, is also eliminated by linearizating the plasmid with two Pmel flanking sites and releasing the HD-leptin-monomer fragment.

EXAMPLE 2 Propagation of HD-viruses

Propagation of both of the HD-viruses was accomplished using a helper virus system having an El -deleted vector with lox sites flanking the packaging signals (AdLC8clucl) Parks et al, (1996) Proc. Natl. Acad. Sci. USA 93:13565-13570, which is hereby incorporated by reference), and a 293 cell line derivative expressing Cre recombinase (293-cre4) (Lieber et al, 1996 J. Virol. 70:8944-8960 which is hereby incorporated by reference; and Parks, R. J. et al. 1996 supra. The HD-2ψ-leptin vector DNA was excised from the plasmid backbone by Pmel digestion and 4 μg were used to transfect semi-confluent 293-cre4 cells in 6 cm plates. Following an overnight incubation, cells were infected at a multiplicity of infection (moi) of 1 with the helper virus AdLC8clucl.

Cells were monitored for complete cytopathic effect (CPE), at which point cells were collected and lysate was used for serial propagation and expansion of viral stock by modification of the method described in Parks et al, 1996 supra, and Parks et al,1997, J. Virol. 71 : 3293-3298. Briefly, 2 mis of lysate collected from PI (the transfection/infection step) was used to infect 6 cm plates of semi-confluent 293- cre4 for 24 hours, supplemented with 1 ml of fresh media. After the 24 hour incubation, the helper virus AdLCδclucl was added at an moi of 1, dropwise to the cells. P2 lysate was collected upon detection of CPE. The same procedure was serially followed for another three propagations (P3, P4 and P5) infecting 10 cm plates followed by 15 cm plates of semi-confluent 293-cre4 cells, respectively. Lysate collected from P4 was used to infect twenty 15 cm plates (1 ml lysate added to 24 mis of fresh media, the lysate was cesium chloride banded as previously described Graham et al, 1991 supra. The banded viruses were analyzed by restriction mapping and the HD-2ψ-leptin virus was sequenced for verification of structure. The final stock of HD-2ψ-leptin was harvested from approximately 1.2 X 109 293-cre4 cells, and the cesium chloride banded viral stock yield was approximately 8 x 1012 particles (2 X 10l2/ml). For semi-quantitation of infectious particles, COS cells were infected with 10 μl of HD-2ψ-leptin or with Ad-leptin at an moi of 10 or 15. Cells were washed 30 minutes post-infection and serum-free media was added. One-hundred-microliter aliquots of media were collected from infected plates at 24, 30, 48 and 54 hours post-treatment, and compared by western blot analysis for leptin protein levels. The HD-2ψ-leptin mediated expression was equivalent to the 15 moi-infected plates, and based on the plaque-forming-unit (pfu) titer of Ad-leptin, the estimated infectious titer was approximately 1-2 x 10 0/ml with a particle to infectious unit ratio of approximately 1 : 100. The helper virus (AdLCδclucl) content in the HD-2ψ-leptin stock was 1.5 x 107 pfulvcA. Fifty microliters (1- 2 x lθH OD particles/dose, containing approximately 7.5 x 10$ pfu helper, i.e. less than 0.001% contamination with helper/ estimated infectious HD dose) of the stock were diluted with dialysis buffer to 100 μl for the mouse tail vein injections.

EXAMPLE 3 Repeat of HD-2ψ-leptin viral rescue Three independent rescues of the HD-leptin recombinant virus, initiated at the first step (PI), which is the transfection of pΔSTK120-HCMV-mOb-BGHρA resulted in an identical, and stable structure of HD-2ψ-leptin. Seven different enzymes were used for verifying the structures of the virus: Asp718, Eagl, Fsel, Hindlll, Pacl, Smal, and Xhol. Digested viral DNA was analyzed by Southern blot analysis: fragments were radio-labeled using T4 DNA polymerase; DNA was run on a 1.0 or 0.5%» (for sizing purposes in case of undigested DNA extracted from HD-2ψ-leptin and Ad-leptin) agarose gels in TAE buffer; gels were dried and exposed to auto- radiography for detection of radio-labeled, digested bands or stained with ethidium bromide and photographed (in case of undigested viral DNA). Fifty - 100 ng of purified viral DNA were used for each digestion.

Figure IB shows the structure and similarity of vectors according to this invention (33 kb), which are all tail-to-tail concatemerizations (junction is at the 3' ITR ends of ΔSTK120-HCMV-mOb-BGHpA).

In Figure 2, the gel labeled "Vector DNA" shows 0.5 μg of DNA extracted from the HD-2ψ- leptin viral stock (Lane A), Ad-leptin viral stock (Lane B) and the Pmel cut pΔSTK120-HCMV-mOb-BGHpA viral stock (Lane C) compared on a 0.5%) agarose gel for sizing. Both HD-2ψ- leptin (33 kb) and Ad-leptin (34 kb) extracted DNA migrate, as expected, between 38.5 - 29.9 kb, and the cut ΔSTK120- HCMV-mOb-BGHpA (16.7 kb) migrates between 17.1 and 15.0 kb, the smaller band corresponds to the plasmid backbone (2.9kb), and the faint band in lane A represents the trace amount of the propagated unconcatemerized HD-leptin (16.7 kb).

In the remaining gels, structures of ΔSTK120-HCMV-mOb-BGHpA (Lane 1) (gel extracted after separation of plasmid backbone by Pmel digestion) and the three HD-2ψ-leptin vectors (Lanes 2, 3, and 4) are compared by restriction analysis, as described in the Examples.

The expected fragment sizes for HD-2ψ-leptin are: Asp718: 15391-single band (-s), 6296-double band (-d), and 2501-d; Eagl: 20445-s and 6270/6266-d;

Fsel: 16523/16458-d;

Hindlll: 10207/10174-d, 5845-d, and 454/450-d; Pad: 16516/16465-d;

Sma l: 6701-d, 5163-d, 2180-d, 1715-s and 1589-d; Xho 1: 11833-d, 2964/2953-d; and 1701/1697-d bp.

The expected fragment sizes for ΔSTK120-HCMV-mOb-BGHpA are: Asp718: 7837-s, 6296-s, and 2501-s; Eagl: 10364-s and 6266-s; Fsel: 16458-s and 172-s; Hindlll: 10174-s, 5848-s, 450-s and 158-s; Pacl: 16465-s and 165-s;

Sma l: 6701-s, 5163-s, 2180-s, 1589-s and 997-s;

Xho 1: 11833-s, 2964-s, 1697-s and 136-s bp. Ml and M2 are DNA markers ( 8-48 Kb, Bio-Rad Laboratories, Hercules, CA and 1 kb DNA ladder, Life Technologies / GIBCO BRL, Gaithersburg, MD, respectively ). Detection of leptin protein expression mediated by the HD-leptin recom-binant virus in vitro. COS cells were infected at 15 and 10 multiplicities of infection (moi) of Ad- leptin, and HD-2ψ- leptin with 10 μl of the viral stock. Media were collected (100 μl) at 24, 30, 48 and 54 hours post-infection and from uninfected control cells (C-) and examined by western blotting for secreted leptin, using a polyclonal anti-leptin antibody (Santa Cruz Biotech.). Leptin was detected as a single band (approximately 16 kD).

EXAMPLE 4 PCR amplification the junction fragment and sequencing

A primer flanking the junction fragment was used for PCR, primer J4-F: 5'- CTC TTC TTC TGT CAC ACC CCT CCC-3' (SEQ.ID.NO. 1) was used individually to amplify the junction-fragment of HD-leptin. The fragment generated was approximately 300 bp, and was cloned into PCR 2.1 vector (Invitrogen, Carlsbad, CA) and sequenced.

EXAMPLE 5 Construction of HD viruses differing in the number of packaging signals. The structure of HD-2ψ-hVEGFl45 plasmid is a pBluescriptllKS based plasmid that contains (in the following order): a) the Ad5 inverted terminal repeat sequences (ITR) and the packaging signal ψ, 440 base pairs (bp), (nucleotides (nt.) 1-440); b) "stuffer DNA" which is a 16054 bp fragment of hypoxanthine guanine phosphoribosyl transferase (HPRT) (nt. 1799-17853 in Genbank gb:humhprtb; c) the expression cassette of the gene encoding hVEGF (human vascular endothelial growth factor —145 amino acids) , containing 1301bp, including the HCMV promoter, the cDNA and the SV40 early polyadenylation signal sequence; d) a Hindlll 9063 bp fragment of: C346 cosmid (nt.12421-21484 in Genbank gb :L31948); and e) a second packaging signal followed by a second left end ITR 440 base pairs (bp), (nucleotides (nt.) 1-440) with the intervening multiple cloning sites between junctions of the different fragments.

The total size of this construct is about 29 kb, which includes 2.9 kb of the pBluescriptllKS. The 2.9 kb of pBluescriptllKS is eliminated prior to HD-vector rescue by linearizing the plasmid with two Pmel flanking sites.

The structure of the HD-lψ-hVEGFl45 plasmid is identical to that of the HD-2ψ-hVEGFl45 plasmid except for its right end. It lacks the second packaging signal, but contains the 3' ITR sequence, 117bp (nt. 35818-35935). The plasmid also includes 2.9 kb of the pBluescriptllKS, which as in the case of HD-2ψ-leptin plasmid, is also eliminated by linearizating the plasmid with two Pmel flanking sites and releasing the HD-vector.

EXAMPLE 6 Rescue and propagation of HD vectors with 1 or 2 packaging signals (1 or 2ψ) Plasmids were transfected into 293 cre4 cells constitutively expressing the ere recombinase using the Calcium phoshate coprecipitation technique. Cells were infected with a helper virus (El -deleted vector with lox sites flanking the packaging signals, AdLC8BHG10lucl) 36 h after transfection. Both viruses were continuously passaged three times in 6 cm plates followed by 15 cm plates 3 times by infecting 293- cre4 cells with 1 ml (6 cm plates) or 5 ml (15 cm plates) cell lysate from the previous passage and the helper virus (1 plaque forming unit (pfu) per cell). Propagation of both viruses was carried out side by side. Cells from passage 6 (two 15 cm plates) were harvested, lysed and the lysate was cesium chloride banded as previously described Graham et al, 1991 supra. HD-vector and Helper were not separated during purification. The DNA was extracted from banded viruses and analyzed by restriction mapping using the restriction enzyme Hind III. Fragments were labeled using Klenow-Enzyme before electrophoresis and the gel was exposed to a film (Figure 4A). The restriction pattern for both HD-vectors is in agreement with the expected one. Note that a contamination with helper DNA is visible only for the vector containing one packaging signal.

HD-vector preparations were further analyzed. The amount of HD- vector particles produced per cell was determined by measurements of the optic density (OD) at 260 nm. Calculation was performed using the equation 1 OD= 1.1 x 1012 particles.

Productivity was 4.3 times higher for the vector containing 2ψ. 6.9 x 104 particles/cell of the HD hVEGFl45 2ψ were produced. The contamination of the HD-vectors by helper virus was determined by quantitative PCR using Taqman Technology. Helper contamination was 5-fold lower for the vector containing 2ψ (contamination 0.4%). Helper contamination was further reduced to 0.2% after three more passages.

EXAMPLE 7

In Vivo Studies Mouse colony, ob/ob (C57BL/J6-ob/ob) mice and homozygous normal lean (C57BL/J6) litter mates (age-matched females), were purchased from Jackson Laboratories (Bar Harbor, ME) for use in this study. Animals were free of all common murine pathogens. Eight-to twelve-week-old mice (ob/ob approximately 70 g and lean approximately 28 g) were re-distributed based on equal representation of weight and caged in groups of five on day 0, immediately preceding treatment. After a series of baseline blood samples were obtained by tail incision from conscious mice, animals were divided into four groups and received by tail vein injection a single 100 μl aliquot containing 1-2 XlOl particles of HD-2ψ-leptin, Ad-leptin, Ad-β-gal (control), or dialysis buffer (control). Body weight and food intake were measured daily and blood was collected 2-3 times weekly, pre- and post-treatment. Animals were killed by carbon dioxide inhalation and organs removed for immuno-histochemistry and RNA analysis. All animals used in this study were maintained in accordance with the "Guide for the Care Use of Laboratory Animals" (DHHS Publication No.(NIH) 85-23, revised 1996). The protocol was approved by the Institutional Animal Care and Use Committee, Merck Research Laboratories, West Point, PA.

Histopathology studies. Mice (n=3 / treatment / time point) were euthanatized, and liver samples were collected and fixed in 10% buffered formalin. Tissues were routinely processed through paraffin, sectioned at 5 microns, and stained with hematoxylin and eosin. Replicate unstained slides were also prepared using standard procedures for immuno-histochemistry and stained for the presence of CD3 (T cell) and CD45R (B cell) determinants on infiltrating or intrinsic cells (not shown).

Blood measurements. Blood samples were obtained by tail incision and collected into heparinized microhematocrit tubes (VWR) every 2-3 days during the course of the study. Tubes were centrifuged at 13,700 g for 2 minutes and hematocrit values were monitored. Plasma was collected for measurement of aspartate aminotransferase (AST), alanine amino-transferase (ALT), leptin, glucose and insulin levels. ALT and AST were measured by using the ALT/SGPT and AST/SGOT DT slides, respectively (Vitros Chemistry Products, Johnson & Johnson Clinical Diagnostics Inc., Rochester, NY). Leptin and insulin levels were measured by a radio-immunoassay performed by Linco Research, Inc. Glucose levels were measured using the Kodak Ektachem DT slides (Eastman Kodak Com.).

Northern and Southern blot analysis. For northern blot analysis, total RNA was extracted (Trizol, Gibco) from livers of Ad-leptin-treated and HD-leptin-2-treated mice at 1-, 2-, 4-, and 8-week intervals, and untreated mice. Leptin message was detected by northern blot analysis (Maniatis et al, 1982 Molecular Cloning: A Laboratory Manual Cold Spring Harbor Laboratory, Cold Spring Harbor, NY) using leptin cDNA as a probe which recognizes a single approximately 500 bp band Morsy et al, 1997 supra. A probe for β-actin was used as the internal control (approximately 1 kb) (Biochain Inc.). Southern blot analysis (Maniatis et al, 1982 supra) was used to investigate the stability of vector DNA.

Genomic DNA was extracted from livers (pooled DNA, n=3 / treatment / time point) of ob/ob and lean mice treated with Ad- vectors, Ad-β-gal, Ad- leptin or HD-2ψ-leptin. Control animals were injected with similar volumes of dialysis buffer. Pooled (n=3) genomic DNA was digested with Hindlll restriction enzyme, and 20 μg of digested DNA were loaded on 0.8 % TAE agarose gels.

For copy number estimation, 20 μg control DNA were spiked with HD vector DNA equivalent to 2.0, 1.0, 0.2 and 0.1 copies per cell, and the mixture digested with Hindlll restriction enzyme followed by Southern blotting. The filters were hybridized with a mouse leptin cDNA (approximately 500 bp) probe, which hybridized to a single Hindlll fragment containing the leptin insert in both the HD-2ψ- leptin (approximately 6 kb), and Ad-leptin (approximately 1.2 kb) vectors. Developed autoradiographs were scanned (Personal Densitometer SI, Molecular Dynamics, Sunnyvale, CA) and the relative band densities quantitated (Image QuaNT software, Molecular Dynamics). For estimation of relative vector DNA stability between treatment time points at each time point, a detected internal leptin genomic DNA signal was used as control both for DNA loading normalization. Copy number equivalence were assigned based on comparisons to the relative density ratio between the internal genomic signal and the leptin signal of spiked vector DNA in a copy number control experiment.

DETAILED DESCRIPTION OF THE INVENTION

This invention is related to improved adenoviral vectors which are useful as delivery systems in gene therapy. Specifically, it has been found in accordance with this invention that an adenoviral vector comprising a packaging signal sequence at either end has improved characteristics for its propagation and use as a vector. One embodiment of this invention is a helper-dependent (HD) adenoviral stock comprising less than about 0.5% helper virus contamination, wherein the HD adenovirus is devoid of adenoviral DNA except for inverted terminal repeat and packaging signal sequences. In another embodiment, the HD adenoviral stock is free from detectable levels of helper virus contamination.

Another embodiment of this invention is a concatamerized vector comprising the following elements (in 5' to 3' order): a first 5 '-inverted terminal repeat sequence (ITR), a first packaging signal, a first heterologous expression cassette, a first 3' ITR an optional second 3' ITR, a second heterologous expression cassette, a second packaging signal and a second 5' ITR and which has an overall size of about 26 to about 38 kb, and wherein the only adenoviral sequences present in the vector are ITRs and packaging signals.

Another embodiment of this invention is a single copy vector comprising the following elements (in 5' to 3' order): a first 5'-inverted terminal repeat sequence (ITR), a first packaging signal, one or more heterologous expression cassettes, a second packaging signal and a second 5 'ITR and which has an overall size of about 26 to about 38 kb, wherein the only adenoviral sequences present in the vector are ITRs and packaging signals, and wherein there are no bacterial plasmid based sequences (such as an origin of replication or bacterial marker genes) present in the adenoviral vector.

Throughout this specification and claims, viral vectors of this invention will be referred to as HD-2ψ vectors.

In addition to the above elements, the vectors may comprise additional, non-expressed portions of DNA, termed "stuffer DNA", which may be inserted at one or more locations between the above-recited elements. A vector containing only a heterologous expression cassette, adenoviral ITR and packaging signals may be quite small, in general about 5-10 kb, depending on the size of the heterologous expression cassette. Such a small virus does not package efficiently. Even if duplicated, the virus may be still too small to package efficiently. However, if the size of the vector is more than about 12 kb and vector is duplicated to generate an HD virus of an overall sized of more than about 25 kb, then efficient packaging can occur. Also, if non-expressed "stuffer" DNA is added to the vector so that its final size is at least about 75% of a wild-type adenovirus, then efficient packaging will occur. Stuffer DNA thus may be any DNA used to increase the size of a vector. It has been found, in accordance with this invention, that populations of concatamerized viruses are remarkably homogeneous, and importantly, have a very low level of helper virus contamination. Thus, another aspect of this invention is a population of concatamerized adenoviral vectors which comprises less than about 0.1% per infectious HD unit and less than 1 pfu of helper virus/100, 000 OD particles (minimum estimated HD infectious unit: OD particle is 1 : 100).

Another aspect of this invention is the accelerated accumulation of propagated HD vectors at early passages. HD vectors require more passages than first generation adenoviruses until a stock is suitable for purification and characterization. At early passages only a small percentage of cells is infected with the HD vector. Because the remaining cells are still infected with helper virus in the absence a packageable HD genome, these cells are likely to generate packagable variants of the helper. Vectors with the described features (2ψ-vectors and concatamerized vectors) reach high titer already at passage 5 or 6, thereby minimizing the risk of helper variants.

Claims

WHAT IS CLAIMED:

1. A helper-dependent (HD) adenovirus population comprising less than about 0.5% helper virus contamination, wherein each HD adenovirus in the population is devoid of adenoviral DNA except for inverted terminal repeat and packaging sequences.

2. A helper-dependent (HD) adenoviral vector comprising the following elements: (A) a first 5'-inverted terminal repeat sequence (ITR);

(B) a first packaging signal sequence element;

(C) a first heterologous DNA and heterologous expression cassette;

(D) a second packaging signal sequence element; and

(E) a second 5 'ITR; and wherein the vector has an overall size of about 26 to about 38 kb, and wherein the only adenoviral sequences present in the vector are the ITRs and packaging signals.

3. A vector according to Claim 2 further comprising a second heterologous expression cassette (F) between the first heterologous expression cassette (C) and the second packaging signal (element D).

4. A vector according to Claim 2, further comprising stuffer DNA.

5. A vector according to Claim 4 wherein the first heterologous expression cassette is substantially the same as the second heterologous expression cassette.

6. A vector according to Claim 4 wherein the first heterologous expression cassette comprises a first transgene and wherein the second heterologous expression cassette comprises a second transgene, and wherein the first and second transgenes are not the same transgene.

7. A vector according to Claim 3 wherein the first and second heterologous expression cassettes comprise a leptin gene.

8. A concatamerized adenoviral vector comprising the following elements (in 5' to 3' order):

(A) a first 5'-inverted terminal repeat sequence (ITR);

(B) a first packaging signal element;

(C) a first heterologous DNA and heterologous expression cassette;

(D) a first 3' ITR;

(E) an optional second 3 'ITR;

(F) a second heterologous DNA and heterologous expression cassette; (G) a second packaging signal; and

(H) a second 5'ITR; and the vector further comprising stuffer DNA in an amount sufficient so that the vector has an overall size of about 26 to about 38 kb, and wherein the only adenoviral sequences present in the vector are the ITRs and packaging signals.

9. A host cell comprising the vector of Claim 1.

10. A population of helper-dependent adenoviruses which comprises less than 1 pfu helper virus per 100,000 OD particles of helper-dependent viral stock.

11. A single copy adenoviral vector comprising the following elements (in 5' to 3' order):

(A) a first 5'-inverted terminal repeat sequence (ITR); (B) a first packaging signal; (C) one or more heterologous DNA expression cassettes;

(D) a second packaging signal; and

(E) a second 5'ITR wherein the vector has an overall size of about 26 to about 38 kb; wherein the only adenoviral sequences present in the vector are the ITRs and packaging signals; and wherein there are no bacterial plasmid based sequences present in the adenoviral vector.

12. A vector according to Claim 11 wherein element C comprises a first heterologous DNA expression cassette and a second heterologous DNA expression cassette.

13. A vector according to Claim 11, further comprising stuffer DNA.

14. A vector according to Claim 12 wherein the first heterologous expression cassette is substantially the same as the second heterologous expression cassette.

15. A vector according to Claim 12 wherein the first heterologous expression cassette comprises a first transgene and wherein the second heterologous expression cassette comprises a second transgene, and wherein the first and second transgenes are not the same transgene.

16. A vector according to Claim 14 wherein the first and second heterologous expression cassettes comprise a human vascular endothelial growth factor gene.

17. A host cell comprising the vector of Claim 11.

18. A method of transferring a transgene to an animal comprising introducing into the animal the adenoviral vector population of Claim 1.

19. A method according to Claim 18 wherein the animal is human.