Synthesis of adenoviral targeting molecules by intein-mediated protein ligation.

Adenoviral vectors infect cells through the binding of capsid proteins to cell-surface receptors. The ubiquitous expression of adenoviral receptors in human tissues represents an obstacle toward the development of systemically deliverable vectors for cancer therapy, since effective therapy may require delivery to specific sites. For these reasons, major efforts are directed toward the elimination of the native tropism combined with identification of ligands that bind to tumor-specific cell-surface proteins. Highthroughput technologies have identified potential targeting ligands, which need to be evaluated for their ability to retarget adenovirus to alternative receptors. Here, we present a strategy that permits the routine analysis of adenoviral targeting ligands. We use intein-mediated protein ligation as a means to produce functional biological molecules, that is, adenoviral targeting molecules that function as adapters between cellular receptors and the adenovirus fiber protein. We demonstrate the versatility of the present system by conjugating targeting ligands that differ in size and nature including an apolipoprotein E synthetic peptide, the basic fibroblast growth factor and folic acid. The resulting adenoviral targeting molecules mediate adenoviral gene delivery in cells that express the corresponding receptor.

A genetically modified adenoviral vector exhibits enhanced gene transfer of human smooth muscle cells.

Adenoviral vector-based gene therapy is a promising approach for the treatment of restenosis postangioplasty. However, a high concentration of adenoviral vector can cause cellular activation, damage, and an enhanced immune response. One approach to solving this problem is to increase gene transfer efficiency by directing adenoviral vector entry via an alternate receptor system. We have constructed an adenoviral vector, Av9LacZ, that encodes the beta-galactosidase gene and contains a chimeric fiber protein that redirects viral vector binding to the Ad3 adenoviral receptor on the host cell. We examined the ability of Av9LacZ to transduce primary human smooth muscle cells (SMC) and found that it showed a 10- to 15-fold higher transduction efficiency when compared to the prototypic adenoviral vector currently used for preclinical and clinical studies. While both vectors were able to transduce rabbit, pig and monkey SMCs, the genetically modified vector transduced human SMC with much higher efficiency. SMC obtained from the aorta, coronary, renal, popliteal and pulmonary arteries were all efficiently transduced by Av9LacZ. Consistent with the data obtained from cultured cells, Av9LacZ also transduced fresh human arterial tissues considerably more efficiently than Av1LacZ. We conclude that the large discrepancy between transduction of animal and human cells by conventional vectors supports a cautious extrapolation of the results of in vivo animal studies to man. Furthermore, the genetically modified AV9 vector may deliver better efficacy and studies in large animal models with this vector could be more predictive of therapeutic efficacy in the treatment of human restenosis.

Enhanced gene transfer to rabbit jugular veins by an adenovirus containing a cyclic RGD motif in the HI loop of the fiber knob.

Gene therapy using recombinant adenoviral vectors represents a promising therapeutic tool to prevent vein graft stenosis, the main complication of coronary artery bypass grafting. However, the low transduction efficiency of vascular smooth muscle cells and endothelial cells (EC) is a potential limitation, presumably due to the low levels of functional adenovirus receptor (coxsackie:adenovirus receptor; CAR). Designing vectors specifically targeted to alpha(v) integrins is a strategy that might overcome the poor expression of CAR in vascular smooth muscle cells and EC. RGD, a receptor-binding motif that can interact with alpha(v) integrins, was inserted into the HI loop and at the C-terminus of the adenoviral fiber protein in two separate adenovirus vectors encoding a beta-galactosidase reporter gene. Av1nBgCRGD (C-terminus) and Av1nBgHIRGD (HI loop) were evaluated in EC in culture and in jugular vein organ culture. Transduction of primary rat and rabbit EC with Av1nBgHIRGD was significantly more efficient when compared to Av1nBgCRGD or Av1nBg. Transduction of mouse, rat and rabbit jugular veins in organ culture using Av1nBg showed that adenovirus-mediated gene expression was greatest in rabbit jugular veins compared to rat and mouse veins. Av1nBgHIRGD augmented gene expression approximately four-fold in rabbit jugular veins when compared to Av1nBg. Histochemical analysis showed that numerous EC but few smooth muscle cells were transduced at all vector concentrations. A substantial number of adventitial fibroblasts were transduced only at the highest vector concentrations of Av1nBgHIRGD. These findings demonstrate that integrin-targeted vectors allow for enhanced gene delivery to veins and strengthen the viability of adenoviral-mediated gene transfer of therapeutic transgenes to human veins prior to vein grafting.

Adenovirus type 5 viral particles pseudotyped with mutagenized fiber proteins show diminished infectivity of coxsackie B-adenovirus receptor-bearing cells.

A major limitation of adenovirus type 5 (Ad5)-based gene therapy, the inability to target therapeutic genes to selected cell types, is attributable to the natural tropism of the virus for the widely expressed coxsackievirus-adenovirus receptor (CAR) protein. Modifications of the Ad5 fiber knob domain have been shown to alter the tropism of the virus. We have developed a novel system to rapidly evaluate the function of modified fiber proteins in their most relevant context, the adenoviral capsid. This transient transfection/infection system combines transfection of cells with plasmids that express high levels of the modified fiber protein and infection with Ad5.beta gal.Delta F, an E1-, E3-, and fiber-deleted adenoviral vector encoding beta-galactosidase. We have used this system to test the adenoviral transduction efficiency mediated by a panel of fiber protein mutants that were proposed to influence CAR interaction. A series of amino acid modifications were incorporated via mutagenesis into the fiber expression plasmid, and the resulting fiber proteins were subsequently incorporated onto adenoviral particles. Mutations located in the fiber knob AB and CD loops demonstrated the greatest reduction in fiber-mediated gene transfer in HeLa cells. We also observed effects on transduction efficiency with mutations in the FG loop, indicating that the binding site may extend to the adjacent monomer in the fiber trimer and in the HI loop. These studies support the concept that modification of the fiber knob domain to diminish or ablate CAR interaction should result in a detargeted adenoviral vector that can be combined simultaneously with novel ligands for the development of a systemically administered, targeted adenoviral vector.

Targetable gene delivery vectors.

Adenoviral vectors, which have targeting ligands for tumor cells on the capsid, no natural tropism, and carry a therapeutic payload should be constructed soon and tested in pre-clinical models. Nevertheless, there are still important considerations for the design and therapeutic use of targetable vectors. Perhaps the single greatest challenge in the future, as it was in the past, will be finding ligands that have a higher apparent affinity for tumor and/or tumor endothelial cells then normal cells. However, the advent of many rapidly advancing technologies and information including the sequencing of the human genome, in vivo and in vitro phage display, rapid analysis of gene and protein expression in any context, and new cellular targets such as angiogenic endothelial cells, may provide many opportunities for the discovery of novel and useful ligands. In addition, the interests in targeting vectors are rapidly growing with new journals and meetings solely devoted to this subject increasing annually. Within the next 5 years, we should have meaningful clinical data on targetable vectors to reassess our progress.

Adenovirus vector pseudotyping in fiber-expressing cell lines: improved transduction of Epstein-Barr virus-transformed B cells.

While adenovirus (Ad) gene delivery vectors are useful in many gene therapy applications, their broad tropism means that they cannot be directed to a specific target cell. There are also a number of cell types involved in human disease which are not transducible with standard Ad vectors, such as Epstein-Barr virus (EBV)-transformed B lymphocytes. Adenovirus binds to host cells via the viral fiber protein, and Ad vectors have previously been retargeted by modifying the fiber gene on the viral chromosome. This requires that the modified fiber be able to bind to the cell in which the vector is grown, which prevents truly specific vector targeting. We previously reported a gene delivery system based on a fiber gene-deleted Ad type 5 (Ad5) vector (Ad5.betagal.DeltaF) and packaging cells that express the viral fiber protein. Expression of different fibers in packaging cells will allow Ad retargeting without modifying the viral chromosome. Importantly, fiber proteins which can no longer bind to the producer cells can also be used. Using this approach, we generated for the first time pseudotyped Ad5.betagal.DeltaF particles containing either the wild-type Ad5 fiber protein or a chimeric fiber with the receptor-binding knob domain of the Ad3 fiber. Particles equipped with the chimeric fiber bound to the Ad3 receptor rather than the coxsackievirus-adenovirus receptor protein used by Ad5. EBV-transformed B lymphocytes were infected efficiently by the Ad3-pseudotyped particles but poorly by virus containing the Ad5 fiber protein. The strategy described here represents a broadly applicable method for targeting gene delivery to specific cell types.

Adenovirus-mediated expresssion of the murine ecotropic receptor facilitates transduction of human hematopoietic cells with an ecotropic retroviral vector.

One factor limiting the ability to modify human repopulating hematopoietic cells genetically with retroviral vectors is the relatively low expression of the cognate viral receptor. We have tested sequential transduction of human hematopoietic cells with an adenoviral vector encoding the ecotropic retroviral receptor followed by transduction with an ecotropic retroviral vector. Adenoviral transduction of K562 erythroleukemia cells was highly efficiently with >95% of cells expressing the ecotropic receptor at a multiplicity of infection (MOI) of 103with a correspondingly high transduction with a retroviral vector. Ecotropic receptor expression in CD34+ cells following transduction with adenoviral vectors was increased by at least two-fold (from 20 to 48%) by replacing the RSV promoter with the CMV E1a promoter, resulting in a parallel increase in retroviral transduction efficiency. Replacing the head portion of the fiber protein in conventional adenoviral vectors (serotype 5) with the corresponding portion from an adenoviral 3 serotype resulted in ecotropic receptor expression in 60% of CD34+ cells at an MOI of 104 and a retroviral transduction of 60% of hematopoietic clonogenic progenitors. The sequential transduction strategy also resulted in efficient transduction of the primitive CD34+CD38- subset suggesting that it may hold promise for genetic modification of human hematopoietic stem cells.

Enhanced neuroblastoma transduction for an improved antitumor vaccine.

BACKGROUND: A recent clinical trial of an antineuroblastoma vaccine used adenovirus serotype 5 (Ad5) vectors to transduce autologous tumor cells with the gene encoding IL-2. A method to improve transduction efficiency was sought to enable the use of lower viral titers, especially when in situ adenoviral-mediated tumor cell transduction is considered. MATERIALS AND METHODS: A chimeric adenoviral delivery vector was utilized in which the fiber head from adenovirus serotype 3 was incorporated into the backbone of Ad5. Since the fiber head protein is responsible for viral attachment to target cells, a different spectrum and range of infectivity might result. Both the chimeric (Av9LacZ4) and Ad5 (Av1LacZ4) vectors were constructed to carry a beta-galactosidase transgene. The relative transduction efficiency of these two vectors was then evaluated in five tumor-derived short-term neuroblastoma cultures and four established neuroblastoma cell lines. Enzyme activity was assessed using three different methods: in situ staining, flow cytometric analysis, and a quantitative assay. RESULTS: A significant improvement in transduction efficiency of the short-term neuroblastoma cultures with the new chimeric adenovector was demonstrated. A similar improvement in transduction efficiency was not observed in the established cell lines, suggesting that the cell surface receptor for the Ad 3 serotype had been lost in vitro. Increased transduction of tumor cells with N-myc amplification was also observed. CONCLUSIONS: The newly constructed chimeric adenoviral vector transduces short-term neuroblastoma cultures more efficiently than the standard Ad5 vector. This vector will permit the use of lower viral titers and may be useful in other adenoviral-based gene-therapy protocols. Increased transgene expression in N-myc-amplified cells offers possible selectivity for in situ gene delivery.

Selective targeting of human cells by a chimeric adenovirus vector containing a modified fiber protein.

The adenovirus fiber protein is responsible for attachment of the virion to unidentified cell surface receptors. There are at least two distinct adenovirus fiber receptors which interact with the group B (Ad3) and group C (Ad5) adenoviruses. We have previously shown by using expressed adenovirus fiber proteins that it is possible to change the specificity of the fiber protein by exchanging the head domain with another serotype which recognizes a different receptor (S. C. Stevenson et al., J. Virol. 69:2850-2857, 1995). A chimeric fiber cDNA containing the Ad3 fiber head domain fused to the Ad5 fiber tail and shaft was incorporated into the genome of an adenovirus vector with E1 and E3 deleted encoding beta-galactosidase to generate Av9LacZ4, an adenovirus particle which contains a chimeric fiber protein. Western blot analysis of the chimeric fiber vector confirmed expression of the chimeric fiber protein and its association with the adenovirus capsid. Transduction experiments with fiber protein competitors demonstrated the altered receptor tropism of the chimeric fiber vector compared to that of the parental Av1LacZ4 vector. Transduction of a panel of human cell lines with the chimeric and parental vectors provided evidence for a different cellular distribution of the Ad5 and Ad3 receptors. Three cell lines (THP-1, MRC-5, and FaDu) were more efficiently transduced by the vector containing the Ad3 fiber head than by the Ad5 fiber vector. In contrast, human coronary artery endothelial cells were transduced more readily with the vector containing the Ad5 fiber than with the chimeric fiber vector. HeLa and human umbilical vein endothelial cells were transduced at equivalent levels compared with human diploid fibroblasts, which were refractory to transduction with both vectors. These results provide evidence for the differential expression of the Ad5 and Ad3 receptors on human cell lines derived from clinically relevant target tissues. Furthermore, we show that exchange of the fiber head domain is a viable approach to the production of adenovirus vectors with cell-type-selective transduction properties. It may be possible to extend this approach to the use of ligands for a range of different cellular receptors in order to target gene transfer to specific cell types at the level of transduction.

Human adenovirus serotypes 3 and 5 bind to two different cellular receptors via the fiber head domain.

The adenovirus fiber protein is responsible for attachment of the virion to cell surface receptors. The identity of the cellular receptor which mediates binding is unknown, although there is evidence suggesting that two distinct adenovirus receptors interact with the group C (adenovirus type 5 [Ad5]) and the group B (Ad3) adenoviruses. In order to define the determinants of adenovirus receptor specificity, we have carried out a series of competition binding experiments using recombinant native fiber polypeptides from Ad5 and Ad3 and chimeric fiber proteins in which the head domains of Ad5 and Ad3 were exchanged. Specific binding of fiber to HeLa cell receptors was assessed with radiolabeled protein synthesized in vitro, and by competition analysis with baculovirus-expressed fiber protein. Fiber produced in vitro was found as both monomer and trimer, but only the assembled trimers had receptor binding activity. Competition data support the conclusion that Ad5 and Ad3 interact with different cellular receptors. The Ad5 receptor distribution on several cell lines was assessed with a fiber binding flow cytometric assay. HeLa cells were found to express high levels of receptor, while CHO and human diploid fibroblasts did not. A chimeric fiber containing the Ad5 fiber head domain blocked the binding of Ad5 fiber but not Ad3 fiber. Similarly, a chimeric fiber containing the Ad3 fiber head blocked the binding of labeled Ad3 fiber but not Ad5 fiber. In addition, the isolated Ad3 fiber head domain competed effectively with labeled Ad3 fiber for binding to HeLa cell receptors. These results demonstrate that the determinants of receptor binding are located in the head domain of the fiber and that the isolated head domain is capable of trimerization and binding to cellular receptors. Our results also show that it is possible to change the receptor specificity of the fiber protein by manipulation of sequences contained in the head domain. Modification or replacement of the fiber head domain with novel ligands may permit adenovirus vectors with new receptor specificities which could be useful for targeted gene delivery in vivo to be engineered.

Phenotypic correction of hypercholesterolemia in apoE-deficient mice by adenovirus-mediated in vivo gene transfer.

To investigate the potential use of apoE in gene therapy of hyperlipidemias, an adenoviral vector was constructed that contained the human apoE3 cDNA under the control of the RSV promoter (Av1RE). Transduction of HepG2 cells resulted in the overexpression of human apoE secreted into the culture medium. Intravenous injection of 5 x 10(11) Av1RE vector particles into apoE-deficient mice resulted in expression of human apoE3 in mouse plasma at levels of 1.2 +/- 0.4 micrograms/L (mean +/- SEM, n = 5) 7 days after injection. Mice injected with the control vector Av1Lacz4 did not express detectable levels of human apoE. Average plasma cholesterol concentrations were reduced approximately eightfold from 737.5 +/- 118 mg/dL (mean +/- SEM, n = 6) to 98.2 +/- 4.4 mg/dL (mean +/- SEM, n = 5) and were unaffected in the control vector group. Expression of human apoE resulted in a shift in the plasma lipoprotein distribution from primarily VLDL and LDL in the control mice to predominantly HDL in the Av1RE-treated group. Western blot analysis of fast protein liquid chromatography-fractionated mouse plasma showed that the human apoE protein was associated with VLDL, LDL, and HDL. Correction of the hyperlipidemic condition found in the apoE-knockout mouse strain by direct in vivo gene transfer establishes the potential of this approach for treatment of hyperlipidemia caused by apoE deficiency or malfunction in human disease.

Human plasma cholesteryl ester transfer protein consists of a mixture of two forms reflecting variable glycosylation at asparagine 341.

Plasma cholesteryl ester transfer protein (CETP) mediates the transfer of neutral lipids and phospholipids between the plasma lipoproteins. The deduced M(r) of the CETP polypeptide from the cDNA is 53,000, but in sodium dodecyl sulfate (SDS) gels plasma CETP appears as a broad band containing two different molecular forms of M(r) 65,000-71,000. The purpose of this study was to see if variable N-linked glycosylation could explain the microheterogeneity of CETP. Recombinant CETP (rCETP), derived from stable expression of the CETP cDNA in Chinese hamster ovary (CHO) cells, appeared as a protein doublet comparable to plasma CETP. Digestion of plasma or rCETP with N-glycosidase F (glyco F, to remove N-linked carbohydrates) resulted in the formation of a lower M(r) doublet in which the bottom band approximated the M(r) of the CETP polypeptide. Metabolic labeling of the rCETP with [3H]mannose and [3H]glucosamine, followed by digestion with glyco F, suggested that the top band of the doublet contained residual N-linked carbohydrates resistant to glyco F digestion. To explore this hypothesis further, each of the four potential N-linked glycosylation sites of CETP (at amino acid positions 88, 240, 341, and 396) was eliminated by mutagenesis of asparagine to glutamine. The wild-type (WT) and mutant CETP cDNAs were transiently expressed in COS-7 cells. Each mutant CETP showed a lower M(r) than WT, indicating that all four sites were occupied by N-linked carbohydrate.(ABSTRACT TRUNCATED AT 250 WORDS)

Role of apolipoprotein E on cholesteryl ester-enriched low density lipoprotein particles in coronary artery atherosclerosis of hypercholesterolemic nonhuman primates.

Significant differences among individuals occur in the lipoprotein response to atherogenic diets in cynomolgus and African green monkeys. The range of concentrations of total plasma cholesterol (TPC) was 100-600 mg/dl and of apolipoprotein (apo) E (quantified by enzyme-linked immunosorbent assay) was 3-20 mg/dl in the animal groups of this study. The correlation between the concentrations of TPC and of apo E was r = 0.89 in these animals. To determine which lipoprotein classes contained the majority of apo E, agarose gel-filtration chromatography was used to subfractionate whole plasma. In hypercholesterolemic monkeys, the majority of the apo E and apo B-100 coeluted within the region of low density lipoprotein (LDL). In normocholesterolemic monkeys, the majority of apo E coeluted with apo A-I and high density lipoproteins. A strong positive correlation was seen between the concentrations of plasma apo E and LDL cholesterol (r = 0.9), but there was no significant correlation between high density lipoprotein apo E and either TPC or plasma apo E concentrations. A positive correlation of r = 0.8 was found between the LDL apo E to apo B-100 molar ratio and the average LDL particle size, suggesting an increase in the number of apo E molecules on the larger LDL particles. Within individual animals, LDL were heterogeneous and the LDL subfractions of lower density (1.02 less than d less than 1.03 g/ml) had the highest proportion of apo E, although apo E was present on LDL of all densities. A strong positive correlation between plasma apo E concentration and coronary artery atherosclerosis was identified, and in stepwise regression analysis, plasma apo B concentration and the apo E to apo B molar ratio of LDL together accounted for more than 90% of the variation in the atherosclerosis end point of coronary artery intimal area. These data strongly suggest that the enrichment of LDL with cholesteryl esters and apo E, which occurs in hypercholesterolemic primates, is an atherogenic feature of the plasma lipoproteins.

Fibrinogen and glycoprotein IIb/IIIa localization during platelet adhesion. Localization to the granulomere and at sites of platelet interaction.

Platelet membrane glycoprotein IIb-IIIa plays a focal role in primary hemostasis by serving as the cell surface receptor for fibrinogen. Recent studies by several groups have suggested that GPIIb-IIIa, which is dispersed randomly in the resting cell, undergoes migration leading to receptor clustering after platelet activation. The authors have investigated this activation-dependent relocation of fibrinogen receptors on platelets adherent to a standardized artificial surface. The correlative use of immunogold electron microscopy, ligand-gold binding, and stereo (three-dimensional) electron microscopy (EM) revealed specific localization of fibrinogen and its receptor at points of platelet to platelet interaction. Fibrinogen distribution on the plasma membrane, studied through the use of fibrinogen-gold conjugates with whole-mount adherent platelets, was primarily over the granulomere and at the cell periphery corresponding to sites of platelet-platelet interaction. Compared with the general hyalomere, fibrinogen density over the granulomere and at contact regions was increased 12-fold and 22-fold, respectively, and the specificity of binding at these sites was verified by positive competition with native fibrinogen, one of its degradation products (Fragment D1), and by monoclonal antibodies (HP1-1d and AP-2) specific for GPIIb-IIIa. The distribution of receptor antigens, localized by immunogold EM using antibodies against GPIIb-IIIa, also was localized over the hyalomere, where fibrinogen did not bind. To understand this apparently nonfunctional hyalomere GPIIb-IIIa further, correlative immunocytochemistry was performed using polyclonal and monoclonal antibodies for GPIIb and GPIIIa simultaneously. Colocalization of the antigens was observed consistently over the granulomere and at regions of cell contact, whereas the hyalomere antigens tended to be nonassociated. The studies document GPIIb-IIIa as a function complex at sites of cell interaction where fibrinogen binds.