In vivo studies on non-viral transdifferentiation of liver cells towards pancreatic ß cells.

Transdifferentiation in vivo is an attractive option for autologous replacement of pancreatic ß cells in patients with type 1 diabetes. It has been achieved by adenoviral delivery of genes for transcription factors in the liver and pancreas of hyperglycaemic mice. However, these viral approaches are not clinically applicable. We used the hydrodynamic approach to deliver genes Pdx1, Ngn3 (Neurog3) and MafA singly and in combination to livers of normoglycaemic rats. Five expression plasmids were evaluated. Livers were removed 1, 3, 7, 14 and 28 days after gene delivery and assayed by quantitative PCR, semi-quantitative PCR and immunohistology. Functional studies on hyperglycaemic rats were performed. The highest and most sustained expression was from a CpG-depleted plasmid (pCpG) and a plasmid with an in-frame scaffold/matrix attachment region ((pEPI(CMV)). When Pdx1, Ngn3 and MafA were delivered together to normoglycaemic rats with these plasmids, insulin mRNA was detected at all time points and was ~50-fold higher with pCpG. Insulin mRNA content of livers at days 3 and 7 was equivalent to that of a pancreas, with scattered insulin-positive cells detected by immunohistology, but levels declined thereafter. Prohormone convertase 1/3 was elevated at days 3 and 7. In hyperglycaemic rats, fasting blood glucose was lower at days 1, 3 and 7 but not thereafter, and body weight was maintained to day 28. We conclude that hydrodynamic gene delivery of multiple transcription factors to rat liver can initiate transdifferentiation to pancreatic ß cells, but the process is reversible and probably requires more sustained transcription factor expression.

Critical physiological and surgical considerations for hydrodynamic pressurization of individual segments of the pig liver.

Hydrodynamic gene delivery to the liver is a promising approach for liver gene therapy in the clinic, but levels of gene expression in larger species have been much less than in rodents. The development of surgical techniques for pressurizing individual liver segments and the establishment of whether hepatic vascular anatomy in fact permits pressurization of individual segments are critical issues that need to be addressed. We have evaluated these issues using hydrodynamic delivery to individual segments of the pig liver, via branches of both portal and hepatic veins. Our objective was to develop surgical techniques that achieve elevated vascular pressures within individual liver segments with small volumes, but without interruption of portal blood flow or reduction in venous return to the heart. We report that, without specific surgical interventions to obstruct outflow of DNA solution from the targeted liver segment, little or no increase in intrahepatic vascular pressure occurs. We demonstrate, for the first time, that selective pressurization of individual liver segments is possible without compromising portal venous flow or venous return to the heart. Thus, hydrodynamic gene delivery to individual liver segments is technically achievable in a clinical setting, but will require open abdominal surgery rather than minimally invasive techniques.

Self-assembly of peptides into spherical nanoparticles for delivery of hydrophilic moieties to the cytosol.

We report a novel class of self-assembling peptide nanoparticles formed by mixing aqueous solutions of K(16) peptide and a 20 amino acid peptide of net charge -5 (GLFEALLELLESLWELLLEA). Particle formation is salt-dependent and yields perfectly spherical nanoparticles of approximately 120 to approximately 800 nm diameter, depending on buffer composition and temperature, with a stoichiometry of approximately 1:2.5 for the cationic and anionic peptides. The anionic peptide forms an alpha-helix in aqueous solution, has all five glutamates on one side of the helix, and exists entirely as a discrete oligomer of 9-10 peptides. A rigid oligomer with 45-50 negative charges almost certainly represents the core component of these nanoparticles, held together by electrostatic interactions with the unstructured K(16) peptide. Cells internalize these particles by an endocytic process, and free particles are frequently seen in the cytosol, presumably because of the acid-dependent fusogenic properties of the anionic peptide. Among other applications, these particles have potential for the targeted delivery of single or multiple therapeutic moieties directly to the cytosol, and we report the successful delivery of a K(16)-linked pro-apoptosis peptide.

Hydrodynamic gene delivery to the liver: theoretical and practical issues for clinical application.

Hydrodynamic gene delivery to the liver has potential as a safe and effective approach for clinical liver gene therapy. However, the simplicity of the technique in rodents - an intravenous injection - belies the theoretical and practical complexity for clinical application. A key issue is that outflow obstruction of the DNA solution from the liver is a critical factor for raising intrahepatic vascular pressure, which in turn provides the force to swell the liver and effect gene delivery. For conventional hydrodynamic gene delivery via tail vein injection, this outflow obstruction is provided naturally by the vascular resistance of the gut, spleen and pancreas. For regional hydrodynamic gene delivery to the liver, outflow obstruction to create a closed system requires surgical intervention, making it unlikely that minimally invasive techniques will be possible in the clinic. Intrinsic factors, in particular compliance (elasticity) of the liver are likely to be crucial in determining the degree of swelling for a given level of intrahepatic vascular pressure. Liver compliance is likely to be the major reason for the low level of hydrodynamic gene delivery in the pig model, and will influence the effectiveness of the approach in man, both in general and in different disease states.

Intestinal lactase as an autologous beta-galactosidase reporter gene for in vivo gene expression studies.

Intestinal lactase has potential as an autologous beta-galactosidase reporter gene for long-term gene expression studies in vivo, using chromogenic, luminescent, and fluorogenic substrates developed for Escherichia coli beta-galactosidase. In normal rat tissues, reactivity with a chromogenic fucopyranoside (X-Fuc, the preferred substrate of lactase) was present only at the lumenal surface of small intestine epithelial cells. Full-length lactase (domains I-IV), mature lactase (domains III and IV), and a cytosolic form of mature lactase (domains III and IV, without the signal sequence or transmembrane region) were evaluated. Transfection of HuH-7 cells in vitro, and hydrodynamic gene delivery to the liver in vivo, resulted in excellent gene expression. The full-length and mature (homodimeric, membrane-bound) forms reacted strongly with X-Fuc but not with the corresponding galactopyranoside (X-Gal). However, the presumptively monomeric cytosolic lactase unexpectedly reacted equally well with both substrates. The fluorogenic substrate fluorescein-di-beta-D-galactopyranoside was cleaved by cytosolic lactase, but not by full-length or mature lactase. Full-length lactase, when expressed ectopically in hepatocytes in vivo, localized exclusively to the bile canalicular membrane. Intestinal lactase is highly homologous in mice, rats, and humans and has considerable potential for evaluating long-term gene expression in experimental animals and the clinic.

Low-volume hydrodynamic gene delivery to the rat liver via an isolated segment of the inferior vena cava: efficiency, cardiovascular response and intrahepatic vascular dynamics.

BACKGROUND: Clinical application of hydrodynamic gene delivery to the liver requires the use of small volumes, an evaluation of the cardiovascular consequences of acute volume overload, and a better understanding of the intrahepatic vascular pressures driving gene delivery. Injection of DNA solution into the isolated segment of inferior vena cava (IVC) draining the hepatic veins is a potentially valuable low-volume approach. METHODS: Various volumes of DNA solution (pGL3 plasmid) were injected at 100 ml/min either systemically or into the isolated IVC segment in the DA rat. Arterial pressure, portal venous pressure, heart rate and electrocardiogram, in addition to reporter gene expression in the liver, were monitored. RESULTS: The 2% volume was > 10 000-fold more effective when delivered via the IVC segment than when given systemically, and as effective as 6% systemically. Isolation of the IVC segment caused profound falls in arterial pressure, with electrocardiogram signs of myocardial ischemia. On release of the IVC ties, without DNA infusion (no volume overload), arterial pressure recovered rapidly. However, with DNA infusion (volume overload) there was a brief recovery of arterial pressure, followed by complete heart block and fall in arterial pressure and pulse for several minutes. Portal venous pressure rose steeply to 30-33 mm Hg during the infusion. CONCLUSIONS: The IVC segment approach enables excellent gene delivery to the whole liver with small volumes, but causes severe cardiovascular disturbances in the rat. Portal venous pressures are slightly higher than in the mouse, and suggest functional outflow obstruction by the capillary bed of the intestines.

(LYS)(16)-based reducible polycations provide stable polyplexes with anionic fusogenic peptides and efficient gene delivery to post mitotic cells.

Extracellular stability, endocytic escape, intracellular DNA release and nuclear translocation of DNA are all critical properties of non-viral vector/DNA particles. We have evaluated a (Lys)(16)-based linear, reducible polycation (RPC) in combination with an acid-dependent, anionic fusogenic peptide for gene delivery to dividing and post-mitotic cells. The RPC was formed from Cys(Lys)(16)Cys monomers. Molecular weight was 24,000 Da, corresponding to an average of 10.5 peptide monomers per RPC. Non-reducible polylysine (PLL) (27,000 Da) and monomeric (Lys)(16) peptide were evaluated for comparison. (Lys)(16)/DNA particles were disrupted at fusogenic peptide concentrations well below those used for gene delivery. By contrast, RPC/DNA an PLL/DNA particles were stable in the presence of high concentrations of the anionic peptide. Addition of 10% serum virtually abolished the transfection ability of (Lys)(16)/DNA/fusogenic peptide particles, but had little effect on RPC/DNA/fusogenic peptide particles. RPC/DNA/fusogenic peptide particles were highly effective for gene delivery to both cell lines and post-mitotic corneal endothelium. PLL/DNA/fusogenic peptide particles were moderately effective on cell lines, but gave no gene delivery with corneal endothelial cells. We conclude that (Lys)(16)-based RPC/DNA/fusogenic peptide particles provide a gene delivery system which is potentially stable in the extracellular environment and, on reductive depolymerisation, can release DNA plasmids for nuclear translocation.

Preliminary evidence that the allogeneic response might trigger antitumour immunity in patients with advanced prostate cancer.

OBJECTIVE: To explore the possibility that allogeneic responses might, by chance, encompass cross-reactive T cell clones specific for neo-antigenic tumour determinants, and thereby activate antitumour immunity; such cross-reactions are well documented for antiviral immunity, and genetic instability in developing cancers generates many neo-antigenic determinants as potential targets for immune responses, but the biology inevitably favours tumour progression. PATIENTS AND METHODS: Fourteen patients with hormone-refractory prostate cancer received full-thickness skin allografts from different, unrelated donors (fellow patients) until each had received six grafts. Serum prostate-specific antigen (PSA) level was used as a surrogate for tumour mass. RESULTS: One patient had a remarkable decline in PSA level, with levels at 1 year lower than before grafting. A second patient had stable PSA levels for almost 2 years. A third patient had stable PSA levels for 10-12 months before they resumed an exponential rise. Of four patients with PSA levels of > 10 ng/mL, three required surgery or radiotherapy for obstructive symptoms during or shortly after grafting. CONCLUSION: Transplant rejection involves mechanistically atypical T cell recognition of allogeneic major histocompatibility complex antigens, with massive polyclonal T cell activation. This unique aspect of T cell biology might represent a novel approach for initiating cross-reactive antitumour responses.

Synthetic peptides as non-viral DNA vectors.

The use of multiple peptide motifs to provide effective gene delivery holds great promise as an elegant, non-immunogenic approach to gene therapy. The molecular understanding of cell and viral biology provides a strong foundation on which to pursue this objective. Synthetic peptides containing multiple lysines and/or arginines (occasionally ornithines) provide natural polycations for multivalent electrostatic binding of DNA, and for DNA compaction into particles suitable for gene delivery. These cationic peptides can incorporate additional functional motifs (e.g. for translocating DNA into the nucleus) and they can be linked by disulphide bonds to produce high molecular reducible polycations with superior properties for gene therapy. Many factors influence the size, surface charge and stability of peptide/DNA particles. For in vivo use, uncharged particles resistant to disruption by salt and protein, and targeted to tissue-specific membrane molecules, will be required. Entry into the cell is via one of the endocytic pathways, depending on particle size and (in principle) the target cell surface molecule. Peptide motifs for endocytic escape are based mainly on the anionic fusogenic peptide of influenza virus haemagglutinin and on histidine-rich peptides (where the buffering properties of the imidazole group cause osmotic swelling and probably rupture of endocytic vesicles). Once in the cytosol, translocation of DNA plasmids across the nuclear pore complex into the nucleus is a crucial step, because most target cells for gene therapy are either non-dividing or slowly dividing. Nuclear translocation can be achieved by classical nuclear localising motifs, or more simply by (Lys)16 and other cationic peptides.

In vivo suppression of major histocompatibility complex class II expression on porcine vascular endothelial cells by an HMG-CoA reductase inhibitor.

BACKGROUND: Vascular endothelial cells of man and pig, but not rodents, strongly express major histocompatibility complex (MHC) class II antigens in vivo, probably via the inducible promoter IV of the class II transactivator. There is abundant in vitro evidence that MHC class II positive vascular endothelial cells can activate T cells. Peripheral antigen presentation by endothelial cells is potentially important for organ-specific immunity, for allograft rejection, and possibly for immune responsiveness in general. Given the reported effects of statins on promoter IV of the class II transactivator, we evaluated in vivo expression of MHC class II antigens in pigs treated with atorvastatin calcium. METHODS: Pigs were given 3 mg/kg/day of atorvastatin orally daily for 16 days, and then killed 24 hr after the last dose. Heart, kidney, and liver were removed for immunohistological and quantitative absorption analysis. RESULTS: Double-labeling studies using immunofluorescence on frozen section for Factor VIII and MHC class II showed a marked suppression of MHC class II on vascular endothelial cells in all 4 treated pigs, in comparison with untreated pigs. This was confirmed using immunoperoxidase techniques on frozen sections. Quantitative absorption analysis showed up to 25-fold reduction in MHC class II expression. CONCLUSIONS: Statins substantially suppress endothelial cell MHC class II expression in vivo. This is likely to inhibit organ-specific immune responses, and possibly also general immune responsiveness. In a transplantation setting, in addition to other regulatory effects on the recipients immune system, statins might reduce the long-term capacity of the donor organ to activate rejection mechanisms.

The effect of anti-lymphocyte serum on subpopulations of blood and tissue leucocytes: possible supplementary mechanisms for suppression of rejection and the development of opportunistic infections.

Xenogeneic anti-lymphocyte serum (ALS) remains a major reagent for immunosuppression in clinical practice, but mechanisms of action and risks of opportunistic infection have not been considered in the context of innate immunity and its role in immune responsiveness. Rabbit anti rat ALS was administered intraperitoneally. Blood was taken for flow cytometry to establish absolute counts of leucocyte subsets. Tissues were harvested for immunohistology to evaluate interstitial dendritic cells and tissue macrophages. At day 2 of ALS therapy, T cells are completely depleted, as anticipated. B cells are undiminished and form approximately 90% of blood leucocytes. Monocytes and natural killer (NK) cells are substantially (approximately 80%), but not completely, depleted, and there is a trend for diminished numbers of putative dendritic cells. Neither interstitial dendritic cells nor tissue macrophages in heart are affected. The results at day 7 were very similar to day 2. Substantial depletion of blood monocytes and NK cells might attenuate the innate immune system, and represent a possible supplementary mechanism (in addition to T cell depletion) for suppression of rejection. It might be of particular importance in reducing defences against infections. Monitoring these parameters could be of clinical value.

Exploration of peptide motifs for potent non-viral gene delivery highly selective for dividing cells.

BACKGROUND: The immunogenicity of viral DNA vectors is an important problem for gene therapy. The use of peptide motifs for gene delivery would largely overcome this problem, and provide a simple, safe and powerful approach for non-viral gene therapy. METHODS: We explored the functional properties of two motifs: the (Lys)(16) motif (for binding and condensing DNA, and probably also nuclear translocation of plasmids) and the fusogenic peptide motif of influenza virus (for acid-dependent endocytic escape of peptide/DNA particles). The physical properties and gene delivery efficiencies of (Lys)(16)-containing peptides in combination with free fusogenic peptide were evaluated, and compared with a single composite peptide incorporating both moieties. Post-mitotic corneal endothelial cells and growth-arrested HeLa were included, so as not to neglect the question of nuclear translocation of plasmids. RESULTS: The fusogenic moiety in the composite peptide was able to adopt an alpha-helical configuration unhindered by the (Lys)(16) moiety, and retained acid-dependent fusogenic properties. The composite peptide gave remarkably high levels of gene delivery to dividing cell lines. However, in marked contrast to (Lys)(16)/DNA complexes plus free fusogenic peptide, the composite peptide was completely ineffective for gene delivery to post-mitotic and growth-arrested cells. CONCLUSIONS: Attachment of the fusogenic peptide to (Lys)(16) appears to block (Lys)(16)-mediated nuclear translocation of plasmid, but not fusogenic peptide mediated endocytic escape. This strengthens the experimental basis for (Lys)(16)-mediated nuclear translocation of plasmids, and provides a single peptide with potent gene delivery properties, restricted to dividing cells. This property is potentially useful in experimental biology and clinical medicine.

Regional hydrodynamic gene delivery to the rat liver with physiological volumes of DNA solution.

BACKGROUND: The major barrier to the clinical application of hydrodynamic gene delivery to the liver is the large volume of fluid required using standard protocols. Regional hydrodynamic gene delivery via branches of the portal vein has not previously been reported, and we have evaluated this approach in a rat model. METHODS: The pGL3 plasmid with the luciferase reporter gene was used at 50 micro g/ml in isotonic solutions, and was administered with a syringe pump for precise control of the hydrodynamic conditions evaluated. Gene expression was individually measured in six anatomically distinct liver lobes. The effect of systemic chloroquine to promote endocytic escape and a (Lys)(16)-containing peptide to condense the DNA into approximately 100-nm nanoparticles was also evaluated. RESULTS: Hydrodynamic conditions for excellent gene delivery were obtained by using 3-ml volumes ( approximately 12 ml/kg) of isotonic DNA solution delivered at 24 ml/min to the right lateral lobe ( approximately 20% of the liver mass). Under these conditions, >95% of gene delivery usually occurred in the targeted right lateral lobe. Outflow obstruction was essential for gene delivery, both at optimal and at very low levels of hydrodynamic gene delivery. The use of systemic chloroquine to promote endocytic escape did not augment hydrodynamic gene delivery, while condensation of DNA in non-ionic isotonic solutions (5% dextrose) to nanoparticles of approximately 100 nm completely abolished gene delivery. CONCLUSIONS: Regional hydrodynamic gene delivery via a branch of the portal vein offers a physiological model of liver gene therapy, for experimental and clinical application.

Imaging in solution of (Lys)(16)-containing bifunctional synthetic peptide/DNA nanoparticles for gene delivery.

The physical properties of non-viral vector/DNA nanoparticles in physiological aqueous solution are poorly understood. A Fluid Particle Image Analyser (FPIA), normally used for analysis of industrial and environmental fluids, was used to visualise individual (Lys)(16)-containing peptide/DNA particles. Eight (Lys)(16)-containing synthetic peptides were used to generate peptide/DNA particles at a constant + to - charge ratio of 2.8:1 with 10 microg/ml of plasmid DNA in phosphate buffered saline. Dynamic Light Scattering (DLS) and gene delivery studies were also performed. We present the first images of non-viral vector/DNA nanoparticles in physiological aqueous solution, together with precise measurements of individual particle size and shape in solution and, for the first time, an accurate measure of particle number. Particle size and shape, particle number, and efficiency for gene delivery varied markedly with different peptides. Under standard conditions for in vitro gene delivery, we estimate approximately 60 peptide/DNA nanoparticles per target cell, each containing approximately 70,000 plasmids. This novel capacity to image individual vector/DNA nanoparticles in solution and to count them accurately will enable a more precise assessment of non-viral gene delivery systems, and a more quantitative interpretation of gene delivery experiments.

A synthetic peptide vector system for optimal gene delivery to corneal endothelium.

BACKGROUND: Efficient and non-toxic gene delivery, preferably with non-viral DNA vectors readily transferable to clinical practice, is generally regarded as a major limitation for gene therapy. METHODS: A 31 amino acid, integrin-targeted bifunctional synthetic peptide (polylysine-molossin), and two (Lys)(16)-containing control peptides, were assessed for ex vivo gene delivery to the rabbit cornea. Critical physical properties of polylysine-molossin/DNA complexes were evaluated and both chloroquine and a 20 amino acid fusogenic peptide were used to promote endocytic exit. RESULTS: Polylysine-molossin/DNA complexes and (Lys)(16)/DNA complexes at 10 microg/ml of DNA were much smaller and much more positively charged in non-ionic isotonic medium (5% dextrose or 5% dextrose buffered to pH 7.4 in 10 mM Tris) when compared with culture medium or phosphate-buffered saline (PBS). Addition of the fusogenic peptide (net charge -5) reversed the positive charge of complexes in PBS, and reduced the strong positive charge of polylysine-molossin/DNA complexes in dextrose. Polylysine-molossin/DNA complexes in 5% dextrose were much more effective for gene delivery to the cornea when compared with complexes in culture medium, and the fusogenic peptide was much more effective than chloroquine for promoting gene delivery. The optimal DNA/polylysine-molossin/fusogenic peptide w/w ratio was 1 : 3 : 2 at 10 microg/ml of DNA. Essentially 100% of corneal endothelial cells were transfected under these optimal conditions, without any evidence of toxicity. Integrin-targeting did not contribute significantly to gene delivery in this system. CONCLUSIONS: This DNA vector system, consisting entirely of synthetic peptides, is ideally suited for clinical applications of gene therapy of the corneal endothelium.

Modulation of adenovirus infection in vitro by antisense oligodeoxynucleotides.

OBJECTIVE: Antisense oligodeoxynucleotides (ODNs) may represent a novel, airway directed approach to the treatment of adenovirus infection of the lung, for which no specific therapy exists. This study assessed the efficacy of antisense ODNs in modulating adenovirus infection in vitro. METHODOLOGY: A biological assay, which quantified viral plaque formation by wild type adenovirus 5 in a lung epithelial cell line (A549), was used to evaluate the inhibitory effect of a number of antisense ODNs targeted to the early (E) 1 A and protein IX genes of adenovirus 5. Antisense ODNs (20-21mers, phosphorothioate end-protected) were designed to straddle the initiation of translation (AUG) codon of the mRNA of the targeted gene. RESULTS: There was a consistent and significant (P < 0.005) reduction in viral plaque formation in those cells treated with an E1A antisense ODN, compared with the nonsense control ODN. Neither the addition of a cationic lipid (Lipofectamine), nor increasing the concentration of ODN from 1 micro mol to 15 micro mol enhanced the original inhibitory effect observed with the E1A antisense ODN. CONCLUSIONS: An antisense ODN targeted to the E1A gene can specifically inhibit adenovirus 5 infection in vitro, suggesting a potential therapeutic role for antisense ODNs in adenovirus infection of the lung.

Efficient gene delivery to primary neuron cultures using a synthetic peptide vector system.

A bi-functional, 31 amino acid synthetic peptide (polylysine-molossin) was evaluated for gene delivery to primary cultures of rat cerebral cortex neurons. Polylysine-molossin consists of an amino terminal domain of 16 lysines for electrostatic binding of DNA, and a 15 amino acid, integrin-binding domain at the carboxyl terminal. High levels of gene delivery were obtained with 20-30 microM chloroquine, with a synthetic fusogenic peptide at an optimal DNA:polylysine-molossin:fusogenic peptide w/w ratio of 1:3:0.2, and with the addition of low concentrations of Lipofectamine 2000 at an optimal DNA:polylysine-molossin:Lipofectamine 2000 w/w ratio of 1:3:0.5. With the best combination, >30% of neurons strongly expressed the beta-galactosidase reporter gene, with no observable toxicity. DNA concentrations >2 microgram/ml were essential for efficient gene delivery. This synthetic peptide provides a safe, readily standardised and flexible DNA vector system well suited to ex vivo gene delivery to neurons for experimental and clinical applications.

The in vivo use of chloroquine to promote non-viral gene delivery to the liver via the portal vein and bile duct.

BACKGROUND: Assistance with exit from endocytic vesicles is a key factor for non-viral gene delivery, and is a particular challenge in vivo. We have evaluated the in vivo use of chloroquine administered systemically, orally and/or locally for gene delivery to the liver. METHODS: The DNA vector (polylysine-molossin) is a 31 amino acid bifunctional synthetic peptide, incorporating an amino terminal chain of 16 lysines for electrostatic binding of DNA. Gene delivery was to the right lateral lobes of the liver by branches of the bile duct or portal vein. RESULTS: Single intraperitoneal injections of 8, 25 and 75 mg/kg of chloroquine (the maximum tolerated single intraperitoneal dose) resulted in increasing levels of luciferase reporter gene expression, following gene delivery via the bile duct. 100 mg/kg of chloroquine orally was equivalent to 25 mg intraperitoneally. A 3-day course of intraperitoneal and oral chloroquine gave approximately 10-30-fold higher gene expression than an optimal single dose, and resulted in a scattering of positive hepatocytes in the lobule. Gene delivery via the bile duct was much more effective than via the portal vein. Serum chloroquine levels at the time of gene delivery showed a highly significant correlation with gene expression, but the maximum achievable levels in vivo ( approximately 1-2 micro M) were much lower than those required for optimal in vitro gene delivery. Chloroquine (0.2-5 mM) was also given locally in the bile duct with vector/DNA complexes. Maximum gene expression was obtained with 0.5 mM local chloroquine, but the level of gene expression was only equivalent to the 25 mg intraperitoneal dose. CONCLUSIONS: The in vivo use of chloroquine is effective for promoting gene delivery to the liver, but requires multiple dosing and is limited by systemic toxicity.