Review Gene Therapy

Intravenous, non-viral RNAi gene therapy of brain cancer

non-viral genes 3. Intravenous RNAi and brain cancer: luciferase target 4. Intravenous RNAi and brain cancer: EGFR target 5. Human applications of the PIL gene-targeting technology

RNA interference (RNAi) has the potential to knock down oncogenes in cancer, including brain cancer. However, the therapeutic potential of RNAi will not be realised until the rate-limiting step of delivery is solved. The development of RNA-based therapeutics is not practical, due to the instability of RNA in vivo. However, plasmid DNA can be engineered to express short hairpin RNA (shRNA), similar to endogenous microRNAs. Intravenous, non-viral RNAibased gene therapy is enabled with a new gene-targeting technology, which encapsulates the plasmid DNA inside receptor-specific pegylated immunoliposomes (PILs). The feasibility of this RNAi approach was evaluated by showing it was possible to achieve a 90% knockdown of brain tumour-specific gene expression with a single intravenous injection in adult rats or mice with intracranial brain cancer. The survival of mice with intracranial human brain cancer was extended by nearly 90% with weekly intravenous injections of PILs carrying plasmid DNA expressing a shRNA directed against the human epidermal growth factor receptor. RNAi-based gene therapy can be coupled with gene therapy that replaces mutated tumour suppressor genes to build a polygenic approach to the gene therapy of cancer.

hle

yP

ub

6. Expert opinion and conclusions

ati on sL

intravenous delivery of

Department of Medicine, UCLA Warren Hall 13-164, 900 Veteran Avenue, Los Angeles, CA 90024, USA

lic

2. Pegylated immunoliposomes for

td

William M Pardridge 1. Introduction

Keywords: blood–brain barrier, epidermal growth factor receptor, insulin receptor, liposomes, transferrin receptor

As

Expert Opin. Biol. Ther. (2004) 4(7):1103-1113 1.

Introduction

Co

py

rig

ht

of

RNA interference (RNAi) enables the post-transcriptional silencing of pathological gene expression in cancer, including brain cancer, as well as viral infections and other chronic disease. RNAi is readily induced in cultured cells with the delivery of synthetic RNA duplexes to the cells in culture by cationic lipoplexes. However, cationic lipoplexes have low efficacy in vivo, owing to aggregation in saline [1-3] and sequestration within the lung microcirculation [4]. In addition, RNA is rapidly degraded in vivo and it may not be practical to develop RNA-based RNAi therapeutics for in vivo applications. Synthetic RNA duplexes are inactive as RNAi agents in the brain, despite 3 days of continuous intracerebral infusion via a transcranial route of administration [5]. An alternative to RNA-based RNAi therapeutics is RNAi-based gene therapy, wherein an expression plasmid produces a short hairpin RNA (shRNA). The formation of a shRNA parallels endogenous mammalian RNAi mechanisms mediated via microRNAs (miRNAs) [6,7]. The rate-limiting step in RNAi-based gene therapy, or gene therapy in general, is delivery. The delivery issue for cancer is most severe in the brain because of the restrictive permeability properties of the brain capillary endothelium, which forms the blood–brain barrier (BBB) in vivo. However, with the exception of organs, such as liver or spleen, that have fenestrated microcirculatory barriers, the vascular barrier is an impediment to gene therapy in virtually all organs, including the brain. Existing gene therapy delivery systems use either viral vectors or cationic liposomes.

Ashley Publications www.ashley-pub.com

2004 © Ashley Publications Ltd ISSN 1471-2598

1103

Intravenous, non-viral RNAi gene therapy of brain cancer

yP

ub

lic

ati on sL

td

insulin receptor normally shuttles its endogenous ligand, insulin, to the nuclear compartment [19]. However, the use of insulin, per se, as a targeting ligand is not preferred because this could cause hypoglycaemia. Alternatively, a peptidomimetic monoclonal antibody (mAb) to the human insulin receptor (HIR) is also a targeting agent for the delivery of therapeutics to the brain following an intravenous injection [20]. The therapeutic plasmid DNA and the targeting mAb must be joined together in a formulation that is stable in the circulation. Cationic liposomes, which are formed by mixing cationic lipids or cationic polymers with anionic DNA, are stable in water, but are not stable in physiological solutions, as the saline causes aggregation of the cationic DNA polyplex [1-3]. An alternative approach is to encapsulate the plasmid DNA in the interior of liposomes. It is possible to fit a single plasmid DNA molecule into the interior of an 85 nm liposome. The upper limit in size of the plasmid DNA that can be incorporated inside a liposome of this size is not known, but plasmid DNA of 6 – 11 kb have been incorporated with equal efficiency. A liposome of 85 – 100 nm approximates the size of many viruses and is small enough to undergo receptor-mediated transcytosis across the BBB [21]. However, liposomes are rapidly removed from the bloodstream in vivo by cells lining the reticulo-endothelial system (RES). The surface of the liposome is adsorbed by serum proteins and this triggers clearance of the liposome by the RES. The RES uptake can be reduced by conjugating several thousand strands of 2000 Da polyethylene glycol (PEG) to the surface of the liposome and this pegylation process results in the formation of a pegylated liposome [22]. However, pegylated liposomes, per se, are biologically inert in vivo and are transported poorly across cell membranes [23]. The desired targeting for treatment of brain cancer can be achieved by conjugating the receptor-specific mAb to the tips of 1 – 2% of the PEG strands, which results in the formation of a pegylated immunoliposome (PIL). Because the receptor specificity required to initiate transfer of the PIL across the first or second barrier may be different, it is possible to conjugate multiple mAbs to the PIL so that this nanocontainer can traverse the intended biological membrane barriers that separate the blood from the nuclear compartment of the cancer cell [15]. This article will review RNAi-based gene targeting in two models of experimental brain cancer. In one model, C6 rat glioma cells are implanted in the brain of adult rats [24]. This brain cancer is perfused by vessels of rat brain origin. Therefore, both the first barrier and the second barrier express rat receptors. As the transferrin receptor (TfR) is expressed on both the rat BBB and the rat tumour cell membrane, the murine OX26 mAb to the rat TfR is used as the targeting mAb for both the first and second barriers in this model [24]. In a second model of experimental brain cancer, human U87 glioma cells are implanted in the brain of adult severe combined immunodeficient (SCID) mice [25]. This human brain cancer is perfused by vessels of mouse brain origin and these vessels express the murine TfR [15]. The OX26 mAb to the rat

As

hle

Disregarding the concerns related to immunogenicity [8] and insertional mutagenesis caused by chromosomal integration with viruses such as retrovirus [9] or adeno-associated virus (AAV) [10], viral vectors are not effective in brain cancer after intravenous administration because they do not cross the BBB. Similarly, cationic liposomes, which aggregate in the saline environment in vivo [1-3] and embolise in the pulmonary circulation [4], do not cross the BBB or enter the brain from the blood [11]. Owing to these terminal delivery problems with either cationic liposomes or viral vectors, attempts to treat brain cancer with gene therapy have used craniotomy and intracerebral injection of the gene therapeutic. However, due to limitations of diffusion [12], the effective treatment volume with a transcranial delivery method is localised to the local region of the brain at the tip of the injection device. It is not possible to deliver the therapeutic gene to the majority of cancer cells in the brain with a transcranial delivery approach. The delivery of gene therapeutics to virtually all cells in brain is possible with transvascular delivery [13]. The distance between capillaries in the brain is only ∼ 50 µm. Therefore, in the early stage of brain cancer, cancer cells are in close proximity to a tumour blood vessel, which is co-opted from neighbouring brain cells [14]. With a transvascular delivery approach, it is possible to target virtually all cancer cells in a brain tumour. Early treatment is essential because eventually a cancer grows so fast that it outstrips its vascular supply and large avascular zones of the cancer develop [15]. However, a combination of neurosurgery to remove the majority of the tumour and intravenous, transvascular gene therapy may lead to a cure for brain cancer. However, the achievement of using gene therapy to treat brain cancer may not be realised, until the rate-limiting delivery problem is solved.

Pegylated immunoliposomes for intravenous delivery of non-viral genes

of

2.

Co

py

rig

ht

A gene therapeutic can enter the brain from the blood following transvascular targeting if gene-targeting technology is used. The goal of gene-targeting technology is the reformulation of the gene therapeutic so that the gene is enabled to traverse the multiple biological membrane barriers that are encountered in vivo and this is possible by accessing membrane receptor/transport systems [16]. In treatment of brain cancer, the gene therapeutic must be enabled to first cross the BBB or brain capillary endothelial membrane (the first barrier) and then cross the tumour cell plasma membrane (the second barrier). In addition, the gene therapeutic must be able to target the nuclear compartment and traverse the nuclear membrane (the third barrier). The traversal of all three barrier systems is possible by accessing endogenous receptor/transport systems expressed within these three membrane systems. The insulin receptor is expressed on both the BBB of normal brain [17] and on the BBB of human brain cancer [18]. The insulin receptor is also expressed on the plasma membrane of virtually all brain cancer cells [18] and the 1104

Expert Opin. Biol. Ther. (2004) 4(7)

Pardridge

A

B

C

mAb R

td

mAb

Promoter

mAb

D

E

ati on sL

mAb

20 liver

16 14 12

brain

6 4

0.6 0.5 0.4 0.3 brain

0.2

spleen

lung

yP

0

0.7

lung

lic

10 8

2

0.8

spleen liver

ub

Protein (pg/mg)

18

Protein (pg/mg)

DNA

0.1 kidney heart

heart kidney

0

Rhesus monkey (HIRmAb) Rat (TfRmAb)

ht

of

As

hle

Figure 1. (A) Diagram of a super-coiled expression plasmid DNA encapsulated in an 85 nm PIL targeted to a cell membrane receptor (R) with a receptor-specific, endocytosing mAb. Tissue-specific expression of the plasmid can be controlled by the promoter inserted 5’ of the gene. (B) Transmission electron microscopy of a PIL. The mAb molecule tethered to the tips of the 2000 Da PEG is bound by a conjugate of 10 nm gold and a secondary antibody. The position of the gold particles shows the relationship of the PEG-extended mAb and the liposome. Magnification bar = 20 nm. From [30]. (C) Confocal microscopy of U87 human glioma cells following either a 3-h (left panel) or a 24-h (right panel) incubation of fluorescein-conjugated clone 882 DNA (fluoroDNA) encapsulated within HIRmAb-PILs. The inverted greyscale image is shown. There is primarily cytoplasmic accumulation of the fluoro-DNA at 3 h, whereas the fluoro-DNA is largely confined to the nuclear compartment at 24 h. Fluoro-DNA entrapped within intranuclear vesicles is visible at both 3 and 24 h. From [27]. (D) β-Galactosidase histochemistry of brain removed 48 h after the intravenous injection of a β-galactosidase expression plasmid encapsulated in HIRmAb-PILs in the adult rhesus monkey. From [32]. (E) Luciferase gene expression in the brain and other organs of the adult rhesus monkey (left panel) and adult rat (right panel) measured at 48 h after a single intravenous injection of the PIL carrying the plasmid DNA. Data are mean ± SE. The plasmid DNA encoding the luciferase gene used in either species is clone 790, which is driven by the SV40 promoter [36]. The PIL carrying the DNA was targeted to primate organs with an HIRmAb and to rat organs with a TfRmAb [32].

rig

HIR: Human insulin receptor; mAb: Monoclonal antibody; PEG: Poly(ethylene glycol); PIL: Pegylated immunoliposome; SE: Standard error; TfR: Transferrin receptor.

Co

py

TfR is species-specific and does not react with the mouse TfR [26]. The rat 8D3 to the mouse TfR can be used as a targeting ligand for mouse brain. However, the 8D3 is speciesspecific and would not recognise the TfR on the human tumour cell in brain. Therefore, the murine 83-14 mAb to the HIR is used to trigger transport of the PIL across the second barrier, which is the human tumour cell membrane. The HIRmAb is also a useful targeting ligand for delivering plasmid DNA to the nuclear compartment, as the DNA packaged inside HIRmAb-targeted PILs rapidly enters the nuclear compartment of brain cancer cells [27]. In summary, for the C6 rat glioma/rat brain tumour model, PILs are targeted with a single mAb to the rat TfR [24]. For the U87 human glioma/SCID mouse brain tumour model, PILs are targeted with an mAb to the mouse TfR and a second mAb to the HIR [15,25]. Prior to

discussion of the results with these two models of experimental brain cancer, the general applications of the PIL gene-targeting technology are reviewed. A diagram depicting the supercoiled plasmid DNA in the interior of a PIL is shown in Figure 1A. A receptor (R)-specific mAb is conjugated to the tips of a small fraction of the 2000 Da PEG polymers projecting from the surface of the liposome [28]. The liposome is ∼ 85 – 100 nm in diameter, which is large enough to accommodate a single supercoiled plasmid DNA in the interior [29]. The relationship of the mAb to the liposome sphere is demonstrated in Figure 1B with electron microscopy. In this experiment, the PIL was bound to a secondary antibody that was conjugated with 10 nm gold [30]. The relationship of the spherical liposome and the extended mAb molecules is shown by the position of the 10 nm gold

Expert Opin. Biol. Ther. (2004) 4(7)

1105

Intravenous, non-viral RNAi gene therapy of brain cancer

yP

ub

lic

ati on sL

td

continuous endothelial barrier and the insulin receptor or TfR is not expressed on the vascular barrier in these organs [33]. The lung is perfused by capillaries with continuous endothelial barriers and the intermediate level of luciferase gene expression in the lung suggests that the endothelium may express insulin receptor or TfR. The results with luciferase gene expression (Figure 1E) illustrate two properties of gene delivery with the PIL system. First, ‘ectopic’ gene expression is observed in non-brain organs with this targeting technology, owing to the expression of the insulin receptor or TfR in peripheral tissues. Depending on the disease, expression of a therapeutic gene in peripheral tissues may be desired. However, if the intent is to express the exogenous gene only in the brain, then this organ-specific gene expression can be achieved by replacing the widely read SV40 promoter with a brain specific promoter. Studies in mice show that if a β-galactosidase plasmid is under the influence of a brain specific promoter taken from the 5´ flanking sequence of the glial fibrillary acidic protein (GFAP) gene, then gene expression in peripheral organs is eliminated [34]. Similarly, studies in rhesus monkeys have shown that if the SV40 promoter is replaced with an ocular specific promoter taken from the 5´ flanking sequence of the opsin gene, then gene expression is confined to the eye [35]. Secondly, the level of luciferase gene expression in the rhesus monkey organs is 50-fold greater than luciferase gene expression in the same organs in the rat [32]. This is not to say that the TfRmAb is not a useful targeting ligand for gene delivery. Indeed, targeting the TfR in rats with experimental Parkinson’s disease results in a 100% normalisation of striatal tyrosine hydroxylase (TH) activity, with TH gene therapy using the PIL gene delivery technology [30]. The results in Figure 1E, showing a 50-fold higher level of luciferase gene expression in the primate brain, underscore the high potency of the HIRmAb as a gene delivery system. This high potency may derive from the fact that the insulin receptor, unlike the TfR, normally serves to deliver the endogenous ligand, insulin, to the nuclear compartment of the cell [27,36].

Co

py

rig

ht

of

As

hle

particles, which are approximately the same size as a mAb molecule. The ability of the HIRmAb to target plasmid DNA to the nuclear compartment was demonstrated with confocal microscopy of human U87 glioma cells as shown in Figure 1C. In this experiment, the plasmid DNA was radiolabelled with fluorescein by nick translation prior to encapsulation in the PIL [27]. The PIL was targeted to the human U87 glioma cells with the HIRmAb and the cells were fixed at either 3 h (left panel, Figure 1C) or 24 h (right panel, Figure 1C). At 3 h, the plasmid DNA was primarily in the cytoplasmic compartment of the cell, although a significant amount of the plasmid DNA is already found in intranuclear vesicular structures. However, at 24 h of incubation, the majority of the intracellular plasmid DNA is confined to the nuclear compartment. This result contrasts with cationic liposomes in cell culture, wherein the majority of the DNA is restricted to the lysosomal compartment of the cytoplasm, with minimal entry of the plasmid DNA into the nucleus [31]. The ability of PILs to deliver an endogenous gene to all parts of the adult Rhesus monkey brain following an intravenous administration is shown in Figure 1D. This figure is a β-galactosidase histochemical assay performed on frozen sections of brain of the adult rhesus monkey, which was sacrificed 48 h after the intravenous injection of a PIL targeted with the HIRmAb. This PIL encapsulates a β-galactosidase expression plasmid under the influence of the simian virus 40 (SV40) promoter [32]. The study shows global expression of the transgene in all parts of the brain of the rhesus monkey. This experimental result is only possible with a transvascular delivery technology. The finding of global expression of a transgene in the monkey brain with the PIL delivery system replicates prior results with β-galactosidase expression plasmids performed in either rats or mice [33,34]. In addition, studies have been performed with an expression plasmid producing luciferase, and a comparison of luciferase gene expression in the brain of adult rhesus monkeys and rats is shown in Figure 1E. The same luciferase expression plasmid, under the influence of the SV40 promoter, was encapsulated in PILs that were targeted with either the HIRmAb, for injection into rhesus monkeys, or the TfRmAb, for injection into adult rats [32]. The animals were sacrificed 48 h after a single intravenous injection and organs were removed for measurement of luciferase enzyme activity with a luminometer. The data show high luciferase expression in brain because the targeted receptor (insulin receptor or TfR) is expressed on both the BBB and the neuronal cell membrane. There is also high expression in the liver or spleen because parenchymal cells of these two organs also express the insulin receptor or TfR [33]. Although the insulin receptor or TfR is not expressed on the microvascular endothelium in liver or spleen, there is no need for a receptor-mediated targeting mechanism at the vascular barrier in liver or spleen. These organs have sinusoidal microcirculatory beds that are highly porous to structures the size of an 85 nm liposome. There is no gene expression in kidney or heart because these organs are perfused by capillaries with a 1106

Intravenous RNAi and brain cancer: luciferase target 3.

The feasibility of knocking down gene expression in intracranial brain cancer was initially evaluated with the luciferase gene as the target [24]. The C6 rat glioma cells were permanently transfected with the luciferase gene using the 790 luciferase expression plasmid, which is derived from the pCEP4 expression plasmid under the influence of an SV40 promoter. The C6 cells transfected with the luciferase gene are designated C6-790 cells. The C6-790 cells were implanted in the caudate putamen nucleus of adult rats under stereotaxic guidance, and large intracranial brain tumours formed within 2 weeks of implantation (Figure 2A). These tumours expressed ∼ 200 – 300 pg lucifirease/mg of tumour cell protein. An antiluciferase expression plasmid, under the influence of U6

Expert Opin. Biol. Ther. (2004) 4(7)

Pardridge

A

and injected intravenously into C6-790 tumour-bearing rats at either 5 or 10 days after the intracerebral implantation of the tumour cells. In Study A, where the PILs were administered at 10 days after tumour implantation, the animals were sacrificed at 12 or 14 days after tumour implantation, which corresponds to 2 or 4 days after a single intravenous administration of the PILs carrying clone 952. In Study B, the tumour bearing animals were injected with PILs at 5 days after tumour implantation and were sacrificed at 7 or 10 days after tumour implantation, which corresponds to 2 or 5 days after the single intravenous administration of PILs carrying either clone 952 or 959. The tumour extracts were measured for enzyme activity of luciferase and γ-glutamyl transpeptidase (GTP). GTP is a control enzyme that is expressed in brain as well as cancer cells. The results of the luciferase and GTP enzyme measurements are given in Table 1. In Study A, luciferase gene expression was inhibited by 68 and 64% at 2 and 4 days after intravenous administration of the RNAi plasmid (clone 952) encapsulated in the TfRmAb-PIL (Table 1). In contrast to luciferase, RNAi gene therapy caused no change in tumour GTP gene expression (Table 1). There also was no change in GTP enzyme activity in contralateral brain. In Study A, RNAi treatment was not initiated until 10 days after brain implantation of 500,000 tumour cells. However, by 10 days after implantation of such a large number of tumour cells, the tumour is large, occupies nearly the entire volume of the striatum, and many parts of the tumour are avascular and cannot be targeted from the blood compartment. Therefore, the authors examined whether higher degrees of knockdown of luciferase gene expression in C6-790 brain cancer in vivo could be achieved when intravenous RNAi treatment is initiated earlier in the course of brain tumour development. In a modified treatment regimen, Study B, intravenous RNAi gene therapy was administered at 5 days after implantation of 500,000 C6-790 cells. The tumour luciferase activity in animals treated with the empty 959 plasmid DNA was not significantly different from the tumour luciferase activity in animals treated with saline (Table 1, Study A). However, the tumour luciferase was inhibited 90 and 87% at 2 and 5 days after treatment with clone 952 plasmid DNA encapsulated in the TfRmAb-PIL (Table 1, Study B). This level of gene inhibition in vivo is comparable to the 90% knockdown of luciferase gene expression observed in the C6 cells in tissue culture following co-transfection both with the clone 790 luciferase expression plasmid and the clone 952 antiluciferase shRNA expression plasmid [24]. The results of Study B also provide evidence that the inhibition of luciferase activity in the brain tumour in vivo is due solely to the production of the antiluciferase shRNA encoded by clone 952 plasmid. The only difference in the molecular formulation between the clone 959 DNA encapsulated in the TfRmAbPIL and the clone 952 DNA encapsulated in the TfRmAb-PIL is that the clone 952 encodes for the shRNA (Figure 2B), whereas the clone 959 is the empty expression plasmid. These studies show that intravenous RNAi is possible in vivo with

ati on sL

td

Contralateral brain

Tumour

B luciferase mRNA 954

Clone 952 shRNA

As

hle

yP

Figure 2. (A) Coronal section of autopsy rat brain at 14 days after implantation of C6 rat glioma cells in the caudate putamen of adult Fischer CD344 rats weighing 180 – 200 g. The C6 cells were permanently transfected with clone 790 plasmid DNA and produce high levels of luciferase when grown as brain tumours in vivo [24]. Luciferase gene expression in the contralateral brain is not detectable. (B) Nucleotide sequence of the luciferase mRNA that is targeted by two different shRNAs, produced either by clone 954 or 952 plasmid DNA. The sequence targeted by the clone 952-derived shRNA is underlined and the sequence targeted by the clone 954-derived shRNA is overlined. The sequence and secondary structure of the shRNAs encoded by clone 952 is shown [24].

ub

lic

952

of

shRNA: Short hairpin RNA.

Co

py

rig

ht

promoter, was engineered that produced a shRNA directed against a specific region of the luciferase mRNA [24]. The target sequence within the luciferase mRNA is shown in Figure 2B. Two different expression plasmids, designated clones 952 and 954, were prepared that produce shRNAs which hybridise to overlapping sequences of the luciferase mRNA as shown in Figure 2B. Complementary oligodeoxynucleotides (ODNs) were synthesised to produce the ODN duplexes corresponding to either clone 952 or 954 shRNA. The synthetic ODN duplexes were subcloned into the expression plasmid downstream of the U6 promoter. A T5 terminator sequence for RNA polymerase III was introduced in the ODNs. Clone 952 was much more active than clone 954 based on cell culture studies and luciferase measurements, and clone 952 was used for subsequent in vivo studies in rats with experimental brain tumours. The clone 952 shRNA has the antisense strand on top or 5´ to the loop, a 25-mer stem, and an eight nucleotide loop, as depicted in Figure 2B. The empty expression plasmid is designated clone 959. Clone 952 or 959 were encapsulated in PILs targeted with the TfRmAb

Expert Opin. Biol. Ther. (2004) 4(7)

1107

Intravenous, non-viral RNAi gene therapy of brain cancer

Study

Time of treatment (days)

Form of treatment Time of sacrifice (days)

Luciferase (pg/mgp)

GTP (nmol/min/mgp)

A

10

952/TfRmAb-PIL

14

77 ± 13

1.3 ± 0.1

Saline

14

214 ± 42

td

Table 1. Luciferase and γ-glutamyl transpeptidase activity in brain cancer.

952/TfRmAb-PIL

12

Saline

12

952/TfRmAb-PIL

10

959/TfRmAb-PIL

10

952/TfRmAb-PIL

7

959/TfRmAb-PIL

7

ati on sL

5

87 ± 12

1.1 ± 0.2

275 ± 52

1.2 ± 0.3

31 ± 5

3.2 ± 0.3

248 ± 37

3.3 ± 0.3

24 ± 2

3.5 ± 0.2

235 ± 12

3.7 ± 0.4

lic

B

1.3 ± 0.1

yP

human brain cancer, following weekly intravenous gene therapy [15]. In order to augment the potency of the clone 882 expression plasmid, this vector contained the oriP and Epstein–Barr nuclear antigen (EBNA)-1 elements [27], which allow for a single round of replication of the expression plasmid with each division of the cancer cell. The inclusion of the oriP/EBNA-1 elements within the pCEP4 derived expression plasmid, under the influence of the SV40 promoter, enabled a tenfold increase in the level of gene expression in human U87 glioma cells in cell culture [36]. However, the EBNA-1 gene encodes a tumorigenic transacting factor [39] and this formulation may not be desirable in human gene therapy. It is possible that a comparable level of EGFR knockdown could be achieved without inclusion of the EBNA-1 gene with the use of a more potent form of antisense gene therapy such as RNAi. This was tested by engineering a new expression plasmid, designated clone 967, which lacked the oriP/EBNA-1 elements and which encoded for a shRNA directed against a specific sequence to the human EGFR mRNA [25]. This plasmid DNA, designated clone 967, was incorporated in PILs doubly targeted with both the HIRmAb and the TfRmAb for weekly intravenous gene therapy in SCID mice with human U87 brain cancer [25]. The conjugation of both the HIRmAb and the TfRmAb to the liposome was confirmed with differentially radiolabelled antibodies [15]. The structure of the doubly targeted PIL is shown in Figure 3A. The TfRmAb targets the mouse TfR on the endothelium of the vessels perfusing the brain cancer and this transports the PIL across the vascular barrier. The HIRmAb targets the HIR on the plasma membrane of the cancer cell in brain and this triggers endocytosis into the cell and transfer to the nuclear compartment. The target sequence of the human

As

hle

intracranial brain cancer, providing an effective gene delivery technology is used. Intravenous, non-viral RNAi of the target gene in brain cancer results in a 90% knockdown of the target gene expression, and this effect persists for at least 5 days after a single intravenous injection. Following these results with the luciferase brain tumour model, the authors next examined the ability of intravenous RNAi gene therapy to knock down the human epidermal growth factor receptor (EGFR) in experimental brain cancer and to prolong survival in mice with intracranial human brain cancer.

ub

Data are mean ± SE (n = 4 – 5 rats/point). Clone 952 or 959 plasmid DNA was encapsulated in the TfRmAb-PIL and was designated 952/TfRmAb-PIL or 959/TfRmAbPIL, respectively. Clone 959 is the empty expression plasmid encoding no shRNA [24]. TfRmAb-PIL is a PIL targeted with a mAb to the TfR. The sequence and structure for the clone 952 derived shRNA is given in Figure 2B. Clone 959 is the empty plasmid DNA. GTP: γ-Glutamyl transpeptidase; mAb: Monoclonal antibody; PIL: Pegylated immunoliposome; SE: Standard error; shRNA: Short hairpin RNA; TfR: Transferrin receptor.

Intravenous RNAi and brain cancer: EGFR target

of

4.

Co

py

rig

ht

The human EGFR is expressed in 90% of primary brain cancer [37], such as glioblastoma multiforme (GBM). In addition, EGFR plays an oncogenic role in 70% of solid cancers that originate outside the brain [38] and many of these solid cancers metastasise to the brain. The incidence of primary brain cancer in the US is ∼ 15,000 cases/year and the incidence of metastatic cancer to the brain is ∼ 150,000 cases/year in the US. Therefore, the EGFR plays an oncogenic role in the US in > 100,000 new cases/year of brain cancer. The EGFR gene can be knocked down with conventional antisense gene therapy using the PIL gene-targeting technology [15,27]. A eukaryotic expression plasmid, designated clone 882, encodes a 700 nucleotide RNA that is antisense to nucleotides 2317 – 3006 of the human EGFR mRNA, which encodes the cytoplasmic domain of the EGFR. Encapsulation of clone 882 in PILs that were doubly targeted with both the HIRmAb and the mouse TfRmAb resulted in a 100% increase in survival time in SCID mice with intracranial U87 1108

Expert Opin. Biol. Ther. (2004) 4(7)

Pardridge

A HIR

mAb2 RNAi

mTfR

mAb1

T5

td

U6 pro

Plasmid

mAb1

ati on sL

mAb2

B

hEGFR mRNA (nt 2529 – 2557) Clone 967 shRNA

C

lic

100 80

RNAi ED50 = 32 days

ub

Saline ED50 = 17 days

60 40 20 0 0

5

10

yP

Survivors (%)

120

15

20

25

30

35

40

45

50

hle

Days post-transplantation

rig

ht

of

As

Figure 3. (A) A model of PIL that is doubly targeted to both the mTfR with the 8D3 monoclonal antibody (mAb1) and to the HIR with the 83-14 monoclonal antibody (mAb2). Encapsulated in the interior of the PIL is the plasmid DNA encoding the shRNA, which produces the RNAi. The gene encoding the shRNA is driven by the U6 promoter (pro) and is followed on the 3´-end with the T5 termination sequence for the U6 RNA polymerase. (B) Nucleotide sequence of the hEGFR mRNA sequence between nucleotides 2529 and 2557 is shown on top.The sequence and secondary structure of the shRNA produced by clone 967 is shown on the bottom. The antisense strand is 5´ to the 8 nucleotide loop and the sense strand is 3´ to the loop. The sense strand contains 4 G/U mismatches to reduce the Tm of hybridisation of the stem loop structure; the sequence of the antisense strand is 100% complementary to the target mRNA sequence. (C) Intravenous RNAi gene therapy directed at the hEGFR is initiated at 5 days after implantation of 500,000 U87 cells in the caudate putamen nucleus of SCID mice and weekly intravenous gene therapy is repeated at days 12, 19 and 26 (arrows). The control group was treated with saline on the same days. There are 11 mice in each of the two treatment groups. The time at which 50% of the mice were dead (ED50) is 17 days and 32 days in the saline and RNAi groups, respectively. The RNAi gene therapy produces an 88% increase in survival time [25].

py

ED50: 50% Effective dose; hEGFR: Human epidermal growth factor receptor; HIR: Human insulin receptor; mAb: Monoclonal antibody; mTfR: Mouse transferrin receptor; PIL: Pegylated immunoliposome; RNAi: RNA interference; SCID: Severe combined immunodeficient; shRNA: Short hairpin RNA.

Co

EGFR mRNA and the structure of shRNA produced by clone 967 are shown in Figure 3B. The sequence targeted by clone 967 was isolated following the initial production of multiple expression plasmids producing anti-EGFR shRNAs directed against sequences within six different regions of the 5000 nucleotide human EGFR mRNA [25]. Some of the plasmids produced shRNAs against regions of the human EGFR mRNA, which resulted in no knockdown of EGFR gene expression [25]. Plasmids producing shRNAs that target other sequences within the human EGFR mRNA produced an intermediate level of EGFR gene expression knockdown. However, only clone 967, which targets nucleotides 2529 – 2557,

produced a level of EGFR gene expression knockdown which was comparable to that achieved previously with clone 882 and conventional antisense gene therapy [27]. The ability of clone 967 to knockdown functional EGFR gene expression in human U87 glioma cells in cell culture following targeting with HIRmAb-PILs was demonstrated with fluorescence video microscopy of intracellular calcium transport. These results showed that the functional EGFR was 95% ablated in cultured cells following delivery of clone 967 with HIRmAb-PILs [25]. For the in vivo brain tumour survival study, 500,000 human U87 glioma cells were implanted in the caudate putamen of

Expert Opin. Biol. Ther. (2004) 4(7)

1109

Intravenous, non-viral RNAi gene therapy of brain cancer

Table 2. Capillary density in brain tumour and normal brain. Treatment

Capillary density per 0.1 mm2

Tumour centre

Saline

15 ± 2

Normal brain

RNAi

3±0

Saline

29 ± 4

RNAi

8±1

ati on sL

Tumour periphery

td

Region

35 ± 1

Saline

33 ± 1

RNAi

Mean ± SE (n = 15 fields analysed from three mice in each of the treatment groups). From [25]. These data correspond to the study described in Figure 3. RNAi: RNA interference; SE: Standard error.

yP

ub

lic

the human EGFR. Therefore, the shRNA produced by clone 967 would not be expected to affect endogenous mouse EGFR gene expression. The results of the survival study (Figure 3C) show that there is an 88% increase in survival time with weekly intravenous RNAi-based gene therapy directed against the human EGFR using the PIL gene-targeting technology [25]. This increase in survival time is not a nonspecific effect of PIL administration, as prior work has shown no change in survival with the weekly intravenous administration of mAb-targeted PILs carrying a luciferase expression plasmid [15]. The increase in survival time obtained with the weekly intravenous anti-EGFR RNAibased gene therapy is comparable to the prolongation of survival time in mice treated with high daily doses of the EGFR tyrosine kinase inhibitor, ZD1839 (Iressa®, AstraZeneca) [41]. However, ZD1839 is not effective in the treatment of brain cancer expressing mutant forms of the EGFR [41]. Many primary and metastatic brain cancers express mutations of the EGFR [42], and it is possible to design RNAi-based gene therapy that will knock down mutant, but not wild-type, EGFR mRNAs. The shRNA designed in Figure 3 would target both wild-type and mutant EGFRs.

Co

py

rig

ht

of

As

hle

the brain of SCID mice under stereotaxic guidance [25]. At this dose of tumour cells, the mice die between 14 and 20 days in the absence of treatment (Figure 3C). Weekly intravenous RNAi-based gene therapy was initiated at day 5 post tumour cell implantation, when the size of the tumour has already grown to a point where it occupies the entire volume of the striate body in mouse brain. The mice were treated either with saline or a dose of clone 967 plasmid DNA (5 µg/ week), which was encapsulated in PILs that were targeted both with the HIRmAb and the TfRmAb (Figure 3A). Salinetreated mice died 14 – 20 days postimplantation with an ED50 of 17 days (Figure 3C). The mice treated with intravenous RNAi-based gene therapy directed against the EGFR died 31 – 34 days postimplantation with an ED 50 of 32 days, which represents an 88% increase in survival time compared to the animals treated with saline (Figure 3C). Knockdown of EGFR in the brain tumour in vivo was demonstrated with confocal microscopy of brain tumour sections taken at autopsy. This showed a high level of tumour cell EGFR expression in the saline-treated animals, but minimal EGFR expression in the animals treated with RNAi-based gene therapy. Another functional assay of EGFR knockdown in brain tumours in vivo was the measurement of the vascular density in the brain cancer [25]. The EGFR has a pro-angiogenic effect in brain cancer [40], and knockdown of EGFR in brain cancer should have antiangiogenic effects. This antiangiogenic effect was confirmed as shown in Table 2. These data are the results of quantitation of capillary density following immunodetection of the vessels with TfR immunocytochemistry. The capillaries perfusing the U87 brain tumours were of mouse brain origin and continue to express high levels of murine TfR, which was demonstrated by immunocytochemistry. The antiEGFR RNAi treatment resulted in an 80 and 72% decrease in vascular density in the tumour centre and tumour periphery, respectively, as compared to the saline-treated animals (Table 2). There is no reduction in vascular density in control mouse brain (Table 2). A BLAST analysis of the nucleotide sequence of the human EGFR mRNA (accession number X00588) and the mouse EGFR mRNA (accession number AF275367) shows that there is only a 76% identity in the mouse sequence corresponding to nucleotides 2529 – 2557 of 1110

Human applications of the PIL gene-targeting technology 5.

The therapeutic efficacy of the PIL gene transfer technology is possible because this approach delivers therapeutic genes to the brain via the transvascular route. The PIL gene transfer technology is effective in primates (Figure 1), and human applications of this new gene transfer technology for the treatment of brain cancer are now possible. Targeting mAbs of murine origin cannot be used in humans. However, the HIRmAb has been genetically engineered to produce a chimeric HIRmAb and the chimeric antibody has the same activity as the original murine antibody, both with respect to binding to the HIR in vitro and to transport across the primate BBB in vivo [43]. Future clinical applications of the PIL gene transfer technology in humans will require repeated administration owing to the episomal nature of gene expression with this technology. There is no permanent integration

Expert Opin. Biol. Ther. (2004) 4(7)

Pardridge

Bibliography

PLANK C, TANG MX, WOLFE AR, SZOKA FC JR: Branched cationic peptides for gene delivery: role of type and number of cationic residues in formation and in vitro activity of DNA polyplexes. Hum. Gene Ther. (1999) 10(2):319-332.

5.

6.

td

DYKXHOORN DM, NOVINA CD, SHARP PA: Killing the messenger: short RNAs that silence gene expression. Nat. Rev. Mol. Cell. Biol. (2003) 4(6):457-467.

hle

7.

8.

9.

CHIRMULE N, PROPERT K, MAGOSIN S et al.: Immune responses to adenovirus and adeno-associated virus in humans. Gene Ther. (1999) 6(9):1574-1583.

13.

PARDRIDGE WM: Drug and gene delivery to the brain: the vascular route. Neuron (2002) 36(4):555-558.

14.

FIDLER IJ, YANO S, ZHANG RD, FUJIMAKI T, BUCANA CD: The seed and soil hypothesis: vascularisation and brain metastases. Lancet Oncol. (2002) 3(1):53-57.

15.

ZHANG Y, ZHU C, PARDRIDGE WM: Antisense gene therapy of brain cancer with an artificial virus gene delivery system. Mol. Ther. (2002) 6(1):67-72.

NAKAI H, MONTINI E, FUESS S et al.: AAV serotype 2 vectors preferentially integrate into active genes in mice. Nat. Genet. (2003) 34(3):297-302.

16.

PARDRIDGE WM: Drug and gene targeting to the brain with molecular Trojan horses. Nat. Rev. Drug Discov. (2002) 1(2):131-139.

11.

OSAKA G, CAREY K, CUTHBERTSON A et al.: Pharmacokinetics, tissue distribution, and expression efficiency of plasmid [33P]DNA following intravenous administration of DNA/cationic lipid complexes in mice: use of a novel radionuclide approach. J. Pharm. Sci. (1996) 85(6):612-618.

17.

PARDRIDGE WM, EISENBERG J, YANG J: Human blood–brain barrier insulin receptor. J. Neurochem. (1985) 44(6):1771-1778.

18.

PARDRIDGE WM: Brain Drug Targeting: The Future of Brain Drug Development. Cambridge University Press, Cambridge, UK (2001):209.

19.

SHAH N, ZHANG S, HARADA S, SMITH RM, JARETT L: Electron microscopic visualization of insulin translocation into the cytoplasm and nuclei

HOFLAND HE, NAGY D, LIU JJ et al.: In vivo gene transfer by intravenous administration of stable cationic lipid/DNA complex. Pharm. Res. (1997) 14(6):742-749. ISACSON R, KULL B, SALMI P, WAHLESTEDT C: Lack of efficacy of ‘naked’ small interfering RNA applied directly to rat brain. Acta Physiol. Scand. (2003) 179(2):173-177.

WU X, LI Y, CRISE B, BURGESS SM: Transcription start regions in the human genome are favored targets for MLV integration. Science (2003) 300(5626):1749-1751.

••

vector-producing cells. Nat. Med. (1997) 3(12):1354-1361. This study shows limitations in transcranial delivery of genes to brain cancer.

10.

Co

4.

py

rig

3.

MAHATO RI, ROLLAND A, TOMLINSON E: Cationic lipid-based gene delivery systems: pharmaceutical perspectives. Pharm. Res. (1997) 14(7):853-859.

enzyme complex. Science (2002) 297(5589):2056-2060.

of

2.

NIIDOME T, OHMORI N, ICHINOSE A et al.: Binding of cationic alpha-helical peptides to plasmid DNA and their gene transfer abilities into cells. J. Biol. Chem. (1997) 272(24):15307-15312.

ht

1.

These features are provided with the PIL non-viral gene transfer technology.

As

Papers of special note have been highlighted as either of interest (•) or of considerable interest (••) to readers.

non-toxic with repeat administration non-viral episomal, not chromosomal able to target distant sites in the body following intravenous administration • potential to eliminate ectopic gene expression with the use of organ-specific gene promoters

yP

Once gene delivery problems are solved, RNAi gene therapy of brain cancer, and gene therapy of cancer in general, can move from monogenic to polygenic forms of therapy. Cancer

ati on sL

Expert opinion and conclusions

• • • •

lic

6.

arises from the mutation of multiple genes within a single cell. Therefore, it is unlikely that the targeting of a single mutated gene will result in a cancer cure. Rather, gene therapy should be directed at multiple gene targets simultaneously, just as chemotherapy of cancer employs multiple drugs simultaneously. Polygenic gene therapy may simultaneously knock down oncogenic genes, as with RNAi-based gene therapy or replace mutated tumour suppressor genes with replacement gene therapy. However, gene therapy of cancer has not moved beyond the monogenic phase because progress in this area has been slowed by the gene delivery problem. An ideal gene delivery system should have the following characteristics:

ub

in the host genome, as demonstrated by Southern blotting [44], and this feature is considered desirable because there is no risk of insertional mutagenesis. Moreover, the weekly administration of PILs has no toxic effects and causes no inflammation in the brain [44]. With regard to eliminating ectopic expression of exogenous genes in non-cancerous cells, prior work has shown that region specific gene expression is possible with tissue-specific gene promoters in either mice [34] or primates [35]. Alternatively, many cancers express mutant forms of the oncogene [42], which are produced from aberrantly processed mRNAs. These mutant mRNAs contain nucleotide sequences not found in normal cells and are ideal targets for RNAi-based gene therapy in cancer. Expression plasmids can be generated that produce shRNAs that target relatively short 15 – 25 nucleotide sequences that are specific for mutant transcripts found only in cancer cells.

12.

RAM Z, CULVER KW, OSHIRO EM et al.: Therapy of malignant brain tumors by intratumoral implantation of retroviral

HUTVAGNER G, ZAMORE PD: A microRNA in a multiple-turnover RNAi Expert Opin. Biol. Ther. (2004) 4(7)

1111

Intravenous, non-viral RNAi gene therapy of brain cancer

22.

•• 23.

HUWYLER J, WU D, PARDRIDGE WM: Brain drug delivery of small molecules using immunoliposomes. Proc. Natl. Acad. Sci. USA (1996) 93(24):14164-14169. The first demonstration of receptor-mediated transport of PILs across the BBB.

29.

30.

31.

•

25.

rig

py

••

ZHANG Y, BOADO RJ, PARDRIDGE WM: In vivo knockdown of gene expression in brain cancer with intravenous RNAi in adult rats. J. Gene Med. (2003) 5(12):1039-1045. The first study of RNAi in the brain in vivo following intravenous administration of the shRNA-encoding plasmid. ZHANG Y, BRYANT J, ZHANG YF, CHARLES A, BOADO RJ, PARDRIDGE WM: Intravenous RNAi gene therapy targeting the human EGF receptor prolongs survival in intra-cranial

1112

32.

•

Co

24.

33.

34.

SHI N, ZHANG Y, ZHU C, BOADO RJ, PARDRIDGE WM: Brain-specific expression of an exogenous gene after i.v. administration. Proc. Natl. Acad. Sci. USA (2001) 98(22):12754-12759. Expression of exogenous gene in the mouse is confined to the brain with combined use of brain-specific gene promoter and PIL gene transfer technology.

ati on sL

td

•

35.

•

lic

SHI N, PARDRIDGE WM: Non-invasive gene targeting to the brain. Proc. Natl. Acad. Sci. USA (2000) 97:7567-7572.

ZHANG Y, SCHLACHETZKI F, LI JY, BOADO RJ, PARDRIDGE WM: Organspecific gene expression in the rhesus monkey eye following intravenous non-viral gene transfer. Mol. Vis. (2003) 9:465-472. Expression of exogenous gene in the rhesus monkey is confined to the eye with combined use of ocular-specific gene promoter and PIL gene transfer technology.

36.

ZHANG Y, BOADO RJ, PARDRIDGE WM: Marked enhancement in gene expression by targeting the human insulin receptor. J. Gene Med. (2003) 5(2):157-163.

37.

KUAN CT, WIKSTRAND CJ, BIGNER DD: EGF mutant receptor vIII as a molecular target in cancer therapy. Endocr. Relat. Cancer (2001) 8(2):83-96.

38.

NICHOLSON RI, GEE JM, HARPER ME: EGFR and cancer prognosis. Eur. J. Cancer (2001) 37(Suppl. 4):S9-S15.

CORNELIS S, VANDENBRANDEN M, RUYSSCHAERT JM, ELOUAHABI A: Role of intracellular cationic liposomeDNA complex dissociation in transfection mediated by cationic lipids. DNA Cell Biol. (2002) 21(2):91-97. This study shows that plasmid DNA is largely confined to the cytoplasmic compartment following delivery to cells with cationic liposome.

39.

SNUDDEN DK, SMITH PR, LAI D, NG MH, GRIFFIN BE: Alterations in the structure of the EBV nuclear antigen, EBNA1, in epithelial cell tumours. Oncogene (1995) 10(8):1545-1552.

40.

ABE T, TERADA K, WAKIMOTO H et al.: PTEN decreases in vivo vascularization of experimental gliomas in spite of proangiogenic stimuli. Cancer Res. (2003) 63(9):2300-2305.

ZHANG Y, SCHLACHETZKI F, PARDRIDGE WM: Global non-viral gene transfer to the primate brain following intravenous administration. Mol. Ther. (2003) 7(1):11-18. Global expression of transgene in brain of adult rhesus monkey is demonstrated following intravenous administration of gene encapsulated in PILs.

41.

HEIMBERGER AB, LEARN CA, ARCHER GE et al.: Brain tumors in mice are susceptible to blockade of epidermal growth factor receptor (EGFR) with the oral, specific, EGFR-tyrosine kinase inhibitor ZD1839 (Iressa). Clin. Cancer Res. (2002) 8(11):3496-3502.

42.

LAL A, GLAZER CA, MARTINSON HM et al.: Mutant epidermal growth factor receptor up-regulates molecular effectors of tumor invasion. Cancer Res. (2002) 62(12):3335-3339.

43.

COLOMA MJ, LEE HJ, KURIHARA A et al.: Transport across the primate

ub

28.

ht

••

PAPAHADJOPOULOS D, ALLEN TM, GABIZON A et al.: Sterically stabilized liposomes: improvements in pharmacokinetics and antitumor therapeutic efficacy. Proc. Natl. Acad. Sci. USA (1991) 88(24):11460-11464. The initial description of stealth liposomes with improved plasma pharmacokinetics.

•

ZHANG Y, JEONG LEE H, BOADO RJ, PARDRIDGE WM: Receptor-mediated delivery of an antisense gene to human brain cancer cells. J. Gene Med. (2002) 4(2):183-194. This study shows plasmid DNA is largely confined to the nuclear compartment following delivery to cells with PILs.

TAM P, MONCK M, LEE D et al.: Stabilized plasmid-lipid particles for systemic gene therapy. Gene Ther. (2000) 7(21):1867-1874.

yP

••

BICKEL U, KANG YS, YOSHIKAWA T, PARDRIDGE WM: In vivo demonstration of subcellular localization of anti-transferrin receptor monoclonal antibody-colloidal gold conjugate in brain capillary endothelium. J. Histochem. Cytochem. (1994) 42(11):1493-1497. Electron microscopic evidence for receptor-mediated transcytosis of peptidomimetic mAb across the BBB.

27.

LEE HJ, ENGELHARDT B, LESLEY J, BICKEL U, PARDRIDGE WM: Targeting rat anti-mouse transferrin receptor monoclonal antibodies through blood– brain barrier in mouse. J. Pharmacol. Exp. Ther. (2000) 292(3):1048-1052.

ZHANG Y, CALON F, ZHU C, BOADO RJ, PARDRIDGE WM: Intravenous nonviral gene therapy causes normalization of striatal tyrosine hydroxylase and reversal of motor impairment in experimental parkinsonism. Hum. Gene Ther. (2003) 14(1):1-12.

hle

21.

PARDRIDGE WM, KANG YS, BUCIAK JL, YANG J: Human insulin receptor monoclonal antibody undergoes high affinity binding to human brain capillaries in vitro and rapid transcytosis through the blood–brain barrier in vivo in the primate. Pharm. Res. (1995) 12(6):807-816.

26.

As

20.

••

brain cancer. Clin. Cancer Res. (2004) 10(11):3667-3677. The first study of prolongation of survival in cancer with RNAi-based gene therapy.

of

••

of intact H35 hepatoma cells using covalently linked Nanogold-insulin. Endocrinology (1995) 136(7):2825-2835. This study shows that the insulin receptor normally serves to deliver its endogenous ligand, insulin, to the nuclear compartment following uptake from the extracellular space.

SHI N, BOADO RJ, PARDRIDGE WM: Receptor-mediated gene targeting to tissues in the rat in vivo. Pharm. Res. (2001) 18:1091-1095.

Expert Opin. Biol. Ther. (2004) 4(7)

Pardridge

44.

td

ZHANG YF, BOADO RJ, PARDRIDGE WM: Absence of toxicity of chronic weekly intravenous gene therapy with pegylated immunoliposomes. Pharm. Res. (2003) 20(11):1779-1785.

ati on sL

•

blood–brain barrier of a genetically engineered chimeric monoclonal antibody to the human insulin receptor. Pharm. Res. (2000) 17(3):266-274. A genetically engineered chimeric mAb is transported across the primate BBB with the same activity as the original murine mAb.

Affiliation

Co

py

rig

ht

of

As

hle

yP

ub

lic

William M Pardridge Department of Medicine, UCLA, Warren Hall 13-164, 900 Veteran Ave, Los Angeles, CA 90024, USA Tel: +1 310 825 8858; Fax: +1 310 206 5163; E-mail: [email protected]

Expert Opin. Biol. Ther. (2004) 4(7)

1113