¹Center for Molecular Imaging Research, Charlestown, MA, USA

²Molecular Neurogenetics Unit, Neuroscience Center, Massachusetts General Hospital

Correspondence to: R Weissleder, Center for Molecular Imaging Research, Massachusetts General Hospital, Building 149, 13th Street, 5403, Charlestown, MA 02129, USA

^aCurrent address: Department of Neurosurgery, Faculty of Medicine, Martin-Luther-University, Halle (Saale), Germany

A number of different viral vectors have been used for gene therapy of tumors, with many more under construction, ultimately designed to improve tumor targeting and transduction efficiency. It has become apparent that insufficient viral delivery can be a key limitation to treatment efficacy. We have studied the in vivo mass distribution of a herpes simplex virus type 1 (HSV) vector, hrR3, by radiolabeling it with ¹¹¹in-oxine. the virus was administered to intracerebral 9l glioma bearing fisher (f-344) rats by intracarotid and intratumoral injection. the blood half-life of the virus was 1 min (fast component, 10% contribution) and 180 min (slow component, 90% contribution). approximately 20% of activity had been excreted by 24 h. with intracarotid injection, the total amount of virus that accumulated in tumor was 0.10 ± 0.07% of the injected dose (id)/g at 1 h and 0.19 ± 0.01% id/g at 24 h. by comparison, co-injection of rmp-7, a synthetic bradykinin analog, with the virus, resulted in slightly increased tumor delivery of 0.17 ± 0.10% id/g (p 0.05) at 1 h. the 1 h organ distribution after intra-arterial injection (%id/organ) was as follows: liver 27.3 ± 2.86%, lung 2.10 ± 0.68% and kidney 1.78 ± 1.60% with lesser amounts in other organs. when virus was injected directly into the tumor, 71% of virus remained in tumor at 24 h (590 ± 212 %id/g, consistent with the small tumor mass containing most of the virus) with the following distribution regions: tumor > border zone > normal brain (99:40:1). These studies are the first quantitative mass distribution studies of HSV vectors in an experimental brain tumor model. Localization and quantitation of viral accumulation in vivo will enable detailed analysis of viral and organ interactions critical for advancing the therapeutic use of vectors. Gene Therapy (2000) 7, 1648-1655.

gene transfer; gene therapy; herpes simplex virus; imaging

Introduction

There are currently over 300 gene therapy protocols in clinical phase I-III trials in the US alone. These protocols are designed, for instance, to test novel marker proteins,¹ perform gene replacement and/or augmentation trials,² test prodrug activation,³ evaluate cell-based therapies⁴ and investigate oncolytic viruses.⁵ During the last 5 years of vector and transgene development, it has become apparent that insufficient delivery of genetic material to tumor cells in vivo and low and/or transient gene expression are the main limitations to the effectiveness of this new treatment paradigm. As a result, a number of novel approaches are being actively pursued to target vectors to specific cells and to improve the efficiency of gene delivery. In order to compare and optimize different delivery strategies, it is important to assess vector delivery and gene expression quantitatively in animal models and, eventually, in human patients. To date, most of our knowledge about transgene expression has come from techniques such as in situ hybridization, PCR and immunohistochemistry. These methods generally do not allow one to obtain information pertinent to the time course and the spatial distribution of gene expression in a clinical setting.

In vivo imaging techniques have the potential to provide critical information of viral and nonviral vector and transgene expression. With the continued development and sophistication of chimeric, targeted and large capacity 'gutted' virus,^6,7,8,9,10 as well as other in vivo transfection methods, it has become clear that additional strategies are needed to quantitatively monitor and compare gene delivery and expression by different methods. One potentially fruitful avenue has been labeling of virions with isotope tracers.^11,12 In enveloped viruses such as HSV, the lipophilic ¹¹¹In-oxine tracer can transverse the viral envelope and be stabilized, likely by rechelation to cysteine-rich viral capsid proteins, such as glycoprotein D and/or ICP5/VP22.¹¹ In prior studies we have shown the virus-label association to be stable during dialysis over several hours.¹¹ Thus labeled HSV can be used to directly quantitate viral mass distribution over time either by in vivo imaging in intact animals and/or by performing quantitative biodistribution studies at given time points.

The goal of the current study was to extend prior feasibility studies and apply the ¹¹¹In tagging technique to study HSV viral delivery to intracerebral implanted gliomas in a rodent model. The questions addressed were: (1) what is the mass distribution of HSV in gliomas following intra-arterial or intra-tumoral delivery; (2) what are the in vivo organ accumulations at 1 and 24 h after administration; and (3) can the tumoral accumulation of labeled HSV be imaged in vivo? These questions were addressed by administering ¹¹¹In-oxine labeled HSV to rats bearing intracerebral 9L gliosarcomas using either intracarotid or direct intratumoral administrations. Carotid injections were performed with and without infusion of RMP-7, a bradykinin analog which facilitates transfer across the blood-tumor barrier.¹³ The viral distribution was determined by biodistribution studies, while the feasibility of in vivo imaging was tested by gamma camera imaging.

Results

The blood half-life of HSV demonstrates a fast and a slow decay component. The fast component half-life was 1 min (10% contribution), and the slow was 180 min (90% contribution). Excretion of virus-associated activity was by the urinary route, and was approximately 20% at 24 h.

Following intracarotid injection, 0.10 ± 0.07% ID/g (mean ± standard deviation) of HSV had accumulated in the tumor at 1 h after injection, whereas considerably less (0.02 ± 0.01% ID/g) was found in the contralateral normal brain (Table 1). At 24 h after injection, there was 0.19 ± 0.01% ID/g in the tumor and 0.02 ± 0.02% ID/g in normal brain. Brain tissue around the tumor accumulated 0.15 ± 0.18% ID/g at 1 h after injection and 0.26 ± 0.25% ID/g at 24 h. The organ distribution of virus is summarized as per cent injected dose per gram and per cent injected dose per organ in Tables 1 and 2, respectively. The ratios of tumor:tumor periphery:normal contralateral brain was approximately 5:9:1 at 1 h, and 10:13:1 at 24 h (Table 3).

Microscopic viral distribution to tumors was investigated with a dual-labeled virus, containing both a indocyanine fluorochrome (CY3 dye) and ¹¹¹In-oxine, injected into the carotid. A GFP-expressing 9L glioma was used as tumor, facilitating detection of the tumor mass by fluorescent microscopy.¹⁴ The distribution of virus in tumors was markedly heterogeneous. Fluorescent microscopy demonstrated granular clumps of HSV particles (labeled red by the indocyanine dye) primarily within blood vessels in the tumor periphery (Figure 1). Autoradiography showed most of the labeled virus to be within or near to the tumor, as compared with normal background brain, with the highest concentration of virus noted at the brain-tumor interface.

The 1 h viral distribution (after intra-arterial injection) to other organs was most pronounced to the liver at 2.6 ± 0.3% ID/g, the lungs at 1.8 ± 0.6% ID/g, the spleen at 1.2 ± 0.4% ID/g, the kidneys at 0.9 ± 0.8% ID/g and to lesser amounts in other organs (Table 1). By 24 h after injection, the liver distribution had increased further to 4.2 ± 1.0% ID/g, while virus tended to distribute away from the lungs, with 1.2 ± 0.7% ID/g still there. The spleen showed 2.4 ± 0.8% ID/g, slightly less than the kidneys, with viral accumulation at 2.7 ± 1.4% ID/g. Ratios of normal brain/kidneys/liver were 1:51:146 at 1 h and 1:136:210 at 24 h (Table 3). When no correction is made for organ masses, the differences in distribution become even more apparent (Table 4).

Whole body imaging of animals paralleled the above biodistribution data (Figure 2). In some animals (n = 2) a left/right brain activity difference was visually appreciable consistent with the higher uptake in the tumor-bearing hemisphere, while in other animals the difference was less marked or only discernable by region of interest analysis (n = 3). Intense uptake in the liver and spleen on imaging was present both initially and at 24 h (Figure 2). Lung uptake was present, with subsequent redistribution of tracer and lower uptake at 24 h, as compared with 1 h, reflecting the biodistribution data.

When the bradykinin analog RMP-7 was administered intra-arterially along with the virus, a different biodistribution was encountered. 9L gliomas accumulated 0.17 ± 0.10% ID/g within 1 h, 70% higher than without RMP-7. There was a large standard deviation, but this increase was statistically significant when compared with injections without RMP-7 (P < 0.05) (Table 1). With RMP-7, delivery to the tumor periphery was 0.20 ± 0.16% ID/g while delivery to the normal brain was 0.08 ± 0.03% ID/g, as compared with 0.15 ± 0.18% ID/g and 0.02 ± 0.01% ID/g, respectively, without RMP-7 (these numbers did not attain statistical significance, but support increased delivery with RMP-7). In a separate study involving control animals without brain tumors, unilateral intra-arterial RMP-7 and virus co-administration did not significantly alter viral distribution between the two hemispheres, with 0.06 ± 0.01% ID/g in the right hemisphere receiving RMP-7 and virus intra-arterially, versus 0.05 ± 0.01% ID/g on the left side contralateral to the injection (data not shown). RMP-7 had a significant altering effect on the bio-distribution of virus outside the central nervous system also. Uptake was notably lower in the liver (1.5 ± 0.3% ID/g as compared with 2.6 ± 0.3% ID/g without RMP-7). The whole organ data (Table 2) demonstrate the reduction in liver uptake even more dramatically, with 12.7 ± 2.4% of the total dose distributing to liver, compared with 27.3 ± 2.8% without RMP-7. There was a commensurate slight increase in the viral load in all other organs, with the exception of the lung and spleen.

In animals receiving virus injections stereotactically into the tumor, 71.3 ± 35.0% of the total dose was found in the tumor at 24 h, with 28.6 ± 20.8% of the total dose in the brain adjacent to the tumor. Imaging at 1 h demonstrated the very focal tracer accumulation in the brain tumors of these animals to be relatively unchanged compared with 24 h. In mass corrected numbers at 24 h, 590 ± 211% ID/g of activity was found in tumor, and 88.6 ± 64.6% ID/g in the immediately adjacent brain. These numbers appear high, because of the relatively small amount of mass (a few milligrams of tumor) containing the bulk of the radioactivity. The only other organs to show high uptake was the kidney, with 2.9 ± 4.5% ID/g uptake (note the high standard deviation) and contralateral brain at 1.5 ± 1.3% ID/g (Tables 1 and 2). Small amounts of activity were noted in the spleen (0.4 ± 0.2% ID/g), liver (0.3 ± 0.0% ID/g), aorta (0.3 ± 0.3% ID/g), and lesser amounts in other organs (Table 1). In vivo nuclear imaging of tumor-injected animals demonstrated intense focal activity at the injection site, with very little activity elsewhere in the body at either 1 h or 24 h. There was very little redistribution of activity, with imaging at 24 h similar to that done immediately after injection.

Discussion

The current study was designed to determine mass distributions of radioactively labeled herpes simplex virus to different tissues and organs using common administration routes for experimental brain tumor therapy. This type of data is critical in evaluating the baseline viral biodistribution for clinical gene therapy trials. Our current data indicate that intracarotid administration of HSV ipsilateral to an intracerebral glioma resulted in only 0.001% accumulation in the tumor, corresponding to 0.10 ± 0.01% ID/g tumor within 1 h after administration. An additional 0.01% of the dose (0.15 ± 0.18% ID/g) accumulated in the tumor periphery. This interface between normal brain and tumor is where tumor recurrence most often occurs following conventional therapies.¹⁵ When the bradykinin analog, RMP-7, was co-administered with HSV, slightly higher amounts of virus were delivered to tumors (0.10 ± 0.07% ID/g without RMP-7 compared to 0.17 ± 0.10% ID/g with RMP-7; P < 0.05) and to the tumor periphery (0.15 ± 0.18% ID/g without RMP-7 compared with 0.20 ± 0.16% ID/g with RMP-7, difference not statistically significant). The significance of these findings is marginal due to large standard deviations in the amount delivered. Tumoral and tumor periphery accumulations were slightly higher at 24 h after injection, presumably because circulating virus continues to accumulate in tumors due to extravasation across leaky tumor vessels. Although the amount of virus delivered to tumor is low, selective propagation in tumor (but not elsewhere) by a replication conditional vector, such as hrR3, will serve to expand the therapeutic index in vivo, as even a single infective virion could give rise to a 'chain reaction' of cell lysis within the tumor.¹⁶

The major advantage of using a replication-conditional oncolytic vector, such as hrR3, is that it can potentially treat invasive tumor beyond the normal reach of the traditional therapies of surgery and radiation therapy. Most gliomas recur within 3 cm of the neurosurgical resection margin, even after administration of radiation therapy.¹⁵ The intratumor injection paradigm of vector administration could potentially be suitable for the delivery of large quantities of virus to the tumor center. Virus-induced cell lysis would then occur, with propagation of more vector and cell lysis throughout the tumor as the virus continued replication and infected adjacent tumor cells. The wave of cell lysis would theoretically stop at the interface with normal brain, as the vector design does not allow replication in normal brain,¹⁷ or as immune response to the virus curtails its spread.¹⁸

Intra-arterial administration, though having a lower delivery to the tumor mass, may still be effective, provided infection of some tumor cells takes place, allowing the 'chain reaction' of propagation from tumor cells to their neighbors to be primed. Intra-arterial delivery has the major potential advantage of being able to treat invasive tumor cells and satellite lesions beyond the primary site of tumor. The large size, complex geometry, multifocality and diffusely infiltrating margins of human gliomas¹⁶ indicates the need for both diffuse and focal therapies. Intra-arterial delivery appears most suited for the diffuse treatment of the tumor margin, because of its propensity to localize to this area. It is now well accepted that the fastest cell replication, and thus the cells most vulnerable to viral vector lysis, as well as the most permeable neovasculature is at the tumor margin.^19,20 This also happens to be the site of failure of traditional therapies, pointing to the potential usefulness of viral gene therapy to complement or complete standard therapies.

Assuming a single compartment, large fluid model with no barriers to diffusion, the expected theoretical viral distribution to tumor or other tissues in our study would be 0.35% ID/g (total dose divided by mean animal mass) following vascular administration and complete distribution through the tissues. The observed accumulation in tumor of 0.10% ID/g, points to the importance of physiological delivery barriers, primarily the blood-brain barrier (BBB) and the blood-tumor barrier (BTB).²¹ After RMP-7 administration, the consequent 70% increase of viral load to experimental tumors is likely due to a partial disruption of the barrier.²² It is interesting to note that in control animals lacking tumors but receiving unilateral RMP-7, no differences in viral accumulation were observed between ipsilateral right and contralateral left hemispheres. In other words, RMP-7 did not affect the normal BBB, but indeed had a selective effect on the permeability of the BTB. These data are consistent with prior observations.^13,22,23 In one study in humans, intra-arterial RMP-7 increased the transport of a small molecular weight PET tracer by 46 ± 42% into gliomas; and similar to our findings, the normal brain tissue was not significantly affected.²⁴

In addition to CNS effects, RMP-7 also had peripheral effects on viral distribution. Bradykinin, when infused intravenously has been shown to increase vascular resistance to the splanchnic bed and visceral organs (liver, stomach, small and large intestine, pancreas and mesentery). Vasoconstriction was also noted in the epididymides, skeletal muscle and fat.²⁵ In contrast, the brain showed reduced vascular resistance.²⁵ RMP-7, in contrast to the bradykinin parent molecule, is a selective B2 receptor agonist, implying that its effects are exerted mainly on the endothelial cells and much less on blood pressure.¹³ The present study did show slightly reduced delivery to the liver, spleen and lungs with, as compared to without, RMP-7. There was a proportionate increase in delivery to the kidneys with the drug (Table 1).

Virus distribution in tumors on a microscopic level was noted to be very heterogeneous, consistent with the highly variable intratumoral biology reported by other authors, including areas of necrosis, hyperperfusion, non-perfusion and altered vascular permeability.²⁶ However, there was a trend for viral particles to be found in the periphery of tumors, an area known for its increased neovascularity,²¹ and internally in hotspots with marked microscopically apparent vascularity similar to that described in a previous work using this tumor cell line.¹⁴ There are several possible reasons for this finding. First, neo-vessels are known to have increased permeability,²⁷ and the virus may leak out into the perivascular interstitium around the vessels, especially in the context of reduced intra-tumor pressure in the periphery.²⁸ Second, neovascular cells may be more infectable²⁹ than established endothelial cells. Third, some components of the endothelial neovascular lining could be comprised of tumor cells³⁰ and thus be directly infectable. Fourth, the contorted vasculature of tumor may retain the virus in the blood pool for a longer period of time than in normal vessels,²⁶ thus allowing more opportunity for viral binding and sequestration within the tumor. Lastly, the density of tumor capillaries is the highest in the tumor periphery, at the invasion front of the tumor, thus allowing greater vascular access to the injected virus.^21,26

Potential shortcomings of our study include our inability to distinguish infectious from noninfectious labeled virions. It has been demonstrated that plasma inactivates viral particles via complement and non-complement pathways of action.³¹ Plasma modified, non-infectious particles may have a significantly different distribution from infectious particles, pointing to the need for verification studies using histochemistry for the identification of virion components, active infection and marker gene expression studies. Studies to distinguish infectious from noninfectious particles may also become feasible with greater understanding of the molecular biology of viral binding and entry into the cell. Loss of co-localization of the virus and radioactive label can be expected to occur as time progresses. The virus and label have been demonstrated in vitro to be in stable association up to 12 h,¹¹ but it stands to reason that colocalization should be gradually lost in vivo during the processes of infection and/or phagocytosis. The timing of this loss of co-localization in vivo remains to be investigated, but the absence of rapidly excreted ¹¹¹In at 24 h suggests that labeled viral components are retained after infection. The current method is limited to starting (ie < 24 h) observation points. In a clinical setting, imaging of gene delivery following infection would likely be accomplished by imaging gene expression, rather than monitoring vector components. Multiple marker genes are under investigation for in vivo imaging. Optical candidates include luciferase,³² GFP,³³ protease activatable fluoroscopic dyes,³⁴ while MR techniques include engineered internalizing receptors³⁵ or tyrosinase.³⁶ PET agents include reporter probes such as FESP (an agonist for dopamine receptors)³⁷ and 18-FIAU or 18-F gancy- clovir/pencyclovir for imaging thymidine kinase.^38,39

The ideal gene therapy paradigm for brain tumors has yet to be evolved, but may consist of a combination of intratumor injection and intra-arterial administration of vectors bearing therapeutic transgenes. There are a number of alternative ways to improve the delivery of virus to brain tumors. Intra-arterial administration may benefit from tempering of the innate anti-viral activity of plasma by immune suppression.³¹ Alternatively, circulating antibodies could be saturated by the administration of viral glycoproteins before viral vector injection. Delivery vehicles, such as gels or migratory cells could be utilized to gradually release virus into the tumor bed or beyond. Chemical modification of viral particles might be undertaken to enhance delivery to target tissues, including uptake by tumor vasculature. The imaging modality defined in this paper will be an important quantitative component of evaluating and comparing delivery modalities in future gene therapy trials.

Materials and methods

Virus and cell culture

The virus used in this study was the genetically engineered HSV mutant, hrR3, derived from HSV-1 KOS, which has a disruption of the ICP6 gene, coding for one of the subunits for ribonucleotide reductase (RR), through insertion of the lacZ gene under the control of the immediate-early ICP6 promoter.⁴⁰ LacZ serves as a marker gene, facilitating detection of gene expression by microscopy during gene expression experiments. Viral stocks were generated in African green monkey kidney cell culture (Vero) and titered by plaque assays as described elsewhere.⁴¹ Sucrose step gradient centrifugation was used to purify the virus and to separate it from cellular fragments and protein contaminants.⁴²

The 9L rat gliosarcoma cell line, was cultivated in DMEM (Mediatech, Herndon, VA, USA) with 10% fetal calf serum and antibiotics. A 9L clone expressing the marker gene GFP (green fluorescent protein) was used for those tumor implants involving fluorescence microscopy.¹⁴

Labeling

Virus was radiolabeled using ¹¹¹In-oxine, prepared freshly 24 h before each labeling procedure. Incubation with the radiopharmaceutical and dialysis was performed in a manner similar to the radiolabeling procedure described elsewhere.¹¹ Briefly, ¹¹¹In-oxine was synthesized according to the method of McAfee,⁴³ and dissolved in a small quantity of DMSO. This solution was added to the desired quantity of virus and incubated for 30 min. Dialysis was then performed against 1 liter of Hank's balanced salt solution with a 10 kDa cutoff membrane for 2 h to separate unbound radionuclide. In the current experiments, 408 Ci/10⁹ plaque forming units (p.f.u.) initial loading activity yielded 290 Ci/10⁹ p.f.u. retained activity after dialysis, for a labeling efficiency of 71%. The labeled virus had previously been shown to retain infectivity.¹¹ Prior work has also demonstrated the stability of the virus-label association with 95.2% of initial activity still associated with the virus after 12 h of dialysis in one experiment.¹¹

In some experiments we also double labeled the virus by the addition of monofunctional CY 3 dye (Amersham Life Science, Pittsburgh, PA, USA) for fluorescence microscopy. Sufficient dye for 1 mg of protein (as specified by the manufacturer) was dissolved in 100 l of DMSO, 20 l of this solution was added to the incubation mixture together with the radionuclide, followed by dialysis as described previously.

Animal studies

Tumor implantation: A total of 17 tumor bearing Fisher rats (F-344, 150-200 g; Charles River Laboratories, Wilmington, MA, USA) were used for the current study. Three rats were used for intra-arterial injection without RMP-7 for 1 h time-points, three for intra-arterial injection with RMP-7 for 1 h time-points, two for intra-arterial injection for 24 h time-points, three for intratumoral injection of virus, two for histological studies and two for determination of the blood half-life. An additional two non-tumor-bearing animals were used as negative controls for the RMP-7 experiments. For the tumor implantation, animals were anesthetized and implanted stereotactically with a 10 l suspension of 4 ´ 10⁴ 9L gliosarcoma cells in the right frontal lobe using a stereotactic frame (David Kopf Instruments, Tujunga, CA, USA) over 5 min.⁴⁴ Tumors were permitted to grow for 10 days before virus vector application. Animals were maintained and treated according to institutional guidelines as approved by the Subcommittee on Animal Care. Two animals were inoculated with a GFP-expressing 9L cell line according to the same procedure.

Vector administration:

For intra-tumoral vector injection, animals (n = 3) were reanesthetized and placed in the stereotactic frame. Using the previously placed burr hole, 8 ´ 10⁶ p.f.u. (14.6 kBq or 0.4 Ci) of radiolabeled virus in 10 l DMEM was injected stereotactically over 5 min into the tumor according to an established protocol.⁴⁵

For intra-carotid HSV vector injection, animals were re-anesthetized, and the right external carotid artery (ECA) was catheterized essentially as previously described.⁴⁴ Labeled HSV vector (7 ´ 10⁷ p.f.u., 0.435 MBq or 11.8 Ci) in a volume of 100 l was injected into the internal carotid artery (ICA) over 1 min. A similar procedure was followed for animals (n = 2) receiving dual-labeled virus and inoculated with GFP-expressing 9L glioma cells, for histologic studies.

Those animals receiving RMP-7 (Cereport; Alkermes Inc, Cambridge, MA, USA) were injected at a total dose of 1.5 g/kg of drug. The dose was divided equally, with one half the dose injected through the intra-carotid catheter using a infusion pump over 5 min, followed by the administration of virus injected over 1 min, followed by the remaining drug over 5 min.

In vivo tracking

Anesthetized animals were injected with the radiolabeled virus under a gamma camera. Selected dynamic and static images were obtained immediately after viral injection up to 1 h after injection and at 24 h. Scintigraphic images were acquired by using a small field of view gamma camera equipped with a high resolution parallel collimator (Sigma 410, Ohio Nuclear, Solon, OH, USA) driven by a NucLear Mac (Scientific Imaging Inc, Littleton, CO, USA). Images were acquired by using 20% windows over the 247-keV photopeak of ¹¹¹In.

Biodistribution

Animals were killed by a lethal overdose of anesthetic (pentobarbital 200 mg/kg, i.p.). Samples of tumor, adjacent brain tissue, contralateral brain tissue, blood, spleen, liver, lung, heart, aorta, duodenum, stomach, kidney, muscle, fat, testis, thyroid, bone, and of both carotid arteries were excised, weighed and radioactivity measured using a well-type gamma counter (Compugamma; Wallac-LKB, Turku, Finland). The organ samples were counted together, with decay correction, and the dose in each sample was calculated. Biodistribution results were expressed as percentage of the injected dose (ID) per gram of tissue (% ID/g) corrected for radioactive decay. In a similar manner, results were also displayed as injected dose per organ (%ID/organ). Standard data⁴⁶ detailing the body mass percentage of various organs was used to calculate the whole organ masses for those organs not counted in entirety. Blood half-life determination was performed by repeated blood sampling over 60 min. Samples were weighed and counted with decay correction, and activities over time analyzed with statistical software (SyStat, SPSS Science, Chicago, IL, USA).

Histology

The tumor-bearing brains were removed and snap frozen in liquid nitrogen. Serial cryosections were obtained through the tumor (Frigocut; Leica Instrument GmbH, Heidelberg, Germany). Sections were performed in the following repeating sequence of thickness: 8 m, 8 m and 50 m. The first thin section (8 m) was fixed in 10% formalin for 10 min, rinsed in water for 5 min, then stained with hematoxylin-eosin (HE) according to the standard protocol for light microscopy. The second thin section (8 m) was used untreated for fluorescence microscopy using the Zeiss Axiovert 100 fluorescence microscope (Carl Zeiss North America, Thornwood, NY, USA) with an XF-32 filter (535/590 nm) for CY-3 dye and a XF-22 filter (485/530 nm) for green fluorescent protein (Omega Optical, Brattleboro, VT, USA). Dual channel images were obtained and merged using a PowerMac 7600/120 computer (Apple Computer, Cupertino, CA, USA) running IP laboratory spectrum software (Signal Analytic, Vienna, VA, USA). The thick 50 m section was used to obtain an autoradiograph using a 36 h exposure on a phosphor screen (Molecular Dynamics, Sunnyvale, CA, USA). The autoradiographic images were captured using a Power Macintosh 7600/120 computer running PhosphorImager SI (Molecular Dynamics).

Acknowledgements

The help of Dr Cliff Eskey in data analysis is gratefully acknowledged. Dr Anna Moore kindly provided the cell lines used. Ms Deborah Schuback prepared and titered the hrR3 viral stocks. Thanks to Dr Alex Bogdanov for helpful discussions. These studies were supported in part by grants PO1CA48729, PO1CA69246 and RO1NS35258. NGR was supported in part by FKZ 2794A/0087H from the State of Saxony-Anhalt.

1 Sokol DL, Gewirtz AM. Gene therapy: basic concepts and recent advances. Crit Rev Eukary Gene Exp 1996; 6: 29-57,

2 Roth JA, Cristiano RJ. Gene therapy for cancer: what have we done and where are we going? J Nat Cancer Inst 1997; 89: 21-39,

3 Dubowchik GM, Walker MA. Receptor-mediated and enzyme-dependent targeting of cytotoxic anticancer drugs. Pharmacol Ther 1999; 83: 67-123, MEDLINE

4 Stewart AK et al. Engraftment of gene-marked hematopoietic progenitors in myeloma patients after transplant of autologous long-term marrow cultures. Hum Gene Ther 1999; 10: 1953-1964, Article MEDLINE

5 Andreansky SS et al. The application of genetically engineered herpes simplex viruses to the treatment of experimental brain tumors. Proc Natl Acad Sci USA 1996; 93: 11313-11318, Article MEDLINE

6 Bilbao G et al. Adenoviral/retroviral vector chimeras: a novel strategy to achieve high- efficiency stable transduction in vivo. FASEB J 1997; 11: 624-634, MEDLINE

7 Jacoby DR, Fraefel C, Breakefield XO. Hybrid vectors: a new generation of virus-based vectors designed to control the cellular fate of delivered genes (editorial). Gene Therapy 1997; 4: 1281-1283, MEDLINE

8 Bilbao G, Gomez-Navarro J, Curiel DT. Targeted adenoviral vectors for cancer gene therapy. Adv Exp Med Biol 1998; 451: 365-374,

9 Pechan PA et al. A novel 'piggyback' packaging system for herpes simplex virus amplicon vectors. Hum Gene Ther 1996; 7: 2003-2013,

10 Johnston KM et al. HSV/AAV hybrid amplicon vectors extend transgene expression in human glioma cells. Hum Gene Ther 1997; 8: 359-370, MEDLINE

11 Schellingerhout D et al. Mapping the in vivo distribution of herpes simplex virions. Hum Gene Ther 1998; 9: 1543-1549, MEDLINE

12 Zinn KR et al. Imaging and tissue biodistribution of 99mTc-labeled adenovirus knob (serotype 5). Gene Therapy 1998; 5: 798-808, MEDLINE

13 Bartus RT et al. Controlled modulation of BBB permeability using the bradykinin agonist, RMP-7. Exp Neurol 1996; 142: 14-28,

14 Moore A et al. Novel gliosarcoma cell line expressing green fluorescent protein: a model for quantitative assessment of angiogenesis. Microvasc Res 1998; 56: 145-153,

15 Bashir R, Hochberg F, Oot R. Regrowth patterns of glioblastoma multiforme related to planning of interstitial brachytherapy radiation fields. Neurosurgery 1988; 23: 27-30,

16 Kramm CM et al. Gene therapy for brain tumors. Brain Pathol 1995; 5: 345-381, MEDLINE

17 Breakefield XO et al. Herpes simplex virus vectors for tumor therapy. In: Sobel RE, Scanlon KJ (eds). The Internet Book of Gene Therapy: Cancer Gene Therapeutics. Appleton and Lange: Stanford, 1995, 41-59.

18 Hunter W, Rabkin S, Martuza R. Brain tumor therapy using genetically engineered replication-competent virus. In: Kaplitt MG, Leowy AD (eds). Viral Vectors: Gene therapy and Neuroscience Applications. Academic Press: San Diego, 1995, 260-274.

19 Rainov NG et al. Intra-arterial virus and nonvirus vector-mediated gene transfer to experimental rat brain tumors. Front Radiat Ther Oncol 1999; 33: 227-240,

20 Hoshino T. A commentary on the biology and growth kinetics of low-grade and high-grade gliomas. J Neurosurg 1984; 61: 895-900,

21 Jain RK. Delivery of molecular medicine to solid tumors. Science 1996; 271: 1079-1080, MEDLINE

22 Barnett FH et al. Selective delivery of herpes virus vectors to ex-perimental brain tumors using RMP-7. Cancer Gene Ther 1999; 6: 14-20, MEDLINE

23 Bartus RT et al. Permeability of the blood-brain barrier by the bradykinin agonist, RMP-7: evidence for a sensitive, autoregulated, receptor-mediated system. Immunopharmacology 1996; 33: 270-278,

24 Black KL et al. Intracarotid infusion of RMP-7, a bradykinin analog, and transport of gallium-68 ethylenediamine tetraacetic acid into human gliomas. J Neurosurg 1997; 86: 603-609, MEDLINE

25 Thomas GR, Thiemermann C, Walder C, Vane JR. The effects of endothelium-dependent vasodilators on cardiac output and their distribution in the anaesthetized rat: a comparison with sodium nitroprusside. Br J Pharmacol 1988; 95: 986-992,

26 Jain RK. 1995 Whitaker Lecture: delivery of molecules, particles, and cells to solid tumors. Ann Biomed Eng 1996; 24: 457-473,

27 Yuan F et al. Time-dependent vascular regression and permeability changes in established human tumor xenografts induced by an anti-vascular endothelial growth factor/vascular permeability factor antibody. Proc Nat Acad Sci USA 1996; 93: 14765-14770, MEDLINE

28 Helmlinger G, Yuan F, Dellian M, Jain RK. Interstitial pH and pO₂ gradients in solid tumors in vivo: high-resolution measurements reveal a lack of correlation. Nature Med 1997; 3: 177-182, MEDLINE

29 Wickham TJ, Haskard D, Segal D, Kovesdi I. Targeting endothelium for gene therapy via receptors up-regulated during angiogenesis and inflammation. Cancer Immunol Immunother 1997; 45: 149-151, MEDLINE

30 McDonald DM, Munn L, Jain RK. Vasculogenic mimicry: how convincing, how novel, and how significant? Am J Pathol 2000; 156: 383-388, MEDLINE

31 Ikeda K et al. Oncolytic virus therapy of multiple tumors in the brain requires suppression of innate and elicited antiviral responses. Nature Med 1999; 5: 881-887, Article MEDLINE

32 Mayerhofer R, Araki K, Szalay AA. Monitoring of spatial expression of firefly luciferase in transformed zebrafish. J Biolumin Chemilumin 1995; 10: 271-275,

33 Aboody-Guterman KS et al. Green fluorescent protein as a reporter for retrovirus and helper virus-free HSV-1 amplicon vector-mediated gene transfer into neural cells in culture and in vivo. Neuroreport 1997; 8: 3801-3808,

34 Weissleder R, Tung CH, Mahmood U, Bogdanov A Jr. In vivo imaging of tumors with protease-activated near-infrared fluorescent probes. Nature Biotechnol 1999; 17: 375-378, Article MEDLINE

35 Weissleder R et al. In vivo magnetic resonance imaging of transgene expression. Nature Med 2000; 6: 351-355, Article MEDLINE

36 Weissleder R et al. MR imaging and scintigraphy of gene expression through melanin induction. Radiology 1997; 204: 425-429, MEDLINE

37 MacLaren DC et al. Repetitive, non-invasive imaging of the dopamine D2 receptor as a reporter gene in living animals. Gene Ther 1999; 6: 785-791, MEDLINE

38 Tjuvajev JG et al. Imaging the expression of transfected genes in vivo. Cancer Res 1995; 55: 6126-6132,

39 Blasberg RG, Tjuvajev JG. Herpes simplex virus thymidine kinase as a marker/reporter gene for PET imaging of gene therapy. Quart J Nucl Med 1999; 43: 163-169,

40 Goldstein DJ, Weller SK. Herpes simplex type 1-induced ribonucleotide reductase is dis-pensable for virus growth and DNA synthesis: isolation and characterization of an ICP6 lacZ insertion mutant. J Virol 1988; 62: 196-205, MEDLINE

41 Roizman B, Spear PG. Preparation of herpes simplex virus of high titer. J Virol 1968; 2: 83-84, MEDLINE

42 Matis J, Lesso J, Mucha V, Matisova E. Purification and separation of enveloped and unenveloped herpes simplex virus particles. Acta Virol 1975; 19: 273-280,

43 McAfee JG. Techniques of harvesting platelets and neutrophils and labeling them with In-111 oxine. In: Thakur Matthew LGA (ed). Radiolabeled Cellular Blood Elements - Current Accomplishments, Immediate Potential, and Future Possibilities. Plenum: New York City, New York, 1979, 51-57.

44 Rainov NG et al. Selective uptake of viral and monocrystalline particles delivered intra-arterially to experimental brain neoplasms. Hum Gene Ther 1995; 6: 1543-1552, MEDLINE

45 Boviatsis EJ et al. Gene transfer into experimental brain tumors mediated by adenovirus, herpes simplex virus, and retrovirus vectors. Hum Gene Ther 1994; 5: 183-191, MEDLINE

46 Lang PL, White WJ. Growth, development, and survival of the Crl:CD(SD)BR stock and CDF(F344)/CrlBR strain. In: Mohr U, Dungworth D, Capen C (eds). The Pathobiology of the Aging Rat. International Life Sciences Institute: Washington, 1994 (vol 2, 131-139.

Figure 1 Microscopic distribution. (a) Photomicrograph of a rat brain with tumor, stained with hematoxylin-eosin (HE). (b) Autoradiograph of the section adjacent to (a), indicating viral accumulation in the tumor. (c) Higher power view of the HE stain photomicrograph. (d) Dual channel fluorescence microscopy of the brain tumor interface with red representing labeled HSV and green the brain tumor expressing GFP (´200).

Figure 2 In vivo imaging of HSV distribution. Nuclear image of a whole F-344 rat obtained at 1 h after ipsilateral intra-carotid administration of labeled HSV. Note uptake in the glioma (arrow), and a clear difference between the left and right hemispheres. Also note uptake in the lungs and liver.

Table 1 Bio-distribution of ¹¹¹In-oxine-labeled HSV expressed as % injected dose/gram tissue

Table 2 Bio-distribution of ¹¹¹In-oxine labeled HSV expressed as % injected dose/organ

Table 3 Comparative viral concentration in various tissues, by %ID/gram tissue, with ratios scaled to normal brain (normal brain = 1)

Table 4 Comparative viral concentration in various tissues, by %ID/organ, with ratios scaled to normal brain (normal brain = 1)