| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
1 Departments of Cell Biology, 2 Medicine, 3 Biochemistry and Molecular Genetics, and 4 Center for Aids Research Core, University of Alabama at Birmingham, Birmingham, Alabama; 5 Serologicals Corporation, Gaithersburg, Maryland; and 6 Southern Research Institute, Birmingham, Alabama
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The present experiments focused on in vivo activity of a less well studied E. coli PNP substrate, 9-(ß-D-arabinofuranosyl)-2-fluoroadenine (F-araA). E. coli PNP converts F-araA to 2-F-adenine (F-Ade) with a catalytic efficiency that is approximately 1,000-fold less than the prototypic E. coli PNP substrate, 9-(2-deoxy-ß-D-ribofuranosyl)-6-methylpurine (MeP-dR; ref. 2 , 11 ). The clinically used form of F-araA (i.e., fludarabine phosphate, F-araAMP) is rapidly converted in the blood to F-araA. F-araAMP is routinely used in the treatment of chronic lymphocytic leukemia but has no therapeutic role in the treatment of solid tumors such as gliomas.
Fludarabine was selected in these experiments for a number of reasons. First, the compound (without E. coli PNP) has been studied extensively, and pharmacokinetics of the agent in animal models are well defined. Second, a convenient source of fludarabine is available (the drug can be purchased from a hospital pharmacy). Because there is no readily available source, MeP-dR has required chemical synthesis in our laboratories in the past for in vivo studies. A commercial supply of GMP-grade E. coli PNP substrate would be expected to facilitate preclinical testing in animal models. Third, a safety profile already exists for fludarabine in human subjects, a feature that would be expected to promote testing of E. coli PNP in the clinic. Finally, early in vivo experiments suggested that fludarabine might be a useful compound for mediating tumor regressions with E. coli PNP (ref. 11 , see also below).
Little information is available concerning the in vivo limits to bystander killing with E. coli PNP and F-araAMP or the safety of F-araAMP in this setting. Measurements of intratumoral PNP activity after in vivo delivery of E. coli PNP to tumors and studies of PNP transgene expression in non-target tissue such as liver after viral delivery have not been conducted previously. We therefore performed a series of experiments to address the following questions. First, how effectively can F-araAMP mediate antitumor effects in vivo after vector-based delivery of E. coli PNP, and how potent is bystander killing by this approach? We examined intratumoral levels of E. coli PNP, ectopic (intrahepatic) expression, and the dose dependence of these effects. Second, how well do intermediate end points (intratumoral levels of E. coli PNP enzyme, accumulation of radiolabeled F-araA metabolites, and intratumoral enzyme distribution) predict tumor regressions and cures with E. coli PNP/F-araAMP. Because in vivo bystander killing with herpes simplex virus thymidine kinase, cytosine deaminase, and other suicide gene products has been difficult to standardize with different vectoring strategies, we designed the present analysis to evaluate antitumor efficacy using three independent expression constructs (adenovirus, MuLv, and lentivirus). The findings indicate strong bystander killing and tumor regressions in a fashion that exhibits in vivo dose dependence on both intratumoral levels of E. coli PNP and the dose of F-araAMP administered.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Measurement of E. coli PNP Activity.
Crude cell extracts were prepared as described previously (9
, 10) from D54MG (human glioma) cells transduced with the E. coli PNP. The extracts were incubated with various concentrations of MeP-dR, and the formation of product was measured by HPLC analysis of the reaction mixture (16)
. Activity was expressed as PNP units; one unit represents 1 nmol MeP-dR-converted/mg tumor cell extract/hour.
Intratumoral Inoculation of Recombinant Adenovirus and EGFP Detection.
Ad-PNP or Ad-EGFP were inoculated directly into tumors [0.25–2 billion plaque forming units (PFU) in 50 µL, 27-gauge needle] along four separate needle tracks. Tumors were approximately 250 mg at the time of inoculation. A subset of animals inoculated with Ad-EGFP were sacrificed, the tumors were embedded in OCT media (Fisher Scientific, Hampton, NH), frozen in liquid nitrogen, and sectioned. Tumor sections were studied by two independent protocols. In the first, a Leitz Epifluorescent microscope was used to monitor the expression of EGFP based on green fluorescence of the reporter gene. In the second, tumor sections were fixed in 4% buffered formaldehyde at room temperature for 20 minutes and washed with PBS (pH 7.4) three times. After blocking of the nonspecific protein-binding sites with goat serum (diluted 1:20 in PBS) for 30 minutes, monoclonal antibody to green fluorescent protein (Roche Diagnostics, Indianapolis, IN) was added at a dilution of 1:100. After 1 hour of incubation and subsequent washing steps, a secondary antibody (antimouse AlexaFluor 586, Molecular Probes, Eugene, OR) was added at 1:400 dilution and incubated for 1 hour. The sections were then washed three times, and a Leitz Epifluorescent microscope with a Photometrics SenSys digital camera and IPLab Spectrum software (Scanalytics, Fairfax, VA) was used for visualization.
Measurement of Tumor Regressions.
Human glioma (D54MG) tumor cells, with or without stably-transduced E. coli PNP were selected because (a) the tumor is quite slow growing in vivo (doubling time approximately 15 days), making this a stringent model for examining bystander killing of resting phase tumor cells; (b) adenovirus, herpes virus, and other viruses have been used in the past to select human gliomas as a clinical target for suicide gene therapy; and (c) low level PNP expressing cells described previously were available for comparison (11)
. Tumor cells (2 x 107) were injected subcutaneously into the flanks of nude mice (nu/nu) purchased from Charles River Laboratories (Wilmington, MA). Tumors were measured with calipers, and an estimate of weight (mg) was calculated as described previously (11)
. Mice were evaluated for weight loss, tumor mass, and overall appearance twice weekly. MeP-dR was made in our laboratories according to the method described by Montgomery and Rosman (17)
. F-araAMP was obtained from Schering A.-G. (Berlin, Germany). All animal tumor regression studies were conducted according to NCI standard procedures. Each depicted curve represents the median tumor size from six animals, and each tumor regression study has been confirmed.
F-Ade Injection into Pre-established Tumors.
Susceptibility to F-Ade was determined in tumors implanted subcutaneously in the flanks of mice as described above. Tumors of approximately 300 mg were injected intratumorally in a delivery volume of 100 µL.
Intratumoral Measurements of F-araAMP and Metabolites.
D54MG parental or D54MG tumor cells expressing E. coli PNP were injected subcutaneously into the flanks of nu/nu mice, as described above. When the tumors were 300 to 400 mg, the mice received injection intraperitoneally with 100 mg/kg (1.4 Ci/mol) [3H]F-araAMP (Moravek Biochemicals Inc., Brea, CA), and the radioactivity in tumors was determined 4 hours later as described previously (1
, 10)
. We have shown previously (10)
that total radioactivity 4 hours after injection is a reliable indicator of F-Ade metabolites in tumor tissues in vivo and that radioactive compounds recovered from tumors under these conditions represent F-araA metabolites. The labeled compounds in tumor extracts exhibit both specificity and dose dependence on the level of intratumoral E. coli PNP. Further efforts to analyze the tumor extracts did not reveal any more information than was indicated by the HPLC assay as originally described (10)
, because of degradation of relevant compounds during the extraction process.
For experiments in which adenovirus was used, labeled compound was injected 48 hours after viral administration. The tumors were dissolved in 1 ml soluene 350 (Packard Instrument Company, Meriden, CT) by incubating at 55°C for 4 hours and then at room temperature until dissolved. A portion of each extract was mixed with scintillation fluid, and the radioactivity was determined. Four tumors were removed for each measurement.
Production of Recombinant Lentivirus Expressing PNP.
Plasmids generously provided by the Trono laboratory (Geneva, Switzerland) were used to perform lentivirus construction according to the method of Trono et al. (18)
. To establish a lentivirus capable of PNP expression, the gene was PCR amplified by primers 5'-ggatccaccatggctaccccacacattaatg-3' (BamHI site and ATG underlined) and 5'-cctcgagtcactctttatcgcccagcag-3' (XhoI site underlined). The resulting product was subcloned into Zero-Blunt vector (Invitrogen, Carlsbad, CA). After digestion with BamHI and XhoI, the luciferase gene in the pHR'CMVLuc W Sin-18 lentivirus vector was replaced with E. coli PNP. Correct insertion was verified (a) by sequencing the entire PNP gene and the ligation sites; (b) by transfecting the resulting plasmid, which encodes a CMV promoter-driven PNP gene, into 293T cells; and (c) by verifying E. coli PNP enzymatic activity by HPLC in vitro (see above). A total of 40 µg of plasmid DNA was used for calcium-phosphate transfection of one 10-cm dish. The DNA mixture contained 5 µg of envelope-coding plasmid pMD.G; 15 µg of the packaging plasmid pCMVDR8.91, which expresses Gag, Pol, Tat, and Rev; and 20 µg of transfer vector plasmid. Replication-deficient viral particles encoding E. coli PNP were collected from tissue culture supernatant after transfection, and a Beckman SW-28 rotor sucrose cushion centrifugation was used to concentrate lentivirus stock 1,000-fold with one round of centrifugation at 26,000 RPM for 90 minutes at 8°C. After 2 hours of incubation on ice, the virus was resuspended into serum-free medium. Titer of lentivirus reaches approximately 1 x 109 infectious particles/ml following concentration by this method, and replication-deficient viral particles encoding E. coli PNP were collected from tissue culture supernatant after transfection.
Recombinant lentivirus encoding PNP was used to generate clonal D54MG human glioma cells with high level transgene expression. Briefly, 5 x 104 D54MG cells were seeded into individual wells of 6-well plates and infected with recombinant lentivirus. Clonal cells were obtained by limit-diluting the pool of infected cells into 96-well plates. Cell clones were tested and confirmed for E. coli PNP activity both by HPLC and a conventional cell-killing assay in the presence of MeP-dR in vitro.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Ad-PNP conferred dose-dependent PNP expression in human glioma tumors after direct intratumoral inoculation of the virus. Injection of adenovirus through needle tracks also delivered PNP to other (extratumoral) tissues, as indicated by E. coli PNP measurements in liver. To examine antitumor effects conferred by E. coli PNP, F-araAMP was administered systemically to animals bearing tumors expressing recombinant enzyme. The results shown in Fig. 1B
indicate that tumor growth slows significantly because of E. coli PNP expression in solid tumors followed by prodrug dosing.
|
. Negligible evidence of expression was observed outside areas of direct inoculation. Overall, <10% of the tumor cells expressed the reporter gene. Tumors treated with control (no EGFP) adenovirus demonstrated no fluorescence signal (data not shown).
|
and 2
indicate F-araAMP-dependent killing of tumor cells not expressing E. coli PNP (bystander killing). To further delineate limits of bystander killing with F-araA and to test specificity of these results, we studied stably-transduced D54MG tumors. Figure 3
describes substantial activity of fludarabine against MuLv-transduced tumors, in which E. coli PNP is expressed from an SV40-driven promoter (11)
. We next expressed E. coli PNP, using a third, independent (lentiviral) vector system. Lentivirus confers a very high level expression in epithelia compared with first-generation adenoviral constructs, as well as those derived from MuLv. In Fig. 4A
, the lentiviral expression system is used to depict a study in which 100% of cells express E. coli PNP at high levels (126,000 units). Complete regression and cures of tumors (12 of 12) were obtained with F-araAMP under these conditions. Tumors established from lower proportions of PNP-expressing cells (10%, 5%, 2.5%) exhibited dose dependence on both the amount of prodrug added (Fig. 4B)
and the intratumoral PNP activity (Fig. 4C)
. Antitumor effects were observed with F-araAMP when as few as 2.5% of tumor cells in vivo expressed PNP (average PNP activity 
3600 units).
|
|
. As shown in Fig. 4D
, treatment by direct intratumoral inoculation of F-Ade failed to slow tumor progression at a maximally tolerated dose. Accumulation of F-araA metabolites in glioma tumors transduced with E. coli PNP is shown in Table 1
.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
One goal of the present experiments was to evaluate a recombinant virus and examine the efficiency of E. coli PNP transduction. The results demonstrate that adenovirus confers dose-dependent expression of functional enzyme in vivo (Fig. 1)
. The data depicted in Figs. 1
and 2
indicates that adenovirus encoding E. coli PNP can elicit substantial tumor cell killing and slow tumor growth after F-araAMP administration, despite the expected tendency to express transgene primarily along inoculation sites. Ectopic expression of E. coli PNP (e.g., adenoviral infection of liver) was well tolerated after systemic dosing of F-araAMP in this setting. The data in Fig. 1
therefore points to both safety and substantial bystander killing by E. coli PNP/F-araAMP.
Antitumor effects with E. coli PNP exhibit dependence on both the level of prodrug administered (Fig. 4B)
and the total amount of PNP activity present within solid tumors (Fig. 4C)
. As few as 2.5% of tumor cells expressing the PNP gene at high levels led to strong tumor responses. On the basis of adeno- and lentiviral studies in D54MG tumors, a transfection rate in which 2–5% of cells express PNP at sufficient levels of activity (e.g. 2,000–3,000 units) can be considered a benchmark for impairing tumor growth. The lentivirus protocol used at higher levels of expression led to tumor cures (for example, when 5–10% of cells express the transgene, Fig. 4C
). Intratumoral F-araA metabolites above 400 nmol/g tumor tissue were also associated with potent tumor regressions and cures (Table 1)
. The data in Table 1
demonstrates specificity and dose dependence on intratumoral levels of E. coli PNP expression, and describes a useful end point for monitoring PNP therapy in vivo.
Transduction of D54MG cells with a first-generation retrovirus (MuLv) led to E. coli PNP expression levels approximately 500-fold less than the levels observed with lentivirus. In tumors established from MuLv and comprised of 100% PNP expressing cells (i.e., no bystander killing required), substantially lower PNP levels than those described above (and correspondingly less toxin accumulation) effectively impaired tumor growth (Fig. 3)
. This is likely because of more potent killing of PNP-transduced (compared with bystander) cells, as we have reported earlier for other E. coli PNP substrates (ref. 6
).7
Adenovirus and lentivirus (3,000–8,000 PNP units, 2–10% transduced cells) were used to compare tumors with similar levels of PNP activity and percentages of cells transduced and similar delays in tumor growth were observed (Figs. 1
and 4C
).
Fludarabine has been reported to confer antitumor activity in an earlier study of E. coli PNP expression in hepatic tumors (23)
and in glioma tumors transduced ex vivo (11)
. The results (Fig. 1
, 3
, and 4)
establish that toxic compounds such as F-Ade can be liberated in tumor tissues by the action of PNP and cure malignancies without untoward toxicity to the host. Tumor regressions with F-Ade cannot be attained simply by intratumoral injection (Fig. 4D)
, presumably because of escape of the compound to the systemic circulation, leakage onto skin at the inoculation site, limited distribution of F-Ade along needle tracks, or some combination of these factors. Intracranial tumors such as glioma are difficult to access for repeat administration of anticancer agents in vivo. At least insofar as a single administration is concerned, the results in Fig. 4
support F-araA activation by E. coli PNP (rather than F-Ade injection) in glioma and other deep parenchymal tumors not well suited for repetitive instillations in human subjects in vivo.
Our studies in mice indicate (a) continued weight gain and normal growth curves in treated animals with otherwise substantial tumor burdens; (b) absence of lethality with any of three vector systems (MuLv, lentivirus, or adenovirus) used to deliver E. coli PNP; and (c) normal activity levels, grooming characteristics, and so forth in large groups of mice treated and/or cured of their tumors. If toxicity because of expression of E. coli PNP occurs, it has not limited the ability to obtain durable antitumor responses without evidence of untoward sequelae by these criteria. However, it should be noted that fludarabine and metabolites, like most clinically approved anticancer agents, cause bone marrow suppression at standard doses. Adenoviral vectors can elicit hepatocellular injury in vivo. Formal toxicology will therefore be necessary before applying the E. coli PNP strategy to cancer therapy in human subjects.
In conclusion, this study establishes a scientific foundation for understanding and optimizing tumor cell bystander killing with E. coli PNP, including variables that help predict tumor regressions and cures. The data demonstrate antitumor activity with three different vectoring systems and point to new use for a clinically approved chemotherapy (fludarabine) in solid tumors. Bystander killing of the magnitude described here has been difficult to establish for any suicide gene in the past. A variety of emerging gene transfer vehicles (e.g., replicating adenovirus, herpes virus, retrovirus, and measles virus) may be capable of transducing 2 to 10% of tumor cells in vivo. The data presented here provide a new means for understanding bystander killing with E. coli PNP and F-araAMP and for applying E. coli PNP to systems such as these in the future.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Eric J. Sorscher, 1918 University Boulevard (798 MCLM), Birmingham, AL 35294-0005. Phone: 205-934-9640; Fax: 205-934-7593; E-mail: sorscher{at}uab.edu
7 E. Sorscher and W. Parker, unpublished observations. 
Received 1/ 5/04. Revised 7/ 8/04. Accepted 7/19/04.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Afshar, T. Asai, and S. L. Morrison Humanized ADEPT comprised of an engineered human purine nucleoside phosphorylase and a tumor targeting peptide for treatment of cancer Mol. Cancer Ther., January 1, 2009; 8(1): 185 - 193. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
) or saline (
). Ad-PNP together with F-araAMP treatment conferred slowing of tumor growth (
). Effects on tumor growth were observed in these studies without excessive weight loss (<5%) or other undesired sequelae. F-araAMP therapy was administered 48 hours after viral inoculation.
