| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Experimental Therapeutics |
Division of Human Gene Therapy, Departments of Medicine, Pathology, and Surgery, and the Gene Therapy Center, University of Alabama at Birmingham, Birmingham, Alabama 35294-2172
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
24 and AdTyr224, by replacing the adenoviral E1A promoter with a cassette containing a polyA sequence and a human tyrosinase enhancer/promoter construct (hTyr2E/P). The small size of this cassette allows retention of the E3 region within these CRAds, which was shown to enhance viral spread and oncolysis. In addition, we introduced mutations (24 and 224) into the viral E1A gene, which attenuate adenoviral replication in quiescent cells. The cell cycle pathways mediating this attenuation are defective in melanoma cells. By analysis of E1A expression, we prove fidelity of hTyr2E/P in the adenoviral genome and in the context of viral replication when an upstream polyA was included. We further show efficient cytotoxicity of AdTyr24 and AdTyr224 in melanoma cell lines and a 100-1000-fold attenuation in cell lines derived from various nonmelanocytic tissues. Virus replication and progeny production of these viruses were similarly selective, resulting in 200–800-fold higher virus yields in melanoma cells versus control cells, thus establishing viral cytolysis and spread as the cause of the observed cell killing. Cytotoxicity of AdTyr24 for normal fibroblasts and keratinocytes was strongly attenuated, and this virus caused selective killing of melanoma cells but not surrounding keratinocytes in a coculture system. Progeny production and cytotoxicity of AdTyr24 in melanoma cells were similar to matching viruses containing the stronger cytomegalovirus enhancer/promoter instead of hTyr2E/P. Furthermore, AdTyr24 showed a cytopathic effect similar to the wild-type E1A containing AdTyrwt and only minimally reduced compared with wild-type adenovirus. We conclude that the generated CRAds AdTyr24 and AdTyr224 constitute novel targeted agents for gene therapy and viral oncolysis of metastatic melanoma. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Previously, gene therapy has been successfully applied for killing of tumor cells in vitro and for treatment of cancer in animal models (4) . Adenoviruses are widely used as gene transfer vectors, because they possess the critical basic properties required for this application (5) . These include their highly evolved gene transfer machinery, the stability of virus particles, and the ease of virus production at high titers. For safety reasons, first-generation adenoviral vectors were rendered replication deficient by deleting essential viral genes, such as those of the E1 region. However, clinical cancer gene therapy trials have suffered from limited gene transfer efficiencies, even when adenoviruses were applied. Indeed, the recognition of gene transfer limitations as a critical problem for clinical gene therapy has led to the development of replication-competent viral vectors (6) . As such, CRAds are considered particularly promising (7 , 8) , because adenoviruses are of low pathogenicity and their above-mentioned properties are similarly favorable in the context of replicating vectors. In addition to the amplification effect of CRAds for therapeutic gene transfer, the intrinsic adenoviral ability to kill cells provides for a remarkable therapeutic mechanism in itself, i.e., viral oncolysis.
Specificity of viral replication is a fundamental requirement for the concept of CRAds and likewise for other replicating viruses. Particularly for systemic applications, as required for treatment of metastatic melanoma, detrimental side effects caused by viral replication and spread in healthy tissues must be avoided without sacrificing replicative efficacy in target cells.
In this study, we describe a strategy to generate CRAds as novel therapeutic agents for treatment of melanoma by combining transcriptional targeting with attenuating mutations of the viral E1A gene. CRAds based on specific promoter activity were described recently (for references, see Ref. 8 ). In these studies, promoters were incorporated into the adenoviral genome to drive targeted expression of essential viral genes, principally E1A. We generated novel viruses that incorporate an optimized transcription regulatory sequence of the melanoma differentiation marker tyrosinase, hTyr2E/P, containing a tandem enhancer fused to the core promoter (9) along with an upstream polyA transcription termination signal to drive expression of E1A deletion mutants. These mutants were reported to restrict adenoviral replication to cells with inactive pRb (10) . Mutations within the pRb pathway are frequent in melanomas (11) , but also normal cells inactivate pRb during proliferation (12) . Herein, we demonstrate fidelity of hTyr2E/P in the context of CRAds resulting in melanoma-specific expression of E1A mutants. Furthermore, we show that melanoma cells are more sensitive to lysis by tyrosinase enhancer/promoter-controlled CRAds relative to both cell lines and nontumorigenic primary cells derived from several other tissues. Moreover, these CRAds selectively kill melanoma cells in cocultures with normal keratinocytes. We also confirm that specific cytotoxicity of the engineered CRAds correlates with selective viral replication. Finally, we demonstrate that deletion of the pRb-binding ability of hTyr2E/P-driven E1A in our CRAds does not compromise its potency to kill melanoma. Thus, we developed oncolytic adenoviruses as a novel strategy for targeted melanoma therapy.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Recombinant Adenoviruses.
Adenoviruses genomes were constructed as follows. Plasmids pScswt, pScs24, and pScs224 were generated by introducing into the corresponding restriction sites of pShuttle (13)
a fragment containing a XhoI restriction site followed by the Ad5 genomic sequence from nucleotide 522 to the MfeI restriction site at position 3924 (nucleotide positions refer to GenBank sequence AD5001). For pScs24 Ad5 nucleotides 923–946 (corresponding to amino acids LTCHEAGF of E1A, see also Ref. 10
) and for pScs224 additionally nucleotides 563–565 (second amino acid of E1A, by site-directed mutagenesis with the Transformer kit from Clontech, Palo Alto, CA using the oligonucleotide 5'-GACTGAAAATGCATATTATCTG) were deleted. To generate pSCMV24, pSCMV224, pSTyrwt, pSTyr24, and pSTyr224, the CMV promoter/enhancer or the synthetic tyrosinase promoter/enhancer, both including an upstream synthetic polyA transcription termination sequence (derived from pGL3CMV and pGL3Tyr; Ref. 14
), were introduced into the corresponding pScs plasmids. These shuttle vectors were then linearized by PmeI digestion and used for homologous recombination with plasmid pTG3602 (obtained from Transgene, Strasbourg, France) as described (13)
to generate plasmids pAdCMV24, pAdCMV224, pAdTyrwt, pAdTyrD24, and pAdTyr224 containing the recombinant adenovirus genomes. Plasmids were validated by PCR and restriction digest. Adenovirus particles were produced by transfection of PacI-digested pAd plasmids into MeWo cells using Lipofectamine (Life Technologies, Inc., Rockville, MD) following the manufacturer’s protocol. Viruses were amplified in MeWo cells and purified by two rounds of CsCl density gradient ultracentrifugation. Verification of viral genomes and exclusion of wild-type contamination were performed by PCR and restriction digest. For PCR, the following oligonucleotides were used: 5'-CMV, 5'-GTCTATATAAGCAGAGCTCTCTGG; 5'-Tyr, 5'-GAATTAAACTATTAATGGTGAATAG; 5'-ITR, 5'-CGGGAAAACTGAATAAGAGGAAGTGA; and 3'-E1, 5'-ATTTTCACTTACTGTAGACAAACAT. To test promoter identity, primers 5'-CMV or 5'-Tyr were used with primer 3'-E1, and to test Ad5 wild-type contamination and E1A mutants, primers 5'-ITR and 3'-E1 were applied. PCR was performed with 4 min at 95°C, followed by 40 cycles of 95°C for 1 min, 55°C for 1 min, 72°C for 1.5 min, and final extension at 72°C for 5 min. For testing of E1A mutants, PCR products were purified with the QIAquick PCR purification kit (Qiagen, Valentis, CA) and subsequently digested with restriction enzymes NsiI (restriction site generated by 2 mutation) or BstXI (restriction site deleted by 24 mutation). Replication-deficient, E1-deleted AdCMVLuc (15)
, wild-type adenovirus (Adwt), and adenovirus containing the 24 mutation (Ad24; Ref. 10
) were amplified in 293 (AdCMVLuc, Adwt) or A549 (Ad24) cells and purified as described. Physical particle concentration (viral particles/ml) was determined by OD260 reading and biological particle concentration (plaque-forming units/ml) was determined by standard plaque assay on 293 cells.
Western Blot.
To determine cellular E1A expression, 1 x 105 SK-MEL-28 or SKOV3.ip1 cells were plated in 24-well plates and infected with the indicated viruses in 200 µl of growth medium containing 2% FBS at a MOI of 500 or mock infected. The infection medium was replaced by growth medium containing 10% FCS after overnight infection. Cells were lysed in SDS sample buffer (containing 10 mM ß-mercaptoethanol) 33 h after infection. Boiled samples were separated by SDS-PAGE in 12.5% gels and transferred to a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA). The membrane was probed with M73 monoclonal antibody specific for E1A (Oncogene Research Products, Cambridge, MA). Bound M73 was detected with a secondary horseradish peroxidase-conjugated antibody (Sigma Chemical Co., St. Louis, MO) and enhanced chemiluminescence (NEN Life Science Products, Boston, MA).
Cytotoxicity Assay.
For determination of virus-mediated cytotoxicity, 1.5 x 104 tumor cells or 3 x 104 primary cells were plated in 24-well plates and infected with adenoviruses in 200 µl of growth medium containing 2% FBS (serum-free medium for keratinocytes) at indicated MOIs or mock infected. The infection medium was replaced with growth medium the next day. When cell lysis was observed for AdCMV viruses at MOI 1 (melanoma cells; SK-MEL-28 at day 14, Mel888 at day 7, and Mel624 at day 10 after infection) or MOI 0.1 (control cells; SKOV3.ip1 at day 18, A549 at day 7, MCF-7 at day 17, NHF at day 10, and NHKs at day 9 after infection), cells were fixed and stained with 1% crystal violet in 70% ethanol for 45 min, followed by washing with tap water to remove excess color. The plates were dried, and images were captured with a Kodak DC260 digital camera (Eastman Kodak, Rochester, NY). All experiments were performed in duplicate wells.
Virus Progeny Production Assay.
To determine virus progeny production, 1.5 x 104 cells were plated in 24-well plates and infected in 200 µl of growth medium containing 2% FBS at MOI 1. The next day, cells were washed three times with HBSS (Mediatech) and further incubated with growth medium containing 10% FBS. Three days after infection, cells and medium were harvested and freeze/thawed three times, and virus titers were determined by standard plaque assay on 293 cells. Experiments were performed in triplicates.
Melanoma Cell/Keratinocyte Cocultures.
For melanoma cell/keratinocyte cocultures, a mixture of 5 x 103 SK-MEL-28 cells and 1 x 104 NHKs was plated in 24-well plates on coverslips. Cells were infected with adenoviruses in 200 µl of serum-free medium at MOI 10 or mock infected. The infection medium was replaced with serum-free keratinocyte medium the next day. All experiments were performed in duplicate wells.
Immunofluorescence Analysis.
Cocultures grown on coverslips were fixed 8 days after infection with 5% formalin for 15 min. Melanoma cells were stained with a rabbit anti-S100 antibody (Nova Castra Laboratories, Newcastle upon Tyne, U.K.) at 1:200 dilution, followed by Alexa-Fluor 488-labeled donkey antirabbit secondary antibody (Molecular Probes, Eugene, OR) at a 1:100 dilution. The secondary antibody dilution contained Hoechst 33342 (Sigma) at a dilution of 1:2000 for nuclear staining. Photomicrographs were captured with an Olympus IX70 inverted fluorescence microscope.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
depicts the genomes of the generated recombinant adenoviruses. The tyrosinase gene is specifically expressed in melanocytes and melanomas. A synthetic construct composed of the 260-bp core promoter fused to a dimer of the 200-bp distal enhancer, both from the human tyrosinase gene (in the following designated hTyr2E/P), has been shown previously to mediate highly specific transcription within plasmids (9
, 14) . These properties make hTyr2E/P an interesting candidate for deriving CRAds. For this purpose, we replaced Ad5 nucleotides 346–521, containing the adenoviral E1A promoter, by hTyr2E/P or the ubiquitous CMV enhancer/promoter as a control. Furthermore, we incorporated a synthetic polyA site upstream of these promoters. This element was shown to avoid nonspecific gene expression attributable to transcription initiation in the Ad5 left ITR (Ref. 16
and data not shown). The E1A mutants introduced in the generated CRAd genomes were E1A24 or E1A224. Both mutants contain a 24-bp deletion within conserved region 2 of the E1A gene that prevents the mutant proteins from binding and inactivating pRb (10)
. E1A224 contains an additional deletion of the second amino acid, generating a mutant incapable of binding and inactivating p300 (data not shown; see also Ref. 17
). CRAds containing these mutants were described previously to confer replication specificity to proliferating cells or tumor cells (10
, 18
, 19)
. The corresponding viruses AdTyrwt, AdTyr24, AdTyr224, AdCMV24, and AdCMV224 were generated after transfection of the recombinant genomes into melanoma cells, which were also used for virus amplification. This strategy prevents the generation of wild-type Ad5 as described for 293 packaging cells. Purified virus preparations were analyzed by PCR and restriction digest (see "Materials and Methods") to confirm modifications to the adenoviral genome as well as absence of Ad5 wild-type contamination (data not shown).
|
24, and AdTyr224; the ubiquitous promoter controls AdCMV24 and AdCMV224; wild-type adenovirus (Adwt); and an adenovirus containing the 24 mutation only (Ad24) as further controls. SK-MEL-28 and other melanoma cell lines applied in this study were determined to express tyrosinase by reverse transcription-PCR analysis (data not shown). E1A expression was assessed by Western blot analysis. The results (Fig. 2)
demonstrate similar E1A expression in both cell lines for Adwt and Ad24. Infection with the CMV adenoviruses also resulted in similar E1A expression in both cell lines but at higher levels relative to Adwt and Ad24. Most importantly, there was no detectable E1A expression for tyrosinase adenoviruses AdTyrwt, AdTyr24, and AdTyr224 in SKOV3.ip1 cells (signal similar to noninfected control), whereas E1A expression for these viruses was similar to Adwt and Ad24 in melanoma cells. These findings clearly demonstrate melanoma-restricted activity of hTyr2E/P within the adenovirus genome and furthermore in the context of replicating adenoviruses.
|
24 and AdTyr224 Induce Melanoma-specific Cytotoxicity.24, the ubiquitous control virus AdCMV24, and the first-generation, E1-deleted AdCMVLuc at MOIs of 10, 1, and 0.1. Cytotoxicity was detected by staining of adherent cells with crystal violet (Fig. 3A)
. As expected, no cell killing was observed for AdCMVLuc. AdCMV24 induced complete cell killing in all cell lines at MOI 10. In contrast, AdTyr24 showed selective cytotoxicity to melanoma cells. Complete cell killing was seen for all melanoma cell lines but no (SKOV3.ip1, A549) or only slight (MCF-7) cell killing for control cells at MOI 10. Cytotoxicity of AdTyr24 was similar to AdCMV24 at all titers for melanoma cells, whereas an at least 100-fold attenuation of cell killing by AdTyr24 was observed in control cells. Similar results were obtained with AdTyr224 versus AdCMV224 (Fig. 4A)
. Here, attenuation of AdTyr224 relative to AdCMV224 was <10-fold in melanoma cells but more than 100-1000-fold in control cells. These results demonstrate a clear melanoma-selective cytotoxicity of tyrosinase-controlled viruses AdTyr24 and AdTyr224. Furthermore, these viruses showed similar (AdTyr24) or only slightly lower (AdTyr224) cytotoxicity compared with the control viruses AdCMV24 and AdCMV224, respectively, which contain the CMV promoter/enhancer, which was much stronger than hTyr2E/P (Fig. 2)
.
|
|
24 and AdTyr224 in Melanoma Cells.24, AdCMV24, or AdCMVLuc (Fig. 3B)
, and virus titers generated after 3 days were determined by plaque assay on 293 cells. As expected, no virus replication was observed for AdCMVLuc in either cell type. Whereas AdTyr24 and AdCMV24 generated similar titers of progeny viruses in SK-MEL-28 cells, AdTyr24 titers were >2 orders of magnitude reduced in SKOV3.ip1 cells relative to AdCMV24. These results demonstrate a tissue selectivity of virus replication of >200-fold with similar replication efficiency of AdTyr24 and AdCMV24 in melanoma cells. AdTyr224 showed slightly higher replication selectivity with almost 1000-fold attenuation in SKOV3.ip1 cells and viral progeny titers similar to AdCMV224 in SK-MEL-28 cells (Fig. 4B)
. Thus, the replication properties of AdTyr24 and AdTyr224 correlate remarkably with their cytotoxicity.
AdTyr24 Is Attenuated in Nontumorigenic Cells and Selectively Kills Melanoma Cells but not Surrounding Normal Cells in a Coculture System.
Having established a melanoma-selective mechanism for adenoviral replication, we further analyzed this strategy in clinically more relevant substrates. For this purpose, we infected primary, nontumorigenic NHFs and NHKs with AdTyr24, AdCMV24, and AdCMVLuc at MOIs of 10, 1, and 0.1. Crystal violet staining 9 days (NHKs) or 10 days (NHFs) after virus infection revealed that cytotoxicity of AdTyr24 to normal cells is attenuated more than 10- or 100-fold compared with AdCMV24 in NHKs or NHFs, respectively (Fig. 5)
. Together with the cytotoxicity data of the same viruses for melanoma cell lines (Fig. 3A)
, these results prove selectivity of tyrosinase-controlled CRAds for melanoma versus normal cells. In a further experiment, we assessed selectivity of AdTyr24 in a coculture system of SK-MEL-28 and NHKs (Fig. 6)
. In these cocultures, spindle-like SK-MEL-28 cells form cell clusters within a cobblestone-like monolayer of keratinocytes (Fig. 6A
, no virus), as demonstrated by immunostaining with an antibody to the melanoma surface marker S100 (Fig. 6B
, no virus), which binds SK-MEL-28 cells but not keratinocytes. In addition, nuclei of tumor cells stained stronger with Hoechst 33342. The coculture pattern was not changed after infection with nonreplicating adenoviruses (Fig. 6A
, AdCMVLuc; immunostaining was similar to mock-infected cocultures and is not shown). In contrast, melanoma cell clusters were disrupted and eventually disappeared after infection of cocultures with AdTyr24, leaving a keratinocyte monolayer (Fig. 6A)
. Residual melanoma cells within a layer of keratinocytes were detected by immunofluorescence staining (Fig. 6B)
. These cells show a round morphology characteristic for adenovirus-induced cytopathic effect. Keratinocytes within the monolayer stained with Hoechst 33342 but not with the S100 antibody. These results demonstrate selective killing of SK-MEL-28 cells by AdTyr24. Of note, AdCMV24 killed both melanoma cells and keratinocytes within cocultures (Fig. 6)
, revealing sensitivity of keratinocytes to adenoviral cell lysis in this experimental setting. Thus, selective cell killing of SK-MEL-28 by AdTyr24 in cocultures results from hTyr2E/P dependence of viral replication.
|
|
24 and AdTyr224-mediated Oncolysis.24 and AdTyr224 to a matching tyrosinase-controlled virus with wild-type E1A gene, AdTyrwt, and ultimately to wild-type adenovirus Adwt (Fig. 7)
. Our results show similar cell killing by AdTyr24 and AdTyrwt but slightly reduced cytotoxicity of AdTyr224. Furthermore, AdTyr24 was only marginally attenuated relative to Adwt.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Tyrosinase, the enzyme that catalyzes the rate-limiting step in melanin synthesis, is a melanocyte differentiation marker. Most malignant melanomas are pigmented (melanotic) and thus express tyrosinase. Even unpigmented, amelanotic melanoma cells have been shown to express tyrosinase, although in this case the enzyme was inactive because of defects in posttranslational processing. As such, tyrosinase has been used for diagnostic purposes, such as PCR-based detection of circulating metastatic melanoma cells (20) , and various therapeutic strategies including tumor vaccination and prodrug activation for treatment of melanoma (21) . Furthermore, regulatory sequences of the human and murine tyrosinasegene have been successfully implemented for transcriptional targeting of gene expression in the context of gene therapy (22 , 23) . The application of cellular regulatory elements for transcriptional targeting can benefit from optimization strategies that derive promoters of higher specificity, activity, and/or smaller size (24) . In this regard, artificial tyrosinase constructs containing enhancer dimers, normally located several kilobases upstream of the transcription start site, directly fused to the core promoter showed promising activity and specificity (9 , 14) . A human tyrosinasegene-derived construct, hTyr2E/P, combines small size (<700 bp) with high activity in melanoma cells and complete lack of activity in control cells. These rare features are particularly favorable for an application in the context of CRAds. In contrast to other promoters applied for CRAd targeting (25 , 26) , the small size of hTyr2E/P allowed us to retain the adenoviral E3 region within the recombinant viral genomes. Of note, retaining the E3 region within CRAds has been reported to enhance adenoviral replication efficiency, spread, and oncolysis (27) .6
A further important component of our transcriptional targeting strategy is the polyA transcription termination signal introduced upstream of hTyr2E/P. We found that read-through transcription after initiation within the adenoviral left ITR (16) resulted in loss of specificity of cellular promoters incorporated into the E1 region. Importantly, promoter fidelity could be restored by upstream polyA-mediated transcription termination.7
Our CRAds include deletions within the adenoviral E1A gene, which result in mutant proteins unable to bind pRb and other pocket proteins (24; Ref. 10
) or to p300 (2; data not shown). Sequestration of pRb and p300 by wild-type E1A is a mechanism by which adenoviruses induce cellular DNA synthesis in quiescent cells, as required for viral replication (5)
. An E1A24 mutant adenovirus was shown to replicate selectively in cells with inactive pRb (10)
. Mutations within the pRb pathway are frequent in melanomas. These include activating point mutations of the upstream regulatory kinase CDK4, overexpression of CDK-binding cyclin D2, or inactivation of the CDK4-inhibitor p16INK4a (11)
. These mutations result in inactivation of pRb in most if not all melanomas, thus providing a rationale for application of E1A24 mutant adenoviruses in CRAd therapy of this malignancy. In contrast, p53 mutations are rare in melanoma (28)
, suggesting that CRAds targeting this defect (29
, 30)
are not advantageous to treat this type of cancer. Furthermore, we generated an E1A224 mutant adenovirus based on recent reports that showed improved tumor specificity of adenoviruses encoding an E1A mutant unable to bind both p300 and pRb, as compared with a mutant impaired for pRb-binding alone (19)
. Importantly, efficiency of viral replication and cytolysis of these E1A mutant adenoviruses were shown to be similar to wild-type adenovirus in cancer cells, whereas E1B mutant viruses were reported to be strongly attenuated even in target cells (18
, 19)
. The conserved functions of E1A domains other than those required for pRb and p300 binding might be crucial in this regard. These functions include transactivation of viral and cellular genes critical for productive viral replication (5)
. Despite these favorable qualities of E1A mutant CRAds, side effects attributable to replication in proliferating normal cells (18
, 31)
are of potential concern. By driving expression of these E1A mutants from hTyr2E/P, we propose a strategy to avoid this problem, thus deriving highly selective CRAds for application in melanoma therapy.
By Western blot analysis, we prove melanoma-specific E1A expression by the generated tyrosinase viruses AdTyrwt, AdTyr24, and AdTyr224. The demonstration of promoter fidelity in the context of replicating adenoviruses is critical for the development of promoter-driven CRAds, because interference of adenovirus sequences and/or proteins with specific promoter activity were shown previously (32
, 33) . Of interest for our approach is the recently reported reduced specificity for tyrosinase enhancer/promoter sequences in first-generation adenoviruses (34)
. Our strategy to incorporate a transcription termination signal upstream of hTyr2E/P might be critical in this regard. A further concern was promoter interference by the E1A enhancer, which cannot be deleted from CRAd genomes because it overlaps with the adenoviral packaging signal. Insulator elements can avoid interference of the E1A enhancer with heterologous promoters (35
, 36)
. However, specificity of the hTyr2E/P element was not lost by enhancer interference within the CRAd genome, as our E1A expression analysis clearly demonstrates. We conclude that the potency of the E1A enhancer to activate heterologous promoters depends on the promoter identity and might be determined by individual promoter elements and the corresponding proteins that bind to these elements. In view of these concerns, it is important to note that we could not detect any E1A expression in nonmelanoma cells, demonstrating the "tightness" of the polyA-hTyr2E/P cassette in the CRAd genome.
AdTyr24 and AdTyr224 induced highly melanoma-selective cytotoxicity with a 100-1000-fold attenuation in nonmelanoma cell lines of diverse tissue origin. This demonstration of CRAd-dependent cytotoxicity is decisive, because virus production does not necessarily result in evident cytopathic effect, as shown for other cell types (8)
. Conversely, cell death after CRAd infection can result from direct effects of viral proteins rather than viral cell lysis. However, cell lysis is most relevant for the oncolytic potency of CRAds, because it results in release and spread of progeny viruses. Therefore, we analyzed AdTyr24 and AdTyr224 in a progeny production assay. Normalized to the ubiquitous promoter control viruses AdCMV24 and AdCMV224, the tyrosinase viruses resulted in 212-fold (24 viruses) or 857-fold (224 viruses) lower titers in nonmelanoma versus melanoma cells. We conclude that specific expression of the E1A mutants results in selective viral replication, progeny production, and cytotoxicity. Qualitatively and quantitatively similar results were reported for CRAds with wild-type E1A expressed from the PSA or human kallikrein promoter in prostate cancer, from the alpha fetoprotein promoter in hepatocellular carcinoma, or from the midkine promoter in pediatric tumors (25
, 26
, 37
, 38)
.
Strong differences in promoter activity, as shown by Western blot for hTyr2E/P and CMV enhancer/promoter in melanoma cells, did not result in different replication efficiencies of the corresponding viruses. Thus, E1A expression at wild-type levels, as mediated by hTyr2E/P, is sufficient for optimal replication of heterologous promoter-driven CRAds. These observations are in accordance with other reports that show little effect of wide variations in E1A expression on adenovirus replication (39)
. The reason for reduced cytotoxicity to melanoma cells of AdTyr224 versus AdCMV224, despite similar replication efficiency, is currently under investigation. Importantly, AdTyr24 showed cell killing similar to AdCMV24. Our results indicate that promoter qualities required by transcriptionally targeted CRAds differ from those required by gene therapy. For the latter, it is well reported that increased promoter strength and transgene expression translate to increased therapeutic efficacy. Thus, the weak activity of most specific cellular promoters represents a major drawback for their application in gene therapy (40)
. In contrast, our studies show that weaker promoters are sufficient to drive E1A expression at levels required for efficient viral replication and oncolysis. Furthermore, these observations imply that promoter "tightness" as described here for hTyr2E/P is critical for CRAd applications, where minimal E1A expression in healthy tissues can result in undesirable viral replication.
A key requirement for clinical applications of CRAds is their attenuation in non-target tissues. Therefore, we tested our concept of hTyr2E/P-controlled CRAds in the most meaningful substrate, normal human cells. In this regard, it is important to note that the evaluation of CRAd specificity and toxicity in animal studies is problematic because adenovirus replication is species specific. Importantly, our study demonstrates attenuated cytotoxicity of AdTyr24 in NHFs and NHKs. More than 10-fold (keratinocytes) or 100-fold (fibroblasts) higher titers of AdTyr24 were required to achieve a cytopathic effect similar to AdCMV24. In contrast cytotoxicity of AdTyr24 in melanoma cells was similar or superior to AdCMV24. Furthermore, we demonstrate that AdTyr24 kills selectively melanoma cells but not surrounding keratinocytes in a coculture system. In this experimental setting, keratinocytes are constantly exposed to adenoviruses released from neighboring melanoma cells. However, AdCMV24 killed both melanoma cells and keratinocytes in cocultures. Thus, specific expression of E1A by hTyr2E/P rather than a different susceptibility to adenoviral infection or replication per se between SK-MEL-28 and NHK is responsible for the observed melanoma selectivity of AdTyr24 in the coculture system. We conclude that our results reveal substantial selectivity of hTyr2E/P-controlled CRAds for melanoma cells versus normal cells in a clinically relevant assay system.
In light of recent clinical trials with dl1520 (ONYX-015), efficacy of CRAd replication turned out to be limiting for this approach to translate into clinical responses. As discussed above, hTyr2E/P-driven expression of E1A mutants does not reduce adenoviral replication and progeny production relative to viruses controlled by the stronger CMV enhancer/promoter, indicating that the strength of hTyr2E/P is not limiting. However, AdTyr24 and AdTyr224 showed slightly reduced killing of melanoma cells relative to wild-type adenovirus Adwt. This attenuation might result from constitutive activity of hTyr2E/P. Prolonged expression of E1A and E1A-dependent viral and cellular genes potentially interferes with late viral replication, a scenario that relates to constitutive promoters per se rather than specifically to hTyr2E/P. AdTyr24 and AdTyrwt showed similar cytotoxicity to melanoma cells. This observation is consistent with previous reports that demonstrated similar cytopathic effects of an adenovirus containing the E1A24 mutant driven from the endogenous E1A promoter relative to the wild-type control (18)
. However, for the double mutant AdTyr224, we report a reduced cytotoxic effect compared with AdTyr24 and AdTyrwt. We conclude that the introduction of E1A mutants, especially of 24, was not deleterious to the potency of the CRAds generated herein. Future studies have to analyze in adequate model systems and ultimately in a clinical setting whether these mutations translate into reduced toxicity to normal tissues. In addition, animal studies have to prove efficiency of these CRAds in vivo.
The melanoma-targeted CRAds described in this study embody a platform for further therapeutic development:
(a) Incorporation of therapeutic genes can generate an efficient gene therapy vector, or "armed CRAd," which amplifies gene transfer to melanoma by specific generation and spread of progeny viruses/vectors within the tumor. This strategy has been implemented with suicide genes for other replicating adenoviruses (41 , 42) . In this context, the timing of prodrug application is critical to prevent adverse effects attributable to killing of virus-producing cells (43 , 44) . In light of recent results for immunotherapy of melanoma (45) , the expression of immunostimulatory genes by CRAds might be a favorable approach for combined therapy. However, space limitations for incorporation of transgenes into CRAds must be considered. In this regard, deletion of the adenoviral E3 region to allow transgene incorporation into CRAds (41 , 42) is problematic. However, a partial E3 deletion might be feasible, because one E3-encoded protein, the adenovirus death protein, was shown to be critical for viral release (46) .
(b) Although the CRAds generated in this study show considerable melanoma-targeted replication and cytotoxicity with a selectivity of more than two orders of magnitude, further increased specificity might be advantageous for therapeutic applications. This goal has been achieved for other transcriptionally targeted CRAds by driving the expression of further viral genes from a specific promoter. Strategies to this end include the application of an internal ribosome binding site or of additional specific promoters (38 , 47) . These strategies can be similarly applied to the tyrosinase CRAds described in this report.
(c) Finally, the critical aspect of adenovirus infectivity for the target tissue must be considered in the context of CRAds. Low infectivity of primary melanoma cells attributable to negligible expression of the adenovirus receptor CAR has been reported (48) . In consequence, future studies should pursue strategies for infectivity enhancement and transductional targeting of CRAds to melanoma. For this purpose, genetic capsid modification (49) or conjugation of targeting molecules to the viral capsid (50) might be applied to the melanoma-restricted CRAds described herein.
In conclusion, we developed highly selective CRAds as novel therapeutic agents for treatment of melanoma by amplified gene transfer or viral oncolysis.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by Deutsche Forschungsgemeinschaft Grant NE832/1 (to D. M. N.), United States Department of Defense Contracts DAMD 17-00-1-0002 and DAMD 17-98-1-8571, National Cancer Institute Grants R01 CA83821 and P50 CA83591, Lustgarten Foundation Grant LF043, the CapCURE Foundation, and the Center for AIDS Research Core Support Grant P30-AI-27767 from the National Institute of Allergy and Infectious Diseases. 
2 To whom requests for reprints should be addressed, at Division of Human Gene Therapy, Gene Therapy Center, BMR II 901 19th Street S, University of Alabama at Birmingham, Birmingham, AL 35294-2172. Phone: (205) 975-0171; Fax: (205) 975-7949; E-mail: dirk.nettelbeck{at}ccc.uab.edu
3 Present address: Almirall Prodesfarma, Barcelona Science Park, 08028 Barcelona, Spain.
4 Present address: Gene Therapy Unit, Institut Català d’Oncologia, 08907 L’Hospitalet Barcelona, Spain.
5 The abbreviations used are: CRAd, conditionally replicative adenovirus; pRb, retinoblastoma protein; ATCC, American Type Culture Collection; NHF, normal human fibroblast; NHK, normal human keratinocyte; CMV, cytomegalovirus; MOI, multiplicity of infection; ITR, inverted terminal repeat; wt, wild type.
6 K. Suzuki, personal communication.
7 D. M. Nettelbeck, unpublished observation.
Received 2/12/02. Accepted 6/ 6/02.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. M. Hine, A. Seluanov, and V. Gorbunova Use of the Rad51 promoter for targeted anti-cancer therapy PNAS, December 30, 2008; 105(52): 20810 - 20815. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. J. Bauerschmitz, T. Ranki, L. Kangasniemi, C. Ribacka, M. Eriksson, M. Porten, I. Herrmann, A. Ristimaki, P. Virkkunen, M. Tarkkanen, et al. Tissue-Specific Promoters Active in CD44+CD24-/low Breast Cancer Cells Cancer Res., July 15, 2008; 68(14): 5533 - 5539. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Tsuruta, L. Pereboeva, M. Breidenbach, D. T. Rein, M. Wang, R. D. Alvarez, G. P. Siegal, P. Dent, P. B. Fisher, and D. T. Curiel A Fiber-Modified Mesothelin Promoter-Based Conditionally Replicating Adenovirus for Treatment of Ovarian Cancer Clin. Cancer Res., June 1, 2008; 14(11): 3582 - 3588. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Leja, H. Dzojic, E. Gustafson, K. Oberg, V. Giandomenico, and M. Essand A Novel Chromogranin-A Promoter-Driven Oncolytic Adenovirus for Midgut Carcinoid Therapy Clin. Cancer Res., April 15, 2007; 13(8): 2455 - 2462. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Hesse, D. Kosmides, R. E. Kontermann, and D. M. Nettelbeck Tropism Modification of Adenovirus Vectors by Peptide Ligand Insertion into Various Positions of the Adenovirus Serotype 41 Short-Fiber Knob Domain J. Virol., March 15, 2007; 81(6): 2688 - 2699. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Kamizono, S. Nagano, Y. Murofushi, S. Komiya, H. Fujiwara, T. Matsuishi, and K.-i. Kosai Survivin-Responsive Conditionally Replicating Adenovirus Exhibits Cancer-Specific and Efficient Viral Replication Cancer Res., June 15, 2005; 65(12): 5284 - 5291. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. T. Rein, M. Breidenbach, T. O. Kirby, T. Han, G. P. Siegal, G. J. Bauerschmitz, M. Wang, D. M. Nettelbeck, Y. Tsuruta, M. Yamamoto, et al. A Fiber-Modified, Secretory Leukoprotease Inhibitor Promoter-Based Conditionally Replicating Adenovirus for Treatment of Ovarian Cancer Clin. Cancer Res., February 1, 2005; 11(3): 1327 - 1335. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. L. Chu, D. E. Post, F. R. Khuri, and E. G. Van Meir Use of Replicating Oncolytic Adenoviruses in Combination Therapy for Cancer Clin. Cancer Res., August 15, 2004; 10(16): 5299 - 5312. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. S. Banerjee, A. A. Rivera, M. Wang, L. T. Chow, T. R. Broker, D. T. Curiel, and D. M. Nettelbeck Analyses of melanoma-targeted oncolytic adenoviruses with tyrosinase enhancer/promoter-driven E1A, E4, or both in submerged cells and organotypic cultures Mol. Cancer Ther., April 1, 2004; 3(4): 437 - 449. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. J. Waxman and P. S. Schwartz Harnessing Apoptosis for Improved Anticancer Gene Therapy Cancer Res., December 15, 2003; 63(24): 8563 - 8572. [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 |
, packaging signal; hTyrE, human tyrosinase enhancer; hTyrP, human tyrosinase promoter; CMVE/P, cytomegalovirus enhancer/promoter; N, E1A NH2-terminal domain; CR2, E1A conserved region 2; E1A wt, E1A wild-type; E1A
