Cyclophosphamide Increases Transgene Expression Mediated by an Oncolytic Adenovirus in Glioma-Bearing Mice Monitored by Bioluminescence Imaging

ARTICLE

doi:10.1016/j.ymthe.2006.08.008

Cyclophosphamide Increases Transgene Expression Mediated by an Oncolytic Adenovirus in Glioma-Bearing Mice Monitored by Bioluminescence Imaging Martine L. M. Lamfers,1,2,* Giulia Fulci,2,3 Davide Gianni,2 Yi Tang,4 Kazuhiko Kurozumi,3 Balveen Kaur,3 Sharif Moeniralm,1 Yoshinaga Saeki,2,3 Jan E. Carette,5 Ralph Weissleder,4 W. Peter Vandertop,6 Victor W. van Beusechem,5 Clemens M. F. Dirven,1 and E. Antonio Chiocca2,3 1

Department of Neurosurgery and 5Department of Medical Oncology, 6Neurosurgical Center Amsterdam, VU University Medical Center, Amsterdam 1007 MB, The Netherlands 2 Molecular Neuro-oncology Laboratories, Neurosurgery Service, and 4Center for Molecular Imaging Research, Massachusetts General Hospital, Harvard Medical School, Charlestown, MA 02129, USA 3 Dardinger Center for Neuro-oncology, Department of Neurological Surgery, The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210, USA *To whom correspondence and reprint requests should be addressed. Fax: +31 0 20 444 4975. E-mail: [email protected].

Available online 22 September 2006

Approaches to improve the oncolytic potency of replication-competent adenoviruses include the insertion of therapeutic transgenes into the viral genome. Little is known about the levels and duration of in vivo transgene expression by cells infected with such ‘‘armed’’ viruses. Using a tumorselective adenovirus encoding firefly luciferase (AdD24CMV-Luc) we investigated these questions in an intracranial mouse model for malignant glioma. Luciferase expression was detected by bioluminescence imaging, and the effect of the immunosuppressive agent cyclophosphamide (CPA) on transgene expression was assessed. Intratumoral AdD24CMV-Luc injection led to a localized dose-dependent expression of luciferase. Surprisingly, this expression decreased rapidly during the course of 14 days. In contrast, mice injected with nonreplicating Ad.CMV-Luc demonstrated stable transgene expression. Treatment of mice with CPA in combination with AdD24CMV-Luc retarded the loss of transgene expression. Staining of mouse brains for inflammatory cells demonstrated decreased tumor infiltration by immune cells in CPA-treated mice. Moreover, in immunodeficient NOD/SCID mice loss of transgene expression was less rapid and not prevented by CPA treatment. Together, our data demonstrate that transgene expression and viral replication decrease rapidly after intratumoral injection of oncolytic adenovirus in mouse brains and that treatment with the immunomodulator CPA prolongs viral-mediated gene expression. Key Words: glioma, oncolytic adenovirus, transgene expression, bioluminescence imaging, immune suppression

INTRODUCTION Replication-competent (oncolytic, replication-conditional, replication-compromised) viruses are being investigated as novel anti-cancer agents for various types of cancers [1], including malignant glioma [2–4]. One feature of such viruses is related to the capacity to engineer therapeutic transgenes into their genome to allow for multimodal therapy. Such ‘‘armed’’ therapeutic viruses (ATVs) allow the action of therapeutic proteins to be combined with the anti-tumor properties of the viral

MOLECULAR THERAPY Vol. 14, No. 6, December 2006 Copyright C The American Society of Gene Therapy 1525-0016/$30.00

infection. Several oncolytic herpes and adenoviruses that ferry such therapeutic transgenes have been described [5– 10]. However, to date, little is known about the levels and duration of transgene expression in cells infected by such armed viruses. AdD24 is an oncolytic adenovirus that is being widely used as a backbone in the construction of various ATVs [11–14]. The tumor specificity of AdD24 is based upon a partial deletion in the CR2 domain of the adenoviral E1A gene, thereby abrogating the bind-

779

ARTICLE

ing of the E1A protein to pRB [15]. The adenoviral E3 region has been removed in AdD24, allowing the insertion of transgene expression cassettes. AdD24 and modifications of this backbone have shown therapeutic efficacy in various in vitro and in vivo models of malignant glioma [14–18]. In the current study, we investigated the in vivo kinetics of transgene expression by this type of oncolytic adenovirus and strategies to increase transgene levels. We employed an AdD24 virus encoding the firefly luciferase marker gene (AdD24CMV-Luc) [19]. The levels of luciferase expressed in tumors infected with this virus were studied in mice bearing intracranial glioma xenografts. Luciferase expression was monitored noninvasively using a cooled charge-coupled device camera. A time-dependent decrease in luciferase gene expression in tumors injected with this oncolytic virus in vivo was found. Concomitant administration of the immunomodulating agent cyclophosphamide (CPA) retarded the observed time-dependent decrease in luciferase expression and decreased the inflammatory reaction mediated by the oncolytic virus in the athymic nude mouse strain. This effect was not observed in nonobese diabetic/severe combined immune deficiency (NOD/SCID) mice, which harbor severe defects in both innate and adaptive immunologic function. Our data indicate that immune-mediated factors play a role in the rapid loss of transgene expression and viral replication in mouse brains and that pretreatment with immunomodulating agents can retard this process.

RESULTS Intratumoral Transgene Expression from an Oncolytic Adenovirus in a Mouse Glioma Model is Transient We injected athymic nude mice bearing intracranial human gliomas with a dose range (106, 107, or 108 plaque-forming units (pfu)) of the replication-competent oncolytic adenovirus encoding luciferase (AdD24CMVLuc) or with 107 or 108 pfu of a replication-defective adenoviral vector encoding firefly luciferase (Ad.CMVLuc). We monitored transgene expression noninvasively by bioluminescence imaging up to day 14 postinfection (Fig. 1). Further follow-up was not possible due to mice becoming symptomatic from tumor growth. The replication competence of the virus leads one to expect a time-dependent increase in luciferase activity as a function of increased viral titers by ongoing replicative cycles. However, despite evidence of ongoing viral replication at all three dosages of AdD24CMV-Luc as detected by hexon staining on cryosections at time of sacrifice (not shown), we observed a rapid decrease in intracranial luciferase expression. On the other hand, luciferase expression from both dosages of the replication-defective Ad.CMV-

780

doi:10.1016/j.ymthe.2006.08.008

Luc remained high throughout the course of the experiment. Quantitative analysis of luciferase signals demonstrated that by day 14 luciferase expression had decreased 23.2-, 21.3-, and 29.4-fold relative to the first measurements taken at 24 h postinfection for the 106, 107, and 108 dose, respectively (Fig. 2). These results thus show that transgene expression mediated by an oncolytic adenovirus in human gliomas in mice is transient and limited in time. Intratumoral Virus Replication Rapidly Decreases with Time To visualize the kinetics of viral replication and spread in the intracranial tumors, we injected nude mice bearing U-87DEGFR xenografts with 5  107 pfu AdD24CMV-Luc and sacrificed them on day 3, 5, or 7 postinfection and at the time of manifestation of symptoms (day 16). Analysis of viral replication by immunohistochemical staining for adenovirus hexon proteins demonstrated that virus replication was distributed in a patchy manner throughout the tumor. There were a large number of uninfected tumor cells in relation to the cells that were infected. A rapid decrease in hexon-positive areas of the tumor was seen as the tumor continued to grow over time (Fig. 3). These findings are in accordance with the loss of transgene expression from a replicating adenovirus as detected by bioluminescence imaging. Cyclophosphamide Retards Loss of Transgene Expression in the Athymic Mouse Glioma Model We have previously reported that cyclophosphamide limits the effects of the host response against an oncolytic HSV1 in human and rat gliomas in rats, allowing prolonged viral replication and transgene expression to take place [20,21]. We thus sought to determine if the addition of CPA enhanced luciferase activity in AdD24CMV-Luc-infected human gliomas in mice. To this end, we injected intracranial U-87DEGFRbearing nude mice with 5  107 pfu AdD24CMV-Luc with or without CPA pretreatment. We chose this viral dose based on the observation that mice are not cured at this dose and hence tumor load would not be a limitation for the replication of the virus, as demonstrated in Fig. 3. Fig. 4 depicts the color-coded images of the individual mice at days 4, 9, and 12 post-virus injection. Control mice (left column) again demonstrate a gradual loss of transgene expression with time in tumors infected with AdD24CMV-Luc. However, CPA treatment (right column) appeared to retard the decrease in luciferase activity over time. In fact, image analysis revealed that CPA allowed for a relatively small decline (6-fold) in luciferase activity in tumors compared to untreated tumors (33-fold), and by the 14th day there was an order of magnitude difference in tumors treated with CPA + virus compared to PBS + virus (Fig. 5). Significant differences in total luciferase

MOLECULAR THERAPY Vol. 14, No. 6, December 2006 Copyright C The American Society of Gene Therapy

doi:10.1016/j.ymthe.2006.08.008

ARTICLE

FIG. 1. In vivo imaging of luciferase activity in mice bearing U-87DEGFR intracranial tumors treated with an intratumoral injection of a dose range of AdD24CMV-Luc (106, 107, 108 pfu) or Ad.CMV-Luc (107 or 108 pfu). Individual mice are shown with color-coded overlay on days 1, 7, and 14 after virus injection.

expressed between these groups were detected on days 5, 9, 12, and 15 ( P b 0.05). The Activated Metabolite of CPA Does Not Affect Transgene Expression Mediated by AdD24CMV-Luc in Vitro One potential mechanism that could explain the in vivo CPA effect might relate to the cytotoxic mode of action of the drug. In fact, adenoviral oncolysis has been reported to be enhanced in combination with various cytotoxic agents [22–26]. If such a mechanism were operative, one would expect to see an increase in viral-mediated transgene expression in vitro, too. To assess whether the observed in vivo enhancement of luciferase expression resulted from direct effects of CPA on transgene expression by the AdD24CMV-Luc-infected tumor cells, we also treated human U-87DEGFR cells with this combination in vitro. We incubated cells with a concentration of the

MOLECULAR THERAPY Vol. 14, No. 6, December 2006 Copyright C The American Society of Gene Therapy

activated metabolite of CPA (4-hydroperoxy-CPA; 4-HO2CPA) that was not toxic to tumor cells and subsequently infected them with m.o.i. 1, 4, and 16 of AdD24CMV-Luc. On days 4 and 7 postinfection, we analyzed the cells for viability and luciferase expression. Treatment with a nontherapeutic dose of 4-OH-CPA (3 AM) had no significant effect on the viability of glioma cells at day 4 or 7 (m.o.i. 0). Whereas the oncolytic activity of AdD24CMV-Luc was not detectable at day 4, a dosedependent cytotoxic effect of the virus was evident by day 7. This effect was not significantly altered by 3 AM 4OH-CPA (Fig. 6A). This dose of 4-OH-CPA also did not significantly alter luciferase expression at day 4 or 7 (Fig. 6B). As expected, when we administered higher doses of 4-OH-CPA, we found an enhanced cell killing effect in combination with the virus, although transgene expression was not significantly altered (data not shown). These results indicate that the activated metabolite of CPA does

781

ARTICLE

doi:10.1016/j.ymthe.2006.08.008

pared to 26.2 F 1.2 and 22.8 F 2.1% positive cells, respectively, in the PBS + AdD24CMV-Luc tumors (Fig. 7B, top). Moreover, at this time point, adenoviral hexon protein immunopositivity was quantitatively and qualitatively increased in CPA-treated animals compared to the PBS controls. This increase correlated with enhanced bioluminescence signals in CPA-treated mice compared to controls starting at this time point. Analysis of tumors harvested at time of sacrifice (days 16–21) revealed no detectable differences in CD68 or CD45 immunopositivity and only very few hexon-positive cells within each tumor (not shown). These results provided suggestive evidence that CPA induced a transient inhibition of inflammatory cell infiltration in tumors treated with the oncolytic AdD24CMV-Luc and that this inhibition is associated with an increase in adenovirus replication. FIG. 2. Quantitative image analysis of intensity of light emitted from intracranial tumor sites demonstrates loss of transgene expression in mice treated with replication-competent adenovirus and endurance of transgene expression in mice treated with replication-deficient virus. Shown are the mean total relative light units (RLU) per mouse F SD.

not affect transgene expression levels by direct action on tumor cells. Decreases in Inflammatory Cell Infiltrates in Gliomas Treated with CPA and Oncolytic Virus Since the effect of CPA did not appear to be related to the drugTs direct action on the cell, we sought to determine if effects on inflammatory infiltrates were present. It is important to note that athymic nude mice still produce small numbers of T cells, as well as still being able to produce interferons, natural killer cells, and macrophages [27]. These residual components of the immune system can contribute to an inflammatory response to oncolytic adenoviruses. We harvested mouse brains 72 h and 5 days after AdD24CMV-Luc infection and at the time of occurrence of symptoms due to tumor growth (days 16– 21). Analysis of tumors of animals treated with PBS + AdD24CMV-Luc demonstrated the presence of numerous CD45- and CD68-positive cells at 72 h postinfection, 20.9 F 3.3 and 17.2 F 3.7%, respectively (Fig. 7A, top). CD45 is expressed on cells of hematopoietic lineage, especially lymphocytes, and to a lesser extent on microglial cells [28]. CD68 is expressed at high levels on monocytic cells, macrophages, and microglial cells. Adenoviral hexon immunopositivity was detected in patches throughout the tumor. When we administered CPA in combination with AdD24CMV-Luc (Fig. 7A, bottom) CD45 and especially CD68 immunopositivity was decreased to14.2 F 5.6 and 4.4 F 1.9%, respectively, while hexon immunopositivity remained similar at this time point. At day 5 postinfection CD45 and CD68 immunopositivity in tumors of animals treated with CPA + AdD24CMV-Luc, 16.1 F 4.5 and 11.1 F 2.6%, respectively, was still decreased (Fig. 7B, bottom), com-

782

Cyclophosphamide Does Not Prevent Loss of Transgene Expression in the NOD/SCID Mouse Glioma Model To verify that an immune response and not other physiological components are being affected by the CPA treatment in the nude mouse glioma model, we repeated this experiment in the NOD/SCID strain. These mice are characterized by a functional deficit in T, B, and NK cells; absence of circulating complement; and defective antigen-presenting cells [29]. We monitored transgene expression noninvasively by bioluminescence imaging up to day 9 after AdD24CMV-Luc injection. Further follow-up was not possible due to mice becoming symptomatic from tumor growth, which occurred more rapidly than in the nude mice. Data analysis of acquired

FIG. 3. Immunohistochemical staining for adenovirus hexon proteins in cryosections of brains of U-87DEGFR tumor-bearing mice treated with 5  107 pfu AdD24CMV-Luc and harvested at day 3, 5, 7, and 16 after virus injection. Tumor borders are indicated with dashed lines. Scale bar represents 1 mm.

MOLECULAR THERAPY Vol. 14, No. 6, December 2006 Copyright C The American Society of Gene Therapy

doi:10.1016/j.ymthe.2006.08.008

ARTICLE

FIG. 4. In vivo imaging of luciferase activity in mice bearing U-87DEGFR intracranial tumors treated with intratumoral injection of AdD24CMV-Luc alone (left column) or in combination with CPA treatment (right column). Shown are the individual mice with color-coded overlay on days 4, 9, and 12 after virus injection.

images revealed a relatively small but significant decline (fivefold, P b 0.05) in luciferase activity in controls by day 4 postinfection, which then stabilized up to day 9. Importantly, CPA treatment did not prolong high-level transgene expression in these mice (Fig. 8A). Immunohistochemical analysis of cryosections of the brains of these animals harvested on day 9 postinfection demonstrated the presence of CD45- and CD68-positive cells in PBS + virus-treated tumors, 9.3 F 3.4 and 10.9 F 1.1%, respectively, albeit in lower amounts and smaller in size than we detected in virus-treated tumors of athymic mice. In addition, the staining pattern for adenovirus hexon protein was more pronounced and widespread than seen in athymic mice on day 7. CPA treatment of NOD/SCID mice did not affect CD45, CD68, and hexon staining patterns in virus-treated tumors (Fig. 8B).

DISCUSSION The use of ATVs is being exploited as a novel therapeutic modality for tumors. One assumption in this therapy has

MOLECULAR THERAPY Vol. 14, No. 6, December 2006 Copyright C The American Society of Gene Therapy

been that viral-mediated transgene expression would increase as long as tumor burden persists due to the recurring cycles of viral replication with viral progeny production and subsequent infection and replication in additional tumor cells. Indeed, in vitro studies in glioma cells have shown a 3 log higher level of luciferase expression 48 h after infection with replication-competent adenovirus encoding luciferase compared to replication-deficient control vectors [30]. Moreover, in primary glioma spheroids a ninefold increase in luciferase expression was found during the course of 9 days using replication-competent adenovirus, compared to a fourfold decrease using replication-deficient adenovirus encoding luciferase [31]. However, the level and duration of in vivo transgene expression from an oncolytic adenovirus have not been monitored thus far. Using bioluminescence imaging, we assessed the pattern of transgene expression from an oncolytic adenovirus delivered to a tumor. Others have reported the utility of bioluminescence imaging for monitoring transgene expression from implanted luciferase-expressing (tumor)

783

ARTICLE

doi:10.1016/j.ymthe.2006.08.008

FIG. 5. Quantitative image analysis of intensity of light emitted from intracranial tumor sites demonstrates prolonged high-level transgene expression in mice treated with CPA + AdD24CMV-Luc (CPA) compared to PBS + AdD24CMV-Luc (Control). Shown are mean total relative light units (RLU) per mouse F SEM, *P b 0.05 compared to controls.

cells [32,33] or after systemic or intratumoral administration of luciferase-encoding nonreplicating (targeted) viral vectors or naked plasmid DNA [34–36]. Moreover, viral infections such as encephalomyelitis by Sindbis virus [37] and spread to various anatomic sites of herpes simplex virus type 1 [38] can be detected using bioluminescence imaging. Furthermore, Yu et al. have employed this system to detect tumors and metastases by injection of btumor-seekingQ bacteria or vaccinia virus encoding luciferase [39]. Thus far, bioluminescence imaging has not been applied for direct in vivo visualization and monitoring of transgene expression from an oncolytic virus.

We demonstrate here that intratumoral injection of the firefly luciferase-encoding tumor-selective adenovirus AdD24CMV-Luc leads to dose-dependent expression of luciferase at the tumor site. This expression can be detected for at least 15 days. However, we also demonstrate that transgene expression actually decreases fairly precipitously upon intratumoral injection of this oncolytic virus. This decrease was unexpected and distinct from the luciferase expression pattern after infection with the nonreplicating vector Ad.CMV-Luc, which remained high throughout the course of the experiment. The decrease in transgene expression from AdD24CMV-Luc may result from less efficient viral spreading and reinfection within the tumor than was predicted by in vitro studies [30,31]. This is supported by studies on wild-type adenovirus spreading in subcutaneous xenografts [40,41]. In addition, immune-mediated factors also appear to play a role in this phenomenon. It has been reported that in the immunoprivileged environment of the brain, nonreplicating adenoviral vectors provide stable and prolonged gene expression [42,43] by comparison to peripheral organs [44]. However, above a certain threshold, acute vector-mediated toxicity and chronic inflammation lead to a decline in transduction persistence [45]. These results are congruent with the stable luciferase expression in brains of mice injected with 107 or 108 pfu of the replication-defective Ad.CMV-Luc. The decline in luciferase expression after intratumoral injection of the replication-competent AdD24CMV-Luc in athymic mice is associated with a CD45+ and CD68+ cellular infiltrate, suggestive of an inflammatory reaction. Bioluminescence imaging of mice

FIG. 6. (A) Viability of U-87DEGFR glioma monolayers after treatment with a dose range of AdD24CMV-Luc alone ( ) or in combination with a nontoxic dose of 4-HO2-CPA (n) was determined using the WST-1 assay on days 4 and 7 postinfection. Results are expressed as mean percentage of untreated controls F SD. (B) Luciferase expression in U87DEGFR glioma cells was determined on days 4 and 7 postinfection with a dose range of AdD24CMV-Luc alone (gray bars) or in combination with a nontoxic dose of 4-HO2-CPA (black bars). Results are expressed as mean relative light units (RLU) F SD.

.

784

MOLECULAR THERAPY Vol. 14, No. 6, December 2006 Copyright C The American Society of Gene Therapy

doi:10.1016/j.ymthe.2006.08.008

ARTICLE

FIG. 7. Immunohistochemical staining for common leukocyte antigen (CD45), monocytic cells/macrophages (CD68), and adenoviral hexon proteins in intracranial U-87DEGFR tumors of athymic mice treated with AdD24CMVLuc + PBS (top) or AdD24CMV-Luc + CPA (bottom) at (A) 72 h postinfection or (B) 5 days postinfection. Original magnification 20.

treated with a combination of AdD24CMV-Luc and the immunosuppressive agent cyclophosphamide demonstrated that CPA can retard the loss of transgene expression, resulting in prolonged high-level luciferase expression at the tumor site. Immunohistochemical analysis of tumor sections confirmed the bioluminescence data by detecting the presence of ongoing viral replication and reduction of immune cell infiltrate in CPA-treated mice. Further support for the hypothesis that an inflammatory response contributes to the inhibition of oncolytic adenovirus-mediated transgene expression and replication in intracranial tumors was obtained using NOD/ SCID mice. The results from this experiment demonstrate that in a background of multiple defects in innate and adaptive immune function, transgene expression from an oncolytic adenovirus declines much less rapidly. Moreover, treatment with the immunomodulating agent CPA has no significant effect on levels of transgene expression.

MOLECULAR THERAPY Vol. 14, No. 6, December 2006 Copyright C The American Society of Gene Therapy

The finding that in vivo transgene expression mediated by an oncolytic adenovirus decreases after tumor injection is in agreement with our previous findings [46,47] with an oncolytic herpes simplex virus (HSV). CPATs ability to retard the loss of transgene expression mediated by an adenovirus also confirms the previous findings with an oncolytic HSV. In addition, for the first time, we also show that CPATs ability to do this is not dependent on species since all our previous studies were reported in rats, while the current experiments were performed in mice. The mechanism of CPA action in providing for improved transgene expression does not depend on a direct effect on cellular metabolism since it does not occur in vitro. Its effect thus must involve an effect on the host or on the tumor stroma. We observed a rapid upregulation of cells expressing CD45 and CD68 in virally treated tumors of athymic mice. These cells represent lymphocytic and monocytic/microglial populations, which have been shown to be involved in inflammatory reactions against

785

ARTICLE

doi:10.1016/j.ymthe.2006.08.008

FIG. 8. (A) Quantitative image analysis of intensity of light emitted from intracranial tumor sites in NOD/SCID mice showing transgene expression levels of mice treated with CPA + AdD24CMV-Luc (CPA) or PBS + AdD24CMVLuc (Control). Shown are mean total relative light units (RLU) per mouse F SEM. (B) Immunohistochemical staining for CD45, CD68, and adenoviral hexon proteins in intracranial U-87DEGFR tumors of NOD/SCID mice treated with AdD24CMV-Luc+PBS (top) or AdD24CMVLuc + CPA (bottom). Original magnification 20.

replication-defective adenoviral vectors as well [48]. Interestingly, a phase I clinical trial of an oncolytic adenovirus in humans with glioma also showed a lymphocytoid/plasma cell infiltrate in injected tumors [3]. Taken in conjunction with our data, this is suggestive evidence of a rapid host response to the oncolytic virus infection. The immunosuppressive action of CPA may thus provide an avenue to circumvent the host response and enhance viral oncolysis and the therapeutic effects of transgenes delivered by oncolytic viral agents. However, the fact that CPA can only partially rescue the decrease in luciferase signal suggests that either the effect of CPA on the immune response is partial or other factors also contribute to the rapid decrease in transgene expression from the oncolytic adenovirus. Such factors may include ineffective viral dissemination within the tumor caused by murine stroma or tumor hypoxia [40,41]. The design of oncolytic viruses or combination therapies that overcome these barriers, together with concomitant immune

786

suppression, is expected to lead to more effective oncolytic viral therapies.

MATERIALS

AND

METHODS

Cell culture. The Ad5 E1-transformed human embryonal kidney cell line 293, and the human lung carcinoma cell line A549 were purchased from the ATCC (Manassas, VA, USA). The malignant glioma cell line U87DEGFR was a generous gift from Dr. H. J. Su Huang (University of California at San Diego, La Jolla, CA, USA). All cells were cultured in DMEM supplemented with 10% FBS and antibiotics (Life Technologies, Paisley, UK). Recombinant adenoviruses. A recombinant E1-deleted, replication-deficient adenovirus expressing the cytomegalovirus (CMV) immediate early promoter-driven firefly luciferase, Ad.CMV-Luc, was provided by Dr. R. D. Gerard (University of Texas Southwestern Medical Center, Dallas, TX, USA). Construction of the conditionally replicative adenovirus with an expression cassette for firefly luciferase in the E3 region, AdD24CMV-Luc, has been described previously [19]. Viruses were plaque purified, propagated on 293 cells (Ad.CMV-Luc) or on A549 cells (AdD24CMVLuc), and purified by CsCl gradient according to standard techniques. The E1D24 mutation and/or CMV-Luc insertion was confirmed by PCR on the final products. Particle titer was determined by absorbance measurements

MOLECULAR THERAPY Vol. 14, No. 6, December 2006 Copyright C The American Society of Gene Therapy

ARTICLE

doi:10.1016/j.ymthe.2006.08.008

at 260 nm and the functional titer (pfu) was determined by end-point dilution titration on 293 cells according to standard techniques. The particle/pfu ratio was 11 for AdD24CMV-Luc and 12.5 for Ad.CMV-Luc. In vitro assays. Glioma cells were plated subconfluently at 2.5  103 cells per well in 96-well plates. After 24 h culture medium was replaced by medium containing a dose range of 4-HO2-CPA (kindly provided by Dr. S. M. Ludeman at Duke University, Durham, NC, USA). In solution, 4-HO2CPA is spontaneously converted to 4-hydroxy-CPA, which is the first active metabolite formed in the metabolism of CPA. After another 24 h cells were infected in triplicate with a dose range of AdD24CMV-Luc. Viability of cells was assessed by WST-1 assay (Roche Diagnostics, Mannheim, Germany) at days 4 and 7 postinfection. Results are presented as percentage of uninfected controls. In parallel plates, luciferase expression in cell extracts was measured using the luciferase chemiluminescence assay system (Promega Benelux, Leiden, The Netherlands) on days 4 and 7 after infection. After the culture medium was removed, lysis buffer was added to the wells and the whole plate was snap-frozen on dry ice. After the samples thawed, the luciferase activity in the cell lysates was measured using a Lumat LB 9507 spectrofluorometer (EG&G Berthold, Bad Wildbad, Germany). In vivo tumor model. Intracranial glioma xenografts were established in adult female athymic mice (CD1 nude; Charles River, Wilmington, MA, USA) or NOD/SCID mice (NOD.SCID/NCr; NIH–National Cancer Institute, Frederick, MD, USA) by stereotactic injection of 105 U-87DEGFR cells in 3 Al HanksT buffer into the right frontal lobe, 2.5 mm lateral to the bregma, at a depth of 3 mm. Injections were performed under anesthesia by intraperitoneal injection of 2 mg ketamine and 0.4 mg xylazine in 0.9% saline. In the first experiment mice were inoculated stereotactically into the same coordinates as the tumor cells on day 5 post-tumor cell injection with a dose range of 106, 107, or 108 pfu AdD24CMV-Luc or with 107 or 108pfu replication-defective Ad.CMV-Luc in 3 Al PBS. Two mice per virus dose were monitored by bioluminescence imaging three times per week up to day 14. Animals were monitored daily and sacrificed upon appearance of symptoms evident for moribund decline due to intracerebral tumor growth, such as paralysis and lethargy. For analysis of viral spreading, an additional four animals were injected with 5  107 pfu AdD24CMV-Luc and sacrificed on day 3, 5, or 7 or at the time of appearance of symptoms. The brains of these animals were snap-frozen for sectioning and immunohistochemical analysis. For the second experiment, nude mice were randomly divided into two groups after receiving stereotactic tumor cell injection. On day 3, a single intraperitoneal injection of either PBS or 200 mg/kg CPA (Cytoxan; Mead Johnson) diluted in PBS was administered. On day 5 animals received 5  107 pfu AdD24CMV-Luc in 3 Al PBS inoculated stereotactically into the same coordinates as the tumor cells. Two animals from each group were sacrificed at 72 h and at 5 days postinfection. Brains were removed and frozen for (immuno)histological analysis. Bioluminescence imaging was performed on the remaining four animals every other day up to day 16. Mice were monitored daily and sacrificed upon appearance of symptoms evident of moribund decline due to intracerebral tumor growth, such as paralysis and lethargy. Brains were removed and frozen for (immuno)histological analysis. For the third experiment in NOD/SCID mice, the same protocol for tumor cells, virus, and CPA injection as in the second experiment was applied. Bioluminescence imaging was performed on the PBS- (n = 6) and CPA- (n = 7) treated mice on days 1, 2, 4, and 9 after AdD24CMV-Luc injection. Mice were monitored daily and sacrificed on day 9 after imaging due to the appearance of symptoms evident of intracerebral tumor growth. Brains were removed and frozen for (immuno)histological analysis. Bioluminescence imaging. Imaging of AdD24CMV-Luc-induced luciferase expression in athymic mice was performed with a cryogenically cooled high-efficiency CCD camera system (Roper Scientific, Duluth, GA, USA) custom developed at MGH using protocols similar to those described previously [49,50]. Mice were injected intraperitoneally with 4.5 mg d-luciferin substrate (Molecular Imaging Products, Ann Arbor, MI, USA)

MOLECULAR THERAPY Vol. 14, No. 6, December 2006 Copyright C The American Society of Gene Therapy

and images were acquired 15 min after d-luciferin administration. A surface image of each animal was acquired by dim polychromatic illumination. The luciferase activity within the brain of each animal was then measured by recording photon counts in the CCD with no illumination. Imaging of NOD/SCID mice was performed similarly using the Night OWL LB981 bioluminescence imaging system (Berthold Technologies, TN, USA). After data acquisition, postprocessing and visualization were performed with a home-written program (CMIR-Image) with integrated image display and analysis suite. Regions of interest were defined and the sum and standard deviation of the photon counts within regions of interest were then calculated. For visualization purposes, the bioluminescence images were fused with the corresponding white light surface images as a transparent pseudocolor overlay. Statistical analysis between virus and virus + CPA groups at each time point was conducted with the two-tailed Student t test. Immunohistochemistry. Immunohistochemical staining for adenovirus hexon proteins and mouse macrophages and leukocytes was performed on acetone-fixed cryosections obtained from brains of U-87DEGFR tumorbearing mice. Briefly, immunohistochemical staining was performed for adenovirus using the goat antiadenoviral hexon protein antibody 1056 (Chemicon Europe, Hampshire, UK), for monocytes/macrophages using the rat anti-mouse CD68 antibody (Serotec, Oxford, UK), and for hematopoietic cells using the rat anti-mouse CD45 antibody to leukocyte common antigen (ITK, Uithoorn, The Netherlands). Negative controls were assessed by omitting the primary antibody. Bound primary antibodies were detected using HRP-conjugated rabbit anti-goat or rabbit antirat antibodies and exposure to the chromogen diaminobenzidine (DakoCytomation, Glostrup, Denmark), followed by counterstaining with MayerTs hematoxylin. Quantification of CD45 and CD68 immunopositivity was performed in four representative areas of the tumor using ImageJ software (NIH, http://rsb.info.nih.gov/ij). Statistical analysis. Data from the first in vivo and in vitro experiments are presented as means + SD. Data from the second and third in vivo experiments are shown as means F SEM. Statistical analysis between groups was conducted with the two-tailed Student t test.

ACKNOWLEDGMENTS The authors thank Ida van der Meulen-Muileman for technical assistance and Ted Graves for assistance in quantitative analysis of imaging data. This work was supported by the Dutch Cancer Society, Grant VU2002-2594; The Ter Meulen Fund of the Royal Netherlands Academy of Arts and Sciences and the VUMC Institute for Cancer and Immunology to M.L.M.L.; and NIH Grants P01 CA69246, NS41571, and R01 CA85139 to E.A.C. In vivo imaging experiments were supported in part by R24 CA92782 and P50 CA86355 to R.W. RECEIVED FOR PUBLICATION OCTOBER 3, 2005; REVISED AUGUST 7, 2006; ACCEPTED AUGUST 17, 2006.

REFERENCES 1. Kirn, D. (2001). Clinical research results with dl1520 (Onyx-015), a replication-selective adenovirus for the treatment of cancer: what have we learned? Gene Ther. 8: 89 – 98. 2. Chiocca, E. A. (2002). Oncolytic viruses. Nat. Rev. Cancer 2: 938 – 950. 3. Chiocca, E. A., et al. (2004). A phase I open-label, dose-escalation, multiinstitutional trial of injection with an E1B-attenuated adenovirus, ONYX-015, into the peritumoral region of recurrent malignant gliomas, in the adjuvant setting. Mol. Ther. 10: 958 – 966. 4. Markert, J. M., Parker, J. N., Gillespie, G. Y., and Whitley, R. J. (2001). Genetically engineered human herpes simplex virus in the treatment of brain tumours. Herpes 8: 17 – 22. 5. Boviatsis, E. J., et al. (1994). Long-term survival of rats harboring brain neoplasms treated with ganciclovir and a herpes simplex virus vector that retains an intact thymidine kinase gene. Cancer Res. 54: 5745 – 5751. 6. Wong, R. J., et al. (2004). Angiogenesis inhibition by an oncolytic herpes virus expressing interleukin 12. Clin. Cancer Res. 10: 4509 – 4516. 7. Tyminski, E., et al. (2005). Brain tumor oncolysis with replication-conditional herpes simplex virus type 1 expressing the prodrug-activating genes, CYP2B1 and secreted human intestinal carboxylesterase, in combination with cyclophosphamide and irinotecan. Cancer Res. 65: 6850 – 6857. 8. Freytag, S. O., Rogulski, K. R., Paielli, D. L., Gilbert, J. D., and Kim, J. H. (1998). A novel

787

ARTICLE

9.

10. 11.

12.

13.

14.

15. 16.

17.

18.

19.

20. 21.

22.

23.

24.

25. 26.

27. 28.

three-pronged approach to kill cancer cells selectively: concomitant viral, double suicide gene, and radiotherapy. Hum. Gene Ther. 9: 1323 – 1333. Sova, P., et al. (2004). A tumor-targeted and conditionally replicating oncolytic adenovirus vector expressing TRAIL for treatment of liver metastases. Mol. Ther. 9: 496 – 509. Bristol, J. A., et al. (2003). In vitro and in vivo activities of an oncolytic adenoviral vector designed to express GM-CSF. Mol. Ther. 7: 755 – 764. van Beusechem, V. W., van den Doel, P. B., Grill, J., Pinedo, H. M., and Gerritsen, W. R. (2002). Conditionally replicative adenovirus expressing p53 exhibits enhanced oncolytic potency. Cancer Res. 62: 6165 – 6171. Oosterhoff, D., Pinedo, H. M., Witlox, M. A., Carette, J. E., Gerritsen, W. R., and van Beusechem, V. W. (2005). Gene-directed enzyme prodrug therapy with carboxylesterase enhances the anticancer efficacy of the conditionally replicating adenovirus AdDelta24. Gene Ther. 12: 1011 – 1018. Conrad, C., et al. (2005). Delta24-hyCD adenovirus suppresses glioma growth in vivo by combining oncolysis and chemosensitization. Cancer Gene Ther. 12: 284 – 294. Lamfers, M. L., et al. (2005). Tissue inhibitor of metalloproteinase-3 expression from an oncolytic adenovirus inhibits matrix metalloproteinase activity in vivo without affecting anti-tumor efficacy in malignant glioma. Cancer Res. 65: 9398 – 9405. Fueyo, J., et al. (2000). A mutant oncolytic adenovirus targeting the Rb pathway produces anti-glioma effect in vivo. Oncogene 19: 2 – 12. Fueyo, J., et al. (2003). Preclinical characterization of the antiglioma activity of a tropism-enhanced adenovirus targeted to the retinoblastoma pathway. J. Natl. Cancer Inst. 95: 652 – 660. Lamfers, M. L., et al. (2002). Potential of the conditionally replicative adenovirus Ad5Delta24RGD in the treatment of malignant gliomas and its enhanced effect with radiotherapy. Cancer Res. 62: 5736 – 5742. Geoerger, B., et al. (2004). Oncolytic activity of p53-expressing conditionally replicative adenovirus AdDelta24-p53 against human malignant glioma. Cancer Res. 64: 5753 – 5759. Carette, J. E., Graat, H. C., Schagen, F. H., Abou El Hassan, M. A., Gerritsen, W. R., and van Beusechem, V. W. (2005). Replication-dependent transgene expression from a conditionally replicating adenovirus via alternative splicing to a heterologous spliceacceptor site. J. Gene Med. 7: 1053 – 1062. Ikeda, K., et al. (1999). Oncolytic virus therapy of multiple tumors in the brain requires suppression of innate and elicited antiviral responses. Nat. Med. 5: 881 – 887. Ikeda, K., et al. (2000). Complement depletion facilitates the infection of multiple brain tumors by an intravascular, replication-conditional herpes simplex virus mutant. J. Virol. 74: 4765 – 4775. Heise, C., Sampson-Johannes, A., Williams, A., McCormick, F., Von Hoff, D. D., and Kirn, D. H. (1997). ONYX-015, an E1B gene-attenuated adenovirus, causes tumorspecific cytolysis and antitumoral efficacy that can be augmented by standard chemotherapeutic agents. Nat. Med. 3: 639 – 645. You, L., Yang, C. T., and Jablons, D. M. (2000). ONYX-015 works synergistically with chemotherapy in lung cancer cell lines and primary cultures freshly made from lung cancer patients. Cancer Res. 60: 1009 – 1013. Bernt, K. M., Steinwaerder, D. S., Ni, S., Li, Z. Y., Roffler, S. R., and Lieber, A. (2002). Enzyme-activated prodrug therapy enhances tumor-specific replication of adenovirus vectors. Cancer Res. 62: 6089 – 6098. Yu, D. C., et al. (2001). Antitumor synergy of CV787, a prostate cancer-specific adenovirus, and paclitaxel and docetaxel. Cancer Res. 61: 517 – 525. Li, Y., et al. (2001). A hepatocellular carcinoma-specific adenovirus variant, CV890, eliminates distant human liver tumors in combination with doxorubicin. Cancer Res. 61: 6428 – 6436. Budzynski, W., and Radzikowski, C. (1994). Cytotoxic cells in immunodeficient athymic mice. Immunopharmacol. Immunotoxicol. 16: 319 – 346. Sedgwick, J. D., Schwender, S., Imrich, H., Dorries, R., Butcher, G. W., and ter Meulen, V. (1991). Isolation and direct characterization of resident microglial cells from the normal and inflamed central nervous system. Proc. Natl. Acad. Sci. USA 88: 7438 – 7442.

788

doi:10.1016/j.ymthe.2006.08.008

29. Shultz, L. D., et al. (1995). Multiple defects in innate and adaptive immunologic function in NOD/LtSz-scid mice. J. Immunol. 154: 180 – 191. 30. Nanda, D., Vogels, R., Havenga, M., Avezaat, C. J., Bout, A., and Smitt, P. S. (2001). Treatment of malignant gliomas with a replicating adenoviral vector expressing herpes simplex virus-thymidine kinase. Cancer Res. 61: 8743 – 8750. 31. Grill, J., et al. (2002). The organotypic multicellular spheroid is a relevant threedimensional model to study adenovirus replication and penetration in human tumors in vitro. Mol. Ther. 6: 609 – 614. 32. Rehemtulla, A., et al. (2000). Rapid and quantitative assessment of cancer treatment response using in vivo bioluminescence imaging. Neoplasia 2: 491 – 495. 33. Soling, A., and Rainov, N. G. (2003). Bioluminescence imaging in vivo—Application to cancer research. Expert Opin. Biol. Ther. 3: 1163 – 1172. 34. Liang, Q., Yamamoto, M., Curiel, D. T., and Herschman, H. R. (2004). Noninvasive imaging of transcriptionally restricted transgene expression following intratumoral injection of an adenovirus in which the COX-2 promoter drives a reporter gene. Mol. Imaging Biol. 6: 395 – 404. 35. Liang, Q., Dmitriev, I., Kashentseva, E., Curiel, D. T., and Herschman, H. R. (2004). Noninvasive adenovirus tumor retargeting in living subjects by a soluble adenovirus receptor–epidermal growth factor (sCAR–EGF) fusion protein. Mol. Imaging Biol. 6: 385 – 394. 36. Honigman, A., et al. (2001). Imaging transgene expression in live animals. Mol. Ther. 4: 239 – 249. 37. Cook, S. H., and Griffin, D. E. (2003). Luciferase imaging of a neurotropic viral infection in intact animals. J. Virol. 77: 5333 – 5338. 38. Luker, G. D., Bardill, J. P., Prior, J. L., Pica, C. M., Piwnica-Worms, D., and Leib, D. A. (2002). Noninvasive bioluminescence imaging of herpes simplex virus type 1 infection and therapy in living mice. J. Virol. 76: 12149 – 12161. 39. Yu, Y. A., et al. (2004). Visualization of tumors and metastases in live animals with bacteria and vaccinia virus encoding light-emitting proteins. Nat. Biotechnol. 22: 313 – 320. 40. Sauthoff, H., et al. (2003). Intratumoral spread of wild-type adenovirus is limited after local injection of human xenograft tumors: virus persists and spreads systemically at late time points. Hum. Gene Ther. 14: 425 – 433. 41. Hay, J. G. (2005). The potential impact of hypoxia on the success of oncolytic virotherapy. Curr. Opin. Mol. Ther. 7: 353 – 358. 42. Akli, S., et al. (1993). Transfer of a foreign gene into the brain using adenovirus vectors. Nat. Genet. 3: 224 – 228. 43. Davidson, B. L., Allen, E. D., Kozarsky, K. F., Wilson, J. M., and Roessler, B. J. (1993). A model system for in vivo gene transfer into the central nervous system using an adenoviral vector. Nat. Genet. 3: 219 – 223. 44. Parr, M. J., Wen, P. Y., Schaub, M., Khoury, S. J., Sayegh, M. H., and Fine, H. A. (1998). Immune parameters affecting adenoviral vector gene therapy in the brain. J. Neurovirol. 4: 194 – 203. 45. Thomas, C. E., Birkett, D., Anozie, I., Castro, M. G., and Lowenstein, P. R. (2001). Acute direct adenoviral vector cytotoxicity and chronic, but not acute, inflammatory responses correlate with decreased vector-mediated transgene expression in the brain. Mol. Ther. 3: 36 – 46. 46. Ichikawa, T., and Chiocca, E. A. (2001). Comparative analyses of transgene delivery and expression in tumors inoculated with a replication-conditional or -defective viral vector. Cancer Res. 61: 5336 – 5339. 47. Wakimoto, H., Fulci, G., Tyminski, E., and Chiocca, E. A. (2004). Altered expression of antiviral cytokine mRNAs associated with cyclophosphamideTs enhancement of viral oncolysis. Gene Ther. 11: 214 – 223. 48. Lowenstein, P. R., and Castro, M. G. (2003). Inflammation and adaptive immune responses to adenoviral vectors injected into the brain: peculiarities, mechanisms, and consequences. Gene Ther. 10: 946 – 954. 49. Tang, Y., Shah, K., Messerli, S. M., Snyder, E., Breakefield, X., and Weissleder, R. (2003). In vivo tracking of neural progenitor cell migration to glioblastomas. Hum. Gene Ther. 14: 1247 – 1254. 50. Shah, K., Tung, C. H., Yang, K., Weissleder, R., and Breakefield, X. O. (2004). Inducible release of TRAIL fusion proteins from a proapoptotic form for tumor therapy. Cancer Res. 64: 3236 – 3242.

MOLECULAR THERAPY Vol. 14, No. 6, December 2006 Copyright C The American Society of Gene Therapy