A novel E1A–E1B mutant adenovirus induces glioma regression in vivo


Malignant gliomas are the most frequently occurring primary brain tumors and are resistant to conventional therapy. Conditionally replicating adenoviruses are a novel strategy in glioma treatment. Clinical trials using E1B mutant adenoviruses have been reported recently and E1A mutant replication-competent adenoviruses are in advanced preclinical testing. Here we constructed a novel replication-selective adenovirus (CB1) incorporating a double deletion of a 24 bp Rb-binding region in the E1a gene, and a 903 bp deleted region in the E1b gene that abrogates the expression of a p53-binding E1B-55 kDa protein. CB1 exerted a potent anticancer effect in vitro in U-251 MG, U-373 MG, and D-54 MG human glioma cell lines, as assessed by qualitative and quantitative viability assays. Replication analyses demonstrated that CB1 replicates in vitro in human glioma cells. Importantly, CB1 acquired a highly attenuated replicative phenotype in both serum-starved and proliferating normal human astrocytes. In vivo experiments using intracranially implanted D-54 MG glioma xenografts in nude mice showed that a single dose of CB1 (1.5 × 108 PFU/tumor) significantly improved survival. Immunohistochemical analyses of expressed adenoviral proteins confirmed adenoviral replication within the tumors. The CB1 oncolytic adenovirus induces a potent antiglioma effect and could ultimately demonstrate clinical relevance and therapeutic utility.


Most malignant primary cerebral neoplasms are gliomas, which grow locally but are highly resistant to conventional treatment with surgery, radiotherapy and chemotherapy. In fact, the median survival of patients with malignant gliomas is 1–2 years (Maher et al., 2001). Gliomas are particularly prone to local recurrence, generally within 1–2 cm of the primary site. Their relentlessly inevitable regeneration produces neurologic dysfunction, ultimately culminating in patient death. Their intrinsic characteristics and surrounding milieu make human malignant gliomas suitable targets for replication-competent adenoviral-based therapeutic strategies. A potential advantage of oncolytic adenoviral vectors over conventional antitumor agents is that viral replication in tumor cells amplifies the input dose of virus, leading to its spread throughout the tumor. Gliomas consist of local tumor masses that never metastasize and are easily accessible by stereotactic techniques for delivering intratumoral treatments. The surrounding tumor milieu consists of arrested neurons and glial cell populations, and astrocytes and stem cells actively proliferating or having the potential for proliferation.

Although the adenoviral life cycle has not been completely elucidated, it clearly involves redirecting the biochemical machinery of host cells by viral gene products. This is particularly true for interactions between Rb and p53 with adenoviral E1A and E1B proteins, pivotal events in adenoviral replication (Levine, 1990). The discovery of complementary interactions between viral genes and the cellular pathways involved in tumorigenesis provided biologic justification for using replication-conditional viruses as anticancer agents. Viral E1A protein associates with Rb, resulting in the release and activation of the E2F transcription factor, inducing cell cycle progression. Cellular expression of E2F-dependent S-phase genes and viral DNA replication proteins promotes S-phase induction and cellular and viral DNA synthesis. To prevent apoptosis by unscheduled DNA synthesis, the E1B-55 kDa protein interferes with p53-mediated apoptosis and inhibits cellular replication by binding to p53. These observations eventually promulgated the design of E1b and E1a mutant tumor-selective adenoviruses with replication restricted in normal cells by functionally active p53 and Rb pathways, respectively (Alemany et al., 2000).

Our group was the first to introduce the Delta-24 adenoviral construct. Delta-24 replication is restricted to dividing cells or to Rb-inactive arrested cells (Fueyo et al., 2000). More recently, modifying the tropism of Delta-24 resulted in maximum infectivity and enhanced oncolysis in human glioma models with a paucity of adenoviral receptors (Fueyo et al., 2003).

Improved infectivity should safely progress along with an improved tumor-selective adenoviral phenotype. Here we report on a double mutant adenovirus constructed on the backbone of the mutant-E1a Delta-24 (Fueyo et al., 2000) adenovirus and incorporating mutant-E1b (Bischoff et al., 1996). Our data showed that this double mutant adenovirus (CB1) replicated efficiently in human glioma cells in vitro and in vivo. Treatment of glioma-bearing mice with intratumoral injections of CB1 resulted in significantly increased survival (P=0.01).


The CB1 conditionally replicating adenovirus

The CB1 double mutant adenovirus is characterized by partial deletions of the E1a and E1b adenoviral genes (Figure 1a). The E1a deletion comprises amino acids 122–129 (Whyte et al., 1988, 1989), the same area deleted in the Delta-24 adenovirus (Fueyo et al., 2000), abrogating the ability of the mutant E1A protein to bind Rb (Whyte et al., 1989; Fueyo et al., 2000). The E1a and E1b deletions were confirmed by polymerase chain reaction (PCR) amplification of the modified regions and subsequent enzyme digestion analyses, as depicted in Figure 1b. Sequencing analyses identified the mutations in the E1a and E1b regions (data not shown). The deletion of the E1b region from 2426 to 3328 bp prevents the expression of wild-type E1B-55 kDa protein in CB1-infected cells (Figure 1c).

Figure 1

CB1 and Delta-24 oncolytic adenoviruses. (a) Nonscaled schematic representation of the E1 region of the CB1 and Delta-24 adenoviruses. The 24 bp deletion in the E1a and the 903 bp deletion in the E1b genes are shown. E1a, dotted area; E1b, hatched area. (b) PCR amplification and enzyme digestion analyses of E1a (left) and PCR amplification of E1b (right) sequences confirming the presence of the deletions in Delta-24 and CB1 adenoviruses. The 24 bp deletion was confirmed by PCR amplification of the E1a region and subsequent BstXI digestion. As the deleted 24 bp fragment eliminated a BstXI restriction site, the digestion generated two, 710 and 207 bp, products. The E1b partial deletion was confirmed by PCR amplification. In this analysis, the reverse primer was complementary to the deleted sequence of E1b, a region resulting in lack of amplification of any product in the CB1 sequence. M, marker. (c) Western blot analyses of the E1B-55 kDa and E1A proteins in U-251 MG human glioma cells 24 h after mock infection or infection with Ad300, Delta-24, CB1, or UV-inactivated Ad300 (UVi) at an MOI of 100. Actin expression is shown as a loading control. Note the higher levels of E1A in cells infected with CB1 adenovirus. This is explained by the fact that one of the functions of E1B-55 kDa protein is to downmodulate E1A (Babiss et al., 1985; Pilder et al., 1986)

Anticancer effect in vitro

In this experiment human glioma cells were infected with CB1, wild-type, or ultraviolet (UV)-inactivated wild-type adenoviruses at doses of 1, 2, 5, and 10 multiplicities of infection (MOIs). Cell viability was first assessed by crystal violet assay and then quantified with MTT tests. Both analyses showed a consistent dose–response effect of CB1 on the three human glioma cell lines (Figure 2). Daily monitoring of the infected cultures by light microscopy also showed a time-related effect. CB1-infected cells demonstrated a weak cytopathic effect (CPE) at 3 days postinfection and a strong CPE at 8–9 days postinfection, as evidenced by detachment of the monolayer cultures and the formation of round, light-refractil cells (data not shown). Furthermore, when the CB1-induced CPE was monitored by staining the remaining attached cells with crystal violet at the end of the experiment, CB1 exhibited an increasing CPE with increasing MOIs in the cell lines tested (Figure 2a). MTT analyses further showed that the decreased viability observed in the crystal violet assays was highly reproducible and dose dependent in the three cell lines tested (Figure 2b). The oncolytic effect was noticeable with a 1-MOI dose and was higher than 75% with a dose of 10 MOIs on the seventh day after infection in each cell line. As expected, UV-inactivated adenovirus did not induce CPE in the infected cells. These experiments also showed that D-54 MG was the most resistant cell line to the oncolytic effect of CB1.

Figure 2

CB1 anticancer effect in vitro. (a) Monolayers of human glioma cells were infected at the indicated MOIs with CB1, Ad300, or UV-inactivated Ad300 (UVi). Viable cells were stained with crystal violet when 10 MOIs of CB1 produced more than 75% of the CPE. (b) In parallel, cell viability was analysed with an MTT colorimetric assay. Cells were infected at different MOIs of 0–10PFU/cell of CB1, 10 MOIs of UV-inactivated Ad300 (UVi), or they were mock infected, and cell viability was measured by MTT assay 7 days after infection. Data are means±s.d. of triplicate determinations, and are represented as cell viability relative to mock-treated cells (equal to 100%). Note that despite slight differences in sensitivity, the dose of 10 MOI resulted in more than 75% of the decreased viability in the three cell lines. These data are consistent with those depicted in Figure 2a

Viral replication profile in human glioma cells

To ascertain if the CPE was due to adenovirus replication, we performed a Tissue Culture Infection Dose50 (TCID50) replication assay. At 3 days after the infection of glioma cells with 1 MOI of CB1, Delta-24, wild-type adenovirus, or UV-inactivated wild-type adenovirus, media and cell lysates were collected and the titer of adenovirus was measured in 293 cells. As expected, no new viral progeny was detected in cells infected with UV-inactivated wild-type adenovirus. Two independent experiments showed that CB1 replicated in U-251 MG, D-54 MG, and U-373 MG cells. Furthermore, the results indicated that CB1 replicated with an efficiency similar to Delta-24 in the three cell lines (P>0.1; t-test, double sided). Thus, the CB1 and Delta-24 titers differed by only 0.4±0.1 orders of magnitude (Figure 3). Although the replication ability of Delta-24 has been already reported (Fueyo et al., 2000, 2003), to our knowledge, these experiments constitute the first demonstration that a double E1a/E1b mutant adenovirus is able to acquire a replication phenotype in cancer cells.

Figure 3

Replication efficiency of CB1 in human glioma cells. Results from cancer cells infected with CB1, Delta-24, or Ad300 at a dose of 1 MOI are shown. At 3 days after infection, the viral titers were determined by the TCID50 method, and expressed as PFU/ml. The means of two independent experiments are shown

Viral replication profile in serum-starved normal human astrocytes

To examine CB1's ability to acquire a replication phenotype in normal cells, normal human astrocytes (NHA) were grown as a low-confluency monolayer and arrested by serum starvation. The arrested cells were infected with CB1, Delta-24, wild-type, or UV-inactivated wild-type at an MOI of 10. Daily monitoring showed an increasing CPE in wild-type adenovirus-treated cells with a striking decrease in cell viability by day 10 after infection. In contrast, the CPE did not appreciably escalate in cells treated with CB1 or Delta-24 by day 10, the latest time point examined. Consequently, most cells retained their morphology and were attached by the end of the experiment, suggesting that the activities of the two mutant adenoviruses were greatly attenuated under these experimental conditions. These results were quantified by trypan blue-based viability assays and are shown in Figure 4a. There were no significant differences in the viability of cells infected with the UV-inactivated, and those cells infected with CB1 or Delta-24 adenoviruses (P>0.5; t-test, double sided). As expected, wild-type adenovirus induced a marked CPE by the end of the experiment.

Figure 4

CPE and viral replication assays on NHA. (a) Cell viability analyses of quiescent and actively dividing NHA. NHA were infected with CB1, Delta-24, Ad300, or UV-inactivated Ad300 (UVi) at an MOI of 10. At 10 days after infection cell viability was assessed by a trypan blue exclusion test. Data are means±s.d. of triplicate determinations, and are represented as cell viability relative to UVi-treated cells (equal to 100%). *P>0.5; +P<0.001 (t-test, double sided). (b) Replication efficiency of CB1 in NHA. Quiescent and replicating NHA were infected with CB1, Delta-24, or Ad300 at a dose of 1 MOI. After 3 days, the final viral titers were determined by the TCID50 method in 293 cells, and expressed as PFU/ml. Data are presented as the means of at least three independent experiments. *P>0.5; +P<0.001; **P<0.005 (t-test, double sided)

To determine virion production, serum-starved NHA were infected with 1 MOI of CB1, Delta-24, wild-type, or UV-inactivated wild-type adenoviruses. At 3 days postinfection viral titers were measured in TCID50 assays. The results showed that CB1 and Delta-24 replicated equally poorly (P>0.5, t-test, double sided) under these conditions with final titers of approximately two orders of magnitude lower than the initial dose. The results obtained with Delta-24 are consistent with our previous observations in quiescent human lung fibroblasts and normal astrocytes (Fueyo et al., 2000, 2003). In contrast, wild-type adenovirus replicates more efficiently than CB1 or Delta-24 (P<0.001; t-test, double sided), and the final titers were 50-fold higher than the initial dose (Figure 4b). As expected, UV-inactivated Ad300-infected cells did not induce new viral production. These experiments showed that CB1 did not efficiently replicate in nonproliferating normal cells.

Viral replication profile in proliferating NHA

To ascertain whether the double mutant adenovirus induced cell death in normal dividing cells, we infected actively dividing NHA with 10 MOI of CB1, Delta-24, wild-type adenoviruses, or UV-inactivated Ad300. Light microscopic monitoring of the cultures revealed that infection with wild-type and Delta-24 adenoviruses resulted in an increasing CPE that was conspicuous by day 5 and dramatic by day 10, the latest time point examined. In marked contrast, infection with CB1 resulted in a moderate CPE (approximately 25%) on day 10 postinfection, results that were quantified by trypan blue-based viability assays and depicted in Figure 4a. These experiments revealed an attenuated lytic phenotype of Delta-24 compared to wild-type adenoviruses (P<0.001, t-test, double sided). More importantly, infection of actively dividing NHA with CB1 resulted in significantly less cell death than infection with Delta-24 (P<0.001, t-test, double sided).

To determine if increased cell death reflected adenoviral replication, we infected actively dividing NHA with a dose of 1 MOI of CB1, Delta-24, wild-type or UV-inactivated wild-type adenovirus. Examination of the viral titers 3 days after infection revealed that Delta-24 and wild-type adenoviruses replicated strikingly better in proliferating than in quiescent astrocytes (P<0.001; t-test, double sided; and P<0.001; t-test, double sided, respectively). Notably, these experiments showed that the final titer of CB1 was slightly inferior to the initial dose (5 × 104 plaque-forming units (PFU)), signifying CB1's greatly attenuated replicative ability in a dividing normal cell population (Figure 4b). These experiments also show that the ability of CB1 adenovirus to acquire a replication phenotype in normal dividing cells is significantly impaired compared to Delta-24 (P<0.001; t-test, double sided). Interestingly, Delta-24 replicated equally well in glioma cells as in dividing NHA (P>0.5; t-test, double sided). Importantly, CB1 replicated 100–10 000 times greater in glioma cell lines than in dividing NHA (P<0.001; t-test, double sided).

Anticancer effect in vivo

We examined the therapeutic efficacy of CB1 in vivo by implanting xenografts of D-54 MG human glioblastoma multiforme cells into the right basal ganglia of athymic mice. We selected the D-54 MG cell line because it was the most resistant to CB1 in vitro. When D-54 MG tumors were treated with CB1, animal survival (mean=51.86 days) was significantly improved compared with UV-inactivated adenovirus treatment (mean=27.72 days) (P=0.01, log-rank test). Delta-24 also produced a significantly improved survival (mean=59.29 days) (P<0.001, log-rank test). The survival curves obtained from Delta-24 versus CB1 treatments were not significantly different (P=0.28, log-rank test) (Figure 5). These results confirmed the antiglioma effect of Delta-24 in vivo shown in previous data from our laboratory using subcutaneous D-54 MG xenografts. Those data revealed an antiglioma effect after intratumoral administration of single and multiple doses of Delta-24 (Fueyo et al., 2000). In addition, multiple doses of Delta-24 induced an anticancer effect in U-87 MG glioma model (Fueyo et al., 2003). These results with CB1 are the first known to demonstrate the antiglioma effect of a double mutant E1A/E1B adenovirus in vivo, and therefore in an intracranial model.

Figure 5

Survival analyses of glioma-bearing animals. Data are represented as a Kaplan–Meier survival curve from the day of D-54 MG intracranial implantation following intratumoral injection with CB1, Delta-24, or UV-inactivated adenovirus (UVi). The P-values (determined by log-rank test) show significant differences in survival between CB1 and control-treated animals

CB1 replication in vivo

We examined the expression of hexon protein, a structural adenoviral protein expressed during active replication of adenoviruses, to determine if CB1 could replicate in the human xenografts. Immunohistochemical analyses revealed hexon in the cytoplasm of the infected cells (Figure 6a), confirming adenoviral replication in vivo. In addition, cells counterstained with hematoxyline, showed ground, glass-like intranuclear viral inclusion bodies (Figure 6b). Neither hexon-positive cells nor inclusion bodies were observed in D-54 MG xenografts infected with UV-inactivated wild-type adenovirus (Figure 6c). The inclusion bodies together with the positive expression of hexon protein constituted strong evidence of adenoviral replication in the D-54 MG human glioma xenografts. As expected, due to the inability of human adenoviruses to replicate in mouse cells, we did not observe expression of late adenoviral genes in the tumor surrounding normal rodent cells.

Figure 6

Immunodetection of CB1 in vivo. (a) Immunohistochemistry analyses of hexon in a section of a D-54 MG xenograft infected with CB1. Hexon protein is detected in the cancer cells, indicating adenoviral replication in vivo (× 100). (N), necrosis. (b) Higher magnification showing the presence of adenoviral inclusion bodies in cells positive for hexon protein (× 400) (arrows). The morphology and staining of the inclusion bodies are characteristic and indicate the presence of adenoviral replication. (c) Section of a D-54 MG xenograft infected with UV-inactivated wild-type adenovirus stained for antihexon protein (× 400). As expected, cells are negative for hexon after immunostaining and there are no evident inclusion bodies


This work characterizes the antiglioma effect of CB1, a novel double E1a–E1b mutant oncolytic adenovirus, and demonstrates its antiglioma effect in vitro and in vivo. Since our laboratory characterized the oncolytic effect of the Delta-24 adenovirus (Fueyo et al., 2000), several publications from other laboratories corroborated its ability to infect and replicate in cancer cells (Cripe et al., 2001; Suzuki et al., 2001; Bauerschmitz et al., 2002; Lamfers et al., 2002; van Beusechem et al., 2002). The downside is that E1A mutants such as Delta-24 are able to replicate in cycling normal cells (Heise et al., 2000) and Delta-24 promoted a widespread CPE in organotypic keratinocyte cultures (Balague et al., 2001). Clearly, the E1A-based Delta-24 oncolytic strategy needed improvement to decrease toxicity in normal dividing cells while maintaining efficacious replication and lysis in cancer cells. Consequently, we impaired Delta-24's ability to replicate in normal dividing cells by including an E1b deletion in the adenoviral genome. Attenuating the function of two critical adenovirus proteins in CB1 makes the construct less efficient at replicating in normal dividing cells. Our data specifically show that the replication phenotype of CB1 is highly attenuated both in nondividing and dividing astrocytes. Although the effect of CB1 was no different from Delta-24 in nondividing astrocytes, in actively proliferating astrocyte cultures CB1 produced significantly less toxicity than Delta-24. Importantly, similar differences were observed in replication studies of CB1, which showed that the end amount of viral progeny did not significantly differ from the titer of CB1 used to infect the astrocytes in the first place. The avid pursuit of much needed strategies for greatly enhanced adenoviral infectivity (Suzuki et al., 2001; Lamfers et al., 2002; Fueyo et al., 2003) underlines the relevance of adenoviruses like CB1, because tropism-enhanced constructs will succeed only if a high degree of infectivity is coupled with a minimal replicative ability in a dividing cell population. Clearly, CB1 and constructs with similar restrictions on their replication phenotype are preferable for systemic delivery where the oncolytic agent unavoidably infects actively dividing normal cells as well as the targeted cell population. Multicomponent targeting is an optimal strategy for effectively generating efficacious, tumor-selective lytic adenoviruses that are designed to be delivered to extratumoral sites without producing unacceptable levels of viral replication and toxicity in untargeted, normal cells.

CB1 is the first adenovirus construct to encompass deletions of E1A and E1B proteins. In earlier work, E1a was partially deleted so that Rb-binding region deletions were combined with the deletion of the p300-binding domain. However, the requisite number of deletions in the E1A sequence required to bind p300 results in low E1A activity (Yoshida et al., 1995), and some double E1A mutant adenoviruses replicated with a low efficiency in cancer cells (Howe et al., 2000). To circumvent these restrictions, E1A and E1B protein modifications have a greater potential for producing anticancer activity than the double E1A strategy previously embraced. Using such reasoning, Ramachandra et al. (2001) generated a conditionally replicating adenovirus that takes advantage of the wild-type p53 status of normal cells to express an E2F antagonist. Their replication-competent adenovirus did not replicate in normal cells, and demonstrated an anticancer effect that was independent of the status of p53 in cancer cells. Importantly, of the transactivation pathways through which E1A acts, that involving CR3 is the most potent, and most of the activation of viral early genes and of exogenous cellular genes by E1A is brought about by this region (Shenk and Flint, 1991). This region alone is sufficient to activate transcription (Lillie et al., 1987). Relevant to our study, this region is intact in the CB1 construct. Other strategies used to target the E2F1 pathway are based on modifying the transcriptional control of adenoviral genes, including E1a (Johnson et al., 2002; Tsukuda et al., 2002; Jakubczak et al., 2003). The tumor selectivity of these strategies, at least conceptually, is not expected to produce superior results to those from Rb-targeted deleted E1A mutants.

Taken collectively our data showed that it is feasible to diminish the ability of E1A mutant adenoviruses to replicate in normal cells by combining E1A and E1B mutations. We also proved that these double mutants possessed an antiglioma effect in vivo that was similar to that produced by E1A single mutants. These data form the basis for further development of E1A/E1B mutants with the ultimate goal of entry into clinical trials.

Materials and methods

Cell lines and culture conditions

U-373 MG and U-251 MG human glioma cell lines were purchased from ATCC (Manassas, VA, USA). The D-54 MG human glioma cell line was a generous gift from Dr Bigner (Duke University, NC, USA). 293 cells were obtained from Microbix Biosystem Inc. (Ontario, Canada). These cell lines were maintained in Dulbecco's modified Eagle/F12 medium (1 : 1, vol : vol) supplemented with 10% fetal bovine serum (FBS) in a humidified atmosphere containing 5% CO2 at 37°C. The NHA cell line is commercially available from Clonetics/BioWhittaker (Wakersville, MD, USA). NHA cultures were maintained using an Astrocyte Growth Medium BulletKit from Clonetics/BioWhittaker.

Adenoviral constructs and infection conditions

The previously described Delta-24 construct has a 24-bp deletion of the E1a region (nt 923–946, both included), corresponding to amino acids L122TCHEAGF129, a region required for Rb protein binding (Fueyo et al., 2000).

Combining the E1a deletion with an E1b-55K deletion created a double deletion in the CB1 virus. Briefly, the deletion in E1b was made by deleting a fragment ranging from the Sau3AI restriction site to the BglII restriction site (2426–3328 bp, both included) in pXC1 (Microbix) originating pXC1-RA55. This deletion eliminates the E1B p53-binding domain, sparing the E1b-19K region. To combine the double deletion, the KpnI/BglII fragment in pXC1-Delta-24, the shuttle vector containing the E1a deletion in Delta-24 (Fueyo et al., 2000), was replaced with the corresponding region in pXC1-RA55. The resulting shuttle vector (pXC1-CB1) was cotransfected with pBHG10 (Microbix) into 293 cells to generate the CB1 adenoviral construct through homologous recombination. Virus stocks were titered by plaque assay in 293 cells.

As controls, we used the wild-type adenovirus Ad300 (Jones and Shenk, 1979), virus inactivated by ultraviolet light (via exposure to seven cycles of 125J UV light), and mock infections with culture medium.

Infection of the cell lines has been described previously (Fueyo et al., 2000). Briefly, viral stock was diluted to specific concentrations, added to cell monolayers, and cells were incubated at 37°C for 30 min with brief agitation every 5 min. This was followed by the addition of culture medium and the return of the infected cells to the incubator at 37°C.

PCR and restriction analyses

Deletions in E1a and E1b were confirmed by PCR. A 941-bp fragment of the adenoviral E1a region was amplified by PCR using sense (5′-IndexTermTTCCGCGTTCCGGGTCAAAGTTG-3′) and antisense (5′-IndexTermTGCATTCTCTAGACACAGGTGATG-3′) primers. The PCR products were subjected to BstXI (Life Technologies, Inc., Rockville, MD, USA) restriction enzyme digestion at 37°C for 1 h. The BstXI enzyme recognizes a restriction site within the targeted 24-bp E1a region. The fragments and 1-kb ladder marker (Life Technologies, Inc.) were resolved on a 2% agarose gel (Sigma, St Louis, MO, USA) and visualized by UV fluorescence. To confirm the E1b deletion, PCR was performed using sense (5′-IndexTermGAATGAATGTTGTACAGGTGGCT-3′) and antisense (5′-IndexTermAGGAAAACCGTACCGCTAAAATTG-3′) primers, which amplified a 536-bp region of the wild-type E1b region. The antisense primer targeted a DNA sequence that is only present in wild-type E1b. Samples were analysed with agarose (Sigma, St Louis, MO, USA) gel electrophoresis.

Western blotting analyses

U-251 MG cells were infected with 100 MOIs of CB1, Delta-24, Ad300, UV-inactivated Ad300, or were mock infected. After 24 h, cells were collected and resuspended in PBS with a protease inhibitor cocktail (Sigma, St Louis, MO, USA), then lysed by adding equal volume of 2 × SDS loading buffer. The lysates were then heated at 95°C for 10 min. The proteins from the lysates were fractionated by SDS–PAGE and transferred to a Hybond ECL nitrocellulose membrane (Amersham Life Science, Arlington Heights, IL, USA). Membranes were probed with the following antibodies: adenovirus 5 E1B-55 kDa (Ab-1; diluted 1 : 100), hexon (diluted 1 : 100) (Chemicon International Inc., Temecula, CA, USA), and anti-human β-actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA). The secondary antibodies were horseradish peroxidase-conjugated anti-goat or anti-mouse IgG antibodies (Santa Cruz Biotechnology). The membranes were developed according to Amersham's ECL protocol.

Cell viability assays

We seeded cells at 105 cells per well in six-well plates and allowed them to grow for 20 h. The cells were infected with CB1, Ad300 or UV-inactivated Ad300 at MOIs of 0, 1, 2, 5 and 10. We concluded the experiment when an MOI of 10 of one of the constructs produced more than 75% of the CPE. The cell monolayers were then washed with PBS, and fixed and stained with 0.1% crystal violet in 20% ethanol. Several rinses with water were used to remove excess dye.

In vitro cytotoxicity was quantified with MTT (Sigma) to measure cell viability. For this assay, 2 × 103 cells were seeded in 96-well microtiter plates and infected 24 h later with 0, 1, 2, 5 or 10 MOIs of CB1, 10 MOIs of UV-inactivated Ad300, or cells were mock infected. Quadruplicate wells were used for each condition. A total of 16 wells were seeded with untreated glioma cells as a viability control, and 16 wells containing only complete medium were used as a control for nonspecific dye reduction. Medium was removed 7 days after treatment, and 100 μl/well of MTT solution (2 mg/ml) was added to each well. The plates were incubated for an additional 4 h, then read on a microplate reader at a test wavelength of 570 nm.

Trypan blue experiments were performed as previously described (Fueyo et al., 2000). Briefly, NHA cultures were infected with the adenoviral constructs at MOI of 10, and 10 days later the viable cells were counted using a hemocytometer.

Viral replication assays

We seeded human glioma and NHA cells at 5 × 104 cells/well in six-well plates and 20 h later infected them with CB1, Delta-24, Ad300, or UV-inactivated Ad300 at an MOI of 1. At 3 days after infection, we scraped the cells into culture medium and lysed them with three cycles of freezing and thawing. We used the TCID50 method to determine the final viral titration. Briefly, the cell lysates were clarified by centrifugation and the supernatants were serially diluted in medium for infecting 293 cells in 96-well plates. We analysed the cells for CPE 10 days after infection. Final titers were determined as PFU, using the validation method developed by Quantum Biotechnology (Carlsbad, CA, USA).

For the viral replication assay in quiescent NHA, we grew NHA cells at a low density (2 × 104/per well in a six-well plate) in the kit's medium with 0.5% FBS and without growth supplements. These conditions inhibited cell growth without evidence of cell death.

Animal model

To study the antitumor effect of the virus in vivo we used an intracranial human glioma xenograft model. A screw-guided system was used to implant tumor cells and to treat the intracranial tumor (Lal et al., 2000). The system consists of a 2.6-mm guide screw with a central 0.5-mm diameter hole that accepts the 26-gauge needle of a Hamilton syringe. The screw is implanted into a small drill hole made 2.5 mm lateral and 1 mm anterior to the bregma. A stylet is used to cap the screw between treatments. Tumor cells or therapeutic agents are injected in a freehand fashion by using a Hamilton syringe and a 26-gauge needle fitted with a cuff to determine the depth of injection. In this study, 5 × 105 cells of the D-54 MG human glioma cell line were engrafted into the caudate nucleus of athymic mice. On day 3 after cell implantation, animals were treated with a single intratumoral injection of CB1, Delta-24, or UV-inactivated adenovirus (n=7 per treatment group). The viral dose was 1.5 × 108 viral particles in 5 μl. Animals showing general or local symptoms of toxicity were killed. Animal studies were performed in the veterinary facilities of The University of Texas MD Anderson Cancer Center in accordance with institutional guidelines.


The presence of adenoviral hexon proteins in the treated xenografts was assessed through immunohistochemistry. Paraffin-embedded sections from the mice brains were deparaffinized and rehydrated with xylenes and ethanol in PBS. Endogenous peroxidase activity was quenched by incubation for 30 min in 0.3% H2O2 in methanol. Sections were treated with goat antihexon (Chemicon Inc., Temecula, CA, USA). Immunohistochemical staining was performed according to the manufacturer's instructions with diaminobenzidine by using Vector Laboratories ABC kits (Amersham).

Statistical analyses

For the in vitro experiments, statistical analyses were performed using a two-tailed Student's test. Data are represented as mean±standard deviation. The in vivo anticancer effect of different treatments was assessed by plotting survival curves according to the Kaplan–Meier method, and groups were compared using the log-rank test.


  1. Alemany R, Balague C and Curiel DT . (2000). Nat. Biotechnol., 18, 723–727.

  2. Babiss LE, Ginsberg HS and Darnell Jr JE . (1985). Mol. Cell. Biol., 5, 2552–2558.

  3. Balague C, Noya F, Alemany R, Chow LT and Curiel DT . (2001). J. Virol., 75, 7602–7611.

  4. Bauerschmitz GJ, Lam JT, Kanerva A, Suzuki K, Nettelbeck DM, Dmitriev I, Krasnykh V, Mikheeva GV, Barnes MN, Alvarez RD, Dall P, Alemany R, Curiel DT and Hemminki A . (2002). Cancer Res., 62, 1266–1270.

  5. Bischoff JR, Kirn DH, Williams A, Heise C, Horn S, Muna M, Ng L, Nye JA, Sampson-Johannes A, Fattaey A and McCormick F . (1996). Science, 274, 373–376.

  6. Cripe TP, Dunphy EJ, Holub AD, Saini A, Vasi NH, Mahller YY, Collins MH, Snyder JD, Krasnykh V, Curiel DT, Wickham TJ, DeGregori J, Bergelson JM and Currier MA . (2001). Cancer Res., 61, 2953–2960.

  7. Fueyo J, Alemany R, Gomez-Manzano C, Fuller GN, Khan A, Conrad CA, Liu TJ, Jiang H, Lemoine MG, Suzuki K, Sawaya R, Curiel DT, Yung WK and Lang FF . (2003). J. Natl. Cancer Inst., 95, 652–660.

  8. Fueyo J, Gomez-Manzano C, Alemany R, Lee PS, McDonnell TJ, Mitlianga P, Shi YX, Levin VA, Yung WK and Kyritsis AP . (2000). Oncogene, 19, 2–12.

  9. Heise C, Hermiston T, Johnson L, Brooks G, Sampson-Johannes A, Williams A, Hawkins L and Kirn D . (2000). Nat. Med., 6, 1134–1139.

  10. Howe JA, Demers GW, Johnson DE, Neugebauer SE, Perry ST, Vaillancourt MT and Faha B . (2000). Mol. Ther., 2, 485–495.

  11. Jakubczak JL, Ryan P, Gorziglia M, Clarke L, Hawkins LK, Hay C, Huang Y, Kaloss M, Marinov A, Phipps S, Pinkstaff A, Shirley P, Skripchenko Y, Stewart D, Forry-Schaudies S and Hallenbeck PL . (2003). Cancer Res., 63, 1490–1499.

  12. Johnson L, Shen A, Boyle L, Kunich J, Pandey K, Lemmon M, Hermiston T, Giedlin M, McCormick F and Fattaey A . (2002). Cancer Cell, 1, 325–337.

  13. Jones N and Shenk T . (1979). Cell, 17, 683–689.

  14. Lal S, Lacroix M, Tofilon P, Fuller GN, Sawaya R and Lang FF . (2000). J. Neurosurg., 92, 326–333.

  15. Lamfers ML, Grill J, Dirven CM, Van Beusechem VW, Geoerger B, Van Den Berg J, Alemany R, Fueyo J, Curiel DT, Vassal G, Pinedo HM, Vandertop WP and Gerritsen WR . (2002). Cancer Res., 62, 5736–5742.

  16. Levine AJ . (1990). Bioessays, 12, 60–66.

  17. Lillie JW, Loewenstein PM, Green MR and Green M . (1987). Cell, 50, 1091–1100.

  18. Maher EA, Furnari FB, Bachoo RM, Rowitch DH, Louis DN, Cavenee WK and DePinho RA . (2001). Genes Dev., 15, 1311–1333.

  19. Pilder S, Moore M, Logan J and Shenk T . (1986). Mol. Cell. Biol., 6, 470–476.

  20. Ramachandra M, Rahman A, Zou A, Vaillancourt M, Howe JA, Antelman D, Sugarman B, Demers GW, Engler H, Johnson D and Shabram P . (2001). Nat. Biotechnol., 19, 1035–1041.

  21. Shenk T and Flint J . (1991). Adv. Cancer Res., 57, 47–85.

  22. Suzuki K, Fueyo J, Krasnykh V, Reynolds PN, Curiel DT and Alemany R . (2001). Clin. Cancer Res., 7, 120–126.

  23. Tsukuda K, Wiewrodt R, Molnar-Kimber K, Jovanovic VP and Amin KM . (2002). Cancer Res., 62, 3438–3447.

  24. van Beusechem VW, van den Doel PB, Grill J, Pinedo HM and Gerritsen WR . (2002). Cancer Res., 62, 6165–6171.

  25. Whyte P, Ruley H and Harlow E . (1988). J. Virol., 62, 257–261.

  26. Whyte P, Williamson N and Harlow E . (1989). Cell, 56, 67–70.

  27. Yoshida K, Higashino F and Fujinaga K . (1995). Curr. Top. Microbiol. Immunol., 199, 113–130.

Download references


We thank Joann Aaron for editorial assistance (Department of Neuro-Oncology, MD Anderson Cancer Center). This work was supported by the Pediatric Brain Tumor Foundation of the United States, The University of Texas MD Anderson Cancer Center, the Anthony Bullock Foundation and the National Institutes of Health R01CA90879.

Author information



Corresponding author

Correspondence to Juan Fueyo.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Gomez-Manzano, C., Balague, C., Alemany, R. et al. A novel E1A–E1B mutant adenovirus induces glioma regression in vivo. Oncogene 23, 1821–1828 (2004). https://doi.org/10.1038/sj.onc.1207321

Download citation


  • glioma
  • therapy
  • E1A
  • E1B
  • adenovirus

Further reading


Quick links