Evaluation of adenovirus vectors containing serotype 35 fibers for tumor targeting


There is growing evidence from in vitro studies that subgroup B adenoviruses (Ad) can overcome the limitations in safety and tumor transduction efficiency seen with commonly used subgroup C serotype 5-based vectors. In this study, we confirm that the expression level of the B-group Ad receptor, CD46, correlates with the grade of malignancy of cervical cancer in situ. We also demonstrate the in vivo properties of Ad5-based vectors that contain the B-group Ad serotype 35 fiber (Ad5/35) in transgenic mice that express CD46 in a pattern and at a level similar to humans. Upon intravenous and intraperitoneal injection, an Ad5/35 vector did not efficiently transduce normal tissue, but was able to target metastatic or intraperitoneal tumors that express CD46 at levels comparable to human tumors. When an oncolytic Ad5/35-based vector was employed, in both tumor models antitumor effects were observed. Furthermore, injection of Ad5/35 vectors into CD46 transgenic mice caused less innate toxicity than Ad5 vectors. Our data demonstrate that Ad vectors that target CD46 offer advantages over Ad5-based vectors for treatment of cancer.


There are more than 50 serotypes of human adenoviruses (Ad); many of them differ in their tissue tropism.1 Ad serotype 5 (Ad5) has been intensively studied over the past 20 years and has been modified as a gene transfer vector. Key features that make Ad vectors an attractive vehicle for gene transfer in vitro and in vivo include the ability to easily prepare high-titer stocks of purified virus and the remarkable efficiency of each step in the Ad cell/nucleus entry process leading to high-level gene expression. Attachment of Ad5 to the primary cell surface receptor, the coxsackievirus-adenovirus receptor (CAR), is mediated by the fiber protein (for a review see Law and Davidson2). After binding, Arg-Gly-Asp (RGD) motifs in the penton base interact with cellular integrins, which function as secondary Ad5 receptors.3 This interaction triggers cellular internalization. The efficiency of Ad5 infection in vitro depends on the level of CAR and integrin expression. Recently, relative resistance of malignant tumor cells to infection with Ad5-based vectors has been connected to downregulation of CAR and/or integrin expression on target cell membranes.4, 5, 6, 7 Group B Ad serotypes (Ad3, 7, 11, 14, 16, 21, 34, 35, 50) use a receptor other than CAR. We and others recently identified CD46 as a cellular receptor that is used by most B-group Ads.8, 9, 10 In humans, CD46 is expressed on all nucleated cells at a low level. However, RNA and protein studies with biopsy material have shown that CD46 expression is greatly upregulated in malignant tumor cells, including breast, colon, liver and endometrial cancers.11, 12, 13 This makes CD46-targeting vectors potential tools for tumor gene therapy.

In addition to problems that limit the efficacy of Ad5 vectors for tumor targeting, intravenous (i.v.) infusion of Ad5 into animals causes systemic toxicity characterized by complement activation, cytokine release and consequent vascular damage that can result in a systemic inflammatory response (for a review see Muruve14). Recent studies indicate that the severity of innate toxicity can be enhanced by pre-existing Ad5 immunity.15

To target tumor cells through CD46, we plan to use vectors possessing fibers derived from B-group Ad such as serotype 35 (Ad5/35). Recent data show that Ad5/35 vectors efficiently transduce human cell types that are relatively refractory to Ad5 infection, including primary tumor cells.16, 17 Furthermore, we demonstrated that i.v. injected Ad5/35 vectors transduce Kupffer cells less efficiently than Ad5 vectors and therefore cause significantly less innate toxicity in mice18 and baboons.19

Whereas the CD46 gene sequences from humans and monkeys are highly homologous, the homology of rodent CD46 to human CD46 is less than 40%. Furthermore, whereas baboons and cynomolgus monkeys express CD46 in a pattern similar to humans,20 CD46 expression in rodents is restricted to the testis. This clearly limits the value of in vivo studies with Ad5/35 in normal mice.21, 22, 23, 24, 25, 26 From other areas of research (in particular, studies with measles virus), a number of CD46 transgenic mouse strains are available. In this study, we used a C57Bl/6-based strain that ubiquitously expresses CD46 at levels comparable to humans.27 To study tumor targeting in these mice with Ad5/35 vectors, we modified a C57Bl/6-derived tumor cell line to overexpress human CD46. We analyzed biodistribution, safety and tumor targeting in CD46 transgenic mice upon i.v. and intraperitoneal (i.p.) delivery of Ad5/35 vectors.

Materials and methods


TC1-CD46 cells: TC1 cells (syngeneic to C57Bl/6 mice) (ATCC# CRL-2785) were used in these studies. Cells were maintained in RPMI 1640+10% fetal bovine serum and supplemented with L-glutamine, penicillin and streptomycin. A plasmid containing the cDNA for CD46-C2 under the control of the CMV promoter pCD4628 as well as a puromycin transferase expression cassette was transfected into TC1 cells using Lipofectamine-Plus (Invitrogen, Carlsbad, CA) and clones were selected in the presence of 7.5 μg/ml of puromicin. Fifty clones were analyzed for CD46 expression by flow cytometry with anti-CD46 antibodies (J4-48; Beckman Coulter, Fullterton, CA). The clone with the highest CD46 levels was expanded and used for all subsequent experiments (TC1-CD46). The copy number of CD46 mRNA in total RNA from TC1-CD46 cells/tumors and liver tissue was determined by Real-time quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR), using the primers F: 5′-IndexTermTGTTTGGGTCATTGCTGTGATT-3′, R: 5′-IndexTermATCTGTACGGGACAACACAAATTACT and PCR conditions: 95°C 10 min and 40 cycles of (95°C 15 s, 60°C 1 min). Primary tumor cells: Cervix biopsies were microdissected by an experienced pathologist and tumor tissue and normal tissue were separated. Tumor tissue and normal tissue were digested with collagenase/trypsin/DNase as described in an earlier study17 to obtain cell suspensions. Cells were plated in keratinocyte growth medium (KGM) (Invitrogen, Carlsbad, CA), supplemented with 70 μg/ml bovine pituitary extract, 0.5 μg/ml hydrocortisone, 10 ng/ml epidermal growth factor, 5 μg/ml insulin, 0.1 mM phosphoethanolamine, Pen/Strep/fungizone and 10% fetal calf serum (FCS). Medium was changed for FCS-free medium 2 days later to suppress the outgrowth of fibroblasts.

CAR/CD46 flow cytometry

Immunofluorescence analysis of CAR and CD46 expression was performed by flow cytometry analysis using RmcAb against CAR29 or J4-48mAB against CD46 (Research Diagnostics, Flanders, NJ).


Ad5.bGal, Ad5.GFP (green fluoresencent protein), Ad5/35.bGal and Ad5/35.GFP are first-generation E1/E3-deleted adenovirus vectors expressing Escherichia coli β-galactosidase (bGal) or GFP.30 Ad5/35.IR-E1A/TRAIL is an oncolytic adenovirus that expresses Ad E1A and TRAIL (tumor necrosis factor-related apoptosis-inducing ligand) in a replication-dependent, tumor-specific manner.17 Viral vectors were produced and propagated following standard procedures.31 Viruses were banded in CsCl gradients, dialyzed and stored in aliquots as described previously. Titers were determined by plaque titration in 293 cells as described31 and the contamination level of wild-type (wt) virus was examined by real-time PCR as described.30 Only vector preparations that contained less than one wt viral genome in 106 genomes were used in these studies. All vectors were free of endotoxin.

Detection of CpG methylation

Methylation of a CpG island associated with the CAR gene promoter was evaluated by bisulfite genomic sequencing.32 CpG island sequences were obtained from the UCSC genome browser (http://genome.ucsc.edu/) and primers were designed using MethPrimer software. Primers were 5′-IndexTermGGTTAGGTTATTTTTAGGAAAGAGG-3′ (forward) and 5′-IndexTermACAAAAACCCACTTTTAATTTTCCAC-3′ (reverse). DNA isolated from cervical biopsies was first converted using the sodium bisulfite method and recovered in 100 μl TE (2 mM Tris-Cl, pH 8.0, 0.3 mM ethylenediaminetetraacetic acid). One microliter of this solution was used as a template for PCR in a Prism HT7900 instrument (Applied Biosystems Inc., Foster City, CA, USA) using 1 U AmpliTaq Gold (Applied Biosystems), in a mix containing 1.5 mM MgCl2, 62.5 μ M each dNTP and 150 nM each primer. Selected PCR products were cloned into pCR2.1 vector (Invitrogen) and sequenced using M13 primers.

Immunohistochemistry analyses on Papanicolaou smears

Papanicolaou (Pap) smears were collected from women who had been referred to the University of Dakar Tumor Institute from the community health clinics in the Dakar region, because of physical examination findings and/or symptoms suggestive of cervical cancer. All study procedures were approved by the institutional review boards of the University of Washington (Seattle, WA) and University of Dakar (Dakar, Senegal). Slides were prepared from residual ThinPrep Pap Test media (Cytyc Corp., Boxborough, MA). Endogenous peroxidase activity was blocked by incubation in 3% H2O2 in methanol. For CAR detection, the monoclonal antibody RmcB was used at a 1:5 dilution. The anti-CD46 monoclonal antibody clone, J4-48 (Beckman Coulter, Fullterton, CA), was used at a 1:50 dilution to detect the Ad35 receptor. Additionally, Ad5 and Ad35 receptors were also identified using biotinylated recombinant adenoviral knobs (100 μg/ml) as detection reagents. Ad5 and Ad35 knobs were produced in E. coli and purified as described elsewhere.33 The knob domains were dialyzed against 5 mM KCl, 10% glycerol and 10 mM MgCl2. Knobs were biotinylated using the Bio-Tag Biotinylation kit (Sigma-Aldrich, St Louis, MO). Binding of primary antibodies and knobs were detected with Envision+ (DAKO Corporation, Carpinteria, CA) and Elite ABC (Vector Laboratories, Burlingame, CA), respectively.


All experiments involving animals were conducted in accordance with the institutional guidelines set forth by the University of Washington. All mice were housed in specific pathogen-free facilities. C57Bl/6 mice (‘C57’) were obtained from Charles River (Wilmington, MA). The CD46 transgenic C57Bl/6 mice line MCP-8B (C57-CD46) was generously provided by Dr Branka Horvat (INSERM, Paris, France).27 This line has been crossed into the C57Bl/6 background for nine generations and these mice express CD46 at levels similar to human cells. The transgene is the CD46 C1 isoform under the control of the ubiquitously active hydroxymethyl-glutaryl coenzyme A reductase promoter. CD46 DNA-positive mice were identified by PCR on tail DNA using the following primers: F-CD46: 5′-IndexTermGGTCAAATGTCGATTTCCAGT-3′, R-CD46: 5′-IndexTermAATCACAGCAATGACCCAAA-3′. To establish mouse models with liver metastases, animals were infused with 2 × 106 TC1-CD46 cells through a permanently placed portal vein catheter as described elsewhere.34 To establish peritoneal tumors, animals were i.p. injected with 1 × 107 TC1-CD46 cells in 500 μl of phosphate-buffered saline (PBS). For i.v. application, Ad vectors in 100 μl of PBS were injected through the tail vein. Intraperitoneal injection was performed either in 2 ml of PBS or 2 ml of 4 or 15% icodextrin (a generous gift from ML Laboratories PLC, Liverpool, UK).

Quantitative PCR for vector genomes

Three days after Ad injection, mice were anesthetized, the heart was canulated and mice were exsanguinated and flushed with 20 ml PBS. Tissue samples were harvested under conditions that minimize cross-contamination. Organ samples were processed for histological analyses or snap frozen in liquid nitrogen. DNA was subsequently extracted from snap-frozen tissues and analyzed for viral DNA using primers specific for viral DNA. Samples were analyzed using a Lightcycler and the SYBR Green kit, and the following primers were used for viral DNA: F-L4: 5′-IndexTermTGCAAGATACCCCTATCCTG-3′, R-L4: 5′-IndexTermCCTGTTGCAGAGCGTTTGC-3′. Purified Ad DNA was used as external standards. Samples were equalized for DNA input using control primers against mouse GAPDH: F-mGAPDH: 5′-IndexTermATCACTGCCACCCAGAAGAC-3′, R-mGAPDH: 5′-IndexTermCACATTGGGGTAGGAACAC-3′.

Analysis of β-galactosidase and CD46 expression on liver sections

At 72 h after Ad5/35 injection into CD46 transgenic mice, organs were frozen in OCT compound (Miles, Elkhart, Indiana), sectioned (6 μm) and analyzed by staining for β-galactosidase activity using X-gal. The number of b Gal-positive cells located within tumors was counted using a microscopic grid. A total of 10 independent 1 mm2 areas were counted for each group. For CD46 immunofluorescence analyses on frozen liver sections, we used polyclonal anti-CVD46 antibodies from Santa Cruz (Santa Cruz, CA).


Serum samples were analyzed using Cytometric Bead Array technology, ‘Mouse Inflammatory Cytometric Bead Array’ (BD Biosciences, Palo Alto, CA). Briefly, 10 μl of mouse plasma was diluted five times and mixed with cytometric beads capable of binding mouse tumor necrosis factor (TNF)α, interleukin (IL)-6, monocyte chemoattractant protein-1 (MCP-1), IL-12 and interferon (IFN)-γ. Bound cytokines/chemokines were detected with corresponding secondary phycoerythrin-conjugated antibodies, and analyzed by flow cytometry along with provided standard proteins. The collected data were processed using the manufacturer's software.

Results and discussion

Studies on human tumor material

To validate the use of Ad5/35 vectors for gene transfer into tumor cells, we studied CD46 and CAR expression on tumors in situ, on Pap smears of cervical cancer patients, using immunohistochemistry with anti-CAR and anti-CD46 antibodies (Figure 1a and Supplementary Figure 1). In normal cervix and low-grade samples, CD46 expression was absent or low, whereas the majority of these samples stained positive for CAR. In contrast, all 16 samples from patients with invasive cervical cancer (ICC) stained positive for CD46 and none of these samples was positive for CAR. High-grade samples grouped between the normal and ICC samples, in regard to CD46 and CAR expression. We have performed Mantel–Haenszel tests for correlation of both CAR and CD46 expression with cytology results. This test is appropriate for looking at associations in ordered categories, such as cytology and level of staining.35 The P-value for the comparison of CD46 and cytology group is <0.0001. The interpretation of this is that CD46 staining level (neg, 1+, 2+, 3+) increases significantly with increasing level of cervical disease (from LSIL to HSIL to ICC). The P-value for the comparison of CAR and cytology group is 0.008, which implies that CAR staining level decreases significantly with increasing level of cervical disease.

Figure 1

Assessment of Ad5/35 vectors for transduction of tumor cells. (a) Left panel: Immunohistological staining for CAR and CD46 expression on Pap smears from patients with low-grade and high-grade/invasive cervical neoplasia with α-CAR and α-CD46 antibodies. In high-grade specimens, both malignant (M) and normal (N) cells are present. Right panel: Summary of studies on Pap smears from patients with low-grade squamous intraepithelial lesions (n=13), high-grade squamous intraepithelial lesions (n=15) or invasive cervical carcinoma (n=16). The signal was scored by an experienced pathologist in a blinded study. (+), (++) and (+++) represent increasing levels of positive staining. Each Pap smear was a different patient. (b) Detection of adenovirus knob binding on Pap smears. Biotinylated Ad5 and Ad35 knobs were added to Pap smears. Binding was developed with avidin–peroxidase conjugates. (c) Paraffin sections of human tissue were incubated with biotinylated Ad35 knobs as described in (b). Shown are sections from ICC, normal cervix and liver. Note positive (brown) staining in tumor-associated vascular endothelial cells. (d) Receptor expression and Ad transduction studies in primary cells cultures. Cells were isolated from biopsies of pre-malignant lesions (LGSIL) and ICC. CAR and CD46 expression was analyzed by flow cytometry (left panel). Cultures were transduced with Ad5.GFP or Ad5/35.GFP at an MOI of 10 PFU/cell and mean GFP fluorescence was measured 24 h later (right panel). N=10/group.

To better predict the infectivity of Ad5/35 in tumor patients, we studied the binding of recombinant Ad fiber knob, which is the Ad capsid moiety that mediates attachment to receptors on the cell surface of Pap smears. Figure 1b (and Supplementary Figure 1) shows that binding of recombinant Ad5 and Ad35 fiber knobs mirrors the expression patterns of CAR and CD46. Binding of Ad35 knobs was also studied on ICC sections. Ad35 knob bound to tumor cells and interestingly, to endothelial cells of tumor-feeding blood vessels but not of blood vessels present in non-malignant tissues (Figure 1c). Notably, the spectrum of surface markers in neoangiogenic vasculature differs from that of tissue blood vessels.36 We also studied Ad receptor expression and Ad transduction on primary cell cultures derived from cervix biopsies (Figure 1d). Overall, the level of CAR and CD46 on cultured cells correlated well with transduction efficiency of Ad5 and Ad5/35. The differences in CD46 expression between cells isolated from premalignant lesions and ICC-derived tumor cells were not as apparent as in situ, indicating that the regulation of CD46 expression changes if cells are taken out of the tissue context and cultured, a phenomenon also described for CAR expression.37 As with Pap smears, CAR expression was lower in cultures of malignant cells than pre-malignant cells. In an attempt to delineate the reasons for low CAR expression, we analyzed CAR mRNA and the methylation status of CpG islands present in the promoter of the CAR gene. CAR mRNA levels were lower in ICC biopsies than in normal cervix tissue (P=0.045) (Supplementary Figure 2a). For methylation studies, DNA was isolated from six cervical biopsies with normal histology and eight biopsies with ICC. Bisulfite genomic sequencing suggested that both normal and ICC samples contained all eight CpG dinucleotides in unmethylated form (Supplementary Figure 2b). This indicates that promoter/gene methylation is not involved in downregulation of CAR expression in ICC. In conclusion, our studies on Pap smears suggest that Ad5/35 vectors are able to preferentially target tumor cells, in contrast to commonly used Ad5 vectors. Clearly, in humans, other factors, most importantly transendothelial transport and accessibility of tumor cells in situ, will affect tumor cell infection upon i.v. Ad5/35 application.

Biodistribution of Ad5/35 after i.v. or i.p. vector injection into CD46 transgenic mice

In our in vivo studies with CD46 transgenic mice, we focused on application of Ad5/35 for metastatic or peritoneal tumors, which would require i.v. and i.p. vector application, respectively. In this context, we studied the biodistribution of transgene expression and vector genomes upon i.v. and i.p. injection of Ad5/35-bGal into C57-CD46 mice. Three days after i.v. injection of 5 × 109 plaque forming units (PFU), bGal expression was found in sparse hepatocytes, predominantly in the periportal regions, which are exposed to the highest dose of incoming virus (Figure 2a). Ad5/35 transduction was also found in the spleen, specifically in the marginal zone of the red pulpa, in CD11c+ cells that most likely represent antigen-presenting cells (data not shown). Only single bGal-positive cells were found in sections of lung, brain (Figure 2a), kidney and lymph nodes (data not shown). After i.p. injection of 5 × 109 PFU Ad5/35 bGal, transgene expression was found in the liver and spleen periphery (Figure 2b). Transduction of central liver and spleen regions was minimal, indicating that only a minor virus fraction was taken up into the blood circulation. The amount of viral genomes detected by qPCR in total DNA isolated from organs 72 h after Ad is shown in Figure 2c. The highest numbers of vector genomes were found in the liver and spleen, and about 10-fold less vector copies were found in all organs after i.p. injection than i.v. injection. Notably, qPCR also detects extracellular vector DNA, which accounts for the relatively high background levels observed. Although we tried to minimize the presence of extracellular or intravascular viral DNA by aggressively flushing the vascular system with normal saline at necropsy, these potential confounding factors have to be kept in mind when interpreting the results. Vector copy numbers measured by qPCR are therefore informative only in comparison.

Figure 2

Biodistribution of Ad-mediated transgene expression and vector genomes in C57-CD46 mice after i.v. and i.p. injection of Ad5/35.bGal. A total of 5 × 109 PFU were injected either i.v. or i.p. into C57-CD46 mice. Seventy-two hours post virus injection, organs were harvested and tissue samples were processed for immunohistochemistry or snap frozen for qPCR analyses. Tissue sections were stained with X-Gal, followed by hematoxylin. (a) Representative sections from mice i.v. injected with Ad5/35.bGal. (b) Representative sections of central and peripheral liver and spleen areas from mice that received i.p. Ad5/35.bGal injection. (c) Quantitative comparison of viral genomes present in major organs at 72 h after i.v. and i.p. injection of Ad5/35.bGal. Three independent tissue samples were analyzed and the mean and standard deviation are shown. The genome concentration was expressed as the number of viral genomes per 107 cells (assuming that the mass of a diploid human genome is 6 pg). qPCR results for vector genomes were equalized based on qPCR data for an endogenous (two copies/genome) mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene.

Serum levels of pro-inflammatory cytokines and chemokines after i.v. and i.p. Ad injection

The extent of innate toxicity upon Ad injection is well reflected in elevation of serum cytokine/chemokine levels. Previous studies in C57 mice showed lower TNFα, IL-6, IFNγ and MCP-1 levels upon i.v. injection of Ad5/35 vectors compared to Ad5 vectors or vectors that contained the long Ad5 fiber shaft and the Ad35 fiber knob (Ad5/35L).18 We have demonstrated that transduction of Kupffer cells (which are critically involved in mediating Ad toxicity) involves binding of Ad particles to blood factors and requires a long fiber shaft, but does not occur with Ad vectors containing short-shafted fibers such as Ad5/35.38, 39 Here we tested whether the presence of CD46 in mice affects the safety profile of Ad5/35 vectors (Figure 3). C57-CD46 mice were injected i.v. or i.p. with 5 × 109 PFU of Ad5-GFP, Ad5/35L-GFP or Ad5/35-GFP and serum cytokine/chemokine levels were measured 6 h later. Whereas after i.v. or i.p. injection of Ad5/35, levels were not (TNFα, IL-6) or only slightly (IFN-γ, IL-12, MCP-1) elevated above levels of mock-injected mice, Ad5 or Ad5/35L injection caused a significant increase in cytokine and chemokine levels.

Figure 3

Innate immune responses upon i.v. and i.p. injection of Ad5/35.GFP into C57-CD46 mice. Plasma levels of MCP-1, TNF-α, IFN-γ, IL-12 and IL-6 at 6 h post Ad injection. C57-CD46 mice were i.v. or i.p. injected with 5 × 109 PFU of Ad5/35.GFP (Ad5/35s). For comparison, mice received an i.v. injection of 5 × 109 PFU of Ad5.GFP (Ad5) or Ad5/35L-GFP (containing the long Ad5 shaft and the Ad35 knob). Plasma samples were collected and analyzed in duplicate by cytometric bead array for cytokine and chemokine levels. For Ad5 and Ad5/35L, the average and standard deviation are shown. For Ad5/35s, values of individual mice are shown. Star means that the levels were below the detection limit.

Tumor-specific infection after i.v. and i.p. injection of Ad5/35 vectors in C57-CD46 mice

To ultimately show that Ad5/35 is able to preferentially target tumors in CD46 transgenic mice, we modified a mouse tumor cell line to express human CD46 at levels comparable to human tumors. We used TC1 cells, which is an immortalized, C57Bl/6-derived, mouse epithelial cell line that expresses HPV E6 and E7 (as a model for cervical carcinoma and as an easy means to study antitumor immune responses). TC1 cells were modified to stably express 5 × 105 RNA copies per cell (TC1-CD46). Notably, human tumors express 105–106 CD46 molecules per cell.40 In vitro transduction studies showed that TC1-CD46 cells can be efficiently transduced with Ad5/35 vectors, whereas unmodified TC1 cells were refractory to Ad5/35 infection (Figure 4a).

Figure 4

Targeting to TC1-CD46 cell-derived liver metastases upon i.v. injection of Ad5/35.bGal into C57-CD46 mice. (a) In vitro transduction of TC1 and TC1-CD46 cells with Ad vectors. TC1 and TC1-CD46 cells were infected in vitro with Ad5.bGal and Ad5/35.bGal at the indicated MOI and bGal expression was analyzed 48 h later by X-Gal staining. Shown is the average of three independent experiments. The standard deviation was less than 10% for all samples. (b) Immunofluorescence analysis of CD46 expression on liver sections from C57-CD46 mice with TC1-CD46-derived metastases. A total of 2 × 106 TC1-CD46 cells were intraportally injected into C57-CD46 mice. Two weeks later, mice were killed and liver sections were analyzed using polyclonal anti-CD46 antibodies. Positive staining appears in green. The white line depicts the border between tumor and liver parenchyma. (c) b-Gal expression in TC1-CD46-derived liver metastases after i.v. injection of Ad5/35.bGal. Two weeks after intraportal TC1-CD46 injection, when liver metastases were established, mice received a tail vein injection of 5 × 109 PFU of Ad5/35.bGal. Three days after virus injection, livers were harvested and sections were stained with X-Gal (upper panel: magnification × 4; lower panel: magnification × 20). Representative sections are shown. (d) Targeting to TC1-CD46 cell-derived i.p. tumors upon i.p. injection of Ad5/35.bGal into C57-CD46 mice. A total of 107 TC1-CD46 cells in a volume of 1.0 ml of Dulbecco's modified Eagle's medium were injected i.p. Fourteen days later, when peritoneal tumors were detectable macroscopically, 5 × 109 PFU of Ad5/35.bGal in 2 ml of PBS, 4% icodextin, or 15% icodextrin were injected i.p. Seven days later, mice were killed and i.p. tumors as well as organs were harvested and sections were stained by X-Gal staining.

The growth kinetics of TC1-CD46 cells as metastatic tumors (upon intraportal tumor cell injection) and peritoneal tumors (upon i.p. injection) was comparable to that seen for tumors derived from the parental TC1 cells (data not shown), indicating that C57-CD46 mice were tolerant to human CD46 expressed on TC1 cells. Immunohistochemical analysis of liver sections from tumor-bearing mice with CD46-specific antibodies revealed intense CD46 staining of TC-CD46-derived liver metastases (Figure 4b). To quantify CD46 expression in the liver from CD46 transgenic mice, we measured CD46 mRNA levels in microdissected CD46-TC1 metastases and surrounding tumor tissue by qRT-PCR. The ratio of CD46 mRNA per gram tissue between tumor and liver was 57±13.

To analyze transduction of TC1-CD46-derived liver metastases with Ad5/35 vectors, mice bearing pre-established metastases were injected via the tail vein with 5 × 109 PFU of Ad5/35.bGal (Figure 4c). Three days later, bGal expression was found in all liver metastases with about 5% of all tumor cells transduced per given metastasis. The density of bGal-expressing cells in liver parenchyma was about 20-fold lower.

In a peritoneal tumor model, 14 days after i.p. injection of TC1-CD46, 5 × 109 PFU of Ad5/35.bGal were injected i.p. in PBS (Figure 4d, left column). Seven days later, transgene expression was found in about 1% of tumor cells. Single bGal-positive cells were found in the liver and spleen. To increase transduction of peritoneal tumors, we injected Ad5/35.bGal together with 4 or 15% icodextrin, a polysaccharide developed for peritoneal dialysis41 that is able to retain and protect the virus in the peritoneal cavity.42 Injection of Ad5/35 in PBS, 4% icodextrin or 15% icodextrin resulted in 15(±3), 25(±6) or 189(±45) β-galactosidase-positive cells per mm2, respectively (Figure 4d, right column). (The P-value for the PBS vs 4% icodextrin groups is 0.02.) Transduction of liver and spleen parenchyma was less with Ad5/35.bGal+15% icodextrin than Ad5/35.bGal+PBS.

Absence of serum factors that interfere with Ad5/35 infection

A recent study suggested that cultured tumor cells shed a soluble form of CD46, which can be detected in the culture supernatant.43 Along this line, Seya et al.44 reported increased concentrations of specific CD46 isoforms in cancer patient sera. Potentially, soluble CD46 present in serum could affect tumor targeting of Ad5/35 upon i.v. injection. To assess this, we performed in vitro transduction studies with Ad5/35 and TC1-CD46 cells in the presence of serum from mice with TC1-CD46 metastases. In these studies, we did not observe an inhibitory effect of serum on Ad5/35 infection. This observation is corroborated by previous studies with human sera, where we found that the CD46 interacting serotype Ad Ad11 infected 293 cells in the presence of sera from cancer patients with equal efficiency as in the presence of sera from normal donors.45 Notably, in enzyme linked immunosorbant assay studies, we did not detect sCD46 in serum from tumor-bearing mice or cancer patients (data not shown). It is possible that CD46 concentrations in our serum samples were below the detection limit and that sCD46 at this level did not critically affect Ad5/35 infection. This is further supported by the fact that Ad5/35.bGal was able to target CD46high liver metastases after i.v. injection into CD46 transgenic mice.

Antitumor effect of Ad5/35-based oncolytic adenoviruses

Our lab has previously developed oncolytic vectors based on Ad genomes deleted for E1A and E1B genes. Tumor cells but not normal cells support DNA replication of these vectors. This tumor-specific replication of viral genomes can be converted into tumor-specific transgene expression utilizing homologous recombination in Ad genomes (Ad.IR system). We have used the Ad.IR system to generate an Ad5/35-based oncolytic vector that tumor-specifically expresses Ad E1A and TRAIL (Ad5/35.IR-E1A/TRAIL) upon homologous recombination.46 We have previously shown that TC1 cells support DNA replication of Ad.IR vectors,47 which is probably owing to transcomplementation by HPV E6 and E7 expressed in TC1 cells. Efficient replication-activated transgene expression was also seen in TC1-CD46 cells upon Ad5/35.IR infection and infection with Ad5/35.IR-E1A/TRAIL at a multiplicity of infection (MOI) of 200 PFU/cell killed 100% of TC1-CD46 cells within 4 days. However, when we tested for PFU present in the lysates and culture supernatants, we found that the production of Ad5/35.IR-E1A/TRAIL progeny virus in TC1-CD46 cells was more than one order of magnitude less efficient than upon infection of human cervical carcinoma (HeLa) cells. This is in agreement with earlier studies showing that the late events in replication of human Ads in mouse cells are inefficient.48

The antitumor efficacy of Ad5/35.IR-E1A/TRAIL was tested in both the metastasis model (Figure 5a) and the i.p. tumor model (Figure 5b). I.v. injection of 5 × 109 PFU of Ad5/35.IR-E1A/TRAIL into C57-CD46 mice with pre-established TC1-CD46 metastases resulted in a significant reduction in tumor burden compared to mock or control virus injection (P=0.038). A significant anti-tumor effect on i.p. tumors was also observed when mice were i.p. injected with 5 × 109 PFU of Ad5/35.IR-E1A/TRAIL mixed with 15% icodextin (P=0.018; no Ad/15% icodextrin vs Ad5/35.IR-E1A/TRAIL/15% icodextrin). Notably, limited de novo virus production in mouse tumors implies limited spread and anti-tumor efficacy of Ad5/35.IR-E1A/TRAIL in TC1-CD46 tumors. Therefore, studies with this oncolytic vector in mouse tumor models most likely underestimate the efficiency of the vector in cancer patients.

Figure 5

Anti-tumor effect of Ad5/35.IR-E1A/TRAIL in mouse tumor models. (a) Liver metastasis model. A total of 5 × 109 PFU of Ad5/35.bGal (control) or Ad5/35.IR-E1A/TRAIL were injected into C57-CD46 mice 2 weeks after intraportal transplantation of 2 × 106 TC1-CD46 cells. Control mice received PBS injections (‘Mock’). Seven days later, mice were killed and tumors were microdissected under a stereomicroscope and weighted. The total tumor mass for individual mice is shown. (b) Mice with i.p. tumors were injected as described in Figure 4d. Seven days after Ad injection, mice were killed, tumors microdissected and the tumor burden for each mouse was measured.

The discovery that B-group viruses exploit a ubiquitous molecule as an attachment receptor may explain the broad tissue tropism of B-group viruses seen in vitro, and the diverse clinical manifestation of B-group virus infection. On the other hand, ubiquitous CD46 expression questions the tumor selectivity of Ad5/35 vectors after systemic vector administration. However, in agreement with previous biodistribution studies in mice and baboons,8, 19 we found that Ad5/35 only inefficiently transduced organs such as liver and lung which are efficiently targeted by Ad5 vectors. The following two factors might account for this observation: (i) the avidity of Ad5/35 to CD46 is relatively low and (ii) Ad5/35 transduction directly correlates with CD46 density on the cell surface, whereby cells expressing less than 100 CD46 molecules per cell cannot be transduced.40 The low levels of CD46 on normal cells might account for the inefficient transduction of normal tissues by Ad5/35. Previous studies from others and our lab suggest that the in vivo tropism of Ad vectors is not only determined by the levels of primary receptor (for a review see Nicklin et al.49), and that interaction with blood factors can target long-shafted Ads to Kupffer cells or hepatocytes independently of CAR or CD46 expression.50 Notably, the latter mechanism does not function for short-shafted Ad vectors such as the Ad5/35 vectors used in our study. Several lines of in vitro data point towards the existence of B-group Ad receptors other than CD46.45 A recent study in rat (non-transgenic for CD46) demonstrated Ad5/35 transduction of endothelial cells of blood vessels associated with rat liver tumors,51 indicating that Ad5/35-mediated transduction of tumor blood vessels might be CD46 independent. Although we detected binding of Ad35 fiber knob to human endothelial cells of tumor blood vessels on ICC sections, no transduction of tumor vasculature was found in mouse TC1-CD46 metastases. The clarification of the discrepancy between our studies in mice and Shinozaki's studies in rats, as well as the impact of additional B-group Ad rectors on in vivo tropism of Ad5/35, requires further research. Importantly, the biodistribution of transgene expression and vector genomes after i.v. injection of Ad5/35 vectors into C57-CD46 mice was comparable to that seen in baboons,19 where, as outlined earlier, there is high homology between human and monkey CD46, and the level and pattern of CD46 expression is similar. This further corroborates that C57-CD46 mice represent an adequate model for in vivo studies with Ad5/35 vectors. As tumor induction/transplantation and subsequent tumor targeting studies in monkeys are impossible for ethical reasons, our finding that an Ad5/35 vector is able to target CD46high liver metastases and peritoneal tumors upon i.v. and i.p. injection, respectively, is of practical value for the application of these vectors in tumor gene therapy.


  1. 1

    Shenk T . Adenoviridea. In: Fields BN, Knipe DM, Howley PM (eds) Fields Virology. vol.2 Lippincott-Raven Publisher: Philadelphia, 1996 pp 2111–2148.

  2. 2

    Law LK, Davidson BL . What does it take to bind CAR? Mol Ther 2005; 12: 599–609.

  3. 3

    Wickham TJ, Mathias P, Cheresh DA, Nemerow GR . Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 1993; 73: 309–319.

  4. 4

    Okegawa T, Li Y, Pong RC, Bergelson JM, Zhou J, Hsieh JT . The dual impact of coxsackie and adenovirus receptor expression on human prostate cancer gene therapy. Cancer Res 2000; 60: 5031–5036.

  5. 5

    Li Y, Pong RC, Bergelson JM, Hall MC, Sagalowsky AI, Tseng CP et al. Loss of adenoviral receptor expression in human bladder cancer cells: a potential impact on the efficacy of gene therapy. Cancer Res 1999; 59: 325–330.

  6. 6

    Bergelson JM, Cunningham JA, Droguett G, Kurt-Jones EA, Krithivas A, Hong JS et al. Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 1997; 275: 1320–1323.

  7. 7

    Miller CR, Buchsbaum DJ, Reynolds PN, Douglas JT, Gillespie GY, Mayo MS et al. Differential susceptibility of primary and established human glioma cells to adenovirus infection: targeting via the epidermal growth factor receptor achieves fiber receptor-independent gene transfer. Cancer Res 1998; 58: 5738–5748.

  8. 8

    Gaggar A, Shayakhmetov D, Lieber A . CD46 is a cellular receptor for group B adenoviruses. Nat Med 2003; 9: 1408–1412.

  9. 9

    Segerman A, Arnberg N, Erikson A, Lindman K, Wadell G . There are two different species B adenovirus receptors: sBAR, common to species B1 and B2 adenoviruses, and sB2AR, exclusively used by species B2 adenoviruses. J Virol 2003; 77: 1157–1162.

  10. 10

    Segerman A, Atkinson JP, Marttila M, Dennerquist V, Wadell G, Arnberg N . Adenovirus type 11 uses CD46 as a cellular receptor. J Virol 2003; 77: 9183–9191.

  11. 11

    Kinugasa N, Higashi T, Nouso K, Nakatsukasa H, Kobayashi Y, Ishizaki M et al. Expression of membrane cofactor protein (MCP, CD46) in human liver diseases. Br J Cancer 1999; 80: 1820–1825.

  12. 12

    Murray KP, Mathure S, Kaul R, Khan S, Carson LF, Twiggs LB et al. Expression of complement regulatory proteins-CD 35, CD 46, CD 55, and CD 59-in benign and malignant endometrial tissue. Gynecol Oncol 2000; 76: 176–182.

  13. 13

    Fishelson Z, Donin N, Zell S, Schultz S, Kirschfink M . Obstacles to cancer immunotherapy: expression of membrane complement regulatory proteins (mCRPs) in tumors. Mol Immunol 2003; 40: 109–123.

  14. 14

    Muruve DA . The innate immune response to adenovirus vectors. Hum Gene Ther 2004; 15: 1157–1166.

  15. 15

    Vlachaki MT, Hernandez-Garcia A, Ittmann M, Chhikara M, Aguilar LK, Zhu X et al. Impact of preimmunization on adenoviral vector expression and toxicity in a subcutaneous mouse cancer model. Mol Ther 2002; 6: 342–348.

  16. 16

    Nilsson M, Ljungberg J, Richter J, Kiefer T, Magnusson M, Lieber A et al. Development of an adenoviral vector system with adenovirus serotype 35 tropism; efficient transient gene transfer into primary malignant hematopoietic cells. J Gene Med 2004; 6: 631–641.

  17. 17

    Sova P, Ren X-W, Shaoheng N, Bernt KM, Mi J, Kiviat N et al. A tumor-targeted and conditionally replicating oncolytic adenovirus vector expressing TRAIL for treatment of liver metastases. Mol Ther 2004; 9: 496–509.

  18. 18

    Shayakhmetov DM, Li ZY, Ni S, Lieber A . Analysis of adenovirus sequestration in the liver, transduction of hepatic cells, and innate toxicity after injection of fiber-modified vectors. J Virol 2004; 78: 5368–5381.

  19. 19

    Ni S, Bernt K, Gaggar A, Li ZY, Kiem HP, Lieber A . Evaluation of biodistribution and safety of adenovirus vectors containing group B fibers after intravenous injection into baboons. Hum Gene Ther 2005; 16: 664–677.

  20. 20

    Hsu EC, Dorig RE, Sarangi F, Marcil A, Iorio C, Richardson CD . Artificial mutations and natural variations in the CD46 molecules from human and monkey cells define regions important for measles virus binding. J Virol 1997; 71: 6144–6154.

  21. 21

    Barouch DH, Pau MG, Custers JH, Koudstaal W, Kostense S, Havenga MJ et al. Immunogenicity of recombinant adenovirus serotype 35 vaccine in the presence of pre-existing anti-Ad5 immunity. J Immunol 2004; 172: 6290–6297.

  22. 22

    Ophorst OJ, Kostense S, Goudsmit J, De Swart RL, Verhaagh S, Zakhartchouk A et al. An adenoviral type 5 vector carrying a type 35 fiber as a vaccine vehicle: DC targeting, cross neutralization, and immunogenicity. Vaccine 2004; 22: 3035–3044.

  23. 23

    Kostense S, Koudstaal W, Sprangers M, Weverling GJ, Penders G, Helmus N et al. Adenovirus types 5 and 35 seroprevalence in AIDS risk groups supports type 35 as a vaccine vector. AIDS 2004; 18: 1213–1216.

  24. 24

    Sakurai F, Mizuguchi H, Yamaguchi T, Hayakawa T . Characterization of in vitro and in vivo gene transfer properties of adenovirus serotype 35 vector. Mol Ther 2003; 8: 813–821.

  25. 25

    Seshidhar Reddy P, Ganesh S, Limbach MP, Brann T, Pinkstaff A, Kaloss M et al. Development of adenovirus serotype 35 as a gene transfer vector. Virology 2003; 311: 384–393.

  26. 26

    Shayakhmetov DM, Li ZY, Ni S, Lieber A . Targeting of adenovirus vectors to tumor cells does not enable efficient transduction of breast cancer metastases. Cancer Res 2002; 62: 1063–1068.

  27. 27

    Marie JC, Astier AL, Rivailler P, Rabourdin-Combe C, Wild TF, Horvat B . Linking innate and acquired immunity: divergent role of CD46 cytoplasmic domains in T cell induced inflammation. Nat Immunol 2002; 3: 659–666.

  28. 28

    Buchholz CJ, Schneider U, Devaux P, Gerlier D, Cattaneo R . Cell entry by measles virus: long hybrid receptors uncouple binding from membrane fusion. J Virol 1996; 70: 3716–3723.

  29. 29

    Hsu KH, Lonberg-Holm K, Alstein B, Crowell RL . A monoclonal antibody specific for the cellular receptor for the group B coxsackieviruses. J Virol 1988; 62: 1647–1652.

  30. 30

    Bernt K, Liang M, Ye X, Ni S, Li ZY, Ye SL et al. A new type of adenovirus vector that utilizes homologous recombination to achieve tumor-specific replication. J Virol 2002; 76: 10994–11002.

  31. 31

    Carlson CA, Shayakhmetov DM, Lieber A . An adenoviral expression system for AAV Rep78 using homologous recombination. Mol Ther 2002; 6: 91–98.

  32. 32

    Grunau C, Clark SJ, Rosenthal A . Bisulfite genomic sequencing: systematic investigation of critical experimental parameters. Nucleic Acids Res 2001; 29: E65.

  33. 33

    Shayakhmetov DM, Li ZY, Ternovoi V, Gaggar A, Gharwan H, Lieber A . The interaction between the fiber knob domain and the cellular attachment receptor determines the intracellular trafficking route of adenoviruses. J Virol 2003; 77: 3712–3723.

  34. 34

    Vrancken Peeters MJ, Perkins AL, Kay MA . Method for multiple portal vein infusions in mice: quantitation of adenovirus-mediated hepatic gene transfer. Biotechniques 1996; 20: 278–285.

  35. 35

    Mantel N, Haenszel W . Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22: 719–748.

  36. 36

    Arap W, Kolonin MG, Trepel M, Lahdenranta J, Cardo-Vila M, Giordano RJ et al. Steps toward mapping the human vasculature by phage display. Nat Med 2002; 8: 121–127.

  37. 37

    Anders M, Hansen R, Ding RX, Rauen KA, Bissell MJ, Korn WM . Disruption of 3D tissue integrity facilitates adenovirus infection by deregulating the coxsackievirus and adenovirus receptor. Proc Natl Acad Sci USA 2003; 100: 1943–1948.

  38. 38

    Shayakhmetov DM, Li Z-Y, Ni S, Lieber A . Analysis of adenovirus liver sequestration, transduction of hepatic cells, and innate toxicity after injection of fiber-modified vectors. J Virol 2004; 78: 5368–5381.

  39. 39

    Shayakhmetov S, Li Z-Y, Ni S, Lieber A . Interference with the IL-1 signaling pathway significantly improves the toxicity profile of systemically applied adenovirus vectors. J Immunol 2005; 174: 7310–7319.

  40. 40

    Anderson BD, Nakamura T, Russell SJ, Peng KW . High CD46 receptor density determines preferential killing of tumor cells by oncolytic measles virus. Cancer Res 2004; 64: 4919–4926.

  41. 41

    Frampton JE, Plosker GL . Icodextrin: a review of its use in peritoneal dialysis. Drugs 2003; 63: 2079–2105.

  42. 42

    Engler H, Machemer TR, Schluep T, Wen SF, Quijano E, Wills KN et al. Development of a formulation that enhances gene expression and efficacy following intraperitoneal administration in rabbits and mice. Mol Ther 2003; 7: 558–564.

  43. 43

    Hakulinen J, Junnikkala S, Sorsa T, Meri S . Complement inhibitor membrane cofactor protein (MCP; CD46) is constitutively shed from cancer cell membranes in vesicles and converted by a metalloproteinase to a functionally active soluble form. Eur J Immunol 2004; 34: 2620–2629.

  44. 44

    Seya T, Hara T, Iwata K, Kuriyama S, Hasegawa T, Nagase Y et al. Purification and functional properties of soluble forms of membrane cofactor protein (CD46) of complement: identification of forms increased in cancer patients’ sera. Int Immunol 1995; 7: 727–736.

  45. 45

    Stone D, Ni S, Li ZY, Gaggar A, DiPaolo N, Feng Q et al. Development and assessment of human adenovirus type 11 as a gene transfer vector. J Virol 2005; 79: 5090–5104.

  46. 46

    Sova P, Ni S, Ren X-W, Bernt K, Lieber A . Efficacy of tumor-targeted and conditionally replicating oncolytic adenovirus expressing Apo2/TRAIL in treatment of liver metastases in a mouse xenograft model. Mol Ther 2004; 9: 496–509.

  47. 47

    Di Paolo N, Tuve S, Ni S, Hellström K, Hellström I, Lieber A . Effect of adenovirus mediated heat shock protein expression and oncolysis in combination with low-dose cyclophosphamide treatment on anti-tumor immune responses. Cancer Res 2006; 66: 960–969.

  48. 48

    Bernt KM, Ni S, Tieu AT, Lieber A . Assessment of a combined, adenovirus-mediated oncolytic and immunostimulatory tumor therapy. Cancer Res 2005; 65: 4343–4352.

  49. 49

    Nicklin SA, Wu E, Nemerow GR, Baker AH . The influence of adenovirus fiber structure and function on vector development for gene therapy. Mol Ther 2005; 12: 384–393.

  50. 50

    Shayakhmetov DM, Gaggar A, Ni S, Li ZY, Lieber A . Adenovirus binding to blood factors results in liver cell infection and hepatotoxicity. J Virol 2005; 79: 7478–7491.

  51. 51

    Shinozaki K, Suominen E, Carrick F, Sauter B, Kahari VM, Lieber A et al. Efficient infection of tumor endothelial cells by a capsid-modified adenovirus. Gene Therapy 2005; 13: 52–59.

  52. 52

    Sova P, Feng Q, Geiss G, Wood T, Strauss R, Rudolf V et al. Discovery of novel methylation biomarkers in cervical carcinoma by global demethylation and microarray analysis. Cancer Epidemiol Biomarkers Prev 2006; 15: 114–123.

Download references


We thank Steve Roffler and Daniel Stone for critical discussions and Steve Hawes for help with statistical analyses. We are grateful to Branka Horvat for providing the transgenic mice. This study was supported by NIH Grants CA080192, HLA078836 and HL-00-008 and grants from the Doris Duke and Avon Charitable Foundations.

Author information



Corresponding author

Correspondence to A Lieber.

Additional information

Supplementary Information accompanies the paper on Cancer Gene Therapy website (http://www.nature.com/cgt.

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Ni, S., Gaggar, A., Di Paolo, N. et al. Evaluation of adenovirus vectors containing serotype 35 fibers for tumor targeting. Cancer Gene Ther 13, 1072–1081 (2006). https://doi.org/10.1038/sj.cgt.7700981

Download citation


  • CD46
  • adenovirus
  • serotype 35
  • tumor targeting
  • transgenic mice
  • CAR

Further reading