Introduction

As the fourth most common malignancies, bladder transitional cell carcinoma seriously risks health of the people both in the United States and the rest of the world [1]. Mutation and deletion of tumor suppressor genes have been well established as the major causes of bladder carcinogenesis [2]. Gene therapy aimed to restore the expression of wide-type tumor-inhibiting genes in cancer cells or stimulate host immune response against tumor has shown certain anti-tumor outcomes and minimal side effects according to preclinical and clinical data [3–6].

However, two major problems always limit the clinical application of adenoviral-vector gene therapy. Firstly, the existing adenoviral vectors for bladder cancer treatment are developed based on adenovirus of serotype 5 (Ad5), which require the coxsackie virus and adenovirus receptor (CAR) localized on cell surface to enter target cancer cells [7]. CAR level is documented to vary greatly among human bladder cancer cell lines and even decline in clinical bladder cancers, not ensuring effective viral infection [8, 9]. In addition to loss of adenoviral receptor, the efficacy of gene therapy needs to be further improved even in the CAR-positive bladder cancer [10]. Almost no objective response was observed even in some trials [3]. Thus, development of novel adenoviral vector for enhanced expression of therapeutic genes is always of interests for bladder cancer therapy [11].

Adenovirus of serotype 35 (Ad35) belongs to group B adenovirus and infects target cells in a CRA-independent mechanism (dependent on CD46) [12]. Researchers altered tropism of the existing Ad5-derived adenoviral vectors by modifying their fiber region. This new type of adenoviral vectors contain a hybrid fiber protein consisting of knob and shaft domain derived from Ad35 and base domain from Ad5, thereby infecting cancer cells in a CD46-dependent way. Their infectivity is shown to be significantly increased to a wide range of cancer cells. Consistently, expression level of gene transferred by these fiber-modified vectors is also increased, resulting in an enhanced anti-tumor effect [13–16]. However, the capability of the 5/35 fiber-modified adenovirus has not been determined to enhance gene transfer in bladder cancer.

Here, we substituted knob and shaft regions of Ad35 fiber for that of Ad5-derived replication-deficient adenovirus to generate a novel 5/35 fiber-modified adenoviral vector. Furthermore, TNF-related apoptosis-inducing ligand (TRAIL) gene was inserted into the fiber-modified adenoviral vector to determine its ability to express exogenous gene. First of all, expression level of CAR and CD46 in bladder cancer cells was quantified in our experiments. The infectivity of 5/35 fiber-modified adenoviral vector was also determined. Subsequently, expression levels of TRAIL mediated by this novel gene transfer vectors were quantified in bladder cancer cells. Finally, we investigated the tumor-suppressing capacity of 5/35 fiber-modified adenoviral vectors armed TRAIL, both in vitro and in mice.

Results

CAR and CD46 expression in bladder cancer cells

CAR is the primary cellular receptor for Ad5 to attach and enter target cells, while CD46 allows Ad35 to infect cells. We investigated expression level of CAR and CD46 in bladder cancer cells. As CAR is highly expressed in Hep3B human hepatocellular carcinoma cell line, we selected Hep3B as positive control in CAR detection. As the same time, we used lentivirus-mediated RNA interfering to knockdown the expression of CAR in Hep3B cell lines (Hep3B-shCAR) and selected it as negative control. Quantitative real-time PCR (qRT-PCR) assay showed that CAR is significantly reduced in T24, TCCSUP, 253J, and RT4 cells than Hep3B (Fig. 1). Meanwhile, all these cancer cells tested in our experiment had higher or similar expression levels of CD46 comparative to CHO cell lines, which has been well established to highly express CD46 (Fig. 1) [12].

Fig. 1
figure 1

Expression level of CAR and CD46 in bladder cancer cells. Expression levels of CAR and CD46 mRNA in T24, TCCSUP, 253J, and RT4 were quantified by qRT-PCR. mRNA level of CAR and CD46 was normalized to GAPDH. Hep3B and CHO were selected as standards. The bars represent mRNA level of CAR and CD46 relative to Hep3B and CHO, respectively. Median values from three independent experiments were shown and error bars show ±SD

Full size image

Given the fact of low CAR expression and increased CD46 expression in bladder cancer cells, we predicted Ad5/35 may probably be more suitable as an expression vector for gene therapy against bladder cancer.

Infectivity of 5/35 fiber-modified adenoviral vector into bladder cancer cells

Next, we compared the infectivity between Ad5-derived and 5/35 fiber-modified adenoviral vector into bladder cancer cells. Ad5-EGFP and Ad5/35-EGFP at MOI of 10 and 1 were used to infect T24, TCCSUP, 253J, and RT4 cells followed by fluorescent activated cell sorter (FACS) analysis. We observed elevated EGFP-positive cell population in all Ad5/35-EGFP-infected cancer cells compared to Ad5-EGFP-infected groups (Fig. 2).

Fig. 2
figure 2

Infectivity of Ad5-EGFP and Ad5/35-EGFP to bladder cancer cell lines. Bladder cancer cells, T24, TCCSUP, 253J, and RT4 were infected by Ad5-EGFP and Ad5/35-EGFP at MOI of 10 and 1. 48 h later, the percentages of EGFP-positive cells were quantified by FACS analysis. The uninfected cells were chosen as negative control (not shown). The bars represent median values from two independent experiments and error bars represent ±SD. **P < 0.01

Full size image

The results demonstrated that 5/35 fiber-modified adenoviral vector infect bladder cancer cells more efficiently than Ad5-derived vector.

Ad5/35 vector-mediated TRAIL expression in bladder cancer cells

To confirm that 5/35 fiber-modified adenoviral vector expressed transgenes more effectively, we subsequently examined TRAIL expression of Ad5/35-TRAIL-infected bladder cancer cells with ELISA assay. Consistently, ELISA data showed that Ad5/35-TRAIL-infected cancer cells secreted increased amount of TRAIL, than Ad5-TRAIL-infected ones (Fig. 3).

Fig. 3
figure 3

TRAIL expression in fiber-modified adenoviral vector-infected bladder cancer cells. 10 MOI of Ad5/35-TRAIL or Ad5-TRAIL were administrated to T24, TCCSUP, 253J, and RT4 cells. 48 h later, expression level of TRAIL gene from adenovirus-infected bladder cancer cells was quantified by ELISA assay. Ad5/35-mediated TRAIL expression was represented relative to the average expression level of TRAIL mediated by Ad5. The bars represent median values from three independent experiments and error bars represent ±SD. *P < 0.05, **P < 0.01

Full size image

In conclusion, we confirmed that 5/35 fiber-modified adenoviral vector was a more suitable expression vector for therapeutic transgene in bladder cancer cells than Ad5-derived vector.

In vitro tumor-suppressing ability of 5/35 fiber-modified adenovirus

We explored the tumor-suppressing capacity of Ad5/35-TRAIL against bladder cancer cell lines. The data showed that Ad5/35-TRAIL had enhanced tumor-inhibiting effect to all the tested cancer cells than Ad5-TRAIL (Fig. 4). More effective anti-tumor treatments against these bladder cell lines were found in 5/35 fiber-modified adenoviral vector-administrated group (Fig. 4). It is worthy noting that less than 1 MOI of Ad5/35-TRAIL showed a significant tumor-suppressing effect, especially in 253J and RT4 cells, compared to Ad5-TRAIL of the same amount.

Fig. 4
figure 4

Anti-tumor capability of Ad5/35-TRAIL to bladder cancer cells. The tumor-suppressing abilities of Ad5/35-TRAIL and Ad5-TRAIL to bladder cancer cells were explored by MTT assay. T24, TCCSUP, 253J, and RT4 cells were infected with Ad5-TRAIL or Ad5/35-TRAIL of the indicated MOIs and absorbance at each MOI was determined. The absorbance of uninfected group was chosen as standard. The dots showed the average absorbance of each group relative to uninfected group and error bars represent ±SD. These experiments were performed at least two times. The data from one experiment were presented

Full size image

The data demonstrated that Ad5/35-TRAIL had a more potential tumor-inhibiting effect on bladder cancer cells than Ad5-derived competent adenoviral vectors in vitro.

Growth-inhibiting effect of 5/35 fiber-modified adenoviral vector to bladder cancer in nude mice

Finally, we investigated the growth-inhibiting effect of Ad5/35-TRAIL to bladder cancers in a nude mice model. T24 and 253J cells were inoculated subcutaneously in BALB/c nude mice. The tumor diameter was measured regularly after PBS, Ad5-TRAIL or Ad5/35-TRAIL were intratumorally administered. The data showed that Ad5/35-TRAIL was able to inhibit the growth of T24 cancer xenograft, whereas Ad5-TRAIL had little effect on these tumors (Fig. 5). For 253J tumor xenograft treatment, Ad5/35-TRAIL also showed an enhanced inhibiting capability (Fig. 5).

Fig. 5
figure 5

Growth-inhibiting effect of Ad5-TRAIL and Ad5/35-TRAIL on bladder cancer xenografts in mice. T24 and 253J cells were inoculated subcutaneously in BALB/c nude mice. PBS, 1 × 109 pfu of Ad5-TRAIL or Ad5/35-TRAIL was intratumorally injected after xenograft tumors formed. The tumor volumes (mm3) were measured at the indicated time points. The dots showed the average values and the error bars showed ±SD. *P < 0.05

Full size image

We observed an increased tumor-suppressing efficacy of 5/35 fiber-modified adenoviral vectors armed with proapoptotic cytokines to bladder cancer compared with Ad5-derived vectors in vivo.

Discussion

In the past two decades, biological therapy was gradually recognized by researchers and therefore added to the therapeutic options for urological cancers. Adenoviruses have been widely studied for bladder cancer therapy in pre-clinical experiments as well as in clinical trials [17–20]. In addition to serve as gene transfer vector, virus administration may promote the adaptive tumor-specific immune in patients, because vaccinia, also used as an anti-tumor agent, has been reported to induce dendritic cell migration to virus-infected sites in bladder cancers [21]. However, the application of conventional Ad5-derived vectors was severely limited.

Because of low or no CAR expression in some kinds of cancer cells and the existence of antibody to Ad5 in cancer-suffering patients, modification of the tropism of adenoviral vectors are greatly needed. Many researchers focused on fiber protein engineering, constructing 5/3 fiber-modified and 5/11p fiber-modified adenoviral vector [22, 23]. Although 5/35 fiber-modified adenoviral vector has been widely used to treat many types of human cancer [24, 25], the efficacy of this tropism-redirecting vector in urological cancers, especially bladder cancers, has not been studied yet. Here, we provided evidence that 5/35 fiber-modified adenoviral vector is more suitable for gene transfer against bladder cancers, compared with traditional Ad5-derived adenoviral vector.

Previous studies have reported that bladder cancer cells had a variable expression level of adenoviral receptor CAR [8]. Our data also reinforces this finding that T24 and TCCSUP has little expression of CAR, while CAR was detected in 253J and RT4. But CAR level of bladder cancer are all below that of Hep3B. Lack of CAR expression may explain why Ad5-based vector is difficult to infect T24 and TCCSUP cells.

Because Ad35 can infect cells in a CD46-dependent mechanism [12], down-regulation of CD46 was documented to reduce infectivity of 5/35 fiber-modified adenoviral vector [26]. We found that CD46 is expressed in all tested bladder cancer cells on a level similar to CHO. These results may ensure effective infection of 5/35 fiber-modified adenoviral vector to these cancer cells. Consistently, we also utilized flow cytometry to confirm that this fiber-modified vector has a higher infectivity to bladder cancers than traditional Ad5-derived vector, even in CAR-positive cells, 253J, and RT4.

Enhanced expression of TRAIL gene mediated by 5/35 fiber-modified replication-deficient adenovirus was not beyond our expectation. MTT assay also confirmed a more effective survival-suppressing effect of gene therapy after this fiber substitution. In agreement with previous data, T24 was the most vulnerable cell lines from TRAIL administration mediated by 5/35 fiber-modified vector. In animal model study, 253J xenograft growth is almost completely inhibited by fiber-modified vector-mediated TRAIL administration. Similarly, T24 tumors had a significant response to Ad5/35-TRAIL while Ad5-mediated gene therapy appeared no effect.

As well as protein-encoding genes, adenoviral vector can also be used to transfer microRNAs or shRNAs which have tumor-inhibiting effect for bladder cancer therapy. The strategy of adenoviral vector-mediated microRNA has been reported to be effective in cancer treatment [27].

Here, we reconstructed traditional Ad5-derived adenoviral vector by replacing Ad5 fiber with Ad35 fiber. Armed with TRAIL gene, this fiber-modified adenovirus inhibits growth of both CAR-negative and CAR-positive bladder cancer cells more effectively. We provided a new promising expression vector for effective gene transfer in bladder cancer cells. We strongly hope this vector can facilitate clinical application of gene therapy.

Materials and methods

Cell cultures

Human bladder transitional cell carcinoma cell line T24, RT4, and TCCSUP were all purchased from the American Type Culture Collection (Manassas, VA). Human bladder transitional cell carcinoma cell line 253J, human hepatocellular carcinoma cell line Hep3B and Chinese hamster ovarian cell line CHO was obtained from the Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China). T24, RT4, TCCSUP, and 253J were grown in McCoy’s 5a Medium Modified (Life Technologies, Rockville, MD) supplemented with 10% (v/v) fetal bovine serum (Life Technologies, Rockville, MD). Hep3B, CHO, and HEK-293 were cultured in Dulbecco’s Minimum Essential Medium (Life Technologies, Rockville, MD) with fetal bovine serum to a final concentration of 10%. All media was supplemented with 4 mM glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin. All cells in this experiment were cultured under a 5% CO2 and humidified atmosphere at 37°C.

Adenoirus construction and production

E1A-null adenovirus Ad5-EGFP and Ad5/35-EGFP were generously gifted from Doctor Hoffmann. The Ad5-derived replication-deficient adenoviral vector armed with TRAIL gene (Ad5-TRAIL) was generated as follows:

The TRAIL gene was synthesized and inserted into pShuttle-CMV at site of SalI/EcoRV, generating pShuttle-CMV-TRAIL. pShuttle-CMV-TRAIL and pAdEasy were co-transfected into E. coli BJ5183 to obtain recombination plasmid pAd-TRAIL, which was subsequently transfected into HEK-293 cells using the Effectene Transfection Reagent (QIAGEN) to produce the recombinant adenovirus Ad5-TRAIL.

To generate 5/35 fiber-modified adenoviral vector, DNA fragment containing 5/35 chimeric adenovirus fiber gene was synthesized followed by digestion with restriction endonucleases BamHI and NheI and insertion into a plasmid containing both homologous arms. Subsequently, this 5/35 fiber containing plasmid was transfected together with pAdEasy into E. coli BJ5183 for homologous recombination, generating an adenovirus-producing plasmid that bears Ad5/35 chimeric fiber gene, pAdEasy-35. Finally, pShuttle-CMV-TRAIL and pAdEasy-35 were co-transfected into E. coli BJ5183 to obtain recombination plasmid pAd35-TRAIL, which was subsequently transfected into HEK-293 cells using the Effectene Transfection Reagent (QIAGEN) to produce the recombinant adenovirus Ad5/35-TRAIL. The fiber structure of Ad5-TRAIL and Ad5/35-TRAIL were shown in Fig. 6.

Fig. 6
figure 6

Illustration of adenoviral fiber regions. The knob and shaft domain of fiber derived from Ad5 was substituted with that from Ad35 to generate a hybrid fiber (5/35 fiber)

Full size image

After plague purification for three times and PCR-based identification, the adenoviruses were harvested and purified with the CsCl gradient centrifugation method. The titers of the involved adenoviruses were quantified with the Tissue Culture Infectious Dose 50 (TCID50) method on HEK-293 cells and shown as plaque-forming units per milliliter (pfu/ml).

Lentiviral vector-mediated gene silencing

Lentiviral vector that expresses short hairpin RNA targeting CAR and CD46 were purchased from Sigma-Aldrich. According to the manufacturer’s instructions, 10 MOI of lentivirus was used to infect the cell lines, Hep3B and CHO.

Quantitative real-time PCR

To determine mRNA level of bladder cancer cells, 3 × 104 cells per well were grown in 24-well plates. 48 h later, cells were harvested to isolate total RNA. The RNA was reversely transcribed into cDNA using Rever Tra AceR qPCR RT Kit (Toyobo, Japan) according to manufacturer’s instructions. qRT-PCR was performed with StopOne™ Real-Time PCR system (Applied Biosystems). The involved primers for qRT-PCR were described in other researchers’ publication [24]. Data were analyzed with StepOne Software v2.0 (Applied Biosystems).

Fluorescent activated cell sorter (FACS) analysis

Bladder cancer cells were grown in 6-well plates at a concentration of 2 × 105 cells/well. 24 h later, the media were replaced with serum-free media, and then, Ad5-eGFP or Ad5/35-eGFP was added into the media at the MOI of 10 and 1. 48 h after infection, cells infected by adenoviral vector were digested and suspended with PBS at the density of 1 × 106 cells/ml. EGFP-positive population in the samples was quantified with FACS (BD Biosciences, USA).

Detection of TRAIL by ELISA

The two-antibody sandwich ELISA was used to detect human TRAIL expression. The used antibodies are monoclonal mouse anti-human TRAIL antibody (R&D Systems), peroxidase-conjugated rabbit anti-goat IgG (H&L) and goat anti-human TRAIL antibody (R&D Systems). The absorbance was read at a 450 nm wavelength. A standard curve was used to determine the concentration of TRAIL.

MTT assay

Cancer cells were plated in 96-well plates at a density of 1 × 104 cells/well. Overnight, a series of MOI of adenoviral vectors were added to growth media. Seven days later, 50 μl of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; 1 mg/ml) was added to cell media. Four hours later, MTT was removed and replaced with 150 μl DMSO. The spectrophotometric absorbance of the samples was measured with Microplate Reader Model 550 (Bio-Rad Laboratories, Japan) at 570 nm with a reference wavelength of 655 nm. Cell proliferation rate is equal to the ratio of absorbance value of adenovirus-infected cells to that of uninfected cells. Median value and ±SD from six reduplicate was shown at each MOI. Viability of cancer cell lines were all performed at least twice and data from one experiment was presented.

Animal model study

Procedures for animal experiments were all approved by the Committee on the Use and Care on Animals (General Hospital of Chengdu MAC, Chengdu, China). T-24 tumor xenografts were established through injection of 1 × 106 cells at the right flanks of 4-week-old BALB/c nude mice (Institute of Animal Center, Chinese Academy of Sciences, Shanghai, China). When tumors reached nearly 8 mm in diameter, 18 mice were randomly split into PBS-administered group, Ad5-TRAIL-infected group and Ad5/35-TRAIL-infected group. 21 mice were subcutaneously injected with 2 × 106 of 253J cells per mice, and then, were split into the three groups. The mice were injected with 100 μl PBS with or without 2 × 108 pfu of Ad5-TRAIL or Ad5/35-TRAIL. The injections were repeated five times every other day, finally reaching a dosage of 1 × 109 pfu of adenoviruses. Tumor diameter was measured by periodic measurements with calipers and volume was calculated using the following formula: tumor volume (mm3) = maximal length (mm) × (perpendicular width) (mm)2/2. Animals were killed once the tumor volume exceeded 2000 mm3. No mice were observed to be died of tumor loading.

Statistical analysis

The statistical tests in our work were all two-tailed. Data were all recognized as statistically significant (*) when P < 0.05 and statistically very significant (**) when P < 0.01.