Construction of an MUC-1 promoter driven, conditionally replicating adenovirus that expresses the sodium iodide symporter for gene therapy of breast cancer

The breast cancer cell lines T47D and MDA-MB-231 were used to examine CRAd Ad5AMUCH_RSV-NIS specificity. Adenovirus infection was performed for 4 hours in serum-free media. Cells were then washed once with PBS and replenished with fresh culture media. The human embryonic kidney cell line stably expressing E1A (HEK 293) was obtained from Cell Biolabs, Inc., San Diego, CA. Cells were cultured as described [8, 23].

The E1A gene flanked by a 5' blunt end, and an EcoRI 3' end was isolated from plasmid pCD512_F2 (a gift of Dr. Richard Vile, Mayo Clinic) and cloned into the Promega vector pGL3-Basic to yield pGL3E1. The MUC-1 promoter (a gift of Dr. Sandra Gendler, Mayo Clinic Arizona) was PCR amplified by using the forward primer AAAAAGGTACCGGACCCTAGGGTTCATCGGAG and the reverse primer AAAAAAGATCTGATTCAGGCAGGCGCTGGCTGC from the plasmid pcDNA3 [8]. The resulting PCR fragment 0.726 Kb spanning nt2188 to nt2914 was cloned into pGL3E1 to yield pGL3MUCH. The MUC-1 promoter-E1A-SV40 PolyA fragment was excised from pGL3MUCH by digestion with EcoRI/KpnI and cloned into the shuttle vector pVQA-PB-NIS [23] restricted by KpnI/EcoRI and dephosphorylated with CIP to generated pVQAMUCH. The human NIS cDNA under Rous Sarcoma Virus LTR promoter control was subcloned into an Ad5 genomic vector containing an E3-deleted version (ViraQuest, North Liberty, IA, USA). The pVQAMUCH shuttle vector and the NIS-containing Ad5 genomic plasmid were recombined to yield the CRAd Ad5AMUCH_RSV-NIS. Plaques were grown, purified, and characterized by ViraQuest as previously described [23]. The replication-incompetent Muc promoter-driven luciferase expression vector, AdEasyMucLuc, was constructed through homologous recombination in Escherichia coli by using the AdEasy system [24]. This vector has the transgene cassette in the E1 deleted region of an adenoviral vector backbone. The vector was constructed as follows: the MUC_1 promoter was derived from a plasmid pGL3MUCH and cloned into plasmid pShuttleGL3B [25] by using KpnI and HindIII sites. The resultant plasmid, pShuttleMuc_GL3B, was linearized with PmeI digestion and subsequently co-transformed into E. coli BJ5183 with the pAdEasy-1 Ad backbone plasmid. After selection of recombinants in these bacteria, the recombinant DNA was linearized with PacI digestion and transfected into 293 cells to generate AdEasyMucLuc. The virus was propagated in 293 cells, dialyzed in phosphate-buffered saline (PBS) with 10% glycerol, and stored at -80°C. Titering was performed with a plaque-forming assay by using 911 cells and optical density-based measurement.

The 10⁶ exponentially growing cells were washed with PBS containing 1 mmol/L EDTA and resuspended in PBS. Cells were incubated with 20 μg/ml mouse anti-human MUC-1 antibody (Cell Signaling Technology Inc. Beverly, MA, USA) 30 min at RT. Controls cells were processed equally, except that they were incubated with 20 μg/ml mouse anti-human (IgG₁) TNP anti-isotype antibody control (BD Biosciences Pharmingen, SanDiego, CA, USA). Cells were then washed in PBS and incubated with FITC-conjugated goat anti-mouse IgG secondary antibody (BD Biosciences Pharmingen) (1:100) for 30 min at RT. Cells were then washed, resuspended in PBS, and immediately analyzed with a Becton Dickinson FACScan by using CELLQuest software.

For all Western blots, the protein concentration was measured in triplicate, and two gels were run. The first was developed with silver staining by using the PlusOne Silver Staining Kit, Protein (GE Healthcare Bio-Sciences, Uppsala, Sweden) according to the manufacturer's instruction and quantitated by using the ImageJ software [26]. This gel served as loading control (not shown). The second gel was blotted onto nitrocellulose and Western blotted. E1a and hexon extracts were prepared from 4 × 10⁶ cells grown on 100-mm tissue-culture plates and infected at MOI 20. At noted time points, cells were rinsed twice in HBSS before lysis in 1 ml of the complete lysis-M reagent with supplied protease inhibitors (Roche Applied Science, Indianapolis, IN, USA) on ice with shaking for 5 minutes and then quantified by using the DC protein assay (Bio-Rad Hercules, CA, USA). Membrane proteins extracts were prepared for NIS and CAR Western blotting as described [27]. Lysate aliquots, 1 μg of protein, were electrophoresed by using I a X cell II Blot Module (Invitrogen, Carlsbad, CA, USA) on 10% BIS-TRIS polyacrylamide gels under reducing conditions at 200 V for 1 hour. Proteins were then blotted to nitrocellulose membranes at 25 V for 1 hour and processed with the ECL advanced Western blot detection kit (GE Healthcare, Uppsala, Sweden) according to the manufacturer's instructions. E1A primary antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) was diluted 1:20,000, and the anti-Ad5 hexon IgG (Santa Cruz Biotechnology), 1:100,000. Anti-CAR analyses were performed by running 5 μg of membrane protein or 1 μg of cell extract on a 10% Bis-Tris polyacrylamide gels under nonreducing conditions. Blocking steps were performed as described earlier. Blots were incubated with primary mouse Anti-CAR, clone RmcB (Upstate Cell Signaling Solutions Inc., Danvers, MA, USA). The antibody was diluted 1:10,000 in TBS-T with 0.1% blocking reagent for 90 minutes. NIS Western blots were performed as previously described [27].

The 1 × 10⁶ cells were infected at MOI 1. At noted time points, the cells were photographed under a light microscope at 100× magnification.

The 5 × 10⁴ cells were seeded in 96-well tissue-culture plates and incubated at 37°C, 5% CO₂ in triplicate. Twenty-four hours after plating, the cells were infected at the indicated MOI. The virus-containing medium was exchanged with fresh medium, and at each time point, 50 μl of CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega Corporation, Madison, WI, USA) was applied and then incubated at 37°C, 5% CO₂ for 2 hours. Absorbance was then read at 490 nm by using a plate spectrophotometer.

The 4 × 10⁶ cells were plated in 150-cm² culture flasks and infected at MOI 1 and MOI 0.1. Seventy-two hours after infection, virus was harvested from media and cells separately and purified by using the ViraBind Adenovirus Purification Kit (Cell Biolabs, San Diego, CA, USA). Virus titers were quantified by plaque assay on HEK 293 cells.

Uptake of ¹²⁵I by virus-infected cells was measured as described [10, 28]. In brief, 1 × 10⁶ cells/well were plated on 36-well tissue-culture plates and infected at MOI 0, 0.5, 5, 10, 20, and 50. At indicated time points, iodide-uptake measurement was performed in triplicate in HBSS supplemented with 10 mmol/L HEPES, pH 7.3, 10 μmol/L NaI, and 1 × 10⁵ cpm of ¹²⁵I. Control wells were incubated in the same buffer plus 10 μmol/L KClO₄. Plates were gently mixed and incubated at 37°C, 5% CO₂, for 45 minutes. After incubation, buffers were aspirated and the cells gently washed once with 4°C 10 mmol/L HEPES, pH 7.3 buffer. Cells were lysed in 1 ml 1 N NaOH with shaking for 20 minutes at RT, and lysates were assayed for ¹²⁵I on a gamma counter.

All experimental protocols were reviewed and approved by the Mayo Clinic Institutional Animal Care and Use Committee (IACUC). Xenografts derived from the T47D cell line were established on each flank of 6-week-old athymic nude Foxn1nu mice (Harlan, Madison, WI, USA) by subcutaneous injection of 4 × 10⁶ cells resuspended in 0.125 μl media and 0.125 μl of BD Matrigel basement membrane matrix (BD Biosciences, Bedford, MA). Mice were maintained on a low-iodine diet and T₄ supplementation (5 mg/L) in their drinking water throughout the duration of the experimentation to maximize radioisotope uptake in the tumor and minimize uptake by the thyroid [8‐10]. The mice were examined daily for tumor growth. Tumor volume was measured with calipers twice weekly and calculated by the formula:

When tumors reached 100 mm³, mice received one intratumoral (IT) injection of 10¹¹ vp of the AdAMUCH_RSV-NIS CRAd on the left flank; the right flank served as control. Two days later, an intraperitoneal dose of 0.5 mCi of ¹²³I sodium iodide was given, and 1 hour later, radioiodine imaging was performed by using a γ-camera equipped with a low-energy high-resolution collimator (VG System; GE Healthcare, Milwaukee, WI). Regions of uptake were quantified and expressed as a fraction of the total amount of the applied radioiodine.

The conditionally replicating adenoviruses (CRAd) Ad5AMUCH_RSV-NIS harbors the transcriptional cassette RSV promoter-human NIScDNA-bGH polyA inserted at the E3 region. In this CRAd, the E1a gene is driven by a 0.726-Kb version of the tumor-specific MUC-1 promoter spanning nt2188 to nt2914 [17] (Figure 1).

Two breast cancer cell lines, T47D and MDA-MB-231, were characterized for their expression of MUC-1. No difference in the flow-cytometry histogram was found between the isotype control and the MUC-1 antibody in MDA-MB-231, confirming that this cell line is MUC-1^- (data not shown). We then zeroed the FACScan by using this cell line and analyzed the T47D under this new setting. This analysis revealed that T47D is MUC-1 positive, whereas MDA-MB-231 is MUC-1 negative (Figure 2a). Moreover, when these two cell lines were infected with the nonreplicating virus AdEasyGL3B-Muc, in which the Luc gene is driven by the MUC-1 promoter, only the T47D cell line supported MUC-1-driven Luc activity, whereas MDA-MB-231 was completely negative (Figure 2b). To exclude the possibility that inability to support viral replication was due to lack of CAR-receptor expression, CAR-receptor expression was assayed with Western blot of membrane extracts prepared from the two cell lines. To control for protein loading, identical amounts of protein were loaded on a second gel that was silver stained for protein assessment (not shown). The insert in Figure 2a supports the notion that both cell lines readily express the CAR receptor at the cell plasma membrane. We conclude that T47D is MUC-I positive, whereas MDA-MB-231 is MUC-1 negative, and that both cell lines express CAR and thus are capable of being infected by wild-type capsid Ad5.

We examined the specificity of Ad5AMUCH_RSV-NIS replication by using several methods. First, we examined the specificity of viral proteins expression. MUC-1-driven E1A and Hexon expression were examined with Western blotting. T47D and MDA-MB-2321 cells were infected with Ad5AMUCH_RSV-NIS at MOI 20. Figure 3 shows two major differences between the permissive T47D cell line and the nonpermissive MDA-MB-231. Only cells infected with Ad5AMUCH_RSV-NIS expressed viral proteins. Most important, a marked difference was noted in the quality and quantities of viral protein expressed in the permissive cell compared with the nonpermissive cell line (Figure 3a). We also examined hexon expression, a product of late protein synthesis and a very good indicator of viral assembly [29]. Here again, the amount of hexon protein produced 48 hours after infection in the permissive cell line T47D outperformed that produced in nonpermissive cell lines MDA-MB-231 (Figure 3b).

Cells infected with Ad5PB_RSV-NIS at MOI 1 were monitored for viral-induced cytopathic effect (CPE) by light microscopy. Four days after infection, the morphology of the MDA-MB-231 cells was unchanged; however, clear signs of CPE in T47D are detected under the same conditions of infection (Figure 4).

To address the issue of virus specific cell killing, the permissive cell line T47D was infected with the Ad5AMUCH_RSV-NIS CRAd at increasing MOIs and then assayed daily by MTS. The results were compared with those obtained with the nonpermissive cell line MDA-MB-231. The T47D cells showed a reduced viability that was both time and dose dependent (Figure 5a). Conversely, MDA-MB-231 viability was completely refractory to virus-mediated cell killing (Figure 5b).

Virus-progeny production after infection with the Ad5AMUCH_RSV-NIS CRAd was measured in T47D, and MDA-MB-231 cells 3 days after infection at MOI 1 or 0.1 by burst assay (Figure 6). To distinguish between released and intracellular virus, media and cell were separated, and virus was purified separately from each. The amount of virus produced by the T47D cells was 6 orders of magnitude greater than the amount produced by the MDA-MB-231 cells.

Taken together, these data demonstrate that replication of the Ad5AMUCH_RSV-NIS virus is stringently restricted to MUC-1-positive breast cancer cells, defining this virus as a conditionally-replicating adenovirus or CRAd.

Expression of the NIS protein was examined with Western blotting by using a mouse monoclonal anti-human NIS antibody [30]. Figure 7a shows that T47D and MDA-MB-231 cells transfected with the Ad5PB_RSV-NIS CRAd expressed NIS, which was detected as a broad band spanning from 130 Kd to 110 kDa. Visualization of the NIS band as a smear is expected; it reflects differences in NIS glycosylation [30, 31]. Expression of the NIS gene in the nonpermissive cell line MDA-MB-231 is supported by the RSV ubiquitous promoter and shows that this cell line is susceptible to Ad5 infection.

Having shown that infected cells are capable of expressing membrane-bound NIS, we next investigated whether the protein was functional. Dose-response and kinetic curves of iodine uptake were constructed. Infection of T47D cells with Ad5AMUCH_RSV-NIS at MOI 20 showed maximal uptake at 24 hours after infection, followed by a twofold decrease in ¹²⁵I uptake at later time points (Figure 7b). Radioiodine uptake was inhibited by KClO₄, a well-known inhibitor of NIS activity [3]. We infer from this result that an optimal time window exists for NIS-mediated radioiodine uptake in the permissive cells, after which, uptake, and as a consequence, NIS expression declines. This might be due to virus replication and induction of CPE. Twenty-four hours after infection, a radioiodine uptake dose-response to increasing MOI concentrations was built. Figure 7c shows that radioiodine uptake was linear over the range of MOI used in the experiment.

Having shown in cell culture that Ad5AMUCH_RSV-NIS can infect, multiply, and direct NIS expression in permissive cells, we wished to confirm that viral infection and NIS expression also take place in vivo. Figure 8 reveals strong ¹²³I uptake by the T47D tumor that has been injected with the Ad5PB_RSV-NIS CRAd 48 hours after infection. Quantitation of the image yielded an uptake in infected tumors of more than 30% of the injected dose at 1 hour. The stomach and thyroid also were imaged because of native NIS expression, and the bladder, because of tracer excretion [9].

Taken together our results demonstrate that Ad5AMUCH_RSV-NIS directs NIS expression in permissive cells before the onset of cell lysis.