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

Introduction The sodium iodide symporter (NIS) directs the uptake and concentration of iodide in thyroid cells. This in turn allows radioiodine imaging and therapy for thyroid cancer. To extend the use of NIS-mediated radioiodine therapy to other types of cancer, we successfully transferred and expressed the sodium-iodide symporter (NIS) gene in prostate, colon, and breast cancer cells both in vivo and in vitro by using non-replicating adenoviral vectors. Methods To improve virotherapy efficiency, we developed a conditionally replicating adenovirus (CRAd) in which the transcriptional cassette RSV promoter-human NIScDNA-bGH polyA was also inserted at the E3 region. The E1a gene is driven by the tumor-specific promoter MUC-1 in the CRAd Ad5AMUCH_RSV-NIS. Results In vitro infection of the MUC-1-positive breast cell line T47D resulted in virus replication, cytolysis, and release of infective viral particles. Conversely, the MUC-1-negative breast cancer cell line MDA-MB-231 was refractory to the viral cytopathic effect and did not support viral replication. The data indicate that Ad5AMUCH_RSV-NIS activity is stringently restricted to MUC-1-positive cancer cells. Radioiodine uptake was readily measurable in T47 cells infected with Ad5AMUCH_RSV-NIS 24 hours after infection, thus confirming NIS expression before viral-induced cell death. Conclusions This construct may allow multimodal therapy, combining virotherapy with radioiodine therapy to be developed as a novel treatment for breast and other MUC1-overexpressing cancers.


Introduction
The American Cancer Society's most recent estimates for breast cancer in the United States for 2009 are as follows: 192,370 new cases of invasive breast cancer and 40,170 deaths of breast cancer. Breast cancer is the most common cancer among women in the United States, other than skin cancer. It is the second leading cause of cancer death in women, after lung cancer [1]. Breast cancer is particularly dif-ficult to treat when it metastasizes and becomes resistant to antiestrogen therapies [2].
The sodium iodide symporter (NIS) is a transmembrane glycoprotein that mediates uptake of iodide into cells, especially thyroid follicular cells [3,4]. The presence of NIS on the basolateral membrane of thyroid cells has been exploited for many years for diagnostic imaging purposes, as well as for ablative therapy of differentiated thyroid cancer by using radi- oactive iodide ( 131 I). This noninvasive therapy has proven to be a safe and effective treatment for thyroid cancer, even in advanced, metastatic disease [5,6]. To extend the use of NISmediated radioiodine therapy to other types of cancer, we successfully transferred and expressed the sodium-iodide symporter (NIS) gene in prostate, colon, and breast cancer cells, both in vivo and in vitro, by using adenoviral vectors. Our experience with adenovirus-mediated NIS transfer and radioiodine therapy was confirmed in a large animal model and has culminated in the opening of a phase I trial for prostate cancer that is currently in progress [7][8][9][10][11][12].
MUC-1 is a highly glycosylated transmembrane mucin. Although MUC-1 is expressed at very low levels in normal tissues; it is overexpressed by most carcinomas, including breast cancers [13]. Enhanced expression of MUC-1 is regulated mostly at the level of transcription [14]. The pronounced differential expression of MUC-1 in tumor versus normal tissues has been used in experimental developments of antitumor therapies, including MUC-1 vaccines and MUC-1 promoterrestricted antitumor-specific viruses [15][16][17].
All gene-therapy approaches depend on the ability to deliver therapeutic genes to target cells. However, the limited ability to transduce tumors efficiently with effective levels of therapeutic transgenes has been identified as the fundamental barrier to effective cancer gene therapy [18,19]. To address this issue, conditionally replicating viruses, including adenovirus, have been constructed, and their efficacy, evaluated [20][21][22].
Our approach to the current problems associated with virotherapy/gene therapy has been the development of tumor-specific, conditionally replicating adenoviral vectors that also harbor the NIS gene. We report here the development of a conditionally replicating adenovirus (CRAd) in which the E1a gene is driven by the tumor-specific promoter MUC-1. In addition, the transcriptional cassette RSV promoter-hNIScDNA-bGH polyA was inserted at the E3 region. Our results suggest that this CRAd may represent a novel approach to breast cancer gene therapy.

Cell culture
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].

CRAd construct and cell lines
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 AAAAAGGTACCGGACCCTAGGGT-TCATCGGAG and the reverse primer AAAAAAGATCTGAT-TCAGGCAGGCGCTGGCTGC 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 pVQA-MUCH. 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 NIScontaining 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.

MUC-1 flow cytometry
The 10 6 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 (IgG 1 ) 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.
(page number not for citation purposes)

Protein isolation and Western blots
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 × 10 6 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].

Cytopathic effect assays
The 1 × 10 6 cells were infected at MOI 1. At noted time points, the cells were photographed under a light microscope at 100× magnification.

MTS assay
The 5 × 10 4 cells were seeded in 96-well tissue-culture plates and incubated at 37°C, 5% CO 2 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% CO 2 for 2 hours. Absorbance was then read at 490 nm by using a plate spectrophotometer.

Replication assay of the CRAd vector
The 4 × 10 6 cells were plated in 150-cm 2 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.

I uptake
Uptake of 125 I by virus-infected cells was measured as described [10,28]. In brief, 1 × 10 6 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 × 10 5 cpm of 125 I. Control wells were incubated in the same buffer plus 10 μmol/L KClO 4 . Plates were gently mixed and incubated at 37°C, 5% CO 2 , 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 125 I on a gamma counter.

Noninvasive imaging of xenografted tumors
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 × 10 6 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 T 4 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][9][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 mm 3 , mice received one intratumoral (IT) injection of 10 11 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 123 I 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.

Virus construction
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).  (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.

Conditional replication, oncolysis, and spread of Ad5AMUCH_RSV-NIS
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]. 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.

NIS expression and radioiodine uptake
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 125 I uptake at later time points (Figure 7b). Radioiodine uptake was inhibited by KClO 4 , 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 123 I 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.

Discussion
Although a number of gene-therapy approaches have been developed for many cancers, clinical trials completed to date have fallen short of expectations [32][33][34][35][36], largely due to limited tumor transduction, and this limitation has been defined as the most important fundamental barrier to effective cancer gene therapy [37]. Our approach to addressing the problem of low tumor transduction is to allow vector replication by using conditionally replicating adenovirus.
We used a transcriptional-regulation method placing the E1A gene under control of the tumor-specific promoter MUC-1. The rationale for this choice is based on the observation that transcriptional-dependent regulation seems to be the most efficient strategy for CRAd development [38]. In addition, we inserted the NIS gene under control of the powerful but nonspecific RSV promoter. We used a truncated version of the MUC-1 promoter that was previously shown to be specifically active in MUC-1-positive breast cancer cell lines [17]. The rationale and utility for use of this promoter is that more than 90% of breast tumors overexpress the MUC-1 gene [13]. By this method, viral replication, and consequently extended NIS  Cell-proliferation assay Cell-proliferation assay. (a) T47D and (b) MDA-MB-231. Cells were infected with Ad5PB_RSV-NIS, and cell viability was monitored daily by using MTS assay. expression, will target a wide spectrum of MUC-1-positive breast tumors.
In the present study, we confirmed that this MUC-1 promoter version restricts viral protein synthesis, cytopathic effect, cell lysis, and infectious viral progeny production to MUC-1-positive breast cancer cells. Moreover, the bulk of viral progeny produced by the MUC-1-positive breast cancer cell line T47D was found in the media, indicating that viral infection of T47D cells with the Ad5AMUCH_RSV-NIS CRAd resulted in cell lysis and subsequent release of infectious viral particles. We hypothesize that the release of infectious viral particles should allow wide viral spread within tumors. Virus specificity was demonstrated by using the MUC-1-negative breast cancer cell line MDA-MB-231. The MDA-MB-231 cell line expresses cellsurface CAR receptor because we visualized CAR expression in membrane-bound electrophoresed proteins. It supports limited E1A and hexon expression and supports NIS expression. All these data confirm that MDA-MB-231 is infected by the Ad5AMUCH_RSV-NIS, yet it is resistant to CRAd-induced  oncolysis or viral replication or both. These data demonstrate that the Ad5AMUCH_RSV-NIS CRAd is stringently dependent on MUC-1 overexpression and should not replicate within normal tissue, where MUC1 expression is minimal.
The NIS protein must be properly targeted to the cell plasma membrane to be functional [39]. Our results also showed that the Ad5AMUC_RSV-NIS CRAd directs correct NIS expression and thus radioiodine uptake in vitro and in vivo. Under the control of the RSV promoter, the NIS protein was correctly targeted to the cell plasmamembrane, thereby promoting high levels of iodine uptake in T47D cells. Our results also show that in MUC-1-negative CAR-positive breast cancer tumors, the NIS gene also can be expressed, because it is under the control of the ubiquitous RSV promoter. Here, the cell line MDA-MB-231 showed very strong NIS protein expression by Western blot, which translated to robust radioiodine uptake (data not shown). Hence the Ad5AMUCH_RSV-NIS CRAd could be used as a non-replicating NIS-expressing vector.
However, in this case, NIS gene expression was transient, thus narrowing the window for radioiodine therapy and loss of the combined radiovirotherapy.
A discrepancy is apparent in the kinetics of NIS expression and luciferase expression that can be explained by several influencing factors; First, the luciferase gene was cloned at the E1 region of a nonreplicating adenovirus, under the control of the MUC-1 promoter, whereas the NIS gene was cloned at the E3 region of a replicating adenovirus under the RSV promoter. Second, the stability of two very different mRNAs and proteins cannot be assumed to be equal.
Finally, NIS is an integral membrane protein, whereas luciferase remains cytoplasmic. Therefore, the pattern of expression of luciferase and NIS under these conditions is not expected to overlap precisely.
NIS is already a therapeutic gene, in that its native expression in thyroid cells is used for therapy for thyroid cancer and hyperthyroidism, where its efficacy is very high [6]. NIS expression could represent an ideal therapeutic gene for breast cancer therapy, because it allows a large bystander effect. Because of the physics of 131 I decay, which occurs primarily through emission of a β particle that traverses up to 2 mm within the tissue. Thus, not every cell must be transfected and express NIS to be affected by radioiodine therapy. This is critical to the effectiveness of any cancer gene-therapy strategy because of the difficulty in achieving 100% transfection with the therapeutic gene, even in the case of a replicating virus. Moreover, NIS expression can be directly monitored and quantitated noninvasively. The uptake of radioiodine in tissues can be quantitated, and both the distribution and the quantity of NIS expression can be easily and safely monitored noninvasively.
The limited efficacy observed in clinical trials of cancer gene therapies to date has suggested that combinatorial therapies to treat cancer will likely be more effective and possibly required for complete tumor eradication. The hypothesis developed from this observation is that attacking tumor cells through different mechanisms of action may prevent tumor cells from developing resistance to the treatment combination and may more effectively induce immunity against the tumor [40]. We hypothesize that, because Ad5AMUC_RSV-NIS CRAd harbors the NIS gene, it should improve the therapeutic value of the CRAd by allowing multimodal therapy in a single agent combining viral-mediated tumor lysis as well as radioiodine-mediated therapy.

Conclusions
In conclusion, the features of Ad5PB_RSV-NIS CRAd address favorably the major hurdles identified in gene-therapy clinical trials for anticancer treatments. Further evaluation, including efficacy studies in animal models, is needed to determine the potential of this radio/virotherapy approach to affect breast cancer treatment.
AdEasyGL3B-Muc with JD. JCM conceived of the study, participated in its design and coordination, and helped to draft the manuscript.