Genetic Incorporation of a Herpes Simplex Virus Type 1 Thymidine Kinase and Firefly Luciferase Fusion into the Adenovirus Protein IX for Functional Display on the Virion

Abstract

An advantage of the adenoviral vector is its molecular flexibility, which allows for vector tropism modifications for the purpose of cell targeting. In addition to targeting ligands, the capacity to incorporate heterologous peptides has allowed capsid incorporation of other functionalities. We have defined the minor capsid protein IX (pIX) as a locus capable of presenting incorporated ligands on the virion surface. Thus, we sought to exploit the possibility of incorporating functional proteins at pIX. In our current study, we sought to expand the potential utility of our capsid labeling strategy by developing simultaneous imaging capacity for dedicated small animal positron emission tomography and bioluminescence imaging on a single adenoviral vector. Therefore, we constructed an adenovirus that incorporates a fusion protein of herpes simplex virus type 1 thymidine kinase and firefly luciferase (Luc) (TK-Luc) into adenovirus capsid pIX. Our study herein clearly demonstrates our ability to rescue viable adenoviral particles that display functional TK-Luc as a component of their capsid surface. Most importantly, Ad-pIX-TK-Luc retained dual enzymatic functions in vitro and in vivo. This dual-modality approach will allow dynamic or real-time imaging analysis of adenovirus-based interventions with maximized analytic flexibility and enhanced resolution potential.

Adenoviral vectors have been used for a variety of gene therapy applications.¹ Their utility is attributed to the unparalleled efficiency of gene transfer in both in vitro and in vivo contexts.^2,3 An additional advantage of the adenoviral vector is its molecular flexibility, which allows for vector tropism modifications for the purpose of cell targeting.⁴ This capacity is based on genetic modification of the capsid proteins involved in the key steps of target cell binding and entry.⁵ The foundation of these targeting paradigms stems from the recent identification of capsid locales permissive for the incorporation of large and/or more complex heterologous proteins.⁶ In addition to targeting ligands, the capacity to incorporate heterologous peptides has allowed capsid incorporation of other functionalities. In this regard, we and others have shown that the fluorographic reporter enhanced green fluorescent protein (EGFP) incorporated at the minor capsid protein IX (pIX) is fully functional and allows labeled viral particles (VPs) to be visualized for analytic purposes.^7,8

This strategy of capsid incorporation of imaging motifs has provided a novel approach for analytic monitoring of virion trafficking at the cellular and subcellular levels of resolution. In our recent studies, we incorporated herpes simplex virus type 1 thymidine kinase (HSV-TK) at the pIX locus.⁹ In this regard, HSV-TK has been used for both molecular chemotherapy approaches in the context of cancer gene therapy applications and for positron emission tomography (PET) schemas.^10,11 PET has been widely exploited for imaging analysis of gene therapy strategies in small animal models^12–14 and, more recently, in human clinical trials.^15,16 Of note, this imaging modality has unique applications and advantages, such as high sensitivity and good spatial resolution, as well as its tomographic characteristic. On the other hand, there are also defined disadvantages to PET, such as its requirements for relatively short-lived radioactive tracers and its high operation cost.

As an alternative to PET, optical imaging techniques such as bioluminescence imaging (BLI), which uses luciferase as a reporter, may offer potential advantages. Luciferase imaging is very sensitive, with possibly 10⁻¹⁵ to 10⁻¹⁷ mole of luciferase/L detectable in vivo.¹⁷ It has been shown that tumor cells can be detected with BLI-based techniques at an early stage where PET and radiology cannot.^18,19 Further, the background luminescence associated with BLI is negligible compared with the background observed with fluorescent imaging.^11,17 On the other hand, there are also defined disadvantages to BLI, such as high tissue absorption, scattering, and the inability to scale up to larger animals and humans. Based on the distinct utility profiles of PET and optical imaging, it is clear that the employment of combined modalities may offer optimal imaging capabilities.

Our genetic capsid labeling strategy provides a direct approach for dynamic assessment of conditionally replicative adenoviruses, as validated in our recent study.²⁰ This strategy provides a valid index of viral parameters that cannot be elucidated via imaging methods based on transgene expression. Our current study expands the range of utility of this capsid-labeling strategy by incorporating a dual-imaging modality (herpes simplex virus type 1 thymidine kinase–luciferase fusion [TK-Luc]) at adenoviral protein IX. This approach provides the potential for simultaneous imaging capacity for microPET and BLI on a single adenoviral vector. We therefore constructed an adenovirus that incorporates a fusion protein of HSV-TK and firefly luciferase (TK-Luc) into adenovirus capsid pIX. Our study herein clearly demonstrates our ability to rescue viable adenoviral particles that display functional TK-Luc as a component of their capsid surface. The display of TK-Luc on the capsid may offer advantages with respect to direct functional applications of vector imaging analysis. Furthermore, the determination of an expanded upper limit of incorporable proteins on pIX highlights its unique utility as a locale for functional vector components.

Results

Construction of an Adenovirus with Capsid-Incorporated TK-Luc

In our recent studies, we incorporated HSV-TK at pIX.⁹ To further expand our capsid incorporation strategy,⁹ a unique adenovirus was created that contains a dual-imaging cassette at pIX. We constructed an adenovirus genome containing a pIX-TK-Luc carboxy-terminal fusion gene in place of wild-type pIX. Between the pIX and TK-Luc coding regions we incorporated an 18–amino acid linker ADDYKDDDDKLAGSGSG containing the octapeptide DYKDDDDK FLAG-tag sequence. On transfecting this genome into the complementing cell line HEK293, we observed plaque formations after 11 days. This result predicates the rescue and production of an adenovirus that contains a TK-Luc motif genetically incorporated into the adenoviral capsid surface at pIX (Figure 1).

Analysis of the TK-Luc Fusion Protein into the Adenovirus Genome and Virion Capsid

To validate expression of the pIX-TK-Luc protein from the adenoviral genome, we performed Western blot analysis on cells infected with Ad-pIX-TK-Luc. HEK293 cells were infected with a virus expressing thymidine kinase (TK) as a transgene (under the control of the cytomegalovirus promoter), Ad-CMV-TK, or our previously constructed virus with TK incorporated into pIX, Ad-wt-pIX-TK, or Ad-pIX-TK-Luc. After 48 hours, infected cells were harvested and lysates were analyzed via Western blot using an anti-TK polyclonal antibody. As shown in Figure 2A, cells infected with Ad-CMV-TK yielded a protein band of approximately 40 kDa, corresponding to the molecular weight of the TK protein. Cells infected with Ad-wt-pIX-TK (see Figure 2A) showed a major protein band that resolved at ≈65 kDa, which corresponded to the expected size of the pIX-TK fusion protein. Lysates of Ad-wt-pIX-TK infected cells also showed a second protein band that resolved at ≈40 kDa, corresponding to the molecular weight of TK protein. This protein band likely represents a second translated protein product initiated at the TK coding region or peptide cleavage products. Importantly, cells infected with Ad-pIX-TK-Luc (see Figure 2A) showed a major protein band that resolved at ≈120 kDa, which corresponds to the expected size of the pIX-TK-Luc fusion protein. We speculate that the expression differences seen with pIX-TK-Luc and pIX-TK are a result of size differences between the proteins; therefore, pIX-TK-Luc has a lower yield compared with the smaller pIX-TK. A plausible answer with respect to cytomegalovirus–thymidine kinase (CMV-TK) versus pIX-TK-Luc is that CMV-TK has a higher gene expression yield compared with pIX-TK-Luc because CMV-TK is driven by a stronger promoter in comparison with pIX-TK-Luc. These results demonstrate that the pIX-TK-Luc fusion protein was expressed from the adenovirus genome.

Figure 1.

Strategy for the construction of an adenovirus with a herpes simplex virus 1 thymidine kinase and firefly luciferase (TK-Luc) fusion reporter incorporated at the pIX capsid protein. Human adenovirus serotype 5 was genetically engineered to express a modified capsid protein IX. An E1/E3 deleted adenovirus was constructed containing a pIX-TK-Luc carboxy-terminal fusion gene in place of wild-type pIX (Ad-pIX-TK-Luc). Between the pIX and TK-Luc coding regions is an incorporated 18–amino acid linker SADDYKDDDDKLAGSGSG containing the octapeptide FLAG-tag sequence (underlined).

To verify that the pIX-TK-Luc protein was incorporated into the adenoviral capsid, we analyzed gradient-purified Ad-wt-pIX-TK and Ad-pIX-TK-Luc viruses using Western blot analysis for FLAG expression, which is genetically incorporated between pIX and the protein of interest. Analysis of Ad-wt-pIX-TK yielded a major protein band that resolved at ≈65 kDa, which corresponded to the expected size of the pIX-TK protein (Figure 2B). An additional band that resolved at ≈126 kDa was detected; we speculate that this additional band is due to inefficient protein termination. This is possible because the C-terminus of pIX has been modified and the endogenous stop codon has been placed further downstream; thus, protein termination may be somewhat inefficient, resulting in multiple polypeptides. Analysis of Ad-pIX-TK-Luc yields a major protein band that resolved at ≈120 kDa, which corresponded to the expected size of the pIX-TK-Luc protein (see Figure 2B). We know that the size of a protein can negatively impact efficient capsid incorporation; therefore, we speculate that pIX-TK-Luc has a lower yield compared with pIX-TK owing to inefficient capsid incorporation. These results demonstrate that the TK-Luc fusion protein was incorporated into pIX within Ad5 virions. Taken together, these data validate the expression of the pIX-TK-Luc protein and incorporation into the virion capsid.

Figure 2.

Expression of capsid-incorporated TK-Luc fusion protein. A, Western blot analysis of thymidine kinase (TK) expression in cells infected with TK-expressing virions. HEK293 cells were infected with 100 viral particles (VPs)/cell of Ad-CMV-TK, Ad-wt-pIX-TK, or Ad-pIX-TK-Luc. Cell lysates were collected and resolved on a sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) gel and transferred to polyvinylidene difluoride (PVDF) membrane. Staining was performed with a rabbit polyclonal anti-TK antibody. The arrow indicates pIX-TK-Luc protein. B, Western blot analysis of Ad-pIX-TK-Luc for TK-Luc protein incorporation. In this assay, 10¹⁰ VPs of CsCl gradient purified Ad-wt-pIX-TK and Ad-pIX-TK-Luc were resolved on an SDS-PAGE gel and transferred to PVDF membrane. Staining was performed with a monoclonal anti-FLAG antibody. The arrow indicates pIX-TK-Luc protein. MW = molecular weight.

Analysis of In Vitro Capsid-Associated Luc Activity of a Capsid-Incorporated TK-Luc Fusion Protein

Enzymatic activity of the firefly luciferase associated with a known concentration of purified Ad-pIX-TK-Luc virions was determined by measuring relative light units (RLUs) that were obtained by the conversion of d-luciferin substrate into the luminescent product. For this experiment, we used an additional control virus, Ad5Luc1. The Ad5Luc1 virus contains the luciferase gene in the E1 region and has the ability to produce the luciferase gene only after cellular infection, in direct contrast to the Ad-pIX-TK-Luc virus, which expresses luciferase on its capsid and is active in the presence of substrate. CsCl gradient-purified Ad5Luc1, Ad-wt-pIX-TK, or Ad-pIX-TK-Luc virions were added to a 96-well plate. RLUs were measured for each sample following the addition of d-luciferin. As shown in Figure 3, the addition of d-luciferin to purified Ad-pIX-TK-Luc virus resulted in 1.1 × 10⁴ RLU, compared with purified Ad5Luc1 or Ad-wt-pIX-TK virions, which resulted in ≈20 RLU. Thus, these results demonstrate that the luciferase incorporated into the adenoviral capsid in the form of the pIX-TK-Luc fusion protein was functional and luciferase activity could be detected without expression in the cell.

Validation of BLI via the TK-Luc Fusion Protein

To determine the functionality of the virion-incorporated pIX-TK-Luc fusion protein as a reporter gene for in vitro BLI, Ad-pIX-TK-Luc or control viruses were used to infect HEK293 cells, followed by BLI. For this, three 60 mm tissue culture dishes were plated with 1 × 10⁶ HEK293 cells and infected with Ad-pIX-TK-Luc (A), Ad-CMV-luc (B), or Ad-CMV-GFP (C). Equal quantities of virus were added to each plate (1 × 10⁸ VPs). BLI was measured after 24 hours after the addition of 1 mM d-luciferin to each dish. The three dishes were placed individually in a Xenogen (Xenogen Bioscences, Cranbury, NJ) IVIS bioluminescence/fluorescence scanner, and image data were acquired with a cooled bioluminescence camera. Significant amounts of luciferase activity were observed in cells infected with Ad-pIX-TK-Luc (Figure 4A). Similar results were obtained in cells infected with a positive control virus encoding luciferase as a transgene, Ad-CMV-luc (Figure 4B). No luciferase activity was observed in cells infected with the negative control Ad-CMV-GFP (Figure 4C). This result confirms the functionality of the luciferase component of the capsid-incorporated TK-Luc fusion protein, which allows BLI analysis of virus in infected cells.

Figure 3.

In vitro analysis of luciferase activity of an adenovirus with a capsid-incorporated TK-Luc fusion protein. Luciferase activities were measured from 10¹⁰ viral particles of CsCl gradient purified Ad5Luc1 virions, Ad-wt-pIX-TK virions, and Ad-pIX-TK-Luc virions. Values are expressed as mean ± standard deviation of three replicates.

Figure 4.

Validation of an adenovirus with a capsid-incorporated TK-Luc fusion protein via bioluminescent imaging. HEK 293 cells were infected with 10⁸ viral particles of (A) Ad-pIX-TK-Luc, (B) Ad-CMV-luc, or (C) Ad-CMV-GFP. After 24 hours, 1 mM d-luciferin was added to each dish and luciferase activity was detected using a cooled bioluminescence camera. The signal bars indicate 0 to 100% activity.

Figure 5.

Validation of an adenovirus with a capsid-incorporated TK-Luc fusion protein via micro–positron emission tomography (microPET). HEK 293 cells were infected with 10⁸ viral particles of (A) Ad-pIX-TK-Luc, (B) Ad-CMV-TK, or (C) Ad-CMV-GFP and after 24 hours were labeled with 50 μCi of [¹⁸F]-FHBG. The quantitative labeling through the planes of the dishes is shown. MicroPET image data were acquired for 15 minutes. The signal bars indicate 0 to 100% activity.

Validation of TK Activity Associated with the TK-Luc Fusion Protein via MicroPET Analysis

Next, functionality of the TK component of the pIX-TK-Luc fusion protein incorporated within the adenoviral virions as a capsid-associated reporter for microPET was determined. For this, Ad-pIX-TK-Luc and its controls were used to infect HEK293 cells, after which microPET imaging was performed (Figure 5). Specific labeling and incorporation with 9-[4-[¹⁸F]-fluoro-3-hydroxymethylbutyl]guanine ([¹⁸F]-FHBG) were observed in cells infected with Ad-pIX-TK-Luc (see Figure 5A). Similar results were seen in cells infected with a positive control virus encoding TK as a transgene, Ad-CMV-TK (see Figure 5B). However, only nonspecific background labeling was observed in cells infected with the negative control Ad-CMV-GFP (see Figure 5C). Thus, PET-based imaging analysis confirmed that the TK component of the fusion protein was also functional.

Analysis of the TK-Luc Fusion Protein In Vivo via MicroPET and BLI

To substantiate that the TK-Luc fusion protein was incorporated within the adenoviral virions as a reporter in vivo, microPET or BLI was performed using Ad-pIX-TK-Luc virions after intratumoral administration in mice. For this, athymic nude mice were implanted with subcutaneous xenografts of a human head and neck squamous cell carcinoma cell line (FaDu). After 3 weeks, tumor nodules were injected intratumorally with adenovirus and subsequently subjected to imaging analysis. To analyze imaging using PET, animals received an injection of Ad-pIX-TK-Luc in one tumor nodule and an injection into the opposing tumor nodule with Ad-CMV-TK as a positive control. At 48 hours postinjection, [¹⁸F]-FHBG was intravenously injected into the animals, followed by microPET (Figure 6A). Tumor nodules infected with Ad-pIX-TK-Luc virions yielded TK activity that was comparable to tumor nodules infected with a positive control virus encoding TK, Ad-CMV-TK. This result demonstrates that the TK component of the capsid-incorporated TK-Luc fusion gene functions for in vivo PET-based imaging analysis.

Another set of animals received an injection of Ad-pIX-TK-Luc into one tumor nodule or an injection into the opposing tumor nodule with Ad-CMV-luc as a positive control. d-Luciferin was injected into the animals intravenously 48 hours later followed by bioluminescent tumor imaging (Figure 6B). These data illustrate that in vivo BLI of tumor nodules infected with Ad-pIX-TK-Luc yields signal comparable to that of tumor nodules infected with a positive control encoding luciferase, Ad-CMV-Luc. This result confirms that the luciferase component of the capsid-incorporated TK-Luc fusion gene also functions in in vivo BLI analysis. These data illustrate that both components of the TK-Luc fusion retain their respective enzymatic functions at the pIX locale and function in an in vivo context, thus making imaging analysis feasible using a repertoire of techniques.

Materials and Methods

Cell Culture

Human embryonic kidney epithelial cells (HEK293), head and neck squamous cell carcinoma (HNSCC) cell line (FaDu), and human embryonic retinoblasts (911)²¹ were obtained from and cultured in the medium recommended by the American Type Culture Collection (Manassas, VA).

Figure 6.

Micro–positron emission tomography (microPET) and bioluminescent imaging analysis of mice infected with an adenovirus containing a capsid-incorporated TK-Luc fusion protein. Athymic nude mice were implanted with subcutaneous xenografts of the human head and neck squamous cell carcinoma cell line (FaDu). Tumor nodules were injected with adenovirus and then subjected to imaging analysis. A, MicroPET. Animals were injected intratumorally with control adenovirus encoding TK, Ad-CMV-TK (1 × 10¹⁰ viral particles [VPs] in 0.1 mL phosphate-buffered saline [PBS]) and in an opposing tumor with the dual-modality imaging Ad-pIX-TK-Luc (1 × 10¹⁰ VPs in 0.1 mL PBS). Mice were subjected to microPET analysis 48 hours after the injection. B, Bioluminescence imaging. Animals were injected intratumorally on one flank with control adenovirus encoding luciferase, Ad-CMV-luc (1 × 10¹⁰ VPs in 0.1 mL PBS), and intratumorally on the other flank with the dual-modality imaging Ad-pIX-TK-Luc virus (1 × 10¹⁰ VPs in 0.1 mL PBS). Mice were subjected to bioluminescent analysis 48 hours after the injection. The signal bars indicate 0 to 100% activity.

All cell lines were incubated at 37°C and 5% CO₂ under humidified conditions.

Recombinant Adenovirus Construction

All viruses were constructed by homologous recombination in Escherichia coli using methods previously described.²² The HSV-TK and firefly luciferase fusion construct used to create Ad-pIX-TK-Luc was a kind gift from Dr. Ariane Söling (United Kingdom).¹¹ The TK construct used in this study is the wild-type form,²³ and the luciferase is the “humanized” firefly luciferase gene from the pGL3 vector (Promega, Madison, WI). To construct the Ad-pIX-TK-Luc virus, NheI sites were introduced into the TK-Luc construct via polymerase chain reaction of the TK-Luc plasmid (pAS24Dr). Primers used to introduce NheI sites in pAS24DR were (forward) 5′-GATGACAATAAGCTAGCGATGGCTTCG-3′ and (reverse) 5′-CTAGCTAGCTTACACGGCGATCTTTCCG-3′. The resulting TK-Luc was inserted in-frame at the carboxyl terminus of the adenovirus pIX gene (NheI site) within pShuttle-IX-flag⁷ after an 18–amino acid linker, SADDYKDDDDKLAGSGSG, containing the octapeptide DYKDDDDK FLAG-tag sequence. The resulting plasmid was named pShuttle-pIX-TK-Luc. Homologous recombination was performed with Ad5 backbone (Stratagene, La Jolla, CA, BJ5183-AD-1), and the resultant recombinant plasmid was linearized with PacI and transfected into HEK293 cells. To construct the Ad-wt-pIX-TK virus,⁹ the HSV-TK gene was inserted in-frame at the carboxyl terminus of the adenovirus pIX gene (NheI site) within pShuttle.wt.E1-IX-flag,⁷ which contains the same linker sequence described above. Homologous recombination was performed with the resulting pShuttle.wt.E1-pIX-flag-TK and with an E3 deleted Ad5 backbone, and the resultant recombinant plasmid was linearized with PacI and transfected into 911 cells to generate the Ad-wt-pIX-TK virus. To construct the Ad-CMV-TK, Ad-CMV-GFP and Ad-CMV-Luc viruses, CMV-TK, CMV-GFP, or CMV-Luc was cloned into the E1 region of an adenoviral shuttle vectors and then homologously recombined into an Ad Easy-1 (Stratagene) system as described above.

Virus Propagation and Purification

Ad-wt-pIX-TK viruses were propagated in 911 cells and Ad-wt-pIX-TK-Luc viruses were propagated in HEK293 cells, which do not express wild-type pIX. The absence of native pIX is important so that recombinant pIX proteins can be produced. Viruses were purified by double cesium chloride (CsCl) ultracentrifugation and dialyzed against phosphate-buffered saline (PBS) without Mg²⁺ or Ca²⁺ and 10% glycerol. Viruses were stored at −80°C until use. Final aliquots of virus were analyzed for viral particle titer using absorbance at 260 nm. The viral titer (infectious units per milliliter) was determined by tissue culture infectious dose on 293 cells.

Western Blot Analysis

To analyze, HEK293 cells were infected with 100 VPs/cell of Ad-CMV-TK, Ad-wt-pIX-TK, or Ad-pIX-TK-Luc. After 48 hours, cell lysates were collected and resolved on a 4 to 15% gradient SDS-PAGE gel and transferred to polyvinylidene difluoride (PVDF) membrane. Staining was performed with a rabbit polyclonal anti-TK antibody (1:1,000) dilution, followed by a secondary horseradish peroxidase (HRP)-linked anti-rabbit immunoglobulin G antibody (1:1,000) dilution.

To analyze, CsCl gradient purified Ad-wt-pIX-TK or Ad-pIX-TK-Luc (10¹⁰ VPs) were resolved on a 4 to 15% gradient SDS-PAGE gel and transferred to PVDF membrane. Staining was performed with a monoclonal anti-FLAG antibody (1:2,000; Sigma-Aldrich, St. Louis, MO), followed by a secondary HRP-linked antimouse antibody (1:2,000; Biorad, Hercules, CA). Proteins of interest were detected using a chemiluminescent ECL kit (Amersham Pharmacia, Pittsburgh, PA).

In Vitro Assay of Luciferase Activity

CsCl gradient purified Ad5Luc1, Ad-wt-pIX-TK, and Ad-pIX-TK-Luc all (10¹⁰ VPs) were added directly into a 96-well plate. Eighty microliters of luciferase substrate was added to the plate according to the published protocols (Promega), and luciferase activity was measured as RLUs in the samples on a multiplate luminescent reader (Orion and MPL2 Microplate Luminometer, Berthold Detection Systems, Pforzheim, Germany).

In Vitro BLI in Adenovirus-Infected HEK293 Cells

For in vitro experiments, 1 × 10⁶ of HEK293 cells were plated in 3 mL of complete medium in three 60 mm tissue culture dishes (Costar, Corning, NY). At 24 hours prior to BLI, 1 × 10⁸ VPs of Ad-pIX-TK-Luc, Ad-CMV-luc, or Ad-CMV-GFP were added to respective dishes. On the day of BLI, 1 mM of d-luciferin was added to each dish. The three dishes were placed individually in a Xenogen IVIS bioluminescence/fluorescence scanner, and imaging data were acquired with a cooled bioluminescence camera.

In Vivo BLI

For in vivo experiments, athymic nude mice were injected subcutaneously with 1 × 10⁷ FaDu (HNSCC) cells on each shoulder. After tumors formed to a size of approximately 1 cm in diameter, the tumor on the left side was directly injected with 1 × 10¹⁰ infectious particles of Ad-CMV-Luc and the tumor on the right side was injected with 1 × 10¹⁰ infectious particles of Ad-pIX-TK-Luc. Forty-eight hours later, the mice were injected in the tail vein with d-luciferin and the tumors were imaged by a Xenogen IVIS bioluminescence/fluorescence scanner. The imaging data were acquired with a cooled bioluminescence camera.

MicroPET Scan of Adenovirus-Infected HEK293 Cells

PET studies were performed using a microPET rodent four-ring system (model R4) from CTI Concorde Microsystems, LLC (Knoxville, TN). In short, the system operates in three dimensions and is composed of 6,144 lutetium oxyorthosilicate crystal detector elements, with a 7.8 cm axial and a 10 cm transaxial field of view. The [¹⁸F]-FHBG has been used as a reporter probe to image the expression of the HSV-TK reporter gene in living organisms and has been recommended as a more effective probe for in vivo imaging than other acycloguanosine analogues.²⁴ The TK phosphorylates [¹⁸F]-FHBG to its monophosphate form, leading to intracellular accumulation.²⁵ Cellular retention of radioactivity is, therefore, an indicator of TK gene expression. Synthesis and purification of [¹⁸F]-FHBG have been described.²⁶

For in vitro experiments, three 60 mm tissue culture dishes (Costar) were plated with 1 × 10⁶ of the HEK293 cell line in 3 mL of complete medium. At 24 hours prior to BLI, 1 × 10⁸ VPs of Ad-pIX-TK-Luc, Ad-CMV-TK, or Ad-CMV-GFP was added to respective dishes. On the day of microPET, 50 μCi of [¹⁸F]-FHBG was added to each dish in 1 mL of PBS. The plates were incubated for 1 hour at 37°C, followed by three washes with 5 mL of PBS. The plates were placed in the microPET scanner, and image data were acquired for 15 minutes.

In Vivo MicroPET Experiments

For in vivo expression experiments, athymic nude mice were injected subcutaneously with 1 × 10⁷ FaDu (HNSCC) on the shoulder. After tumors formed to a size of approximately 1 cm in diameter, the right tumor was injected with 1 × 10¹⁰ infectious particles of Ad-pIX-TK-Luc and the left tumor was injected with 1 × 10¹⁰ infectious particles of Ad-CMV-TK. Forty-eight hours later, the mice were injected in the tail vein with 150 μCi of [¹⁰F]-FHBG. After 60 minutes, the animals were imaged in a microPET scanner, and image data were acquired for 15 minutes. MicroPET images were reconstructed by using an iterative reconstruction technique.

Discussion

We have developed a novel imaging approach based on capsid incorporation of imaging motifs. This method provides the possibility of dynamic and real-time monitoring of replicating adenovirus-based virotherapy agents for cancer treatment. In addition, direct analysis of vector particle localization dynamics in vivo can be endeavored via this approach. Our previous studies have established the compatibility of this method with fluorescent and PET imaging modalities. Our goal in this study was to establish capsid labeling to allow simultaneous imaging analysis via luminescent and PET-based methods. To achieve the dual-modality imaging, we incorporated a fusion protein consisting of luciferase and TK to allow bioluminescent and PET-based imaging analysis, respectively. Our studies have validated that the TK-Luc fusion can be successfully incorporated at the pIX capsid protein without deleterious effects on virion integrity and functionality. Of note, both components of the imaging fusion gene exhibit full functionality in the capsid-incorporated context. Furthermore, pilot studies demonstrated the utility of this method for dual-imaging analysis in murine model systems. These studies establish the feasibility of dual-modality imaging via our method of capsid incorporation of imaging modalities.

The strategy we pursued involved the fusion of TK-Luc¹¹ with the capsid protein pIX. The reasons for selecting pIX as a locale are as follows: (1) as a minor capsid protein, it is not absolutely required for viral replication; (2) its carboxy-terminus is externally exposed, and proteins may be fused to it without affecting the interior of the capsid; and (3) it has a remote location in the group of nine, which would not affect the penton base or the fiber knob, the major structures that are involved in key adenoviral entry mechanisms.²⁶ The carboxy-terminus of pIX has been well defined as a locus for a variety of ligands,^27,28 as well as imaging modalities such as EGFP and TK.^7,8,9 We have provided relative comparisons at the protein level between Ad-wt-pIX-TK and Ad-pIX-TK-Luc (see Figure 2). We were unable to provide direct comparisons between the relative activities of the two viruses based on confounding factors, such as unequal capsid incorporation efficiency owing to the size difference of pIX-TK and pIX-TK-Luc proteins. Our current pIX-TK-Luc represents the largest fusion protein to have been functionally incorporated in the adenoviral capsid context; the 1,000–amino acid TK-Luc protein sequence and an 18–amino acid linker sequence added to the carboxy-terminus of the 140–amino acid pIX protein sequence represents 1,158 amino acids in the aggregate. Thus, we have expanded the upper size limit of pIX fusion proteins, which has yet to be determined and may be limited only by what the adenovirus genome constraints dictate.

When evaluating any imaging platforms, the threshold or output signal is of great importance for assessing imaging utility. When designing our dual-imaging capsid-containing viruses, we could have tried a variety of genetic capsid modification approaches. In this regard, location and sequence of the imaging components in the adenoviral genome can greatly affect reporter gene function. For example, Söling and colleagues commented that, in general, luciferase activity was ≈50-fold higher in cells expressing a TK-Luc construct compared with cells expressing a Luc-TK construct.¹¹ Therefore, we used the wild-type/humanized TK-Luc Söling construct, creating the virus Ad-pIX-TK-Luc. On the other hand, TK activity may be improved by constructing another configuration of this virus whereby TK is the carboxyl component (Ad-pIX-Luc-TK) and is not “sandwiched” between pIX and Luc. Therefore, the TK activity in our current virus configuration may fall short of its maximal potential owing to steric hindrance. Also, it may be possible to improve on the sensitivity of imaging by constructing viruses with mutations in TK. Mutating certain amino acids within HSV-TK has been shown to create an enzyme that has higher TK activity.^29–31 In addition, based on the successful rescue of this complex virus, there are multiple possibilities by which various mosaic viruses could be constructed. For instance, a mosaic virus could be constructed with a fusion of TK-Luc to pIX at its native locale in combination with the addition of recombinant pIX containing a fusion of TK-Luc at other locales of the adenoviral genome. In addition, viruses containing dual-imaging motifs in their capsid could be made by creating mosaics with separate reporter gene function. For instance, a mosaic virus can be constructed that contains a pIX fusion of TK at the wild-type pIX locale in combination with a recombinant pIX-Luc fusion at other genome locales. Empiric evaluation would be needed to find the best strategy.

In addition, this imaging platform may be further extended via construction of viruses that have a triple-fusion imaging protein at pIX since the upper size limit of pIX incorporation has yet to be defined. Recently, the generation of various triple-fusion proteins for imaging with different reporters has been cited by two groups.^31,32 These triple fusions consist of wild-type or mutated TK, a fluorescent protein (EFGP, DsRed2, or monomeric red fluorescent protein [mRFP]), and Renilla³² or firefly luciferase,^31,32 respectively. Both groups have demonstrated that these constructs can be used for simultaneous imaging in vivo with bioluminescence, fluorescence, and PET.

In our more recent study, we illustrate dynamic monitoring of oncolytic adenovirus in vivo by genetic capsid labeling. We constructed an oncolytic virus that contained mRFP1 at pIX.²⁰ Although data with this virus were promising, there have been limitations associated with fluorescence-based imaging. Therefore, it is necessary to create oncolytic viruses that encompass additional labeling motifs and thus allow more stringent or dynamic readouts compatible with a human context, such as an oncolytic vector that contains TK on pIX. In this regard, our current pIX-TK-Luc labeling system can be applied to an oncolytic context and allow dynamic imaging of adenoviruses in small animals by using TK and luciferase simultaneously. In addition, the TK component of this system allows compatibility within the human context. In summary, we have created a virus that allows for dynamic simultaneous imaging. We plan to transition this paradigm to create vectors that express PET and fluorescence-based (EGFP) imaging fusions in the pIX locus. These dual-imaging methods will provide means for noninvasive real-time tumor imaging of adenovirus-based virotherapy interventions with the possibility of imaging analysis of both localized and disseminated neoplastic disease contexts.

Footnotes

Acknowledgments

We gratefully acknowledge Drs. Long P. Le and Ariane Söling for their valuable reagents, pIXshuttle plasmids, and HSV-TK-Luc constructs. We also acknowledge Drs. Anton V. Borovjagin and Maaike Everts for their critical reading of the manuscript.

References

Glasgow

Bauerschmitz

Curiel

Hemminki

. Transductional and transcriptional targeting of adenovirus for clinical applications. Curr Gene Ther 2004;4:1–14.

Barnett

Crews

Douglas

. Targeted adenoviral vectors. Biochim Biophys Acta 2002;1575:1–14.

Kay

Glorioso

Naldini

. Viral vectors for gene therapy: the art of turning infectious agents into vehicles of therapeutics. Nat Med 2001;7:33–40.

Rots

Curiel

Gerritsen

Haisma

. Targeted cancer gene therapy: the flexibility of adenoviral gene therapy vectors. J Control Release 2003;87:159–65.

Everts

Curiel

. Transductional targeting of adenoviral cancer gene therapy. Curr Gene Ther 2004;4:337–46.

Volpers

Kochanek

. Adenoviral vectors for gene transfer and therapy. J Gene Med 2004;6 Suppl 1:S164–71.

Everts

Dmitriev

. Fluorescently labeled adenovirus with pIX-EGFP for vector detection. Mol Imaging 2004;3:105–16.

Meulenbroek

Sargent

Lunde

. Use of adenovirus protein IX (pIX) to display large polypeptides on the virion—generation of fluorescent virus through the incorporation of pIX-GFP. Mol Ther 2004;9:617–24.

Sibley

. Genetic incorporation of HSV-1 thymidine kinase into the adenovirus protein IX for functional display on the virion. Virology 2005;338:247–58.

10.

Herschman

. Non-invasive imaging of reporter genes. J Cell Biochem Suppl 2002;39:36–44.

11.

Söling

Theiss

Jungmichel

Rainov

. A dual function fusion protein of herpes simplex virus type 1 thymidine kinase and firefly luciferase for noninvasive in vivo imaging of gene therapy in malignant glioma. Genet Vaccines Ther 2004;2:7.

12.

Ray

Bauer

Iyer

. Monitoring gene therapy with reporter gene imaging. Semin Nucl Med 2001;31:312–20.

13.

Inubushi

Sundaresan

. Optical imaging of cardiac reporter gene expression in living rats. Circulation 2002;105:1631–4.

14.

Bhaumik

Gambhir

. Optical imaging of Renilla luciferase reporter gene expression in living mice. Proc Natl Acad Sci U S A 2002;99:377–82.

15.

Jacobs

Voges

Reszka

. Positron-emission tomography of vector-mediated gene expression in gene therapy for gliomas. Lancet 2001;358:727–9.

16.

Yaghoubi

Barrio

Dahlbom

. Human pharmacokinetic and dosimetry studies of [(18)F]FHBG: a reporter probe for imaging herpes simplex virus type-1 thymidine kinase reporter gene expression. J Nucl Med 2001;42:1225–34.

17.

Massoud

Gambhir

. Molecular imaging in living subjects: seeing fundamental biological processes in a new light. Genes Dev 2003;17:545–80.

18.

Wetterwald

Van Der

Que

. Optical imaging of cancer metastasis to bone marrow: a mouse model of minimal residual disease. Am J Pathol 2002;160:1143–53.

19.

Lewis

Gambhir

. Noninvasive imaging of lentiviral-mediated reporter gene expression in living mice. Mol Ther 2003;7:681–91.

20.

Dmitriev

. Dynamic monitoring of oncolytic adenovirus in vivo by genetic capsid labeling. J Natl Cancer Inst 2006;98:203–14.

21.

Fallaux

Kranenburg

Cramer

. Characterization of 911: a new helper cell line for the titration and propagation of early region 1-deleted adenoviral vectors. Hum Gene Ther 1996;7:215–22.

22.

Zhou

da Costa

. A simplified system for generating recombinant adenoviruses. Proc Natl Acad Sci U S A 1998;95:2509–14.

23.

Lyons

Forry-Schaudies

Otto

. An improved retroviral vector encoding the herpes simplex virus thymidine kinase gene increases antitumor efficacy in vivo. Cancer Gene Ther 1995;2:273–80.

24.

Alauddin

Shahinian

Gordon

Conti

. Direct comparison of radiolabeled probes FMAU, FHBG, and FHPG as PET imaging agents for HSV1-tk expression in a human breast cancer model. Mol Imaging 2004;3:76–84.

25.

Green

Yap

Nguyen

. Indirect monitoring of endogenous gene expression by positron emission tomography (PET) imaging of reporter gene expression in transgenic mice. Mol Imaging Biol 2002;4:71–81.

26.

Parks

. Adenovirus protein IX: a new look at an old protein. Mol Ther 2005;11:19–25.

27.

Campos

Parrott

Barry

. Avidin-based targeting and purification of a protein IX-modified, metabolically biotinylated adenoviral vector. Mol Ther 2004;9:942–54.

28.

Vellinga

Rabelink

Cramer

. Spacers increase the accessibility of peptide ligands linked to the carboxyl terminus of adenovirus minor capsid protein IX. J Virol 2004;78:3470–9.

29.

Gambhir

Bauer

Black

. A mutant herpes simplex virus type 1 thymidine kinase reporter gene shows improved sensitivity for imaging reporter gene expression with positron emission tomography. Proc Natl Acad Sci U SA 2000;97:2785–90.

30.

Luker

Sharma

Pica

. Noninvasive imaging of protein-protein interactions in living animals. Proc Natl Acad Sci U S A 2002;99:6961–6.

31.

Ponomarev

Doubrovin

Serganova

. A novel triple-modality reporter gene for whole-body fluorescent, bioluminescent, and nuclear noninvasive imaging. Eur J Nucl Med Mol Imaging 2004;31:740–51.

32.

Ray

Min

. Imaging tri-fusion multimodality reporter gene expression in living subjects. Cancer Res 2004;64:1323–30.