|Home | About | Journals | Submit | Contact Us | Français|
Current gene therapy technology is limited by the paucity of methodology for determining the location and magnitude of therapeutic transgene expression in vivo. We describe and validate a paradigm for monitoring therapeutic transgene expression by noninvasive imaging of the herpes simplex virus type 1 thymidine kinase (HSV-1-tk) marker gene expression. To test proportional coexpression of therapeutic and marker genes, a model fusion gene comprising green fluorescent protein (gfp) and HSV-1-tk genes was generated (tkgfp gene) and assessed for the functional coexpression of the gene product, TKGFP fusion protein, in rat 9L gliosarcoma, RG2 glioma, and W256 carcinoma cells. Analysis of the TKGFP protein demonstrated that it can serve as a therapeutic gene by rendering tkgfp transduced cells sensitive to ganciclovir or as a screening marker useful for identifying transduced cells by fluorescence microscopy or fluorescence-activated cell sorting (FACS). TK and GFP activities in the TKGFP fusion protein were similar to corresponding wild-type proteins and accumulation of the HSV-1-tk-specific radiolabeled substrate, 2′-fluoro-2′-deoxy-1β-d-arabino-furanosyl-5-iodo-uracil (FIAU), in stability transduced clones correlated with gfp-fluorescence intensity over a wide range of expression levels. The tkgfp fusion gene itself may be useful in developing novel cancer gene therapy approaches. Valuable information about the efficiency of gene transfer and expression could be obtained by non-invasive imaging of tkgfp expression with FIAU and clinical imaging devices (gamma camera, positron-emission tomography [PET], single photon emission computed tomography [SPECT]), and/or direct visualization of gfp expression in situ by fluorescence microscopy or endoscopy.
Experimental and clinical cancer gene therapy strategies were first developed by using transduction of the thymidine kinase gene of the herpes simplex virus type 1 (HSV-1-tk) into proliferating tumor cells [1–5]. Further approaches for gene therapy of cancer have used genes encoding other prodrug activating enzymes, such as cytochrome P450 ; cell-cycle regulating proteins, such as p53 and pRB ; inhibitors of angiogenesis ; or cytokines  to enhance the immune response to tumor antigens or have taken advantage of the selective oncolysis mediated by mutant herpes or adenoviruses [10–19] [20,21]. Successful gene therapy of tumors in a clinical setting appears to be limited by insufficient and nonspecific gene delivery. Noninvasive imaging technologies that provide information about the localization, magnitude, and duration of therapeutic transgene expression in vivo would greatly facilitate assessment and implementation of novel gene therapy vectors with better gene delivery and expression characteristics.
Currently, several approaches are being developed to establish clinically applicable noninvasive imaging technologies of transgene expression [22–29]. These imaging technologies are based on the detection of marker enzyme-mediated accumulation of specific radiolabeled marker substrates [22–25], binding and accumulation of marker receptor-specific compounds [26–28], or binding and accumulation of gene-specific oligonucleotides  with gamma camera, positron emission tomography (PET), and magnetic resonance imaging (MRI).
The vast majority of therapeutic genes, however, do not have appropriate marker substrates that can be radiolabeled or would accumulate in transduced tissues to a level that could be detected by noninvasive imaging. Therefore, it is necessary to develop indirect imaging strategies for monitoring the expression of different therapeutic genes. This can be achieved by functional coexpression of marker and therapeutic genes. The most important requirement of this approach is a proportional and constant coexpression of the marker gene (HSV-1-tk or gfp) and “therapeutic” gene over the wide range of expression levels. Proportionality of coexpression may be achieved by using an internal ribosome entry site (IRES) element of picornaviruses  to link the two genes in an expression cassette under control of a single promoter [31–35]. A more direct approach, is to produce a fusion gene that would encode for a “marker/therapeutic” fusion protein [36,37].
In this report, we describe a paradigm for monitoring the expression of therapeutic transgenes that retain functional activity when coexpressed as a fusion protein. This can be achieved when both the therapeutic and marker subunits of the fusion protein retain functional activity. In this context, we constructed a tkgfp fusion gene and assessed the proportionality of HSV-1 thymidine kinase (TK) and green fluorescent protein (GFP) activity over the wide range of expression levels in transduced cells in culture.
Rat 9L gliosarcoma cells  were grown as monolayers in Dulbecco's modified Eagle medium (DMEM; Mediatech, Washington, DC) supplemented with 10% fetal bovine serum (FBS; Sigma, St. Louis, MO) and 100 U/mL penicillin and 100 µg/mL streptomycin (P/S; Sigma). The RG2 and W256 cell lines were grown in MEM supplemented with 10% FBS and antibiotics. All cells were grown at 37°C in a 5% Carbon dioxide/95% air atmosphere.
The cDNA encoding the thymidine kinase gene of HSV-1 was obtained by digestion of pLTKRNL  with BamHI. To facilitate further cloning, the 2.2-kb fragment encoding the open reading frame of the HSV-1-tk was isolated and ligated into plasmid, pcDNA3.1/Zeo (Invitrogen), after digestion with BamHI. The resulting plasmid (pTKZeo3.2; 7.8 kb) was sequentially digested with BamHI and XmaI. BamHI cuts 57 base pairs (bp) in front of the start codon of tk and XmaI cuts 22 bp in front of the STOP codon. The resulting 1.2-kb tk fragment was isolated and ligated to the 4.7-kb DNA fragment obtained after plasmid pEGFP-N1 (Invitrogen) was digested sequentially with BgIII and XmaI. After ligation, the open reading frames of both tk and gfp are in-frame within the pTKGFP expression plasmid (5.9 kb), placing the tkgfp cDNA under control of the cytomegalo-virus (CMV) immediate early (IE) 1 promoter (Figure 1).
A retroviral vector encoding the fusion protein TKGFP was constructed in the Moloney murine leukemia virus-based SFG vector backbone , as previously described . pTKGFP was digested with Not I, then treated with Klenow to blunt the fragment, and then digested with MIu I. The vector MoT (Figure 2A), which encodes HSV-1-tk alone , was digested with BamHI, blunted with Klenow, and then digested with MIu I. The construct was named as SFG-TKGFP (Figure 2B). Ten micrograms of plasmid DNA was transfected by calcium phosphate coprecipitation as previously described  into the GPG29 packaging cell line . VSV-G-pseudotyped particles were used as cell free viral stocks  to transduce tumor cell lines.
The rat RG2 glioma and rat W256 carcinoma cells were exposed to the medium containing SFG-TKGFP retrovirus (~106/mL) in presence of polybrene (8 µg/mL) for 8 hours. GFP-positive transduced RG2TKGFP + and W256TKGFP + cells were detected by fluorescence microscopy. Single-cell-derived clones were obtained from a mixed population of transduced RG2TKGFP + and W256 TKGFP + cells by seeding the cells into the 96 well plates at a density of 0.1 cell/well. GFP-positive RG2TKGFP + and W256 TKGFP + colonies were identified by fluorescence microscopy and used for further propagation into single-cell-derived clonal populations.
The initial functional assessment of the tkgfp gene construct included the determination of green fluorescence and ganciclovir (GCV)-sensitivity in transiently transfected rat 9L gliosarcoma cells in culture. For both evaluations, 80% confluent monolayers of 9L cells (600,000/35mm dish) were transfected with the plasmid pTKGFP or control plasmids pHSV-TK and pHSV-GFP (see the next section) or no vector with lipofectamine (GIBCO BRL), according to the manufacturer's protocol (1 µg DNA +6 µl lipofectamine per 35-mm dish). For fluorescent detection of gfp-expressing cells, culture plates were examined with a standard fluorescence microscope (Zeiss IM35; Carl Zeiss Inc, Thornwood, NY) 24 hours after transfection. For determination of the ganciclovir sensitivity, GCV-treatment was carried out at final concentrations of 0, 0.1, 0.3, 1, 3 and 9 µg/mL medium for half of the dishes starting 24 hours after transfection. Four days later, surviving 9L cells were determined as a percentage of non-GCV-treated transfected cells. A tk-bearing HSV-1 amplicon plasmid (pHSV-TK, 11.8 kb) featuring the tk gene under control of the CMV IE1 promoter (Figure 1) and an HSV-1 amplicon plasmid containing gfp under control of the CMV IE1 promoter (pHSV-GFP; 10.4 kb; Figure 1) served as positive controls. All experiments were performed in triplicate and repeated 3 times.
The expression of GFP in retrovirally transduced single-cell derived clones of RG2TKGFP + and W256 TKGFP + cells was assessed by fluorescence-activated cell sorting (FACS) analysis of GFP fluorescence on a FAC-Sstar Plus analyzer (Becton Dickinson, NJ). Cells were detached by trypsinization, counted suspended in phosphate-buffered solution (PBS) at 4°C at a density of 106 cells/mL. GFP fluorescence was measured at the wave length of 488 nm. Parental (nontransduced) RG2 and W256 cells were used as negative controls. Gates were designed to include viable single cells only. The range of fluorescence was set to include no more than 1% of corresponding control cells. The mean GFP fluorescence per cell was used as a measure of GFP expression in a clonal population of transduced cells.
The single-cell-derived clones of RG2TKGFP + and W256 TKGFP + cells were seeded in 150 x 25-mm culture dishes (at 5000 cells per dish) and grown until 50% confluent. The incubation medium was replaced with 14 mL of medium containing 2-[14C]-FIAU (fialuridine) (56 mCi/mmol) and methyl-[3H]-TdR (65.4 mCi/mmol) (Moravek Biochemicals, Brea, CA). Radiochemical purity of each compound was checked in our laboratory by high performance liquid chromatography (HPLC) and found to be more than 97% pure. The concentrations of 2-[14C]-FIAU and methyl-[3H]-TdR (thymidine deoxyribose) were 0.01 and 0.1 µCi/mL, respectively. The cells were harvested by using a scraper after various periods of incubation (10, 30, 60, 90, and 120 minutes), weighed, and assayed with a Packard B1600 TriCarb β-spectrometer and standard 3H and 14C dual-channel counting techniques. The medium was counted before and after incubation. The data were expressed as a harvested cell-to-medium concentration ratio (dpm/g cells)/(dpm/mL medium) and plotted against time . The steady-state accumulation rate of FIAU, normalized by that of TdR, was obtained from the slopes of the plot and used as a measure of HSV-1-tk gene expression in culture (FIAU/TdR ratio).
The expression of TK and GFP were evaluated by Western blotting with a rabbit polyclonal HSV-TK antiserum and a rabbit anti-GFP antibody (CLONTECH, Palo Alto, CA). Twenty-four hours after transfection of 4 µg each of purified pTKGFP, pHSV-TK (positive control for TK), or pHSV-GFP (positive control for GFP) into 9L cells, monolayers were harvested and resuspended in phosphate buffered saline (1 x PBS) and processed as described previously . In brief, after cell lysis, samples were boiled for 5 minutes and electrophoresed on denaturing 10% acrylamide gels overnight. Proteins were transferred to a Hybond-ECL membrane (Amersham, Arlington Hts., IL) by standard tank transfer in Tris-glycine transfer buffer at 45 V (0.5 mA) for 3 hours. Nonspecific membrane binding was blocked by incubating the membrane in 10% nonfat dry milk in PBS-Tween overnight. After washing (3x), the membrane was incubated with primary antibodies (anti-TK 1:1000 and anti-GFP 1:1000, respectively, in 1% nonfat dry milk/PBS-Tween) for 1 hour at room temperature. After washing (3x), the membrane was incubated with horseradish peroxidase-conjugated horse antirabbit (1:4000) immunoglobulin in blocking solution for 1 hour, then washed, and developed by using the ECL reagents (Amersham).
The mean values ± SD of cell counts and the percent differences between GCV-treated and nontreated dishes were calculated and results plotted with the software package SigmaPlot 1.02. The analysis of coexpression of HSV-1-tk and gfp genes was done with StatView 4.57 (Abacus Concepts, CA) and Kaleidagraph 3.08 Synergy Software, CA).
The expression plasmid pTKGFP bears the HSV-1-tk gene in-frame to the gfp gene under control of the CMVIE promoter (Figure 1). Twenty-two base pairs (bp) are missing at the 3′ end of the open reading frame of tk to remove the stop codon. Eighteen base pair in-frame encoding Pro-Gly-Ser-Ile-Ala-Thr were introduced between the open reading frames for tk and gfp. HSV-1 amplicon plasmids, which feature the tk gene or the gfp gene alone under the same CMV promoter (CMV IE1; Figure 1) served as positive controls for subsequent experiments. In the retroviral vector, SFG-TKGFP, the tkgfp fusion gene is under control of the 5′-retroviral LTR promoter (Figure 2).
In order to compare the levels of gfp-expression, plasmids pTKGFP and pHSV-GFP were transfected into rat 9L gliosarcoma cells. The levels of gfp-expression were assessed 1, 2, 3, and 5 days after transfection with a standard fluorescence microscope and were visually of similar intensity in tkgfp and native gfp transfected cells (Figure 3). The data indicate that tkgfp transduced cells can be readily identified by fluorescent light at least to the same extent as gfp transduced cells. As was pointed out by Loimas et al. , the intracellular expression pattern of the TKGFP seems different than the native GFP. TKGFP is mostly concentrated in the nucleus, whereas the native GFP displays a predominantly cytoplasmic distribution.
In order to investigate the ability of the TKGFP fusion protein to mediate significant sensitivity towards the pro-drug ganciclovir, pTKGFP and pHSV-TK transduced 9L glioma cells (positive control) were treated with different amounts of GCV ranging from 0 to 9 µg/mL over a 5-day exposure period. After the treatment period, surviving cells were determined as a percentage of non-GCV-treated controls. In contrast to pHSV-GFP and nontransfected negative controls, the pTKGFP and pHSV-TK transfected 9L cells were both sensitive to GCV in a similar dose-dependent manner (Figure 4).
To investigate the spectrum of transgene coexpression that is achieved in polyclonal cell populations, we generated two series of clones to examine coexpression of both genes in the context of fusion gene expression. We obtained multiple retrovirally transduced RG2TKGFP + and W256TKGFP + clones with different levels of transgene expression (due to different genomic sites of retroviral integration) to explore coexpression over a broad range of transgene expression levels. The proportionality of the HSV-1-tk and gfp gene expression was assessed in culture by using the radiotracer and FACS assays, respectively. For FACS analysis of GFP, the range of fluorescence was set to include no more than 1% of corresponding control cells (Figure 5). The mean GFP fluorescence per cell was used as a measure of GFP expression in a clonal population of transduced cells. Regression analysis demonstrated a strong relationship between the levels of expression of gfp and HSV-1-tk genes in corresponding clones of retrovirally transduced RG2 glioma cells (RG2TKGFP +; Figure 6A) and in W256 carcinoma cells (W256TKGFP +; Figure 6B). Combining the RG2TKGFP + and W256TKGFP + data sets yields a similar relationship over a wide range of transgene expression levels (Figure 6C). Statistical analysis of the combined data also revealed that there was a significant correlation between the HSV-1-tk and gfp expression (r = 0.894; P < .0001; and t = 2.477; P < .05 based on paired t-test analysis). This suggests that the relationship of fusion gene-mediated coexpression is independent of tissue type.
In order to investigate the tkgfp fusion gene expression at the protein level, cellular proteins of pTKGFP, pHSV-GFP, and pHSV-TK-transfected 9L cells harvested at 24 hour after transfection were resolved by SDS-PAGE and analyzed on Western blots by using antibodies to TK and GFP (Figure 7). The predominant band recognized by both the TK and the GFP antibodies is 66 kD, the appropriate size of the TKGFP fusion protein. TK and GFP proteins are, as expected, recognized only by their specific antibodies and maintain their native size (approximately 46 kD and 27 kD, respectively).
The paradigm for noninvasive imaging of therapeutic gene expression that we propose in this report relies on proportional coexpression of a “marker” gene and a “therapeutic” gene products. Usually, the coexpression of two gene products is achieved by using a dual-promoter approach. However, on an individual cell level, as opposed to cell populations in which total expression is averaged, proportional coexpression may be compromised because of promoter interference or transcriptional repression. The proportionality of IRES-mediated coexpression of two genes is determined at the translational level and depends on protein synthetic activity and cell cycle phase of the transduced cells . Strictly proportional (equimolar) coexpression of two genes under control of a single promoter can be achieved by using fusion gene approach.
To assess whether the fusion gene approach is applicable to the described imaging paradigm, we constructed the tkgfp fusion gene consisting of the HSV-1-tk marker gene (for noninvasibe imaging) and the gfp gene, which served as model gene for different therapeutic transgenes. Transduction of different types of tumor cells with tkgfp allowed the selection of transduced living cells by fluorescence microscopy and quantitation of tkgfp expression with FACS analysis. Expression of tkgfp induced sensitivity of the transduced tumor cells to GCV similar to that resulting from transduction of the cells with the same vector bearing the HSV-1-tk gene alone. The levels of HSV-1-tk gene expression in tkgfp-transduced clones, as measured by the FIAU accumulation, were comparable to those observed by us previoulsy in RG2 and W256 tumor cells transduced with HSV1-1-tk gene alone [22,23] and were within the range that is adequate for noninvasive imaging with radioiodinated FIAU and gamma camera, single photon emission computed tomography (SPECT), or PET [22–24]. Previously, we have shown that the FIAU/TdR accumulation ratio correlates highly with other independent measures of the HSV-1-tk gene expression: namely, ganciclovir sensitivity (IC50) and concentration of HSV-1-tk mRNA .
Most importantly, we demonstrated that fusion gene approach provides a constant relationship between the expression of the HSV-1-tk and gfp genes over a wide range of expression levels. Thus, noninvasive images of HSV-1-tk expression should reflect distribution and level of expression of gfp or different therapeutic genes. Parametric images of therapeutic gene expression could be generated if the relationship between the levels of FIAU accumulation (HSV-1-tk expression) and functional measures of therapeutic gene expression in transduced tissues are known.
Limitations of the fusion gene-fusion protein approach is that is cannot be applied to proteins that may loose functionality as the result of fusion or will not be able to localize to the appropriate subcellular compartment. This approach cannot be applied to secretory proteins (e.g., cytokines, growth factors) because the HSV-1-tk marker enzyme should remain intracellular in order to catalyze the intracellular accumulation of radiolabeled FIAU. Potentially higher immunogenicity of fusion proteins may also limit their application in in vivo studies.
The tkgfp fusion gene can also be considered as therapeutic-marker gene construct, where tk is used as therapeutic gene and gfp as marker gene for in situ detection or direct visualization. GFP is the most widely used marker protein for direct visualization of transgene expression in living cells and transparent tissues and organisms in vivo [46–54]. Recently, Flotte et al.  implemented a videoendoscopy technique for direct visualization of gfp-transduced airway epithelium. Similar endoscopic technique for visualization of tkgfp fusion gene expression could be useful for gene therapy directed to the gastrointestinal tract by using the HSV-1-tk/GCV paradigm. Therefore, tkgfp fusion gene could be used as a dual-marker gene for noninvasive imaging, direct endoscopic visualization, and in situ detection with fluorescence microscopy.
From a vector design prospective, the tkgfp fusion gene has several positive features. A relatively small coding region of tk (1.2 kb) and gfp (750 bp) results in a tkgfp fusion gene that is less than 2 kb in size and allows for easy incorporation into virus vectors with small transgene capacities (e.g., adeno-associated virus vector). The transgene capacity of adeno-associated virus vectors is less than 5 kb, which makes cloning of two transgenes (therapeutic and selection gene) within the same vector difficult. The GFP-fluorescence is not dependent on substrates or cofactors, which enables direct detection of transgene expression in living cells. Therefore, the need for time-consuming immunocytochemistry and histochemistry for the detection of transduced cells can be avoided. This is particularly important because of the lack of a readily available antibody for HSV-1-tk. Furthermore, GFP-expressing cells can be rapidly isolated from a pool of transduced cells by FACS. This avoids time-consuming drug selection and allows for fractionation of cells with respect to the level of GFP expression and, in turn, the level of tk expression.
The proportionality of coexpression of two genes in fusion, the HSV-1-tk and gfp, was demonstrated over the wide range of expression levels. Fusion gene approach should be applicable to the paradigm of monitoring therapeutic transgene expression by noninvasive imaging of marker gene coexpression. Noninvasive imaging of transgene expression should facilitate assessment of the efficacy and target specificity of novel gene delivery vectors by defining the distribution, level, and temporal profile of expression of different therapeutic genes in vivo. It would contribute significantly to further development of vector technology and gene therapy in general.
We thank Dr. David Jacoby for the kind gift of plasmid construct pHSV-GFP. We thank Dr. Ronald G. Blasberg for helpful discussions during preparation of this manuscript. The RG2 rat glioma cells were kindly provided by Dr. D. Bigner (Duke University Medical Center, Durham, NC). The W256 rat mammary carcinoma cells were obtained from American Type Tissue Culture Collection. The rabbit polyclonal HSV-TK antiserum was kindly provided by Dr. W. Summers, Yale University School of Medicine. Support for this work was received from the Gertrud Reemtsma Foundation (S 104), Ministerium SWWF/LNRW (516-40000299), ZMMK (TV46), the Max-Planck-Society and the Center of Molecular Medicine, Cologne, Germany (A.J.); from NCI grant CA69246 (X.O.B.); from James McDonnell Foundation Award and NIH RO1 CA59350 M.S.; and from NIH RO1 CA76117 (J.G.T.).