Introduction

Molecular imaging encompasses several frequently used methods that allow researchers to track specific molecular markers in vivo, as well as follow physiological processes or determine functional status. These techniques have developed to provide not only morphological data [1–5], but also quantitative or other digital information [6–11]. The ideal modality for in vivo imaging requires a sensitive, highly stable, biologically compatible, and relatively nontoxic molecular tracer [12, 13]. In addition, imaging molecules for cytological applications should have low elimination ratio in living cells [14], and should be minimally affected by cell viability or proliferation [15]. One of the most important types of molecular imaging is optical imaging, which is generally used to visualise cells and animal models [16]. For example, researchers use optical imaging to follow the migration, development, differentiation, and therapeutic effects of stem cell transplants [17, 18].

Optical imaging methods are typically based on fluorescence or bioluminescence [19, 20]. Fluorescent proteins are used in the first approach, usually to study morphological structures. On the other hand, by introducing luciferase gene into cells or tissues, bioluminescence imaging generates not only qualitative but also quantitative data. Luciferase uses oxygen, magnesium, and adenosine triphosphate (ATP) to emit light that can be detected by a charge-coupled device [21]. Since mammalian animals or cells do not express luciferase endogenously, the background is much lower. For the same reason, bioluminescence imaging is advantageous in terms of signal to noise ratio, practicality, and specificity [22, 23].

We have developed a dual-reporter method that exploits both fluorescence from enhanced green fluorescent protein (eGFP) and bioluminescence from luciferase. By integrating a fusion of Luc2 and eGFP into host chromosomal DNA, bioluminescent and fluorescent markers maybe expressed as required in regions or structures of interest, both in cells or living animals. To demonstrate this technique, we transformed the dual-reporter construct into mouse brains using HIV-1 lentiviral particles, which were prepared from HEK293T cells by a three-plasmid lentiviral system. We evaluated multiple technical properties of the tracer molecule, including sensibility, biocompatibility, and lifetime. For instance, we found that quantifiable optical signals from luciferase [21, 24] are sustained for at least 28 days in the mouse brain. Furthermore, by modifying the lentiviral vector, other genes could potentially be introduced to enable gene function studies in specific tissues or cells in vivo.

Materials and Methods

Construction of FUGW-Luc2-eGFP

A 1652-bp fragment of luc2 was amplified by PCR from the plasmid pGL4.17 [25] using primers with sequence 5′-GAC TCT AGA GCC ACC ATG GAA GAT GCC AAA AAC ATT AAG AA-3′ and 5′-CGG GGA TCC TTC ACG GCG ATC TTG CCG CCC TT-3′. The fragment was then subcloned into a hUbC-driven eGFP lentiviral plasmid, FUGW vector [26] using XbaI and BamHI. The final construct FUGW-Luc2-eGFP (reporter construct) was verified by sequencing.

Lentivirus Preparation

We used a three-plasmid system to prepare lentiviral vectors [27–29] from HEK293T cells. Cells were transiently transfected with Lipofectamine™2000 (Invitrogen) containing the reporter construct FUGW-Luc2-eGFP, the packaging plasmid pCMV-deltaR8.2 (Addgene, USA), and the envelope plasmid pCMV-VSV-G (Addgene, USA) in order to produce pseudotyped viral vectors that are secreted into the culture supernatant. The plasmids were transfected at molar ratio 4:3:2. The culture supernatant was harvested 48 h after transfection, and gently mixed with PEG8000 at a final concentration of 20 %. Vectors were then precipitated by centrifugation for 30 min and 4 °C at 1500×g, and re-suspended in 25 mM HEPES pH 7.05 containing pre-cooled, sterilized phosphate buffer saline (PBS) pH 7.4. Vectors were re-suspended in 1/100 volume of the original culture supernatant.

Infective Titre

The tenth echelon dilution method was used to measure infective virus titer [30]. Briefly, HEK293T cells were plated in a 96-well plate at a density of 2 × 104 cells/100 μL in each well, and infected with a serial dilution of lentiviral vectors. Culture media were replaced with fresh media 6 h after infection, and the number of fluorescent cells in each well was determined 96 h after infection. Viral titer was calculated based on the highest dilution factor that produced fluorescent cells.

Infectivity in Cells

Frozen endothelial progenitor cells (EPC) and GL261 glioma cells were thawed and cultured. Cells were then dissociated using 0.25 % (w/v) trypsin, and transferred to serum-free media containing 5 mg/mL polybrene (Sigma-Aldrich) as well as lentiviral particles at multiplicity of infection (MOI) 100. Transduction was analysed by immunocytochemistry and flow cytometry 72 h after transfection.

For immunocytochemistry, cells were mounted on slides, fixed with pre-cooled acetone, and incubated for approximately 30 min in PBS containing 3 % bovine serum albumin, 0.1 % gelatine, and 0.1 % Triton X-100. Samples were probed with 1:500 anti-GFP (Life Technologies) in PBS, and then labelled for 60 min at room temperature with 1:2000 FITC-conjugated rabbit anti-chicken IgY (Life Technologies) in PBS. Cells were examined through an upright fluorescence microscope (Carl Zeiss, Jena, Germany). For flow cytometry, approximately 107 cells were suspended in 1 mL culture media and sorted based on eGFP fluorescence. Cells were analyzed with a FACScan instrument (FACSCalibur, BD Biosciences, USA). A total of 5000 events were analysed for each sample.

Transduction in Mice

Mice were anaesthetized for 3 min with 3.5 % isoflurane in a semi-airtight chamber, and placed in a stereo-taxic frame. During surgery, animals were administered 1 % (v/v) isoflurane, 67 % (v/v) nitrogen, and 32 % (v/v) oxygen at standard atmospheric pressure through a respiratory mask. A small hole was drilled into the skull, 1.5 mm to the lower right of the bregma midpoint. A virus suspension (2 μl) was injected 1.5 mm underneath at 0.4 μL/min. The needle was kept in place for 3 min after injection. After surgery, mice were monitored until consciousness was regained.

In Vivo Imaging

Mice were anesthetized, and placed in an IVIS Spectrum in vivo preclinical imaging system (Perkin Elmer Corporation, USA) instrument, a bioluminescence imaging system [31]. Each mouse was analysed at days 1, 3, 7, 14, 21, and 28. Images of regions of interest (ROI) were collected 10 min after intraperitoneal injection of D-luciferin, the luciferase substrate, at 3.75 mg per 25 g mouse weight.

Ethics Statement

Laboratory mice were provided care according to “The Care and Use of Laboratory Animals” by the Laboratory Animal Center of Southeast University.

Analysis of Protein Structures

Protein structures were obtained from the PDB (http://www.rcsb.org/pdb/home/home.do), and PyMol was used for structural analysis and alignment.

Results and Discussion

Generation of Lentiviral Vectors Based on HIV-1

To insert two reporter genes in chromosomal DNA, we used a three-plasmid packaging system to produce lentiviral vectors based on HIV-1. The first plasmid, the reporter construct FUGW-Luc2-eGFP (Online Resource 1), contains a fusion of luciferase and eGFP, and was generated by subcloning luc2 into an empty FUGW plasmid. Clones were verified by restriction enzyme digestion and PCR to contain fragments with nearly identical electrophoretic mobility of about 1652-bp (Fig. 1). Clones were sequenced for further verification.

Fig. 1
figure 1

Agarose gel electrophoresis of reporter constructs. Lanes 1 and 3, PCR product from pGL4.17 (lane 1) and FUGW-Luc2-eGFP (lane 3); lane 2, FUGW-Luc2-eGFP digested with XbaI and BamHI

Full size image

FUGW-Luc2-eGFP was then transfected into HEK293T cells, along with the plasmids pCMV-deltaR8.2 and pCMV-VSV-G, to produce lentiviral vector particles, which were secreted into the culture media [32] and harvested 48 h after transfection. The particles were then assayed for infectivity.

Infectivity Titre

To measure infection titre, we used the simplest available method, in which the viral stock is serially diluted until the number of host cells far exceeds the number of active particles. At this point, every single active particle present would have infected a cell. Thus, HEK293T cells were infected with a serial dilution of lentiviral particles, and the number of resulting fluorescent cells was counted The infection titre of the lentiviral particles was 109 transducing units per mL, based on the number of fluorescent HEK293T cells after infection (Fig. 2).

Fig. 2
figure 2

Infectivity titre. HEK293T cells were infected with different volumes of lentiviral stock. The minimum volume that produced a fluorescent cell was 10−6 μL (red arrow head). Bright-field images are also shown. Scale bar, 20 μm. Magnification: ×100

Full size image

The infection titre is typically considered to be a measure of the concentration of active lentiviral particles. Nevertheless, it has little real validity, but is only used to calculate MOI [33], which is the ratio of active viral particles to host cells.

Infectivity in Cells

The infectivity of viral preparations was tested on endothelial progenitor cells (EPC) from human umbilical cord blood [34] and GL261 glioma cells from a murine brain tumour [35]. Both cell lines were infected at MOI 100. Immunocytochemistry indicates that a significant majority of endothelial progenitor and GL261 cells expressed eGFP after infection (Fig. 3). The fraction of fluorescent endothelial progenitor and GL261 cells was 84.8 and 91.69 %, respectively, based on flow cytometry (Fig. 3).

Fig. 3
figure 3

Transduced cells. Epithelial progenitor (EPC) and GL261 cells were analysed 72 h after transfection by immunocytochemistry against eGFP, and by flow cytometry based on GFP. Scale bar, 50 μm. Magnification: ×200

Full size image

Dual-reporter imaging enables the use of trackable, quantifiable, and biocompatible reporters to investigate differentiation, cell cycling, drug effects, and processes such as synaptogenesis. For example, a potential application is to track endogenous or implanted stem cells. Hence, we tested endothelial progenitor because of their stem cell-like properties.

However, we foresee the dual-reporter system to have wider application in the development of drugs, especially antitumor agents. Therefore, we also tested GL261, a mouse glioma cell line. To assess the antitumor properties of a compound, researchers may generate tumour models by implanting neoplastic cells or by using genomic mutants. Data are then obtained, traditionally through analysis of the morphological features of a tumour, or through biochemical analysis of samples from the model. In contrast, dual-reporter imaging will allow experimentalists to visualize and quantify the effects of drug intervention in vivo, as well as tumour progression and other carcinogenic processes.

In Vivo Bioluminescence Imaging

To test if the dual-reporter system is suitable for use in an animal model, we infected mouse brains by orthotopic injection. As a result, luc2-eGFP was introduced in part of the cerebral histocyte, although the infection efficiency was unknown. Bioluminescence was evident beginning at day 3. The signal peaked to 105 p/s/cm2/sr at day 14, and remained substantial through day 28 and beyond (Fig. 4). These results indicate that the system can be used in animal models through direct infection of viable tissue. Alternatively, in vivo bioluminescence may be introduced in other ways, such as by implanting cells that are already infected [36]. Notably, the results also imply that D-luciferin crosses the blood-brain barrier, and will probably reach any other region of interest. Taken together, the data suggest that the dual-reporter system is a powerful, widely applicable tool.

Fig. 4
figure 4

Luciferase imaging in vivo. An infected mouse was analysed at days 1, 3, 7, 14, 21, and 28 in an IVIS Spectrum instrument. Bioluminescence intensity is coloured according to the colour scale (×105 p/s/cm2/sr)

Full size image

Comparison of the Dual-reporter Fusion with eGFP and Luciferase

EGFP fluorescence is easily detected with a flow cytometer or a fluorescence microscope. However, because fluorescence does not penetrate tissues, eGFP is unsuitable for viviperception of animal models, especially when the region of interest is internal. In addition, anti-eGFP may be required to label and enhance eGFP fluorescence. On the other hand, luciferase cannot be directly tracked, because it is not fluorescent and requires luciferin. However, the enzyme emits photons that penetrate tissues, are quantifiable, and enable tracking and visualization of biological processes in vivo. The fused dual reporter, Luc2-eGFP, combines the capabilities of both reporters, and, consequently, acquires significant advantages over each.

A main concern with protein fusions is the possibility of loss of function from either or both components. The beta-barrel structure of eGFP (PDBID: 4EUL) is rigid (Online Resource 2A), and should not be significantly affected by N- or C-terminal fusions. Unfortunately, the available structure for Luc2 (PDBID: 1BA3) does not clearly show the cofactor-binding pocket, which is critical for activity. Therefore, we used the structure of the homolog (PDBID: 2D1R) from the Japanese firefly to locate the cofactor-binding pocket. The two structures are very similar and align well (Online Resource 2B) except at the C-terminus, making the root mean square deviation for global alignment was as high as 1.059. As the cofactor-binding pocket is not at the C-terminus, we concluded that the pocket in Luc2 is formed by the corresponding motifs in the Japanese firefly luciferase (Online Resource 2C). The entrance to this pocket is 29.37 Å from the C-terminus. Hence, fusion of eGFP to the C-terminus should not block the pocket and inhibit enzyme activity. In summary, the Luc2-eGFP fusion should have fully functional Luc2 and eGFP, although we are unable to provide quantitative evidence.