Noninvasive Imaging of Cationic Lipid-Mediated Delivery of Optical ...

4 downloads 0 Views 300KB Size Report
1The Crump Institute for Molecular Imaging, 2Department of Molecular & Medical ... 4UCLA–Jonsson Comprehensive Cancer Center, School of Medicine, ...
doi:10.1006/mthe.2002.0700, available online at http://www.idealibrary.com on IDEAL

METHOD

Noninvasive Imaging of Cationic Lipid-Mediated Delivery of Optical and PET Reporter Genes in Living Mice Meera Iyer,1 Manijeh Berenji,1 Nancy S. Templeton,3 and Sanjiv S. Gambhir1,2,4,* 1

The Crump Institute for Molecular Imaging, 2Department of Molecular & Medical Pharmacology, and 3Center for Cell and Gene Therapy, Baylor College of Medicine, Houston, Texas 77030, USA 4 UCLA–Jonsson Comprehensive Cancer Center, School of Medicine, University of California, Los Angeles, Los Angeles, California 90095, USA *

To whom correspondence and reprint requests should be addressed. Fax: (310) 209-4655. E-mail: [email protected].

Gene therapy involves the safe and effective delivery of one or more genes of interest to target cells in vivo. The advantages of using nonviral delivery systems include ease of preparation, low toxicity, and weak immunogenicity. Nonviral delivery methods, when combined with a noninvasive, clinically applicable imaging assay, will greatly aid in the optimization of gene therapy approaches for cancer. We demonstrate cationic lipid-mediated noninvasive monitoring of reporter gene expression of firefly (Photinus pyralis) luciferase (fl) and a mutant herpes simplex virus type I thymidine kinase (HSV1-sr39tk, tk) in living mice using a cooled charge coupled device (CCD) camera and positron emission tomography (PET), respectively. We observe a high level of fl and tk reporter gene expression predominantly in the lungs after a single injection of the extruded DOTAP:cholesterol DNA liposome complexes by way of the tail vein, seen to be time- and dose-dependent. We observe a good correlation between the in vivo bioluminescent signal and the ex vivo firefly luciferase enzyme (FL) activity in different organs. We further demonstrate the feasibility of noninvasively imaging both optical and PET reporter gene expression in the same animal using the CCD camera and microPET, respectively. Key Words: cationic lipid, firefly luciferase, CCD camera, thymidine kinase, microPET, reporter gene imaging

INTRODUCTION Efficient delivery of one or more genes of interest to target cells or tissues is a prerequisite for the success of human gene therapy and requires the development of vehicles that achieve safe and selective gene transfer. Different routes are currently being pursued to achieve optimal gene delivery and expression in vivo. Main categories of gene delivery vehicles include viral and nonviral vectors. Viral vectors, such as recombinant adenoviruses, are popular because of their high efficiency following systemic administration in vivo [1]. However, antiviral immune response presents a considerable risk to patients and has limited the use of these vectors. In recent years, considerable research efforts have been directed toward developing and improving nonviral vectors for successful gene transfer in vivo. Among the nonviral vectors, cationic lipids in particular represent an attractive category. The advantages of using cationic lipid-mediated gene transfer systems include ease of production and low toxicity. Clinical trials for the treatment of melanoma [2] and cystic fibrosis [3] have used cationic lipids as gene transfer vectors.

MOLECULAR THERAPY Vol. 6, No. 4, October 2002 Copyright © The American Society of Gene Therapy 1525-0016/02 $35.00

Cationic lipids associate with negatively charged DNA to form complexes that can bind to cell surfaces by way of electrostatic interaction, thereby delivering DNA into the cells. Among the cationic lipids, different formulations have been evaluated for their gene transfer ability in vivo [4–9]. Due to a lack of understanding of the underlying mechanisms governing cationic lipid-based gene delivery, the formulations being used in different studies vary greatly in the nature of the lipid. In any cancer gene therapy protocol, knowledge of the location, magnitude, and time variation of therapeutic gene expression is critical to optimizing treatment. We and others have previously described noninvasive assay methods to image adenoviral mediated herpes simplex virus type I thymidine kinase (HSV1-tk) reporter gene expression in living animals using positron emission tomography (PET) [10–13]. Because most therapeutic genes lack ligands that can be radiolabeled and subsequently imaged, it is not feasible to determine the magnitude of therapeutic gene expression. Therefore, indirect imaging approaches using a marker/reporter gene along with the therapeutic gene are being investigated. These approaches are based on similar coexpression of both genes over a wide range of expression

555

METHOD

doi:10.1006/mthe.2002.0700, available online at http://www.idealibrary.com on IDEAL

A

FIG. 1. Small animal imaging systems used for imaging reporter gene expression. (A) MicroPET imaging system for high-resolution PET imaging. The scanner has a transverse field of view of 11 cm and an axial field of view of 1.8 cm. The spatial resolution of the scanner is 23 mm3. An animal is anesthetized, injected with a positron-emitting tracer in nonpharmacological doses, and, after a period of time (to allow the tracer to be cleared from blood and various tissues), scanned for a period of 15–45 minutes. Then, three-dimensional images are reconstructed and distribution of radioactivity can be quantitated. (B) Xenogen in vivo imaging system (IVIS) for optical imaging. The system consists of a cooled charge coupled device (CCD) camera (Roper Scientific EB1300) mounted on a light-tight imaging chamber, a cryogenic refrigeration unit, a camera controller, and a computer system for data analysis. The calibration standards show a picture of a hockey puck drilled with four holes and covered with glass (inset). Each hole contains 14C isotopes (t1/2 = 5730 years) of various amounts (100 ␮Ci, 10 ␮Ci, 100 nCi, and 10 nCi) immersed in scintillation cocktail. Light emitted from ␤-decay is detected by the cooled CCD camera, converted into photons, and transformed to RLU/minute on images. The CCD camera detects low levels of light emitted from within the animal after injection of D-Luciferin into the animal. An animal can typically be scanned over a period of 1–5 minutes. Projection (non-tomographic) images of bioluminescence can be superimposed on digital visible images of the mouse.

B

levels. Currently, there are several methods being pursued to monitor reporter gene expression in living animals [13–16]. Small animals, such as mice, provide a convenient platform to study gene expression and models of human disease because of their genetic homology with human and their extensively characterized genetics [17]. PET is one of the most quantitative techniques for imaging gene expression in vivo (Fig. 1A). Conventional PET scanners have resolutions of (5–6)3 mm3 and provide information about both the spatial and temporal distribution of the radiotracer in a single study. The extension of PET technology to image small animals has resulted in the development of small animal PET scanners [18]. The microPET is an example of a

556

dedicated small animal PET system [19,20]. The microPET system has a resolution of 23 mm3, is highly quantitative, and supports a wide range of studies that include wholebody imaging of mice. In our laboratories, noninvasive methods using radiolabeled reporter probes to monitor PET reporter gene expression in vivo have included the use of adenovirus-based approaches [11] and tumor xenograft models [12]. In parallel with these approaches, we have also been evaluating nonviral delivery systems to transfer reporter genes in a living animal and monitoring the expression over time. Non-radionuclide methods being used by researchers to track reporter gene expression in vivo include the use of an optical bioluminescent reporter gene such as firefly (Photinus pyralis) luciferase (fl) and a reporter probe (D-Luciferin) [14,21,22], a fluorescent reporter gene such as green fluorescent protein (GFP) with external light stimulation [23], and magnetic resonance imaging (MRI) techniques [24,25]. The ability to detect firefly luciferase enzyme (FL) activity in vivo has been made possible with the availability of more sensitive light detection equipment, such as cooled charge coupled device (CCD) cameras (Fig. 1B). Note that fl refers to the gene and FL to the enzyme. Luciferases are photoproteins that provide an internal source of light (wavelength range 490–560 nm), which can be measured externally as indicating the expression of the reporter gene [26].

MOLECULAR THERAPY Vol. 6, No. 4, October 2002 Copyright © The American Society of Gene Therapy

doi:10.1006/mthe.2002.0700, available online at http://www.idealibrary.com on IDEAL

CCD cameras have been used in many applications that include tracking tumor cells in mice [27] and virus-mediated gene delivery [22,28]. In the present study, we used the cationic lipid 1,2dioleoyl-3-trimethylammonium-propane (DOTAP) to condense plasmid DNA. Cholesterol was added as a helper lipid, as it is known to stabilize the lipid bilayers against degradation by plasma components and to enhance the efficiency of gene transfer [8]. Cholesterol-containing formulations are reported to have the optimal balance between the initial rapid aggregation and the subsequent slow disintegration, thereby leading to a high level of transfection efficiency [6]. These DOTAP:cholesterol (1:1 mixture) complexes are also stable at high concentrations of serum between 70 and 100% (N.S.T., unpublished data). The extruded, bilamellar, invaginated (BIVs) DOTAP:cholesterol liposomes efficiently encapsulate nucleic acids [8] and adenoviruses [29]. Therefore, these complexes produce greater transfection efficiency in vivo and greater efficacy in small animal models for cancer [30] than unilamellar (SUVs) or multilamellar (MLVs) vesicles. To combine the cationic lipid-mediated gene delivery with noninvasive imaging, we monitored the expression of fl reporter gene in living mice using a cooled CCD camera following systemic injection of the fl DNA–DOTAP:cholesterol complexes. We further used the cationic lipid-based delivery approach to image the expression of a mutant tk reporter gene, HSV1-sr39tk (tk) [31], in living mice using microPET. We further evaluated the dose- and time-dependent kinetics of fl and tk gene expression following systemic administration of the DNA–lipid complexes. We also report on the feasibility of imaging both optical and PET reporter genes in the same animal.

RESULTS

RLU/minute/␮g protein

Systemic Administration of fl DNA–Lipid Complexes Leads to Predominant Expression in Mouse Lungs Plasmid DNA (50–150 ␮g), in which a cytomegalovirus (CMV) promoter drives the expression of fl (CMV-fl), was

50 ␮g 75 ␮g 100 ␮g 150 ␮g

METHOD

complexed with DOTAP:cholesterol and administered to CD-1 mice by way of the tail vein. FL activity in the different organs was analyzed 24 hours after injection. The lungs show the highest level of fl gene expression at all concentrations following systemic injection of the complexes (Fig. 2). All other organs show very low levels of fl gene expression. At higher amounts of DNA, fl expression in the heart and liver are 10% and 60%, respectively, of that in the lungs. Expression of fl Is Dependent on DNA Dose and Imaging Time We examined the extent of fl gene expression as a function of DNA concentration and time. Mice injected with different concentrations of DNA complexed with DOTAP:cholesterol were imaged 5 hours and 24 hours after injection using the CCD camera. A significantly high level of fl gene expression is detected in the lungs of mice carrying fl gene as early as 5 hours after injection (Fig. 3A). The control mouse (injected with DOTAP:cholesterol only) shows background level of fl gene expression. At 5 hours, fl expression in the right lung appears to be greater than in the left lung. However, analysis of the data indicates that the difference is not statistically significant (P = 0.15). Expression of fl in the lung 24 hours after injection is lower than that observed at 5 hours, but still is significantly greater than background (Fig. 3B). At both time points, the lung remains the organ with the highest level of fl expression. We next examined the DNA dose-response relationship by injecting different amounts of plasmid DNA complexed with DOTAP:cholesterol. Mice injected with 10 ␮g and 25 ␮g of DNA show background levels of fl expression in all organs (relative light units (RLU)/minute < 50, data not shown). In mice injected with 50 ␮g or more of DNA, fl expression in the lung increases with an increase in the injected dose (Figs. 3A and 3B). At higher DNA concentrations (> 100 ␮g), some fl expression can be detected in the heart, liver, and spleen. FL expression in the heart at the higher DNA amounts is 10% of the lung activity as measured by a luminometer. Due to the significantly high bioluminescent signal in the lung, it is difficult to separate cardiac activity from lung activity based on the in vivo CCD images alone. During the course of the study, we also observed that a DNA dose greater than 100 ␮g was toxic to the animals. A few of the animals injected with 125 ␮g and 150 ␮g of DNA died between 5 hours and 24 hours after injection.

FIG. 2. Biodistribution of fl DNA–lipid complexes following systemic injection in CD-1 mice. A plasmid carrying fl DNA complexed with DOTAP:cholesterol (amounts ranged from 50 ␮g to 150 ␮g) was injected into CD-1 mice by way of the tail vein. The mice were sacrificed 24 hours later and the different organs were harvested and analyzed for FL activity using a luminometer. Error bars represent SEM for triplicate measurements.

MOLECULAR THERAPY Vol. 6, No. 4, October 2002 Copyright © The American Society of Gene Therapy

557

METHOD

doi:10.1006/mthe.2002.0700, available online at http://www.idealibrary.com on IDEAL

RLU/minute

A

14,000 12,000 10,000

5 hours

RLU/minute

B

24 hours

FIG. 3. In vivo optical CCD imaging in CD-1 mice following injection of fl DNA–DOTAP:cholesterol complexes. All images shown are the visible light image superimposed on the optical CCD image with a scale in RLU/minute as shown. CMV-fl DNA–DOTAP:cholesterol complexes were prepared using amounts ranging from 50 ␮g to 150 ␮g DNA. The complexes were injected into CD-1 mice. Each animal received an injection of D-Luciferin 10 minutes before imaging. One animal in each group served as a control and was injected with DOTAP:cholesterol only. (A) Mice were imaged 5 hours after injection of the complexes. Mice injected with increasing amounts of DNA show increasing levels of fl gene expression predominantly in the lungs and some expression in the spleen (at 50 ␮g). The RLU/minute ranged from 400 to 7500. The control mouse shows only background levels of fl gene expression (RLU/minute < 50). (B) Mice imaged 24 hours after injection of the complexes show decreased levels of fl gene expression in the lung with RLU/minute ranging from 100 to 4800.

In Vivo CCD Relative Light Units Demonstrate a Good Correlation to ex Vivo FL Enzyme Assay To determine whether the bioluminescent signal observed in the lung in living mice does, in fact, correlate with the FL activity in the lungs as measured ex vivo, two groups of CD-1 mice were injected with the DNA–DOTAP:cholesterol complexes with varying amounts of DNA and imaged 5 hours and 24 hours later. The fl expression in the heart, kidney, liver, and spleen was minimal. Because the maximal level of expression was always detected in the right lung, we correlated the light intensity as measured by the RLU/minute in the right lung to the FL activity at 24 hours (Fig. 4). We observed a good correlation both at 5 hours (data not shown) and at 24 hours (r2 = 0.70 and 0.90, respectively). These results indicate that the bioluminescent signal (RLU/minute) from the CCD camera images can be used as a measure for ex vivo FL activity. MicroPET Imaging of Cationic Lipid-Mediated Delivery and Expression of tk Demonstrates Timeand Dose-Dependent Kinetics MicroPET imaging of CD-1 mice injected with tk DNA and

558

DOTAP:cholesterol complexes shows low accumulation of activity in the lungs at 5 hours (Fig. 5A), but the expression increases with increase in time up to 24 hours (Fig. 5B). MicroPET images were analyzed for 9-(4-[18F]-Fluoro3-hydroxymethylbutyl) guanine percent-injected dose per gram lung by analyzing regions of interests (ROIs) [10]. The [18F]FHBG % ID/g in the lung at 24 hours is observed to increase with increase in DNA amounts (0.2, 0.4, and 1.3 for 50 ␮g, 75 ␮g, and 100 ␮g of DNA, respectively). In the control mouse, tk gene expression is minimal ([18F]FHBG % ID/g lung = 0.01 at 24 hours). Low levels of tk gene expression are detected in the kidney at higher DNA amounts. The signal from the kidney most likely indicates the elimination of [18F]FHBG from this route. As in the case of fl expression, the tk expression in the lung is observed to be higher than in all other organs. The level of tk gene expression, which was observed to be low in the lungs at 5 hours, gradually increased with time up to 24 hours. Due to the low levels of [18F]FHBG retention in the lungs at early time points (5 hours), no significant difference between the lungs and surrounding tissue or between the left and right lungs is noted. Analysis of the % ID/g at

MOLECULAR THERAPY Vol. 6, No. 4, October 2002 Copyright © The American Society of Gene Therapy

METHOD

doi:10.1006/mthe.2002.0700, available online at http://www.idealibrary.com on IDEAL

300,000

RLU/minute/␮g protein

FIG. 4. Correlation of RLU/minute in the lung (from CCD camera) with the ex vivo FL activity in the lung. DOTAP:cholesterol complexes containing increasing amounts of fl DNA were prepared and injected into CD-1 mice. The mice were imaged with D-Luciferin using a CCD camera 24 hours after injection. Following the scan, the mice were sacrificed and the lungs were assayed for FL activity. RLU/minute and ex vivo FL activity are plotted on the x and y axes, respectively (r2 = 0.90).

200,000

100,000

24 hours showed no significant difference in [18F]FHBG accumulation between the left and right lungs (P = 0.71). Toxicity was still an issue at amounts greater than 100 ␮g, as a few of the mice in each study group injected with 125 ␮g and 150 ␮g tk DNA died between 5 hours and 24 hours after injection. In an attempt to correlate the in vivo imaging signal with the ex vivo thymidine kinase enzyme (TK) activity, we observed that the TK activity levels for the mice carrying DNA–lipid complexes were low. Note that tk refers to the gene and TK to the enzyme. With increase in DNA concentration, we did not observe significant increase in TK activities. This is reflected in the small increase in the % ID/g in the lung at various amounts. Expression of Optical (fl) and PET (tk) Reporter Genes Can Be Imaged in the Same Animal Because many applications may benefit from the ability to repeatedly monitor the expression of two reporter genes in the same animal, we further explored the feasibility of imaging both an optical and a PET reporter gene in the same animal using two different imaging modalities. Plasmid DNA carrying fl and tk reporter genes, both complexed with DOTAP:cholesterol, was injected by way of the tail vein in a CD-1 mouse. Background optical images obtained before injection of D-Luciferin showed minimal levels of fl expression (RLU/minute < 50; Figs. 6A and 6B). The images obtained after injection of D-Luciferin show a high level of fl expression in the lungs (RLU/minute = 2500). The mouse injected with 75 ␮g DNA showed higher fl expression when compared with the mouse carrying 50 ␮g of DNA, supporting the earlier observation of the increase in gene expression with increasing DNA amount (RLU/minute = 3668). The same mice were imaged on a microPET scanner 24 hours later following injection of the PET reporter probe, [18F]FHBG. HSV1-sr39tk gene expression is detected in the lungs in both mice, with the [18F]FHBG % ID/g lung being 0.50% and 1% for the mice carrying 50 ␮g and 75 ␮g of DNA, respectively. The signal in the kidneys of both the mice indicates clearance of the tracer from this route.

DISCUSSION In clinical gene therapy applications, one of the many challenges is the ability to control and effectively deliver genes to target cells. Therefore, the development of noninvasive imaging technologies that will allow monitoring of the location, magnitude, and time-variation of gene

MOLECULAR THERAPY Vol. 6, No. 4, October 2002 Copyright © The American Society of Gene Therapy

10,000

Maximum RLU/minute (R. lung)

expression is highly desirable and will lead to quantitative monitoring of gene expression in vivo. In this study, we report that cationic lipid-mediated delivery and expression of fl and tk reporter genes can be imaged in living mice using a cooled CCD camera and microPET, respectively. CD-1 mice injected with fl DNA–DOTAP:cholesterol complexes showed a high level of fl gene expression in the lung after a single, tail-vein injection of DNA. Previous studies using lipid–DNA complexes have showed gene expression predominantly in the lung, with low levels of gene expression in the heart, liver, spleen, and kidney as determined by FL activity [7,32]. This was attributed to the pulmonary vasculature being the first capillary bed encountered by the lipid complexes following tail-vein injection. Gene expression in the lung was detected as early as 5 hours and persisted for up to 24 hours. It is interesting to note the initial higher level of fl gene expression in the right lung across several mice in different experimental groups. We reasoned that this may be due to the attenuation of the left lung signal by the heart. However, ex vivo analysis of the left lung tissue in a luminometer also showed low levels of FL activity, consistent with the bioluminescent signal from the CCD camera. Therefore, the observed lower signal in the left lung is not due to attenuation by the heart. This leads us to believe that the preferential accumulation of fl gene expression in the right lung at early times may be related to physiological events within the animal, such as blood flow. We do not have a simple way to quantitate potential differences in blood flow between the two lungs. A microsphere study to assess blood flow may prove to be helpful but is beyond the scope of this study. Future studies will need to address this issue in greater detail. Optical imaging using the cooled CCD camera provides a convenient, fast, and reproducible method to monitor fl gene expression. Previously, cationic lipid-mediated expression of fl gene required ex vivo analysis of FL activity in the different organs to study the pattern of fl gene expression. With the availability of the CCD camera, it is now possible to noninvasively track the expression of fl gene in a dose- and time-dependent manner in vivo and to obtain real-time functional data about the location, magnitude, and duration of gene expression.

559

METHOD

doi:10.1006/mthe.2002.0700, available online at http://www.idealibrary.com on IDEAL

FIG. 5. In vivo tk gene expression as a function of DNA dose after systemic delivery of tk DNA–lipid complexes in CD-1 mice. HSV1-sr39tk DNA–DOTAP:cholesterol complexes were prepared using different amounts of DNA (50 ␮g, 75 ␮g, and 100 ␮g). HSV1-sr39tk gene expression was monitored using microPET 5 hours (A) and 24 hours (B)after injection of the complexes. The mice were injected with 200 ␮Ci [18F]FHBG 1 hour before imaging. The control mouse shows only background levels of tk expression at 5 hours and 24 hours (% ID/g lung = 0.01 at 24 hours). At 5 hours, the tk mice show low levels of gene expression in the lung. At 24 hours, the tk gene expression in the lung is observed to increase in all mice injected with DNA–lipid complexes (% ID/g lung = 0.18, 0.36, and 1.3 at 24 hours).

A 5 hours

B 24 hours

Lungs are the primary organs for transgene expression following systemic administration of DNA–lipid complexes in mice. However, it is also possible to target gene expression to a specific organ using cell-based receptors. Using this approach, Templeton et al. [8] have demonstrated that chloramphenicol acetyltransferase (CAT) gene expression can be targeted to the liver using asialoglycoprotein receptor ligands. This should allow for targeted gene transfer to other tissues using cell-specific receptors. We further extended the cationic lipid-based gene delivery approach to noninvasively image the transfer of tk reporter gene in CD-1 mice using [18F]FHBG and microPET. Previous studies in our laboratories comparing the different reporter probes for HSV1-sr39tk showed that [18F]FHBG is a significantly improved reporter probe to noninvasively monitor tk reporter A gene expression using microPET [33]. Therefore, [18F]FHBG was used to image DNA–lipid-mediated tk reporter gene expression in vivo. As with fl, tk gene expression was

FIG. 6. Dual imaging of optical and PET reporter genes in the same animal. CD-1 mice were first imaged 5 hours after injection using the CCD camera in the absence of D-Luciferin (control), and then imaged again with D-Luciferin using a CCD camera. The same mice were imaged 24 hours later with [18F]FHBG using microPET. (A) A CD-1 mouse carrying 50 ␮g each of fl and tk DNA–lipid complex. The control image shows background levels of fl expression in the lung area (RLU/minute < 50). At 5 hours, the fl gene expression is significantly greater than that of the control (RLU/minute = 2958). The same mouse imaged at 24 hours shows low levels of tk gene expression in the lung (% ID/g = 0.46). (B) A CD-1 mouse was injected with 75 ␮g each of fl and tk DNA–lipid complex. The control image shows background levels of fl expression (RLU/minute < 50). In contrast, the image acquired after injection of D-Luciferin shows a high level of fl expression in the lung (3668 RLU/minute). MicroPET imaging using [18F]FHBG shows increased levels of tk gene expression in the lung at 24 hours (% ID/g = 1.0).

560

B

mainly detected in the lung. To determine whether the [18F]FHBG signal observed in the kidneys indicates tracer clearance or tk gene expression, the different organs were harvested and assayed for TK enzyme activity following the microPET scan (data not shown). The TK activity in the kidneys was found to be minimal and comparable to control, thereby suggesting that the [18F]FHBG signal observed in the kidneys indicates elimination of the tracer by this route and does not indicate significant levels of HSV1sr39tk reporter gene expression. The level of tk gene expression was low 5 hours after injection but increased with increase in time (up to 24 hours). This is in contrast to fl gene expression in the lung, which was maximum at 5 hours and decreased by 24 hours. The observed differences between fl and tk gene expression can be attributed to differences in the half lives of mRNA and protein. Firefly luciferase is known to be highly susceptible to proteases in vivo and has a half-life of 3 hours [34]. In contrast, tk mRNA has a longer half life of 8.3 hours [35]. Future studies aimed at determining mRNA levels using quantitative

5 hours

RLU/minute

24 hours

5 hours

RLU/minute

24 hours

MOLECULAR THERAPY Vol. 6, No. 4, October 2002 Copyright © The American Society of Gene Therapy

METHOD

doi:10.1006/mthe.2002.0700, available online at http://www.idealibrary.com on IDEAL

RT-PCR will be useful in elucidating the role of mRNA half-lives (fl and tk) in the kinetics of reporter gene expression. Future studies in quantitating the reporter protein half lives of FL and TK should also be useful. We also found that mice injected with large doses of DNA died between 5 hours and 24 hours after injection. The toxicity observed at high DNA amounts can be attributed to the presence of contaminants in plasmid DNA preparations. These contaminants can not be removed by endotoxin removal. To safely deliver large doses of DNA, it is essential to remove these contaminants from DNA preparations. The use of clinical grade DNA (which is now available) devoid of contaminants will increase levels of safety and gene expression [36]. We have also demonstrated the ability to image cationic lipid-mediated delivery and expression of optical (fl) and PET (tk) reporter genes in the same animal using the CCD camera and microPET, respectively. Noninvasive imaging of dual reporter genes in the same living subject will prove to be useful to monitor the expression of a therapeutic gene based on the expression of the reporter gene when the reporter gene is coupled to a therapeutic gene [37–39]. Approaches to couple a reporter gene with a therapeutic gene include fusion gene technology, the use of double promoters, co-vector administration, and a bidirectional, transcriptional approach [40]. It is important to bear in mind the relative merits and demerits of the two imaging techniques. Optical imaging provides a low-cost, convenient alternative to microPET for real-time analysis of gene expression in living animals. However, the CCD camera imaging has certain limitations that include the two-dimensional nature of the images and lack of depth information, the inability to distinguish signals in close proximity due to the relatively poor spatial resolution (for example, heart versus lung signals), and the inability to generalize to human applications. The distinct advantage in using the microPET is that it provides three-dimensional information and animal studies can then be extended to human subjects using a clinical PET scanner. This, however, requires a greater expense and the ability to produce the radiolabeled reporter probe. Further, the overall sensitivities of the two approaches are dependent on several factors. In the optical approach, the bioluminescent signal can be affected by its location, the concentration of reporter protein, the dose and delivery of reporter probe, and the regional hemoglobin concentration (which absorbs emitted bioluminescence). In the PETbased approach, the [18F]FHBG signal is less dependent on tissue depth and is not altered by regional hemoglobin levels. Also, the background signal in the optical approach is low, as the reporter probe does not become active until it comes into contact with the reporter protein. In contrast, a high background signal is observed in PET imaging as the radioactive probe is detected while it is being cleared from various tissues. Another vital difference between the two approaches is the amount of reporter probes used. PET

MOLECULAR THERAPY Vol. 6, No. 4, October 2002 Copyright © The American Society of Gene Therapy

imaging involves the use of trace levels (ng) of [18F]FHBG. In contrast, for optical imaging, mass levels (mg) of D-Luciferin are injected into the animal, leading to greater production of photons. In the current study, fl and tk reporter gene expression was detected at DNA amounts greater than 50 ␮g, suggesting that the two approaches exhibit comparable sensitivity. Further studies involving these issues will be needed to compare the absolute sensitivities of the two approaches. We have demonstrated that a high level of fl and tk gene expression achieved after a single intravenous injection of DNA–lipid complexes can be imaged using a cooled CCD camera and microPET, respectively. The efficiency of gene transfer can be increased further by developing improved lipid formulations with increased stability. As demonstrated in this study, the CCD camera and microPET can be immensely useful, individually and together (by linking the reporter gene to a therapeutic gene), to monitor cationic lipid-based gene transfer in a noninvasive and repetitive manner in small animal models.

MATERIALS

AND

METHODS

Chemicals. 1,2-Dioleoyl-3-trimethylammonium-propane (DOTAP) and cholesterol (20 mM, 1:1 mixture) was purchased from the Center for Cell and Gene Therapy (Baylor College of Medicine, Houston, TX). We purchased 5% dextrose (D5W) from Abbott Laboratories (North Chicago, IL). Luciferase assay kit was obtained from Promega (Madison, WI). 9-(4-18F-fluoro-3-hydroxymethylbutyl) guanine ([18F]FHBG; specific activity, 5–10 Ci/␮mol) was synthesized by a modification of the procedure reported by Yaghoubi et al. [41]. Radiochemical purity of [18F]FHBG exceeded 99%, as determined by high-pressure liquid chromatography (HPLC). D-Luciferin was obtained from Xenogen Corporation (Alameda, CA). Purification of plasmid DNA. CMV-fl plasmid DNA encoding the firefly (Photinus pyralis) luciferase gene was a gift from Christopher Contag’s laboratory (Stanford University, Stanford, CA). The fl and tk plasmid DNA were amplified in DH5␣ strain of Escherichia coli (GIBCO-BRL, Gaithersburg, MD) extracted by alkaline lysis and purified according to the Qiagen Endofree plasmid purification protocol (Qiagen, CA). The purity and identity of the DNA samples was confirmed by absorbance measurements at 260 and 280 nm and by agarose gel electrophoresis. Endotoxin levels were reduced to less than 0.1 endotoxin units/mg plasmid DNA. Preparation of DNA–lipid complexes. The DNA–lipid complexes were prepared 24 hours before injection in mice. CMV-fl DNA was diluted in D5W and a stock solution of DOTAP:cholesterol (20 mM) was diluted in D5W. Equal volumes of the DNA and DOTAP:cholesterol solutions were used. The DNA solution was gradually added at the surface of the lipid solution. The solution was mixed rapidly and stored at 4⬚C. The amount of DNA used ranged from 10 to 150 ␮g. The tk DNA–lipid complexes were prepared in a similar manner. In vivo gene delivery. Each 6-week-old, female CD-1 mouse was injected in the tail vein with the fl DNA–DOTAP:cholesterol and/or tk DNADOTAP:cholesterol complexes using a 27-gauge syringe needle. The complexes were allowed to warm to room temperature before injection. Optical imaging of fl expression in vivo. Animal care and euthanasia were performed with the approval of the University of California Animal Research Committee. Female CD-1 mice were anesthetized with ketamine–xylazine (4:1). The mice received an intraperitoneal injection of DLuciferin (150 mg/g body weight; Xenogen, CA) 5 minutes later. The animals were placed in a light-tight chamber and a gray-scale reference image was obtained under low-level illumination. Photons emitted from within the mouse and transmitted through the tissues were collected using a

561

METHOD

doi:10.1006/mthe.2002.0700, available online at http://www.idealibrary.com on IDEAL

cooled charge coupled device (CCD) camera (Xenogen IVIS, Xenogen Corporation, Alameda, CA) and integrated for a period of 2 minutes. Images were obtained using Living Image software (Xenogen Corporation, Alameda, CA) and Igor Image Analysis Software (Wavemetrics, Seattle, WA). For quantitation of transmitted light, ROIs were drawn over the lung region and maximum RLU/minute measurements were obtained. Each experimental group included five animals, and a total of three groups were studied for optical imaging. MicroPET imaging of tk reporter gene expression in vivo. Female CD-1 mice were injected with tk DNA–DOTAP:cholesterol complexes. The mice were imaged 5 hours and 24 hours following injection of the complexes. The mice were anesthetized with ketamine–xylazine (4:1) and injected with 200 ␮Ci of [18F]FHBG 1 hour before imaging. The animals were placed in a supine position and scanned with microPET with an acquisition time of 56 minutes. For tk, each study group consisted of four mice, and two groups of mice were used for microPET imaging. MicroPET image analysis. The microPET images were reconstructed by using filter-back projection and an iterative three-dimensional reconstruction technique [42] with an isotropic image resolution of 1.8 mm and a volumetric resolution of ~ 8 mm3. ROIs were drawn over the lung area. The ROI counts were converted to % ID/g as described [31]. Optical and microPET imaging of dual reporter genes in the same animal. The complexes carrying both fl and tk DNA were prepared 24 hours before injection in animals. CMV-fl and CMV-tk DNA (50 ␮g each) were diluted in D5W. DOTAP:cholesterol (20 mM) was diluted in D5W. The DNA solution was added dropwise at the surface of the lipid solution and the solution was mixed by rapidly pipetting a few times. The complex was stored at 4⬚C overnight. The procedure was repeated using 75 ␮g of CMVfl and CMV-tk DNA. On the day of injection, the complexes were allowed to warm to room temperature and subsequently injected in female CD-1 mice using a 27-gauge syringe needle. The mice were imaged 5 hours after injection for fl using the CCD camera after receiving an intraperitoneal injection of the substrate D-Luciferin 10 minutes before scanning. The same mice were imaged again with microPET for tk 24 hours later after receiving an injection of 200 ␮Ci [18F] FHBG 1 hour before the scan. This experimental group included two mice.

ACKNOWLEDGMENTS We thank David Stout, Khoi Nguyen, and Xiaoman Lewis (Crump Institue for Molecular Imaging) for technical assistance. This work was supported in part by funding from Department of Energy contract DE-FC03-87ER60615 (S.S.G.), NIH RO1 CA82214-01 (S.S.G.), and SAIRP R24 CA92865 (S.S.G.). RECEIVED FOR PUBLICATION MARCH 25; ACCEPTED JULY 25, 2002.

REFERENCES 1. Berkner, K. L. (1988). Development of adenovirus vectors for the expression of heterologous genes. BioTechniques 6: 616–629. 2. Nabel, G. J., et al. (1993). Direct gene transfer with DNA-liposome complexes in melanoma: expression, biologic activity, and lack of toxicity in humans. Proc. Natl. Acad. Sci. USA 90: 11307–11311. 3. Caplen, N. J., et al. (1995). Liposome-mediated CFTR gene transfer to the nasal epithelium of patients with cystic fibrosis. Nat. Med. 1: 39–46. 4. Felgner, J. H., et al. (1994). Enhanced gene delivery and mechanism studies with a novel series of cationic lipid formulations. J. Biol. Chem. 269: 2550–2561. 5. Li, S., and Huang, L. (1997). In vivo gene transfer via intravenous administration of cationic lipid-protamine-DNA (LPD) complexes. Gene Ther. 4: 891–900. 6. Li, S., et al. (1999). Dynamic changes in the characteristics of cationic lipidic vectors after exposure to mouse serum: Implications for intravenous lipofection. Gene Ther. 6: 585–594. 7. Liu, Y., et al. (1997). Factors influencing the efficacy of cationic liposome-mediated intravenous gene delivery. Nat. Biotechnol. 167–173. 8. Templeton, N. S., et al. (1997). Improved DNA:liposome complexes for increased systemic delivery and gene expression. Nat. Biotechnol. 15: 647–652. 9. Zabner, J., Fasbender, A. J., Moninger, T., Poellinger, K. A., and Welsh, M. J. (1995). Cellular and molecular barriers to gene transfer by a cationic lipid. J. Biol. Chem. 270: 18997–19007. 10. Gambhir, S. S., et al. (1998). Imaging of adenoviral directed herpes simplex virus type 1

562

thymidine kinase gene expression in mice with ganciclovir. J. Nucl. Med. 39: 2003–2011. 11. Gambhir, S. S., et al. (1999). Imaging adenoviral-directed reporter gene expression in living animals with positron emission tomography. Proc. Natl. Acad. Sci. USA 96: 2333–2338. 12. Iyer, M., et al. (2001). 8-[18F]-Fluoropenciclovir: an improved reporter probe for imaging HSV1-tk reporter gene expression in vivo using positron emission tomography. J. Nucl. Med. 42: 96–105. 13. Tjuvajev, J. G., et al. (1998). Noninvasive imaging of HSV1-tk marker gene with FIAU for monitoring transfer and expression of other therapeutic genes by multi-gene delivery vectors. J. Nucl. Med. 39: 130P. 14. Contag, P. R., Olomu, I. N., Stevenson, D. K., and Contag, C. H. (1998). Bioluminescent indicators in living mammals. Nat. Med. 4: 245–247. 15. Gambhir, S. S., Barrio, J. R., Herschman, H. R., and Phelps, M. E. (1999). Assays for noninvasive imaging of reporter gene expression. Nucl. Med. Biol. 26: 481–490. 16. Tjuvajev, J. G., et al. (1997). Quantitative PET imaging of HSV1-TK gene expression with [I-124]FIAU. J. Nucl. Med. 38: 239P. 17. Cherry, S. R., and Gambhir, S. S. (2001). Use of positron emission tomography in animal research. ILAR J. 42: 219–232. 18. Bloomfield, P. M., et al. (1997). Three-dimensional performance of a small-diameter positron emission tomograph. Phys. Med. Biol. 42: 389–400. 19. Chatziioannou, A. F., et al. (1999). A high-resolution lutetium oxyorthosilicate PET scanner for animal imaging. J. Nucl. Med. 40: 1164. 20. Cherry, S. R., et al. (1997). MicroPET: A high resolution PET scanner for imaging small animals. IEEE Transactions Nuclear Science 44: 1161–1166. 21. Contag, C. H., Jenkins, D., Contag, P. R., and Negrin, R. S. (2000). Use of reporter genes for optical measurements of neoplastic disease in vivo. Neoplasia 2: 41–52. 22. Wu, J. C., Sundaresan, G., Iyer, M., and Gambhir, S. S. (2001). Noninvasive optical imaging of firefly luciferase reporter gene expression in skeletal muscles of living mice. Mol. Ther. 4: 297–306, doi:10.1006/mthe.2001.0460. 23. Yang, M., et al. (2001). Whole body and intravital optical imaging of angigenesis in orthotopically implanted tumors. Proc. Natl. Acad. Sci. USA 98: 2616–2621. 24. Louie, A. H., et al. (2000). In vivo visualization of gene expression using magnetic resonance imaging. Nat. Biotechnol. 18: 321–325. 25. Weissleder, R., et al. (2000). In vivo magnetic resonance imaging of transgene expression. Nat. Med. 6: 351–355. 26. Gould, S. J., and Subramani, S. (1988). Firefly luciferase as a tool in molecular and cell biology. Anal. Biochem. 175: 5–13. 27. Sweeney, T. J., et al. (1999). Visualizing the kinetics of tumor-cell clearance in living animals. Proc. Natl. Acad. Sci. USA 96: 12044–12049. 28. Lipshutz, G. S., et al. (2001). In utero delivery of adeno-associated viral vectors: intraperitoneal gene transfer produces long-term expression. Mol. Ther. 3: 282–292, doi:10.1006/mthe.2001.0267. 29. Yotnda, P., et al. (2002). Bilamellar cationic liposomes protect adenoviruses from pre-existing humoral immune responses. Mol. Ther. 5: 233–241, doi:10.1006/mthe.2002.0545. 30. Ramesh, R., et al. (2001). Successful treatment of primary and disseminated human lung cancers by systemic delivery of tumor suppressor genes using an improved liposome vector. Mol. Ther. 3: 337–350, doi:10.1006/mthe.2001.0266. 31. Gambhir, S. S., et al. (2000). 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. USA 97: 2785–2790. 32. Mahato, R. I., et al. (1998). Biodistribution and gene expression of lipid/plasmid complexes after systemic administration. Hum. Gene Ther. 9: 2083–2099. 33. Iyer, M., et al. (2000). Comparison of FPCV, FHBG and FIAU as reporter probes for imaging herpes simplex virus type I thymidine kinase reporter gene expression. J. Nucl. Med. 41: 80P. 34. Thompson, J. F., Hayes, L. S., and Lloyd, D. B. (1991). Modulation of firefly luciferase stability and impact on studies of gene regulation. Gene 103: 171–177. 35. Henley, D. C., and Wir, J. P. (1991). Stability of selected herpes simplex virus type I mRNAs. Virus Res. 20: 121–132. 36. Templeton, N. S. (2001). Developments in liposomal gene delivery systems. Emerging Biotherapeutic Technologies 1: 1–4. 37. Yu, Y., et al. (2000). Quantification of target gene expression by imaging reporter gene expression in living animals. Nat. Med. 6: 933–937. 38. Sun, X., et al. (2001). Quantitative imaging of gene induction in living animals. Gene Ther. 8: 1572–1579. 39. Yaghoubi, S., et al. (2001). Direct correlation between positron emission tomographic images of two reporter genes delivered by two distinct adenoviral vectors. Gene Ther. 8: 1072–1080. 40. Ray, P., et al. (2001). Monitoring gene therapy with reporter gene imaging. Sem. Nuc. Med. 31: 312–320. 41. Yaghoubi, S., et al. (2001). Human pharmacokinetic and dosimetry studies of [18F]FHBG: A reporter probe for imaging herpes simplex virus type I thymidine kinase (HSV1tk) reporter gene expression. J. Nucl. Med. 42: 1225–1234. 42. Qi, J., Leahy, R. M., Cherry, S. R., Chatziioannou, A., and Farquhar, T. H. (1998). Highresolution 3D Bayesian image reconstruction using the microPET small-animal scanner. Phys. Med. Biol. 43: 1001–1013.

MOLECULAR THERAPY Vol. 6, No. 4, October 2002 Copyright © The American Society of Gene Therapy