Systemic delivery of a breast cancer-detecting adenovirus using targeted microbubbles

Abstract

One of the major limitations of cancer gene therapy using recombinant human adenovirus (Ad) is rapid Ad inactivation from systemic delivery. To eliminate this, biotin-coated ultrasound contrast agents, or microbubbles (MBs), were streptavidin-coupled with biotinylated antibodies to three distinct tumor vasculature-associated receptors (αVβ3 integrin, P-selectin and vascular endothelial growth factor receptor-2) for systemic targeting of a previously generated vector Ad5/3-Id1-SEAP-Id1-mCherry. This cancer-specific, dual-reporter vector was loaded in the targeted MBs and confirmed by confocal microscopy. MB loading capacity was estimated by functional assays as 4.72±0.2 plaque forming unit (PFU) per MB. Non-loaded (free) Ad particles were effectively inactivated by treatment with human complement. The Ad-loaded, targeted-MBs were injected systemically in mice bearing MDA-MB-231 tumors (Grp 1) and compared with two control groups: Ad-loaded, non-targeted MBs (Grp 2) and free Ad (Grp 3) administered under the same conditions. Two days after administration the blood levels of secreted embryonic alkaline phosphatase (SEAP) reporter in Grp 1 mice (16.1 ng ml–1±2.5) were significantly higher (P<0.05) than those in Grp 2 (9.75 ng ml–1±1.5) or Grp 3 (4.26 ng ml–1±2.5) animals. The targeted Ad delivery was also confirmed by fluorescence imaging. Thus, Ad delivery by targeted MBs holds potential as a safe and effective system for systemic Ad delivery for the purpose of cancer screening.

Introduction

Ultrasound (US) contrast agents, or microbubbles (MBs), are perfluorocarbon gas-filled lipid micelles with a size on the order of a micrometer. Their use during contrast-enhanced US has advanced a number of vascular imaging technologies, such as: vessel density characterization,1 determination of microvascular flow rates2 and evaluation of antiangiogenic therapy.3, 4 In addition, the intrinsic stability and non-immunogenicity of MBs has led to their use in MB-based strategies for tumor delivery of therapeutic compounds,5 plasmids6 and viral vectors.7 When exposed to an US pressure field, MBs physically resonate. Upon being systemically circulated and under certain US exposure conditions, it has been shown that vibrating MBs can transiently improve vascular permeability and chemotherapeutic drug uptake.8 Alterations in the malleable lipid exterior cause physical association of vectors with the MB surface resulting in augmentation of transgene expression during plasmid-based therapy.9 Although targeting of these therapies has been limited to site-directed US, receptor targeting of MBs has been shown to further improve the efficacy of this drug delivery system.10

Targeting MBs to receptors commonly overexpressed by a tissue of interest has been shown to improve contrast-enhanced US and overall MB accumulation at target sites.11 The active targeting of MBs is achieved by conjugating receptor-specific ligands to the outer shell through either biotin–avidin chemistry or covalent linkage.12 Ligand-modified MBs bind specifically to molecular targets providing enhanced visualization during contrast-enhanced US through localized MB accumulation, while unbound MBs are cleared from the circulation.13 Improved MB aggregation using targeted strategies has been demonstrated in the molecular imaging of inflammation,14, 15, 16 intravascular thrombi17, 18 and tumor angiogenesis.19, 20, 21 MB targets have evolved from single to multiple receptor moieties with well-characterized synergistic benefits of multi-target strategies.22, 23, 24 In the current study, MBs were targeted to αVβ3 integrins, P-selectin and vascular endothelial growth factor receptor-2 (VEGFR2) in order to improve MB accumulation in tumor vasculature and payload delivery. Targeting of these receptors in combination using a triple-targeted strategy has been shown to improve MB accumulation in tumor vasculature over single- and dual-targeted moieties.25

Various MB-based approaches have been explored to improve cancer therapy. One strategy of US-mediated drug delivery involves a combination of MBs, US and a therapeutic compound without physical association of the latter with MBs.26 In this instance, an augmented delivery is achieved through US-mediated MB destruction. Other mechanisms of US-assisted payload delivery involve physical association between MBs and a therapeutic compound.6, 27 One of those involves labeling compounds, such as hydrophilic pDNA attached to the exterior of protein-shelled MBs by non-covalent interactions.28 Other strategies have capitalized on the unique structure of the MB lipid shell to enable its association with lipophilic compounds, such as paclitaxel, through physical incorporation into the MB core.29, 30 Additional approaches involve double-emulsified MBs that physically encapsulate hydrophilic macromolecules such as pDNA31 or doxorubicin.32 In the latter studies, complete encapsulation of the payload compound was proven beneficial for systemic delivery, considering that the compound was shielded from immune response and nonspecific sequestering mechanisms, while the MB shell remained innocuous. Recently, a novel method of MB packaging was demonstrated where adenovirus (Ad) particles were encapsulated by reconstituting lyophilized MBs with a concentrated Ad solution in the presence of perfluorocarbon gas.33 When the freeze-dried MBs were reconstituted, the lipid shell encapsulated the concentrated Ad particles present in the solution, effectively loading intact particles. Human complement was then used to inactivate non-loaded Ad particles. This approach of MB packaging was further evaluated using a therapeutic Ad in a pre-clinical study.34

The success of breast cancer (BC) therapy shows a direct correlation with the stage and grade of the tumor at the time of diagnosis.35 We recently constructed a diagnostic Ad agent, Ad5/3-Id1-SEAP-Id1-mCherry, with BC-specific induction of dual-reporter expression: a secreted embryonic alkaline phosphatase (SEAP) enzyme for blood-based cancer screening and a fluorescent reporter mCherry for localized tumor imaging.36 The tumor-specific promoter Id1, driving expression of each reporter in the above diagnostic Ad, possesses a cancer-selective (tumor-on, liver-off) activation profile with its transcriptional activity directly correlating with the aggressiveness of cancer phenotype.37 To achieve augmented infectivity and detection sensitivity for BC, the Ad was genetically modified to encode a human Ad serotype chimera (Ad5/3) fiber, altering viral tropism to Ad3 receptor(s)38 and thereby circumventing deficiency of the native coxsackie-and-Ad receptor, commonly downregulated in cancer cells.39 This diagnostic Ad agent has earlier demonstrated sensitivity for BC detection.36

Clinical applicability of cancer therapy involving MB packaging and tumor targeting has been limited by the site-directed nature of the US-triggered MB payload release mechanism, which has utility only when the exact location of malignancy is known. The mechanism of MB payload release in cancer screening applications must, therefore, be independent of site-directed US induction. The systemic half-life of MBs (on the order of minutes) in circulation is limited by their nonspecific disruption from pulmonary capillary passage, turbulent flow pressure and hydrostatic forces.6 When a MB ruptures under hydrostatic pressure, the perfluorocarbon gas diffuses out, while the lipid cortex begins to bleb. It is hypothesized that during this blebbing process, the payload containment fails and loaded compounds are released.27 Therefore, when the loaded MB is bound to the target site its molecular content can be naturally released over time without need for the triggering step. This payload delivery method combined with the shielding factor of MB packaging make this approach a viable alternative to direct systemic administration of free vectors.

In this study, we explored a novel strategy of MB-mediated payload delivery to the tumor vasculature for the BC screening application. The benefits of payload encapsulation by MBs include the shielding from the natural Ad-sequestration/neutralization mechanisms that present a major challenge for the effective systemic Ad delivery. In addition, the payload encapsulation protects the patient from immunogenic complications associated with traditional therapy. These benefits may allow lowering the therapeutic dose and extending pharmacological half-life of targeted therapeutics leading to improved treatment safety.

In this report, the dual-reporter Ad was loaded in MBs, using the lyophilized reconstitution method, and targeted to tumor vasculature via three distinct receptors αVβ3, P-selectin and VEGFR2 simultaneously. After systemic administration and subsequent binding of the targeted MBs to tumor vasculature, the payload was spontaneously released as a result of MB dissipation leading to Ad infection of the neighboring cancer cells. Following the Ad cell infection the cancer-specific activation of both Id1 promoters induced expression of a fluorescent reporter mCherry and tumor-secreted enzyme SEAP, which enabled a localized tumor imaging and blood-based BC detection, respectively.

Materials and methods

Culture methods and in vitro experiments

MDA-MB-231 human BC and mouse angiosarcoma SVR (SVEN 1 RAS) endothelial cell lines were obtained from the American Tissue Type Collection (Manassas, VA). The MDA-MB-231 cell line was maintained in Dulbecco's modified Eagle's medium, 10% fetal bovine serum and 1% L-glutamine. The mouse SVR cell line was maintained in Dulbecco's modified Eagle's medium, 5% fetal bovine serum and 1% L-glutamine. All cells were cultured to 70–90% confluence before passaging. Cell lines were grown at 37 °C and 5% CO2. Cell numbers were determined with hemocytometer and Trypan blue dye exclusion.

MB loading and targeting

Ad loading was performed using custom lyophilized, biotin-coated MBs (Targestar-B, Targeson, San Diego, CA). Vials (3 × 108 MB/vial, average MB size distribution 1.52 μm each) were reconstituted with a 0.7 ml (7 × 1010 plaque forming unit (PFU)) solution of replication-incompetent Ad5/3-Id1-SEAP-Id1-mCherry,36 Ad5/3-CMV-luc40 or Ad5/3wt-pIXmRFP1 with the capsid protein IX (pIX)-fused mRFP1.41 Ad-loaded MBs were then treated with 10 × volume human complement sera (60 mg ml–1, S1764, Sigma-Aldrich, St Louis, MO) followed by 30-min incubation at room temperature. Ad-loaded MBs were then washed by 3 × centrifugation (400 g for 3 min) with 10 × volume phosphate-buffered saline (PBS) to remove unloaded particles. Infranatant from the final wash was used for background (control) measurements. Following washes, MBs were conjugated in a triple-targeted motif using biotinylated, rat immunoglobulin G antibodies against mouse αVβ3 (13-0512, ebioscience, San Diego, CA), mouse P-selectin (553743, BD Pharmingen, San Diego, CA) and mouse VEGFR2 (13-5821, ebioscience) according to the manufacturer's protocol. Briefly, biotin-coated MBs were incubated with streptavidin followed by PBS centrifuge wash. MBs were then incubated with biotin-conjugated antibodies (13.3 μg each) for 20 min followed by 3 × centrifuge washing to remove excess antibody. Following each conjugation, MB concentration was determined by hemocytometer.

In vitro experiments

All in vitro experiments were conducted in triplicates. Cells were plated 24 h before Ad infection. Experimental multiplicity of infection ratios were calculated based on PFU titers determined by a standard agarose-overlay plaque assay. Background values for the SEAP expression analysis were obtained from culture media collected in parallel from uninfected cells at the corresponding time points. A total of five in vitro experiments were performed: Figure 1a Ad5/3-CMV-Luc was aliquoted (4.4 × 108 PFU per tube in 0.1 ml) while a 1 : 2 serial dilution of human complement sera (60 mg ml–1) was prepared in 0.4 ml PBS with dilutions ranging from 15 to 0.218 mg and sequentially added to Ad tubes. Ad solutions containing complement were then incubated 30 min at room temperature or 37 °C. After incubation, Ad solutions containing serial dilution of human complement were added to adherent MDA-MB-231 BC cells (1 × 106 cells per well) and incubated for 30 min at 37 °C. Cells were then washed 3 × with PBS followed by addition of a complete medium. After 24 h, luciferin substrate (Caliper Life Sciences, Mountain View, CA) was added (15 μg per well) and plates were imaged for 60 s at the binning of 8 with an IVIS-100 CCD imaging system (Caliper Life Sciences). Matched region of interest analysis was performed using instrument software (Living Image 3.2, Caliper Life Sciences) to quantify mean luciferase (Luc) counts per well. In Figure 1b, three vials of lyophilized, biotin-coated MBs were reconstituted with 7 × 1010, 3.5 × 1010 or 1.75 × 1010 PFU of Ad5/3-Id1-SEAP-Id1-mCherry followed by complement inactivation. Three additional vials were prepared in parallel without complement inactivation. Following 3 × centrifuge wash of all vials, Ad-loaded MBs were aliquoted into tubes (1 × 106 MB per tube) followed by the addition of a monoclonal mouse antibody against the Ad5 hexon (LF-MA0177, Thermo-Scientific, Rockford, IL). A phycoerythrin (PE)-conjugated anti-mouse immunoglobulin G (1030-09, Southernbiotech, Birmingham, AL) was used as a secondary antibody. Ad-loaded MB groups were analyzed for fluorescent counts (10k event minimum) using an Accuri C6 flow cytometer (Accuri Cytometers, Ann Arbor, MI). Tubes with MBs and secondary antibody (immunoglobulin G) alone were used to establish background. In Figures 1c and d, lyophilized, biotin-coated MBs were reconstituted with Ad5/3wt-pIXmRFP1 followed by 3 × PBS washes to remove free particles. Ad-loaded MBs were then applied dropwise to microscope slides and a glass cover slip was added. Ad-loaded MBs were imaged for RFP1 fluorescence using a solid state red laser mounted on a Leica SP2 confocal laser-scanning microscope (Leica Microsystems, Bannockburn, IL) at 63 × oil immersion objective. PBS reconstituted MBs mixed with Ad5/3wt-pIXmRFP1 were used as control. In Figure 2, MDA-MB-231 cells (3.5 × 105 cells per well) were plated in 24 black-well plates and incubated overnight to allow adherence. Ad-loaded MBs were prepared with Ad5/3-CMV-Luc followed by complement inactivation. After the final wash, the infranatant was added to the wells in equal volumes as the Ad5/3-CMV-Luc loaded MB addition. Ad-loaded MBs (6.15 × 107 in 0.1 ml) were added to adherent MDA-MB-231 cells. Subsequently, 0.1 ml of the final wash infranatant was added to separate wells. Free Ad5/3-CMV-Luc was then added to separate wells in a 1 : 2 titration beginning at 1.3 × 109 PFU and ending at 1.3 × 106 PFU to serve as an Ad infectivity standard. Plates were then incubated overnight at 37 °C. At 24 h after infection, luciferin substrate was added (15 μg per well) and plates were imaged for 60 s at a binning of 8 with an IVIS-100 CCD imaging system. Matched region of interest analysis was performed to quantify mean Luc counts per well. In Figure 3, triple-targeted, Ad5/3-Id1-SEAP-Id1-mCherry loaded MBs were generated and complement inactivated. Ad-loaded targeted MBs (1.7 × 106 MBs per well) were then added in triplicate to wells in six-well plates containing 1 : 1 co-cultured mouse SVR and MDA-MB-231 cells (plated 24 h prior, 2.5 × 105 cells per well). Mouse SVR cells express the mouse αVβ3, P-selectin and VEGFR2 receptors required for targeted MB binding.25 MDA-MB-231 cells express the human Ad5/3 receptor and Id1 cancer promoter.36 Ad-loaded non-targeted MBs (1.7 × 106) and free Ad particles (8.5 × 106 PFU, multiplicity of infection=21) were prepared using the same protocol and served as controls. Treated wells were incubated while rocking for 30 min at room temperature. After incubation, plates were washed 3 × with PBS. After final wash, each well was viewed at 40 × with an Olympus IX70 microscope (Olympus America, Melville, NY) followed by complete media addition (Dulbecco's modified Eagle's medium, 7.5% fetal bovine serum and 1% L-glutamine) and placed in incubator for 24 h. Media from all wells was then removed and screened for SEAP. Wells were also imaged for mCherry fluorescence (see fluorescence imaging methods).

Figure 1
figure1

Complement inactivation and visualization of adenovirus (Ad)-loaded microbubbles (MBs). (a) MDA-MB-231 breast cancer cell luciferase (Luc) counts from titrated human complement treated Ad5/3-CMV-Luc Ad. Complement inactivation was performed at either 37 °C or room temperature. Data are means±s.d. (b) Phycoerythrin (PE) fluorescent counts from anti-Ad5 hexon antibody binding titrated Ad-loaded MBs with and without complement inactivation. Data are means±s.d. (c) Representative fluorescent confocal microscopy image (matching 0.5 μm z-plane slice) of Ad5/3wt-pIXmRFP1 loaded MBs and (d) unloaded MBs.

Figure 2
figure2

Quantification of functional plaque forming unit (PFU) per microbubble (MB). (a) Luciferase (Luc) counts from MDA-MB-231 breast cancer cells treated with Ad/3-CMV-Luc loaded MBs and wash infranatant. Background (infranatant) subtracted Luc counts (8.13 × 105) generated from 6.15 × 107 MBs were matched with Luc counts generated from (b) Ad5/3-CMV-Luc infectivity standard. Data are mean Luc counts±s.d.

Figure 3
figure3

In vitro analysis of adenovirus (Ad)-loaded targeted microbubbles (MBs). (a) Bright field images (40 × ) of MDA-MB-231 and mouse SVR endothelial cells co-cultured cells treated with Ad-loaded targeted MBs and (b) co-cultured cells treated with Ad-loaded MBs. (c) Secreted embryonic alkaline phosphatase (SEAP) amounts from co-cultured cells treated 24 h prior with either free Ad, Ad-loaded MBs or Ad-loaded targeted MBs. Asterisk denotes P<0.05. Data are mean SEAP (ng ml–1)±s.d. mCherry fluorescent images (20 × ) and inset bright field of co-cultured cells treated 24 h prior with either (d) free Ad, (e) Ad-loaded MBs or (f) Ad-loaded targeted MBs.

In vivo evaluation

Athymic female nude mice were obtained from Frederick Cancer Research (Frederick, MD) and implanted with 4 × 106 (0.1 ml) MDA-MB-231 cells in the mammary fat pad. Four weeks after implant, mice were sorted into three groups (n=5) based on tumor size (average tumor size: 271.3±72 mm3). Targeted, Ad5/3-Id1-SEAP-Id1-mCherry loaded MBs (complement-inactivated) were prepared and intravenously injected (tail vein) into five mice at 1 × 107 MBs per mouse in 0.1 ml dose (0.05 ml MBs+0.05 ml saline). Subsequently, non-targeted Ad5/3-Id1-SEAP-Id1-mCherry loaded MBs (complement-inactivated) were prepared and intravenously injected (1 × 107 MBs per mouse) into a separate mouse group. The final group received free Ad5/3-Id1-SEAP-Id1-mCherry (5 × 107 PFU per mouse in 0.05 ml) without complement inactivation. Simultaneously, this control group received non-targeted, unloaded MBs (1 × 107 MBs per mouse in 0.05 ml), which were reconstituted with PBS. A background control group was then prepared with five non-tumor-bearing mice that received targeted, Ad5/3-Id1-SEAP-Id1-mCherry loaded MBs (1 × 107 MBs per mouse in 0.1 ml dose). Blood was then collected on days 1, 2 and 4 post-injection from anesthetized mice by retro-orbital collection with heparin-treated capillary tubes. Collections were then spun and blood plasma was obtained for SEAP analysis. After day 4, mice were killed and tumors were excised and imaged for mCherry fluorescence. Institutional Animal Care and Use Committee at the University of Alabama at Birmingham approved all animal procedures.

SEAP analysis

Blood-based reporter analysis was conducted using a Great EscAPe Fluorescence Detection Kit (Clontech, Mountain View, CA, USA). All steps were performed according to the established protocol (New Fluorescent Great EscAPE SEAP Assay (January 1997) CLONTECHniques XII (1): 18–19).

Fluorescence imaging

In vitro fluorescent images (20 × ) were rendered using an mCherry 587 nm excitation/600 nm long pass filter on an inverted microscope with a halogen light source and Nuance multi-spectral camera. A liquid-crystal tunable wavelength filter in the camera was set for collection of emission images from 600 to 720 nm in 5 nm increments. Composite images (unmixed composites) were generated for each image cube by unmixing the spectral signature of the mCherry reporter proteins from those of background auto-fluorescence using a spectral library that was compiled from numerous spectral profiles collected from uninfected control cells under the same conditions. For tumor fluorescence images, mice were euthanized and tumors collected and imaged on day 4 post-injection with a Leica stereomicroscope (Model MZ-FLIII, Vashaw Scientific, Norcross, GA).

Statistical analysis

Statistical comparisons were performed using Minitab 15 statistical software (State College, PA). In Figure 1a, dose titration groups (15–0.437 mg) were compared with an analysis of variance test of mean Luc counts. An unpaired two-sample t-test was then used to compare mean Luc counts from the 0.437 mg dose versus the 0.218 mg dose group to determine the minimum amount of complement needed to inactivate Ad. In Figure 1b, an unpaired two-sample t-test was then used to compare PE counts from the 7 × 1010 PFU MB group and both the 3.5 × 1010 and 1.75 × 1010 PFU MB groups. In addition, an analysis of variance test of PE counts was performed on all complement-inactivated MB groups. In Figure 2a, mean Luc counts from the Ad-loaded MB treated group were compared with mean Luc counts from the infranatant-treated group using an unpaired two-sample t-test. For Figure 3c, an unpaired two-sample t-test was used to compare SEAP amounts from the Ad-loaded targeted MB group versus the Ad-loaded non-targeted MB group. The test was repeated to determine the significant difference between the SEAP amounts from the free Ad and the Ad-loaded targeted MB groups. For the in vivo analysis, an unpaired two-sample t-test was performed to determine significance of background-subtracted SEAP amounts between the Ad-loaded targeted MB groups and both the Ad-loaded non-targeted MB and free Ad groups. A P-value of <0.05 was considered statistically significant.

Results

Complement inactivation of Ad

Systemically administered Ad is known to undergo rapid sequestration from human circulation and functional inactivation by numerous factors including pre-existing anti-Ad antibody and the complement system. To determine the amount of human complement required for full inactivation of Ad particles, an in vitro complement inactivation assay was performed using a firefly Luc reporter-expressing E1-deleted vector (replication-incompetent) Ad5/3-CMV-Luc. The latter Ad (4.4 × 108 PFU) was pre-incubated with various amounts of human complement ranging from 15 to 0.218 mg either at room temperature or 37 °C before in vitro infection of MDA-MB-231 BC cells. At 24 h after infection, bioluminescent imaging of the above cell culture was used to quantify Luc expression that was proportional to the residual Ad infectivity, shown in Figure 1a. Although no significant change (P>0.05) in Ad inactivation efficiency was observed between samples receiving from 15 to 0.437 mg of complement, dilutions containing 0.218 mg of complement showed a significant decrease in Ad inactivation (P<0.05), suggesting that 1 mg of human complement is sufficient to fully inactivate 1 × 109 PFU of the Ad vector. The experiments also showed that the 0.218 mg dose was sufficient to reduce the Ad transduction efficiency 3-fold relative to the complement-untreated Ad (Figure 1a). Of note, no significant difference in Ad inactivation rates was observed between the room temp and the 37 °C incubation conditions (P>0.05).

Anti-Ad antibody binding of complement-inactivated Ad-loaded MBs

Susceptibility of the Ad-loaded MBs to neutralization with pre-existing anti-Ad antibodies is a critical biological issue and an important step in the biological evaluation of the MB-based Ad delivery systems. To assess the ability of an Ad-specific antibody to bind (and potentially neutralize), the Ad-loaded MBs, an antibody-binding assay was performed by probing the Ad-loaded MBs with a monoclonal antibody against the Ad5 capsid protein hexon. To this end MBs were reconstituted in the presence of 7 × 1010, 3.5 × 1010 or 1.75 × 1010 PFUs of the Ad5/3-Id1-SEAP-Id1-mCherry BC-detection vector, followed by in vitro Ad inactivation with human complement. The Ad-loaded MBs were then evaluated for binding to the Ad5 capsid (hexon)-specific primary antibody and a PE-conjugated secondary antibody by flow cytometry technique. In parallel, the same PFU-reconstitution MB groups were evaluated for the efficiency of PE/antibody labeling without prior complement inactivation of the MBs. The efficiency of Ad5-specific antibody binding to the complement-inactivated MB groups was on average 18-fold lower than to the complement-untreated MB groups (Figure 1b). In the absence of complement inactivation, the mean fluorescent counts (MFCs) for the 7 × 1010 PFU-reconstituted MB group (MFC=1.3 × 104±1.0 × 103) was significantly greater than for the 1.75 × 1010 PFU-reconstituted MB group (MFC=1.0 × 104±1.3 × 103) (P<0.02), indicating a dose-dependent MB binding of the primary (Ad capsid-specific) antibody. In contrast, after complement inactivation no significant difference in hexon antibody binding could be observed between the 7 × 1010 PFU-reconstituted MB group (MFC=5.8 × 102±7.2 × 101) and other groups including the 1.75 × 1010 PFU-reconstituted MB group (MFC=6.4 × 102±4.2 × 101) (P>0.30), reflecting the background interaction with the secondary (PE-conjugated) antibody.

The results of this experiment confirm complete Ad loading in the MB core. In addition, the results suggest that partial encapsulation of the vector takes place during MB lyophilized reconstitution with some Ad particles exposed to the surface by association with the lipid layer. Accessibility of the MB surface-associated Ad particles for interaction with human complement is evidenced by the dramatic (18-fold) loss in anti-Ad antibody binding following complement treatment. This result indirectly suggests that MB-shielding of the fully encapsulated Ad particles is likely to also prevent systemic immunogenicity. Thus, following complement inactivation of the MB surface-exposed (partially encapsulated) Ad particles the loaded MBs could become invisible to (non-reactive with) pre-existing anti-Ad5 antibodies commonly present in patient's serum.

Visualization of the Ad-loaded MB

To confirm feasibility of Ad encapsulation into the targeted MBs by the lyophilized reconstitution method, we used fluorescently labeled particles of a unique Ad5/3wt-pIXmRFP1 virus with mRFP1-tagged capsids41 to allow a direct visualization of MB Ad-loading by confocal fluorescent microscopy. The MB core localization of the mRFP1-labeled particles, evidenced by a representative composite (merged) image (Figure 1c), suggests capability of MBs to fully encapsulate Ad particles. In contrast, PBS-reconstituted non-loaded MBs mixed with Ad5/3wt-pIXmRFP1 were negative for core-localized fluorescence (Figure 1d).

Quantification of functional Ad loading per MB

In order to estimate functional Ad loading capacity of an average MB, an in vitro Luc assay was performed on MDA-MB-231 BC cells using complement-inactivated Ad5/3-CMV-Luc virus-loaded MBs. Luc counts obtained from the BC cells treated with wash infranatant were subtracted from readings obtained for the 6.15 × 107 loaded MBs (calculated from hemocytometer-determined MB concentration) to account for the background activity of non-encapsulated (free) Ad particles remaining after MB reconstitution, complement inactivation and subsequent washes. The background-corrected Luc counts obtained in the assay for a given number of Ad-loaded MBs (Figure 2a) were matched with the assay standard curve, plotted in parallel for serial dilutions of the same free Ad5/3-CMV-Luc virus (Figure 2b), to determine the corresponding infectious unit (PFU) equivalent of the free Ad standard (2.9 × 108 PFU) (Figure 2b). Dividing the above experimentally determined PFU (corresponding to the same Luc readings of the Ad standard) by 6.15 × 107 (number of Ad-loaded MBs used) allowed determining an average number of infectious viral particles per MB as 4.72±0.2 PFU per MB. Given that the Luc reading (MFC=1.02 × 106±8.0 × 104) produced by 6.15 × 107 Ad-loaded MBs was significantly greater than that (MFC=2.07 × 105±2.0 × 104) of the background (infranatant) control (P<0.01) and that equal volumes of infranatant from the last wash and the complement-inactivated Ad-loaded MBs were used in the experiment, this difference in Luc expression provided another (functional) line of evidence for the ability of MBs to fully package intact infectious Ad particles. The functional nature of the assay permitted evaluation of functional particles loaded within the MBs while excluding non-functional particles loaded within the MBs.

In vitro analysis of Ad-loaded targeted MBs

To evaluate the potential of the MB-based Ad delivery system for systemic detection of BC, we utilized a previously reported Ad5/3-Id1-SEAP-Id1-mCherry BC diagnostic vector in the context of MBs triple-targeted to mouse endothelial cells. The Ad-loaded targeted MBs were added to co-cultured mouse SVR and human BC cells. The mixed cell population enabled evaluation of MB targeting selectivity for the target receptor-expressing mouse endothelial cells. Treatment of co-cultured cells with the Ad-loaded targeted MB preparations resulted in localized clustering of the targeted MBs around mouse SVR cells expressing αVβ3, P-selectin and VEGFR2 receptor targets (Figure 3a). No MB clustering around cells was observed when Ad-loaded non-targeted MBs were used instead (Figure 3b). Binding of the Ad-loaded targeted MBs to the cell monolayer was also evidenced by the SEAP detection assay results: the levels of the SEAP reporter expression were significantly greater for the Ad-loaded targeted MB group (7.1 × 102 ng ml–1±31.0) than for the non-targeted MB group (4.6 × 102 ng ml–1±48.7) (P<0.01). On the contrary, there was no significant difference between the free (7.7 × 102 ng ml–1±63.1) and the Ad-loaded targeted MB group (7.1 × 102 ng ml–1±31.0) (P>0.05), demonstrating an enhanced in vitro performance of the complement-inactivated targeted MBs in protecting Ad particles and endothelial cell targeting. The pattern of SEAP reporter expression was validated by mCherry fluorescence expression in human MDA-MB-231 cells because of the human-specific Ad induction of this fluorescent reporter expression for the free Ad (Figure 3d), Ad-loaded MBs (Figure 3e) and the Ad-loaded targeted MB experimental groups (Figure 3f).

In vivo analysis of Ad-loaded targeted MBs

MDA-MB-231 tumor-bearing mouse groups (n=5) were prepared and IV injected with free Ad5/3-Id1-SEAP-Id1-mCherry, Ad-loaded MBs or Ad-loaded targeted MBs. Figure 4a compares SEAP levels obtained from the above groups on days 1, 2 and 4. The shown SEAP concentrations were calculated by subtracting the background SEAP concentrations measured in non-tumor-bearing mice (n=5) that received Ad-loaded targeted MB. The SEAP levels from Ad-loaded targeted MB group on day 2 (16.1 ng ml–1±2.5) were significantly greater than those from the Ad-loaded MB group (9.75 ng ml±1.5) or free Ad group (4.26 ng ml–1±2.5) (P<0.05) (Figure 4a). However, there was no significant difference between the Ad-loaded MB (non-targeted) and the free Ad groups (P>0.05). A representative overlay (composite) image of sliced BC tumor xenografts isolated from a mouse of the Ad-loaded targeted MB group on day 4 after injection, showed the strongest tumor-localized mCherry fluorescence (Figure 4d) relative to tumors from the other groups (Figures 4b and c). This result showed an advantage of MB targeting to tumor vasculature as well as presents proof-of-principle for the MB-based systemic delivery of Ad vectors with the purpose of cancer detection.

Figure 4
figure4

In vivo analysis of adenovirus (Ad)-loaded targeted microbubbles (MBs). (a) Background-subtracted (secreted embryonic alkaline phosphatase (SEAP) amounts from Ad-loaded targeted MB injected non-tumor-bearing controls) SEAP amounts from MDA-MB-231 tumor-bearing mice intravenously injected with free Ad, Ad-loaded MBs and Ad-loaded targeted MBs. Asterisk denotes P<0.05. Data are mean SEAP (ng ml–1)±s.d. Fused representative bright field and mCherry fluorescence images of excised tumors from mice receiving (b) free Ad, (c) Ad-loaded MBs and (d) Ad-loaded targeted MBs.

Discussion

Described here is a strategy for systemic delivery of Ad vectors for the specific purpose of BC detection. The diagnostic, dual-reporter Ad, Ad5/3-Id1-SEAP-Id1-mCherry, was loaded in MBs, which were targeted to endothelial receptors overexpressed on the tumor vasculature. Using a fluorescently tagged Ad as a diagnostic vector analog, Ad-loaded MBs were imaged using fluorescent confocal microscopy to confirm complete packaging, that is, enclosure of the fluorescent Ad within the MB core. In order to assess the full potential of the Ad-loaded MBs and remove residual unloaded Ad particles, human complement was used to bind to and inactivate unloaded particles. To evaluate the potential susceptibility of the MB delivery system to pre-existing anti-Ad antibody binding in circulation, Ad-loaded MBs were probed with a monoclonal antibody against the Ad5 capsid protein hexon. For this experiment, complement inactivation was shown to reduce antibody binding 18-fold over untreated groups, demonstrating the reduced immunoreactivity of the complement-inactivated MBs. In addition, the number of functional Ad particles loaded within each MB was determined using a Luc reporter expressing Ad and was found to be 4.7±0.2 PFU per MB. In vitro analysis of the Ad-loaded targeted MBs was performed using co-cultured mouse endothelial (SVR) and human BC (MDA-MB 231) cells. The Ad-loaded targeted MBs were shown to cluster around the endothelial cells, while non-targeted MBs failed to adhere to cells and were removed during wash steps. Complement-inactivated Ad-loaded targeted MBs were shown to express SEAP and mCherry reporters at the levels equivalent to those of the unloaded (free) Ad in co-cultured cells. In vivo experiments using MDA-MB-231 tumor-bearing mice demonstrated the benefit of targeted MBs in systemic delivery Ad to tumor with SEAP levels being significantly greater than both the Ad-loaded non-targeted MBs and the free Ad receiving experimental groups (P<0.05). Non tumor-bearing control mice injected with Ad-loaded targeted MBs provided the reporter background levels for each time point measured. A detectable expression of the reporters within 24–48 h after virus administration was consistent with our previous studies using this dual-reporter system.36 The enhanced tumor delivery afforded by the Ad-loaded targeted MBs was additionally evidenced by in vivo imaging of the fluorescent reporter expression in tumors.

Confocal fluorescence imaging of MBs reconstituted with mCherry-labeled Ad particles demonstrated localization of the mCherry fluorescence to the gas-filled, hydrophobic core of MBs. This result was counter-intuitive, considering that Ad particles are hydrophilic. It is hypothesized that during the lyophilized MB reconstitution, the high concentration of Ad in solution prompts uptake of droplets of Ad solution along with an ambient perfluorocarbon gas within the lipid micelle. The hydrophilic solution is repelled from the hydrophobic lipid tails and localized to the inner core forming hydrophilic micelles. The fluorescent structures inside the MB core visualized in our experiment are likely to represent those micelles of Ad solution. Once the Ad-loaded MB is in circulation, the natural blebbing that occurs as MB containment fails, releasing the captive Ad micelles, formerly held within in the core by hydrophobic interactions.

To estimate the number of infectious Ad particles fully loaded within an average-size MB, a concentrated solution (1 × 1011 PFU ml–1) of non-replicative Ad5/3-CMV-Luc virus was used to reconstitute a lyophilized vial containing 3 × 108 MBs. The background-subtracted Luc expression readings produced by MDA-MB-231 BC cells, treated with the Ad-loaded MBs were matched with Luc counts generated from a free Ad5/3-CMV-Luc infectivity standard. Using this strategy, it was found that a single MB can on average encapsulate 4.7±0.2 functional equivalents (PFUs) of Ad particles. Quantifying the level of Luc expression on the Ad-loaded MB-treated cells allowed calculating the total number of functional Ad particles, excluding non-infectious and/or complement-inactivated ones. During MB preparation, the integrity of the Ad particles may have been compromised, leading to a loss of infectivity. An alternative method to quantify the total number of Ad particles loaded within the MBs would involve quantification of total Ad DNA encapsulated by the reconstituted Ad-loaded MBs. However, this approach would simply reveal the total number of loaded Ad particles: both functional and non-functional. Thus, quantification of fully infectious Ad particles per MB was essential for subsequent determining the equivalent amount of the same Ad needed for treatment of the control groups. The determined value (4.7±0.2 PFU per MB) was used to assure equal dosing of unloaded and loaded Ad particles. A possible limitation of the MB loading method is the titer-dependent MB loading. It is hypothesized that a greater Ad titer may increase the PFU/MB ratio thereby improving the overall performance of the Ad-loaded MBs. In addition, a larger MB size distribution may permit a greater uptake of intact Ad particles.

MBs with a perfluorocarbon gas interior improve MB stability in an aqueous environment.6 The longer MB half-life enables the US contrast agent to remain in circulation and maintain the bubble echogenicity required for contrast-enhanced US. Previously investigated methods for drug or vector delivery using MBs rely on the natural effect of MB destruction during high frequency US to nonspecifically deliver a payload to the site of interest. For payload delivery applications independent of US, the improved stability of a perfluorocarbon MB adversely affects the delivery of loaded payload. In the absence of US, once the targeted MB is bound to the endothelium expressing the target receptors, the breakdown of the MB shell must occur to release the payload before the fastened MB is forced off because of the shear forces of circulation. In this regard, less stable MB would be beneficial for those applications where US triggering of payload release is technically unfeasible, particularly for generalized cancer detection applications. Future strategies of US-independent MB-based delivery may include less stable nitrogen or air-filled MB formulations, which may allow a more readily releasable payload.

This proof-of-principle study demonstrates the potential of targeted MB to deliver encapsulated Ad safely and more efficiently than traditional systemic delivery. Although the in vivo results are modest, they are significant, and will only be improved with future advancements in MB payload capacity and encapsulation techniques leading to greater Ad transduction efficiency. Future studies will include these advancements along with tissue biodistribution analysis, which would quantify liver uptake and circulatory persistence, and toxicity assays that would further quantify the improved patient safety afforded using encapsulated Ad.

Although this Ad delivery strategy is useful for cancer detection, the model can be extended to other applications where US-based payload release is not practical such as in atherosclerotic plaque detection or chemotherapeutic drug delivery for metastatic lesions. In addition, specific and nonspecific sequestering mechanisms, which normally reduce the compound's pharmacological half-life, would be circumvented using the MB delivery system. This would lead to a prolonged delivery of a compound or Ad, increasing the half-life and allowing lower doses to be administered. For example, inactivation of MB surface-associated Ad particles was shown to reduce anti-Ad5 hexon antibody binding, that is, the MB immunoreactivity. Elimination of these extraneous and detrimental particles would reduce first-line Ad sequestration by liver that normally hinders a systemic virus delivery.

Targeting Ad-loaded MBs to tissue for payload delivery without the requirement of external US triggering represents a novel use of MBs. The unique ability of MBs to package macromolecules and target systemic tissue makes them a viable alternative to traditional adenoviral therapy. Here, the MB was shown to fully encapsulate the Ad particles localizing them to the MB core where their immunoreactivity potential is the lowest. The enhanced tissue delivery and immunogenic protection afforded with MB loading permits reduced payload dosage, potentially reducing its clinical toxicity for the patients and thereby circumventing the complications traditionally associated with gene therapy.42 The safety of MBs for systemic application has been well documented43 and the Food and Drug Administration approval for human use further augments the auspicious potential for immediate clinical use.

References

  1. 1

    Du J, Li FH, Fang H, Xia JG, Zhu CX . Correlation of real-time gray scale contrast-enhanced ultrasonography with microvessel density and vascular endothelial growth factor expression for assessment of angiogenesis in breast lesions. J Ultrasound Med 2008; 27: 821–831.

  2. 2

    Hudson JM, Karshafian R, Burns PN . Quantification of flow using ultrasound and microbubbles: a disruption replenishment model based on physical principles. Ultrasound Med Biol 2009; 35: 2007–2020.

  3. 3

    Hoyt K, Warram JM, Umphrey H, Belt L, Lockhart ME, Robbin ML et al. Determination of breast cancer response to bevacizumab therapy using contrast-enhanced ultrasound and artificial neural networks. J Ultrasound Med 2010; 29: 577–585.

  4. 4

    Guibal A, Taillade L, Mule S, Comperat E, Badachi Y, Golmard JL et al. Noninvasive contrast-enhanced US quantitative assessment of tumor microcirculation in a murine model: effect of discontinuing anti-VEGF therapy. Radiology 2010; 254: 420–429.

  5. 5

    Tartis MS, McCallan J, Lum AF, LaBell R, Stieger SM, Matsunaga TO et al. Therapeutic effects of paclitaxel-containing ultrasound contrast agents. Ultrasound Med Biol 2006; 32: 1771–1780.

  6. 6

    Ferrara K, Pollard R, Borden M . Ultrasound microbubble contrast agents: fundamentals and application to gene and drug delivery. Annu Rev Biomed Eng 2007; 9: 415–447.

  7. 7

    Taylor SL, Rahim AA, Bush NL, Bamber JC, Porter CD . Targeted retroviral gene delivery using ultrasound. J Gene Med 2007; 9: 77–87.

  8. 8

    Sorace AG, Warram JM, Umphrey H, Hoyt K . Microbubble-mediated ultrasonic techniques for improved chemotherapeutic delivery in cancer. J Drug Target 2012; 20: 43–54.

  9. 9

    Hauff P, Seemann S, Reszka R, Schultze-Mosgau M, Reinhardt M, Buzasi T et al. Evaluation of gas-filled microparticles and sonoporation as gene delivery system: feasibility study in rodent tumor models. Radiology 2005; 236: 572–578.

  10. 10

    Willmann JK, Deshpande RH, Lutz NAM, Cochran JR, Gambhir SS . Targeted contrast-enhanced ultrasound imaging of tumor angiogenesis with contrast microbubbles conjugated to integrin-binding knottin peptides. J Nucl Med 2010; 51: 433–440.

  11. 11

    Klibanov AL . Microbubble contrast agents: targeted ultrasound imaging and ultrasound-assisted drug-delivery applications. Invest Radiol 2006; 41: 354–362.

  12. 12

    Lindner JR . Microbubbles in medical imaging: current applications and future directions. Nat Rev Drug Discov 2004; 3: 527–532.

  13. 13

    Klibanov AL . Ligand-carrying gas-filled microbubbles: ultrasound contrast agents for targeted molecular imaging. Bioconjugate Chem 2005; 16: 9–17.

  14. 14

    Klibanov AL, Rychak JJ, Yang WC, Alikhani S, Li B, Acton S et al. Targeted ultrasound contrast agent for molecular imaging of inflammation in high-shear flow. Contrast Media Mol Imag 2006; 1: 259–266.

  15. 15

    Behm CZ, Kaufmann BA, Carr C, Lankford M, Sanders JM, Rose CE et al. Molecular imaging of endothelial vascular cell adhesion molecule-1 expression and inflammatory cell recruitment during vasculogenesis and ischemia-mediated arteriogenesis. Circulation 2008; 117: 2902–2911.

  16. 16

    Lindner JR . Contrast ultrasound molecular imaging of inflammation in cardiovascular disease. Cardiovasc Res 2009; 84: 182–189.

  17. 17

    Kaufmann BA, Lindner JR . Molecular imaging with targeted contrast ultrasound. Curr Op in Biotech 2007; 18: 11–16.

  18. 18

    Unger EC, McCreery TP, Sweitzer RH, Shen D, Wu G . In vitro studies of a new thrombus-specific ultrasound contrast agent. Am J Cardiol 1998; 81: 58–61.

  19. 19

    Leong-Poi H, Christiansen J, Christiansen J, Klibanov AL, Kaul S, Lindner JR . Noninvasive assessment of angiogenesis by ultrasound microbubbles targeted to alpha-v integrins. Circulation 2003; 107: 455–460.

  20. 20

    Weller GE, Wong MK, Modzelewski RA, Lu E, Klibanov AL, Wagner WR et al. Ultrasonic imaging of tumor angiogenesis using contrast microbubbles targeted via the tumor-binding peptide arginine-arginine-leucine. Cancer Res 2005; 65: 533–539.

  21. 21

    Korpanty G, Carbon JG, Grayburn PA, Fleming JB, Brekken RA . Monitoring response to anticancer therapy by targeting microbubbles to tumor vasculature. Clin Cancer Res 2007; 13: 323.

  22. 22

    Willmann JK, Lutz AM, Paulmurugan R, Patel MR, Chu P, Rosenberg J et al. Dual-targeted contrast agent for US assessment of tumor angiogenesis in vivo. Radiology 2008; 248: 936–944.

  23. 23

    Weller GE, Villanueva FS, Tom EM, Wagner WR . Targeted ultrasound contrast agents: in vitro assessment of endothelial dysfunction and multi-targeting to ICAM-1 and sialyl Lewisx. Biotechnol Bioeng 2005; 92: 780–788.

  24. 24

    Rapoport N, Gao Z, Kennedy A . Multifunctional nanoparticles for combining ultrasonic tumor imaging and targeted chemotherapy. J Natl Cancer Inst 2007; 99: 1095–1106.

  25. 25

    Warram JM, Sorace AG, Saini R, Umphrey H, Zinn KR, Hoyt K . Triple-targeted US contrast agent provides improved localization to tumor vasculature. J Ultrasound Med 2011; 30: 921–931.

  26. 26

    Lindner JR, Kaul S . Delivery of drugs with ultrasound. Echocardiography 2001; 18: 329–337.

  27. 27

    Tinkov S, Bekeredjian R, Winter G, Coester C . Microbubbles as ultrasound triggered drug carriers. J Pharm Sci 2009; 98: 1935–1961.

  28. 28

    Korpanty G, Chen S, Shohet RV, Ding J, Yang B, Frenkel PA et al. Targeting of VEGF-mediated angiogenesis to rat myocardium using ultrasonic destruction of microbubbles. Gene Ther 2005; 12: 1305–1312.

  29. 29

    Tartis MS, McCallan J, Lum AF, LaBell R, Stieger SM, Matsunaga TO et al. Therapeutic effects of paclitaxel-containing ultrasound contrast agents. Ultrasound Med Biol 2006; 32: 1771–1780.

  30. 30

    Unger EC, McCreery TP, Sweitzer RH, Caldwell VE, Wu Y . Acoustically active lipospheres containing paclitaxel: a new therapeutic ultrasound contrast agent. Invest Radiol 1998; 33: 886–892.

  31. 31

    Qiu L, Zhang L, Wang L, Jiang Y, Luo Y, Peng Y et al. Ultrasound-targeted microbubble destruction enhances naked plasmid DNA transfection in rabbit Achilles tendons in vivo. Gene Ther; advance online publication, 10 October 2011 [E-pub ahead of print].

  32. 32

    Mayer CR, Geis NA, Katus HA, Bekeredjian R . Ultrasound targeted microbubble destruction for drug and gene delivery. Expert Opin Drug Deliv 2008; 5: 1121–1138.

  33. 33

    Howard CM, Forsberg F, Minimo C, Liu JB, Merton DA, Claudio PP . Ultrasound guided site-specific gene delivery system using adenoviral vectors and commercial ultrasound contrast agents. J Cell Physiol 2006; 209: 413–421.

  34. 34

    Greco A, Di Benedetto A, Howard CM, Kelly S, Nande R, Dementieva Y et al. Eradication of therapy-resistant human prostate tumors using an ultrasound-guided site-specific cancer terminator virus delivery approach. Mol Ther 2010; 18: 295–306.

  35. 35

    Smith RA, Cokkinides V, Brawley OW . Cancer screening in the United States, 2009: a review of current American Cancer Society guidelines and issues in cancer screening. CA Cancer J Clin 2009; 59: 27–41.

  36. 36

    Warram JM, Borovjagin AV, Zinn KR . A genetic strategy for combined screening and localized imaging of breast cancer. Mol Imaging Biol 2011; 13: 452–461.

  37. 37

    Perk J, Iavarone A, Benezra R . Id family of helix-loop-helix proteins in cancer. Nat Rev Cancer 2005; 5: 603–614.

  38. 38

    Borovjagin AV, Krendelchtchikov A, Ramesh N, Yu DC, Douglas JT, Curiel DT . Complex mosaicism is a novel approach to infectivity enhancement of adenovirus type 5-based vectors. Cancer Gene Ther 2005; 12: 475–486.

  39. 39

    Kanerva A, Mikheeva GV, Krasnykh V, Coolidge CJ, Lam JT, Mahasreshti PJ et al. Targeting adenovirus to the serotype 3 receptor increases gene transfer efficiency to ovarian cancer cells. Clin Cancer Res 2002; 8: 275–280.

  40. 40

    Matthews QL, Sibley DA, Wu H, Li J, Stoff-Khalili MA, Waehler R et al. 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. Mol Imag 2006; 5: 510–519.

  41. 41

    Le LP, Le HN, Dmitriev IP, Davydova JG, Gavrikova T, Yamamoto S et al. Dynamic monitoring of oncolytic adenovirus in vivo by genetic capsid labeling. J Natl Cancer Inst 2006; 98: 203–214.

  42. 42

    Raper SE, Yudkoff M, Chirmule N, Gao GP, Nunes F, Haskal ZJ et al. A pilot study of in vivo liver-directed gene transfer with an adenoviral vector in partial ornithine transcarbamylase deficiency. Hum Gene Ther 2002; 13: 163–175.

  43. 43

    Jakobsen JA, Oyen R, Thomsen HS, Morcos SKR . Members of Contrast Media Safety Committee of European Society of Urogenital. Safety of ultrasound contrast agents. Eur Radiol 2005; 15: 941–945.

Download references

Acknowledgements

This work was supported the UAB Small Animal Imaging Shared Facility NIH Research Core Grant (P30CA013148) and the Department of Defense (BC050034).

Author information

Affiliations

Authors

Corresponding author

Correspondence to K R Zinn.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Warram, J., Sorace, A., Saini, R. et al. Systemic delivery of a breast cancer-detecting adenovirus using targeted microbubbles. Cancer Gene Ther 19, 545–552 (2012). https://doi.org/10.1038/cgt.2012.29

Download citation

Keywords

  • adenovirus
  • detection
  • fluorescence
  • microbubble

Further reading