Analysis of Targeted Viral Protein Nanoparticles Delivered to HER2+ Tumors

Published 6/18/2013

Your institution must subscribe to JoVE's Bioengineering section to access this content.

Fill out the form below to receive a free trial or learn more about access:


Enter your email below to get your free 10 minute trial to JoVE!

By clicking "Submit", you agree to our policies.



This article details the procedures for optical imaging analysis of the tumor-targeted nanoparticle, HerDox. In particular, detailed use of the multimode imaging device for detecting tumor targeting and assessing tumor penetration is described here.

Cite this Article

Copy Citation

Hwang, J. Y., Farkas, D. L., Medina-Kauwe, L. K. Analysis of Targeted Viral Protein Nanoparticles Delivered to HER2+ Tumors. J. Vis. Exp. (76), e50396, doi:10.3791/50396 (2013).


The HER2+ tumor-targeted nanoparticle, HerDox, exhibits tumor-preferential accumulation and tumor-growth ablation in an animal model of HER2+ cancer. HerDox is formed by non-covalent self-assembly of a tumor targeted cell penetration protein with the chemotherapy agent, doxorubicin, via a small nucleic acid linker. A combination of electrophilic, intercalation, and oligomerization interactions facilitate self-assembly into round 10-20 nm particles. HerDox exhibits stability in blood as well as in extended storage at different temperatures. Systemic delivery of HerDox in tumor-bearing mice results in tumor-cell death with no detectable adverse effects to non-tumor tissue, including the heart and liver (which undergo marked damage by untargeted doxorubicin). HER2 elevation facilitates targeting to cells expressing the human epidermal growth factor receptor, hence tumors displaying elevated HER2 levels exhibit greater accumulation of HerDox compared to cells expressing lower levels, both in vitro and in vivo. Fluorescence intensity imaging combined with in situ confocal and spectral analysis has allowed us to verify in vivo tumor targeting and tumor cell penetration of HerDox after systemic delivery. Here we detail our methods for assessing tumor targeting via multimode imaging after systemic delivery.


Tumor-targeting of chemotherapy has the potential to eliminate cancer cells at lower dose compared to untargeted drugs because more of the delivered therapy can accumulate at its intended destination rather than distribute to non-tumor tissue. As the latter situation would dilute out the efficacy of the drug and thus require higher doses to be effective, tumor-targeting has both therapeutic and safety advantages over standard non-targeted treatment.

Targeting chemotherapy by encapsulation in self-assembled nanoparticles allows the drug to remain chemically unmodified in contrast to drugs that are covalently linked to targeting molecules. As such linkage has the potential to alter the activity of both the drug and the targeting molecule; non-covalent assembly allows drug potency to be retained.

We have previously shown that the novel three-component, self-assembled complex, HerDox, targets HER2+ tumors in vivo and elicits tumor-growth ablation while sparing normal tissue, including the heart 1. HerDox is formed through non-covalent interactions between the receptor-binding cell-penetration protein, HerPBK10, and the chemotherapeutic agent, doxorubicin (Dox), via a small nucleic acid linker. HerPBK10 binds the human epidermal growth factor receptor (HER) and triggers receptor-mediated endocytosis 2-4, while endosomal membrane penetration is accomplished through incorporation of the adenovirus-derived penton base capsid protein 4-6. A positively-charged domain on the protein enables nucleic acid binding 4, 5, through which DNA-intercalated Dox can be transported for targeted delivery. Electrophilic, intercalation, and possibly protein oligomerization interactions facilitate self-assembly into round 10-20 nm particles that are stable in blood and under extended storage at different temperatures 1. Preferential targeting to HER2+ tumor cells is facilitated by the enhanced ligand affinity when HER2 is elevated.

Our previous studies have shown that systemic delivery of HerDox yields preferential accumulation in tumors over non-tumor tissue and in comparison to untargeted Dox 1, and penetration into tumor cells in vivo 7. We have observed that HerDox releases Dox after tumor cell entry, allowing Dox accumulation into the nucleus 1. Tumor-accumulation appears to correlate with receptor level, as relatively low HER2-expressing tumors accumulate less HerDox compared to those with comparatively higher HER2 levels 1. Moreover, the effective cell death concentration exhibits an inverse correlation with HER2 display on tumor cell lines expressing differing cell surface HER2 levels 1. HerDox exhibits a therapeutic and safety advantage over untargeted Dox, as tumor killing occurs at over 10-times lower dose compared to the untargeted drug and yields no detectable adverse effect on heart (detected by echocardiography and histological stain) or liver (detected by TUNEL stain) tissue, in contrast to untargeted Dox 1. Despite its derivation from a viral capsid protein, HerPBK10 exhibits no detectable immunogenicity at therapeutic levels 2. Whereas pre-existing antibodies to whole adenovirus can recognize HerPBK10, they are unable to prevent cell binding 2.

Tumor volume measured over time is a standard method of assessing therapeutic efficacy of targeted therapeutics, and has been employed for assessing the therapeutic efficacy of HerDox. Supplementing this approach with in vivo and ex vivo fluorescence intensity imaging has allowed us to better assess targeting efficiency 7. We have specifically integrated in situ confocal imaging of excised tumors with spectral analysis of Dox fluorescence to verify that HerDox not only accumulated at tumors in vivo but penetrated into tumor cells and delivered Dox into the cytoplasm and nucleus 7. Spectral analysis furthermore enabled us to distinguish Dox fluorescence from autofluorescence 7.

Here we demonstrate in greater detail our approach for assessing HerDox in vivo after systemic delivery, and most importantly, for assessing targeting through multimode imaging methods and analyses.

Subscription Required. Please recommend JoVE to your librarian.


1. Systemic Delivery In vivo

  1. Mix enough HerDox with sterile saline to equate 0.2 ml of a 0.004 mg/kg dose of HerDox per injection for a 6-8 week old NU/NU mouse bearing subcutaneous bilateral flank xenograft tumors.
  2. Gently draw the HerDox mixture into a 3/10 cc insulin syringe fitted with a 29G needle, avoiding bubbles.
  3. Anesthesia is induced by brief isoflurane exposure in an induction chamber equipped with a gas scavenging system (Oxygen flow rates: 0.5-1 L/min, isoflurane concentration: 3-4% (or lower).
  4. Inject the entire mixture into the tail vein of an anesthetized mouse (0.2 ml per injection). IV injection can also be performed in a restrained, unanesthetized mouse.
  5. Repeat injections on the same mouse for six more sequential days, once per day.

2. Fluorescence Imaging In vivo

The accumulation of HerDox fluorescence in tumors can be detectable by the last day of injection (Day 7) using a multimode imager. The procedures below entail the use of a customized macro-illumination and detection system (Figure 1) 8.

  1. Turn on the Multimode In vivo Optical Imager.
  2. Select an emission bandpass filter (590 nm ± 30 nm) suited for doxorubicin fluorescence detection.
  3. Turn on Argon-Krypton laser and place an excitation bandpass filter (488 nm ± 10 nm) at the laser optical path.
  4. Turn on the anesthesia system and then place a mouse into the anesthetizing chamber (Oxygen flow rates: 0.5-1 L/min, isoflurane concentration: 3-4% (or lower)equipped with a gas scavenging system.
  5. Transfer the mouse from the anesthetizing chamber to the imaging chamber of the Multimode In vivo Optical Imager when the mouse is anesthetized.
  6. Place a nosecone over the nose of the mouse and open the flow to administer continuous anesthesia during image acquisition (Oxygen flow rates: 0.5-1 L/min, isoflurane concentration: 2-3% (or lower).
  7. Acquire fluorescence images using an exposure time of 5-15 sec.
  8. Perform image analysis and processing including background correction or contrast adjustment.

3. Fluorescence Imaging Ex vivo

HerDox fluorescence can be imaged in tumors and specific organs (including the liver, kidney, spleen, heart, and skeletal muscle) harvested from euthanized mice at 24 hr after the final (Day 7) injection of HerDox.

  1. Turn on the Multimode In vivo Optical Imager.
  2. Select an emission bandpass filter (580 nm ± 20 nm) for Doxorubicin fluorescence detection.
  3. Turn on Argon-Krypton laser and place an excitation bandpass filter (488 nm ±10 nm) at the laser optical path.
  4. Place tumors and specific organs arranged on a Petri-dish into the imaging chamber of the Multimode In vivo Optical Imager.
  5. Acquire fluorescence images of tissues using an exposure time of 5-15 sec. An example of initial fluorescence image acquisition is shown in Figure 2a. Repeat the same using an empty Petri-dish, which will serve as the Background (Figure 2b).
  6. Perform image analysis and processing including background correction or contrast adjustment. An example of a corrected image is shown in Figure 2c, resulting from subtraction of Figure 2b from Figure 2a.

4. In situ Confocal Imaging of Tumors

In situ confocal imaging allows detection and analysis of HerDox tumor accumulation at the cellular level.

  1. Turn on a Leica SPE confocal microscope.
  2. Select 488 nm laser light for excitation of doxorubicin and emission wavelengths (560-620 nm) for doxorubicin fluorescence detection.
  3. Select a 40X or 63X objective and drop immersion oil on the objective lens.
  4. Extract fresh tumors from euthanized mice that previously received HerDox and mock treatments as described in procedure 2.
  5. Place tumors on a Petri-dish on ice to avoid tissue degradation, and then transfer the tumors to a Delta T chamber for confocal imaging.
  6. Acquire confocal images of the tumors at sequential focal depths (step size: 1 μm, thickness: 20 μm). An example of sequentially-acquired images along the z-axis is shown in Figure 3, left panel.
  7. Perform maximum intensity z-projection of the images. A maximum intensity projection of z-stacked images is shown in Figure 3, right panel.
  8. Calculate mean fluorescence intensities of the maximum intensity Z-projection images. Mean fluorescence intensities of the images over the overall field of view were calculated using ImageJ.

5. Ratiometric Spectral Imaging and Analysis

Ratiometric spectral imaging and analysis allows discrimination between Dox fluorescence and autofluorescence.

  1. Power on a laser scanning fluorescence confocal microscope.
  2. Acquire 15 images of the HerDox-treated and untreated tumors at a specified depth within the spectral range of 510-650 nm, with a step size of 10 nm, and excitation at 488 nm light using a Leica SPE confocal microscope.
  3. Prepare a 100 μM solution of doxorubicin.
  4. Perform spectral imaging of the 100 μM doxorubicin solution to obtain the pure spectral signature of Dox fluorescence (spectral range: 510-650 nm, a step size: 10 nm). Typical results of image acquisition from spectral imaging and resulting fluorescence spectrum plotted as a graph are shown in Figure 4.
  5. Acquire the autofluorescence spectral signature from an image cube (spectral range: 510-650 nm, a step size: 10 nm) obtained by spectral imaging of untreated tumors. Typical results of image acquisition from spectral imaging and resulting fluorescence spectrum plotted as a graph are shown in Figure 5.
  6. Generate four reference spectral signatures (pure autofluorescence, 0.1.doxorubin+0.9.autofluorescence, 0.2. doxorubicin+0.8.autofluorescence, 0.3.doxorubin+0.7.autofluorescence) using the program we developed 9. A typical curve showing four reference spectral signatures is shown in Figure 6.
  7. Perform spectral classification of the images as defined by the reference spectral signatures through Euclidean distance measure using the program we previously developed 9.
  8. Perform linear spectral unmixing of those images by using a spectral unmixing program (plug-in in ImageJ) we developed 7, 10, for comparison to the ratiometric spectral imaging and analysis. An example of separating HerDox fluorescence from autofluorescence by linear spectral unmixing is shown in Figure 7.

Subscription Required. Please recommend JoVE to your librarian.

Representative Results

Figure 1 shows the in vivo optical imager prototype, which was built for the purpose of image acquisition under multiple modalities, including fluorescence intensity, spectral, lifetime, 2-photon, intra-vital confocal, and bioluminescence imaging. In addition, the cooled high sensitive camera and high power laser lines incorporated in this system yields higher contrast fluorescence images compared to commercial optical imaging systems 11, especially for the in vivo detection of doxorubicin fluorescence. Hence, this system was critical for use in the present study to acquire multiple complementary data points in the assessment of HerDox in vivo.

The high tumor fluorescence shown in Figure 2 is typical after systemic delivery of HerDox, and indicative of its tumor-preferential targeting. The fluorescence detected in the skeletal muscle is atypical and may result from anomalies in the injection procedure (i.e. injecting too close to the body). Qualitatively, the image shown in Figure 2a has low background. However, to perform a quantitative analysis of the relative tissue fluorescence intensities, it is necessary to perform a background correction using an empty dish, as shown in Figure 2b. This can result in a "background corrected" image shown in Figure 2c that has a higher background signal at the edge of the image compared to the original (Figure 2a) because the background image (Figure 2b) has lower intensities at the edge of the image. Thus, while background correction can yield a signal increase at the edge of the image, this is a necessary step for subsequent quantitative analysis.

In order to obtain a quantitative assessment of HerDox accumulation in tumors, confocal fluorescence images were obtained at different focal depths in situ followed by maximum intensity z-projection of the images (Figure 3), and acquisition of the mean fluorescence intensity of the projected image. The resultant image contains the HerDox accumulation information over the indicated z-depths at every pixel, and thus the average of the fluorescence intensities reflects overall HerDox accumulations in tumors at different depths quantitatively.

To distinguish HerDox fluorescence from autofluorescence in tumors and quantitatively discriminate the two with high specificity, ratiometric spectral imaging and analysis was performed. The pure spectral signatures for doxorubicin and autofluorescence obtained by spectral imaging are illustrated in Figures 4 and 5, respectively. Reference spectral signatures (Figure 6) were also created by mixing different ratios of the pure spectral signatures using the program we developed 9. Each reference spectral signature here represents relative fluorescence of HerDox accumulated in tumors with respect to autofluorescence. The spectral classification images (data was not shown in this paper) showed better quantitative HerDox accumulations 7 compared to the spectral unmixed image obtained by using the pure reference spectral signatures (Figure 7). Figures 4-7 represent typical findings when performing ratiometric spectral imaging and analysis.

Figure 1
Figure 1. In vivo optical imager. An optical imaging system with fluorescence intensity, spectral, fluorescence lifetime and intravital confocal imaging capabilities must be used. The multimode optical imaging system used here is shown, and has been described previously 8. Image provided with kind permission from Springer Science & Business Media B.V.: Hwang, J.Y., et al., Mol. Imaging Biol., 14, 431-42 (2012).

Figure 2
Figure 2. Representative fluorescence intensity images of internal organs and tumors extracted from a mouse receiving HerDox. Tissues and tumors were harvested from euthanized mice at 24 hr after the final (Day 7) injection of HerDox Fluorescence intensity images before and after background correction (b) are represented by (a) and (c), respectively. Click here to view larger figure.

Figure 3
Figure 3. Maximum intensity z-projection of the images obtained at sequential z-depths. The stacked images shown at the left are representative of the confocal scans of tumor tissue obtained at sequential 1 μm depths. The image at the right shows the compilation of the stacked images on the left.

Figure 4
Figure 4. Acquisition of a pure spectral signature for doxorubicin. After the images of doxorubicin solution were obtained at sequential wavelengths within 510-650 nm, with a step-size of 10 nm (left), the pure spectral signature for doxorubicin was obtained using the program we developed. Click here to view larger figure.

Figure 5
Figure 5. Acquisition of a pure spectral signature for autofluorescence. After the images of untreated tumor were obtained at sequential wavelengths within 510-650 nm, with a step-size of 10 nm (left), the pure spectral signature for autofluorescence was obtained using the program we developed. Click here to view larger figure.

Figure 6
Figure 6. Four reference spectral signatures for ratiometric spectral imaging and analysis. The reference spectral signatures were created by mixing the pure spectral signatures using the program we developed. The green-colored solid line represents pure autofluorescence, the red-colored solid line represents 0.1.doxorubin+0.9.autofluorescence, the blue-colored solid line represents 0.2. doxorubicin+0.8.autofluorescence, and the cyan-colored solid line 0.3.doxorubin+0.7.autofluorescence, respectively. Figure is reproduced from the authors' previous work: Hwang, J.Y., et al., PLoS One. 7 (4), (2012).

Figure 7
Figure 7. HerDox fluorescence in tissue before and after spectral un-mixing. After obtaining the spectral signatures of Dox and autofluorescence, the Dox signals are separated by using linear spectral un-mixing 12,13. The figure shows the autofluorescence and HerDox mixed image before spectral un-mixing and the separated HerDox image after spectral un-mixing.

Subscription Required. Please recommend JoVE to your librarian.


Dox fluorescence can be detectable in vivo using the multimode imager when tumors are subcutaneous. However, the therapeutically effective dose of HerDox (0.004 mg/kg) is below the detection threshold after a single dose. In contrast, after 7 daily injections (1x/day for 7 days), the tumor accumulation and retention of the particle is sufficient to enable visualization of Dox fluorescence.

It is critical when working with Dox or any other fluorophore for in vivo imaging that clean technique is used. Drips on the outside of the mouse should be avoided, as the imager will detect the fluorescence on the outside of the skin without discriminating whether the Dox is external or internal. Similarly, Dox contamination on the gloves can leave fluorescent marks on the outside of the mouse during animal handling, thereby causing erroneous fluorescence detection.

In the multimode optical system, a laser beam is utilized for excitation of HerDox. The laser beam typically has a Gaussian profile. Hence, for uniform excitation of specimens of interest, optical diffusers are utilized. Nevertheless, the diffused laser light somewhat preserves a Gaussian beam profile. Thus, for ideal quantitative analysis, background correction must be performed. The before-corrected and background images typically have a depth of 16 bits. Therefore, before the background correction is performed, the image depth should be converted to 32 bits in order to avoid unwanted errors from the image depth discrepancy between resultant and input images.

In the acquisition of a background image, a black paper (Artagain) was previously put on an imaging plate in order to prevent illumination reflections before performing fluorescence imaging. However, since the camera is highly sensitive, low background signals from the black-paper was detected by the camera. Therefore, the background signals shown in Figure 2 came from the black paper rather than from an empty Petri-dish.

A high concentration of doxorubicin solution was utilized to obtain pure spectral signatures with high accuracy. We further examined spectral resolutions at which the system is capable of detecting enough fluorescence from the specimen to produce reliable spectral signatures. As a result, the spectral resolution was tuned to 10 nm. Here we found that the microscope under the setting of higher spectral resolutions (<10 nm) could not detect sufficient fluorescence to yield reliable spectral signatures. With the pure spectral signatures, we could generate several ratiometric spectral signatures (Figure 6). In particular, we could quantitatively distinguish HerDox fluorescence signals from autofluorescence by the respective ratiometric spectral signatures 7. The spectral unmixed image obtained with the pure spectral signatures (Figure 7) was similar to the classified image by the two reference spectral signatures (blue: 0.2.Dox+0.8.Autofl. and Green: Autofl.). Altogether, the ratiometric confocal/spectral imaging provided more quantitative information on HerDox accumulation in tumor cells in situ with high resolution, thus verifying its usefulness in the analysis of the nanoparticles delivered to HER2+ tumors.

Subscription Required. Please recommend JoVE to your librarian.


The author, Daniel Farkas, is Chairman of Spectral Molecular Imaging. The remaining authors have no competing interests.


This work was funded by grants to LKM-K from the National Institutes of Health/National Cancer Institute (R01CA129822 and R01CA140995). Dr. Medina-Kauwe thanks C. Rey, M. M-Kauwe and D. Revetto for continued support.


Name Company Catalog Number Comments
Fluorescence laser scanning confocal microscope Leica SPE
In Vivo Optical Imager Spectral Molecular Imaging Multimode In Vivo Optical Imager
Doxorubicin-HCl Sigma-Aldrich D4035
Nude (NU/NU) mouse, female, 6-8 week Charles River Strain code 088
MDA-MB-435 human HER2+ tumor cells NCI-Frederick Cancer DCTD Tumor/Cell Line Repository 0507292
3/10 cc insulin syringe U-100 with 29G x 1/2" Ultra-FineIV permanently attached needle BD 309301
Delta T chamber Bioptechs 04200417B



  1. Agadjanian, H., Chu, D., et al. Chemotherapy Targeting by DNA Capture in Viral Protein Particles. Nanomedicine. 7, (3), 335-352 (2012).
  2. Agadjanian, H., Ma, J., et al. Tumor detection and elimination by a targeted gallium corrole. Proc. Natl. Acad. Sci. U.S.A. 106, (15), 6105-6110 (2009).
  3. Agadjanian, H., Weaver, J. J., et al. Specific delivery of corroles to cells via noncovalent conjugates with viral proteins. Pharm. Res. 23, (2), 367-377 (2006).
  4. Medina-Kauwe, L. K., Maguire, M., et al. Non-viral gene delivery to human breast cancer cells by targeted Ad5 penton proteins. Gene Therapy. 81753-81761 (2001).
  5. Medina-Kauwe, L. K., Kasahara, N., et al. 3PO, a novel non-viral gene delivery system using engineered Ad5 penton proteins. Gene Therapy. 8795-8803 (2001).
  6. Rentsendorj, A., Xie, J., et al. Typical and atypical trafficking pathways of Ad5 penton base recombinant protein: implications for gene transfer. Gene Ther. 13, (10), 821-836 (2006).
  7. Hwang, J. Y., Park, J., et al. Multimodality Imaging In vivo for Preclinical Assessment of Tumor-Targeted Doxorubicin Nanoparticles. PLoS ONE. 7, (4), e34463 (2012).
  8. Hwang, J. Y., Wachsmann-Hogiu, S., et al. A Multimode Optical Imaging System for Preclinical Applications In Vivo: Technology Development, Multiscale Imaging, and Chemotherapy Assessment. Mol. Imaging Biol. (2011).
  9. Hwang, J. Y., Gross, Z., et al. Ratiometric spectral imaging for fast tumor detection and chemotherapy monitoring in vivo. J. Biomed. Opt. 16, (6), 066007 (2011).
  10. Fujimoto, J. G., Farkas, D. L. Biomedical Optical Imaging. Oxford University Press. New York. (2009).
  11. Hwang, J. Y., Moffatt-Blue, C., et al. Multimode optical imaging of small animals: development and applications. Proc. of SPIE. 6411, (2007).
  12. Ducros, M., Moreaux, L., et al. Spectral unmixing: analysis of performance in the olfactory bulb in vivo. PLoS One. 4, (2), e4418 (2009).
  13. Zimmermann, T. Spectral imaging and linear unmixing in light microscopy. Adv. Biochem. Eng. Biotechnol. 95245-95265 (2005).



    Post a Question / Comment / Request

    You must be signed in to post a comment. Please or create an account.

    Video Stats