LETTERS PUBLISHED ONLINE: 30 MARCH 2015 | DOI: 10.1038/NNANO.2015.25
In situ conversion of porphyrin microbubbles to nanoparticles for multimodality imaging Elizabeth Huynh1,2, Ben Y. C. Leung3, Brandon L. Helfield2,3, Mojdeh Shakiba1,2, Julie-Anne Gandier4, Cheng S. Jin1,5,6, Emma R. Master4, Brian C. Wilson1,2, David E. Goertz2,3 and Gang Zheng1,2,5,6* Converting nanoparticles or monomeric compounds into larger supramolecular structures by endogenous1,2 or external3,4 stimuli is increasingly popular because these materials are useful for imaging and treating diseases. However, conversion of microstructures to nanostructures is less common. Here, we show the conversion of microbubbles to nanoparticles using low-frequency ultrasound. The microbubble consists of a bacteriochlorophyll–lipid shell around a perfluoropropane gas. The encapsulated gas provides ultrasound imaging contrast and the porphyrins in the shell confer photoacoustic and fluorescent properties. On exposure to ultrasound, the microbubbles burst and form smaller nanoparticles that possess the same optical properties as the original microbubble. We show that this conversion is possible in tumour-bearing mice and could be validated using photoacoustic imaging. With this conversion, our microbubble can potentially be used to bypass the enhanced permeability and retention effect when delivering drugs to tumours. The intrinsic conversion of supramolecular structures involves the activation of shape transitions. Of growing interest in the biomedical field is the conversion of perfluorocarbon nanodroplets as activatable ultrasound contrast agents. These nanometre-sized droplets expand to form micrometre-sized bubbles upon heating5,6. Despite a growing number of investigations utilizing the conversion from nano- to micro-sized structures, to our knowledge there have been no studies exploiting the advantages of a micro-to-nano conversion. The most conventional micrometre-sized structure investigated for biomedical applications is the microbubble, a gas-filled microsphere formed with a biocompatible shell composed of lipids, proteins or polymers. In addition to acting as ultrasound contrast agents, microbubbles have also been studied for their bursting behaviour in response to destructive ultrasound for the measurement of blood flow parameters7 and drug and gene delivery8. The interaction of a lipid microbubble with ultrasound pulses of sufficient amplitude can result in its fragmentation or shrinkage, which may be accompanied by ‘shedding’ of shell material9. These processes lead to the formation of structures that significantly decrease the ultrasound imaging contrast relative to that of the initial microbubble10. Furthermore, the remnants of the microbubbles, the lipids themselves, do not serve any imaging purpose. Recently, we discussed the concept of intrinsically multimodal contrast agents11 and developed multimodal ultrasound contrast agents in which the building blocks of the microbubble had optical and metal chelation properties. We used a porphyrin–lipid to form a shell around a perfluorocarbon gas, forming porphyrin
microbubbles (pMBs). The encapsulated gas provided the ultrasound imaging contrast, and the high density of porphyrins provided photoacoustic and fluorescence contrast12,13. Here we utilize these properties to investigate the response of pMBs to lowfrequency ultrasound. The unique feature of forming the multimodality microbubble with building blocks that have intrinsic optical properties ensures that the responses observed are properties of the pMB after destruction and the formation of porphyrin nanoparticles (pNPs) from the pMB (Fig. 1a,b), and not simply the release of imaging agents from the microbubble as others have investigated14. Although previously reported pMBs were formed using pyropheophorbide–lipid, the pMBs used here were formed using a a
Perfluorocarbon gas
b Transducer Conversion ultrasound
Nanoparticles 5–500 nm
Microbubble 1–10 µm
Figure 1 | Schematics of porphyrin microbubbles (pMBs) and their micro-to-nano conversion. a, pMBs consist of a BChl-lipid shell encapsulating perfluorocarbon gas. b, Conversion of pMBs to porphyrin nanoparticles (pNPs) via sonication with low-frequency, high-duty-cycle ultrasound (conversion ultrasound).
1
Princess Margaret Cancer Center and Techna Institute, University Health Network, Toronto, Ontario M5G 2M9, Canada. 2 Department of Medical Biophysics, University of Toronto, Toronto, Ontario M5G 1L7, Canada. 3 Sunnybrook Health Sciences Center, Toronto, Ontario M4N 3M5, Canada. 4 Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto, Ontario M5S 3E5, Canada. 5 Department of Pharmaceutical Sciences, University of Toronto, Toronto, Ontario M5S 3M2, Canada. 6 Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, Ontario M5G 1L7, Canada. *e-mail: [email protected] NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology
© 2015 Macmillan Publishers Limited. All rights reserved
1
LETTERS
NATURE NANOTECHNOLOGY a
b
Contrast mode (nonlinear)
B-mode (linear)
0.35
DOI: 10.1038/NNANO.2015.25
Attenuation (dB mm−1)
0.30 0.25
Saline
0.20 0.15 0.10 0.05
pMBs
0.00 0
10
20
30
3.0 × 102 2.5 × 102 2.0 × 102 1.5 × 102 1.0 × 102 5.0 × 101 0.0 × 100 0
5
10
15
d
16
16
14
14
12
12
10
10 8
8
pMBs pNPs
6 4
4
2
2
0
0
0 pulses 1 pulse 3 pulses 10 pulses
7 6 5 4 3 2 1 0 0.0
0 1 2 3 4 5 6 7 8 9 10 Number of pulses
Diameter (µm)
g
f
6
e 8 Concentration (×106 NPs ml−1)
0 pulses 1 pulse 3 pulses 10 pulses
Concentration (×108 NPs ml−1)
Differential volume (µm3)
c 3.5 × 102
Concentration (×106 MBs ml−1)
Frequency (MHz)
10 µm
0.2
0.4 0.6 0.8 Diameter (µm)
1.0
h
500 nm
500 nm
Figure 2 | Characterization of the conversion of pMBs to pNPs. a, Acoustic attenuation measurement of pMBs with a resonance frequency of 4.5 MHz using two transducers (black line, transducer 1, 1.5–12 MHz; grey line, transducer 2, 7–27.5 MHz). Mean ± 1 s.d. (n = 3). pMBs possess a resonance attenuation peak at 4.5 MHz. b, Linear and nonlinear ultrasound properties of pMBs. A tissue-mimicking flow phantom composed of agar and graphite was used with a wall-less vessel in which pMBs or saline were allowed to flow through and imaged with a clinical ultrasound scanner. B-mode (linear) and contrast mode (nonlinear) ultrasound imaging of saline (top) and pMBs (bottom). The phantom produced linear backscatter (shown in the B-mode image), similar to tissue, but did not possess nonlinear properties, indicated by the lack of contrast in the contrast mode image. pMBs generate both linear and nonlinear ultrasound signals. Scale bars, 5 mm. c, Size distribution of pMBs before and after application of conversion ultrasound pulses according to volume distribution. After interaction with conversion ultrasound pulses, the pMB volume population decreased. d, Concentration of pMBs and pNPs before and after application of conversion ultrasound pulses. pMB concentration decreased and pNP concentration increased with an increased number of conversion ultrasound pulses. Mean ± 1 s.d. (n = 3). e, Size distribution of pNPs after conversion ultrasound pulses were applied. f, Light microscopy image of pMBs. g, TEM image of pNPs formed from pMBs after ten ultrasound pulses. h, TEM image of liposome-like structures formed from pNPs after placing pNPs in a vacuum.
bacteriochlorophyll–lipid (BChl-lipid) (Fig. 1a), which has more favourable optical properties in the near-infrared window. The pMBs had a final concentration of (8.4 ± 0.4) × 107 MB ml–1, were acoustically responsive with a resonance attenuation peak at 4.5 MHz (Fig. 2a), and were able to generate both linear and nonlinear responses to ultrasound. A tissue-mimicking flow phantom containing a wall-less vessel was imaged with a clinical ultrasound scanner and filled with saline or pMBs. The phantom produced linear backscatter in the B-mode image, similar to tissue, but did not possess nonlinear properties. When the vessel was filled with pMBs, they were able to produce both linear backscatter and nonlinear responses (Fig. 2b). The pMB peak size distribution was 2
between 2 and 8 µm by volume (Fig. 2c) with 99.9%
In situ conversion of porphyrin microbubbles to nanoparticles for multimodality imaging Elizabeth Huynh1,2, Ben Y. C. Leung3, Brandon L. Helfield2,3, Mojdeh Shakiba1,2, Julie-Anne Gandier4, Cheng S. Jin1,5,6, Emma R. Master4, Brian C. Wilson1,2, David E. Goertz2,3 and Gang Zheng1,2,5,6* Converting nanoparticles or monomeric compounds into larger supramolecular structures by endogenous1,2 or external3,4 stimuli is increasingly popular because these materials are useful for imaging and treating diseases. However, conversion of microstructures to nanostructures is less common. Here, we show the conversion of microbubbles to nanoparticles using low-frequency ultrasound. The microbubble consists of a bacteriochlorophyll–lipid shell around a perfluoropropane gas. The encapsulated gas provides ultrasound imaging contrast and the porphyrins in the shell confer photoacoustic and fluorescent properties. On exposure to ultrasound, the microbubbles burst and form smaller nanoparticles that possess the same optical properties as the original microbubble. We show that this conversion is possible in tumour-bearing mice and could be validated using photoacoustic imaging. With this conversion, our microbubble can potentially be used to bypass the enhanced permeability and retention effect when delivering drugs to tumours. The intrinsic conversion of supramolecular structures involves the activation of shape transitions. Of growing interest in the biomedical field is the conversion of perfluorocarbon nanodroplets as activatable ultrasound contrast agents. These nanometre-sized droplets expand to form micrometre-sized bubbles upon heating5,6. Despite a growing number of investigations utilizing the conversion from nano- to micro-sized structures, to our knowledge there have been no studies exploiting the advantages of a micro-to-nano conversion. The most conventional micrometre-sized structure investigated for biomedical applications is the microbubble, a gas-filled microsphere formed with a biocompatible shell composed of lipids, proteins or polymers. In addition to acting as ultrasound contrast agents, microbubbles have also been studied for their bursting behaviour in response to destructive ultrasound for the measurement of blood flow parameters7 and drug and gene delivery8. The interaction of a lipid microbubble with ultrasound pulses of sufficient amplitude can result in its fragmentation or shrinkage, which may be accompanied by ‘shedding’ of shell material9. These processes lead to the formation of structures that significantly decrease the ultrasound imaging contrast relative to that of the initial microbubble10. Furthermore, the remnants of the microbubbles, the lipids themselves, do not serve any imaging purpose. Recently, we discussed the concept of intrinsically multimodal contrast agents11 and developed multimodal ultrasound contrast agents in which the building blocks of the microbubble had optical and metal chelation properties. We used a porphyrin–lipid to form a shell around a perfluorocarbon gas, forming porphyrin
microbubbles (pMBs). The encapsulated gas provided the ultrasound imaging contrast, and the high density of porphyrins provided photoacoustic and fluorescence contrast12,13. Here we utilize these properties to investigate the response of pMBs to lowfrequency ultrasound. The unique feature of forming the multimodality microbubble with building blocks that have intrinsic optical properties ensures that the responses observed are properties of the pMB after destruction and the formation of porphyrin nanoparticles (pNPs) from the pMB (Fig. 1a,b), and not simply the release of imaging agents from the microbubble as others have investigated14. Although previously reported pMBs were formed using pyropheophorbide–lipid, the pMBs used here were formed using a a
Perfluorocarbon gas
b Transducer Conversion ultrasound
Nanoparticles 5–500 nm
Microbubble 1–10 µm
Figure 1 | Schematics of porphyrin microbubbles (pMBs) and their micro-to-nano conversion. a, pMBs consist of a BChl-lipid shell encapsulating perfluorocarbon gas. b, Conversion of pMBs to porphyrin nanoparticles (pNPs) via sonication with low-frequency, high-duty-cycle ultrasound (conversion ultrasound).
1
Princess Margaret Cancer Center and Techna Institute, University Health Network, Toronto, Ontario M5G 2M9, Canada. 2 Department of Medical Biophysics, University of Toronto, Toronto, Ontario M5G 1L7, Canada. 3 Sunnybrook Health Sciences Center, Toronto, Ontario M4N 3M5, Canada. 4 Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto, Ontario M5S 3E5, Canada. 5 Department of Pharmaceutical Sciences, University of Toronto, Toronto, Ontario M5S 3M2, Canada. 6 Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, Ontario M5G 1L7, Canada. *e-mail: [email protected] NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology
© 2015 Macmillan Publishers Limited. All rights reserved
1
LETTERS
NATURE NANOTECHNOLOGY a
b
Contrast mode (nonlinear)
B-mode (linear)
0.35
DOI: 10.1038/NNANO.2015.25
Attenuation (dB mm−1)
0.30 0.25
Saline
0.20 0.15 0.10 0.05
pMBs
0.00 0
10
20
30
3.0 × 102 2.5 × 102 2.0 × 102 1.5 × 102 1.0 × 102 5.0 × 101 0.0 × 100 0
5
10
15
d
16
16
14
14
12
12
10
10 8
8
pMBs pNPs
6 4
4
2
2
0
0
0 pulses 1 pulse 3 pulses 10 pulses
7 6 5 4 3 2 1 0 0.0
0 1 2 3 4 5 6 7 8 9 10 Number of pulses
Diameter (µm)
g
f
6
e 8 Concentration (×106 NPs ml−1)
0 pulses 1 pulse 3 pulses 10 pulses
Concentration (×108 NPs ml−1)
Differential volume (µm3)
c 3.5 × 102
Concentration (×106 MBs ml−1)
Frequency (MHz)
10 µm
0.2
0.4 0.6 0.8 Diameter (µm)
1.0
h
500 nm
500 nm
Figure 2 | Characterization of the conversion of pMBs to pNPs. a, Acoustic attenuation measurement of pMBs with a resonance frequency of 4.5 MHz using two transducers (black line, transducer 1, 1.5–12 MHz; grey line, transducer 2, 7–27.5 MHz). Mean ± 1 s.d. (n = 3). pMBs possess a resonance attenuation peak at 4.5 MHz. b, Linear and nonlinear ultrasound properties of pMBs. A tissue-mimicking flow phantom composed of agar and graphite was used with a wall-less vessel in which pMBs or saline were allowed to flow through and imaged with a clinical ultrasound scanner. B-mode (linear) and contrast mode (nonlinear) ultrasound imaging of saline (top) and pMBs (bottom). The phantom produced linear backscatter (shown in the B-mode image), similar to tissue, but did not possess nonlinear properties, indicated by the lack of contrast in the contrast mode image. pMBs generate both linear and nonlinear ultrasound signals. Scale bars, 5 mm. c, Size distribution of pMBs before and after application of conversion ultrasound pulses according to volume distribution. After interaction with conversion ultrasound pulses, the pMB volume population decreased. d, Concentration of pMBs and pNPs before and after application of conversion ultrasound pulses. pMB concentration decreased and pNP concentration increased with an increased number of conversion ultrasound pulses. Mean ± 1 s.d. (n = 3). e, Size distribution of pNPs after conversion ultrasound pulses were applied. f, Light microscopy image of pMBs. g, TEM image of pNPs formed from pMBs after ten ultrasound pulses. h, TEM image of liposome-like structures formed from pNPs after placing pNPs in a vacuum.
bacteriochlorophyll–lipid (BChl-lipid) (Fig. 1a), which has more favourable optical properties in the near-infrared window. The pMBs had a final concentration of (8.4 ± 0.4) × 107 MB ml–1, were acoustically responsive with a resonance attenuation peak at 4.5 MHz (Fig. 2a), and were able to generate both linear and nonlinear responses to ultrasound. A tissue-mimicking flow phantom containing a wall-less vessel was imaged with a clinical ultrasound scanner and filled with saline or pMBs. The phantom produced linear backscatter in the B-mode image, similar to tissue, but did not possess nonlinear properties. When the vessel was filled with pMBs, they were able to produce both linear backscatter and nonlinear responses (Fig. 2b). The pMB peak size distribution was 2
between 2 and 8 µm by volume (Fig. 2c) with 99.9%
