Vascular bursts enhance permeability of tumour blood vessels and improve nanoparticle delivery Yu Matsumoto 1,2‡ , Joseph W. Nichols 3‡ , Kazuko Toh 1 , Takahiro Nomoto 4 , Horacio Cabral 5 , Yutaka Miura 1 , R. James Christie 1† , Naoki Yamada 1 , Tadayoshi Ogura 6 , Mitsunobu R. Kano 7 , Yasuhiro Matsumura 8 , Nobuhiro Nishiyama 4 , Tatsuya Yamasoba 2 , You Han Bae 9 * and Kazunori Kataoka 1,5 * Enhanced permeability in tumours is thought to result from malformed vascular walls with leaky cell-to-cell junctions 1,2 . This assertion is backed by studies using electron microscopy and polymer casts that show incomplete pericyte coverage of tumour vessels and the presence of intercellular gaps 3 . However, this gives the impression that tumour permeability is static amid a chaotic tumour environment. Using intravital confocal laser scanning microscopy 4,5 we show that the per- meability of tumour blood vessels includes a dynamic phenom- enon characterized by vascular bursts followed by brief vigorous outward flow of fluid (named ‘eruptions’) into the tumour interstitial space. We propose that ‘dynamic vents’ form transient openings and closings at these leaky blood vessels. These stochastic eruptions may explain the enhanced extravasation of nanoparticles from the tumour blood vessels, and offer insights into the underlying distribution patterns of an administered drug 6,7 . Studies of nanoparticle distribution in solid tumours have revealed that large particles suffer from limited penetration and are heterogeneously concentrated at perivascular regions 6–8 . To pursue a solution to this distribution pattern, we intravitally imaged the transport of two different sizes (30 and 70 nm) of fluorescent-labelled polymeric nanoparticles at 10 min intervals to evaluate eruption and distribution phenomena in hypovascular human pancreatic BxPC3-GFP tumours implanted in BALB/c nu/ nu mice using intravital confocal laser scanning microscopy (IVCLSM) (Fig. 1, Supplementary Fig. 1 and Supplementary Movies 2–4). These polymeric nanoparticles are well established and described elsewhere 6 . Here we define ‘dynamic vent’ as the time-limited formation of an opening in the vessel wall and ‘erup- tion’ as vigorous extravasation through the vent. The eruptions occur stochastically throughout the 10 h observation time period, suggesting that neither the injection volume nor the excitation lasers were responsible for the event (Fig. 1c). Growth of the obser- vable eruption plumes generally happened early, followed by a slower dispersal phase (Fig. 1a,b). The random temporal distri- bution of the eruptions is evidence that they are not an artefact of the experimental protocols, such as a spike in pressure following injection, which would be more likely to cluster the eruptions closer to the beginning of the experiment, nor a byproduct of exposure to the laser or nanoparticles, which would increase the fre- quency of the eruptions as the experiment progressed and as cell damage accrued. All treated groups showed the presence of eruptions, but 30 nm nanoparticles showed shorter-lived eruption plumes than those created by 70 nm nanoparticles ( p = 0.0017) (Fig. 1d). The size of the eruption plumes created by the 70 nm nanoparticles were slightly larger than those created by 30 nm nanoparticles ( p= 0.018) (Fig. 1e). When the 70 nm nanoparticles were administered, the eruption areas fell into a right-skewed distribution with an average area of 571 μm 2 and standard deviation of 213 μm 2 (Fig. 1e). The eruption persistence was more heavily right-skewed and averaged 148 min (Fig. 1d). Eruption plumes of 30 nm nanoparticles averaged 316 μm 2 in area (Fig. 1e) and 22 min in persistence time (Fig. 1d). Presumably the vessel dynamics are unaffected by nanoparticle size and the differences in eruption plume size and duration are due to the more rapid dis- persion of the smaller nanoparticles. When nanoparticles become sufficiently diffuse, the fluorescence intensity falls below the detection threshold level of our analysis. Background fluorescence generally rose throughout the experiment, and qualitatively seemed to increase more rapidly when 30 nm nanoparticles were used in place of 70 nm nanoparticles, sometimes making eruptions difficult to detect. This again indicates that smaller nanoparticles are able to distribute throughout a tumour more easily. During the initiation phase, however, the eruptions containing the different-sized nanoparticles proceeded similarly. Comparing the rate of plume growth in both sets showed that the smaller nanoparticles travelled faster than larger ones as the eruption began ( p = 0.006) (Fig. 1f). This indicates that during the early phase, nanoparticle transport is primarily convection driven with larger nanoparticles experiencing slightly more transport resistance from the tissue structures. Once fully developed, plume dispersal is dominated by diffusion and is much more limited for larger nano- particles (Fig. 1b,d). Thus, vascular eruptions could allow for deep penetration of nanoparticles, followed by limited dispersion from the initial plume. Eruption distances from GFP-labelled tumour cells were measured for the 70 nm nanoparticles (Fig. 1g and Supplementary Movie 4). 1 Division of Clinical Biotechnology, Center for Disease Biology and Integrative Medicine, Graduate School of Medicine, The University of Tokyo, 113-8655, Japan. 2 Department of Otorhinolaryngology and Head and Neck Surgery, Graduate School of Medicine and Faculty of Medicine, The University of Tokyo, 113-0033, Japan. 3 Department of Bioengineering, University of Utah, Utah 84112, USA. 4 Polymer Chemistry Division, Chemical Resources Laboratory, Tokyo Institute of Technology, Kanagawa 226-8503, Japan. 5 Department of Bioengineering, Graduate School of Engineering, The University of Tokyo, 113-8656, Japan. 6 Nikon Instech Company Limited, Tokyo 108-6290, Japan. 7 Department of Pharmaceutical Biomedicine, Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama University, Okayama 700-8558, Japan. 8 Investigative Treatment Division, Research Center for Innovative Oncology, National Cancer Center Hospital East, Chiba 277-8577, Japan. 9 Department of Pharmaceutics and Pharmaceutical Chemistry, University of Utah, Utah 84112, USA. † Present address: Antibody Discovery and Protein Engineering, MedImmune, LLC, Gaithersburg, Maryland 20878, USA. ‡ These authors contributed equally to this work. *e-mail: you.bae@utah.edu; kataoka@bmw.t.u-tokyo.ac.jp LETTERS PUBLISHED ONLINE: 15 FEBRUARY 2016 | DOI: 10.1038/NNANO.2015.342 NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology 1 © 2016 Macmillan Publishers Limited. All rights reserved