Effect of Shock-Induced Cavitation Bubble Collapse on the damage in the Simulated Perineuronal Net of the Brain

The purpose of this study is to conduct modeling and simulation to understand the effect of shock-induced mechanical loading, in the form of cavitation bubble collapse, on damage to the brain’s perineuronal nets (PNNs). It is known that high-energy implosion due to cavitation collapse is responsible for corrosion or surface damage in many mechanical devices. In this case, cavitation refers to the bubble created by pressure drop. The presence of a similar damage mechanism in biophysical systems has long being suspected but not well-explored. In this paper, we use reactive molecular dynamics (MD) to simulate the scenario of a shock wave induced cavitation collapse within the perineuronal net (PNN), which is the near-neuron domain of a brain’s extracellular matrix (ECM). Our model is focused on the damage in hyaluronan (HA), which is the main structural component of PNN. We have investigated the roles of cavitation bubble location, shockwave intensity and the size of a cavitation bubble on the structural evolution of PNN. Simulation results show that the localized supersonic water hammer created by an asymmetrical bubble collapse may break the hyaluronan. As such, the current study advances current knowledge and understanding of the connection between PNN damage and neurodegenerative disorders.

from prior literatures that the physics of shock wave formed by reflection is same as conventional piston-driven shock wave formation. The shock velocity generated by our system agrees with experimental results and other simulation work.

Appendix 2: Cavitation Formation and Nucleation Site
In this section, we have investigated the distance between cavitation-induced bubble and hyaluronan to determine the minimum distance at which hyaluronan exhibit stable configuration. The simulation box containing water, hyaluronan, and ions are slowly stretched up to 23% in volume, which enables nano cavitation to form.
This set of simulation is aimed to obtain the characteristic distance between nanobubble and hyaluronan ( Figure A1). The initial size of the simulation box is 16 nm, 16 nm, and 25 nm containing a total of 599,205 atoms with a 12 fold hyaluronan (~500 atoms), which is the same equilibrated box in the main article. We stretched the entire simulated domain, including all of the atoms, for 10.78% in volume (3.47% in length) within 10 ps. Then we continuued the simulation for another 40 ps with fixed doamin size. During the entire simulation, the temperature is fixed at 300K with the damping constant of 12.5 fs. Figure A1 shows the results. It can be observed that when the box started to stretch, pressure starts to go down below atmospheric (i.e., building up of negative pressure or tensile pressure). Once the negative pressure is reached the critical pressure for bubble nucleation (in our case it is ~ -170 MPa, which is close to the experimental result from 7 ), bubbles start to form, and the pressure stops to build further even when the box is continued to expand. After the box stop stretching (10 ps), the pressure start to rise up. It can also be observed that the near-spherical bubbles are formed from many nucleation sites. Most bubbles, with an average size of ~2 nm, are seen to nucleate about ~3-5 nm away from the hyaluronan-liquid interface. The nearest bubble nucleation site we captured has the distance ~0.5 nm from bubble surface to the hyaluronan.
At the end of the simulation (50 ps), all the cavitation-bubble tended to merge into one, which is the lowest energy state at given constraint.
By observing the interaction between the newly nucleated bubble and hyaluronan, it is clear that hydrophilic hyaluronan tends to repel the bubble away. The bubbles are only stable at a minimum distance of 0.5 nm from the hyaluronan implying the presence of at least one or two layers of water between the bubble surface and the hyaluronan. The minimum distance of 0.5 nm is later used in the main article.

Appendix 3: Visual comparison of the three sets of simulation (with rotated HA as initial condition)
In Fig. A2, results of all three sets of simulations have been shown. The threshold of HA breaking are the same in all three sets. Since the HA of the three sets are rotated by 90 degrees each, the broken segment of the HA are at different location but following the same pattern of breaking glyosidic bond and breaking glucose structure (mostly for 7.2 km/s cases).