Development and optimisation of in vitro sonodynamic therapy for glioblastoma

Abstract

Sonodynamic therapy (SDT) is currently on critical path for glioblastoma therapeutics. SDT is a non-invasive approach utilising focused ultrasound to activate photosensitisers like 5-ALA to impede tumour growth. Unfortunately, the molecular mechanisms underlying the therapeutic functions of SDT remain enigmatic. This is primarily due to the lack of intricately optimised instrumentation capable of modulating SDT delivery to glioma cells in vitro. Consequently, very little information is available on the effects of SDT on glioma stem cells which are key drivers of gliomagenesis and recurrence. To address this, the current study has developed and validated an automated in vitro SDT system to allow the application and mapping of focused ultrasound fields under varied exposure conditions and setup configurations. The study optimizes ultrasound frequency, intensity, plate base material, thermal effect, and the integration of live cells. Indeed, in the presence of 5-ALA, focused ultrasound induces apoptotic cell death in primary patient-derived glioma cells with concurrent upregulation of intracellular reactive oxygen species. Intriguingly, primary glioma stem neurospheres also exhibit remarkably reduced 3D growth upon SDT exposure. Taken together, the study reports an in vitro system for SDT applications on tissue culture-based disease models to potentially benchmark the novel approach to the current standard-of-care.

Figure 1

Illustration of the in vitro experimental setup: water bath, needle hydrophone, plate holder and 96-well plate, transducer and fixture, sensor motor slides for the X, Y, and Z axis, and water heater and pump. See also Supplementary Table S1.

Figure 2

Comparison between ultrasound translucent plate NH mapping and simulation for no well plate, in-well (submerged), and in-well (water level) setups and standard polystyrene plate in-well (submerged setup). 0.4 W/cm² continuous wave. Simulations completed in COMSOL. (A) NH field mapping: no well plate. (B) Field Simulation: no well plate. (C) NH field mapping: in-well (submerged) in an ultrasound translucent plate (190 µm film-base). (D) Field Simulation: in-well (submerged). (E) NH field mapping: in-well (water level). (F) Field simulation: in-well (water level). (G) NH field mapping: in-well (submerged) in a standard polystyrene plate (1 mm base thickness).

Figure 3

Single axis pressure profiles to compare standard polystyrene and ultrasound translucent (MicroClear) plates. (A) Z-axis profiles intercepting the peak pressure point. (B) X-axis profiles intercepting the peak pressure point. (C) X-axis profiles at Z = 0.1 mm, the height of the cells in the well.

Figure 4

Comparison of in-well (submerged) pressure field mappings at increasing emitted intensity, emitted generator amplitude, and target peak pressure of continuous wave focused ultrasound, with and without standardised scale bars (i and ii subfigures, respectively): (Ai and Aii) 5.4 W/cm² (2.5 V, 0.402 MPa), (Bi and Bii) 7.0 W/cm² (2.75 V, 0.458 MPa), (Ci and Cii) 9.9 W/cm² (3.25 V, 0.545 MPa), (Di and Dii) 12.2 W/cm² (3.5 V, 0.603 MPa).

Figure 5

X and Z Pressure profile comparisons illustrating the effects of altered frequency and adjacent structures on pressure distribution. (A) X-axis pressure profiles comparing lateral dampening from below and above the water level of adjacent wells, 0.3 mm and 2.3 mm high, respectively, in an in-well (water level) setup. (B) Z-axis pressure profiles at 0.1, 0.6, 1.1, 1.6, and 2.1 mm from well central axis for an in-well (water level) setup. (C) Comparisons of 666 kHz and 1.2 MHz pressure profiles for 0.4 W/cm² continuous wave focused ultrasound in an in-well (submerged) setup for Z-axis profiles intercepting the peak pressure point. (D) Comparisons of 666 kHz and 1.2 MHz pressure profiles for 0.4 W/cm² continuous wave focused ultrasound in an in-well (submerged) setup for X-axis lines at Z = 0.1 mm, the height of the cells in the well.

Figure 6

COMSOL Simulations of in-well (submerged) and no well plate setups for 666 kHz and 1.2 MHz transducers. Acoustic intensity and total acoustic pressure are shown for each condition: (A) 666 kHz No Well Plate Setup, (B) 666 kHz In-Well (Submerged) Setup, (C) 1.2 MHz No Well Plate Setup, (D) 1.2 MHz In-Well (Submerged) Setup.

Figure 7

SDT induces cell death in primary patient derived glioma cells in vitro. (A) Representative bright field images of GBM22 cells 1.5 h post-treatment with/without 1 mM 5-ALA, with/without FUS (0.4 W/cm², 30 s cumulative sonication, 10% DC, 90 ms pulse length), and both (SDT). Scale bar = 125 μm. (B) Representative fluorescent images only of GBM22 cells, pre-treated with Annexin V-FITC apoptosis marker, 1.5 h post-treatment with/without 1 mM 5-ALA, with/without FUS (0.4 W/cm², 60 s cumulative sonication, 10% DC, 90 ms pulse length), and both (SDT). Scale bar = 125 μm. Also see Supplementary Fig. S6. (C) Representative overlaid bright field and fluorescent images of GBM22 cells, pre-treated with CellROX dye, 1.5 h post-treatment with/without 1 mM 5-ALA, with/without FUS (0.4 W/cm², 30 s cumulative sonication, 10% DC, 90 ms pulse length), and both (SDT). Scale bar = 125 μm. (D) Representative bright field images of GBM120 3D neurospheres. The dissociated GBM120 cells were pre-treated with indicated conditions as in A and allowed to form neurospheres over 21 days. n = 3 biological replicates. Scale bar = 200 μm. (E) Representative bright field images of GBM120 3D neurospheres. The indicated treatments as in A were carried out on pre-formed 3D neurospheres and allowed to grow over a further 21 days. n = 3 biological replicates. Scale bar = 200 μm. (F) Quantification of D across all n = 3 replicates. The diameter of the neurospheres were quantified using ImageJ. The significance of the differences was measured using one-way ANOVA with Dunnett’s multiple comparisons. ****p < 0.0001; ns: not significant. (G) Quantification of the diameter of formed neurospheres on Day 0 and Day 21 across all n = 3 replicates. The diameter of the neurospheres were quantified using ImageJ. The significance of the differences was measured using two-way ANOVA with Bonferroni’s multiple comparisons. ****p < 0.0001; *p < 0.05; ns: not significant.