Scalability of spheroid-derived small extracellular vesicles production in stirred systems

Abstract

Introduction: Small extracellular vesicle (sEV)-based therapies have gained widespread interest, but challenges persist to ensure standardization and high-scale production. Implementing upstream processes in a chemically defined media in stirred-tank bioreactors (STBr) is mandatory to closely control the cell environment, and to scale-up production, but it remains a significant challenge for anchorage-dependent cells.

Methods: We used a human β cell line, grown as monolayer or in suspension as spheroid in stirred systems. We assessed the consequences of culturing these cells in 3D with, or without fetal bovine serum in a chemically defined medium, for cell growth, viability and metabolism. We next explored how different scale-up strategies might influence cell and spheroid formation in spinner flask, with the aim to transfer the process in instrumented Ambr®250 STBr. Lastly, we analyzed and characterized sEV production in monolayer, spinner flask and STBr.

Results and discussion: Generation of spheroids in a chemically defined medium allowed the culture of highly viable cells in suspension in stirred systems. Spheroid size depended on the system's volumetric power input (P/V), and maintaining this parameter constant during scale-up proved to be the optimal strategy for standardizing the process. However, transferring the spinner flask (SpF) process to the Ambr®250 STBr at constant P/V modified spheroid size, due to important geometric differences and impeller design. Compared to a monolayer reference process, sEV yield decreased two-fold in SpF, but increased two-fold in STBr. Additionally, a lower expression of the CD63 tetraspanin was observed in sEV produced in both stirred systems, suggesting a reduced release of exosomes compared to ectosomes. This study addresses the main issues encountered in spheroid culture scale-up in stirred systems, rather conducive for the production of ectosomes.

Keywords: bioprocess; bioproduction; bioreactor; extracellular vesicles; scale-up; shear stress; spheroid.

FIGURE 1

Kinetics of 1.4E7 cells cultured in 2D static T-flasks as monolayers (ML, black dot) in medium with FBS or in 3D in suspension in spinner flasks (SpF) in medium with (white square) or without FBS (white crossed dot). (A) Growth curves (continuous line) and viability determined using trypan blue exclusion (dashed line). (B) Growth rate and (C) cell density at 72 h. (D) Intracellular lactate dehydrogenase (LDH) content of 1.4E7 cells grown as ML (black dot) or spheroids (crossed square). (E) Evolution of dead cells in culture, determined from LDH measurements in the cell culture supernatants. (F) Viability was determined by assessing dead cells concentration based on LDH measurements in the cell culture supernatants. Results are expressed as means ± SD. Numbers of biological replicates from independent experiments were as follows: 2D (n = 16), 3D with FBS (n = 7), and 3D without FBS (n = 8). Statistical analyzes were performed using the Kruskal–Wallis test with Dunn’s multiple comparisons test for three-group comparisons, and the Mann-Whitney test for two-group comparisons (**p < 0.01; ****p < 0.0001).

FIGURE 2

Metabolite profile of 1.4E7 cells cultured in 2D static T-flasks as monolayers (ML, black dot) in medium with FBS or in 3D in suspension in spinner flasks (SpF) in medium with (white square) or without FBS (white crossed dot). (A) Glucose, (B) lactate, (C) glutamine and (D) ammonium concentrations measured in the cell culture supernatant. (E) Glucose and (G) glutamine specific consumption rates (q _S ), (F) lactate and (H) ammonium specific production rates (q _P ). (I) Lactate-to-glucose and (J) ammonium-to-glutamine metabolic yields (Y _Lac/Glc and Y _NH4/Gln , respectively). (K) Cell-to-glucose and (L) cell-to-glutamine growth yields (Y _cells/Glc and Y _cells/Gln , respectively). Results are expressed as means ± SD. Numbers of biological replicates from independent experiments were as follows: 2D (n = 11), 3D with FBS (n = 7), and 3D without FBS (n = 7). Statistical analyzes were performed using the Kruskal–Wallis test with Dunn’s multiple comparisons (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001).

FIGURE 3

Characteristics of 1.4E7 spheroids generated in 0.125 L spinner flasks set at 90 rpm in complete medium containing 10% FBS or in production medium in the absence of FBS. (A) Representative images of the 1.4E7 spheroids at 24 h, 48 h, and 72 h. Scale bar = 100 µm. Evolution of the spheroids (B) concentration and (C) size. (D) Evolution of the free cell percentage (continuous line in spheroid) cultures and their viability determined using trypan blue exclusion (dashed line). (E) Number of cells per spheroid across the culture duration. Results are expressed as means ± SD. Numbers of biological replicates from independent experiments were as follows: 3D with FBS (n = 7) and 3D without FBS (n = 8). Statistical analyzes were performed using the Mann-Whitney test (***p < 0.001).

FIGURE 4

Impact of hydrodynamics and scale-up strategies on spheroid formation in spinner flasks (SpF) in production medium. (A) Representative images of the 1.4E7 spheroids obtained after 72 h in a 0.125 L SpF at agitation rates (N) of 60 rpm (2.6 W/m³), 90 rpm (8.5 W/m³), and 120 rpm (20.5 W/m³). Scale bar = 100 µm. (B) Correlation between the P/V and the size (continuous line) or the concentration (dashed line) of 1.4E7 spheroids in a 0.125 L SpF after 72 h of culture. (C) Size of 1.4E7 spheroids after 72 h of culture in a 0.125 L SpF at 90 rpm compared to those generated in a 0.5 L SpF at a constant N (90 rpm, black square), at a constant P/V of 8.6 W/m³ (75 rpm, crossed dot), or at a constant impeller tip speed (υ _tip ) of 0.19 m/s (61 rpm, white square). Each culture was performed at 80% of the SpF working volume. Results are expressed as means ± SD. Numbers of biological replicates from independent experiments were as follows: (A,B), 60 rpm (n = 5), 90 rpm (n = 8), 120 rpm (n = 6); - (C), 0.125 L (n = 5), 0.5 L υ _tip (n = 3), 0.5 L P/V (n = 4), 0.5 L N (n = 3).

FIGURE 5

Kinetics of 1.4E7 cell 3D cultures in 0.5 L spinner flasks (SpF) or stirred-tank bioreactors (STBr) in production medium at a constant P/V of 8.5 W/m³. 1.4E7 cells were cultured in an initial volume of 430 mL in SpF or 245 mL in STBr to form spheroids. After 48 h, the production medium was renewed for a subsequent 24 h period of sEV production in a final volume of 400 mL in SpF and in 200 mL in STBr. (A) Representative images of the 1.4E7 spheroids obtained after 72 h in either SpF or STBr. Scale bar = 100 µm. Evolution of the spheroids (B) concentration and (C) size. (D) Number of cells per spheroid throughout the culture duration. (E) Schematic representation of the flat paddle impeller in the SpF and the dual 30° three-segment pitched-blade impellers in the STBr, along with their corresponding flow regime patterns. (F) Violin plot showing the size of the 1.4E7 spheroids after 72 h of culture in SpF and STBr at 8.5 W/m³, including the corresponding Kolmogorov length scale. (G) Growth curves (continuous line) and viability assessed using trypan blue exclusion (dashed line). (H) Cell density at 48 h prior to medium renewal. (I) Viability at 72 h assessed based on the quantification of dead cells, extrapolated from LDH measurements in cell culture supernatants. (J) Percentage of free cells in cultures. (K) Interfacial area per mL, calculated from spheroid and free cell size and concentration. Results are expressed as means ± SD. Numbers of biological replicates from independent experiments were as follows: SpF (n = 4) and STBr (n = 8). Statistical analyzes were performed using the Mann-Whitney test (*p < 0.05; **p < 0.01; ****p < 0.0001).

FIGURE 6

Characterization of 1.4E7-derived sEV from monolayers (ML), 0.5 L spinner flasks (SpF), or stirred-tank bioreactors (STBr) cultures. (A) Representative cryo-transmission electron microscopy images of 1.4E7-derived sEV from two independent experiments/group. Scale bar = 200 nm. (B) Size distribution of tetraspanin-labeled particles measured using nanoparticle tracking analysis. (C) Purity of sEV batches expressed as the ratio of tetraspanin-positive particles per µg of protein. 1.4E7-derived sEV yields at 24 h expressed as the number of tetraspanin-positive particles per (D) producing cell or (E) interfacial area. Results are expressed as means ± SD. Numbers of biological replicates from independent experiments were as follows: ML (n = 5), SpF (n = 4) and STBr (n = 4). Statistical analyzes were performed using the Mann-Whitney test (*p < 0.05). (F) Representative western blot of sEV-associated proteins, ML (n = 4), SpF (n = 3), STBr (n = 3).

FIGURE 7

Quantification of CD63 ⁺ 1.4E7-derived sEV from monolayers (ML), 0.5 L spinner flasks (SpF), or stirred-tank bioreactors (STBr) cultures and localization of tetraspanins CD63, CD9 and CD81 in 1.4E7 cells. (A) Representative images of sEV captured on CD63 ExoView™ human tetraspanin chips and labeled with fluorescent antibodies directed against CD9 (blue), CD63 (red), and CD81 (green). (B) Percentage of sEV captured on CD63 chips relative to the total number of sEV captured on CD9, CD63, and CD81 chips. (C) Percentage of CD63⁺ sEV (including CD63⁺, CD63+/CD81+, CD63+/CD9+, and CD63+/CD9+/CD81+ labeled) captured on CD9 chips or (D) CD81 chips. Results are expressed as means ± SD and are based on 4 samples per group from two independent experiments. Statistical analyzes were performed using the Mann-Whitney test (*p < 0.05). Confocal imaging of the cellular localization of (E) CD63, (F) CD9 and (G) CD81 in 1.4E7 cells, with (bottom) of without (top) white light. Nuclei were counterstained with Hoechst 33342. Scale bar = 10 µm.