Miniaturization of hiPSC-derived 3D neural cultures in stirred-tank bioreactors for parallelized preclinical assessment of rAAV

Abstract

Introduction: Engineered 3D models employing human induced pluripotent stem cell (hiPSC) derivatives have the potential to recapitulate the cell diversity and structure found in the human central nervous system (CNS). Therefore, these complex cellular systems offer promising human models to address the safety and potency of advanced therapy medicinal products (ATMPs), such as gene therapies. Specifically, recombinant adeno-associated viruses (rAAVs) are currently considered highly attractive for CNS gene therapy due to their broad tropism, low toxicity, and moderate immunogenicity. To accelerate the clinical translation of rAAVs, in-depth preclinical evaluation of efficacy and safety in a human setting is primordial. The integration of hiPSC-derived CNS models in rAAV development will require, amongst other factors, robust, small-scale, high-throughput culture platforms that can feed the preclinical trials. Methods: Herein, we pioneer the miniaturization and parallelization of a 200 mL stirred-tank bioreactor-based 3D brain cell culture derived from hiPSCs. We demonstrate the applicability of the automated miniaturized Ambr^® 15 Cell Culture system for the maintenance of hiPSC-derived neurospheroids (iNSpheroids), composed of neuronal and glial cells. Critical process parameters were optimized, namely, cell density and agitation mode. Results: Under optimized conditions, stable iNSpheroid cultures were attained in the microbioreactors for at least 15 days, with high cell viability and astrocytic and neuronal phenotype maintenance. This culture setup allowed the parallelization of different rAAVs, in different multiplicity of infections (MOIs), to address rAAV-host interactions at a preclinical scale. The iNSpheroids were exposed to rAAV2- and rAAV9-eGFP in the microbioreactors. Transgene expression was detected 14 days post-transduction, revealing different astrocyte/neuron tropism of the two serotypes. Discussion: We advocate that the iNSpheroid cultures in miniaturized bioreactors are reliable and reproducible screening tools for addressing rAAV transduction and tropism, compatible with preclinical demands.

Keywords: 3D (three dimensional) models; CNS modeling; gene therapy; human induced pluripotent stem cell (hiPSC); miniaturization; recombinant adeno-associated viruses (rAAVs); stem cell bioengineering; stirred-tank bioreactor.

FIGURE 1

Establishing 3D human iNSpheroids cultures in the Ambr® 15 culture system (Ambr15). (A) Schematic representation of the experimental layout, comprising the transition of the iNSpheroid bioprocess to the Ambr15 system, from the 200 mL stirred-tank bioreactor (STB) to the Ambr15, with a parallel control culture in the STB. (B) Representative live/dead images of the iNSpheroids at the day of transfer from the STB to the Ambr15 (Day 0), and at 7 and 15 days after transfer. INSpheroids were stained with fluorescein diacetate (FDA, live cells, green) and propidium iodide (PI, dead cells, red). Scale bar, 100 μm. (C) LDH leakage assay, by measurement of LDH activity in culture supernatant; (D) Cell and (E) aggregate concentration throughout the different timepoints. Data are represented as mean ± SD of three independent experiments. (F) normalized cell death by assessing the ratio between LDH leakage in culture and total cell LDH, calculated by a calibration curve of total LDH in iNSpheroids.

FIGURE 2

Neuron and astrocyte phenotype within iNSpheroids cultured in the Ambr^® 15 culture system (Ambr15). (A) Transcriptional profiles of neurons (i) and astrocytes (ii) after 15 days of culture in the Ambr15.; genes coding for neuronal β3-tubulin (TUBB3), glutamate decarboxylase (GAD1), and synaptophysin (SYN); and genes coding for astrocytic glial fibrillary acidic protein (GFAP), S100 calcium-binding protein B (S100B), glutamate-aspartate transporter 1 (SLC1A3), aquaporin-4 (AQP4), vimentin (VIM) and glutamine synthetase (GLUL) were analyzed by RT-qPCR and the comparative cycle threshold values method (2^−ΔΔCT). Data are represented as mean ± SD of three independent bioreactor experiments. For comparing three or more groups two-way ANOVA test with Šídák’s multiple comparison test was used: *p < 0.05; **p < 0.01, ***p < 0.001. (B) Immunofluorescence detection of MAP2 (green, marker for somatodendritic compartments in mature neurons), (SYN) (magenta, marker for presynaptic vesicles of neurons) and vimentin (VIM) (white, marker for intermediate filaments in astrocytes) in the 0.5, 1.0 and 3.0 Ambr15 conditions introduce: right column correspond to the superimposition of the VIM, MAP2, SYN and DAPI signal. Scale bar, 50 μm; 2 μm for zoom-in insets.

FIGURE 3

rAAV transduction of the iNSpheroids cultured in the Ambr^® 15 culture system (Ambr15). (A) Schematic representation of the experimental layout. INSpheroid were inoculated in the miniaturized system 1 day prior to the exposure to rAAV2-eGFP and rAAV9-eGFP and monitored for 14 days after transduction. (B) Immunofluorescence detection of eGFP (green), and the proteins microtubule-microtubule-associated protein 2 (MAP2, marker of the somatodendritic compartment of mature neurons, white), and vimentin (VIM, marker for intermediate filaments in astrocytes, red). Scale bar, 50 μm and 10 μm for zoom-in insets. Immunofluorescence-based quantification of eGFP, and MAP2 (neurons, white) or eGFP and VIM (astrocytes), using (C) hAdv5, (D) rAAV2 and (E) rAAV9 at the respective MOIs. Data are represented as mean ± SD, of three independent experiments. Statistical significance was evaluated by paired t-test: ns, non-significant; *p < 0.05.

FIGURE 4

Neural cell phenotype modulation upon rAAV transduction in the Ambr^® 15 culture system (Ambr15). (A) Transcriptional profiles of (i) neurons and (ii) astrocytes, after 14 days of culture in the Ambr15; genes coding for neuronal β3-tubulin (TUBB3), glutamate decarboxylase (GAD1), and synaptophysin (SYN); and genes coding for astrocytic glial fibrillary acidic protein (GFAP), S100 calcium-binding protein B (S100B), glutamate-aspartate transporter 1 (SLC1A3), aquaporin-4 (AQP4), vimentin (VIM) and glutamine synthetase (GLUL) were analyzed by RT-qPCR and the comparative cycle threshold values method (2^−ΔΔCT). Data are represented as mean ± SD of three independent bioreactor experiments. For comparing three or more groups two-way ANOVA test with Šídák’s multiple comparison test was used: *p < 0.05; **p < 0.01, ***p < 0.001. (B) Immunofluorescence detection of eGFP (green), GFAP (magenta) and DAPI (blue) in non-transduced control neurospheroids and neurospheroids transduced with the different viral vectors and MOIs. Scale bar, 50 μm; 2 μm for zoom-in insets.