A human brain microphysiological system derived from induced pluripotent stem cells to study neurological diseases and toxicity

Abstract

Human in vitro models of brain neurophysiology are needed to investigate molecular and cellular mechanisms associated with neurological disorders and neurotoxicity. We have developed a reproducible iPSC-derived human 3D brain microphysiological system (BMPS), comprised of differentiated mature neurons and glial cells (astrocytes and oligodendrocytes) that reproduce neuronal-glial interactions and connectivity. BMPS mature over eight weeks and show the critical elements of neuronal function: synaptogenesis and neuron-to-neuron (e.g., spontaneous electric field potentials) and neuronal-glial interactions (e.g., myelination), which mimic the microenvironment of the central nervous system, rarely seen in vitro before. The BMPS shows 40% overall myelination after 8 weeks of differentiation. Myelin was observed by immunohistochemistry and confirmed by confocal microscopy 3D reconstruction and electron microscopy. These findings are of particular relevance since myelin is crucial for proper neuronal function and development. The ability to assess oligodendroglial function and mechanisms associated with myelination in this BMPS model provide an excellent tool for future studies of neurological disorders such as multiple sclerosis and other demyelinating diseases. The BMPS provides a suitable and reliable model to investigate neuron-neuroglia function as well as pathogenic mechanisms in neurotoxicology.

Keywords: 3D culture; CNS; brain; microphysiological system; myelination.

Fig. 1:. Characterization of BMPS during differentiation I

(A) Diagram of the differentiation protocol. (B) Co-immunostaining of MAP2⁺ neurons with the maturation marker Nestin at 2, 4 and 8 weeks. Scale bar: 20 μm. (C) Size of aggregates measured during the 3D neuronal differentiation. Aggregates (n = 20) from three independent experiments were randomly selected per time point for obtaining pictures and measuring size using SPOT software 5.0. The blue bar (NPC med. 2d) represents the size of aggregates cultured in 3D for two days in NPC medium while from day 1 the cells were cultured in differentiation medium. Results are expressed as mean ± SD. Cells were kept for two days in NPC medium, indicated as “NPC med. 2 d”. Inserted phase contrast image shows several aggregates at 13 days of differentiation with an average diameter of 310 μm. (D) BMPS mRNA and miRNA expression of different markers during differentiation (b: right y-axis relative quantification of S100B and d: right y-axis relative quantification of GABRA1, TH; e: FOXA2, FOXO4 and MBP). (E) Flow cytometry population analysis of BMPS at different stages of differentiation.

Fig. 2:. Characterization of BMPS during differentiation II

(A) Comparison of expression of neuronal and glial markers at 2 and 8 weeks. At 2 weeks, oligodendrocytes (O1, CNPase and NOGOA) were identified without a preferential localization (a, b, e, i, j), later they resemble human oligodendrocytes and localize in close proximity with axons (c, d, h, k, l). At 2 weeks, there are few MAP2⁺ cells and they lack an identifiable neuronal shape (i,j) whereas at 8 weeks the MAP2⁺ cells acquire a well-defined dendritic network (k, l). The number of astrocytes (GFAP⁺) and density of the astroglial network increased with time of differentiation (e, f vs. g, h). (B) Co-immunostaining of neurons (MAP2) with cell-division marker Ki67 showed that some cells are still dividing at 8 weeks (a, b). There was also a small degree of apoptosis demonstrated by positive staining with CASP3 at 8 weeks (c). CASP 3-positive nuclei did not co-localize with mature neurons (d). (C) Ultrastructure analysis by electron microscopy at 8 weeks showed evidence of cell-to-cell junctions demonstrating functional interactions between the cells (arrows, a, b). Nuclear variation was confirmed by the presence of a few apoptotic nuclei (c) compared to normal healthy nuclei (d). NF, Neurofilament-heavy-chain; MAP2, microtubule-associated-protein 2; GFAP, glial-fibrillary-acidic protein; O1, Olig1 ; CNPase, 2’,3’-cyclic- nucleotide-3’-phosphodiesterase; CASP3, caspase-3.

Fig. 3:. Morphologic characterization of mature human BMPS

(A) At 8 weeks, neuronal populations exhibited a diversity of neurotransmitter identities as shown by identification of dopaminergic TH⁺ (a, b), glutamatergic VGLUT1⁺ (c, d) and GABAergic calbindin⁺ (e, f) neurons. Neurons disclosed characteristic axons (NF) and synapsins (SYN) (g, h). (B) Two distinctive glial populations were identified in close interaction with neuronal populations, GFAP⁺ astroglia, and CNPase⁺, O1⁺, NOGOA⁺ oligodendroglia. O1⁺ oligodendrocytes were closely associated with axonal processes (NF) (a, b), CNPase⁺ oligodendroglia appeared mixed among GFAP⁺ astroglia (c, d) and exhibited the characteristic multipolar glial processes, which extended from the perykaria (e, f). NOGOA⁺ cells were associated with MAP⁺ neurons (g, h). (C) Example of custom algorithm created using the Cellomics Target Activation image-analysis software package to study astrocytes and oligodendrocytes (a, b, c, d). Quantification of cell populations as a percentage of the total nuclei count of GFAP, Calbindin, CNPase, NOGOA, Olig1, Ki67 and Caspase positive cells at 8 weeks (e). Randomly selected pictures from three experiments were acquired. The algorithm associated an astrocyte or oligodendrocyte cell body with a nucleus and quantified with respect to the total nucleus count (D) Co-expression of mature oligodendroglia markers (MBP and O2). (E) Expression of neuronal markers (VGLUT, TUJ1, SYN).

Fig. 4:. Myelination in BMPS

(A) Co-immunostaining of neurons (NF) and the myelin marker MBP at 2, 4 and 8 weeks of differentiation (a, b, c, respectively) showed progressive increase of MBP⁺ cells in association with axonal processes. (B) 3D-reconstruction based on confocal z-stacks at 8 weeks demonstrating a “wrapping” myelinating process, which resembles the myelination of axons in human CNS. (C) MBP⁺ oligodendrocytes (green) issued processes in close association with axons (red) and seemed to enwrap them at 8 weeks (a, b, c). Myelination calculated as the mean percentage MBP positive oligodendrocyte processes coverage of NF-positive axons (a, b, c) at 2, 4 and 8 weeks in at least 18 microscopy fields from at least 3 individual BMPS in 2 independent experiments showed significant increase of myelination observed with time of differentiation (d) (***, p < 0.001; Kruskal-Wallis test). (D) Electron microscopy analysis of BMPS at 4 (a) and 8 (b) weeks of differentiation identified morphology of axonal structures and cells, which appeared to be oligodendrocytes. Myelinating-like processes, which closely resembled cross-sections of myelinated axons of the CNS, were identified at 8 weeks of differentiation.

Fig. 5:. Electrical activity as a neuronal function of the BMPS

Cells were cultured in 3D for 8 weeks and then cultured in 48-well MEA plates for 2 more weeks. (A) Heat map recordings from a 48-well plate. Illustration of an active well showing (B) spike morphology and (C) spike activity. (D and E) Phase-contrast imaging of the BMPS on MEAs, electrode diameter is 40–50 μm and inter-electrode space is 350 μm. (F) Activity pattern recordings over 0.05 spikes/sec of the electrode over 10 min.