Floor plate differentiation of iPSC-derived dopaminergic neurons in 3D culture
The drug development path—from hit discovery and lead advancement to the clinical space and commercialization— relies on accurate experimental data that are predictive of outcomes in downstream preclinical and clinical settings. Increasing the complexity of disease models with 3D organoids in the early stages of drug discovery is an approach that provides a more relevant cellular context, potentially more accurate physiological data in orthogonal experiments, and improved chances of eliminating false-positive hits during high-throughput screening. To realize these benefits, however, complex disease models must exhibit a similar cellular makeup and function as the organ under investigation. The development of 3D organoids that better represent the corresponding organ complexity enables researchers to better scale experiments, make data-driven decisions, and systemize those processes to speed discovery.
Highlighted here are some of the first steps in generating a functional 3D ventral midbrain neural organoid for modeling Parkinson’s disease (PD). Achieving an organoid model of the brain—with the correct cell types that interact with each other and develop toward mature functionality—depends on many carefully defined parameters. These include physical factors such as those defined by basement membrane matrices, culture conditions, and spatial properties, as well as experimental factors such as how to best leverage the capabilities that advanced gene editing and imaging technologies provide. In the case of neurological diseases such as PD, research is hindered by a lack of access to diseased tissue. To model PD, we have developed a method for differentiating human induced pluripotent stem cells (iPSCs) to midbrain dopaminergic (DA) neurons, while also incorporating CRISPR technology to engineer iPSC-derived organoids such that they harbor either the PD-linked α-synuclein A30P mutation or its unaltered wild-type counterpart.
Comparison of 2D and 3D floor plate specification
Optimizing the conditions for organoid growth and structure is not a straightforward task, but with the right tools and technologies, in vitro disease models can be reproducibly generated at large scale. The Gibco PSC Dopaminergic Neuron Differentiation Kit enables the specification of iPSCs to midbrain floor plate cells. This kit is a set of three components (Floor Plate Specification Supplement, Floor Plate Cell Expansion Base and Supplement, and Dopaminergic Neuron Maturation Supplement) optimized for 2D midbrain floor plate specification, expansion, and maturation. In this discussion we demonstrate the use of this kit with and without an extracellular matrix (ECM) to specify floor plate cells and further differentiate DA neurons in 3D suspension culture. The schematic in Figure 1 describes the parallel floor plate derivation and DA maturation processes for 2D and 3D cultures.
The initial 3D workflow was kept as similar as possible to the optimized 2D schedule of passages and medium changes. In the 3D scenario, human iPSC spheroids in rotating suspension culture were dissociated and seeded into low-attachment 96-well, U-bottom (96U-well) microplates for floor plate specification in static suspension, then changed into expansion medium and maturation medium while in suspension without further passaging. As shown in Figure 1, the time requirements for specification and expansion in 3D were significantly reduced, with equivalent expression of floor plate markers. As expected, iPSCs that underwent 3D neural differentiation—based on their self-organization of progeny cells into organoids with brain-like structures and function—exhibited phenotypes not observed in 2D culture.
Figure 1. Comparison of 2D and 3D floor plate derivation processes using three different media prepared with the components provided in the PSC Dopaminergic Neuron Differentiation Kit. Starting with a human iPSC culture growing in Gibco Essential 8 Medium, we used the Gibco PSC Dopaminergic Neuron Differentiation Kit for the 2D and 3D floor plate derivation process. The top workflow shows 2D floor plate specification of an attached human iPSC culture, followed by multiple passages in expansion medium until day 21. Floor plate cells are then passaged onto poly-D-lysine (pDL) and laminin for differentiation of DA neurons up to day 35. The bottom workflow shows 3D midbrain organoid formation, beginning with human iPSCs in rotating suspension culture. These cells are dissociated and seeded into suspension culture for 3D floor plate specification. The floor plate specification medium is sequentially replaced with floor plate expansion medium and DA maturation medium while in suspension, without further passaging. As compared with 2D cultures, the duration of specification and expansion for 3D cultures can be shortened while maintaining equivalent expression of floor plate markers.
Organoid culture in the absence and presence of extracellular matrices
Attempts to improve the complexity of brain-like organoids are often accompanied by decreases in throughput and reproducibility, both of which impact research results and the scalability of disease models for drug discovery. As described in Figure 2, a comparison of four different 3D organoid culture methods was performed to assess early spheroid morphologies during differentiation (i.e., in rotating suspension versus U-well microplates; and without ECM, encapsulated in ECM, or suspended in ECM). To this end, cultures were grown in the absence of ECM and in the presence of 50% Gibco Geltrex LDEV-Free Reduced Growth Factor Basement Membrane Matrix (encapsulated in ECM) or a dilute ECM (suspended in 2% Geltrex matrix), and evaluated based on growth patterns, morphology, and maturation to understand the effects of different growth environments on the physical properties of the organoids.
Culture conditions that included ECM encapsulation or U-well microplates changed the morphology and complexity of midbrain floor plate organoids. Importantly, static suspension cultures grown in U-well microplates with addition of dilute ECM could reproduce some of the known benefits of ECM encapsulation, such as organoid complexity and faster neuronal maturation, without the difficulty and low throughput of encapsulation methods (method D, Figure 2). Surprisingly, U-well microplates increased the outgrowth of neural epithelia and yielded complex organoids, and the combination of U-well microplates and 2% Geltrex matrix produced a regular shape to the complex organoids (method D).
Figure 2. Organoid formation is boosted by ECM. The numbering of days corresponds to the 3D workflow shown in Figure 1. Floor plate specification and expansion of human iPSCs in rotating suspension (A) without ECM addition or (B) with ECM encapsulation at day 2 of floor plate specification in Gibco Geltrex LDEV-Free Reduced Growth Factor Basement Membrane Matrix diluted 1:1 in DMEM/F-12. (C) Floor plate specification and expansion of human iPSCs in static suspension culture in Thermo Scientific Nunclon Sphera 96U-Well Microplates without ECM; organoids were transferred to rotating culture for maturation. (D) Floor plate specification and expansion of human iPSCs in static suspension culture in U-well microplates using floor plate specification medium supplemented with 2% Geltrex matrix at day 2; all other medium changes matched those in (C). Overall, we found that ECM encapsulation or a U-well microplate changes the morphology and complexity of midbrain floor plate organoids. Scale bar = 1,000 μm.
Confirmation of organoid morphology using high-content imaging and analysis
New instruments for imaging the whole brain, coupled with fluorescent gene reporters and reagents for optical clearing of tissue, can help shed light on neurodegenerative disease states . The Thermo Scientific CellInsight CX7 LZR High-Content Analysis (HCA) Platform provides a powerful combination of fluorescence microscopy, image processing, automated cell measurements, and informatics tools to characterize the physical and biochemical properties of 3D organoids using a broad range of multiwell plate formats. When paired with onstage incubation, robotic plate handling, and the Thermo Scientific HCS Studio Cell Analysis Software, the CellInsight CX7 LZR platform can help to make 3D organoid research more scalable by taking advantage of rapid acquisition of z-stacks from multicellular structures.
To best understand the architectural effects that ECM has on floor plate specification in microwell-grown organoids, organoid culture methods C and D (Figure 2) were compared using HCA. Day 7 organoids grown without ECM or with dilute laminin or Geltrex matrix were cleared using the Invitrogen CytoVista 3D Cell Culture Clearing Reagent, immunostained for N-cadherin and counterstained with DAPI nuclear stain, and then imaged on the CellInsight CX7 LZR platform (Figure 3). N-cadherin (neural cadherin) is a transmembrane protein found to play a role in neural crest development, cell-to-cell adhesion, differentiation, and signaling . N-cadherin antibody and DAPI staining showed that the addition of ECM increased the appearance of rosette-like structures on the surface of day 7 organoids.
Figure 3. ECM addition during floor plate specification supports prominent rosette-like structures in early organoids. After specification by method C (Figure 2), static U-well microplate organoids (no ECM, left panel) were supplemented at day 2 with 200 μg/mL laminin (middle panel) or 2% Gibco Geltrex LDEV-Free Reduced Growth Factor Basement Membrane Matrix (right panel). Organoids were fixed on day 7 and labeled using Invitrogen N-cadherin monoclonal antibody (clone 3B9), in conjunction with Invitrogen Alexa Fluor 488 donkey anti–mouse IgG antibody (green) and DAPI (blue). Shown are maximal intensity projections of 8 μm optical sections, captured on the Thermo Scientific CellInsight CX7 LZR High-Content Analysis Platform.
Modeling Parkinson’s disease with 3D organoid models
PD is characterized by the selective loss of DA neurons in the substantia nigra of the midbrain. This loss of DA neurons is observed in postmortem tissue along with Lewy bodies, which contain aggregates of phosphorylated α-synuclein protein. Lewy body formation in neurons has been described as the causative factor in DA neuron degeneration and the progressive loss of motor function associated with PD [3-5, also see BioProbes 80 "Elucidate the underlying mechanisms of Parkinson's disease and other neurological disorders"]. Although only a minority of patients have a family history of PD, a growing number of genetic risk loci have been linked to sporadic cases and appear to influence susceptibility to environmental triggers.
In vitro models that reproduce the genetic basis of human disease can now be obtained by reprogramming patient cells to create iPSCs. Additionally, advances in gene editing technologies have led to the ability to create iPSC lines with any disease-relevant genome alteration. To apply these promising steps toward a reproducible 3D PD model, PD-associated single-nucleotide polymorphisms (SNPs) were engineered into iPSCs by CRISPR gene editing. Genes were edited in a stable Cas9-expressing iPSC line for high efficiency of cleavage and homology-driven repair. CRISPR editing was followed by one round of isolation by fluorescence-activated cell sorting (FACS), after which single cells showed robust clonal survival and growth when plated onto Gibco rhLaminin-521 in Gibco StemFlex Medium with Gibco RevitaCell Supplement. The precision of the mutations and clonality of the Cas9-expressing mutant and wild-type (WT) cell lines were verified by next-generation sequencing (NGS). One of the engineered SNPs creates the PD-linked A30P mutation in α-synuclein (SNCA). These SNCA mutant and WT control iPSC lines were differentiated toward midbrain organoids in U-well microplates with 2% Geltrex matrix in solution.
Early differentiation of the 3D cultures is marked by morphological change and expression of microtubule-associated protein 2 (MAP2) in neurons at the organoid surface. A portion of these are tyrosine hydroxylase (TH)-positive DA neurons (Figure 4). Comparison of organoid culture methods A, B, C, and D (Figure 2) demonstrates that midbrain organoids formed in free suspension without ECM (method A, Figure 4A) have a simple architecture with a single layer of neuronal cell bodies, whereas encapsulation in Geltrex matrix (method B, Figure 4B) increases neuroepithelial folding and the outgrowth of DA neurons. These effects are partially mimicked by organoid formation in low-attachment U-well microplates without (method C, Figure 4C) or with (method D, Figures 4D and 4E) a suspension of low-concentration Geltrex matrix.
Growth in this dilute ECM suspension (method D), however, outperforms encapsulation (method B) in promoting the maturation of midbrain organoids. The reddish-brown pigment neuromelanin is a byproduct of dopamine synthesis that gives dark coloration to the substantia nigra in vivo . Midbrain organoids that have been encapsulated in ECM are dotted with pigment after many weeks of differentiation (Figure 5); organoids formed in U-well microplates with diffuse Geltrex matrix reach this milestone in about half the time, beginning to show neuromelanin pigmentation within 5 weeks.
Importantly, we generated midbrain organoids from WT and SNCA A30P iPSC lines using U-well microplates with dilute ECM and saw similarly complex epithelial morphology, earlier outgrowth of DA neurons, and evidence of rapid dopamine synthesis (Figures 4 and 5). Given the advantages in ease of use and faster maturation, we have chosen to continue midbrain organoid formation in U-well microplates with diffuse Geltrex matrix for our downstream functional and neurotoxicity studies.
Figure 4. ECM encapsulation and U-well microplates increase organoid complexity and DA neuron yield. (A) Rotating suspension organoid (method A, Figure 2) at day 19 (early maturation): maximal intensity projection of Invitrogen MAP2 antibody and Invitrogen tyrosine hydroxylase (TH) antibody staining and DAPI nuclear staining, as imaged on the Thermo Scientific CellInsight CX7 LZR High-Content Analysis Platform. (B) Encapsulated organoid (method B, Figure 2) at day 19: maximal intensity projection of MAP2 and TH antibody and DAPI staining. (C) Static U-well microplate organoid (method C, Figure 2) at day 19: maximal intensity projection of MAP2 and TH antibody and DAPI staining. (D) 23-day-old static U-well microplate organoid specified in the presence of 2% Gibco Geltrex LDEV-Free Reduced Growth Factor Basement Membrane Matrix at day 2 (method D, Figure 2): maximal intensity projection of TH antibody and DAPI staining. Midbrain organoids formed with ECM or in U-well microplates have a more complex epithelial morphology and earlier outgrowth of DA neurons. (E) 3-week-old static U-well microplate organoid specified in 2% Geltrex matrix: single optical section of TH and Invitrogen FOXA2 antibody antibody and DAPI staining (left) and two images of hematoxylin/eosin–stained organoid sections (middle, right) from a SNCA wild-type (WT) or SNCA A30P mutant CRISPR iPSC line, which reveal thick bands of cells at the outside of the organoids surrounding a dense core of degrading cells. For immunodetection, primary antibodies were detected with Invitrogen Alexa Fluor 488 donkey anti–mouse IgG, Invitrogen Alexa Fluor 594 donkey anti–rabbit IgG, or Invitrogen Alexa Fluor 647 donkey anti–rabbit IgG secondary antibody.
Figure 5. ECM promotes maturation of floor plate organoids. (A) Encapsulated organoid (method B, Figure 2) was imaged at day 73 without stains or dyes. The reddish-brown color suggests the presence of neuromelanin, a pigment that is a byproduct of dopamine synthesis . (B,C) 23-day-old static U-well microplate organoid specified in the presence of 2% Gibco Geltrex LDEV-Free Reduced Growth Factor Basement Membrane Matrix at day 2 (method D, Figure 2) (B) from a SNCA wild-type (WT) CRISPR iPSC line or (C) from a SNCA A30P mutant CRISPR iPSC line. Earlier detection of neuromelanin in U-well microplate organoids with dilute Geltrex matrix suggests more rapid maturation of DA neurons.
Detecting action potentials in neural organoids using multielectrode arrays
Multielectrode arrays (MEAs) measure extracellular field potentials and are useful for studying neural circuit connectivity in monolayer cultures or organoids . To detect spontaneous network activity in our midbrain model, we allowed single organoids to attach to an MEA with 16 electrodes, monitoring a surface area of 1.2 mm2. Midbrain organoids generated by our method displayed spontaneous coordinated activity across the recording area in as little as 5 weeks of total differentiation time (Figure 6). In vivo, action potentials in the DA neurons of the substantia nigra are inhibited by excess dopamine . Dopamine addition silenced the coordinated bursts we detected in midbrain organoids, confirming that these action potentials are driven by DA neurons.
In short, we have generated midbrain organoids from iPSCs by modifying the 2D protocol of the PSC Dopaminergic Neuron Differentiation Kit. This user-friendly method hastens functional maturation of DA neurons and makes promising steps toward a reproducible disease model for PD.
Figure 6. Floor plate organoids mature functionally. (A) Brightfield images show a static U-well microplate organoid specified in the presence of 2% Gibco Geltrex LDEV-Free Reduced Growth Factor Basement Membrane Matrix at day 2 (method D, Figure 2), cultured in suspension until day 21, and then allowed to attach in culture on a poly-D-lysine (pDL) and laminin–coated multielectrode array (MEA) for 14 days. (B) Raster plots of MEA recordings were derived from the plated U-well microplate organoid in (A). Each plot shows 300 sec of activity, first in maturation medium, second after addition of 100 μM dopamine, and third following washout of the dopamine. Vertical pink bars indicate detected network bursts. Midbrain organoids formed in U-well microplates with dilute Geltrex matrix can produce coordinated DA neuron activity in as little as 5 weeks.
Advancing neuroscience research with 3D cell models
3D in vitro models of the brain and its disease states have become the focus of neuroscience research, due in part to the disappointing responsiveness of 2D culture models, but also to recent technological improvements that allow neuronal structures and functions to be observed in dense cell assemblies. Here we show our initial progress in generating organoids that show relatively rapid differentiation of DA neurons and capture several developmental events of the substantia nigra. We have engineered PD-relevant mutations into human iPSCs with CRISPR technology and are applying the derived midbrain organoids to model critical PD events. Ultimately, the use of iPSCs to build 3D brain models promises to advance our understanding of human disease mechanisms that are important to the development of therapeutic drugs and the implementation of cell and gene therapies.
- Zhu X, Xia Y, Wang X et al. (2017) Neurosci Bull 33:95–102. PMID 27535148
- Derycke LD, Bracke ME (2004) Int J Dev Biol 48:463–476. PMID 15349821
- Irvine GB, El-Agnaf OM, Shankar GM et al. (2008) Mol Med 14:451–464. PMID 18368143
- Braak H, Del Tredici K (2016) Cold Spring Harb Perspect Biol 8:a023630. PMID 27580631
- Moussaud S, Jones DR, Moussaud-Lamodière EL et al. (2014) Mol Neurodegener 9:43. PMID 25352339
- Jo J, Xiao Y, Sun AX et al. (2016) Cell Stem Cell 19:248–257. PMID 27476966
- Spira ME, Hai A (2013) Nat Nanotechnol 8:83–94. PMID 23380931
- Lacey MG, Mercuri NB, North RA (1987) J Physiol 392:397–416. PMID 2451725
For Research Use Only. Not for use in diagnostic procedures.