Probabilistic cell seeding and non-autofluorescent 3D-printed structures as scalable approach for multi-level co-culture modeling

Abstract

To model complex biological tissue in vitro, a specific layout for the position and numbers of each cell type is necessary. Establishing such a layout requires manual cell placement in three dimensions (3D) with micrometric precision, which is complicated and time-consuming. Moreover, 3D printed materials used in compartmentalized microfluidic models are opaque or autofluorescent, hindering parallel optical readout and forcing serial characterization methods, such as patch-clamp probing. To address these limitations, we introduce a multi-level co-culture model realized using a parallel cell seeding strategy of human neurons and astrocytes on 3D structures printed with a commercially available non-autofluorescent resin at micrometer resolution. Using a two-step strategy based on probabilistic cell seeding, we demonstrate a human neuronal monoculture that forms networks on the 3D printed structure and can establish cell-projection contacts with an astrocytic-neuronal co-culture seeded on the glass substrate. The transparent and non-autofluorescent printed platform allows fluorescence-based immunocytochemistry and calcium imaging. This approach provides facile multi-level compartmentalization of different cell types and routes for pre-designed cell projection contacts, instrumental in studying complex tissue, such as the human brain.

Keywords: Astrocytes; Calcium imaging; Co-culture models; IP-Visio; Neurons; Two-photon polymerization.

Graphical abstract

Fig. 1

Scalable approach for parallel seeding and characterization of neuronal-astrocytic co-culture models. a) Schematic illustration (side view) of the co-culture model. Neurons (red) and astrocytes (blue) cover the glass substrate surface. The elevated 3D printed structures (in green) feature only seeded neurons, forming connections between the neuronal population on the 3D printed pillar cavities and the co-culture population on the glass substrate surface through the ramps connecting the glass substrate surface and the cavity on top of the pillar. b) Process flow for the realization of the co-culture models. A commercially available low-autofluorescent resin, IP Visio, is printed using a 25 ​× ​0.8 NA objective lens in resin immersion configuration. After development, a PDMS well is placed on the glass substrate. After Matrigel® coating, we first seeded astrocytes with low density, cultured the cells to reach confluency, and then seeded neurons with high density to obtain two levels with different cell populations. The transparency of the substrate and resins allow both upright and inverted microscopy characterizations. c) Fluorescence images (substrate and pillar cavity planes) of the 3D printed structures after two-step cell seeding. The blue and red cell tracker colors visualize the presence of astrocytes and neurons, respectively. The glass substrate plane image shows a co-culture of neurons and astrocytes, while only neurons are present on top of the structures (pillar cavity plane). Scales bars, 100 ​μm. d) Indirect visualization of electrophysiological activity of neurons by calcium imaging recorded with an inverted wide-field microscope. Each recording site is indicated with dashed lines in the pillar cavity. Scales bar, 100 ​μm. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 2

Design, optimization, and characterization of the printed structures. a) SEM image (top view) showing the printed platform composed of two pillar arrays with and without ramps connecting the pillar cavities with the substrate. Scale bar, 200 ​μm. b) 3D model of the structures with nominal dimensions. c) Illustration of post-printing structure shrinkage. The values are based on comparing the dimensions and distances of the 3D model with the measurements in SEM images of printed structures.

Fig. 3

Neurite outgrowth guided by the 3D printed structure. Confocal images (top view and cross-sectional view) of pillar cavities populated with differentiated LUHMES cells (day 8) for pillars a) with and b) without ramp connection, showing how neurites extend down the ramps or are isolated inside the pillar cavities based on the presence or absence of the ramp structure, respectively. Samples stained with anti-TUBB3 (red) and Hoechst/Nuclei (blue). The images on the lower panel show the side view of the areas highlighted with a white dashed line in the images of the upper panel. Scales bars, 100 ​μm. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4

Two-step seeding approach to generate neuronal monoculture inside the pillar cavities and astrocytes-neurons co-culture on the substrate surface. a) Schematic illustration of the two-step seeding method. Astrocytes (in blue) are first introduced with a low seeding density (2500 ​cells/cm²) not to get astrocytes in the pillar cavities. Next, the astrocytes are cultured until they reach the desired level of confluency. Then, neurons (in red) are seeded with high-cell density (150000 ​cells/cm²) populating the substrate, forming a co-culture with the astrocytes. The pillar cavity features neuronal populations isolated from the underlying co-culture or confined contact points in the presence of the ramps. Figures b) to g) are top-view microscopy images composed using the brightfield signal and the fluorescence signal from the cell tracker dyes (top view). b) Astrocytes (blue cells) on the glass substrate 2 ​h after astrocyte seeding. The initial area coverage on the glass substrate is low (around 2%), in line with the low seeding density, and increases over time, reaching around 10% in d). f), g) Neurons (red cells) 4 ​h after neuron seeding. No astrocytes are found on top of the structures at any stage (c, e, g). Scale bars, 100 ​μm. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5

Viability assay and optimization of the co-culture models. a) Viability analysis co-culture of LUHMES cells (differentiation day 9) and human astrocytes in a conventional well plate with no printed structures (control) and in the PDMS well close (<1 ​mm) and far (>2 ​mm) from the printed structure. No significant difference in viability was found, indicating good biocompatibility of the printed structures. b) Viability assay of differentiated LUHMES cells as monoculture and co-culture with human astrocytes. We found the viability of the co-culture to be significantly lower (p«0.005), suggesting a loss of viability due to the high cell density seeding. c) Cell culture protocol with the optimized co-culture protocol (DM+). Supplementing standard (DM) differentiation media with ROCK inhibitor and astrocyte growth factor supplement significantly increases the viability of both neuron monocultures and neuronal-astrocytic co-cultures (p«0.0005). One-way ANOVA test and two-sample t-test were performed to obtain the p values. Error bars show the standard deviations, with the centerline being the median value, and box plots corresponding to 25 and 75 percentiles.

Fig. 6

TH Protein expression. ICC-based TH protein expression analysis in a) differentiated LUHMES cells from monoculture and in co-culture with human astrocytes showing downregulation of TH in the co-culture condition, b) TH expression in LUHMES cells differentiated inside the printed pillar cavities and on the bottom substrate layer showing similar expression, and c) TH expression in LUHMES cells differentiated in co-culture condition with human astrocytes growing on the substrate layer directly on top of the astrocytes or inside the pillar cavities with and without ramp connection to the astrocyte co-culture. One-way ANOVA test and two-sample t-test were performed to obtain the p values. Error bars show the standard deviations, centerline the median value, and box plots 25 and 75 percentiles.

Fig. 7

Calcium imaging of rat neurons and rat astrocytes. a) Calcium imaging of isolated neuron population growing inside the pillar cavities (green calcium dye Calbryte™ 520 AM). Scale bars, 100 μm. b) Characterization of the spontaneous firing rate of rat neuron monocultures on the glass substrate layer and inside the pillar cavities with and with our ramps. c) Characterization of the spontaneous firing rate of rat neurons in co-culture with rat astrocytes, growing on the substrate layer directly on top of the astrocytes or inside the pillar cavities with and without ramp connection to the astrocyte co-culture. P values were calculated using Linear Mixed Models. Red and black marked data points represent two experimental rounds (see Fig. S6b and materials and method section). Error bars show the standard deviations, centerline the median value, and box plots 25 and 75 percentiles. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)