Tunable Superhydrophobic Microwell Array Devices for High-Throughput Culture of 3D Cancer Models

Introduction:: Manipulation of the biological-material interface is performed in mammalian cell culture to control cellular adhesion to the substrate or to other cells [1]. Cancer experimentation is typically performed using a 2D tissue culture model which are easy to use, cost effective, and uniform, but greatly reduce cell:cell interactions, making them less physiologically relevant. 3D cancer models (spheroids) offer increased physiological relevance by allowing for increased cell-to-cell signaling and diffusion of nutrients through the organized layers of cells to study the primary tumor [2]. An example of a low attachment plate to grow spheroids is the Aggrewell^TM800 (Stem Cell Technologies). This grows many spheroids but depends on the user coating with an Anti-Adherence rinsing solution which can be non-uniform. The primary tumor is not the only instance of cancer cells aggregated into 3D clusters that requires further study. During the metastatic cascade, cancer cells will travel through the bloodstream as circulating tumor cells (CTCs). Cancer cells can cluster in circulation, conferring higher shear stress, treatment resistance, and increased metastatic potential [3]. In vitro methods of modeling CTC clusters consist in growing cells until they are a confluent monolayer, adding trypsin, and then gently pipetting to not disrupt floating cell aggregates. Drawbacks of this method include: lack of control over the size of aggregates and weakening of cell-cell junctions during trypsinization [4]. In this study, we successfully designed, fabricated, and validated an easily tunable superhydrophobic array device to culture a wide range of 3D cancer models, from spheroids to CTC clusters.

Materials and Methods:: Two superhydrophobic array devices (SHArD) were designed: one to grow cancer spheroids (SHArD-S) and the other to culture smaller cell aggregates which mimic the clustered CTCs found in patient blood samples (SHArD-C). The SHArD-S microwell dimensions were chosen to match those of the Aggrewell^TM800 plates in order to allow for direct comparison of spheroids between the two culturing methods. The SHArD-C was cultures smaller cells aggregates, with a wall-to-wall distance of 100µm and wall depth of 75µm. Both SHArDs were fabricated using a multi-step additive process as outlined in Figure 1A. The only step in the microfabrication protocol that is changes is the SU8 lithography, making the device alteration for different 3D model culture straight forward. Microfabrication was done at the VINSE and the CNMS Oak Ridge National Laboratories cleanrooms. The colorectal cancer cell line HCT116 was used to test the device. For spheroid formation, cells were cultured for 6 doubling periods with in the at 2.5k, 5k, and 7.5k cells per microwell in the SHArD-S and the Aggrewell^TM800. Spheroids were exposed to physiological fluid shear stress (FSS) with a cone-and-plate viscometer and observed via confocal microscopy by staining with DAPI to quantify compactness. For CTC cluster formation, cells were incubated for 3 doubling periods in the SHArD-C at 3 (small), 5 (medium), and 7 (large) cells per microwell. The monolayer aggregate protocol was used as a control. Clusters were exposed to FSS as described above and confocal microscopy was used to assess E-Cadherin expression.

Results, Conclusions, and Discussions:: Superhydrophobic surfaces have two essential components: a hydrophobic coating and nano-scale surface roughness. The ZnO nanoparticle layer and the ZnO nanorod layer were uniform and easily reproducible (Figure 1 B&C). The average nanorod length was measured to be 469nm ± 171nm and randomly oriented. In both SHArD designs, SU-8 was chosen to create the microwell walls as it is known to be a biocompatible material and a highly tunable method for easily creating high aspect ratio micro-scale structures. SU-8 structures for the SHArD-S and SHArD-C can be seen in Figure 1 D&E. Thin and even coating of the device with the non-wetting polymer was confirmed via SEM (Figure 1F). The static WCA of the nanorods coated in the non-sticky polymer was 167°. The dynamic WCA were 165.2° and 168.6° for the advancing WCA and receding WCA, respectively (Figure 1G).

It was observed that spheroids cultured in the SHArD-S were less round and more compact than those cultured in the Aggrewell^TM800, more closely resembling tumors in the body (Figure 1H&I). Spheroids grown in the Aggrewell^TM800 displayed a large decrease in volume after FSS, while those grown in the SHArD-S showed no significant change in cross-sectional area (Figure 1J&K). HCT116 spheroids grown in the SHArD-S had, on average, 30% higher compactness than those cultured in the Aggrewell^TM800 in each of the plating densities (Figure 1L).

When cancer cell clusters were cultured in the SHArD-C, cluster size was easily controlled by varying plating density of the cells within each microwell (Figure 1M). The coefficient of variation in cluster size was larger in clusters grown in the SHArD-C in comparison with the monolayer control protocol (Figure 1N). Additionally, clusters grown in the SHArD-C did not disaggregate when exposed to FSS in a cone-and-plate viscometer (Figure 1O). Clusters cultured using the monolayer protocol showed significantly reduced E-cadherin expression, an important cell junction protein (Figure 1P). We thus provide a novel method to reliably produce highly tunable 3D cancer models from CTC clusters to cancer spheroids, allowing us to simulate cell:cell interactions throughout the metastatic cascade.


Acknowledgements (Optional): :

References (Optional): : [1] M. Ferrari, F. Cirisano, and M. C. Morán, “Mammalian Cell Behavior on Hydrophobic Substrates: Influence of Surface Properties,” Colloids and Interfaces, vol. 3, no. 2, Art. no. 2, Jun. 2019, doi: 10.3390/colloids3020048.

[2] K. A. Grayson, N. Jyotsana, N. Ortiz-Otero, and M. R. King, “Overcoming TRAIL-resistance by sensitizing prostate cancer 3D spheroids with taxanes,” PLoS One, vol. 16, no. 3, p. e0246733, 2021, doi: 10.1371/journal.pone.0246733.

[3] E. Schuster, R. Taftaf, C. Reduzzi, M. K. Albert, I. Romero-Calvo, and H. Liu, “Better together: circulating tumor cell clustering in metastatic cancer,” Trends in Cancer, vol. 7, no. 11, pp. 1020–1032, Nov. 2021, doi: 10.1016/j.trecan.2021.07.001.

[4] H.-L. Huang et al., “Trypsin-induced proteome alteration during cell subculture in mammalian cells,” Journal of Biomedical Science, vol. 17, no. 1, p. 36, May 2010, doi: 10.1186/1423-0127-17-36.