1. Introduction
As researchers have attempted to better replicate in vivo cell biology using in vitro systems, 3D cell culture systems have developed rapidly.(Baker & Chen, 2012; Edmondson, Broglie, Adcock, & Yang, 2014) One of the key goals of 3D cell cultures is to reproduce the spatial organization and function of cells in the human body. There are a variety of formats including spheroids, organoids and engineered tissues. In 2D a cell becomes polarized as surface integrins enable it to attach to the substrate. In 3D, the cell surface integrins are distributed more randomly as the cell attaches to the scaffold/matrix in multiple directions. The physiological differences induced by the different attachment modes and the resulting cell/media interactions mean that cellular functions such as proliferation, differentiation, survival and mechanical signaling vary between 2D and 3D cell culture. The details of these variations have been reviewed extensively,(Duval et al., 2017) and many studies have shown variations in metabolism between 2D and 3D systems(Brajša, Trzun, Zlatar, & Jelić, 2016; Verjans, Doijen, Luyten, Landuyt, & Schoofs, 2017) leading to a focus of these systems for early stage drug testing and disease progression.
As experiments migrate from 2D to 3D models, traditional monitoring techniques need to evolve to accommodate the needs of researchers and the specific questions being asked of the 3D models. Currently, it may take up to 30 days to grow a fully differentiated and mature 3D tissue such as skin.(Carlson, Alt-Holland, Egles, & Garlick, 2008; V. Lee et al., 2014) During this time and after it has matured, researchers need to ensure the tissue has grown correctly and is biologically sound for experimental use. However, at this point in time, there are few convenient, bench-top platforms to monitor this process without physically disrupting the model. In 2D culture, bright field microscopy is often used in real‑time to monitor cell growth, but this is generally not possible with non-transparent 3D tissues. To date, the most common techniques used to visualize and validate 3D models are end-point histology and fluorescence microscopy.(Dmitriev, 2017) Even though these techniques are gold standards, they are time consuming and in most cases requires the destruction of the sample for sectioning and staining. While live cell approaches abound, reporter constructs and stains have a limited lifespan and are not targeted to 3D cultures.(Walker-Daniels, 2012) Introducing new standardized methodologies to monitor 3D cultures in real-time would save time, reducing research cost and providing a quality assurance pathway for manufacturing.
Cell visualization is the most intuitive way of identifying cell health, but a range of other methods have been developed to indirectly assess cells in culture. Physical and chemical probes that measure the levels of oxygen, glucose, CO2 and pH of cell culture media can determine cell metabolism and health status.(Bavli et al., 2016; Hossein Mahfouzi, Amoabediny, Doryab, Hamid Safiabadi-Tali, & Ghanei, 2018; Jenkins, Dmitriev, Morten, McDermott, & Papkovsky, 2015; Shaibani, Etayash, Naicker, Kaur, & Thundat, 2016; Weltin et al., 2014; Weyand et al., 2015) Traditional techniques that measure parameters such as glucose and oxygen concentrations take advantage of ion permeable membranes and selective enzymes to determine concentration levels.(Oliver, Toumazou, Cass, & Johnston, 2009) However, a broad range of transducers can also be used to indirectly measure these parameters of cell health.(Modena, Chawla, Misun, & Hierlemann, 2018) Field effect transistors (FET) and light-addressable potentiometric sensors (LAPs) have emerged as useful sensors capable of detecting cellular metabolic products and provide effective spatial resolution within 2D cell culture systems.(Dantism, Takenaga, Wagner, Wagner, & Schöning, 2015; Poghossian, Ingebrandt, Offenhäusser, & Schöning, 2009) Most of these culture characteristics are measured in the media downstream of cells and they give real-time information without destruction of the culture or tissue. These sensors are commercially available and are usually incorporated within bioreactor systems that are used to grow and monitor spheroids or cells in suspension over time.(Alexander, Eggert, & Wiest, 2018; Weyand et al., 2015) This approach assumes that the media is reflecting the behavior inside the tissue. With 3D cultures such as spheroids, this approach may become problematic as spheroids display diffusion limits to nutrients leading to cell death at the center of the culture.(Mehta, Hsiao, Ingram, Luker, & Takayama, 2012) Further, co-culture systems can lead to local variations in behavior(Goers, Freemont, & Polizzi, 2014) which are difficult to detect in the whole system by media sampling.
An alternative approach involves the characterization of electrical properties of cells in culture. Trans‑epithelial electrical resistance (TEER) measurements are widely used in biology and tissue engineering to determine epithelial and endothelial cell health and membrane integrity. These measurements give an indirect, non-invasive assessment of the permeability of cellular tight junctions and ultimately the barrier function of a cellular monolayer or 3D tissues.(Chen, Einspanier, & Schoen, 2015; Schmitz et al., 2018) There are several variables known to affect the final readout including electrode size, measurement temperature, media formulation, time in culture and cell passage number, making it difficult to standardize measurements.(Elbrecht, Long, & Hickman, 2018; Srinivasan et al., 2015) A detailed review of TEERs and a guide to laboratory standardization is presented by Srinivasan and colleagues.(Srinivasan et al., 2015)
Derived from the same principles as TEER, electrical impedance spectroscopy (EIS), is a well-established tool for monitoring 2D cell cultures. EIS measurements are acquired with electrodes positioned beneath the culture and changes in current/voltage through the cells are measured as a frequency sweep is undertaken. This process allows the electrical resistance and capacitance of cells to be monitored.(Benson, Cramer, & Galla, 2013; Groeber et al., 2015) While TEER measures currents at one frequency, sweeping the frequency allows EIS to obtain more detailed information about cellular health status over time. Currently, there are a range of commercially available TEER and EIS devices used for assessing cell barrier function, integrity and cell proliferation in real time. These include OrganoTEER (MIMETAS, Leiden, The Netherlands), xCELLigence Real Time Cell Analysis (ACEA Biosciences, Inc., CA, USA), EVOM2 (World Precision Instruments, FL, USA) and Millicell‑ERS2 (EMD Millipore Corporation Billerica, MA, USA). However, these devices are mainly designed for monitoring adherent cells grown in 2D or transwell based co-culture systems. Electrodes have also been integrated into a range of different 3D cell culture models and microfluidic platforms for EIS monitoring.(V. F. Curto, Ferro, Mariani, Scavetta, & Owens, 2018; S.-M. Lee et al., 2016; Lei, Liu, & Tsang, 2018a). In particular, cell proliferation within spheroids have been successfully monitored using this technique.(Bü, Diener, Frey, Kim, & Hierlemann, 2016; Lei, Lin, & Tsang, 2017; Lei, Liu, & Tsang, 2018b; Pan et al., 2019)
Additionally, EIS has also been incorporated into 3D culture systems for multi-parametric monitoring including bioelectrical sensing and the detection of cell metabolites. Despite the need for the integration of in-line sensing and optical monitoring tools for real time and rapid biological readouts, little research exists where these have been successfully integrated. This is especially evident with highly modular, sensor-based organ-on-chip systems that are difficult to integrate with current microscopy approaches due to design incompatibilities.(Vincenzo F. Curto et al., 2017; Esch, Ueno, Applegate, & Shuler, 2016; Zhang et al., 2017) Multi-parametric approaches incorporating sensing systems, optical readouts and techniques that can provide better spatial resolution are currently needed. This review focuses on the varied methods that have been used, the challenges and limitations associated with different strategies and the factors that need to still be addressed to enable real time, quantifiable results from these systems.