Intervertebral Disc-on-a-Chip as Advanced In Vitro Model for Mechanobiology Research and Drug Testing: A Review and Perspective

Abstract

Discogenic back pain is one of the most diffused musculoskeletal pathologies and a hurdle to a good quality of life for millions of people. Existing therapeutic options are exclusively directed at reducing symptoms, not at targeting the underlying, still poorly understood, degenerative processes. Common intervertebral disc (IVD) disease models still do not fully replicate the course of degenerative IVD disease. Advanced disease models that incorporate mechanical loading are needed to investigate pathological causes and processes, as well as to identify therapeutic targets. Organs-on-chip (OoC) are microfluidic-based devices that aim at recapitulating tissue functions in vitro by introducing key features of the tissue microenvironment (e.g., 3D architecture, soluble signals and mechanical conditioning). In this review we analyze and depict existing OoC platforms used to investigate pathological alterations of IVD cells/tissues and discuss their benefits and limitations. Starting from the consideration that mechanobiology plays a pivotal role in both IVD homeostasis and degeneration, we then focus on OoC settings enabling to recapitulate physiological or aberrant mechanical loading, in conjunction with other relevant features (such as inflammation). Finally, we propose our view on design criteria for IVD-on-a-chip systems, offering a future perspective to model IVD mechanobiology.

Keywords: degenerative disc disease (DDD); intervertebral disc; mechanical loading; mechanobiology; microphysiological device design; organ-on-a-chip.

FIGURE 1

Intervertebral disc (IVD) anatomy and physical stimuli. (A) Vertebral column and the IVD (created with BioRender.com). (B) Structure, composition and main stresses acting on the components of the IVD. An overall vertical compression (C) of the IVD results in an increase of the hydrostatic pressure (P) in the nucleus pulposus (NP), which in turns tends to expand laterally causing an increment in the circumferential tension (T) experienced by the annulus fibrosus (AF), typically composed of 20 lamellae, each constituted by roughly 40 fibers disposed with a 30° angle (Adams and Roughley, 2006). (C) Schematization of the IVD in rest condition (i.e. when no stimuli are applied), following compression leading to a decrease in height and an outward expansion of the AF, and following a flexional slate (e.g. the one arising from someone bending their back). Bending in particular results in a complex stimulation state in the AF with the fibers on one side experiencing compression and on the other side experiencing tension. (D) Stress in human AF and in the NP of a grade 0 and a grade II (cadaveric IVD, as determined by stress profilometry) (McNally and Adams, 1992b; Adams, 2004). Images (B) adapted from Adams and Roughley (2006), (D) from McNally and Adams (1992b) and Adams (2004), reprinted with publisher permissions (SAGE Publications and Wolters Kluwer Health, Inc., respectively).

FIGURE 2

Cytokine-based IVD-on-a-chip models. (A–D) The layout and the possible conceptual adoption of the device presented by Hwang et al. (2017). (A) The device is composed of three chambers designed to culture AF, NP, and nerve or endothelial cells (Chamber 1, 2, and 3 respectively), and also comprises a gradient generator. Chambers 1 and 2, and 3 were coated respectively with fibronectin and with poly-D-lysine to facilitate cellular adhesion; the small channels connecting Chambers 2 and 3 were filled with a collagen-based hydrogel. (B) Conceptual use of the device. Different disease states are mimicked in the model as the cytokine concentration in the chambers increases (as indicated by the purple color). (C) Functional validation of the model. The gradient generator allows a concentration gradient in different culture chambers. (D) Design of an experiment investigating the effects of a gradient of macrophage-derived factors on IVD cells behavior. (E) The device and the experimental timeline of the study performed by Hwang et al. (2020). The authors used a simple device characterized by three channels separated by pillars. The central channel was filled with a collagen hydrogel while the lateral channels were filled with IVD (AF or NP) cells and human endothelial cells respectively. IVD cells were exposed to IL-1β and the effect of the stimulus on IVD cells and on the endothelial compartment in terms of cellular migration was studied. Images (A–D) adapted and reprinted from Hwang et al. (2017), (E) from Hwang et al. (2020), with publisher (API Publishing) permissions and based on http://creativecommons.org/licenses/by/4.0/, respectively.

FIGURE 3

IVD devices providing physical stimuli. (A,B) Layout and the conceptual usage of the mouse disc-on-a-chip with controlled flow proposed by Dai et al., 2019). (A) The device is composed of four identical chambers, with three IVDs located in each chamber. A matrix of micropillars at the inlet of each chamber reduces the shear flow to which the whole mouse IVDs are exposed. (B) The system allows to keep constant levels of nutrients and metabolites resulting in the higher IVD culture times. (C) Layout of the shear stress device proposed by Chou et al. (2016). The top part shows the pre-culture chamber, while the bottom part shows the complete PDMS device. (D,E) The electrical stimulation device proposed by Shin et al. (2019). (D) Description of the experimental procedure from cellular extraction and stimulation to the compartments of the device. (E) Depiction of the different layers composing the electrically active device. Images (A,B) adapted from Xing et al. (2019) (Dai et al., 2019), (C) from Chou et al. (2016), (D,E) from Shin et al. (2019), reprinted with publisher (ACS Publications) permissions and based on http://creativecommons.org/licenses/by/4.

FIGURE 4

Possible technological transfer, introducing mechanics to IVD-on-chip models. (A,B) devices that allow subjection of 3D constructs to defined levels of confined compression or stretching, introduced by Occhetta et al. (2019) and Marsano et al. (2016). (A) Device layout and functioning principle. The device contains two chambers (a culture camber and an actuation chamber), divided by a flexible membrane. When a positive pressure is applied to the actuation chamber, the membrane bends upwards until it reaches the mechanical stop provided by two rows of overhanging pillars in the culture chamber. Regulating the distance between the pillars and the membrane, it is possible to apply a defined compression or stretching level. (B) Compression device and experimental evaluation of the lateral expansion of the device proposed by Occhetta et al. (2019). The T shaped pillars limit the lateral expansion upon compression resulting in an almost ideal confined compression state. (C) Stretching device and experimental evaluation of the lateral expansion of the device introduced by Marsano et al. (2016). Using hexagonal pillars with a wider spacing in between them upon compression the hydrogel in the central chamber expands laterally providing laden cells with a 10% stretching level. (D,E) The device layout and the displacement field of the device proposed by Gizzi et al. (2017) allowing complex displacements stimulation states of a porous membrane. (D) A central porous membrane is connected by four pneumatic chambers (that can be actuated electively to produce complex strain fields) and to four perfusion channels (highlighted in purple from left to right). The final device is obtained by bonding of two halves. (E) Evaluation of the strain field: (E.a) Displacement field induced on the porous membrane (PM) under uniaxial (left), equibiaxial (center) and biaxial 3:5 (right) loading patterns for a maximum pressure p = −500 mbar. (E.b) Color map and isolevel contours of the first invariant of deformation for the corresponding loading patterns. (E.c) Color map of the von Mises stress distribution for the three loading patterns. Images (A–C) adapted from Occhetta et al. (2019), (D,E) from Gizzi et al. (2017), reprinted with publisher (Springer Nature) permissions and based on http://creativecommons.org/licenses/by/4.