Tunable osteoarthritis in vitro model based on equine osteochondral plugs

INTRODUCTION: Mechanical overloading of synovial joints
is known to cause cartilage lesions that can lead to chronic diseases such as
osteoarthritis (OA) if left untreated. Several in vitro OA models are
being developed on osteochondral plugs with numerous mechanical loading
protocols, culture durations and culture medias. However, high sample-to-sample
variability is a reoccurring limitation in published work. Thus, the purpose of
this study was to investigate whether employing nondestructive assessment of
cartilage electromechanical properties to guide plug extraction could reduce
sample-to-sample variability and lead to a robust and tunable OA model.

METHODS: Equine femoral condyle blocks (N=4) were harvested and submitted
to electromechanical quantitative parameter (QP) assessment with a hand-held
probe (Benchtop Arthro-BST, Biomomentum) used on 40+ positions of the surface. (Figure
1A). Bone-cartilage plugs (D=4.8mm) were then extracted from each block
with a drill press under constant irrigation (Figure 1B). Based on internal
quality scores, 17 plugs were retained and distributed in 3 test groups: Control (n=5), Injured (n=6), and Injured+TNFα (n=6). Plugs were fitted
in custom 2-chamber systems (Figure 1C) for 6 days incubation at 37oC
and 5% CO₂. Both chambers were supplied with DMEM/F12 media with 0.1%BSA
and 50ng/mL Dexamethasone, and TNFα (100 ng/mL) was only added to the cartilage
chamber of the Injured+TNFα samples. Culture media was sampled at day-3
and day-6 to quantify levels of glycosaminoglycans (GAG) released. To determine
the percentage change in cartilage mechanical properties, a sequence of
unconfined compression and rotational friction-on-glass was performed
aseptically at day-0 prior to injury and again at 6-days post-injury. Mechanical
injury was inflicted on samples (n=12) at day-0 via a compressive deformation of
50% strain at 100%/s rate, except for the control group. Stereomicroscope
images were used to measure the cartilage thickness and record macroscopic
alterations to the cartilage surface. Once the final mechanical tests were completed,
cartilage samples were taken to determine cell viability and GAG tissue content,
and to generate Safranin-O/Fast Green-stained paraffin sections for GAG. ANOVA
with Tukey Post-Hoc test was used for statistics.

RESULTS:
The electromechanical quantitative parameter (QP) of cartilage confirmed a high variability
between osteochondral blocks harvested and between plugs from the same block (Figure 1). Images of the cartilage surfaces taken immediately after injury revealed
that the severity of cracks was strongly correlated to the plugs’ QP values (Figure 2) with deeper cracks apparent in plugs with High-QP compared to
superficial ones for plugs with Low-QP. Percentage changes in cartilage
mechanical properties in unconfined compression and friction were also
accentuated in High-QP plugs compared to Low-QP plugs. However, the most drastic
changes were observed for Permeability (Figure 3A) which increased
by 67-80% in Low-QP plugs but increased by 221-300% in High-QP plugs. High-QP
plugs were also more responsive to the effect of TNFα that lead to a 2-fold
higher in GAG release (Figure 3B) compared to injury alone and translated in
a visible gradient of GAG depletion in histology (Figure 3CD).

DISCUSSION:
As expected various levels of alterations were observed for 1) the mechanical
and tribological properties, 2) the GAG levels released in the media and 3) the
GAG matrix distribution visible in histology sections. Moreover, these
alterations were not only dependent on the mechanical injury and the presence
of TNFα in the culture media but also strongly dependent on the electromechanical quantitative parameter (QP) measure before the
cartilage plug extraction.
Consequently, the addition of the QP measure to our protocol successfully helped
lower sample-to-sample variability. Future studies will investigate the
effect of longer culture duration, various levels of mechanical injuries & the
use of other pro-inflammatory cytokines.

SIGNIFICANCE: OA is a debilitating disease with a slow progression where the
patient’ diseased joints advance through different stages of cartilage
degradation. A tunable OA model would help scientists to develop therapeutics
that could target specific stages of cartilage degradation.