Characteristics of Distal Femoral Articular Cartilage in Six Weeks Post-Traumatic Osteoarthritis by a Sub-critical Impact

Abstract

Traumatized knee greatly contributes to Osteoarthritis of the knee in young adults. To intervene effectively before the onset of severe structural disruption, detection of the disease at the early onset is crucial. In this study, we put together the findings for the detection of osteoarthritis from the femoral knee joint cartilage of the rabbit at six weeks post-trauma. Articular cartilage samples are taken from the impacted and non-impacted joints at zero week (serving as the control group) and at six weeks post-trauma by minimal force. The samples were imaged using μMRI at 11.7 μm/pixel and Polarized Light Microscopy (PLM) at 1 μm/pixel. In addition, an Inductively Coupled Plasma - Optical Emission Spectrometry (ICP-OES) analysis was performed using the adjacent cartilage samples. The outcomes of this study demonstrate an increase in T2 values in 6weeks samples compared to the 0week samples by μMRI technique, indicating a general increase of tissue hydration within cartilage. PLM detects a decrease in the average thickness of the SZ in the post-traumatic osteoarthritis (PTOA) samples, significant in the impacted femurs. There was an average increasing trend of maximum retardation in the tide mark in comparison to the reported calcium concentration (mg/L) in impacted samples suggesting a possible rise in mineralization in the 6weeks samples. Qualitatively, physical observation of the joint after 6weeks showed signs of reddening in the anterior femur suggesting the disease process is a localized phenomenon. Through microscopic imaging, we are able to detect these changes at 6weeks post-trauma qualitatively and quantitatively.

Graphical Abstract

μMRI imaging of bulk 6-week femoral articular cartilage post trauma to the cartilage shows an increase in T2 relaxation values (p=0.06), an indication of hydration in the tissue. The superficial zones (SZ) and Transitional zones (TZ) were the major contributors to cartilage hydration. Polarized light microscopic imaging of the cartilage histological sections showed reduced SZ cartilage thickness indicating slight tissue disruption at the cartilage surface.

1 |. INTRODUCTION

Post-traumatic osteoarthritis (PTOA) is a type of osteoarthritis that arises after traumatic injury including fractures and tears, which accounts for approximately 12% of all symptomatic osteoarthritis (OA) cases in the United States^1,2. The symptoms of PTOA are similar in characteristics to osteoarthritis (OA) which is characterized by damage to articular cartilage and subchondral bone, pain, and stiffness in the joint^3,4. The most frequently injured joints in PTOA are the knee and ankle joints³. In this study, we focus on the sub-critical damage to the articular cartilage of the knee joint using a post-traumatic disease progression rabbit model.

Covering the ends of the femur, tibia, and patella surface in the joints, articular cartilage is composed mainly of water, collagen fibers, proteoglycan, and chondrocytes⁵. Structurally, the non-calcified articular cartilage is commonly considered to have three sub-tissue zones: a superficial zone (SZ), a transitional zone (TZ), and a radial zone (TZ)⁶. The superficial zone has mostly the parallelly-oriented fibers and constitutes about 2–8% of the total thickness (height, depth). The transitional zone has mostly randomly oriented fiber and constitutes about 5–15% of the tissue thickness. The radial zone has mostly the perpendicularly-oriented fibers and constitutes about 80–90% of the thickness⁷. The non-calcified cartilage is attached to the subchondral bone via a thin layer of the calcified cartilage⁸.

The pathogenesis of PTOA is not yet fully understood^9,10. For example, the onset of PTOA is unpredictable clinically as it is diagnosed at the symptomatic phase which may occur early or over a long period of time after being triggered by the initial trauma¹¹. In addition, some anti-inflammatory therapeutics were recently proposed for the prevention and delay of the disease¹², which would require the successful validation of predictive markers to the progressive nature of PTOA¹¹. It is therefore crucial to carry out the preclinic PTOA study during its early stages, to fully understand the characteristic pathogenesis of PTOA.

In this study, we show both qualitative and quantitative characteristics of articular cartilage during the early stages of degradation through the application of one sub-critical impact to the knee of the rabbit joint. We achieved this goal using microscopic Magnetic Resonance Imaging (μMRI) and Polarized Light Microscopy (PLM) techniques. As we know, MRI is a non-invasive procedure known to be exceptional in the detection of pathological processes of the knee joint at higher resolutions^13,7,14,15. We utilized the anisotropy of T2 relaxation in this study, which has the sensitivity to the dynamics of water molecules in the tissue and reflects the organization of the collagen matrix in the tissue. In addition, PLM is the gold standard in clinical histology, which can examine the fibril structure in connective tissues via their birefringent properties¹⁶. Quantitative measurement by the Inductively Coupled Plasma-optical emission spectrometry (ICP-OES) was also performed to determine the concentration of minerals in the cartilage, which complements the imaging results¹⁷. Since μMRI shares the same physics principles and engineering architectures with the whole-body MRI scanners in the hospitals, our combined μMRI and PLM study that uses quantitative protocols has the potential to provide a direct translational pathway between bench and bed.

2 |. METHOD

2.1 |. The Impact Model

A total of twelve female mature rabbits were used in this study. The ages of the rabbits were 6 months and older and each weighed approximately 3–4kg. The use of the animal model and the experimental procedures were approved by the Institutional Animal Care and Use Committee in the relevant Institutions before performing the study. The rabbits were randomly grouped into two: 0week post-surgical group (control group, N=6) and 6week post-surgical group (N=6). An impact device with a metal hemispherical indenter tip was built in-house to deliver an impact to the knee¹⁸. The amount of a single impact was calibrated using a number of pressure sensor films (Fujifilm Prescale, Sensor Products Inc., Madison, NJ), to be approximately 30Mpa. This impact was sub-critical and did not cause any visible damage to the cartilage tissue¹⁹. Under the surgical condition, an incision was made to the joint capsule following a medial parapatellar incision to the skin. The patella was subluxated and a single impact was applied to the medial aspect of the femur as shown in Fig 1A (inside the dotted rectangular shape on the femur). For the 0week (control) group, one knee from each animal (right knee) was impacted and the contralateral knee (left knee) was non-impacted. The rabbits from this control group were sacrificed immediately after the impact and the joints excised. For the 6weeks group, an identical surgical and impact procedure was performed; the impact site was sutured back subsequently. The rabbits were sacrificed six weeks after the impact. Notes and pictures on the visual observations of the joints were taken during sample preparation and before imaging.

FIGURE 1.

Photo of A) the cut slice of the femur with the dotted box showing where cartilage block was taken from the medial side, B) the anterior surface of the femur of a 0week post-impact joint, and C) is a photo of a similar location of a 6week post-impact from impact and non-impacted joint. The white arrows in C) show color changes on the joint surface.

2.2 |. Specimen Preparation

During specimen harvesting, blocks of articular cartilage still attached to the underlying bone were cut from 12 impacted and 12 contralateral femurs in both 0week and 6weeks groups (dotted box in Fig 1A). A total of 24 samples were harvested for this study: 6 impacted femurs (0week), 6 each contralateral femur from the 0week group, 6 impacted femur (6weeks), and 6 contralateral femurs from the 6weeks group. All cartilage blocks were taken consistently and approximately at the same location on the medial femoral condyles for all joints. The cartilage-bone blocks, each measuring approximately 2×3×2 mm, were sealed in individual glass tubes that contained saline with protease inhibitor.

2.3 |. Imaging Techniques

All cartilage-bone specimens in the glass tubes were imaged individually using Bruker AVANCE IIIHD μMRI imager, with a 7-Tesla/89-mm magnet. The quantitative T2 imaging experiments used a magnetization-prepared sequence⁷. The measurements were made while the normal axis of the cartilage surface was set to 0°and 55° to the external magnetic field (B₀), which specifies the maximum and minimum influences of the dipolar interaction to spin relaxation⁷. For all experiments, the slice thickness was 0.8mm; the field of view (FOV) was 3×3mm. The matrix size for imaging was 256×128 which was reconstructed during the quantitative analysis to 256×256, giving the transverse resolution at 11.7μm/pixel. The image-formation segment had a constant echo time (TE) of 7.19ms and a repetition time of 1800ms; the T2 contrast segment had a series of varying TEs, which were 2, 4, 6, and 12ms for the 0° orientation and 2, 8, 16, and 32ms for the 55° orientation.

All specimens after μMRI were sent to an external histological service for tissue process (Yale Pathology Tissue Services, New Haven, CT). Briefly, the specimens were fixed in 10% buffered formalin and treated with the paraffin method. Three sections (6μm thick) from each specimen were made in the middle of the tissue block, approximately at the same slice location in μMRI. The sections were imaged using a PLM system, which was described previously¹⁶. The intensity images from PLM system are sensitive to the birefringent property of the cartilage. Two 2D quantitative images of each section can be constructed, one azimuth image in the unit of degree and one retardation image in the unit of nanometers.

2.4 |. ICP-OES Experiments

To determine the glycosaminoglycan (GAG) concentration in cartilage, additional forty-eight cartilage specimens from the locations adjacent to the impact/imaging sites were harvested from all 0week and 6weeks joints, approximately at the anterior and posterior locations of both impacted and non-impacted femurs. These adjacent specimens were weighed wet (ww) using an analytical balance three times. The specimens were then dried in an oven for three hours, and subsequently weighed again as dry weights (dw). The water percentage was calculated using the weights ((ww-dw)/ww %). The tissue was then liquefied by concentrated 100μl of nitric acid overnight and subsequently diluted to 4.85 ml of ultra-pure water. An internal standard was added to the aqueous samples (50 uL) making a total volume of 5 ml. An inductively coupled plasma optical emission spectrometer (Optima 7000 DV, Perkin Elmer, Waltham, MA) was used to measure sodium concentration in mg/L for the calculation of GAG concentration (mg/ml)^17,20–23. In all cases, a calibration curve, and QC samples was running for every ten samples with the curve calibration of 0.999 correlation averagely.

2.5 |. Data and statistical analysis

In the analysis of quantitative 2D T2 images from μMRI, the depth-dependent profiles from a 10-pixel width of the region of interest (ROI) from the middle of the tissue were extracted and averaged using a public-domain software ImageJ (v1.25a) and a commercial software KaleidaGraph (v 4.5.4). In the analysis of quantitative 2D angle and retardation images from PLM, similar approaches and software were used with an ROI of 117 pixels at similar locations as in the μMRI procedure. The widths of the tissue averaged in both μMRI, and PLM were approximately the same (117μm). The total thickness of cartilage was measured from the 2D intensity images in μMRI at 55° and the zonal thickness by the 1D T2 profiles at 0° based on the established criteria²⁴. For more accurate comparison, the radial zone was divided into two (RZ1 and RZ2) and the bulk T2 relaxation times were obtained over the entire tissue from the selected zones. A one-way ANOVA with Bonferroni correction was performed to test a pair-wise difference between control, OA, impact, and nonimpact for bulk and zonal parameters. A p-value of less than 0.05 was considered statistically significant. To enable direct comparison of the articular cartilage with variations in thickness, the depth-dependent profiles were normalized to 0 to 1, with 0 indicating the surface of the cartilage and 1 indicating the cartilage bone boundary.

3 |. RESULTS

3.1 |. Visual Observation

The average width of the femur across the condyles at the point where the cartilage block was measured was approximately 1.5cm, and the medial condyle was approximately 0.6cm. The femoral cartilage of the impacted joint after 6week post-surgery was intact without significant sign of destruction, similar to the cartilage in the impacted 0week joint (control). The femoral cartilage of the 6weeks joint showed localization of ‘dark red to purple’ color change on the anterior aspect of the femur in both impacted and non-impacted joints (Fig 1C). It should be noted that Fig 1B of the 0week joint has an overall ‘blood-red’ appearance resulting from fresh oxygenated blood of immediate post-surgical procedure.

3.2 |. μMRI Measurement

Fig 2 shows the 2D MRI intensity images at both 0° and 55° orientations together with one corresponding T2 map at 0° orientation. Visually, the impact and non-impacted cartilage of the 0week (Fig 2A) and 6weeks (Fig 2B) samples looked similar with little changes in their thickness (summarized in Table 1). From the quantitative T2 images, the depth-dependent T2 profiles were extracted from all images, which were subdivided into the cartilage structural zones based on our published methods^14,24,25. Fig 3 shows the correlation among the whole-profile averaged T2 values between the impacted and non-impacted samples for both 0week and 6weeks at 55°. While there was no difference in T2 between the non-impact and the impact control samples, there was a clear increase in T2 in the 6weeks samples, where the impact T2 is higher than the non-impact T2 (p=0.06).

FIGURE 2.

2D intensity images from μMRI T2 experiments. A) shows the intensity images of 0week cartilage blocks from impact and non-impact joint and B) shows the intensity images from 6weeks of cartilage blocks from impact and non-impact joint. The quantitative T2 maps are shown on the rightest panel. Similar cartilage characteristics are observed in both sample groups

Table1.

Correlation of cartilage thickness between 0week and 6weeks samples from the impacted and non-impacted femurs. The p-values are for each comparison (the bold font indicates the significant difference).

	SZ (μm)			Total Tissue (μm)
	0week	6weeks	PV	0week	6weeks	PV
μMRI Impact	14.6±2.3	11.7±0.01	1.0	298.2±17.4	286±17.0	1.0
μMRI Non-impact	17.5±2.7	19.4±1.2	1.0	313.9±24.0	306.1±16.6	1.0
PV	1.0	0.09		1.0	1.0
PLM Impact	19.2±0.4	7.7±1.3	0.0002	433.0±19.4	412.0±13.0	1.0
PLM Non-impact	16.5±1.2	12.3±0.4	0.2	400.0±16.8	394.0±25.0	1.0
PV	0.1	0.1		1.0	1.0

FIGURE 3.

Averaged bulk T2 values of cartilage between impact and non-impact joints for 0week and 6weeks samples with their surface oriented at a magic angle (55°) to the external magnetic field (B_o). Note a general increase in 6weeks samples (fitted with a black solid line) in bulk than in 0week samples (fitted with a grey solid line). Corresponding dashed lines represent the 95% confidence intervals for the average T2 mean values of the individual samples (markers).

Fig 4 shows the zonal averaged T2 correlation between the impacted and non-impacted samples for both 0week and 6weeks. At the 55° orientation (Fig 4A), the zonal correlations show a significant T2 increase in both SZ and TZ (P=0.003) and a less significant increase in RZ. At the 0° orientation (Fig 4B), although the zonal T2 have lower values due to the influence of dipolar interaction, a similar trend in the T2 correlation can be found. (RZ 2 at 0° was not analyzed due to extremely short T2 values causing an increase in noise influences in that region²⁶). Table 2 summarizes the correlation in T2 between the impacted and non-impacted cartilage for both 0week and 6weeks samples. Table 2 also summarizes are the slopes of the fitted lines for both groups of samples, where a steeper slope means T2 has increased. The slope of the solid diagonal in both Fig 3 and Fig 4 is 1.

FIGURE 4.

Averaged zonal T2 values of cartilage between impact and non-impact joints for 0week and 6week samples with their surface oriented at A) at magic angle (55°) to the external magnetic field (B_o), B) at 0° to B_o. The 6weeks samples are fitted with the black solid lines and 0week with grey solid lines. Corresponding dashed line fits represent the 95% confidence intervals for the average T2 mean values of the individual samples (markers).

Table 2.

Comparison of the averaged zonal T2 values between 0week and 6weeks, and between impact and non-impact femoral cartilage in μMRI. The p-values are for each comparison (the bold font indicates the significant difference).

ZONE	IMPACT T2 (ms)			NON-IMPACT T2 (ms)			SLOPE
	0week	6weeks	PV	0week	6weeks	PV	0week	6weeks
A) 55°
BULK	30.15±0.84	32.66±0.51	0.06	31.89±0.98	30.23±0.40	0.56	1.02	1.10
SZ	26.56±1.48	30.48±4.3	0.003	35.98±1.47	29.89±2.30	0.002	0.68	0.94
TZ	26.30±1.34	31.39±1.04	0.003	34.77±1.14	28.81±0.47	0.003	0.75	1.13
RZ1	34.73±0.77	39.02±1.22	0.005	32.64±1.20	35.77±1.80	0.01	1.06	1.07
RZ2	32.54±0.48	29.98±0.60	0.06	24.62±0.40	24.70±0.41	1.0	1.02	1.10
B) 0°
SZ	7.81±0.30	9.47±0.41	0.05	9.35±0.55	10.54±0.37	0.3	0.67	0.82
TZ	8.15±0.21	10.34±0.50	0.001	10.44±0.47	10.20±0.34	1.0	0.68	0.94
RZ	3.44±0.11	3.66±0.09	1.0	3.81±0.10	4.23±0.23	0.3	0.74	0.83

3.3 |. PLM Measurement

Fig 5 shows the quantitative 2D angle (Fig 5A) and retardation (Fig 5B) images at a 1μm/pixel resolution, which assesses the architectural integrity of collagens in cartilage. One visual difference between the 0week and 6weeks groups can be noted: when compared to the 0week samples, the cartilage surface in almost all the 6weeks impacted samples shows reduced total thickness and less uniform fibril orientations. This is illustrated visually that the 2D angle image of the 6weeks impacted sample has the thinnest blue region in the surface (i.e., the least amount of uniform fibers in parallel with the articular surface) as well as the 2D retardation image of the 6weeks impacted sample has the lowest gray values in the surface (i.e., reduced retardation values), which collectively indicates the fibril structures in the surface cartilage have changed in the 6weeks impacted group. These changes are statistically significant (Table 1) in the impacted joint as compared to the contralateral joint. In addition, there seems to be an increase in the volume of the individual chondrocytes in the majority of the samples in the 6weeks groups. Note that there was no significant difference in the weights and ages between these two groups of rabbits.

FIGURE 5.

2D images of angle (A) and retardation (B) PLM technique at higher resolution 1μm/pixel for 0week and 6weeks samples of impact and non-impact joints. SZ, TZ, and RZ represent superficial, transitional, and Radial zones.

Fig 6 compares the angle profiles between the 0week and 6week of the impacted (Fig 6A) and non-impacted cartilage (Fig 6C), which further supports the observations in the 2D images (Fig 5) by showing the diminishing in the thickness at the superficial zone in the 6week impact samples as indicated in the arrow in Fig 6A. The corresponding retardation profiles show the reduction of the retardation values at the surface for the 6week impact samples (Fig 6B, pointed by the surface arrow), consistent with the observations in the 2D images (Fig 5) and supporting the observation of a reduced/disrupted surface structure. There is also a noticeable increase in the retardation values around the tidemark region, pointed out by the upwards arrow in Fig 6B. The values of the minimum and maximum retardations are summarized in Table 3. While the maximum retardation values are comparable, the 6weeks impacted samples have smaller values in the minimum retardation region.

FIGURE 6.

Averaged depth-dependent profiles of the angle in the impacted A) and non-impacted C) joints for 0week and 6weeks cartilage blocks. Arrow in A) shows a reduction in SZ thickness of the impact SZ thickness. Averaged depth-dependent profiles of the retardation in the impact B) and non-impacted D) for 0week and 6weeks of cartilage blocks. In B), the surface arrow points to the reduction of the retardation in the 6week group; while the deep arrow points to the estimated tide mark of the cartilage.

Table 3.

Comparison between 0week and 6weeks impacted, and non-impacted femoral cartilage (n=24) for PLM parameters.

	IMPACT			NON-IMPACT
Parameter	0week	6weeks	PV	0week	6weeks	PV
Min retardation(nm)	1.66±0.14	1.09±0.08	0.4	1.56±0.18	1.57±0.32	1.0
Max retardation(nm)	10.20±0.27	11.52±0.23	0.8	8.21±0.82	8.09±1.27	1.0

Table 4 summarized the ICP analysis for the cartilage harvested from the locations adjacent to the impact/imaging locations from all 0week and 6weeks joints, approximately at the anterior and posterior locations of both impacted and non-impacted femurs. There were no significant differences among the bulk measurements of water percentage, GAG and calcium concentration, which further indicates the tissue degradation in this study was minor, i.e., at the very early stages of the progression.

Table 4.

Biochemical analysis by the ICP-OES method, between impacted and non-impacted 0week, and 6weeks femoral cartilage (n=10). PV corresponds to the p-value.

	IMPACT	NON-IMPACT	PV	IMPACT	NON-IMPACT	PV
	0week	0week		6weeks	6weeks
Water (%)	74.5 ± 1.7	77.4±2.0	0.3	77.8±1.4	76.3±2.5	0.6
GAG (mg/ml)	88.5±8.4	79.5±6.1	0.4	68.2±6.1	71.7±7.8	0.7
Calcium (mg/L)	3.2±1.0	1.73±0.4	0.2	4.3±0.7	2.5 ±0.7	0.09

4 |. DISCUSSION

In musculoskeletal applications, clinical MRI lacks sufficient resolution to detect the early stages of cartilage degradation in OA. μMRI, which shares the same physics principles and engineering architecture with clinical MRI, has the ability to study intact tissue blocks and live specimens non-invasively and non-destructively, hence is widely used in preclinical studies. In this project, we combined μMRI with the quantitative PLM to study the early stages of cartilage degradation that was triggered by a sub-critical impact, using a rabbit model. The use of μMRI is essential in this project due to the rather thin thickness of the rabbit cartilage (300–400 μm). The use of PLM further enhances the μMRI data with its optical resolution^{27, 28}.

4.1 |. Early PTOA

When the external trauma (e.g., a direct impact) is substantial but not catastrophic, the relevant tissues and joints in individual animals in an identically specified group could develop different and diverse responses in the subsequent post-traumatic period at the molecular and chemical levels²⁹. In this study, the trauma to the femoral head was caused by a single and direct hit with deliberate low pressure (~30 MPa). The design of this study was to use a pressure that would not cause significantly visible damage (hole) to the cartilage. After the impact, no substantial damage was visible on the articular surfaces of the impacted in the 0week group (Fig 2A). At the 6weeks post-trauma time point, however, there were signs of degeneration in the anterior aspect of the femur, around the patellar groove in both impacted and non-impacted joints, as reddening and signs of bruising are observed (Fig 2C). These observations are consistent with the understanding to the causal role of inflammation in OA progression³⁰.

Since both impacted and non-impacted femurs showed signs of bruising, the observation cannot be placed solely on the unintended damage and internal bleeding from the surgical procedure, but possibly to a reactive or degenerative response to the immediate or acute phase of OA degradation, related to the mechanical stimuli through intraarticular bleeding³¹. The signs of degradation in the non-impacted femurs might also suffer from the consequence of a changed gait patterns in the animals (i.e., changes in loading of the contralateral due to pain in the injured limb). In addition, we noted that the degenerative sign did not occur on the central aspect of the medial condyle where the impact was made but rather on the anterior femur, which indicates OA as a localized process. That is, the early response and progression of OA does not occur uniformly over the entire cartilage surface at any time. This topographical variation in the response to impact inflammation could be due to the fact that loading is known to affect the metabolic activities of chondrocytes³² as well as from the motional overloading of the non-impacted joint after any single-side injury. This topographical variation in the response to impact inflammation also increases the challenges in an accurate determination of early degradation in the affected joints.

4.2 |. OA Detection by MRI

MRI intensity images between the 0week and 6weeks groups showed no significant changes in terms of cartilage lamination and thickness. This lack of visual differences illustrated a minimal damage to cartilage, which was the consequence of our intentionally sub-critical impact. Even using one of the highest transverse resolutions (11.7μm/pixel) in this μMRI study, the voxel-averaged collagen fibers remained unchanged during early osteoarthritis²⁷. The topographical nature of cartilage on the femoral joint further masks any detection sensitivity of MRI, when the intensity images are examined⁸.

In contrast, our quantitative T2 results showed subtle changes in cartilage due to the impact. Based on extensive previous studies from many research groups^33,34, the T2 values have been linked to the progression of osteoarthritis, especially its values at the magic angle (55°) due to the minimization of the dipolar interaction⁷. In this study, the bulk T2 value at 55° for cartilage was higher in the 6weeks group than in the 0week group (Fig 3). This supports the understanding that an increase in hydration in cartilage is an early event in the cartilage fatigue that signals the disruption of the collagen matrix³⁵. The changes in the surface tissue hydration were the major contributing factor to the differences in the bulk cartilage (Fig 4), which illustrates a non-uniform degradation over the tissue thickness – i.e., the surface tissue degrades earlier than the deep tissue in this type of impact-induced osteoarthritis. The deep (radial) zones, which are the thickest sub-tissue zone in non-calcified cartilage, is the most difficult region to be examined by both clinical MRI and μMRI, largely due to its collagen structure and molecular composition. Past studies in the canine OA model also showed that the deep cartilage would be the last region to be degraded in OA²⁷.

An additional possible finding is that the applied impact, although sub-critical, had caused some minor damage to the subchondral bone plate, which could weaken the integrity of non-calcified cartilage. Any degradation in the subchondral bone plate could have complex consequences to the health of cartilage, including the increased mineralization and other degradation^{36,37 38}. For example, the degradation of the cartilage matrix is known to be a possible critical factor for matrix calcification and a crystal deposition may occur without severe cartilage degradation³⁹. The potential mineralization of the deep cartilage, which shortens the T2 values, unfortunately increases the difficulties in the detection of the deep tissue degradation by MRI. Further studies of this deep region may require the use of some additional tools, such as microscopic computer tomography (μCT), which could sense the variations of the minerals in calcified cartilage and the subchondral bone plate.

4.3 |. OA Detection by PLM

PLM can have higher spatial resolution than any μMRI, and PLM uses optical birefringence to examine quantitatively the collagen network in articular cartilage⁴⁰. In this study, the surface cartilage in the 6weeks impact group was found to have a reduced zonal thickness in the angle map (thinner and noisy blue colors in the top right image in Fig 5A) in addition to a variable retardation value in the surface region (the reduced brightness in the gray-scale image on the top figure of Fig 5B). These two measurements illustrate collectively the most profound disruption of the surface collagens in the 6weeks impact group. This disruption of cartilage surface in early OA is a significant biomarker and agrees with some previous OA studies in canine cartilage²⁷. Our study showed a strong statistical difference in the SZ thickness between the control and OA, which was more significant in the impacted joint (P=0.0002) than in the non-impacted joint (P=0.2). This is expected due to direct injury applied to the impacted joint. This result attests to the high ability of PLM to detect minute structural changes even when the trigger event (i.e., impact) is sub-critical.

In addition to the observed disruption in the surface collagen, the 2D PLM images in the 6weeks group qualitatively showed increased cellular volume, especially in the deep cartilage (comparing the two left images with the two right images in Fig 5B). It is interesting to note that this increase in the cellular volume could also be observed in the non-impact joints (between the two lower images in Fig 5B). It is believed that in PTOA progression, there is the initiation of intra-articular pathogenic processes, such as apoptosis of articular chondrocytes during the initial traumatic event¹¹. Although we do not know the precise nature of this cellular change in the 6week groups, any change in the chondrocyte environment can affect the maintenance of the cartilage through mediation in homeostasis and reduce the ability of cells to repair⁴¹. Irrespective of minor disruption occurring in the 6weeks cartilage tissue, there could still be the possibility of apoptosis occurring through the caspase pathway⁴².

Summarizing the discussion in the last two sections, we note that the tissue degradation after a single impact to a particular location in a joint can be complex and topographically distributed -cartilage tissue in the nearby locations can also degrade due to the single impact. Despite lacking visual changes to the cartilage, the high sensitivity of our combined μMRI and PLM techniques was able to pull quantitative changes out of such subtly damaged cartilage.

4.4 |. Experimental Notes and Limitations

Any study of a group of individuals (animals or humans) must face a number of challenging factors that are difficult to control or uniformize. In the animal model studies of OA, these factors could include age, weight, genetics, surgical comparables, and difference in behavior towards repetitive use of the affected joints; all of which could influence the outcome of any quantitative study. In PTOA studies, additional factors include the precise amount and topographical location of the impact, as well as the shape of the indenter tip. The use of female rabbits in this study was due to the availability of the rabbits during the performance of the experimentation (the beginning of the pandemic), when the supply of the retired females was far greater than the male supply. To the best of our ability and control, we randomly assigned the animals to different groups, and we monitored the specifics of the individual animals. We do not observe any trend in the influence of the age and weight of the animals on the experimental data in this study. By definition, the animals of the control group were six weeks younger than the 6week group. Although Fig 5b does show that the 6week group has low cell density, the thinning of the superficial zone in the four angle images (Fig 5a, between 0week and 6week, and between impact and non-impact) suggests clearly that the observed and measured differences between the two groups in this study is not due to the age difference of the two groups, but the impact procedure on the joint.

The fundamental limitation for any imaging study of cartilage degradation is the mismatch between the thickness of articular cartilage and the spatial resolution in imaging. Although the transverse resolution in our μMRI is likely the highest in any MRI study of rabbit cartilage, it is still barely sufficient to study the zonal variations in the femoral cartilage in the rabbit model. When larger animals are used, the resolution in MRI can be relaxed, as was elaborated in the scaling law in MRI of cartilage⁴³.

In addition, topographical variation in cartilage characteristics over any single joint surface is also a challenging factor in any study of OA⁸. Some joints have more complicated topographical variations than some other joints (e.g., the knee joint has more complex topographies than the same in humeral and hip heads). Although we tried our best to impact at and harvest from the same location in the femur, minor uncertainty must exist, which could further mask the detection sensitivity when a group of animals are studied individually.

Finally, we acknowledge that our inclusion of only a healthy contralateral control without a sham procedure is a limitation of the study. The decision to utilize a healthy contralateral control limb in lieu of a sham limb or sham group allowed us to significantly reduce the number of rabbits needed for our experimentation. We note that healthy contralateral limbs are commonly utilized for comparison in surgical models of OA in many published studies in literature.

5 |. CONCLUSION

Using a single sub-critical impact to the femoral cartilage, we confirmed the development of cartilage degradation in a rabbit model, which can be detected by both μMRI and PLM. Quantitatively, there was an increase in the bulk T2 values in cartilage which can be related to an increase of water mobility and content in the cartilage matrix, especially at the surface. PLM complements the μMRI in the detection of changes in the surface cartilage thickness and collagen organization. The data confirmed that the degradation to cartilage was localized to the surface tissue in this low-impact PTOA model of rabbit.