If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence and reprint requests to: S. Taheri, Department of Trauma Surgery, Orthopaedic Surgery and Plastic Surgery, Universitätsmedizin Göttingen, Robert-Koch-Str. 40, 37075 Göttingen, Germany. Tel.: 49-(0)-551-39-62613.
University of Göttingen Johann-Friedrich-Blumenbach-Institute for Zoology & Anthropology, Department of Historical Anthropology and Human Ecology, Göttingen, Germany
University of Göttingen Johann-Friedrich-Blumenbach-Institute for Zoology & Anthropology, Department of Historical Anthropology and Human Ecology, Göttingen, Germany
We have identified a 3D network of subchondral microchannels that connects the deep zone of cartilage to the bone marrow (i.e., cartilage-bone marrow microchannel connectors; CMMC). However, the pathological significance of CMMC is largely unknown. Here, we quantitatively evaluated how the CMMC microarchitecture is related to cartilage condition, as well as regional differences in early idiopathic osteoarthritis (OA).
Methods
Two groups of cadaveric female human femoral heads (intact cartilage vs early cartilage lesions) were identified, and a biopsy-based high-resolution micro-CT imaging was employed. Subchondral bone (SB) thickness, CMMC number, maximum and minimum CMMC size, and the CMMC morphology were quantified and compared between the two groups. The effect of joint's region and cartilage condition was examined on each dependent variable.
Results
The CMMC number and morphology were affected by region of the joint, but not by cartilage condition. On the other hand, the minimum and maximum CMMC size was changed by both the location on the joint, as well as the cartilage condition. The smallest CMMC were consistently detected at the load-bearing region (LBR) of the joint. Compared to non-pathological subjects, the size of the microchannels was enlarged in early OA, most noticeably at the non-load-bearing region (NLBR) and the peripheral rim (PR) of the femoral head. Furthermore, subchondral bone thinning was observed in early OA as a localized occurrence linked with areas of partial chondral defect.
Conclusion
Our data point to an enlargement of the SB microchannel network, and a collective structural deterioration of SB in early idiopathic OA.
Osteoarthritis (OA) is a prevalent joint disorder with an increasing public health challenge that profoundly impacts individuals as well as the society in general
. Although originally labeled as a “wear and tear” process mainly characterized by articular cartilage (AC) degradation, OA is now widely considered to be a progressive joint failure involving all periarticular tissues
AC and SB are intimately interlocked, creating a complex unit called the AC-SB interface, where the functionality and homeostasis of the adjacent unit can be modulated by mediators from the other tissue
. Through the course of bone remodeling, leukotrienes, prostaglandins, and several growth factors that are released in vivo by osteoblasts can reach the overlying cartilage
. However, the direct connective pathways allowing this communication between AC and SB are poorly understood. Since the second half of the last century, smaller fractures or microchannels in subchondral bone were observed
. Our recent studies have revealed that cartilage and bone marrow are connected by a three-dimensional network of microchannels (i.e., cartilage-bone marrow microchannel connector; CMMC), which are microarchitecturally different in number, size, and morphology depending on the maturation phase of the bone
, there is surprisingly a lack of studies on their (patho)physiological significance. In normal human femoral heads, the age-related reduction of vessel-containing SB perforations has been interpreted as an early degeneration marker
. Given our recent findings on the CMMC network, and accumulating evidence that the (micro)structural changes of SB precede full-focal cartilage defects
, one essential question is if/how these hierarchical microarchitectures change in early OA.
Here, we extended our established methodology to quantitatively evaluate the CMMC network of human femoral heads that exhibit partial-thickness defects and mild chondral fibrillations associated with idiopathic early OA. We hypothesized that the CMMC metrics change based on the health-state of the overlying cartilage, as well as areas of the joint that represent regional differences in physiological loading.
Materials and methods
Preparation and selection criteria of human bone specimens
The subjects used in this study were anonymous human femurs granted by the anatomical gift program of the Hannover Medical School (MHH). As no information was provided regarding the medical history, sex, and age of the cadaver donors, the OA status, sex, and the biological age of the subjects were determined post-mortem as outlined in Section 1 of the Supporting Information (SI). Five healthy and five early arthritic female donors (right leg) were selected for measurement and analysis. The majority of subjects had a biological age of 40–60 years at death (Table S2).
Assigning of the measuring points and biopsy
To acquire high-resolution images, 2.00-mm cartilage-bone cylinders were extracted from the surface of femoral heads. In healthy subjects, the measuring points were intersections of twelve concentric lines and four parasagittal planes that were outlined on the surface of the joint using a template grid [Fig. 1(A); details in Ref.
]. 43 measuring points were extracted from the surface of each healthy subject [215 biopsies in total; yellow dots in Fig. 1(B)]. These geometrical shapes divide each femoral head into several sectors, one of which is marked by black crosshatched lines in Fig. 1(B). A maximum of four measuring points per sector was extracted from early OA subjects in case the entire surface of the sector was covered by early chondral defects. With the aid of a template grid that incorporated the described shapes, the position of each set of biopsy was standardized taking into consideration the normalized size of each joint. Additionally, a severity mapping of local partial-thickness defects was created by a sequence that comprised several photogrammetry procedures [Fig. 1(C)–(E); details in the SI]. In total, 116 cartilage-bone biopsies were drilled out and scanned from regions where the overlying cartilage showed early pathological features, and compared with the 215 biopsies of healthy femurs.
Fig. 1The system for assigning measuring points on the femoral head. (A) For healthy subjects, measuring points were defined as the convergence points of twelve concentric lines (yellow dashed lines) and four parallel parasagittal planes (Roman numerals; cyan) on the surface of the joint. (B) In each healthy subject, 43 cartilage-bone cylinders were extracted for high-resolution scanning from the designated locations (orange dots). The aforementioned geometrical shapes divided each femoral head into 42 sectors, one of which was marked by black crosshatched lines. For early OA subjects, 1–4 biopsies per sector were extracted and scanned depending on the expansion of early OA cartilage characteristics (i.e., partial-thickness defects and fibrillations) over each individual sector. To create a severity mapping of the most prevalent early OA regions of the joint, areas of early cartilage degeneration were superimposed by a photogrammetry technique that consisted of photo alignment and point cloud generation. (C) processing of dense could and continuous mesh over the surface of the model, and (D) integration of the texture maps over each mesh to generate realistic 3D models of each subject.
. The SB thickness was measured in sagittal cross-sections of the 3D-reconstructed models, while the CMMC metrics were quantified on a layer-by-layer basis (in the transverse plane relative to the long axis of the biopsy) using an ImageJ macro. They were characterized by their local density per mm2 (CMMC number), maximum (Feret) and minimum size (MinFeret), as well as the morphological index, Circularity (Circ.
). For schematic descriptions and exemplary values of Feret, MinFeret, and Circ., see Fig. S3.
By estimating reaction forces at the right hip from the gait analysis on a separate healthy cohort, we have previously demonstrated that regional differences in the microchannel architecture of SB may reflect regional differences in loading
. The results were used as a physiological reference to categorize the output of the algorithm (i.e., CMMC metrics). Similarly, the results here are categorized and reported for the identical load-bearing region (LBR), non-load-bearing region (NLBR), and the peripheral rim of the joint (PR; see middle panel in Fig. 4). The number of measuring points for each region and health-states are reported in Table I.
We evaluated the effect of region and cartilage condition on SB thickness and each CMMC metric using generalized estimating equation (GEE) models. A detailed description of the methods, including statistical analyses and histological techniques is provided in the SI.
Results
Early subchondral thinning and cartilage fibrillation occur prevalently at the inferoposterior portion of the femoral head
The SB thickness measurements indicated a strong effect of region (95% Wald Confidence Interval; 95% WCI = 87.13 to 169.46, p < 0.0001). Irrespective of the health-status of the cartilage, the thickest subchondral bone was detected at LBR of the femoral head (mean = 239.41 μm, SEM = 14.90, 95% WCI = 210.21 to 268.61), while the SB thickness was significantly reduced at NLBR (β = −82.52, 95% WCI for difference = −121.13 to −43.93, P = 0.005), and the peripheral rim (β = −107.58, 95% WCI for difference = −139.04 to −76.15, p < 0.0001) of the joint [Fig. 2(A)]. We also observed a milder, albeit significant effect of cartilage condition on SB thickness (95% WCI = 4.99 to 66.76, p = 0.023). At the LBR, the SB thickness was not affected by cartilage condition (mean difference = 4.90, 95% CI for difference = −82.55 to 92.35). Pairwise comparisons, however, showed that early OA subchondral thinning was detected at the NLBR (mean difference = 56.435; 95% WCI for difference = 16.8129 to 96.0571; p = 0.004). At the rim, a trend toward the reduction of SB thickness was observed, which was not significant (95% WCI for difference = −6.598 to 99.214; P = 0.153).
Fig. 2Changes of SB thickness and characteristics of early cartilage defects in human femoral head. (A) Changes of SB thickness in different areas of the femoral head. Data are presented as box plots with whiskers. The triangles overlying the boxplots signify the mean values of corresponding subjects in each region and cartilage condition. Subjects 1–5 were healthy while subjects 6–10 had early osteoarthritic chondral characteristics (n = 5 per group). Each circle in close proximity to the boxplots represent the mean value of a measuring point, color-coded to its corresponding subject (m; number of measuring points = 215 for healthy subjects; 116 for early OA subjects). Analysis was conducted using generalized estimating equation followed by the Bonferroni test for pairwise comparison. Significant P-values are presented as ∗P <0.05; ∗∗P <0.01. (B) Colormap of the distribution of early cartilage lesions in five subjects with a mean Outerbridge score between 1.5 and 2.5. The color scale is based on the total number of measuring points per sector in the five subjects, where a maximum of four measuring points per sector was extracted in case the entire surface of the sector was covered by early OA chondral defects. (C) Masson-Goldner staining of a representative sagittal section of a healthy biopsy showed a preserved articular surface with no signs of cartilage degradation, a zone-dependent chondrocyte distribution, and a single microchannel with an approximate size of 25 μm. (D) The microchannel (marked by asterisk) was engulfed by a sheath of subchondral bone lamellae (turquoise), which penetrated through the calcified cartilage layer (CC; pale green). (E) In early osteoarthritic cartilage, partial degradation of the superficial cartilage layer was observed, tidemark duplication was detected (open triangles), and chondrocytes lost their hierarchical distribution within the extracellular cartilage matrix. Additionally, cartilage discoloration associated with ectopic cartilage calcification was found to be a prominent phenomenon (F; yellow triangles).
The colormap of partial thickness defect showed that early cartilage lesions and fibrillations were most prevalent inferoposteriorly in the vicinity of the fovea capitis [Fig. 2(B)]. Consequently, the majority of early OA measuring points were located at the NLBR (63%), and the outermost PR of the joint (27.6%), while the least frequency of partial thickness defect (9.4%) was observed at the load-bearing, anterosuperior portion of the joint.
Masson-Goldner's trichrome staining of the osteochondral junction supported our macroscopic observations and sample selection. Healthy subjects were characterized by a regular cartilage surface, as well as a hierarchical chondrocyte distribution in different cartilage zones [Fig. 2(C)]. CMMCs could occasionally be detected as two-dimensional vessels in the sagittal plane [one of which is magnified image in Fig. 2(D)]. They advanced through the calcified cartilage layer at the contact point of cartilage and subchondral bone, and were surrounded by the subchondral bone plate. On the other hand, the early OA biopsies were signified by cartilage erosion and partial thickness defects down to the mid-zone of cartilage as expected [Fig. 2(E)], a loss of morphological variation in chondrocytes in different cartilage zones, tidemark duplication, and discoloration at the uppermost cartilage surface due to ectopic cartilage calcification [Fig. 2(F)].
Histological examination of the CMMC network in transverse cuts
An exemplary microCT-acquired 2D image and the corresponding binarized image is depicted [Fig. 3(A) and (B) ]. Two neighbor transverse cuts from the same cross-section of the biopsy were stained by Giemsa and Masson-Goldner trichrome [Fig. 3(C) and (D), respectively], revealing identical microporous structures compared to the micro-CT image, which demonstrated the utility of transverse cuts for histological examination of the CMMC network, as well the accuracy of the applied threshold setting in micro-CT analysis. Most CMMCs were surrounded by a concentric sheath of appositional lamellar bone that did not comprise an endosteum, and therefore, seemed to be different from classic mature osteons. Additionally, the microchannels contained thin-walled vessel-like membranes that were lined with intravascular-, as well as extravascular cells in the encompassing subchondral bone plate (higher magnification images in Fig. S6). The exact cellular and molecular contents of the CMMCs are unknown, even though morphological assessments strongly indicated that SB microchannels were composed of diverse microvasculature structures (see Section 4 of the SI).
Fig. 3Histological examination of the CMMC network in transverse cuts. (A) a 2D slide of an exemplary biopsy obtained from micro-CT imaging before and (B) after segmentation and binarization. Transverse sectioning of histological cuts allow for a direct correlation between CMMC structures derived from micro-CT with corresponding histological slides. Two consecutive sections situated at ∼100 μm below the tidemark were stained by (C) Giemsa and (D) Masson-Goldner's trichrome, showing identical microporous features to the micro-CT image, as well as a concentric lamellar bone structure that surrounds each CMMC. Higher magnification Alcian blue, Giemsa, and Masson-Goldner stainings of a selected region of interest (the dotted orange rectangle in A) are shown in Fig. S4(D–F) of the supporting information and illustrate more details.
General overview of the regional CMMC characteristics based on the health-status of the overlying cartilage
The representative 3D-reconstructed models showed that the region-specific distribution of the microchannels that we previously observed in the healthy, control group persisted in early-OA group as well
. i.e., In the LBR, SB in both healthy and early OA subjects is perfused with abundant small microchannels, which form several coalescent finger-like interconnections [Fig. 4(A); Supplemental Videos 1 and 2]. On the other hand, NLBR is characterized by sporadic ampulla-like canals with a varying combination of small, medium, and large channel sizes [Fig. 4(B); Supplemental Videos 3 and 4]. The largest CMMCs are typically detected at the PR of the joint as irregularly-shaped gap formations with sizeable contact areas with cartilage [∼0.003–0.012 mm2; Fig. 4(C); Supplemental Videos 5 and 6]. Observations indicated that the CMMC size was generally larger in early OA subjects compared to their region-specific healthy counterparts [Fig. 4(A)–(C)].
Fig. 43D representation of the SB microchannel network in human femoral head. (A) The superior view of typical samples associated with the load-bearing region (LBR; marked in orange in the central figure) revealed abundant microchannels reaching to the uppermost superficial surface of SB. (B) At the non-load-bearing region of the joint (NLBR; marked in blue), the local density of the CMMC appeared to be reduced compared to the LBR with a varying combination of small and medium-sized channels, while (C) the peripheral rim of the joint (PR; marked in green) was characterized by large and irregularly-shaped gap formations in the SB. In each identified area, an enlargement of the CMMC size was observed in early arthritic biopsies compared to the corresponding healthy specimens (A–C). (D) The cross-section of the healthy, LBR model was cut at a random plane of sectioning (yellow dotted line in A). Yellow arrowheads signify selected CMMCs that reached the SB surface (E) The negative 3D model of the same LBR cross-section showed the AC in orange, osseous structures in transparent, and the CMMC in blue-grey. The microchannels were visible throughout the entire thickness of the SB, while discrete microchannels cut by the random plane of sectioning were marked in red. The white dotted lines in D and E denote the lower margin of the subchondral bone plate.
3D model of an exemplary cylindrical biopsy (original size 2.00 mm), extracted from the load-bearing region (LBR) of human femoral head covered by intact cartilage (mean Outerbridge score < 0.5). The direction of the camera starts from the trasverse view of the subchondral bone’s uppermost surface and ends with a 360° rotation around the sagittal plane below the tidemark. In the LBR, subchondral bone is perfused with abundant microchannels, which form several coalescent finger-like interconnections.
3D model of an exemplary cylindrical biopsy (original size 2.00 mm), extracted from the load-bearing region (LBR) of human femoral head covered by early OA cartilage lesions (mean Outerbridge score 1.5 < Grade < 2.5). The direction of the camera starts from the trasverse view of the subchondral bone’s uppermost surface and ends with a 360° rotation around the sagittal plane below the tidemark. CMMCs maintain a similar morphological characteristic previously observed in the LBR (i.e. finger-like coalescent), but the channels are enlarged compared to healthy biopsies.
3D model of an exemplary cylindrical biopsy (original size 2.00 mm), extracted from the non-load-bearing region (NLBR) of human femoral head covered by intact cartilage (mean Outerbridge score < 0.5). The direction of the camera starts from the trasverse view of the subchondral bone’s uppermost surface and ends with a 360° rotation around the sagittal plane below the tidemark. NLBR is characterized by sporadic ampulla-like canals with a varying combination of small, medium, and large channel sizes.
3D model of an exemplary cylindrical biopsy (original size 2.00 mm), extracted from the non-load-bearing region (NLBR) of human femoral head covered by early OA cartilage lesions (mean Outerbridge score 1.5 < Grade < 2.5). The direction of the camera starts from the trasverse view of the subchondral bone’s uppermost surface and ends with a 360° rotation around the sagittal plane below the tidemark. Compared to the equivalent healthy biopsy, the CMMCs are larger.
3D model of an exemplary cylindrical biopsy (original size 2.00 mm), extracted from the peripheral rim (PR) of human femoral head covered by intact cartilage (mean Outerbridge score < 0.5). The direction of the camera starts from the trasverse view of the subchondral bone’s uppermost surface and ends with a 360° rotation around the sagittal plane below the tidemark. Typically, the largest CMMCs are detected at the peripheral rim of the joint and appear as irregularly-shaped gap formations.
3D model of an exemplary cylindrical biopsy (original size 2.00 mm), extracted from the peripheral rim (PR) of human femoral head covered by early OA cartilage lesions (mean Outerbridge score 1.5 < Grade < 2.5). The direction of the camera starts from the trasverse view of the subchondral bone’s uppermost surface and ends with a 360° rotation around the sagittal plane below the tidemark. Typically, the largest CMMCs are detected at the peripheral rim of the joint and appear as irregularly-shaped gap formations. At PR, CMMC size was generally larger in early OA subjects compared to its healthy counterparts.
To demonstrate the intricate microstructure of the CMMCs, an exemplary model was virtually cut and viewed from an arbitrary coronal plane [yellow dash line in Fig. 4(A), viewed in Fig. 4(D)], which revealed isolated microporous structures in the plane of sectioning. Nonetheless, the inverted 3D-representation showed that these micropores were part of a dense CMMC network that existed throughout the entire subchondral bone.
The local density of CMMC does not change in early OA but their size is increased
The type III test of model effects revealed no effect of the joint's cartilage condition on the CMMC number (95% WCI = −2.703 to 3.585, P = 0.883), but a significant impact of region (95% WCI = 5.669–8.464, P < 0.0001). The estimated marginal means, SEM, and the 95% WCI of all the CMMC metrics across the levels of each factor are presented in Table II. Irrespective of cartilage condition, the CMMC number was lower in the NLBR (β regression coefficient = −4.486, 95% WCI for difference = −6.567 to −3.684, P < 0.0001), and the PR (β = −5.381, 95% WCI for difference = −6.648 to −3.637, P < 0.0001) compared to the LBR as reference. In early OA samples, in particular, a consistently higher CMMC number was detected in the LBR than the NLBR (95% WCI for difference = 3.162 to 5.809; P = 0.036) and the PR (95% WCI for difference = 4.789 to 5.974; P < 0.01). Other significant pairwise comparisons are marked in Fig. 5(A).
Table IIRegion-specific CMMC metrics in healthy and early OA human femoral heads
Fig. 5Quantitative analysis of the CMMC metrics in healthy and early OA human femoral heads. The changes of the (A) CMMC number, (B) maximum caliper diameter (Feret), (C) minimum caliper diameter (MinFeret), and (D) the circularity index (Circ.) at the uppermost 50 μm of the SB are depicted as boxplots. The triangles overlying the boxplots denote the mean values of each subject in different loading areas. Subjects 1–5 were healthy while subjects 6–10 had early osteoarthritic chondral characteristics (n = 5 per group). Each circle in the immediate proximity to the boxplots represents the mean value of a measuring point, color-coded to its corresponding subject (m; number of measuring points = 215 for healthy subjects; 116 for early OA subjects). Analysis was conducted using generalized estimating equation followed by the Bonferroni test for pairwise comparison. Significant p-values are presented as ∗p <0.05; ∗∗p <0.01; ∗∗∗p <0.001.
For the maximum size of the microchannels, it was observed that both cartilage condition (95% WCI = 8.682–31.770, P = 0.002) and region (95% WCI = 64.198 to 86.330, < 0.0001) had significant effects on the Feret diameter. Compared to the non-pathological subjects, the maximum diameter of early OA microchannels were enlarged in all regions; Namely, 36.7% (mean difference = 20.226; 95% WCI for difference = 2.938 to 37.514; P = 0.018), 77.1% (mean difference = 57.271; 95% WCI for difference = 33.234 to 81.308; P = 0.0009) and 78.8% (mean difference = 71.953; 95% WCI for difference = 21.8775 to 122.0295; P < 0.01) increases were observed in the LBR, NLBR, and the PR, respectively. Interestingly, the spread around the Feret's population means were wider in early OA (i.e., higher standard of error) compared to the non-pathological group [Fig. 5(B)]. Regardless of the health status of the cartilage, the smallest microchannels were consistently detected at the LBR of the femoral head, while the CMMCs were significantly larger in the NLBR (β = 37.79, p < 0.0001), and the rim (β = 62.12, p < 0.0001) of the joint. The distribution difference of Feret between NLBR and PR was narrower, but significant nevertheless (β = 24.32, p = 0.023).
Similar observations were made for the influence of cartilage condition (95% WCI = 0.381 to 23.226, p = 0.0045) and region (95% WCI = 36.744 to 59.306, p < 0.0001) on the minimum CMMC caliper diameter. The MinFeret showed a clear increasing trend in early OA compared to the corresponding healthy groups, which was significant at the NLBR (mean difference = 37.81; 95% WCI for difference = 16.73 to 58.89; p < 0.001) and the PR (mean difference = 35.40; 95% WCI for difference = 13.11 to 57.70; p = 0.0014), but not significant at the LBR [p = 0.64; Fig. 5(C)]. Additionally, the MinFeret diameter was generally smallest at the LBR (the reference category in parameter estimation), and was progressively increased at the NLBR (β = 23.31, p = 0.007), and the PR (β = 33.13, p < 0.0001).
The morphology of the SB microchannels was impacted by the region (95% WCI = 0.74 to 0.844, p < 0.0001), but not by the cartilage condition (95% WCI = −0.017 to 0.098, p = 0.078). The CMMCs were generally round and circular in the LBR, while getting increasingly elongated and irregular at the NLBR (β = −0.081, 95% WCI for difference = −0.1136 to −0.0474, p < 0.0001), and especially the rim of the joint (β = −0.131, 95% WCI for difference = −0.1648 to −0.0969, p < 0.0001). Again, we found a higher variance in the early OA subjects compared to their healthy counterparts. In particular, the LBR measuring points were two times more spread out from the mean, and from one another (SEM: 0.026 vs 0.0126) compared to the healthy subjects. When early OA groups were compared to their respective healthy regions, decreasing trends were observed for all groups, which were, however, not significant (LBR: p = 1.0; NLBR: p = 0.052; PR: p = 0.67) owing to the aforementioned variance [Fig. 5(D)].
Discussion
OA is a leading cause of chronic pain and disability worldwide where early pathological changes occur well before the disease is readily diagnosable in clinical settings
. This is due to the low sensitivity of current non-invasive analytical methods, which makes them inadequate to quantify subtle microarchitectural changes
. In subchondral bone, particularly, much is unknown about the direct microstructural pathways that connect AC to the trabecular spacing, and to our knowledge, this is the first study examining early osteoarthritic changes of SB microchannel network in human femoral heads.
High-resolution 3D mapping of SB reveals a complex network of CMMCs that directly connects the medullary cavity and the deep zone of cartilage, where regional differences in the CMMC number and size—irrespective of the health-status of cartilage—seem to reflect regional differences in habitual loading on the joint
. This observation may have several biological interpretations: (1) There is evidence that in normal, non-arthritic joints, the areas of higher cartilage thickness and subchondral bone thickness are colocalized
. In such regions, the length of the diffusion pathway from the subchondral region to the basal cartilage layer is shorter than that from the cartilage surface
, which makes it conceivable that the CMMCs have nutritive functions (i.e., O2 and glucose) at least for deep-lying chondrocytes adjacent to the tidemark
. The abundance of the CMMCs in the central areas of the joint that are generally subject to a higher compressive stress can then be interpreted as a functional adaptation for providing adequate nutritive support
. Interestingly, the density of microcracks—another possible communication pathway for nutrition and cartilage—bone crosstalk—has been correlated to improved cartilage homeostasis as well
. (2) The exchange of pore fluid between the (micro)porous structures of bone due to cyclic mechanical loading and blood pressure has been deemed crucial in nutrient transport
. The effect of such pressure fluctuations could, on the one hand, influence the extent and type of nutrient supply, and on the other hand, the formative stimulus on bone cells in the basal layer of the AC
. It is therefore plausible that the CMMCs are passages for the interstitial fluid movement between the cartilage and the medullary cavity.
In early OA, the minimum and maximum CMMC size is evidently increased compared to intact-cartilage subjects. At the NLBR and the PR, the significance levels are particularly high, and more pronounced for the maximum size than the minimum size. Given that the CMMC number is unchanged compared to the healthy groups, it can be inferred that the porosity of the SB (number per area∗size) is increased in early OA. This result is consistent with reports demonstrating elevation of subchondral plate porosity during early stages of OA development
Osteoarthritis induction leads to early and temporal subchondral plate porosity in the tibial plateau of mice: an in vivo microfocal computed tomography study.
. The increase in the size of SB perforations can be interpreted as a compensatory mechanism to enhance bone-cartilage crosstalk for diffusion of small molecules
Our data suggest that subchondral bone thickness is more affected by the region of the joint than the cartilage condition. One possible reason for this may be the lower number of early OA measuring points in the LBR, which can reduce the overall effect of cartilage condition. Nonetheless, the reduction of SB thickness is mostly local at the NLBR and the PR of the joint, which can be interpreted as evidence for the increased deterioration of bone in early stages of OA. The early phase of OA is known to be associated with structural deterioration, and an early loss of bone due to elevated bone remodeling
Tibiofemoral contact pressures and osteochondral microtrauma during anterior cruciate ligament rupture due to excessive compressive loading and internal torque of the human knee.
We found that only ∼10% of the affected areas on the femoral heads characterized by early cartilage lesions were located at the LBR, which may be surprising considering the fact that increased chronic loads are seen as major contributors to cartilage damage and degeneration
. Nonetheless, it is reported that early cartilage fibrillation and splitting is more commonly seen at the edges of the joint contact regions or at the inferomedial portion of the human femoral head, even though SB exposure in end-stage OA patients usually develops on or near the central, LBRs
. This paradox has puzzled some scientists and led them to believe that focal chondral defects on the inferomedial portion have a small capacity to progress into full SB exposure, while those located superocentrally, have a higher osteoarthritic potential, even though they are less frequent
. It is, thus, equivocal whether partial-thickness chondral lesions are as relevant in the context of OA progression as full-thickness defects. Nevertheless, it has been demonstrated that regardless of defect depth, the prevalence of focal cartilage defects in an isolated sub-region of the tibiofemoral joint increases the risk for development of new chondral defects in unaffected sub-regions of the same joint
Partial- and full-thickness focal cartilage defects contribute equally to development of new cartilage damage in knee osteoarthritis: the multicenter osteoarthritis study.
. Local investigation of partial-thickness chondral lesions and its underlying SB are, therefore, important when exploring preventive measurements or repair strategies of degenerative changes, which possibly should target non-load-bearing and the peripheral areas of the joint
Our preliminary histological analysis shows that most CMMCs have some type of soft-tissue content, e.g., tunica intima with cell linings at their basement membrane. It is frequently reported that at least some of the SB perforations are conduits for blood vessels that are lined by endothelium
. These thin-walled vessels contain elements of the peripheral blood such as erythrocytes [Fig. S6(G)], and are engulfed by concentric layers of bone. Our histological observations, however, point to a much more heterogeneous profile of possible CMMC contents such as small (12–25 μm) and medium-sized (35–55 μm) capillaries, sinusoids, post-capillary venules and arterioles-like structures. These contents may have important implications in bone (patho)physiology, e.g., to allow for the diffusion-driven exchange of nutrients, mediators, growth factors, and waste products in blood and/or the interstitial fluid
. Dedicated and comprehensive studies are needed to evaluate the precise CMMC contents and their quantitative variations regarding the effect of region and cartilage condition. Additionally, it would be interesting to examine possible correlations of CMMCs with those characteristics of the overlying cartilage that typify OA progression, e.g., the percentage of the tidemark duplication length to that of local cartilage thickness.
A limitation of our study is the small sample size in each group (n = 5) and the unknown medical history of the human cadavers. Nonetheless, the biopsy-based mapping of the CMMCs across the entire femoral head at an isotropic voxel size of 1.2 μm was inherently restricting. i.e., for a reasonably higher sample size (e.g., n = 10 per group), either twice as much biopsies would have had to be scanned, which was simply inconceivable given the time needed for scanning and more importantly data processing. Or a compromise would have had to be made in terms of the distance between individual measuring points (i.e., the mapping detail). Here, we decided to give the priority to the mapping detail owing to the high intra-regional variance that was expected, as well as the fact that an imprecise perception of the CMMC characteristics could potentially jeopardize ongoing/future studies on pathological joints, or cause misinterpretations. Therefore, we hope that the small sample size of this study can be evaluated in a context-dependent manner. Another limitation is the evaluation of the CMMC metrics in sequential 2D layers rather than in three-dimensional models. Hence, Feret. and Circ. may also reflect the angle at which the microchannels run through the SB, as slanted channels are cut by a transverse plane of sectioning, which can possibly lead to an overestimation of the values. However, at the uppermost 50 μm of the AC-SB interface where the values are computed, the channels have to merge to the SB surface and are therefore perpendicularly extended, which minimizes this potential effect.
In conclusion, we have profiled and evaluated the influence of cartilage condition as well as local regions of human femoral head on SB microarchitecture in early idiopathic OA. Our findings point to a previously undescribed enlargement of the SB microchannel network especially in the unloaded region of the joint, and a collective structural deterioration of SB in early stages of OA. Still, the AC-SB interface seems to be even more intricate than originally thought, and several questions regarding the functionality and the cellular/molecular content of the CMMCs remain to be answered. Likewise, it would be interesting to assess the evolution of CMMCs during the late-stages of the pathological cycle. Given the biphasic nature of OA progression
, a remodeling sequence may be hypothesized in which the subchondral densification and excessive bone matrix formation occlude the connective pathways between the medullary cavity and the basal layer of cartilage. This can be used as a direction for future research.
Data availability
The data that support the findings of this study are available from the corresponding author (ST) upon reasonable request.
Informed consent
Bone tissue samples were derived from the anatomical gift program of the MHH, and thus all the participants signed the consent form for the whole-body donation executed by the prospective donor.
Author contributions
Conceptualization, ST and AFS; Data curation, ST, TY, RHF, LJ, and ALF; Formal analysis, ST, TY, KOB, RHF, LJ, ALF, and BG; Funding acquisition, WL and AFS; Investigation, ST, TY, KOB, RHF, LJ, ALF, and BG; Methodology, ST, TY, KOB, ALF, BG, TH, WL, and AFS; Project administration, ST and AFS; Resources, WL and AFS; Supervision, AFS; Validation, ST, TY, ALF, TH, BG; Visualization, ST and TY; Writing – original draft, ST, TY, ALF; Writing – review & editing, ST, TY, KOB, RHF, LJ, ALF, BG, TH, WL and AFS.
Conflict of interest
Shahed Taheri and Arndt Friedrich Schilling have filed a patent application with the University Medical Center Göttingen that is partially based on the results reported in the current manuscript. All other authors have nothing to disclose.
Funding
This study was supported by the German Research Foundation (DFG) as part of subproject 5 [SCHI 857/9-1/SCHI 857/9-2] of the Research Consortium ExCarBon FOR2407.
Acknowledgments
The authors would like to thank the anonymous body donors and the Prosektur of the Institute for Functional and Applied Anatomy for their donation of the bone samples. Likewise, we would like to thank Timo Beil and Jan Hubert for classification of the bones according to the Outerbridge system, and Andreas Buchhorn (MHH) for his support with the samples.
Appendix A. Supplementary data
The following are the Supplementary data to this article:
Osteoarthritis induction leads to early and temporal subchondral plate porosity in the tibial plateau of mice: an in vivo microfocal computed tomography study.
Tibiofemoral contact pressures and osteochondral microtrauma during anterior cruciate ligament rupture due to excessive compressive loading and internal torque of the human knee.
Partial- and full-thickness focal cartilage defects contribute equally to development of new cartilage damage in knee osteoarthritis: the multicenter osteoarthritis study.
00879-2&id=gr1.jpg){kind=link}
00879-2&id=gr2.jpg){kind=link}
00879-2&id=gr3.jpg){kind=link}
00879-2&id=gr4.jpg){kind=link}
00879-2&id=gr5.jpg){kind=link}