Alkaptonuria is a genetic disorder of tyrosine metabolism, resulting in elevated circulating concentrations of homogentisic acid. Homogentisic acid is deposited as a polymer, termed ochronotic pigment, in collagenous tissues, especially cartilages of weight-bearing joints, leading to a severe osteoarthropathy. We undertook this study to investigate the initiation and progression of ochronosis from the earliest detection of pigment through complete joint failure.
Nine joint samples with varying severities of ochronosis were obtained from alkaptonuria patients undergoing surgery and compared to joint samples obtained from osteoarthritis (OA) patients. Samples were analyzed by light and fluorescence microscopy, 3-dimensional scanning electron microscopy (SEM), and the quantitative backscattered electron mode of SEM. Cartilage samples were mechanically tested by compression to determine Young's modulus of pigmented, nonpigmented, and OA cartilage samples.
In alkaptonuria samples with the least advanced ochronosis, pigment was observed intracellularly and in the territorial matrix of individual chondrocytes at the boundary of the subchondral bone and calcified cartilage. In more advanced ochronosis, pigmentation was widespread throughout the hyaline cartilage in either granular composition or as blanket pigmentation in which there is complete and homogenous pigmentation of cartilage matrix. Once hyaline cartilage was extensively pigmented, there was aggressive osteoclastic resorption of the subchondral plate. Pigmented cartilage became impacted on less highly mineralized trabeculae and embedded in the marrow space. Pigmented cartilage samples were much stiffer than nonpigmented or OA cartilage as revealed by a significant difference in Young's modulus.
Using alkaptonuria cartilage specimens with a wide spectrum of pigmentation, we have characterized the progression of ochronosis. Intact cartilage appears to be resistant to pigmentation but becomes susceptible following focal changes in calcified cartilage. Ochronosis spreads throughout the cartilage, altering the mechanical properties. In advanced ochronosis, there is aggressive resorption of the underlying calcified cartilage leading to an extraordinary phenotype in which there is complete loss of the subchondral plate. These findings should contribute to better understanding of cartilage–subchondral interactions in arthropathies.
Alkaptonuria is a rare autosomal-recessive condition caused by deficiency of a single enzyme in the tyrosine metabolic pathway (1, 2). The deficiency of the enzyme homogentisate 1,2-dioxygenase results in an inability to break down homogentisic acid (3, 4). Even with efficient renal clearance there are elevated circulating levels of homogentisic acid in the blood and body tissues (5). This results in a triad of clinical features. The earliest and most obvious presentation is a darkening of urine due to the presence of homogentisic acid. Over many years, homogentisic acid polymerizes and becomes deposited by an unknown mechanism in collagenous tissues, a process termed ochronosis (4–6). Recent work has shown that initial pigmentation is associated with the periodicity of fibrillar collagen, possibly suggesting that there is a preferential binding site for the ochronotic pigment on the collagen fibers (7). Although cartilage is a preferential site for ochronosis, pigmentation is observed in other tissues, such as the submandibular gland, even in the absence of cartilaginous matrices (5, 8, 9). With increased exposure to homogentisic acid and continued polymerization in joint tissues, the end stage of the disorder is a severe osteoarthropathy. Although there is no proven link between the presence of homogentisic acid and shortened lifespan, the quality of life is severely affected, and the only current treatments to have any proven clinical benefit are either liver or renal transplantation or more commonly joint replacement surgery (10–14).
Herein we describe the progression of articular cartilage degeneration in ochronotic arthropathy from initiation of pigmentation to complete joint failure. We have previously suggested that the presence of homogentisic acid alone is not the determining factor in pigment deposition (8). Here we provide evidence that intact tissues are initially resistant to pigmentation and that focal cellular and extracellular matrix (ECM) abnormalities cause joint tissues to be susceptible to pigmentation. Pigmentation initially manifests at the boundary of the subchondral bone and calcified cartilage before proceeding toward the articular surface.
MATERIALS AND METHODS
Study samples and patient demographics.
All tissues (9 joint tissues from 4 hips and 5 knees of 8 alkaptonuria patients) were obtained as surgical waste under ethical approval from the Liverpool Research Ethics Committee (REC) with informed consent from alkaptonuria patients undergoing joint replacement (further details are available at http://findakure.org/calcified-cartilage/). A separate sample of osteoarthritic (OA) tissues was obtained as surgical waste under ethical approval from the Liverpool REC with informed consent from patients with clinically diagnosed OA undergoing joint replacement (further details are available at http://findakure.org/calcified-cartilage/).
Load-bearing joint tissues (hips and knees) were fixed immediately at the time of surgery in 10% phosphate buffered formol saline. Pieces of tissues were processed routinely for histology by paraffin embedding. Five-micrometer sections were cut and stained with hematoxylin and eosin (H&E) or left unstained. Sections were then dehydrated through graded alcohols and mounted in DPX (Sigma). Direct-view stereoscopic fluorescence microscopy was conducted on histologic sections and scanning electron microscope (SEM) blocks (see below) using an Edge 3-dimensional (3-D) microscope (Edge Scientific Instrument Corporation) fitted with twin Canon digital cameras (15).
Mechanical loading of cartilage.
Cartilage cubes were obtained at the time of surgery from the medial femoral condyles of alkaptonuria and OA patients undergoing joint replacement for ochronotic osteoarthropathy or osteoarthropathy. Unfixed tissues were washed in sterile phosphate buffered saline (PBS) immediately following surgery. Cubes of articular cartilage were dissected from the articular surface using a No. 10 scalpel (Swann Morten). Cartilage cubes were removed from the sample as close to the subchondral bone as possible and were then measured using Vernier calipers. The dimensions of the samples obtained were 3 mm × 3 mm × 4 mm (depth × length × width). The dimensions were recorded and used in the stress (length × width) calculation as area, and the depth measurement was used to further calculate the strain. Samples were stored overnight at 4°C in PBS (pH 7.4). Wet samples were placed on a 9-cm petri dish. Excess saline was removed from the surface and around the periphery of the tissue. Samples were mechanically tested by compression on a Nene M5 tensile tester using Nene software with a 500N-load cell. Samples were loaded at a rate of 0.5 mm/minute. Data were recorded as load displacement. The stress- strain curves were plotted for each sample, and Young's modulus was calculated.
Bone slices were embedded in polymethylmethacrylate; the block surfaces were trimmed and polished and coated with carbon by evaporation. They were imaged using backscattered electrons in a Zeiss DSM962 SEM with external computer control. Images were recorded at 20 kV, 0.5-nA probe current, and 17-mm working distance using halogenated dimethacrylate standards to calibrate the electron backscattering (16). Other slices were macerated in 4% Tergazyme (alkaline bacterial pronase enzyme detergent; Alconox) to remove all cells and normal hyaline cartilage matrix and leave all calcified matrices. After washing and drying, these slices were carbon coated from both sides and imaged using 3-D backscattered electron mode of SEM.
Statistical differences were determined by applying the Kruskal-Wallis test with Dunn's multiple comparison test using GraphPad Prism 5 software.
Examination of samples revealed a spectrum of pigmentation from focal ochronosis to complete pigmentation across all zones of cartilage. In the samples that were not completely pigmented, the focal deposits were in regions known to be subjected to the greatest load bearing during locomotion (e.g., knee medial tibial plateau compared to lateral tibial plateau) (17). We also observed that pigmentation of the articular surface was central rather than peripheral in advanced ochronosis (Figure 1). These are the region(s) associated with initiation and severe pathology in primary knee OA (18). In this series, hip samples showed more severe pigmentation than knee samples. The medial tibial plateau of heavily pigmented samples showed high mottling and fissuring of the articular surface (results not shown). Routine processing of samples for SEM revealed that the enzymatic action of Tergazyme, which effectively digests noncalcified ECM, was unable to break down the pigmented hyaline articular cartilage. Dissection revealed that the pigmented cartilage was difficult to remove from the bone and was much stiffer than nonpigmented cartilage (findings that were subsequently confirmed by mechanical testing) (Figure 1E).
Loading of articular cartilage cubes from freshly surgically resected medial femoral condyles revealed large differences in Young's modulus between heavily pigmented cartilage and nonpigmented and OA cartilage (Figure 1E). The average Young's modulus of pigmented cartilage (156 MPa) was >5-fold greater than that of nonpigmented cartilage and OA cartilage (22.4 MPa and 26.3 MPa, respectively; P < 0.01). This demonstrates a large increase in stiffness attributed to the presence of ochronotic pigment in the cartilage of patients with ochronotic osteoarthropathy.
Histologic analysis of H&E-stained sections confirmed macroscopic observations regarding the presentation of pigmentation. In partially pigmented samples, ochronotic pigmentation was associated with individual chondrocytes in the calcified cartilage and included both intracellular and pericellular pigmentation of the chondrocyte lacunae and territorial matrix (Figures 2A and C). At this stage, there was no pigmentation in other cartilage zones and no evidence of advanced arthritic changes at the articular surface (Figures 2A and B).
Samples with more widespread ochronosis had a greater frequency of chondrocyte intracellular pigmentation in the calcified cartilage and territorial matrix, but pigment was absent from the interterritorial matrix. As pigmentation of chondrocytes and territorial matrix increased further, there was still no evidence of pigmentation in the hyaline cartilage and at the articular surface. When the pigmentation front extended into the noncalcified hyaline cartilage matrix, the deep and middle zones had blanket pigmentation (areas where the cartilage matrix had no nonochronotic regions), and the territorial matrix in these zones was even more intensely pigmented. However, alongside this widespread pigmentation the interterritorial matrix of the calcified cartilage matrix was nonpigmented. Pigmentation was also absent from the superficial zone and articular surface.
Hyaline cartilage began to show fissuring at the deep surfaces close to the subchondral interface. Pigmentation of the interterritorial matrix and the articular cartilage was observed only in samples with the most advanced and widespread pigmentation. In addition, the most pigmented samples had dense pigmentation of the ECM in all deep and middle zones of the hyaline cartilage. No calcified cartilage was present, and the subchondral bone was absent from almost all of the section in these most pigmented samples. In regions where subchondral bone was still present, osteoclasts could clearly be seen resorbing the thin bone matrix to the point of contact with the pigmented hyaline articular cartilage (Figure 3A). Pigmented cartilage was adjacent to, and impacted onto, the trabecular structures, which were thin and showed evidence of osteoclastic resorption. The pigmentation in the calcified cartilage was consistently associated with the territorial matrix and at no stage showed the pattern of blanket pigment deposition observed in the noncalcified hyaline cartilage matrix.
There was a general lack of pigment in bone matrix in all samples, but intracellular ochre-colored deposits were observed in osteocytes and their lacunae (Figure 3B). Pigmentation could also be seen in the canalicular network. In some sections, osteoclasts were seen engulfing pigmented osteocytes as they were released from the surrounding matrix (Figure 3C). Pigmentation was seen in the osteocytes from all samples showing cartilage matrix pigmentation.
In samples with advanced ochronosis, shards of pigmented cartilage were embedded in the marrow space cavity between trabeculae. The ochronotic fragments were surrounded by fibrous tissue, mononuclear cells, and multinuclear cells all showing intracellular ochronotic pigmentation. Shards of pigmented cartilage, lost from the articular surface, could be observed embedded in synovium; this was present in 8 of 8 samples with synovial tissue (Figure 3D).
Examination of the most severely pigmented samples revealed 2 types of pigmentation in the matrix (Figure 4). There was blanket pigmentation in the deep zone, which was more intense around the territorial matrix (Figure 4D). In some samples in the superficial and middle zones, there was a granular rather than blanket pigmentation with a ground pepper–like appearance (Figure 4C). This was also observed in the chondrocyte lacunae and the chondrocyte cytoplasm (Figure 4).
Fluorescence microscopy findings.
Consistent with findings on histologic examination, fluorescence microscopy confirmed the presence of ochronotic pigment in the articular cartilage, commencing in the calcified zone and being absent from the articular surface, with no notable arthritic changes. There was also confirmation of the lack of calcified cartilage and loss of subchondral bone in the most pigmented sample(s). In some regions, pigmented ochronotic cartilage projected deep into the bone domain. The cartilage at the articular surface exhibited high fluorescence as pigmentation progressed, while the deeper zones, where pigment had been deposited, hardly fluoresced and appeared dull ochre in color. Bone exhibited a high autofluorescence except in the osteocytes, which exhibited the same coloration as the pigmented chondrocytes and their pericellular matrix (Figure 3E). Other osteocytes exhibited no fluorescence. Interestingly, in the transmitted light examination of the unstained histologic sections, there appeared to be no pigment associated with the bone matrix, but regions of the bone matrix exhibited coloration similar to the ochre-colored matrix under fluorescent light, indicating the presence of pigment.
Fluorescence microscopy also confirmed the presence of ochronotic shards of cartilage in the marrow cavity, surrounded by fibrous tissue that appeared to engulf the fragmented shards. The fragments of cartilage were often seen enclosed within thick fibrous bands of tissues bridging from one trabecular structure across to another.
Three-dimensional backscattered electron mode of SEM analysis in the most pigmented, macerated sample showed the complete absence of the subchondral bone plate. There was no calcified cartilage present in the sample (Figures 1C and 5A). Pigmented hyaline articular cartilage appeared as the dark (because it is not mineralized) substance attached to the trabecular bone structure that had survived intact through the routine processing of the sample (Figures 5A and B).
Two-dimensional compositional quantitative backscattered electron mode of SEM imaging of the most pigmented sample revealed mineralized matrix with numerous dead osteocytes, identified due to their dense lacunar mineralization. The bone outline could clearly be observed, but no mineralized cartilage or subchondral bone plate was present. Overlaying topographic and compositional images showed the presence of nonmineralized bone where the subchondral junction should have been (Figure 5C). The bone–cartilage interface appeared as noncalcified cartilage in contact with nonmineralized bone.
Quantitative backscattered electron mode of SEM analysis revealed a range of mineralization densities within normal limits in regions of trabecular bone (Figure 5D) as well as completely nonmineralized bone matrix patches. In the most severely pigmented sample, there was no mineral content in any region of the cartilage, confirming the absence of cartilage mineralization and loss of any preexisting calcified cartilage. The mineralized bone trabeculae showed regions of lower or zero mineralization density that were in contact with the nonmineralized tissues (Figure 5D) and also showed evidence of osteoclastic action in the form of resorption lacunae (Figure 5D). We conclude that pigmented, nonmineralized hyaline articular cartilage was confluent with nonmineralized bone.
Three-dimensional backscattered electron mode of SEM of femoral condyles from alkaptonuria and OA patients revealed an interesting comparison. An OA femoral condyle showed little evidence of damage to the articular cartilage with no sign of fissures or abrasions on the surface (results not shown). The subchondral architecture revealed widespread sclerosis with subchondral trabeculae showing distinctive thickening, particularly in the midline. There was no evidence of bone marrow lesions in this sample (Figure 6A). A femoral condyle from an alkaptonuria patient showed incomplete cartilage pigmentation, with the pigment present extensively in the cartilage matrix but not completely up to the articular surface (results not shown). The subchondral architecture showed evidence of thinning with a bone marrow lesion in the periphery of the condyle. The trabeculae showed no sign of sclerosis comparable to that seen in the OA sample (Figure 6B). The most severely pigmented sample from another alkaptonuria patient was a femoral head that showed a focal bone marrow lesion toward the periphery of the sample. There was sporadic evidence of sclerosis, but the most interesting finding was the complete absence of the subchondral plate (Figure 6C).
We describe the deposition of ochronotic pigment in articular cartilage from the initial focal deposition associated with individual chondrocytes in calcified cartilage, through proliferation of pigment throughout the hyaline cartilage and associated resorption of the subchondral plate ending in complete joint destruction. Our findings add weight to the theory that ECM is normally resistant to ochronosis but may become susceptible to pigmentation in response to tissue injury. The type of injury might include biochemical or mechanical damage including microtrauma (19). Pigmentation appears to protect collagen fibers from the action of proteolytic enzymes; however, the increased stiffness due to the presence of pigment is likely to make the matrix more susceptible to mechanical damage through normal loading, causing shards to break off and become embedded in the synovium (Figure 3D).
Although it has previously been suggested that the calcified matrices do not undergo ochronosis (20), our results suggest that the initial pigmentation starts deep in the cartilage, at or close to the mineralizing front, associated with individual chondrocytes and their territorial matrices. The conditions that allow pigmentation have not been identified but could be related to mechanical or oxidative damage or to alteration in chondrocyte gene expression. The relationship between intracellular and extracellular pigmentation is not understood. Focal pigmentation is always associated with pigmented cells, indicating that the cellular action is required for the initiation of pigmentation. However, even though pigmentation might commence intracellularly, there is no doubt that it can proceed extracellularly, and the vast majority of ochronosis arises from the spread of pigmentation to established matrix—not by secretion of newly pigmented matrix. Once pigmented, ochronotic matrix appears to be resistant to turnover by matrix-degrading enzymes, as demonstrated by its resistance to Tergazyme and the presence of ochronotic cartilage shards embedded in the synovial tissues, with multinuclear cells present in the vicinity but without evidence to suggest that they are degrading the ochronotic tissue.
The route by which homogentisic acid reaches the site of initial pigmentation is not clear. It could diffuse from the synovial fluid, where its concentration is thought to be similar to that in plasma, through the hyaline cartilage, or it could permeate through the calcified cartilage from the underlying subchondral bone, consistent with the transit of other low molecular weight compounds (21).
Observations of samples with more advanced ochronosis indicate that once the initial pigment has been deposited, there is proliferation to other chondrocytes and then to interterritorial matrix. This proliferation of pigmentation suggests that focal ochronosis may alter the biomechanical and/or structural properties of adjacent tissue, making it susceptible to pigmentation. This additional pigmentation leads to further damage and a downward spiral of tissue destruction. Dissection revealed that pigmentation alters the mechanical properties of cartilage, making it stiffer and hard to cut, indicating an increase in Young's modulus that has been confirmed by mechanical testing (Figure 1E). We speculate that the initial, focal changes in mechanical properties may lead to stress concentrations or risers, which could consequently induce further mechanical damage in adjacent tissue (22, 23). Mechanical damage and/or inflammation is known to up-regulate stress genes in early OA (24). Pigmentation is shown to begin in the pericellular and territorial matrices of individual chondrocytes, and thus matrix turnover events in these regions could be an important factor for initiation of pigmentation (25) and may provide evidence for why pigmentation commences around the cells prior to moving into the interterritorial matrices. This is one possible mechanism for the proliferation of pigmentation, but we cannot rule out other mechanism(s) including chemical damage to the matrix by, for example, oxidative stress resulting from the initial deposition of pigment (26).
In association with extensive hyaline cartilage pigmentation, there is remodeling of the calcified cartilage and underlying bone leading to a presentation reminiscent of bone marrow lesions. Eventually the aggressive resorption leads to complete loss of the subchondral plate. This extraordinary phenotype has not previously been reported in pathologic conditions. This is possibly a consequence of stress shielding of the underlying calcified cartilage beneath the stiff shell of pigmented cartilage, with enhanced osteoclastic activity resorbing the unloading subchondral bone and calcified cartilage. There is thinning and reduced mineralization of some underlying trabeculae and sclerotic changes in others similar to that previously reported in bone marrow lesions in OA (27–29). Although the aggressive resorption appears to be focal, it is noteworthy that enhanced urinary excretion of crosslinked N-telopeptides of type I collagen has been reported in alkaptonuria patients (30).
Although remodeling of subchondral bone and articular cartilage has been reported in rheumatoid arthritis (31), the complete loss of the subchondral bone seen in alkaptonuria is exceptional. Although in a previous study it was concluded that bisphosphonates were not beneficial in patients with advanced ochronosis (32), our findings suggest that bisphosphonate therapy may be worthy of reconsideration, with administration either before significant subchondral bone and calcified cartilage loss (i.e., primary prevention) or during the early phase of subchondral bone and calcified cartilage loss (secondary prevention/treatment).
As a consequence of the subchondral bone loss, the stiff pigmented cartilage becomes impacted onto the less well mineralized trabecular bone. Fracturing of the cartilage causes fragments to be displaced into the marrow space, engulfed by the marrow cells, and embedded in the synovium and capsular tissues (12, 33). The end stage of the disorder shows complete loss of pigmented cartilage from the articular surface (Figures 1D and 6B and C).
Pigmentation was not detected in mineralized bone matrix, but osteocytes, osteoclasts, and osteoblasts all displayed intracellular pigmentation, as did the canalicular network and osteocyte lacunae. Osteoclasts were also seen phagocytosing pigmented osteocytes. Although generalized rates of bone formation have been observed to be almost normal in ochronotic individuals (30), we observed focally lower bone mineralization density on the periphery of trabeculae. There was no evidence of generalized osteomalacia, and previous studies show no evidence of vitamin D deficiency in alkaptonuria patients (34).
This research clearly highlights the central role that the bone–cartilage interface plays in the initiation of ochronosis in alkaptonuria, and the importance of the calcified cartilage and subchondral bone in the subsequent arthropathy. The absence of the calcified cartilage itself is of major pathologic significance. Calcified cartilage has often been neglected in the study of joint degeneration in OA, but it is clearly the tissue through which loads are distributed from the hyaline articular cartilage to the underlying subchondral bone (35). There are, however, many reports highlighting a role for subchondral bone in the pathogenesis of OA (36). It has been suggested that the structural integrity of articular cartilage relies on normal subchondral bone turnover, intact chondrocyte function, and normal biomechanical stresses (36–38). Alkaptonuria has already been documented as mimicking or causing premature OA (39, 40).
The examination of our results shows that there may be a large overlap of OA with the pathogenesis of alkaptonuria, and there is also growing evidence that in OA, alterations in the composition and biomechanical properties of cartilage are facilitated by increased bone turnover (36, 38, 41). It is tempting to speculate that the initial pigmentation is associated with preexisting, possibly OA-related changes that make the cartilage matrix susceptible to pigmentation, and that homogentisic acid is acting as an endogenous marker of degenerative changes. Joint arthropathy in alkaptonuria has some parallels with other joint diseases, including OA, but also has some unique characteristics. Further study of this rare disorder may contribute to the understanding and treatment of joint degeneration, not only in alkaptonuria but also in OA.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Taylor had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Taylor, Boyde, Hunt, Ranganath, Gallagher.
Acquisition of data. Taylor, Boyde, Wilson, Hunt, Gallagher.
Analysis and interpretation of data. Taylor, Boyde, Wilson, Jarvis, Davidson, Hunt, Ranganath, Gallagher.
The authors are extremely grateful to Mo Arora and Brenda Wlodarksi for their technical assistance in sample preparation. We would like to thank Dr. Kirstie Andrews for her help in mechanical testing of the cartilage samples and Jane Dillon for reading the manuscript. Finally, we are extremely grateful to the patients who were so generous in donating their tissues for this research.