SEARCH

SEARCH BY CITATION

Keywords:

  • osteoarthritis;
  • non-enzymatic glycation;
  • canine;
  • pentosidine;
  • proteoglycan

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Osteoarthritis is a highly prevalent disease, age being the main risk factor. The age-related accumulation of advanced-glycation-endproducts (AGEs) adversely affects the mechanical and biochemical properties of cartilage. The hypothesis that accumulation of cartilage AGEs in combination with surgically induced damage predisposes to the development of osteoarthritis was tested in vivo in a canine model. To artificially increase cartilage AGEs, right knee joints of eight dogs were repeatedly injected with ribose/threose (AGEd-joints). Left joints with vehicle alone served as control. Subsequently, minimal surgically applied cartilage damage was induced and loading restrained as much as possible. Thirty weeks after surgery, joint tissues of all dogs were analyzed for biochemical and histological features of OA. Cartilage pentosidine levels were ∼5-fold enhanced (p = 0.001 vs. control-joints). On average, no statistically significant differences in joint degeneration were found between AGEd and control-joints. Enhanced cartilage pentosidine levels did correlate with less cartilage proteoglycan release (R = −0.762 and R = −0.810 for total and newly-formed proteoglycans, respectively; p = 0.028 and 0.015 for both). The current data support the diminished cartilage turnover, but only a tendency towards enhanced cartilage damage in AGEd articular cartilage was observed. As such, elevated AGEs do not unambiguously accelerate the development of early canine OA upon minimal surgical damage. © 2012 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 30:1398–1404, 2012

Osteoarthritis (OA), with a high prevalence and increasing incidence due to the aging population, having a large impact on the patient's quality of life, is characterized by progressive cartilage damage, bone changes, and secondary synovial inflammation.1

As yet, the pathogenesis of OA is largely unknown. Several factors have been reported to predispose to the development of OA, such as genetic background, overweight, joint laxity, and muscle weakness.2 However, undisputedly, the most important risk factor for development of OA is age.3 The incidence of OA increases strongly with age: >50% of the population over 60 years of age is affected.4, 5 However, there are still many uncertainties how age contributes to the onset and progression of OA. Age-related changes in the articular cartilage are suggested to play an important role in the susceptibility of cartilage to OA.

One of the major age-related changes in articular cartilage is the spontaneous modification of proteins by non-enzymatic glycation resulting in the accumulation of advanced glycation endproducts (AGEs). Non-enzymatic glycation is a post-translational modification of proteins by reducing sugars. The spontaneous condensation of reducing sugars with free amino groups in lysine or arginine residues on proteins leads to the formation of AGEs.6 Pentosidine, a fluorescent AGE formed between lysine and arginine residues, is frequently used as marker for AGEs. AGEs are formed in all proteins, and can only be removed from the tissue when the protein is removed. The low turnover of proteins in articular cartilage results in an abundant accumulation of AGEs in this tissue with increasing age.7–9

AGEs are known to affect physical and chemical properties of proteins. Tissue strength is dependent upon the amount of crosslinks present10 and accumulation of AGEs is correlated with increased tissue stiffness of cartilage. Moreover, an increase in AGEs makes the cartilage more brittle,9, 11 making the tissue more prone to mechanical damage. In addition to mechanical changes AGEs also interfere with cellular processes. It has been demonstrated that increased AGE levels lead to decreased proteoglycan turnover (synthesis and release) of articular cartilage.12, 13 Altogether, these data suggest that increasing levels of AGEs in cartilage tissue, renders the tissue more prone to damage and limits repair activity, adding to development and progression of cartilage damage as seen in osteoarthritis.

Studying the effect of AGEs on cartilage damage in vivo, independent of age, necessitates animal models in which AGE levels of articular cartilage can artificially be enhanced by supplying high amounts of reducing sugars. This method was used in the canine anterior cruciate ligament transection (ACLT) model of OA.14 AGE levels in the knee cartilage of beagle dogs were artificially enhanced ∼5-fold by repeated intra-articular ribose injections. In this model, all joints developed OA due to joint instability but cartilage damage was more severe in the joints with artificially enhanced AGE levels than in the PBS-injected joints. As such this study supported the role of AGEs in progression of cartilage damage.

In the present study the role of AGEs in the development (initiation) of OA was studied. For this purpose minimal surgically applied chondral damage was applied. By use of minimal chondral damage and restraining joint loading it was anticipated that OA in normal joints would not develop spontaneously. Joints with artificially enhanced cartilage AGE levels were hypothesized to be more sensitive to the OA-inducing trigger and were expected to develop OA.

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Animals

Thirteen female Beagle dogs (age 16.4 ± 3.5 months and weight 9.5 ± 1.7 kg) were obtained from the animal laboratory of the Utrecht University, the Netherlands. They were fed a standard diet and had water ad libitum.

Enhancement of AGE Levels

After 1 week of acclimatization, the right knee of eight animals was injected with a combination of 300 mM ribose and 5 mM threose in PBS (total volume of 2 ml; twice weekly for 7 weeks; further referred to as “AGEd-joint”). The left contra-lateral knee was injected at the same moment with same volumes of PBS (further referred to as “PBS-injected joints”). All intra-articular injections were performed under a short sedation (Dormitor®/AntiSedan® Pfizer Animal Health, Louvain-la-Neuve, Belgium). After the last intra-articular injection, dogs were allowed to recuperate for 2 weeks.

Induction of OA

Subsequently, the femoral cartilage of both joints was surgically damaged according to procedures used for the canine Groove model, but less severe.15–18 Surgery was carried out through a 2–2.5 cm medial incision close to the patellar ligament of the knee. Bleeding and soft tissue damage was prevented as much as possible to avoid dominance of a surgically induced inflammatory component. In utmost flexion, four grooves (instead of normally at least 10 in the classical Groove model) were applied on the weight-bearing parts of the femoral condyles without damaging the subchondral bone. The tibial plateau was left untouched. The animals were permanently housed individually in indoor pens (3.5 m2 per animal) to keep movement (and with that loading) of the joints to a minimum. To put in perspective, normally the animals are led out on a patio of 200 m2 for 4 h/day in large groups. Cartilage pentosidine levels (as a measure of AGE) and severity of joint degeneration were evaluated 30 weeks after surgery.

For negative and positive controls, in the five additional dogs (not artificially AGEd), OA was induced according to the classical Groove model in the right joint using a minimum of 10 grooves and forced loading of the affected joint by fixing the control joint (further untreated) to the trunk of the animal for a few hours a day, 3 days a week. Only biochemical features of joint degeneration were evaluated 20 weeks after induction.15–18

The Utrecht University Medical Ethical Committee for animal studies approved the study.

Outcome

After euthanizing, both hind legs were amputated immediately postmortem. High resolution photographs were taken of cartilage and synovial tissue for macroscopic evaluation. Subsequently synovial tissue and cartilage of both joints were collected and processed within 2 h. Procedures were carried out under laminar flow conditions.

Pentosidine Levels

Cartilage pentosidine levels as a measure of AGE accumulation were measured as described previously.19 In short, at least three randomly taken cartilage samples of each joint were pooled and reduced overnight, sequentially treated with L-cysteine (5 mM, Sigma), EDTA (50 mM, Sigma), NaOH (1 M, Merck), and papain (3% v/v. Sigma). Collagen pentosidine content and amino acid composition were determined by high-performance liquid chromatography (HPLC) according to standard procedures.20, 21 The pentosidine content of collagen samples is expressed as millimoles per mole collagen, assuming 300 hydroxyproline residues per triple-helical collagen molecule.

Macroscopic cartilage damage and synovial tissue inflammation were evaluated on digital high-resolution photographs, by two observers unaware of the source of the photographs, using criteria as described previously.17 Severity of cartilage damage was graded from 0 to 4: 1 = smooth surface, 2 = roughened, 3 = slightly fibrillated, 4 = fibrillated. The scores of the two observers were averaged for the tibial joint surface (maximum score 4) and synovial tissue (maximum score 6); average values subsequently used for statistical evaluation.

Histological cartilage damage and synovial tissue inflammation were evaluated using samples from the weight bearing tibial plateau, and from the infra-patellar synovium, from predefined locations as described before.15 Fixed samples were sectioned and stained with safranin-O-fast-green-iron-hematoxylin and hematoxylin–eosin for cartilage and synovium, respectively. Sections were scored blinded and in random order by two independent observers using the OARSI score for cartilage (max. = 36) and synovium (max. = 18).22 The scores were averaged for the specimens from the tibial surface (n = 4, each) and synovial tissue (n = 3) and for the two observers; average values of each joint were subsequently used for statistical evaluation.

Chondrocyte Activity

Cartilage samples of the tibial plateau were cultured according to standard procedures.17 Cartilage proteoglycan content, -synthesis, -retention of newly formed proteoglycans, and -release were determined for eight explants per parameter of fixed weight bearing locations with identical locations at the contralateral joint.17 All samples were handled individually. The average result of the eight samples was taken as representative of that tibial joint surface and was used for statistical analysis.17

PG Content

To measure the PG content of the cartilage samples, the amount of tissue glycosaminoglycans (GAGs) was determined. The GAGs in the papain digest of cartilage samples were precipitated, stained with Alcian Blue, and quantified photometrically by the change in absorbance at 620 nm with chondroitin sulphate as reference. Values were expressed per wet weight of the cartilage tissue (mg/g).

Proteoglycan Turnover

As a measure of proteoglycan (PG)-synthesis, the rate of sulphate incorporation was determined ex vivo. After 1 h of pre-culture, 370 kBq equation image in 10 µl DMEM was added to each sample. After 4 h labeling, the cartilage samples were washed three times with medium for 45 min. Subsequently, samples were cultured for 3 days without label. Cultures were stopped by washing two times with cold [phosphates-buffered saline (PBS)] and freezing of the samples.

GAGs in a papain digest of cartilage samples were precipitated with Alcian Blue and equation image labeled GAGs were measured by liquid scintillation counting. The total sulphate incorporation rate of each cartilage sample was calculated using specific activity of the medium and was normalized for labeling time and wet weight of the explants (nmol/h/g).

Release of Newly Formed Proteoglycans

To determine the release of newly synthesized PGs as a measure of the retention of these PGs, the release of equation image labeled GAGs in 3-day culture medium was measured. GAGs were precipitated from the medium with Alcian Blue, as described.17 The equation image labeled GAGs were measured by liquid scintillation counting and the release was calculated using the specific activity of the medium normalized to the wet weight of the explants. The release of newly formed PGs is corrected for the synthesis rate and expressed as percentage release of newly formed PGs in 3 days (% new PG release).

PG Release

For the total release (loss) of PGs, GAGs in the culture medium were precipitated and stained with Alcian Blue and quantified as described above. The total amount of GAGs released is expressed as a percentage of the PG-content (% GAG release).

Calculations and Statistics

For each animal, a single value was obtained for each parameter by averaging the multiple analyses. In total, eight animals were used and as such this study might be considered explorative. Importantly, for cartilage parameters only the surgically untouched tibial plateau cartilage was analyzed to prevent interference of the surgical damage, which was applied to the femoral cartilage only. These values of each animal were averaged and mean values of eight (or 5 for the positive and negative control) animals ± SD are presented. To analyze differences between the PBS-injected and AGEd-joints a non-parametric paired Wilcoxon test was used. To analyze differences with the (positive and negative) control joints the non-parametric non-paired Mann–Whitney U-test was used. For comparison of differences in chondrocyte activity parameters with differences in pentosidine levels spearman correlation was used; all two-sided p-values ≤ 0.05 considered statistically significant.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Pentosidine Levels

Intra-articular injections with ribose/threose resulted in enhanced pentosidine levels in each individual dog. On average there was a fivefold higher cartilage pentosidine level in the AGEd-joints compared to PBS-injected joints (p < 0.001, Fig. 1 solid line), with a clear variation between dogs.

thumbnail image

Figure 1. Cartilage pentosidine levels of the joints. Each line represents the difference in pentosidine levels for the PBS-injected joints (left) and sugar injected joints (right). The solid line represents the change in mean value (n = 8) between PBS and sugar injected joints. The amount of pentosidine is statistically significant (p < 0.001) higher in the AGEd-joints compared to the PBS-injected joints.

Download figure to PowerPoint

Cartilage Integrity

Thirty weeks after induction of experimental OA, the affected knees of all animals showed minimal macroscopic cartilage damage (Fig. 2A, light and dark grey bars; score of 1 of a max score of 4). Figure 2B shows a representative micrograph of cartilage from the surgically untouched tibial plateau of the PBS (top) injected and AGEd (bottom) joints. Figure 2C shows representative histology of the tibial plateau of the PBS (top) and AGEd (bottom) joints. The average microscopic OARSI score is shown in Figure 2D (light and dark grey bars). Also microscopy showed a minimal joint degeneration, OARSI score of 3–5 of a max score of 36. No difference between the macroscopic and microscopic scores of the AGEd and the PBS-injected joints were found (Fig. 2A,D, respectively). Both macroscopy and histology of the cartilage tissue was compared with data of the classical Groove model (white and black bars). For both scores the outcome lies in between the control and experimental joints.

thumbnail image

Figure 2. Cartilage integrity. (A) Average macroscopic cartilage score; (B) representative photographs of tibial surface obtained from PBS-injected (top) and AGEd (bottom) joints of tibial plateau; (C) representative micrographs of cartilage histology obtained from PBS-injected (top) and AGEd (bottom) joints of tibial plateau. Note the mild fibrillated surface, mild loss of safranine-O staining and minimal cell clustering as characteristics of very mild degree of joint degeneration; and (D) microscopic cartilage score, of tibial cartilage from PBS-injected and AGEd-joints. Bars represent mean ± SD of n = 8 animals per group. Data of the present experiment are presented in the context of a negative control (left white bar; Groove model) and experimentally induced OA as appositive control (right and black bar; Groove model; n = 5). Light grey and dark grey bars in the middle represent the PBS-injected and AGEd-joints of the present experiment, respectively (n = 8). p-values between the PBS-injected and the AGEd-joints are indicated, considering p ≤ 0.05 as statistically significant different.

Download figure to PowerPoint

Synovial Inflammation

The macroscopically judged synovial inflammation in the PBS-injected and the AGEd-joints was on average low to moderate (Fig. 3A, [light and dark grey bars]; score of 2 of a max score of 5), indicating that synovial inflammation was minimal, slightly enhanced expectedly due to the repeated intra articular injections. Also the microscopic histological score of the synovium was low to moderate (Fig. 3D, [light and dark grey bars]; score of 4–5 of a max score of 18). Representative micrographs are given in Figure 3B for the macroscopy and in Figure 3C for histology. On average no differences in macroscopic and microscopic score (Fig. 3A,D, respectively) were found for the AGEd and the PBS-injected joints. Both macroscopy and histology of the synovial tissue was compared with data of the classical Groove model (white and black bars). For both scores, the outcome lies in between the control and experimental joints, although for the macroscopic score this is less evident.

thumbnail image

Figure 3. Synovial inflammation. (A) Average macroscopic cartilage score; (B) representative photographs of synovial tissue obtained from PBS-injected (top) and AGEd (bottom) joints. Note the mild synovial inflammation; (C) representative micrographs of synovium histology obtained from PBS-injected (top) and AGEd (bottom) joints of tibial plateau; and (D) microscopic cartilage score. Bars represent mean ± SD of n = 8 animals per group. Data of the present experiment are presented in the context of a negative control (left white bar; Groove model) and experimentally induced OA as appositive control (right and black bar; Groove model; n = 5). Light grey and dark grey bars in the middle represent the PBS-injected and AGEd-joints of the present experiment, respectively (n = 8). p-values between the PBS-injected and the AGEd-joints are indicated, considering p ≤ 0.05 as statistically significant different.

Download figure to PowerPoint

Cartilage Biochemistry and Chondrocyte Activity

As was seen for the cartilage integrity and the synovial tissue inflammation also proteoglycan (PG) content, -synthesis, newly formed-release, and total-release were not statistically significant different for the PBS-injected and the AGEd-joints (Fig. 4, light and dark gray bars in the middle of each set of four bars).

thumbnail image

Figure 4. Biochemical markers of cartilage matrix integrity and chondrocyte activity of PBS-injected and AGEd-joints. (A) total proteoglycan content; (B) proteoglycan synthesis; (C) % release of newly formed proteoglycans (measure of retention of newly formed proteoglycans); and (D) % total proteoglycan release. Data of the present experiment are presented in the context of a negative control (left white bar; Groove model) and experimentally induced OA as appositive control (right and black bar; Groove model; n = 5). Light grey and dark grey bars in the middle represent the PBS-injected and AGEd-joints of the present experiment, respectively (n = 8). Bars represent mean ± SD. Correlation coefficients for the stepwise change in biochemical parameters (control, PBS-injected, AGEd, and OA joints) are shown.

Download figure to PowerPoint

When the biochemical data are compared to control and OA joints from the classic Groove model (Fig. 4 the white bars; left, representing unoperated control joints and Fig. 4 the black bars; right, representing the OA joints) it is clear that the severity of cartilage damage in the present study is positioned in-between normal healthy and early OA joints. In fact it also demonstrates that the AGEd-joints are in general in between the PBS-injected joints and the early OA joints. The PBS-injected joints are in-between the normal healthy joints and the AGEd-joints. Assuming a stepwise increase in severity of damage there was a correlation between severity of damage and the different conditions from healthy control joints, via PBS-injected joints with minimally surgically damage, AGEd-joints with minimal surgical damage, to experimentally induced OA joints (correlation coefficients above 0.8, p-values < 0.001 except for PG synthesis (p = 0.20), for all biochemical parameters; Fig. 4). Similarly, this was also seen for histology (Figs. 2D and 3D).

Relationship with Pentosidine

The significant variation in pentosidine levels in the AGEd-joints, allowed us to analyze possible relationships between the pentosidine levels and the outcome parameters. The difference between the pentosidine levels of the AGEd and PBS-injected joints of each animal is correlated with the percentage change of cartilage parameters between AGEd and PBS-injected joints of the same animal. For the release of total and newly formed PGs a strong negative relationship was found with the enhanced pentosidine levels between AGEd and PBS-injected joints (R = −0.762, p = 0.028 and R = −0.810, p = 0.015, respectively; Fig. 5A,B). For proteoglycan synthesis, no statistically significant relationship with pentosidine levels was found, although the tendency was negative as well (high pentosidine correlated with low synthesis and vice versa; R = 0.389, data not shown).

thumbnail image

Figure 5. Correlation between the differences in cartilage pentosidine between PBS-injected and AGEd-joints and percentage change between both contra-lateral joints, for cartilage newly formed (A) and total (B) proteoglycan release. Proteoglycan release is negatively correlated with the change in pentosidine content. Correlation coefficients (R) and p-values are indicated, considering p ≤ 0.05 as statistically significant different.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The present study could not demonstrate a clear difference in cartilage damage in AGEd-cartilage compared to control cartilage, upon minimal surgically applied damage. Compared to the positive control (clear OA features of the original Groove model) and negative control (healthy cartilage), artificially aging of the cartilage appeared to accelerate development of cartilage damage (OA), not reaching statistical significance. Additionally, high cartilage AGE levels demonstrated a low cartilage proteoglycan release, corroborating a diminished turnover of proteoglycans due to the AGEing of the tissue.

In this study, the pentosidine levels in the AGEd-joints were approximately fivefold increased compared to the PBS-injected joints. PBS-injected joints had AGE levels expected for dogs of this age. Only once before, artificial enhancement of AGEs was used in vivo in a canine model; also demonstrating a fivefold increase in cartilage AGE levels.14 It has been demonstrated that cartilage pentosidine in humans from young (about 20 years) to old (about 80 years) also increased fivefold.12 This means that the outcome of these experiments can be translated into the human situation so leading to more information about the development of OA and with that potential therapeutic strategies. As such it is concluded that the artificial aging was successful and of relevance to human conditions.

In the present study minor changes, characteristic of joint degeneration, were found in the PBS-injected joints, as compared to the negative controls (no treatment). The triggers for induction of joint degeneration were purposely kept to a minimum. The surgically induced cartilage damage was restricted to a maximum of four grooves on femoral condyles only, whereas in the regularly performed Groove model OA at least 10 grooves are made.18 Additionally, loading of the joints was minimized, knowing that loading adds to development of OA.23 In the present study, the animals were permanently housed individually in indoor pens (3.5 m2), the minimal reasonably acceptable space, to keep movement and with that loading of the joints to a minimum. This in clear contrast to active exercise in large groups on a large patio in case of the classical Groove model. Apparently, the minimal surgically applied damage of the femoral condyles and the remaining minimal loading is sufficient for minimal development of joint damage in the PBS-injected tibial joints surfaces. It should be kept in mind that the time for development of joint degeneration in the classic Groove model is 20 weeks,15 whereas it was 30 weeks for the AGEd animals in the present study. This prolonged follow-up was specifically chosen to provide an increased time window to enhance the chance for differences in the development of joint damage between the PBS-injected and AGEd-joints.

Most interesting, the minimal condylar surgical-induced damage and minimal joint loading did not result in spontaneous healing but actually resulted in development of damage on the untouched tibial plateau. Studies on the spontaneous healing or progression of damage of articular cartilage defects have been performed in several animal models, including dogs.17, 24, 25 Clearly even the minimal joint damage of the femoral condyles in the present study, with minimal loading is not healed in a period of 30 weeks but even slightly progressed to the tibial plateau.

However, the artificial aging of the cartilage only marginally accelerates the cartilage degeneration. No significant differences were observed although compared to the control group a tendency towards more damage was seen at a biochemical level, although less than in the classic Groove model without enhanced cartilage AGEing. It might well be that this minor difference was not related to development (initiation) of joint degeneration due to the artificial aging, but that it was due to slight progression of the degeneration already started. This corroborates the previous canine study where artificial AGEing was able to accelerate progression of OA study upon ACLT.14 As such in the present study the original approach to study development of OA was hampered by the minor development of joint degeneration in the control PBS-injected joints.

Interestingly, it appeared that within the group of AGEd-joints, proteoglycan release (turnover) was the lowest in those joints with the highest cartilage AGE levels. This corroborates previous findings that increased AGE levels in cartilage lead to increased cross-linking of proteins and with that to diminished release of these macromolecules.10 Also degradation of AGE-modified collagen by matrix metalloproteinases is impaired compared to unmodified collagen.26 This will impair repair activity of the cartilage. Despite this phenomenon, also observed in the artificially AGEd-joints in this study, changes were insufficient to clearly drive the development of joint degeneration in the present study set-up.

In conclusion, despite the fact that enhanced cross-linking of macromolecules by the AGEs restrains loss of proteoglycans, corroborating the diminished turnover of old cartilage, and a tendency toward enhanced cartilage damage in the artificially AGEd-joints, the present data do not support a predominant role for enhanced cartilage AGE levels in development of joint degeneration in early OA. As such the clinical relevance of AGEing of cartilage in development of OA remains obscure.

Acknowledgements

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

This study was supported by the Dutch Arthritis Association.

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES