Collagen Growth Pattern in Human Articular Cartilage of the Knee

Abstract

Objective

During skeletal growth, the articular cartilage expands to maintain its cover of bones in joints, however, it is unclear when and how cartilage grows. We aim to determine the expanding growth pattern and timing across the tibia plateau in human knees.

Design

Six human tibia plateaus (2 healthy, 2 with osteoarthritis, and 2 with posttraumatic osteoarthritis) were used for full-depth cartilage sampling systematically across the joint surface at 12 medial and 4 lateral sites. Methodologically, we took advantage of the performed nuclear bomb tests in the years 1955 to 1963, which increased the atmospheric ¹⁴C that was incorporated into human tissues. Cartilage was treated enzymatically to extract collagen, analyzed for ¹⁴C content, and year at formation was determined from historical atmospheric ¹⁴C concentrations.

Results

By age-determination, each tibia condyle had central points of formation surrounded by later-formed cartilage toward the periphery. Furthermore, the tibia plateaus contained collagen with ¹⁴C levels corresponding to mean donor age of 11.7 years (±3.8 SD). Finally, the medial condyle had lower ¹⁴C levels corresponding to formation 1 year later than the lateral condyle (P = 0.009).

Conclusions

Human cartilage on the tibia plateau contains collagen that has experienced little if any turnover since school-age. The cartilage formation develops from 2 condyle centers and radially outward with the medial condyle finishing slightly later than the lateral condyle. This suggests a childhood programmed cartilage formation with a very limited adulthood collagen turnover.

Introduction

Bones in joints are lined with articular cartilage (AC) that allows for near frictionless movement. ¹ During childhood and adolescence, bones grow in length and width. The longitudinal growth is well studied: it occurs by cartilage expansion and ossification in the epiphysis and metaphysis until these fully mineralizes into bone, and the skeleton reaches maturity. ² AC never mineralizes and it is produced by a different pool of chondrocytes; however, the AC growth, and expansion laterally in the joint, during postnatal life is not well understood. ³ Lateral cartilage expansion is essential for maintaining cover of the entire joint surface as growth occurs, and further, a larger surface area of AC would limit the increasing pressure from the rising body weight as the individual grows.

Healthy AC contains around 60 % collagen type II (dry weight). ⁴ Studies on animals show that apart from the most superficial layer of AC, the initial immature AC is replaced during the skeletal growth prepuberty,^5,6 and during puberty the collagen fibers further develop with a change in orientation and in enzymatic cross-link content into the final mature collagen mesh.^7-9 Similar studies have not been done in humans, but if similar development takes place, only a minimum of new collagen should be made in adulthood in the mature AC, and the amount of cartilage collagen is determined at some point(s) in time before that. Indeed, human studies show that mature AC has at best a very limited collagen turnover.^10-12 Previous work in our lab has shown limited collagen turnover after the age of approximately 20 years both in healthy and osteoarthritis (OA) affected AC of the tibia plateau. ¹⁰ However, as only 2 samples from each donor were analyzed, there are no details available regarding timing and the lateral expanding growth pattern. Thus, we aim to determine whether there are certain time points during growth where a major part of the collagen mesh is laid down, and to reveal the expanding growth pattern across the tibial cartilage in human knees. We hypothesize that the cartilage develops before the age of 20 years along with skeletal maturity, and that central cartilage areas are laid down before peripheral.

Materials and Methods

In the years 1955 to 1963, testing of nuclear bombs led to a dramatic increase in atmospheric ¹⁴C levels. After the Test Ban Treaty, a subsequent decline followed, leaving a spike in atmospheric ¹⁴C level, known as the “bomb pulse” ( Fig. 3 ). Plants incorporate ¹⁴C from atmospheric ¹⁴CO₂, which then accumulates in animals and humans that ingest these plants. Tissues with rapid turnover such as muscle have a content of ¹⁴C corresponding to the current level in the atmosphere. ¹³ On the other hand, tissues that are not turned over after the initial formation such as the eye lens ¹⁴ will contain ¹⁴C levels corresponding to the level present in the atmosphere when the tissue was formed. Therefore, it is possible to determine the turnover of tissues from organisms born in the years spanning the bomb-pulse due to the incorporation of ¹⁴C starting from embryogenesis to tissue sampling.

Figure 3.

¹⁴ C content in cartilage collagen. The ¹⁴C bomb-pulse curve shows the chronological atmospheric concentration of ¹⁴C (black line) as percent modern carbon (pMC). The horizontal dashed line indicates the approximate atmospheric ¹⁴C level at the period of tissue sampling (February to December 2013). Vertically aligned symbols represent ¹⁴C concentrations in treated cartilage from the same individual at the different locations on the tibia plateau. Several donors were born the same year, but for clarity the donors’ birth years have been plotted months apart to separate the vertically aligned symbols.

Human Donors

Donors born just after the peak of the bomb pulse were selected, as they would have experienced declining levels of atmospheric ¹⁴C since birth. Consequently, a high level of ¹⁴C in the collagen matrix would indicate that the matrix is “old,” that is, laid down early in life, while a lower level would indicate that the matrix is “young” and therefore laid down later in life. We used 6 whole tibia plateaus for cartilage sampling collected in a previous study ¹⁰ and kept at −80 °C ( Table 1 ). Two plateaus were collected as waste tissue from patients with healthy cartilage, who had undergone surgery due to primary malignant bone tumors of the distal femur, thus neither affecting the tibia bone nor the cartilage (donors 1 and 2). Four plateaus were from patients with OA: 2 primary and 2 posttraumatic OA (PTOA) collected as waste tissue from total knee replacement surgery (donors 3 to 6). The study was conducted in accordance to the Declaration of Helsinki, and ethical approval for obtaining waste tissue was obtained from the Ethical Committee of the Capital Region of Denmark (H-4-2012-131), and all participants gave written informed consent.

Table 1.

Donor Characteristics.

Donor	Birth Year	Tissue Type	Tibia Plateau Cartilage Condition
1, male	1971	Healthy cartilage	Normal. No sign of OA.
2, male	1971	Healthy cartilage	Normal. No sign of OA.
3, female	1968	Primary OA	Moderate medial and lateral OA.
4, female	1967	Primary OA	Moderate medial and lateral OA.
5, female	1967	Posttraumatic OA	Moderate medial and lateral OA. Medial meniscus lesion in the twenties (specific age unknown).
6, male	1967	Posttraumatic OA	Moderate medial and lateral OA. Severe patellofemoral OA. Patella fracture and posterior cruciate ligament rupture at age 30 years.

OA = osteoarthritis.

Cartilage Sampling

Using a 5-mm stainless steel punch, 16 cylindrical full depth cartilage samples were taken from each tibia plateau. To standardize sampling, we used a template to obtain 8 biopsies systematically across the entire plateau with additional 8 samples from the larger medial condyle to enhance details in the anterior to posterior direction ( Fig. 1 ). Cartilage was separated from the underlying subchondral bone by cutting as close as possible to the bone with a scalpel. Next, the biopsies were cut in thin slices with a scalpel perpendicular to the surface. All biopsy samples were weighed, freeze-dried, and reweighed. A slice of each biopsy was kept as a raw control, that is, without any enzymatic treatment.

Figure 1.

Systematic placement of biopsy locations across the tibia plateau. (Left) Photograph of a right tibia plateau after sampling of cartilage tissue. (Right) Schematic representation of the same plateau. Biopsies were sampled as follows: The anterior cruciate ligament (ACL) and the intercondylar area (ICA) were located (black pins) marking the anterior-posterior line. A perpendicular line was made giving the longest possible span across (blue pins). Four samples were collected from the most central (C1 and C10) and peripheral (P3 and P30) location possible on each side, and 2 additional samples evenly distributed in between, both medially (C2 and M3) and laterally (C20 and M30). On the medial condyle, we located the points where the cartilage tissue curved from the central to anterior/posterior side near the horns of the medial meniscus (white pins). Two lines between the white pins and sample C2 were drawn and biopsies were taken the most peripheral location possible (P1 and P5) and another evenly between these and C2 (M1 and M5). Two bisecting lines between samples P1-C2 and C2-P3 and between P5-C2 and C2-P3 were marked (red pins). Biopsies were taken the most peripheral location possible (P2 and P4) with another evenly distributed between these and C2 (M2 and M4). This led to systematic sampling of 12 samples medially and 4 laterally.

Collagen Purification

We aimed at studying the ¹⁴C content in collagen, thus we removed glycosaminoglycans (GAGs) and other noncollagenous substances from the remaining slices of the cartilage samples using a collagen extraction procedure as previously described. ¹⁰ By using enzymatic treatment, the GAGs can be removed with minimal collagen loss. ¹⁵ Freeze-dried samples were treated overnight with hyaluronidase (H3506, Sigma) (5 U/mL in 0.05 M sodium acetate and 0.15 M NaCl [pH 6]) at 37 °C. Then, the samples were centrifuged, supernatant removed, and the cartilage slices were washed with isotonic NaCl. Next, trypsin (T8802, Sigma) (1 mg/mL in phosphate buffered saline [PBS]) was added and samples were incubated at 37 °C overnight. To remove any trace of carbon from the added enzymes and the acetate, the samples were washed with PBS, 0.7 M KCl, and 3 times with distilled water before being freeze-dried, weighed, and kept at −80 °C for later analyses. As trypsin is not able to cleave native triple helical collagen, ¹⁶ our protocol would leave behind an indigestible cartilage matrix of predominately fibrillar collagen type II, while unincorporated collagen would be removed.

Isotope Analyses

From the enzymatic treated samples, 2 to 3 mg of cartilage slices were sent for isotope analyses (¹³C, ¹⁴C, and ¹⁵N) at the AMS ¹⁴C Dating Centre, Aarhus University, Denmark. The samples for accelerator mass spectrometry (AMS) were combusted with CuO in sealed combustion tubes at 950 °C and converted to graphite prior to ¹⁴C analysis at the 1 MV Tandetron accelerator. ¹⁷ The radiocarbon dating results are reported according to international convention ¹⁸ and ¹⁴C content are given as percent modern carbon (pMC) based on the measured ¹⁴C/¹³C ratio corrected for the natural isotopic fractionation by normalizing the result to the standard δ¹³C value of −25‰ VPDB (Vienna Pee Dee Belemnite: a δ¹³C calibration standard). Stable isotope values of δ¹³C, δ¹⁵N, carbon and nitrogen fraction (by dry weight), and carbon/nitrogen (C/N) atomic ratios were measured at the Aarhus AMS Centre by continuous-flow isotope-ratio mass spectrometry.

Hydroxyproline and Glycosaminoglycan Assay

To verify the extraction of collagen and removal of GAGs, we performed assays detecting hydroxyproline (HYP) and GAGs on all of the remaining enzymatic treated slices and the raw control slice. All were treated with papain (P3125, Sigma) (papain [0.125 mg/mL] in 100 mM sodium phosphate buffer, 10 mM Na₂-EDTA, and 10 mM l-cysteine [pH 6.5]) at 60 °C overnight. For GAG quantification, the diluted papain digest was mixed with 1,9-dimethylmethylene blue (DMMB) (Sigma 341088) solution (38 µM DMMB in 40 mM NaCl, 40 mM glycine, pH 3), and absorbance were read at 595 nm and 540 nm wavelengths (subtracted) and compared with a known standard curve of chondroitin sulphate C (Sigma C4384). For HYP quantification, the papain digest was hydrolyzed in 6 M HCl overnight at 110 °C, dried, rehydrated with distilled water, and dried again before being resuspended in an acetate-citrate buffer (0.6% acetic acid, 130 mM citric acid, 440 mM sodium acetate, 425 mM NaOH, pH 6). A chloramine-T solution (60 mM chloramine-T, 50% 1-propanol) was added, and samples incubated at room temperature for 20 minutes. Next, an aldehyde perchloric acid solution (1 M 4-dimethylaminobenzaldehyde, 60% 1-propanol, 22% perchloric acid [70%-72%]) was added and incubated for 25 minutes at 60 °C before the reaction was stopped in an ice-bath. The samples were read at 570 nm wavelength and compared with a known standard curve of HYP (Sigma, H1637). As collagen has been found to contain around 13.4% HYP, the concentrations were converted to collagen concentration by multiplying with a factor of 7.5. ¹⁹

Statistics

All statistical analyses were performed using GraphPad Prism version 8.4.1 (GraphPad Software, San Diego, CA, USA). Repeated measures 1-way analysis of variance (ANOVA) or mixed-effects analysis (if values were missing) with the Geisser-Greenhouse correction for sphericity were used. Within areas of each condyle, Holm-Sidak’s multiple comparisons test was used post hoc.

Friedman test was used for differences in the ranking order of growth with Dunn’s multiple comparisons test post hoc within areas on the medial condyle.

Paired t test was used for difference between medial and lateral condyles.

Statistical significance was defined as P values <0.05 for all analyses.

Results

Tissue Composition of Raw and Treated Cartilage

The isotope measurements were made on extracted collagen. Thus, to evaluate the effectiveness of the extraction protocol, the content of GAGs and collagen were analyzed for every location. All cartilage samples were freeze-dried initially: The water content in raw cartilage was 70.9% ± 2.7% without any detectable difference between locations (Supplementary Fig. S1A). However, between areas a difference was found (P = 0.022), but multiple comparisons test was insignificant ( Fig. 2A ). The extraction protocol successfully reduced the GAG content by 92%: Raw cartilage samples contained 137.3 ± 31.8 µg/mg fry weight (d.w.) GAGs, which decreased to 10.7 ± 2.8 µg/mg d.w. GAGs in treated cartilage without any significant difference between locations (Supplementary Fig. S1B). However, in raw cartilage only, a difference between areas was found (P = 0.047), with multiple comparisons test showing difference between M1-M5 and P1-P5 (P = 0.016) on the medial condyle ( Fig. 2B ). The extraction protocol expectedly increased the total collagen content: Raw cartilage samples contained 395.9 ± 95.3 µg/mg d.w. collagen, while treated cartilage contained 697.7 ± 111.9 µg/mg d.w. collagen, without any significant difference between locations (Supplementary Fig. S1C). However, in raw cartilage only, a difference between areas was found (P = 0.038), but multiple comparisons test was insignificant ( Fig. 2C ). Thus, as expected the treated cartilage samples used for isotope measurements contained high levels of collagen and very low levels of GAGs regardless of location on the plateau.

Figure 2.

Tissue composition. The composition in raw (gray bars) and treated cartilage (open bars) for every areas on the plateau: Central (C1-C2), middle (M1-M5), and peripheral (P1-P5) on the medial condyle, and central (C10-C20), middle (M30), and peripheral (P30) on the lateral condyle. (A) Water content. (B) Glycosaminoglycan (GAG) content. (C) Collagen content. Values are mean of the areas for each donor, and error bars represent overall area mean ± SEM for all donors. ^#Equals P < 0.05 for analysis. *Equals P < 0.05 between areas.

Growth Pattern

The ¹⁴C levels in cartilage collagen from each individual donor at each anatomical location are shown with the historical atmospheric concentrations of ¹⁴C for comparison^20,21 ( Fig. 3 ), and in details in Supplementary Table S1. To examine differences across the plateau, the growth pattern is presented for each donor at every location ( Fig. 4A ) showing a significant difference across the plateau (P = 0.049). As the medial condyle had eight additional samples taken, the means from every area are presented ( Fig. 4B ): A significant test for linear trend was found (P = 0.007), and 1-way ANOVA showed a strong trend (P = 0.056). To further clarify the growth pattern visually, we used the pMC values (Supplementary Table S1) to rank every sample location from high to low, for each of the condyles: For each donor, the samples were ranked 1 to 12 on the medial condyle, and 1 to 4 on the lateral condyle (with the mean rank used for similar levels), thus 1 being made first/earliest in life and 12 last/latest in life (Supplementary Fig. S2). The median rank for each location is presented ( Fig. 4C ) showing a significant difference across the plateau (P = 0.015). We then compared the mean ranks of the areas on the medial condyle, which showed a trend (P = 0.072) ( Fig. 4D ). To illustrate the growth pattern on the plateau, a schematic overview of all donors is presented ( Fig. 4E ). Both tibia condyles have central points of formation containing older collagen with younger collagen radially outward from here.

Figure 4.

Growth pattern. (A) Tibia plateau growth pattern. Values are individual percent modern carbon (pMC) levels. (B) Growth pattern between central (C1-C2), middle (M1-M5), and peripheral (P1-P5) areas on the medial condyle. Values are mean pMC levels for each donor. (C) From each donor, the samples were ranked 1 to 12 on the medial condyle, and 1 to 4 on the lateral condyle using the mean rank for similar levels, with 1 being made first and 12 last. Values are individual ranks, and bars represent median rank for each location. (D) The mean ranks between central (C1-C2), middle (M1-M5), and peripheral (P1-P5) areas on the medial condyle. Values are mean ranks for each donor, and error bars represent mean rank for each area ± SEM. (E) Schematic presentation of the growth pattern on the tibia plateau. Values are median ranks at each location. (F) Location legend. ACL, anterior cruciate ligament; ICA, intercondylar area. ^#Equals P < 0.05 for analysis.

¹⁴ C Levels and Estimation of Tissue Formation Year

By comparing the ¹⁴C levels of each cartilage biopsy with the historical atmospheric ¹⁴C levels, a year of tissue formation can be estimated. By relating this year of tissue formation with the birth year of the donor, the age of the donor at the time of cartilage formation can be estimated. For example, for donor 1, tissue in location C1 had a ¹⁴C level of 131.54 pMC (Supplementary Table S1), corresponding to the atmospheric level between the years 1978 and 1979 and thus to a donor age of 7.5 years (1978.5-1971). Importantly, this does not necessarily mean that all the tissue in site C1 for this donor was formed at the age of 7.5 years, but merely that the mass weighted average of the ¹⁴C levels accumulated during growth of collagen in this site corresponds to the atmospheric ¹⁴C levels at that age. Thus, the actual collagen formation may have occurred during some years before and after the age of 7.5 years.

Estimated Age at Formation

A detailed overview with photos and estimated donor age at tissue formation for all donors is provided (Supplementary Fig. S2). The mean age was 11.7 ± 3.8 years ( Fig. 5A ). A schematic overview of the average age at formation for each anatomical location is presented for all donors together ( Fig. 5C ), showing collagen with ¹⁴C levels corresponding to atmospheric levels present when the donors were 7 to 16 years of age. Between each condyle, the four samples at similar locations were compared (i.e., C1, C2, M3, and P3 medially vs. C10, C20, M30, and P30 laterally) showing higher age at formation medially at 11.43 ± 1.2 years than laterally at 10.33 ± 1.1 years (P = 0.009) ( Fig. 5B ).

Figure 5.

Age at formation. (A) Corresponding age at formation at every location. Values are individual data points in years, with lines representing the area means. (B) The age at formation between each condyle. Values are means in years of four samples on each condyle for each donor (n = 5 due to missing values in donor 1) and error bars represent condyle means ± SEM. (C) Schematic presentation of the age at formation on the tibia plateau. Values are mean years. (D) Location legend. ACL, anterior cruciate ligament; ICA, intercondylar area. *Equals P < 0.05.

Discussion

The Tibia Plateau Contains 2 Central Areas with Older Matrix and Expands Outward

We measured the lifelong incorporation of ¹⁴C in enzymatically treated cartilage samples across the entire tibial condyle to characterize the developing growth pattern and timing of cartilage collagen formation. The enzymatically treated cartilage samples had high collagen content (70% ± 12%) and negligible GAG content (1% ± 0.3%) regardless of tibia location, and the ¹⁴C data presented on the trypsin-resistant matrix are thus largely representative of fibrillar collagen. Our data demonstrate that the collagen at the center is generally oldest (5 out of 6 plateaus) and contains younger collagen radially outward toward the periphery.

Previous work showed a correlation between years at formation and distance to the edge of the tibia plateau regardless of OA, ¹⁰ and a cell-lineages tracking study by Kozhemyakina and colleagues ²² found that the greatest apparent expansion of progenitors in the knee joint of mice occurred in the tibial articular cartilage, near the central/middle region of the condyles—that is, in the regions where this tissue experiences the highest levels of mechanical loading. ²² It could seem that mechanical loading by movement of the joint is a stimulus for the AC progenitor cells, and several groups speculate that a population of cells in the superficial zone have properties of stem cells that proliferate and expand laterally.^5,6,22-25 By applying cell-lineages tracking studies in mice, it has been observed that some superficial cells divided along the joint surface and then remained there, and further that superficial cells gave rise to clonal clusters, which might facilitate lateral cartilage expansion during juvenile growth. ²³ Our data on the cartilage collagen matrix are in line with the data on cell proliferation with two central areas providing lateral expansion across each plateau. Unfortunately, due to methodological limitations it was not possible to separate superficial from more deep zones of the cartilage to provide zone-specific ¹⁴C content measurements in the present study.

In addition to the central area being made first, we found strong trends suggesting that 2 concentric areas are developed radially outward. However, as one donor showed different growth pattern out of the limited number of donors in the current setup, further statistical significance was unattainable. Nonetheless, this intriguing finding would seem to indicate a childhood programming of cartilage formation and turnover. On the other hand, the opposite growth pattern found in one donor does demonstrate that an exception to the suggested programming is possible. This donor did experience major knee trauma at age 30 years (anamnestic information in patient history) leading to both a patella fracture and a lesion of the posterior cruciate ligament, which would lead to significant joint bleeding. This could induce cartilage changes as evidenced by hemophilic arthritis with frequent joint bleedings. ²⁶ Furthermore, the posterior cruciate ligament lesion would change the mechanics in the joint and likely lead to different loading pattern on the medial condyle. As the growth on the lateral condyle (i.e., with less weight bearing) is similar to the pattern observed in the other donors, this further suggests a possible influence of changed mechanics due to the trauma. Since the pMC levels are very similar across the plateau ( Fig. 4A ), even minor turnover could tip the preprogrammed growth pattern to the one observed in this donor.

The Growth Pattern Forms within a Decade during School Years

The present study showed that the medial condyle contained slightly lower levels of ¹⁴C corresponding to formation 1 year later than the lateral condyle. As the medial condyle is bigger, it could simply take longer to develop, alternatively, as it is loaded more, and as discussed above regarding cell behavior, mechanical loading could be a stimulus for formation, and the difference found between condyles could thus be due to mechanics. In tendons (also rich in fibrillar collagen), stable isotope infusion technique has indicated that a smaller fraction of collagen can be turned over more acutely in response to mechanical loading in adults.^27,28 To our knowledge, a loading effect specifically on collagen content or formation in cartilage has never been examined in children or in adolescents, and the implications for cartilage collagen made predominately at an age interval of 7 to 16 years are unclear. In young adult athletes aged 18 to 20 years, increased markers of collagen degradation were found. ²⁹ In school children, vigorous physical activity was associated to a greater amount of cartilage by magnetic resonance imaging scans, but did not provide details on cartilage composition. ³⁰ However, when the same population of children was followed longitudinally, the association to physical activity became insignificant, possibly due to the small sample size of predominantly very active children. ³¹ Still, any effect on collagen by exercise or loading in childhood cannot be ruled out, although, in young animals and adult humans, the effect of exercise is primarily found on the GAGs and not on collagen itself.^32,33 We were not able to collect information regarding activity, however, when analyzing cartilage samples from donors over a large age-span, the ¹⁴C levels do approximately follow the bomb-pulse curve. ¹⁰ This suggests that the donors follow similar turnover kinetics with only a small amount of collagen being turned over despite having different exercise/loading history. Considering these findings, it thus seems unlikely that any difference in activity levels between the donors would change the interpretation of the current data presented on collagen.

Cartilage Maintenance Potential in Adulthood

On average, we found that the trypsin-resistant matrix of cartilage collagen contained ¹⁴C levels corresponding to levels in the atmosphere at donor age of 7 to 16 years with a mean age of 11.7 ± 3.8 years. This suggests very limited turnover in adulthood, where most of the collagen mass is retained after skeletal maturation. This confirms previous findings in cartilage by using ¹⁴C levels ¹⁰ and by Banks et al., ⁷ who showed accumulation of advanced glycation end-products (AGEs) after the age of 20 years in human femoral (hip) cartilage. As AGEs are nonenzymatic cross-links only removed by tissue breakdown, AGEs accumulate in tissues with low turnover. However, as shown in Figure 5A , some samples in one donor contained pMC levels corresponding to a formation age around 30 years. Although the peripheral areas in this donor were macroscopically colored violet, they were not osteophytes, which is a fibrocartilaginous outgrowth containing newer matrix. ³⁴ Because the plateau contained darkened areas of cartilage in the periphery, an extra sample was taken (Supplementary Fig. S2: white circle on photo) to confirm the high values in this posterior part of the medial plateau. Consistently with the surrounding biopsies, this had an age at formation at 39 years. Thus, it appears that factors can intervene resulting in regional matrix incorporation of newer carbon later in life. By chance, we were able to capture this unique phenomenon, although it must be stressed that none of the other donors in the current nor the previous study ¹⁰ showed any sign of newer collagen synthesis. Finally, due to discoloration of the plateau, we cannot rule out an unidentified underlying condition.

As the measured pMC level is a mass-weighted average of each sample biopsy, minor amounts of newer collagen could be incorporated into the matrix throughout life, and thus, it cannot be concluded that any collagen syntheses later in life is impossible. Furthermore, cartilage could be maintained by other parts of the matrix not analyzed in the current setup. In mice, it has been demonstrated that the persistent collagen mesh of incorporated collagen fibers experience rhythmic circadian maintenance from a pool of newly synthesized soluble collagen, ³⁵ and a recent study in human cartilage not using enzymatic treatment did show soluble collagen type II turnover. ³⁶ Our results are on the trypsin-resistant collagen matrix, and any new unincorporated collagen would not have been measured, but removed by the extraction protocol. Thus, a steady synthesis of soluble collagen not incorporated in the cartilage matrix cannot be ruled out. Furthermore, collagen maturations with cross-links could still influence the cartilage strength later in life. However, the ¹⁴C method cannot provide answers regarding cross-links, as carbon from the collagen helixes greatly outnumber carbon found in cross-links. Finally, a recent study using continuous infusion of stabile isotopes hours prior to knee replacement surgery showed turnover in raw cartilage, ³⁷ and this result is therefore likely to be carried by proteins with a short half-life, for example, proteoglycans. Thus, future studies could use isotope infusion over a longer period combined with enzymatic treatment of cartilage to examine if the stable matrix of predominately collagen does indeed have any turnover.

Cartilage Turnover in Early Childhood (Preschool)

In the present study, we did not find any samples containing ¹⁴C levels corresponding to atmospheric levels present before a donor age of 5 years, which indicates that the cartilage matrix studied is either completely renewed, or that the samples contain embryonic collagen mixed with collagen formed later in life. Both would still lead to the average ¹⁴C levels found in the samples. In tendons, it has been shown that during embryonic development, the collagen fibrils increase in both number and length, but during postnatal growth they remain at a constant fibril number, of increasing length and diameter. ³⁸ This could imply that the collagen fibrils are made during embryogenesis and not replaced, but instead, newer collagen is made during growth. Postnatal collagen could also be formed by remodeling, meaning that the embryonic fibrils are replaced by thicker and longer collagen fibrils. Our data fit both possibilities, as an addition of new collagen or gradual replacement of embryonic collagen would lead to a dilution of the embryonic fibrils’ ¹⁴C levels resulting in ¹⁴C levels corresponding to atmospheric levels some years after birth as found in the current study—though some or even all the embryonic collagen might remain.

In addition to the overall common finding in the different cartilage donors, clearly individual differences were present.

In this current explorative study, we are limited by a relatively low number of donors with different clinical conditions, and additional donors would have provided more strength to the analyses.

In conclusion, by using the bomb pulse our data show that human cartilage on the tibia plateau contains collagen that has experienced little turnover since adolescence. Furthermore, the tibia plateau develops centrally on each condyle, expanding radially outward, with the lateral condyle made slightly earlier than the medial. We suggest a possibility for a childhood programmed cartilage formation, and a very limited adult turnover kinetic only partly susceptible to minor increases with disease.