Femur Shape Changes in Prg4‐Deficient Mice: Morphological Insights Into Joint Well‐Being

ABSTRACT

Many studies have reported on the role of Proteoglycan‐4 (PRG4, aka lubricin) in the reduction of friction between cartilage surfaces with a specific focus on chondroprotection within the joint. Disruption of the Prg4 gene in humans and mice leads to premature joint failure, hallmarked by synovial hyperplasia and premature articular cartilage fibrillation. Our group has published extensively using Prg4 knockout mice and has consistently noticed variable distal femoral morphology in these animals when compared to Prg4 ^+/+ wild‐types (WT). This prompted us to undertake a quantitative study examining joint element size and shape to elucidate if this phenotype was consistent in a larger sample size. High‐resolution X‐ray microscopy (XRM) images were obtained from WT and Prg4 ^−/− mice between 8‐ and 36 weeks of age. We then employed geometric morphometrics to characterize mouse femora shape changes, which were correlated to cross‐sectional histological findings. We find that Prg4 ^−/− femora vary in size and shape compared to WT controls; distal femora in Prg4 ^−/− mice are enlarged, extended (anteroposterior) and narrower (mediolateral), with the largest regional deviations being traced to the trochlear groove, epicondyles, and medial condyle. Additionally, quantifiable changes in condylar articular cartilage thickness were associated with abnormal compressive biomechanical properties. Collectively, these data suggest that PRG4 loss extends beyond joint homeostasis and critically impacts joint morphology.

Using X‐ray microscopy, histology, and geometric morphometrics, Prg4 ^−/− mice were shown to have enlarged, elongated, and narrowed femora with marked alterations in trochlear groove shape. These shape changes are accompanied by thicker but biomechanically compromised articular cartilage that deteriorates with aging. The findings highlight the critical role of PRG4 in postnatal joint morphogenesis and long‐term joint integrity.

1. Introduction

The knee joint is an inherently complex anatomical structure. Its geometry in relation to the varied properties and structural requirements of its constituent tissues is of critical importance for locomotion and daily activities. There are two distinct articulating surfaces on the distal femur. These are the patellofemoral (PF) and tibiofemoral (TF) articulations. The morphology of these surfaces helps determine the mechanical properties and stability of the knee joint [1]. Knee joint formation begins during embryonic development. Proper patterning and specification of individual joint structures rely on a series of coordinated cellular and molecular events along with mechanical signals. However, the mature morphological features of each knee joint element (e.g., compartments) emerge postnatally. Likewise, the articular cartilage acquires its specialized internal structure and mature shape through postnatal remodeling [2, 3]. The precise regulatory mechanisms driving morphogenesis of the knee joint and articular cartilage, as well as postnatal growth and maturation, remain poorly understood.

Mutations in genes responsible for synthesizing, assembling, and modifying cartilage extracellular matrix components have been implicated throughout the gradient of known skeletal phenotypes. Disruptions in these developmental and remodeling processes can lead to structural and functional issues in multiple tissues within synovial joints, from bones to cartilage, ligaments, and menisci, resulting in non‐physiological loading of the knee joint, increasing the risk of injury and degenerative conditions like osteoarthritis (OA). One gene of interest is PRG4 (Proteoglycan‐4), which codes for a mucin‐like glycoprotein mostly known as a lubricating and chondroprotective molecule in synovial joints. In situ hybridization and genetic lineage tracing studies have established that Prg4‐expression is upregulated during the later stages of mouse embryogenesis (gestational day (E)17.5) and is enriched within the newly forming synovial membrane and articular cartilage [2, 4, 5]. Its expression is retained in the synovium and articular cartilage surfaces of adult joints, and Prg4 protein products are found at high levels in the synovial fluid and bound to the surface of the articular cartilage and synovium, where it is believed to be essential for joint tissue homeostasis and functionality [5, 6, 7].

Not surprisingly, it has been demonstrated that Pgr4 loss‐of‐function mutations impact joint pathology in mice and humans. In humans, PRG4 mutations have been linked to camptodactyly‐arthropathy‐coxa vara‐pericarditis (CACP) syndrome [8, 9], a rare genetic condition with various joint‐related clinical features. Disease manifestation includes congenital or early childhood‐onset flexion deformity of phalangeal joints (camptodactyly) and arthropathy of larger joints, such as the knee, associated with swelling, restricted range of motion, non‐inflammatory synovial hyperplasia, and articular cartilage degeneration [9, 10]. Moreover, decreased levels of PRG4 have been observed in the synovial fluid in patients with OA or rheumatoid arthritis (RA) degenerative joint diseases [11]. In mice, Prg4‐null (Prg4 ^−/− ) mutations recapitulate many of the phenotypic changes identified in CACP patients, including synovial hyperplasia, abnormal cartilage surface, restricted range of motion, and joint failure [5].

Given the onset expression of Prg4 during embryonic development, it is surprising that knee joints have been reported to look morphologically/overtly normal at birth [5, 12]. Yet, previous investigations during early postnatal development have demonstrated microstructural changes in subchondral bone and articular cartilage, including increased bone porosity, disrupted collagen structure and arrangement in articular cartilage, which is also thickened [12, 13, 14, 15]. Prg4 expression has also been proposed to influence bone skeletogenesis in mice [16, 17]. However, to date, no study has employed a quantitative approach to empirically determine the impact of Prg4 in overall femoral morphology.

Building upon anecdotal evidence gained from multiple investigations on Prg4 ^−/− mice, we employed histological cross‐sectional evaluation and landmark‐based geometric morphometrics to identify, quantify, and characterize variations in the distal femoral morphology of Prg4 ^−/− mice in adulthood compared to age‐matched controls. Since the geometry and anatomy of articular surfaces relate to their function, we also examined regional differences in condyle cartilage thickness and indentation biomechanics. Although the published literature recognizes that loss of Prg4 does not impede developmental knee joint formation, its role in tissue patterning and joint structure/geometry/anatomy remains obscure and could have implications for proper maintenance and function of these elements.

2. Materials and Methods

2.1. Animals

All animal procedures were performed in accordance with the Canadian Council of Animal Care guidelines and approved by the University of Calgary Animal Care Committee (protocols AC16‐0043 and AC20‐0042). Prg4 ^−/− mice (Prg4 ^tm1Mawa/J, stock # 025737), whose generation has been described previously [5], and wild‐type (WT) mice (C57BL/6J, stock #000664) were obtained from The Jackson Laboratory and maintained in‐house at the University of Calgary. Prg4 ^−/− mice were backcrossed to C57BL/6J for ten generations, and genotyping from ear biopsies was carried out to confirm homozygous Prg4 ^−/− status. Age‐matched C57BL/6J (backcrossing colony littermates, referred to as WT) mice were used as controls. All animals were housed under a standard light cycle and had access to food and water ad libitum. Mice were euthanized at postnatal Weeks 8, 12, 16, 20, 25, and 36. Hindlimb specimens from at least four animals per group were used to ensure reproducibility in histological, morphometrical, and biomechanical analyses.

2.2. Histology

Hindlimbs were fixed in 10% neutral formalin buffer (NBF) solution (Fisher Scientific), decalcified using 10 w/v EDTA (pH 7.4) for 2.5 weeks with solution changes every other day. During paraffin embedding, specimens were oriented such that the typical knee joint flexion was preserved, and the tibia was kept parallel to the sectioning surface. Serial sections (10 μm) were mounted onto Superfrost Plus glass slides (Fisher Scientific) for further analysis. To examine the gross joint morphology and degree of cartilage degeneration, slides were stained in Safranin‐O/Fast‐green and scored using the Osteoarthritis Research Society International (OARSI) mouse histopathological scoring system [18]. Articular cartilage in the mid‐to‐posterior load‐bearing regions of the knee, identified by the presence of cruciate ligaments, was scored in each animal, and the sum of the maximum histopathological score of two independent scorers (one blinded) per knee joint femoral compartment was reported.

2.3. 3D X‐Ray Microscopy (XRM) Imaging

To obtain a three‐dimensional representation of the complex femoral shape and determine the effect of Prg4 loss‐of‐function on morphology and associated age‐related changes, right murine distal femurs (n = 4–5 per genotype) were scanned at a 4.4 μm isotropic voxel resolution using the X‐ray microscopy (XRM) imaging system (Versa 520, Carl Zeiss X‐ray Microscopy, USA). Specimens were fixed in 10% NBF for 24 h and preserved in 70% ethanol until the day before scanning (no more than 3 days), when they were incubated in 1% phosphotungstic acid (PTA) solution for 18–24 h prior to scanning for optimal contrast enhancement [19, 20]. Specimens were enclosed in a Kapton straw chamber (5 mm diameter, GoodFellow Cambridge) vertically (Figure 1a) with PTA solution on the bottom of the chamber to minimize specimen desiccation. Scan parameters were: 40 kVp voltage, 3 W power, 2001 projections. Image projections were reconstructed with the ZEISS XMReconstructor software.

FIGURE 1.

Prg4 ^−/− mice exhibit morphological changes in the distal femur and articular cartilage. (a) General appearance of distal femora at 8‐ and 36‐weeks of age in WT and Prg4 ^−/− mice. Note the abnormal shape of the trochlear groove at 8 weeks, with protrusion of trochlear ridges and deepening of the trochlear groove, still identifiable by 36 weeks of age, wherein the medial condyle appeared to extend posteriorly. Scale bars, 500 μm. (b) Schematic of sample collection timeline and further characterization through histology and 3D X‐ray microscopy (XRM) imaging. (c, d) Representative images showing sagittal and transverse cross‐sections of PTA‐stained femora and comparative knee joint safranin‐O histology across ages for both genotypes, highlighting notable morphological alterations in trochlear and condylar cartilage thickness (arrowheads) in Prg4 ^−/− mice, as well as epiphyseal osteopenia (asterisk) at 36 weeks of age. Scale bars = 1000 μm (XRM); 400 μm (safranin‐O). Ant, anterior; GP, growth plate; lat, lateral; med, medial; pos, posterior; pr, proximal; TG, trochlear groove.

2.4. Morphology Data Acquisition

To quantify phenotypic changes related to Prg4 loss‐of‐function, we adapted a registration‐based approach [21, 22] to the femur to collect anatomical landmarks and segmentations for morphometric analysis. Since the femora were stained and thus susceptible to differential staining artifacts, we corrected for intensity non‐uniformity using the N4 algorithm [23] and normalized the intensities. Then, a subsample (n = 25) of the femur scans was funneled through a computer‐automated workflow to generate a study‐specific atlas or anatomical average. The atlas was labeled with 43 anatomical landmarks that provided sparse yet comprehensive coverage of distal femoral shape (Figure 2a). Also, atlas segmentation of cartilage, bone (including marrow space), and growth plate was carried out for volume measurements (Figure S2). Next, each specimen was non‐linearly registered to the atlas. While the affine alignment was computed with a multi‐resolution framework [24], the subsequent non‐linear alignment was performed with the SyN (Symmetric Normalization) algorithm [25]. Afterwards, the affine and non‐linear transformations were concatenated and inverted to propagate the atlas labels to each individual image. All image processing was performed on the ARC compute cluster at the University of Calgary using the open‐source MINC (Medical Imaging NetCDF) software (https://github.com/BIC‐MNI/minc‐toolkit‐v2).

FIGURE 2.

Prg4 ^−/− distal femur changes in size and shape. (a) Representation of distal femur global reference atlas and the forty‐three anatomical landmarks configuration used for geometric morphometrics analyses. (b) Scatter plots of the first two principal component analysis (PCA) of Procrustes shape variables show how WT and Prg4 ^−/− femora exhibit distinct shapes along the major axis of variation (PC1), as represented by respective morphs (bottom) along the extremes of PC1 (exaggerated by a factor of 2). PC1 min: WT femur; PC1 max: Prg4 ^−/− femur. (c) 3D distance heatmaps represent the mean shape deviations between genotypes for each time point and indicate how far inward (blue) or outward (red) the mean WT femur shape needs to be displaced to match the mean Prg4 ^−/− femur shape at each respective age of assessment. (d) Scatter plot showing the relationship between fitted PC1 scores (i.e., shape scores predicted by regression of shape on size) and centroid size (log) for each experimental group.

2.5. Size and Shape Comparisons

To understand how genotype and age influence size variation across the femur, centroid size of each landmark configuration and volume of each segmented compartment were computed. Centroid size was calculated as the square root of the sum of squared distances of all landmarks from their centroid [26]. For shape comparisons, all landmark configurations were superimposed into a common shape space via Generalized Procrustes Analysis (GPA) [26, 27]. A series of geometric morphometric (GM) tests were then undertaken on the Procrustes shape data to examine the effects of genotype and age on femoral morphology. To account for confounders, femoral shape was regressed on sex to acquire sex‐adjusted Procrustes shape variables. Group differences were evaluated using Procrustes analysis of variance (ANOVA) [28] and further visualized using principal component analysis (PCA) and distance heatmaps. While the PCA visualizations involved deforming the mesh of the mean shape to the extremes of each PC via thin‐plate spline, the heatmap visualizations involved deforming the mesh of the WT mean to the Prg4 ^−/− via thin‐plate spline [29]. Then, the magnitude and direction of shape deviations at each point in the deformation were quantified. Comparison between WT and Prg4 ^−/− mean femoral shape was carried out for each age (Figure 2d) and comparisons within each genotype across timepoints were carried out relative to their respective 8‐week‐old baseline (Figure S2e,f). Geometric morphometric analyses were performed in R with the geomorph [30], RRRP [28, 31] and Morpho packages [32]. Additional morphometric analyses (linear distance/angle) were carried out between WT and Prg4 ^−/ regarding specific traits of interest: trochlear groove depth (TD) and sulcus angle (SA), anteroposterior distance (length) of medial (M_AP) and lateral (L_AP) compartments, bicondylar width (BCW), intercondylar notch width (IN), and condylar width (medial—MCW, lateral—LCW).

2.6. Condylar and Trochlea Cartilage RNA‐Seq Data Collection and Analysis

The articular cartilage from the trochlea and condyles of C57BL6 mice were dissected under a light microscope in PBS on ice and separately collected in 2 mL tubes containing 200 μL of TRIzol and a 5 mm stainless steel bead. Samples were obtained at postnatal stages P0, P30 and 1 year age (n = 4/group). Right and left sides of each specimen were pooled and sample processing was carried out as previously described [33]. In brief, each sample underwent a homogenization process at 50 Hz frequency for 2 min. After a brief cooling period (1 min), samples were subjected to a final homogenization at the same frequency for an additional 2 min. Samples were stored at −80°C until RNA extraction. A phenol‐chloroform reaction was initially employed for RNA extraction; thus, each sample was transferred into a new microcentrifuge Eppendorf tube with 200 μL of chloroform for every 1 mL of TRIzol present in the sample. After being thoroughly vortexed and incubated at room temperature for 2 min before being centrifuged at 4°C, 12 000 g force for 5 min. Post centrifugation, the aqueous layer was carefully separated and moved into a new Eppendorf tube. Final RNA extraction was performed using the Zymo Direct‐zol RNA MicroPrep kit according to the manufacturer's instructions, and the final RNA sample was eluted in 15 μL of nuclease‐free water. The resultant RNA was quantified using a Qubit per the manufacturer's protocols. Samples were also nano‐dropped to determine 260/230 and 260/280 values and stored at −80°C. Aliquots were also run on a TapeStation to determine their RNA integrity number (RIN) value, and only samples with RIN scores of 7 or higher were used for subsequent steps. Samples were normalized to a single concentration for cDNA library generation and sequencing, and libraries were prepared using KAPA mRNA Directional Library Preparation methods following the manufacturer's protocols. Twenty libraries were then run as quality control on a TapeStation. Then, a quantitative polymerase chain reaction (PCR) was performed on the library pool before sequencing the library on NextSeq High 2 × 38. The library was sequenced repeatedly using paired‐end sequencing on three lanes to obtain a minimum of 10 million reads for each sample, with some samples requiring only one or two lanes. Computational analysis of RNA‐seq data began with running Fast QC on each Fastq file to determine the per‐base sequence quality, GC content, etc., to ensure that all files met our standards. Because the samples had each been run on three lanes to achieve the desired number of reads, the reads for each sample were concatenated into a single file for R1 and a second file for R2. STAR version 2.6.0 was then used to map reads to the mouse genome, GRCm39. DESeq2 was then used to quantify differential Prg4 expression and downstream comparisons.

2.7. Biomechanical Assessment

Femurs (n = 5/group) were isolated under a dissection microscope (Leica). With condyles facing up, the femur bone shaft was placed into a pipette tip filled with cyanoacrylate glue. The construct was secured to a customized sample holder and affixed to the Mach‐1 (v500csst Biomomentum) mechanical tester within a plexiglass chamber that contained PBS. Thirty‐one measurements were performed across the medial and lateral femoral condyles by fast (0.02 mm/s) indentation of articular cartilage with a 0.3 mm diameter indenter probe to a displacement of 0.02 mm [20]. Indentation tests were performed under stress‐relaxation, after which a 30G needle replaced the indenter probe, and thickness was determined at immediately adjacent locations by needle probing technique, as previously described [20, 34]. Measurement points were performed in the same chronological order for all samples. Load–displacement curves were recorded for indentation and needle probing assessments and used for instantaneous modulus and cartilage thickness determination. Analysis was performed on Automated Indentation and Thickness Batch Analysis software (v.2.0.2 Biomomentum) to obtain cartilage thickness and instantaneous moduli (Poisson's ratio ν = 0.5), which was determined by fitting the Hayes model [35] to the force/displacement curves. Heatmaps for data display were generated using Analysis‐MAP software (v 1.0.0.2).

2.8. Statistical Analysis

Statistical analyses aside from geometric morphometrics were performed using GraphPad Prism v.8. Unless stated otherwise, results are presented as mean ± standard error of the mean (SEM). For multiple comparisons, Holm‐Šidák or Bonferroni correction was applied after performing Mann–Whitney non‐parametric test or ANOVA. Statistical tests are noted in the figure legend. A p‐value less than 0.05 was regarded as statistically significant.

3. Results

3.1. Phenotypic Abnormalities in Prg4^−/− Distal Femur

Gross morphological differences in distal femora were obvious at 8 weeks of age in Prg4 ^−/− mice (Figures 1a and S1). At the macroscopic level, they exhibited misshaped trochlear grooves with a deepened center sulcus (anterior view) and outward projection of its medial and lateral facets (medial side view and distal views). Prg4 ^−/− femora were also generally enlarged compared to age‐matched wild‐type (WT) femora (distal view). The posterior view demonstrated no noticeable morphological differences in femoral condyle shape (Figure 1a). With aging, some of the described morphological traits appeared to change. For instance, posterior extension of the medial condyle (medial side view) and flattening of the trochlear groove proximal center sulcus (anterior and distal views) were qualitatively observed in aged (36‐week‐old) femora. To further characterize these spatiotemporal phenotypic changes, we isolated the femur at various postnatal ages and carried out histopathological and X‐ray microscopy (XRM) imaging (Figure 1b). Each femur was stained with the contrast agent phosphotungstic acid (PTA) before being imaged to differentiate murine soft tissues from bone structures [19, 20]. Tissue contrast was readily apparent in the sagittal plane, highlighting visibly enlarged articular cartilage in the trochlear region and, to a lesser extent, on the condyles of Prg4 ^−/− mice (Figure 1c, arrowheads). Closer examination of comparative XRM and histological transverse cross‐sections confirmed an overall dysmorphic femur in Prg4 ^−/− mice, with abnormalities in both soft and hard tissues compared to WT controls across all time points. In line with previous reports, an osteopenic phenotype was observed in Prg4 ^−/− mice [5, 16], with visible reduction of subchondral trabecular bone by 36 weeks of age (Figure 1c). Together, this data indicated that postnatal femoral morphology is impacted in the Prg4 loss‐of‐function mice.

3.2. Distal Femur Size and Shape Are Altered in Prg4^−/− Mice

Next, morphometric analyses were employed to quantitatively assess the phenotypic femoral abnormalities related to Prg4 loss‐of‐function. Landmark‐based geometric morphometrics is commonly employed in studying anatomical variation between natural populations or experimental groups due to its ability to capture and quantify subtle and complex shape alterations [21, 36]. Here, a study‐specific atlas (reference average volume) was generated and labeled with a standardized set of 43 anatomical landmarks to comprehensively represent the distal femur geometry (Figure 2a). The morphological displacement of the landmark configuration of each specimen was assessed to understand the effects of genotype and age on femoral size and shape. A landmark‐based centroid size comparison was carried out since Prg4 ^−/− mouse femora appeared visually larger. On average, Prg4 ^−/− femora exhibited a 6.8% increase in centroid size compared to WT femora (WT: 9.47 ± 0.153 mm vs. Prg4 ^−/− : 10.11 ± 0.162 mm; p < 0.001), holding sex and age constant. We also investigated differences in the size of specific tissues between mouse strains and found that Prg4 ^−/− mice displayed substantially more cartilage, bone (with marrow space) and growth plate volume (Figure S2a–d), yielding 20%, 13.4%, and 22.1% respective increases on average compared with the WT group. Of note, sex differences did not significantly contribute to variations in overall femur size (i.e., centroid size) (p = 0.627). This finding is further supported by individual assessments of bone, cartilage, and growth plate volumes, which also showed no significant differences between the sexes. Although femoral shape did differ significantly between males and females (R ² = 3.7%, p < 0.001)– driven mainly by modest widening of the male distal femur, especially around the condyles at early timepoints (Figure S3a,b)—these differences were relatively subtle when compared to those arising from genotype or timepoint. As our primary focus was on genotype, timepoint, and size‐related variation, the effect of sex was regressed out and not explored further.

To visualize the major sources of femoral shape variation among genotypes and ages, PCA was performed on sex‐adjusted shape variables (Figure 2b). The first principal component (PC1) accounted for 40% of the total variation in femoral shape. Moreover, distinct clusters between genotypes could be observed along PC1, with Prg4 ^−/− femora scoring more positively and WT femora scoring negatively. Shape morphs along the extremes of PC1 (i.e., PC1 min: WT femur and PC1 max: Prg4 ^−/− femur) (Figure 2b) helped visualize significant sources of shape variation between genotypes, including mediolateral narrowing and proximodistal/anteroposterior elongation. Notably, many of these changes were recapitulated in distance heatmaps comparing the WT and Prg4 ^−/− femoral mean shapes (i.e., how far the mean WT femur shape needs to be displaced to match the mean Prg4 ^−/− femur shape) for each time point (Figure 2c). Regions of outward expansion (red) were mainly traced back to the medial condyle and proximal region of the trochlear groove, and to a lesser extent, the lateral condyle and outer ridges of the trochlear groove. Conversely, areas of inward contraction (blue) were of greater magnitude at the central sulcus of the trochlear groove and lateral/medial epicondyles.

One essential aspect of shape variation that PC1 often captures is the allometric effect of size [37]. Allometry refers to the correlation between size and shape, and despite being a ubiquitous phenomenon in nature, its importance in biomedical studies of morphology is underappreciated [37, 38]. To investigate how shape correlates with size, fitted PC1 scores (i.e., predicted shape score by regression of shape on size) were compared to centroid size (log‐transformed) for each experimental group (Figure 2d). At any given centroid size, Prg4 ^−/− femora scored more positively in predicted shape and deviated from the allometric pattern expected based on WT femora. Collectively, these results demonstrate that Prg4 deficiency impacts both the size and shape of the distal femur, leading to larger femora on average, which are not simply geometrically scaled versions of their WT counterparts, but also display distinct shapes (narrower, extended, etc.), and these effects covary.

3.3. Changes in Individual Anatomical Features

Next, individual anatomical features of interest and clinical relevance, such as the intercondylar notch width (IN) and trochlear groove sulcus angle (SA) [39, 40, 41] were quantified (Figure 3a–j). At 8 weeks of age, Prg4 ^−/− mouse femora had a deeper trochlear groove associated with lower values of sulcus angle, indicating a steeper and more pronounced groove compared to the WT group. As the mice aged, Prg4 ^−/− samples gradually approximated the trochlear morphology of WT femora (Figure 3b,c) by becoming shallower and flatter. This phenomenon was associated with abnormal soft and hard tissue remodeling, as seen histologically (Figure 3j). In contrast, the WT group demonstrated relatively stable trochlear features. In addition to the differences in trochlear groove, Prg4 ^−/− mice displayed a distinct and consistently narrower bicondylar width (Figure 3c). Interestingly, as the WT mice aged, they showed an increase in intercondylar width while simultaneously experiencing a decrease in both medial and lateral condyle widths, resulting in minimal changes in the bicondylar width (Figure 3d–g). Linear distance measurements also demonstrated an anteroposterior elongation in Prg4 ^−/− mice compared to WT in both the medial and lateral compartments (Figure 3h,i), being more prominent on the former. Histological examination suggested that this elongation may be influenced by micro‐environmental constraints (Figure 3k) and likely associated with synovial hyperplasia and contracture phenotypes observed and previously described for Prg4 ^−/− mice [5].

FIGURE 3.

Quantitative morphological analysis of individual anatomical features. (a) XRM measurements of femoral anatomical features in WT and Prg4 ^−/− mice (n = 4–5 mice/group) of interest included (b) trochlear groove depth (TD) and (c) sulcus angle (SA), (d) bicondylar width (BCW), (e) intercondylar notch width (IN) and the (f, h) width and height of the lateral condyle (LCW and L_AP) and (g, i) medial condyle (MCW and M_AP). Data presented as mean ± SEM (n = 4–5 mice/group). Analysis of variance ANOVA control vs. Prg4 ^−/− mice for individual endpoints with Bonferroni correction for pair‐wise comparison. p‐values reported. (j, k) Representative histology demonstrating changes in trochlear groove and anteroposterior extension of the femur in Prg4 ^−/− vs WT mice. (l, m) Expression of Prg4 (by RNAseq) varies with age in the femoral condyle articular cartilage in WT mice. (n) Heatmaps represent the mean shape deviations for the Prg4 ^−/− mice at 36w age relative to its respective 8w old femur. Note the inward contraction regions at the trochlear groove ridges and medial condyle (dark blue), and to a lesser extent on lateral condyle.

The overt phenotype observed between trochlea and condyle cartilage in Prg4 mutant versus WT femora prompted us to investigate the transcript levels of Prg4 through RNAseq at different postnatal stages (P0, P30, P1 year). The analysis revealed increased Prg4 expression in condyles over time, in contrast to relatively higher expression levels at P0 for trochlear cartilage. These findings suggest that Prg4 loss‐of‐function may have compartment‐specific effects that are also age‐dependent, impacting trochlear morphogenesis to a higher degree than condylar cartilage, but important for maintenance of condylar cartilage with aging. In this regard, mean shape deviations were visually apparent at the trochlear groove ridges and medial condyle and to a lesser extent on the lateral condyle for the Prg4 ^−/− mice with aging (36 weeks compared to 8‐weeks of age) (Figures 3n and S2f). The presence of areas of contraction in these specific regions could be associated with biomechanical stress and imbalances in the joint, possibly contributing to the degradation of the articular cartilage over time in the Prg4 ^−/− mice, which has been extensively described in the literature [5, 13, 14].

3.4. Prg4 ^−/− Morphological Changes Associate With Condyle Articular Cartilage Degeneration

Since articular cartilage in Prg4 ^−/− mice has been previously characterized as being thicker and mechanically inferior [14, 20, 42], we wanted to test the hypothesis that the localized shape changes with advanced age seen at Prg4 ^−/− condyle regions would correlate with spatial changes in thickness and mechanical properties. We employed a previously validated protocol for indentation and cartilage thickness mapping of murine articular cartilage [20]. The data demonstrated regional differences between genotypes, with consistently thicker articular cartilage in Prg4 ^−/− condyles, associated with lower modulus, irrespective of age (Figure 4a,b). However, with aging, thinning of Prg4 ^−/− cartilage was observed in the mid‐outer portion of the condyles, particularly prominent in the medial side, despite no corresponding apparent increase in instantaneous modulus (Figure 4a,b arrowhead). These findings were also observed histologically, wherein areas of cartilage fibrillation and erosion primarily localized to the mid‐to‐outer portions of the medial condyle in Prg4 ^−/− mice were observed (Figure 4c, arrowheads), which worsened with aging (Figure 4c,h,i). Site‐specific measurements in the mid‐outer regions of both lateral (Figure 4d,f) and medial condyles (Figure 4e,g) highlighted differences in thickness and instantaneous modulus. These findings provide valuable insights into the biomechanical alterations and histological changes associated with the absence of Prg4 in the femoral condyles.

FIGURE 4.

Abnormalities in articular cartilage structure and biomechanics of Prg4 ^−/− mice with aging. Color maps of averaged values for lateral and medial condyle cartilage thickness (a) and instantaneous modulus (b) as determined by Hayes et al. [35] elastic model at 20% strain. Data highlights regional differences between genotypes for both parameters, with articular cartilage being consistently thicker in Prg4 ^−/− condyles, and of lower stiffness, irrespective of age. Note, however, thinning of Prg4 ^−/− cartilage at mid‐outer portion of condyles with aging, being more prominent in the medial side (arrowhead) and corresponding to an increase in stiffness. In contrast, condyle cartilage surfaces in WT mice demonstrated little variation in thickness with aging, albeit matrix stiffening with peak at mid‐age (16w and 20w) (c) Representative images of safranin‐O/fast‐green staining of articular cartilage of the medial femoral condyles of WT (top) and Prg4 ^−/− mice (bottom) from 8‐ to 36‐weeks of age, highlight areas of cartilage fibrillation and erosion (arrowhead) in Prg4 ^−/− mice primarily on mid‐to‐outer portions of condyles. Scale bars, 100 μm. Site‐specific measurements for mid‐outer regions (schematic bottom corner) wherein differences in thickness and instantaneous modulus were identified during mapping for medial (d, f) lateral and (e, g) medial condyles. Thickness and instantaneous modulus presented as mean ± SEM (n = 5 mice/group, 8–10 positions/mouse) are shown. Student's t‐test control vs. Prg4 ^−/‐  mice for individual endpoints with Holm‐Šidák correction. p‐values reported. OARSI histopathological scoring of (h) lateral and (i) medial condyles cartilage damage. Data represented as mean ± SEM (n = 5 mice/group), Mann–Whitney tests with Holm‐Šidák correction. p‐values reported. MFC, medial femoral condyle; LFC, lateral femoral condyle.

4. Discussion

This study investigated phenotypic abnormalities in adult joint tissues associated with Prg4 loss‐of‐function, specifically the impact on distal femur shape, size, and articular cartilage degeneration with aging. The data presented here provide empirical evidence that the loss of Prg4 profoundly impacts distal femur geometry and the anatomy of its structural components. At the tissue level, changes in size and shape within Prg4 ^−/− femora are directly associated with corresponding changes in bone and cartilage tissues (bulk and internal structure/composition), resulting in distal femur enlargement, increased anteroposterior extension, and reduced mediolateral width and misshaped trochlear groove when compared to WT controls.

Changes in condylar articular cartilage thickness associated with inferior compressive biomechanical properties in young mice (up to 16 weeks of age) are in keeping with previous reports [14, 20]. Here, we add to the body of knowledge by demonstrating that with aging, the disruption in cartilage biomechanical properties and joint lubrication due to Prg4 loss is associated with cartilage thinning predominantly in the medial condyle and in its mid‐anterior aspect, which is also noticeable at the histological level. This suggests that mechanical joint loading factors are at play within the progressive degenerative phenotype that spontaneously occurs within these Prg4 ^−/− mice. Of note, loss of Prg4 has been shown to alter ankle range of motion; mice develop an abnormal hopping gait [5, 16], strengthening the link between joint geometry and function within these mice. Notably, this regional variation could also be captured in distance heatmaps generated via landmark‐based geometric morphometrics analysis, highlighting the ability of this technique in quantifying size and shape phenotypes in even relatively small‐scale investigations [21, 22].

In vivo studies in Prg4 ^−/− mice have underscored the importance of this molecule as a boundary lubricant and inhibitor of cell adhesion. These lubrication and chondroprotection properties are crucial for maintaining tissue homeostasis within the adult joint and preventing spontaneous degenerative cartilage changes [43, 44]. However, the implications of the absence of Prg4 on joint formation and patterning, given the onset expression of Prg4 occurs during embryonic development [5] are less understood. While Prg4 deficiency does not prevent knee joint formation, it appears to influence the morphology and structural components of the joint during early postnatal development [15, 42, 45].

The early postnatal period is a crucial time for the growth and maturation of joint tissues, including bone and cartilage, which attains its characteristic zonal structure [2, 46]. It is possible that the lack of proper lubrication due to Prg4 deficiency can hinder smooth sliding yet increase adhesion and shear between opposing cartilage surfaces, thereby creating abnormal stresses and strains in both the developing and maturing joint tissues—particularly in areas of cartilage‐cartilage contact. This disruption in normal movement and mechanical signaling during critical growth and maturation phases could result in abnormal chondrocyte activity, matrix organization, and overall joint structure seen in Prg4 ^−/− mice, ultimately manifesting into the dysmorphology observed by 8 weeks of age. The complexity of joint development and the specification of joint shape arise from the interplay of genetic factors, mechanical loading, and biochemical signaling that influence growth and remodeling. While originally researched due to its lubricating properties, over the past decade Prg4 has been implicated in various biological processes, such as regulation of inflammation [47, 48], adhesion‐dependent proliferation [5], fibrosis [49], angiogenesis and wound healing [50], and bone homeostasis [12, 16]. Many of these processes are critical to the developing limb, and our study raises important questions about the role of Prg4 in regulating these developmental processes. However, it is essential that we examine these functions in the context that Prg4 deficient animals are born with functional joints. We know that these joint tissues (bone, cartilage and synovium) in Prg4 deficient animals are biased/primed to degeneration, which is reminiscent of specific populations of patients suffering from osteoarthritis. This begs the question of what role could Prg4 be playing during early human development and could its differential regulation across a population increase the risk of joint disease later in life in specific individuals.

To address some of these questions, it would be interesting to gain insight into the extent of dysmorphologies in Prg4 ^−/− mice and infer possible effects on joint kinematics and contact mechanics. An inherent limitation of Prg4 ^−/− mouse studies is that this model has a global gene deletion of Prg4. Since Prg4 expression has been shown in other musculoskeletal tissues, like bone and tendons [51, 52], as well as in other tissues like the liver, possible local and systemic confounding factors cannot be excluded.

In summary, this study provides new insights into the intricate role of Prg4 loss‐of‐function on overall joint morphology and articular cartilage health and function in adulthood. Our findings shed light on the intricate relationship between joint geometry and function, reinforcing the importance of Prg4 in maintaining joint homeostasis, but also at potential implications of Prg4 deficiency during early postnatal development, where disrupted lubrication may affect joint maturation and morphology.

Author Contributions

Conceptualization: A.O.M., R.J.K. Methodology: A.O.M., J.D., N.D., C.R.C., Z.L., B.H. Investigation: A.O.M., J.D. Visualization: A.O.M., J.D. Supervision: B.H., T.D.C., J.A.B., W.B.E., R.J.K. Writing – original draft: A.O.M. Writing – review and editing: A.O.M., J.D., N.D., C.R.C., B.H., Z.L., T.D.C., J.A.B., W.B.E., R.J.K.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: fsb270974‐sup‐0001‐FigureS1.docx.

Masson A. O., Devine J., Das N., et al., “Femur Shape Changes in Prg4‐Deficient Mice: Morphological Insights Into Joint Well‐Being,” The FASEB Journal 39, no. 16 (2025): e70974, 10.1096/fj.202501775R.

Data Availability Statement

All data are available in the main text or the Supporting Information.