Keywords

1 Introduction

Articular cartilage is the soft tissue at the end of bones that enables joint locomotion, energy dissipation, and lubrication [40]. These functions are endowed by the specialized extracellular matrix (ECM) of cartilage, a hydrated composite of collagens and proteoglycans that are synthesized by residing chondrocytes [26] (Fig. 5.1a). Osteoarthritis (OA), the most prevalent musculoskeletal disease, is characterized by the irreversible breakdown of cartilage ECM, resulting in severe joint pain and limited motion [45]. Due to its avascular and dense nature, cartilage has very limited self-repair capabilities, and regenerative therapies often fail to restore the structure and function of healthy tissue [29]. This renders it crucial to understand the establishment, homeostasis, and disease-induced degeneration of cartilage ECM. The ECM mainly consists of ~65–70% w/w water, ~20–30% collagens, ~5–10% proteoglycans [40] as well as DNAs and other minor proteins/proteoglycans (Fig. 5.1a–c) [28]. In vivo, the collagen network is primarily responsible for cartilage tensile stiffness, while aggrecan and its negative fixed charges are the key determinants of cartilage compressive resistance and fluid flow-associated energy dissipative properties [40].

Fig. 5.1
A set of illustrations, S E M images, and electron microscopy image labeled a, b, and c. A has cylindrical strands of collagen with horizontal lines. It has fibrous structures labeled aggrecan, with small tubules called hyaluronan. There is link protein on the tubules. B has 3 S E M images at 200 nanometers. The images have strand-like interconnected structures for superficial, middle, and deep regions. C is at the scale of 200 nanometers. It has tiny extensions that extend from a horizontal central strand.

An overview of articular cartilage extracellular matrix (ECM). (a) Schematic illustration of the major constituents in cartilage ECM: type II/IX/XI collagen fibril network and aggrecan-hyaluronan (HA) aggregates. (b) Depth-dependent nanostructure of collagen fibril network visualized by helium ion microscopy on proteoglycan-removed rabbit cartilage. Adapted with permission from Ref. [56]. (c) Nanostructure of aggrecan-HA aggregates imaged by transmission electron microscopy. (Adapted with permission from Ref. [3])

Full size image

In the past few decades, there have been many attempts in understanding cartilage biomechanics in health and disease, in the hope to gain new insights into disease progression and functional regeneration. In the 1980’s, Mow and colleagues applied the biphasic poroelasticity theory [2] to interpret the time-dependent, energy dissipative properties of cartilage. This theory addressed the key role of molecular friction arising from water-solid matrix interactions during fluid flow in contributing to energy dissipation [42]. Grodzinsky and colleagues further integrated the concept of electrical streaming potential with cartilage mechanical deformation and underscored the crucial role of fixed charges in overt tissue biomechanics [16, 17]. These seminal studies established the foundation of modern cartilage biomechanics theory. At the same time, numerous experimental tools have been implemented to delineate the contributions of each matrix constituent to cartilage mechanical properties, including confined and unconfined compression, indentation, dynamic oscillatory loading, and shear [39]. For instance, investigating the biomechanics of normal and glycosaminoglycan (GAG)-depleted cartilage have shown that aggrecan and its fixed charges directly contribute to ~50% compressive modulus of cartilage [60].

Despite these advances in tissue-level studies, disease intervention and regeneration remain elusive. This is at least partly because the tissue-level investigatory approaches are unable to account for the salient heterogeneity and a high level of complexity of cartilage structural hierarchy from nm-to-mm scales. For example, the collagen network is dominated by types II/IX/XI collagen heterotypic fibrils [15], with a minor amount of type III collagen co-assembling on the surface of collagen II fibrils [58]. The collagen fibrils (diameter ~ 30–80 nm) vary in orientation and diameter with depth in the tissue [9], from being transverse in the superficial layer, to random in the middle layer, then predominantly perpendicular in the deep layer, accompanied by an increasing gradient of fibril diameter and proteoglycan concentration (Fig. 5.1b) [56]. Aggrecan, the major proteoglycan, has a “bottle-brush” structure, and is decorated with highly negatively charged chondroitin sulfate (CS) and keratan sulfate (KS) GAGs along its core protein [26]. In vivo, aggrecan is end-attached to the linear hyaluronan (HA) molecules via its G1 domain at the N-terminal [27], and this interaction is further stabilized by link protein [4]. In the ECM, these supramolecular aggregates are entrapped within ~100 nm nanopores between collagen fibrils at ~50% molecular compressive strain [59], thereby adopting a highly compacted configuration that endows the tissue with its high fixed charge density and osmotic swelling pressure. The electron microscopy study by Buckwalter and Rosenberg highlighted the complexity in the assembly and retention of aggrecan in vivo, and provided the first direct visual evidence of the aggrecan-HA aggregation (Fig. 5.1c) [3]. Given these complexities, understanding the ECM from the molecular level is necessary for developing effective disease intervention and tissue regeneration.

Dr. Grodzinsky is the pioneer in studying the molecular mechanics of cartilage ECM constituents. Through collaboration with Dr. Christine Ortiz, a world-renowned scientist in polymer nanomechanics and atomic force microscopy, this team has made numerous transformative discoveries on the nanostructure and nanomechanics of cartilage, with a focus on the major proteoglycan, aggrecan. This chapter provides a brief summary of Grodzinsky’s contributions to the understanding of aggrecan within the context of cartilage function, disease, and regeneration. This chapter begins with the summary of the ultrastructural and nanomechanical studies of native aggrecan (Sect. 5.2), followed by the overview of applying the knowledge of aggrecan to understanding cartilage aging, disease and tissue engineering (Sect. 5.3), and then, the discussion of more recent studies on other native and biomimetic proteoglycans that were directly inspired by the Grodzinsky’s work (Sect. 5.4), and finally, concludes with a summary and future outlook (Sect. 5.5).

2 Ultrastructure and Nanomechanics of Aggrecan

One seminal contribution by Grodzinsky and colleagues is the direct visualization of the ultrastructure of aggrecan and its GAG side chains, which was the first of its kind [43]. In this study, Ng et al. deposited aggrecan molecules extracted from fetal epiphyseal and mature nasal bovine cartilage samples onto 3-aminopropyltriethoxysilane (APTES)-treated, positively charged, atomisticly flat mica surfaces. The nanostructure of aggrecan was then revealed via tapping mode AFM-imaging at a spatial resolution of ~2 nm. Imaging aggrecan at such unprecedented resolution enabled not only visualization of its “bottle-brush”-like molecular architecture, but also direct quantification of its structural parameters (Fig. 5.2a). This includes the core protein contour length, Lc, end-to-end distance, Ree, degree of extension, Ree/Lc, and the packing density and length of GAG bristles. In turn, the persistence length, Lp, as calculated from the worm-like chain model (~110 nm for fetal epiphyseal aggrecan), illustrated the highly extended conformation of aggrecan in its equilibrium state. Conversely, this study also showed that in vivo, aggrecan adapted a highly compacted conformation, for that the in vivo concentration is at least 40× higher than the densely packed form imaged on mica (Fig. 5.2b). To this end, the contrast between aggrecan from fetal epiphyseal versus mature nasal cartilage highlighted the large variation of its ultrastructure with tissue source and age (Fig. 5.2a). Furthermore, these high resolution images clearly illustrated the dominating role of longer CS-GAGs in contributing to the molecular conformation of aggrecan. Indeed, a follow-up study by Lee et al. compared the structure and conformation of aggrecan from a 29-year-old human donor subjected to chondroitinase ABC and keratanase II treatment, and confirmed that the extension and conformation of aggrecan is predominantly governed by the longer CS-GAGs, rather than the shorter KS-GAG chains [36].

Fig. 5.2
A set of 3 A F M images, a and b. A has 2 images at 100 nanometers of fetal epiphyseal and mature nasal, respectively. It is a fibrous and strand like structure on the surface of mica. It has labels for N and C. B is at the scale of 100 nanometers and has textured and rough surface.

Ultrastructure of aggrecan via tapping mode atomic force microscopy (AFM) imaging. (a) Nanostructure of individual fetal bovine epiphyseal and mature bovine nasal aggrecan deposited on atomicly flat mica surface. (b) Nanostructure of densely packed fetal bovine epiphyseal aggrecan monomers, illustrating the highly compressed conformation of aggrecan in vivo (~ 40× higher than the aggrecan packing imaged here). (Panels (a) and (b) are adapted with permission from Ref. [43])

Full size image

In addition to nanostructure, Grodzinsky and colleagues also, for the first time, assessed the nanomechanics of aggrecan under multiple deformation modes, including compression, energy dissipation, shear, and adhesion. Building on earlier work of CS-GAG nanomechanics [51,52,53], Dean et al. chemically functionalized aggrecan with thiol-groups at its N-terminal, and end-attached thiol-functionalized aggrecan onto gold-coated planar silicon substrates and microspherical AFM colloidal tips (R ≈ 2.5 μm). This set-up established a 2D biomimicry assembly of aggrecan at ~50 mg/mL, near its physiological packing density (~20–80 mg/mL), thus enabling the studies of aggrecan interactions under in vivo-like conditions. With this biomimetic system, the team performed an in-depth analysis of key molecular mechanical behaviors of aggrecan. First, compressive nanomechanics was quantified using force spectroscopy and contact mode AFM imaging in aqueous solutions with varied ionic strength (IS) conditions [10, 11]. As expected, the long-range repulsion force between two opposing aggrecan layers extended to >1 μm in IS = 0.001 M solution, while the distance and magnitude of compression resistance decreased drastically with increasing IS from 0.001 to 1.0 M (Fig. 5.3a). Applying Poisson-Boltzmann-based models, this study confirmed that the electrical double layer (EDL) repulsion interactions arising from CS-GAGs play a dominating role in the compressive nanomechanics of aggrecan, in comparison to other non-electrostatic factors such as steric hindrance and conformational entropy. Importantly, under in vivo conditions (IS = 0.15 M), given that the Debye length, κ−1 ≈ 0.8 nm, is at the same order of GAG-GAG packing distance (~ 2–3 nm), the salient heterogeneity in electrical potential was a key factor in determining the magnitude of repulsion. While the continuum Donnan model substantially overestimated the repulsion force, both the unit cell model [5] and charged rod model [12], which accounted for the nanoscale spatial heterogeneity of electrical potential, accurately predicted the aggrecan-aggrecan repulsion [10].

Fig. 5.3
A set of illustrations and graphs, a, b, c, and d. A. Illustration for downward tip motion with graph that plots force versus distance. The trend is declining. B. Illustration for tip motion profile with graph for phase angle versus frequency. The peaks are between 100 and 1000 hertz. C. Illustration for horizontally bidirectional tip motion with graph that plots lateral force versus normal force. the trend is inclining. D. illustration for upward tip motion with graph for adhesion energy versus surface dwell time. the trend is inclining except for one value.

Nanomechanics of aggrecan measured between two opposing layers of aggrecan via atomic force microscopy (AFM)-based nanomechanical modalities. Experiments were performed applying aggrecan functionalized colloidal tips (R ≈ 2.5 μm, except that R ≈ 22.5 μm for panel (b) to aggrecan-functionalized planar substrates in NaCl aqueous solutions at different ionic strengths (IS = 0.001–1.0 M). Left panels: Schematics of experimental set-ups. For panel (c), the substrate was prepared via micro-contact printing to form micropatterned surface of hydroxyl-terminated self-assembled monolayer (OH-SAM, HS(CH2)11OH) and aggrecan monolayer. Right panels: (a) Compressive force-versus distance curves as a function of bath IS via colloidal molecular force spectroscopy. Adapted with permission from Ref. [10]. (b) The magnitude of phase angle, ϕ, of newborn human aggrecan as a function of dynamic frequency and bath IS via AFM-based nanorheometric test (mean ± 95% CI, n = 6). Adapted with permission from Ref. [44]. (c) Lateral versus applied normal force curves as a function of bath IS via lateral force microscopy (mean ± SD, n = 8). Adapted with permission from Ref. [24]. (d) The total aggrecan-aggrecan adhesion energy as a function of surface dwell time and bath IS via molecular force spectroscopy (mean ± SEM, n ≥ 30). (Adapted with permission from Ref. [23])

Full size image

Following the studies of elasticity, Nia et al. probed the energy dissipative, poroelastic nanomechanics of aggrecan using the custom-built AFM-nanorheometer [44]. Similar to the case of elastic modulus, the energy dissipation of aggrecan layer is also largely governed by the EDL repulsion, as illustrated by the salient dependence of phase angle on bath IS (Fig. 5.3b). Meanwhile, the dynamic oscillatory loading responses of three specimens were compared: normal cartilage, GAG-depleted cartilage, and the end-attached aggrecan monolayers. In the low frequency elastic domain, the modulus of GAG-depleted cartilage, EL, was ~1.5 × lower than that of the normal cartilage, while the modulus of aggrecan layer is ~7 × lower. Despite this much lower modulus, the aggrecan monolayer had comparable hydraulic permeability, k, to the native cartilage, while that of the GAG-depleted cartilage was ~24 × higher. Such contrast underscored the direct contribution of aggrecan and its sGAGs to cartilage fluid flow and pressurization capabilities. This is because the closely spaced GAG chains of aggrecan, with an effective pore size ~2–4 nm, provide the main resistance to intra-tissue fluid flow in cartilage, as manifested in the GAG-GAG nanomolecular model of hydraulic permeability [14], while the collagen network has a much larger pore size (~100 nm), resulting in elevated hydraulic permeability. Similar to the case of elasticity, a more complex structural model was needed to capture the magnitude of aggrecan-endowed energy dissipation. The fiber-reinforced model [55] or transversely isotropic model [54], which accounted for cartilage tension-compression asymmetry, were able to quantitatively capture the degree of energy dissipation, while an isotropic poroelastic model would markedly underestimate these values.

Applying lateral force microscopy (LFM), Han et al. elucidated the shear nanomechanics of both single and two opposing aggrecan layers [24, 25]. The shear resistance of aggrecan was quantified as a function of aggrecan layer height and applied normal force under varied IS. The lateral linearity ratio, μ (= dFlateral/dFnormal), was found to vary significantly with both IS (Fig. 5.3c) and lateral displacement rate, suggesting that the shear resistance was also largely governed by both EDL repulsion and fluid flow. At lower IS, given the dominance of EDL repulsion, aggrecan exhibited a more extended conformation. Therefore, a lower lateral proportional coefficient could be attributed to the minimal interdigitation between opposing aggrecan and strong water hydration effects surrounding negative charges, similar to the highly lubricative case of negatively charged synthetic polyelectrolytes. To this end, the shear of two opposing aggrecan layers also yielded a lower lateral coefficient, due to stronger EDL repulsion than just one single layer. In addition, divalent Ca2+ ions (~2–4 mM in cartilage [40]) were also found to mediate the shear behavior through extra screening of EDL repulsion and potentially the ion bridging effect. Collectively, EDL repulsion dominates not only the compressive, but also shear nanomechanics of aggrecan. It is also worth noting that the low lateral coefficient, μ, observed here does not imply a role of aggrecan in cartilage lubrication, as the concentration of aggrecan on cartilage surface is very low, and the synovial fluid is dominated by fragmented aggrecan, which lacks the G1-domain that enables its binding to HA [50].

Interestingly, despite the dominance of strong EDL repulsion, aggrecan also exhibited marked adhesive interactions with adjacent aggrecan molecules [23], and with collagen II fibrils [49]. When compressed at physiological-like molecular strain (~50% [59]) for 0–30 seconds, aggrecan was found to undergo pronounced adhesion, with a magnitude at ~1 pN between per pair of aggrecan-aggrecan, and ~ 0.3 pN per aggrecan molecule versus collagen II fibrils in physiological-like solution. Such adhesion was attributed to non-specific interactions, such as hydrogen bonding, hydrophobicity, ionic interactions as well as physical entanglement. Increasing EDL repulsion by lowering IS effectively limited the intermolecular contact and reduced the adhesion between the layers (Fig. 5.3d). On the other hand, Ca2+-mediated ion bridging further enhanced the adhesion by providing additional ionic linkage. Given that the highly compressed state of aggrecan mimics the physiological molecular strain in unloaded cartilage, these non-specific interactions were hypothesized to be an important biophysical factor that helps stabilizing the retention of fragmented aggrecan in healthy cartilage and contributes to the integrity of cartilage ECM.

3 Implications for Aging, Disease and Regeneration

Following these fundamental studies of aggrecan, Grodzinsky and colleagues further applied the experimental paradigm to gain new molecular insights into cartilage disease pathogenesis and regenerative medicine. One important application was to assess age-associated changes of aggrecan polymorphism using human cartilage samples. Applying tapping mode AFM imaging and force spectroscopy, Lee et al. compared aggrecan molecules from newborn and 38-year-old adult donors [36]. Aggrecan from newborn cartilage exhibited superior nanostructure and compressive nanomechanics relative to that from adult cartilage. First, the adult aggrecan population consisted of substantially more fragmented monomers that did not have the G1 or G3 globular domain (Fig. 5.4a, left panel). Such observation illustrated that aggrecan fragmentation could be a normal homeostasis process during growth and aging, and is a prevalent feature even in healthy adult cartilage. It also indicated that the retention of aggrecan in vivo may require additional mechanisms beyond the aggrecan-HA association, such as the aggrecan-aggrecan and aggrecan-collagen II adhesion [23, 49]. And, even for the sub-population of full length aggrecan, the newborn aggrecan exhibited longer core protein length, longer CS-GAG length, and denser packing of CS-GAGs, contributing to much stronger compression resistance (Fig. 5.4a, right panel). Taken together, these results provide direct molecular-level evidence about the effects of age on cartilage matrix changes, which may assist the intervention of age-associated cartilage degeneration and OA initiation.

Fig. 5.4
A set of A F M images, graph, and immunofluorescence images, a, b, and c. A. 2 A F M images at the scale of 500 nanometers for newborn and 38 years old adult. The first image has fewer strands than the second. A graph plots force versus distance. The trend is declining. B. 2 immunofluorescence images from different planes with R approximately equal to 2.25 micron. C. 2 A F M images for adult B M S Cs and adult chondrocytes.

Applications of AFM in the studies of aggrecan in cartilage aging, disease initiation and tissue engineering. (a) Left panel: Tapping mode AFM height images of newborn and adult (38-year-old) human aggrecan monomers. Arrow heads: globular domains. Right Panel: Compression resistance curves of end-attached newborn and adult human aggrecan monolayer measured at 0.01 M ionic strength via colloidal force spectroscopy. (Adapted with permission from Ref. [36]). (b) Left panel: Schematic illustration of immunofluorescence (IF)-guided AFM nanomechanical mapping on mature murine cartilage cryosection using a microspherical colloidal tip; the pericellular matrix (PCM) is immunolabeled with collagen VI. Right panel: Representative indentation modulus maps show the early reduction of PCM and territorial/inter-territorial extracellular matrix (T/IT-ECM) modulus at 1 week after applying the destabilization of the medial meniscus (DMM) surgery to 3-month-old male wild-type mice, relative to the Sham control. (Adapted with permission from Ref. [6]). (c) Tapping mode AFM height images of aggrecan ultrastructure synthesized by adult equine bone marrow stromal cells (BMSCs) and chondrocytes. (Adapted with permission from Ref. [33])

Full size image

Despite being part of the natural homeostatic process, aggrecan fragmentation is more aggravated during the initiation of OA. Inspired by the molecular-level studies by Grodzinsky, Chery et al. investigated how aggrecan degeneration alters the micromechanics of pericellular matrix (PCM), the immediate microenvironment of chondrocytes, and in turn, the mechanotransduction of chondrocytes in post-traumatic OA. In the destabilization of the medial meniscus (DMM) murine model [18], the micromodulus of PCM was measured by immunofluorescence (IF)-guided AFM nanomechanical mapping (Fig. 5.4b), and was found to show significant reduction as early as 3 days post-surgery relative to the Sham control. This reduction preceded both changes of overt tissue-level mechanical properties measured by classical AFM-nanoindentation (1 week after) [13], and appearance of histological cartilage damage (4–8 weeks after) [18]. This weakening of the PCM can be attributed to accelerated aggrecan degradation in OA, as the aggrecan neo-epitopes (e.g., VDIPEN) were found to be mainly localized in the pericellular domain at this early stage. In alignment with the PCM degeneration, at 3 days after DMM, chondrocytes also exhibited demoted intracellular calcium signaling, [Ca2+]i, activities, one of the earliest, fundamental cell responses to mechanical stimuli [8]. This effect was most pronounced under hypo-osmotic stimuli, which simulate the amplified GAG-GAG EDL repulsion and increased cell strain during compressive joint loading. Conversely, when the aggravated catabolism was attenuated by small molecule inhibitor, GM6001, the reduction of PCM modulus and disruption of [Ca2+]i activities could be effectively rescued. Thus, aggravated aggrecan degradation represents a key molecular event in the initiation of OA, which not only impacts the tissue-level mechanical properties, but disrupts chondrocyte mechanotransduction by impairing the PCM.

Understanding the molecular aspects of aggrecan in normal aging and disease initiation also shed light on the development of novel tissue engineering and regeneration strategies. In tissue engineering, the use of primary chondrocytes is challenged by the limited amount of cells and donor site morbidity. Bone marrow stromal cells (BMSCs) are often used as the alternative cell source [41]. When undergoing chondrogenesis, BMSCs were found to synthesize full length aggrecan within 1–2 weeks of chondrogenic culture [33, 37]. Adult equine BMSCs undergoing chondrogenesis within hydrogel cultures could synthesize aggrecan molecules having CS-GAG chains that were almost 2× longer than the CS-GAGs synthesized by primary chondrocytes harvested from those same horses (Fig. 5.4c). Importantly, it was also discovered via fluorophore-assisted carbohydrate electrophoresis (FACE) analysis that the aggrecan made by these adult BMSCs demonstrated CS-GAG sulfation patterns typical of those observed in newborn growth cartilage, even though these cells were originated from adult animals. The BMSC-derived aggrecan also showed higher compressive stiffness, close to that of newborn human aggrecan as seen in Fig. 5.4a. On the other hand, in comparison to primary chondrocytes, BMSCs had a lower synthesis rate of collagen and proteoglycans, as well as a lower retention rate of newly synthesized aggrecan in its neo-matrix. This resulted in a lesser assembled matrix with lower sGAG content in the BMSC neo-matrix [1]. When assessed via AFM-nanorheometer, BMSC neo-matrix showed a higher degree of energy dissipation and similar elastic modulus at lower frequencies, but lower modulus at high frequency relative to that of chondrocyte neo-matrix [34, 35]. Therefore, despite its capability of synthesizing more superior aggrecan, BMSCs may also have inferior capabilities in biosynthesis and neo-matrix assembly. These factors need to be considered and modulated at both molecular and cellular levels to enhance the quality of regenerative tissues.

Fig. 5.5
A set of images, a to e. A. Set of 4 immunofluorescence images for S a f-O, fast G, and aggrecan. B. 2 graphs for E i n d and k. W T is higher for first, and D c n is higher for second. C. 2 models for pore pressure at the scale 0 through 20 for W T and D c n. D. 2 A F M images with strands for aggrecan plus decor in and aggrecan only. First is denser. E. 3 illustrations for decor in-mediated aggrecan adhesion and structural integrity of the aggrecan network in cartilage E C M with link protein, H A, collagen fibrils, decor in, and aggrecan.

Decorin regulates the integrity of aggrecan and biomechanics of cartilage ECM. a-c) Structural and biomechanical phenotype of Dcn−/− murine cartilage relative to the wild-type (WT) control at 3 months of age. (a) Safranin-O/Fast Green histology and IF images show the reduction of sulfated glycoaminoglycans (sGAGs) and aggrecan in Dcn−/− cartilage. (b) AFM-based nanoindentation and nanorheometric tests show the reduction of indentation modulus, Eind, and the increase of hydraulic permeability, k, of cartilage. (c) Maximum pore pressure calculated from the fibril-reinforced poroelastic finite element model at the peak frequency (∼10 Hz) corresponding to maximum phase angle. (d) Tapping mode AFM height imaging shows the formation of interconnected supramolecular network when aggrecan is reconstituted with free decorin protein, and individual aggrecan monomers when reconstituted without. (e) Schematic illustration of the structural role of decorin in regulating the molecular adhesion of aggrecan-aggrecan and aggrecan-collagen fibrils, and thus, the integrity of cartilage ECM. (Panels (a)–(e) are adapted with permission from Ref. [21])

Full size image

4 Other Native and Biomimetic Proteoglycans

In addition to aggrecan, cartilage matrix also consists of many other proteoglycans and glycoproteins, including small leucine rich proteoglycans (SLRPs), perlecan, lubricin, matrilins and cartilage oligomeric matrix protein (COMP) [28]. These molecules are present at minor quantities, and thus, do not directly contribute to tissue biomechanics. However, they could have important roles in regulating matrix assembly or cell-matrix interactions through specific interactions with other matrix molecules, cell surface receptors, and/or cytokines [32]. Regulatory roles of individual proteoglycans have been studied by assessing the phenotype of various genetic knockout murine models [28]. Previously, analysis of murine cartilage phenotype has been mainly limited to gross-level assays, such as bulk chemistry, histology, immunohistochemistry, and micro-computed tomography (μCT). Assessment of the functional relevance of these molecules was challenged by the small volume and irregular shape of murine cartilage, which renders conventional biomechanical tools not applicable.

The nanomechanical paradigm established by Grodzinsky and colleagues enabled direct quantification of murine cartilage biomechanical properties, providing a new path for pinpointing the activities of individual proteoglycans [22]. For example, a recent study by Han et al. investigated the role of decorin in cartilage biomechanical function and OA progression [21]. Decorin is a class I SLRP containing ~40 kDa leucine rich core protein with one CS- or dermatan sulfate (DS)-GAG chain attached near its N-terminal. In human cartilage, the concentration of decorin is ≈ 15 nmol/ml, comparable to that of aggrecan (≈ 20 nmol/ml) [47], which implies its potential importance to cartilage integrity. In both decorin-null (Dcn−/−) and inducible decorin knockout mice (Dcnf/f/Rosa26CreER, or DcniKO), loss of decorin resulted in reduced aggrecan and sGAG content in the ECM (Fig. 5.5a). When tested under the AFM-nanorheometer, decorin-deficient cartilage demonstrated compromised biomechanical properties, including lower modulus, higher hydraulic permeability and reduced fluid pressurization (Fig. 5.5b, c). These observations highlighted a crucial role of decorin in regulating the integrity of aggrecan in cartilage ECM. This hypothesis was supported by molecular-level nanomechanical experiments. First, when free decorin protein was added to the solution, molecular adhesions between two opposing aggrecan layers, and between aggrecan and collagen II fibrils, were both significantly increased. Second, when decorin protein and aggrecan monomers were reconstituted on a mica surface, they formed interconnected supramolecular networks, despite the presence of strong EDL repulsion (Fig. 5.5d). These results corroborate the observation that Dcn−/− chondrocytes synthesized a similar amount of sGAGs, but a lesser portion was retained in the neo-matrix. Therefore, in cartilage, decorin could serve as a “physical linker”, which in turn, strengthens the aggrecan-aggrecan and aggrecan-collagen II molecular adhesion, enhancing the integration of aggrecan network in cartilage (Fig. 5.5e) [21].

The impact of decorin on aggrecan integrity also regulates chondrocyte mechanotransduction. Applying immunofluorescence (IF)-guided AFM, Chery et al. showed that the PCM of Dcn−/− cartilage was impaired during post-natal growth, leading to demoted chondrocyte intracellular calcium signaling, [Ca2+]i, activities in situ [7]. This study further confirmed that such impairment can be attributed to the reduction of aggrecan and sGAG content in the PCM, supporting the role of decorin in mediating chondrocyte mechanobiology through regulating the integrity of aggrecan in the PCM. In the DMM model, both Dcn−/− and DcniKO mice exhibited accelerated loss of sGAGs and fibrillation of cartilage surface, contributing to more severe OA relative to the control [38]. The mediation of aggrecan assembly was also found to be specific to decorin. Biglycan is another class I SLRP, whose core protein has ~57% structural homology to that of decorin, but harbors two, rather than one, CS/DS-GAG side chains near its N-terminal [30]. In contrast, such aggravated OA was not detected in biglycan inducible knockout mice (Bgnf/f/Rosa26CreER) subjected to DMM surgery [20]. Therefore, building on the foundation established by Grodzinsky, these recent studies highlighted an indispensable role of decorin in regulating the integrity of aggrecan network in cartilage matrix, and thus, the ECM biomechanics and chondrocyte mechanotransduction. Meanwhile, decorin also contributes to the slowdown of OA progression by attenuating the loss of fragmented aggrecan and inhibiting cartilage fibrillation.

Besides decorin, the impact of perlecan on cartilage development and homeostasis has also been studied from the nanomechanics perspective. Perlecan is a basement membrane-specific heparan sulfate proteoglycan (HSPG, Mw ~ 470 kDa), and contains three heparan sulfate (HS)-GAG or CS-GAG chains near its N-terminal. In cartilage, perlecan is localized in the PCM, and is suggested to interact with collagens VI and XI to stabilize the matrix compartment [62]. It also directly regulates cell surface mechanosensing [19] and activation of fibroblast growth factor-2 (FGF-2) [57]. Applying IF-guided AFM, Wilusz et al. demonstrated direct contribution of perlecan and its HS-GAGs to PCM integrity. Heparinase III digestion was shown to increase the micromodulus of porcine cartilage PCM, but not that of the bulk ECM [61]. It was hypothesized that the HS-GAG chain of perlecan could contribute to the local fixed charges and osmotic swelling pressure, while its enzymatic removal may reduce the swelling of PCM and in turn, increase the apparent local modulus. Furthermore, in newborn perlecan knockdown mice (Hspg2+/−), Xu et al. observed reduced cartilage matrix stiffness as well as defective PCM formation. Production of an abundance of matrix proteins was elevated, including atypical sGAGs, which was hypothesized to compensate for the loss of perlecan, illustrating an important role of perlecan in mediating initial matrix assembly [63].

Marcolongo and colleagues synthesized a family of biomimetic proteoglycans (BPGs) to recapitulate the biophysical characteristics of native aggrecan at the molecular level [48]. Specifically, BPG10 is a synthetic polymer consisting of a ~ 10 kDa synthetic poly(acrylic acid) (PAA) core, decorated with ~5–7 CS-GAG bristles (Fig. 5.6a). Similar to that of aggrecan, BPG10 exhibits the “bottle-brush” architecture, with CS-GAGs packed at 3–4 nm spacing along the PAA core [48], comparable to the 2–3 nm spacing of CS-GAGs along aggrecan core protein [43]. When infiltrated into bovine cartilage explants, these BPGs were able to localize in the PCM and territorial domain [46] (Fig. 5.6b). Kahle et al. applied IF-guided AFM to BPG10-augmented bovine cartilage explant, and showed that its localization increased the micromodulus of PCM without altering properties of the matrix bulk (Fig. 5.6c) [31]. Such effect was attributed to the increased fixed charge density within the PCM due to the localization of BPG10. In turn, residing chondrocytes in BPG10-augmented PCM exhibited enhanced in situ [Ca2+]i activities (Fig. 5.6d). When tested by molecular force spectroscopy, these BPG10 molecules demonstrated the capability of undergoing molecular adhesion with other BPG10 molecules and with native aggrecan, at a similar adhesion magnitude as aggrecan-aggrecan self-adhesion. Thus, it was hypothesized that by mimicking the “brush-like” ultrastructure and polyanionic nature of aggrecan, BPG10 can integrate with aggrecan in native cartilage through biophysical adhesions (Fig. 5.6e), and thus, has the potential to be used for harnessing cell mechanoresponses and modifying disease progression [31].

Fig. 5.6
A set of images, a to e. A. Chemical structure of B P G 10 made of P A A core and C S G A G bristle. B. 3 immunofluorescent images at the scale of 10 microns. C. 2 images for control and B P G 10 on the scale of 0 to 0.3 of E i n d. D. 2 I F images at the scale of 20 micron.

Biomimetic proteoglycan, BPG10, strengthens cartilage pericellular matrix (PCM) and modulates chondrocyte mechanotransduction through integrating with the aggrecan network in the PCM. (a) BPG10 is synthesized by grafting natural chondroitin sulfate glycosaminoglycan (CS-GAG) bristles to an enzymatically resistant, synthetic poly (acrylic acid) (PAA) core. (b) IF images of adult bovine cartilage explants infiltrated with fluorescently-labeled BPG10 and co-stained with collagen VI demonstrate the preferred distribution of BPG10 within the PCM and nearby territorial domain. (c) Representative indentation modulus, Eind, maps of control and BPG10-treated cartilage in 20 × 20 μm2 regions of interest (ROIs) containing well-defined PCM rings (40 × 40 indents) via IF-guided AFM nanomechanical mapping illustrate the increase of PCM micromodulus by the infiltration of BPG10. (d) Left panel: Representative IF images of intracellular calcium signaling, [Ca2+]i, of adult bovine chondrocytes in situ. BPG10 enhances mechanosensing of chondrocytes in both isotonic and hypotonic (osmotically-simulated compression) conditions, as illustrated by an increase in the percentage of responding cells, %Rcell (mean ± 95% CI, ≥ 445 cells from n = 3 animals). (e) Schematic illustration of biophysical adhesion interactions between BPG10 and aggrecan, which enables the integration of BPG10 with the aggrecan-enriched cartilage PCM, and thus, the preferred localization of BPG10 in the PCM. (Panels (a)–(e) are adapted with permission from Ref. [31])

Full size image

5 Summary and Outlook

This chapter summarized the transformative impact of Dr. Grodzinsky’s contributions to the understanding of aggrecan molecular mechanics at the nanoscale. By developing and applying an array of AFM-based nanomechanical modalities to cartilage molecules, cells, and tissues, this body of work established a new front in understanding the origins of cartilage ECM functions, cell-ECM interactions, and disease initiation events. In addition, as discussed in this chapter, this nanotechnology paradigm established by Dr. Grodzinsky opened the door for further in-depth studies on the roles of other minor proteoglycans and proteins in cartilage biomechanics and mechanobiology, as well as the evaluation of novel molecular therapeutic strategies for OA treatment. It is expected that many future studies will benefit immensely from this molecular foundation established by Dr. Grodzinsky, which is one of the many fronts that he has contributed in musculoskeletal research.