Interactions between Lubricin and Hyaluronic Acid Synergistically

Subscriber access provided by UNIV OF LOUISIANA

Biological and Medical Applications of Materials and Interfaces

The Interactions between Lubricin and Hyaluronic Acid Synergistically Enhance Anti-Adhesive Properties Huijun Ye, Mingyu Han, Renliang Huang, Tannin A Schmidt, Wei Qi, Zhimin He, Lisandra L. Martin, Gregory D. Jay, Rongxin Su, and George W. Greene ACS Appl. Mater. Interfaces, Just Accepted Manuscript • Publication Date (Web): 26 Apr 2019 Downloaded from http://pubs.acs.org on April 26, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

The Interactions between Lubricin and Hyaluronic Acid Synergistically Enhance Anti-Adhesive Properties Huijun Ye,† Mingyu Han,‡ Renliang Huang,§ Tannin A. Schmidt,# Wei Qi,† Zhimin He,† Lisandra L. Martin,Δ Gregory D. Jay,|| Rongxin Su,†,* George W. Greene‡,* †State

Key Laboratory of Chemical Engineering, Tianjin Key Laboratory of Membrane Science

and Desalination Technology, School of Chemical Engineering and Technology, Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), Tianjin University, 92 Weijin Road, Nankai District, Tianjin 300072, China. ‡Institute

of Frontier Materials, Australian Centre of Excellence in Electromaterials Science,

Deakin University, 75 Pigdons Road, Waurn Ponds VIC 3216, Australia §School

of Environmental Science and Engineering, Tianjin University, 92 Weijin Road, Nankai

District, Tianjin 300072, China #Biomedical

Engineering Department, University of Connecticut, 263 Farmington Avenue,

Farmington, CT 06030, USA ||Department

of Emergency Medicine, Warren Alpert Medical School, Division of Biomedical

Engineering, School of Engineering, Brown University, Providence, RI 02912, USA ΔSchool

of Chemistry, Monash University, Wellington Rd, Clayton, VIC 3800, Australia

. * Author to whom any correspondence should be addressed: E-mail: [email protected] (R. S.), [email protected] (G. W. G.)

ACS Paragon Plus Environment

1

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 43

ABSTRACT Preventing the unwanted adsorption of proteins and cells at articular cartilage surfaces plays a critical role in maintaining healthy joints and avoiding degenerative diseases such as osteoarthritis. Immobilized at the surface of healthy articular cartilage is a thin, interfacial layer of macromolecules consisting mainly of hyaluronic acid (HA) and lubricin (LUB; a.k.a. PRG4) which is believed to form a co-adsorbed, composite film now known to exhibit synergistic tribological properties. Bioinspired by the composition of cartilage surfaces, composite layers of HA and LUB were grafted to Au surfaces and the anti-adhesive properties were assessed using surface plasmon resonance and quartz crystal microbalance. A clear synergistic enhancement in anti-adhesive properties was observed in the composite films relative to grafted HA and LUB layers alone. AFM normal force measurements provide insight into the architecture of the HA/LUB composite layer and implicate a strong contribution of hydrophobic interactions in the binding of LUB end-domains directly to HA chains. These AFM force measurements indicate that the adhesion of LUB to HA is strong and indicate that the hydrophobic coupling of LUB to HA shields the hydrophobic domains in these molecules from interactions with other proteins or molecules. KEYWORDS: synergy, anti-adhesive, hyaluronic acid, lubricin, PRG4, coating


2

Page 3 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


 INTRODUCTION Osteoarthritis (OA) is a complicated, degenerative disease of the joint that affects hundreds of millions of people worldwide. The factors triggering the early onset of OA disease remains unclear and ambiguous; however, numerous genetic,1 mechanical,2 and inflammatory factors3 are known to contribute to the development of the disease. One recognized factor in OA initiation and progression are inflammatory mediators, including cytokines,4 chemokine,5 and prostaglandins,6 which are released locally from joint cells (e.g., chondrocytes, synoviocytes) and/or other tissues (e.g., adipose tissue) following injury. These mediators activate chondrocytes for metalloproteinase synthesis which accelerates the cartilage degradation and can also induce synovial angiogenesis and the increase in inflammatory cytokine and metalloproteinase synthesis within synovial tissues. The potential of these abnormal components in synovial fluid to induce damage and degradation of cartilage tissue inspired us to study the anti-fouling role of hyaluronic acid and lubricin, two major macromolecule components of the cartilage surface.

Lubricin (LUB), encoded by the proteoglycan 4 (PRG4) gene, is a special type of mucin-like glycoprotein with an extended mucinous domain flanked at either end by globular end-domains at the N- and C-terminus.7-8 These two globular domains, which possess subdomains similar to somatomedin-B (SMB) in the N-terminus domain and hemopexin (HPX) in the C-terminus domain are very adhesive and have a strong affinity to a wide variety of surfaces. LUB selfassembles to form a ‘telechelic’ polymer brush layer7 by adsorbing to surfaces via these enddomains. The mucin-like domain is highly hydrated which leads to excellent lubrication8 and antiadhesive properties that resist protein binding9-11 and cells/bacteria adhesion12-13. Additionally, the absence of LUB in synovial joints has been shown to result in abnormal protein deposits and cell


3


Page 4 of 43

overgrowths on cartilage surfaces of lubricin-null mice (e.g., Prg4-/-mice) as the mice aged14 indicating an important role in controlling unwanted adhesion of proteins and cells to the cartilage surface.

Besides LUB, hyaluronic acid (HA), an anionic glycosaminoglycan, is also abundant in synovial fluid and has long been considered a boundary lubricant.15 It has been discovered that only HA strongly immobilized at the cartilage surface provides effective lubrication as opposed to the ‘free’ HA within the synovial fluids.16-19 More importantly, HA contains a large number of amide (CO−NH) and carboxyl (COOH) groups, which provide high surface hydration strength that contribute to the prevention of protein adsorption and microbial adhesion.20-24

With respect to joint protection and health, it is becoming increasingly apparent that HA and LUB can no longer be treated as separate entities. Significant evidence now suggests that LUB interacts with HA chains to form a macromolecular complex both within synovial fluids25 and at cartilage surfaces26. However, the strength and nature of this interaction remains unknown. As a complex, the HA/LUB system leads to synergistic lubrication properties, which reduce friction coefficients and enhanced wear resistance than either component alone on both cartilage27 and model surfaces28. Since healthy joints must remain both undamaged and ‘contamination’ free, this study investigates whether HA/LUB also function synergistically in controlling surface adhesion, which might implicate an expanded role of the LUB/HA complex on protecting joint surfaces.

Surface fouling by the nonspecific adsorption of proteins and other large biomacromolecules is a persistent and often damaging problem in numerous fields including medical implants,


4

Page 5 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


microfluidics, biosensors, and medical diagnostics,

Approaches to controlling non-specific

adsorption at surfaces generally consist of the grafting, often covalently, anti-adhesive macromolecules to the surfaces. Over the decades, numerous anti-adhesive coating technologies have been developed which generally involve the grafting of hydrophilic, charge neutral polymers (e.g. polyethylene glycol (PEG),29 polyacrylamide30), synthetic and natural polysaccharides (e.g. glycopolymers31, dextran32), anti-adhesive peptides and peptoids (e.g. poly(β-peptoid)s,33 zwitterionic poly(sulfobetaine methacrylate)34 and poly(carboxybetaine methacrylate)35. Many of these coatings involve harsh synthetic processes, toxic monomers, and/or complex coupling chemistries to graft coating molecules to surfces. Since HA and LUB are biopolymers distributed widely throughout the body, coatings of these materials are innately biocompatible. This high biocompatibility, combined with a coating process that achieves composite films via self-assembly, provides a novel strategy for controlling adhesion in fields like microfluidics,10, 36 biosensors,37 medical devices,38 and surgical implants such as bionic electrode interfaces.11

To mimic the structure of HA/LUB composite at the surface of cartilage26-28 , thiolated HA was chemically grafted to Au surfaces followed by the physical adsorption LUB to the chemically immobilized HA layer. The non-specific adsorption from solutions of bovine serum albumin (BSA) and undiluted human blood serum was evaluated on Au surfaces modified with HA/LUB composite layers and compared against Au surfaces modified with either single layers of chemically grafted HA or physically grafted LUB using complementary techniques of surface plasmon resonance (SPR) and quartz crystal microbalance (QCM). For these anti-adhesive studies, native LUB (i.e. LUB) purified from bovine synovial fluid was used to better approximate the composition of the cartilage surface. The method used for the purification of LUB (i.e. affinity


5


Page 6 of 43

chromatography) isolates not only LUB, but also a number of non-LUB molecules and so is only able achieve LUB purities of around 90%. Although the specific impurities are not known or quantified, the majority is believed to be degraded glycosylated protein ‘debris’ with a low surface binding affinity; however, additional impurities consist of a number of identified LUB binding ligands39-41 found associated with LUB in synovial fluid and at the surface of cartilage and so are also expected to be present in the purified LUB. A previous study comparing the anti-adhesive performance of native LUB and recombinant LUB (i.e. R-LUB) coatings on electrochemical sensors indicate that the presence of these impurities can have a negative influence on native LUBs protein resistance.37 Since these LUB binding ligands are present in mammalian joints, native LUB was used in the HALUB anti-adhesive synergy studies as it better represents the state of LUB within joints.

AFM normal force measurements (i.e. the interaction forces acting perpendicular to the surface being probed) are utilize extensively in the investigation of protein-protein interactions,42-44 intraprotein interactions and secondary structure,45-46 and protein-surface interactions.47-49 AFM interaction force studies typically rely on the analysis of the adhesion forces measured during separation to obtain adhesion energies (via integration of the force measurements) and analyzing the interaction energy landscape through the specific peaks in the measured force-distance curves.50 To probe the interaction between HA and LUB, AFM normal force measurements using both hydrophilic and hydrophobic AFM probe tips were conducted to assess the adhesive forces between HA, LUB, and HA/LUB to the probe surface chemistries. Because the presence of impurities co-purified with native LUB may obscure the subtle interaction forces (particularly


6

Page 7 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


between HA and LUB), these AFM normal force measurements were conducted with highly pure (recombinant) R-LUB.

RESULTS AND DISCUSSION Characterization of HA, LUB and HA/LUB coatings. This work seeks to achieve bioinspired anti-adhesive synergy via the modification of surfaces with a chemically grafted HA/LUB composite that mimics, in many respects, the surface of articular cartilage (see Scheme 1).

Scheme 1. Schematic illustration of (A) the joint cartilage surface, (B) the molecular structure of the disaccharide in HA, (C) the structure of LUB, and (D) HA/LUB modified Au (Au-HA/LUB) surfaces.

The assembly and adsorbed/grafted masses of HA, LUB and HA/LUB coatings onto Au surfaces (see “Fabrication of Au-HA, Au-LUB and Au-HA/LUB surfaces”) were monitored in situ by SPR (Figure 1A-C) as well as QCM (Figure 1D-F). Figure 1A shows the SPR signal response during the assembly of HA surfaces, achieved in stages, through the sequential injection of 1 mg/mL solutions of thiolated-HA and rinse in PBS. The SPR signal response over a single injection-rinse cycle can be divided into three distinct stages. Upon initial injection of the thiolated-HA solution,


7


Page 8 of 43

the SPR signal drops sharply due to the lower refractive index of the HA solution compared to the PBS. As HA binds to the Au, the surface refractive index increases causing the SPR signal to rise. Upon rinsing with PBS, the rising signal plateaus to a steady value. The difference between the initial and final values of the SPR signal (i.e. before adsorption and after rinsing) is proportional to the mass of HA adsorbed to the surface. With each additional injection-rinse cycle, the increase in HA adsorption was found to gradually diminish until no significant difference between cycles was observed indicating surface saturation. This saturation is reflected in the signal of the final injection-rinse cycle where the signal was found to plateau prior to the PBS rinse. Figure 1B shows the SPR signal response during the self-assembly of LUB onto the Au surface which, again, was performed using the sequential injection of 100 µg/mL solution of LUB followed by a PBS rinse. The initial, sharp increase in the signal upon the injection of LUB solution is due to rapid adsorption of LUB on Au surface with the majority of LUB adsorption occurring over the first few minutes. The SPR signal continued to rise, albeit more slowly, even after the flow of the LUB solution was halted and increased again when the flow was restarted. The LUB injection-rinse cycle was continued until the SPR signal again reached a plateau, indicating LUB saturation of the Au surface. The assembly of the HA/LUB composite in the SPR is shown in Figure 1C and achieved by first adsorbing a layer of thiolated-HA followed by the adsorption of LUB sequentially. In both the HA and LUB adsorption stages, the SPR signal responses were similar to those observed in Figure 1A and 1B; although, the mass of LUB adsorbing to HA was approximately 1/2 of that directly adsorbing to Au. Figure S1A shows the total change in the average SPR signal observed during the assembly of the HA, LUB, and HA/LUB surfaces over 3 separate experiments.


8

Page 9 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In a complimentary study, QCM measurements were performed on the same HA, LUB, and HA/LUB systems previously utilized in the SPR experiments. As shown in Figure 1D, a representative frequency (ΔF) and dissipation (ΔD) response was observed in the QCM measurement during the assembly of a HA layer on the Au surface. After exposing the Au surface to a 1 mg/ml solution of thiolated-HA in PBS, a rapid decrease in frequency and increase in dissipation was observed that decays roughly exponentially in time consistent with the adsorption and assembly of an interfacial HA layer. After approximately 30 min, an apparent stable frequency was achieved after which the flow of additional HA solution produced no further decrease in ΔF. At this point, the surface was rinsed using a continuous flow of PBS. Typically, during the rinse of a surface following adsorption, an increase in ΔF is observed as weakly adsorbed and entangled (but unabsorbed) molecules are removed reducing its mass. Unusually, during the PBS rinsing stage following HA adsorption, a gradual but rather large (~ -10 Hz) decrease in ΔF was observed over a period of approximately 30 minutes. Superficially, a decrease in ΔF would suggest that the mass of the HA layer was increasing; however, viscoelastic dampening, ΔD, also decreased significantly over this same period of time indicating that the HA is becoming mechanically stiffer. Since the ΔF in QCM can underestimate the true mass of ‘soft,’ viscoelastic films, the observed decrease in ΔF combined with a concomitant decrease in ΔD should be interpreted as a conformational rearrangement of the HA molecules within the layer into a mechanically stiffer and denser film; most likely via the attachment of previously ‘free’ thiol groups within the HA chains to the Au surface. Figure 1E shows the frequency and dissipation responses during LUB adsorption, which is consistent with previous QCM LUB adsorption studies.9 A representative frequency and dissipation response during assembly of the HA/LUB composite layer, which was consistent across 4 separate measurements, is shown in Figure 1F. It can be seen that during the


9


Page 10 of 43

initial assembly of the HA layer, the frequency and dissipation response observed is similar to that shown in Figure 1D. During the LUB adsorption stage in Figure 1F, a large decrease in ΔF was observed indicating the binding of LUB molecules to the HA chains and the formation of a HA/LUB composite. As also seen in SPR (Figure 1C), the LUB adsorbing to HA was roughly ½ that adsorbing to Au. The ΔF caused by the self-assembled HA, LUB, and HA/LUB composite layer from 4 separate experiments is shown in Figure S1B, which exhibits similar trends as SPR experiments.


10

Page 11 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1. Representative examples of (A-C) the SPR sensorgrams showing the change in SPR angle, Δθ, over time and (D-F) the QCM sensorgrams showing the change in frquency, ΔF, and dissipation, ΔD, over time during the adsorption of HA (A and D) , LUB (B and E) and HA/LUB (C and F) on Au substrates. In both SPR and QCM experiments, 1mg/ml solutions of thiolated HA in PBS and 100µg/ml LUB in PBS were utilized for HA and LUB adsorption stages, respectively. Details of the respective adsorption experiments are provided on the Materials and Methods under the heading “Fabrication of HA, LUB, and HA/LUB coated Au surfaces.

To characterize the wettability of modified Au surfaces, the static water contact angles of bare Au, and HA, LUB, and HA/LUB coatings on Au were measured and summarized in Table S1. The bare Au was found to be slightly hydrophilic with a water contact angle of 73.2º. After adsorption of HA, the contact angle decreased notably to 13.1º, demonstrating the high hydrophilicity of HA as has also been reported elsewhere.20 The adsorption of LUB on Au also decreased the contact angle to 40.6º, which was notably less hydrophilic than the HA surface. Interestingly, after modifying Au with HA and LUB stepwise, the contact angle was reduced to 22.9º, whose value is intermediate of the HA and LUB coated Au surfaces.


11


Page 12 of 43

Figure 2. Flattened topographic AFM images of Au SPR sensor surfaces coated with adsorbed layers of (A) HA (B) LUB, and (C) HA/LUB which were measured in PBS solution using a silicon nitride cantilever with a normal spring constant of 0.7 N/m (Bruker), Inserts in (A-C) show higher magnification image of each surface.

Scale bar in inserts correspond to 300 nm.

(D)

Representative height traces of the surface topographies of each modified surface corresponding with the respective white horizontal line displayed in figures (A-C). (E) The root-mean-square (RMS) roughness values calculated from a total of traces collected from the AFM images shown in (A-C), The error bars in (E) depict standard deviations of the mean of three measurements in three different regions/images of the surface . Surface Morphologies of the HA, LUB, and HA/LUB composite. The surface morphologies of HA, LUB, and HA/LUB composite layers were further investigated by an atomic force microscopy (AFM) in PBS. As shown in Figure 2A and 2E, the HA surface in the wet state has a relative smooth surface with a characteristic root-mean-squared roughness of 0.79 nm, which indicated conformal coverage. In Figure 2B, the texture of the LUB modified surface was found to be


12

Page 13 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


similarly ‘smooth’, yielding only a marginally higher root-mean-squared roughness of 1.09 nm (Figure 2E), consistent with previous reports and attributed to the polymer brush-like structure of LUB on Au47 and other substrates (e.g. SiO2,47 mica,7,

51

graphite,52 etc.). For the HA/LUB

composite modified surface (shown in Figure 2C), the root-mean-squared roughness is 2.06 nm due to the presence of larger, randomly distributed asperities. For comparison, an AFM image of the surface topography of an uncoated Au SPR sensor surface is provided in Figure S2 of the Supplemental materials which shows an RMS roughness of 1.11 ± 0.16 nm. The authors would like to note that the uncoated SPR sensor surface shown in Figure S2 was subjected to 5 RCA cleaning cycles while the coatings shown in Figures 2A-C were deposited on brand new Au surfaces that had only been subjected to 1 RCA cleaning cycle. The additional cleaning cycles may or may not have an impact upon the measured RMS roughness value reported for the uncoated Au surface. This observed increase in surface roughness, particularly given the relative ‘smoothness’ of the HA and LUB modified surfaces (see Figure 2C and 2D), suggests that the adsorbing LUB fails to self-assemble into a polymer brush-like layer on top of the HA. The texture and root-mean-squared roughness of the HA/LUB layer is similar to that observed with non-brush forming pig gastric mucin layers on Au.9 The failure of LUB to self-assemble is further supported by the fact that the mass of LUB adsorbing onto the HA is less than half of that observed on Au (see Figure S1).

Nonspecific protein adsorption on HA, LUB and HA/LUB surfaces. The anti-adhesive properties of the HA/LUB composite was assessed and compared with that of LUB or HA layers upon exposure to solutions of BSA and undiluted human blood serum. SPR was chosen as the primary technique for quantifying the protein adsorption because of its high mass sensitivity and


13


Page 14 of 43

relative insensitivity to the mechanical/viscoelastic properties and strongly bound hydration layers around molecules which is a source of error in QCM quantitative mass measurements.53

Figure 3. SPR sensorgrams showing the nonspecific adsorption mass on Au sensor surfaces coated with HA, LUB, or HA/LUB from (A) BSA solution (2mg/mL-1) and (B) undiluted blood serum. (C) Quantified non-specific adsorbed mass per unit area measured after exposure to BSA solutions and undiluted serum on bare Au, and Au surfaces coated with HA, LUB and HA/LUB. The error bars presented in (C) are standard deviations of the averaged mass per unit area from 3 individual measurements of each surfaces. (D) Representative plots of the change in frequency (ΔF) measured by QCM for the adsorption of BSA (2mg mL-1) on Au sensor surfaces coated with adsorbed layers of HA, LUB and HA/LUB.


14

Page 15 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3A shows the real-time signals (denote as the change in SPR angle, ∆θ) of BSA adsorbed on HA, LUB and HA/LUB composite modified surfaces. Compared with the signal caused by BSA on Au surfaces shown in Figure S3A (Supplementary Materials), all three modified surfaces tested exhibited a significantly lower response, indicating they all have excellent anti-adhesive properties against protein adsorption. However, the change in the SPR angle due to BSA adsorption on HA/LUB composite surfaces indicates a significantly smaller angle shift of just Δθ = 0.005º compared with the HA (Δθ = 0.031º) and LUB (Δθ = 0.021º) modified surfaces, respectively. Although much larger than what was observed with BSA, a similar trend in the angle shift was also observed when the HA/LUB, HA, LUB modified surfaces and Au surfaces were exposed to undiluted human blood serum (see Figure 3B and S3B, Supplementary Materials). Exposure of the HA/LUB surface to blood serum produced an angle shift of Δθ = 0.048º compared to HA (Δθ = 0.120º) and LUB (Δθ = 0.098º) surfaces, again indicating a significantly lower mass of non-specifically adsorbed material.

The SPR angle shifts for BSA and blood serum (shown in Figure 3A and 3B respectively) have been converted to adsorbed masses and plotted, together, in Figure 3C. In these SPR measurements, an angle shift of 0.1 º is equal to 59.24 ng/cm2 of adsorbed protein.22 The total mass of BSA adsorbing on unmodified Au, and Au surfaces modified with HA, LUB and HA/LUB composite layers were found to be 144.55 ng/cm2, 16.68 ng/cm2, 12.62 ng/cm2 and 2.57 ng/cm2 respectively. These results indicate that the Au surface coated with the HA/LUB composite layer exhibited a nonspecific adsorption of BSA at least four times lower than either the HA or the LUB coating alone; falling within the accepted criterion of ‘ultralow’ adhesion (i.e. 100 nm from the hard-wall repulsion was observed indicating hydrophobic HA-tip interactions. Similar hydrophobic interactions with HA chains have been reported and are attributed to the presence of the methine groups on the polymer’s N-acetylglucosamine residues.5556

Measurements performed on R-LUB coated surfaces with hydrophilic (Figure 4B) and

hydrophobic (Figure 4B’) tips show a longer range steric repulsion on approach compared to HA coated surfaces that extends 50-80 nm from the hard-wall repulsion. On separation, both the hydrophilic and hydrophobic tips against R-LUB coated surfaces produce a long range adhesive force extending 200-350 nm in most measurements with a few outliers on the low and high end of this range in both systems. In previous work47, this adhesive interaction was attributed to adhesion between the tip and the LUB end-domains which are accessed by the ‘sharp tip’ probe penetration of the mucin domain ‘brush’. The long range of the interaction is attributed to the unraveling and stretching of these R-LUB end-domains. No such adhesive interactions are reported in similar


18

Page 19 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


surface forces apparatus measurements using bare mica surfaces against LUB coated mica surfaces where this mucin domain region cannot be penetrated.7, 51 Finally, in measurements performed on HA/R-LUB coated surfaces with hydrophilic (Figure 4C) and hydrophobic (Figure 4C’) tips, a similar long range steric repulsive force was measured on approach as observed in the R-LUB coated surface extending out 50-80 nm from the hard-wall repulsion. Again, upon separation, the HA/R-LUB coated surfaces produce long range adhesive forces similar in range to those measured with the R-LUB coated surfaces. Comparing the hydrophobic tip measurements against HA coated surfaces (Figure 4A’) and HA/R-LUB coated surfaces (Figure 4C’), noticeably absent from the majority (i.e. 87 out of 95 total force measurements) of the HA/R-LUB force curves is the presence of the deep but short range adhesive interaction peak observed in the HA coated surfaces. The absence of this short range peak suggests that the binding of R-LUB to the HA prevents strong hydrophobic interactions from being formed between the hydrophobic tips and surface grafted HA chains. The adhesive forces observed with hydrophobic tips against HA/R-LUB coated surfaces are therefore dominated by tip-LUB hydrophobic interactions most likely formed through the RLUB end-domains.

Figure 5 provides a summary of the adhesive interactions measured using hydrophilic tips against HA, R-LUB, and HA/R-LUB coated surfaces. Figure 5A summarizes the average peak (i.e. maximum) adhesive force and the adhesion energy obtained via the integration of the separation force curves for the HA, R-LUB, and HA/R-LUB coated surfaces respectively. While no adhesive force or energy was measured with HA coated surfaces, large adhesive forces and adhesive energies with a broad distribution were measured for R-LUB and HA/R-LUB coated surfaces. Although the average adhesive force of the HA/R-LUB coated surface (0.51 ±0.34 nN) is slightly


19


Page 20 of 43

smaller than the R-LUB coated surface (0.88 ±0.40 nN), the average adhesion energy of the HA/RLUB coated surface (9.2 ±4.2 x10-17 J) is slightly larger than the R-LUB coated surface (7.7 ±3.8 x10-17 J). Histograms of the adhesion energy measurements for the R-LUB and HA/R-LUB coated surfaces are shown in Figure 5B and 5C respectively. Both the R-LUB and HA/R-LUB coated surface exhibit equally broad distributions in the measured adhesion energies with the mode of the adhesion energy distribution for both surfaces falling within the 6.0-7.2 x 10-17 J bin. In neither of the hydrophilic tip experiments with R-LUB or HA/R-LUB coated surfaces was any evidence observed suggesting R-LUB detachment from the surface and accumulation on the AFM tip; e.g. a gradual increase in the range of the steric repulsion on approach or a gradual decrease in the magnitude or range of adhesive forces measured on separation as observed and reported in Ref. 47 The absence of adhesive interactions between the hydrophilic tip and HA and the similarity in the adhesive energy distributions of the R-LUB and HA/R-LUB coated surfaces thus reflects the energy required to detach the R-LUB end-domains from the hydrophilic tip rather than from the Au surface or HA chains to which they are bound.

Probability plots of the measured R-LUB, and HA/R-LUB adhesion energies using hydrophilic tips (see Figures S5 and S6, Supplementary Materials) were used to determine the type of distribution that the respective data conform to and was subsequently used to fit the histograms in Fig. 5B and C using the appropriate statistical distribution. The adhesion energy distributions measured with the hydrophilic tips against the R-LUB (see Fig. 5B) and HA/R-LUB (see Fig. 5C) coated surfaces were both found to be well fitted by two-parameter Weibull distributions possessing very similar shape parameters (2.17 for R-LUB and 2.32 for HA/R-LUB) indicating similar populations of energies in the two experiments. Although the distributions have very


20

Page 21 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


similar shapes, the scale parameter of the HA/R-LUB distribution was found to be larger (10.41) compared with the R-LUB distribution (8.75). The larger scale of the HA/R-LUB distribution likely reflects the additional energy associated with the stretching of HA-chains to which the RLUB is attached as well as the slightly increased penetration depth (i.e. approximately equal to the separation distance between the onset of the repulsive force and the hard-wall repulsion in the approach force curves in Figure 4). The effect of both HA chain stretching and increased penetration depth will broaden the distribution and shift the mean to higher values as observed in Fig. 5A. The stretching of HA chains is likely to have the largest impact upon the measured adhesion energy as all R-LUB molecules binding to the AFM tip will also be bound to HA chains. HA chain stretching is thus present in all measurements. Since the AFM tip is able to form adhesive interactions with only the two small end-domain regions of the LUB molecule (~2-6 nm in diameter; see refs.

47, 51),

the slight increase in the penetration depth will impact the adhesion

energy measurements only if/when a larger (average) number of end-domains interact with the AFM tip during the measurement compared with the LUB system. The AFM tip interacting with higher end-domain numbers is unlikely to occur in every measurement. The similarity in the RLUB and HA/R-LUB distribution’s shape parameter supports the notion that the bulk of the adhesion energy measured is derived from the same dominant interactions in both systems; i.e. adhesive interactions between the R-LUB end-domains and the hydrophilic AFM tip.


21


Page 22 of 43

Figure 5. (A) The averaged peak adhesive force and corresponding adhesion energy measured between hydrophilic Silicon nitride AFM tips with Au surfaces coated with either HA, LUB, or HA/R-LUB. (B and C) Histograms of the adhesion energy measured for the Silicon nitride AFM tips interacting with Au surfaces coated with (B), R-LUB, and (C) HA/R-LUB. The adhesion energies in (B and C) were obtained by integrating the adhesion force-distance curves measured during the retraction of the AFM tip from 71 (B) and 111 (C) individual force measurements. The histograms in (B and C) have been fitted to a Weibull distribution determined from probablility plots found in Figures S5 and S6 of the Supplementary Materials.


22

Page 23 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 6 summarizes the adhesive interactions from experiments using the hydrophobic tip against HA, R-LUB and HA/R-LUB coated surfaces. The average adhesion forces and adhesion energies of HA, R-LUB, and HA/R-LUB coated surfaces are shown in Figure 6A. The HA coated surface exhibited the highest average adhesion force (0.95 ±0.36 nN) but lowest average adhesion energy (2.2 ±2.2 x10-17 J) of the three surfaces tested. This disparity between the average adhesive force and energy exhibited by the HA surface reflects the combination of the large magnitude of the adhesive forces but short range over which the adhesive interaction acts. Measurements performed with hydrophobic tips against R-LUB and HA/R-LUB coated surfaces produce similar average adhesive forces (0.45 ±0.19 nN and 0.57 ±0.31 nN, respectively) and energies (2.9 ±1.7x10-17 J and 4.3 ±2.4 x10-17 J, respectively) which were both significantly lower than the hydrophilic tip measurements (see Figure 5A). Histograms showing the distribution of adhesion energy measurements for HA, R-LUB, and HA/R-LUB coated surfaces are shown in Figure 6B, 6C and 6D, respectively. The adhesion energy measured with the hydrophobic tip against the HA coated surface showed a narrow distribution with a very prominent mode within the 1.2-2.4 x10-17 J bin. As observed previously with hydrophilic tips (Figure 5B and C), the adhesion energy distributions measured with the hydrophobic tips against the R-LUB and HA/R-LUB coated surfaces were again very similar with an equally broad distribution and similar mode adhesion force falling within adjacent bin ranges; 1.2-2.4 x10-17 J for R-LUB and 2.4-3.6 x10-17 J for HA/R-LUB coated surfaces, respectively. Again no evidence was observed during the course of the force measurements consistent with the detachment of R-LUB from either the Au or HA chains. The absence of any short range (i.e. >100 nm) peak in the repulsive force upon separation in the force curves of the hydrophobic tip against the HA/R-LUB coated surface (see Figure 4C’) as was observed in the HA coated surface (see Figure 4A’) indicates that the bound R-LUB prevents the formation of


23


Page 24 of 43

strong hydrophobic interactions between the hydrophobic tip surface and HA chains. The adhesive forces and energy measured between the hydrophobic tip and HA/R-LUB coated surface is thus predominantly due to hydrophobic interactions between the tip and R-LUB end-domains. The similarity between the adhesion energy distributions of the R-LUB and HA/R-LUB coated surfaces also supports the notion that HA-tip hydrophobic interactions contribute very little to the overall measured adhesion in the HA/R-LUB surface.

Probability plots of the measured HA, R-LUB, and HA/R-LUB adhesion energies using hydrophobic tips (see Figures S7-S9 Supplementary Materials) were used to determine the type of distribution that the respective data conform to and was subsequently used to fit the histograms in Figure 6B-D using the appropriate statistical distribution. The adhesion energy measured with the hydrophobic tip against the HA coated surface (see Figure 6B) showed a narrow lognormal distribution. As observed previously with hydrophilic tips (Figure 5B and C), the adhesion energy distributions measured with the hydrophobic tips against the R-LUB (see Figure 6C) and HA/RLUB (see Figure 6D) coated surfaces were again very similar and both well fitted to Weibull distributions possessing very similar shape parameters (1.71 for R-LUB and 1.84 for HA/R-LUB) indicating similar populations of energies in the two experiments. As previously observed with the hydrophilic tip measurements, while the distributions have very similar shapes, the scale parameters of the HA/R-LUB distribution is larger (4.86) compared with the R-LUB distribution (3.28) which likely reflects the additional energy associated with the stretching of HA-chains to which the R-LUB is attached as well as the small, increase in penetration depth in the HA/R-LUB system. The similarity in the R-LUB and HA/R-LUB distribution’s shape parameter supports the notion that the bulk of the adhesion energy measured is derived from the same dominant


24

Page 25 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


interactions in the two systems; i.e. hydrophobic interactions between the R-LUB end-domains and the modified AFM tip. The different types of distributions observed between the HA and HA/R-LUB surfaces indicate that the contributions of hydrophobic interactions between HA and the AFM tip in the HA/LUB experiments are negligible.

Figure 6. (A) The averaged peak adhesive force and corresponding adhesion energy measured between CH3-SAM modified AFM tips with Au surfaces coated with either HA, R-LUB, or HA/RLUB. (B-D) Histograms of the adhesion energy measured for the CH3-SAM interacting with Au surfaces coated with (B) HA, (C), LUB, and (D) HA/R-LUB. The adhesion energies in (B-D) were obtained by integrating the adhesion force-distance curves measured during the retraction of the AFM tip from 99 (B), 82 (C) and 95 (D) individual force measurements. The histograms in (B) has been fitted to a lognormal distribution determined from probablility plots found in Figure S7 of the Supplementary Materials. The histograms in (C and D) have been fitted to a Weibull


25


Page 26 of 43

distribution determined from probablility plots found in Figures S8 and S9 of the Supplementary Materials.

Discussion on the binding of LUB to HA chains. There is debate among lubricin and joint tribology researchers about the nature of the interaction between HA and LUB with some arguing that LUB binds strongly (and directly) to HA chains26-28 forming a macromolecular ‘complex’ and others arguing that LUB does not bind directly (or at least, very weakly) to HA and thus HA and LUB function as separate entities, albeit synergistically.57-59 Our SPR and QCM measurements clearly show LUB does bind to HA in significantly high densities. Since thiolated HA was used in these experiments, it could be argued that LUB binding is mediated by disulfide bonds formed between free thiols (i.e. not bonded to Au) in the HA and cysteine located within both LUB enddomains.51 A series of 50 AFM normal force measurements were performed using an unmodified Au coated cantilever tip against the grafted HA coated Au surface and showed no measurable adhesive force (see Figure S4, Supplementary materials) which would be expected if free thiol groups were present in the HA layer in significant numbers. The strong hydrophobic adhesive interactions between the LUB end-domains (see Figure 4B’, 6A and 6C) and the hydrophobic tip, when viewed together with the disappearance of any contribution to the adhesive force by strong (but short ranged) hydrophobic interactions between the tip and HA, strongly implies that hydrophobic interactions between LUB end-domains and HA chains play a significant role in the binding. As the LUB molecules bind to the HA chains, hydrophobic domains within the LUB enddomains and HA chains (predominantly the methyl groups on the N-Acetylglucosamine residues) become shielded and so are no longer able to participate in additional interactions with proteins or other molecules. This hydrophobic mediated coupling between LUB and HA thus contributes to the anti-adhesive synergy described in this report as well as the lubrication synergy described


26

Page 27 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


previously.27-28 AFM adhesion measurements conducted using the hydrophilic tip and HA/LUB coated surfaces (see Figure 5A and 5C) provides a lower limit for the strength of the LUB-HA binding interaction since these force measurements probe the strength of LUB adhesion without any measurable HA-tip adhesion. Since the evidence indicates that the LUB detaches primarily from the hydrophilic tip rather than from the HA, the adhesion energy of the bonds binding LUB to the HA chains is at least 9.2 ±4.2 x 10-17 J in magnitude; i.e. the average adhesion energy measured.

The importance of the HA/LUB anti-adhesive synergy to healthy joint function can be seen in previous tribology experiments performed with genetically modified mice that were LUB deficient at birth, but had their LUB expression restored enzymatically 3 weeks after birth.60-61 The restoration of LUB expression in the LUB deficient mice resulted in an initial restoration of the low coefficient of friction (COF) values to levels comparable to LUB sufficient (i.e. healthy) mice in cartilage friction testing performed after the mice had reached 2 months in age. However, over the course of 1 hr under cyclic loading conditions, the initially low COF of the cartilage surfaces in the restored LUB mice population was found to increase significantly61; something that was not observed in the healthy, LUB sufficient mice. This gradually increasing COF could be understood as an effect of surface fouling by HA in the LUB deficient mice that occurred over the 3 week period preceding the restoration of their LUB expression. The increasing COF, in this case, is symptomatic of the presence of bound, fouling molecules on HA reducing the levels of LUB binding to HA chains and/or significantly weakening the LUB-HA bond strength by reducing the number of hydrophobic sites on the HA. As ex situ HA/LUB tribology experiments have shown, the binding of LUB to grafted HA surfaces dramatically reduces the observed COF values


27


Page 28 of 43

compared to HA alone28 and a reduction or loss of bound LUB would, ipso facto, increase friction. Alternatively, the LUB deficient mice were engineered to exhibit symptoms analogous to CACP disease which, in addition to the absence of LUB, is also associated with increased concentrations of HA in the synovial fluid52 and at the cartilage surface. Upon restoration of LUB expression, there may be an imbalance in the relative abundances of LUB and HA from healthy levels. This imbalance can result in cartilage surfaces enriched in HA and reduced in LUB that are more prone to fouling by other synovial fluid constituents in (e.g. albumin) which could also contribute to the increasing COF values. Reduced LUB concentrations resulting from trauma61 can likewise lead to imbalances in the relative abundance of LUB and HA with consequences to joint fouling resistance and health.

 CONCLUSIONS Inspired by nature, the experiments described in this report clearly demonstrate that LUB anchored to HA chains in a composite-like layer generates a strong and significant synergistic enhancement in the anti-adhesive properties against proteins and blood constituents relative to either HA or LUB acting alone. This anti-adhesive synergy, which complements the recently discovered lubrication synergy, highlights the importance of the HA-LUB interaction in the protection of cartilage surfaces from unwanted adhesion and the preservation of joint health. To form this synergistic composite film, LUB molecules are found to couple directly to HA chains through strong intermolecular interactions that include a significant hydrophobic component. It is postulated that the hydrophobic bonds LUB end-domains form with HA chains shield these domains from further interactions with other biomolecules (or surfaces) giving rise to the observed anti-adhesive synergy from the composite structure and, likely, the observed lubrication synergy26-28 as well. In addition to enhancing the anti-adhesion of joint surfaces, the anti-adhesive synergy exhibited by HA/LUB


28

Page 29 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


composite films promise improved performance and efficacy in a number of non-biological applications for which HA and/or LUB coatings are currently being developed which include microfluidics10, 62, biosensors20, 37, contact lenses63, and biomedical implants64, among others. The addition of LUB also provides a means for improving the fouling resistance of HA hydrogel materials and HA-polyelectrolyte multilayer films.

 MATERIALS AND METHODS Lubricin materials. Native lubricin protein (LUB) was purified using the procedure described in Greene et al9 (a slightly modified method based upon the protocols8) from ~500 ml of bovine synovial fluid sourced from MC Herd (Corio; Victoria, Australia). The synovial fluid was collected percutaneously from the radiocarpal joints of freshly slaughter cattle (~1-year-old; male and female) using a sterile 18-gauge hypodermic needle and stored in a polypropylene bottle, on ice, until the time of processing (approximately 2 hours after collection).

The extracted and purified LUB was analysed for purity using a density gradient SDS-PAGE Biorad gel subsequently stained with coomassie blue. The relative purity of the LUB (as a fraction of the total protein content) was assessed using a Biorad imager and spectroscopic analysis and was found to be approximately 89%. The LUB band appeared on the SDS-PAGE gel at approximately the 280 kDa region, consistent with previous reports.8-9 The concentration of LUB in the extracted solution was determined using the Biorad protein assay, with BSA as the standard. After the concentration of LUB was assayed, the solution was concentrated using a Millipore Amicon Ultra Centrifugal Filter with a 100 NMWL membrane to yield a final concentration of 400 µg/ml of protein in PBS buffer at pH 7.4 with an additional 0.5 mM CaCl2 and 0.2 mM alpha


29


Page 30 of 43

lactose as stabilizing agents. For experiments, this solution was diluted to working concentration of 100 µg/ml with PBS. AFM force and adhesion measurements were performed using full sequence recombinant human lubricin (R-LUB) in PBS obtained from Lubris Biopharma (Weston, MA USA) at a purity of

Interactions between Lubricin and Hyaluronic Acid Synergistically

Recommend Documents